United States                  EPA-600/9-81-Qf?
                Environmental Protection
                Agency                     March 1981 v
<>EPA       Research  and
               Development
                PROCEEDINGS: SYMPOSIUM ON

                IRON AND STEEL POLLUTION ABATEMENT

                TECHNOLOGY FOR 1980

                (Philadelphia, PA, 11/18-11/20/80)
               Prepared for
               Office of Environmental Engineering and Technology
               Prepared by

               Industrial Environmental Research
               Laboratory
               Research Triangle Park NC 27711

-------
                  RESEARCH REPORTING SERIES


 Research reports of the Office of Research and Development, U.S. Environmental
 Protection Agency, have been grouped into nine series. These nine broad cate-
 gories were established to facilitate further development and  application of en-
 vironmental  technology.  Elimination of  traditional grouping was consciously
 planned to foster technology transfer and a maximum interface in related fields
 The nine series are:

    1. Environmental Health Effects Research

    2. Environmental Protection Technology

    3. Ecological Research

    4. Environmental Monitoring

    5. Socioeconomic Environmental Studies

    6. Scientific and Technical Assessment Reports (STAR)

    7. Interagency  Energy-Environment Research and Development

    8. "Special" Reports

    9. Miscellaneous Reports


This report has been assigned to the MISCELLANEOUS REPORTS series. This
series is reserved for reports whose content does not fitinto one ofthe other specific
series. Conference proceedings, annual reports, and bibliographies are examples
of miscellaneous reports.
                        EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commerciat products constitute endorsement or recommendation for use.

This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.

-------
                              EPA-600/9-81-017
                              March 1981
    Proceedings:  Symposium  on
Iron and Steel  Pollution Abatement
         Technology for  1980
(Philadelphia,  PA,  11/18-11/20/80)
                Franklin A. Ayer, Compiler

                Research Triangle Institute
                   P. 0. Box 12194
              Research Triangle Park, NC 27709
                Contract No. 68-02-3152
                    Task No. 3
               Program Element No. 1BB610
            EPA Project Officer: Robert V. Hendriks

           Industrial Environmental Research Laboratory
        Office of Environmental Engineering and Technology
             Research Triangle Park, NC 27711
                    Prepared for

          U.S. ENVIRONMENTAL PROTECTION AGENCY
             Office of Research and Development
                 Washington, DC 20460

-------
                            PREFACE

  These proceedings for the "Symposium on Iron and Steel Pollution Abate-
ment Technology for 1980"  constitute  the final report submitted to the
Industrial Environmental Research Laboratory, U.S. Environmental Protec-
tion  Agency  (IERL-EPA),  Research  Triangle  Park, NC.  The symposium
was conducted at the Benjamin Franklin Hotel in Philadelphia, PA, Novem-
ber 18-20, 1980.
  This symposium was convened to provide participants an opportunity to
exchange information on technology problems related to air, water, and solid
waste  pollution  control in  the  iron and steel  industry, the program
included a Keynote Address, presentations  on the environmental aspects of
a proposed formcoke demonstration plant, and the future of steel technol-
ogy and the environment. Sessions were conducted on air pollution abate-
ment, covering coke plant emission control, fugitive emission control, inno-
vative air pollution technology, iron and steelmaking emission control, and
inhalable particulates; water pollution abatement, covering recycle/reuse of
water, coke plant wastewater treatment and coke plant wastewater new
developments; and a  session on solid waste pollution abatement.
  Mr. Robert  V. Hendriks, Chemical Engineer, Chemical Processes Branch,
Industrial Environmental Research Laboratory, U.S. Environmental Protec-
tion  Agency,  Research Triangle Park, NC, was Project Officer and General
Chairman for  the symposium.
  Mr. Franklin A. Ayer, Manager, Technology and Resource Management
Department,  Center fors Technology Applications,  Research Triangle Insti-
tute, Research Triangle Park, NC, was symposium coordinator and compiler
of the proceedings.
                                 ii

-------
                           TABLE OF CONTENTS
                                                                            Page

OPENING SESSION	  1
  Robert V. Hendriks, Chairman

Statement of Symposium Objectives	  3
  Robert V. Hendriks

Smarter Regulation: Getting Business to Find a Better Way	  5
  William Drayton

Environmental Aspects of the Proposed Inland Steel
Formcoke Demonstration Plant	  11
  Donald C. Lang, Michael 0. Holowaty,* and Harold L Taylor

Future Steel Technology and the Environment	  25
  Joel S. Hirschhorn

Session 1: AIR POLLUTION ABATEMENT	  35
  Thomas J. Maslany, Chairman

COKE OVEN EMISSION CONTROL

A Model for Coke Oven Charging Emissions	  37
  William B. Tucker

The SCAT System for Fugitive Particle Emission Control	•    53
  Shui Yung, Richard Parker*,
  Seymour Calvert, and  Dennis Drehmel

FUGITIVE EMISSION CONTROL

Fugitive Emission Control of Open Dust Sources	• .  .  71
  Thomas A. Cuscino, Jr., Chatten Cowherd, Jr.,*
  and Russel Bohn

Wind Velocity Distribution Over Storage Piles
and Use of Barriers	, .  .    86
  S. I. Sou/ J,  C. P«rw, fiiui S. Hnzakrwiy

Particulate and SO2 Emission Factors for
Hot Metal Desulfurization	   107
  Jim Steiner* and B. J. Bodnaruk

INNOVATIVE AIR POLLUTION TECHNOLOGY

Demonstration of the Use of Charged Fog
in Controlling Fugitive Dust from Large-Scale
Industrial Sources	   129
  Edward T. Brookman

* Denotes speaker

                                        iii

-------
                                                                                Page
 A Gravel Bed Filter with Fluidized-Bed
 During Reverse Cleaning	  145
   Van Xingzhong* and Wang Nengqin

 Fine Particle Control at High Gas Temperature	  161
   Michael A. Shackleton

 IRON AND STEELMAKING EMISSION CONTROL

 BOF and Q-BOP Hot Metal Charging Emission Comparison	  177
   C. W, Westbrook

 Field Evaluation of Fugitive Emissions from BOF
 Steelmaking Shops	  197
   Peter D. Spawn, Thomas J. Nunno,*
   Stephen G.  Piper, and Larry F. Kertcher

 Status of Casthouse Control Technology in
 the United States, Canada, and West Germany in 1980	  217
   Peter D. Spawn,* Thomas J. Maslany, and Richard Craig

 Efficient and Economical Dust Control System for
 Electric Arc Furnace	 237
   Leon Hutten-Czapski

 INHALABLE PARTICULATES

 Iron and Steel Inhalable Particulate Matter
 Sampling Program: An Overview Progress Report	  249
   Robert C. McCriilis

 Session II: WATER POLLUTION ABATEMENT  	 267
   George F. Haines, Jr., Chairman

 RECYCLE/REUSE OF WATER

 Applying Greenfield Water System Design and
 Management Techniques to Exfsting Steel Plant Facilities	 269
   Harold J. Kohlmann  and Harold Hof stein *

 The Regeneration  of Nitric and Hydrofluoric
 Acids from Waste Pickling Liquid	 283
   Hu Deiu, Liang Xiuchung, and Wang Ch ing wen"

 Steel Industry Pickling Waste and  Its Impact
 on Environment	 293
   S. Bhattacharyya

'Denotes speaker
                                        iv

-------
COKE PLANT WASTEWATER TREATMENT

The Effects of Pretreatment on Coke Plant
Waste water Biological Treatment Systems	  307
  Bernard A. Bucchianeri *, Leon W. Wilson, Jr.,
  and Kenneth D. Tracy

Process Control for Activated Sludge Treatment
of Coke Plant Wastewaters	  321
  Andrew C. Middleton

Biological Treatment of Coke Plant Waste
Using an Integral Clarification Concept	  343
  Myrl R. Wear,* James A. Grantz,
  and Ronald J. Thompson

COKE PLANT WASTEWATER NEW DEVELOPMENTS

Treatment of Coke Plant Wastewater Using
Physical-Chemical and Biological Techniques	  379
  Richard Osantowski* and Anthony Geinopolos

Single Stage Nitrification of Coke Plant Wastewater	  395
  George Wong Chong and John D. Hall*

Nitrogen and Contaminant Control of Coke Plant
Effluents in an Upgraded Biological System	  457
  T. R. Bridle, H. Melcer,*
  W. K. Bedford, and B.  E. Jank

An  Investigation of Foreign By-Product Coke Plant
and Blast Furnace Wastewater Control Technology	  479
  Harold Hofstein* and Harold J. Kohlmann

Factors Influencing Biological Nitrification of
Steel Industry Wastewaters	  497
  Ronald D. Neufeld

Flotation of Iron-Cyanide Complexes from
iron and Coke Plant WastiWattrs	   .  	  &01J
  R. 0.  Hues!i,  G. W. Lower.* and  D. J, Spettliw&od

Session 3: SOLID WASTE POLLUTION ABATEMENT	  521
  John S. Ruppersberger

Impact of the Resource Conservation and Recovery
Act (RCRA) on the Steel Industry	  523
  Penelope Hansen* and William J. Kline
"Denotes speaker

-------
Spent Sulfuric Pickle Liquor Recovery
Alternatives and By-Product Uses	  551
  Wayne C. Micheletti,* Peter A. Nassos,
  and Koren T. Sherrill

Environmental Appraisal of Reclamation Processes
for Steel Industry, Iron-Bearing Solid Waste	  577
  A. 0. Hoffman,*  E. J. Mezey, J. Varga, Jr.,
  W. G. Steedman, and R. D. Tenaglia

Handling and Deoiling of Rolling Mill Scale and Sludge-
A Profit Centerfrom a Problem	  607
  L. A. Duvai

SYMPOSIUM SUMMARY	  619

Closing Remarks  	• •  •  621
  Robert V. Hendriks

APPENDIX: Attendees	  623

* Denotes speaker
                                          vi

-------
                   OPENING SESSION

Chairman:   Robert V. Hendriks
            Industrial Environmental Research Laboratory
            U.S. Environmental Protection Agency
            Research Triangle Park, NC

-------
                     STATEMENT  OF  SYMPOSIUM OBJECTIVES

                            Robert   V. Hendriks
               Industrial Environmental Research Laboratory
                        Research Triangle Park, NC


Being here  in Philadelphia, I  am reminded of one  of  this  country's most
remarkable citizens - Benjamin Franklin.  If Franklin were alive, he would
share many  interests  with those of us  here today.   In a sense,  he was a
composite of all of us in  this  room.  He had a keen interest in science and
technology and made many important scientific contributions  in his day.  He
was an inventor, with several notable inventions to his credit.   He was an
author with papers  on many subjects.  He  was  a dedicated public servant.  Yet
the most remarkable thing about Franklin was his vision and courage and his
desire to make life better for his fellow man.
Franklin can well  serve as a symbol for this Symposium as we, too, try to make
life better through an improved environment.   During the next few days we
will  focus  our  attention  on  technological  solutions   to  environmental
problems  in  the  iron  and steel  industry.   Air,  water, and  solid  waste
pollution control  technology will be covered.  You will hear about innovative
technologies and ideas, improvements in design and operating procedures, and
the  results  from  the  latest  assessment  programs.    We  have  a  number of
excellent papers and excellent speakers representing a wide cross section of
industry, government,  and contractor viewpoints.  This year, the symposium
even takes on  an  international flavor with  speakers from Canada  and the
Peoples Republic of China.  But  the papers not  only convey information, but
serve  as  a means  of  generating  questions  and  discussion  from you,  the
audience.

In additon  to  this opening session,  the Symposium  is divided into  three
sessions - Air, Water,  and Solid Waste.  Each speaker will have 20 minutes for
his presentation.   To encourage questions, we have grouped papers by similar
topics to get several views of  the  same topic.  After presentation of all the
papers in the group, the speakers will sit on a panel to answer any questions
that you might  have concerning their papers or particular  area of expertise.

I see a number of familiar faces here today.  Based on last  year's attendance,
almost one-fourth  of us work for federal and state regulatory and enforcement
agencies,  one-fourth  of  us  work  for  steel-producing companies,  and the
remaining  one-half   of  us  represent  contractors  and  other  interests.
Although we frequently sit on opposite sides  of the table,  we all have at
least  one  similar  goal  - to improve  environmental control  in  the  steel

-------
industry.  I think our diversity is a, good thing.   This  is an open forum in
which we encourage a variety of viewpoints.  We invite you to ask the tough
questons and bring important issues to the forefront.

I welcome you here on behalf of my colleagues in the  Industrial Environmental
Research Laboratory and hope you enjoy the symposium.

-------
          SMARTER REGULATION:   GETTING BUSINESS TO FIND A BETTER WAY
                              BY WILLIAM DRAYTON

     Rules, even the very best possible rules,  are crude and wasteful.   They
can't do what the manager on the spot can:   find the most efficient way of
getting whatever the public wants done done.   Worse, an unrelieved regime of
rules takes away a manager's incentive to find solutions, let alone to innovate.
     No one can measure just how wasteful our rule-dominated approach to
regulation is.  We don't have a practical alternative in place with which to
compare it.  However, it's clear that the cost is enormous.
     In 1977 EPA stopped to ask if all the regulations it was writing one by
one* made sense when applied to specific plants.  At the first plant it found
that the cost of removing one pound of dust from the emissions of different
regulated processes ranged from less than 25$ to over one hundred dollars.
Subsequent studies—both of particular plants and of the comparative cost
effectiveness of the Agency's many regulations—confirm that wide variations
between the costs per pound removed are common.
     Obviously, we'd all be better off if we could control more 25$, or even
$1 pounds and forget about an equal number of the very expensive hundred
dollar pounds.  A DuPont study of fifty of its plants suggests that the company
could cut its $106 million annual hydrocarbon pollution control bill over
60 percent if it were allowed to make one simple change to the mix of its
controls.
     However, under the traditional "command and control" regulatory regime
neither the company or those charged with regulating it could make this sort
of common sense trade.  Every process must remove a specific percent of the
pollution it would otherwise emit by a specific, legally binding deadline.  A
rule is a rule.
*Like many other regulatory agencies EPA writes regulations fitted to the
particular characteristics of specific production processes, e.g., paint spray
booths, degreasing operations, printing presses, oil storage tanks.

-------
     EPA is changing that.  Its air pollution rules are no longer chiseled in
concrete.  If a company can make a more sensible counterproposal, one that
will save the company money and still get the environmental job done, EPA will
happily accept.
     Instead of "command" (rule writing) and "control" (enforcement), EPA is
moving to a quite fundamentally different "command, counterproposal, and
control" approach.
     Here's how EPA's new counterproposal approach works.
     EPA begins this new approach much as it did before:   it writes the best,
most responsible rules it knows how.  However, it then encourages business to
trade reduction requirements, substituting more efficient, sometimes innovative
controls for relatively expensive requirements, both within and across plant
lines.   To provide the necessary incentive EPA's policy leaves the savings to
business.  And the Agency is working to create new market structures to facili-
tate this trading.
Offset Trading
     It is illegal to build a new facility in an area where the air is too
filthy to meet basic health standards, if doing so will aggravate the area's
air quality problem.  To prevent an intolerable conflict between the need for
local economic growth and modernization on the one hand and this statutory
public health requirement on the other, EPA first adopted the idea that one
source of pollution could meet its environmental obligation by getting another
source to take additional control actions.
     The offset policy requires the new facility to control its emissions as
tightly as possible and then to offset what it can't control with reductions
of the same pollutants elsewhere in the community.  Thus,  for example:
          When the State of Pennsylvania successfully attracted Volkswagen to
          a Western Pennsylvania site, one element in its package of inducements
          was the provision of sufficient offsetting reductions of the pollutants
          that cause photochemical oxidants in the area to meet this requirement.
          It provided these offsets itself in large part by shifting to water-
          based rather than petroleum-based asphalt in its road building and
          repair work in 16 nearby counties.   The VW plant now needs additional
          offsets and it and the state are exploring several prospects at
          nearby steel facilities.

-------
          Similarly, the Chambers of Commerce in Shreveport,  Louisiana,  and
          Oklahoma City, Oklahoma, made room for new GM plants in their  com-
          munities by inducing local oil companies to clos6 marginal  facili-
          ties and to reduce storage tank emissions;

          General Portland Cement paid Parker Brothers $520,000 to install
          dust collectors on its facility in New Braunfels, Texas, so that
          General Portland could add a new coal-fired preheater to its plant
          there without pushing the area over the health standard for partic-
          ulates.

     Even during our first two years'  experience with this new tool,  it  has

made a significant difference.  Despite the barrier of novelty and the difficulty

of knowing what possible, economically attractive control alternatives exist

outside any one manager's sphere of control, business has completed 650  such

offset trades during these start-up years.  The banking reform outlined  below

will make trading much easier and no doubt increase its volume substantially.

The Bubble

     The "bubble" allows managers to escape the narrow vision of existing

process-by-process regulation and to look at their facilities as a whole.  As

they plan their counterproposal they can imagine that their facilities are

covered by an enormous plastic bubble.  As long as the bubble doesn't let more

pollution escape than the sum of all the process regulations it covers would,

the manager can go about controlling the several sources of pollution under

the bubble as he thinks makes most sense.  As long as air quality is protected,

the bubble can stretch well beyond one plant's boundaries.  It covers all

existing processes.

     Although the bubble policy was only announced nine months ago, industry

has already proposed over seventy, quite varied bubble trades.  For example:

          3M's Bristol, Pennsylvania, plant has proposed to bubble ten tape
          coating lines.  It would save several million dollars by using
          water-based coatings and a new solventless "hot melt" process  to
          reduce emissions more than EPA's standard requires on some lines and
          cut back controls on others.

          Weyerhauser proposes to save $5 million in capital costs, $200,000
          in annual operating costs, and up to 1 million barrels of oil  a year
          at its Plymouth, North Carolina, plant by increasing its use of wood
          in an existing oil/wood waste boiler and shutting down several oil-
          only boilers.  These trades would hold particulate emissions constant.

-------
          Andre Greenhouses proposes to switch from low sulfur oil to a mix of
          high sulfur oil and natural gas at three locations in Pennsylvania.

     As we gain more experience with this central innovation, as we work out

its bugs and loosen some of its restrictions as we gain confidence, we can

expect it to become more and more important.

Banking

     In order to further facilitate controlled trading in emission reductions

EPA is also assisting states in developing "Banking" regulations.  "Banking"

will provide additional cost savings and incentives for innovation by allowing

businesses to achieve extra control (more emission reduction than otherwise

required) at the most cost-effective time and then using those reductions as

credits toward emission reduction requirements at some later time.
     This new approach promises to be the most significant change in how this

country regulates since the 1930 's.

          It returns to business the flexibility to figure out the best way of
          getting the job done.  (It leaves the regulatory agency with the
          responsibility of checking to ensure performance.)

          It provides for the first time the same powerful, bottom line incen-
          tive for business to find new, more efficient methods of control
          that managers now have to cut production costs.  "Command and control"
          provides no such incentive to innovate.  A plant engineer that found
          a better way of controlling pollution in one area until now could
          not benefit elsewhere.  The much increased rate of control technology
          innovation this reform promises is its most important benefit.
          Without this incentive, the environmental movement is in very deep
          danger.  The volume of pollution compounds every year as population
          and the economy compound.  But there is never any more air, land, or
          water.  Consequently, the cost of any given level of environmental
          safety increases each year.  The only way to avoid this dangerous
          result is the increased rate of control technology innovation this
          new approach wou.ld incite.

          It should go a long way to improving industry's dismal failure to
          operate and maintain control equipment properly.  Now the average
          "complying" source of air pollution emits more than 25 percent more
          pollution than it legally should, even netting out periods when it
          emits less.  This problem, already critical , will get steadily worse
          as equipment ages unless we give plant managers and engineers new
          incentives.  Now they commonly neither understand nor care much
                                      8

-------
          about this equipment.   However, once they design their own controls,
          they will have both the necessary understanding and a paternal
          professional psychic investment in making them work.
          It should soften the adversarial, overly lawyer-dominated relation-
          ship now common between managers and regulators by getting both
          sides to talk more about how to get a job done rather than only
          about whether or when it should be done.
          It will save business a great deal of money.   (With air and water
          pollution cleanup costs alone running at roughly $40 billion per
          year, even a small 10 percent savings would be worth $4 billion a
          year.  The greater efficiencies I think likely and/or applying the
          reform to broader areas of regulation could of course multiply the
          savings.)
The same counterproposal principle EPA is currently demonstrating in the air
pollution area could be applied widely across government's other regulatory
programs.

                                   Addendum
                Freeing the Bubble from Unnecessary Constraints

     Our first eight month's experience with the Bubble has convinced us we
need to relax several unnecessary constraining rules.
     The Bubble policy asked persons who wanted to use the Bubble approach at
their plants to estimate through "models" the impact of their emissions on air
quality by methods more stringent than were used to set existing emission
limits, a common practice when reviewing "variances."  In some cases, the
models showed micro-violations at one or more points, technical violations
which prevented the bubble even when the bubble would reduce overall pollution.
     We are modifying our estimating procedures.  The policy now says that
between similar pollution sources we will ask only if total emissions from the
bubble counterproposal are the same or less than from the existing regulation.
If so, no models are needed.  We will only model in cases where the proposed
changes will increase emissions on balance.
     The bubble policy said that to make the counterproposal federally enforce-
able, each had to be treated as a revision to the applicable State regulation.
Revisions are procedurally slow and require multi-layer government reviews.

-------
Many businesses have claimed the prospect of such delay prevented them from
designing a counterproposal.
     The State of New Jersey proposed In its regulations a generic provision
allowing bubbles under specified conditions.  On October 6, EPA accepted this
New Jersey approach.  EPA specifically announced that it would henceforth
allow states to adopt generic review procedures for bubble applications,
thereby making unnecessary a formal revision to state regulations for each
bubble.
                                      10

-------
          1-NVIROlM-NTAL ASPECTS OF TIE: PROPOSED

       INLAND STEEL FORMCOKE DEMONSTRATION PLANT



       Donald C. Lang, Director, Air and Water Control

       Michael 0. Holowaty, Senior Advisor, Research

       Harold L. Taylor, Senior Advisor, Research


                  Inland Steel Company

                  East Chicago, Indiana
                        ABSTRACT
      A brief history of the development of formcoking processes
for metallurgical coke production is presented and blast furnace
evaluations of formcoke produced with the FMC formcoking process
are reviewed.  The flow sheet for the Inland Steel Formcoke
Demonstration plant is presented and the emissions control
equipment and environmental advantages of the FMC formcoke
process are discussed based on the results of recently completed
engineering studies.
                             11

-------
                  ENVIRONMENTAL ASPECTS OF TIIE PROPOSED
                INLAND STEEL FORMCOKE DEMONSTRATION PLANT

REVIEW OF FORMCOKE PROCESSES

      The concept of a formcoke process was proposed by Dawson in 1856.
At the present time, almost every steel producing country has a project
or an idea for formcoking.  The reason for the drive to drastically change
or to completely replace the conventional by-product coking process lies in
the fact that the formcoke process is better suited for present conditions.
The by-product coking process was satisfactory when the supply of coking coal
was abundant, air pollution standards more lenient,and labor cheap.  This is
no longer the case.  Environmental control is of paramount importance; the
cost of labor and equipment have skyrocketed, and the supply of inexpensive
coking grade coals is being exhausted.

      The first recorded commercial production of partially devolatilized
coal agglomerates in the united States took place in November, 1933, near
the Champion coal preparation plant of the Pittsburgh Coal Company located
approximately 32 km west of Pittsburgh.  Eventually, two more production units
were added in 1934 and 1936, and the average production for the three kiln unit
was about 5900 Mg per week of low temperature coke known as Disco fuel.

      The Disco plant operations were discontinued in 1946.  There was little
activity in the formcoking field until about 1951, when the Belgian Iniex
process and the British National Coal Board process reached the experimental
stage.

      General speaking, the formcoking processes developed since 1951 fall into
one of the following three categories:

      1.  Hot briquetting - coal acting as the binder

      2.  Hot balling - coal or coal tar pitch acting as the binder
      3.  Cold briquetting of char - coal tar pitch acting as the binder

Hot Briquettes - Ancit Process^ '

      The hot briquetting process was originally developed in Europe.  The
Iniex process, the B.B.F.-Lurgi process, the Australian Auscoke, and the
Luxembourg Ancit processes are typical for this approach to the coal agglomerate
problem.  The Ancit process is shown diagramatically in Figure 1.

      The production of formcoke by this process requires approximately 65%
inert low volatile coal (anthracite) or char to be preheated to about 590°C.
At this point, it is thoroughly mixed with a strongly coking coal so that the
temperature of the blend increases to almost the temperature of incipient
fluidity (400-430°C).  At this point, the blend is force fed between two
briquetting rolls and the resulting briquette is green formcoke.  It can be
completely coked and devolatilized by heating to about 870°C for 1 hour.
Subsequently, it is cooled to 120°C and used in the blast furnace.
                                    12

-------
r4-
[F
                       MMMflENv W MflH MMNI^KfW
                       C«rfilt wMh Ita4tat CM! MM
FIGURE 1 FLOWSHEET OF THE ANCIT FORMCOKING PROCESS

-------
      Two formcoke plants were built on the basis of the B.B.];. process -
a 270 Mg/d plant at Prosper in Germany, and a 550 Mg/d plant at Scunthorpe,
England.  Both plants now are inactive because of process difficulties.


Coke Pellet Process

      A second formcoke process being developed in the United States is the
Coke Pellet Process, or the Consol-BNR process.  The unique element of the
process is the mixing of preheated coal, preheated char, and preheated pitch
under controlled conditions in a rotating drum to form a sized agglomerate
(snowballing effect), which is later calcined at an elevated temperature to
produce coke,

      Two of the more important objectives of the Coke Pellet Process are:

      1.  Produce strong, sized blast furnace coke from a single coal
          or from coal that cannot be used as a 100 percent charge to
          a conventional slot oven (due, for example, to unfavorable
          swelling properties).

      2.  Reduce pollution.


Clean Coke Process

      The Clean Coke Process is a development of United States Steel Corporation
and, since 1973, has been developed under a joint project with the Office of
Coal Research (now a part of the Department of Energy).  The objective of this
development is to permit the use of high sulfur, nonmetallurgical, high ash
coals for cokemaking while producing low sulfur liquid and gaseous fuels as
by-products in an environmentally acceptable manner.  Estimates made for the
case of processing 1 Mg of high ash, high sulfur Illinois^coal indicate a
yield of about 0.25 Mg coke, 0.12 Mg chemicals and 110 dm  of liquid fuel.
The process, therefore, is much more than a formcoking process.


FMC Formcoke Process '1>5-*
      FMC started laboratory- scale work on a formcoking process in 1956.,  By
1958, the first few kilograms of coke were produced which appeared strong and
useable.  The basic concept of FMC was directed initially exclusively to the
needs of their Pocatello, Idaho, phosphorus smelting furnaces.  There, the
high delivered cost of the small sized by-product merchant coke was a great
incentive to such an effort.  However, they soon recognized that their process
also could benefit the steel industry; and consequently, a joint venture was
formed between FMC and united States Steel Corporation.  The objective was to
build and operate a formcoking plant at Keumerer, Wyoming, where very extensive
coal reserves of subbituminous coal are available.  This high moisture coal was
regarded for years as excellent steam coal with rather poor storage characteris-
tics because of its strong oxidizing properties.  Extensive tests in the FMC
                                    14

-------
laboratory and later at the  1 Mg/d FMC pilot facility at Philadelphia showed
that this geologically young coal with absolutely no coking properties (in
the conventional sense) yielded carbon agglomerates of apparently, satisfactory
characteristics.  'The Philadelphia coke was first tested in Pocatello in 1959
and 1960 with excellent results.

      The FMC process (Figure 2) utilizes the simple concept that coal having
no agglutinating properties must first be brought to the level of carbon (char
or calcinate) and then blended with binder and agglomerated to the desired
shape and size.

      In the FMC flowsheet, the coal is first crushed to (-3 mm), dried, pre-
oxidized, and preheated in the first process step.  The crucial carbonization
step is carried out in the second fluidized bed vessel in which the previously
preheated coal is partially devolatilized with steam and air.  It is here that
tars evolve which are later used for bonding the carbonaceous particles into a
coke agglomerate.  The discharge temperature from the carbonizing step reaches
480°C and the char, which is somewhat spheroidal in shape, moves to the third
and last process vessel - the calciner, which operates at 870°C.  Here the
volatile matter is reduced from up to 171 to about 3%.  Simultaneously, a
shrinkage of the char particles is effected.  The evolving low heating value
gas (5.7 MJ/m3) is used internally in the plant.

      The calcinate, cooled to approximately 100°C, is thoroughly and intimately
blended with a low temperature pitch binder produced from the raw coal tar and
this blend is briquetted.  The amount of the binder required varies between 8
to 151 by weight, depending on the characteristics of the calcinate.  The
briquettes are subsequently treated by first heating in an oxygen rich atmosphere
to harden them and then devolatilized to a coke product containing about 21 V.M.

      The Kemmerer, Wyoming, formcoke plant, built jointly by FMC and united States
Steel, went on stream in 1961 and quickly started producing formcoke sized at
32 mm x 19 mm.  In appearance, this plant resembles a chemical plant or a
petroleum refinery and not a conventional coke plant.  By the summer of 1962,
a shipment of 1820 Mg of Kemmerer formcoke was tested in the United States Steel
experimental blast furnace at Universal, Pa., with a 1.2 m diameter hearth.
The formcoke performance was compared with that of Gary by-product coke sized
in two ranges:  small coke 19 mm x 5 mm, and large coke 32 mm x 19 mm.

      The furnace experienced very few problems, and for the duration of the
test,  the Gary by-product coke and the FMC coke were easily interchanged.

      In 1966, an agreement was reached between-Armco steel, Inland Steel,  and
the FMC to carry out the first test with formcoke on an Armco commercial blast
furnace at Hamilton, Ohio.
                                    15

-------
                      CLEAN GAS
      COAL -10 MESH
•SH        V

H^TCIS'
                                          FUEL GAS (ISO BTUS)
                            SCRUBBER
                              >

                     DRYER    £'
         VOLATILE MATTER
                                 CARBONIZER
                                 FUEL GAS
                                 (170 BTUS)
                                     BAG FILTER
                                             CAICINER
                                                       COOLER
en
+ AIR 1 ^ ^^
STEAM + AIR ^^k
HEATED AIR
IN

S»
* ^
ERT GAS
          BLENDER
                 BRIQUET rib
                  PRESS ML/
                            HEATED AIR
  PRESS
                         I
                         I
                         l_.
J	f
  CURING
   OVEN
                        O  GAS TO
                           INCINERATOR
                                                                MAKE LIQUOR



                                                                  BLOWING LOSS

                                                            TAR PROCESS
                                                                               I	
                                                                    .FORMCOKE
                                                                       PRODUCT
                         FIGURE  2  FLOWSHEET OF THE FMC FORMCOKE PROCESS

-------
      First, a preliminary test with 500 Mg of 35 mm briquettes was conducted
in October, 1966.  ITie problem of coke fines, which were generated in handling,
was noticed but it appeared that the quality of the briquettes being produced
for the test could be improved.  In the spring of 1967, 11,700 Mp  of 50 mm
formcoke pillow briquettes and about 1,300 Mg of 35 ram pillow briquettes were
shipped to Armco's blast furnaces at Hamilton, Ohio.  The program was to re-
place 100% of,the furnace coke as soon as possible using a stepwise approach.

      After a short period at 50% replacement with 50 ram formcoke in the burden,
the furnace operation became erratic; the furnace went cold, slipped, and
filled the tuyeres and blowpipes with slag.  During the test, as compared to
a normal operating period, the production rate decreased by about 35%, and
the furnace operation was erratic.  On the basis of flue dust analyses, it was
estimated that about 110 kg of formcoke fines were blown out per each Mg of
hot metal produced.

      The Armco-Inland-FMC test was obviously somewhat of a disappointment.
However, it must be borne in mind that in this test, large 50 mm briquettes
were used extensively for the first time.  The test was conducted without
knowledge of what to expect from this new coke and what properties to demand.
The results strongly indicated that the coke must be structurally strong and
abrasion resistant.

      In October of 1970, FMC and Inland Steel undertook a joint study to
determine the feasibility of producing formcoke from No. 6 Illinois coal in
the Kemmerer plant. For this purpose, 820 Mg of Illinois coal were shipped to
Wyoming and processed into 38 mm formcoke briquettes in September of that year.
It was found that the Illinois coal could be used quite easily if it were
properly preoxidized in the first process step of the FMC process.

      To help control the surface abrasion, the finished coke flowing from
the loadout bin into the railroad car was sprayed with a solution of calcium
lignosulfohate, which appeared to easily soak  into the briquettes. A total
of three cars of Illinois formcoke were received at the Indiana Harbor blast
furnaces, of which two were treated with lignosulfonate and one was left
untreated.  The unloading of these cars into highline bins showed that the
treated formcoke generated very little airborne dust, particularly when water
was sprayed during the unloading operation.  The untreated car was quite dusty.

      The lignosulfonate treatment was obviously not the only improvement which
affected the quality of the resulting Kemmerer briquettes.  Changes which greatly
influenced the structure of the briquettes were made in the plant operating
practices and a large scale test on an intermediate size, well operated blast
furnace was considered.

      At about this same time, the British Steel Company and FMC agreed on the
production of 2,730 Mg of FMC formcoke made of Kemmerer coal, which would be
subsequently shipped to England for testing at the East Moors Blast Furnace
No. 3(4).  The formcoke sized 38mmx38mmx25mm was shipped in stages with
transfer points at St. Louis, New Orleans, Barry near Cardiff, and finally
East Moors.
                                    17

-------
      The conclusion reached by the British Steel Corporation can be summarized
as follows:

      1.  The shipment of FMC formcpke, which was treated with ligno-
          sulfonate, was not hampered by dust problems and breakage,
          despite the numerous transfer points and the distance (about
          10,500 km).

      2.  This trial established beyond all reasonable doubt that calcine
          based formcoke could replace a superior quality conventional
          coke at the 100% level with improvements in efficiency and at
          potentially higher production levels.

Consortium Test

      In the spring of 1973, a consortium of five steel companies and FMC
was formed to conduct a large scale blast furnace test with FMC formcoke.
These included Armco Steel Corp., J§L Steel Corp., Inland Steel Company,
McLouth Steel Co., and U. S. Steel Corporation.

      The detailed information developed during the total test program with
almost 18,200 Mg of coke has been presented elsewhere15).  However, in
summary, results were as follows:

      1.  The shipment, storing, reclaiming, and handling of the 50 mm
          formcoke to the plant and in the blast furnace charging system
          presented no difficulties.

      2.  The furnace operation was found to be normal when up to about
          50 percent FMC coke was used.  At high levels of substitution,
          increasing instability of operation was noted, which was attributed
          to factors related to the materials distribution pattern within
          the furnace.

      Because of the limited quantity of formcoke available for the test,
development of operating practices to achieve stable operation at formcoke
substitution levels up io 100% was not possible.  Inland Steel concluded that
the experimentation required a formcoke production plant with a production
rate that would be in balance with an operating blast furnace.  Thus, to
advance the technology. Inland Steel, in 1978, retained Davy McKee to conduct
a preliminary engineering study of a 910 Mg/d formcoke production plant.  The
results of this study were used as a basis for an unsolicited proposal to
the Department of Energy for a cooperative agreement to conduct the Inland
Steel Formcoke Demonstration Project.  This proposal was submitted in April,
1979, and is currently being considered by DOE.
                                     18

-------
ENVIRONMENTAL ADVANTAGES OF THE FMC FORMCOKE PROCESS

     The emission-control problems associated with conventional coking
operations have been well documented and are only too apparent to all of us
who have striven over  the years to improve on coke battery pollution abate-
ment.  The great difficulties which we have experienced in controlling
battery emissions are caused by the complexity of operations inherent to
batch production of coke.

     We won't attempt to review those operations in any great detail, but it
may suffice to say that the charging, coking, pushing and quenching steps
associated with operating multiple-slot oven batteries create a great number
of potential sources of emissions.  For example, a typical coke battery might
consist of 87 ovens which--in one complete operating cycle of from 12 to 24
hours--requires at a minimum, opening and closing 174 doors, 435 topside lids,
87 standpipe lids, to say nothing of maintaining seals on these closures during
the actual coking period.  Recognizing that a coke battery constructed of
large quantities of brick is subjected to considerable thermal stress during
its lifetime, which exacerbates the leakage problem, it is no wonder that
pollution abatement requires an intensive round-the-clock maintenance effort.
Even when emissions are minimized, the major problem remains of collecting or
capturing these emissions where possible from the multiple sources for subse-
quent removal in pollution abatement equipment.  While great strides have been
made to date, coke batteries remain one of the toughest problem areas and the
last to yield to innovative technology.  The best equipment and practices
available today simply cannot prevent intermittent emissions from coke oven
batteries.

     The FMC formcoke process, by its very conception as a continuous production
sequence, offers excellent opportunities to reduce emissions.  The coal feed is
processed continuously in fluid bed reaction vessels which isolate the coal
from the atmosphere.  As can be seen in the formcoke flowsheet in Figure 3, the
coal once fed into the first fluid bed vessel is not exposed to the atmosphere
until it exists as calcined char which is mixed with binder and briquetted at
midstream of the process.  The green briquettes enter again an enclosed process-
ing equipment and reappear as cooled coke briquettes being delivered to the
storage silo.  All of the process vessels are made from continuous steel shells,
rather than the nonmonolithic masonry in conventional coke batteries.

     Prevention and repair of leaks in steel casings are relatively simple.
The vessels are not routinely opened and closed as an integral part of the
production process, therefore, there are no charging, door, topside, pushing
or quenching emissions characteristic of slot coke oven batteries.

     As a result of:  1) the Inland-Davy McKee Engineering study;  2) the
proposal evaluations conducted for DOE by The Ralph M. Parsons Companyt°J and
for the Environmental Protection Agency by Research Triangle Institute (?); and
3) data obtained from the FMC Kemmerer, Wyoming formcoke plant, considerable
information regarding the environmental aspect of the FMC formcoke process are
now available.
                                     19

-------
     To control airborne emissions ficm an BIG formcoke facility, we would
expect to install LAER (lowest achievble emission rate) technology for a new
facility on the few point sources which exist and BAT (Best Available Technology)
to control waterborne pollutants.  The present process as shown in Figure 3
indicates the following pollution controls:

     1.  Coal handling and preparation is a separate step in the Inland
         Formcoke configuration.  Gases and dust off the coal dryer are
         scrubbed prior to discharge.

     2.  A front end incinerator would combust emissions from:

         (a)   the catalyzer
         (b)   the briquetting step
         (c)   the curing oven, and
         (d)   the binder preparation area

         Cyclonic separators and scrubbing equipment would precede or follow
         the  incinerator as required.   In addition to controlling process
         emissions,  the incinerator would furnish the inert gas required in
         the  process.

     3.  The  off-gases from the Carbonizer and the Calciner which contain coal
         fines in addition to volatile organic compounds distilled from the
         coal are stripped in cyclones, venturi scrubbers, wash towers and
         electrostatic precipitatprs as required.  The tar removed in this
         sequence is sent to the binder preparation plant.  The off-gases are
         combined and after steps to remove naphthalene, ammonia and light
         oils (not shown) sulfur is removed in a desulfurization step--possibly
         a Stretford unit as shown.  The desulfurized gas is-used'to genera'e
         steam in a  boiler for use throughout the steel plant.

     4.  A fabric filter serves the calcinate storage silo to remove fine dust.
         This material can likely be added to the subsequent mixer step prior
         to th<5 briquetter and thus recycled.

     5.  Emissions from the Briquette Curing Oven contain dust and some vojatiles
         and  are combusted in the front end incinerator as previously mentioned.

     6.  Off-gas from the coking kiln passes through a cyclone, venturi scrubber
         and  an electrostatic precipitator to remove dust and tar.   The tar is
         returned to the binder preparation plant.  Some of this low Btu off-
         gas  is returned to fuel the kiln, whereas, the rest is sent through
         its  own desulfurizing step.  The resulting clean kiln gas is then used
         as fuel in  the curing oven and the coal dryer.

     7.  Fugitive omissions from the briquette screening step and from the handling
         and  distribution steps are removed in cyclonic separators or fabric filters
         as required.   The dust is recycled back to the mixer preceding the briquett*

-------
             HAW GUI
        --4-.
      JCOMHAHDUM


      |MWMATtOH
      I
             WfTCOAL
            COAL MVIM
-*|   K«URM«  |—«•
           ranMcoKi MiourrTH
                                          (Winffl TO INiAND
FIGURE  3 PROCESS FLOW DIAGRAM FOR PROPOSED FORMCOKE
          DEMONSTRATION PLANT

-------
     8.  Generally, all scrubber liqur;r would be recycled after solids
         removal in the calcinate thickener.  The sludge could be dried
         and recycled back into the process or combusted in a separate
         system.  Blowdown from the system would be physically-chemically
         treated to remove the small quantities of cyanide and ammonia prior
         to biological treatment.

     In addition to the pollution control that can be achieved with the FMC
formcoke process, recent tests at the FMC Kemmerer plant established that
the plant is in compliance with the 1978 OSHA emission standards controlling
coke oven worker exposure (8).  The results indicated no worker was exposed
to ambient air Benzene Soluble Fraction of Total Particulate Matter (BSFTPM)
concentration as high as the OSHA limit of 150
     In summary, the environmental and health/safety advantages of the FMC
formcoke process are considerable and, while specific emission and effluent
limits have not been set, we are confident that they can be met and at a cost
substantially below that for conventional batteries.  This would include the
abatement equipment, labor and repair requirements and energy consumption.
                                    22

-------
                             BIBLIOGRAPHY


1.   "Formcoke - Still Waiting in the Wings" by J. Dartnell; Iron and
     Steel International; June, 1978.

2.   "Clean Coke Process: Process Development Studies; Final Feport";
     Volume II - Stannary of PDU Studies; FE-1220-39 (Vol. 2) November, 1978,

3.   "FMC Formcoke - Furnace Testing" by Wendell L. Darrow, Eric Sailer,
     and John Poast; Iron and Steel Engineer; August 1973.

4.   "Use of Formed Coke: BSC Experience 1971/1972"; by J. K. Holgate and
     P. H. Pinchbeck, Journal of the Iron and Steel Institute; August, 1973.

5.   "Blast Furnace Test with 20,000 Net Tons of FMC Formcoke  at Inland's
     No. 5 Blast Furnace", by P. K. Strangway and M. 0. Holowaty; Proceed-
     ings of the 37th ABffi Ironmaking Conference; April 16-20, 1978.

6.   "Proposal Review - Inland Steel Formcoke Demonstration Project - Final
     Report"; prepared by The Ralph M. Parsons Company for the U. S.
     Department of Energy under Contract No. DE-AC07-80CS40448.

7.   "Preliminary Environmental Assessment on Formcoke Making Process";
     by C. W. Coy, C. C. Allen and B. H. Carpenter; Contract No. 68-02-
     3170, Task 17; Prepared for Industrial Environmental Research
     Laboratory, U. S. Environmental Protection Agency; June, 1980.

8.   "Meeting Emission Standards with the FMC Coke Process" by W. R. King,
     R. R. Severns and H. A. Kulberg; 39th AIME Ironmaking Conference;
     March 23-26, 1980, Washington, D.C.
                                     23

-------
                FUTURE STEEL TECHNOLOGY AND THE ENVIRONMENT
                            Joel S. Hirschhorn

                      Office of Technology Assessment
                          United States Congress
                         Washington, D. C.  20510
ABSTRACT

     According   to   the  OTA   report   TECHNOLOGY   AND  STEEL   INDUSTRY
COMPETITIVENESS,  there  are  considerable   opportunities  for  major   new
steelmaking technologies to be  created and introduced in the domestic steel
industry during the coming decades.   During the decade of the  1980»s there
will  be continued  increases  in  the  use of  scrap based electric furnace
steelmaking by  both nonintegrated  and integrated  steelmakers.    Moreover,
there  will be  substantial increases  in  the use  of continuous casting of
steel.  One of  the important  impacts  of  these  changes will  be the need to
use  direct reduced  iron  as  a  complement to ferrous  scrap in  electric
steelmaking furnaces.   The shift  to  greater  scrap use  and the  gradually
increasing  use  of  direct   reduced  iron  signifies  less  dependence  on
ironmaking in  blast furnaces fueled primarily  by coke.   Thus,  pollution
should  be  abated.   Although there may be very  limited introduction of coal
based direct reduction during  the 1980's in the United States, in the 1990's
we may see large  scale direct  reduction plants based  on cbaligasification
and  more  DR  plants  using coal  directly.    Federal  policies  shape  the
development and  use of new technology.   Those  policies that aid and reward
the  companies  with  the  poorest  performance  may  be  detrimental  to  a
competitive  domestic  steel  industry,  such  is  the  nature  of  President
Carter's proposal  to  grant extensions to  the  compliance schedule  for  the
Clean Air Act to certain qualifying plants.
                                     25

-------
 FUTURE  STEEL TECHNOLOGY AND THE  ENVIRONMENT

      In his recently  announced  program1  for  the domestic  steel industry,
 President  Carter   linked  the   goals  of  modernization  and  environmental
 protection  in a more  precise  manner than ever before.   The key proposal is
 an  amendment  to the reauthorization Of  the  Clean Air Act which would allow
 the EPA  Administrator to use   discretion,  on  a  case-by-case basis,  for
 extending  by  up  to  three   years the  deadline  for  compliance with  the
 requirements of the Act.   A  number of conditions must be met,  with the key
 point being the requirement  to  modernize  existing steel facilities with the
 funds that  would have been spent  to meet  the original compliance schedule.
 Also  included  in  this   policy  package  for  the  steel  industry  is  the
 acceptance  of greater  use  of the  "bubble  concept"  and the  stabilization of
 discharge permits  under  the  Clean Water Act until EPA's revised regulations
 take  effect in  1981.   How will  changing steelmaking  technology  affect  the
 environment,  and  how  will these  policy changes  affect  the introduction of
 new technology in the  domestic steel  industry?  These are  the  key issues I
 would like  to explore.

 How Much  of a Problem Have Environmental Regulations Been for the Industry?

     The  premise  behind  the  new  policy  thinking described  above is  that
 compliance  with environmental regulations is costly.   As  President Carter
 has said, "Full compliance will require large additional expenditures at the
 very  time  that   the  industry  must   also  make   major  investments  for
 modernization."!    The implication is  that  the  industry   cannot  do  both
 modernization   and   compliance   at  the   same  time  for  all  steelraaking
 facilities.   Just  how easy  a case can  be made  by companies  for  specific
 facilities  for  extension of  the compliance schedule is  not entirely clear.
 But based on the  strong  positions taken  by individual  companies and  the
 American  Iron  and  Steel Institute that  insufficient  profitability  and
capital formation have been debilitating, there is some indication that past
 industry  interpretations of their financial problems will continue, and that
many requests for compliance  stretchouts would be made.  To what extent have
 past environmental regulatory costs been a severe burden to the industry and
 an obstacle to modernization?

     During the period 1971 to 1978 pollution abatement capital expenditures
 equaled  14.5$ of total capital  spending  for the  domestic steel  industry
 according  to  AISI  data.   In that same period, using  data from the  same
 source,  spending  On  non-steel   investments   (diversification)  amounted  to
 16.0%  of total capital  spending.2   Both diversification  and  reinvestment
strategies  have been used  successfully  and  unsuccessfully  to  increase
 profitability.   However,  regulatory  compliance  investments have not  been
 sufficiently large  to prevent diversification  when that was desired,  nor
 have  the considerable diversification  efforts of  the past decade  brought
about any remarkable improvement in steel industry profitability.   Moreover,
the  industry  has  maintained  paying  relatively  high  dividends  to  its
stockholders,  regardless of the  generally low profitability it has  usually
had.  During  this  same period of  1971 to  1978 dividends amounted to 4631 Of
net income and was equivalent to 23.9% of total capital spending.   There  is,
of  course,  a responsibility  to  make a return on stockholders investments as
well as to the public to  maintain a clean  environment.   Nevertheless,  steel
industry management have perpetuated  a  climate in which the value Of owning

                                    26

-------
stock  in  most large steel companies is  based on current return rather than
on  appreciation of  the  stock  which  is hinged  to  perceptions  of  future
success.  Both dividends and non-steel investments are discretionary uses of
available capital, and  it is, perhaps,  an  ethical  issue and a question  of
values  as well as anything else as to whether these uses have priority over
the corporate responsibility to refrain  from polluting the environment.

     Certainly,  diversification  is  a  rightful  perogative  of  corporate
managers, but interestingly Hall's  recent  study of  eight mature  domestic
industries,  including  steel,  revealed  that diversification  has not  been
successful because:  "By waiting too long to  begin  diversification efforts,
most lack the capital and managerial skills  to  enter new markets and/or to
grow businesses successfully  in  these markets.  Thus their  diversification
efforts to date have been too small or have been managed in too conservative
a fashion to  obtain  sustainable performance  improvements, as witnessed  by
the  very minor performance  contribution of  U.S.  Steel's  diversification
program into  chemicals..."3   The exception  for the steel  industry has been
Armco,  the  most diversified  and profitable  "steel  company", which  used a
strategy  of investment in  low  cost steel  production in  selected regional
segments as well as early and well managed diversification.  And it has been
Armco which has lead the effort to use the  "bubble  concept"  to economically
meet emission standards for nontoxic pollutants.

     Since  the   domestic  steel  industry   makes  much   importance  over
comparisons   between   itself   and  the  Japanese  steel  industry,   it  is
interesting to note how to Japanese have responded to the social demands for
a clean environment.  During the period  from 1971 to 1978 the Japanese spent
13.4ft of their total capital spending on pollution abatement.  Although this
is  slightly more  than  what the  domestic industry spent  as a fraction  of
total capital spending, it does not fully capture the difference between the
industries.  By normalizing actual pollution abatement spending on the basis
of the  amount of  raw steel produced during  this period, it is possible  to
see that the Japanese have had a greater intensity Of spending.  The per ton
(metric) capital spending on pollution abatement during  this  period  was 50%
greater for the Japanese as compared to domestic  spending.4   This was true
even though the Japanese had  the advantage Of  new plant  construction,  for
the most part, as compared to retrofitting in the United States.

     The  Japanese  have  put  much  effort  into  pollution  abatement  technology
because Japanese ambient  air  quality standards are generally more stringent
than those in the United States.  For  example,  the 24 hour S02 standard  in
Japan  is  0.04 ppm, as  compared to  0.14 ppm in the  United States,  the one
hour standard for photochemical oxidants is 0.06 ppm in Japan and  0.12 ppm
domestically, and  the standard  for  particulate matter in Japan for 24 hours
is 100 ^g/M3 and 260 yfcg/M3 for the United States.5

     There  is no  denying  the  reality  that compliance with environmental
regulations has been expensive for the domestic steel industry.  However,  it
can only  be viewed as a contributory factor to the industry's problems.  By
itself,  regulatory   spending  cannot   explain  the   industry's   declining
profitability.  Thus, while Federal policies aimed at reducing the costs of
compliance  are useful  and  appropriate -  if  they  do not  also  lead  to
significant health threats  to the public and the workers  of the industry -
they are not likely going to  lead to some swift or meaningful turnabout  in

                                     27

-------
the  performance  of  the  industry  in  general,  or  of  the companies  and
facilities with the poorest technological and economic performance.

     One   factor   generally   overlooked  when  considering  the  burden  of
complying  with   environmental  regulations  is  the  substantial   use   of
Industrial   Development   Bonds.     Such  financing  reduces  the  need  for
internally  generated  capital and  makes  large  amounts  of outside  funds
available  at low  cost and for long  periods of time.  Since State and local
governments make tax-exempt revenue bonds available to companies,  this is a
form of public subsidy for what is reasoned rightfully to be a public good.
During  the  past  decade  the   domestic  steel  industry  has used  IDB's  to
generate  nearly  half  the  capital  requirements  for  pollution  abatement
investments."

     In  the  AISI  analysis of what  the  domestic steel  industry's capital
needs are  for the next decade, the  money  needed for regulatory compliance
($800 million annually in 1978$) represents 12.3% of the total, the same as
the fraction  designated  for  diversification efforts.   In actual  fact $700
million  of this  is  for pollution abatement  (10.8% Of the  total)  and $100
million is  for meeting OSHA regulations.   An  EPA analysis  forecasts  about
$500 million  annually to meet pollution control  needs  during  this period.7
Thus, with the likely lower level of environmental spending, diversification
would outweight regulatory compliance.

Trends and Prospects for New Technology

     During  the  past decade  there have been  two trends  which undoubtedly
will continue and which reduce pollution.  These are the greater adoption of
electric furnace steelmaking and the use of continuous casting.

     The fraction of steel made in electric furnaces rose  from  about 15% to
25% during the decade of  the 1970's.  The greater use  of domestic ferrous
scrap in electric  furnaces signifies that less primary  ironmaking  in  blast
furnaces and less cokemaking  are  going on.  This shift  form  integrated to
nonintegrated  steelmaking reduces  pollution  and the  costs  of  pollution
abatement substantially.  Although there are limits to both the total amount
of ferrous  scrap  available and to the  amount  Of high quality  scrap,  there
will be a continued  increase in electric  furnace  steelmaking,  probably to
the 35% to 40% level in the 1980's.

     The  second  most important  trend is  the  greater  use  of the highly
efficient continuous  casting  method  of converting  molten steel into  solid
shapes  rather than the  multiple step  ingot  casting approach.   Continuous
casting reduces  pollution directly  because of the  elimination of  soaking
pits and  furnaces, and  indirectly because of the substantial  increase in
yield of the process.  The increase in yield means that more finished  steel
is produced from a given amount of raw steel, usually in the order of 10% to
15% more finished steel.  This means that less steel and less iron has to be
made and thus pollution is reduced.  During the past decade domestic use of
continuous casting,  although   low  compared to most  other  industries,  has
increased dramatically  from just  a few percent to Over  15%.  By the end of
the decade it should approach 40% to 50%.   Modernization  of  existing plants
will be based on retrofitting existing facilities with continuous casting to
a large degree because the many benefits  of continuous  casting,  including

                                     28

-------
reduced energy  consumption,  lead to return on  Investments  of Over 2056, and
even more for alloy and specialty steels.

     Both  electric  furnace  steelmaking and  continuous  casting  are  well
proven technologies.  Considerable attention, both here and abroad, is being
given  to the  development  of truly  new technologies  that must  be proven
technically  and economically  on a  pilot  and  demonstration  level  first.
According to the  recently completed study of the domestic steel industry by
the Office of  Technology Assessment, the most  important  new technology for
the domestic  steel industry  during the next  several decades  will be coal
based direct reduction.  Unlike natural gas based direct  reduction which is
undergoing  phenomenal  expansion throughout the  world,  coal  based  direct
reduction has been used in only a few places in relatively small operations.
Newer  forms  of direct  reduction  that  can use  cheap grades  of  coal and
possibly coal  gasification to  produce  either  conventional direct reduced
iron (DRI) or molten iron are being developed.

     For the United  States with abundant supplies of low grade  coals, the
prospect  of coal based  direct  reduction  offers  a number  of  potential
advantages.  Capital and operating costs may be substantially lower than the
blast  furnace  and coke  oven route once the technology  is  fully developed.
Most coal  DR technologies  are relatively  simple,  one step  processes  that
offer a closed  system approach with very little  pollution.   Moreover, like
natural gas based  direct reduction,  a modular rather than  economy of scale
approach can be used.  For the  cyclic steel industry with capital problems
as well as considerable uncertainties  for  demand,   foreign  competition and
government  policies,  the  ability  to  construct  relatively  small  coal  DR
modules is most attractive.

     Since the increasing use of electric furnaces will put much pressure on
ferrous scrap supplies, in  terms of quantity and quality, there  will be an
increasing need for direct reduced iron to be used as a complement to scrap.
Although DRI is already becoming a world  traded  commodity because  of the
rapidly increasing DR capacity in natural gas rich nations, such as Mexico,
Venezuela and Saudi  Arabia, other  factors  are likely  to make domestic  DR
plants economically  viable.   During the 1980's we  will  likely see several
different coal based  DR technologies both proven and adopted  in  the United
States  on  a small   scale,  possibly  at nonintegrated  steel  mills or  as
merchant DR plants serving a geographical region with limited  ferrous scrap
availability and   a  relatively  high rate  of  growth in  steel  consumption.
During  the  1990's   it  is  quite  conceivable  that  larger   scale  coal
gasification DR plants will  be constructed.   The  integrated  plant Of the
future may be based on DR rather than the blast furnace.

     Critics of DR rightfully  note that at  present large  scale DR plants
could not be justified.   However, it Is crucial to  understand  that coal  DR
technology is in its infancy and that many improvements are likely to result
from the substantial  amount of  R&D activity in  this area.   The  two  most
important  driving  forces  for  developing  coal  DR  technology  are  the
reductions in  capital  costs  and  pollution.   Creating  and  adopting  new
technology is  facilitated by  both  rapid company growth  and profitability.
In the United States there is a segment of the domestic steel industry which
satisfies these conditions.   This  is the  nonintegrated  segment,  the scrap
based minimills, midimills  or  market mills which have undergone  tremendous

                                     29

-------
growth in the past decade, from only a few percent of the domestic market to
about 15% today.  The OTA study forecasts that by 1990 these companies could
account for at least 25% of the market.  Since the growth of these companies
must depend on broadening the product mix to  include  higher quality steels,
the need  to  introduce virgin iron in the form of DRI will be critical.  And
many Of  these companies  have  already demonstrated  their  inclination  and
ability  to innovate  quickly in  new technology.   Even with  the increased
capital costs  of constructing  DR facilities, the  total capital  costs  for
combined  DRI-scrap  based  mills  could  be  less  than  the  costs  for  a
conventional greenfield  integrated plant Or  even the  costs  for  extensive
modernization of  existing  integrated plants.   For example, according to the
OTA study, the capital costs of a combined DR-scrap  plant would  likely be
approximately $500  per metric  ton Of annual  steel product capacity, about
one third that for a conventional greenfield integrated plant.

     A  number  of  other  major changes  in  technology  are  likely  to  be
developed commercially within the next decade, these include:   formcoking,
plasma  steelmaking, direct  casting  and  a host of  relatively incremental
technical improvements which when applied collectively to a particular plant
represent  a  substantial  overall  improvement  in efficiency.   Virtually of
these changes Imply reductions in pollution because they, make  use of closed
systems  and  reduce  the  dependence  on  conventional  cokeroaking,  either
directly or indirectly by improving efficiency and yield  in the steelmaking
portion of the mill.

     Contrary to  some concepts popularized  about the  steel industry, iron
and steelmaking  technology is  far from static.   Technology  is  a  problem
solving  tool  that  bold  and  risk  taking  managers  can use  to  deal  with
problems of limited capital availability, rising labor  costs,  environmental
regulations and raw material  constraints.   Future steel demand is uncertain
and any increase  in demand  will be small.   Nevertheless,  steel will  remain
an  absolutely critical  material  for  all  societies.     Recent  oversupply
conditions  have  taught   valuable   lessons   to  most   steel  industries,
particularly those  in Europe.   There is a good possibility of a close match
between world steel supply and demand leading to higher  profits in the mid-
1980 's.   It  is this perception by many  people both in the industry now and
examining the industry from afar that is stimulating the development  of new
technologies,  particularly those  such as direct  reduction that  will make
entry into the industry easier.  All these changes, including  the influx of
increasing amounts of foreign capital into the domestic steel industry, will
likely lead to reduced environmental pollution in the years ahead.

Federal Policies Affect Modernization

     Although Federal policies concerned  with the environment may relieve
some of the immediate problems of capital availability,  they are not  likely
to serve  as  a major stimulus for modernization based on the most innovative
technologies or  for modernization  of plants  which are  already  relatively
efficient.

     Policies which more  directly deal with capital formation,  RD&D,  and
prices  have  greater  ability  to  influence  modernization  based  on  new
technology.  With the exception of the,growing nonintegrated segment Of the
industry   and   the   highly  efficient   and   competitive   alloy/specialty

                                     30

-------
steelmakers,  the absence  of a  coordinated set of  policies for  the  steel
industry  which  leads to  a substantial  increase in  profitability  for  the
large  integrated companies will likely result  in the continued contraction
of  this  segment of the  industry.    If  the  nonintegrated  segment  can
compensate  for  the  loss in capacity of the integrated companies, Or even if
greater levels  of  imports  result,  the net effect  for  the. nation  will  be
reduced  pollution.    There  are,  of  course,  a  number of  undesirable side
effects of  increased  imports,  including loss of employment  and  a threat to
our  national  security  should  imports  rise  above   about  20%  of domestic
consumption.

     The important issue is not how to reduce pollution by reducing domestic
steelmaking capacity and increasing our dependence on foreign steel.   It is
how  do we  both decrease pollution and maintain competitive and profitable
steelmaking capacity.  The key here is a combination Of Federal policies and
changes in  the  industry itself which lead to the greatest  use  of the most
innovative  technology by the  best  managed companies rather than policies
which tend to protect the poorest managed and performing companies.

     If one object were to insure a viable  competitive  steel industry, then
the  President's idea of three year extensions for some plants could aid the
wrong  facilities,  and  the companies  that  have  delayed  compliance  and
modernization  could be  rewarded.   There  are,  after all,  a number of oH
inefficient and  poorly located steel mills that are  not truly competitive.
The  spectre of concentrated losses  of jobs  from  plant closings could be
dealt with  by  appropriate worker retraining and relocation  prograns.   Like
other  policies,  such  as  the  recent  EDA  loan  guarantee program,  this
regulatory  approach  tends  to give advantage to  those firms with  the  least
competitiveness,  rather than rewarding those better  managed companies that
would benefit from incentives for still greater  risk  taking and investments
in  the future.   Greater  application  of the bubble concept  appears to be a
more acceptable approach to meeting the objectives of reduced  pollution and
freeing more money for modernization by companies.

     The skewing of competitiveness by Federal policy is illustrated by the
situation of  Inland,  Bethlehem  and U.S.  Steel,  all with  large integrated
plants in the Chicago region.  By investing in steelmaking, using innovative
technology  and  having good strategic planning and management, Inland became
the  lowest  cost  domestic  producer  (or large  integrated  steelmakers)  and
reduced their  pollution.   But Bethlehem and U.S. Steel plants serving the
same market are far less efficient and, if President Carter's  proposal is
accepted  by Congress,  they  could  use  capital  to invest  in  modernization
rather  than pollution  abatement at  a  time when they  will  receive  even
greater tax benefits from modernization capital investment  than Inland did
some time ago.   Interestingly, the extension proposal asks  only that "funds
which  would have been spent to comply with the deadline will be expended in
the  same time period for modernization," but there is little indication that
such a  level   of  spending  would  normally  be  sufficient  to  modernize
qualifying  plants to  a significant  degree.  Moreover, the  maximum of  three
years  for extension is not long enough to plan and implement major forms of
modernization.    Cash  flow from  such  plants  could  still  be  used  for
diversification  out  of steelmaking.  Federal policy, such as the three year
extension,  therefore,  gives Inland's  higher cost competitors  an advantage,
but  in the long run it  cannot  make these less efficient  and  more poorly

                                    31

-------
managed plants independently competitive.

     Moreover, Federal  policies need more comprehensiveness.   For example,
although  modernization  is  a  worthy goal  and new technologies  may indeed
reduce air  and  water pollution, we  know that  in many cases  the pollution
will merely be  shifted to the  solid  waste category.   The proposed policies
may demonstrate  to companies that they  are  better off in the long  term by
fighting  compliance  with  RCRA  and  the  regulations affecting  hazardous
wastes, and by using available capital for purposes such as diversification.

     What is  needed are incentives for developing, testing and adopting new
technologies  that offer less  pollution of  any  kind  (including  noise  for
example)  together  with sufficient   economic  benefits  to  justify  private
investment.    Rather   than   tying   delayed  spending  on   compliance   to
modernization, it may  be useful to consider linking  it to high risk R&D on
innovative steelmaking (including pilot and  demonstration  activities).   TWO
reasons support this point of view.  First,  the level Of R&D spending by the
industry as a whole and by individual companies  is closer to the level of
spending on compliance with  environmental regulationsj that is, hundreds of
millions of dollars annually for the industry  for both cases, rather  than
billions of dollars needed every year for capital spending on modernization.
Thus, the impact would be reater on  R&D than on  modernization.   Second, as
the OTA study showed,  a relationship between environmental capital spending
and R&D spending exists for the past decade.  For example, from 1969 to 1973
the average  ratio of  R&D  to environmental  capital spending  was  unity,  but
from 1974 to  1978 the  average ratio  dropped to one-half.   The increase in
environmental capital  spending  from  the earlier period to the later one was
from an average  of $157 million  to  $440  million annually.   Environmental
capital  spending  appears  to  be  influencing  the  discretionary  use  of
corporate funds on R&D in the domestic steel industry.  R&D  spending  is  not
linked to total capital spending by the industry.

     This policy approach of fostering R&D rather than modernization appears
to have  the  disadvantage of shifting the  use Of  funds  from the plant to
corporate level.   However,  by promoting  intrinsically  profitable,  efficient
and clean technology,  it has the potential  to  foster  legitimate,  long term
industry competitiveness.  Using the bubble approach to deal with short term
needs  and  the  freeing  of  capital  for  modernization,  and the  compliance
schedule extension for supporting long term R&D, federal policies could help
the domestic steel industry in a fair and comprehensive manner.
                                    32

-------
References

1.   "A   Program   For  The   American  Steel   Industry,   Its  Workers  and
     Communities," The White House, September 30, 1980.

2.   D. F.  Barnett, "The  American Steel  Industry  in the  1980's:   Capital
     Requirements  for  Modernization,"  The  Atlantic  Economic  Conference,
     October 12,  1979,  Washington, D.C.   (note:  data reported are capital
     expenditures  for  productive  steelrriaking  facilities;  these have  been
     used  with  AISI  data  On  total  capital   spending  and  environmental
     spending to obtain the nonsteel spending.)

3.   W.  K.  Hall,  "Survival  Strategies In A Hostile  Environment,"  Harvard
     Business Review, September-October, 1980.

4.   H. Mueller and K.  Kawakito, "The International Steel  Market:   Present
     Crisis and Outlook for the  1980's," Middle  Tennessee State University,
     1979.

5.   A.  Mukaida,   "Environmental  Control  Measures  in  the  Japanese  Steel
     Industry," Japan Steel Bulletin, vol. M, no. 2, 1980.

6*   Technology and Steel  Industry  Competitiveness,  Office  of Technology
     Assessment, Washington, D.C., June, 1980.

7«   The  Cost  of  Clean  Air  and  Clean  Water, Report  to  Congress,  EPA,
     Washington, D. C.  1979.
                                    33

-------
Session 1: AIR POLLUTION ABATEMENT

Chairman:   Thomas J. Maslany
           Region HI
           U.S. Environmental Protection Agency
           Philadelphia, PA
                     35

-------
Title:    A MODEL FOR COKE OVEN CHARGING EMISSIONS

Author:   William B. Tucker
          Assistant to the Director
          Environmental Control Department
          Republic Steel Corporation
          P.O. Box 6778
          Cleveland, Ohio  44101

Abstract: This paper discusses coke oven charging emissions from well
          controlled coke oven batteries.  They are found to follow a
          statistical distribution that is badly skewed toward long
          times, although the mode and mean times are short.  Causes
          of occasional long emission times are explained.  It is
          suggested that charging emission standards of the type that
          permit a limit of L seconds visible emission per N charges,
          such as 125 seconds per five charges, will inevitably be
          violated and are inappropriate.  The paper suggests that the
          standard should be a limit on per cent of visible charge
          emissions exceeding 25 seconds.  The limit should depend on
          battery age and features.  The paper suggests compliance be
          determined from an average derived from a month or more of
          observations.
                                   37

-------
                A MODEL FOR COKE OVEN CHARGING EMISSIONS
I.    Introduction
      The author has participated in several  dialogues of the following
      type:
      Agency Person:  "The standard for coke  oven charging emission should
                      not exceed 125 seconds  for five successive charges."
      Industry Rep:    "That is too strict.  It can't be done."
      Agency Person:  "It is being done and it is actually a generous
                      standard.  We have data to show batteries meeting
                      35 seconds for five charges."
      Industry Rep:    "But you ignored other  data that showed the same
                      battery exceeding 125 seconds for five charges."
      Agency Person:  "That was because of equipment breakdown or human error."
      Industry Rep:    "But there are always breakdowns and human errors at
                      coke ovens."
      The participants in this dialogue are looking at the same animal from
      two different  perspectives and are focusing on two different extremes
      of its possible behavior.  I believe the divergence in viewpoint stems
      from the two parties having different "models", if you will, of the
      behavior of coke oven charging emissions.  I believe the way to
      progress toward the goal of an equitable standard for coke oven charging
      emission performance will be to first find the right model, one that
      faithfully describes its behavior.  Then one can determine useful
      things about measuring and controlling  coke oven charging emissions.
II.    Data
      A.  Data Sources
          I  have collected data on visible emissions during charging from
          ten different batteries operated by five different steel companies
          in five different geographic areas.  This was not a random sample.
          These are  not average coke batteries from the standpoint of charging
          emission control.  Nine of the ten  were the best performing batteries
          out of an  aggregate of about 90 coke batteries collectively owned
          by their companies.  The tenth battery is an old one which was
                                   38

-------
    included to give some perspective to the performance of the
    other nine.  Most of this data was measurements of charging
    emissions made on a self-monitoring basis by representatives
    of the individual companies (not necessarily the battery
    operators) who were required to submit monthly or quarterly
    reports to state or local agencies.  I don't believe there was
    any tendency to overemphasize long emission times because these
    reports of long emission times are self-reported failures to
    meet regulatory requirements.  I picked batteries where much
    effort has gone into equipping the batteries properly and
    training the operators to perform good stage charging.  Among
    these are two which U.S. EPA describes as "exemplary" for stage
    charging control.  Several others come very close to the exemplary
    batteries in their performance.  One old battery does not.  Even
    this old battery performs better on charging emissions than many
    others, including some much newer batteries.  In this paper I
    have called them batteries 1, 2, 3, ... 10.  (See Table I).
                   Table I - Battery Characteristics
               Height   Collecting
               Meters     Mains
    Battery 1
    Battery 2
    Battery 3
    Battery 4
    Battery 5
    Battery 6
    Battery 7
    Battery 8
    Battery 9
4
3>s
3%
6
4
4
4
3%
3*5
3J<
1
1
1
2
1
2
1
2
2
2
Period Observed      Notes
9/12/79 to 4/2/80
11/1/79 to 6/30/80
11/1/79 to 6/30/80
11/1/79 to 6/30/80
11/1/79 to 6/30/80
6/1/79 to 11/30/79    (1)
1/1/77 to 2/12/80     (2)
1/1/80 to 4/30/80
1/1/80 to 4/30/80     (1)
1/1/80 to 4/30/80
    Battery 10
    (1) Battery described by EPA as "exemplary" for charging emissions
    (2) Battery 7 is more than 35 years old
B.  Histograms ;and Cumulative Frequency Plots
    I have plotted two types of histograms of the charging emission
    times.  When data on individual charges were provided, (all
    batteries except No. 6), I plotted histograms of individual charge
                             39

-------
    times (see Figure I).  For example, in Figure I the histogram
    for Battery 1 is constructed from 745 observations of individual
    oven charges.  When data were available as cumulative time, i.e.,
    emission time cumulated from four to seven successive charges,
    I plotted histograms of average charge time (see Figure II).
    For example, in Figure II the histogram for Battery 1 is constructed
    from 149 averages of the cumulated sum of emissions from five
    successive charges.  Therefore, Figure II is arrived at as if
    149 compliance tests had been performed at that battery.  Also
    I plotted the individual oven charge emission data on a cumulative
    frequency plot where the per cent of observations less than a
    certain time is plotted on a normal probability scale against
    the logarithm of the time (see Figure III).
C.  Characteristics of the Histograms
    Each battery has its own unique histogram.  However, as a class,
    the histograms have common features.  They are badly skewed
    distributions with the skew toward the long time.  The mode
    and mean times are short, mainly in the range of 5 to 25
    seconds in the individual charge histograms.  However, each
    of the batteries produced some very lengthy times exceeding ten
    times the mean.  In the average time histograms the distribution
    is not as badly skewed because the cumulation of four to seven
    consecutive emission times tends to average longer and shorter
    times.  Even so, on each battery there are a few observations
    that exceed four times the mean.
D.  Characteristics of the Cumulative Frequency Plots
    In Figure III Ithe data from all ithe batteries seemed.to be
    reasonably representable by a straight line.  However, each
    battery has its own line and the slopes of the lines are
    not all equal.  Thankfully the short charge emissions greatly
    predominate.  But the lengthy emissions are also there, and
    from Figure III one can determine the percentage of them.
    The following performance characteristics were determined
    from the data.  (See Table II).
                             40

-------
                             HISTOGRAMS  OF  INDIVIDUAL  COKE OVEN  CHARGE  EMISSIONS
                  COKE OVEN CHflRClNC DflTfl
                         BflTTERY I
        k,

-------
                                                        FIGURE  1
                            HISTOGRAMS  OF  INDIVIDUAL COKE  OVEN  CHARGE EMISSIONS
                  COKE OVEN  CHflRGING DflTfl
                          BflTTERY 7
                  COKE  OVEN  CHflRCINC OOTA
                          BATTERY 9
                                            -*a-i	R—
00    30.0    40.0    10.0    M.O    100.0   190.0   IW.O   m.o   >!•••

ou uiom I me.         MCOMK MR CHABM
0.0    10.1
                                    IM.I   tm.t   MD.I   •!»'
                  COKE OVEN CHflRGINC OflTfl
                         BflTTERY  8
                                                                 1 «l.
                                                                 nn.
                                                                 10.0,
                  COKE OVEN CHflRCINC OflTfl
                         BflTTERY 10
0.0    JO.O    «0.0    M.O    10.0    100,0   UO.I
                                                100.0   >1«0.0
                                                                                            mo    to.*   imi   Mat   •>.•
                                                                    OBi. lllfIN i.< BE.
                                                             42

-------
                                 HISTOGRAMS  OF  AVERAGE  COKE  OVEN  CHARGE  EMISSIONS
COKE  OVEN CHflRCINC DflTfl
                                                                                           COKE OVEN  CHflRCINC DflTfl
na.'
90.0,

ii. a



13.0.

1.0.

*.o

0.0













r
rl























-
















-




u.














1










































—













































































BflTTERY 1











-j


in.
nu.
10. a.

mean »•"
OO0NC
M xeao 10.4

ii. a

ii.a

1.0.
n
M.O n-0 >M.O














U 1

















_



















_







BflTTERY 4



nun
OEX01MB
MttCMM 1 •






f^^
10.0 ».o 10.0 40.0 m.0 M.O mo »-o >«.«
   "^ KWIM ».»sic.
o.o    jo.o    ta.o    u.o    10.0    TO.O
                     iMpua atcomrai i
COKE OVEN  CHflRCINC DflTfl
       BflTTERY 5
                                                                                               bQ=,
                                                                         °-°     >"'a    M-°    10-0    *•"    *•*    "••    M-0    •5-°    *•>••
                                                                         cm. uwm 2.s au.
   .
mo.
 Ka
 ii.a
 IJ.O,
      !
COKE OVEN  CHRRCINC DflTfi
        BflTTERY 3
                                                    M KCIMB 7. 1
         '0.0    M.0    M.O    40,0    M.O

   1111 »imt j.i sir.
                                        M.O    TO.O    n.o
                                                                      >«..
                                                                      f«£0.
COKE OVEN CHflRCINC DflTfl
        BflTTERY 6
M
M.a
ie a
1.0.
0.0.


J
-


roam
DCDMn


0.0


1 i 1 	 1 	 1 	 1 1 1 	 —
10.0 M.0 u.o to.o RLI la.o n.0 m.o -to.a
ecu. NIOIN >.s K. MUM *nx» m CIMOC
                                                                 43

-------
                                                          FIGURE  II
                                HISTOGRAMS OF AVERAGE COKE  OVEN CHARGE EMISSIONS
                     COKE  OVEN  CHflRGINC DflTfl
                            BflTTERY  7
COKE OVEN CHflRC-INC- DflTfl
       BflTTERY 9
                                                 fOCUT
                                                 EXXOIINC
   0.0    10.0    20.0    10.0    40.0    50.0    10.0


   CCU WIDTH 2.9 XC.
                                          in.n
                                            W.O   M.O    -»-0
                                                                nra.
                                                                 •o.o.
                                                                   0.0    10.0    20.0    30.0
                                                                   COL UIOIM 2.s sc.
                                                                                                            n.0   m.a    ••>.'
««tL.
mo.
 so.o.
                    COKE OVEN CHflRCINC DflTfl
                           BflTTERY 8
 COKE OVEN CHflRGINC DflTfl
       BflTTERY 10
                                                 M 9CCOMK 1.0
   0.0     10.0    20.0    10.0    40.0    50.0   M.O    10.0    10.0   >«O.U
   COl UIOIH 2.S ICC.
                                                                   o.o    IO.D    90.0    10.0
                                                                     1 UIOIH 2.S 9C£.
                                                                                                so.o    10.0    n.0
                                                             44

-------
                             FIGURE III


CUMULATIVE FREQUENCY -  PER CENT OF CHARGE EMISSIONS SHORTER THAN OBSERVED TIME
    oz
    Z HI
    «o o
    IM O
    gw

    <*
    3 ui

    5s
    _iO
                                                            /to
              INDIVIDUAL CHARQE EMISSION TIME - SECONDS
                                 45

-------
Collecting
Mains
1
1
1
2
1
2
1
2
2
2
Average
Charge-Sec.
23.5
9.7
11.2
8.3
16.0
10.0
42.5
5.7
4.2
5.2
% Exceeding
25 Sec.
31.6
6.2
9.3
3.6
14.3
3.2
83.3
0.8
0.4
0.8
% Exceeding
34 Sec.
19.4
4.4
7.1
1.8
6.7
1.0
65.0
0.6
0.4
0.6
               Table II - Average Emission Time Characteristics
    Battery 1
    Battery 2
    Battery 3
    Battery 4
    Battery 5
    Battery 6
    Battery 7
    Battery 8
    Battery 9
    Battery 10
E*  Discussion of Charge Time Distributions
    For a coke battery to meet a standard of 125 seconds visible
    charging emission per five consecutive charges, it would have
    to average less than 25 seconds pep charge, and similarly to
    meet a standard of 170 seconds per five charges, which is the
    standard in two important coke industry states, it would have
    to average less than 34 seconds per charge.  In Figure I and
    Table-II it is apparent that all the batteries in this survey,
    except Battery 7, average well under 25 seconds per charge.
    However, in Figure I it is apparent that every battery also
    produced some long emissions.  In Figure II the emission times
    were cumulated for 4, 5 or 7 charges, whichever was required
    by the local  regulation.  Then the cumulated time was divided
    by the number of charges to get the average charge emission
    length.  This figure shows that on all batteries except 8 and
    9 a not inconsiderable number of inspections would have been
    failed.  For  example, exemplary battery 6 exceeded 25 seconds
    average emission time in 3.2% of the observations and 34 seconds
    in 1.0% of the observations.  Even batteries 8 and 9 had a
    rare long emission that would have caused failure of the in-
    spection.  None of these well controlled batteries would
    always have passed a compliance inspection.
                             46

-------
          Although none of the  batteries  in  this  survey is  brand  new,
          several  employ all  the  equipment that is  presently believed
          to be LAER technology.   The  data suggest  any battery using
          this  technology will  probably not  be  able to consistently
          achieve  the LAER standard of 55 seconds per  five  successive
          charges  visible emission.
      F.   Effect of Battery Physical Features on  Change Emissions
          In this  survey the best performance is  achieved by Batteries
          6, 8, 9  and 10 which  are short  batteries  with double collector
          mains.  The 6 meter battery  (No. 4),  which also has double
          collector mains, performs almost as well  as  the short batteries
          with  double collector mains. Batteries 1, 2, 3,  5 and  7 which
          have  single collector mains, perform  well on the  average, but
          are not  as consistent as the double main  batteries.  They
          have  more long charge emissions than  the  double main batteries
          have.
III.   The Coke  Battery Operator's Problem With  Chance
      A.   The Equipment Problem
          If everything is working well,  the steam  aspiration is  strong,
          the "tunnel head" space above the  coal  is open, and inleakage
          through  oven openings is controlled,  then the charging emissions
          will  be  brief or even entirely  captured within the oven.
          However, a typical battery has  thousands  of  parts that have
          some  effect on charging emission time.  All  of these parts
          must  function well for  every charging emission to be captured
          and held within the oven.
      B.   The Personnel Problem
          In addition to the equipment having  to  work  right, the performance
          depends  on people, to operate the  equipment  in the right sequence,
          to carry out every step correctly, and, to maintain the equipment
          diligently.
          Automation can eliminate some human errors,  but automatic equip-
          ment is  very vulnerable to the  severe heat,  cold  and moisture
          plus  the pounding it will receive at the coke battery.  Therefore,

                                   47

-------
          machinery that  is too automatic seldom works.  The optimum
          amount of automation on a coke battery is far less than It
          would be in a more temperate location.  The optimum automation
          for a coke battery still relies heavily on human operators.
      C.  The Long Emission Time
          There are some  types of equipment malfunctions which should
          be apparent (but not necessarily correctable immediately)
          before a charge begins, but others are not.  For example, a
          misaligned drop sleeve should be apparent.  But, a shorted
          electrical wire might not be noticed before there is a loss
          of electrical function that has already resulted in a long
          charge emission.  There are also "malfunctions" caused by
          changes of the flow properties of the coal.  Sometimes coal
          moisture or other characteristics can change the coal flow
          behavior enough to result in a long emission time.  Freezing
          weather can prevent the coal from flowing freely as needed
          for good stage charging,  These coal flow problems are likely
          to cause a string of long charge emissions.
      D.  Operators Viewpoint
          Taking all of these factors together, unpredictable equipment
          malfunctions, unpredicat&ble human errors, and variable coal
          properties,, the operator usually doesn't know In advance when
          a long charge emission is coming.  But, he knows by the law
          of averages that some long charge emissions are going to come.
          Consequently, when an Agency Persons says you may not exceed
          125 seconds for five charges, the operator feels it is an
          inequity.  It is like demanding that a baseball player bat
          .300 in every ball game.  A good player can bat .300 for a
          season, but he knows he can't do it in every game.  Similarly,
          a good coke battery operator can average less than 125 seconds
          for five charges but he will go over 25 seconds on a significant
          fraction of the charges.  Occasionally just one charge will
          exceed 125 seconds.
IV.    Emission Quantity Versus  Allowable  Emission Time
      A.  Operators Want Zero  Emissions
          An agency engineer testified in a  public  hearing that shortening
                                  48

-------
    the  emission  limit from 170  seconds  per  five  charges  to  125
    seconds  per five  charges would  reduce coke  oven  charging
    emissions  by  45/170ths  or 26%.   This would  be true  only  if
    the  emission  on each  charge  occurred for the  full time
    allowed  by the standard.  This  in turn implies a degree  of
    control  over  charging emissions which does  not exist  in
    practice.   In reality the operators  think this way: "A bad
    charge can happen at  any time,  and when  it  occurs it  is
    likely to  use up  our  whole allowable emission time.  Therefore,
    we had better make each charge  emission  as  close to zero as
    we can."
    Figure II  corroborates this. These  batteries are required
    in one case to average under 34 seconds, some to average
    under 19 seconds, and one to average under 11 seconds.   All
    but one do much  better than  required.  These batteries  are
    not just barely meeting their allowable emission.
B.  Comparison of 170 Seconds to 125 Seconds Standard
    Operators  I have  questioned  felt that for all practical  purposes
    a standard of 170 seconds holds emissions down as well  as a
    standard of 125  seconds per  five charges.  They must use the
    same equipment,  work practices, and training to attempt to
    meet either standard.  As discussed in the preceding paragraph,
    there won't be any difference in the average charge emission
    time.  The only difference is that the battery will violate
    the 170 second standard somewhat less frequently.
Recommending a More Equitable Kind of Charging Emission Standard
A.  The Batting Average Measures Ability in Baseball
    Returning to the baseball analogy, baseball fans recognize
    that the all around best measure of a player's batting ability
    is his medium to long term batting average.  They  realize
    that hitting a baseball under game conditions is a statistically
    variable function.  Even though the player yearns  to hit,
    he cannot do it every time or even once in every game.  A
    single game batting average is not a reliable measure of
    batting ability.  A manager who punished players for failure
    to hit  in each game would be thought capricious and would
                             49

-------
    probably provoke a confrontation with the players union.
    Similarly, there is no :way humanly possible to eliminate
    long charge emissions from a coke battery.  Although the
    operators know that with good equipment, good maintenance
    and careful attention to work practices they can increase
    their chances to pass inspection against a limit of 125
    seconds for five charges, still in their perception it is
    inevitable that there will be violations.  If they are to
    be cited for violations, they perceive this" to be inequitable
    and they find themselves in a disagreeable position.
B.  A Batting Average for Coke Oven Charging Emissions
    I suggest that a different kind of charging emission standard
    is needed.  It should be more in the nature of a medium to
    long term batting average and not have the nature of a single
    game batting average,  A good coke battery operation will
    have the great majority of its charging emission time short.
    But even the best operated batteries also will have unpredictable
    occurrences of long charging time.  The coke battery has to be
    allowed occasional bad charges, and even a string of bad
    charges.  My suggestion is purely a personal one.  This 1s
    not an industry opinion nor a Republic Steel opinion.  Hy
    suggestion is that the charging emission time for single
    charges ought not to exceed 25 seconds more than X% of the
    time.  The data suggest X could be larger for single main
    batteries than for double main batteries,.and larger for
    tall batteries than for short ones.  In general, X should
    take into account design features, condition, age and
    remaining life of the battery.  The very select batteries
    studied in this paper (with one exception) had X' in the range
    of 5 to 35.  One must bear in mind that whereas Table II
    deals with average emissions, the X I am suggesting would
    be derived from Individual emissions and might be larger
    than the values in Table II.  For example, on battery 5,
    while 14.3% of the daily averages exceed 25 seconds, 17.2%
    of the individual charge emissions exceed 25 seconds.  Additionally,
    X should be adjusted upward or downward according to the number
                           50

-------
          of observations  fri the sample.   It  should be larger for 50
          observations than for 500 observations.
          I  believe the "batting average"  for determining compliance
          should be based  on at least a month of data (preferably a
          quarter)  where five consecutive  charges are timed daily,
          five days a week.  This,  of course, would require the company
          observers to inspect themselves  because no agency can afford
          the manpower to  inspect so frequently.  But, I  believe the
          experience already obtained in several states with self-
          reported  coke battery observations  shows the majority of
          company observers to be trustworthy.  I believe the agencies
          will easily discover which batteries can seldom pass inspection
          when agency personnel are present to observe.
          All Around Ability Versus Batting .Average
          As important as  hitting is to baseball, there are other abilities
          that a player needs.  Since no player is equally as strong at
          all facets of the game^ a manager may be willing to forego a
          certain amount of hitting ability to get outstanding fielding
          or base running.  Though I have not studied door leaks and
          topside emissions as thoroughly as I have studied charging
          emissions, I have data (not presented in this paper) that
          indicates exemplary charging emission control seldom combines
          with exemplary door leak control and topside leak control on
          the same  coke battery.  There may be interactions between one
          and another.  Agency people may have to adapt a pragmatism
          toward balancing the various facets of coke battery emissions
          control.
VI.   Conclusion
      This paper discusses the problem of establishing an equitable and
      effective coke oven charging emission standard.  The solution depends
      on having a model that describes accurately the behavior of coke
      oven charging emissions.
      Data collected by the author from ten of the best controlled coke
      oven batteries in the industry, two of which are considered exemplary
                                   51

-------
batteries by U.S. EPA reckoning, are skewed distributions.  These
data imply that even the best controlled batteries will have some
charges with long emission times.  Nevertheless* the average charge
emission time is short.
The inappropriateness of present standards of the type which limit
visible emissions to L seconds per N charges is discussed.  This
paper suggests that an equitable emission standard ought to permit
occasional long charge emissions.  It is recommended that the
standard measure performance over a month or more of inspections
and require that no more than X% of the individual oven charges
exceed 25 seconds emission time.  For the very select batteries
studied in the paper, X could be in the range of 5 to 35 depending
on the features of individual batteries including the number of
collection mains, and whether the battery is short or tall.  Also
for statistical reasons X should be adjusted to accord with the
number of observations in the sample.
                           52

-------
             The SCAT System for Fugitive Particle Emission Control

                                       by
                  Shui Yung, Richard Parker and Seymour Calvert
                         Air Pollution Technology, Inc.
                                       and
                                 Dennis Drehmel
                                   U.S.-E.P.A

                                 A B S T R A C T
     The Spray, Charging and Trapping (SCAT) scrubber system is unique
fugitive emission control system being developed by Air Pollution Technology,
Inc.  It has many potential applications in the Iron and Steel industry
including major sources such as coke ovens and blast furnaces.  The SCAT uses
air curtains and push jets to contain, divert and convey the fugitive emissions
into a charged spray scrubber.
     Experiments were performed on an 8,000 CFM bench-scale spray scrubber to
verify the theory and demonstrate the feasibility of collecting fugitive
particles with charged sprays.  The effects of charge levels, nozzle type, drop
size, gas velocity, and liquid-to-gas ratio were determined experimentally.
The experimental data and theoretical predictions are presented in this paper.
     A prototype SCAT system was built and tested on a crosswind and on a hot,
buoyant smoke plume.  Theoretical predictions and experimental data are
presented.
                                    53

-------
 INTRODUCTION
     Controlled disposal is the permanent way to control fugitive particle
 emissions because it prevents the redispersal of particles.  This technique
 involves gathering and conveying the fugitive emissions to particulate control
 devices.  Present methods used to contain fugitive emissions, such as total
 Gilding enclosure and evacuation, or secondary hooding at the local source of
 emissions, are ineffective and costly.
     The SCAT (Spray, Charging/and Trapping) scrubber system is a novel,
 controlled disposal scrubber system for controlling industrial process fugitive
 emissions.  It uses air curtains and/or air push jets to contain, divert, and
 convey fugitive particles into a fine particle scrubber.  The SCAT scrubber
 uses charged water sprays for removing particles entrained in the gas stream.
     A schematic drawing of the SCAT system is shown in Figure 1.  The SCAT
 system uses four major components:  (1) air curtain or air push jets,
 (2) particle charger and charged water sprays, (3) entrainment separator, and
 (4) water treatment and recycling system.  There is no one fixed design or
 configuration for the SCAT and it is not necessary for a SCAT to have all four
 components.  Its design varies from source to source.
     The most important features and advantages of the SCAT system are listed
 below.
     1.  Inexpensive, simple design.
     2.  Portable, does not interfere with process equipment.
     3.  Minimum use of solid boundaries enabling access to the source.
     4.  Minimum use of ducting and hooding.
     5.  Deflects crosswinds.
     6.  Contains hot buoyant plumes.
     7.  Enables controlled disposal of particles.
     8.  Flexible design to suit individual  process.
     9.  Energy efficient fine particle collection.
     The purpose of this work was to evaluate available engineering design
models for the SCAT system.   Specifically this included evaluating air curtains
and charged spray models.   Although air curtains and  charged sprays have been
                                      54

-------
                                   AIR
                                 CURTAIN
en
en
PUSH FAN
   OR
PUSH JET
FUGITIVE
EMISSION
                                   AIR
                                 CURTAIN
                                                   SPRAY
                                                 SCRUBBER
                                       ENTRAPMENT
                                        SEPARATOR
                                       Figure  1.   Example  of  SCAT  system  arrangement.

-------
studied and used in industry for years, there are surprisingly few useful design
models and few studies reporting sufficient data to evaluate these models.
     This work was broken into two phases: the air curtain study and the
charged spray study.  The goal was to arrive at an understanding of the SCAT
system which was sufficient to enable the design of a SCAT system for any
specific industrial installation.  Details of the engineering models and
calculations are reported by Yung, et al (1980).

AIR CURTAIN STUDY
     An air curtain is a sheet of air blown out of circular or rectangular
nozzles at high speed.  Air curtain design for the SCAT system requires
information on several parameters of the jet stream including the velocity
distribution, jet expansion angle, the air entrainment ratio, and the effects
of hot buoyant plumes and crosswinds.  The jet expansion angle determines the
overall cross-sectional dimensions of the receiving hood.  The air entrainment
ratio determines the volumetric flowrate of the gas to be cleaned.  The cross-
wind and/or buoyant plumes dictate  the placement of air curtains and the
receiving hood.  Design models for these parameters were evaluated experimentally
in this study.
Air Curtain Manifold
     The air curtain distribution manifold should give a uniform discharge
velocity distribution along the manifold length to perform properly.  In
addition, the discharge direction should be as close to perpendicular to the
flow direction in the manifold as possible.
     To obtain a uniform discharge, a constant static pressure must be maintained
along the duct.  This was done by using a tapered duct to counterbalance static
regain in the manifold.
     The discharge angle was maintained near 90° by using a continuous slot
nozzle divided by deflector vanes which protruded 9 inches from the air
curtain duct.  This is illustrated in Figure 2.
Jet Expansion Angle and Air Entrainment Ratio
     The jet expansion angle and air entrainment ratio were determined experi-
mentally.  They were calculated from velocity distribution measurements at
several downstream locations from the manifold.  The calculated jet expansion
                                     56

-------
                  AIR INLET
      1-8 cm
                                   DIAGONAL
                                    PLATE
30 cip
                         36 cm
Figure' 2.  Air curtain distribution duct.
                  57

-------
angle was about 25° which is identical to that predicted for a pure momentum
jet, but it is smaller than the 30°-40° reported in the literature (Juve and
Priester, 1944 and Hemeon, 1963).
     The measured jet stream center!ine axial velocity decay agrees with that
predicted from Prandtl's eddy viscosity model for turbulent flow as shown in
Figure 3.  The total air flow increases with distance from the air curtain
manifold because the air curtain jet entrains surrounding air.  The air
entrainment ratio was measured and the data agree fairly well with predictions
as shown in Figure 4.
Crosswind Experiments
     Experiments were carried out to evaluate the effect of crosswinds on the
trajectory of an air curtain.  The air curtain jet may be used either to deflect
the crosswind so that it bypasses the scrubber or to entrain and deflect the
wind and fugitive emissions into the scrubber downwind from the emission source.
In either case the design of a specific installation will require the ability
to predict the momentum balance between the wind and air curtain and thus the
resultant air flow trajectory.
     Wind deflection depends on several SCAT operating parameters such as the
distance between the air curtain and the scrubber, air curtain slot width, slot
exit velocity, the incidence angle between the wind and air curtain jet stream,
and the wind speed.  The incident angle at which the air curtain meets the wind
is the most important parameter.  For maximum deflection, the jet stream should
be at 45° opposing the wind.
     The purpose of these experiments was to determine the requirements for
deflecting crosswind and to verify published correlations.  For the SCAT
system, complete wind deflection is required to avoid disturbances in dust
containment.  Complete wind deflection is defined as the condition for which
the resultant air flow of the crosswind and the SCAT air curtain bypassed the
SCAT scrubber.
     Several formulae for correlating the parameters mentioned above are
available in the literature.  Instead of fixing the incident angle at 135°,
which gives the maximum range, wind deflection experiments were performed by
fixing the range (distance from air curtain to scrubber), wind speed, air
curtain slot width, and slot exit velocity and varying the Incident angle for
                                     58

-------
in
to
          1.0
       § 0.5
       UJ
          0.2
           0.1
               10
                                     I     I     I    I   r   IT
                                .PREDICTION
20
J	1	1   I   I  I I
40  5010
                                         I
                                                         I     I    I   I   I  I  I _
                                                               SLOT WIDTH
                                                                 1.3 cm
                                                                 2.5 cm
                                                            D   3.8 cm
30     40  50            100           200
       AXIAL DISTANCE/JET WIDTH RATIO,
                                                                             I  i
                                                                                                             TOGO
                               Figure 3.  Measured and predicted center!ine velocity decay.

-------
    100
                 I      I    I   I  I I I I
    50
A
a
o
    20
    10
a:
       10       20
                                     III  I  I  I I l_
                            FROM PRANDTL
                            VISCOSITY
                        MCELROY
                                               ABRAMOVICH
                 t     I   I  I  Mill	I     I    I   I   I  II  I
                 50        100     200

            AXIAL DISTANCE/JET WIDTH,  m/m
500        1000
           Figure  4.   Measured  and  predicted air entrainment ratio.
                                    60

-------
wind deflection.  The measured angle was then compared with predictions.  It
was found, as shown in Figure 5, that the measured angle agrees with that
calculated from theory (for details on the theory see Abramovich, 1963).
Thus the deflection of a crosswind can be accurately predicted from the air
curtain design and operating parameters.
Hot Source Experiments
     In many metallurgical processes very hot fugitive plumes containing high
concentrations of particles are released.  The most efficient and economical
way to clean these plumes would be to capture them at the source where the
concentration is highest and gas volume lowest.  However, for practical reasons
such as the presence of overhead cranes, it is impossible in many cases to
capture the plume at the source or even vertically above the source with fume
hoods.  In these situations a SCAT system could be applied to control the
emissions.  An air curtain could be used as the ceiling to contain the fumes
and dust and horizontally displace the plume into a receiving hood or directly
into a scrubber if space is available for its installation.
     Experiments were done to study the feasibility of containing hot plumes
with air curtains.  The hot source was simulated with an open top furnace which
had a natural gas, open flame burner.  The furnace was a rectangular box lined
with Insulating fire-bricks.  There was an opening on top of the furnace for the
hot gas to exit.
     The furnace was located at the center of the SCAT system.  The distance
between the air curtain assembly and the spray scrubber was 3.1 m.  Since the
operation of the burner was fixed, the "ceiling" air curtain location was
adjustable so that the jet stream could meet the hot plume at different
temperatures.  At a height of 60 cm above the furnaces the hot plume had a
temperature of 500°C and a velocity of 2 m/s.
     The effect of buoyancy on the air curtain flow field was deduced by
comparing the velocity and temperature profiles before and after the burner
was turned on.  Fly ash particles were also injected into the furnace with the
celling air curtain off and on for hot plume trajectory observations.
     A correlation for predicting the hot plume trajectory was derived.  It is
based on a momentum balance which accounts for the buoyancy and momentum of
the plume.  Figure 6 shows the experimental layout and predicted and observed
                                      61

-------
    150
 
-------
I— 300 cm
                                                         OBSERVED
                                                           PLUME
 — 200 era
                  AIR CURTAIN
                  MANIFOLD
 — 100 cm
                                                    FURNACE
                                                                    PREDICTED
                                                                    CENTERLINE
                                                                    TRAJECTORY
                                                                                            SCRUBBER
                   0
100
 I
200
300 cm
                              Figure 6.  Hot plume trajectory and flow pattern.

-------
plume center line at the scrubber inlet,  predicted plume rise is slightly
higher than observed.

CHARGED SPRAY STUDY
     For a spray scrubber collection by drops is the principal collection
mechanism and the particle penetration for a given size particle nay be
calculated from theory (Calvert, et al, 1975).
     The charged spray scrubber has been studied theoretically and experimentally
by a number of researchers, including Melcher and Sacher (1974), Leart et al,
(1975, 1976), and Pilat, et al, (1975, 1976, 1978 a, b).  However, data reported
by these researchers were not sufficient to verify the theory.
     An experimental charged spray scrubber system was built to obtain design
data under well defined conditions.  The scrubber consisted of a flow
straightening section, an inlet particle sampling section, a particle charging
section, a spray section, an entrainment separator, and an outlet sampling
section.  The spray section consisted of two removable spray banks and their
configuration varied depending on nozzle type.
     The drops were charged by induction.  A high voltage grid assembly was
placed in front of nozzles to induce charge on the water drops.  The distance
between the grid and nozzle was adjustable to allow for maximizing the drop
charge level.
     Two types of nozzles were tested.  The hook type nozzle (type "A")
produced drops with mean diameter around 240 urn at a nozzle pressure of 430 kPa.
The discharge coefficient and spray angle for this nozzle were 0.63 and 100°,
respectively.
     The particle collection efficiency of the spray scrubber was measured for
the following four conditions:
     1.  ND/UP (Neutral drops/uncharged particles),
     2.  CD/UP (Charged drops/uncharged particles),
     3.  ND/CP (Neutral drops/charged particles), and
     4.  CD/CP (Charged drops/charged particles).
Note that the "uncharged" particles may not be neutral because they may carry
some charge when they enter the scrubber.
     Figures 7, 8 and 9 show typical data along with theoretical predictions,
The particle collection efficiency of the scrubber was found to Improve by
                                      64

-------
   1.0
   0.5  _
c
o
O

-------
    1.0
                 I      I
                         CD/UP
                                      IIII  I I 1.

                                      PREDICTION       '.
    0.5
    0.2
O
(O
    0.1
I   0.05
    0.02
    0.01
              NOZZLE "B"
I      I    I   I  I  I I I I
i      i    i  i  i i i i
        0.1     0.2
            0.5       1.0       2

          AERODYNAMIC DIAMETER, ymA
                       10
         Figure 8.   Charged spray performance for image charge with
                    nozzle B.
                                   66

-------
    1.0
    0.5
                 1     I   I  I  1  I  I  I
                      PREDICTION
                          I   I   I  II I !..
    0.2
c
o
•r~
4J
U

£   o.i
    0.05
LU

Q.
                           CD/CP
    0.02
              NOZZLE "B1
    0.02
         j|
I      L   I   I  f I  I  I
        0.1     0.2
0.5       1.0       2              5

 AERODYNAMIC DIAMETER, ymA
                       10
         Figure 9.  Charged  spray  performance for coulombic force with

                    nozzle B.
                                      67

-------
charging either the water drops or the particles.  Measured drop charge and
particle charge levels were +5.8 x 10"7 C/g and -1.5 x 1(T5 C/g, respectively.
Test particles were, hydrated lime with mass median diameter of 3.5 yraA and
geometric standard deviation around 2.  Further improvement was obtained when
the drops and particles were oppositely charged.  The improvement was more
with submicron particles.  For particles larger than 5 umA diameter, charging
the water and/or particles has little effect on efficiency.
     Hook type nozzle gave a higher collection efficiency than the pigtail type.
This is consistent with the smaller drops produced by hook type nozzles.
     The data agree with predictions for the ND/UP conditions (spray scrubber
only).  When either or both the particles and drops were charged, the measured
collection efficiency was higher than that predicted.  In fact the theory
predicted that there should be no improvement in efficiency for ND/CP and
CD/UP conditions.  Slight improvement was predicted for the CD/CP condition.  There-
fore a better theoretical design model is required for predicting performance
of charged spray scrubbers.

CONCLUSIONS
     A simple technique for controlling fugitive process emissions has been
developed.  The technique involves the use of air curtains and air jets to
contain and convey the emissions into a nearby spray scrubber.
     The collection efficiency of a spray scrubber was investigated experimentally.
The collection efficiency was improved by charging the water and/or the particles.
The measured particle penetration agrees with theoretical predictions for the
ND/UP condition.  For the electrostatically augmented scrubber, the measured
penetration is lower than that predicted.
     Air curtains have been used in industry to contain dust but no carefully
performed study has been reported in literature.  The air curtain developed
in the present study can achieve a smaller expansion angle and a lower entrain-
ment ratio than those reported in the literature.  Small expansion angles and
entrainment ratios are beneficial to the control of fugitive process emissions
with the SCAT system.
     The effects of crosswind and the containment of a hot buoyant plume were
also studied.  It has been shown that available theory (Abramovich) gave
reasonable predictions for the air curtain range, trajectory of the air curtain
                                     68

-------
axis in the presence of a crosswind, and the required spray scrubber rotation
angle for intercepting the jet stream.  Therefore, in operating the SCAT system
under crosswind conditions, there is a rational basis for locating the air
curtain and spray scrubber relative to the crosswind and emission source.
     The air curtain was successful in containing a hot buoyant plume.  At
an air curtain/spray scrubber separation of 3 m, an air curtain with slot width
of 5.1 cm and air exit velocity of 20 m/s can contain a hot plume which is rising
at a velocity of 200 cm/s and has a temperature of 470°C.  The trajectory of
the plume can be predicted from a plume rise formula.
     We have done most of the necessary basic research in the present study.
The general SCAT system can be applied to many kinds of sources.  The next step
is to select a source and demonstrate the feasibility of the SCAT system.

REFERENCES
Abramovich, G.N.  The Theory of Turbulent Jets.  The MIT Press, 1963.
Hemeon, W.C.L.  Plant and Process Ventilation. 2nd Ed.  The Industrial Press,
     New York, 1963.
Lear, C.W., W.F. Krieve, and E. Cohen.  "Charged Droplet Scrubbing for Fine
     Particle Control", T. of A.P.C.A., 25_: 184-189, 1975.
Lear, C.W.  "Charged Droplet Scrubber for Fine Particle Control-Laboratory
     Study."  NTIS PB258823, 1976.
McElroy, G.E.  "Air Flow at Discharge of Fan-Pipe Lines in Mines."  Part II,
     U.S. Bureau of Mines Report of Investigation 3730, November 1943.
Melcher, J.R., and R.S. Sacher.  "Charged Droplet Technology for Removal of
     Particulates from Industrial Gases."  NTIS PB205187, 1971.
Tuve, G.L. and G.B. Priester.   "Control of Air Streams in Large Spaces."
     Heating, Piping and Air Conditioning, ASHVE Journal, 50:39, 1944.
Yung, S.C., J. Curran and S. Calvert.   "Spray, Charging,  and Trapping
     Scrubber for Fugitive Particle Emission Control."  Final  Report to
     U.S.-E.P.A. for Contract  68-02-3109.  To be published.
                                      69

-------
              FUGITIVE EMISSION CONTROL OF OPEN DUST SOURCES

                                    By

                          Thomas A.  Cuscino, Jr.,
                           Chatten Cowherd, Jr.,
                             and Russel Bohn

                        Midwest Research Institute
                           425 Volker Boulevard
                       Kansas City,  Missouri  64110


                                 ABSTRACT
     This paper presents empirically developed predictive emission factor
equations for open dust sources in iron and steel plants.  The ranges of
applicability and the precisions of the equation are discussed.   Presently,
the equations for the two open dust sources of greatest magnitude (unpaved
and paved roads) have good precision, with 68% of the predicted values lying
within factors of 1.21 and 1.53 of the measured values for unpaved and paved
roads, respectively.

     Also presented are the results of tests performed on control techniques
to mitigate fugitive dust from vehicles traveling on unpaved roads.  Limited
testing of chemical dust suppressants for industrial unpaved roads indicates
a high initial control efficiency (exceeding 90%) which decreases more than
10 percentage points within about 24 hr after application.  The emission
factor equations are shown to be useful in estimating control efficiencies
in the absence of adequate efficiency test data.
                                     71

-------
                         FUGITIVE EMISSION CONTROL
                           OF OPEN DUST SOURCES
INTRODUCTION

     Two types of fugitive emissions occur in the iron and steel industry—
process fugitive emissions and open dust source fugitive emissions.  Pro-
cess fugitive emissions include uncaptured particulates and gases that are
generated by steel-making furnaces, sinter machines, and metal forming and
finishing equipment, and that are discharged to the atmosphere through
building ventilation systems.  Open dust sources are those that entail the
generation of fugitive particulate emissions by the forces of wind and ma-
chinery acting on exposed raw, intermediate and waste aggregate materials
during storage, transfer and disposal.

     The ranking of the emissions potential of open dust sources at a given
iron and steel plant or across the industry is an important tool in deciding
where controls may be needed.  This requires the development of emissions
inventories, i.e., calculation of average emission rates for all significant
sources at one or more plant sites.

     Calculation of the emission rate for a given source requires data on
source extent, uncontrolled emission factor and control efficiency.  The
mathematical expression for this calculation is as follows:

          R = Me (1 - c)

where     R = mass emission rate
          M = source extent
          e = uncontrolled emission factor, i.e., rate of uncontrolled
                emissions per unit of source extent
          c = fractional efficiency of control

     Because of the wide range of particle size associated with fugitive
particulate emissions, it is important that the applicable particle size
range be specified for the calculated emission rate.  The particle size
range should be that for which the uncontrolled emission factor and the
fractional efficiency of control apply.

     In a recent study of fugitive particulate emissions from integrated
iron and steel plants, Midwest Research Institute1 (MRI) determined that
open dust sources (specifically, vehicular traffic on unpaved and paved
roads and storage pile activities) ranked with fugitive emissions from
steel-making furnaces and sinter machines as sources which emit the largest
quantities of fine and suspended particulate matter, taking into account
typically applied control measures.  It became evident that open dust
sources should occupy a prime position in control strategy development for
fugitive particulate emissions within integrated iron and steel plants.
Moreover, preliminary analysis of promising control options for both pro-
cess sources of fugitive emissions and open dust sources indicated that
control of open dust sources has a highly favorable cost-effectiveness
ratio for particulate matter.
                                     72

-------
     The industry-wide open dust source emissions inventory compiled by MRI
utilized predictive emission factor equations developed by MRI based on
field tests of a variety of uncontrolled sources at iron and steel plants.
The most significant limitation to the reliability of the emissions inven-
tory resulted from a lack of detailed knowledge of control efficiencies for
various open dust source control techniques.

     The following sections of this paper discuss:  (a) the ranges of appli-
cability and the precisions of the empirically developed predictive emis-
sion factor equations for uncontrolled open dust sources; and (b) the use
of the emission factor equations to estimate the efficiencies of open dust
source controls.  In addition, the results of the tests performed on con-
trol techniques to mitigate fugitive dust from vehicles traveling on un-
paved roads are presented.

     A mixture of metric and English units was used in this paper.  The
symbol "T" refers to short ton, which is equivalent to 2,000 Ib.  The sym-
bol "t" refers to the metric tonne, which is equivalent to 2,200 Ib.  An
English-to-metric conversion table is presented at the end of this paper.

EMISSION FACTORS FOR UNCONTROLLED OPEN DUST SOURCES

     The emission factor equations empirically developed by MRI1'2 for uncon-
trolled open dust sources are shown in Table 1.  The equations describe emis-
sions of particles smaller than 30 pro in diameter based on a particle density
of 2.5 g/cm3.  Although the equations represent anthropogenically uncontrolled
emissions, most of the equations do incorporate natural control due to pre1
cipitation.

     The precision of each equation in predicting measured emission factors
over given ranges of independent variables has been calculated for those
open dust sources which have been tested frequently enough to support a sta-
tistical analysis.  Table 2 shows the one-sigma precision factors for the
predictive equations.  The one-sigma precision factor (fj) is defined such
that 68% of the predicted emission factors will be with a factor of fx of
the measured values.  The two-sigma precision factor (f2) is defined such
that 95% of the predicted emission factors will be within a factor of f2 of
the measured values.

     The precision factors given in Table 2 are applicable only when the
predictive equations are used with values of the independent variables that
are within the ranges tested, which are also shown in Table 2.   The equa-
tions are of undetermined precision when applied to sources for which in-
dependent variables lie outside of the ranges tested.   Tables 3, 4, and 5
list the parameter ranges that have been measured by MRI3 for road'surface
dust and aggregate materials within the iron and steel industry.

     The precisions of the various equations differ extensively.  The one-
sigma precision factor for unpaved roads is 1.21, while the two-sigma pre-
cision factor is 1.46 based on 23 tests.   Because unpaved roads a^oft-en
the largest open dust source in an iron and steel plant, more effort has
been placed on the testing of this source in relation to the others  with
                                     73

-------
                          TABLE  1    OPEN  DUST SOURCE  EMISSION FACTOR  EQUATIONS
         Source category
     Measure of extent
           Emission factor
      (Ib/unit of  source extent)
                                                                                     Correction parameters
1.  Unpaved roadi
2.  Paved roads
3.  Batch load-in (e.g.,  front-
      end loader, railcar duap)
 .  Continuous load-in (e.g.,
      •lacker, transfer station)
5.  Active storage pile aainte
      nance and traffic
6.  Active storage pile wind
      erosion
7.  Batch load-out (e.g.  froMt-
      end loader
8.  Wind erosion of exposed areas
Vehicle-Biles traveled
                                      Vehicle-siiles traveled
Tons of Material  loaded in
Tons of siaterial loaded in
Tons of Material  put  through
  storage
Tons of Material put  through
  storage
Tons of Material loaded out
Acre-years of exposed land
     '.     S   /W\°-7/w\°-5   d
5'»   12    30  (3/   («/     365
                                  •«•   S    fo   i*i.  (!)
                                                                                                     0.7
                                 0.0018
                                 0.0018
         s    0    h
         555
          /u.2  /wi 0.33
                                                                                s   U   h
                                                                                5   5  10
°-IOK   O   235
••«   A   215    ff  io
                                 0.0018
3.400
                                                                               e_  -s_   f_
                                                                               50   15   25
  s = Silt  content of aggregate or road
        surface material (%)

  S = Average vehicle speed (raph)

  W = Average vehicle weight (tons)

  L = Surface dust loading on traveled
        portion of road (lb/siile)

  V = Mean  wind speed at 4 • above
        ground (oph)

  M = Unbound aoisture content of
        aggregate or road surface
        swterial (X)

  J = Dmping device capacity (yd*)

  K = Activity factor6

  d = Kusiber of dry days per year

  f - Percentage of tin* wind speed ex-
        ceeds 12 «ph at 1 ft above the
        ground

  D = Duration of Material storage (days)

  e - Surface erodibility (tons/acre/year)

P-E at Thornthwaite's Precipitation-
        Evaporation Index

  H e Nuaber of active travel lanes

  I = Industrial road augmentation
        factor

  w = Average nuaber of vehicle wheels

  h = Drop  height (ft)
a  Represents particnlate sanller than 30 pa in diaawter baaed on particle density of 2.5 g/ca1.

b  Equals 1.0 for front-cad loader Maintaining pile tidiness and 50 round trips of custoaer tracks per day in the storage area.

c  *  Equals 7.0 for  tracks coming tram unpaved to paved roada and releasing dust froa vehicle underbodies;
   *  Equals 3.5 when 201 of the vehicles are forced to travel teaporarily with one set of wheels on an unpaved road  bem while passing on narrow roads;
   *  Equals 1.0 for  traffic eatirely on paved surfacea.

-------
              TABLE 2.  PREDICTIVE EQUATION PRECISION FACTORS
                          AND RANGES OF APPLICABILITY
  Source category
                                    Tested range of
                                 independent variable
                                  Precision factor
                                 for 68% confidence
                                      interval
1.   Unpaved roads
3.
4.
5.
6.
7.
    Paved roads
    Batch load-in an.d
      load-out (e.g.,
      front-end loader,
      railcar dump)
    Continuous load-in
      (e.g.,  stocker,
      transfer station)
    Active storage pile
      maintenance and
      traffic

    Active storage pile
      wind erosion
    Wind erosion of
      exposed areas
  s:   4.3-68%
  S:   22-64 km/hr (13-40 mph)
  W:   3-142 t (3-157 T)
  w:   4-12 wheels

  I:   1-7
  N:   2-4 travel lanes
  s:   5.1-13.2%
  L:   42-629 kg/km
        (150-2,230 Ib/mile)
  w:   3-12 t (3-13 T)

  s:   1.3-7.3%
  U:   2.1-22 km/hr (1.3-14 mph)
  h:   1.5 m (5 ft)
  M:   0.25-0.7%
  Y:   1.5-7.7 m3 (2-10 yd3)

  s:   1.9-19.1%
  U:   0.8-2.7 m/s (1.8-6.0 mph)
  h:   1.5-12 m (5-40 ft)
  M:   0.64-4.8%

  K:   1.0
  s:   1.5%
  d:   235 days

  s:   1.5%
  d:   235 days
  f:   15%
  D:   90 days

  e:   2.24 x 107 kg/km2 yr
        (100 T/acre/yr)
  s:   8.5%
  f:   100%
P-E:   40
                                                                   1.21
                                                                   1.53
4.9
5.4
source:  References 1,
a
                          and
   Limited number of tests combined with theoretical development.
                                     75

-------
                TABLE 3.  SILT CONTENT VALUES APPLICABLE IN
                            THE IRON AND STEEL INDUSTRY
                                              Range of silt
Source
Unpaved roads
Paved roads
Number
of tests
12
9
content
00
4 -13
1.1-13
Average silt
content (%)
7.3
5.9
Material handling activities and
  storage pile wind erosion

  Coal                                 7         2-7.7          5.0
  Iron ore pellets                    10         1.4-13            4.9
  Lump iron ore                        9         2.8-19            9.5
  Coke breeze                          1            -              5.4
  Slag                                 3         3.0- 7.3          5.3
  Blended ore                          1            -             15.0
  Sinter                               1            -              0.7
  Limestone                            1            -              Q.4
  Flue dust                            2        14  -23           18.0

Source:  Reference 3.     ~~  ~~~
                                    76

-------
              TABLE 4.  SURFACE MOISTURE CONTENT VALUES APPLICABLE
                          IN THE IRON AND STEEL INDUSTRY
             Source
       Number
      of tests
                                              Range of surface
                                              moisture content
       Average
  surface moisture
    content (%)
Material handling activities and
  storage pile wind erosion
Coal
Iron ore pellets
Lump iron ore
Coke breeze
Slag
Blended ore
Flue dust
6
8
6
1
3
1
1
2.8 -11
0.64- 3.5
1.6 - 8.1
-
0.25- 2.2
-
•
4.8
2.1
5.4
6.4
0.92
6.6
12.4
source:  Reference 3.
                 TABLE 5.  SURFACE LOADING ON TRAVELED LANES OF PAVED
                             ROADS IN IRON AND STEEL PLANTS
    Number
   of tests
Range of surface
     loading
    (Ib/mile)
Average surface
    loading
   (Ib/mile)
                                 65-17,000
                                       2,700
   bource:  Reference 3.
                                        77

-------
the result that equation for unpaved roads is the most precise.  By com-
parison, the one-sigma precision factor for paved roads is 1.53, while the
two-sigma precision factor is 2.34 based on 10 tests.

     To illustrate the effectiveness of the predictive emission factor equa-
tion for unpaved roads, comparisons may be made between predicted emission
factors and the corresponding measured values in the supporting data base.2
As shown in Figure 1 (Case 3), the measured emission factors for unpaved
roads, which span two orders of magnitude, are predicted using the emission
factor equation with a two-sigma precision factor of 1.46, i.e., the 95%
confidence interval for a predicted emission value, P, extends from P/1.46
to 1.46 P.  If the average of all emission factor measurements is used in-
stead of the equation, the two-sigma precision factor escalates to 5.2.  In
the other case shown in the figure, the average of measurements at a given
site predicts the measurements at that site with a precision factor of 2.3.

     The one-sigma precisions of the equations for batch and continuous ma-
terial handling are 4.9 and 5.4 based on 8 and 9 tests, respectively.  The
diversity of materials handled and handling operations necessitates that
more tests be performed and that the predictive equations be modified to
achieve a higher degree of precision.  However, the measured emission fac-
tors for typical material handling operations indicate that these opera-
tions are small sources in most industrial settings and may not merit fur-
ther testing.

     The precision factors for the storage pile maintenance, storage pile
wind erosion and exposed area wind erosion equations are unknown since only
limited testing of these sources has been performed to date.  Wind erosion
is currently being investigated by MRI using a portable wind tunnel with
the goal of developing a statistically precise predictive emission factor
equation for the wind erosion of storage piles and exposed areas.

EFFICIENCIES OF OPEN DUST SOURCE CONTROLS

     Only limited testing results are available to define the efficiencies
of various open dust source control techniques.  However, MRI is currently
engaged in a project funded by the U.S. Environmental Protection Agency
(EPA Contract No. 68-02-3177, Task 4) and entitled "Iron and Steel Plant
Open Source Fugitive Emission Control Evaluation," which is directed to the
quantification of control efficiencies for open dust sources.

     In the absence of adequate test data on control efficiency, the pre-
dictive equations shown on Table 1 can be used to estimate control effi-
ciencies, if the changes in the independent variables affected by these
control techniques can be quantified.  Table 6 identifies the independent
variables in the predictive equations in Table 1 which are affected by var-
ious open dust source control techniques.

     Table 7 shows the results of 8 control technique quantification tests
for emissions from unpaved roads.  Two tests were performed on roads treated
with Coherex® (petroleum resin), three on roads treated with TREX (ammonium
lignin-sulfonate), and two roads were tested after a precipitation event.
                                     78

-------
                                  95  PERCENT CONFIDENCE  INTERVALS FOR
                                      UNPAVED ROAD  EMISSION FACTORS
 i
 I
 i
                   CASE  1 - Use of
                   Overall Average
                   Measured Emission
                   Factor to Predict
                   Individual Measure-
                   ments at All Test  Sites
 c
 o
-o
 0)
o
o
CASE 2 - Use of
Average Measured
Emission Factor for
Each Test Site to
Predict Individual
Measurements at
Only That Site
                                                                                 CASE 3 - Use of
                                                                                 Emission Factor
                                                                                 Equation to Predict
                                                                                 Individual Measure-
                                                                                 ments at All Test Sites
                         Figure  1.   Comparison of unpaved road emission factor precisions.

-------
                    TABLE 6.  EFFECT OF CONTROL TECHNIQUES ON
                                INDEPENDENT VARIABLES
     Source
    Control technique
Independent variable affected
Unpaved roads
Paved roads
Watering (anthropogenic)

Precipitation

Chemical dust suppressants

Vehicle speed and type
  control


Flushing


Broom and vacuum sweeping
                     Vehicle speed and type
                       control
Material handling    Windbreaks
                     Watering (anthropogenic)

                     Precipitation

                     Chemical dust suppressants
                       Water extenders
                       Agglomerators

                     Reduce aggregate material
                       drop height

                     Use larger capacity
                       equipment for batch
                       handling
None

Number of dry days (d)

Silt content (s)

Vehicle speed (s)
Number of wheels (w)
Vehicle weight (W)

Silt content (s)
Total loading (L)

Silt content (s)
Total loading (L)

Vehicle speed (s)
Number of wheels (w)
Vehicle weight (W)

Mean wind speed at 4 m (U)
Percent of time wind exceeds
  12 mph at 1 ft above
  eroding surface (f)

Moisture content (M)

Moisture content (M)
                               Moisture content (M)
                               Silt content (s)

                               Drop height (h)
                               Dumping device capacity (Y)
                                        80

-------
                              TABLE 6.  (continued)
     Source
    Control technique
Independent variable affected
Wind erosion of
  storage piles
Wind erosion of
  exposed areas
Watering (anthropogenic)

Precipitation

Chemical dust suppressants

Windbreaks
Crusting of inactive piles
  (induced by watering,
  precipitation, and chem-
  ical suppressants)

Watering (anthropogenic)

Precipitation

Chemical dust suppressants


Windbreaks
                     Crusting of inactive
                       surfaces (induced by
                       watering, precipitation,
                       and chemical dust
                       suppressants)
Hone3:

Number dry days (d)

Silt content (s)

Percent of time wind exceeds
  12 mph at 1 ft above
  eroding surface (f)

Silt content (s)
None3

PE Index (PE)

Silt content (s)
Erodibility (e)

Percent of time wind exceeds
  12 mph at 1 ft above
  eroding surface (f)

Silt content (s)
Erodibility (e)
   Watering does not affect silt content since s is determined by dry sieving.
                                        81

-------
                             TABLE 7.   SUMMARY OF CONTROL EFFICIENCY TESTS FOR UNPAVED ROADS
oo
fsj
Time after
Uncontrolled chemical
emission application
factor or rainfall
Road surface (Ib/VMT) Control type cessation (hrs)
Dirt/slag 2.3 (M)a 10% Coherex® in water 20
10% Coherex® in water 22
Crushed rock 21.5 (P) 25% Trex in water; 0.1 gal/yd2 24
26
27-1/2
Crushed rock 21.6 (M) Rainfall of 1.13 in. on two 23
& glacial till preceding days
Crushed lean 21.5 (P) Rainfall of 1.13 in. on two 28-1/2
taconite rock preceding days
29-1/2
Source: References 2 and 4.
, H - measured
• D — nr-^A-l ft-** A
Measured
Controlled
emission Control
factor efficiency
(Ib/VMT) (%)
0.073 97
0.36 84
2.0 91
2.3 89
3.6 83
2.3 89

11.6 54

11.6 54

-

-------
Consistent with the emission factor equation for unpayed roads, the lowering
of emissions on the Coherex-treated road was reflected by a reduction in
silt content from 9.0% to 0.03%.

     It is evident from Table 7 that in all cases the decay in control ef-
ficiency with time after application and road usage was dramatic.   The con-
trol efficiency of Coherex decayed from a high initial value (exceeding 90%)
to 84% efficiency within 22 hr after application, while TREX similarly de-
cayed to 83% efficiency within 27-1/2 hr after application.  The natural
control of precipitation decayed to 54% within 29 hr after the rain ended.

CONCLUSIONS

     Emission factor equations which are applicable to uncontrolled open
dust sources in iron and steel plants have been developed and are being im-
proved through the generation of a more extensive and reliable data base.
The use of predictive emission factor equations provides for much greater
precision than single-valued emission factors by incorporating correction
parameters which account for source variability.  Presently, the equations
for the two open dust sources of greatest magnitude (unpaved and paved roads)
have good precision, with one-sigma precision factors of 1.21 and 1.53 for
unpaved and paved roads respectively.  The tested ranges of independent vari-
ables which enter into the equations generally encompass the uncontrolled
source conditions found at iron and steel plants.

     The lack of quantitative data on open dust source control efficiencies
indicates a strong need for more source testing.  As an interim measure,
the predictive emission factor equations may be used to estimate control
efficiencies.  This entails the measurement or estimation of the independent
variables under controlled conditions rather than the more difficult mea-
surement of the reduction in particulate emissions.

     Limited testing of chemical dust suppressants for industrial unpaved
roads indicates a high initial control efficiency (exceeding 90%)  which de-
creases by more than 10% within about 24 hr after application.  Consistent
with the emission factor equation, the lowering of emissions is reflected
by the reduced silt content of the road surface material after the applica-
tion of chemical dust suppressants.  Additional testing is needed to better
quantify the performance of road dust suppressants.
                                      83

-------
ACKNOWLEDGEMENT

     The work upon which this paper is based was performed in part pursuant to
Contract No. 68-02-3177 (Task 4) with the U.S. Environmental Protection Agency.
                  ENGLISH TO METRIC UNIT CONVERSION TABLE
               English unit    Multiplied by    Metric unit
Ib/T
Ib/vehicle mile
Ib/acre yr
Ib
T
mph
mile
ft
acre
0.500
0.282
112
0.454
0.907
0.447
1.61
0.305
0.00405
kg/t
kg/vehicle km
kg/km2 yr
kg
t
m/s
km
m
km2
REFERENCES
     1.   Bohn, R., T. Cuscino, Jr., and C. Cowherd, Jr.  Fugitive Emis-
          sions from Integrated Iron and Steel Plants.  EPA-600/2-7S-050,
          U.S. Environmental Protection Agency, Research Triangle Park,
          North Carolina, March 1978.  276 pp.

     2.   Cowherd, Chatten, Jr., Russel Bohn, and Thomas Cuscino, Jr.   Iron
          and Steel Plant Open Source Fugitive Emission Evaluation.  EPA-600/
          2-79-103, U.S. Environmental Protection Agency, Research Triangle
          Park, North Carolina, May 1979.  139 pp.

     3.   Cuscino, Thomas A., Jr.  Particulate Emission Factors Applicable
          to the Iron and Steel Industry.  EPA-450/4-79-028, U.S. Environ-
          mental Protection Agency, Research Triangle Park, North Carolina,
          September 1979.  83 pp.

     4.   Cuscino, Thomas, Jr.  Taconite Mining Fugitive Emissions Study.
          Minnesota Pollution Control Agency, Roseville, Minnesota,.June
          1979.  75 pp.
                                      84

-------
   WIND VELOCITY DISTRIBUTION OVER STORAGE PILES AND USE OF BARRIERS
                                  by
               S. L. Soo, J. C. Perez, and S. Rezakhany
          Department of Mechanical and Industrial Engineering
              University of Illinois at Urbana-Champaign
                           Urbana, IL  618Q1
                                  ABSTRACT
     To reduce the wind blown dust from storage piles of coal or other
bulk materials, wind tunnel tests of scale models are being made to
determine the effect of pile configuration and use of wind barriers to
reduce wind penetration.

     Preliminary results show that a wind break or barrier simulating a
snow fence whose height is 1/2 the pile height with a solidity of 2/3
may reduce the penetration velocity of wind by one-half when it is
placed three pile-heights away as an optimum.  A similar barrier of 1/4
pile height located 2.5 pile-heights away reduces the wind velocity by
23 percent.  The effect of leeward barriers is less obvious, however.
Reduced wind penetration into a pile also conserves the moisture in the
pile thus improving the adhesion .of dusts.
                                  85

-------
                             NOMENCLATURE





SYMBOLS



C    Pressure coefficient
 P

h    Height of pile

                              2
k    Permeability of a pile, m



L    Characteristic length, m



N_   Reynolds number, dimensionless
 Ke


P    Pressure, Pa



V    Wind velocity, m/s



p    Density of fluid, kg/m



y    Viscosity of fluid, kg/ms





Subscripts



o    Characteristic or freestream quantitites



p    Quantities inside the pile
                                   86

-------
INTRODUCTION
     The significance of the abatement of fugitive emission the form of
wind blown dust from storage piles of bulk materials such as coal and
ores is seen from the measurements reported by Cowherd and Hendricks
(1977) .  The annual wind loss of particles below 30 ym size from a
single steel plant may run into thousands of tons from storage piles
alone.  Nationally, this mode contributes to 10 percent of total
suspended particle emissions of particles below 30 ym (Cowherd, et al.
     2
1979)  thus constituting an important pollution source.  Estimated
typical control efficiency is around 40 percent.
     Many factors influence this wind erosion loss such as wind velocity,
moisture content, rainfall, duration of storage, compaction of pile and
the amount and size of fines, and the pile configuration.  To this, we
may add the effectiveness of means of abatement.  Wind erosion tests
were made in wind tunnels to simulate both stationary piles and rail
cars (Nimerick, et al 1979) .  Similarity relations are, however,
complicated in that wind velocity distribution over a pile is affected
by the characteristic Reynolds number of the pile while the lifting of
particles from a pile is influenced by the size of particles and shear
flow (Soo and Tung 1972)4.
     The present research concerns abatement of such a pollution source
as well as conservation of resources.  The task consists of quantifi-
cation, simulation, and optimization of the shape and orientation of a
porous storage pile and the strategic use of economically feasible
barriers (windbreaks) or covering.  The desirable condition includes
minimum penetration at the windward end of the pile by dynamic pressure
and the leeward end of the pile by separated flow and trailing vortices.
This part of the study has been made by wind tunnel testing and numerical
modeling of the fluid mechanics involved in wind penetration.   The
results will facilitate prediction of the behavior of the dust plume.
This will lead to the desirable external protection and pile configuration
and orientation and accurate estimation of dust production by wind based
on the characteristics of particles.
     To insure accurate prediction, strict modeling criteria have been
followed.
                                   87

-------
 SIMILARITY  RELATION

     Similarity of flow  over  a porous  specimen or model to that of the
 prototype storage pile is  achieved by  keeping their characteristic
 Reynolds numbers equal.  The  latter  is given by:

                 NRe  = Lo P V*
 where  LQ is a  characteristic  length  such as pile height, V  is the
 characteristic  wind  velocity, p and  u  are the density and viscosity of
 air, respectively.   For  the present  study, control of the boundary layer
 thickness of the approaching  flow is also needed.  Typical relations are
 seen in the following example:
                                   Prototype            Model
     Pile Height			---3.05 m-	76 mm
     Air Velocity (VQ)			1.14 m/s		46 ffl/s
 Because of  the  large wind  velocity needed in the model for similarity,
 the above pile  of simulating  bulk material has to be confined by a wire
 screen.

     Similarity in wind penetration into a porous pile is characterized
 by a parameter  which is the ratio of viscous resistance to external
 freestream  to the resistance  to flow through the pores:

                 *Vkp p Vo
 where k  is the permeability of the pile given by the Kozeny-Carman
       P              532
relation (Carman 1956) :  k  = e /S  a,  v/here e is fraction void called
porosity, S is the surface area per unit volume,  and a is constant which
is nearly 5.  The basis of choosing k  is given by the relation for the
flow velocity in the porous body V  is given by:

               Vp » (k/y)  7vP

-------
where  VP denotes gradient of pressure in the porous body.  For similar
materials and chips of coal of similar size distributions, typical size
relations are:
     Prototype;  10 x  0 on;    Model: 2.54 x 0 mm (No. 8 Sieve)

EXPERIMENTAL FACILITY
     For the quantitative modeling of a coal pile under the influence of
wind, our present multiphase wind tunnel facility of 305 mm square
cross-section have been modified to  accommodate a specimen as in Figure
1 showing traversing stations.  This device make's possible testing of
two-dimensional porous specimens, simulating a coal pile.  Wind velocity
distributions are determined by traversing of pi^ot-static and yaw
probes outside of the pile model.  From these measurements, penetration
of wind into the inside of the pile  was deduced.

     Models were designed to simulate prototype storage piles.  Field
observation of storage piles of coal was made by J. D. Tyrrel at the
Consumer Power Company, J. C. Weadock Plant, Essexville, Michigan.
Measurements from photographs taken  gave angles of 37.4 and 37.7 degrees
from the horizontal.  These coal piles had an average height of 3.05 m
(10 ft).  Variables affecting coal pile configurations and packing
frequencies were amount and frequency of coal supply, area of coal pile,
coal usage demands, hopper location, stacking capabilities, and manpower.

     Based on these observations, a  76 mm and a 51, mm thick porous piles
of haydite* 305 mm wide and 610 mm long (base) were prepared with pile
angles of 40 and 35 degrees on the two ends for the preliminary measure-
ments. These models can be reversed  in the wind tunnel to test effects
of pile angles.

     Preparation of the porous specimen included crushing solids, haydite
(sp. gr. 1.31) in this case, to the  desjlred sizes followed by sieving
and mixing to achieve the desired porosity for the purpose of simulation.

     A 3 mm square wire screen has been used to cover the particles of
3 mm by 6 mm to maintain the prescribed geometry and porosity.  This
 A high grade shale product which is kiln processed and crushed to nearly
 JLU Hull siz€*
                                   89

-------
screen cover is needed because of the high wind velocity used in the
simulation.  The pile has a porosity (void) of 0.40 and a permeability
of  1.77 x  10~8 m2.
     Figure 1 shows that for each given particle size constituting the
bulk of the pile, the variables of geometry are the height  (h), length
(L), and angles (B^ and 92) of the pile.  Simulation of actual wind
profiles is accomplished by a long inlet pipe length to give a fully
developed  boundary layer.  Standard pitot and yaw probes were used with
fluid and  electronic manometers.  A hot wire anemometer was used to
determine  the velocity fluctuations.
     Figure 1 also shows the model and the stations at which air velocity
was measured.  At stations A through I, a pitot probe was inserted
vertically to determine the flow configuration surrounding the pile.  At
points 1 through 11, the yaw probe was inserted horizontally in order to
measure the air velocity and angle at the surface of the pile.
     For these experiments, average wind velocities were chosen by
damper adjustment and fan speed.
     Type  of barriers tested include 25.4 mm barriers, solid and slotted
to a solidity of 2/3, and 13 mm barriers, solid and slotted to 2/3
solidity.  All these were tested for optimum locations ahead of the pile
(windward) and behind the pile.  All locations are in terms of nominal
pile heights.
     When  testing the 51 mm pile model, the 25.4 mm barrier corresponds
to 1/2 the pile height in order to simulate a likely real situation.
The objective was to simulate simple fences that could be used as a
means to reduce wind penetration and lift of dust.   In all cases, the
barriers were set parallel to the front or back edge of the pile and
perpendicular to the bottom plate of the wind tunnel.

MEASUREMENTS AND EXPERIMENTAL RESULTS
     Tests have been made principally with a pile of 51 mm height and a
pile of 76 mm height.   Extensive measurements have been made both to
determine the reference condition of flow over the pile with an inlet
boundary layer simulating the atmospheric wind tests were made first
without protective barriers and subsequently with various forms of
barriers.
                                   90

-------
Figure 1  Measuring stations of wind velocity over
          the porous specimens (all dimensions in mm)

-------
1. Tests without Barriers
   A typical set of test results is shown in Figure 2.  Figure 2a shows
the wind velocity contour over the pile in the wind tunnel at a nominal
wind velocity of 30 m/s based on pitot static tube traverses.  Figure 2b
shows the result of a yaw prove survey giving velocity and direction of
wind in the vicinity of the pile.  The increase in air velocity above
the pile needs correction for comparison to free flow above ground.
Figure 2 provides for a general understanding of the flow field around
a pile.
     Detailed typical measurements in Figure 3 show the evolution of the
air velocity profile around the porous pile specimen.  As is readily
seen, Fig. 3 shows a typical turbulent flow profile for the approaching
flow.  Velocities at stations D through G represent clearly the regions
where separation of the flow occurs while velocities at stations G and H
show the reattachment at the boundary layer.  Figure 4 shows the results
of traverse across the pile specimen with the velocity profiles of the
flow at the testing points near the surface.  It should be noticed that
even when the limits of the regions of flow separation cannot be exactly
determined due to the finite number of probe points, these separation
zones appear clearly defined by zero velocity readings at the surface of
the pile as shown in Figure 3.
2.  Windward Barriers
    The use of barriers in front of the pile model appear to be the
simplest way to control wind penetration.   In order to test how effective
this method is, two kinds of barriers of two different heights were
tested (solid and slotted).   The probe was placed, in turn, at stations
3 and 9 (see Figure 1)  which were located in regions where flow penetration
was likely to occur.   At those points, the air velocity was measured at
different distances from the front glass of the tunnel and the barrier
was placed at different distances from the edge of the pile.   The
experimental results for a solid barrier placed in front of the pile
model are shown in Figure 5.
                                   92

-------
              Wind Tunnel
                             'Porous Pile of Haydite Fastened to
                              Wind Tunnel
                         (a)  Air Velocity (Horizontal Component) Over the Porous
                              Pile-Pitot Static Tube Survey
                                                                                                      305 mm
                 37.3 m/s    23.9 m/s    14.4 m/s    18 m/s
19.5 m/s   16 m/s
           -16.7°
      25.4 m/s
      28.6
14.4 m/s
25.7° From
Horizontal
                         (b)  Flow in the Vicinity of Pile-Yaw ProbeSurvey
                       ^Fluctuating
                        Flow Between:
                        -55° and 4.4°
      Figure 2  Wind tunnel measurement of  flow over  and  into  a  76 mm  high  porous pile

-------
                                            100     150
                                                  Bottom
Figure 3  Typical  air  velocity distribution over a porous pile
                              94

-------
8
5
      Separation
        Region
       Separation
         Region
                                                               Station No.
                    Rear
                           20     18    16    14    12     TO    Front
                                           t
                                    Direction of Flow
                      (Positions  in cm from  front glass)
      Figure 4   Mow velocity over the  porous pile  from horizontal  truvorsos
                                        95

-------
      1.0
                     Remarks:   Station  3, Temperature = 24°C

                               Solid  Barrier,  Low Speed,  in Front

                                   Barrier's Distance from Pile
                               D  =
                             ref
                Height of the Barrier

           = 33 m/s
      0.9
                                                                   No  Barrier
 01
n
*
      0.8
      0.7
      0.6
      0.5
      0.4
                                                                  0  =  0
      0.3
                                                                      1.0
      0.2
         1
      0.0
           o   10
12
14
16
18
20
22
                 Horizontal  Distance  from  Front  Glass,  cm


                  Figure 5  Velocity profile in the horizontal plane at
                            various forward barrier locations
                                        96

-------
      It  is  seen that at a distance equivalent to three pile heights
measured from  the bottom edge, the air velocity at point 3 is a minimum.
In other words, the barrier becomes most effective at this distance.
Furthermore, it was found that this distance is optimum for different
wind  velocities and also for the slotted barrier.
      Figure 6  shows the effect of the 25.4 mm solid barrier at three
pile  heights ahead of the pile, the limiting streamline and the velocity
profiles outside of this limiting streamline.  The fluctuating wind
velocity behind the barrier was measured with a hot wire anemomter at
stations 1  (Figure 1) and 2 to be 12.7 m/s and 10.6 m/s, respectively,
for a free  stream velocity of 33 m/s.  This shows the effectiveness of a
forward  barrier.
      Figure 7  shows the limiting streamlines produced by the 13 mm solid
and slotted barriers.  They correspond to a barrier height 1/4 of the
pile  height.
      Figure 8  shows the velocity at station 3 (Figure 1) for various
barriers and locations in terms of number of pile-heights for solid
barriers and 2.5 pile-heights for slotted barriers of heights equal to
1/4 of the nominal pile-height, and three pile-heights for solid and
slotted  barriers of heights equal to 1/2 of the nominal pile-height.
The scaling relation is that the model corresponds to an actual pile of
3 m high under 20 km/hr wind arid a 1.5 m high snow fence reduces the
wind  approaching the pile to 9 km/hr when placed 9 m from the foot of
the pile.  Other pile heights and velocities remain to be tested.
      Reduction in wind velocity also contributes to conserving the
moisture  in the pile, thus improving adhesion of fine particles.
3.  Leeward Barriers
    Expecting the back part of the pile to be a region for flow pene-
tration, the air velocity at station 9 (Figure 1)  was measured with a
barrier placed at various distances from the back edge.   Test results of
the downwind barrier to this stage appears inconclusive, although nondetri-
mental if the windward barriers become downwind due to a reverse in wind
direction.
                                   97

-------
Center-line
                                                            Velocity Profile
                                                            at STA.  B
                   Limiting Streamline
                     Figure 6  Velocity profiles and  limiting  streamline  for  flow  over  a
                               25.4 mm barrier located  at  three  pile  heights  in  front of  the  pile

-------
                               imiting Streamline
13 mm Slotted Barrier
 2.5 Pile Heights
     From Foot
                                      Foot of
                                       Pile
                               Limiting
                               Streamline
            13 mm Solid
            Barrier
           2 Pile Height B
           From Foot
                                      Foot of Pile
         Figure  7   Limiting streamlines for barriers of 1/4 nominal pile
                    height near optimum locations  (velocity along centerline
                    of duct is 31 m/s)
                                        99

-------
                        51mm Pile Height
                25.4 mm Solid  Barrier
                25.4 mm Slotted Barrier
                13 mm Solid Barrier
                13 mm Slotted Barrier
      Barrier Location ,No. of  Pile Heights
      from Foot of Pile
Figure 8  Effect of windward barrier location in
        reducing wind penetration--station 3
                       100

-------
     Figure 9 shows the limiting streamlines of flow separation for the
condition of no barrier (solid line) and the wind velocity distribution,
the use of 25 mm solid barrier (dotted lines for the barrier and the
limiting streamline) and the use of 13 mm solid barrier (dashed lines
for the barrier and the limiting streamline).
     Figure 10 shows the wind velocities at stations 9, 10, and 11 for
25 mm solid barrier at various pile heights from the foot of the pile.

DISCUSSION
     Experimental study to this stage shows that, wind barriers such as
snow fences, when properly deployed, can reduce wind penetration into a
storage pile by one-half.  This reduces blowing of dust as well as loss
of moisture from the pile.  Loss of moisture tends to increase the
amount of fine  dust in loose form.  It appears desirable to have the
length of a pile normal to prevailing wind, with snow fences along both
sides.
     It appears that wind loss of dust from storage piles occur in three
modes (2): blowing due to wind penetration, erosion of the flat top as
an effective exposed area, and during load-in and load-out.  The first
two modes occur over long durations.  Wetting down with water as a dust
suppressant and reducing wind penetration will reduce the loss signifi-
cantly.  The eventual computer program will include estimation of loss
of dust and dispersion according to existing models.  The former concerns
entrainment and blowing of dust from the passages of the porous pile
(Soo 1967, Settari 1975)6'7 and the latter concerns diffusion of dust
plume from nonpoint areas (Neuman 1975)8.  These will also be covered in
future tests.  An input to the emissions factor equations such as in
Cowherd, et al. (1979)2 is expected with the presence of barriers.
     Application of coating or other dust suppressant on storage piles
is also within consideration.  Economic justification and cost effective-
ness of protective devices are essential.
                                   101

-------
       L. S. w/  25 mm
       Barrier
   25 mm Solid
   Barrier
Figure 9    Limiting streamlines for flow downwind of the storage
            pile for the conditions of: without ;i barrier, with a
            I 3 mm solid harrier, and a 25. I mm solid barrier.
            Velocity profiles are for the ca.se without a harrier.

-------
   25
                                                                  STA 9
x
o
o
(U
>
                                                                           STA 11
                    1234



                   D, No of pile heights from the foot of the pile
        Figure  10   Effect  of  leeward  barrier  (25.4 ram  solid  barrier)

                    location on  wind velocity  behind  the pile (STA's  in  Fig.  2)
                                          103

-------
                                 CONCLUSIONS
     To this stage of an experimental program, we conclude, in a prelim-
inary sense, that barriers such as snowfences or height 1/4 to 1/2 of
that of the storage pile, if properly deployed in the windward direction
of the pile at two to three pile heights away, can reduce the wind
penetration and the resulting lift of dust by 25 to 60 percent,not
counting the reduced rate of drying of the materials in the pile.  If
these barriers are located also in the downwind location, no harmful
effect is noted.  A tentative recommendation is to have the pile protected
in the directions of prevailing wind as an inexpensive means of reducing
wind blown dust.
     Continuing study will include the effect of changes in pile heights
and geometry (angles, etc.), pile porosity and amount of fines.  Wind
velocity patterns behind a windward barrier and in front of a trailing
barrier will be studied in detail with a hot wire anemometer.
                                 104

-------
                              ACKNOWLEDGEMENTS
     The authors wish to express their deep appreciation to the co-
sponsors of this research, notably Messrs. N. Plak, R. V. Hendricks, and
J. S. Ruppersberger of the Industrial Environmental Research Laboratory,
U.S. Environmental Protection Agency, and Messrs. E. Kirkendall and
E. A. Veel of the American Iron and Steel Institute.
                                   105

-------
                                 Klil'lJkliNCliS

1.   Cowherd, C., Jr., and K. V. llendricks,  "Fugitive  Emissions  from
     Integrated Iron and Steel Plants—Open  Dust Sources,"  Paper 77-6.2,
     presented at the 70th Annual Meeting of APCA,  20-24  June  1977,
     Toront o, Canada.

2.   Cowherd, C., Jr., R. Bonn, and T. Cuscino, Jr., "Iron  and Steel
     Plant Open Source Fugitive Emission Evaluation,"  EPA-600/2-79-103,
     May 1979.

3.   Nimerick, K. H., and G. P. Laflin, "In-Transit Wind  Erosion Losses
     of Coal and Methods of Control," Soc. of Mining Engr., August  1979,
     pp. 1236-1240.

4.   Soo, S. L., and S. K. Tung, "Deposition and Entrainment in  Pipe
     Flow of a Suspension," J. Powder Tech., Vol. 6, No.  5, 1972, pp.
     283-294.

5.   Carman, P. C., The Flow of_ Gases through Porous Media, Butterworth,
     London, 1956.

6.   Soo, S. L,, Fluid Dynamics of Multiphase Systems, Blaisdell Pub.
     Co., 1967, Chap.  6, p. 411.

7.   Settari, A., and K. Aziz, "Treatment of Nonlinear Terms in  the
     Numerical Solution of Partial Differential Equations for  Multiphase
     Flow in Porous Media," Int. J. Multiphase Flow, Vol. 1, pp.  817-
     844, 197S.

8.   Neuman, J., "Turbulent Diffusion of Pollutants from  Some  Plane Area
     Sources," Atmospheric Environment, Vol. 9, pp. 785-792, 1975.
                                   106

-------
PARTICULATE AND S02 EMISSION FACTORS
    FOR HOT METAL DESULFURIZATION
             Jim Steiner
     Source Evaluation & Analysis
   A Division of Acurex Corporation
           485 Clyde Avenue
   Mountain View,  California 94042
            B.  J.  Bodnaruk
 U.S.  Environmental  Protection Agency
          230  South  Dearborn
       Chicago,  Illinois  60604
                107

-------
INTRODUCTION

      The Source Evaluation and Analysis Division (SEA) of Acurex
Corporation undertook a series of tests for EPA Region 5 at the hot metal
desulfurization plant of Kaiser Steel located in Fontana, California.  The
tests were performed to develop emission factors (particulate mass,
particle size, SOg) for uncontrolled and controlled emissions from this
process.

PROCESS DESCRIPTION

      Kaiser Steel owns and operates an external hot metal desulfurization
(HMDS) plant at its Fontana Works located in Fontana, California.  The
plant began operation in the spring of 1978 and uses technology based on
the Nippon Steel Corporation desulfurization process.  Hot metal from the
blast furnace arrives at the HMDS plant in torpedo cars which are rolled
into a partially open shed attached to the HMDS building.  Lances with
stoppers are inserted into as many as three torpedo cars at one time and a
predetermined amount of calcium carbide (CaC^) and calcium carbonate
(CaCOs) is blown into the hot metal using nitrogen (200 cfm at 30 to 40
psi).  The calcium reacts with the sulfur in the metal to form a slag
which floats to the top of the torpedo car.  The hot metal is usually
desulfurized to <0.03 percent sulfur but can be reduced to as low as 0,003
percent sulfur for special low sulfur steel production.  The stopper on
the lance fits into the mouth of the torpedo car to minimize emissions
during the desulfurization process which lasts approximately 8 to 16
minutes.  Emissions which do escape around the stopper are captured by a
local hood and ducted to a six compartment, positive pressure
Wheelabrator-Frye baghouse.  The design capacity of the baghouse is
100,000 scfm and it contains 1728 Dacron bags, each 5.25 inches in
diameter and 14 feet 3 inches long.  The air-to-cloth ratio of the unit  is
3:1 at a design pressure drop of 7 to 8 inches w.g.  Bag cleaning is
accomplished by mechanical shaking and the cleaned gases are exhausted to
the atmosphere via six stacks — one for each compartment.  The dust
collected by the baghouse (~2 tons per day) is trucked to landfill for
disposal.  After desulfurization is complete the torpedo cars are taken  to
the skimming station of the Basic Oxygen Plant (BOP) where the slag is
skimmed from the hot metal.  This hot metal is now suitable for use in the
Basic Oxygen Furnace (BOF).

SAMPLING LOCATIONS

Uncontrolled Emissions Location

      Samples of the uncontrolled emissions from the HMDS process were
collected at the inlet to the baghouse as indicated in Figure 1.  The
particle size sampling ports (six total, three equidistant ports on each
                                    108

-------
                         DeMsMng station
                         duct
Hot metal
desulfurlzation
station hood
                              r
                                    Damper
                                                                           'Pacific exhauster' fan
                            Reliance 200 HP
                            motor
                                       DeMshlng
                                       station
                                         .uct
                                     6"
                                     sampling ports
                                                                  4"
                                                                  sampling ports
  \
Transition
niece  In duct-
work	
                  Figure 1.    Uncontrolled emissions sampling port locations.
                                                   109

-------
vertical side of the rectangular inlet duct which was 63 inches high and
72 inches wide) were located 24 inches downstream from the beginning and
72 inches upstream from the end of the transition piece in the ductwork.  The
particulate mass sampling ports (three total, each port equidistant on one
vertical side of the rectangular inlet duct which was 57 inches high and
80 inches wide at this point) were located 48 inches downstream of the
beginning and 48 inches upstream of the end of the transition piece in the
ductwork.  Particle size and mass measurements were made at nine sampling
points at each of these locations (three points per sampling port).

Controlled Emissions Location

       Samples of the controlled emissions from the HMDS process were
collected at the outlet of the baghouse as indicated in Figure 2.  The
particle size sampling ports (one 6-inch diameter port located at 45° to the
mass sampling ports on stacks 2 and 5 which were 30 inches in diameter) were
located 45 inches downstream of the top of the baghouse and 27 inches upstream
of the stack exits.  The particulate mass sampling ports (two 4-inch diameter
ports located at 90° to each other on stacks 2 and 5 which were 30 inches in
diameter) were also located 45 inches and 27 inches downstream of the baghouse
and upstream of the stack exits respectively.  S02 tests were also conducted
in these sampling ports.  Particulate mass measurements were made at twelve
sampling points along two diameters of each stack.  Particle size measurements
were made at a single point of average velocity as were the S02 measurements.

SAMPLING EQUIPMENT

Particulate Mass

       All particulate mass measurements were made with the Acurex High Volume
Stack Sampler (HVSS) which is an EPA Method 5 sampler.  Figure 3 illustrates
the heated oven containing a 3 ym cyclone and a 142 mm filter holder.  The
cyclone was used for all the inlet tests to capture the large, abrasive
particles greater than 3 ym in size to prevent damage to the glass fiber
filter in the filter holder.  No cyclone was used for the outlet tests, just
the filter holder.

Particle Size

       Two different particle size sampling devices were used to measure the
size distribution of the uncontrolled and controlled emissions from the HMDS
process.  Uncontrolled emission size distributions were measured with a SoRI
in-stack, 2-cyclone train (15 ym and 2.5 ym) with a 2.5-inch backup glass
fiber filter.  Controlled emissions were measured with an Andersen Mark III
in-stack cascade impactor equipped with a 15 ym cyclone precutter with
straight nozzles.  All filter media was Reeve Angel 934 AH.  Figures 4 and 5
illustrate these devices.

       Since both of these particle sizing devices were used in-stack, these
devices were mounted directly on the end of a 5-foot long stainless steel
probe.  The remainder of the sampling train components were identical to a
Method 5 train.
                                    110

-------
      Reliance ZOO hp TEFC
      460V
01
1 II IJJ


	 1
>
_l 11 II-
~T

                      CD    CD    O.)
V
Isolation
joint
   Figure 2.    Controlled emissions sampling  port locations.

-------
Figure 3.   Cyclone and filter holder in heated oven of Method 5  train

-------
Figure 4.   SoRI in-stack 2-cyclone train with 2.5-inch backup filter.

-------
Figure 5.    Andersen Mark III cascade impactor with 15 ym cyclone precutter,

-------
 Sulfur Dioxide

        An EPA Method 8 sampling train was used to sample the controlled HMDS
 process emissions for S02 to maximize the volume of gas sampled since the
 concentration of $03 was expected to be very low.

 SAMPLING PROCEDURES

 Particle Size Device Comparisons

        Since the SoRI 2-cyclone train was a relatively new sampling device,
 duplicate comparison tests of the SoRI 2-cyclone train and the Andersen Mark
 III impactor were carried out at the inlet location prior to the test
 program.   Each device was compared against the other to determine if there was
 any difference in the performance of these devices.   For example, the SoRI
 2-cyclone train was inserted into one sampling port on one vertical side of
 the inlet duct while the Andersen Mark III impactor was inserted into the
 sampling  port on the opposite vertical side of the inlet duct.   Hence,  both
 devices were in the same measurement plane 6 inches apart (to minimize
 interference effects of one device on the other but close enough to see the
 same particle concentration and size distribution).   The results of these
 comparison tests are presented in Table 1 and indicate reasonable agreement
 between the devices.

        The repeatability of each measurement device was also determined  in a
 similar manner.   For example, two Andersen Mark III  impactors were positioned
 in  the  duct work 6  inches apart and  were  used to  simultaneously sample  the
 uncontrolled emissions.   The samples were recovered  and weighed in an
 identical  manner to  determine each device's  measurement repeatability.   The
 results of these precision  tests are presented  in Figures  6  and 7  and indicate
 good  measurement repeatability is  possible with careful  operation  of  each
 device.

 Particulate  Mass

        The  participate mass  tests  were  basically  conducted in accordance with
 EPA Method  5 procedures.  Sampling involved  careful  timing and  coordination
 since particle  size  and mass  determinations  were  made  simultaneously  and were
 dependent  on the  desulfurization time  (varied with the  number of  torpedo cars
 to  be desulfurized,  the degree  of  desulfurization required,  and  the start  of
 each  torpedo car  desulfurization).   Since  the time of desulfurization varied
 from 5  to  18 minutes, each  sampling  point  in  the  inlet  duct  (nine  total) was
 sampled for  30  seconds for  a  total sampling  time  of  4.5 minutes.   The volume
 of  gas  sampled per test varied  from  10 to  15  ft3.  It should  be  noted that
 the inlet  test  location was  under  considerable  negative pressure (upstream of
 baghouse fan) and the sampling  probe was  inserted into the duct -15 seconds
 before  the  start  of  a test with  the  vacuum pump drawing ~0.05 cfm  through the
 train to prevent  the filter paper  from lifting off the support  screen and
 possibly tearing  and to prevent  the  impinger  liquids from being sucked forward
 through the  train.  At the completion of a test,  the flowrate through the
 train was again reduced to ~0.05 cfm as the probe was withdrawn from the
duct.  The pump was then shut off when the probe was outside the duct.
                                    115

-------
TABLE 1.   COMPARISON TEST RESULTS FOR PARTICLE SIZING DEVICES
Particle Size
Device
SoRI 2-cyclone train
with 2.5 inch backup
filter
Andersen Mark III
impactor with 15y
cyclone precutter
Volume of
Gas
Sampled
(dscf)
4.375
4.237
0.859
0.857
15 ym
Less than
Stated Size
(%) (gr/dscf)
31.7 0.733
34.4 1.120
27.1 0.763
32.2 1.090
2.5 ym
Less than
Stated Size
(%) (gr/dscf)
13.3 0.308
12.6 0.410
11.3 0.318
14.3 0.484
Total
Part icu late
Mass
Concentration
(gr/dscf }
2.313
3.256
2.815
3.385

-------
    20
    10
     9
u
i

5


I
   1.0
    .9


    .7
    .5
                                
-------
      20
      10
       9

       7


       5
      1.0
       .9
                       
-------
       At the outlet test  location, several desulfurizations  had  to  be  sampled
consecutively to collect enough particulate matter for  sample evaluation.
Hence, each point on two traverses  (IE points total) was  sampled  for  1  minute
for a  total sampling time  of  20 to  24 minutes.  The volume  of gas sampled  per
test varied from 47 to  75  ft3.  It  took a minimum of two  days to  complete
one test at the baghouse outlet.  The outlet test location  was  at a  slight
positive pressure and was  sampled in a normal manner.   Only stacks 2  and 5
were tested simultaneously (for economical reasons) for particulate  mass and
size.

Particle Size

       As a result of the  comparison tests, it was shown  that the performance
of the devices (SoRI 2-cyclone train, Andersen Mark III)  were equivalent.
Hence, the SoRI 2-cyclone  train was used for all inlet  tests  (high grain
loadings) and the Andersen Mark III impactor with a 15  ym cyclone precutter
was used for the outlet tests (very low grain loadings).

       Prior to conducting a  particle size test, each device  (SoRI 2-cyclone
train and Andersen Mark III impactor) was thoroughly brushed  and  cleaned with
reagent grade acetone.  In addition, both the cyclones  and  the  impactor were
cleaned in an ultrasonic bath (using a liquid cleanser) after every  second or
third test.

       The sampling procedures for the particle size sampling trains  were
identical to those of the  particulate mass sampling trains  except for the
number of points sampled during a given desulfurization.  Tests conducted  with
the SoRI device consisted  of  sampling in one sample port  for  equal times
(2.0 minutes) at each of the  three sampling points.  Approximately 2.5  ft3 .
of gas was sampled per  test which provided a more than  adequate amount  of
particulate matter in each collection stage.  Three points  per test were
selected to minimize the number of discrete tests required  to traverse  the
entire duct based on the preliminary velocity measurements  which  indicated
very little horizontal  stratification.  All three points  in all three sample
ports were sampled repeatedly to account for particulate  stratification (in
the vertical direction) over  the entire duct.  The tests  conducted with the
Andersen Mark III impactor at the baghouse outlet were  done at a  single point
of average velocity in  stacks 2 and 5 since the preliminary measurements
indicated very uniform  profiles in a given stack although velocities
varied from stack to stack.   Since the outlet grain loadings  were  so  low
(<0.01 gr/dscf), several desulfurizations were tested in  order to  collect
5 to 6 ft3 of gas for a weighable sample.

RESULTS

Uncontrolled Emissions

       Table 2 summarizes  the sampling data obtained with the EPA  Method 5
particulate mass train.  Table 3 provides a summary of  the  HMDS production
data obtained during these  tests.   The emission factor  data (based on Method 5
measurements) are dependent on the degree of sulfur removal (higher for larger
sulfur removal) from the hot metal but does not appear  to correlate with the
rate of desulfurization; the  desulfurization agents injection rate used
                                    119

-------
TABLE 2.    INDIVIDUAL PARTICIPATE MASS TEST SAMPLING  DATA
           (Uncontrolled  Emissions)
Test
Number
14
18
19
20
22
23
25
26
27
28
Volume
of Gas
Sampled
(dscf)
14.064
14.340
11.854
14.514
13.314
14.154
13.960
12.912
11.710
10.908
Parti cul ate
Mass
(mg)
631.54
2,705.84
2,805.15
696.11
2,129.50
2,549.61
622.60
1,028.64
2,290.14
2,294.71
Particulate
Concentration
(gr/dscf)
0.6930
2.9919
3.6519
0.7401
2.4683
2.7798
0.6882
1.2294
3.0191
3.2323
Particulate
Emission
Rate
(Ib/hr)
605
2,266
2,633
605
1,862
2,151
547
904
1,977
2,017
                          120

-------
TABLE 3.   HMDS PRODUCTION DATA (Uncontrolled Emissions)
Test
No.
14
18
19
20
22
23
25
26
27
28
Desulfurization
Time
(mln)
5.50
9.53
9.97
5.45
5.45
6.93
13.55
10.40
9.03
5.97
12.53
13.48
14.02
19.00
8.97
7.35
9.23
8.75
6.60
Desulfurization
Agents
CaC?
(lb)
333
1063
1033
470
316
613
1043
814
565
342
1006
1095
1191
2073
691
551
570
558
314
CaC03
(lb)
32
33
46
35
26
51
46
42
52
61
41
53
113
100
58
96
64
87
40
Participate
Mass .
Emission
Rate
(Ib/min)
10.08
37.77
43.87
10.08
31.07
35.85
9.07
15.08
32.95
33.61
Average
Desulfurization
Rate
( tons/mi n)
21.82
12.31
20.19
8.86
12.87
9.82
6.37
11.15
14.54
15.18
Participate
Emission
Factor
(Ib/ton)
0.46
3.07
2.17
1.14
2.41
3.65
1-.42
1.35
2.27
2.21
i
                       121

-------
during these tests was much higher (average 83.3 Ib/min) than 50 to 65
Ib/min cited by Kaiser and the CaC2 usage averaged 6.4 Ib/ton hot metal
to obtain an average sulfur reduction of 65.2 percent compared to the
cited 3.6 Ib CaC2/ton hot metal for 60 percent sulfur removal.

Controlled Emissions

      Table 4 sunmarize the sampling data obtained with the EPA Method 5
particulate mass train.  Table 5 provides a surmiary of the HMDS production
data obtained during these tests.  The emission factor data (based on
Method 5 measurements) are dependent on the degree of sulfur removal
(higher for larger sulfur removal) from the hot metal; the average
desulfurization injection rate was 83.9 Ib/min and the CaC2 usage
averaged 3.95 Ib/ton hot metal to obtain an average sulfur removal of
50 percent.

SUMMARY OF TEST RESULTS

Emission Factor Data

      The average emission factor (based on Method 5 measurements) for
uncontrolled HMDS process emissions was 1.09 j^O.44 Ib/ton for one torpedo
car (TC), 2.53 +0.47 Ib/ton for two TC's, and 2.74 +0.79 Ib/ton for three
TC's.  Based on the particle size measurements made with the SoRI
2-cyclone train, approximately 25 percent of these particles are less  than
15 ym in size and 12 percent are less than 2.5 urn in size.

      The average emission factor (based on Method 5 measurements for
controlled HMDS process emissions was 0.009 +0.003 Ib/ton.  Based on the
Andersen impactor measurements, the Dp5Q for these emissions was 3.4 ym.

      The average mass removal efficiency of the baghouse based on
Method 5 measurements (not simultaneous) was 99.36 percent and no SOg
was detected in the baghouse exhaust.

Comparison of Sampling Train Data

      Tables 6, 7, 8 and 9 compare the Method 5 train test Results with
those obtained by the particle size trains.  All inlet particulate mass
tests traversed the entire inlet duct; all particle size tests were done
at three points only as follows:  tests 18, 19, and 26 were done in the
top sampling port (points 1, 2, 3); tests 20, 23, and 25 were done  in  the
middle sampling port (points 4, 5, 6); tests 22, 27, and 28 were done  in
the bottom sampling port (points 7, 8, 9); Method 5 train  (front half
only) concentrations were 21 percent greater than the SoRI train
concentrations for the top port, 8 percent greater for the middle port,
and 23 less for the bottom port; there is obviously particulate
stratification from top to bottom in the inlet duct.

      All outlet particulate mass tests traversed both baghouse  stacks;
all particle size tests were done at a single point of average velocity;
in all but one case, the Method 5 particulate mass concentrations were
                                    122

-------
TABLE 4.   INDIVIDUAL PARTICIPATE MASS TEST SAMPLING DATA
           (Controlled Emissions)
Test
Number
15
16
17
Volume
of Gas
Sampled
(dscf)
65.909
75.296
47.718
51.900
57.159
57.355
Par ticu late
Mass
(mg)
31.09
104.54
33.71
71.75
17.84
60.20
Par ticu late
Concentration
(gr/dscf)
0.0073
0.0214
0.0109
0.0213
0.0048
0.0162
Participate
Emission
Rate
(Ib/hr)
0.96
2.65
1.31
2.70
0.63
2.17
                          123

-------
            TABLE  5.  HMDS PRODUCTION DATA  (Controlled  Emissions)
Test
No.
15
16
17
Desulfurization
Time
(rain)
6.30
4.48
6.40
6.37
6.32
7.02
7.42
6.97
6.97
8.48
7.77
8.13
3.77
6.22
4.83
5.93
6.73
7.13
4.72
4.13
4,12
3.83
3.73
4.12
Desulfurization
Agents
CaC?
Ob)
475
263
409
413
426
464
628
568
595
929
736
774
296
475
356
537
562
540
391
216
268
300
266
254
Catth
Ob)
34
28
75
36
35
51
35
49
59
34
47
62
33
44
31
34
47
63
32
26
60
33
33
30
Participate*
Mass
Emission
Rate
(Ib/min)
0.180
0.202
0.132
Average
Desulf urination
Rate
(tons/min)
17.60
18.31
29.01
Participate
Emission
Factor
(Ibyton)
0.010
0.011
0.005
i
Computed using average Method 5 concentration  for two stacks and  average
 volumetric flowrate from inlet tests
                                      124

-------
  TABLE 6.   COMPARISON OF METHOD 5 AND SoRI TRAIN DATA (Top Port)
Sampling
Device
Acurex Method 5 high
volume stack sampler
SoRI 2 -eye lone train
with 2.5 inch backup
filter
Sampling
Location
(Inlet Duct)
All 9 points
Points 1,2,3
Volume of Gas
Sampled
(dscf)
14.340
11.854
12.912
2.550
2.694
2.785
Particulate Mass
Concentration
(gr/dscf)
2.9119
3.6519
1.2294 ,
2.3575
2.8104
0.9740
TABLE 7.   COMPARISON OF METHOD 5 AND SoRI'TRAIN DATA (Middle Port)
Sampling
Dev i ce
Acurex Method 5 high
volume stack sampler
>
SoRI 2-cyclone train
with 2.5 inch backup
filter
Sampling
Location
(Inlet Duct)
All 9 points
Points 4,5,6
Volume of Gas
Sampled
(dscf)
14 .'5 14
14.154
13.960
2.679
2.651
2.749
Particulate Mass
Concentration
(gr/dscf)
0.7401
2.7798
0.6882
0.4496
2.6186
0.8102
                                125

-------
TABLE 8.   COMPARISON OF METHOD 5 AND SoRI TRAIN DATA (Bottom Port)
Sampling
Device
Acurex Method 5 high
volume stack sampler
SoRI 2-cyclone train
with 2.5 inch backup
filter
Sampling
Location
(Inlet Duct)
All 9 points
Points 7,8,9
Volume of Gas
Sampled
(dscf)
13.314
11.710
10.908
2.644
2.742
2.722
Participate Mass
Concentration
(gr/dscf)
2.4683
3.0181
3.2323
2.8942
2.9927
5.3917
        TABLE 9.    COMPARISON OF METHOD 5 AND ANDERSEN MARK III
                   IMPACTOR DATA (Stacks 2 and 5)

Sampling
Device
Acurex Method 5 high
volume stack sampler




Andersen Mark III
imp actor with 15 ym
cyclone precutter



Sampling
Location
(Outlet Stacks)
All 12 points





Average point





Volume of Gas
Sampled
(dscf)
65.909
75.296
47.718
51.900
57.159
57.355
5.868
6.030
6.265
5.435
6.249
5.983
Parti cul ate Mass
Concentration
(gr/dscf)
0.0073
0.0214
0.0109
0.0213
0.0048
0.0162
0.0078
0.0060
0.0054
0.0084
0.0044
0.0066
                               126

-------
considerably greater (average 50 percent) than corresponding impactor
concentrations: the volumetric flowrates through stacks 2 and 5 showed
very little difference, but the mass concentrations were significantly
different (stack 5 always greater than 2).
                                    127

-------
                   DEMONSTRATION OF THE USE OF CHARGED FOG
                         IN CONTROLLING FUGITIVE DUST
                     FROM LARGE-SCALE INDUSTRIAL SOURCES

                                     by

                             Edward T. Brookman
                              Project Engineer

                     TRC-Environmental Consultants, Inc.
                           Wethersfield,  CT  06109
                                  ABSTRACT

    A unique  device  for the control of  particulate emissions works on  the
principle that  most  industrial pollutants  acquire  an electrostatic charge
as they  are dispersed into the air.   If  this charged airborne material  is
exposed  to  an.oppositely charged  water fog, the charges act to enhance  the
contact  between the  particulates and  the  fog droplets,  resulting in  rapid
agglomeration and  particle fallout.   A  device that  generates  charged  fog
has now  been substantially developed  and  is being  offered commercially  by
The Ritten Corporation.

    TRC-Environmental Consultants, Inc. has been contracted by EPA/IERL/RTP
to test  the Ritten Corporation's  Fogger  IV on several large-scale fugitive
dust sources.   This  paper  discusses the initial test at  a sand and gravel
operation  and  presents  preliminary  test  results  in   terms  of  percent
reduction in TSP.  The changes in fogger effectiveness due to variations  in
operational parameters  are discussed.   The  initial tests indicate overall
fogger efficiencies of approximately 70 percent.
                                      129

-------
                   DEMONSTRATION OF THE USE OF CHARGED FOG
                         IN CONTROLLING FUGITIVE DUST
                     FROM LARGE-SCALE INDUSTRIAL SOURCES
INTRODUCTION

    A spray.of  fine water droplets is a  well-known means of airborne  dust
removal.   Various  types  of  scrubbers  rely  on  water  droplets  to  sweep
particles from  the  inlet  gasses and water sprays are  often used  in  mining
and material  handling  for  dust suppression.   Unfortunately,  water  sprays
are not very efficient in removing dust from the ambient air.

    One means of  improving the  efficiency  of water  sprays is by applying  a
charge to the spray  that  is opposite  in  polarity  to the charge of  the  dust
to be  suppressed.  It has been found that  most  industrial pollutants and
naturally occurring  fugitive  dusts  acquire an electrostatic charge as  they
are dispersed into  the  air.   If this charged, airborne material is exposed
to an oppositely  charged  water  spray there is enhanced contact between the
particulates and  the water  droplets.  After  contact  is  made,  the  wetted
particulates agglomerate rapidly and fall out of the atmosphere.

    The  charged  sprays  can  be further  improved  by  atomizing  the  water
droplets so that  a  fog  is produced.   The  fineness of the  particles enhances
the  charge  carrying  capabilities   in   the   spray.    Furthermore,  Hoenig
(1977)l  has  demonstrated  that  the  greatest  effectiveness is obtained  when
the water  droplets  are of  a  similar size  to the dust particles  to  be
controlled.  Lastly, less water is  required when fog is used,  thus reducing
operating costs.

    A device capable of producing  this fine spray and applying a  charge  to
it is known as a  charged fogger.  A charged  fogger  uses a nozzle to produce
the fog, an induction ring to charge  the fog as it  leaves the  nozzle,  and a
fan to  transport  the fog  to  the dust source.   The  operation of the  fogger
requires a  water  supply,  a pressurized  air supply, and  power.    There are
several  such devices  on  the market,  tailored to the  size  and  type  of
industrial application.

    The charged fogger  is intended  primarily  for  fugitive dust sources that
cannot  reasonably be controlled  via  conventional  means  such  as  hooding.
Such  sources include  materials handling  operations  (transfer  points and
conveyors),  truck  and  railroad  car  loading and unloading,  front end
loaders, ship loading, grain  silos, and mining operations.  The charged fog
concept has  been  applied  to  operations  as small  as a  hand  grinder and  as
large as a quarry.

    Although the  charged  fog  concept has been widely applied  to industrial
souces of fugitive  dust,  little data is available  regarding fogger control
efficiency.  To  obtain such  data,  the  Industrial Environmental  Research
Laboratory  of   the  Environmental  Protection  Agency  at  Research   Triangle
Park,    North   Carolina,    (IERL/EPA/RTP)    contracted    TRC-Environmental
Consultants, Inc.  (TRC) to conduct  a full scale demonstration of  a charged
                                       130

-------
fogger  on several  appropriate  industrial  fugitive emission  sources.   In
particular/   IERL/EPA   was   interested  in  testing  the   largest  fogger,
designated "Fogger  IV",  manufactured by the Ritten Corporation of Ardmore,
Pennsylvania,  on several sources  within the  iron  and  steel  and  sand  and
gravel industries.

    The   sources   considered  for  testing  included   materials   transfer,
conveying, grinding,  crushing,  and  truck  and  railroad  car  loading  and
unloading.   The requirements  for a  test  site  included  isolability  from
other   dust   sources,    availability   of   necessary   utilities,   relative
difficulty of control  by other  methods,  representativeness to the general
industry, relatively continuous operation, and  fairly heavy  dust  production
to facilitate sampling.

    Following numerous visits  to iron and steel and sand and  gravel sites,
several  suitable  sources  were   selected  for  field  testing  the  charged
fogger.   The  source chosen  for the first test was the primary rock crusher
operation at a sand and gravel site in Connecticut.

DESCRIPTIONS OF SITE AND TEST EQUIPMENT

    Figure 1  is a plot  plan of  the primary crusher  operation showing  the
locations and dimensions of  the various structures.   Descriptions of  the
site and test equipment are given in the following subsections.

Test Site

    The inital  fogger test  site  was a primary  rock crusher.   Approximately
100 dump  trucks per day,  each carrying  loads  of approximately  45 Mg  (50
tons) of  quarry rock  (basically basalt) mixed with  dirt,  back  up to  the
crushing pit  to unload.  Unloading times  vary  from  30-60 seconds,  depending
on conditions in  the  pit.   The  pit  itself is  roughly  8 meters long and  6
meters  wide.   The crushing  is  done  by a  Superior  4265  gyrotory  rock
crusher.  There is a two story computer control building to the  north  side
of the crushing pit, a  control shed  to the  east,  and  a large  paved area to
the south side.  All approach roads  and areas around the buildings and  pit
are  paved  and  kept   reasonably  clean  through   frequent  sweepings   and
waterings.

    Fugitive   dust  emissions   result   from   the  dumping   and   crushing
operations.   The  truck  unloading  is  the primary source of dust with  the
major portion coming  from dust  boil-up  at  the rear of  the pit.   There  is
also dust at the rear of the  truck during the dump.  The crushing  procedure
itself also  produces dust,  but to a  much  lesser  degree than  the  unloading
process.

Charged Foggers

    Two identical  foggers were specially designed  for  TRC  and EPA by  the
Ritten Corporation of Ardmore, Pennsylvania.   Ritten's standard Fogger  III
was modified  and  upgraded  in order to  allow for  variations of  parameters.
The final configuration, designated  "Fogger  IV",  is shown schematically  in
Figure 2.
                                      131

-------
    /v
           12m
COMPUTER
 CONTROL
 BUILDING
   PRIMARY
   CRUSHING
      PIT
BREAKER ARM

  PRIMARY
  CRUSHER
        DUMP
        TRUCK
           15m
                    CEMENT
                    BLOCKS
                   PAVED AREA
                       EDGE OF
                     EMBANKMENT
                        CRUSHER
                        CONTROL
                          SHED
                      CRANE FOR
                      DISLODGING
                      JAMS AND
                      REMOVING
                      OVERSIZE
                      MATERIAL
                       12m
            Figure 1.   Primary crusher plot plan.
                       132

-------
co
co
                                                    BELT GUARD
                                                           ^JUNCTION BOX  (MOTOR)
                                                           1OKI

                                                                      22.9 cm
                                                                            5HP MOTOR
                                                                                BELT DRIVEN CENTAX1AL FAN

                                                                         -48.3  cm
                                                                              WEATHER-PROOF CONTROL
                                                                                 PANEL ENCLOSURE
                                                                                  CONTROL PANEL
   15.2 cm INDUCTION RING
    AIR ATOMIZING  NOZZLE
          NOSECONE
5 * \    WATER  LINE
       COMPRESSED AIR LINE

        HIGH VOLTAGE  LINE
                             UTILITY BASKET
      ONTROL BOX
          AIR  AND WATER INPUT
           CONNECTION PORTS
         230 VAC RECEPTACLE
       230 VAC MAIN
     DISCONNECT SWITCH

          CONTROL CABINET


f-LIFTING  EYE  FOR SKID JACK
                                       Figure 2.     Schematic of the Ritten  Corporation's  Fogger  IV.

-------
    In  the  generation of  the  charged fog  by  the  Fogger  IV,  water  is
atomized via a compressed air supply and ejected  from  a  nozzle.   As the fog
leaves  the  nozzle,  it passes  through an induction  ring where either  a
positive  or negative  charge,  depending   on  the  nature  of  the  dust,  is
applied  to the  spray.  A  flow of air around the nozzle,  provided by  a
centaxial fan, projects the  fog towards the dust source.   A  control panel,
located  on  the  back   of   the   fogger,  allows  for  fogger  operation  and
parameter  variability.  A   schematic  of  the  control  panel  is  shown  in
Figure 3.

    The requirements for and capabilities  of  the  operational  parameters  are
as follows:

    •  Air supply to nozzle - A compressed air supply of 5.6 -  8.8 kg/cm2
       (80-125 PSI)  is required.  For the  tests the air  was supplied by a 2
       hp compressor.  The air  flow through the nozzle is  variable  from 0 -
       11.3 mVhr (0-400 SCFH).

    •  Water flow  - The water  supply  to  the fogger  should  be around  3.5
       kg/cm2  (50 PSI) which is typical  "shop" water pressure.  The water
       flow through the nozzle  is variable from 0 - 151 1/hr (0-40 GPH).

    •  Power - The  foggers  require a  power supply  of  230  volts,  single
       phase, 60 Hz.  The current requirements do not exceed 35 amps.

    •  Centaxial fan - The  fan,  driven by  a  5  hp explosion-proof  motor,
       operates  at  a   maximum  of  79  m3/min  (2800  SCFH).   The  maximum
       output air velocity  is  approximately  3048 m/min   (10,000  FPM).   The
       fan flow rate is variable from 0 - 100% of capacity.

    9  Charge per drop -  Assuming an average droplet  size of approximately
       60  ym,  the  average   number  of  elementary charges  per  droplet  was
       calculated to be approximately 8 x 101*  for 75 1/hr  (20  GPH)  water
       flow.

    •  Flow spectra  -  Two  different flow  nozzles were used for  the tests,
       both  manufactured  by Delavan   in  Des  Moines,  Iowa.   One  nozzle
       produced a conical spray of droplets estimated  to be in the  50  - 70
       ym  size range . while  the other  had  a  heavier  flow  capacity  and
       produced a conical spray of droplets estimated  to be in the  60  - 80
       ym  size range.   A  third type  of  nozzle,  which  produces  a  flat
       spray, was not  yet available for  these  tests, but  will be  used  at
       subsequent locations.

    The  two  foggers were  tested at  various locations  around  the pit  to
determine  the  arrangement for  optimum dust  control.   Placement was  also
dependent on wind direction.  The exact positions are  described in  the next
section.

Sampling Equipment

    The  equipment   used   for   particulate   measurements  included   seven
hi-volume  samplers  and  a   wind   recording   system.   The  hi-vols   were
manufactured by  Misco  Scientific  and  had automatic flow control.   This
                                      134

-------
01
                21.6
                 cm
                            400
                                                  110 VAC YELLOW INDICATOR LIGHT-


                                           230 VAC RED PUSHBUTTON SWITCH-

                                    230 VAC BLACK PUSHBUTTON SWITCH-
                                                              -230 VAC  GREEN  INDICATOR LIGHT
 GPH
WATER
   40
                                                    | ON j OFF]
                       POWER

                       INDUCTION
                         RING
              COMPRESSED
                 AIR
ON      OFF
FAN      POWER
                                                      ON { OFF
                                         OFF
                                                       WATER
                                    INDUCTION
                                      RING
SHEAR
 o
 PIN
                                                       POWER
                                                        O
                                                        ON
             FAN SPEED
                                                                                                         A.C.  VOLTS
                                                                           76.2 cm
                                                                                     L
                                                                                                                             OHMS
 VARIABLE
TRANSFORMER
                                                   INDICATION LIGHTS FOR THE
                                                   PARAMETRIC MOTOR SPEED CONTROLLER
                                            Figure  3.    Schematic  of  the  Fogger  IV  control  panel

-------
enabled  the mass  flow  rate  to  be  held  constant  irrespective  of  filter
loading,  atmospheric conditions,  and  line  voltage  changes.   Two  of  the
hi-vols were fitted  with Anderson Model 7000 Size  Selective Inlets  (SSI's)
which are designed to remove all participates larger than  15 pm from  the
sampled  air before  filtering  the remaining  participates onto a  standard
hi-vol  filter.   Two  other  hi-vols   were  fitted  with  Sierra  Instruments
Series 230  4-stage cascade  impactors  (CI's).   By using the SSI's and  CI's,
the charged  fogger efficiency  could be  examined for various  particle size
ranges.

    The  wind velocity  and direction  measurements  were  recorded  using  a
Climatronic Mark  III Wind  system.  Wind  speed is  measured  with  a  3-cup
anemometer coupled to a  light chopper.   The chopper output is converted to
DC voltage  and  recorded on a chart.   The wind  direction  is  measured by  a
wind  vane  coupled to  a  precision  low-torque  potentiometer.   The  wiper
voltage of the potentiometer is  recorded on another chart.

    The  hi-vols  were  positioned  at  various  locations  and  in  various
combinations around  the pit,  depending on  wind  direction.   The  sampling
array for each test is described in the next section.

TEST PROGRAM AND PROCEDURE

    The  test program consisted of 32  runs during  6  days of  testing.   The
test conditions  are  presented  in  Table 1 and  the equipment  positions  for
each  set  of runs are  shown  in Figure  4.    Conditions  at  the  crusher
prevented  extensive  parameter  variations.   Water  was  provided  by  a  tank
with a small pump  which  limited  nozzle flow to  approximately  80  1/hr.   Fan
speed  was  reduced  to  80% of   capacity  to  help  reduce  excessive  dust
reentrainment in the pit.

    The  sampling  procedure was  essentially  the same  for each test.   Upon
arrival at  the test  site the wind recording  system was set up and  the wind
direction  determined.    The hi-volume  samplers  were then positioned in  a
sampling  array downwind  of the  crushing  pit.   The foggers were  positioned
to control the  dust cloud  while not spraying  directly into  the samplers.
Once the  equipment was  positioned, the pre-weighed hi-volume  filters  were
placed into the  samplers.   The  samplers were then  turned  on simultaneously
just prior  to  the first truck dump  of a  predetermined sequence of  trucks
(typically  8 trucks provided  sufficient  material  for sampling  purposes).
For the runs with the foggers in  operation, the  foggers were also turned on
at this  time and adjusted  to the predetermined fogger operational parameter
conditions.  After the  last truck of the  sequence  had dumped  into  the  pit
and crushing was  completed, the  samplers  and  foggers were all stopped  and
the  filters removed.   At  the  end of  the  day, all  of the  filters  were
returned   to  TRC's  chemistry   laboratory where   they   were   subsequently
desiccated and weighed.

PRELIMINARY RESULTS AND  DISCUSSION

    The  majority  of the test  runs at  the primary crusher,  numbers  7-31,
were  completed  before  the final filter  weights  were  available from  the
chemistry  laboratory.   Upon examining  the data, several important  factors
came to  light.   in  almost  all  cases,  the TSP  levels,  as  measured by  the
various  samplers,  showed  increases  above  the uncontrolled  levels  when
                                       136

-------
Table 1.   TEST CONDITlQMS - PRIMARY CRUSHER

Run Aabient
Designation Equipment • Temp.
No. Positions Date Time |°C)
7 Fig. 4a 10-13-80 0938-1000 9
8 j 1 1050-1129 10
9 1 1 1300-1381 13
10 1 \ 1326-1350 13
11 T t 1355-1434 13
12 Fig. 4b 10-14-80 0833-0915 6
13 II 0933-1005 6
1< || 1022-1050 7
IS 11 1059-1125 7
16 | 1 1245-1305 9
17 f| 1313-1346 10
18 Fig. 4c 10-15-80 0949-1026 9
19 1
20 \
22 Fig. 4d 10-1
23 ,
24
1-25 |
W 26
27 |
1039-1113 9
1116-1156 11
-80 0940-1003 12
1021-1038 12
1056-1127 16
1251-1314 20
1323-1345 20
1350-1412 20
28 Fig. 4e 10-17-80 0850-0927 21
29 1 1 0936-0927 21
30 1 1 1016-1045 21
31 T t 1105-1135 21
32 Fig. 4C 10-24-80 925- 943 4
33
34
35
36
37
38
39 t
950-1004 4
1010-1025 5
1027-1040 5
1045-1112 5
1120-1138 6
1244-1325 11
1334-1403 11
* Type Is low flow
Type 2: heavy flow
Fogger 803018
Relative
Htmldlty
Hind
Hind Speed Ho. of
(l| Direction (B/sec) Trucks
77
77
70
70.
70
72
72
72
72
72
72
36
36
36
57'
57
57
52
52
52
55
55
55
55
82
82
82
82
82
82
68
68


M-E 2-5 8
NNH 4
\ 1
8
8
8
Hater
Flow
(1/hrl

61
68
72
Air
Flow
<»3/hr )

4.2
4.2
4.0
Fan
Speed
(»)

80
80
80
Sign
of
Charge

<0)
<;'!

Hoi lie*
Type

1
1
1
Hater
Flow
U/hrJ

68
68
68
Pogger 803019
Air
Flow
(Bl/hr)

4.2
4.2
4.0
Fan
Speed
(»)

80
80
80
Sign
of
Charge

(«)
';'

Nozrle*
•type

i
l
1
HNH Vac . 8
Iw/gusts 8
to
1

\
9 a
18
8
a
60
57
53
53

2.3
2.3
2.7
2.4

80
80
80
80

(0)
(+1
(-)
<-)

1
1
1
1

68
72
64
61

2.3
1.6
1.8
1.4

80
80
80
80

lu)
(+)
(-)
(->

i
i
A
1

CalB CalB 10
t j
1 '°.
68
76
4.0
4.1
50
50
(0)
2
2
66
76
4.0
3.6
50
50
to!
2
2
CalB CalB a
I 1
t \
4
a
SSH 2-5 8
1
*
8
8
76
72
72
77

2.8
2.0
2.6
2.7

80
80
80
80

(*)
(->
(*)
(0}

2
2
2
2

76
76
72
77

4.4
2.6
3-1
2.8

80
80
80
80

(+i

<+j
(0)

2
2
2
2

CalB CalB 8
I j
1 i
76
80
76
3.4
2.2
3.6
CalB CalB 6
\
m-E
I

t ^
6
6
6
6
C
76

76
74

4.2

4.7
4.8

SE-S 1-2 10
SE-S 1-2 10




78


4.2


70
70
70
80
80
80
80
80
80
80
80


(OJ
(*>

(0)

(0)



(0)


2
2
2

2

2
2


2


76
80
76

74

76
78


78


4.0
2.8
2.8

4.2

4.4
4.1


4.2


70
70
70
80
80
80
80
80
80
BO
80


(0)

(0)

(0)



to)


2
2
2

2

2
2


2



-------
CO
00
                                                               LEGEND

                                                        HI-VOLS-

                                                        •  7084 STANDARD
                                                        O  7112 STANDARD
                                                        O  71O6 STANDARD
                                                        A


                                                        A
7101 CASCADE
IMPACTOR

7094 CASCADE
IMPACTOR

7105 SIZE
SELECTIVE INLET

7092 SIZE
SELECTIVE INLET
                                                        FOGGERS-

                                                        •  803019

                                                        O  803018
                               Figure 4.   Test equipment positions.

-------
 uncharged  fog was applied to the  crushing  operation.   Although this result
 was unexpected, further analysis  soon  found the problem.  The  fans in the
 foggers  which  create  the airflow that projects  the  fog  toward  the  dust
 source are so powerful that they  were  actually  creating  an artificial wind
 effect.  The uncontrolled dust plume was only subject to  the  ambient  wind
 whereas  the controlled plume was  being radically "directed" by the fogger
 air  jets.   This  discovery  produced the need  for a  final  series  of  tests
 wherein  the  uncontrolled  baseline  TSP levels were recorded  with the fans on
 with no  water added.

    Another  concern that developed was with the intermittent nature of the
 truck  dumps.   In  some  cases,  eight  trucks  would  unload  within  twenty
 minutes  while at other  times  it would take thirty  or forty minutes.   The
 data  was  therefore  reduced  on a  per-truck basis since  the unloading  and
 crushing times, the times when the vast majority of  the dust  is produced,
 were  essentially  the  same  for  all  dumps.  The  data were  also  slightly
 adjusted to account for  deviations  of  the  actual sampler  flow  rates  from
 the design flow of  1.1 m3/min (40 cfm).

    While  the  data from  runs  7-31  did  not  reveal  information  regarding
 overall  fogger efficiency,  it did  provide  insight  into  the   increase  in
 efficiency due to  charging  the  fog versus uncharged  fog.   This efficiency
 could be further  examined with regards  to  particle  size, distance  from the
 pit,  and  positive charge  versus  negative  charge.    Not all  of   the  runs
 produced usable data  since  the  fog  impinged  on  the  samplers   during  some
 tests.

    The  data from runs 32-39 were used to  determine fogger efficiency  with
 respect  to uncharged fog.  The spacing  of the  samplers also allowed for the
 examination  of efficiency versus  distance  from the pit.   This information
 was then combined with the  data  from runs  7-31  to calculate  overall fogger
 efficiency.

    Attempts  were made at obtaining  information on  visibility improvement
 via EPA  Method  9  (visual  determination  of  opacity),   it was found  that  the
 opacity  of  the  fog was  similar  to  the  opacity  of the  uncontrolled  dust
 plume so that no real visibility improvement was noted.

    Figure 5  presents  the preliminary test  results from the initial charged
 fogger tests at  the primary crusher  site.  The  left side  of the  figure
 shows the  percent reduction in TSP levels  when  an uncharged water  fog  was
 used to  control the fugitive dust.  The right  side of the  figure shows  the
 additional percent  reduction in TSP levels  when  a  charge  was applied to the
 fog.   The   data  from  this  figure  reveal   several  important  results, as
 discussed  in the following subsections.

 Uncharged Fog Efficiency

    Based on  the  limited  amount of data for fan only  versus uncharged  fog,
 it appears that a  water  spray  alone is approximately 30-40% efficient in
 reducing  the  fugitive dust levels  from  the  primary  crusher.   It  also
appears  that  this efficiency  is  independent  of  particle  size  and   the
distance from the pit.
                                     139

-------
1OO


o
gso
Z uj
O C9
o<
si60
OC ^^f
^^^ f^M
gf ^j ^**
III ^
OL O


IL
20

O
_ i i ' t-

O STANDARD HI-VOL
A SIZE SELECTIVE INLET
^B




O -



_
H

^M

till
1 — -is 4 A. . io
100


O
u. 8O
O
III
oa
Pz
00 6O
•» •
Q (0
z2
K« *o
OQ
KUJ
UIO
ouoe
z

z 20

0
- O STANDARD1 HI-VOL ' '"
n CASCADE IMPACTOR -
U TOTAL LOADING
A SIZE SELECTIVE INLET
OPEN SYMBOLS: NEG. FOG
SOLID SYMBOLS: POS. FOG
a
o
i •
o ft . •
• A
A
^ ^^ ^







A
• 	 1 	 L
23 5 7.5 IO
•^•4«v & *i^»v> v»«%^ka I*I^P l~~.\
  DISTANCE FROM PIT (m)
Figure 5.   Percent reduction in TSP levels due to fogging of primary crusher.

-------
Efficiency Increase Due to Charging of Fog

    By applying a charge  to the water spray, the  fugitive  dust levels were
reduced 40-70% over the levels  recorded  using uncharged fog.  There appears
to  be a  trend  of increasing  reduction  with  increasing distance  from the
pit.  This apparent phenomenon  is  not explainable at this  time,  but it may
have  something  to  do  with  agglomeration  and  particle  fallout.   This
possible distance factor will be examined further in future tests.

    Figure 5  also shows that TSP  reduction due to charging is essentially
the same  regardless of whether a positive  or a negative charge  is applied
to  the  spray.  This  indicates that  the dust cloud  contains a  mixture  of
part ides ^ some with  negative charge  and  some with positive  charge.   This
is  consistent  with  the  findings  of  other  researchers,  namely  Hoenig
(1977)l and Kunkel (1950)2.

    TSP  reduction appears  to be  the same  for the  respirable  size  range
(<_15um),  as measured with  the  hi-vols  with  size selective  inlets,  as
for  the size  range  sampled  with  the standard  hi-vol  (£ 30ym).  it  was
hoped  that  the  use  of  the  cascade impactors  would  provide  additional
information  on efficiency  versus  particle size,  but  the results  proved
unusable.  Almost all  of  the material collected by the  hi-vols fitted with
the  impactors was collected on the  back-up  filter.   This  indicates that
there  was severe  particle  bounce between  the  impactor   stages.   Perhaps
tests at sources with finer dust will yield more useful information.

Overall Fogger Efficiency

    By  combining  the  results presented in  Figure  5,  it  is possible  to
calculate an  overall collection efficiency for the charged foggers.   Based
on  the  preliminary data,  the use  of charged fog can  reduce  the fugitive
dust  levels   that  result  from   the  primary  rock   crushing   operation
approximately 65-75%.  It is  felt  that this reduction could be even greater
through the  use of additional  foggers,  wind baffles to reduce turbulence,
and increased water flow.

FUTURE WORK

    The two  foggers are currently  being tested at the  second  source  which
is a  secondary rock crusher  at  another site in Connecticut.   At  this  site,
fugitive dust results from the fall of the  crushed material onto a conveyor
belt.  This dust is released to the atmosphere  through  openings at the base
of the  operation.  The foggers have been positioned  so as  to  blanket this
area  with a  cloud  of charged  fog.   Testing  should be  completed at  this
location by the end of November.

    Following the  tests at the  secondary  crusher, the  next sources  to  be
tested  will  be within  the  iron  and 'steel industry.   Negotiations are
currently underway for testing at  two different steel  companies - one  in
the eastern  United States  and  one in Canada.   At the  one steel  company,
there are  two possible sources for testing the foggers.   One source  is  a
sinter  plant pug  mill  which  mixes  water  with  baghouse dust  for   dust
suppression  before  recycling the  material  back  through  the plant.   Even
with  the  water addition  there is  a  significant  amount of fugitive  dust
                                      141

-------
around the  source.   The second  source is  the  hot fume  that  results in  a
cast house  from the filling of  a  ladle car with  molten iron from a  cast.
At  the other  steel  company/  a possible  source   for   testing  is  a coke
screening  operation.   Coke  is  transferred from  a  conveyor  belt  onto  a
shaker screen where it  is  sorted by size.   The shaking results in copious
amounts of  dust.  There  is also the possibility of testing some  limestone
handling operations,  such  as transfer points and  truck unloading, at this
location.

    The steel company tests mentioned above will  most  likely be  performed
in the winter and early spring.  Following  completion of all field tests,  a
final report will be prepared for the EPA presenting the results.
                                      142

-------
REFERENCES

1.  Hoenig,  S.A.  Use  o£  Electrostatically  Charged  Fog  for  Control  ot
    Fugitive  Dust Emissions.   Prepared  for  U.S.  EPA.   EPA-600/7-77-131.
    November 1977.

2.  Kunkel,  W.B.    The  Static  Electrification   of   Dust  Particles   on
    Dispersion  into  a  cloud.   Journal  of  Applied  Physics.   Volume  21.
    August 1950f pp.  820-832.
                                      143

-------
                         A GRAVEL BED FILTER WITH FLUIDIZED-BED
                                DURING REVERSE CLEANING
                               By Yan Xingzhong
                                     Vice chief Engineer
                                  Wang Nengqin
                                     Engineer
                          Safety Technology Research Institute
                          Ministry of Metallurgical Industry
                          Wuhan
                          People's Republic of China

                                      ABSTRACT
    In order to meet the needs for cleaning large quantity of high tem-
perature fume from iron and steel plants. Safety Technology Research
Institute of Metallurgical Industry Ministry of China has developed a
new type of gravel bed filter in which the gravel bed fluidizes during
reverse cleaning.  Being designed on the principle of bulb forming and
fluidiaation, this kind of filter can clean dust remained in the filtrating
bed by fluidized backflushing without the rake stirring mechanism.  This
helps the filter very much to be of the towery type with multilayers.

    In this paper, through theoretical analysis and calculations the
authors provide necessary data of fluidizing the gravel bed.  The appli-
cation of this filter in industry and the comparision of its technology
with that of ordinary gravel-bed filters and other types of efficient
dry filters have shown this filter is simple in structure and easy for
maintenance.  It takes up less space and requires less capital cost as
well.  To the present , -hls *yPe of *ilt«* with gas volume I20000m^/hr
has been run in practice and its performance of reverse cleaning is sa-
 isfactory.
                                    145

-------
                        A GRAVEL BED FILTER WITH FLUIDIZED-BED
                              DURING REVERSE CLEANING
INTRODUCTION
    Since the 1960s the gravel bed filters have been used for removal of
dust from Oas stream in some industries, such as in clinker coolers, lime
kilns and sinter machines.  The gravel bed filter uses solid granular
materials such as granular quartz, pebble-stone, etc. as its filtrating
medium that has following advantages: low cost, high temperature resistance,
good wear resistance and durability.  In comparison with some efficient
dry filters, such as electrostatic precipitators and bag houses, the
gravel bed filter is simpler in construction and is easier for main-
tenance, its capital cost is comparatively low.  So the gravel bed filter
is a kind of dust collector which promises well and will be further de-
veloped.

    In iron and steel plants a large quantity of high temperature dust
laden gases must be cleaned.  It is important to have an economic and
effective method to filtrate those dust laden gas.  The gravel bed fil-
ter is especially suitable for cleaning the high temperature gases, there-
fore this method is well worth developing.

    Safety Technology Research Institute has engaged in developing the
gravel bed filter since I974-.  From the beginning of the research we tried
to suit this kind of filter to filtration of large quantity of high tem-
perature dust laden gases.  Under such circumstance the principal method
is to place the filter layers on top of each other to form a filtrating
tower, which would occupy less space.  In our first pilot test case pi$*B
were used as the shaft of turning the rake stirring device in each layer,
as shown in Figure I.  In the second test a shaft was set at one side of the
towery filter to drive the rake stirring device in each layer with sprockets
and a roller chain, as shown in Figure 2.  Both of tests failed because
it was too difficult to bring this construction into reality and we could
only overlap 3-5 layers of filter together at most.  We carried on our
test, trying to construct a filter, every layer having its own motor,
reductor, sprocket and roller chain that would drive the rake stirring
                                    146

-------
Figure I.  Filter with case
pipes as shaft to drive
the rake stirring device
Figure 2.  Filter with a
side shaft to drive the
rake stirring device
                    Figure 3.  Filter with individual
                    motors to drive the rake stirring device
                    in each layer
                            147

-------
device separately, as shown in Figure 3»  However, it was still impossible
for MS to build such a multilayer filter tower, for the intervals between
layers were too large.

    Finally, by applying the principle of bulb forming and fluidization
in the gravel bed, we developed in 1977 a multi-layers gravel bed filter
of towery type which could clean dust remained in the filtrating bed by
fluidized backflushing, so that the rake stirring mechanism could be left
out.  This kind of filter is much simpler in structure, easier for main-
tenance, occupying less space and requiring less capital cost.  Now, a
gravel filter of this type with gas volume I20000m^/hr has been run in
practice and its performance of reverse cleaning is satisfactory.

PRINCIPLE AND ANALYSIS
    During the backflushing of the gravel bed filter with stirring rake
devices, the rakes stir the granules and make them rub mutually.  In this
process, the backflushing air separates the dust from granules and sends
the dust out of the filter layer.  With the help of rake stirring, the
surface of filter layers may be kept plane*

    For a gravel bed filter with fluidized bed during reverse cleaning,
the fluidization in the filter layer can get the same effect as the rake
stirring.  During the period of backflushing, the reverse cleaning air
will reach a definite velocity and the filter layer will be fluidized.
In the fluidized bed the granules will move up and down and rub mutually
as boiling water does.  In such case, dust which adheres to the granular
surface may be separated and brought away by the flushing air.

    To achieve ideal effect, the reverse cleaning process should be con-
trolled as follows:
    I.  The bulb forming must be uniformly distributed in the whole area
of the gravel bed layer, there must not be any dead corners or local
blowing-off.
    2.  After reverse blowing, the filtrating layer must have a plane sur-
face, on which there must be no hills, no cavities or no uneven area.
Therefore the filter can achieve good filtrating effect.

                                    148

-------
     3.  The parts which can meet the above-mentioned requirements should
 be simple in construction, easy for maintenance and hare high reliability
 and low pressure drop*
         filter layer possesses a good appearance, as shown in Figure 4 and
 Figure 5.  In Figure 4 the filter layer is in a state of reverse cleaning.
 The granules move up and down uniformly as the boiling water does.  Figure
 5 shows that after reverse cleaning the surface of the layer of the gravel
 filter is quite plane*

     The forming of fluidization and good reverse cleaning are determined
 by the size of granules, the properties of dust, the velocity of reverse
'air blowing, the pressure drop as well as the interval time of reverse
 blowing.
 The Diameter of Granules
     In the gravel bed filter, it is important to choose granules of pro-
 per sizes.  The diameter of granules has direct influence on the effi-
 ciancy of dust collection and reverse blowing.  In general the finer the
 filtrating granules, the higher the dust collection efficiency and more
 easily the fluidization will be formed.  But there are some limits to
 the fineness of the filtrating granules.  If the granules are too fine,
 they would drop through the screen.

     It is impossible for all the granules to have same diameter.  For
 the purpose of denoting the size of granules, the "average equivalent
 diameter" Dp (mm.) is used.
     Dp is determined from the following equation:
                     I   = 2-2L-
                    Dp        di
 Where     xi = the percentage of the weight of the granules whose diameter
           is di (mm.)
     The granules are generally of irregular shape.  The average equivalent
 diameter must be multiplied by a form coefficient pn to get "calculating
 average equivalent diameter " Do (mm.) which is expressed as
                                     149

-------
Figure 4-.  Filter layer in bulb
forming and fluidization
 Figure 5.  Filter layer after
            backflushing
          150

-------
                        DO = #s '  DP
     The form coefficient is the root square of the ratio of spherical
 surface area Sg to irregular surface area S^,   It applies

                        *
     Generally,  the form coefficient of the granular quartz  is  close  to
 0.5 and 1.3-2.2 mnu of the average  equivalent  diameter  of granules is
 used.

 Properties  of Dust
     The properties of  dust,  especially the adhesion, affect  the result
 of backflushing.   It is easy to blow off  the dust  deposited  on the sur-
 face of the filtrating layer.   But  to  blow off the dust  adhering to  the
 surface of  the  granules requires some  specific conditions.   To elutriate
 this kind of fine  dust there must be a greater blowing-off velocity  and
 every  granule must be  brought  into  touch  with the  backflushing air jet.
 Nowadays it has succeeded  in blowing off  some low  adhesive dusts which
 adhere to the surface  of the granules.  But some of dust of metal oxides
 cannot be wholly blown off from the  filter layer due to  its large cohesive
 force.

     The size  of dust which can  be blown off during reverse cleaning has
 a  close connection with the  velocity of reverse blowing and the density
 of dust itself.  It may be considered that the reverse blowing velocity
must be larger  than the terminal settling velocity of dust particles.
Only on this  condition can the  dust particles be blown off from the filter
layer during  reverse cleaning.   For example, when the average reverse
blowing velocity is Im/sec,  only the dust whose diameter is  less  than
0.15 mm. can be blown off from  the filter layer,  and the dust larger  than
0.13 mm. must be deposited in the settling chamber beforehand.

Critical Fluidized Velocity
    When the volume of  reverse  cleaning air flowing through  the filtrating
layer is very small, the air will flow through the small porosity  and the
                                   151

-------
granules remain steady.  This kind of filtrating layer is called steady
bed.  When this velocity increases, the granules begin to lift by the
air.  This is called primary fluidized bed.  If the velocity would in-
crease continuously, air bulb and violent stirring will occur in the fil-
trating filter.  That is called bulbing or fluidized bed.  The minimum
reverse cleaning velocity which changes the filtrating bed from steady
state into fluidized state is called critical fluidized velocity.

    The critical fluidized velocity can be calculated by using the fol-
lowing equation [ij:
        Vfo    B  Jl_ .       Ar
Where   V  - kinematic viscosity (m2/sec)
        Do = calculated average diameter of granules (ra)
        £  n porosity of granules layer
        Ar = Archimedes number
                           ;
    For the porosity of granules, the following equation applies
                                  Vr - Yb
                                    Yr
Where   Yb = bulk specific gravity
        Yr = real specific gravity of granules
    The Archimedes number applies
                                   •
                     Ar
                                V2  «Ya
Where    g  =  gravitational  acceleration  (m/sec2)
         Yr  =  real  specific  gravity  of  granules
         Ya  =  specific  gravity  of  air
         V =  kinematic viscosity  (m/sec)
         Do  s  calculated  average diameter of granules  (m)
    While the average  equivalent  diameter  of  granular quartz is within
the range of  i.3-2.2 mm., the  critical fluidized velocity will be 0.68-
I.I9m/sec.
                                    152

-------
The Maximum Reverse Blowing Velocity
    When the reverse blowing velocity is larger than the critical flui-
dized velocity, the bulbing effect will be intensive and the backflushing
will achieve good cleaning result.  But if the velocity of reverse blowing
is too large, the granules may be blown off from the filter bed.  Then
there should be the maximum permissible reverse blowing velocity.
    In the steady air the granules dropping due to the gravitational
force has a terminal velocity.  When air flow reaches this velocity, the
resistance of air acting on the granules equals the gravitational force.
The dropping velocity of granules would not increase further.  In the
field of air flow, the granules would be blown off if the velocity of air
flow exceeds the terminal velocity of falling granules.  So there should
be the maximum reverse blowing velocity Wt(m/sec) (2) , which can be found
from the relationship:

             —  Do3(rr -ra) =
              6
                              g
                                       Do
                                                  wy
                                      4          2g
In the form, £g is related to the Reynolds number   Re =
                                                            Do
    When   Re <  0.4
                    24
                    Re
    Then
          wt  =
               18 rav
                              Do'
    When   0.4 < Re < 500
                         10
    Then
            Wt =
• Re
4
225
0.5
C'r
•
- ra)2g2
r 2v
                                                 Do
When    500 < Re < 200000

        8     f  3.1g (*r - ra)
    Then
           Wt =
                                 Do
                                          si
                                            2
    For  granular quart* its average  equivalent  diameter  is 1.3-2.2 mm.
                                    153

-------
               Wt = 5.4-8.76m/sec

The Resistance of Reverse Bloving
    When the reverse cleaning air keeps a low flow rate, the resistance
of air flow is proportional to the air flow velocity.  When the velocity
of reverse cleaning air rises to the fluidizing condition, the granules
will be blown up and will float in the air.  The resistance of reverse
blowing will be a constant.  It will not increase even if the velocity
of reverse blowing rises further*

    In the fluidizing condition the resistance of reverse blowing "P"
is close to the weight of the granular filtrating layer in unit area(l).
It is of the form
              P s 


-------
CONSTRUCTION
    A typical multilayer gravel bed filter with fluidized-bed during re-
verse cleaning is shown in Figure 6.

    In Figure 6, the raw dust laden gas enters the settling chamber (2)
from the inlet duct (I) to remove the coarse dust.  The dust not deposited
in the settling chamber will go into the filtrating chamber (3) with the
gas, where the gas flows through the filter layer (4) and the remaining
dust is removed.  Filtrated gas then will enter the clean gas outlet duct
(9) through the screen (5) and the exhaust port.  The process described
above continues until the filtrating bed is fully laden with dust.  Then
a new cleaning cycle begins.  At this time the layer of filter is isolated
from the cleaned gas stream by actuation of the backflushing control valve
(10) driven by the valve cylinder.  Then this layer has come into the bed
cleaning or backflushing period.

    In the lower filtrating layer in Figure 6, fresh air enters the unit
through the backflushing air inlet (8) and is carried upward and runs
through tfie filtrating layer (4-).  During this period the gravel bed is
fluidized so that stirring bulbs are formed.  The dust is removed from
the filter medium.  Some of the agglomerated dust particles which are
significantly larger than the original fine dust will be led down to the
settling chamber and some other dust is reintroduced to the filtrating
layer together with the raw dust laden gas.

    The backflushing valves are powered by the double-acting compressed
air cylinders, which are controlled by an adjustable interval timer.

    A section of filter layer consists of two filtrating areas (A.B) se-
parated by an insulating board (II).  A1J the layers are of the same con-
struction.  Consequently it is easy to place one layer over another.
Generally 4 layers are combined to be a set, and a multilayer towery type
gravel bed filter consists of several sets.

    The chief advantages of the gravel bed filter with fluidized-bed over
the ordinary gravel bed filters are as follows:

                                     155

-------
                                     Section I—I

                                     I.   inlet  duct
                                     2.   settling  chamber
                                     3.   filtrating chamber
                                     4.   filter layer
                                     5.   screen
                                     6.   dust chamber
                                     7.   dust outlet
                                     8.   back flush air inlet
                                     9.   clean  gas duct
                                    10,   back flush control
                                         valve
                                    II,   insulating board

                                     •A.B  filtrating area
Figure 6.  Schematic diagram of multilayer gravel bed filter
           with fluidized bed during reverse cleaning
                       156

-------
     Due to the  omission of the rake stirring mechanism the construction
 of the filter is simplified and the height of every layer is effectively
 reduced.   Then  it is easy  to build the towery type  of gravel bed filter.
 In the filtrating chamber,  the temperature resistivity is improved  and
 the maintenance needs less  labour.  Now the principal part of a  filter
 with gas  filtrating volume  30000mVhr  onlY occupies an area of I6m2,  and
 only costs RMB  1200 Yuan per 1000 cubic meters  of dust laden gas an hour.

 APPLICATIONS
     The gravel  bed filters  with fluidized  bed during reverse cleaning have
 been installed  in the following areas.

 Sintering Plant
     The first gravel  bed filter of this type  was installed  in  a  sintering
 plant in  1979,  as is  shown  in  Figure 7.  This filter  is used to  filtrate
 the dust  laden  gas in an agglomerat transportating  and lifting system.
 There are altogether  10  filtrating layers.  This filter can  treat 15000
 cubic meters of dust  laden  gas  an hour.  For  the dust of  agglomerat 75?6
 of  which  was less than 400  in  diameter, the  filter can have 95-97.8
 percent filtrating efficiency.   The temperature of raw gas was I30-I80°C.
 The pressure drop of  the filter is about 80mm* water.

     The second  set of  this  type filter  installed in 1980  is used for con-
 trolling  the smoke of  a  sintering machine, Figure 8.  It  has 22 filtrating
 layers  and  can  treat  25000-30000  cubic metqrs of duct laden gas an hour.
 For  the dust 72.8$ of  which  is  less than 40n  in diameter, the filter can
 have  96 percent collection  efficiency.  The temperature of raw gas is
 I50-I70°C.  The pressure drop of  the filter is about 80-90 mm. water.

 Steel Heating Furnace
    In  October  1978, a filter of  this 'type with 18 layers was installed
 in a  steel heating furnace.  In April 1979 three more filters of  the same
 type were installed at the same place, as is shown in Figure 9.  The total
volume  of dust  laden gas filtrated by these four filters is I20000m^/hr.
For the dust 9I.2# of which is less than 40//  in diameter, 95 percent col-
                                    157

-------
Figure 7.   Filter  installed  in
an agglomerat transportation
and lifting system
 Figure 8.  Filter  installed
 in a sintering machine
Figure 9.   Filter installed
in a steel heating furnace
Figure iu.  Filter installed
ir. -^ clay drying machine
                                 158

-------
lection efficiency can be reached.  The temperature of raw gas is 2I?C.
The filtrating medium is granular quartz.  The filter layer is IIO-I50
mm* high.  The pressure drop during filtration is 70mm. water.

Clay Drying Machine
    A gravel bed filter of fluidized backflushing with 10 layers was in-
stalled in a clay drying machine in 1978, as is shown in Figure 10.  The
gas volume of this filter is I5000nrVhr.  The collection efficiency is
98-98.5$.

CONCLUSION
    From the experiments and practical operation of the gravel bed filter
with fluidized-bed during reverse cleaning, the authors have come to the
following conclusions:

    I.  When the velocity of backflushing is slightly greater than the
critical fluidizing velocity, air bulbs and violent stirring take place
in the granular filtrating layer.  After backflushing the surface of fil-
trating layer can keep plane and the dust which sticks to the granular
surface can be removed.   The primary pressure drop of the filter recovers.

    2.  By applying the theory of fluidization to backflushing,  the gravel
bed filter can be constructed without the rake stirring mechanism.  It
is easy for this kind of gravel bed filter to be of multilayer towery type,
more applicable to the filtration of large quantity of high temperature
dust laden gas in iron and steel smelting plants.

    3.  Compared with the conventional gravel bed filters,  the gravel bed
filter with fluidized-bed during reverse cleaning is much simpler, re-
iuires less capital cost and occupies less space.  It is easier  for main-
tenance and owns good heat resistance.

    To raise the dust collection efficiency and to filtrate adhesive
dust with this kind of filter,  wj still have much to do.
                                     159

-------
REFERENCES

                              *                    . 10.
         Daizo Kunii and Octave Levenspiel: "Fluidization Engineering"
                                    John Wiley & Sons, Inc. 1969
                                                1976.3.

                                    The End
                                     160

-------
                FINE PARTICLE CONTROL AT HIGH GAS TEMPERATURE
                            Michael  A.  Shackleton
                              Acurex  Corporation
                               485  Clyde  Avenue
                       Mountain  View,  California  94042
                                  ABSTRACT
     A new filter media made from ceramic fibers offers  the  potential  for
fine particulate control at gas temperatures up to  1,000°C (1,800°F).
These filter media represent an emerging technology under development  for
application to hot gas cleaning in pressurized fludized  bed  combustion.
However, the ability to control particles at extreme temperatures will
offer benefits to other industries as well.  For example, the 1,000°C
filter will reduce the need to dilute hot gas streams in the iron and
steel industry prior to particle removal.  The resulting clean hot gas can
then be used in a heat recovery system to offset the cost of pollution
control.  Progress to date in the development of this new filtration
device is reviewed in this paper.
                                   161

-------
               FINE PARTICLE CONTROL AT HIGH GAS TEMPERATURE
INTRODUCTION
     Pressurized fluidized bed combustion and some other advanced coal
utilization processes contain particle laden hot gas streams (800°C).
Before energy can be extracted from these gas streams, the particles must
be removed to protect the energy extraction equipment — usually a
turbine.  Because of their simplicity, cyclones are a preferred device to
accomplish this particle removal and up to three cyclones in series have
been used with promising results on cascade erosion tests.  But, cyclones
do not remove fine particles (<5 ym) which are a potential health hazard
and will need to be removed before release to the atmosphere.  In
addition, potentially corrosive alkali metals are concentrated in the fine
particle fraction of the dust distribution and could damage metal turbine
components through deposition on the blading.  Thus, if it were economical
to do so, there are technical advantages to removing all particles,
including fine particles, from the hot gas stream.  The ceramic fiber
filter media under development at Acurex offers the potential to
accomplish this needed fine particle control at high temperatures.

     The ceramic filter represents a system approach to filter design.
That is, it is not simply a material substitution of ceramic fibers for
conventional fibers.  The entire filter unit including media support and
cleaning techniques are designed for the unique properties of the ceramic
fibers.  This approach maximizes the advantages of ceramic fibers while
minimizing the effects of their weaknesses.  The media employs fine
diameter fibers (3 ym nominal) which provide improved collection
efficiency.  Fibers are arranged in a mat having a low solidity so that
they can move relative to one another.  This prevents breaking of
individual fibers during cleaning.  A loosely woven ceramic cloth on the
inner and outer surfaces of the media provides a strengthening screen
which contains the filtration layer of fine fibers.  Because the media is
inherently more efficient than conventional media, dust tends to deposit
and form a cake on its surface where it is most easily removed.  It can
also collect particles at relatively high face velocity 10 cm/sec
(20 ft/min).  Operation at high face velocity requires offline cleaning to
prevent reintrainment of the dust removed by the cleaningipulse.  Cleaning
offline also provides the most effective cleaning at minimum energy.  To
provide continuous filtration only a portion of the filter unit is cleaned
at a given time.

     Significant features of the ceramic filter are:

     •   Operation at high temperatures and pressures (800°C nominal, up
         to 10 atm or more pressure)
     t   High efficiency collection of fine and submicron particles offers
         potential for alkali removal
     •   High face velocity operation offers potential for compact size
         and lower costs
     •   Pressure drop is determined primarily by dust cake properties and
         face velocity
                                    162

-------
 The Acurex ceramic filter has been shown to be a promising concept for hot
 gas cleaning in a series of feasibility tests sponsored by the
 Environmental Protection Agency (EPA).  The results of those tests are
 briefly reviewed in this paper.

 THEORY DISCUSSION

      The equations predicting filter performance show that fiber diameter
 is an important parameter.  Small  diameter (3 ym) fibers improve
 performance but, except for very high efficiency filters as used in
 biological applications, fine fibers have generally not been used in
 filter media design.   The principle reason is that fine fibers are not
 commercially available as readily as coarser (20 to 50 ym) fibers, such as
 cotton or cellulose.   Since the textile industry dominates the use of
 fibers,  most artificial fibers have been made to simulate natural ones.
 Also  small fibers are more expensive to produce and existing fibers
 perform satisfactorily in most cases.  So, even though seldom used, fiber
 diameter is a powerful tool to manipulate filter media performance.

      Space limitations do not permit a complete analytical discussion of
 temperature and  pressure effects upon filter performance and of how these
 effects  can be  overcome in the design of a filter media.   These analyses
 have  been made,  however, and  the results are summarized below.

      Because of  the  increased viscosity of gases at high temperatures,
 particle collection  by inertial  impaction is reduced  10 to 15  percent for
 the particle size range of 0.5 to  5.0ym.   Thus,  an inertial  device
 collecting 2.0 ym particles  at 90  percent efficiency  in a room ambient
 test  may collect the  same particles at only 75  percent  efficiency at
 high-temperature and  high-pressure  (HTHP)  conditions.

      Filters  employ  inertial "impaction,  direct  interception,  and  diffusion
 mechanisms in collecting particles.   High-temperature  and  pressure
 influence  collection  by diffusion  only slightly and may,  in  fact,  improve
 performance.  Direct  interception  is  not a function of  temperature  and
 pressure.   All three  of these particle collection mechanisms,  however,  are
 strong functions  of fiber diameter.   If  the  predicted  performance of  a
 fiber  bed  composed of  20ym diameter  fibers  is  compared with an equal
 weight per unit  area of 3.0 ym diameter  fibers,  collection efficiency will
 significantly improve.   This  improvement  in  performance will be much
 larger than the  adverse  effects resulting  from  operation  at  high
 temperatures.  For example, such a  change  in fiber  diameter will  improve
 collection of 0.5 ym particles from 20 percent  using 20 ym fibers,  to
 almost 90  percent using  3.0 ym fibers,   Filters  achieving  20 percent
 collection of 0.5 ym particles are typical of those in  commercial use and
 their  adequate performance can be attributed to the additional filtration
 efficiency of the dust  cake which forms  on the  filter surface.  Thus  by
 using small fibers and by  relying upon dust cake filtration, it is
possible to make a filter media which will give good performance  at
viscosity  conditions similar  to those  at high temperatures.

     Fortunately ceramic fibers are available which have diameters of
3.0ym.  Figure 1 presents the performance predicted for a bed of 3.0ym
                                    163

-------
            3.0*1 m OIA FIBERS
            2.8 g/cm3 FIBER DENSITY
                                             0.5 u m DIA PARTICLE
                                             1.5 g/cm3
                                             815°C
                                             10 ATM
   90
O
i
o

O
o
70



60


50


40



20


 0
              (5 ft/min)
             2.54 cm/sec J
                                           (25 ft/min)
                                           12.7 cm/sec
                                            Typical  16 oz  felt.
                                                20 ym fibers
                                           (16 oz/yd=)
                                            540 g/m2
a _ Fiber volume
     Bed volume
                                                                     J_
           100   200   300   400    500   600    700    800    900   1000   1100   1200

                                 BASIS WEIGHT, g/m2
   Figure 1.   Calculated  performance of a ceramic fiber  bed  composed
                of 3.0 vim diameter  fibers.
                                        164

-------
ceramic fibers and 16-ounce felt collecting 0.5 ym particles under HTHP
conditions.  The felt  is a conventional filtration media with 20 ym
fibers.  It is apparent that an equal weight of fine fibers causes
dramatic improvement in collection efficiency.

     Some other useful observations can be made from Figure 1.  The effect
of filter face velocity and filter media solidity on efficiency are small
compared to the effects of the change in fiber diameter.  In other words,
a filter composed of 3.0 ym fibers could have higher collection efficiency
than currently used commercial filter media, even if operated at filter
face velocity as high as 25 ft/min and if only 2 percent of the bed were
occupied by fibers (98 percent open area).   The figure also shows that
collection efficiency can be increased by adding fibers (increasing the
basis weight).  This is, in effect, what happens when a dust cake is
formed — the dust cake gets thicker and collection efficiency and
pressure drop increase.

     In summary, currently available fiber filtration models show that
fine fibers employed in a mat filter will be able to achieve high
efficiency collection of fine particles under high-temperature and
pressure conditions.  Further, that a low solidity (fluffy) bed of these
fibers can achieve high efficiency fine particle collection even at filter
face velocities as high as 25 or 30 ft/min.  Our test programs have
verified these theoretical results.

     This discussion has only addressed particle collection.  Other
important questions concerning operating pressure drop, durability, and
cleanability are best examined by experiment and are discussed in the
following section.

SUMMARY OF TEST EXPERIENCE

     Four categories of feasibility tests have been performed:

     •   Room ambient tests — These tests confirmed the theoretical
         analysis and examined a broad spectrum of ceramic filter media
         candidates
     0   High-temperature and pressure mechanical durability tests —
         These tests determined that the ceramic filter media candidates
         could survive the mechanical stresses associated with 50,000
         cleaning pulses (about 1 year of service)
     •   High-temperature and pressure filtration tests — These tests
         measure the filtration characteristics of the filter media in
         200-hour dust feeding tests
     •   Slipstream tests at PFBC — These  tests showed high collection
         efficiency and cleanability of a ceramic fiber filter in tests at
         the EPA/Exxon Miniplant using as-generated PFBC flyash which
         passed the secondary cleanup cyclone
                                   165

-------
Room Ambient Filter Media Tests

     A large number of ceramic fiber filter media candidates were
subjected to a series of filtration tests at room ambient conditions.
These tests included some examples of conventional filter media for
comparison.  The type of tests performed were:

     •   Dioctylphtalate (OOP) smoke penetration as a function of airflow
         velocity
     •   Determination of maximum pore size (in micrometers)
     •   Measurement of permeability
     •   Flat-sheet dust loading tests using AC fine test dust (a standard
         0 to 80 ym classified Arizona road dust).  Overall collection
         efficiency and dust loading required to develop 3.7 kPa
         (15 inches H20) pressure drop are determined from this test
         which is run at 10 cm/sec (20 ft/min) face velocity.

     The test data revealed which of the available ceramic media
candidates would most likely provide good filtration performance.  These
data are shown in detail in EPA-600/7-78-194.  A summary of findings from
these tests follows:

     •   Several of the ceramic paper and felt materials are capable of
         removing fine particles at high efficiency without excessive
         filter basis weights.  OOP tests tended to confirm the
         performance predicted by analysis as shown in Figure 1.
     •   The ceramic paper and felt materials have filtration
         characteristics and performed similarly to paper and felt
         commercial filter media in a series of filter media tests
     •   The ceramic woven materials were characterized by large pores and
         low collection efficiency in the dust loading tests.  The range
         of parameters exhibited by the various woven materials indicates
         that an acceptable woven ceramic filter media can probably be
         fabricated.  However, acceptable performance would only occur at
         low air-to-cloth ratios.
     •   "Blanket" ceramic fiber materials (felts) consisting of small
         diameter fibers (3.0 ym) appear to be the most promising
         materials for high temperature and pressure tests because of
         their combination of good filtration performance and relatively
         high strength

High-Temperature and Pressure Mechanical Durability Tests

     The mechanical durability tests answered the following questions:

     •   How durable are ceramic fiber structures when subjected to
         environmental conditions associated with filtration applications?
     •   How well do ceramic fibers perform as filters in the HTHP
         environment?

     Concerning the first question, three ceramic filter media
configurations survived a test during which the filter elements were
subjected to 50,000 cleaning pulses.  The objective of these tests was to
                                    166

-------
 simulate approximately 1 year of operation of mechanical  cleaning loads on
 the  media at  high-temperature and pressure.   These tests  showed that the
 low  solidity, fine-fiber filters were undamaged by pulse  cleaning loads.
 They also showed  that the flyash dust cake was deposited  mostly on the
 surface  of the media.  Details of these  tests were also reported earlier
 and  in EPA-600/7-78-194.

 High-Temperature  and  Pressure Filtration Tests

      Filter performance at high temperatures  and pressure was tested for a
 period of 200 hours.   The filter media configuration  which was selected as
 most promising consists of an approximately 1 cm thick  layer of Saffil
 alumina  blanket insulation material.   This ceramic material  was contained
 between  two layers  of knit 304 stainless steel screen and provided a basis
 weight of approximately 500 g/mz.   Tests were performed at three airflow
 velocities to determine performance as a function of  both time and filter
 media face velocity.   Reintrained  flyash from the EPA/Exxon  Miniplant was
 used as  the test  dust.   Figure 2 presents average outlet  concentration as
 a function of time  for  the three tests which  were performed.   The
 designation "Turbine  Limit" corresponds  to the turbine  tolerance of
 0.002 gr/scf  reported by Sverdrup  in  EPA-600/9-78-004.  During the three
 tests, between 40 and 100 kg of dust  was fed  to each  test filter.   Each
 test filter contained only 1.5 ft2 of filter  media area,  so  dust loading
 was  considerably  higher than that  expected from the second stage cyclone
 of a typical  PFBC.  During these tests,  cleaning pulses occurred about
 once every 10 minutes and pressure drop  was maintained  at less than
 10 inches  of  water.   The test conducted  at 2.5 cm/sec airflow velocity
 showed an  increased rate of penetration  after about 50  hours  of
 operation.  This  condition evidently  was caused by a  defect  in the filter
 media although subsequent visual  examination  could not  positively locate
 this  defect.

      Outlet concentration as a function  of face velocity  (air-to-cloth
 ratio) is  plotted on  Figure 3 for  three  time  periods  of 50,  100,  and
 200  hours.  If we assume the filter used  in the first test at 2.5  cm/sec
 developed  a leak  and  extrapolate expected  performance (dotted lines),  it
 is apparent that  outlet concentration is  reduced  as a function of  time at
 all  velocities.   This result is  similar  to that of using  conventional
 filter media  in a room  ambient dust feeding test.

      Overall  particle collection efficiency is plotted  as  a  function of
 face  velocity (air-to-cloth  ratio)  on  Figure  4 for  three  time periods  of
 50,  100,  and  200  hours.   Again,  if  the two discrepant data points  are
 ignored,  collection efficiency is  essentially independent  of  face  velocity
 in the range  tested.  This is  consistent with  a hypothesis which  holds
 that  filter penetration  occurs primarily during cleaning.  The filter  was
cleaned  at  zero forward  flow in  all three  tests  (offline).

Slipstream  Tests  at the  EPA/Exxon Miniplant

     A 1.5  ft2 ceramic  bag  filter was evaluated/at  the  EPA/Exxon
Miniplant PFBC.   The  results  summarized  below  were  reported  in the
January  1979  Monthly  Progress  Report  No.  107  for  EPA  Contract 68-02-1312,
under which Exxon  Research  and Engineering Company  operated the miniplant
test rig.


                                    167

-------
 Exxon miniplant fly ash
 air-to-cloth ratio:

G 2.5 cm/sec
^4.8 cm/sec

Q 9.0 cm/sec
800°C
10 atm
'0
          50          TOO

                 Time -- hours
             150
200
    Figure 2.  Average outlet concentration.
                     168

-------
         0   50 hours

         Q  100 hours

             200 hours
                  Filter face velocity cm/sec
                  (air-to-cloth' ratio)
Figure 3.  Outlet concentration as a function of face velocity.
                             169

-------
                ®  50 hours

                Q100 hours

                0200 hours
        100
OJ
(J

OJ
Q.
0)


(J
•P*
H-
«4-



C


•i— «/>

(J  E
       99.99
       99.98
       99.97
       99.96
                                468

                         Filter face  velocity cm/sec
                         (air-to-cloth  ratio)
                                                              10
 Figure 4.  Collection efficiency  as a function of face  velocity.
                               170

-------
      The pressure drop across the filter bags varied with time in the
 classical manner for fabric filters Figure 5 which demonstrates that the
 bags could be cleaned.  The effect of coal type on cleanability seemed to
 be fairly small, and could not be determined from the relatively few tests
 completed at the miniplant.

      Filtration efficiencies for the Acurex ceramic bag filter were all
 over 90 percent, generally ranging from 96 to 99.5 percent.   An exact
 filtration efficiency was difficult to determine because of  problems in
 measuring the filter inlet particle concentration.  Filter inlet
 particulate concentrations were measured or calculated by three methods:
 (1)  Balston total  filter  catch on an extracted sample, (2) mass balance
 around  the third miniplant cyclone, (3) mass balance around  the ceramic
 bag  filter.

      The bag filter  outlet particulate concentration was determined by
 passing the entire filtered gas flow through a large Balston total
 filter.   The total particulate concentration was obtained  by weighing the
 filter  before and  after  collection.  A particle size distribution could
 not  be  obtained.   The concentration of particles on  the total  filter was
 so low  that insufficient  material  was available as a filter  cake for
 Coulter Counter analysis.   The filters were washed off with  a  solvent
 (Isoton  II) in  an  attempt  to remove particles  without mechanical
 brushing.   This  method allowed enough Balston  filter material  to be washed
 into  solution to completely obscure the flyash particulates.   A clean
 Balston  total  filter,  not  exposed  to  any flyash but  also washed  with
 Isoton  II  gave  a sample which  had  a size distribution similar  to that
 obtained  from a used  filter.

      During the tests  at  the miniplant,  one double and  eight
 single-thickness bags  were exposed  to PFBC  conditions.   Most bags were
 exposed  for 6 hours  or more.   By averaging  the face  velocity and exit
 particulate concentration  over the  first 6  hours  of  new bag exposure a
 plot  of  velocity against  loading yield points  closely clustered  around a
 line  Figure 6.   The effect of  coal  type  on  outlet  loading  was
 insignificant.

      The outlet particulate  loading tended  to  decrease with  increasing
exposure time.   Figure 7 shows  the  change in outlet  loading with time.
Along with  the  decrease in  filter particulate  outlet  loading,  baseline
pressure drop increased from 0.1 to 3.0  kPa  as  expected.  The  filter  cake,
which was not completely removed with  cleaning, caused  both the  lower
outlet loading and the higher  pressure drop.

      Bag cleaning  at  ambient conditions  after  a run was not a problem.
Several  of  the filter elements  were cleaned by  passing a vacuum hose  over
a strip of  the element.  This strip had  the appearance of a virtually new
bag,   indicating very little dust penetration through the bag fibers.
                                   171

-------
o
 CM
O.
O

-------
    0.025
u
V)
o>
C7>
s
(O
i-

-------
              Bag No. 5 paniculate penetration history
0.015


^_^
f>
E
en
O
S o.oio
c
OJ
O
^
O
O
01
HJ
•g 0.005
0
t

"^^ *"•—•—•*


^
t 1 1 f 1 1 1 1 1 1 f f 1 1 1
t
X
t*
y
*
















5              10

     Bag Age (Hours)
              Run 90
 pun 91 .l|
Run 91.2
                                                       15
Run 91.3
Figure 7.  Acurex ceramic bag filter
           penetration history.
            — bag no.  5 particulate
                            174

-------
      At its conclusion, the Exxon report said:

      "Based on these preliminary tests, high-temperature, high-pressure
 ceramic bag filtration appears to be feasible.  Pressure drops of under
 2 kPa have been maintained for over 6 hours at face velocities of
 4.6 m/min with efficiencies of 95 to 99 percent.  The short evaluation
 tests have yielded results far superior to granular bed filtration which
 never achieved efficiencies greater than 95 percent and which also had
 baseline pressure drops in excess of 14 kPa.  Further testing is required
 to determine maximum economic face velocity, and bag life under optimum
 cleaning conditions."

      A larger ceramic filter test unit containing  five filter elements
 each 1.5 m (5 ft) long and having 2.3 m2 (50 ft2) of filter area has
 been built for testing at Westinghouse under an EPRI-sponsored test
 program.   Results of this testing will be reported at a future date.

      Through the cooperation of the EPRI, a swatch of the ceramic filter
 media was  exposed to the hot dusty gas from the PFBC at the CURL facility
 in Leatherhead, England during a 1,000-hour test there.  While this was
 not a filtration test,  subsequent examination of the fibers revealed that
 no damage  resulted from 1,000 hours of exposure to the  chemical  products
 of combustion of coal.

 SUMMARY  AND CONCLUSION

      Through a series  of tests conducted  at laboratory  bench-scale and  on
 a slipstream of gas  at  the EPA/Exxon Miniplant,  the Acurex ceramic fiber
 filter has  continued to show that a cleanable filter capable of  operation
 at up to 800°C (1,500°F)  can be developed  to commercial scale.
 Evidence  in support  of  this  conclusion includes:

      •   A  theoretical  basis  exists  for  it
      •   Room ambient  tests  showed  high efficiency collection of fine
         particles
      •   Room ambient tests  showed  dust cake formation  and  dust  loading
         characteristics  similar  to  conventional  media
      •   Accelerated pulse cleaning  tests  showed  the media  could survive
         mechanical  flexing  associated with 50,000 cleaning  pulses  at
         800°C  (1.5000F)
      •   Dust  loading tests  at  high  temperature  and  pressure  with  flyash
         showed  that high  collection  efficiency  and  good  cleanability
         (control of pressure  drop)  could  be maintained for  200  hours at
         face  velocity  up  to  10 cm/sec (20  ft/min)
      •   Tests  on a  slipstream of gas  at the EPA/Exxon  PFBC miniplant
         showed  the  filter could  perform satisfactorily in  removing
         particles of flyash from an  as generated  gas stream
      •   1,000  hours of exposure  to hot PFBC gas  and dust showed the  media
         could  survive the chemical environment of the  PFBC exhaust

     Continued development of the ceramic filter  is  planned under a
DOE-sponsored program which will culminate  in a subpilot-scale
                                   175

-------
demonstration at the experimental PFBC operated by Curtiss Wright
Corporation in Woodridge, New Jersey.  Successful completion of this
program will provide a capability to control particles including fine
particles at temperatures of 800° to 900°C (1,500° to 1,650°F) and
pressures of atmospheric to over 10 atmospheres.  This will provide
protection for gas turbines in PFBC applications and may eliminate the
need for further cleanup of the effluent gas stream and result in lower
total installation and operating costs.  In addition, the filter should
find application in heat recovery and in industrial processes for
high-temperature particle control.
                                   176

-------
               EOF AND Q-BOP HOT METAL CHARGING EMISSION COMPARISON
                                 C. W. Westbrook
                            Research Triangle  Institute
                          Research Triangle Park, N. C,
                                    ABSTRACT

     Fumes generated during hot metal charging of a EOF and a Q-BOP were
sampled and analyzed for total particulates, particulate mass in four size
ranges, and inorganic and organic compounds.  The data indicate that the Q-
BOP generates three times as much particulate and 15 times as much organic
matter per megagram of hot metal charged as does the EOF.  Polynuclear
aromatic hydrocarbons (PNA) were found in the Q-BOP fume but not in the EOF
fume.  No carcinogenic PNAs were detected.
     The differences found are probably due to the additional time required
to charge hot metal into the Q-BOP at the particular plant sampled as com-
pared to the EOF (2.2 minutes for the Q-BOP versus 1.0 minutes for the EOF)
and to blowing of nitrogen gas into the bottom of the Q-BOP during the
charging operations.
                                     177

-------
              EOF AND Q-BOP HOT METAL CHARGING EMISSION COMPARISON
INTRODUCTION
                                                          1 2
     This paper summarizes results obtained in two studies '  of fumes gene-
rated during hot metal charging of basic oxygen furnaces.  Two types of
vessels were tested.  One, referred to herein as the BOF, is of typical
design in that oxygen is introduced during the "blow" period through a lance
suspended above the metal.  In the second vessel, referred to as the Q-BOP,
oxygen is blown into the vessel through tuyeres in the bottom of the vessel.
A gas, usually nitrogen, is blown into the vessel through the tuyeres at all
other times to prevent their plugging.
     The purpose of the testing was to determine if, and in what amounts, a
wide variety of inorganic and organic materials might be generated during the
hot metal charging and to determine if there is a significant difference
between vessel types.  Although both furnaces tested do have equipment ope-
rating to collect and control the emission of these fumes, no testing was
done on the outlets from the control equipment.  All testing was of the
uncontrolled fume going to the emission control equipment.
     Samples of the fume generated during hot metal charging were collected
using the Source Assessment Sampling System (SASS).  This system, shown in
Figure 1, collects particulate matter in four size fractions (>10 pm, 3-10 ym,
1-3 um, and <1 urn).  It also traps organic vapors with an organic resin (this
section of the train is referred to as the organic module) and certain metals
(mercury, antimony, and arsenic), that may be in the vapor phase, using a
series of impingers.
     I will first briefly discuss the two vessels and the testing conducted
on each.  A direct comparison of the results obtained will then be presented.
Q-BOP TESTING

     The Q-BOP shop tested contains two vessels each rated at 225 tons/heat.
Six to eight heats can normally be completed in an eight hour shift.  Shown
in Figure 2 is a schematic of the Q-BOP vessel and ancillary equipment.  The
Q-BOP vessel differs from the conventional BOF in that oxygen is introduced
through tuyeres in the bottom of the vessel rather than through, a lance above
the charged metal.  When the vessel is being charged nitrogen gas is blown
                                     178

-------
                                                      FILTER
                                                                          GAS COOLER
                             COHVECTION OVEN
                                                                  GAS
                                                                  TEMPERATURE
                                                                  T.C.
                                SORBEMT
                                CARTRIDGE
IMP/COOLER
TRACE ELEMENT
COLLECTOR
                                                 COMPENSATE
                                                  COLLECTOR
DRY 6AS METER ORIFICE METER
 CENTRALIZED TEMPERATURE
  AND PRESSURE READOUT
     CONTROL MODULE
                                            TWO IHrVmi. VACUUM PUMPS
                     Figure 1.  Source assessment sampling system.

-------
                                         BUMPER
CO
o
                               SECONDARY HOOD
                  HOT METAL CHARGING LADLE
                     FURNACE CHARGING DOORS
                     (RETRACTABLE)
                    SLAG POT
WATER COOLED HOOD

HOOD TRANSFER CAR

ADJUSTABLE SKIRT

TAPPING EMISSIONS DUCT
SEAL RING


FURNACE ENCLOSURE
                                                                                                  OPERATING
                                                                                                  FLOOR
                                                                                                     TEEMING
                                                                                                     LADLE
                                                                                                         SHOP AIR IN DRAFT
                                                                                                         DURING SLAGGING
                                                                                                         AND TAPPING
                                               Figure 2.  Schematic of Q-BOP vessel.

-------
 through the tuyeres to prevent their becoming plugged.  The nitrogen flow
 rate is higher during hot metal charging than during scrap charging.
      Shown in Figure 3 is a schematic of the fume control systems.  The
 entire vessel and the fume collecting hoods are in an enclosure (referred
 to as a "doghouse").  The bottom side of the enclosure is open to allow
 tapping and slag dumping.  All fumes generated during the oxygen blow are
 collected with the primary hood.  When the vessel is tilted more than 20°
 (during charging, for example) the secondary hood collection system is
 automatically actuated and the primary system decreased to about 10 to 20
 percent of full capacity (by damper adjustments).  Design flow rates under
 this condition with both fans in operation is 10,600 m /min through the
                                  3
 secondary hood system and 1,250 m /min through the primary system.
      Testing was done on this furnace only during the periods  that  hot
 metal was being charged into the vessel.   Two ladles of hot metal are
 charged for each heat.   The average time  of hot metal addition (2 ladles)
 during testing was 2.2  minutes.   Sampling was in  the secondary fume control
 duct just before the downcomer to the quencher (refer to Figure 3).   Eight
 separate periods of hot metal addition (4 heats)  were sampled  in three
 hours.   The overall sampling rate was near isokinetic (104  percent)—
 single point,  no duct traverse.   Gas  volume collected was  1.027  dry stand-
                                                      3
 ard  cubic  meters.   Throughout  this text the symbol Nm  is used  to mean a
 cubic meter corrected to  20°C, 76.0 cm Hg,  and 0% moisture.
      Given in  Table 1 are the  pertinent process and  sampling data.  The
 scrap and  hot  metal figures  are  the sums  for  the  four heats.  Scrap is
 about 20 percent  of  the total metallics charged.  The results obtained will
 be presented later.
 BOF TESTING
      The BOF shop  in which the testing was conducted contains two vessels,
 each  rated at  about 250 tons/heat.  Normally 6 to 8 heats can be completed
 on each vessel in  an eight hour period."  Shown in Figure 4 is a schematic
 of one of  the vessels in  the shop.  Although this schematic is very similar
 to that for the Q-BOP, some important differences should be noted.   First,
 and most important, there is no injection of gases (oxygen or nitrogen)
 through the vessel bottom.  Thus, during scrap and hot metal charging, no
gas (nitrogenl is blown into the metal.  You will also note that secondary
                                     181

-------
     0 BOP NO. 1 FURNACE ENCLOSURE


     SECONDARY HOOD NO. 1
    0-IOP NO. 2
    FURNACE ENCLOSURE
                                               FAN NO. I
            (ELL VALVE NO. 2

    SHUTOFF NO. 2 CLOSED
                                                         STACK NO. 1
                                                          STACK NO. 2
                                                 FAN NO. 2
SCRUBBER NO. 2
Figure 3.  Gas collection system for Q-BOP.

-------
                   TABLE 1.  PROCESS AND SAMPLING DATA, Q-BOP
Charging Data


                              Scrap, tons (Mg)          Hot Metal,  tons (Mg)


     Total, 4 heats             206.5 (187.3)                822.7  (746.3)


     Average/heat                51.6 (46.8)                 205.7  (186.6)


Sampling Data


     Volume Gas Sampled                                       1.027  Nm


     Stack Gas, Temperature                                  66.5°C


     Stack Gas, Velocity                                     31.9  m/s

                                                                  3
     Stack Gas, Flow Rate                                 11,491 Nm  /mln


     Sampling Rate,  % Isokinetic                             104
                                     183

-------
I
                      CHARGING FUME HOOD
              HOT METAL CHARGING LADLE
         FURNACE CHARGING DOORS
         (RETRACTABLE)
              SLAG POT
                                                                                        TAPPING FUME HOOD
                                                                                         FURNACE ENCLOSURE
                                                                                       WATER COOLED PRIMARY HOOD
OPERATING FLOOR


   TEEMING LADLE
                                                Figure 4. Schematic of BOF vessel.

-------
 fume collection hoods are on both the charging and on the  tapping side of  the
 enclosure.
      Shown  in Figure 5 is a schematic of  the secondary fume  collection system.
 Interlocks  in the system prevent suction  on the tapping hood or  primary hood
 during the  charging operations.   Also,  suction at  the hot  metal  transfer and
 reladling station and on the charging fume  hoods of the second vessel  is pre-
 vented during charging the test  vessel.   During the tests, the second  vessel
 was not in  operation.   Thus,  only hot metal charging fumes were  contained  in
 the gas sampled.
      Testing was  conducted only  during  periods  of  hot metal  charging.  At
 this plant  all hot metal needed  for the heat is charged using only one ladle.
 Average time required  for hot metal charging during the testing was 1.0
 minutes.  In order to  obtain sufficient sample,  sampling was  carried out
 during 24 heats over a five day  period.   The sampling system was sealed
 between test periods.   Only the  second  and  third impinger  solutions (which
 contain unstable  reagents)  were  recovered each  day.   Therefore, data for
 individual  heats  are not available.   Shown  in Figure  5  is  the sampling point
 for these tests.   The  overall sampling rate was  115.5 percent isokinetic.
 Total gas collected was 2.795 dry standard cubic meters.
      Given  in Table 2  are the pertinent process and sampling data.  Both the
 total amount of scrap  and hot metal charged  for  the 24 heats and the average
 amount per  heat are given.  The  scrap charge  for individual heats did not
 vary from the average  by more than ±  19 percent.  Hot metal charge for indi-
 vidual heats did  not vary from the average by more  than ± 7 percent.  Scrap
 accounted for about  26  percent of  total metallics charged.
      For  the sampling  data, note  that samples were  taken from only one of the
 two  charging fume  collection  ducts.  The gas flow rate in each duct was
 measured  and found to be  essentially the same.  Therefore,  to calculate total
 flow rate and particulate mass, the results obtained were multipled by two.
 DATA COMPARISONS
 Particulates
      In both tests the  total particulate concentration in the gas was deter-
mined using  the SASS train.  The SASS train collects the particulate in four
 size  fractions as  given earlier.   Additional particulate, probably containing
                                      165

-------
ENCLOSURE
          \                             CON
TAPPING AND SLAGGING FUME
     CONTROL HOODS
                                                r
                                                                                   ENCLOSURE






RELADLIMG



HOT METAL
TRANSFER






DAM




DAMPER





L
IPERS
\
V
1






oc
a
*














1




Jl


.

\ J
^ 7^
CHARGING FUME
EONTROL HOODS




f
DAMPER
Y
X












I
*












J









IT
a
*







\
DAMPER
Y
A














X J.
\ /
CHARGING FUME
CONTROL HOODS



















cc
^







1
1
1
1
1
1
1
1
1
1
1






                                    TO BAGHOUSE
                    Figure 5.  BOF shop ncondary fume control tystam.

-------
                    TABLE 2.  PROCESS AND SAMPLING DATA, EOF
Charging Data


                              Scrap, tons (Mg)         Hot Metal, tons


     Total, 24 heats            1598 (1450)                 4494 (4077)


     Average/heat                 66.6 (60.4)                187.2 (169.8)



Sampling Data


     Number of Heats Sampled                                 24

                                                                      3
     Gas Volume Sampled                                       2.795 Nm


     Stack Gas, Temperature                                  85.3°C


     Stack Gas, Velocity                                     26.5 m/s


     Stack Gas, Flow Rate—Total Both  Ducts              11,509 Nm3/min


     Sampling Rate,  % Isokinetic                            115.5
                                    187

-------
some of all size fractions, is trapped in the probe.  Given in Table 3 are
the total particulate and size fraction data.
     The particulate concentration in the gas from the Q-BOP was substantially
higher than from the EOF (1298 versus 788 mg/Nro3).  There is also a signifi-
cant difference in the particulate size distribution.  For the ROF, 27
percent of the particulate was less than 3 ym in size versus 16 percent of
the Q-BOP particulate in this size range.  About 70 percent of the Q-BOP
particulate was larger than 10 ym whereas only 35 percent of the EOF partic-
ulate fell in this size range.
     Shown in Table 4 is the mass of particulate generated in each size
range per ton (or megagram) of hot raetal charged.  The important point to
consider is that the Q-BOP appears to generate about three times as much
particulate as does the EOF (per unit of mass of hot metal charged).  Thus,
about 17.6 kg of particulate is generated by the Q-BOP per 100 Mg of hot
metal (35 lb/100 tons hot metal) versus about 5.3 kg of particulate per 100
Mg hot metal (10.6 lb/100 tons hot metal) generated by the EOF.
Inorganics
     Given in Table 5 are the concentrations of a number of elements in the
uncontrolled fume from the two vessels and the amount generated per megagram
(1.1 tons) of hot metal added.  There appears to be no substantial difference
in the amounts of the elements contained in the fume.  The calculations for
the BOF are clouded somewhat since some of the analyses for individual SASS
train components were reported only as MC (Major Components).  For the BOF
data, a > (greater than)  symbol indicates that at least one of the SASS
component samples was analyzed as a major component.  MC indicates that the
element was a major component in all SASS samples.
     For the Q-BOP, nickel, iron, chromium,  calcium, arsenic, lead, and
possibly sulfur and phosphorous are at sufficiently high concentrations that
the fume should not be emitted uncontrolled.  For the EOF, the concentrations
of barium, cadmium, selenium, arsenic, chromium, nickel, and possibly lead,
calcium,  and iron are sufficiently high that fume control is needed.  Both
plants tested do control the fume from hot metal charging.  No testing was.
done on the outlet from the control device.
                                     188

-------
TABLE 3.  PARTICULATE SIZE DATA FOR EOF AND Q-BOP

EOF 3
Size Range mg/Nm
<1 y 85
1-3 y 124
3-10 y 206
>10 y 274
Probe, Cyclone Washes 99
Totals: 788
%
11
16
26
35
12
100
Q-BOP-
mg/Nm
26
182
69
892
129
1298
%
2
14
5
69
10
100
                       189

-------
                TABLE 4.  PARTICULATE MASS DATA FOR EOF AND Q-BOP*
EOF
Size Range Ib/ton
<1 pm 11.6
1-3 ym 16.8
3-10 vm 28.0
>10 ym 37.0
Probe, Cyclone Washes 13.4
Totals: 10.6
x IO"3
x 10~3
x 10~3
x 10" 3
x 10~3
x IO"2
5.
8.
14.
18.
6.
5.
kg/Mg
8 x
4 x
0 x
5 x
7 x
3 x
io-3
io-3
10'3
io-3
io-3
IO-2
7
49
18
242
35
35
Q-BOP
Ib/ton
.0
.4
.6
x
.0
.2
x 10
x 10
x 10
10"3
x 10
x 10
-3
-3
-3

-3
-2

3
24
9
121
17
17
kg/Mg
.5
.7
.3
x
.5
.6
x 10"3
x 10"3
x 10"3
ID'3
x 10~3
x 10"2
*Calculated on the basis of hot metal added.
                                     190

-------
              TABLE 5.   SELECTED INORGANICS IN EOF AND Q-BOP FUME
Element
Aluminum
Antimony
Arsenic
Barium
Bismuth
Cadmium
Calcium
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
Phosphorous
Selenium
Silicon
Strontium
Sulfur
Zinc
BOF
Concentration
In Gas
MC
0.006
50.05
>0.11
0.003
0.077
MC
0.84
0.18
>2.7
>0.02
>0.55
>0.11
0.0008
0.31
MC
0.087
>1.1
0.016
>0.06
MC
Total Generated
mg/Mg Hot Metal

0.4
£3.4
>7.4
0.2
0.47

57
12
>183
>1.4
>37
>7.4
0.05
21

5.9
>74
1.1
>4

Q-BOP
Concentration Total Generated
+
In Gas mg/Mg Hot Metal
0.43
<0.001
<0.02
0.02
<0.0007
0.002
64
0.26
0.1
85.3
0.41
2.3
3.8
>0.0031
0.18
0.53
<0.033
4.2
<0.056
7.9
0.14
58
0.14
<2.7
2.7
0.09
0.27
8671
35
14
11,557
56
312
5]5
0.42
24
72
4.5
569
7.6
1070
19
Concentrations in mg/Nm .
                                      191

-------
 Organics
      Given in Table  6 are  the  total organics generated during hot metal
 charging  of the  EOF  and Q-BOP.  Note that the total organics generated per
 megagram  of hot  metal added  is about 15 times higher for the Q-BOP than for
 the BOF.   Also note  that about 2/3 of the Q-BOP organic is adsorbed on the
 particulate matter but less  than 10 percent of the organic from the BOF is
 adsorbed  on the  particulate.
      Shown in Table  7 is a breakdown of this organic matter into the major
 organic compound categories.   The major categories for BOF organics are
 aliphatic and aromatic hydrocarbons, and esters.  Note that no evidence for
 fused aromatics—polynuclear—was found for the BOF.  For the Q-BOP, the
 major categories are aliphatic hydrocarbons and fused (or polynuclear)
 aromatics.   A variety of compound types are found in the Q-BOP organic.
      A low resolution mass spectrographic analysis of the Q-BOP organic
 indicated the presence of organics with masses equivalent to the masses of
 known carcinogens.   The sample was further analyzed by GC/MS.  No carcino-
 genic organic compounds were found in this analysis.
 CONCLUSION
      In this  paper a comparison has been made for the fumes generated during
 hot metal charging of a BOF and a Q-BOP.  The data indicate that particulate
 generated in  this operation is substantially greater for the Q-BOP than for
 the BOF.  This is probably directly related to blowing gas into the bottom
 of the Q-BOP  and through the metal charged.  Data have been presented to
 show  that there  is not a significant difference in the types of inorganics
 in the fume.
     The data also indicate that the fume from the Q-BOP contains substanti-
 ally more, and significantly different types, of organic matter than fume
 from the BOF during hot metal charging.   Although fume from the Q-BOP does
contain polynuclear aromatic hydrocarbons,  and none of these materials was
 found in fume from the BOF, no evidence  of carcinogenic organic compounds
was found.
                                      192

-------
TABLE 6.  TOTAL ORGANICS IN BOF AND Q-BOP FUME
                            BOF
Q-BOP
In SASS Participates, mg/Nm
                             +>
In SASS Organic Module, mg/Nm"

                    3
Total Organic, mg/Nm


kg Organic/Mg Hot Metal


Ib Organic/ton Hot Metal


kg Organic/heat


Ib Organic/heat
    0.7


    7.9

    8.6


 5.8 x 10

11.6 x 10


 9.9 x 10

19.8 x 10
                                 -4
                                 -4
                                 -2
                                 -2
    43.6


    20.5

    64.1


 8.7 x 10


17.4 x 10

     1.62

     3.24
      -3
                      193

-------
              TABLE 7.  MAJOR ORGANIC COMPOUND CATEGORIES
              CONCENTRATIONS IN UNCONTROLLED FUME,  mg/Nm3
                                        EOF                 Q-BOP
Aliphatic Hydrocarbons                  0.3                  7.7
Aromatic Hydrocarbons                   0.2                  1.4
Fused Aromatics                         0                    4.9
Alcohols                                0                    2.6
Amines                                  0                    1.3
Amides                                  0                    1.3
Esters                                  0.9                  2.6
Carboxylic Acids                        0                    3.0
                                  194

-------
                                 ACKNOWLEDGMENT

     This work was supported by the U.S. Environmental Protection Agency,
Industrial Environmental Research Laboratory,  Research Triangle Park,  N. C.
under EPA Contracts 68-02-2630 and 68-02-3152.  Mr.  R. V.  Hendriks and Mr.
R. C. McCrillis were the EPA Project Officers  for the Q-BQP and EOF studies,
respectively.
                                      195

-------
                                    REFEENCES
1.   Westbrook, C. W.  Level 1 Assessment of Uncontrolled Q-BOP Emissions.
     EPA-600/2-79-190, September 1979, 85 pp.

2.   Westbrook, C. W.  Hot Metal Desulfurization, EOF Charging and Oxygen
     Blowing:  Level 1 Environmental Assessment, October 1980.  In pre-
     paration for publication.
                                      196

-------
                         FIELD EVALUATION OF FUGITIVE
                     EMISSIONS FROM EOF STEELMAKING SHOPS

                                 Prepared by:

             Peter D. Spawn, Thomas J. Nunno and Stephen G. Piper

                            GCA/Technology Division
                              213 Burlington Road
                         Bedford, Massachusetts  01730

                                      and

                               Larry F. Kertcher
                     U.S.  ENVIRONMENTAL PROTECTION AGENCY
                           230 South Dearborn Street
                           Chicago,  Illinois  60606
                                   ABSTRACT


      This  paper presents  the  findings  of week-long  field  evaluations  of
emission controls  of  six  (6)  EOF steelmaking  shops.   The  field  evaluations
included shops  with complete  furnace enclosure,  Gaw damper  plates,  and  tap-
side  enclosures.   For each  evaluation, process engineers  stationed  inside
the furnace control room  and  on  the operating floor carefully documented
process operations.   Simultaneous with in-shop process evaluation,  outside
observers  recorded visible  emissions from  the shop  roof monitors in accordance
with  EPA Method 9.  FM radio  contact between  in-shop and  outside observers
enabled documentation of  the  source of all roof monitor emissions.  Perfor-
mance of each control system  is presented on  the basis of average roof monitor
opacity for each furnace  operation, i.e., charge, oxygen blow,  tap, etc.  Also
investigated were  the potential causes of roof monitor emissions such as
variations in steelmaking procedures and/or emission  control operational
vn-ri flfil oa
variables.
                                      197

-------
     FIELD EVALUATIONS OF FUGITIVE EMISSIONS FROM EOF STEELMAKING SHOPS
INTRODUCTION

     GCA is conducting field evaluations of secondary emission controls at
EOF shops for EPA Region V.  The primary objective is to document the opacity
of roof monitor visible emissions (RMVE) for each vessel operation; i.e.,
scrap charge, hot metal charge (HMC), oxygen blow and reblow, turndowns and
tapping.  A secondary objective is to evaluate all process data that describes
how the vessels and the emission control systems were operated in order to
achieve the performance levels observed.  This second objective has become
an important aspect of the program since operational practice has such a
strong impact on emissions that escape capture and are emitted from shop roof
monitors.

     This paper summarizes the highlights of six (6) EOF studies conducted
between June 1979 and August 1980.  When reviewing these data, the reader
should recognize that the control system and emissions data describe the sys-
tem performance observed by GCA during the week(s) of testing.  The final
reports prepared for each field test contain a complete record of steelmaking
process variables and emission control operation during testing.  Evaluations
of other EOF shops are currently in progress.

DATA COLLECTION AND ANALYSIS METHODOLOGY

     Field evaluations are conducted by GCA engineers familiar with EOF shop
operations.  Each team member is certified to observe visible emissions (VEs)
in accordance with EPA Method 9.  RMVEs are continuously recorded by outside
observers while inshop observers document shop operations.  Continuous FM
radio contact between all observers enables determination of the origin of
each RMVE.  This approach is discussed further below.

Visible Emissions Observations

     VE data are collected in accordance with EPA Reference Method 9 as pub-
lished in the Federal Register, 40 CFR Part 60.275, Appendix A.  Ground level
positions are normally used to evaluate emissions against the sky.  Occasion-
ally, RMVE are blown horizontally across the length of a shop roof monitor.
As directed by Method 9, these emissions are observed against the best avail-
able contrasting background which is usually the roof monitor area.  Normally,
however, VE observers are able to view emissions at a point just above the
roof monitor.  This approach avoids the possibility of inaccurate measurements
in the event the plume travels horizontally for a brief period before rising
upwards.  The densest portion of the plume is observed, as specified in
Method 9.

     Constant FM radio contact enables precise documentation of which shop
operation caused each visible emission.  By receiving instant radio confirma-
tion of what process is occurring, VE observers know when emissions of no
interest occur; i.e., hot metal transfer, skimming, and teeming.
                                       198

-------
 Inshop Process Observations

      Process observers are stationed at strategic locations within the shop
 to allow observation of all emission-causing operations.  An engineer is
 assigned to each operating furnace to record process data and the precise
 time each operation occurs - scrap charge, HMC, oxygen blow, turndowns, etc.
 This observer closely monitors the control room instrumentation and documents
 the steelmaking techniques in use.  He also observes each vessel operation from
 the working floor and estimates the approximate magnitude of emissions that
 escape capture.  Constant radio contact with outside VE observers enables
 determination of the origin of each RMVE.   Process data collected in the con-
 trol room generally consists of:  oxygen blow rates and net quantity blown,
 lance height, at each interval, tons of steel and scrap charged, hot metal (and
 tap) temperature and chemistry, time of flux additions, etc.

      Emissions escaping capture at the furnace are evaluated by one of two
 techniques.   When the uncontrolled emission can be observed, a percent capture
 is estimated.   If the total uncontrolled emission is not visible,  the observer
 records the  magnitude of emissions escaping capture by the following criteria:
 light,  5 to  25 percent opacity, (as viewed within the shop); moderate,  30 to
 60 percent opacity;  and heavy,  65 to 100 percent opacity.   Also recorded  are
 the duration of each vessel operation and  any unusual operating practices.

      When necessary,  another process engineer records process data at the gas
 cleaning device,  to  document control system operation and  the exhaust flow
 rate applied to the  furnaces.   Often,  these data are available in  the EOF con-
 trol room, or  continuously recorded at the gas cleaning device.

      In some shops,  observer(s)  are required to document the precise  time of
 non-furnace  operations such as  teeming,  hot metal  transfer,  or skimming.
 Often,  emissions  from these operations can be observed  by  one of the  furnace
 observers.

 Data Handling  and Assessment

      After each day,  observers review  their  data sheets for  clarity and con-
 sistency.  The Lead Engineer collects  all  data  and reviews them once again.
 VE data  sheets are handled by the  chain-of-custody technique  and kept in a
 secure area  at GCA until the final reports are  submitted to EPA.

      In  the  office, the average opacity  for each vessel operation  is calcu-
 lated.   If requested, compliance with  the  applicable opacity  standard is also
 determined.  When multiple shop or furnace operations occurred, the corres-
 ponding VE observations are not included in the summary tables.  Only the
 HMVE whose origin can be absolutely determined by the radio communication
 technique are used to compile the  data presented in this paper.

 BOF SHOP AND EMISSION CONTROL SYSTEM DESCRIPTIONS

     The six (6) BOF shops and their control systems are briefly described
below.  Additional data is available in the GCA report for each shop and also
in the open literature.
                                      199

-------
Republic Steel, South Chicago, Illinois

     The two 204-m ton* (225-ton) Q-BOP furnaces at RSC/Chicago replaced four
204-m ton  (225-ton) open hearth furnaces in 1977 and operate within the old
open hearth shop.  The eight Q-BOPs currently operating in the U.S. represent
more than  half the number of basic oxygen pjrocess furnaces entering service
in the last 7 years.  The six other Q-BOPs are operated by US Steel, three
apiece at  Fairfield and Gary.

     The Q-BOP differs from the conventional BOF in that instead of top
blowing oxygen through a water-cooled retractable lance, oxygen and fluxes
are blown  through the bath via tuyeres in the furnace bottom.  Because the
vessel is  bottom blown, the Q-BOP has some unique emission control problems
when compared to a conventional BOF.

     Primary emissions generated during the oxygen blow are captured by a
retractable water-cooled skirt that is lowered over the mouth of the furnace,
acting as  a conventional, suppressed combustion hood.  Furnace off-gases are
cooled to  1093°C (2000°F) in water-cooled ductwork before entering the
quenchers.  The quenchers consist of two parallel Venturis followed by impact
separators that cool the gas stream to 80°C (175°F).  Cooled gases pass
through the venturi scrubbers which operate at a pressure drop of 150 mm Hg
(80 in. W.C.) during the oxygen blow.  Particulate entrained in the water
droplets exit the venturi section and enter a fixed vane separator before the
exhaust passes through the induced draft (ID) fan and stack.

     At RSC, only one furnace is normally on-line since the enclosure system
is designed so the fan on the down furnace supplements the operating furnace's
fan for more draft during charging and tapping.  Captured secondary emissions
are exhausted to two venturi scrubber systems that operate in parallel.
However, the quenchers are bypassed since the secondary exhaust stream is
much cooler than the primary exhaust.

     All secondary emissions are captured through a charge side hood located
inside the enclosure.  During scrap and hot metal charges, the enclosure door
opening is minimized, and an exhaust flow rate of about 170 m3/sec (360,000
acfm) is provided to the charge-side hood by both fans.  During a tap, the
charge-side doors remain closed, and the charge-side hood operates at 170 m3/
sec (360,000 acfm).  Total exhaust is reduced to about 144 m3/sec (305,000
acfm), during turndowns for sampling.  During idle periods; i.e., waiting for
the chemistry analysis, the system operates at about 106 m3/sec (225,000 acfm).

Republic Steel, Gadsden, Alabama

     Primary emissions generated during the oxygen blow are captured by a
water-cooled, full combustion hood located 0.75 m (2-% ft) above the vessel
mouth.  Before entering the ESP, furnace exhaust gases are cooled by a cas-
'cade-type  evaporative chamber which also improves particle resistivity for
better collection.  Full system draft of approximately 30 mm Hg (283 m3/sec or
600,000 acfm measured at the stack) is used during the HMC and oxygen blow.
Partial draft of approximately 18.7 mm Hg (165 m3/sec or 350,000 acfm measured
at the stack) is used during scrap charging, turndowns, tapping, and slagging.
Lance hole covers were present on one vessel during GCA's evaluation.

*Metric ton.


                                       200

-------
      The  Gaw damper (closure plate)  is a water-cooled  steel plate which slides
 in a horizontal plane across the opening of  the primary hood.   The  damper  re-
 duces the hood  opening and increases the capture velocity of the primary ex-
 haust system.   The Gaw damper was designed to  close off 75 to 80 percent of
 the hood  opening during the HMC according to design data.   The damper  is de-
 signed for a furnace tilt  angle not  exceeding  45 degrees from vertical.

      The  first  Gaw damper  was retrofit to the  No.  1 vessel in June  1977 at a
 cost of $350,000 according to RSC.   The leading edge of the plate warped badly
 after 6 to 9 months of operation, and water  cooling was installed on this
 edge of the plate.   The maintenance  cost of  the water  system proved less than
 the replacement cost of the plate, and a similar,  water-cooled Gaw  damper  was
 installed on the No.  2 vessel in November 1978 at  a reported cost of $425,000.

      During tapping,  the rotation of the vessel places the mouth of the furnace
 a  considerable  distance from the hood opening.   To help contain tapping emis-
 sions and divert them into the primary hood, a tap-side enclosure was  construc-
 ted on each furnace.   The  Gaw damper is  not  used for tapping emissions  control
 because the necessary I-beam supports would  interfere  with charging at  the
 other side of the furnace.

 U.S.  Steel,  Gary No.  1 BOF

      This shop  is an open  hearth conversion  and the  three  195-m ton (215-ton)
 vessels entered service in 1965.  Annual production  capacity of carbon  and
 alloy steels is 4.0 MM m tons (4.4 million tons) for the three vessels
 combined.

      U.S.  Steel considers  this  shop  a "one-and-a-half"  vessel operation.  Norm-
 ally,  two furnaces  are available for.steel production,  with  the third either
 on standby or down for reline.   One  vessel normally  operates at maximum pro-
 duction to supply the  continuous caster, producing  9 to 10 heats per 8-hour
 turn.   The second vessel usually operates in a manner not  interfering with the
 principal vessel,  and  produces  5  to  10 heats per turn.   Normal shop production
 is  51 heats  per 24-hour  period.

      Primary emissions are collected  by  conventional water-cooled,   full com-
 bustion hoods at  each  of the  three vessels.   Each hood  discharges into a com-
 mon exhaust  manifold  that  leads  to the scrubber system.  Isolation dampers close
 off the off-line  furnace(s).
     The gas cleaning system consists of two identical scrubbers operating in
parallel.  Furnace exhausts pass through a variable throat venturi, through
a flooded elbow, and into a separation tower.  Exiting the tower, gases pass
through a 4100 kW (5500 hp) fan (at -103 to -150 mm Hg or -55 to -80 in.  W.G.
static pressure) and into a final separator.

     The emission control system was designed for two furnaces on oxygen blow
and one furnace on HMC, with a total exhaust flow of 307 m3/sec (650,000
scfm), sat.  Induced draft fans operate in three distinct modes, drawing 1800,
2400 or 3800 kVa (each).  Fan load is mainly controlled by upstream louvers
                                       201

-------
which operate at 50, 80, and 100 percent open for the three modes.  Fan loads
are also a function of the automatic control of the variable throat Venturis.

     Exhaust flow rates applied to each furnace depend on the isolation damper
position, the fan load and venturi pressure drop, and the number of furnaces
on-line.  The flow rates at each furnace are measured in terms of static pres-
sure in the quench section and recorded on continuous strip charts.  The isola-
tion damper set points are 20 percent open for scrap charging, turndowns and
tapping, and 100 percent open for HMC and oxygen blows.

     The Gaw dampers at Gary were designed to close off 70 percent of the hood
opening during HMCs, according to design data.  The damper is not used for
scrap charging control because of possible damage from large scrap pieces, and
because the furnace mouth is tilted too far away from the exhaust hood.

     Furnace tilt during HMC is limited to 35° from vertical.  Charging ladle
spout extensions enable positioning of the ladle well into the furnace, bring-
ing the point of emissions closer to the primary hood.  Sheet metal curtain
walls constructed above the furnace and Gaw damper, on the charge side, help
minimize adverse effects of cross-drafts.  The crane operator has instructions
to slow or stop pouring should HMC emissions escape capture.

     Total installed cost of the three Gaw damper systems in late 1978 and
early 1979 was $750,000 according to U.S. Steel engineers.  To date, there
have been no major operation or maintenance problems with the dampers.  The
plates are constructed of one-quarter inch carbon steel and have not become
warped.  Only the rails supporting the wheels upon which the plate rides into
position are water-cooled.  The original wheel design had to be modified to
avoid binding from build-up of dusts in the rails.  The damper is installed
6-inches below the primary exhaust hood because of the proximity of a major
structural beam.

     To help contain tap fumes and direct them into the primary exhaust hood,
a tap-side enclosure was installed in 1977 on each furnace.  The furnace iso-
lation damper is automatically set at .20 percent open during tapping.  The
Gaw damper is not used.  No significant maintenance or operational problems
exist with the current enclosures according to U.S. Steel engineers.

National Steel, Great Lakes Steel Division

     Primary emissions generated during the oxygen blow are captured by full
combustion hoods located directly over each furnace mouth.  Air is pumped into
the hood system, approximately 0.33 m (10 feet) above the hood mouth, at a rate
of 28.3 m3/sec (60,000 scfm) to help combust carbon monoxide.  Temperature
activated water sprays in the exhaust hood ducting control moisture and temp-
erature.  The oxygen lance opening in the primary exhaust hood is fitted with
a sliding disk seal to reduce the escape of emissions and loss of draft.

     To help contain tapping emissions, doors (referred to as garage doors)
were installed on the tap-side of the furnace.  The doors retain tap emissions
long enough to allow fumes to be carried into the primary exhaust hood.
                                       202

-------
 Inland Steel, No. 4 EOF Shop

      Inland Steel operates two basic oxygen furnace (EOF) shops at the Indiana
 Harbor Works.  Two 231-tn ton (255-ton) top blown furnaces at the No.  4 shop
 entered service in July 1966, and presently have an annual rated capacity ex-
 ceeding 3.6 million m tons (4.0 million tons) of steel.

      Primary emissions generated during the oxygen blow are captured by a full
 combustion-type hood over each furnace.  No secondary emission controls are pre-
 sent for charging, turndowns, tapping, or slagging.  Steam rings located on the
 circumference of the lance holes above each combustion hood, help to control
 the escape of oxygen blow emissions.

      Furnace exhaust passes from the primary exhaust hood through the spark
 box for water spray cooling,  and into a quench tower.   The gas stream is then
 spilt into four identical Buell venturi scrubbers operating in parallel.   Water
 laden gas enters a moisture separator prior to the four parallel, 1300 kW
 (1750 hp) ID fans.  The four  streams are combined and discharged through a
 single stack.

      The emissions control system is designed to operate with one furnace on
 oxygen blow.   Guillotine dampers isolate the off-line vessel.   According to
 stack tests performed in 1974,  the exhaust flow through the scrubbers is
 229 mVsec (484,500 acfm), at 58°C (128°F).

 Ford Steel Division EOF Shop, River Rouge Plant

      The two 227-m ton (250-ton)  top blown BOFs,  installed in  1964, have  com-
 bined annual production capacity of 2.7 million m tons (3 million tons) per
 year.   The shop produces primarily low carbon sheet for use in automobile
 manufacturing.

      The scrap  charge,  composed  of home and purchased  scrap,  is predominantly
 clean No.  1 bundles  from nearby  auto stamping facilities.   The two BOFs employ
 a  three-holed  lance  with an oxygen blow rate of 10.4 m3/sec (22,000 scfm).

      Primary emissions  are captured by  a full  combustion,  water-cooled hood.
 Exhaust  gases.are  cooled in a cascade-type evaporative  chamber prior  to
 entering an ESP.   Steam injection at the evaporator chamber improves  particle
 resistivity.  An isolation damper regulates exhaust rates  during various  fur-
 nace  operations.   The ESP  is  equipped with four fans,  dividing the gas flow
 among eight ESP banks with four fields  per bank.   The pressure drop measured
 across all  compartments  is  about  5.6  mm Hg (3  in. W.C.).

      Secondary  emissions during HMC  are  controlled  by Gaw  damper plates in-
 tailed  in  1973.  GCA observed the plate  to  close off approximately 50 percent
 of the primary  hood opening.  The vessel  is  tilted  30°  from vertical at the
beginning of the HMC and lowered  to 45°  from vertical at the end of the HMC.
 Since GCA's evaluation, Ford has  removed  the Gaw damper and is experimenting
with a modified doghouse enclosure to control  secondary emissions,
                                       203

-------
ANALYSIS OF FUGITIVE EMISSIONS FROM EOF SHOP OPERATIONS

     The following section characterizes the performance of six EOF secondary
emissions control system relative to the following EOF process operations:

     •    Scrap Charge (SC)                •    Turn Down (TD)

     •    Hot Metal Charge (HMC)           •    Tapping

     •    Oxygen Blow (02 Elow)            •    Slagging

Note that the findings presented herein are based on detailed analyses of pro-
cess information available for six BOF studies.  In some cases, the analyses
are somewhat limited by the extent of process data supplied to GCA, and monetary
constraints imposed by each project.

     For comparison of RMVEs between shops, the authors employed an analysis
tool termed the opacity-duration product.  The opacity-duration product is de-
fined as the product of:  (1) the percentage (in fractional form) of operations
showing RMVEs greater than zero percent opacity; times (2) the average opacity"
of RMVEs greater than zero; times (3) the average duration of those RMVEs.  The
opacity-duration product essentially represents an average visible emissions
flux for each furnace operation.

Scrap Charge

     Each heat begins with the addition of cold scrap to the empty vessel.
Many of the shops evaluated were equipped with hydraulic (Calderon type) scrap
charging systems although only GLS and Inland regularly used their systems.
RSC/Gadsden and GLS employed two scrap charges while the other shops charged
scrap in a single operation.

     RMVEs from scrap charging represented a relatively small portion of BOF
secondary emissions in these shops.  SC emissions resulted from Ignition of
oily material and other combustibles as the scrap contacts the hot vessel.
Control of SC emissions was minimal at the shops evaluated except at the
RSC/Q-BOP where SC emissions are captured by the secondary hood within the
enclosure.  At most of the BOF shops, primary exhaust drafts were dampered off
to between 10 to 30 percent open, during the SC.  Gaw Dampers are not designed
to control SC emissions since they are not effective in controlling scrap
charge emissions.  Also dampers could be damaged by large pieces of scrap.

     Table 1 summarizes the RMVE data collected (for SC) at the six BOF shops.
GLS/Ecorse and the RSC/Q-BOP showed the lowest RMVEs for scrap charging.  Note
that while the estimated capture of SC emissions at the RSC/Q-BOP was 90 per-
cent, GLS/Ecorse showed zero RMVEs with reduced hood capture of about 50 percent.
The data indicate that other factors such as scrap composition and quality can
play an Important role in reducing SC emissions.

     Five process parameters typically evaluated for their relationship with
scrap charge RMVEs were:  (1) scrap composition and quality; (2) scrap quan-
tity; (3) duration of scrap charge; (4) quantity of slag in vessel; and (5)
temperature of vessel.
                                      204

-------
                                             TABLE  1.   SUMMARY  OF  SCRAP CHARGE  DATA FOR  SIX BOF SHOPS
tn
BOF shop
RSC/Chlcago
Q-BOP


RSC/Gadsden
(Avg. of 2
vessels)
USSC/Gary
No. 1 BOF
(Avg. of 3
vessels)
Inland Steel
No. 4 BOF
GLS/Ecorse
No. 2 BOF

Ford Steel


SC
emission
control
Doghouse
enclosure
w/ local
hood
None


None



None

None


None


Uncontrolled
magnitude8
Light



Moderate
to heavy

Light



Moderate

V. light


Light


Percent
of scrap
charges
showing Average Average
Estimated KMVE > 0 opacity of duration of Peafe RMVEs
capture (Z) KHVE* > 0 RMVEs > 0 (Z opacity)
90 18 6 15 10



50 48 8 62 40


10 11 15 50 40



0-10 13 5.0 45 5.0

~50 0000


~50 20 14 95 50


Opacity-
duration
productb
(Z opacity-
seconds)
17



230


83



29

0


270


Comments
Very clean scrap, minimal fines,
mostly ingots, butte and reclaim.
Dolomite sprayed on scrap prior to
charging.
Two scrap charges by transfer crane.
Hone scrap and No. 1 and No. 2 pur-
chased scrap charged.
Vessel isolation dampers closed to
20Z open during SC. 60 to 70Z clean
scrap, ~10T bundles, 12Z pit, —I 23
misc. Generally low oil, clean scrap.
Medium quality scrap, few bundles,
sotae oil in scrap.
Two scrap charges, clean bundled scrap,
probably stampings from auto mfg.
Mostly purchased scrap.
Clean bundled scrap, probably stampings
from auto mfg. Home and purchased
scrap .
              lilght - 5 to 25 percent opacity for emissions escaping the vessel, observed inside the shop.
              Moderate - 30 to 60 percent opacity for emissions escaping the vessel, observed inside.
              Heavy • greater than 60 percent opacity for emissions escaping the vessel, observed Inside.

              Product " portion of SC showing emissions z average opacity x average duration of emissions.

-------
     At one shop, correlations between scrap composition and RMVEs from scrap
charging were reflected in the data as shown in Table 2.  Generally, clean home
scrap (Grade 1) from the Gadsden mills generated somewhat lower RMVEs compared
to purchased scrap (No. 2 heavy melt, consisting of recycled machinery).

       TABLE 2.  SCRAP COMPOSITION AND EMISSIONS GENERATED DURING SCRAP
                 CHARGING AT RSC/GADSDEN EOF SHOP
              Scrap type
                             Scrap
                          composition
  RMVEs**
 range of
(% opacity)
          Home Scrap
          Clean home scrap
                    Blooming mill scrap
                    Rotary shear scrap
                    Teeming mold scrap
          Medium grade scrap  Strip mill scrap

          Galvanized scrap    Galvanized sheet metal

                              Iron bearing BOF slag
Reclaim

Purchased Scrap

No. 1 heavy melt


No. 2 heavy melt
                              Recycled machinery
                                (clean)

                              Recycled machinery
                                (low quality, with oil)
  0 to 5



  0 to 5

  5 to 10

  0 to 35



  0 to 5


  5 to 40
          aFor all heats* both vessels, measured at BOF shop roof
           monitor.

     At RSC/Gadsden, strip mill scrap appeared to contain the greatest amount
of oil, while blooming mill scrap (low oil content), charged alone, generated
almost no emissions.  Cha-fging of galvanized metal generated white zinc oxide
RMVE ranging from 5 to 10 percent opacity.  Reclaim, a low quality scrap, often
generated fine-sized particulate emissions during charging.  Emissions from
purchased scrap (No. 1 and No. 2 heavy melt) are attributable to foreign, non-
ferrous material present in the scrap.

Hot Metal Charging

     In all shops evaluated, hot metal is charged to the vessel by overhead
cranes.  The slowest HMC rate was 63.6 m tons/min (70 tons/min) at RSC/Gadsden
while the highest was 319 m tons/min (351 tons/min) at GLS/Ecorse.  At the
RSC/Chicago Q-BOP, two HMCs were required per heat because of the small shop
ladle size (91 m ton).  All other shops used one HMC per heat.  Furnace tilt
angles ranged from 30° to 45° from vertical.  In most cases, cranemen tried to
keep the charging ladle as close as possible inside the vessel mouth.
                                       206

-------
      In many shops, RMVEs from HMCs represent a major portion of emissions from
 BOF process operations.   As shown in Table 3, four of the shops showed RMVEs in
 the moderate to heavy range (roughly 50 to 70 percent opacity,  observed inside
 the shop)  while two of the shops generated very heavy HMC emissions  (always
 100 percent opacity as seen inside the shop).  These differences in  uncontrolled
 emissions  may be due to  scrap quality arid hot metal charge rates, as discussed
 later.

      HMC controls assessed included two shops with no controls,  three shops
 with retrofit Gaw dampers, and one shop with  an enclosed and  secondary hood.
 Republic Steel's Gadsden BOF and the Chicago  Q-BOP, which employed substantially
 higher  evacuation rates, showed the best capture of HMC  emissions.   USSC/Gary's
 No.  1 BOF, employing a lower design HMC evacuation rate  (GCA  engineers believe
 that actual rates were much higher) showed generally lower HMC  emissions cap-
 ture.   However, capture  did vary at Gary,  generally increasing  as the number
 of  vessels on-line decreased.   HMC evacuation rates at the remaining shops were
 not available,  but were  thought to be lower as reflected by in-shop  observations.
 GCA observers noted that the evacuation system at  Inland appeared to be inopera-
 tive during HMC's.   (Recall that there were no secondary controls at Inland's
 No.  4 BOF  shop.)

      A  comparison of RMVE data in Table 3  clearly  shows  that  the Republic  Steel
 Chicago Q-BOP and Gadsden BOF produced the lowest  RMVE during HMC of all the
 shops evaluated.   The USSC/Gary No.  1 BOF  also showed relatively low HMC emis-
 sions,  while the  Ford BOF and the two uncontrolled shops showed  substantially
 higher  emissions.   Note  that at RSC/Gadsden and USSC/Gary,  exhaust drafts  may
 have been  increased to improve HMC capture (these  steps  may have reduced gas
 cleaning efficiency).  The other shops operated exhaust  flows under  standard
 procedures.   The  RMVE analysis suggest that properly  designed and operated re-
 trofit  Gaw dampers  can approach the degree of  HMC  emissions control provided by a
 secondary  hood  and  enclosure.

      The following  process variables  which might effect  HMC emissions were
 evaluated:   (1) scrap quality;  (2)  ladle handling  and  positioning; (3)  hot
 metal charge rate;  (4) hot metal temperature  and chemistry; and  (5) hot  metal/
 scrap charge ratio.   In  general,  the  minimal  variations  of  these process param-
 eters within a  single shop did not  provide adequate data to show any  substan-
 tial correlations with RMVEs.   Also,  the masking effect  of  other process varia-
 tions further complicated  the  analyses,  A data base  on  the order of 100 heats
 per  operating vessel  would likely be  required to conduct an in-depth  evalua-
 tion employing multiple  regression  analysis techniques.

      General trends noted  at RSC/Gadsden indicated  that  low quality reclaim,
 No.  2 heavy  melt, and medium quality  strip mill scrap produced the highest
 HMC  emissions at that  shop.  High quality home  scrap  (blooming mill scrap and
 '.ngots)  showed no RMVEs  during  HMC.   The above  trends were exactly reversed
 for  data collected at USSC/Gary.  Similar contradictions were noted for shop
 ladle handling practices.  At Gadsden, nine cases of poor ladle handling re-
 sulted  in  only one case  of RMVEs  greater than 0.  However, at USSC/Gary, where
HMC  emissions abatement  procedures required attenuation of  the HMC pour rate
when emissions were heavy, in-shop observers noted decreased emissions when
ladles were  carefully poured.
                                       207

-------
                                     TABLE 3.    SUMMARY OF HOT METAL CHARGE  (HMC) DATA FOR SIX  EOF  SHOPS
ro
o
00



Control
system
Percent
evacuation of HMCs
Secondary Uncontrolled Estimated
BOF
shop
RSC/Chicago
Q-BOP


RSC/Gadsden
BOF

USSC/Gary
No. 1 BOF

Inland Steel
No. 4 BOF

GLS/Ecorse
No. 2 BOF
Ford Steel
BOF
^ight • 5 to 25
Moderate - 30 to
Heavy " greater
Tf assured during
emission emissions
control magnitude
Doghouse Moderate
enclosure
w/local
hood

Gav damper Moderate


Caw damper Moderate
to heavy

None Moderate
to. heavy

None Very heavy
Gav damper" Very heavy

capture
(t)
90




80


40-100


0-10


-30-40
-30-40

percent opacity for emissions escaping
60 percent opacity for
than 60. percent opacity
stack tests.
rate
m'Vsec
(dscfm)
156 b
(330,900)°



230 c
(490, 000) c

44 d
(94,000)d

-Oe

C
NAf
NAf

the vessel,
showing
RMVE >0
(Z)
57




59


53


90


100
100

observed
Average
opacity
of
RMVE >0
(Z
opacity)
8




7


15


34


53
31

inside the
emissions escaping the vessel, observed inside
for emissions

escaping the

Average
duration
of
RMVEs >0
(seconds) (Z
42




54


81


112


132
143

shop.
the shop.


Peak
RMVEs
opacity)
25




30


75


90


100
95



vessel, observed inside the shop.




From data collected at shop.
A • _______ 	
                               Flow rate provided by USSC design drawings, generally thought to be higher as observed by GCA.


                               Control system appeared inoperative.


                              £Not available.


                              *A representative  from Ford indicated that the Gaw Damper system at Ford has been removed in order

                               to experiment with "Dog House" enclosures for collection of secondary emissions.

-------
        Evaluation of HMC pour rates for individual EOF shops showed no substan-
   tial correlations with RMVEs.   However,  a collective evaluation of HMC data
   for all six EOF shops, presented in Table 4,  clearly shows that shops employ-
   ing higher HMC pour rates produced substantially greater RMVEs.  Note that
   the masking effect of process  variations is minimized in Table 4 because the
   data base represents over 200  heats of data.   Deviations from the general
   trend are attributable to differences in secondary emissions controls at each
   shop.

                TABLE 4.   COMPARISON OF HMC POUR RATES AND  RMVEs
   Shop
Secondary
emissions
 control
   HMC
pour rate
 (t/min)b
   (1)
Percent of
   HMCs
  showing
 RMVEs >0
                                                  (2)
                                               Average
                                               opacity
                                               of RMVEs
                                                  >0
   (3)
Duration
of RMVEs
    >0
(seconds)
Opacity-
duration
product3
RSC/Chicago  Secondary
               hood       92

RSC/Gadsden  Gaw
               damper     70
                          57
                          59
                                    42
                                    54
                                    190
                                    220
USSC/Gary

Inland
Ford
GLS
Gaw
damper
None
Gaw damper
None

190
184
216
351

53
90
100
100

15
34
31
53

81
112
143
132

640
3400
4400
7000
Product of columns labeled (1) (2) and (3) showing relative magnitude of RMVEs.
 m tons/min can be obtained by multiplying t/min by 0.908.


  Oxygen Blow

       The basic oxygen process converts a charge of molten iron and scrap to
  steel by blowing large quantities of oxygen through the charge.  The oxida-
  tion process lowers carbon and silicon content and provides heat for melting
  the scrap.  Supplementary fuel sources such as silicon carbide were added to
  the vessel in some of the six shops.   Oxygen blow rates ranged from 15,000 to
  24,000 scfm at the six shops evaluated.   Hot metal to scrap charge ratios
  ranged from 1.8 to 3.6.   Oxygen blows at the EOF shops evaluated ranged from
  13 to 22 minutes in duration.   For the top blown vessels,  fluxes are charged
  through overhead chutes  after ignition of the off gases has been achieved.
  Fluxes are charged to the Q-BOP through the center tuyeres.
                                        209

-------
     Origin of RMVE During Oxygen Blows.  At many EOF shops, oxygen blow emis-
 sions  are a major contributor  to RMVEs.  The heaviest KMVEs  generally  occur
 towards the middle of the oxygen blow  (approximately 12  to 13 minutes  into
 the blow) when a critical period is reached as  the  scrap becomes suspended in
 the bath and melts completely.  During this period, the  scrap overturns and
 the bath erupts violently, occasionally splashing hot metal  from the vessel.
 At this point, carbon and silicon are  oxidized  at the greatest  rate, increasing
 fume generation and the potential of fume escape from primary hoods.

     RMVEs from the oxygen blow at the top blown BOF shops originated  from
 three  main sources; (1) leakage from cracks and expansion gaps  in  the  primary
 hood and exhaust ductwork; (2) leakage from the lance hole in the  primary hood;
 (3) fumes escaping capture at  the vessel mouth.

     The five top blown BOF shops employed open, full combustion primary hoods.
 The vessels at three of the top blown  BOF shops were fitted with lance hole
 covers or steam rings at. the lance hole to contain  oxygen blow  emissions.  On
 vessels with lance hole emission control, emissions from the lance hole area
 were significantly reduced.  The reduction of oxygen blow RMVEs through the
 use of lance hole covers is illustrated in Table 5  which shows  RMVEs from
 Vessel 1 (no lance hole cover) and Vessel 2 (with lance  hole cover) at
 RSC/Gadsden.

     Oxygen blow emissions originating from cracks  and gaps  in  the primary
 hood and ductwork are largely  related  to the level  of maintenance  of the
 primary emission control system at each shop.   This point is illustrated in
 Table  5 by comparing oxygen blow RMVEs from Inland's No. 4 shop to RMVEs
 from other shops evaluated.  In-shbp engineers  noted that lance hole emis-
 sions  were minimal, due to the steam ring controls, but  holes and  gaps, in
 the primary control system ductwork were responsible for most of the RMVEs
 at Inland.  Conversely, the doghouse enclosure  and  ducting at the  Q-BOP showed
 no signs of leaks, and RMVE were 0 for all oxygen blows.  Oxygen blow  RMVEs
 from the other top blown BOF shops, where primary exhaust, ducting was  better
 maintained, were significantly reduced.

     Primary Emissions Control of Oxygen Blow Emissions.  Data  presented in
 Table  5 shows that the Q-BOP with suppressed combustion  and wet scrubber
 emission controls showed the lowest oxygen blow RMVEs (0 opacity)  of all shops
 evaluated.  The Q-BOP also employed the lowest  control system exhaust  i-ates
 (primary and secondary combined) in part due to the use  of suppressed  combus-
 tion system.  The remaining top-blown  shops all employed full combustion hoods
 with either ESP or wet scrubber controls.  Lack of  correlation  between;oxygen
 blow RMVEs and control system exhaust  rates is  likely due to the effect; of
 quencher and scrubber spray additions  which effect  the actual exhaust  rate
 at the hood.  However, Ford's BOF shop, which produced the lowest  oxygen blow
 RMVEs  among the top-blown shops, employed the highest exhaust rate per operating
vessel of all top-blown shops.
                                       210

-------
             TABLE 5.   SUMMARY OF OXYGEN BLOW DATA  FOR SIX  BOF SHOPS
BOF jjhop am)
vessel I.D.
KSC/CHICMIO
Q-BOP
2 Vessel «


RSC/GADSDEN
BOF
No. 1 Vessel


No. 2 Vessel



USSC/GARY
NO. 1 EOF

Msry
Evelyn
Delay
INLAND/ !ND.
IHKBOR
NO. 4 BOT
Vessel 50
Vessel 60






6LS/ECORSE
(2 V«»»«ls)


FORD STEEL
BOF
(2 Vessels)
Control ay»ta» Oxygen
Emissions control evacuation rate blowret*
system mVssc  254 (540,000) 7.1 (15,000)
W/lance hoi* covers 1



Full coBbustlon to
quenchers and wet
scrubbere
\
} 307d (650,000) 9,4 < 20, 000)
1
Full combustion to
wet ucTubbers
steam ring on
Lance hole* an |
poth vessels 228* (484,000) 12.3(26,000)






Full conbuitlon to HA* 9.* (20,000)
f.EP. W/ Lance hole
coven. Both
veeaela)
Full conbuaclon to 2208 (465,000) 11.3 (24,000)
ESP

Percent
of heats
Hot n»tal showing Avenge
to scrap RHVE opacity ot
charj* > OX 1HVE> > 0
ratio (t) (X opacity)
3.6 0 0






94 18

1,8
68 7.6






57 24
2.7 45 20
63 28



100 30
2'9 100 19






HA 100 14.3



2.0 69 8.0


Avairage
duration
of Peak
RHVla > 0 RHVZe (S
 and atmospheric pressure.
Measured at ESP stack (saturated) at 163°C (325°F> and ataniphcrlc pressure.
dlnstall«d (an c.paclty for all three vesesls »t scrubb« outltt (saturatad) -146 ooHg (-78 in. U.C.) at 110°F.
'Stack flow measurements «t wt scrubber outlets {.aturlt«i « 53°C (128°F) and at^apheric pnuura.

 Reliable estimates not available.
^assured at outlet (tack (dacFn) March 1979.
                                                                           i «ad opacity of
      Process Observations and Oxygen Blow RMVEs.  Even though  the comparison
of  primary  exhaust  flow rates for all shops  did not  show strong correlations
with oxygen blow RMVEs, it  is generally accepted that  higher exhaust rates
should produce lower RMVEs  in most situations.   At USSC/Gary,  the No. 1  BOF
control system is shared by three vessels labled Daisy,  Mary and Evelyn.
During oxygen blow  periods  on Daisy,  the isolation damper opened to only
85  percent  of maximum.  The  lower exhaust flow on yessel Daisy corresponded
to  increased oxygen blow RMVEs compared to the other two vessels as shown
in  Table 6.

      Spray  Patching Impacts  on Oxygen Blow RMVEs.  RMVEs were  observed to
increase significantly when  a vessel  was spray patched prior to a heat.   For
example, RMVEs from the oxygen blow on one vessel at Gary that  had just been
extensively patched were two  to three times higher in  opacity and duration as
compared to  nonpatched heats.   At Gadsden,  the  lining  on No. 1  vessel was
nearing the  end of  its campaign and was frequently patched.   More oxygen  was
                                           211

-------
blown to account for the greater heat loss through the worn lining.  However,
lack of lance hole covers on this vessel caused heavy oxygen blow emissions,
masking an increase that may have been attributable to patching.

            TABLE 6.  COMPARISON OF PRIMARY EXHAUST HOOD DRAFT WITH
                      OXYGEN BLOW RMVEs AT USSC GARY No. 1 BOF

                             Average            Opacity -
                           hood drafta      duration product^
               Vessel       (in. HaO)        (% opacity-sec)
Daisy
Mary
Evelyn
6.2
8.2
8.3
5,045
3,450
1,904
            Q
             For all heats with available data; static pressure
             measured in quench section.

             Opacity - duration product is the product of the
             percent of oxygen blows showing emissions >0 times.
             the average opacity and average duration of emissions
             from those oxygen blows.

     Effect of Steel Chemistry.  Oxygen blow RMVE may also be affected by the
final steel chemistry.  For example, high carbon steel heats at Gary showed
RMVE for 80 percent of all oxygen blows, compared to the shop average of
50 percent of all heats.  Conversely, low carbon heats produced the lowest
RMVE.  The oxygen flowrates were about the same for both high and low carbon
heats, but the blow duration was less for the high carbon heats.  One would
almost expect higher emissions for heats where more carbon was oxidized
(i.e., low carbon heats), but the exact opposite was observed at Gary;

     Hot Metal to Scrap Ratios.  Several BOF operating personnel have commented
that high hot metal to scrap ratios generally cause more bath turbulence and
fume escape during a blow.  However, a review of process data for individual
heats within each shop, and comparison of each shop to the others found no
correlations.                                                        ;

     Manual versus Automatic Control of Blow Rate and Lance Height.  Manual
control of lance height and oxygen blow rates was observed to significantly
reduce bath turbulence and fume escape at Gadsden.  All of the other shops
relied on the normal operating procedure of automatic control, regardless
of hood capture efficiencies.  At Gadsden, increased turbulence during the
critical period contributed substantially to oxygen blow RMVEs.  On several
occasions the melter reduced emissions during the critical period by decreasing
the oxygen blow rate and raising the lance to retard the reaction rate in the
vessel.  When these emissions abatement practices were employed, GCA visible
emissions observers verified substantial decreases in RMVEs by radio contact
with in-shop observers.  RSC/Gadsden indicated^ that oxygen blow rates ^t
that shop can be reduced to approximately 12,000 scfm without causing Exces-
sive cooling of the bath.
                                       212

-------
 Vessel Turndowns

      After the oxygen blow, vessels are turned down  for sampling of tempera-
 ture  and bath chemistry.  Additives are sometimes  placed in the slag layer
 of  the bath to retard slag foaming and make sampling easier.   Slag-conditioning
 addititives at various shops included:  (1) blocks of green wood,  (2) "Slag-
 Mag"  cartridges, and (3) elemental sulphur.  After sampling,  vessels return to
 the vertical position for:  (1) reblow if the bath temperature is  low and/or
 the carbon is too high, (2) vessel rocking if the  temperature is too high,
 or  (3) idle mode awaiting return of the chemistry  analysis.   The TD duration
 was usually limited to 2 to 3 minutes at the six shops,  although some IDs
 lasted over 15 minutes.

      At the five top-blown EOF shops, TD emissions control  was limited to use
 of  the primary hood systems (often dampered to 20  to 30  percent open).  Only
 the Q-BOP employed full exhaust rates and use of a secondary collection hobd
 during TD.   For all six shops, capture of TD emissions was  less than 50 per-
 cent,  and in most cases, zero capture was observed.

      Turndown RMVE data in Table 7 indicates greatest TD emissions  at USSC/
 Gary  and Inland's No.  4 shop.   The lowest turndown RMVEs were observed at RSC/
 Gadsden.   At Gadsden,  RMVE occurred for only 1 TD  of  60  observed.   Turndown
 RMVEs from the Q-BOP were moderate as compared to  the other  shops.   Increased
 TD  emissions capture at the Q-BOP was offset by increased emissions created
by  purge gases (nitrogen,  oxygen and natural gas) blown  through the tuyeres.
RSC reports that nitrogen is normally used as a purge gas, but oxygen and
natural gas were blown for 24 of 33 TDs observed.  The data  in Table 7 indi-
cate  that TD: RMVEs were heavier when oxygen and natural  gas were used.

                TABLE 7.  VESSEL TURNDOWN DATA AT SIX  BOF SHOPS


BOF
shop
RSC/Chlcago
Q-BOP

RSC/Gadsden
BOF
USSC/Gary
No. 1 BOF
Inland Steel
No. 4 BOF
GI.S/Ecnrse
No. 2 BOF
Ford/Dearborn
BOF


Emissions
control during
turndown
Primary and
secondary
hoods operating

Primary hood
only
Primary hood
only
Primary hood
only
Primary hood
only
Primary hood
only
Percent of
turndowns
showing
RMVEs > 0
W
78
83

1.7

73

67

21

33


Average
opacity of
RMVEs > 0
(X opacity)
13
17

5

18

18

16

10

Average
duration
of RMVEs
> 0
(seconds)
42
43

30

98

142

73

83



Peak
RMVE
(X opacity)
30
65

10

75

65

100

20


Opacity
duration
product*
(X opacity-sec) Cements
440 Nitrogen blown through tuyeres.
620 Oxygen and natural gac b:owr EhnMi^.
tuyeres.
2.5 Wood blocks added.

1300 Slag-mag added.

1700

250 Elemental sulphur added.

270

"opacity duration product Is the product of the percent of turndowns with RMVEs > 0 and the average opacity and duration
 of those RMVE«.                               '"
                                       213

-------
       A review of process variations which might affect turndown RMVEs  included
  the  following;  (1)  vessel additions; (2) turndown duration;  (3)  slag  condi-
  tioning additives.   No correlations were evident between turndown RMVEs and
  vessel additions or  TD duration.

       A moderate correlation between slag conditioning additives and  turndown
  RMVEs was observed  from data collected at USSC/Gary.  The slag conditioning
  additives used at the Gary No. 1 shop were a mixture of rock  salt, sawdust
  and  sulfur,  called  "Slag-Mag."  In-shop observers counted the number of car-
  tridges thrown into  the vessel for several TDs.  Figure 1 shows the  duration
  of RMVE as a function of "Slag-Mag" cartridge additions, indicating  that  RMVE
  durations increased  with increased "Slag-Mag" additions.  It  was  not clear,
  however, whether the Slag-Mag reduces flames and foaming thereby  creating
  heavier (smoke) emissions, or whether more cartridges are added to TDs with
  heavier emissions.
   4.0
   3.3
 E
   3.0
                                                   O - Itt TO
                                                   A - 2nd TO
               O           O
m
in

-------
      Tapping emissions control at  the  six  shops  include two shops with no con-
 trols, three shops with tap-side enclosures,  and the Q-BOP which employs the
 charge-side hood for tapping control.  The data  in Table 8 show the highest
 RMVE for shops with no tap-side enclosures, and  the Q-BOP.  Fume capture is
 reduced at Ford and Inland because control systems operate at reduced draft
 and the lack of tap-side enclosures or other  containment devices.  Capture of
 tapping emissions at the Q-BOP was somewhat reduced because the single secondary
 hood is located on the charging side.  Tapping fumes escaped through a loose
 fitting door on the tap-side of the doghouse  enclosure in spite of an exhaust
 rate of nearly 123 m3/sec or 260,000 scfm.  Recent designs of doghouse enclo-
 sures include tap-side hoods to improve tap control.

      The data in Table 8 indicate that the  three top-blown EOF shops with tap-
 side encloFures showed the lowest tapping  RMVEs.   RSC/Gadsden produced the
 lowest RMVEs, probably because of the  relatively high exhaust rates (47.7 m3/
 sec or 101,000 scfm).   RMVE at USSC/Gary and  GLS/Ecorse resulted from the
 escape of emissions out the charge side of  the vessel.   Increased exhaust rates
 may increase tapping control at these  latter  two shops.   However, engineers
 at  Gary indicated that introduction of cold air  would adversely affect emis-
 sions capture at other online vessels.

                    TABLE 8.  TAPPING DATA FOR SIX BOF SHOPS

BOF shop
DSC/CHICAGO
Q-BOP
RSC/GADSDEN

USSC/CARY
NO. 1 BOF
INLAND STEEL
NO. 4 BOF
CLS/ECORSE
NO. 1 BOP

FORD/DEARBORN
BOF

Enlsslona control
during tapping
Secondary hood
only
Prlnary hood w/
tapaide enclosure
Primary hood w/
tapslda encloaure
Primary hood only

Primary hood w/
garage door tsp-
alda encloaura
Primary hood only

Estimated
emissions
capture
U>
0-50

65

0-80

0-20

50-70


<50

Tap ladle
of tapa
showing
additions RMVEa > 0
Kg (Ib) (X)
3350 (7*00)

2450 (5400)

620 (1350)

HA*

NA


1600 (3530)

75

10

74

100

58


83

opacity
of
RMVEa > 0
 0
(seconds)
368

116

175

223

174


275

Peak
Opacity*
duration
product

RMVE (Z opaclty-
(X opacity) seconds) Cements
85

20

60

50

60


70

2700

68

1800

3600

1400


2500

Deaulfex ladle additive
N; gas blown through tuyeres
Exhaust rate 60-701 of MX.

Isolation damper 20X open

Minimal or zero exhauat




Isolation daBper 301 open

"opacity duration product la the product of the percent of turndowns with RMVEs > 0 and the average opacity «nd duration of those RMVEa.
bNot available.

     Factors  that may affect tapping emissions were assessed at  three of  the
shops.  Tapping  emissions  at RSC/Gadsden and USSC/Gary showed no strong cor-
relations between RMVEs  and the following process parameters:   (1)  total  tap
duration; (2) tap ladle  additions;  (3) steel temperature; (4) quantity of
steel tapped.

     For the  Q-BOP,  the  analysis showed a strong correlation between tapping
RMVEs and the addition of  a desulfurizing compound to the tap ladle.  Tapping
RMVEs reached 85 percent opacity and  averaged 20 percent opacity, for heats
receiving this compound.   By comparison,  tapping RMVEs averaged less than
5 percent opacity for heats not receiving this  ladle additive.
                                       215

-------
Slagging

     After tapping, furnaces are  rotated back to the charge side to dump slag
Into pots located below  the charging floor.   Slagging generally lasted about
1 minute.  Occasionally,  the vessel  was  rocked back and forth to coat the
vessel lining if linings  were  thin.

     Control of slagging  emissions  is difficult because the vessel mouth is
directed away from the primary hood.   At the top-blown BOFs., almost zero
capture of slagging emissions  was observed.   Emissions capture was more sub-
stantial at the Q-BOP because  the vessel remains Inside the doghouse.  Table 9
summarizes the slagging RMVE data,  showing the lowest emissions (0 opacity)
were observed for the Q-BOP.

     The following were assessed  for possible relationships to slag RMVE:
(1) slag duration; (2) steel tap  temperature; and (3) quantity and type of
flux additions.  No significant correlations between slagging RMVEs and these
process parameters were observed.

                  TABLE  9.  SLAGGING DATA FOR SIX BOF SHOPS




BOF shop
RSC/Chicago
Q-BOP
RSC/Gadsden
BOF
USSC/Gary
No. 1 BOF
Inland Steel
No. 4 BOF
GLS/Ecorse
No. 2 BOF
Ford/Dearborn
BOF
Secondary
emissions
control
during
s lagging
Primary and
secondary hood
None

None

None

None

None

Percent
of slags
showing
RMVE > 0
(%)
0

10

33

100

13

5

Average
opacity of
RMVEs > 0
(^
opacity)
0

5.9

16

14

20

5


Average
duration of
RMVEs > 0
(seconds)
0

55

85

56

113

30


Peak
RMVE
(X
opacity)
0

15

50

45

50

5

Opacity
duration
product8
(% opacity-
sec)
0

34.2

449

7B4

294

7.5

 30pacity duration product is the product of the percent of slags with RMVEs > 0
  and the average opacity and duration of those RMVEs.
                                        216

-------
                   STATUS OF CASTHOUSE CONTROL TECHNOLOGY
                      IN THE UNITED STATES, CANADA, AND
                            WEST GERMANY IN 1980

                                Prepared by:

                               Peter D.  Spawn

                           GCA/Technology Division
                             213 Burlington Road
                        Bedford, Massachusetts  01730

                             Thomas J. Maslany
                    U.S.  Environmental Protection Agency
                          Sixth and Walnut  Streets
                      Philadelphia, Pennsylvania  19106

                                    and

                                Richard Craig
                    U.S.  Environmental Protection Agency
                               26 Federal Plaza
                          New York, New York   10007
                                  ABSTRACT
     This paper provides an up to date review of recent developments in blast
furnace casthouse control technology in the U.S.  Six casthouse control sys-
tems were installed in the U.S. on a permanent basis by the fall of 1980.
Three additional continuous-service casthouse control systems are operating
in Canada.  A number of prototype systems are under development and demonstra-
tion in the U.S.  At this time the U.S.  steel industry has made committments
to install controls on at least 41 casthouses.  The paper discusses current
trends in the U.S. and also describes operating control systems in this
country, Canada, and West Germany.
                                   217

-------
                   STATUS OF CASTHOUSE CONTROL TECHNOLOGY
                      IN THE UNITED STATES, CANADA, AND
                            WEST GERMANY IN 1980
INTRODUCTION

     Casthouse emission control systems are installed on a permanent basis
at nine North American blast furnaces as of the fall of 1980.  Demonstration
tests at four other casthouses are being conducted in order to obtain EPA
approval of proposed systems.  By the fall of 1980, the steel industry has
made committments to EPA to retrofit controls to at least 41 furnaces, and
discussions are presently being concluded on controls for another 17 furnaces.
This paper provides the current status of casthouse controls in the U.S. and
also discusses controls in Canada and West Germany.  Descriptions of Japanese
casthouse controls have appeared elsewhere in the literature.1

     The principal problem with casthouse control technology, as stated by
the industry, is the retrofit of existing furnaces.  At this time, seven con-
ventional furnaces in North America have retrofit control systems—DOFASCO's
Nos. 1, 2 and 3, and four Bethlehem Steel furnaces.  DOFASCO uses total build-
ing evacuation (TE) for three of their furnaces, but is planning to convert
Nos. 2 and 3 to a local hood (LH) system.  The Bethlehem B, C, D and E furnace
have large canopy-type hoods in the roof trusses above the taphole and runners.
All existing systems in North America use baghouses for gas cleaning.

     Industry's main concern with retrofitting TE technology center around
the costs of moving large air volumes, the difficulty in completely sealing
a casthouse to prevent fume escape, and the need to structurally reinforce
the roof.  The LH option relies on relatively manageable air volumes, but the
difficulty of fitting local hoods and routing ductwork within the confines of
an existing casthouse varies from casthouse to casthouse.  A secondary concern
with the LH option is achieving efficient capture of casting emissions from
the numerous potential sources, i.e., notch and troughs, runners, ladles, slag
spouts, and slag pits or pots.

     A number of experimental LH systems have been tried in the past at U.S.
casthouses.  This fall, several prototype systems are being demonstrated at
the Edgar Thomson Works of U.S. Steel (LH), Wheeling-Pittsburgh Steel,
Monessen Works (LH) and at three J&L Steel plants—Aliquippa, Indiana Harbor
Works and Cleveland (noncapture shrouding technique).  These systems are all
retrofit to existing furnaces.  The J&L technology is of interest nationwide,
because it does not use exhaust fans or gas cleaning devices.

     The two newest blast furnaces in the U.S. were designed with emissions
control as one objective.  Bethlehem's L furnace at Sparrows Point and.the
No. 7 at Inland represent the state-of-the-art of both iron-making and .emis-
sions control.  Both furnaces have evacuated, covered runners, local hoods .
over the trough and iron spouts, and a baghouse.  These new furnace controls
are similar in concept to the casthouse controls developed in Japan and also
used in West Germany.  Certain design aspects of these large furnaces -
                                     218

-------
 multiple tapholes, tilting Iron runners, and ductwork running underneath the
 casthouse floor—are Instrumental In achieving gbod control performance.
 These features are not present on most furnaces in the U.S. and could be added
 only during a major rebuild.

 TOTAL CASTHOUSE EVACUATION (TE)

      The first continuous-service TE control system on a basic iron-producing
 furnace in North America was  retrofit to the DOFASCO No, 1 casthouse in 1975.
 A similar system entered service on DOFASCO's Nos. 2 and 3 casthouses in 1978.
 A ferromanganese-producing blast furnace operated by Bethlehem Steel in Johns-
 town, PA was fitted with a 400,000 acfm TE system in the mid-seventies.  How-
 ever, this furnace has been out of service since 1977.

      U.S. Steel was committed to six TE systems in the Monongahela Valley (PA).
 However, the company has recently proposed local hoods instead, for these six
 blast furnaces.  At DOFASCO,  the Nos.  2 and 3 TE systems are being converted
 to the LH option.   However, the conversion from TE is primarily to better
 utilize the existing 400,000  acfm baghouse which is shared by the  two furnaces.
 At Bethlehem's Bethlehem, PA  plant,  partial evacuation  systems were installed
 on four furnaces in the summer of 1980.

      Although the  current trend seems  to favor the LH option,  the  TE concept
 may still be a favored alternative for smaller casthouses,  as evidenced by
 DOFASCO's TE system on No.  1  furnace and the four Bethlehem partial evacuation
 systems.  DOFASCO  reports they are satisfied with their 5 years experience
 with the No.  1 TE  system, and several  U.S.  companies are considering options
 for TE or partial  evacuation  at this time.

 Existing TE Systems:   DOFASCO Nos.  1,  2,  and 3

      The only online  North American  casthouses  employing total  evacuation are
 DOFASCO's Nos.  1,  2,  and 3 furnaces.   Table 1  provides  a quick summary  of cast-
 house size  and evacuation rates.

            TABLE 1.   CHARACTERISTICS OF THE DOFASCO  CASTHOUSES  AND
                      CONTROL  SYSTEMS2
Furnace
number
1
2
3
Rated
capacity,
nthm
per day
2,140
2,015
2,150
Casthouse
floor
dimensions
(ft)
62 x 65
91 x 100
91 x 67
Total
enclosed
volume
(ft3)
150,000
350,000
365,000
Design
evacuation
rate
(acfm)
300,000
400,000
400,000
Air
changes
per
minute
2.00
1.14
1.10
     DOFASCO began experimenting with the No. 1 casthouse in 1975 with a TE
exhaust rate of 250,000 acfm.  A progression of improvements to increase per-
formance were summarized as follows:2
                                     219

-------
     •    Improved sealing between the furnace and the bustle pipe to
          make the curtain wall separating the casthouse from the
          furnace shell fully effective.

     •    Sliding cover at the scrap hole opening over the railroad
          tracks to prevent short circuiting by air entry at this
          point.

     •    Sliding doors over the slag pit openings.

     •    Counterweighted panels along the western (prevailing wind)
          casthouse wall to allow closure of sidewall openings
          during casting.

     •    Considerable flashing to seal the building and prevent
          fume leakage.

     •    Increasing the fan exhaust flow rate to 300,000 acfm (flow
          rate was measured at 340,000 acfm, maximum in 19761).

     Before the above steps were implemented, DOFASCO reported that the bag-
house catch averaged 0.3 Ib/ton of iron cast.  After the casthouse was
"tightened up," DOFASCO measured 0.6 Ib/ton (5-month average).2  Method 5
testing conducted for EPA in 1976 measured 0.6 Ib/ton at the baghouse inlet
on No. 1 furnace.1

     Maximum temperature on the cast floor during casting ranges from 37° to
48°C, a level acceptable to DOFASCO.3  A study of personnel exposure to res-
pirable suspended particulate ( 5 microns), total suspended particulate, sulfur
dioxide, and carbon monoxide found that the casthouse control system on fur-
naces Nos. lt 2, and 3 did not change exposure levels, compared to the
uncontrolled casthouses.2

     The No. 1 baghouse is a single compartment, positive pressure unit with
a design air-to-cloth ratio of 2.5:1 at 250,000 acfm.  It has 2,160 polyester
bags measuring 8 inches in diameter and 22.5 feet in length.  A bypass stack
between the fan and the baghouse protects bags in case of excessive inlet
temperatues, and also provides the exhaust during bag cleaning.

     After the No. 1 TE system was completed, a similar IE system was installed
at the Nos. 2 and 3 blast furnaces in 1978.  DOFASCO anticipated that casting
could be sequenced so only one furnace was casting at any one time.  Thus,
one baghouse and fan serves both furnaces as shown in Figure 1.  The Nos. 2
and 3 control systems are similar to the No. 1 furnace, with minor differences
noted below.

     At the No.  3 casthouse, the curtain wall, roof plenum and roof reinforce-
ment were installed during a furnace reline period, in a fashion similar to
the No. 1 casthouse.  However, construction of the No. 2 system was simplified
over Nos. 1 and 3, and completed while No.  2 furnace was in service.  The No. 2
system has a lighter roof load and lighter weight curtain wall.  This weight
                                     220

-------
                                             FAN
               Figure 1.   Plan view of  No.  2  and  No.  3  casthouse
                          fume control  system.2


 reduction,  and corresponding  cost  reduction, was accomplished by  the  following
 des ign  changes:2

      •    Supporting the  10-foot diameter  exhaust duct  from a structural
          steel bridge spanning the  slag pit, as opposed to laying  the
          duct on  the casthouse roof as was  done on Nos. 1 and 3.   This
          reduced  the extent  of additional roof  reinforcement.

      •    The  roof plenum intake duct  on No. 2 was kept as small  as
          possible and fabricated  from 1/4-inch  plate  to eliminate
          additional structural support steel on the casthouse roof.

      •    The  curtain wall on No.  2  was constructed of  prefabricated
          framing  panels  that were presteeled with corrugated stain-
          less steel prior to installation.  Curtain walls on Nos.  1  and
          3 were framed in placed  and  covered with 1/4-inch plate.

 The above weight reduction steps also  reduced No. 2 furnace shutdown  require-
 ments to 21 8-hour days,  compared  to 46 8-hour shutdowns which would  have been
 required to construct a TE system  identical  to Nos. 1 and 3.

     The single compartment,  positive  pressure baghouse serving Nos.  2 and 3
 casthouses  has an air-to-cloth ratio of 3.3:1 at 400,000 acfm.  The 1,360
 polyester bags each  measure 11-3/4 inches  in diameter and 30 feet 7 inches
 in length.  The exhaust fan is designed for 400,000 acfm at a static  pressure
 of 16 in. W.G., using a 2,500 hp,  900  rpm  direct drive motor.  When not cast-
 ing, the fan idles with dampers partially  closed.

     The Nos.  2 and  3 casthouse control system entered service in November
 1978.  DOFASCO's data for  28  February  1979 to 18 March 1979 shows the average
 baghouse catch was 0.702 Ib per ton of hot metal.2

 Existing Partial Evacuation Systems:   Bethlehem's B,  C, D,  and E Furnaces

     A partial evacuation  (PE) control system was installed in the summer of
 1980 at blast  furnaces B,  C, D, and E at Bethlehem Steel's  Bethlehem PA plant.
 Each PE system consists of a single canopy hood located in  the roof trusses
 above the iron notch, trough,  and runner areas.   The  system design is based
on experiments conducted in 1977 on Bethlehem's E furnace.   The first genera-
 tion collection hood  installed on the E furnace in 1976 proved unworkable.k
                                      221

-------
 A  second  generation  collection hood  installed  in  early  1977 and  tested  in late
 1977  provided  the basis  for  the  current control system.

      The  capture hoods at  the four furnaces consist of  a  21 x 56-foot hood
 in the  roof-truss bay, directly  above the taphole, main trough,  skimmer, and
 upper slag  runner.   Permanent closure of selected openings in casthouse side
 walls controls  cross drafts  which could reduce hood capture.

      A  single baghouse provides  gas  cleaning for  all four furnaces.  When one
 furnace is  casting,  a single fan provides the design evacuation  rate of 390,000
 acfm.   The  other three furnaces  are  dampered off.  When two furnaces are cast-
 ing,  two fans provide a design evacuation rate of 330,000 acfm per furnace.
 A backup fan of 200,000 acfm was also installed.  Normal hot metal production
 is as follows:  B -  2,900 tpd, C - 2,700 tpd, D - 3,200 tpd, and E - 2,150 tpd.

LOCAL HOODS RETROFIT TO CONVENTIONAL FURNACES

      Interest in local hoods (LH) has stemmed from the premise that exhaust
 flowrates (and, thus, capital and operating costs) would be considerably less
 than  for the TE option.  This is especially important for larger casthouses that
inay require above half a million acfm to provide the number of casthouse air
 changes thought to be necessary  for adequate control (i.e., 1.1 to 2.0 air
 changes per minute as at DOFASCO).  The primary industry concerns with retro-
 fit LH systems are ductwork routing within the available space, interference
with  casthouse operations (drilling, mudgun, working of runners, runnesr main-
 tenance), decreased  capture efficiency because of cross drafts, and possible
 lack of space for adequately-sized hoods and/or ductwork.

     At present, no  retrofit LH systems are operating in the U.S. on a continu-
 ous basis.  However, Wheeling-Pittsburgh Steel's Monessen plant has a planned
 start-up date of November 1980 with local hoods on one  furnace.  In addition,
several steel companies are committed to retrofitting LH on a number of casthouses.

Planned LH Systems:  Wheeling-Pittsburgh, Monessen, PA

     Wheeling-Pittsburgh Steel will be demonstrating a LH system at the Monessen
plant,  the proposal calls for local hoods over the iron trough area and four
local hoods above each of the four iron spouts.  The hoods will be evacuated to
a 130,000 4cfm baghouse.   A demonstration-type system is scheduled to-start  up
in November 1980 on the No. 3 blast furnace and will be evaluated by EPA at  that
      C    " '                           '                 ''               "
time.    If;approved, this system will be retrofit to all three Wheeling-
Pittsburgh facilities.

Proposed LH Systems:  U..S. Steel

     U.S.  Steel is proposing to  retrofit LH controls with noncapture•techniques
on a number of their casthouses.   The control plan is to retrofit LH to seven
plants in the Monongahela Valley and to three furnaces at the Geneva,  UT plant.
Discussions with EPA for'LH controls at other U.S. Steel plants are currently
underway.
                                     222

-------
      In the Monongahela Valley complex,  U.S.  Steel has proposed to use LH
 instead of TE for seven casthouses.  The company is currently experimenting
 with a local hood at the Edgar - Thomson No.  1 blast furnace.  The proposed
 system consists of a notch-area hood under the bustle pipe,  extending over
 most of the iron trough.  A blower system is  planned to help direct trough-area
 emissions into the hood.  Noncapture techniques are being proposed for con-
 trolling the rest of the casthouse.  Total exhaust flowrate  for the trough
 area hood is presently 140,000 acfm, with one baghouse per furnace.  EPA is
 presently evaluating the proposed hood at the Edgar Thomson  plant.

 Other Proposed or Planned LH Systems

      National Steel is proposing to use  local hoods above the iron spouts and  a
 trough area enclosure for controlling casthouses at the Weirton,  Granite City
 and Great Lakes plants.  Additional information is not available at this time.6

      American Air Filter recently announced receipt of a turnkey order from
 U.S. Steel to install a LH control system on  the No.  8 blast furnace at
 Fairfield, Alabama.  Local hoods and runner covers will be installed on this
 two-taphole furnace.  Total evacuation rate is reportedly 300,000 acfm.   An
 off-line cleaned baghouse designed for an outlet grain loading of 0.010 gr/ft3
 will be used.7

 DOFASCO's Conversion of Nos.  2 and 3 to  LH from TE

      DOFASCO is studying the  conversion  of the Nos.  2 and 3  TE systems  to  an LH
 system to avoid the occasional operation of an uncontrolled  furnace.  As men-
 tioned previously,  Nos.  2 and 3 furnaces share a baghouse.   The original plan
 was  to stagger casting schedules so only one  furnace  was  cast at  any  one time.
 Casts overlap to some extent, and whichever furnace  is cast  second  operates
 uncontrolled.

      DOFASCO is planning to use LH over  the taphole and torpedo cars with
 a net evacuation rate of 200,000 acfm per furnace.  This would  allow the
 400,000 acfm baghouse,  originally designed to  handle  one  furnace with TE con-
 trol,  to handle both furnaces simultaneously.   Hooding is currently being
 installed on the No.  2  furnace which is  down for  reline  (Fall  1980).  Both
 furnaces will  be converted to two-taphole runner  systems.

 Stelco's  Experience  with  LH

      Stelco  operates four blast  furnaces  at the Hilton Works in Hamilton,
Ontario.  A  prototype system was  recently installed on one furnace, consisting
of an iron trough hood and baghouse.   Stelco is committed to installing con-
trols on the remaining three furnaces, although there is no installation
 chedule at  present.

     Stelco's Lake Erie plant has one new, 5500 tpd blast furnace with local
hoods over the iron  trough and iron spouts, with runner covers.  Apparently,
the casthouse is considered in compliance with applicable regulations, although
formal compliance tests have not yet been conducted.8
                                     223

-------
LOCAL HOODS ON NEW FURNACES AND LARGE RETROFIT FURNACES

     Although this paper primarily discusses casthouse control technology in
the U.S., Canada, and West Germany, brief mention will be made of other
country's experiences.  Virtually 100 percent of all Japanese blast furnaces
have casthouse controls, of Which some are retrofit and some are new fur-
naces .x»9  Four large furnaces in the Rhine-Ruhr region of Germany were re-
trofit with LH and evacuated runner cover controls in the late seventies.
Two other furnaces also have control systems.  The same type of technology
was employed at the two newest blast furnaces in the U.S.—the L furnace at
Sparrows Point, and the No. 7 furnace at Inland.  DOFASCO is installing sim-
ilar systems on their 4,000 tpd No. 4 furnace.  The newer Japanese systems
differ somewhat from the West Germany and new U.S. furnaces in that the new
Japanese systems control drilling and plugging emissions by a scavenger hood
with curtains that lower around the taphole area.

     Casthouse controls are also being developed in France, Italy, England
and Sweden; at least one controlled casthouse operates in each country.
Italsider in Taranto, Italy operates a system consisting of side draft hoods
on the iron notch area, runner covers, and iron spout hoods.  The British
Steel Corporation operates a control system at the Middleboro plant consist-
ing of runner covers and iron spout hoods, complemented with curtains.  In
France, the Usinor-Dunkerque plant is also installing casthouse controls.6

     The new and retrofit LH systems on the large blast furnaces in the U.S.,
West Germany, Canada, and Japan operate in relatively large casthouses which
generally have adequate internal space for hoods and ductwork.  These cast-
houses have several features not generally present on older U.S. furnaces;
i.e., multiple tapholes, tilting iron and slag runners, and casthouse floors
suspended on columns which allow for ductwork routing underneath the floor.
Applicability of these LH systems to older U.S. casthouses may be limited
without extensive rebuild.  Applicability must be judged on an individual
basis.

New Furnace Control in the U.S. - Sparrows Point and Inland

     The two newest blast furnaces in the U.S. recently entered service with
casthouse controls consisting of local hoods and evacuated runner covers.
Bethlehem's L furnace at Sparrows Point and Inland's No. 7 furnace differ
from older furnaces in the U.S. in production capacity, charging and casting
methods, and top pressure.  Where older furnaces are rated at about 1000 to
2000 tons/day of hot metal, the new furnaces produce about 10,000 tons/day.
Most older furnaces have one taphole located in a single casthouse, while the
new approach is four tapholes located within two casthouses.

     The new furnaces are cast almost continuously, rotating hour-long casts
between tapholes.  The fourth runner system is usually off-line for relining
or maintenance.  The new furnaces operate under a top pressure of several
atmospheres.  Iron is cast into multiple ladles through tilting iron runners.
Burden is charged continuously through a bell-less top via a belt conveyor.
                                     224

-------
      The new furnace's emission control systems are similar in concept to the
 West German plants, described below.  Troughs and runners are covered, and
 interconnected to exhaust hoods located at key points:  taphole, skimmer/dam,
 and tilting iron runners.  The exhaust ducts pass from hoods directly down-
 wards through the casthouse floor, and join common ducts leading to the bag-
 houses.  The casthouse floors rest on columns, and torpedo cars travel under-
 neath.   This contrasts to some older furnaces where the casthouse is back-
 filled; i.e., there is no open space underneath the floor.

      The iron runners are fairly deep at the new furnaces.  This allows for
 flat runner covers, with an arched section used on the trough and/or spouts.
 The arched section is removed for drilling and plugging.   The total exhaust
 flowrates are reportedly 300,000 acfm at Sparrow's Point,  and at Inland's
 No.  7 furnace,  320,000 acfm.

 Retrofit LH Systems in West Germany

      All of the six controlled casthouses  operating in the Rhine-Ruhr region
 of West Germany use the LH option.   Only the larger furnaces  (5,000 to 10,000
 tonnes/day) are controlled.   The six controlled furnaces represent  25 percent
 of West Germany's  iron production.

      GCA observed  four of the six controlled casthouses.   The two controlled
 furnaces not observed are operated  by Thyssen;  one produces ferromanganese,
 while the other has a control system not yet completed.  Table 2 provides an
 overview comparison of the four  casthouses of interest.  It is important to
 note that all four furnaces observed in  West Germany are the  pressure-type;
 i.e., the top pressures are 2+ atmospheres.   Highlights of the plant  visits
 appear  below.

     Krupp-Rheinhausen No. 1  Furnace.  This  retrofit casthouse has  side draft
 hoods on the taphole  and  the  torpedo  car.  A hinged taphole hood extension
 lowered  over the iron pool is  raised  for drilling  and plugging.  The slag
 runner,  slag pots,  skimmer/dam area,  and the  iron  runner are uncontrolled.
This furnace has two  tapholes  and two  casthouses.

     According to regulatory officials,  about 80 percent total capture is
normally obtained with  this system.  Regulatory officials feel this is the
least effective design  in the  region because of the following factors:

     •    Low exhaust flowrate.

     •    No control of skimmer and end of iron pool.

     •    Uncovered iron and slag runners.

     •    No slag pot control.

     •    Casthouse sides are open and cross  drafts tend to
          reduce capture efficiency.
                                      225

-------
                                             TABLE 2.    SUMMARY OF WEST GERMAN CASTHOUSE  CONTROL  SYSTEMS
                  Firm
                                Plant
                                        Year
                    Daily       Year   control
                  •production,  furnace system             No. of
                    metric     entered entered   Ho. of   cast-
                     tpd       service service  tapholes  houses
                                                                                                    Capture system
                                                                                       Total
                                                                                      exhaust
                                                                                      applied
                                                                                      to hood
                                                                            leaning   system     capture
                                                                            system    m*/hrp   efficiency"1
                                                                                                                             Gas      to hood   Estimated
                                                                                                                           cleaning   system     capture
               Krupp       Rheinhausen
                     5,000      1976     1977
               Thyssen
               Thyssen
Hamborn No. 4
5,000
Schwelgern No. 1    10.000
 1964*
(1975)
           1973
                                         1978
         1975
ro
N9
cr»
               Kannesaann  No.  A
                     6,000      1973     1978
Side-draft hood,  both sides
of taphole.  Side-draft  hood,
both sides of torpedo car.
No control for slag runner,
iron runner or skimmer.  Mo
runner covers.

Three hoods, over taphole,
dam and skimmer,  and tilt-
Ing iron runner.   No slag
runner or slag pot control.
No runner covers.

Hoods above taphole and
iron pool.  Hood over dam,
skimmer,  tilting  iron
runner.   Iron runner covers.
No control of slag runner.

Runner covers over iron  and
slag runners and  iron pool.
Covers connected  to three
hoods, at dam/skimmer,
tilting iron runner, and
tilting slag pot  runner.
Scavenger offtakes in roof
canopy above cast  floor.
                                                                            Scrubber1-  320,000
                                                                                         80
Baghouse   580,000
Baghouse   700,000
                                                                                                                           ESP
                                                                                      716,000
80
                                                                                                   95
                                                                                                                                                   99+
               "Conflicting data.

               b"Normal."  standard m'/hr.

               °Scrubber reportedly selected over baghouse or ESP due to space restrictions and availability of excess sludge handling capacity.
                Estimates based regulatory officials' assessment of normal operation, supplemented by GCA's observations.

-------
      Thyssen - No.  4 Hamborn Furnace.   This furnace has two relatively small
 casthouses,  and both the regulatory authorities and Thyssen feel the system
 is the best  achievable due to retrofit problems.   Thyssen would prefer to in-
 crease the taphole  flowrate, but space restrictions limit the ductwork diam-
 eter.   Observation  of one cast found the following.  The tilting iron runner
 and  the skimmer/dam hood both showed 95+ percent  capture.  The slag runner,
 slag pot and iron runner were uncontrolled, but showed only light emissions
 of 10  to 20  percent opacity as observed inside the casthouse.   Most of the
 emissions escaped from the taphole area.   Regulatory officials feel that this
 system normally provides capture of about 80 percent of total  casting emissions.

      Thyssen-Schwelgern No. 1 Furnace.  Considered the second-best control
 system in the region, this furnace is  controlled  by local hoods above the tap-
 holes, the dams and skimmers, and the  tilting iron runners.  Refractory-lined
 covers completely enclose the iron runners and connect with the three hoods.
 The  slag runner is  uncontrolled.   The  system layout appears in Figure 2.
                                   CROSS-SECTIOH

   TO
BAGHOUSE
                                      PLAN VIEW
 Figure 2.  Emission control system of Thyssen-Schwelgern No.  1  casthouse.
                                     227

-------
     Regulatory officials reported that capture efficiency is normally 95+
percent, but there are usually some small leaks in the hood system.  Thyssen
would like to increase the total exhaust from 700,000 N m^/hr to about
1.0 x 10G to improve taphole control, and use Mannesmann-type SiC runners to
decrease runner maintenance.  The runner material currently used at Schwelgern
requires relining once per week.  Other than these modifications, both Thyssen
and the regulatory authorities are satisfied with the system.

     Mannesmann A Furnace.  This retrofit system is the best-performing cast-
house control system in West Germany according to regulatory officials.  Obser-
vation of one complete cast, tap to plug, found almost no visible emissions
escaping the runners or taphole, as viewed inside the casthouse.  The only
emissions that escaped the control system were as follows:

     •    One to two minutes of emissions when the iron trough cover
          was removed for a drill, redrill or plug.

     •    Minor puffs that occur infrequently through joints in the
          runner covers.  Puffs rapidly dissipated to 0 opacity
          within the casthouse, or were immediately drawn back into
          the system.

     Figure 3 shows the Mannesmann system of arched, refractory-lined covers
on all portions of the iron pool, iron runner and slag runner.  Casting emis-
sions are contained by the runner covers and drawn off by three hoods, one
each over the tilting iron runner, the skimmer area and the tilting slag
runner.   There is no taphole hood, and no visible emissions escaped the tap-
hole area during normal operation.  Should any emission escape the system,
the exhaust flow can be manually diverted to three or four scavenger ducts
located in an overhead canopy formed by steel sheeting on the roof trusses.
Mannesmann reported they will use an almost identical control system when the
B furnace is built.

     An important feature of the Mannesmann system is the precast SiC iron
and slag runners which last for 200,000 tonnes of hot metal before replacement.
Runner covers are moved weekly for runner maintenance.  Slag buildup on the
SiC runners is reportedly lessened compared to conventional silica sand lining
material.   Excessive downtime during reline is avoided because the precast
runners  are removed in sections, and replaced with new sections.

DOFASCO  No. 4 LH System

     The No.  4 blast furnace has two tapholes within a single casthouse that
has a much larger volume than Nos. 1,  2, or 3.   DOFASCO reported that the TE
option would require a 600,000 acfm system and would likely cause crossflow
of fumes from the casting runner to the furnace crew cleaning the other
runner.2
                                    228

-------
                                         IRON TROUGH COVER
                                         EMERGENCY IRON SPOUT
                                         TILTING IRON RUNNER
                                         TILTING SLAG RUNNER
                                         EMERGENCY SLAG SPOUT
              Figure 3.  Plan view of the controlled casthouses at
                         the Mannesmann "A" furnace.


      DOFASCO selected a system of localized hoods over the two tapholes  and
 the  two  tilting runners as shown in Figures 4, 5 and 6.  These four hoods
 will be  connected to a common duct leading to a fan and a baghouse when  the
 system is  completed in 1981.   The original concept called for an air curtain
 to push  iron trough emissions into the taphole hood.  DOFASCO's current  plan
 (Fall 1980)  is  to use covers  for iron runner control.

      The tilting runner design allows for a shorter iron runner length com-
 pared to a fixed runner design.   DOFASCO reports that  without the tilting
 runner,  total iron runner  length would be 75 feet per  taphole compared to
 24 feet  per  taphole with the  use of the tilting runners.   The shorter runners
 theoretically reduce fume  generation by reducing hot metal exposure to air.
 Runner maintenance is  also reduced.   A captive locomotive was necessary  to
 place the  additional torpedo  car under one side of the tilting runner.    The
 locomotive normally used for  casting positions ladles  on the other side  of the
 •unner.

      DOFASCO  reports several  advantages with the LH approach in a large  cast-
house, compared  to  TE;  i.e.:

     •     Fumes  are collected  at  the source and better working conditions
           are achieved.
                                      229

-------
                                   FUMBMOOO
                                        TAPMOUMQOO
    Figure 4.   No.  4 blast furnace casthouse fume collection.
                  WINCH

               CYLINDER
HIAT SHICLD


     CASTHOUSE FLOOR
                                                 HOT MCTAL CAR
         Figure  5.   Tilting runner-No.  4  blast furnace.
                                             TUVCMC
                                              IRON
                                              NOTCH
Figure 6.  Local  taphole and air curtain  on No.  4 blast furnace.
                                 230

-------
      •    The fume collection system is smaller, with respect to gas
           volume and capital construction costs.

      •    There is minimal interference with furnace operation.

      DOFASCO indicated that efforts to reduce generation of iron pool emissions
 have included the consideration of reducing the taphole angle from 18 degrees
 to 12 or 15 degrees (measured from the horizontal)  and to reduce the drill bit
 size to 1-7/8 inches.  Both of these modifications  would tend to reduce iron
 pool turbulence and theoretically reduce uncontrolled emissions.

 NONCAPTURE TECHNOLOGY DEMONSTRATED BY J&L STEEL

      J&L Steel is presently demonstrating a proprietary technology involving
 suppression of casting emissions by noncapture shrouding techniques.   EPA
 contractors are presently evaluating emissions control performance of this
 technology at three J&L plants—Aliquippa,  Indiana  Harbor Works, and  at
 Cleveland (Fall of 1980).  All three systems were retrofit.   The technology
 has generated interest because it does not  rely on  evacuation fans or gas
 cleaning devices.   No further discussion is provided because the system design
 and operation details are considered proprietary.

 SUMMARY OF AVAILABLE EMISSIONS DATA

      Visible emissions (VE)  data are available to describe the control  per-
 formance of DOFASCO's three  TE systems,  and J&L's prototype  systems.  Limited
 VE data are available for the Bethlehem E furnace during demonstration  tests.
 More VE data will  become available as additional retrofit  control  systems  come
 online in 1981-82.   Mass emissions data  describing  the uncontrolled emission
 rates are available from DOFASCO,  the Bethlehem E furnace  and  four West German
 casthouses.

 Visible Emissions  Data

      Table 3 summarizes  VE data  for DOFASCO, Bethlehem E,  and  the  three J&L
 casthouses.   A more detailed  emissions breakdown will be provided  in the field
 test  report  for each plant.   '   *

      The 31  casts at DOFASCO  were  observed  by  test  teams consisting of three
 environmental engineers,  all  certified to record VE by EPA Reference Method 9
 (40  CFR 60.275 Part A).   One  engineer  stationed inside the casthouse documented
 the origin and magnitude of all  emissions generated during the entire cast.
 A  second  engineer recorded VE escaping the  casthouse in accordance with
 Method  9.  The third  test team member  continuously walked around the casthouse,
 searching  for VE that may not be visible to  the main VE observer.  Constant FM
 radio contact enabled  the test teams to document whether the Method 9 observer
 was viewing  the greatest emissions.

     For all  casts observed at DOFASCO's Nos. 1, 2 and 3 furnaces, the hot
metal sulfur averaged 0.054 percent by weight, ranging from 0.030 to 0.088.
More detailed process data and emissions assessment  appear in the Final Report.11
                                     231

-------
    TABLE 3.  SUMMARY OF VISIBLE EMISSIONS DATA FOR CONTROLLED CASTHOUSES
Percent of total Method 9


Casthouse
DOFASCO No. I3
DOFASCO No. 2a
DOFASCO No. 3a
Bethlehem E


J&L-Indiana Harbor
J&L-Clevelandf
J&L-Aliquippa

Test
date
5/80
5/80
5/80
10/77C
10/77d
10/766
9/80
10/80
11/80
Nn of
li O » OH
casts
observed
10
9
12
2
1
6
30
50
21

0-15%
opacity
93
95
74
90
8
39
94
95
97

20-60%
opacity
7
5
26
10
88
58
5
5
3
observations

>60%
opacity
0
0
0
0
4
3
0.5
0.03
0.1
 *he DOFASCO data provided here excludes the last 5 minutes of each cast due
 to heavy plugging emissions that escaped capture.  VE escaping the casthouses
 during plugging were > 20 percent opacity for about 60 percent of all casts
 observed.

 Method 9 observations recorded at 15-second intervals.

CAt ~300,000 acfm.

dAt ~150,000 acfm.
n
 Experimental notch area hood, roof monitor emissions read by one observer.
 Results of preliminary data assessment.  About 75 percent of emissions
 >60 percent opacity occurred during plugging and/or abnormal operations.

     The J&L VE data were collected in a fashion similar to the DOFASCO data.
The process data are considered confidential and are not available at this
time.

Mass Emissions Data

     Table 4 summarizes available mass emissions data for uncontrolled cast-
house emissions that were captured by each control system.  The DOFASCO data
are based on several months of weighing the baghouse catch and is considered
representative of their operation.  The DOFASCO IE systems have been observed
by GCA to capture on the order of 95+ percent of the total casting emissions.
A testing program on the No. 3 casthouse is underway in November 1980 to obtain
additional uncontrolled mass data and inhalable particulate data.
                                     232

-------
 TABLE  4.  MASS  EMISSIONS  DATA FOR UNCONTROLLED CASTHOUSE EMISSIONS CAPTURED
           BY  CONTROL  DEVICES
Date of Measurement Uncontrolled emission rate,
Plant measurement technique Ib/ton of hot metal
DOFASCO No. 1
DOFASCO No. 1
DOFASCO Nos. 2 & 3
DOFASCO No. 3
Bethlehem E
Four West
.German Plants6
1977
1976
1979
1980
1976
1978-79
fl.
Baghouse catch
Method 5
Baghouse catch
Method 5, IP
Method 5
VDI, mass
balance
0.60
0.60
0.70
c
0.24
2.0
aFive-month measurement; Reference 2.
bNineteen-day measurement; Reference 2.
CMethod 5 and inhalable particulate tests conducted in November 1980; data
 not yet available.
Tlood capture unknown; see text.
eData for large, pressurized furnaces, which likely generate more emissons
 than conventional  furnaces; see text.  VDI method is West Germany's standard
 mass emissions measurement technique.

     The mass emissions data shown for Bethlehem's E furnace are the average
of nine Method 5 tests on the duct serving the partial evacuation canopy
hoods.  The exhaust flowrates were varied from 175,000 acfm to 358,000 acfm
during these tests.  It cannot be determined what percent of total casthouse
emissions were captured and sampled because the Bethlehem partial evacuation
systems are not designed to capture emissions from all casthouse emission
sources.  VE data provided previously in Table 3 indicate some emissions
escape this system  and are emitted to the atmosphere.

     The data from West Germany was obtained by direct sampling of inlets to
control devices for three plants, and by measurement of scrubber solids for
the fourth plant.  The West Germans consider these data unrepresentative of a
smaller, 2000 tpd plant for two reasons:  (1) the data represent large (6,000
to 10,000 tpd) furnaces operating with a top pressure of several atmospheres,
and these furnaces are considered to generate more notch area emissions than
furnaces with about one atmosphere of top pressure; and (2) the high exhaust
flowrate applied to the local hoods and covered runners is felt to increase
"he^ncon^olled emission rate by increasing hot metal/air mixing.>°

     Not included in Table 4 are results of testing for uncontrolled casthouse
emissions by using high volume samplers suspended in roof monitors.  Because of
several uncertainties associated with the accuracy of this test method, it is
quite difficult to determine whether this technique can develop an accurate
uncontrolled emission factor.
                                     233

-------
CONCLUDING COMMENTS

     The diversity of concepts being evaluated by DOFASCO, J&L, U.S. Steel,
National Steel, Bethlehem Steel and Wheeling-Pittsburgh Steel indicate there
are many options for solving blast furnace casthouse problems, and that emis-
sions controls may not be limited to only a few options.  Furthermore, dis-
cussions with workers and operating personnel at DOFASCO, U.S. Steel and J&L
indicate that a problem of worker and operator acceptance of control apparatus
may not exist.

     Future trends in iron-making are likely to change emissions characteristics
of blast furnaces in the U.S.  Large new furnaces, along with changes in
smaller, existing furnaces such as use of external desulfurization, modest
increases in hearth diameter during reline, improvements in burden and fuels,
installing second tapholes and increasing the wind, will increase the potential
casting emissions from individual furnaces.  The decrease in total number of
blast furnaces, and the associated increase in productivity per furnace,
coupled with the industry's research of new control techniques, has resulted
in a crucial economic turn-around in the cost of controls per ton of hot metal.

REFERENCES

 1.  May, William P.  Blast Furnace Casthouse Emission Control Technology
     Assessment.  Betz Environmental Engineers, Inc.  EPA Publication No.
     600/2-77-231.  November 1977.

 2.  Samson, D. H.  Dominion Foundries and Steel, Limited.  Hamilton, Ontario.
     Blast Furnace Casthouse Emission Control.  Paper from DOFASCO, Hamilton,
     Ontario.  July 1979.

 3.  Private communcation.  Peter Spawn, GCA Technology Division to DOFASCO.
     October 1980.

 4.  Application for Plan Approval to construct a blast furnace casthouse
     particulate emissions control system for the Bethlehem Plants Blast
     Furnaces B, C, D and E.  Bethlehem Steel to Pennsylvania DER.  June 16,
     1978.

 5.  Application for Installation of an Air Cleaning Device/Operating Permit
     for the Blast Furnace No. 3 Casthouse.  Wheeling-Pittsburgh Steel Corp.
     to Pennsylvania DER.  September 28, 1979.

 6.  Private communication.  Tom Maslany, U.S. Environmental Protection Agency,
     Region 3, to Peter Spawn, GCA/Technology Division.

 7.  Announcement by American Air Filter, as published in the Mcllvaine Fabric
     Filter Newsletter, October 10, 1980.

 8.  Private communication.  Richard Craig, U.S. Environmental Protection
     Agency, Region 2, to Stelco.

 9.  Maslany, Thomas J.  Japan Trip Report.  Spring 1979.
                                     234

-------
10.  Spawn, Peter D.  West Germany Trip Report.  GCA for U.S. EPA.
     June 1980.

11.  Spawn, Peter and Stephen Piper.  DOFASCO, Hamilton, Ontario.  Blast
     Furnace Casthouses—Visible Emission Assessment.  GCA Draft Final
     Report No. TR-80-55-G.  Volume 1.  July 1980.  Final Report due
     December 1980.

12.  Field evaluation of demonstration tests at J&L Indiana Harbor Works and
     Cleveland.  GCA/Technology Division.  September-October 1980.  Report due
     December 1980.

13.  Field evaluation of demonstration tests at J&L Aliquippa.  JACA Corp.
     .November 1980.  Report due December 1980.
                                     235

-------
              EFFICIENT AND ECONOMICAL DUST CONTROL SYSTEM
                       FOR ELECTRIC ARC FURNACE
              by:  Leon Hutten-Czapski, M.Sc. P. Eng.
                   Technical Assistant to the Director of Engineering
                   Sidbec-Dosco, Usine de Contrecoeur,
                   C.P. 1000, Contrecoeur, Que\
                   Canada JOL ICO
                   Tel:  (514) 587-2091  ext. 363
                              ABSTRACT


          The total efficiency of a dust control  system is a product
of collecting and filtering efficiencies.  Most of existing dust
control systems for electrical arc furnaces have inefficient collect-
ing and very efficient filtering components.  The poor efficiency of
collecting hoods is a result of their distance from the source,  exist-
ing cross wind in the melt shop and lack of sufficient exhaust capacity
even at very high gas flow rate.

          The only rational and economic solution is to capture  the  fumes
at or close to the source.  The system installed in Sidbec-Dosco Contre-
coeur melt shop is employing retractable collecting hoods  located close
to the ladle.  It achieves high collecting efficiency at low exhaust
rate of gas flow without interference with operation of the overhead
crane.

          The paper in question provides technical  and economical  ana-
lysis of several  alternative solutions of dust collecting  system for
electric arc furnaces in steel making plants.
                                 237

-------
               EFFICIENT AND ECONOMICAL  DUST CONTROL  SYSTEM
                        FOR ELECTRIC ARC FURNACE
          During its operating cycle the modern high  powered  electric
arc furnace releases about 1% of the weight  of the molten  steel as  fume
and dust.

          This emission contains a variety of metal oxides including
heavy metals.

          For health reasons of the workers  employed  in  the melt  shop and
the general  population living around the plant, the emissions must  be
limited to a degree balanced between the cost of control and  social bene-
fits resulting from this control.

          At the present time legislature and regulations  in  North  America
and Europe require that approximately 94 to  96% of all emissions  from the
electric arc furnace operation are to be captured.  Only 4 to 6%  can be
allowed to exit to the atmosphere.

          There are also health regulations  in force, which limit the
dust concentration inside the plant.

          The task of the project engineer is to provide control  systems
which will satisfy the regulations at minimum investment and operational
costs.

          In order to find the best solution, it is necessary to  quantify
the emissions from the furnace for each  phase of the  operating cycle.  The
average of data obtained are shown in table  # 1.
ACTION:
INSPECTION AND FETTLING
CHARGING
MELTING AND REFINING
TAPPING AND DEOXIDIZING
SLAGGING
TOTAL EMISSION
WEIGHT OF NON
CONTROLLED
EMISSION
KG /TONNE
0.06
0.24
9. 25
0.40
0.05
TIME OF DURATION
OF EMISSION
IN MINUTES
3 TO 20
1.5 TO 3
120 TO 150
5 TO 11
3 TO 5
10.0 KG/TONNE OF MOLTEN STEEL
TABLE 1 WEIGHT AND DURATION OF DUST EMISSION
                                  238

-------
          A  survey of available data shows a large spread of specific
weight of uncontrolled emissions ranging from 6 to 24 kg per tonne of
molten steel.

          The  factors influencing the high rate of particulate emission
are  shown in table # 2,
              OILY SCRAP

              LIGHT GAUGE SCRAP

              PAINTED SCRAP

              ZINC COATED SCRAP

              PNEUMATIC FEED OF LIME AND DOLOMITE

              MORE THAN ONE BUCKET CHARGE

              DOORS OF THE FURNACE OPEN

              LEAKY LINTEL AROUND THE ROOF OF THE FURNACE

              WORN OUT REFRACTORY REQUIRING REPAIR

              ACCIDENTS LIKE:  BROKEN ELECTRODES

              INTERRUPTION  OF POWER SUPPLY
       TABLE  2 - FACTORS INCREASING THE PARTICULATE EMISSIONS

                OF FUME AND DUST FROM THE ELECTRICAL ARC FURNACE
          Beside the  particulate  emission there is also an emission of
gasses:  carbon dioxide  C02,  carbon monoxide CO, nitrogen oxides NO ,
hydrogen H2, water vapour  HpO  and  a  vast variety of hydrocarbons aftd
other inorganic and organic compounds.

          Of particular  interest  for  the project engineer are emissions
of combustible gasses since they  may  cause explotions.   In order to reduce
teh chances of such an accident,  additional  air, necessary for combustion,
is introduced into the ducts, and appropriate dilution is added to remove
the mixture of gasses and air from  the  explosive range.

          Another cause  of explosions is:   closed vessels containing gasses
and liquids, for instance:  hydraulic cylinders, pumps  and valves.  They
burst during heating, suddenly  releasing their content, which expands, or
ignites instantly.  In winter months, there is a danger of explosion by
ice frozen into crevices of the scrap.   Contact of water and molten metal
dissociates water into hydrogen and oxygen which recombine again with
explosive force.
                                  239

-------
          The rate of participate and gasous emissions varies in a wide
range during the cycle of operation of the electric arc furnace.

          The total efficiency of a dust control  system is a product of
collecting and filtering efficiencies.  Most of existing dust control
systems for electrical arc furnaces have inefficient collecting and
very efficient filtering components.  The poor efficiency of collecting
hoods is a result of their distance from the source, existing cross wind
in the melt shop and lack of sufficient exhaust capacity even at very high
gas flow rate.
                       flfiURE 1 : COIVWTIOHA1.0UJT COITHOLJTJTU
          The new cycle of operation begins with inspection of the refractory.
The electrodes and the roof are raised and moved off the furnace.  Since
the furnace  is nearly empty, there is very little particulate emission.  If
the refractory needs repair, a procedure called fettling is performed.  A
plastic  refractory is blown, or thrown by centrifugal pump toward the places
needing  repair.  The moisture of the plastic refractory evaporates quickly
and takes some of the refractory into a column of raising vapour.  This
first  emission is partially captured by an exhaust hood located at the roof
truss  level  in the charging bay.  The second part of the emission falls
down mostly  along the walls of the building.  The third part exits  by
the roof ventilation into the ambient atmosphere.

          When the refractory of the furnace is in good condition, there is
no need  for  inspection and fettling.  The cycle begins with charging:  the
roof of  the  furnace  is opened, the charging bucket filled with scrap carried
by the overhead  crane arrives, centers with the axis of the electric arc
                                  240

-------
furnace and releases its load into the furnace.   In order to  save wear
of the refractory at the bottom of the furnace the charging bucket  is
loaded first with light scrap, and next, with heavy scrap.  Light scrap
has a large surface to volume ratio, and is more contaminated than  heavy
scrap.  The sudden contact of the light scrap with hot refractory heats
the thin scrap layer rapidly, evaporating water  first, and  hydrocarbons
(oil, plastics, rubber) second.  The vapours of  hydrocarbons  ignite within
few seconds.

          The intensity of the flame depends on  the hydrocarbon content in
the scrap.  Shavings contain up to 5% of oil by  weight.  Large amounts of
oily scrap in the charge may result in flames up to 20 meter  high reaching
the roof of the building.  In these conditions,  gas emissions rate  may be
more than 15 m /sec. for each square meter of the furnace hearth area.
With limited content of oily scrap in the charging bucket the gas emission
rate is about 5 m /sec. for each square meter of furnace hearth horizontal
projection.

          The total  charging operation can be performed by  skilled  operators
in less than 90 seconds.  The large volume of gas and  fume  released is
normally beyond the capturing capacity of the exhaust  system  located over
the furnace.

          The result is a visible emission escaping by the  melt shop roof
ventilators into the atmosphere.

          After charging the melting opeartion begins.  Three electrodes
are lowered into the furnace and electric power  is applied  gradually until
full power is reached.  The cold scrap is therefore heated  simultaneously
from outside by the walls, bottom and roof of the furnace and from  inside,
by radiant heat developped by the electric arc.   The process  of:  vapori-
zation melting and sublimation reaches throughout the  total scrap mass.
Since the temperature in different parts of the  charge is not uniform, a
mixture of vapours and gasses escapes from the furnace.

          In the present state of technology, the common method used to
capture emissions issued during the melting process is direct evacuation.
This is done by applying sufficient suction at the 4th hole (3  holes for
electrodes) in the roof of the furnace.  The practice  shows,  that the
exhaust rate of 0.14 kg/sec, of gas for each Megawatt  of the  main trans-
former rating is sufficient for normal  conditions.  Since the temperature
of the gaseous emissions rises up to 1875 C during the melting  period the
gasses are cooled first in a water cooled duct reaching  the roof of the
melt shop and diluted to provide 100% of excess  air for  combustion  of CO
and H« and later diluted for protection of the ducts with air drawn from
the hoods located over the furnace.

          In some installations, an additional cooling of gasses is performed
in a radiant cooled consisting of a series of tubes exposed to  the  ambient
air.
                                 241

-------
                                                                  GAS
                          WATER
                          COOLED
                          DUCT J
                              EMERGENCY
                               DILUTION
                                 AIR
                                                          LARGE
                                                        PARTICLES
                                                                 SMALL
                                                                PARTICLES
       FIGURE 2
ELECTRIC ARC FURNACE ECONOMY DUST CONTROL SYSTEM SCHEMATIC
            (WHEN  CHARGING)
          The gas dilution factor is about 8.  It means that about 1.12
kg/sec, of gas and air per each Megawatt of the furnaces main transformer
rating reaches the dust collector, which is normally a bag house.  The
temperature in the baghouse should not exceed 135 C for Dacron bags.
New materials for bags allow for higher operating temperature, which in
turn reduces the quantity of dilution air, power drawn by fans, dimensions
and cost of installation.

          The gas pressure in the furnace should be kept slightly negative
for the melting of carbon steel and slightly positive for the melting of
alloy steels.

          This control of pressure is necessary to avoid excess infiltration
of ambient air into the furnace since the excess air increases burn out
of the electrodes, increases emissions and increases electric power consunipt-
tion.  The positive pressure in the furnace results in emission around
electrode holes in the roof.  The fugitive emissions can be captured by
the hood over the furnace but at the cost of drawing a large quantity of
infiltrated ambient air into the column of rising gas.

          The gas velocity in the ducts carrying dust is set at 15 m/sec.
for each meter of duct diameter, and shall not be less than 15 m/sec. for
small dia. ducts.  The common operating problems with direct evacuation
systems are:  burning of holes in the bags by incandescent material, build-
up of cake on the bag surface, erosion of fan blades and occasional explos-
ions.
                                 242

-------
           Due to the short gas  transfer  time  from the furnace to the bag-
 house (about 5 to 9 seconds)  some  incandescent particles may reach the
 bags resulting in a burning of  holes.  In order to avoid the transfer
 of large particles, a cyclone or a  plenum chamber with reduced gas velocity
 should be provided to eliminate particles larger than 0.5 mm of diameter.

           Build-up of cake on the bags is a result of excessive moisture
 content in the gas at the  start of  the operation and of moisture entering
 the baghouse after the end of the operation.  The captured dust contains
 large amounts of burnt lime,  which  remains as a fine, submicron size, powder
 when it is dry.   In the presence of moisture  this burnt lime becomes a
 mortar and clogs the fine  passages  between the  yarns of the woven fabric
 of the bags.  Shaking or back flow  of gas used as a means to clean up the
 bags may be not fully effective to  dislodge this mortar.

           In time the layer of  mortar builds and the resistance to the
 flow of gasses increases.   The  stresses due to increased pressure may
 rupture the bags.   Keeping the  bags dry between operations by applying
 external  heating seems  to  be  too expensive, but operation of the direct
 evacuation system for a  half  an hour after stopping of operation will
 dislodge most of the lime  from  the  bags, thus at least partially preventing
 the cake build-up.

           In the Sidbec-Dosco steel making practice, when the first charge
 is partially melted,  an  addition of continuously fed pre-reduced iron
 pellets begins.   The rate  of  feed is adjusted to the quantity of the
 electrical  power supplied  to  the electrodes in order to  maintain constant
 temperature of the  molten  bath.  Burnt lime and/or dolomite and alloys  are
 fed with  pre-reduced  pellets, into  the furnace from the  bins located on
 the roof of the  melt shop.  All continuously fed materials  enter through
 a  stainless  tube through the  5th hole in the roof of the furnace.   The
 downward  motion  of  the falling material  creates movement of the air in
 the feeding  duct directed  toward the furnace.   An opposite  movement of
 the hot gas  from the  furnace counteracts this  movement of air.   As  a
 result  of these  two actions, an escape of gas  from the furnace through
 the feeding,duct is observed.

           The  emission through the feeding pipe can be controlled by reducing
 its diameter,  and by  applying suction at the funnel  where the stationary
 and movable  parts of  the feeding pipe join.

           In  electric arc melt shops which do  not  employ continuous feed  of
 pre-reduced  pellets,  an additional  one,  or more,  bucket  charges  are used.
 The  second and following charges of scrap fall  into  already molten  metal,
 therefore, the instant burst of gasses,is more violent than at  the  first
 charge.   In order to reduce those violent emissions,  the second  and
 consecutive charges should consist  of clean  scrap.

          Tapping into the ladle begins when the charge  is  melted,  the
 furnace is deslagged, and the batch  brought  to the  desired  temperature  and
chemical content.
                                 243

-------
          The tapping ladle may be suspended on the hoods of the overhead
crane, or may be stationary, positioned on a stand located in front of
the spout of the electric arc furnace.
                                                           GAS
             FIBURE  1
                                                    LARGE
                                                   PARTICLES
                     ELECTRIC ARC FURNACE ECONOMY OUST CONTML SYSTEM SCHEMATIC
                                 (•HEN MUTIN8)
          The suspended ladle enables the falling height of the molten
steel to be minimized, therefore, reduces generation of oxide fumes.
However, this practice uses valuable time of the overhead crane, and
makes confinement of escaping emissions more difficult.  The stationary
ladle, being fixed in place, provides the possibility to collect the
fumes through the hood which is close to the ladle.  This is far more
efficient than by the hood located at the roof truss level of the melt
shop, normally used for the suspended ladle fume control.  Direct measure-
ments of the rate of emissions during the tapping operation in Sidbec-Dosco,
Contrecoeur Melt Shop revealed that the gas emission for each square
meter of the ladle's horizontal projection area is 3.5 m^/sec. at the
ladle rim level, and 40 m3/sec. at the roof truss level.  This is due
to entrapment of the ambient air by the column of highly turbulent
gasses rising from the ladle.

          The hot gas column starting from the ladle is deflected by
horizontal air currents in the casting bay of the melt shop.  Deflection
up to 45° were observed in some cases.  These were not constant, being
influenced by changes of wind velocity and direction, the opening of
doors and by activating the on/off switches of the preheating furnaces
for the ladles.
                                  244

-------
          In these conditions,  if the evacuation  hood  system located on
the roof truss level  was selected its capacity  would be  about 450 m3/sec.
(950,000 ACFM) to obtain an average of 85% dust capturing  efficiency
(for tapping).

          The estimated cost in 1977 for this type  of  dust control system
of this capacity was  6.5 to 8 million U.S. dollars.

          In order to save investment and operating costs  and to achieve
better fume control during the  tapping operation, an idea  was deyelopped
to use the existing capacity of the direct evacuation  system, which remains
idle during tapping time.

          A swinging  hood located just over the stationary ladle was
designed.  This hood  has a cut-out permitting passage  of the stream of
molten metal pouring  from the spout of the electric furnace.

          The ladle hood would  have been suspended  on  a  structure which
would pivot and rotate to permit access of the  hoods of  the overhead crane
to the trunnions of the ladle when the ladle was  filled  with molten steel.

          There was an opinion  that a swinging  hood over the ladle could
be easily damaged, and would take too much valuable space  in the casting
bay of the melt shop.

          A new design was prepared:  a horizontal  retractable hood with
a blanket of horizontal air currents created by fans located opposite to
the intake of the hood.  All mechanisms were located under the tower support-
ing the additive bins.  Being hidden under the  structure of the tower the
mechanisms were safe  from accidental striking by  the hooks of the overhead
crane and no additional space was required in the casting  bay.

          A prototype was made  and installed for one of  the furnaces.  The
tests showed that fans pushing  the air were not required and the ladle
fume evacuation system could collect about 85%  of the  fumes issued during
tapping operation.

          The success of the prototype was followed  by  installation of
similar units for all four furnaces in the Contrecoeur Melt Shop.

          The total cost of four ladle fume control systems was $500,000
which is much less than 6.5 to  8 million dollar estimate for an evacuation
system with similar efficiency  comprising hoods located  at the roof truss
level.

          Retractable hoods work in automatic or manual  mode.  If automatic
mode is selected then the signal  from the tilting furnace  actuates the
extension of the hood toward the ladle and simultaneously  closes the
damper of the direct  evacuation system and opens the damper of the ladle
                                 245

-------
fume evacuation system.  When, after  filling  the  ladle  the  furnace returns
to its vertical position another signal causes  retracting of  the  hood
and reversal of the position of the dampers.

         . Figure # 4 shows a schematic of  the combined  system.
                                                         GAS
                                                   LA HUE
                                                  PARTICLES
                                                       PARTIClEl
           FIGURE 4 :  ELECTRIC ARC FURNACE OUST CONTROL JWTEB SCHEMATIC
                                (I HEN TAPPING]
          At the present time  in our Contrecoeur Melt Shop  two major
sources of air pollution from  the operation of the electric arc  furnaces
are under control:  emissions  when melting and refining, and when  tapping.
A third, however much smaller  emission, when charging,  requires  about
200 m3/sec. exhaust capacity in a hood located above the furnace at
the roof truss level.

          Experimental work is being committed to develop an improved
method of controlling emission during operations of charging, inspection
and fettling.

          Development of an efficient and economical fume and dust control
system for electric arc furnaces have been done in an existing operating
Melt Shop.  Valuable information has been gathered concerning:   engineering,
operational, maintenance and human relation factors.

          Some of ^these observations are shared here.

          For an existing melt shop with multiple electric arc furnaces the
most economical solution of fume and dust control is the full utilization
of capacity and time of an existing direct evacuation system, in order to
control emissions during melting, tapping and charging  periods.
                                  246

-------
           There should be a centralized manual or automatic dispatch
 directing  necessary flow rate for different emission sources in order
 to obtain  the best total dust collecting efficiency at a minimum cost.

           At the  present time there are no integrating instruments avail-
 able on the market which can provide a signal proportional  to dust mass
 transfer rates in different ducts of a complex integrated dust collecting
 system.  Such signals could be processed in a real time computer to open
 the dampers of the ducts carrying the dust, and close dampers on ducts
 which momentarily carry little or no dust at all.

           Manual  operation of dust control systems calls for continuous
 attention  sound judgement and quick response and requires additional
 personnel.

           On the  basis of observations made of the operation of many steel
 melt shops, one can say that the dust control equipment can be easily
 damaged by the build-up of slag, overheating, clogging by dust and mechanic-
 al abuse.  Therefore, it has to be continuously supervised by trained
 personnel  in order to ensure its availability all the time.

           Since production always has maintenance priority even a well
 designed dust control system left without sufficient supervision and proper
 care for long period of time, becomes quite often inoperative resulting
 in excessive emissions into the atmosphere and excessive dust concentration
 in the working area.

          When a  "green-field" installation of an electric  arc melt shop
 is in the  preliminary planning stage, a total  enclosure over the furnace
 and ladle  could be an alternative solution.  However, in an existing melt
 shop, there is rarely a chance of implementing this idea, due to space
 limitations.  In addition, the possibility of frequent damage to the
 doors of the furnace envelope 1s a factor which restrains the acceptability
 of total enclosure of an electric arc furnace.

           In summary, one can say the present trend in the  design of dust
 control systems for electric arc furnaces is to utilize a minimum of  exhaust
 capacity,  but to use it fully in all  stages of the operating cycle.

          Since the dust collecting hoods located close to  the dust  source
 are nearly ten times more efficient than the hoods located  on the roof
 truss level, their high collecting efficiency reduces the necessary  exhaust
 gas flow rate and associated cost of investment and operation of a  dust
control system.   For this  reason, one can expect that in the future  design
of dust control  systems, the systems with hoods close to dust source will
be employed more frequently.
                                 247

-------
ACKNOWLEDGMENT
          The statistical  data re emissions  were compiled  from  reports
of Environment Canada APCD subcommittee electric arc  furnace headed  by
Mr. V.C. Havelock in 1977.

          The measurements of gas flow rate  when tapping and charging
were made by Mr. M. Bender of Hatch & Associates, Toronto.

          The design and development of retractable hood for tapping
was done by Sidbec-Dosco engineering group headed by  Mr. R.  Sevigny.

          The practical  implementation of the idea of close  to  dust
source exhaust system was  encouraged by Messrs:   T.E.  Dancy, senior
vice-president of technical  studies and development and G.H. Laferriere,
director of engineering.
                                 248

-------
     IRON AND STEEL  INHALABLE PARTICULATE
        MATTER SAMPLING PROGRAM:  AN
          OVERVIEW  PROGRESS REPORT
                     By
             Robert  C. McCrillis
       Metallurgical Processes Branch
Industrial Environmental Research Laboratory
    U.S. Environmental Protection Agency
           Research Triangle Park
           North Carolina   27711
                    249

-------
      Iron and Steel Inhalable Particulate Matter Sampling Program
                       An Overview Progress  Report
                                Abstract

EPA's Office of Research and Development has  entered  into a major program
to develop inhalable particulate matter (IPM) emission  factors, where IPM
is defined as airborne particles of <\5 ym aerodynamic  equivalent diameter.
The Metallurgical Processes  Branch  of  EPA's  IERL-RTP is responsible for
the iron and steel industry segment of  this program.  This paper presents
a summary of efforts to date.  Implementation has proceeded along two major
lines of action.  The first follows the classical route:  literature review,
prioritization of sources,  identification of sources for  which existing
data  are  adequate,  selection of plants, testing,  and  finally reporting
results.  The other aspect consists of meshing the IPM requirements with
those of other EPA sampling programs, thus reducing overall  cost to EPA and
minimizing inconvenience to the host plants.

A review of  existing particle  size  data showed not only that relatively
little data exists, but also that most existing data are of questionable
quality.  Therefore, the field test program entails  sampling virtually all
significant sources.  To date, tests have been completed of basic oxygen
furnace  (EOF)  charging  and tapping,  hot  metal desulfurization,  blast
furnace cast  house  (building evacuation approach), uncontrolled paved and
unpaved  roads,  and  EOF  main  stack   (limited  combustion system  after
scrubber). Discussions are now underway with  several  plants to test other
high priority sources.'   It is anticipated that  funds will allow duplicate
tests at another plant of at least the highest priority sources.
                                   250

-------
INTRODUCTION

The U.S.  Environmental Protection  Agency is required,  under  the amended
Clean Air Act of 1977, to review the scientific basis for the total suspended
particulate ambient  air quality  standard.   Major  consideration  is being
given to a size-specific particulate standard focusing on inhalable parti-
culate matter  (IPM),  defined as airborne particles  of <15 ym aerodynamic
equivalent diameter.^  EPA has initiated an extensive program to compile
and review existing  data  and,  where  necessary,  conduct field  sampling
programs from which reliable emission factors can  then  be determined. These
emission factors will be available to the States  for  the purpose of revising
State Implementation Plans if an ambient  standard  based on particle size is
promulgated.

EPA's Office of Research and Development is responsible  for developing these
IPM emission factors.   A major part  of  this  effort  is  directed toward the
steel  industry.   This  paper discusses  briefly the rationale  behind the
selection of 15 ym as  the upper cut point in the IPM definition and why a .field
sampling program must be undertaken.  Sampling protocols are described for
ducted condensible and noncondensible emissions  and for fugitive emissions.
The rationale  and approach being  followed  to  select  test sites  are also
discussed.  Results obtained to date are summarized and  conceptual plans for
the remainder of the program are presented.

DISCUSSION

The human  respiratory  tract consists  of three main areas:(1>   the upper
respiratory tract, conducting airways, and gas-exchange  area.  All available
data  demonstrate that  direct  health  effects  from  particles  >_15 y m are
primarily restricted to  the upper respiratory  tract.  IPM is the  term used to
represent  airborne  particles  capable  of affecting  the lower  respiratory
tract which includes the  last two of  the three general areas.   Thus the
definition of IPM has been taken to mean those  particles  of <15ym aerodynamic
equivalent diameter.   Although discussion continues among health effects
experts as to  the  specific  upper cut point for an ambient air particulate
standard based on size, the definition of IPM remains unchanged.  A secondary
cut point  of  <2.5 ym has  been  recommended  to estimate  that  fraction of
particles capable of penetrating  to the gas exchange  area of the respiratory
system.   Sampling data generated under  this program will include several
intermediate cut points  to enable  accurate interpolation once  an ambient air
standard cut point is selected.

Sampling Protocols

Sampling protocols have been developed  to permit  IPM measurement  of duct-
ed^,3,4) an(j fugitive^) emissons.

Ducted Sources  (Non-Condensible  Emissions) - Initially, EPA  felt  that the
main interest  lay in  determining total IPM of  j<15 ym and  also the quantity of
IPM <2.5 ym.   Since existing  cascade  impactors are limited  both  by small
                                    251

-------
sample  collection capability (hence  very short sampling  time  in heavily
loaded streams) and in collecting particles >10 ym without particle "bounce,"
EPA decided to develop a sampling system  specifically for the IPM program.

The result of this effort, depicted in Figure 1, consists of two in-series
cyclones inserted directly into the stack using a modified EPA Method 5
probe and associated equipment^). Operating at a nominal flowrate of 23
1/min, the large cyclone has a 050 of 15  ym and the small one, 2.5 ym. The
filter (not shown) collects all particles <2.5 ym which, when added to the
small  cyclone  catch, equals  total  IPM.   A significant  advantage  is the
capability to collect a fairly large sample, which means a longer sampling
time.

During  this development  program it  was found that  the  button hook nozzle
commonly  used  with  cascade  impactors  was,  by  itself,  a very  effective
collector of particles <15 ym. Thus an important criterion  for any particle
sizing device is that it have  a straight nozzle. A comparison was run between
the so called IPM train described above, an Andersen 2000 Mark III in-stack
cascade impacter equipped with a 15 ym cyclone precutter,  and an Andersen in-
stack high capacity stack sampler (HCSS).  As reported earlier, in the paper
by J. Steiner  and  D.  Bodnaruk^",  test results showed that  all three devices
gave comparable results.

Subsequent to the development of the  IPM  train, controversy occurred con-
cerning the selection of the upper cut point for the purpose of defining a
particle size  ambient particulate standard.   Consequently,  the decision was
made to use the in-stack cascade impactor with a 15 ym cyclone precutter (see
Figure 2).  The data thus generated would  permit  the more accurate nlotting
of curves relating particle size versus cumulative sample weight emitted per
unit process weight.   A straight nozzle  would,  of  course,  be  used in all
tests.  The field  test protocol  requires four measurements each for partic-
ulate mass concentration and for particle size distribution.   The impector
data  yield the  size  distribution  which is  then  applied  to  the  mass
measurement to give  cumulative  emission factors as  a function of particle
size.  An EPA Method 5 sampling train is  used for the mass measurement.

Ducted Sources (Condensible Emissions) -  Some iron  and steel process
emissions"contain  a significant  fraction of condensible compounds.  In this
case, significant is taken  to mean greater  than 10 percent  of the total
particulate emitted.   Typical  sources  in this  category are:  coke oven
charging, quenching,  and battery stacks; sinter windboxes;  and electric arc
furnaces.  The  condensible  fraction may  contain  both organic and inorganic
compounds.

The condensible emission IPM sampling protocol  was prepared  by Southern
Research Institute,  under contract to  the  EPA.  The  general approach is to
introduce a  slip  stream of  the  source  exhaust into a  dilution chamber
supplied with clean ambient air.  The dilution air  flowrate may be on the
order  of  10-30 times  the  slip  stream  flowrate.   Total  particulate and
particle size samples of the diluted slip stream are obtained after complete
mixing.  The samples thus collected would reflect their initial state in
the atmosphere.
                                     252

-------
The condensible  sampling system has been designed and units are now being
fabricated.  As shown in Figure 3,  its  essential parts include a dilution
chamber  approximately 30.5 cm  in  diameter and  183  cm long.   An  ID  fan
supplies filtered air to one end of the chamber.   The  source  slip stream is
extracted isokinetically using a heated probe and is introduced at the same
end.  At  the outlet end of the chamber a high volume impactor filter assembly
collects and sizes the  total particulate.

Fugitive Sources  -  The measurement of  fugitive particulate  emissions by
size requires several sampling devices.  A standard high volume  sampler gives
total  suspended  particulate (TSP);  a  high  volume  sampler (with  a size-
selective head)  and  a cascade impactor on  a high volume sampler  (with a
horizontal elutriator as a 15 ym precutter) both yield particle size data.
One major reason  for  using multiple particle sizing  devices  at this time is
the lack of experience upon which to base the selection of a better device.
As field experience accumulates,  a  decision can be made, thus reducing both
testing and analytical costs.

A  typical  upwind/downwind  fugitive  emission  field  setup,  such  as  for
measuring emissions  from vehicular traffic on a road, would include:

      (1)  selection of a section  of  road  which  was normal to the  wind
           direction and sufficiently remote from buildings,  trees, or other
           roads to avoid interference;

      (2)  a sampling array on the upstream side of the road consisting of
           three high volume samplers — one  equipped  for TSP, one with a
           size-selective  head,  and  one  with  a cascade  impactor,  all
           situated 1 m above ground surface;

      (3)  an array at the downwind site consisting of the same equipment as
           upstream at a 1 m height, plus  two high volume units at a  2  m
           height —  one with a s.ize-selective head, the other with a cascade
           impactor,  and all five units located  on the plume center line;

      (4)  a size-selective high volume sampler  at each plume wing at a 1 m
           height; and

      (5)  for determining verticle profiles, locate  high volume  samplers
           with elutriators on the  plume centerline at heights of  1,  2, 3,
           and 4  m.

 The quasistack method involves hooding the fugitive  emission  source  and
 drawing  off emissions through a duct;  emission  measurements would  then
 follow the ducted emission protocol.   Obviously,  this approach would
 only be  feasible for relatively small  or compact sources  since the expense
 of building hoods and ducts just for  test  ourposes could  be prohibitively
 expensive.   Alternatively, if  a normally  fugitive source,  such as   EOF
 charging,  is found to be well contained at a plant,  then every effort would
 be made  to perform tests at that plant assuming it was  representative of
 the industry in  its  operational  characteristics.
                                    253

-------
 For the measurement of fugitive emissions  exiting  through a roof monitor,
 either a high volume cascade impactor  fitted with  the elutriator head  or
 a fugitive assessment sampling train (FAST) would be used.  The FAST shown
 in  Figure  4 offers  the   advantage of collecting large quantities  of
 material in  selected size  fractions,  permitting subsequent  chemical and
 biological characterization.

Source Selection

At the outset of the iron and steel sampling program, the decision was
made to proceed with field test selection voluntarily, rather than through
the application of Clean Air Act Art. 114,   Industry contact was initiated
through the American Iron and Steel Institute (AISI) who established an ad
hoc coordinating committee.   Meetings with this committee were held to
present an overview of the whole program and, following resolution of
outstanding issues, to review sources selected for testing.

The EPA/AISI cooperative effort has thus far resulted in the mounting of
an extensive field sampling  program at Armco, Inc:'s Middletown Works.  Dis-
cussions will soon be initiated with  several other  companies; several field
test programs should be getting underway soon.

Tfie source selection priority  list,  shown  in Table  1, was developed
based on estimated controlled particulate emissions from each  source on a
nationwide basis.  This prioritization represents  an average of emission
factors  developed  under separate  efforts:  one  represented very good
control efficiency(8)_; the  other,  a  somewhat  lower level which might be
termed typical for non-new installations^),  It is only fair  to say that,
at best,  this procedure  is  still rather arbitrary but does nevertheless
provide a rational approach  to  source selection.

The source selection priority list also considers existing particle size
data and its quality.  Since there in very little  particle size data for
iron and  steel  sources,  this consideration did  not affect the priority
position of any  source. Six data sets are currently contained in EPA's Fine
Particle  Emissions  Information  System (FPEIS)lO.    These   data  sets,
consisting of  three open hearth  stack tests,  two  electric  arc  furnace
tests, and  one  coke  pushing shed test, are summarized in Figures 5-7,
respectively.  Although  these  data are judged  to be good, consideration
must be given to  when they  were obtained (1974-77) and the advancements
made in particle size  sampling technology since then. These three sources
should be tested again.

In addition to the selection  procedure coordinated with AISI, every  effort
has  been  made to  combine measurement  of  IPM  with other EPA  sampling
programs.  Not only  does  this  serve  to reduce EPA expenditures, it also
reduces inconvenience to the host companies.  To date, these "piggy back"
projects constitute the greatest area of activity due primarily to  the fact
that initial ground work  hact already been  laid by the Agency, making it
possible to mount the field effort relatively quickly.
                                    254

-------
 Results To Date

 Kaiser Steel  Corp.  The  first  two  sources  tested under  the  IPM program
 were  the  hot metal desulfurization  (HMDS)  and BOF charging and  tapping
 emission control systems recently installed at Kaiser Steel  Corporation's
 Fontana,  CA plant.   These tests,  performed  in coordination with  EPA's
 Office of Enforcement through  the Region V  office,  consisted of  total
 particulate by EPA Method 5 and particle size before and  after the  control
 device which  was,  in both cases, a baghouse.  EPA's contractor  was Acurex
 Corporation.

 Particle  size before  control was measured  with the EPA two-cyclone  IPM
 train described  previously.   Dtie  to the  low particulate concentration,
 baghouse outlet particle size in both cases was measured with an Andersen
 Mark III impactor  fitted, with  the  15 ym cut point  cyclone precutter.

 BOF emissions from hot metal charging and  tapping  of  finished  steel were
 measured separately.  Charging measurements did not include  the addition
 of scrap.  Results  of the HMDS tests were reported in detail earlier  in this
 Symposium.'^)

 Armco. Inc.. Middletown.  The extensive testing program undertaken  for EPA
 by Midwest Research Institute at Armco,  Inc.'s Middletown Plant  encompas-
 sed both  open sources and ducted  sources; emphasis  was on the  former.
 Specifically, emissions from paved  and unpaved roads were measured before
 and after the initiation of emission reduction schemes.   For  paved roads,
 the emission reduction consisted of flushing with water  and/or  sweeping -
 - vacuuming at regular intervals.  Berms of paved roads  were  treated with
 Coherex®.  The controlled urtpaved road was  first tested  shortly  after ap-
 plication of  the suppressant (Coherex®) and again  several weeks later to
 establish the emission control efficiency decay curve.

 In  addition  to  the  road  tests,  measurements were  made  of  windblown
 emis ions from the  coal pile and emissions arising as a result of coal pile
maintenance.

All of these open source  measurements were  made using  the upwind/downwind
 techniques.    Road  surface  silt content  was sampled to allow correlation
with the measured emission rate.  Numerous samples were  taken from other
 inplant roads to develop an idea of the representatives oi" the  sampled road
 segments and also to allow a more realistic extrapolation of  the  test data
 to the whole plant.

Ducted emissions measured at Middletown were from the  BOF main  stack.
Testing  of the open hearth  stack  is being considered.   Although  the IP
protocol calls for measurements before and after the control device, the
nature of the limited combustion  BOF operation  precluded  measurements
before the  scrubber.    Measurements  after  the scrubber  included total
particulate and  particle size as  per the protocol.  Results  are being
reduced and should be available in December 1980.
                                   255

-------
Testing of the open hearth stack is in the initial  planning  stage.  It is
intended that the field program consist of total particulate and particle
size samples before and after the control device for each major portion of
the furnace cycle.   Each furnace at Middletown has its own control device.

Dominion Foundries and Steel  Company.  Dominion first installed cast house
emissions controls several years ago.  The No. 1 blast furnace cast house
control system,  installed  in 1975,  was  tested by EPA'*-*'  in 1976.  This
system employs the total building evacuation condept.  Although.particle
size measurement of  uncontrolled emissions was  attempted,  'the data were
not reliable due to particle bounce (no precutter was used).  Using 1980
techniques which employ the  15 ym cyclone precutter ahead of the cascade
impactor would have solved this problem.   The combined control system for
cast houses No. 2 and 3 was started  up  in November 1978.  These cast houses
share a common fan  and baghouse and  also employ total building evacuation.
Cast house No. 4  is  currently being used to evaluate concepts for local
emission control.

The IPM emission tests are being run on  the combined system  serving No. 2
and 3; however, measurements  are being made only when No.  3 is casting.
Emission tests follow the protocol  for ducted sources.  Measurements are
being made for EPA by GCA/Technology Divison in the duct upstream of the
baghouse; no attempt is being made  to  measure the discharge  from the open
monitor on the baghouse.  Testing will be  completed  in mid-November 1980;
preliminary results are anticipated in January 1981.

Bethlehem._S_te_e_l_ Corporation,  Sparrows  Point. GCA/Technology  Division will
be conducting emission'tests  for EPA at the new "L"  furnace cast house at
Sparrows Point later this month. This is a large modern  furnace employing
close-fitting hoods and covers over the trough,  iron runners, and spouts,
a practice pioneered in Japan.  Emissions are ducted to  a large baghouse.
Emissions, following the ducted source protocol, will be measured in the
duct upstream of  the baghouse.  Since this baghouse also controls emissions
from numerous other fugitive sources, no attempt will be made  to sample the
open monitor discharge.  Results of  these tests are  also  anticipated early
next year.

Future Tests

EPA is currently reviewing the overall status of the field testing program
in  light  of the source  priority  list.    Future  tests  will  continue to
address the  highest  priority  sources  first.   Testing  of  high priority
sources suspected  of containing a significant  fraction  of condensible
emissions (e.g.,  sinter plant windboxes, electric arc furnaces, and coke
ovens) will  be initiated as soon  as the  condensible emission sampling
protocols are formalized and the necessary  sampling equipment is available,
In  the  meantime,  noncondensible sources  will  continue to  be  tested as
rapidly as possible.
                                   256

-------
ACKNOWLEDGEMENTS

The  three  sampling contractors for EPA's  IP  program arc. GCA/Technology
Division, Midwest  Research  Institute, and  Acurex  Corporation.   The AISI
and  its  member  companies  (in  particular  Armco,  Inc.,  Bethlehem Steel
Corporation, Dominion Foundries  and Steel, Ltd.,  and Kaiser Steel Corpora-
tion) have been most helpful.
                                   257

-------
TABLE 1.   IRON AND STEEL SOURCE PRIORITY RANKING FOR IPM STUDY,
           CONTROLLED EMISSIONS

Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Process
Coke quenching
Blast furnace cast house
BOF stack
Material stockpiles
Roadway travel
Coke combustion stack
BOF charge and tap
Coke pushing
Sinter, misc. fugitives
Sinter windbox
EAF charge, tap, slag
Coal preparation
OH stack
Coke door leaks
EAF stack
Sinter discharge end
Blast furnace top
Teeming
Ore Screening
BOF, misc. fugitives
Coke topside leaks
Industry toLol
part i.cul ul c
emissions, Mg/yr
34,500
22,700
20,000
16,300
16,300
16,300
14,500
8,900
8,700
8,200
7,600
7,400
7,300
7,100
6,600
5,700
3,700
3; 700
3,300
2,200
2,100
                                   258

-------
TABLE 1.   IRON AND STEEL SOURCE PRIORITY RANKING FOR IPM STUDY,
           CONTROLLED. EMISSIONS (Continued)
Rank
22
23
24
25
26
27
28
29
30
31
32
Process
Reheat furnaces
Blast furnace combustion
OH roof monitor
Coal charging
Open area
Machine scarfing
EOF, hot metal transfer
OH, misc. fugitives
Soaking pits
EAF, misc. fugitives
OH, hot metal transfer
Industry total
parti.cul.ate ;
emissions, Mp/yr
2,000
2,000
2,000
1,800
1,100
670
650
640
570
540
190
                                   259

-------
                                        SAMPLING NOZZLE
                                                         PROBE
                         CYCLONE SRI IX
                         °50 " 15+2 Jim
CYCLONE SRI III
OgQ - 2.5 i 0.6 Jim
            (2)
    Figure 1.  Schematic  of dual-cyclone sampler  for  inhalable particles,
                                 SAMPLING NOZZLE
                                                                   PROBE
                CYCLONE
                                            CASCADE IMPACTOR
        (2)
Figure 2.  Schematic of a cascade impactor/precollector cyclone  system.
                                   260

-------
                         PROCESS STREAM
ho
                                        HI-VOL IMPACTOR
                                        FILTER ASSEMBLY,
                                          SAMPLING
                                          CYCLONE
                                                                               EXHAUST BLOWER
/  \
/
1
*
FLOW
-~ 	 — •

|.-H"I'B
PROBE



F
r1— 	 1
f
TO HEATERS, BLOWERS
TEMPERATURE SENSORS
1

\
TO ORIFK
PRESSURE
                                                  MAIN CONTROL
                                                                           IT
                                                                        :©<§>©
       FLOW. PRESSURE
       MONITORS
                                                                                     TO ULTRAFINE
                                                                                     PARTICLE SIZING
                                                                                     SYSTEM (OPTIONAL)
                                                                                        ,DILUTION AIR
                                                                                         HEATER
                                                                                                 CONDENSER
                                                                                         ICE BATH
                                                                                                              DILUTION AIR
                                                                                                              BLOWER
                                            (4)
                                    'igure 3.   Diagram of atmospheric dilution  sampling system.

-------
          INLET
          THERMOCOUPLE
          vv
                        CYCLONE
INLET
                                                      MAIN VACUUM
                                                      BLOWER
             HORIZONTAL
             ELUTRIATOR
                         CYCLONE
      TRAP
      OUTLET
THERMOCOUPLE
                                  VACUUM
                                  PUMP
                                                                   EXHAUST
 Figure  4.   The FAST with a horizontal  elutriator inlet  for collecting
             inhalable  particulate.
                                     262

-------
 99.990

 99.950
  99.90
  99.80
  99.50
     99
     96

     95
     90
UJ
o
ce
u
a.
UJ
O
 80
 70
 60
 50
 40
 30
 20

 10
  5

  2
  I
 0.5
 0.2
0.15
 O.I
 0.0
  10
          AVERAGE
          02 LANCING

          CHARGING
                       ill     i   i  i i 11111
                       10°             10'
               PARTICLE  DIAMETER, micrometers
   Figure 5.  Average size distribution - open hearth  furnace
             emissions, uncontrolled.
                             263

-------
UJ
o
a:
u
0.

UJ
o
 99.990


 99.950

  99.90

  99.80

  99.50

    99

    98


    95


    90


i    80

    70

    60

    50

    40

    30


    20


    10


     5


     2

     I

   0.5

   0.2

  0.15
   O.I
    0.0
    10
              COVERALL  PROCESS AVERAGE

              O MELT


              D TAP-IIELT
              * *  ' «
                   III11	«   '
                                             I  I  I,
                       10
               PARTICLE   DIAMETER,  micrometers
  Figure 6.  Average  size distribution  - Marathorn LeTourneau

            Electric arc facility.
                             264

-------
99.990)
         O Run One
         D Run Two
               Three
                                                        10'
               PARTICLE  DIAMETER,  micrometers
Figure
             Average particle size distribution — coke oven pushing.
                            265

-------
References

1.   Miller, F. J. ,  et al.,  "Size  Considerations  for Establishing a Standard
     for Inhalable Particles," J.  Air Pollut. Contr. Assoc., 29(6):610(1979).

2.   Smith, W.  B. ,  and R. R. Wilson,  Jr.,  Procedures  Manual for Inhalable
     Particulate Sampler Operation (Draft), EPA Contract 68-02-3118, Southern
     Research Institute, November 1979.

3.   Harris, D. B. (Editor),  Procedures for  Cascade Impactor  Calibration and
     Operation in Process Streams  - Revised 1979 (Draft), EPA  Contract 68-02-
     3118 TD 114, Southern Research Institute, May 1980.

4.   Williamson, A.  D. , Procedures Manual for Operation  of  the Dilution Stack
     Sampling  System (Draft), EPA  Contract 68-02-3118,  Southern  Research
     Institute, October 1980.

5.   Protocol for the Measurement  of Inhalable Particulate  Fugitive Emissions
     from Stationary Industrial Sources (Draft), EPA Contract 68-02-3115 TD
     114, The Research Corporation of New England, March 1980.

6.   "Method 5  - Determination  of  Particulate  Emissions from Stationary
     Sources," Federal Register Vol.  42,  No.  160,  August 18,  1977, pp. 41776
     to 41782.

7.   Steiner, J., and D.  Bodnaruk, "Particulate and S02 Emission Factors for
     Hot  Metal   Desulfurization," Symposium  on   Iron  and Steel  Pollution
     Abatement Technology for 1980 (November 1980, PhiladeIpia, PA).

8.   Cuscino, T. A.,  Particulate Emission Factors  Applicable  to  the Iron and
     Steel  Industry,  EPA-450/4-79-028, Midwest  Research  Institute,  August
     1979.

9.   Barber, W.C., Particulate Emissions  from Iron and Steel Mills, EPA/OAQPS
     Internal Memorandum, dated November 6,  1978.

10.   Reider, J. P.,   and  R. F.  Hegarty,  Fine Particle Emissions Information
     System: Annual  Report   (1979),  EPA-600/7-80-092  (NTIS  PB 80-195753),
     Midwest Research Institute,  May 1980.

11.   Fitzgerald,  J. ,  D. Montanaro,  and E.  Reicker,  Development of Size-
     Specific Emission Factor (Draft), EPA  Contract 68-02-3157 TD  3, GCA/-
     Technology Division, October  1980.

12.   Gronberg,  S., Test  Program Summary for Characterization  of Inhalable
     Particulate Matter  Emissions (Draft),  EPA  Contract  68-02-3157 TD 5.
     GCA/Technology  Division, September 1980.

13.   May, W.  P.,  Blast  Furnace  Cast  House  Emission Control  Technology
     Assessment, EPA-600/2-77-231 (NTIS  PB  276999), Betz  Environmental En-
     gineers, November 1977.
                                    266

-------
Session 2: WATER POLLUTION ABATEMENT

Chairman:   George F. Haines, Jr.
           Homer Research  Laboratories
           Bethlehem Steel Corporation
           Bethlehem, PA
                    267

-------
HyDROTECHNIC CORPORATION
          APPLYING GREENFIELD WATER SYSTEM DESIGN AND MANAGEMENT

              TECHNIQUES TO EXISTING STEEL PLANT FACILITIES


                 HAROLD J. KOHLMANN AND HAROLD HOFSTEIN

                        HYDROTECHNIC CORPORATION


                               ABSTRACT


             A great amount of emphasis is being placed on recycle,
       reuse,  cascade,  etc. systems for the purpose of reducing the
       amount of contaminants discharged from industrial facilities.
       These methods can and do reduce volumes of water and masses
       of contaminants  discharged;  however, the application of new
       principles' to existing facilities is usually looked upon with,
       at best,  extreme skepticism.

             This attitude is understandable since,  in many cases,
       an existing system cannot be completely dr conveniently re-
       vamped  to include all the "niceties" that can be designed
       into  a  greenfield site.   A completely closed mind,  however,
       cannot  be tolerated in these days of increasing prices  and
       stricter  environmental controls.   There is no room  for  the
       attitude  of "we've been  doing it  that way for thirty years
       and no  one can make us do it differently".   Attitudes like
       that  can  only force management  to spend much  more money than
       is really needed to comply with the  regulatory requirements for
       pollution control.

            This paper presents various  practices regarding the reuse,
       recycle and cascade  of wastewater  in  steel plants which were
       either developed for greenfield installations  and later applied
       to existing installations or  those that  were  applied to existing
       installations  directly.   It also points  out areas of  difficulty
      which are  encountered when existing  facilities  are upgraded.
      Design parameters must be carefully determined  so that  facilities
      are not grossly  oversized or, more tragically,  undersized.  The
      treatment  and  reuse  compatibility of one wastewater with another
      must also be ascertained  so that incompatible  "mixes" are not
      attempted.

            Segregation of flows within existing mills, although  initially
      costly in most cases, can be the most economic  alternative.   This
      is especially true when large quantities of "clean cooling water"
      combine with "dirty water".  These various methods are presented
      together with descriptions of systems actually  installed and operat-
      ing to show the results that can be obtained with proper study and
      design principles.
                                    269

-------
Due to the consideration of water pollution control, water con-



servation and the recovery of our precious resources, the recycle



and reuse of water in industrial facilities is receiving a tremen-



dous amount of attention.  Various methods exist to effect reduc-



tion of discharges.  These include in-process changes, close con-



trol of water chemistry, increased instrumentation, segregation



of flows, etc.  These methods can and do result in reduced volumes



of water discharged and, in turn, the masses of contaminants dis-



charged.  However, the application of new principles to existing



facilities is usually looked upon with, at best, extreme skepticism.





This attitude is understandable since, in many cases, an existing



system cannot be completely or conveniently revamped to include



all the "niceties" that can be designed into a greenfield site.



A completely closed mind, however, cannot be tolerated in these



days of increasing costs and stricter environmental regulations.



There is no room for the attitude of "we've been doing it that way



for thirty years and no one can make us do it differently".  This



type of attitude causes management to spend more money than is



really needed to comply with the regulatory requirements while



using undue amounts of energy.






Years ago when a new plant was designed the layout of the production



facilities in a most practical manner wa£ the prime consideration.



This practical layout resulted in a smooth and efficient flow of



raw materials to finished product.  Little thought was given to



water systems except to keep them "out of the way" and make sure
                              270

-------
 they caused as few problems as possible.  Scale pits and blast
 furnace thickeners were installed, not for the primary purpose
 of water pollution control but to safeguard against the clogging
 of sewers which could cause production delays and stoppages.
 As environmental regulations became stricter, clarifiers were in-
 stalled after scale pits,  portions of blast furnace gas washer
 water were recirculated,  oils were skimmed and acids neutralized.
 Most wastewater collection systems "grew like Topsy" without over-
 all plant or system-wide  planning which resulted in mixes of
 different,  incompatible wastewaters in common sewers.

 Initially environmental regulations were developed,  in  many  in-
 stances,  through  the mutual cooperation between steel plants,  the
 regulatory  authorities  and engineering consultants  retained  by in-
 dustry  and/or  the authorities.  This process  was necessitated  be-
 cause no  one knew for sure what effluent standards  could  be  met  in
 a  practical and economical way.   Pilot tests  had to  be  conducted
 and  reasonable  standards were  set based on the  ability  of a  treat-
 ment process to produce a  specific  effluent.  Contaminants were
 limited to  those  that were obvious  and readily  detectable.   They
 were, for the most part, suspended  solids, oils, iron pH, etc.
 As new steel plants were planned  and  constructed, more  sophisticated
 treatment schemes were required by  the  regulatory authorities who
 rightly felt that new, greenfield planning could produce  treatment
 systems that would discharge wastewater that contained  lower levels
of contaminants and would achieve these levels economically.
                               271

-------
This tightening of the regulations for new facilities had an
impact, in addition, on the regulations imposed on existing
facilities.  Application of the new technology developed for
greenfield sites to existing sites was a natural consequence.
Some of the ideas that developed are as follows:

Hydrochloric Pickler Cascade Rinse System
In the pickling of steel, a rinse step is necessary to remove
residual pickling acid and iron salts that adhere to the steel
and cause staining and corrosion.  Originally, rinse systems
utilized several rinsing stages, each completely separated
from the others.  The rinse water from each stage was discharged
directly to sewer.  Resultant discharge rates for this type of
rinse system vary from 700 to 1900 1/kkg (200 to 550 gal/ton)
of steel pickled.  Figure I is a schematic diagram of a typical
rinse system of this type.  All the rinse water from this type
of system requires treatment prios to discharge.

A relatively recent development is the cascade rinse system where
the steel leaving the pickler is rinsed with water which is then
cascaded upstream towards the pickling tanks.  The last rinse
tank contains the freshest water.  When spent, the rinse water is
discharged, but because of the cascade configuration, only 50 to
200 1/kkg  (14 to 55 gal/ton) of water is required which consider-
ably reduces the amount of wastewater to be treated prior to dis-
charge.  Figure II is a schematic diagram of a typical pickler
cascade rinse system.
                             272

-------
     FRESH WATER
 ro
 ^j
 CO
           SQUEEGEE
             ROLLS
           PICKLING TANK
RINSE TANK No.l
RINSE TANK No.2
RINSE TANK No.3
                                                                    DISCHARGE TO TREATMENT
                                                                      700 TO 1900 l/kkg
                                                                      (200 to 550 go I/ton)
4YDROTECHNIC CORPORATION
    NEW YORK. N.Y.
                        FIGURE  I - TYPICAL  PICKLER  RINSE  SYSTEM

-------
     SQUEEGEE
       ROLLS-
         PICKLING
          TANK
       DISCHARGE TO
        TREATMENT
     OR REGENERATION
          PLANT
      50 TO 200 l/kkg
      (14 TO 55gal/ton)
                      n

                              T.
RINSE TANK
No. I
RINSE TANK
                   No.2
                                             FRESH
                                             WATER


                                             STRIP
RINSE TANK
                   No.3
                                            RINSE TANK
                   No. 4
                    FIGURE 31 - TYPICAL PICKLER  CASCADE RINSE  SYSTEM
HYDROTECHNIC CORPORATION

     NEW YORK. N.Y.

-------
Cascade rinse systems can and have been retrofitted to existing
pickler systems which, according to reports, are operating satis-
factorily.  It is obvious that the treatment facilities capital
and operating costs will be significantly less for the reduced
flow.

Elimination of Pickling Tank Heating Steam
Pickling baths must be maintained at elevated temperatures to
permit proper pickling of steel.  This elevated temperature is
usually maintained by the injection of live steam into the
pickling baths.  This injection of steam has three results, namely:
the bath is heated and maintained at the desired temperature, the
steam causes vaporization of acid which must be cleaned with fume
scrubbers, and the steam is condensed causing dilution of the
pickling bath.
Temperature maintenance is the desirable aspect but vaporization
and dilution are undesirable aspects of raw steam injection.
If this injection can be eliminated then the two problems could
also be eliminated or eased.  Various means are available to heat
the pickle acid baths such as submerged combustion,  external heat
exchangers and a furnace to heat the strip prior to its entrance
into the pickling bath which would then bring the necessary heat
into the tank to maintain the desired temperature.   However,  the
furnace method requires a supplemental heat source for start-up
conditions.  These three methods are shown schematically on Figure
III.
                              275

-------
                             GAS
    AIR
    GAS
L>
   FLAME
                 PICKLING
                   TANK
                                       FURNACE
          SUBMERGED COMBUSTION
                                                PICKLING TANK
                                               STRIP FURNACE
                                   PICKLING TANK
                                                  PUMP
                                                 HEAT EXCHANGER
                                    CONDENSATE

                                 HEAT EXCHANGER
                 FIGURE II  ALTERNATE  METHODS FOR MAINTAINING
HYDROTECHNIC CORPORATION
    NEW YORK. N.Y.
                     PICKLING  TANK TEMPERATURE
                                                                       STRIP

-------
 Substitution  of  one  of  these methods  can,  for  example,  reduce
 the  flow of waste  pickle  liquor  from  14-18 m3/nr  (60-80 gpm)
 to 7 m3/hr  (30 gpm)  for a pickler processing 1.5  x  106  mtpy.
 This 50% reduction would  greatly reduce treatment and disposal
 costs or the  costs to erect and  operate an acid regeneration
 plant.   Retrofit of  any of these methods could be accomplished
 without  problems during a period when a pickler is  down for
 major maintenance.

 Hot Mill Water Reuse and  Conservation
 Many methods  are available to reduce the discharges of  water
 from hot rolling mills.   Any reduction in  the amount of water
 discharged from  this type of production facility  will be sig-
 nificant since large amounts of  water are  required  for  cooling
 and cleaning  purposes.  Hot rolling facilities are usually
 composed of a reheat furnace, a  forming section,  a cooling sec-
 tion and, in  the case of  strip or skelp mills, a  coiling section.

 Large amounts of water  are needed to protect various parts of re-
 heat furnaces from the  high temperatures encountered, but since
 this water is non-contact cooling water, it will not be discussed
 further  except to  add that the water uses  for this application
can be reduced and in many cases energy can be conserved by using
the furnace cooling system as a  steam boiler.

Coiler water is  used for  cooling, oil removal and for lubrication
and becomes contaminated with oil,  suspended solids and heat.   It
should be collected and treated with the wastewater that discharges
from the mills'   stands.
                              277

-------
Large amounts of water are used to cool the steel strip on the
runout tables of hot strip mills.  The water is used to reduce
the temperature of the steel strip from rolling temperature to
coiling temperature in a relatively short period.  To accomplish
this, large amounts of water are applied.  Although the tempera-
ture of the steel is reduced considerably, the water temperature
does not increase very much due to the large volume necessary to
achieve the rapid cooling effect desired.  For instance, in
cooling strip from 870°C (1600°F) to a coiling temperature of
565°C (105QOF), a change in steel temperature of 305°C  (550°F),
the temperature of the entire water system may increase only by
approximately 10°C (18OF).

Cooling of this large amount of water is extremely expensive and,
since contamination by suspended solids is minimal, elaborate
treatment of the entire flow is not necessary.  Usually the mill
supplier recommends a blowdown from the runout table system to
the mill stands section of between 20 and 30% of the water applied
to the strip.  This blowdown amount is to remove the heat and sus-
pended solids contained in the water to prevent an unacceptable
buildup.  If, instead of blowing down this large amount of water,
the runout table system water is treated in a segregated
system there can be closer control of the water chemistry and, in
turn, the amount of blowdown required may be reduced significantly.

This segregation will enable the operator to control the level of
dissolved material, specifically chlorides, which affect strip
staining.  if a portion  (20 to 30%) of the runout table water is
                            278

-------
filtered and cooled, the temperature will be maintained and
the suspended  solids levels will be acceptable for reuse.
Chloride levels can be continuously monitored and a blowdown
discharged to  maintain the chloride level at an acceptable value.
This is easily done in a greenfield installation and, surprisingly,
in most existing mills it is not difficult to attain,particularly
if the runout  table system discharges to a sewer outside the mill
building and can be intercepted easily.

Figure IV is a schematic diagram showing how these concepts; have
been applied at a greenfield hot strip mill.

General Concepts
At greenfield  sites discussed in previous sections the concepts
for maximum water conservation can, with little extra effort and
cost, be designed into the plant water systems.  However, at
existing plants where production modifications cannot be made
without considerable expense, wastewater treatment systems can be
installed to optimize desired treatment to enable recirculation
and also reduce pollution loads discharged.

Facilities Sizing
When a treatment facility must be installed to treat the waste-
water from a production facility, the first stage is to establish
the volume requiring treatment.  The first source of information
would be the most recent drawings showing water requirements and
discharges.   However, it may be dangerous to use these figures
blindly because in many cases operators may have made modifica-
tions which were never added to "as-built" drawings. It is there-
fore necessary to undertake a field survey using the proper
                             279

-------
oo
o
1
4
}_
™
ipf

b
0
8
i
r
O
0
1


ro
ID
' ^

r
REHEAT FURNACE
SKID
COOLING
i
r
SEAL
TROUGH
a WATER
JETS
i
AIR COOLERS
a
PUMP STATION
i


i
t

DEMINERALIZED
WATER
GENERATION
a STORAGE
— vx 	 '
FROM
iBLE WATER
>LY SYSTEM
HYDRAULIC
a
INSTRUMENT
COOLING
> 100


i


i

r


i
V) If)
W IO
0 *
-1 ^


».
0
o
ro
if)
' ^ '
	 »
i
en
O
" PRODUCT COOLING WATER "
- FURNACE
|_ a ROUGHING
SECTION


1

LOSS
A
.4
ROUGHING
g SECTION

LOSS
4,


A
r
FINISHING 0
SECTION 0
SCALE PIT rt
5278 ^


r
63
k
if)
 r-
r V v
MOTOBROOM M|LL
LUBE COOLING INSTRUMENTS

00
j
26 v

COOLING
*" TOWER 23 ogg
_^ a LWJJ
" PUMP STATION
r
SETTLING
FILTRATION
COOLING TOWER
a PUMP STATION
120
199
i
CVJ
.OSS
L

* 1
                                                           MAKE UP
                                                                                2I-
                                                                                   LOSS
                                                                     NOTE:

                                                                     ALL FLOWS IN m3/hr.
                            FIGURE M - HOT STRIP  MILL WATER  SYSTEM

-------
 techniques to determine what the actual flows being discharged
 are, and to establish the quality of the water so that the
 proper treatment unit operations, properly sized, can be in-
 stalled.  Not to perform the field survey may result in treat-
 ment facilities that are improperly sized.   Oversized facilities
 are wastefut but the apparent economy of undersized facilities
 can also result in wasteful practices.

 Sewer Segregation
 Many of the older steel plants have common  sewers which collect
 clean cooling waters,  contaminated wastewaters  and storm runoff.
 In  the 1950s,  most sanitary sewers had  to be  segregated from the
 other flows.   The remaining combined sewers,  in many  cases,  re-
 sult in the necessity  of installing treatment facilities  that  are
 hydraulically  sized to treat not  only the contaminated  wastewater,
 but  also clean cooling water.   This is  an extremely wasteful
 practice and serious consideration  should be given to sewer  segre-
 gation  prior to construction of oversized treatment facilities.
A properly  performed survey  can determine which flows may be
eliminated  from treatment requirements.

Although sewer segregation can be very costly, in the long run,
the expenditure of monies to  effect  segragation can significantly
reduce  future operating costs.
                             281

-------
Water Reuse
During the course of the recommended field survey, the qualities
of water required for production operations would be identified.
These qualities can be compared with allowable discharge qualities.
A determination can then be made as to whether it is more economical
to treat wastewaters for discharge, or for reuse within the
facilities, or at some other production facility.  Ideally, an
entire plant should be treated as one water system so that treat-
ment facilities can be optimized by combinations of compatible
streams to be treated in one facility.  Reusing water without
treatment by cascading the blowdown from a high quality recirculat-
ing system as makeup to a system requiring lower quality wherever
possible is the ideal.  However, the quality will be degraded
eventually to a point where treatment will be required.  Geological
constraints within individual plants usually preclude that sort of
operation.  It is therefore necessary that each plant be studied
as an individual entity and each plant will be found to have its
own optimum operating conditions.
                              282

-------
THE REGENERATION OF NITRIC AND HYDROFLUORIC
     ACIDS FROM WASTE PICKLING LIQUID
             Hu Delu,  Engineer

   Liang Xiuchung, Head Engineer of the
         Water Quality Department

    Wang Chingwen,  Civil  Engineer,  Head
      of the Water  Quality Department
        General Research Institute
                    of
         Building  and  Construction
                    MMI
        People's Republic of China
                       283

-------
        THE REGENERATION OP NITRIC AND HYDROFLUORIC
             ACIDS PROM WASTE PICKLING
                BY Hu Delu, Liang Xiuchung
                   and Wang Chingwen
INTRODUCTION
The scales formed on the surface of stainless steel during processing
are usually removed by pickling with 7 - 15% nitric acid and 4 -8$
hydrofluoric acid*  This process has the advantages over other pro-
cesses of high pickling rate, no overpickling and keeping the pickled
steel with good appearance.  The pickling liquid will lose pickling
ability and become waste acid when its metal ion contents have reached
a certain concentration' through repeated pickling*  The total acidity
of waste acid is still rather high, which contains a large amount of
iron,  nickel and chromium salts* Nitric and hydrofluoric acids are
strong corrosives, chromium, is a strong toxicant, it should therefore
be forbidden to discharge waste acid without being treated* Besides,
nitric acid being expensive should be recovered and reused from the
economic point of view*
Research work on treatment of waste liquid containing HNO$ and HP has
been carried out and some methods put into operation in the European
countries •• the U. S., Japan, etc.  The chemical method can recover
some useful matters, but it is complicated and needs a lot of equip-
ment. It  is a progress to use fluid! zed bed instead of fixed one in
the ion exchange method, but it needs concentration of hydrochloric
acid and  treatment of dilute acid* The lower energy consumption and
the availability of equipment and material are Strong points  of
solvent extraction method, but further study is needed due to low
                                284

-------
recovery ratio of hydrofluoric acid.  The vacuum evaporation method
makes it possible to use material with high temperature and corrosion
resistance, the acid recovery ratio is high as well.  Up to now,  it
is the more effective recovery method.

Based on the data obtained from pilot tests, a set of semi-continuous
one-step HN03 and HP regenerating device by vacuum evaporation was
designed for a seamless steel pipe shop in 1976 and put into opera-
tion in 1977.  Its main feature is the use of corrosion resistant
heater, evaporator and condenser.  Later, an equipment for treating
residual liquid after acids recovery waa added and put into opera-
tion in 1979.  Now, the waste liquid of the shop is no longer  dis-
charged and has been fully regenerated and reused.  Moreover, nickel
carbonate and ferrous sulfate can be recovered from the residual
liquid.  What la more, there will be no secondary pollution. Prac-
tice has proved that the vacuum evaporation" method has the advan-
tage of simple equipment, easy operation and high, recovery ratio,
it is therefore an effective recovery method.

ONE^STEP VACUUM EVAPORATION
Nitric and hydrofluoric acids are volatile, while sulfuric acid is
not.  The equilibrium temperature under atmospheric pressure of HN03,
HP and H2S04 are 87°C, 20°C and 300°C respectively.  The mixture  of
nitric and hydrofluoric acids can be recovered based on their dif--
f c ""ent partial vapor pressure.  The method consists of adding sul-
furic acid to waste acid, heating and evaporation.  ]>uring evapora-
tion, HN03 and HP are evaporated together with water, since H2304
has much lower partial vapor pressure than HN03 and HP.  Nitrate
                                   285

-------
and fluoric radicals displaced from metal salts toy sulfate  radical
combine with hydrogen ion in sulfuric acid to form HN03  and HP,
which are also evaporated, through condensation and recovery we
obtain the regenerated acids.
The vacuum evaporation method is recommended to lower evaporation
temperature, reduce corrosion: and Increase recovery ratio of HNOj
and HP.
The vacuum evaporation method is of two kinds, i.e. one-step and
two-step evaporation.  Two-step evaporation! comprises evaporation
in two stages.  The waste liquid is first evaporated in the  1st
evaporation system for dehydration^.  The waste liquid concentrated,
with HgSO^ added, is then reevaporated in the 2nd  evaporation
system, HN03 and HP are escaped to be condensed and r«oovered.
It is suited to treat diluted waste acid obtained during pickl-
ing.  One-step evaporation is a process used to recover mixture
of HN03 and HP where the waste liquid flows continuously into
the evaporation system without concentration and evaporates under
vacuum.  The equipment used in one-step evaporation is half of
that in two-step evaporation.  One-step evaporation is suited to
treat undiluted waste liquid obtained during pickling.  The pro-
cess is shown below.
                                286

-------
Waste liquid
       Plow sheet of regeneration of HN03 and HP from
       waste pickling liquid
  1- heater, 2- evaporator, 3- condenser, 4- receiver,
  5- injection, pump, 6- circulating water pond, 7- water
  pump, 8- reactor, 9- centrifugal filter, 10- oxldlzer,
  11- filter, 12- precipitator, 13- waste acid metering
  control device, 14- HgS04 metering control device,
  15- air lift, 16- pressure guage.
 Start injection pump, supply cooling water, suck certain volume
 of waste acid and concentrated HgS04 into evaporator.  Start air
 lift, heat with steam, add waste acid upon "boiling, keep the
 liquid  level in the evaporator unchanged, while HN03 and HP eva-
 porate  continuously and condense in the condenser, distilled
                                 287

-------
 liquid obtained is regenerated acid*  Metal salts in waste liquid
are continuously converted into sulfate, which remains in evapo-
rator.  When ferric ion concentration has reached the controlled
value, stop evaporation, turn off heating steam, remove vacuum
in regenerating system, transfer into reactor the residual li-
quid from evaporator, which is neutralized and reduced by iron
filings.  Nickel sulfide is precipitated by adding sodium poly-
sulfide according to the nickel content and separated by centri-
fugal, the filtrate is used to produce ferrous sulfate•  Clean
nickel sulfide precipitate is oxidized and dissolved, from which
iron is removed by adding sodium carbonate to form basic  nickel
carbonate•

OPERATING CONDITIONS OP VACUUM EVAPORATION PROCESS
Evaporation Conditions
The pressure of the heating steam is determined by the permeabi-
lity and compression strength o£ impermeable graphite heater,
it generally does not exceed 2 kg/cm2 and is controlled at 1 kg/cm2
during long-term operation.
Vacuum is a function, of boiling point*  The lower the vacuum,  the
higher the boiling point, and vice versa*  Operating temperature
is recommended not to exceed 65°C, since fibre-reinforced PVC
plastics is used in evaporator and pipes*  The vacuum should be
660 - 680 MM Hg.

To ensure complete evaporation of HN03 and HP, certain amount  of
                                288

-------
       should be added, i.e. sulfuric acid needed to keep the
 concentration of circulating HgS04 at 12.5 N plus sulfuric acid
 needed to transform nitrate and fluorate into sulfate.

 Acid recovery ratio will decrease if concentration; of circulating
 H2S04 is too low.  If it is too high, iron salts will crystallize
 ahead of time and amount of waste acid to be treated is reduced.

 It is determined by pilot tests that in the process sulfuric acid
should *>e discharged as residual liquid when ferric ion concen-
 tration, in. the circulating H2S04 has reached 75 g/1,  otherwise
 crystallization! will take place and the heater will be blocked
 up.  Impermeable graphite impregnated with fluorine containing
 polymer has been developed to make heat exchanger, the key equip-
 ment for recovering HN03 and HP.  A series of tests show that
 impermeable graphite is chemically stable  and has excellent beat
 conductivity, impermeability, strong oxidation resistance and
 non-fouling property.  It has been successfully used in several
 plants since 1974.  The development of this kind of material has
 opened up a new path for the development of corrosion resistant
 heat conducting material•

 Treatment of Residual Liquid
 Sulfuric acid in discharged residual liquid is neutralized in
 the reactor by iron, filings, while ferric  iom is  reduced to
 ferrous ion.  Iron filings  react intensely with 2<$ H2S04,
 therefore, in order to shorten, the reaction time, the residual
 liquid after discharging into the reactor  should  be diluted with
                                 289

-------
water, then neutralized by adding Iron filings*  In this  way,
pU of the solution can quickly Increase to 2 - 3.
Nickel salt reacts with sodium polysulfide at pH 2  - 3 to form
nickel sulfide precipitate.  Through filtration, the precipitate
is separated from ferrous sulfate solution*  The precipitate can
he oxidized and dissolved by dilute HN03*

The ferrous ion inclusions in nickel sulfide is oxidised  into
ferric ion, which forms Na2Peg(S04)4(OH)12 precipitate with
sodium carbonate added, precipitate is removed by  filtration,
pure nickel nitrate solution is obtained*- when its  pH is  about
5.  At the end of reaction pH is controlled up to 4.  Clean
nickel nitrate solution reacts with sodium carbonate  solution
under heating to form basic nickel carbonate precipitate, which
can be used as raw material for- nickel electrolysis after fil-
tration, washing and drying.

OPERATION PRACTICE
Measurement has been made for waste acid recovery project of a
steel plant after 3 years operation, recovery data  are shown
below.
Items
waste acid
residual liquid
regenerated
acid
sulfuric acid
recovery ratio
I*)
Volume
1140
350
858
150

Content
H+U)
2.58
13.47

32.4

P'(NJ
2.67
0*88
3.29

92.9
NO 3U)
1.70
0.23
2 J.2

93.9
Fe*(g/l.)
21.50
80.05



Nl*(g/l)
3.58
12.30



Cr*(g/l)
4.27
13.7



                               290

-------
 The recovery ratio of nickel la 85$.  Prom the calculation baaed on
 the composition of waste acid, 8 kg of basic nickel carbonate can
 be recovered per M3 of waste acid.

 The recovery of waste acid not only controls environmental pol-
 lution, but also leads to significant economic results. Regenerat-
 ing 1 M3 of ENOj and HP mixture results in 4 saving of 478.1 yuan*
 Details are shown below.
1 1 ems
Quantity
Value
( yuan J
Total
( yuan )
recovery
98$ HMOs
117 kg
58 .5
42fi HP
134 kg
509.2
567.7
consumption
<
electricity
50 KWH
4.2
water
300M3
10.8
•
steam
2 T
14.0
«
H2SOd
270 kg
54.0
•
wages
3.3 people
6*6 *
89.6
The capacity of the regenerating device put into operation in the
steel plant is 1 M3 of regenerated acid per shift.  The capital
cost of the device is 130,000 yuan, which can be paid off in one
year.

CONCLUSION
L. To recover HNOs and HP by one-step vacuum evaporation has the
  advantages of simple equipment, easy operation, high recovery
   ratio and significant economic result.
2. The process has solved environmental protection with no
   secondary pollution.
3. The residual liquid can be treated to recover nickel
   carbonate and ferrous sulfate *
                                ,291

-------
REFERENCES
1. Mixed Acid Pickling Waste "METAL FINISHING"  May,  1963,
3-
4-
                                       1964, NQ 9
                               PPM3C9)1972-
                              292

-------
                            STEEL  INDUSTRY PICKLING WASTE
                            AND  ITS  IlIPACT ON  ENVIRONMENT

                                  By Dr.  S.  Bhattacharyya

                                  IIT  Research  Institute
                                  Chicago,  Illinois  60616
ABSTRACT

     This study was  directed to develop Information
on ferrous sulfate hepcahydrate (copperas) produced
by the steel industry.  Several major aspects of
copperas generation  and disposal were studied, and
these are:
     1.  Pickling waste generation in integrated steel
Industry and by secondary processors, present practice,
and future directions.
     2.  Waste pickle liquor disposal technology and
future development.
     3.  West European pickling technology, centralized
zonal waste pickling liquor treatment concept and co-
treatment of different industrial waste stream.
     4.  Ferrous sulfate heptahydrate production and
end-use.
     The study shows that out of GO million tonnes
pickled, about one-third la pickled with sulfuric acid,
but except for very  small amounts going to the pigment
industry, most of the waste is disposed of without any
recovery of either free acid or dissolved iron units.
Host of the 75)000 tonnes/year of iron sulfate crys-
tals produced are from secondary industries while com-
mercial plants produce a similar quantity for pigment
production starting  from scrap and sulfuric acid.
While pigment production is more than one-half of  the
present market for copperas, potentially the largest
future market for copperas is likely to be sewage
treatment plants.
     The study also  shows that the rapid changeover
from H2SO/ to HC1 pickling has ended and the relative
price structure between the two acids may favor HjSO^
in the future.  However, all green-field plants are
likely to use HC1 pickling because the end product.
 if recycling is practiced,  is Fe203 which can be util-
 ized in the plant itself.  .
     Satisfactory copperas  production technology exists,
 and market expansion into water-sewage treatments may
 require a rapid expansion of  copperas production units
 —particularly, if regional centralized facilities
 are encouraged and established on a cooperative basis,
 each member paying for its  service proportional to
 the volume of acid waste treated and getting commen-
 surate credit for acid recovered.
     Several recommendations  have been made for addi-
 tional research and regulatory modifications to aid
 and enhance greater use of copperas.

 INTRODUCTION

     An estimated 60 million tonnes of steel products are
 pickled every year In  the U.S.A.  resulting  in a pickle
 liquor volume of about one billion gallons  per year.
 Of this volume, over"851 are  generated by 18 large
 steel corporations and the remaining 1ST at the sev-
 eral hundred smaller plants and batch processing units
 located all over the country.  Recycling of waste
 pickle liquor and its  disposal without harming the
 environment are causes of major concern to both the
 industry and EPA.   This study was performed to develop
 Information to minimize the impact of steel Industry
 ferrous sulfate heptahydrate waste on environment.
 Additional information on hydrochloric acid pickling
 technology and its bearing on sulfuric acid pickling
 has been considered where necessary.   The pickling
 practice and waste pickle liquor disposal In foreign
 steel plants has also  been reviewed.   Selected refer-
ences are cited and recommendations made for additional
 research and development.

THE MAGNITUDE OF PICKLING WASTE

Pickling  in the U.S. Steel Industry
     The  bulk of pickling waste is  generated by inte-
grated  steel plants located at 61 sites.1  In addition.
35 specialty steel companies  (48 locations), and 37
                                                 293

-------
scrap/sponge Iron-baaed steel plants  (55 locations),
are widely distributed In the U.S.2 A far larger number
(689) of small manufacturers and fabricators of steel
mill products are distributed over the country.3  These
smaller companies operate primarily via batch process-
es.   The total tonnage treated by them may be as
large as 10 to 15% of total national steel shipment.
     Growth In raw steel production and steel shipment
has shown significant fluctuations during the last 20
years.  Annual steel production and shipment data for
1956 to 1976 are given in Table I.*'5  Various projec-
tions have been made regarding steel consumption and
raw steel production by 1985 and beyond, based on long-
term trends.6'7  Based on 2Z growth rate, the projec-
tion Indicates a raw steel production of about 153 *
10° tonnes by 1985.  To produce 153 * 10* tonnes by
1985, the Installed capacity must increase by about
25 x 10° tonnes.  This additional capacity can be ob-
tained by expansion of present facilities and/or by
constructing new green-field plants.   In green-field
plants, the chances are that the most technologically
advanced pickling systems will be installed, very
likely based on HC1 with regeneration.  This will not
affect the present ferrous sulfate ptclcle liquor dis-
posal problem.   I£, however, a large fraction of the
25 * 10° tonnes additional capacity la added by expan-
sion of existing units (primarily because of capital
shortage), then some or all of the sulfurlc acid pick-
ling units already in use are likely to be retained
and extended, thereby significantly aggravating cur-
rent waste sulfuric acid pickle liquor disposal
problems.
     Table 1 shows that the ratio of steel pickled to
steel shipment ranged from 0.54 to 0.66 with a long-
term average of 0.63.  Unless the product mix changes
significantly during the next few years (due to mate-
rial substitution in automobile, container, and other
industries), a ratio of 0.63 may be expected to be
maintained until 1985.  Thus with 112 x io6 tonnes of
steel shipment by 1985, about 71 * 10° tonnes of steel
will require pickling.
     The steel pickling process generates wastes from
three distinct sources:  (a) waste pickle liquor from
acid baths, (b) acidulated rinse water from washing
acid drag-out from the pickling baths, and (c) acidi-
fied water generated in cleaning acid vapors and mists
from the pickling baths.  The waste pickle liquor is
relatively small in volume when compared with the other
two waste streams.  In l^SO^ pickling (with IS iron
loss), a typical pickle liquor contains 8% free acid
and 8% dissolved iron and is generated at a rate of
105 I/tonne of steel; with HC1, the corresponding fig-
ures are 1/2 to 1% free acid, 10% dissolved iron, and
50 I/tonne.
     Rinse waters pose a different kind of problem
because of the large  quantities needed.  Some large
continuous strip mills may use as much as 3.8 m-Vmln,
but most mills use about 1.5 m3/min.   Batch-type pick-
ling plants use less water than the continuous plants,
i.e., 0.1 to 1.1 m3/min.  Furae scrubbing water is
usually combined with rinse water for treatment.

Present Disposal Techniques for
Haste Pickle Liquor (WPL)
     At present, the 3.8 x 10° m3 of waste pickle liquor
is handled in many different ways, but the least com-
mon method Is recycling and recovery of acid and the
lost iron values.  A loss of II Iron value may amount
to 700,000 tonnes by 1985—almost one-third billion
dollars!  Similarly, the loss of free acid thrown away
amounts to many million dollars.
     Haste pickle liquor (WPL) is disposed by:
     1  Contract hauling.
     2  Deep will injection.
     3  Neutralization/oxidation and lagoonlng,
     4  Dumping In ocean/lake/river/alkaline beds.
     5  Co-treatment with municipal waste.

Contract Hauling.  Contract hauling avoids the immed-
iate problem but appears to be a relatively expensive
means of disposal, I.e., $36/m3.8  Several large steel
plants and oany batch processors use contract hauling.
Industrial waste disposal by contract hauling Is an
established Industry and may become a more significant
industry in the near future.9  However, additional
studies are needed in co-treatment of different wastes
to recover valuable metallies and reduce expenses.  Or,
the waste may be heated and converted to a form accept-
able for sewage conditioning in a municipal waste sys-
tem.  The complexity of treatment needed dictates the
total cost of hauling.  For example, a simple treat-
ment may be as low as $26/m3, a more complex process-
ing- $48/m3, and difficult wastes, such as mixed chlor-
inated hydrocarbons, as high as $79/m3.9   (All these
costs are based on ref. 9, escalated to reflect the
1970-1978 price trend.)

Deep Well Injection.  There are several hundred Injec-
tion wells in the U.S.,10'11  some accept pickle liquor
at the rate of as much as  5 ra3/min and are as deep as
3660 meters.  Only a  few of then are used  for disposal
of waste pickling liquor.   Before disposal, the pickle
liquor requires careful  filtration because otherwise
the well pores clog,  blocking infiltration.   Several
large steel plants use deep wells for disposal  of
waste pickle liquor.  Operating  costs vary up  to
$2.60/m3.12  This is  much  less expensive  than  contract
hauling.  However,  this  disposal method  Is being
phased out.

Neutralization/Oxidation and  Lagoonlng.  Neutralization
of  the highly acidic  WPL with lime,  soda ash,  or caus-
 tic  soda was practiced  for a  long  time.   The  treatment
 Increased  the pH to neutral,  and Iron  precipitates  as
a gelatinuous  Iron hydroxide  sludge which may not
settle  in  20 years'   Usually, large  lagoons were
 created,  e.g.,  WPL  from one million tonnes of steel
 on neutralization results  in  about  200,000 tonnes of
 sludge  requiring a  lagoon  one meter  deep spread over
an area  200 m by 1  km.   The cost for simple neutral-
 ization may range up  to $13/m3.12   In  addition to the
 simple  cost of  neutralization and  lagoonlng,  the cost
 of total  energy requirement In the  form of lime pro-
 duction must be considered.   It  requires about 2.78
 million kcal   to produce 1 tonne of quicklime (CaO),
 and the cost  of lime, which went fro* $20/tonne in
 1972 Co $3S/tonne In 1976,8 may increase significantly
 by 1905.
      This  simple lime neutralizatlon/lagooning process
 can,  however,  be modified  to  incorporate vacuum fil-
 tration and disposal  of a  wet cake,  62Z water, as
 practiced by H. H.  Robertson  Co.,  Arabridge,  Pennsyl-
 vania.13  This  waste disposal technique minimizes
 space requirements  (at the expense of additional  cap-
 ital and operating costs), but neither energy saving
 nor recovery of any metallic  or other product values
 is obtained.
      A DuPont process12 uses  controlled neutralization
 and oxidation with air to yield a mixture of magnetite
 (Fe30ii)  and CaS04,  which can be easily, dewatered.
 The CaSOi, may find a market as wallboard raw material.
 Bethlehem steel plant at Burns Harbor, Indiana, orig-
 inally used the DuPont neutralization processes to
 treat H^SO^ WFL.
                                                    294

-------
                                                   Table 1

                                  STATISTICAL DATA, UNITED STATES INDUSTRY*»5
Raw Steel,
Year 10 tonnes*
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1985

al tonne
1 ton
104.5
102.3
77.4
84.8
90.1
88.9
89.2
99.2
115.3
119.2
121.7
115.4
119.3
128.2
119.3
109.2
120.9
136.8
132.2
105.8
116.2
113.7
124.0
153

- 1000 kg
- 2000 Ib
Steel
Shipment,
10 tonnes
75.6
72.5
54.4
... 6.3.0
64.6
60.0
64.1
68.6
77.0
84.2
81.7
76.1
83.4
85.2
82.4
78.9
83.3
101.1
99.4
72.6
81.1
82.7
87.1
112

• 1.102 ton

Steel
Shipment/
Raw Steel
0.72
0.71
0.70
0.74
0.72
0.67
0.72
0.69
0.67
0.70
0.«7
0.66
0.70
0.66
0.69
0.72
0.69
0.74
0.75
0.69
0.70
0.73
0.70
0.73



Steel
Fielded,
(Estimates)
10 tonnes8
45
39
35
41
43
39
42
45
49
53
50
47
53
54
51
51
55
67
62
43
(51)b
(52)
(55)
(71)



Steel
Pickled/
Steel
Shipment
0.59
0.54
0.65
0.64
0.66
0.65
0.66
0.65
0.63
0.62
0.61
0.62
0.64
0.63
0.62
0.64
0.66
0.66
0.63
0.60
(0.63)
(0.63)
(0.63)
(0.63)



                       bFigures  in parentheses are estimate* based on 63Z of all ship-
                        ment being pickled.
      Interlake  Steel uses  an  Interlake-DuPont modifi-
cation process  at  its plant in  Chicago, Illinois,
which produces  magnetite and  calcium chloride from
HC1 WPL.  The CaCl2 solution  is treated with H2S04 to
regenerate HC112 and to precipitate CaSC>4 which may
be utilized.  In both the  DuPont and modified neutral-
ization processes, iron values  are recovered.  The
70304, can be pelletized and  used as a blast furnace
feed, or it can be converted  and used in ferrites.
magnetic tape,  pigment, and other industries.l*t*»

Dumping in Ocean/Lake/River/Alkaline Beds.  'If WPL
is diluted  sufficiently,  then  it can be tolerated'
Is the principle governing dumping in large bodies of
water such as ocean and river.   Bethlehem's Lacka-
waraia and Sparrow's Point plants had permits to dis-
charge WPL in Lake Erie, Hew York, and Patapsco River.
Maryland, respectively.  Many states and local author-
ities had given permits to steel plants in the past..
With more stringent regulations, the steel Industry
is actively examining alternatives to dumping in pub-
lic rivers and lakes.
     All steel plants have large slag dumps, and the
WPL can be dumped on then.  The alkaline slag neutral-
izes a part of the free acid,  but the dissolved  iron
Is not converted to a stable form.  Runout from  these
 dumps would contain appreciable concentration of iron
 salt* and  have  to be treated by the storm water drain-
 age system, which usually discharges either  to a river/
 lake/ocean or to the local municipal system.   If hot
 slag is  quenched,  acid  pollutants  are transferred to
 the air  and there is more widespread cause for con-
 cern.

 Treatment  of Municipal  Waste Water.   Tertiary water
 treatment  of municipal  waste water for removal of
 phosphates 1* opening up  new markets for  iron and
 aluminum salts.   From 10  plants  in 1968,  the  number
    tfECUry treatment  plants increased to 445 in
 1972io and may have  Increased significantly since
 then.  The common chemicals  that can be used  to react
with phosphates and  to  remove then as salts are  ferric
and ferrous chloride, ferric  and ferrous  sulfate,
aluminum sulfate  (alum),  sodium aluminate, lime, and
WPL  (FeCl2  and FeS04).  Currently  lime, ferric chlor-
ide, and alum are tha main chemicals used.  Lime has
run into worker opposition at sewage plants.    It is
also a highly energy-intensive product.   Between Al
and Fe salts, Iron appear* to be preferred.
     A report by KSF Chemical Processes, Ltd., Cam-
bridge,  Ontario, Canada5 shows that about 1 tonne of
heptahydrate is required to treat 3800 n3 of  sewage.
                                                 295

-------
Sewage treatment plants  around  the Great Lakes area
alone treat approximately 76 million m3/day  and  may
require as much as 20,000 tonnes of heptahydrate per
day.  The  steel  plants in this area, If all plants
use sulfuric acids, can only produce about 2000-4000
tonnes of heptahydrate per day and will be Inadequate
to fill the demand.

ACID REGENERATION AND/OR IRON RECOVERY-
STATE OF ART, U.S.A. AND FOREIGN

Hydrochloric versus Sulfuric Acid Pickling
     In an  EPA  Technology Transfer Capsule Report,
the economics of the different competitive processes
have been carefully analyzed for sulfuric acid pickle
wastes.  On the  basis  of certain assumptions, it was
estimated  that  for plants with  pickling capacity of
45.000 tonnes/yror more. HjSO^ recovery is economical
even when  no credit  is given  for the heptahydrate
crystals.
     Information is, however, quite fragmentary in the
open literature regarding the steel industry pickling
capacities in terms of 112804 vs.  HC1 units with regen-
eration, and the type of regeneration practiced.  Pri-
vate estimates' Indicate that the number of HC1 lines
(about 60)  exceeds  that of  H^SO^ lines  (about  30).
On the other hand, HC1  acid consumption Is less  than
HjSOi acid use.  This IB so because about 19 kg H2SO^
(IOOX) is required to pickle one tonne of steel result-
Ing  In  about 105 liters  of WPL.  The corresponding
quantities for HC1 are about 6.5  kg of 100Z acid/tonne
of steel resulting in about 50 liters of WPL.  It has
been estimated5 that out of about 60 million tonnes of
steel pickled in 1974, about two-thirds were pickled by
HC1 and one-third by 112804.  HC1 and HjSO^ acid costs
increased  during  1966 to  1976 from $48 to $67/tonne
and $10 to $32/tonne,  respectively.17  The rapid  con-
version from HjSO^ to HC1 pickling which occurred  dur-
ing the 1960's has stopped, and  a  relatively cheaper
({2804 may  make it more attractive for pickling.   SO2
emission control regulations may generate a large sur-
plus of S in the 1980's.l8

Hydrochloric Acid Regeneration Processes,
U.S.A. and Canada
     HC1 regeneration is known to be practiced in  only
eight plants in the U.S.A. and Canada.11  Spray roast-
Ing, known as the Woodall-Duckham-Ruthner process, and
fluldized bed roasting, known  as Keramchemie-Lurgi, are
well established in the U.S.A. and  foreign countries,
and are described fully In the literature.  The result-
ing products  are regenerated  HC1 and  granular  iron
oxide,  which  is usually  used as  a sinter  feed for
blast furnace.
     The sliding/vibrating  roaster process,  the ETI
Proceas11 (Environmental Technology,  Inc.),  uses  a
unique roaster reactor design combining a spray roaster
principle with a vibrating moving grate for iron oxide
removal.  The oxide particle size Is larger than  that
produced by both spray and fluldized roasting processes,
and the dust removal problem is minimized.
     The low-temperature PORI*' chemical process  is
completely different from the high-temperature process-
es.  It consists of four major  operational stations:
evaporator, oxldizer,  hydrolyzer, and a falling  film
condenser-absorber system.   At  the  evaporator,   the
FeCl2 concentration is raised from a typical 20Z to  36Z.
In the oxldizer,  Fed2  is  oxidized to FeClj and Fe20*
followed by FeCl3 hydrolysis to BC1  and more FejOj.
The F«20j is dense and can be recycled to steel produc-
tion.  The HC1 produced  Is  30Z concentrated with 99Z
yield and is recycled to the process.
Sulfuric Acid Regeneration Processes,
U.S.A. and Canada	
     All the regeneration systems recover the free acid
by removing the iron salts from the WPL.  There are 32
plants in operation in the U.S.A. and Canada, as shown
in Tables 2 and 3.  The  two  basic processes are  (1)
heating  type,  producing FeSO^ monohydrate (5 plants),
and  (2)  cooling  type,  producing FeS04 heptahydrate
(27 plants).
     In  the  continuous Sulfex process,11 a submerged
zirconium  heat  exchanger heats WPL to 125 C with 690
KPa steam; the  acid concentration reaches 45Z to 50J,
and FeSO^ • H20 precipitates out. After crystal separa-
tion,  the  recovered acid  containing  about  IjZ Fe is
returned for pickling.
     The continuous  Pureco  process1'1  uses submerged
combustion heaters to heat WPL to 95 C and concentrate
the  acid to  35Z.   The  iron precipitates  out as
FeSO^'HjO,  and the recovered acid containing %Z Fe is
recycled.
     In the cooling type continuous Keramchemie pro-
cess j11'20 the WPL is precooled and then flows to  a
crystallizer  where  the temperature is decreased to
5 C  using  high-pressure steam  ejectors.  The acid/
crystal slurry is concentrated and centrifuged.  The
recovered  acid  containing 2-3Z Fe is recycled.  The
large  flow of ejector condensing water  Is  slightly
contaminated with entrained WPL.
     In the batch type  Crown Chemical process,  the
WPL  is chilled through  a  Freon refrigeration system
to about 0 C in about 8 hr.  The heptahydrate Is con-
centrated In settling tanks and filters.
     The24KSF Process units11 are modular batch type.
The  WPL  is cooled with  chilled reclrculatlng water
circuit.   The solution  is  cooled to 7 to IOC, the
heptahydrate  crystals  are discharged  by gravity,
washed, partially dried, and discharged.  The recov-
ered acid Is reheated and recycled at a concentration
of  25Z containing 2 to 3Z Fe.  The  process is "zero
discharge" incorporating full  recycling of all acid
mists, spent acid, and pickling rinse waters.
     The planning for  additional pickling units  in
steel  plants  is  not  readily available  In  the open
literature.  Private Information Indicates that Beth-
lehem  (Sparros Point, Maryland), Jones and  Laugnlin
(Allqulppa, Pennsylvania) and U.S. Steel (Falrfield,
Alabama) may  use HjS04 pickling  more  significantly
during the next five years.  Other steel plants,  such
as Armco (Ashland, Kentucky), National (Weirton,  West
Virginia), Jones and  Laugnlin (Cleveland, Ohio),  and
Republic (Gadsden, Alabama) may significantly enlarge
HC1 pickling practice with  regeneration.  It  is  ex-
pected that  during 1980-85  the  addition of 25 * 106
tonnes steel capacity will require establishment of at
least one (U.S. Steel-Conneaut), and perhaps two green-
field plants in addition to the above and several more
unannounced rounding-off facilities requiring pickling
plant additions.

European Steel Plants
     Western Europe raw steel production was 154  mil-
lion tonnes In 1977.  On an equivalent tonnage basis,
WPL production In European steel plsnts is slightly
less because of the difference In product mix.  In the
U.S.A., the proportion of flat products In the product
mix is relatively more than In Europe, and  WPL gen-
eration Is maximum in flat product pickling.  In ad-
dition to the integrated steel plants, European sec-
ondary and tertiary fabricators and processors us*
similar batch pickling processes and have similar prob-
lems, if not more so because of less disposal elas-
ticity in land surface and water bodies. On the other
                                                296

-------
                       Table 2

     TYPES OF SULFURIC ACID REGENERATION PLANTS..
   AMD THEIR LOCATIONS IN THE U.S.A. AND CANADA
                                                    Table 3

                               LIST OF KSF PICKLING AND ACID RECOVERY SYSTEMS
                                                 11
          Type
By Heating (monohydrate)

   Rulfex Process
   (Sulfex Corporation,
    Maple Heights, Ohio)

   Pureco Process
   (Pureco Systems, Inc.
    Mt. Prospect, 111.)
By Cooling (heptahydrate)

   A.  Vacuum Type

   Keramchemle Process
   B.  Refrigeration Type

   Crown Process
   (Crown Chemical Co.,
    Inc., Indianapolis,
    Ind.)

   KSF Process
   (KSF Chemical Proces-
    ses, Ltd., Cambridge,
    Ontario, Canada)
                                   Plants
Metal Processing Co.
  Maple Heights, Ohio
Joselyn Manufacturing &
Supply Co.
  Chicago, 111.

Empire Galvanizing
  Dlv. of Joslyn Steel
  Chicago, 111.

H. H. Howard Company
  Chicago, 111.

Budd Company
  Frankfort, Ohio
Fltzslmons Steel Company
  Youngstown, Ohio

Slvaco-Ingersoll, Ltd.
  Ingersoll, Ontario,
  Canada

Sldbec-Dosco, Ltd.
  Rexdale, Ontario,
  Canada
Laclede Steel Company
  Alton, 111.
24 Plants (See Table 3
           for locations)
Wlmco  Steel Sales,  Ltd.
1430 Martingrove Road
Rexdale,  Ontario

P.  L.  Robertson Manu-
facturing Co.,  Ltd.
Bronte Street
Milton, Ontario

Atlantic  Wire Company
One Church Street
Branford, Connecticut

Motor  Wheel Industries
(Chatham) Ltd.
650. River view Drive
Chatham,  Ontario

Macwhyte  Company
2906 14th Avenue
Kenosha,  Wisconsin

Electric  Wheel  Company
1120 North 28th Street
Quincy. Illinois

Firestone Street Prod-
ucts of Canada  Ltd.
31  Firestone Blvd.
London, Ontario

National  Fence  Company
Bladenaburg, Maryland

Russell,  Burdaall &
Hard Company
Rock Falls, 111.

Nelsen Steel &  Wire Co.
9400 West  Belmont
Franklin  Park,  111.

Thompson  Steel  Co., Inc.
9470 King  Street
Franklin Park,  111.

Republic Hire Corp.
500 Blair  Road
Carteret, New Jersey

Everlock Division
(Mlcrodot  Inc.)
Detroit, Michigan
 Metal Products Division
 Goodyear Tire & Rubber Co.
 1144 East Market Street
 Akron, Ohio

 Motor Wheel Corporation
 1600 North Larch Street
 Lansing, Michigan

 Motor Wheel Corporation
 Ogletown Road
 Newark, Delaware

 Igoe Brothers, Inc.
 234 Poinier Street
 Newark, New Jersey

 Walker Hire & Steel Co.
 660 East 10 Mile Road
 Ferndale, Michigan

 New York Wire Mills Corp.
 3937 River Road
 Tonawanda, New York

 Bethlehem Steel Corp.
 Lackawanna, New York

 American Chain & Cable
   Co.,  Inc.
 American Chain Div.
   (Office)
 East Princess  Street
 York, Pennsylvania

 Atlantic Steel Company
 16th &  Holly Streets
 Atlanta,  Georgia

 Krueger & Company, Inc.
 900 Industrial  Drive
 Elmhurst, 111.

 Slvaco Wire: & Nail Co.
 800 Quellette Street
 Marlevllle, Quebec

Boric Kldric Steelworks
Niksic
Yugoslavia
                                                   297

-------
hand,  the close proximity  of many secondary plants
allows  development of  central treatment facilities
with economy-ln-size  benefits which  their  counter-
parts in  the U.S.A. nay not enjoy. At Altena in West
Germany,  an  HC1 acid recovery plant is owned  and
operated  Jointly by 30 smaller companies drawing wire
and/or  galvanizing.   Tankers collect  spent  liquors
and return recovered  acid to  the works. The oxides
recovered are sent to the  steelmakers.  The companies
so served by this central  recovery plant are situated
within an  80 km (50 mile) radius.21
     A summary of the pickling processes practiced in
the  different  European plants  is given in Table 4.
Two  steel plants have  only sulfuric acid pickling,
four have only hydrochloric acid,  and two have both.
One  plant is changing  over  from l^SO^ to RC1, and
another is planning to change.
     Three sulfuric acid pickling lines recover  hep-
tahydrate and use it Cor water treatment, in the sin-
tering plant for conversion to oxide, as animal feed
additive, and also for discharge on a slag dump.
     Of the  six plants  using HC1 pickling, five re-
generate  the acid and produce Fe2C>3 as a by-product.
Heating for  regeneration is  done using either oil,
natural gas,  or coke oven gas.  Four of the five re-
generation units use the spray roasting technique and
obtain  a  very  fine  reddish-brown Fe203.   One unit
uses a fluidized  bed  roaster and  generates a much
coarser gray iron oxide product.
     Most of the Fe203 is used either in the ferrite
industry  or at the slnter-pelletlzing  plant.   There
seems to be no significant problem in Feo03 disposal.
     The  steel plants consider HC1  pickling to be a
better process  than  HjS04 pickling.    It  is faster,
gives a better looking pickled surface, and the acid
can be regenerated (99*) with a by-product which  is
eminently salable  and  usable.  The process requires
energy, and if  oil and  natural gas  have to be used
Instead of the coke oven gas within the steel plants,
then It will also be more expensive In the near future.
The process also has some maintenance problems.
     If the  H-^SOtt pickling line  has  a  substantial
scale breaker, then H2S04  pickling may become equally
as  fast   as HC1  pickling,  as  opined  by  one steel
plckler.   While HC1 directly dissolves FeO (scale),
HoS04  operates  better  if it  can readily get below
the scale and attack the metal surface and there lies
the significance of scale breaking before pickling.
     HC1 pickling gives a brighter sheet.   However,
if  the  pickled products  are  to  be eventually cold
rolled (as most are) the  initial relative brightness
of HC1  pickled sheet has no  special advantage over
HoS04 pickled sheets before cold  rolling because of
their brightness alone.
     On the other hand,  it was noted  that FeS04-7H20
has a limited market.   It contains 2 to 32 moisture
and has some storage and dumping problems.   At Stora
Kopparberg plant, the heptahydrate crystals were stored
in the open, inside covered steel bins.  Crystalliza-
tion of heptahydrate requires cooling, and additional
energy is required for refrigeration.

Stora Kopparberg.  Domnarvet Steel Works,  Borla'nge.
Sweden.  In addition to treating WPL to obtain hepta-
hydrate crystals,  the wash  water,  vapor  scrubbing
water, and excess  steam  condensate are treated in a
neutralizing plant at a rate of about 20 m-Vhr.  After
treatment iron is reduced to  <2 ppm,  TSS is reduced
from 14 to <0.1  ppm, pR  adjusted  to  7.5 ±  1 with no
unneutralized free acid In the waste.   The sludge cake
contains about 352 solid.  The neutralized water is fed
to 4 sedimentation tanks at the rate of 0.22m3/m2 • hr
with a residence time  of  9 hr.  Flocculent is added
at  the last tank.
     The total production of 7000 tonnes/year  of  hep-
tahydrate  is  completely  utilized.   A Swedish firm
Imports  15,000  tonnes/year of  heptahydrate for water
treatment.

Arbed  Steel  Works. Differedatige. Luxembourg.   This
100,000 tonnes/year continuous spiral  sheet  pickling
plant was designed by Sundwig, of  Germany, with the
regeneration units designed by Falker, a Swiss company.
About 1000 tonnes/year of heptahydrate crystals arc pro-
duced by the refrigeration/crystallization/centrifuge
technique. The crystals have only 1.12 water and 0.6Z
free acid  and  are primarily used for water treatment
and algae growth prevention.  Because of low require-
ments, it is also dumped in the slag yard.
     The wash water is treated with milk of line to pH
8 to 9, settled, decanted, and discharged to a river,
and the sludge is deposited on the slag dump.
     A new HC1 pickling unit with regeneration was due
to begin operation March 1979. The pickling unit was
designed by Benguin (France) having a 25,000 tonnes/
month capacity; the regeneration unit, using natural
gas, follows the Keramchemle-Lurgl system.

Veeet-Alplne Aktiengesellsehaft. Lint. Austria.   Ho. 1
cold rolling mill uses H,S04 pickling,  and No. 2 cold
rolling mill uses BCl pickling.  The^SO^ unit process-
es 600,000 tonnes/year and generates heptahydrate amount-
ing  to 12,000 tonnes/year using the  Lurgl  process.
A small cut sheet plant in the Ho. 1 mill also produces
1200 tonnes/year of heptahydrate crystals. Most (902)
of the heptahydrate crystals go to the sinter plant and
the rest for water treatment and as animal feed additive.
     The No. 2  CR  mill unit  using  HC1  regenerates
acid  with  the  Ruthner design.   The mill  processes
800,000 tonnes/year and  generates 3,600 tonnes/year
of Fe203 for the  ferrite industry.  Wash water from
Ho. 2 CR mill Is neutralized.   Sludge is transported
by dump car for disposal,  and the decanted water  is
run off to the Danube River.  Wash  water from No. 1
CR mill (H2S04) is discharged to the Danube dIreetly.

EHSIDESA. Aviles.  Spain.   Of the two pickling lines,
the H2S04 line was Installed in 1964,  and the HC1 in
1972.  They are considering changing the H2S04 unit
to HC1.  There la  no acid recovery or neutralization
plant.  Between the two mills, the plant pickles 1.5
million tonnes/year. The WPL is  mixed with wash water
and fume scrubber  waste and  diluted to less than 12
acid concentration, conveyed by canal,  and discharged
to the ocean.
     The plant has its own reservoir fed  from the river
and uses three water qualities in the plant.  For all
plant use,  th* industrial quality water is taken from
the reservoir.  There are filtering stations attached
to each plant unit  which  treat  the water and recycle
some  of  it.  Then  the third  system  Is the drinking
water.

Hoogovens Ijmuiden BV. IJmuiden. Holland.  This large
steel plant has a capacity of 6 million tonnes/year.
It now produces 5 million tonnes/year and la planning
to  expand to 11 million tonnes/year.   Only HC1 pickling
is used—two lines without regeneration and one  line
with  a Lurgl regeneration  plant.   They had an H2SO^
line  but  changed over to HC1  without regeneration.
The waste acid and wash water is discharged to the sea
for a fee paid to the government.
                                                  298

-------
                                           Table 4

                  WESTERN EUROPEAN STEEL PLANTS AND THEIR PICKLING PROCESSES
         Steel Plant
                                                Pickling Process
3.
5.
6.
     Sulfurlc Acid Process

     Stoca Kopparberg
     Domnarvet Steelworks,
     Borlange, Sweden
     Voest-Alpine
       Aktlengesellichaft
     Llnz,  Austria
     ENSIDESA,
     Avlles,  Spain
    ARBED Steel Works,
    Differdange,
    Luxembourg
    Hydrochloric Acid

    Uddeholms Aktlebolag,
    Munkfors, Sweden
Fried, Krupp
  Kiittenwerke AC,
Bochum, W. Germany

Voest-Alpine
  Aktlengesellschaft
Linz, Austria

ENSIDESA,
Aviles, Spain
Hoogovens Ijmuiden BV,
Ijmulden, Holland
    British Steel Corp.
    Ebbw Vale Works
    Ebbv Vale, U.K.
                         One continuous pickling  line with FeSO^ • 7H20 as a by-product  and
                         H2S04 recycling:   Pickling capacity,  350,000 t/year.   FeS04*7H20
                         production in 1977—7000 tonnes.  Price  about $24/tonne.  One-half
                         of  ferrous sulfate sold  to water  treatment  plant and as flocculat-
                         ing agent.  Other half sold to  another company which burns off
                         sulfur and mixes  the oxide with fine  ore and pelletizes/sintmrs.
                         Snail quantities  used in concrete and to kill weeds.

                         One continuous pickling  line.   Capacity—600,000 t/year.  PeSO^-
                         7HoO—12,000 t/year,  90Z to sinter plant, 10Z for water cleaning
                         and animal feed.   One small cut sheet batch pickling unit.  FeSOi-
                         7H20—1,200 t/year.

                         On* continuous pickling  line.   Capacity—650,000 t/year.  Waste
                         pickle liquor  (9-10Z  acid)  diluted to less  than IZ, discharged to
                         sea via canal.  Planning to  install RC1  line.

                         One continuous  spiral pickling  line.   Capacity  about 100,000  t/year.
                         FeS04-7H20—1,000  t/year, used  for water treatment, also dumped in
                         slag dump.   New HC1 line with regeneration under  construction
                         (March 1979).
 One continuous  pickling  line with acid  regeneration.   1978 produc-
 tion,  20,000-25,000  t/year.  HC1 recovery,  97-99Z.  Oil consumption,
 35  kg/h.   FejOj by-product, less than 1,000 t/year.  Used in sinter.
 Some stored.

 One continuous  pickling  line with acid  regeneration.  Capacity	
 1,200,000  t/year.  HC1 recovery about 98t.   Fe203 by-product,
 6,000  t/year, used in sintering plant.

 One continuous  pickling  line with acid  regeneration.  Capacity—
 800,000 t/year.  Fe203 production—3,600 t/year, used in the f err ite
 industry.

 One continuous  pickling  line without acid regeneration.  Capacity—
 850,000 t/year.  Acid waste diluted with wash water and fume scrubber
 water  to less than lit and discharged into the sea.

 Three  continuous pickling lines; two older lines without regeneration,
 and  one with regeneration.  Now, as much as possible,  waste liquor
 from older lines  is  regenerated in the double unit regeneration
 plant.  Maximum capacity—1,500,000 t/year.  Current capacity—
 1,100,000 t/year.  Fe203 production about 6,000 t/year. Sold to fer-
rlte industry.  Quality very iuportant.  Bayer oxide—very good
quality—sells for $500/tonne.

One continuous pickling line with regeneration.   Capacity—1,400,000
t/year.  ?e203 production about 12,000 t/year.  Fluidlzed bed  regen-
eration unit produced gray, coarse 100Z >l/2 m spherical particles
well suited to sintering.
                                         299

-------
     The Ho. 2 CR mill was built in 1971. The pickling
line was designed by Wean-Dameron and the regeneration
unit by  Ruthner.  Pickling about 1.1 million tonnes/
year, the regeneration plant produces about 6000 tonnes/
year Fe2<)3 from the WPL.  The  Fe2<>3 is used in the fer-
rite Industry where the quality is  very Important.
Chloride should be less than 0.IX; density (after set-
tling)  0.8 g/cm3; specific surface,  270 to  320 m2/g,
loss on Ignition, 0.31 at 800 C.  A very high quality
oxide,   such  as  that  produced  by Bayer,  sells  for
$500/tonne.
     The wash and fume scrubber water is  partially
used for  LD  (EOF)  gas cleaning.  The  rest of it is
mixed with NaOH and discharged 1 to  2 km out Into the
sea. For the month of May 1978,  the  range of flow was
40 to 580 m3/hr,  well under the  limit of 1000 m3/hr.
The pH ranged 6.8 to 8.5, TSS 16 ppm avg  (5-375 ppm,
range), under the 30 ppm limit.   Iron limit la 3 ppm,
and soluble iron average was 0.34 ppm.
     Government regulations limit waste acid discharge
to the ocean  to a rate of 170m3/day (30 gpm), amount-
ing to 22 tonnes of  Fe++/day.  Other limitations are
2.9 tonnes HCl/day and 7 kg Cu/day.  In  addition, a
penalty equivalent to $1.30/ton of F*4' is charged when
the limit is  exceeded.

British Steel Corporation, Ebbw Vale Works,  Ebbw Vale,
U.K.  The three K^SO^ pickling lines were shut down In
1974 and replaced by one continuous  HC1 line to treat
about 1.4 million tonnes/year and  to regenerate acid
In a Lurgi-deslgned fluldlzed bed roaster.  The  pick-
ling  line is designed by Head Urlghtson. Fej03 produc-
tion amounts  to 12,000 tonnes/year, and the  coarse
(>)j mm) gray particles  are  excellent as sinter feed.
Loss of pickling is about 0.6%.   The WPL  contained 3
to 6.5% HC1 and 90 to 122 g/1 iron.   Rinse water con-
tained HC1  20 g/1 and Fe 10 g/1.  The  Fe2<33  had  a
0.075/5 chloride content.  All wash waters are properly
treated, and  the  effluent  discharged into  the Eddw
River  goes  through  the Bristol Channel  to the sea,
32 km (20 miles)  away.

Fried.  Krupp  Hiittenwerke AC, Bochum,  West Germany.
The plant has  two HC1 regeneration units; the first one
designed by  Dr.  C. Otto  is  not giving  satisfactory
operation now.  The second one designed by Keramchemie
Is mostly  used  to  treat WPL at the rate of 3000. I/hr
(13 gpm).  It  is fired  by  coke oven gas at the rate
of 1400 m3/hr.  ?e2®3 produced at 6000 tonnes/year is
used in the sintering plant. The wash water is  neutral-
ized with  lime water  (pH 7 to 8), stirred with air,
sent to  settling  tanks,  and then  filtered  and  the
cake is dumped.  The effluent is discharged at the rate
of 20 to 40 m3/hr.    New water purchased for pickling
costs DM 0.95/m3 ($2.00/1000 gal).

Ruhrverband. Essen. West Germany; Bmschergenossenschaf t.
Essen,  West Germany.  In the Ruhr district In Vest Ger-
many, where Industry Is highly concentrated, separation
of municipal and Industrial waste  waters is encouraged.
Central treatment plants for each waste stream permits
special water treatment techniques with the additional
prospect of metal recovery and recycle.  All plants in
this area are members of the Ruhrverband and  pay for
water use.   The Ruhr is a drinking water river.  Some
100 water treatment plants are  located in this area.
     In addition to steel plants, there are many coal
mines which generate a lot of coal fines entrained in
wash water amounting to 800-1000 tonnes/day of coal.
These coal fines mixed  with  steel plant  wastes and
biological wastes  combine to form a  sludge .having a
combined carbon of 40Z, which la used in power plants
equivalent to brown coal.  In a recent U.S.  EPA trip
report,22 Craig  has mentioned the  concept of  waste
exchange which  has been successfully implemented In
Europe.
     Emschergenossenachaft  (Emscher  Association) was
created In 1904 to control all water discharges in the
Emscher River  basin.  Recently,   In  order to  reduce
the load on the  Rhine,  a large-scale treatment plant
was built at Emscher, 7.5 km from the Rhine confluence.
It serves a population of  about  2.5  million and can
handle a  mmflimin flow of 30 m3/sec.  The sludge pro-
duced  here  la burned In the  Karnap Power Station of
the KWE AC.
     Centralized treatment plants for Industrial wastes,
both publicly and  privately owned, treat the  wastes
of their member companies on a fee basis.  The Altena
Central HC1 regeneration plant  (30 members)  recovers
BC1 from  WPL and  sells the  FejOo to  use as a land
fill.22

Ferrous Sulfate Heptahydrate
(Copperas). Production and End-Use
     The 33 plants  recovering  copperas  are  given in
Tables 2  and 3.  Only  one  of  them,  Bethlehem's
Lackawanna  plant,  Is  an Integrated steel plant.  It
Is possible that they are using a modified KSF process
with  5 modular  units  each capable of  processing 19
1/min.  In all, there  are about 32 modular units pro-
cessing 570 1/min of  WPL  by the KSF process.  On an
average,about 105 liters of WPL is formed on treating
1 tonne of steel resulting in 30 kg of copperas crys-
tals  at  0.6% iron  loss.   With  a  total  processing
capacity of  570  1/min, copperas production rate from
all KSF units will amount to about 160 kg/mln or about
10 tonnes/hr.  The actual annual production will  de-
pend  on  the number of shifts operating.  Assuming a
standard 2-shift operation of 16 hr/day and 330 flays/
year, the total copperas production from all KSF units
can be estimated at about 55,000 tonnes/year.
     The  three  plants  using  vacuum crystallizing
Keramchemie  process  have a combined capacity of 114
1/min of WPL and may produce as much as 10,000 tonnes/
year  of  heptahydrate  on the basis  of continuous
3-shift operation.
     The batch-type Crown Chemical process  treats 42
1/mln of WPL and on a 2-shift/day, 330 days/year may
produce about  4000 tonnes/year of  the heptahydrate.
     The total  FeSO^-7H20 (copperas)  production from
the 28 plants at  present  may  amount to  as much as
66,000 tonnes/year, if fully operated.
     The four Pureco process plants treat a  total of
80 1/mln of WPL and produce about 10 kg/mln of FeSO^«
H2o—the monohydrate.  The Pureco process is continu-
ous,  and  if operated  on the basis of 20 shifts/week
and 50 weeks/year for a total of 8000 hr, it will pro-
duce about 4800 tonnes/year.
     The single  Sulfex process unit  treats 35 1/min
of WPL and, on 8000 hr continuous basis, is  capable
of producing 2300 tonnes/year of FeSO^HjO.
     The total  FeSO^-HjO (monohydrate) produced from
the five units may  amount to 7000 tonnes/year.  Thus,
a total of about 75,000 tonnes/year of ferrous sulfate
crystals (copperas 68,000 tonnes and monohydrate 7000
tonnes) are available in the market for use or disposal.
     Approximately three times more  HjSO* Is needed
thanHCl to pickle one tonne of steel on a 100X basis.
On this basis, in 1974:the estimated 30 H2S04 pickling
lines in the steel  Industry treated  about one-third
of the total estimated  60 million tonnes  pickled,
or about 20 million tonnes.   At  0.6Z iron loss, this
                                                  300

-------
 amount of steel treated will  generate about  600,000
 tonnes/year of  heptahydrate (60 Ib/ton of  steel)  if
 fully recovered.  To this  amount, about 12 to 15% may
 be added  if recovered fully  from the  secondary and
 tertiary processing plants spread all over the U.S.A.
 Thus  in  1974,  the  total  potential  for  heptahydrate
 production could have been estimated  to be as high  as
 630,000 tonnes/year. Using the 2% growth rate projec-
 tions, by  1985, the  amount of steel pickled may in-
 crease to 72 million tonnes/year.  If the  additional
 pickling capacities added  maintain the same ratio be-
 tween H2S04 and HC1 pickling tonnage  (1:2),  then the
 total potential heptahydrate generation may  Increase
 to about 800,000 tonnes/year by 1985, about  ten times
 the amount that  is currently processed.   That is the
 potential magnitude of copperas that must  be either
 safely dumped,  or  normally  absorbed  in the  marketing
 process which  is preferred.
      Besides steel  plant  pickling,  there are  other
 industrial processes which  generate  ferrous  sulfate
 heptahydrate.   These are (1)  commercial producers who
 deliberately dissolve scrap  in  sulfuric acid  to pro-
 duce the salt,  (2)  a by-product from  titanium  pigment
 producing units,  and  (3)  from sulfuric acid leaching
 of copper-bearing minerals and slags.  The five major
 commercial producers have  a  total copperas  capacity
 of 320,000 tonnes but  produced only  about   180,000
 tonnes in 1967  and  were projected to produce  as much
 as 270,000 tonnes by 1975.*3  The two largest  producers
 are National Lead  Company  and Pfizer, Inc.,  with a
 total installed  capacity of 250,000 tonnes/year.   Thus,
 there exists  a  large  surplus  commercial capacity for
 production of copperas.  The average  price of  commer-
 cially produced  copperas remained about $5 to $6/tonne
 during 1960 to  1967.23
      Ferrous sulfate heptahydrate has been used  for a
 long  time for production of  synthetic iron oxide pig-
 ment,  copperas red.  In a  2-step Keating process, the
 sulfurous  off-gas can  he-used to generate  HjSO^.  A
 wet chemical process  by  Penniman and Zoph is  also
 used  to produce  synthetic  iron oxide from heptahydrate
 solution.
      Iron  oxides used  for  pigments  and  electronics
 can be produced  as a  by-product from  the titanium
 dioxide pigment  industry.1*  The sulfate-process  por-
 tion  of the titanium pigment industry has  so far  been
 a  key source of copperas.   A  shift  in  titanium  pigment
 technology  from  sulfate to  chloride  process will de-
 crease copperas production, but  iron oxide may be
 available as  a by-product  if  FeCl2  is converted to
 Fe2°3'  Apparent domestic demand  of iron oxide pigments
 have  seen  a  steady growth  of  45% between  1964 and
 1974.  The compounded growth  rate of 3.6%, if projected
 to 1985,  indicates  a total  domestic  demand exceeding
 250,000 tonnes valued at $150 million  (U.S. $0.30/lb).
      The market  share of copperas from all sources has
 been  estimated at 55X, iron  oxide  pigments; 30X,  fer-
 rites;  5%, water-sewage treatment;   and  101,  others
 including  fertilizers,  feed  stock,   ink, etc.  While
 innumerable end uses  are  available, most of them are
 small.  Two major consumers are paint  and plastics.
 Sales  of paints  and allied products  increased from
 $2.4 billion  to  $3.6 billion  from 1963 to  1971.  A
 $7  billion market is projected in 1980.2*  The plastic
 resin  market doubled  during the  same period and is
 projected to reach $12 billion by 1980.24
     A trend in automotive finishes to the copper and
 bronzes has created a market for ultrafine iron oxides
 referred  to as  "transparent"  or  "low opacity" pig-
ments.  These pigments are combined with  aluminum to
 produce metallized automotive finishes.25  The present
 market is estimated to be between 1500 and 2500 tonnes/
 year,  worth about $5  million.
     Utilization of the additional property of absorb-
 ing ultraviolet  radiation Is  being investigated  to
 determine the potential of transparent  oxides in con-
 tainers and packaging for food  and wherever  else dur-
 ability,  transparency, and ultraviolet  absorption are
 desired in  a pigment.15
     Steel plant WPL  oxides (from FeClj)  are displac-
 ing some of the copperas oxide  products from the tra-
 ditionally ferrlte market. The soft  ferrlte market
 still  remains a  large consumer of  copperas reds and
 calcined  yellow oxides.   For 1972, the estimated  ship-
 ments  of  permanent magnets, T.V.  yokes, memory cores,
 and ferrite parts  were $192 million,2° and  they  are
 expected  to grow significantly.   A world  ferrite pro-
 duction of 495.,000 tonnes/year is  estimated by 1985.27
     At present, dnly  a very small fraction (5Z)  of
 copperas  is used for  water-sewage treatment.   The re-
 sult of a  study made  by the Fitzsimons  Steel Company,
 Youngstown,  Ohio,23  in 1973 to  market  heptahydrate
 for sewage-water treatment in 13  communities within a
 120 km  (75 mile)  radius was  not  encouraging.    For
 coagulation,  alum is  usually preferred and  the  water
 treatment  plants were unwilling to  try new  products.
 Where  phosphate control is Important, copperas is de-
 sirable but must compete with FeClj.   Detroit receives
 WPL free from HC1 pickling plants of Great Lakes  Steel
 Company and Ford Motor Company.   Milwaukee uses  FeCl3
 buC had started to use  copperas.   Chicago uses  FeCl3
 plus some  ferric sulfate.   They  found WPL from  local
 mills  too  unpredictable and impure for safe usage.23
     On the  other hand,  a more  recent experience of
 Bethlehem  Steel  Plant,  Sparrows  Point, Maryland,  is
 more encouraging.28   The 6.3 x  106  tonne/year  plant
 generates  about 100,000  gallons of WPL per day.   They
 have successfully used this WPL for removing phosphorus
 from municipal  wasteuater  at : the Gity of Baltimore's
 700,000 m3/day  (185   Mgd)  Back River wastewater treat-
 ment plant  and  at the District  of  Columbia's  1.14 x
 1Q6 m3/day  (300 Mgd) Blue Plains
 plant.
     There  is  no doubt  that in  phosphate treatment,
 WPL and copperas—supplied  In  proper condition, and
 with some treatment plant modifications, such  as addi-
 tion of solution tank,  aeration facilities, and larger
 pumps—can  adequately compete  with its close rival,
 FeClj.  With  the  emphasis  on  secondary and tertiary
 water-sewage  treatment,  the  market share of copperas
 should  grow  significantly  from the present 5Z.   And
 even a small  increase,  when treating billions of gallons
 daily, amounts to  a lot of WPL-copperas use potential.
     While worthwhile  and  economic end  uses for re-
 covered  copperas  crystals will be desirable, it must
 be recognized, at present, that  there does not exist a
 large  market for  copperas  and  the market has to be
 created.  During this period of market development, the
 copperas produced from WPL can be handled  in a far  more
 effective way (environmentally speaking)  than the pres-
 ent  techniques of contractor hauling, neutralization/
 lagooning, or deep-well disposal of WPL.
     The essentially acid-frea copperas  crystals  can
be dumped on municipal dumps as  proper fill  material.
They may be combined with other  municipal incinerator
 inert products and then dumped.
     The crystals can  be mixed  with lime,  and  less
lime quantity will be needed because there is no  free
acid to neutralize.  Or,  the  ferrous sulfate can be
converted to magnetic  oxide and  calcium sulfate  with
 lime and oxidation, and  the inert matter can be  used
as a landfill.
sewage  treatment
                                                  301

-------
WHAT'S IN THE FUTURE

     The bulk of copperas is not, at  present, generated
at the Integrated steel plants, but In the hundreds of
secondary  steel processing plants.   Titanium pigment
manufacturers and the commercial producers also generate
a significant quantity of copperas.  While these plants
are dispersed all over the U.S.A., they are, to a cer-
tain extent, clustered in certain Industrially developed
regions,  for example,  the Chicago-East Chicago-Gary
greater metropolitan area. Before trying to solve the
problem  of disposal of copperas yet  to be generated
from steel plant WPL,  it will be necessary to direct
attention and effort to make use of  WPL from the sec-
ondary processing industries to generate copperas, and
either to find a market for it or to dispose of it in
an environmentally safer way.  To achieve this end, it
will be necessary to direct attention to the following:

     1.  A study to define the scope and estab-
         lishment of regional, centralized WPL
         facilities to be funded Jointly by the
         secondary processors and aided by EFA
         or  a similar regulatory body  for ini-
         tial subsidized operation.

     2.  Co-treatment of different waste streams
         to recycle and recover valuable prod-
         ucts.  Also, to develop waste exchange
         Information data with the help of EPA
         to aid industry In this effort.

     3.  Promotion of tertiary sewage treatment
         with particular emphasis on phosphate
         removal.  Also,  to  establish an EPA
         directive to require all EPA-alded sewage
         treatment facilities to utilize recycled
         waste product such as copperas and to
         extend additional aid to them for nec-
         essary equipment modifications to make
         such use possible.

     4.  Research and development to find new
         uses for copperas, to Improve pickling
         technology, to reduce use of acid and
         water, and to minimize iron loss result-
         ing  in process  development  with zero
         liquid discharge (KSF or similar process) .

Centralized WPL Treatment Plant
     The economics of scale will be very favorable for
such a treatment plant.  Similar treatment plants are
in operation in West Germany serving Industries within
a radius of 80 km.   The mechanics of  joint ownership
and Federal  aid must be studied in the U.S. economic
and rezulatory context.  However, a radius of  80 km
will encompass the whole of the Greater Chicago Metro-
politan Area and beyond.  At some point in time, some
of the big steel industries of this  area  using HjSO^
acid pickling may even  participate in its operation.
The members will be charged on the basis of  WPL volume
treated, and a credit will be given for recovered acid.
The heotahvdrate will either be sold to sewage treat-
ment  plants  of the same  area operated with  Federal
aid, or converted to oxide, or treated  and  dumped as
land fill.
     Well-established technology for WPL  treatment
for copperas production is now available  Cone  being
the  KSF-zero discharge process).   Nonetheless,   new
research and development studies are needed  to  develop
alternative economical processes.  The Crown Chemical
ion exchange process3  has demonstrated a  marketable
Fe20j starting  with heptahydrate  crystals.   Using a
double loop counter current flow,  a hydrogen ion ex-
change resin, nitric acid, and a hydrolyzer,  a  bench
scale unit produced ferrite/pigment grade Fe2C>3.  In
1975, about 17,500 tonnes of Fe203 were produced from
steel plant WPL and sold at an average  cost of $557
tonne, principally  to  ferrite manufacturers.  A dem-
onstration  plant based on  the hydrogen ion exchange
resin process will be required  before the economics
of the process can be established.
     There are  many other areas where additional re-
search  will  bear results.   For  example,  several
laboratory-scale electrolytic techniques have  been
developed, but none has attained a demonstration plant
stage.  It  was  demonstrated  that using Hg cathodes,
electrolytic regeneration  of  iron  1 s  both reliable
and economical."

Co-Treatment of WPL and Other Waste Streams
     WFL is still treated with technologies 50 years old:
cooling, crystallization, filtration,  washing, drying
heating—a combination of  fairly  primitive technolo-
gies. Very little attention is given to advanced tech-
nologies  such  as ion exchange, electrolysis,  high-
intensity  magnetic separation, and reverse  osmosis,
to name a few.   Many other metal processing Industries
generate waste liquor/solids which are equally or more
difficult  to  handle.   Electroplating and electronic
industries are two  such groups whose wast* streams may
be effectively co-treated with WPL to benefit both.
     Two printed circuit (PC) shops are presently using
heptahydrate to treat electroless copper and alkaline
etch rinses.  A large West Coast PC manufacturer treats
1200 I/day with 5 to 20 g/1 of heptahydrate.   An East
PC manufacturer treats six electroless copper  rinses
and one alkaline etchant rinse with heptahydrate in a
continuous flow system.  The 100 1/min (26 gpm)  flow
of 20 t'o 30 mg/1 copper is reduced to less than 1 mg/1
after the sulfate treatment.    Copper plating  rinse
waters were treated with WFL to reduce Initial copper
concentration of 1000 mg/1 to O.S mg/1.30   A similar
co-treatment of  ehromate waste  water with WPL and wash
water can reduce Iron and chrome In the final effluent
to O.S and 0.05 mg/1, respectively.31
     While recovery and recycling  of acid and metallic
values are worthwhile under certain conditions, it is
often necessary to treat WPL and acid wash water with
a neutralize to make it environmentally harmless.
Often, lime Is used—a highly energy-Intensive product.
On the other hand, a large quantity of lime residue is
available from the carbide process. Like many unwanted
waste products, the lime residue has a negative value
and, Instead of lime, this lime residue can be effec-
tively and economically co-treated with wash water/WPL.32
     In addition to development of co-treatment process
technology, it will  be very desirable to compile regional/
local waste stream/solid waste generation information.
This information can be made available to the regional
Industries with available technological information re-
garding  the possibilities of  their co-treatment and
disposal.  A Federal agency such as EPA may select one
large region (for example, the Chicago area) as a target
area to initiate this  study.  This pattern  of  waste
Information exchange was Initiated In West Germany,  and
now the industries have developed  their  own contacts
and need very little federal help.

Tertiary Sewage Treatment and Phosphate Removal
     Potentially,  sewage treatment for phosphate re-
moval and water treatment for coagulation are two uses
which can theoretically utilize all possible copperas
(or WPL) production.  However,  long-term agreements,
                                                  302

-------
reluctance  b y  authorities  t o.  experiment with new
chemicals, lack  of compelling enforcing regulations
for phosphate removal, additional expense for equip-
ment modification, and, lastly, no concerted expres-
sion of Interest  from different Federal authorities
for recycling of heptahydrate  are some of the major
reasons for the poor share  (51) of heptahydrate usage
In  water-sewage treatment.  Because  the potential
Is so vast and the application of copperas is techno-
logically  so straightforward, economical, and well-
proven,28' 3 3 it  ts surprising  that  the Federal au-
thorities  who  aid many  of  the urban water-sewage
treatment plants did not yet exert their Influence to
promote the use of copperas in this application.  The
existing Federal regulations which reward use of re-
cycled products can be effectively utilized in water-
sewage  treatment plants.  The  aid  to  these plants
can be  made contingent on  their utilizing recycled
products  such  as copperas  or WPL.  Additional  aid
may be available to these plants for  equipment ad-
ditions and alteration which will be required if
copperas is used to replace FeCl3 and alum.  This use
is potentially  most  effective because  of  the close
proximity  of the copperas (WPL) -generating  Indus-
tries  and the  local  treatment  facilities,  thereby
reducing transportation cost very substantially.

Research and Development of Picklinft Technology
     While pickling is an old technology, many changes
have taken place during the last 20 years to Improve
its application.   One  major shift Is from H2S04 to
HC1 pickling resulting in  a better looking product,
faster pickling, and potential for easier HCl regen-
eration and by-product iron oxide production and usage.
After a rapid growth of HCl pickling,  it has come to
a balance  with  H2S04 pickling on a tonnage basis of
30% H2S04  to 70* HCl.   Further, HCl  pickling conver-
sion of existing H2S04 pickling lines  has  virtually
ceased In the U.S.A.,  whereas it is  still proceeding
In Europe.   The higher HCl cost  relative to H2S04,
and the possible oversupply of H2S04 because of  S02
emission  control  and  S  recovery due  to EPA regula-
tions may stabilize H2S04 cost farther, thereby revers-
ing the trend of HCl  acid pickling.  Any new  green-
field plant is likely to use HCl pickling, but the  ad-
ditional roundlng-of f facilities to existing units will
still be by the H2S04  acid process where already ex-
isting,  and will add to the copperas disposal problem.
     Hew pickling technology using a mixture of H,S04/
HCl acids has been developed bv Wean/KSF5  and has been
tried in continuous/batch pickling operation.  HCl as-
sists H2S04 in the pickling, and H2S04  also acts as a
reagent regenerating HCl from FeCl2 according to  the
equation:
          FeCl
                        -  FeS0
                                   2HC1
Copperas  can be recovered  from  the WPL with little
chloride contamination.  The resulting pickled steel
has the brightness of HCl pickling.
     Additional  studies  are required  in the area of
pickling mechanism  and washing of pickling solution
from the sheet surface.  If a true understanding of the
mechanism involved in the pickling of steel  by HCl and
H2S04 is obtained, anew technology using less acid and
consuming less iron can be realized.  Generally,  at
0.5Z Fe loss, one tonne of steel pickled produced 25 kg
(55 Ib) of FeS04-7H20.   At IX Fe loss,  It is 50 kg of
heptahydrate/tonne  steel.   While these are  average
values, different products, depending on their surface
area/volume ratio and process technology, will show dif-
ferent amounts of Iron losses.   The data show that in
wire pickling about 1Z iron loss is encountered, whereas
In billet pickling it Is only 0.22Z.21  Thus, any effort
in understanding the mechanisms Involved in Iron loss
and the effect of Inhibitor control in minimizing iron
loss will eventually have the highest potential effect
in reducing copperas production.   The industrywide
figure is 0.6ZFe loss, and a reduction to 0.52 Fe loss
means a total reduction of 16Z in WPL/copperas produc-
tion amounting to hundreds of thousands  of  tonnes per
year.
     Considerable Improvement  i n  the technology of
washing  of steel surfaces has taken place in recent
years.  The traditional method of dipping In water bath
and flood cooling requiring vast amounts of water is
giving way to more scientific spray cooling and tempera-
ture control.  Also,  indirect heating  of  baths and
externa.1 heat exchangers are lessening steam condensa-
tion and dilution effects of earlier days.  However,
the mechanisms involved in the removal of thin acid/
water films (water is a polar compound) from a freshly
pickled and highly reactive steel surface as a func-
tion of water temperature, velocity, and other unde-
fined parameters are not adequately understood.  Also,
additional equipment development In squeezing out carry-
over acids will  tremendously help  In reducing wash
wster problems and its subsequent disposal.
     Hew research to expand the use of copperas in new
areas is greatly needed along with effort to hold  the
market for existing uses.  For example, the Introduc-
tion of the Aniline Process for making iron oxide pig-
ment is likely to affect pigment use of copperas and
must be countered by research.

CONCLUSIONS AND RECOMMENDATIONS

     At present, the integrated steel industry does not
feel the economic compulsion for generating copperas
from the several hundred million gallons of waste pickle
liquor it produces annually.  As such, copperas from
the steel Industry does not pose any problem to those
who are try ing to market it.  Copperas has a ready but
limited market.  The largest user is the pigment Indus-
try, and it cannot possibly absorb significantly larger
quantities In the near future.  The most promising area
for expanded use is sewage treatment for  phosphate  re-
moval.  The long-term contracts and traditional resis-
tance to use of an unknown chealcal is,  at present,pre-
venting  its  effective utilization.  Also,  lack of
emphasis on tertiary treatment regulations and positive
incentive in the  form  of additional federal  aid for
equipment modifications and use of recycled  products
are  several  institutional factors  holding back in-
creased used of heptahydrate In sewage treatment.
     Even If heptahydrates are not  all marketable they
are better for disposal than diluted and neutralized
waste liquor.   To economically achieve copperas pro-
duction,  large centralized treatment units  such  as
those that exist in West Germany will  go a long way In
solving the problems  of hundreds  of small  processing
units.  Also,   research  studies  are  needed  on  co-
treatment  of  different waste streams, and  regional
surveys are needed to bring such information to the
attention of companies Interested In waste treatment
facilities.
     Sulfuric acid pickling in the steel Industry  is
anticipated to continue to make up about thirty per-
cent of pickling practice.  Hew plants are likely  to
use hydrochloric acid pickling  combined with acid re-
generation in spite of the fact  that  the process  Is
energy Intensive.   The roundlng-off facilities to be
added to steel plants is likely  to  extend the  existing
pickling practice.  If new regulations force  some  of
                                                  303

-------
the steel plants to switch  from deep-well disposal,
neutrallzatlon/lagoonlng, or discharge Into large water
bodies, then the plants may consider heptahydrate pro-
duction as one of the alternatives.  At  present, only
75,000 tonnes of Iron sulfate  crvetals  are produced
and most of the product IB marketed.
     The bulk of marketed copperas Is not  generated
at the Integrated steel plants which procesa about 20
million tonnea of steel per'year using sulfurIc acid.
If  It  were, then art  additional 600,000 tonnes/year
would be added to a market which does not  have an easy
time in marketing the estimated 75,000 tonnes/year pro-
duced by the snail ateel processing units. Under some
future altered circumstances, production of copperaa
from waste pickle liquor by the integrated  steel plants
has the potential of adversely affecting  the copperas
market.  In order to avoid serious future dislocation
and to promote  use  of copperaa (and VPL) and expand
its market in a positive manner, several suggestions
are made below:

     1.  A study to define the scope for estab-
         lishment of regional centralized waste
         pickle liquor facilities  to be funded
         jointly by  the secondary processors.
         Encouragement, and perhaps Initial sub-
         sidized operations, from government may
         be needed.

     2.  Significant research on co-treatment of
         different waste streams and on solids to
         recycle/recover valuable products and to
         stabilize the  waste for  safe  disposal.
         Establishment of a bank of waste  exchange
         Information on a zonal basis for different
         regional industries to take mutual advan-
         tage of one another's waste products.

     3.  A study of  the use of spent pickle liquor
         and copperas for  wastewater  treatment,
         particularly tertiary treatment for phos-
         phate removal and effluent polishing.  This
         approach uses two environmentally negative
         practices,  the  disposal  of  spent pickle
         liquor and  the discharge of phosphorus, to
         abate each other.  The advantages and dis-
         advantages  should be considered, including
         the problem of heavy metal contamination
         of wastewater treatment plant effluents
         and sludges,  and the question of whether
         spent pickle liquor should be used directly
         or only  the copperas  derived   from  It.
         Methods of  encouragement should also  be
         considered.

     4.  Research and development to find new uses
         and to extend present application of cop-
         peras by Improving its quality.   New stud-
         ies are needed to improve pickling technol-
         ogy so that  less Iron is lost, resulting
         InlessWPL,   Additional knowledge is needed
         regarding mechanisms of  pickling and  the
         mechanism of washing of drag-out acids with
         minimum water.
ACKNOWLEDGMENT

     The program was funded by the D.S. Environmental
Protection Agency.  Mr. J. S.  Ruppersberger of  the In-
dustrial Environmental Research Laboratory, U,S. E.P. A.,
Research Triangle Park, N.C. was the Project Officer,
and his interest and help in this study Is gratefully
acknowledged.  We also gratefully acknowledge the co-
operation of many Individuals and steel planes in the
performance  of  this study.   Mr. R. J. Lackner,  Vice
President,  KSF  Chemical  Processes Ltd., Cambridge,
Ontario, Canada, was very helpful  and  his assistance
Is gratefully acknowledged.

REFERENCES

 1.  Bhattacharyya, S., "Process, Water Quality Re-
     quirements for Iron and Steel Industry," EPA-
     600/2-79-003, January 1979.

 2.  "The Stael Industry In the United States, Plant
     Locations," Institute for Iron and Steel Studies,
     Green Brook, New Jersey.

 3.  Peterson, J. C., "Closed Loop System for the
     Treatment of Haste Pickle Liquors," EPA-600/2-
     77-127, July 1977.

 4.  Annual  Statistical Report, 1969. 1973, and 1976,
     American Iron and Steel Institute, Washington,
     D.C.

 5.  Lackner, R. J., KSF Chemical Processes Ltd.  and
     Wean United, Inc., private conunlcatlon.
 6.  Kotsch, J. A., Labee, C. J., Schmidt, R. L., and
     Samways, N. L., "Developments In Iron and Steel
     Industry U.S. and Canada - 1978," Iron and Steel
     Eng., Vol. 56, No. 2, February 1979, pp. D-l to
     D-38.

 7.  A Delphi Exploration of the U.S. Ferralloy and
     Steel Industries. NMAB-277, July 1971, National
     Materials Advisory Board, Washington, D.C.

 8.  EPA Technology Transfer Capsule Report, "Recovery
     of Spent Sulfurlc Acid fron Stael Pickling Opera-
     tions," EPA-625/2-78-017, 1978.

 9.  Anon., "Central Waste Disposal, New Service Looks
     for Some Action," Environ. Sci. Techno1., Vol. 4.
     No. 3, March 1970, pp. 195-197.

10.  Bayazeed, A. P., and Donaldson, E. C., "Subsur-
     face Disposal of Pickle Liquor." R.I. 7804, U.S.
     Bureau of Mines, Washington, D.C., 1973.

11.  Lackner, R. J., "Acid Recycling Systems for Pick-
     ling Lines," Presented to Association of Iron
     and Steel Engineers, Youngstovn District Section,
     Girard, Ohio, May 6, 1974.
                                                 304

-------
12.  Anon., "Water Pollutant or Reusable Source?"
     Environ. Sci. Technol., Vol. 4, No. 5, May 1970,
     pp. 380-382.

13.  Ulttman, I. E., and Shephard, G. S., "Integrated
     Steel Pickling Rinse Water Treatment Syitam,"
     Iron Steel Eng., Vol. 49, No. 2, February 1972,
     pp. 69-71.

1*.  Hancock, K. R., "Iron Oxide Pigments," Pfizer
     Inc., Minerals, Pigments and Metals Division,
     December 27, 1972.

IS.  Jones, S. T., "Iron Oxide Pigments (in two
     parts), 1. Fine Particle Iron Oxides for Pigment,
     Electronic, and Chemical Use," Inform. Circ.
     8771, 1978, U.S. Bureau of Mines, Washington,
     D.C.

16.  EPA Technology Transfer, March 1, 1973, U.S.
     Environmental Protection Agency, Research Tri-
     angle Park, N.C.

17.  Current Industrial Reports, Inorganic Chemicals
     (1976, 1971, 1970), M28A (76)-14, Aug. 1977;
     M28A(71)-14, Oct. 1972; M28A(70)-14, June 1972;
     U.S. Dept. Com., Bureau of the Census, Washing-
     ton, D.C.

18.  Berry, R, I., "Asphalt Substitutes: The Time
     May be Ripe," Chem. Eng., Vol.  86, Ho. 25,
     November 19, 1979, pp. 98, 100.

19.  Burtch, J. W., "Hydrochloric Acid from Indus-
     trial Waste Streams—The PORI Process," Can.
     Mining Met. Bull., Vol. 68, January 1975,
     pp. 96-100.

20.  Rupay, G. H., and Jewell, C. J., "The Regenera-
     tion of Hydrochloric Acid from Waste Pickle
     Liquor Using Keramchemle/Lurgl Fluidized-Bed
     Reactor System," Bull., Canadian Institution
     of Metallurgy, February 1975, pp. 89-98.

21.  Ing. Wurmbauer,  "A Process for  Regeneration of
     Spent Hydrochloric Pickle Acid," Wire Industry,
     Vol. 45,  January 1978, pp. 43-45.

22.  Craig, Jr., A. B., "Trip Report—Visits to
     Centralized Treatment Plants for Industrial
     Wastes," October 28, 1978, U.S.  EPA,  Cincinnati,
     Ohio.
23.  Sayler, J. K., Thornton, W. E., and Householder,
     M. K., "Sulfuric Acid and Ferrous Sulfate Re-
     covery from Waste Pickle liquor," EPA-660/2-73-
     032. January 1974.

24.  U.S. Industrial Outlook, 1971, p. 188 and p. 199,
     U.S. Dept. of Commerce, Washington, D.C.

25.  Anon., "More Sparkle for Auto  Finishes," Chem.
     Week, Vol. Ill, August 2, 1972, pp. 29-31.

26.  Andrews, P. B. B., "Sales Forecast:  Records Will
     Tumble in Active Year," Ceramic Age, Vol. 88,
     Jan. 1972, pp. 23 ff.

27.  Ruthner, M., and Ruthner, 0.,  "25 Years of Pro-
     cess Development in HC1 Pickling and Acid Regen-
     eration," Iron and Steel Engineer, Vol. 56,
     No. 11, November 1979, pp. 38-41.

28.  Kerecz, B. J., Mohr, R. T., and Bailey, W. F.,
     "Use of Spent Pickle Liquor to Remove the Phos-
     phates in Municipal Sewage Treatment Plants,"
     Presented at the EPA Symposium on Iron and Steel
     Pollution Abatement Technology, Chicago, 111.,
     Oct. 30-Nov. I, 1979.

29.  Jangg,  G., "Electrolytic Regeneration of Sulfuric
     Acid Iron Pickling Solutions," Electrodeposition
     Surf. Treat., Vol. 1, No. 2, November 1972,
     pp. 139-149.

30.  Wing, R. E., "Process for Heavy Metal Removal
     from Plating Waste Waters," EPA-600/8-78-010,
     May 1978,  First Annual Conference on Advanced
     Pollution Control for the Metal Finishing Indus-
     try, Lake Buena Vista, Florida, January 17-18,
     1978.

31.  Cupps,  C. C., "Treatment of Wastes for  Automo-
     bile Bumper Finishing," Ind. Water Wastes, Vol. 6,
     1961, pp.  111-114.

32.  Heise,  L.  W., and Johnson,  M., U.S.  Patent No.
     2,692,229, October 19, 1954, and assigned to
     A. 0. Smith Corporation.

33.  Anon.,"Fe  Wastes Help Remove Phosphates," Can.
     Chem. Process.,  Vol.  58, May 1974,  pp.  44-46.
                                                  305

-------
                         THE EFFECTS OF PRETREATMENT
            ON COKE PLANT WASTE WATER BIOLOGICAL TREATMENT SYSTEMS
                                 Authors:

               Bernard A.  Bucchianeri
               Division Engineer -.Chemical Operations
               U.  S.  Steel - Clairton Works, Clairton, Pa.

               Leon W.  Wilson, Jr.
               Senior Research Engineer
               U.  S.  Steel - Monroeville, Pa.

               Kenneth D.  Tracy
               Principal
               Environmental Dynamics, Inc. -  Greensboro, N. C.
          The United  States Steel Corporation in  conjunction with the Environ-
mental  Protection Agency is conducting an extensive experimental program to
develop input data  relative to BATEA technology for coke plant waste waters.
The  program which  involves testing on both a bench  scale as well as the pilot
scale  is  concerned  with  determining the optimum operating criteria for coke
plant  biological  treatment systems.  Once the optimum conditions are identi-
fied,  further investigations will evaluate the impact of  the addition of
powdered activated  carbon  to an optimized system.

          As an initial  step in identifying optimum operating conditions, a
separate  investigation was  conducted  to evaluate the importance of pre-
cleaning  of the feedwater  to the biological system.  The evaluation program
centered  on those constituents falling  into  the two general categories  of
suspended solids and "oil  and grease".   Various removal techniques were
evaluated and the necessity for achieving specific levels of influent pre-
treatment was addressed.  The validity of the resultant conclusions relative
to precleaning is supported by operational data from a 9,500 m-Vday (2.5 MM
GPD)  coke  plant waste water treating facility.
                                     307

-------
                        THE  EFFECTS OF PRETREATMENT
          ON COKE PLANT WASTE WATER BIOLOGICAL TREATMENT SYSTEMS
BACKGROUND AND INTRODUCTION

         Clairton Works of  the United States  Steel Corporation  is one of the
world's  largest  producers of metallurgical  coke.  The facility consists of
nearly 1,000  ovens which produce sufficient quantities of coke  to supply all
of the corporation's steel-making facilities in the Pittsburgh area with some
excess being available for other corporate locations.  In addition to produc-
ing coke, Clairton Works has a totally-integrated system for recovering and
refining a  full  complement  of coal chemicals.  Unlike traditional by-product
facilities  which utilize a series of  low pressure recovery processes, the
Clairton system  employs elevated pressures 3.52-3.87 Kg/CM2 (50-55 PSIA) and
unique separation and recovery processes.   Naturally-occurring ammonia is
recovered as  an  anhydrous product using the U.  S. Steel patented  Phosam
process.  Light  oil fractions are separated from the gas using a computer-
controlled cryogenic-regenerator system in which the gas is cooled to tempera-
tures of less  than -157* C (-250° F).   This  same regeneration  system yields
an ultra-pure, hydrogen-rich gas for  consumption in the synthetic ammonia
plant while simultaneously providing  a feed gas of exceptional quality to
heat the coke  ovens.   Additional facilities  are operated for the removal of
sulfur from fuel gases leaving Clairton Works  for use in other local U. S. S.
facilities.   In  addition to the primary-recovery facilities,  complimentary
systems  are operated  to produce metallurgical grade anhydrous  ammonia, ben-
zene, toluene  and xylene as well as a complete line of tar based derivatives
and naphthalene.

         At typical operating levels,  Clairton Works generates somewhat in
excess of 9,500 m^/day (2.5 MM GPD) of contaminated water.  Approximately 45%
of the total  generation occurs as a direct  result of the coking operations
with the remaining 55% attributable to chemical processes.  The  typical
composition  of the raw  contaminated water is shown in Table 1,


                                  Table  1


          COMPOSITION OF COKE PLANT CONTAMINATED WATER


              Ammonia                     — 1500-2000 PPM

              Phenol                       - 800-1200 PPM

              Thiocyanate                  — 600 -700 PPM

              Cyanide (Total)                — 200 -400 PPM

              Oil/Grease (Freon Extractibles) — 2000-4000 PPM

              Total Suspended Solids        — 300-1500 PPM
              pH                           -    8-9

              Temperature                  — 130-170O p.
                                     308

-------
          The  principle  elements of  the Clairton  Works contaminated water
 treatment  facilities are shown schematically in Figure 1.   Following  gravity
 separation  of  both solids  and suspended oils  in  the settling  tanks,  the
 contaminated water  is processed through the U. S. Steel patented  Cyam process.
 Here the water steam stripped of so-called "free ammonia",  pH  adjusted by the
                                    Figure 1

                 United States Steel Corporation    Clairton Works
              CONTAMINATED WATER TREATMENT PLANT
                          SETTLING
                           TANK
                                          FREE AMMONIA
                                             STILL 	..
                                                                -PHOS-AM ABSORBER
                                                                    PHOS-AM
                                                                    REGENERATOR
      100.OOO SAL SlTTLIMa TANK*  ADDITIVES
                                       100.000 GAL
                                        FEED TANK
         ACTIVATED SLUDOE AERATION BASINS
                6.9 MM SAL.
                                           100,000 OAL.
                                           COUALIZINO
                                                       LIME V   CLOSED
                                                      SLUDGE    CIRCUIT
                                                       THICKENER  COOLERS
addition  of lime  to liberate  "fixed ammonia"  and finally further  steam
stripped to yield  a biological  feed stream of desired ammonia content.   In
addition to ammonia removal,  the  Cyam system also accounts for the removal of
^0-95%  of  the cyanides as well  as other  "acid" gas fractions.  Following
ammonia removal, the water is cooled  and clarified prior to biological treat-
ment.   Clarification is provided  by the  use  of a conventional center-well
peripheral  overflow clarifier.

          The biological treatment system is a single-stage process consisting
of 2  independent aeration basins  operating  in  parallel.  The total system
volume  is  in excess of 24,600 m3  (6.5 million  gallons) with aeration and
mixing being provided by low speed mechanical surface aerators.


                                      3t)9

-------
          Concerns relative  to  the impact of influent quality on the  overall
performance  of  a biological  system arose as a result of studies to determine
the ability  of  the existing  system  to meet proposed  1984  final effluent
standards.   A literature search  substantiated by in-plant  inspections of
operating  facilities suggested  that  technology developed in the petroleum
industry^  might well have validity in the coking industry.  This technology
stressed  the importance of  extensive precleaning of  the  feedwater to a
biological system as a significant prerequisite to achieving nitrification.
Total suspended solids and  oil/grease concentration each not exceeding 20
mg/L were  stated as constraints.  Because of the qualitative similarities
which exist  between coke plant waste waters and refinery waste waters, con-
sideration was  given as to  the  applicability of this technology.  In-plant
investigations  conducted on  what might be termed a "macro" level concluded
that although coke plant water  contained the same general spectrum of con-
taminants  as refinery wastes,  the use of the previously stated constraints
concerning TSS  and 0/G did  not  appear applicable to  the Clairton system.

          During the third quarter of 1979, the Environmental Protection
Agency  and U. S. Steel agreed to conduct an extensive experimental program to
develop input data relative  to BATEA technology for coke plant wastewaters.
the goals  of the study were  twofold.   First, the conditions which promote
optimum performance of a biological system were to be determined.  Secondly,
the impact of the addition of powdered activated carbon  (PAC) to a biological
system operating at the optimum conditions was to be evaluated.  Environmental
Dynamics Incorporated of Greenville, South Carolina, was selected to work in
conjunction  with U. S. Steel Research and Clairton personnel to conduct the
year-long  evaluation program.   The  complete investigative program involves
extensive  testing of biological  reactors on both the bench-scale as well as
the pilot  scale.  Bench-scale  reactors [.028 m3 (7.5 gal.)] being used to
evaluate individual variables and pilot reactors [3.2m3 (850 gal.)] used for
a parallel evaluation of an optimized biological system in comparison to a
similarly  optimized system with  PAC addition.  As part of the overall program
of determining  optimum operating conditions for a biological system, it was
decided to investigate (on  a "Micro" basis) the importance of influent pre-
cleaning in terms of total suspended solids and oil/grease.

METHODS AND PROCEDURES

          The evaluation  of  pretreatment alternatives was based primarily on
jar tests.  The limitations of using  a batch procedure like the jar test to
evaluate a continuous-flow  treatment unit  was recognized.  It was felt,
however, that since alternatives were being compared,  the results would be
relative,  and the procedure could be  used  to define  optimal conditions.  To
minimize the analytical  load generated by  the large  number of jar testss
qualitative  observations such as floe  size,  supernatant clarity  and relative
settling rate were used to screen alternatives.  When promising  alternatives
had been defined,  quantitative analyses  of  the supernatant oil  and grease  and
suspended solids were used to define the  optimum conditions.

          An  evaluation  of  the existing  lime  sludge  thickener  (LST) was
included in the  pretreatment evaluation.  To  define  the performance potential,
several Class 2 settling analyses were  performed^.   This procedure, designed
to  predict clarification  of flocculent suspensions,  consists  of quiescent
batch-settling tests with periodic measurement of suspended solids at several
depths  in the subsiding column.   The  concentrations  at  the various depths  and

                                      310

-------
times  permit the development of the relationship between overflow rate  and
solids removal.   Such  tests were conducted to estimate performance under  three
different coagulant  schemes.

          In addition to optimizing  the performance of the LSI, alternative
pretreatment processes  were considered.   The dissolved air flotation  (DAF)
process was evaluated  using a batch pressurization tank and separation column.
By varying the pressure and volume in the pressurization tank, it was possible
to investigate a wide variety of air-to-solids ratios.  Additional tests were
also conducted to evaluate  the applicability  of the induced air flotation
(IAF) process.

          Granular-media filtration was considered as a supplemental process
to upgrade the effluent  from the LST.  A 5  cm (two-inch) diameter, dual-media
filter  was used in the  evaluation.   The media consisted of 30 cm (12 inches)
of 0.9 mm anthracite and  30 cm (12 inches) of 0.65 mm of quartz sand over 9 cm
(3-1/2  inches)  of graded gravel.  The length of the filter run was determined
by the  available head of  1.2 m (A feet)  or the breakthrough of solids in
excess  of 20 mg/L.  Performance of the filter was assessed by effluent solids
and oil and grease levels.

RESULTS AND DISCUSSION

Oil and Grease Removal
          In this study, the term "oil  and grease" is used to refer to  those
materials which are  extractable in fluorocarbon-113 and detectable by gravi-
metric  analysis after evaporation at  70°  C.  The nature of the test,  then,
limits  its applicability  to higher molecular-weight hydrocarbons which have
boiling points greater than 70°C.  This  is  a reasonable limitation when using
the  test to assess the  impact of oil and grease on the operation of biologi-
cal  processes since  the lower molecular-weight hydrocarbons,  which are not
detected in the test, tend  to be more  degradable and have  little adverse
impact  on the process.  One group of lower molecular-weight compounds which
the  test does measure is  the organic  acids.   These compounds are normally
dissociated and, therefore,  soluble at typical coke-plant waste water pH.
During  the oil and grease extraction,   the pH is first lowered to less than
2.0 which shifts the ionic  equilibrium resulting in most of the acids being
non-dissociated  and extractable.  The oil and grease procedure,  then,  measures
a class  of compounds which  are normally  soluble.   When using the freon-
gravimetric test to  monitor influents to biological processes,  high influent
levels  and potential operating problems can  be indicated when large quantities
of organic acids are present.   These  acids, in general,  are biodegradable,
which would preclude any  adverse impact on the biological process.

          One of the  goals of the Clairton BATEA study was to optimize the
existing biological  process  to determine  what,  if any, further improvements
would be required to meet future standards.   Similar studies in other in-
dustries1 »3 »4 have indicated that operation at long solids retention time
(SRT) results in the enhancement of effluent  quality.  Successful operation at
long SRT,  however,  requires  an influent which is low in oil and grease since
higher  molecular-weight hydrocarbons are normally removed by physical absorp-
tion on  the sludge rather than biodegradation.   These materials, therefore,
accumulate at  long SRT resulting in deterioration c-f settling characteristics
and eventual process failure.   In light of this background, an influent oil
and grease target of 20 mg/1 was selected for  the pretreatment studies.


                                     311

-------
          Jar tests were  used to evaluate  the effects of a wide variety of
operational  variables on the removal  of  oil and grease by the existing pre-
treatment  units.  These  tests indicated  that  the presently used polymer and
dosage performed  as well as alternative  coagulants or coagulant-flocculant aid
combinations.  Varying the coagulation pH from 8 to 12 and the provision of up
to eight  minutes of supplemental flocculation also failed to provide improve-
ments  over the existing scheme.  In general, the supernatants produced in jar
tests  with optimum coagulant addition were similar to LSI effluent in oil and
grease content indicating  the near-optimal  performance of the existing system.

          The failure of  the jar-test  program to indicate a potential for
significant  enhancement  of LSI performance  for 0/G removal led to the evalu-
ation  of  alternative processes.  Batch  DAF studies showed that the density of
the  solids'and  Freon extractable materials  in  the LST feed were such that
flotation would be very  difficult.  No  single  polymer was able to produce
flotation;  and, although several coagulant  combinations resulted in flotation,
the  supernatant quality  was poorer than the  existing LST effluent.   One
combination  consisting of  a cationic  primary  coagulant and an anionic floc-
culant aid proved effective, but the  dosages required were Economically
prohibitive  and  the float volumes excessive.  The test results indicated that
flotation was not a viable option for  pretreatment of coke plant wastewater.

          The final alternative for attaining the oil and grease target of
20 mg/1 was  the use of granular-media filtration as a supplement to the
existing  LST.   In these  tests, LST effluent, which had been treated with the
optimum polymer  dose prior to settling,  was  passed through a laboratory-scale
sand-anthracite  filter.   The results of the  filter runs summarized in Table  2
indicate very little removal of oil and grease with the effluent far in excess
of the target.
                                    Table 2

               SUMMARY OF BENCH-SCALE FILTRATION TESTS
                                                  AVG.    AVQ.
                                                  TSS    O&G
                          TEMP   RATE            (mg/1)    (mg/1)
             RUN NO.  pH    QC  1/nUr
-------
          The  failure  of the filter  to  reduce oil and grease to 20 mg/1
prompted an investigation of the nature  of the oil and grease in  the LST
effluent.  Duplicate  samples were taken periodically over a two-week interval.
One  sample was analyzed  for total freon  extractables while  the other w.-is
passed  through a 0.45  micron filter  and  analyzed for the soluble fraction.
The average total  oil and  grease for these  samples was 35 mg/1 and the  soluble
fraction was 31 mg/1 confirming that  the  bulk of the oil in the LST effluent
is soluble.   The solubility of the oil,  therefore, makes the pretreatment
target of 20  mg/1  unattainable by conventional technology.

          Using  historical operating data,  a material balance was performed to
assess the fate  and effects of oil and  grease  on the activated sludge system.
The  18-month data base consisted of  freon  extractable values  obtained on a
daily basis.   Figure 2 illustrates the  frequency distribution of  the operating
data.   The 50-percentile values for influent and effluent oil and grease are
25 and  2.6 mg/1 respectively.  Soxhlet extractions of the mixed liquor indi-
cated an average oil and grease accumulation of less than one percent of the
total mixed  liquor on  a  dry weight basis.  This  low level of accumulation
indicates that the bulk of the freon extractables removed from the waste water
were biodegraded.  Since the influent oil  and  grease is  degradable higher
concentrations  will not  adversely effect settling properties  and interfere
with  long-SRT operation.

                                   Figure 2
           O/G (mg/1)
           100
FREQUENCY DISTRIBUTION
  OF OIL & GREASE DATA
               0.1 0.5 1  2   5  10   20 3040506070 80   90  95 9899
                 Percent of Time Less Than or Equal to the Indicated Value
Suspended  Solids Removal

          The Clairton waste water  is  treated with lime to  raise the pH prior
to the fixed-ammonia still.   With the high pH  and excess  calcium present,
large amounts  of  calcium carbonate are precipitated.  This  precipitate is
removed in  the LST which serves  to both clarify  the feed  to the activated
sludge  system and thicken the resultant  inorganic sludge.
                                     313

-------
           In establishing an influent  solids target for a biological  process,
 the major concerns are the nature of the  solids, and their resulting impact on
 the biological  system.  Inert solids, such as the inorganic precipitate in the
 Clairton  feed, constitute a material load which must be transported,  but does
 not contribute to process  performance.  Based  on experience with  long SRT
 operation  in the petroleum industry, an arbitrary pretreatment  target of 20
 mg/1 was  selected.

          The results of jar  tests to optimize suspended solids  removal were
 similar  to  the oil and grease results which indicated that the present opera-
 tional conditions were equal  to or  better than the alternatives.   From the
 many alternatives evaluated,  three were selected for  clarification tests.
 Class  II settling analyses were conducted to evaluate the theoretical clari-
 fication for:   (1)  no coagulant addition, (2) the current polymer  addition
 scheme, and  (3) a metallic coagulant with a polymeric coagulant aid.   Figur'e  3
 illustrates  the  relationship  between overflow rate and solids removal for
 these  three cases.
                                    Figure 3
                Clairifier
                Effluent
                15
                30
                45
                60
                75
                90
               105
               120
EFFECT OF OVERFLOW RATE
     ON CLARIFICATION
         0.75 mg/1 Anionic
         /   Polymere
                   Ferric Chloride
                   Plus Polymere
                                               I
                            I
                  0  200  400  600  800  1000 1200  1400
                Overflow Rate (gal./ft.2 — day)  FEED TSS = 300 PPM
          The tests cannot be directly  compared because they were  based on
different samples.  They  are indicative, however, of the  sensitivity of
performance to overflow rate.  The test  with no coagulant addition proved to
be the least sensitive to overflow rate.   this sample, it should be noted, had
the highest initial solids  concentration, and high-solids samples seemed to
clarify  better throughout  the testing.  The metallic coagulant provided the
highest  overflow rate which would meet the pretreatment target of 20 mg/1
suspended solids  (see Table 3), but proved to be the most sensitive to changes
in overflow rate.  This alternative also generated larger  quantities of a
            d i f f ioul t,-to-dewatfr sludge.
                                     314

-------
                                   Table 3
              OVERFLOW RATES REQUIRED TO MEET TARGET
            Test No.
               1
               2
               3
Initial TSS
 (mg/1)
  492
  298
  366
                                Required
                                Removal
 96
 93
 95
              Overflow Rate
          l/min/m2    (gal/ft2, day)
       8,954
       6,309
      16,077
(220)
(155)
(395)
          In  all  three cases, the required overflow rates  are substantially
less than the  normal  operating rate of 23,199 1/min/m2 (570 gal/ft2 day)  for
the existing clarifier.   This does not  imply  that the LST was improperly
designed, only that the high  removal percentages needed to meet the pretreat-
ment target  of 20  would  require extremely low overflow rates.  If the target
were raised to 40 mg/1,  the existing overflow rate would be adequate.  This is
illustrated by Figure 4 which  shows average effluents of 37 and 39 mg/1 during
the near-optimum operation which prevailed during October and November of
1979.
                                  Figure 4
             Suspended
             Solids (mg/1)
             200 i-
             150
             100
              50
                 J	I
                    MONTHLY AVERAGE LST-EFFLUENT
                           SUSPENDED SOLIDS
 I	I
                         M
M
J   J
1979
N
                                    315

-------
          The jar-test  program and settling tests indicated that  the exist-
ing operational scheme of  the full-scale pretreatment system was achieving
near-optimum results for a gravitational-sedimentation unit.   The resulting
effluent  quality, however, was still short of the target of 20  mg/1.  Supple-
mental  filtration was then evaluated as  a means of attaining the target.  The
results of  granular media  filtration  tests (Table 2) indicated that this
additional process would produce an effluent quality within the  preset target.
Although  filtration appears to be a feasible  technique  for attaining the
desired effluent solids  concentration, the validity of this  target must be
considered in light of full-scale operating experience.

REASSESSMENT OF  PRETREATMENT TARGETS
          The establishment of pretreatment targets must consider  the nature
of the material to be removed.  In the  Clairton case, the oil  and  grease was
found to  be biodegradable with no physical effects on the biomass. The only
effect of  these  materials on the activated sludge process is the oxygen demand
exerted during metabolism.  Since the average influent concentration is only
26 mg/1 or  283 Kg/day  (624 Ibs.), the impact on system oxygen resources is
minor when  compared to that of other contaminants.  Therefore, the original
oil and grease target of 20 mg/1 is not applicable in this instance.

          The relative  impact of  influent  total suspended solids must be
considered  in  light of volatile  solids generation taking place  within the
biological  system.  In the case of the  Clairton system, solids generation is
relatively large.  Typically, 5443-7257 kg (6-8 tons) of biological  sludge are
generated  daily. Assuming that precipitation is not a contributing  factor and
further assuming that all external inputs have been identified, the effect of
influent  TSS on equilibrium levels of mixed liquor volatile suspended solids
(MLVSS) can be readily  calculated.  Figure 5 illustrates the projected impact
on the  Clairton system.   It  is apparent  that little improvement in  % MLVSS is
to be realized in modifying  the existing  system to further reduce influent TSS
from  the  present value of  approximately 50 mg/1 to the previously projected
target  value of 20 mg/1.   Thus although achieving an extemely low level of
influent  TSS is possible using conventional  technology, in this case its
incorporation would certainly not be warranted.
                                    Figure 5

                      EFFECT OF INFLUENT SOLIDS ON
                  AERATION BASIN SOLIDS DISTRIBUTION
                Equilibrium Volatile
                Solids Content (%)
                100
                 90
                 80
                 70
Assume: 1. Volatile Solids Generation of 6 T/D
       2. Feed Flow • 1900 QPM
                                              I	I
                                J	I
                       0  10  20  30  40  50  60  70 80  90 100
                             Aeration Basin Feed TSS (PPM)

-------
IN-PLANT VERIFICATION

          In-plant  investigations  were initiated when the  volatile solids
content  of the mixed  liquor was  less than 45 percent.   Constraints imposed
by the mixing capability of the aerators and the low percentage of volatile
solids  limited the  equilibrium biological population as well as the equi-
librium  solids retention time.  Figure 6 illustrates the Clairton biological
system and its related  auxiliaries.  With the system equilibrated to a feed
TSS content of 50  PPM, the calculated MLVSS content of 90% (see Figure 5)  was
considerably different  than the observed value of 45%.  Obviously, a source of
non-volatile solids  in  addition to that entering with the feed water existed
within the system.   A careful review of operating conditions revealed  the
sludge wasting system to  be the source of the problem.
                                   Figure 6

                United States Steel Corporation  -  Clairton Works
                        CWTP BIOLOGICAL SYSTEM
                     CHEMICAL
                     ADDITIVES
                                               300.000 GAL.
                                               EQUALIZATION TANK
 ACTIVATED SLUDGE AERATION BASINS
           6.5 MM GAL.
                      FINAL EFFLUENT
                      TO RIVER
                                                                   WATER FROM
                                                                   NHjSTILLS
                                                                  TO SLUDGE
                                                                  DISPOSAL
                                                      VACUUM
                                                      FILTERS
GRANULAR
MEDIA
FILTER
          As shown  in Figure 6,  the rotary vacuum filters (RVF)  are used  to
dewater  sludge from  the  lime sludge thickener as well as bio-sludge  from the
aeration basins.   The  piping  for returning filtrate from the RVF system to the
aeration basins established a pathway for nonvolatile solids to enter the
aeration basins.   Although the filtrate  is expected to be low in TSS,  this did
                                      317

-------
"not  prove to be the case.  Solids which were deposited in the RVF cloth, but
which  were not removed from the cloth into the wastage bin were subsequently
washed from the cloth and  returned to the aeration basin.  Although the %
MLVSS  was shown to be relatively insensitive to influent solids (Figure 5),
the  correction of the internal solids-recycle  loop described above had a
pronounced effect  as illustrated in Figure 7.

                                  Figure 7

               AERATION BASIN VOLATILE SOLIDS CONTENT
             % VOLATILE SOLIDS	
               100

                90

                80

                70

                60

                50

                40
                     	—I—L_J_J—L_l—L_J—I-JL
                    J  F M  A
M  J  J
  1979
A S  O N  DIJ
F M
1980
          The  resulting improvement in %  MLVSS dramatically increased the
 total quantity  of  biological material  the  system was capable of retaining in
 suspension.   During the time frame in question, a series of process modifi-
 cations  were made to the treatment system.   These modifications  in conjunction
 with the improved  climate  resulting from the increased volatile solids frac-
 tion produced a significant change in system performance as typified by the
 improvement in nitrification (Figure 8).

                                    Figure 8

                       NITRIFICATION PERFORMANCE*
               % Nitrification
               100

                80

                60

                40

                20

                 0
             Gross Removal of
             NH& SCN- & CM-
                          MAM
        J    J
        1979
                   N
                                      318

-------
 SUMMARY

          Although  the  prescribed limits  of  neither TSS nor 0/6 were met,
 the  system was able  to provide nitrification in a single  stage system.  Is
 is concluded that  the  importance of pretreatment in terms of TSS and 0/G
 removal must be considered on an individual basis for each  application.  The
 nature and composition of  so-called "oil/grease" must be  determined before
 speculating as to  its  impact.   Similarly, the  relative quantity of daily
 sludge generation must  be considered  before establishing what would be termed
 "acceptable" limits  on influent total suspended  solids.   Based upon our
 experience at Clairton Works,  it is concluded that reasonable operation of
 conventional pretreatment equipment  yields an influent of sufficient quality
 to permit maintenance  of an optimum biological system.
                                REFERENCES


JL  Grutsch,  J.  P., and Mallatt,  R.  S., "Optimize the Effluent System",
    Hydrocarbon Processing,  76,  105, (March 1976).

2.  Rich, L. G.,  Environmental Systems  Engineering,  336, McGraw-Hill, New
    York, (1973).

3.  Crame, L. W. ,  "Pilot  Studies on Enhancement  of the Refinery Activated
    Sludge Process",  API Publication 953 (October 1977).

4.  Stenstrom, M.  K.,  and  Grieves, C. G., "Enhancement of Oil Refinery Acti-
    vated Sludge by Addition of  Powdered Activated Carbon", Proceedings of the
    32nd Annual Industrial Waste Conference, Purdue  University, Ann Arbor
    Science Publishers,  Ann  Arbor, MI,  196 (1978).
                                    319

-------
             PROCESS CONTROL FOR ACTIVATED SLUDGE TREATMENT
                        OF COKE  PLANT WASTEWATERS
                                   By
                      ANDREW C. MIDDLETON, MANAGER
                   WATER  QUALITY  ENGINEERING  SECTION
            ENVIRONMENTAL RESOURCES AND OCCUPATIONAL HEALTH
                         KOPPERS COMPANY,  INC.
                            MONROEVILLE, PA
ABSTRACT

Once an activated sludge system has been started up and brought to
steady operation, control methodology must be applied to it to maintain
it in such a state.  Additionally, during the life of the system
the situation may change from that of the original design, and the
control methodology must be capable of maintaining control in such
situations.  This paper presents a straightforward, rapidly interpretable
control methodology based on solids retention time (SRT); a methodology
for determining capacities of an existing system; and, an illustration
of these for a coke plant activated sludge system.

-------
             PROCESS CONTROL FOR ACTIVATED SLUDGE TREATMENT
                       OF COKE PLANT WASTEWATERS

INTRODUCTION

Once an activated sludge system has been  started up and it has pro-
gressed to a controllable system, control methodology must be applied
to maintain it in such a state.  A process control methodology will
be described here that is used at Koppers* facilities.  This method-
ology is not complex and provides a base  to evaluate the operational
state of the system.

PROCESS CONTROL METHODOLOGY

Table 1 is a step-by-step list detailing  the control procedure.
This procedure is based on the use of solids retention time (SRT)
as the primary independent control variable.  The solids retention
time (SRT) is the average amount of time  suspended solids are retained
in the system, and  it has been shown to be a rational, convenient
                                                    1 2
parameter for activated sludge design and operation.    Only the
following measurements are required:  aeration tank mixed liquor
suspended solids (MLTSS and MLVSS), setting tank effluent total suspended
solids (TSS ), and  recycle sludge total suspended solids (TSS ).
With these measured data, the effluent discharge rate, and the aeration
tank volume, the rate of sludge wasting can be computed to maintain
a desired SRT.  A normal operating procedure would be to make these
computations daily  and adjust the sludge wasting rate accordingly
at daily intervals.  In many cases, it is possible to make the adjustment
on five days per week, i.e., Monday through Friday, and leave the
sludge wasting rate constant at the fifth day's setting for the remaining
two days, i.e., Saturday and Sunday.

The procedure also  includes a step for estimating the required rate
of phosphorus addition.  This step is based on the principle of adding
                                   322

-------
                        ACTIVATED SLUDGE PROCESS
                               SRT CONTROL
                                 TABLE  1
1.   Select  the  desired  SRT based on  operational  experience.  Usually
     20 days is  reasonable for BOD, TSS, phenol,  oil & grease, and
     sulfide removal at  temperatures  ranging  from 15 to 20 C.

2.   Calculate the weight of  total suspended  solids (TSS) in  the
     aeration tank daily as:

          M in Ibs = (MLTSS)  (V) (8.34)

     where MLTSS = aeration tank total suspended  solids concentration
     as measured, mg/1; V » aeration  tank volume, million gallons;
     and, 8.34 » conversion factor, mg/1 to Ib/million gallons.

3.   Determine the total weight of TSS to be wasted daily as:

                                    M
          Ibs to waste per day
                                   SRT
     Calculate the total weight of TSS lost daily in the clarifier
     overflow as:

          L in Ibs lost per day - (Q) (TSSe) (8.34)

          where Q = effluent TSS as measured,  mg/1 (composite sample
          if possible).
                                                             continued
                                    123

-------
5.   Determine the weight of TSS to be intentionally wasted daily
     as:

                                                  M
     PI in Ibs to intentionally waste per day "  __- -L

6.   Determine the volume of recycle sludge to be drawn off daily
     to achieve this intentional wasting of TSS as:

          w in million gallons = (TSSJ(8.34)

     where TSS = recycle sludge TSS concentration as measured, mg/1.

7.   Determine the phosphoric acid addition rate based on the production
     of volatile suspended solids (MLVSS) as:
     Phosphoric
     .......   - _ .       - (MLVSS)(V)(8.3A)(P:MLVSS Ratio)
     Addition Kate in       	>_ _—-—-— , „ __ \
                                (P Content of H-PO,)
     Gallons per Day (gpd)
                                                               p
     where, MLVSS » VSS of aeration tank as measured, mg/1;
                                                             MLVSS
     • ratio phosphorus content of VSS, usually 0.024 is reasonable;
     and, P content of phosphoric acid « as specified, e.g., for
     75% H3P04 the P content is 3.13 Ib P/gallon.

-------
 the phosphorus removed by biomass losses from the system.   The biomass
 is estimated by the volatile suspended solids (VSS)  concentration.
 The VSS removal from the system can be estimated, and then based
 on the phosphorus content of the VSS and the phosphoric acid,  the
 required amount of phosphoric acid addition can be computed.   A value
 of 0.024 Ib P/lb VSS is suggested for the phosphorus content  of the
 VSS.   However, this ratio can be different depending on the system,
 and it should be adjusted as necessary based on operating  experiences.
 The soluble effluent phosphorus should be monitored  to provide a
 basis  for adjusting this computation.

 Table  2 is a recommended daily data sheet that facilitates this control
 methodology.   A data sheet similar to this one was originally  proposed
 for use with municipal activated sludge systems.

 CAPACITY ESTIMATION

 Once an activated sludge system has been brought  into steady operation,
 it is  desirable to adjust it to an optimal operation for the specific
 situation.   The operating situation will often be different from
 that of the original design and may change during the life of  the
 system.   Wastewater flow rates and compositions will vary  with production
 rates  and process  changes.

 For an existing system the solids  retention time  (SRT)  and recycle
 ratio  (r)  are  usually the only adjustable parameters.   The hydraulic
 retention time (HRT)  is  fixed by the feed rate  of wastewater to be
 treated and can be adjusted for process  control only if there  are
multiple  aeration  basins  available.  The day-to-day  variation  in
HRT due to  flow variation is  considered  not  to  be part  of  the  process
control  in  this  discussion.

As stated in Table  1,  a 20  day SRT  is normally  satisfactory for phenol,
fiOD, TSS, oil & grease, and  sulfide  removal  at  temperatures ranging
from 15 to 20°C. However,  there may  be cases where both higher  or
                                      525

-------
                        ACTIVATED  SLUDGE  PROCESS

                              SET CONTROL

                                TABLE 2
Operator
                                            Date
                             _days
1.   Operational SRT 	

2.   Mass TSS in aeration tank (M):

     MLTSS » 	mg/1

     M - (MLTSS(V)(8.34)

     M - 	Ibs

3.   Mass TSS that must be removed in M * SRT
 Aeration Tank Volume

 V " 	million
	gallons
                                                       .Ibs/day
4.   Mass TSS lost in settling tank effluent (L):

     Effluent TSS - 	mg/1

     L - (Q)(Effluent TSSX8.34)

     L -	Ibs/day
                                                  Effluent Daily Discharge

                                                  Q - 	million
                                                             gallons
                                                             per day
5.   Mass TSS to intentionally waste (PI):

     PI - (M * SRT) - L

     PI -            Ibs/day

6.   Waste sludge flow rate (from recycle sludge lines)(w):

     Recycle Sludge TSS (TSSr) 	mg/1


               PI
          (TSSr)(8.34)
                         x 1,000,000
                                                     P   Ratio
                                                   MLVSS
                                                  P Content of
                                                  Phosphoric Acid -
                                                  	Ibs/gal
7.   Phosphoric acid addition required:

     MLVSS - 	mg/1

     MLVSS Production - (MLVSS)(V)(8.34) f SRT

     MLVSS Production • 	Ibs/day

     Ib P required - (MLVSS Production) (   P
                                          MLVSS
                                                 Ratio)
     Ib P required
                                    _lb/day
     Phosphoric Acid addition rate - (Ib P required) r ( P content)

     Phosphoric Acid addition rate - 	gpd
                                    326

-------
 in operation, it is ususally best to operate at as high an SRT as
 system capacity permits.  At higher SRT values sludge production
 and hence, sludge wasting and disposal, are less. Thus, any costs
 associated with sludge disposal are reduced.

 In addition to reduced sludge production, operation at longer SRT
 values maintains a higher biomass in the system.  The higher biomass
 provides greater stability to transient conditions because there
 is more biomass to respond to loading increases.  Figure 1 is a graph
 showing the relationship of MLTSS, MLFSS, and MLVSS concentrations
 to SRT for an operation system.   The appropriate equations are shown
 on the graph.   Equations used in this paper have been presented else-
                                                     145
 where and, they are not commented on in detail here.  '  '    The trend
 of higher concentrations at longer SRT values is illustrated by this
 graph.   Finally,  effluent quality due to biological oxidation improves
 with increasing SRT.   In particular,  soluble phenol and BOD concentra-
 tions would decrease  at longer operating SRT values.

 Two factors will  limit the maximum SRT value that  can be achieved
 by an operatng  system for a given wastewater loading: 1) settling
 tank  capacity;  and, 2) aerator capacity.

 The activated sludge  settling characteristics must  be knpwn  to evaluate
 the settling tank capacity.  These characteristics  are  determined
 by performing zone  settling  tests  on  the  activated  sludge  after steady
 system operation has been achieved.  An illustrative example will
 be provided here based on experiences at Koppers facilities.

 Figure 2  shows the zone settling velocity (ZSV) of  activated sludge
 at  two Koppers facilities as a function of its TSS concentration
 as measured by zone settling tests.  Plant A is a by-product coke
 plant producing foundry coke at the time of settling tests.  Wastewater
pretreatment included  ammonia removal from the excess flushing liquor
using a lime still and equalization in a feed tank.  Plant B is a
coal tar distillation plant.

-------
oo
                    c
                    o
                    a>
                    o
                    c
                    o
                    u

                    co
                    •o
                    o
                    en
                    (U
                   -O
                    ex
                    CO
                    3
                   CO


                   •a
                    (0
                   H

                    e
                    o
                    to
                       12
                       10
Equations



  MLTSS = MLFSS + MLVSS

                01

  MLFSS = FSS
              (SRT,
                                                           S -S A
                                                            e  nd
                             MLVSS
                VHRT


          T
SRT(Yk-b)-!
        l+b(SRT)


S_= 1097 mg/1


    = 160 mg/1
                                                                                     MLTSS
   Q


  FSS
  HRT

  Y =

  b -

  k =
    =1.6 days

    0.30     .

    0.06 day'1

    2.6 day"1
                             K = 240 mg/1
                             ,3
   nd
        110 mg/1
                                  10
                 20       30       40        50       60



                    SOLIDS RETENTION TIME (SRT), DAYS
                                                                                            70
                                                                        80
                         Figure 1.  Relationship of MLTSS, MLFSS, and MLVSS To  Solids  Retention Time.

-------
    80—
5
O
O
iH
0>
M
C
0)
W

0)
c
O
                                               Plant A  ZSV = 72i°*23 TSS


                                           Q  Plant RZSV= 27i°-24 TSS
                              A           6            8           10

                                 TOTAL SUSPENDED SOLIDS  (TSS),  g/1
12
14
     Figure 2. Relationship of Activated Sludge Zone  Settling Velocity  To
               Total Suspended Solids Concentrations.

-------
The settling characteristics are described by the equations shown
on the graph.  The ZSV decreases exponentially as the TSS increases.
The difference in the settling characteristics is probably due to
the differing MLVSS/MLTSS ratios.

There are two parameters that can be used to evaluate the loading
condition of a final settling tank:  (1)  the surface overflow rate
(SOR), and (2) the solids loading rate (SLR).

These parameters are defined as follows:

SOR = final settling tank effluent discharge rate, gpd                (1)
                                      2
      effective upflow surface area,ft
SLR = mass flow rate of mixed liquor suspended solids to              (2)
      the final settled tank, Ib/day _
                                         2
      effective downflow surface area, ft
The SOR has a maximum allowable value (SOR   ) that depends on the
                                          max         r
mixed liquor total suspended solids concentration (MLTSS) and the
settling characteristics.  Figure 3 shows the SOR    as a function
of the MLTSS based on settling characteristics described by the equations
in Figure 2. The appropriate equation is shown on Figure 3.

The SLR has a maximum allowable value (SLRmax) that depends on the
final settling tank underflow total suspended solids concentration
(TSS ) that is being attempted to be maintained.  This concentration
depends on the MLTSS and the ratio of the recycle sludge pumping
rate to the final settling tank effluent discharge rate; this ratio
is the recycle ratio (r).  The equation describing the TSSr as a
function of the MLTSS and independent variables is as follows:
                                                                        <3>
     TSSr .        (1+r -
In many specific cases, the ratio HRT/SRT is relatively small compared
to the (1+r) term, and it can be neglected without sacrifice of signif-
                                     #30

-------
             5000
u»
         •O
         a.
         60
          *  4000
         01
             3000
Vi
2
o
         jj   2000
         01
         JO

         I
         -J   1000
                               I
                                                                   -k  (MLTSS)
                                             SOR
                                                                  max
                                                               ZSV e
                                                                  o
                                                                              10
                                                                                   12
02468

                                      MLTSS, g/1

Figure 3. Relationship  of Maximum Allowable Surface Overflow Rate  To MLTSS
14

-------
leant accuracy.  The resulting equation is then as follows:
Figure 4 shows the SLR    as a function of TSS based on settling
characteristics described by equations in Figure 2.  The appropriate
equations are shown on Figure 4.  As can be seen in the graph, the
SLR    is relatively sensitive to the TSS . Hence, reckless changes
   max                                   r
in the recycle ratio can result in unexpected SLR    values that
                                                 max
may result in an overloading of the settling tank.

The effective upflow surface area of a circular clarifier is the
surface area of the annulus between the peripheral overflow weir
and the center feed well corrected for 'non-ideal flow patterns.
For circular center feed clarifiers a correction factor of 0.55 can
be used, i.e., only 55 percent of the actual area is effective. The
effective downflow surface area is the entire surface area of the
final settling tank again corrected for non-ideal flow patterns.

The required final settling tank area (A) is a function of the MLTSS,
flow rate (Q), recycle ratio (r), and settling characteristics (k ,
ZSV ).  The following equations apply:

     A * max (Aclar' AZsr>

     Aclar - <>/SORmax

     Azsr * (Hr)(Q)(MLTSS)/SLRmax                                      (7)

Figure 5 shows the relationship of required final settling tank area
(A) to MLTSS and recycle ratio for settling characteristics and flow
rate applicable to Plant A.  In general, the value of A increases
with increasing MLTSS and decreasing r.  The portions of the graph
not within the capabilities of the system are shaded.  The upper
horizontal boundary is the available final settling tank area (950
  2
ft  upflow).  The curved lower boundary is the minimum area required
                                    332

-------
   250
cfl
TJ
   200
  X
  0)
 e
   150
 00
 e
•H
•a
 ct)
 o
 (0

T)
 O
CD
   100
 §
 XI
 CO
    50
                          -k  (TSS  )

SLR    = (ZSV )(k ), g/1
50
       Figure 4.  Relationship of Maximum Allowable Solids

                 Loading Rate to Recycle Sludge TSS.
                            333

-------
   1200
   1000
         Available Area
HI
s-
•H
3
—
U
    800
    600
H   400

00
c
    200
• -
-
           Recycle
           Ratio
Q = 0.174 MGD
q    = 0.216 MGD
 TU.
               nax
              ISM  =  72e°'23 TSS
                                                                           14
        Figure 5.   Relationship of Required Final Settling Tank Area  to MLTSS  and
                   Recycle Ratio (r).
                                                                     16

-------
 at the maximum recycle ratio (1.24;  the recycle sludge pump is sized
 for 150 gpm).   Any combination of MLTSS and r not lying in the shaded
 portions is within the system's capabilities.  The value of r would
 be adjusted by changing the flow rate of the recycle sludge pump;
 the value of the MLTSS would be adjusted by changing the SRT as re-
 quired according to Figure 1.

 The oxygen uptake rate varies  with the SRT value because the degree
 of biological  oxidation of influent  pollutants and the microbial
 respiration are dependent on SRT.  Hence,  for an activated sludge
 system with fixed aeration capacity,  the aeration tank dissolved
 oxygen (DO) concentration will vary  with SRT.  Figure 6 shows this
 variation for  Plant A and the  conditions listed.   The aeration tank
 DO decreases with increasing SRT.  The decrease is relatively sharp
 at SRT values  less than 10 days,  and  more  gradual at higher values.
 At SRT values  less than 10 days for  this situation,  the degree of
 oxidation increases greatly with  increasing SRT giving a sharp increase
 in oxygen consumption and thereby  the sharp decrease in aeration
 tank DO.   At higher SRT values the increase in oxygen uptake is more
 dependent on microbial respiration which increases with increasing
 bacterial mass  as shown in Figure  1.

 Normally,  for  coke plant wastewaters  a minimum aeration tank DO of
 2.0 mg/1  is recommended to sustain consistent,  reliable phenol and
 BOD. removal.   A vertical boundary is shown on Figure 6 at the SRT
 value  where the aeration tank  DO equals  20  mg/1.   The region of the
 graph  to  the right is  shaded as being beyond the  recommended limits
 of the  system.   Hence,  a constraint is placed on  the allowable operational
 range  of  the SRT in addition to those due  to settling characteristics.

 EXAMPLE

A  series  of  graphs  have  been presented showing various  parametric
relationships for  an existing  activated  sludge  system (Plant A).
This system was  designed  to  operate at a 20  day SRT  and  a  recycle
ratio of  1.0.  For  a hydraulic  retention time  (HRT)  of  1.6  days,
                                      335

-------
o
Q
•o

I
rH
C
00
a
-
i-1

CO
    10
    0
Equations


    D0 = ^
                     K + Q
                                                (P)(Nstd)(1.024)T-20(ct)
             Vr
               DO
                                  1.42Y
                                              PC
                  S   + Ks(l+b(SRT)
                                                                 Q - 0.174 MGD

                                                                 SQ= 1097 mg/1


                                                                 Y = 0.30

                                                                 b = 0.06 da%~

                                                                 k = 2.6  day
                                                                   =  240 mg/1
                                                                 nd
                                                                N
                                              2.5 Ib/HP-hr
               10
                20
30
                                            40
                                             50
                            SOLIDS RETENTION TIME  (SRT),  DAYS
      Figure 6.  Relationship of Aeration Tank Dissolved Oxygen Concentration to

                 Solids Retention Time.

-------
 an influent COD (S ) of 1097 mg/1,  an influent FSS (FSS ) of 160
 mg/1,  and a flow (Q) of 174,000 gpd,  the graphs indicate the system
 to be  operating well within its capabilities.   Figure 1 gives a MLTSS
 of approximately 3600 mg/1 for an SRT of 20 days.   Figure 5 gives
                                                     2
 a required settling tank area of approximately 60  ft  for a MLTSS
 of 3600 mg/1 and a recycle ratio of 1.0.   Figure 6 gives an aeration
 tank DO of approximately 2.6 mg/1 for an SRT of 20 days.

 The available settling tank area is in excess  of that required.
 The difference is due to the measured full-scale settling character-
 istics  being superior than those used for original design.   The  differ-
 ence is apparently due to a large amount  of inert  fines flocculated
 with the sludge at the time of the  settling tests.  These fines  in-
 creased the specific gravity of the floe  particles and hence,  the
 settling velocity.

 Given  this series of graphs,  the following questions could  be  posed:

     1.    What is the maximum SRT and minimum  recycle ratio that
           this system could be expected  to be  operated at?

     2.    What would be  the advantages of making such changes?

 To  begin with,  determine the  maximum  SRT  allowable due to aeration
 constraints.   On Figure  6,  this  value is  approximately 43 days to
 maintain an aeration tank DO  of  2.0 mg/1.   Use  a value of 40 days
 so  that  some excess  remains.   On Figure 1,  an SRT  value of  40 days
 results  in a MLTSS value of approximately 6100  mg/1.   On  Figure  5*
 a MLTSS  value  of 6100  mg/1  gives  a  required  settling tank area of
 950  ft2  at  a recycle  ratio  of  slightly greater  than  0.20.   Use a
 value of  0.30  so that  some  excess remains.   Hence,   it  would not  be
 unreasonable  to  attempt  a state with  an SRT  of  40  days  and  a recycle
 ratio of  0.30. It would  be  recommended that  operational changes  to
 go  from  an  SRT 20 days to 40 days and  a recycle  ratio  of  1.0 to  0.30
be done  in  stages, i.e.,  to 30 days and 0.6  for  several weeks, then
 to 40 days  and 0.3.
                                    337

-------
After each change of  state  is made,  the  system should be  observed
for deleterious effects at  the new  state.  The graphs are approxima-
tions of reality, and unexpected  effects can  occur.  The  following
are some possible effects that may  not be  accounted  for in the  graphs:

     1.   Some nitrification may  occur at  the longer SRT  values re-
          sulting in  increased oxygen uptake  rates and a  more rapid
          decrease  in aeration tank DO values.  Significant nitrifi-
          cation has  not normally been observed to occur  in undiluted
          coke plant  wastewaters  at SRT values less  than  50 days
          without the use of activated carbon addition to the aeration
          tank.

     2.   Reducing  the recyle ratio increases the residence time
          of the recycle sludge in  the final  settling tank.  If some
          nitrification occurs, the longer residence time may allow
          reduction of nitrate to nitrogen gas (denitrification)
          within the  sludge causing it to  float to the surface  of
          the settling tank.

     3.   An overloaded condition for the  aerator or settling tank
          can occur before  it is  expected  because the graphs may
          not be as precise as expected.

If any of the above effects are observed,  the correctional procedure
is to reverse the direction of the  change  until the  effect disappears.
It should also be noted that the  change of state could go beyond
what is predicted by  the graphs.  If there were no deleterious  effects
at a 40 day SRT and a recycle ratio of 0.30,  it would be  recommended
that the SRT be increased and the recycle  ratio be decreased further
in smaller increments until such  effects, were observed.

At the longer SRT values, the following changes  would be  expected:

     1.   improved soluble effluent quality,  i.e., lower  phenol and
          soluble BOD,, concentrations.
                                     338

-------
      2.   decreased sludge mass production.

      3.   decreased phosphoric acid additions*

 At the lower recycle ratio values,  an increased recycle sludge TSS
 (TSS^) would be achieved.   An :
 waste sludge volume flow rate.
(TSS ) would be achieved.  An increased TSSr would result in a decreased
 The improved effluent quality is difficult to quantify,  because of
 the lack of detailed biokinetic studies on coke plant wastewater.
 An example of the possible improvement at these SRT values might
 be a reduction in average effluent phenol concentration  from 0.9
 ppm to 0.6 ppm or thereabouts.

 The decreased sludge production can be estimated using the SRT control
 procedure outlined in Tables  1  and 2.   Assuming no loss  of solids
 in the effluent for this  computation and using SRT values  of 20 and
 40 days,  MLTSS values of  3600 and 6100 mg/1,  an aeration tank volume
 of 0.278  mil gal  in Steps 1,2,3,4,5 of Table  2,  sludge production
 rates  can be estimated.   The  results are 417  Ib/day and  353 Ib/day
 *or the 20 and 40 day SRT values,  respectively.   Using Equation 3,
 the recycle sludge TSS concentrations  can be  estimated.  The results
 «re 6912  mg/1 and 25,620  mg/1 for a 20 day SRT,  1.0 recycle ratio
 and a  40  day SRT,  0.3 recycle ratio, respectively.   The  waste sludge
 flow rates can be estimated using Step 6,  Table  2.   The  results are
 7233 gpd  and 1652 gpd for the two above states,  respectively.   The
 phosphoric acid addition  rates  can be  estimated  using  MLVSS  values
 from Figure 1  in  Step 7,  Table  2.   The results are  1.42  gpd  and 0.93
 gpd for the two above states, respectively.

 CONCLUSION

A  straightforward,  rapidly interpretable control methodology has
 been presented.  The methodology makes possible  the  rapid  identifica-
 tion and  control  of the state of operation of an activated  sludge
 sytem,  i.e.,  the SRT, HRT, and r value.  Procedures have been presented

-------
to estimate the capacity of an existing system for a specific situa-
tion.  The procedures make possible rational changes in the state
of a system, to approach an optimal operational state.

REFERENCES

1.   Lawrence, A.W. and McCarty, P.L., "Unified Basis for Biological
        Treatment Design and Operation."  Jour. San. Eng. Div., Proc.
        Amer. Soc. Civil Engr.. 96, 757 (1970)

2.   Jenkins, D. and Garrison, W.E., "Control of Activated Sludge
        By Mean Cell Residence Time."  Jour. Water Poll. Control
        Fed..    40, 1906 (1968).

3.   Smith, D.A. and Alessi, C.J., "SET Control:  A Case History,"
        in Design and Operation of the Activated Sludge Process,
        WREE Research Report No. 78-2, Dept. Civil Engr., SUNY/Buffalo,
        Buffalo, NY (1978).

4.   Middleton, A.C. and Lawrence, A.W., "Least Cost Design of Activated
        Sludge Systems."  Jour. Water Poll. Control Fed.. 48, 889
        (1976).

5.   Middleton, A.C. and Lawrence, A.W., "Least Cost Design of Activated
        Sludge Wastewater Treatment Systems."  EPM Tech. Rep. 75-
        _!, Dept. Env. Engr., Cornell Univ*, Ithaca, NY (1975).

NOTATION

                                               2
A         - final settling tank surface area, L

b         - decay coefficient, T

 FC                                               3
C         - saturation DO at field conditions, M/L

C*°       - saturation DO at 20°C, M/L3
                                    340

-------
TSS       - total suspended solids concentration of final settling
   r                           3
            tank underflow, M/L


                                   3
V         - aeration tank volume, L
          - oxygen transfer coefficient



          - DO saturation coefficient
                                   341

-------
                                               3
DO        - dissolved oxygen concentration, M/L



FSS       - influent fixed suspended solids concentration) M/L



k         - maximum specific COD utilization rate, T


                                             3
k         - empirical settling coefficient, L /M



K         - Monod coefficient, M/L
 8


MLFSS <    - mixed liquor fixed suspended solids concentration, M/L


                                                                  3
MLTSS     - mixed liquor total suspended solids concentration, M/L


                                                                     3
MLVSS     - mixed liquor volatile suspended solids concentration, M/L
H   .      - standard aerator oxygen transfer capacity, Ib 0_/HP-hr



P         - aerator power



p         - atmospheric pressure, mm Hg



Q         - influent flow rate L3/T



q         - recycle sludge pumping rate, L /T



r         - recycle ration (q/Q)



r_0       - volumetric oxygen uptake rate, M/L -T


                                           3
Sg        - effluent COD concentration, M/L


                                           3
SQ        - influent COD concentration, M/L



T         - temperature, °C
                                   342

-------
BIOLOGICAL TREATMENT OF COKE PLANT WASTE
USING AN  INTEGRAL CLARIFICATION  CONCEPT
              Myrl R.  Wear
            James A. Grants
           Ronald J.  Thompson
              Arroco, Inc.
       Environmental Engineering
            Middletown,  Ohio
                   343

-------
Introduction
     Armco's Hamilton plant is located in New Miami, Ohio, on the Great Miami
River.  The plant produces molten pig iron, metallurgical coke, coke gas, and
coking by-products.
     The Hamilton Coke Plant consists of four Koppers-Becker underjet design
batteries with a, total of 110 ovens.  The oldest battery was constructed in
1928 and the newest started up in 19U7-  During this period the batteries
have been rebuilt several times.
     In 1976, Armco initiated an extensive modification and rehabilitation
program for all four coke batteries.  A major part of this program was the
installation of state-of-the-art air and water pollution control facilities.
The water pollution control program included the collection and treatment of
sanitary sewage, ammonia still waste, benzol plant waste, quench tower waste,
and high temperature noneontact cooling water.  This paper specifically deals
with the treatment of the sanitary sewage, ammonia still waste, and the
benzol plant waste.
Background
     The wastewaters generated at Armco's Hamilton Coke Plant are primarily
from the flushing liquor system and the benzol plant.  The flushing liquor is
hot water which is sprayed directly into the collecting mains to quench coke
oven gas as it leaves the ovens.  A circulating liquor system is used to cool
the gas in direct spray primary coolers.  Water evaporated from the coal is
condensed in the main and primary coolers creating excess flushing liquor.
This highly contaminated excess liquor from the two systems is collected in
storage tanks prior to treatment.
     Several sources of wastewater from the benzol plant are collected in a
common oil separation sump.  The largest source is condensate from wash oil/
crude light oil distillation operations.  The wash oil is purified by steam
stripping to remove crude light oil that was absorbed in the light oil gas
scrubbers.  The steam condensate is discharged to the benzol sump as a con-
taminated waste stream.  See Table I for the design volumes and chemical
composition of the raw excess ammonia liquor and benzol plant wastewater.
     At the outset of the program, a study was conducted to determine the
best approach for treating coke plant wastewater.  Alternatives studied and
                                     344

-------
                                      TABLE I

                       RAW WASTE WATER - DESIGN COMPOSITION
                             Excess  Ammonia Liouor             Benzol Plant
                         mg/1     kg/day     (#/da.y)          mg/1  kg/day

AVERAGE FLOW r,Pn                 58,300                               U3,100

Anmonia-N
Cyanide
Oil & Grease
Phenol
Sulfide
Suspended Solids
COD
TOG
PH
1*625
25
il*oo
10
11l|0
23
60
8180
2350

1020
5.5
310
1.8
250
5
11*
1820
520
9.0
(221*7)
(12)
(680)
(U)
(550)
(11)
(30)
(1*000)
(111*0)

13
19
18
1*5
111*
11
l*o
550
5Uo

2.3
3.2
2.7
7.3
19
1.8
6.1*
90
88
7.2
(5)
(7)
(6)
(16)
(M)
(U)
(1U)
(200)
(195)

                                     345

-------
rejected included physical-chemical treatment with activated carbon and joint
treatment in a publicly-owned treatment works (POTW).  The physical-chemical
scheme offered a lower capital investment but a much higher operating cost
and was thus rejected.  The joint treatment scheme was rejected because of
the remote location of the coke plant and the unique configuration of the
POTW.  Thus, a combination of physical-chemical treatment followed by bio-
logical treatment was chosen.  The major treatment objective was compliance
with NPDES permit requirements.  In order to meet this objective, each major
waste stream had to be pretreated by physical-chemical methods to remove in-
compatible pollutants prior to biological treatment.
Physical-Chemical Pretreatment
     The benzol plant waste contains large quantities of oil.  During the
original survey the majority of this oil was free or floatable oil with less
than 30 mg/1 of emulsified oil.  However, the light oil recovery operation
has since been modified, and the waste stream now contains 800 mg/1 of emul-
sified oil.  The free oil is partly removed in the existing oil separation
sump.  However, this sump is not capable of handling large oil spills.  Be-
cause of the potential detrimental effect of large quantities of free oil on
the biological treatment plant, additional oil removal equipment was
installed to help contain spills.  This equipment consists of a prepackaged
gravity oil/water separator.  The separator is installed in series with the
existing sump and designed to remove free oil that passes through the primary
separation  tank.  After treatment in this separator, the water is pumped to
the biological treatment plant.  To date, the emulsified oil has caused "no
apparent problems at the bioplant.
     Excess ammonia liquor contains large quantities of ammonia, sulfide,
                                                                   Cl 2}
cyanide, and other compounds which can inhibit biological oxidationv  * '.
The ammonia is present in two forms, commonly referred to as "free" and
"fixed" ammonia   •  Free ammonia, including ammonium hydroxide, ammonium
carbonate,  ammonium sulfide, ammonium cyanide, etc., is easily dissociated
and removed by steam stripping.  Fixed ammonia salts, including ammonium
chloride, ammonium thiocyanate, ammonium sulfate, etc.^' are dissociated and
removed by  raising the pH with an alkaline material and steam stripping.  To
enable the  final effluent to meet the NPEES permit requirements a steam
distillation system was installed to remove the free and fixed ammonia,
                                    346

-------
cyanide, and sulfide.  This system was chosen over other concepts being used
in the industry because of economics and site specific factors.
     The alkaline material used to dissociate the fixed ammonia at Hamilton
is caustic soda.  A solution of 50$ sodium hydroxide is injected directly
into the still without additional dilution.  Caustic soda was chosen over the
more traditional material, milk of lime, because of simplified operation and
fewer maintenance problems.  The addition of caustic requires only the in-
stallation of a storage tank and a metering pump, rather than the complex
feeding system required for lime, with a resultant lower capital cost.  The
pH at the top of the fixed still can be controlled, thus eliminating the
swings in pH at the bottom of the still due to the long lag time in the still,
and the problem of fouling the still with line has been eliminated.  Another
advantage, as yet not fully evaluated, is that caustic eliminates suspected
problems caused by high concentrations of calcium in the bioplant feed asso-
ciated with lime stills.  The major disadvantage of caustic is that it costs
several times as much as an equivalent amount of lime.
     An ammonia still originally installed in 1951* as part of the Middle town
Coke Plant and retired in 1976 was relocated to the Bamilton Coke Plant.  The
still contains 5> free ammonia (free leg) trays and 10 fixed ammonia (fixed
leg) trays.  The trays are all standard cast iron single bubble cap plate
sections.  The system is equipped to use 50f£ sodium hydroxide as the alkaline
material necessary to dissociate fixed ammonia.  A second "standby" ammonia
still was .installed for use during maintenance and cleaning of the primary
still.  The second still is a standard pressure vessel column with float
valve type, tower filler trays.  The ammonia still system includes an auto-
matic pH monitoring and control system which measures and records the pH of
the still discharge as well as controlling the amount of caustic fed to the
fixed leg.
     Excess ammonia liquor is injected into the free leg of the ammonia still
near the top of the still column.  After passing through the free leg, the
liquor is removed and caustic added and mixed with a motionless mixer.  The
high pH liquor is then injected into the fixed leg.  Low pressure exhaust
steam is injected at the base of the still, which is bubbled through the
descending flow of liquor to strip the ammonia.  The ammonia vapors and other
acid gases are collected at the top of the still, cooled to condense excess
                                     347

-------
 water*  and discharged into the  coke  gas  downstream of the  primary cooler.  The
 ammonia is later recovered from the  gas  as ammonium sulfate,  a by-product.
 The still wastewater with little remaining ammonia is pumped  to the  biological
"treatment plant.
 Design Considerations
      The biological system was  designed  from actual wastewater flows and
 analysis (Table EL  and III), data  in the literature^1 ^^'6j7>8'9%  and. in-
 formation gathered while inspecting most of the operating coke plant bio-
 logical treatment systems in North America.  The most significant findings of
 this predesign investigation were:  1) the need for extended  equalization,
 primarily to equalize the wide  fluctuation in wastewater chemistry;  2} a
 minimum aeration detention time of 2k to l$ hours; 3) the need for completely
 mixed aeration to minimize the  concentration of toxic parameters; U) the
 possibility of achieving both carbon oxidation and nitrification in a common
 aeration tank; 5) the need to add phosphorus to support bio growth;  6) the
 universal problem with aerator foaming;  and 7) the need to control pH and
 temperature.  In addition, the investigation uncovered an innovative clari-
 fication system that might be used to great advantage in coke plant bio-
 treatment.  The system incorporates an integral clarifier with scraper
 mechanism to direct the settled sludge back into the aeration basin.  In late
 1976, there were no similar operating systems in the United States, although
 several were operating in Canada.  In December 1976, a trip was taken to
 observe operating systems in Quebec^  ' and Ontario, Canada^    .  Following
 this trip, it was concluded that the concept offered several advantages, in-
 cluding potential capital cost savings, elimination of a separate sludge
 return system, potential for improved treatment, and substantial land sav-
 ings.  Therefore, in February 1977t the engineering firm of Burns & McDonnell
 was instructed  to proceed with the design of the biological treatment system
 utilizing this  integral clarifier concept with provisions to achieve nitrifi-
 cation and with the capability to add a second stage nitrification reactor
 should it be required*
      In November 1977* construction commenced with the clearing of the site.
 The work proceeded very slowly through the winter months, and  in March 1978,
 the first major pour of concrete was made.  The work progressed through the
 summer of 1978  and because of many delays caused by a wet spring and summer,
                                      348

-------
                                     TABLE II
                         BIOLOGICAL TREATMENT PLANT INFLUENT
 Flows
   Ammonia Still  Waste
   Benzol  Yard Sump
   Misc. Waste

      Total Process

   Sanitary
      Total Contaminated Waste Water

   Dilution Water
      Total Plant Feed
   Ammonia-N
   Cyanide
   Thiocyanate
   Oil & Grease
   Phenol
   Suspended Solids
   COD
   TOC
   PH
                                Design Composition
                                    kg/day     (#/day)
 61|,600 GPD
 1*3,100 GPD
 10,900 GPD

118,600 GPD

 21.800 GPD
1UO,1;00 GPD

 73.800 GPD
214,200 GPD
                        Actual Composition*
                    mg/1    kg/day     (#/day)
80
5
1;00
12
250
ho
1800
600

70
3.6
320
10
200
32
1500
500
7.0-10.0
           150
             8
           700
            22
            70
          3300
         (1100
115
  3
175
 U5
180
 70
           102,500 GPD

            35.100 GPD
           137,600 GPD

            61.UOO GPD
           199,000 GPD
 85
  2.3
130
 35
135
 55
190
  5
285
 75
300
120
**Detention  time  -  total feed          -  30 hours
**Detention  time  -  process  & sanitary -  U5 hours
                                 32 hours
                                 U6 hours
  Clarifier Overflow Rate-Total Feed
  Clarifier Overflow Hate-Process &
                          Sanitary

  Clarifier Weir Loading-Total Feed
  Clarifier Weir Loading-Process &
                         Sanitary
   300 GPD/SF
   200 GPD/SF
 1*,100 GPD/LF
 2,700 GPD/LF
            275 GPD/SF
            190 GPD/SF
          3,800 GPD/LF
          2,650 GPD/LF
 *0ctober 1979-March 1980:  6-month average of 30-day averages.
**Based on aeration tank volume (including chimney) of 265,000 gallons,
                                      349

-------
                                    TABLE III

                       BIOLOGICAL TREATMENT PLANT EFFLUENT
                   Design Composition              Actual Composition*
                  me/I   kg/day   (#/Day)         mg/1 kg/day    (#/day)

Flow:                  21k,200 GPD                    199,000 GPD
Ammonia-N          180     1l;5     (316)            7     6       (13)
Cyanide              U.5     3-6     (8)            1,5   1        (2.5)
Ihiocyanate          -                              U     3        (7)
Oil & Grease        12      10      (22)            5     U        (8)
Phenol               3-0     2.3     (5)            0.001 0.007    (0.01$)
Suspended Solids    70      60      130            32    25       (55)
pH                        6.0-9.0
*0cto"ber 1979 - March 1980s  6-month average of 30-day averages
                                      350

-------
 construction carried over into the winter of 1979*  Fortunately, the winter
 of 1979 was relatively mild and the plant was ready to be started in March of
 1979-
 Description of Facilities
     The treatment plant process flow schematic is shown on Figure 1.  The
 incoming waste is received at the plant in one of two surge tanks.  The pro-
 cess surge tank,  which is 1*0 feet in diameter by lj.0 feet tall with a working
 capacity of 300,000 gallons, receives waste from the ammonia still system and
 the benzol plant.  The tank provides now and chemical equalization by oper-
 ating  at 50$ full and using the side entering mixer shown in Figure 2.  Pro-
 cess waste is normally stored in the surge tank for approximately one day
 "before it is pumped to the aeration basin.  The sanitary surge tank shown on
 Figure 3 receives the sanitary sewage generated in the coke plant area, and
 gas seal water from the flare stack.  This tank is agitated with air to
 insure a non-septic waste.  After equalization, the sanitary waste is com-
 bined  with the process waste and discharged into the aeration basin.
     The ammonia  liquor from the still is approximately 105°C and must be
 cooled to approximately 25°C»    for optimum ammonia removal in the bioplant.
 This cooling is achieved in part by flash cooling at the still and natural
 heat loss in the  surge tank, with the final cooling achieved in two parallel
 spiral flow heat  exchangers.  To compensate for the hot blower air used in
 the aeration process,  the liquor is cooled to  around 20 C during the summer
 months.
     In order to  maintain, the nutritional balance^  '  of the aerobic system,
 pho'sphoric acid is added to the liquor just after the heat  exchangers.  At
 the same point, sulfuric acid is added for pH  control should the  basin pH
 exceed the desired 7.8 operating point.   Following chemical addition,  the
 total  flow.is discharged into two parallel  aeration/clarification basins.
     The treatment system utilizes  a completely mixed,  activated  sludge
 extended aeration concept with an integral  clarifier.   The  aeration portion
 of  the plant  contains  two cubical concrete  basins each with its own clari-
 fier section.  Bach basin has  one submerged turbine aerator to achieve  com-
plete mixing and  oxygen  transfer.  Three rotary lobe positive displacement
air blowers shown on Figure k are used to generate  the  supply of air to the
submerged turbine aerator.  The aeration-clarification  configuration shown

                                     351

-------
0-Z90S
                                                                                                                                                                       Q
                                                                                                                                                                       (i)   i*»-
                                                           FIGURE  1.  ScheinMic  Diagram -  Treatment  Plant

-------
Figure 2.  Process Surge Tank with Side Entering Mixer


     Figure 3.  Control Building and  Surge Tanks
                          353

-------
Figure i|.  Positive Displacement Air Blowers
                      354

-------
 on Figure 5 allows the aerator to develop a horizontal velocity along the
 surface of the aeration basin.  This, in turn, causes a downward flow of
 approximately 10 times the once-through flow in the "chimney" or the space
•provided between the aeration section and the clarification section.  The
 majority of this flow passes back into the aeration basin, carrying with it
 the solids settled in the clarifier section.  This action provides a theo-
 retical recycle rate of 20096.  The flow into the clarifier is equal to the
 incoming waste.  This flow enters at the bottom along the length of the
 clarifier, passes vertically up through the basin, and is finally discharged
 from the system*  A scraper mechanism is used to move the settled sludge
 down the sloped clarifier bottom to the chimney area where it is resuspended
 and carried back to the aeration basin.
      The treatment plant graphic control panel (see Figure 7) enables the
 operator to monitor the physical operations of the plant including flows,
 temperatures, tank levels, etc., and to make minor corrections as required.
 The system continuously monitors the aeration dissolved oxygen, temperature
 and pH and automatically controls the latter two.  Based on these controls .
 and chemical analysis made in the treatment plant laboratory, adjustments
 are made to the system.
 Shakedown and Start-Up
      The checking and testing of the system conducted during February and
 early Kkrch of 1979 went smoothly.  The only major problems encountered were
 a bad vibration in the south aerator mixer and an excessive pressure drop in
 the air supply system.  The pressure drop problem was easily corrected by a
 modification to the sparge ring, but the vibration problem was not fully
                                                            t
 corrected until late August.  Additional construction delays were encountered
 in the ammonia still area and the benzol yard area so that those areas were
 not fully operational until July and September, respectively.
      On Ifarch 11*, 1979, the north basins and process surge tank were filled
•with clean water.  Waste ammonia liquor was then discharged to the surge tank
 which resulted in an extremely dilute solution of feed stock.  Approximately
 20 gpm of this dilute feed was pumped into the aeration basin which had been
 seeded with approximately 3000 gallons of activated sludge from Middle town,
 Ohio's, POTW.  During the following days, additional truckloads of sludge
 were pumped into the system until the mixed liquor was at 1200 mg/1 MLTSS.
                                     355

-------
HOUR!  5. 8oh«MMo of AMBllon / Oorlftartton
                     356

-------
Figure 6.  Waste Water Treatment Facility
    Figure  7-   Graphic  Control  Panel
                    357

-------
Because the Middle town. POTW treats wastes from Armco's Middle town Coke Plant,
the microorganisms were acclimated to coke plant waste.  As shown on Figure 8,
the total solids in the mixed liquor started to climb on day one and has con-
tinued to increase.  At the end of thirty days, the mixed liquor had increased
to 6000 mg/1 and as shown on Figures 9 and 10, the total solids inventory has
continued to climb while the volatiles have generally been above 70$.  The dip
in MLTSS concentration shown on Figure 10 is caused by the start-up of the
second aeration basin after correction of the vibration problems and install-
ation of a foam spray system.  While operating with only one aeration basin,
the system consistently removed phenol and thiocyanate with little or no
removal of ammonia*
Discussion of Operating Data
     Figure 11 shows the wastewater flows that have been treated in the first
year of operation.  The sharp increase in September is caused by the intro-
duction of the benzol yard waste for the first time and the use of river water
in a foam spray system.  Because the system experienced severe foaming while
operating on one basin, a river water spray header was installed around the
basins to help knock down the foam.  After the second aeration basin was
started, and as the MLTSS continued to climb, the foam subsided allowing
decreased use of river water during early 1980.
     The system has functioned well in treating phenol even during upset or
shock loading conditions.  Figure 12 shows the phenol loadings that have bfeen
treated.  The influent phenol monthly average has varied from a low of 80
kilograms per day (l80#/day) to a high of 230 kg/day (500#/day) with no effect
on the effluent quality.  The highest monthly average discharge to date has
been 30 grams per day (1 ounce per day).
     Thiocyanate removal has not been as spectacular as is shown on Figure
13*  Thiocyanate has proven to be the hardest parameter to remove and the most
sensitive to varying operating conditions.  For this reason, and the fact that
the wet chemical test for thiocyanate is easy, it has been used to determine
the relative health of the system;  Figure 11; shows the daily influent and
effluent concentrations during "upset".conditions.  In May, phosphate concen-
trations in the system were inadvertently depleted.  The phosphate levels were
undetectable for more than a week before corrective measures were taken.
Figure .14 shows the removal of thiocyanate was completely lost and recovered

                                      358

-------
ro
CO
CO
           16    18
II    13
 APRIL
                                                                                                  19   21   23
                                                         1979
                           Figure 8.  Mixed Liquor  Totol  Suspended  Solids During Start-Up

-------
Ol
            o
            X

            £•
            S
          20,000-
          I8.0OO-
          16,000-
          14,000-
          12,000-
   £
   S?
   78


   76


   74


   72


   70


   68


   56
IO,000-(
1.
o
.5"
*"j
^^
•o
5
8,000-
6.00O-
4,000-
)6



          2,000-
MAR
                                        % Volatiies
                       APR
                                       JUNE
  i
JULY     AU6
   1979
 i
SEP
 I
OCT
 I
NOV
 I
DEC
 I
JAN
                                                                                                        FE8     MAR
                                                                                                           1980
 7
APR
                                 Figure  9. Mixed Liquor:  Total BIO Mass Inventory 8%  Volatiles

-------
o
3
IT
16,800-



15,600-



14,400-



13,200-



12,000-



10,800-



9,600-



8,40O-



7,200-



6,000-



4,600-



3,60O-



2,400-
      I.2OO-Y
                    One Bosin Operation
Two  Basin Operation
                   •        i        i        i       i
          MAR     APR     MAY     JUNE     JULY     AU6

                                             1979
                                                       ii        i        i        r^      i        i        i
                                                     SEP     OCT     NOV     DEC      JAN     FEB     MAR      APR
                                                                                                 I960
                            Figure IP. Mixed Liquor  Daily Average Concentration M.L.T.S.S.

-------
Q
Q.
O

O
O
O
O
         MAR
APR
MAY
 1       I
JULY     AUG

   1979
 1       \
FEB    MAR
   1980
 I
APR
                          Figure  II.  Average  Monthly  Waste  Water Flows

-------
«*»
  o
  0)
  a.
                                                               Effluent
                                                               (G/Day)
           MAR
JULY     AUG

   1979
 I
FEB     MAR

   1980
APR
                             Figure 12. Phenols: Influent  ft  Effluent  Average  Daily Loadings

-------
                                                          Influent
o
210-




190-




170-




150-




130-




110-




 90-
o>

o
c
a
>,
o
o
                                                  Effluent
 50-




 40-




 3O-




 2O-




 10-
         MAR
           I

          APR
MAY     JUNE     JULY     AUG

                   1979
 I        I        I       I       I        I        I        I
SEP     OCT     HQV     DEC     JAN     FEB    MAR     APR

                                          1980
                           Figure 13.  Thiocyanate (SCN)= Influent  & Effluent Average  Daily  Loadings

-------
  c
  .2
**» 2
at c
en a
  >.
  u
  .o

  Je
                  6
e
                                                          1979


                             Figure 14.  Upset Conditions Due To  Lack Of Phosphorus  Nutrient

                                       As  Indicated  By  Thiocyonote

-------
slowly.  As Figure 13 shows, the treatment of thiocyarate has stabilized,
possibly due to the start-up of the second basin in September, with the
resultant increased bionass inventories, and no upsets have occurred since.
     During the first seven months of operation, ammonia removal, as shown on
Figure 15» was poor due to the lack of caustic in the ammonia still and opera-
tion of only one aeration basin.  The ammonia still system was started before
the caustic feed system and did not achieve fixed ammonia removal until late
May.  The bioplant operated on one aeration basin until September 15, when
the second basin was put into service.  During that four-month period, June-
September, ammonia removal was erratic and the system difficult to operate.
The plant would achieve nitrification, a drop in the pH would occur, and the
next day there would be no evidence of nitrification.  On September 1$, 1979,
the second aeration basin was put into service which stabilized the system.
Soon nitrification began to occur consistently and the system has achieved
excellent ammonia removal to date.  Since mid-October the system has had an
average influent loading of 85 kg/day (I90#/day) of ammonia nitrogen and 130
kg/day (285#/day) of thiocyanate and has discharged an average for the six
months of only 6 kg/day (l3#/day) of ammonia-nitrogen.  An indicator of the
stability of the system is shown on Figure 16, which shows a shock loading in
early March, 1980.  Around February 29, the last free tray of the ammonia
still became plugged with tar and pitch.  After minor modifications and tun-
ing, the standby still was put on stream March 5 and operated until March 18.
Although the standby still did not exhibit the removal efficiency of the
primary still, it operated well enough to bring ammonia loadings at the bio-
plant back into range, eliminating the need for backup storage lagoons or
other treatment.  During the first week of March, the treatment plant ammonia
feed was more than tripled to 3^0 kg/day (750#/iay)» yet the biosystem
achieved over 90$ removal of *Mrnnnia..  The highest quantity discharged during
this time was 30 kg/day (66#/day).
     Figure 17 shows the food-to-mass ratio expressed in kilograms of BOD5,
phenol, and ammonia fed to the plant over kilograms of mixed liquor volatile
suspended solids.  As is shown, the F:M ratios are currently very low, with
BOD5 in the range of 0.06 kg/kg ML7SS, phenol at 0.01 kg/kg ML7SS and ammonia
in the range of 0.015 kgAs MLYSS.  These low ratios are primarily due to the
high inventory of mixed liquor solids.
                                     366

-------
co
en
   O
   o
   e>
   CO
   a
   0)
   o>
   O
   a
   'E
   o
   E
   E
             MAR
                    APR
                            MAY
                                    JUNE
                                                1979
FEB     MAR

   1980
                              Figure  15. Ammonia  Nitrogen = Influent  and  Effluent  Total Daily Loadings

-------
 o
O
 o>
 o
 o
z
 I
 o
'c
 o
 E
 E
           15
17    19
21    23    25.

 FEBRUARY
                                                           (980
                            Figure 16.  Ammonia  Upset-  Influent  8  Effluent  Total  Daily Loadings

-------

-------
     The suspended solids in the effluent as shown on Figure 18 have gener-
ally been quite low.  With the exception of one peak period, solids in the
effluent have been less than UO kg/day (90#/day)» and have averaged 2$ kg/day
(55#/day) for the last six months.  In August and September, solids were very
high in the one operating aeration basin when flows through the basin were
high.  Subsequently, the sludge blanket in the clarifier section was nearly
at the water surface and a carry-over of solids resulted.  In January, a
sludge wasting program was initiated and approximately 35 kg/day (77#/day) of
solids are currently wasted.  Mass balance calculations indicate that bio-
logical growth has been equivalent to approximately 0.25 kg/kg of phenol
removal.  The excellent performance in the clarifier is in part credited to
the design of the integral clarification concept.  No polymers are added and
the flow receives no mechanical flocculation in the clarifier.  The absence
of sludge recirculation pumps has prevented breakup of the floe as it is
returned to the aeration basin.  The low effluent suspended solids may be
attributable in part to the use of sodium hydroxide in the ammonia still
rather than milk of lime.  This contributes a much lower inorganic solids
loading to the system, thus allowing the microorganisms to form a better floe
with less  "fine" solids carry-through  to the effluent.
Operating Training and Responsibilities
      The successful operation of  this  plant must be  credited  to outstanding
performance by the  operators.   The plant is under  the  direct  control  of one
day foreman with one operator present  around  the clock and an extra operator
on day turns.  All of  these men underwent  an  extensive  three-week classroom
training program,  with additional on-hand  training before  and during  the
start-up stages.   In addition,  the operators  are obtaining the required
experience and reviewing additional  training material  in preparation  for ob-
 taining State certification as licensed operators.   The  operators are conduct-
 ing the majority of the  chemical  analyses  with only cyanide and oil & grease
 analyses contracted to outside labs.   By m^rming the plant around the clock,
 the various  operations can be closely monitored and controlled.   Figures  19,
 20, and 21  show the close control that has been achieved on temperature,  pH,
 and dissolved oxygen.   The average temperature has been controlled at 25  C
 plus or minus 1°C throughout the  last year.   By close observation, the
 operators can make corrections in the cooling water system before troubles
 develop.  Even though the pH on Figure 20 varies from 6.6 to 8.3, the
                                     370

-------
   to
t*»  "O
-J-  —
•-1  o
   CO
   0)
   a.
   
-------
O
£
3
s
0)
                                                                     Daily Maximum
                          ---- x ---- x ----
                                                                 Daily Minimum    /
                                                             ^X ---- X-
                                                /
                                                     X"
         MAR
JUNE
 1       1       1
JULY     AUG     SEP
   1979
                                                             OCT
 1
NOV
 1
DEC
 1        1       1
JAN    FEB     MAR
          I960
 I
APR
                          Figure 19.  Aeration  Basin   Temperature

-------
        8.5-
        8.O-
                                             Maximum
                                               Day
                                                                    /\
                                                                  /     V
                                                                 /        \

to
       7.5-
o
•o
c
o
55
X
0.
70-
x ---- x
6.5-
                                             /
                                                  \
                                   /   Minimum
                                         Day
                                                           >     '
       6.O-
         MAR
          '
         APR
        I
      MAY
 i        1       1
JUNE    JULY     AU6

          1979
                                                      SEP     OCT     NOV    DEC
 1       1        1       1
JAN    FEB    MAR     APR
          1980
                          Figure20. Aeration  Basin Maximum  & Minimum  PH  Data

-------
operators were able to correct this by adding acid or alkaline materials to
bring the system back into specification.  The dissolved oxygen as shown on
Figure 21 is not usually a controlled parameter, but is used as a monitoring
tool.  The operator can detect changes in the D.O. which may signal a pending
upset and take corrective action.
Capital and Operating Cost
     The biological treatment system as originally constructed cost $2.15
million, with an additional $1*35 million for collection and ammonia still
systems.  An additional $1.5 million was spent for miscellaneous sumps, cool-
ing towers, and related projects, bringing the total project cost to $5.0
million.  Included in this figure is an estimated $800,000 for modifications
and relocations to retrofit the existing coke plant to accommodate the new
treatment plant.  Eased on the total contaminated wastewater design flow of
11*0,000 GPD, the cost of the biological treatment plant was $l5/gallon.  The
ammonia stills and collection system add $10/gallon, for a total capital cost
for the bioplant and associated pretreatment systems of $25/gallon.  Direct
operating cost for the treatment plant and the ammonia stills for the period
July-December 1979 was about $18/1000 gallons of process liquor treated, or
approximately $1.30/ton of coke produced.  The added cost of capital recovery
makes the treatment cost $30/1000 gallons of process liquor or $2.20/ton of
coke produced.
Conclusions
     The treatment of coke plant waste liquors to achieve phenol and ammonia
removal in a single stage reactor has proven to be a viable treatment method,
although expensive when used in series with a caustic soda ammonia still.
Control of pH has been the most difficult factor because of the formation of
acid in the treatment process and the destruction of the available alkalinity.
     Negative effects on nitrification or phenol removal by the introduction
of emulsified oil has not been a problem.  Emulsified oil in the effluent is
averaging less than 5 mg/1 with an average inlet loading rate of over UO
mg/1.
     Operating the system with the extremely high mi Ted liquor and long
sludge ages in the aeration basin has not been a problem.  During periods of
high now, some carry-over of solids is evident but there is no indication of
a problem during normal operations.
                                     374

-------
           6.O-
          8.O-
          4.0-
   0)
Co O
-~J C
Cn o
  O
  X
  O

  •o
  0>
  o
  w
  CO
          3.0-
          2.O-
          |JO-
                                                       Average
     Minimum  Day
 '*** ^
/     •"-^
                                        \  ^^
    	1	1	1—
MAR     APR     MAY     JUNE
                                           JULY     AUG
                                               1979
                                                SEP     OCT
                 i — — i - \ — — r - 1 — — r-
                NOV     DEC     JAN     FEB     MAR     APR
                                           1980
                              Figure 21.  Aeration  Basin  Dissolved  Oxygen  Data

-------
     Although the operations of this plant have been extremely smooth, a
degree of caution must be exercised if this data is to be considered for other
treatment plants.  At this writing, the plant has operated thirteen months
with only six months of satisfactory nitrification.  Nitrification has only
occurred during the winter months.  It is essential for complete demonstra-
tion of the plant to obtain a full year of operating data.
                                    376

-------
                                 References


 1.  Barker, J. B. and H. J. Thompson (1973) "Biological Removal of Carton
     and Nitrogen Compounds from Coke Plant Wastes."  Environmental Protec-
     tion Technology Series.  EPA-R2-73-167.

 2.  Eockenbury, M. R. and C. P. L. Grady, Jr. (197?) "Inhibition of
     Nitrification Effects of Selected Organic Compounds," Journal WPCT
     May 1977, 768.

 3.  Wilson, P. J. and J. H. Wells (1950) "Coal,  Coke and Coal Chemicals,"
     McGraw-Hill, New York.

 4.  Wong-Cheng, G. M. (1978) "Design and Operation of Biological Treatment
     for Coke Plant Wastewaters," September 1978 (AISI Study).

 5.  Wong-Chong, G. M. and S. G. Caruso (1977) "Advanced Biological Oxida-
     tion of Coke Plant Wastewaters for the Removal of Nitrogen Compounds,"
     April 1977 (AISI Study).

 6.  Wong-Chong, G. M. and R. C. Loehr (1974) "The Kinetics of Microbial
     Nitrification," July 1974, Presented at ASCE, BED Specialty Conference,
     Pennsylvania State University.

 7.  Ganczarczyk, J. J. (1977) "Pilot Plant Studies on Second Stage
     Activated Sludge Treatment of Coke Plant Effluent," October 1977 Report
     to AISA.

 8.  Ganczarczyk, J. J. and D. Elion (1978) "Extended Aeration of Coke Plant
     Effluents," May 1978, Presented at 33rd Purdue Industrial Wastewater
     Conference.

 9.  Ganczarczyk, J. J. (1978) "Pre-Treatment of Coke Plant Effluents."

10.  Charette, C. and J. Herbineaux (1977) "Chemical Products Plant Solves
     Problem of Contaminated Wastewater," October 1977, Water and Pollution
     Control.

11.  "Clarifier Designed for TTse With Surface Aerator," November 1973,
     Water and Pollution Control.
                                    377

-------
                    TREATMENT OF COKE PLANT WASTEWATER

                       USING PHYSICAL-CHEMICAL AND

                          BIOLOGICAL TECHNIQUES
                                   By
                Richard Osantowski and Anthony Geinopolos
                      Rexnord Inc. Corporate R&D
                          Milwaukee, Wisconsin
ABSTRACT:  Pilot studies were performed concurrently at two coke plants to
investigate the effectiveness of physical-chemical and biological treatment
in meeting steel industry BAT guidelines for the by-product cokemaking
subcategory.
                                     379

-------
                    TREATMENT OF COKE PLANT WASTEWATER
                       USING PHYSICAL-CHEMICAL AND
                          BIOLOGICAL TECHNIQUES
INTRODUCTION

The primary purpose of this project was to investigate the technical and
economical feasibility of treating by-product cokemaking wastewater to
Best Available Technology (BAT) levels using physical-chemical and
biological methods.  The wastewaters generated from the by-product recovery
process include excess ammonia liquor, benzol plant wastes and other
miscellaneous discharges associated with the production of coke.  Pollutants
contained in these wastewaters typically include suspended solids, ammonia,
phenolic compounds, cyanide, sulfide, thiocyanates, oil and greases as
well as many toxic pollutants.  Two plants were studied; the physical-
chemical test work was performed at Shenango, Inc., Pittsburgh, Pa.; the
biological study was conducted at the Wheeling-Pittsburgh Steel Corp.,
Follansbee, W.V..  The plants investigated had operating treatment systems
for upgrading the raw wastewater to a Best Practical Control Technology
(BPT) Currently Available level.

The investigations were conducted using the U.S. Environmental Protection
Agency's (EPA's) mobile physical-chemical and biological treatment systems.
These pilot plants are housed in three semi-trailer vans as shown in
Figures 1-3.


EXPERIMENTAL RESULTS  (PHYSICAL-CHEMICAL RESEARCH SITE)

General

The physical-chemical investigation on Shenango's by-product cokemaking
wastewater was conducted between November 14, 1979 and January  17, 1980.
During the study, coke production averaged 1,673 metric tons  (1,519 tons)
per day, while the average wastewater  flow was 1,025 m3/day  (0.271 mgd).
The corresponding water application rate  (liters of water/metrie  tons of
coke produced) during the study was 743 liters/kkg (178 gal./ton).  Based
on BAT limits, the allowable pollutant concentrations in the  effluent
would be:
                                                   BAT
               	Parameter	Limit
               pH                                6.0-9.0
               Ammonia, mg/1                       13.4
               Cyanide-T, mg/1                     0.33
               Oil and Grease, mg/1                13.4
               Phenol, mg/1                        0.27
               Sulfide, mg/1                        0.4
               Suspended Solids, mg/1                 27
               Thiocyanate, mg/1                    200
                                      380

-------
                                     Figure  1.
              TRAILER
              *5'U M B'W x I1'-6"H
SAMPLE
REFRIGERATOR
                                   CLARIFIER
                                                       REVERSE OSMOSIS
                                                          SYSTEM
       OZONE CONTACT
          TANKS
                                    Figure 2,
                                     SAMPLE
                                   REFRIGERATOR
(IOIOGICAI  TREATMENT
  SYSTEM NO. I
                       TtMPtRATURE
                       CONTROL  SYSTEM
              TRAILER
              *5'L x 8'V x I3'-6"H
         SIOLOCICAL TREATMENT
            SYSTEM NO.  2
                                    Figure  3,
                                           381

-------
The advanced waste treatment trains  that were investigated on a pilot scale
included  the following:
                                                   KEY
           1.   ACL + FIL + AC

           2.   ACL + SBD + FIL
                                        AC:   activated carbon
                                       ACL:   alkaline chlorination
                                       FIL:   dual media filtration
                                       SBD:   sodium bisulfite
                                              dechlorination

In the first  pilot treatment train, the  BPT wastewater was passed through
a two stage alkaline chlorination process for cyanide* phenol,  sulfide,
thiocyanate and  ammonia removal.  The wastewater was then filtered for
suspended solids removal and dechlorinated  on activated carbon.   The second
treatment train  again consisted of alkaline chlorination which was followed
by sodium bisulfite dechlorination and dual media filtration.  The treatment
train arrangements are shown in Figure 4.
         (I)  ALKALINE CHLORINATION, FILTRATION, CARBON ADSORPTION




NaOH -
^-


\l-
\
2?

t



H2SO<
\1
\
2?





        (Z)  ALKALINE CHLORINATION, SODIUM BISULFITE DECHLORINATION, FILTRATION

                       	       H,SOt
                    NaOH      i       '• *
                                              ta
                                                            KEY
                                                     Flu FILTRATION
                                                     AC:  ACTIVATED CARSON
                                                     SBO: SOOIUH BISULFITE
                                                           OCCHLORINATION
      Figure  A.
                  Process trains  investigated for treatment  of by-product
                               coke  plant wastewater.
Wastewater Treatment System

There are four  process water streams associated with the by-product coke-
making operations at the plant investigated; namely (1) final cooler waste-
water; (2) phenolate wastes; (3) light  oil separator effluent;  and (4) hot
oil decanter underflow.

The final cooler waatewater originates  from direct spray cooling of the

                                       382

-------
 coke  oven  gas and  represents about 31 percent of  the  total plant flow.  The
 phenolate  waste  stream  (excess flushing liquor) is passed through a free
 ammonia  still, dephenolizer and a fixed ammonia still prior to mixing with
 other plant wastes.  This stream comprises 47 percent of the combined coke
 plant flow.  The hot oil decanter discharge  (-19% of plant flow) passes
 through  a  dissolved-air flotation unit, containing 4.6 ia2(50 ft2) of surface
 area.  The underflow is then blended with other plant effluents.  The
 fourth major process stream, the light oil separator effluent, accounts
 for approximately  three percent of the coke  plant effluent flow.  The
 discharges from  these four principal coke plant wastewater sources are
 combined and blended in a 643 m3 (170,000 gal.) equalization tank equipped
 with  mechanical  mixers.  The equalized effluent, representing a BPT waste-
 water, is  then fed to a full-scale advanced  waste treatment system.  Feed-
 water to the mobile system was taken out of  the equalization basin during
 the entire study.

 Characteristics  of the wastewater obtained from the equalization basin
 during the study are shown in Table 1.
        TABLE 1.  COMPARISON OF BPT EFFLUENT LIMITATION GUIDELINES TO
             ACTUAL PLANT VALUES OBSERVED DURING THE STUDY PERIOD

Parameter
Cyanide T, mg/1
Phenol, mg/1
Ammonia, mg/1
Oil and Grease, mg/1
Suspended Solids, mg/1
PH
BPT1
Limit
29
2
123
15
49
6.0-9.0
Analytical Value From
the Research Site
Average
85.2
142
506
54
103
—
Range
4.7-267.5
68-850
101-2,150
3-147
26-361
3.8-10.8

       lDev. Doc., By-Product Cokemaking - EPA 440/l-79/024a, Oct. 1979

As shown in the table, concentrations of pilot system feedwater were well
above the BPT limits during the study.

Alkaline Chlorination Results

During the study, 40 alkaline chlorination test runs were performed.  The
pilot test procedure consisted of passing the coke plant wastewater through
a series of completely mixed reaction tanks under alkaline and neutral pH
conditions in the presence of an oxidizing agent (sodium hypochlorite).  In
the first reaction tank, sodium hydroxide and sodium hypochlorite were
added to the wastewater to oxidize the cyanide present to cyanate.  In the
second chamber, the wastewater was neutralized in the presence of excess
chlorine to oxidize ammonia.  The treated wastewater was then filtered and
dechlorinated using either activated carbon or sodium metabisulfite.  As a
byproduct of treatment, other parameters exerting a chlorine demand (sulfide,
phenol, thiocyanate, etc.) were also oxidized.

The alkaline chlorination system was run continuously over the test period
                                      383

-------
 to  provide  24  hour  composite  samples  for both  conventional and  toxic
 analysis.   Feed  ammonia  concentration ranged from 100 mg/1 to a high of
 almost  2,200 mg/1 during the  study.   However,  effluent ammonia  was
 typically less than 10 mg/1 as  shown  in Figure 5.  A summary of selected
 alkaline chlorination test results is shown in Table 2.  Figure 6 indicates
 that  effluent  ammonia was reduced significantly as the oxidation-reduction
 potential (ORP)  setting  was increased.  Operating in the 800-950 mv range
 provided sufficient treatment while maintaining the lowest possible chlorine
 residual.   The data presented in Table 2 indicate that alkaline chlorination
 was effective  in reducing pollutant concentrations to below BAT levels with
 the exception  of total cyanide.  Obviously, the presence of complexing
 agents  in the  coke  plant effluent prevented complete oxidation  bf the
 cyanide by  chlorine.





'1
CE
7
1




cctra
2000
1800
1600

H00
1200

100(3
800

600
400
200
(
=
— }
1_
i
i
~~ \
•
i
- . i
INFLUENT
/ .'
:- / * / \
H /v A/ \
I \ /\ / '""•. . / EFFLUENT \
v '""'•••--/' A 'A y//^
"llAllf\/IMlf'lV(lllJ1 w ' iljl lllllM Ullli
1 5 18 13 20 25 30 35 40
                                   RUN NUMBER
              Figure 5.  Alkaline chlorination effluent ammonia
                       concentration versus run number.
Based on the alkaline chlorination pilot results, a full scale treatment
system could be expected to routinely meet it's NPDES permit limitations for
all parameters except total cyanide.  However, it should be remembered that
the average influent cyanide value of the wastewater tested was approxi-
mately three times the BPT limit (30 mg/1) and cyanide values as high as
nine times the BPT value were observed.  It is unknown if BAT cyanide levels
could be consistently met if better BPT treatment were provided.  Values of
cyanide-A exiting the pilot treatment system were typically <0.05 mg/1.

Dual Media Filtration Results

Dual media filtration tests were performed on the coke plant effluent as a

                                     384

-------
        TABLE 2.  COMPARISON  OF SELECTED ALKALINE CHLORINATION

                   RUN EFFLUENTS TO BAT LIMITATIONS
Run
No.
I .;
11
12
1 1
20
22
ORP
mv
900
960
960
950
9
-------
polishing step from November, 1979 to January, 1980.  Tests were conducted
both with and without polymer.

Filtration results of the effluent without polymer are summarized in Table
3.  Filtration removed significant quantities of suspended solids (712
removal).  Table 3 also shows the average, maximum and minimum influent and
effluent characteristics when chemically pretreated coke plant effluent was
filtered with polymer as a coagulant aid.  Overall removal of suspended
solids (93% removal) was greatly improved with 'addition of the polymer at
a dosage of 3 mg/1.

            TABLE 3.  DUAL MEDIA FILTRATION PILOT STUDY RESULTS


                  Suspended Solids, mg/1Suspended Solids, mg/1
                        No Polymer	          Polymer Added

Average
Maximum
Minimum
Inf.
51
139
7
Eff.
11
36
2
Inf.
71
152
10
Eff.
3
9
1

Activated Carbon Results

The use of activated carbon was investigated to determine its effectiveness
as a dechlorinating agent.  The carbon will convert the excess chlorine,
produced by the alkaline chlorinatlon process into chlorides and other
harmless byproducts.

The average influent chlorine concentration to the carbon system during the
study was 63.6 mg/1.  The average chlorine removal rate across the carbon
bed was 95 percent with a range from 83.9-100 percent.  Removal efficiency
decreased as volume processed increased.  The carbon was also effective In
reducing influent ammonia by 35 percent and TOG by 61 percent.  Effluent
phenol concentrations from the carbon were decreased by 73 percent.

Sodium Bisulfite Dechlorination Results

Sodium bisulfite was investigated as a dechlorinating agent during the pilot
study.  A plot of bisulfite:chlorine ratio and percent: chlorine removal
(Figure 7), shows that 100 percent of the chlorine can be removed at a
bisulfite:chlorine ratio of 2:1.  The figure illustrates that chlorine
removal is a function of bisulfite to chlorine ratio.  The studies were
performed at a wastewater detention time of 30 minutes under complete mixing
conditions.

Priority Pollutant Discussion (Physical/Chemical Test Site)

Priority pollutant analyses were performed on 63 samples of the coke plant
wastewater plus blanks during the Phase I (alkaline chlorination-filtratlon-
activated carbon) and Phase II (alkaline chlorination-sodium bisulfite
dechlorination-filtration) Investigations.

                                      386

-------
                u
                D
                -,:

                :":
                Ul
                u
                I
                     iee.0 p
                      90.a  —
                      80.0  —
70.8 —
60.e  r-
                     30.0  =-
                     10.0
                                 -K-K-
                             III I ill III III II ll III III! ill I  111 ill IIIIlll ll
   0.0    1.0   2.0   3.0   4.8    5.0

           BISULFITE:!CL2  RflTIO
                                                            6.0
      Figure 7.   Plot of percent C12 removal versus bisulfite:C12 ratio.


     Volatile Organics.   Chlorination of the influent resulted  in decreasing
the concentration of benzene, acrylonitrile and toluene by approximately
half.  Chlorination increased chloroform concentrations from the 8 to  170
Ug/1 found in the raw influent to 3,700 - 22,000 yg/1 in the chlorinated
effluent.  Dibromochloromethane, carbon tetrachloride, 1, 2 dichloroethane,
chlorobenzene and bromoform were not detected in the influent samples;
however, significant concentrations were found in  the chlorinated effluent.
Neither Phase 1 or Phase 2 treatment systems were  completely effective  in
the removal of volatile organic priority pollutants.  However, Phase 1
processes were superior, removing 10 to 100 times  more volatile  organics than
Phase 2.  Only negligible volatile organics removals were observed during
Phase 2.

      Semivolatile Organics.   Phase 1 provided a  more complete  semivolatile
 organic priority pollutant removal than Phase 2.  The Phase 1 operation
 removed all semivolatile organic priority pollutants to non-detectable
 limits.  Phase 2 reduced all semivolatile organics to less than 100 yg/1
 except for naphthalene.

      Metals.   Phase 1 final effluent metals concentrations were very close
 to the initial raw influent levels.  Similarly, priority pollutant metals
 were not removed in the Phase 2 operation.
 EXPERIMENTAL RESULTS (BIOLOGICAL RESEARCH SITE)

 General

 The biological research work was performed at the Wheeling-Pittsburgh Steel
 Corporation's Steuvenville East coke plant from October, 1979 to February,
                                        387

-------
1980.  The water application rate  during  the study period was 432 fc/kkg
(104 gal./ton).

As shown in Figure 8,  three treatment trains were investigated.  In the
first treatment train, plant wastewater from downstream of the coke plant
cooling tower was passed  through a mixing tank,  through the first stage
activated sludge system  (carbonaceous removal)  and then through the second
stage activated sludge system  (nitrogen removal).  The second treatment
train consisted of the first treatment train (bio-oxidation) followed by
activated carbon adsorption.   The  third treatment train included the compo-
nents of the first treatment train followed by  dual media filtration.

The mix tank shown in  Figure 8 was used for equalization, pH adjustment,
dilution, and chemical dispersal.   As the wastewater flowed through the
first stage activated  sludge system,  carbonaceous material (BOD, Phenol, etc.)
was removed.  Effluent from the first stage clarifier was pumped through
the second stage activated sludge  system  where  ammonia nitrogen was oxidized
by the nitrification process.   The use of powdered activated carbon to
remove toxics and improve settling was investigated.  Final effluent from the
second stage activated sludge  system was  passed through the activated carbon
or dual media filtration  systems late in  the test program to complete the
second and third treatment train arrangements respectively.
      Ill HCTIVATtQ 1UOCC. HITHIflCATIOII
                                                                      tmutni
      (11
             siuiice.
                             C««MN


AS-2
t
\

Ml
ClAftl
b
                                                               •twin NT
      ())  MTIVATU JIUIWI, MITKIftCATIOIt. OU«l MOIA mtMTIO«
                              ClAHiritK


AS-2
t


UI
CUUMFlIt
^r
           Figure 8,
Process trains investigated  for  treatment of
 by-product coke plant wastewater.

                                        388

-------
Wastewater Treatment System

There are two primary process water streams associated with the by-product
cokemaking operations at the plant investigated:

     1.  Benzol plant effluent.
     2.  Ammonia still excess liquor.

Strong ammonia liquor blowdown is stripped in the ammonia still by steam
and caustic soda.  Due to the steam and caustic soda injection, the
volume of this stream increases and the temperature rises to -94°C (200°F).
Wastewater from the benzol plant is blended with the excess ammonia liquor
from the ammonia still for dilution and cooling purposes.  This mixture,
of which about 25 percent is excess ammonia liquor, thereby is cooled to
44-67°C (112-152°F).  After passing through an equalization tank and a
cooling tower, the wastewater has lost sufficient heat to make it amenable
to microorganism degradation.  However, the waste still contains significant
concentrations of pollutants toxic to biological life.  Therefore, down-
stream from the cooling tower, coal yard runoff and dilution service water
are added to the wastewater stream to make the waste acceptable to the
plant's single stage activated sludge treatment system.

The coke plant's equalized wastewater was used as a source of feed to the
pilot system during the study.  Characteristics of this water are shown
below in Table 4.  BPT and BAT limits are also shown in the table for the
plant investigated.


             TABLE 4.  FEEDWATER CHARACTERISTICS AND EFFLUENT
                LIMITATIONS FOR BIOLOGICAL RESEARCH SITE

Parameter
CN-T, mg/1
Phenol, mg/1
NH3, mg/1
O&G, mg/1
SS, mg/1
Sulfide, mg/1
SCN, mg/1
Diluted1
feed
9.7
657
767
17
70
30.7
451
BPT2
limit
52
3.6
217
26
223
-
-
BAT2
limit
2
0.11
32
20
86
0.6
1

           1 Diluted in the ratio of 3 parts coke plant
             equalized wastewater to 1 part service water.
           2 Dev. Doc., By-Product Cokemaking-EPA440/l-79/024a.,
             Oct., 1979.


As shown above, the pilot system feedwater was well above BPT limits for
phenol and ammonia during the test period.


                                      389

-------
Biological Treatment Results

     First Stage Carbonaceous Removal System.   The first stage pilot
activated sludge system was effective in removing both BOD and phenol.  The
feedwater was pH adjusted and diluted (3 parts coke plant wastewater to
1 part service water) prior to pilot scale treatment.  Phosphoric acid
was also added.

Influent BOD ranged from 1,290 mg/1 to 2,550 mg/1 and averaged 1,800 mg/1.
The BOD removal efficiency was typically 95% with a range from 60 percent
to 99 percent.  Good removals of phenol were.also observed in the first
stage system.  Efficiency ranged from 90% to 100%.  The average diluted
feed phenol concentration was 657 mg/1, with a range from 440 mg/1 to
920 mg/1.  Removals of thiocyanate and TOG were also observed in the first
stage treatment system.

     Second Stage Nitrogen Removal System.   The primary objective of the
second stage bioxidation unit was to reduce influent concentrations of
ammonia through the nitrification process.  The pilot system feed ammonia
during the study was quite variable, ranging from 293 mg/1 to 2,553 mg/1
after dilution.  The average diluted feed ammonia concentration to the
pilot units was 767 mg/1.  The influent wastewater to the second stage was
pH adjusted using sodium carbonate and sodium hydroxide.  Powdered activated
carbon was also added to reduce the effect of toxic shock loads and help
weight the nitrified sludge.  Polymer was also added to the second stage
clarifier to minimize biomass losses over the weir.

Attempts to achieve a substantial population of nitrifiers were unsuccessful
during the first six weeks.of testing.  Dilution to the second stage was
initiated on December 1, 1979 to help reduce the wide fluctuations in
feed ammonia concentration and therefore promote the growth of nitrifying
bacteria.  This had a positive effect on the microorganism population and
nitrification began to take .place in early January, 1980.  Excellent
ammonia removals were then consistently achieved for the duration of the
project as shown in Figure 9.  Test results for a period of time when
nitrification was occurring are shown in Table 5.
                 TABLE 5.  ANALYTICAL RESULTS FOR SELECTED TEST
                        PERIODS - NITRIFICATION SYSTEM
Dae*
1/16/80
1/18/80
1/21/80
1/23/80
1/23/80
1/28/80
SS
Inf.
47
34
26
27
41
57
(OB/1)
til.
56
66
61
31
59
93
NHa
Inf.
206
228
171
160
136
234
(mg/1)
Eff.
17
8
2
1
3
6
CN-
Inf.
2.0
3.6
3.6
8.3
9.5
13.3

-------
2259

£000

17SQ

isaa

1258

1003

 750

 sea

 esa

  a

                        IV
                               INFLUENT
                       EFFLUENT-

                       I I I I I I I I I I I I I II I M I I I I    IT
111 IIUJ 1! Ill I
                 10/31  11/08  11/21  12/03  12/17 12/31  1/14  1/28  2/35

                                      DFtTE

        Figure 9.   Influent ammonia versus AS-II effluent ammonia.
By comparing the Table 5 analytical  results with the BAT limits shown in
Table 4, it is apparent that  BAT  guidelines could be met for all parameters.

     Activated Carbon Results.    The results of the biological test site
activated carbon study determined that carbon was effective in removing TOC
(53% removal), color  (60% removal),  BOD (40%) and SCN (79% removal) from
the second stage activated  sludge effluent.  There were also minor removals
of phenol, oil and grease and TKN.

     Dual Media Filtration  Results.    Dual media filtration tests were also
performed on the nitrified  effluent.  Runs were conducted at filtration rates
of 122, 204, 326 and 407 i/m±nfm2 (3, 5,  8 and 10 gpm/ft2).  During the
study, influent suspended solids  averaged 190 mg/1.  Removal efficiency
ranged from 17 to 64 percent.

During one of the dual media  filtration runs, samples of influent, effluent,
and backwash were collected for metals analysis.  The metals analysis data
are summarized in Table 6.


             TABLE 6.  DUAL MEDIA FILTRATION HEAVY METALS REMOVAL
Sample
Point
Influent
Effluent
Backwash
Heavy Metal Concentration, mg/1
Cu
0.06
0.05
0.86
Cd
0.03
0.03
0.13
Pb Zn
<0.1 2.4
<0.1 0.9
0.20 30.0
As
<0.0005
<0.0005
<0.0005
Se
0.56
0.42
2.33
Sb
<0.05
<0.05
<0.05
                                       391

-------
As seen in Table 6, copper, cadmium and selenium were removed In trace
amounts, while a significant amount of zinc was removed by the filter.
There were no measurable removals or arsenic or antimony.

Priority Pollutant Discussion (Biological Test Site)

Priority pollutant analyses were performed on the coke plant wastewater
treated in the experimental two stage biological system followed by the
pilot activated carbon adsorption unit.  A total of 13 samples plus
appropriate blanks were collected during February, 1980.

     Metals.   The pilot treatment system was not effective in reducing
influent zinc concentrations.  The mean influent concentration of zinc was
443 yg/1 and the effluent 400 yg/1.  Zinc was found in high concentration
(630 yg/1) in the raw water added to the treatment system as makeup water.
Zinc concentrations in the carbon column effluent were reduced threefold
from the activated sludge effluent levels.  Selenium was reduced by 772
from a mean Influent concentration of 1,370 yg/1 to a final effluent concen-
tration of 320 ug/1.  The selenium concentration in the carbon column
effluent decreased approximately 25 percent from the activated sludge
effluent level.  Arsenic was reduced from a mean influent concentration of
360 yg/1 to an effluent concentration of 99 yg/1.  Influent silver concen-
tration was 18 yg/1 compared to an effluent level of 12 yg/1.  All the
other priority pollutant metals were removed to concentration levels close
to or below detection limits.

     Volatile _0_rganics_.   The treatment process was effective in removing
all volatile organic priority pollutants.' Toluene was reduced from a mean
influent concentration of 607 yg/1 to a final effluent of less than 10 yg/1
(on two of the three sampling dates).  While the activated sludge system
influent contained 6,100 yg/1 to 9,800 yg/1 of benzene, no benzene was
detected in the activated sludge or the final carbon column effluents.  All
other volatile organic compounds in the activated sludge and carbon column
effluents were generally below the detection limits.

     Base/Neutral Extractable Organics.   The treatment process was effective
in reducing all monitored base/neutral extractable organics.  The concen-
tration of all nine of the base/neutral compounds found in the influent
was reduced through the treatment process to less than 10 yg/1 in the
activated sludge final effluent.

     Acid Extractable Qrganics.    All acid extractable organic compounds
monitored were effectively removed by the treatment system.  For example,
the system influent contained a mean concentration of 62 mg/1 phenol.  Phenol
was reduced by the treatment process to less than the detection limit in the
final effluent.  All other acid extractable organic compounds in the final
effluent were below the detection limit.


SUMMARY AND CONCLUSIONS

     Physical-Chemical Test Site.  Two physical-chemical treatment trains
were investigated.  Train 1 consisted of alkaline chlorination, filtration


                                       392

-------
and activated carbon.  Train 2 consisted of alkaline chlorination, filtra-
tion and sodium bisulfite dechlorination.

     1.  The results of the pilot program indicated that alkaline chlorina-
         tion was effective in reducing influent concentrations of ammonia,
         oil and grease, phenol, sulfide, suspended solids and thiocyanate
         to below future BAT levels.  The presence of complexing agents in
         the coke plant effluent prevented complete oxidation of the cyanide
         by chlorine and as a result, BAT cyanide-T values could not be
         consistently met.

     2.  Filtration provided effective polishing of the alkaline chlorinated
         coke plant effluent, removing 71 percent of the influent suspended
         solids.  Suspended solids removal could be increased to 93 percent
         with the addition of 3 mg/1 polymer.

     3.  Activated carbon and sodium bisulfite were investigated as dechlori-
         nating agents.  Activated carbon was found td consistently remove
         95 percent of the incoming total chlorine.  Sodium bisulfite
         provided 100 percent chlorine removal at a bisulfite:chlorine
         ratio of at least 2:1.

     4.  During the pilot study, 63 samples were analyzed for priority
         pollutants.  The results concluded that the physical/chemical
         treatment trains investigated created several volatile organic
         priority pollutants.  Phase I technologies removed 73% of the
         volatile organic priority pollutants to non-detectable limits;
         Phase 2 technologies were effective in treating only 17% of
         incoming volatile organic toxics to non-detectable levels.  Semi-
         volatile organics were all effectively reduced for Train 1.
         Train 2 also reduced all semivolatile organics to less than 100
         pg/l except for naphthalene.  The physical/chemical treatment
         trains removed only negligible concentrations of metals.

     Biological Test Site.   A pilot scale two stage activated sludge unit
was investigated for removing coke plant wastewater contaminants to below
BAT values.  Filtration and activated carbon were also studied as polishing
steps.

     1.  The first stage activated sludge unit was capable of removing
         95 percent of the influent BOD and from 90-100 percent of the
         incoming phenol.  Thiocyanate and TOG reductions were also
         achieved.

     2.  Influent ammonia to the second stage activated sludge system was
         quite variable, ranging from 293 mg/1 to 2,553 mg/1.  It was
         necessary to dilute the first stage activated sludge effluent to
         maintain a consistent feed ammonia strength before nitrification
         could be achieved.  After a sufficient population of nitrifiers
         were in the system, ammonia reductions of >97 percent were
         consistently achieved.  Suspended solids, oil and grease, thio-
         cyanate and phenol were also reduced to below BAT levels.
                                       393

-------
     3.  Activated carbon, when used as a polishing step for the nitrified
         effluent, was capable of removing 53 percent of the influent IOC,
         60 percent of the color, 40 percent of the BOD and 79 percent of
         the remaining thiocyanate.

     4.  Dual media filtration was found to remove about 50 percent of the
         suspended solids present in the second stage activated sludge
         nitrified effluent.

     5.  Priority pollutant analyses were performed on 13 samples taken
         from various points in the treatment system.  All priority
         pollutant metals were reduced to less than 100 yg/1 except for
         selenium and zinc.  The biological treatment train was efficient
         in removing all volatile organics, base/neutral extractable
         organics and acid extractable organics to non-detectable levels.
ACKNOWLEDGEMENTS

Rexnord acknowledges the cooperation and support of the U.S.  Environmental
Protection Agency.  The assistance given by Robert Hendriks,  Project
Officer was received with much appreciation.  Deep appreciation is also
extended to Nick Buchko and Jim Zwickel of Shenango, Inc. and Bill Samples
of the Wheeling-Pittsburgh Steel Corporation.
 The information contained in this paper is part of a draft final report
 being prepared for  the U.S. Environmental Protection Agency.  Modifications
 to the enclosed material prior to publication of the final report are
 probable.

                                       394

-------
             SINGLE STAGE NITRIFICATION OF COKE PLANT WASTEWATER
                            Dr. George Wong Chong
                   Environmental Research and Technology
                            Pittsburgh, PA 15219
                                     and

                              Mr. John D. Hall
                        National Steel Corporation
                         Research and Development
                             Weirton, WV  26062
ABSTRACT
A laboratory scale study of single stage phenol oxidation-nitrification
activated sludge treatment of coke plant wastewater was conducted.  The objec-
tives of the study were to determine:

        the operating conditions at which a treated effluent would
        contain an ammonia concentration of 10 mg/1 or less,

        the effects of sudden changes in loadings of certain waste-
        water constituents on the biological process,

        the effects of the process on priority organic pollutants,
        and

        means of enhancing the biological process.

In this study, eight test reactors were used; the feed to these reactors was
undiluted ammonia still waste which was ammended to a constant composition of
ammonia, phenol and thiocyanate.

The results of the study show that the single stage phenol oxidation-nitrifica-
tion process can produce high degrees of treatment for ammonia, free cyanide,
phenol, thiocyanate and sulfide but it was ineffective in treating complex
cyanide.  This process is also effective in controlling priority organic
pollutants found in coke plant wastewater.  Sudden changes in the reactor
loadings of conventional pollutant constituents resulted in neither toxic nor
prolonged inhibitory effects.  However, the process was sensitive in responding
to abrupt changes in feed composition and reactor composition.  These responses
to changes should be tested on a full scale operation such that the true impact
of normal coke plant operations could be assessed.  The preliminary evaluation
of activated carbon addition, carbonate addition and commercial mutant bacteria
addition as means of enhancing (increasing the rate of nitrification) were
inconclusive.

This study was conducted in fulfillment of an EPA Grant (No. R806234-01-2)
which provided partial funding for this program.
                                     395

-------
             SINGLE STAGE NITRIFICATION OF COKE PLANT WASTEWATER

INTRODUCTION

The Environmental Protection Agency's proposed "Best Available Technology
Economically Available," BAT, effluent guidelines for by-product coke plant
wastewaters may include:  ammonia, cyanide (total), oil and grease, phenolic
compounds, sulfides, thiocyanates and priority pollutants.  Alternative techno-
logies for achieving compliance with the proposed effluent guidelines are (a)
physical chemical technology and (b) biological treatment.  The proposed BAT
biological technology suggests a multi-stage biological treatment system which
includes a phenol removal reactor, cyanide-ammonia oxidation reactor followed
by a nitrate reduction reactor and a final step aeration for the reoxidation
of sulfide.  Figure 1 presents a flow diagram of the proposed BAT treatment
system showing a free/fixed leg ammonia still pretreatment stage.  This report
deals with a single reactor alternative for phenol oxidation/nitrification.

Previously, National Steel conducted a preliminary examination of a single stage
biological reactor to determine the feasibility of this technology to produce
compliance with proposed BAT limitations.  Although the evaluation had not
progressed sufficiently to conclude whether the process was a viable treatment
alternative, the results were sufficiently encouraging to warrant further
examination.  The preliminary testing with a single stage pilot reactor,
treating coke plant ammonia still waste showed that both phenol oxidation and
nitrification could be achieved in single reactor system.  However, too often
periods of sustained effective operation were interrupted by unidentified
episodes which completely disrupted the nitrification process; subsequent
start-up after an interruption required an extended period of reacclimation.
These extended periods were as long as seven weeks.

Undoubtedly, those unidentified disruptive episodes are causes for concern in
any decision to apply this technology on a full scale and In many respects they
reflect the level of understanding of this technology.  Consequently, this
program was initiated.  The objectives of the program were:

(1)  To determine the operating conditions necessary to produce an effluent
with an ammonia concentration of 10 mg/1 or less and to determine the corres-
ponding concentration of the other conventional pollutants,

(2)  To determine the effects of certain constituent compounds in coke plant
wastewater on the performance of the single stage phenol oxidation/nitrifica-
tion process,

(3)  To conduct a preliminary examination of methods for enhancing the opera-
tion/performance of the single stage phenol oxidation/nitrification process,

(4)  To determine the effects of the single stage phenol oxidation/nitrification
process on priority organic pollutants in coke plant wastewaters, and
                                         396

-------
     Flushing
     Liquor
     (WAL)
Free/Fixed
Ammonia
Still



Biological
Phenol
Removal
t
-<
                                Biological
                                Cyanide.& Ammonia
                                Oxidation
OJ
10
                             Discharge"
Step Aeration
Sulfide
Control
Biological
Nitrate
Reduction
                                                                                           I
                                                                                                             Methanol
                  Figure 1.  Flow Diagram of Proposed BAT Treatment System for By-Product Coke Plant
                             Wastewaters.

-------
(5)  To develop a better understanding of the operation and performance of the
single stage phenol oxidation/nitrification process.

The program was essentially a laboratory investigation in which actual coke
plant wastewater was examined.  This report presents the findings of this
program.

INVESTIGATIVE FACILITIES AND PROTOCOL

The experimental program was conducted at the National Steel Corporation
Research Center.  Facilities for the biological wastewater treatment experi-
ments included:  1) a sludge bank reactor unit; 2) eight bench scale reactor
units and 3) support analytical facilities.

Sludge Bank Unit (SBR)

The sludge bank unit was a 160 gallon pilot reactor, Figure 2, which served
as a source of readily available acclimated sludge for the experimental pro-
gram.  This pilot unit was used previously for the single stage biological
experiments which led to the study program.  Although the unit was composed
of two discrete vessels, a single reactor effect was achieved by the common
heads between the two vessels and the high recirculatory rate.  This reactor
has been in operation since October 1977.  Feed to the pilot unit came from
the Weirton Steel Brown's Island coke plant ammonia still (ASW).

Because of the long term operation of this reactor, the sludge in the reactor
was fully acclimated to the wastewater.  This readily available source of
acclimated sludge greatly facilitated the experiments which were conducted by
circumventing the long operating time required for sludge acclimation.  Thus,
the experiments could be conducted to satisfy the hydraulic requirements for
steady state.
                                       398

-------
                                         pH
                                      Controller
                                                                Temp.
                                                               Control
                                                               ~85°F
CO

to
2000 Gal
Feed
Storage
Tank
              Batch Filled
                                      Feed  Pump
                                                                               Treated
                                                                               Waste
                                                               Recirculation Pump
                                                                    30 gpm
                                       Figure  2.   Sludge  Bank  Reactor Unit.

-------
Bench  Scale  Reactors

The bench  scale  reactors were  all  activated  sludge  type  systems with a  31
liter  aeration chamber  and  integral  clarifier.  Each unit was  equipped  with
its own  feed system which enabled  independent operation  relative  to operating
condition  or feedwater  composition.  Figure  3 is an isometric  diagram of a
bench  scale  activated sludge unit.   The eight bench scale units were housed
in a temperature controlled cabinet.  The  temperature within this cabinet was
maintained at 85°F plus or  minus 2 degrees.

One of the eight reactors,  number  2, was operated exclusively  as  a control
reactor  for  the  duration of the program.   The other seven reactors were used
to study test variables.  Often, a specific  test condition was examined in
two reactors.  All eight reactors were continuously monitored.  When a  spe-
cific  test condition was not being examined  in any  reactor that reactor was
used for collecting additional control data.

The wastewater feed to  the  bench scale unit was ammonia  still waste with the
major  constituent concentrations adjusted  as follows:  1) ammonia - (NH3) = 150
mg/1,  2) phenol  = 500 mg/1, 3) thiocyanate = 300 mg/1 and 4) alkalinity -
1060 mg/1.   In subsequent discussions the  above ASW feed will be  referred to
as a standard feed.  It is  recognized that in real  practice the wastewater
will not have a  constant composition.  However, the variability in feed com-
position had to  be minimized in order to better evaluate variable effects.
In specific  phases of the program the feed was changed to meet desired
requirements.

Reactor  Monitoring and  Chemical Analyses

All the  bench scale units were monitored daily for  the following:  1) feed
and effluent flow rate, 2)  reactor pH with appropriate adjustments when
necessary, 3) reactor DO, 4) reactor temperature and 5)  reactor sludge volume.

Appropriate  feed and reactor mixed liquor  samples were taken for  analyses as
per Standard Methods.   Table 1 presents a  schedule  of sample analysis.   Addi-
tional analyses were conducted as necessary.  Two sets of samples were taken
for priority organic analyses by gas chromatography-mass spectrograph.

SINGLE STAGE CONTINUOUS FLOW REACTOR KINETICS

Wong-Chong and Caruso(D examined the treatment of  coke plant wastewater in
a single stage reactor  for phenol degradation and nitrification.   Their study
examined the  treatment  in batch and continuous flow reactors; both synthetic
and actual coke plant wastewater were used.  The batch reactor experiments
provided information on the order of the different reactions involved,  some
insight into  interactions which might occur, and the sequence in which  the
reactions occur.   These observations are shown in Figure 4.   These observa-
tions indicate that the nitrification reaction is the process controlling step
and the reaction is of zero order with respect to ammonia nitrogen.   Thus,
the reaction can be mathematically described as

                          ~- - kA/MLSS                            (1)
                                     400

-------
                       Test Feed Reservoir
                                                                              Air
                                                                              Humidification
Seated
Effluent
Reservoir
         Figure 3.   Isometric Diagram of Bench-Scale Activated Sludge Unit
                                             401

-------
   Table 1.   REACTOR AND EFFLUENT ANALYSIS SCHEDULE






Parameter       Monday  Tuesday  Wednesday  Thursday  Friday
Ammonia XX X
Cyanide, Free X X
Cyanide, Total X X
Thiocyanate X X
Sulfides X X
Phenol X X
Phosphates X
Alkalinity X
Nitrite & Nitrate X
Solids X
COD X
Oil and Grease X
X X
X
X
X
X
X
X
X
X
X


                            402

-------
                                                                                     - 250
120  -
                                   6         8         10

                                   Reaction Time hrs.
12
16
    Figure  4.   Reaction Sequence for Ammonia,  Cyanide,  Phenol and Thiocyanate In a Batch
               Reactor.

-------
where EN = concentration of oxidizable nitrogen.
         = NH4+ - N + 0.54 CNF- + 0.24 SCN~

 k^/MLSS = oxidation rate for a given mixed liquor solids concentration
       t = reaction time

Applying Equation 1 to a continuous flow, batch mixed reactor and performing
a material balance around the reactor produces the expression

                  EN = ENt - TkA - Aie-t/1                         (2)

where  EN = effluent oxidizable nitrogen concentration
      EN-i - feed oxidizable nitrogen concentration
       A! = boundary condition factor (constant of integration)
        T - hydraulic residence time

at steady state condition, Equation 2 simplifies to

                  EN - ENi - TkA                                   (3)
and from Equation 3, the reaction rate for a specific mixed liquor sludge can
be expressed

                           =-=                                    («)
From the previous observations^ ' with various levels of mixed liquor solids,
kA was correlated as a function of mixed liquor solids as

                  k -= 15.2 (TVS)                                   (5)

where TVS = mixed liquid volatile solids, g/1.  The control reactors had
measured oxidation rates which had good agreement with those previously re-
ported, as shown in Figure 5.  Details of these test experiments are presented
in Table 2.

Equation 2 can be used to describe both steady state and transient state
operations.  Transient conditions can result from changes in feed flow rates
and substrate loadings.  This occurrence is demonstrated in Figure 6.
The reactor was operating under the following conditions:  1)
reactor mixed liquor solids =4.20 g/1 TVS, 2) hydraulic residence time,
HRT, -3.9 days and 3) nitrogen oxidation rate =55.6 ing/I/day.  The feed
flow rate was increased to produce a hydraulic residence time of 2.2 days.
From the above information, the reactor EN concentration can be predicted by

                   N - 104 - 96e~°'45T                             (6)

It must be noted that a major fraction of the residual oxidizable nitrogen
was due to ammonia, although the thiocyanate concentrations were very high.
However, these high thiocyanate concentrations were well within the range of
predicted values.
                                      404

-------
         Figure  5.  Agreement Between the Oxidation Rates from
                     Control Reactors of this Program and that
                     Previously Reported.
200
                                                          Reported  oxidation rate
                                                          curve as  a function of
                                                          reactor solids
                                                         k=  15-2 TVS (Mellon Institute)
                                                           OControl reactors
                                                             experimental results
                 2.0         4.0         6-°
                     Mixed Liquor Solids, g/1 TVS
                                                                 10.0
                                    405

-------
         Table 2.    DETAILS OF THE OPERATION OF CONTROL REACTORS
                             Feed Composition
Phenol = 500 mg/1
CNF = <0.1 mg/1
NH3 =
SCN =
EN =
= 150
= 300
= 225
Reactor Operating Conditions
Reactor
1
2
3
3A
4(b)
6
7
7A(0
6
7
7
6
7
7
7
7
.5-7.6
.0-7.5
.2-7.7
.9-7.6
.0-7.6
.1-7.6
.1-7.4
.1-7.6
T
days
3.7
3.5
3.3
3.7
3.8
2.7
3.9
5.0
days
192
250
174
177
96
153
156
282
k
mg/1 /day
57.5
60
67
56
59
83
55
54
.9
.0
.2
.9
.0
.6
.9
MLTVS
4.24
5.10
4.65
4.14
3.47
5.05
4.20
3.43
MLSS
g/1
4.05
5.17
4.58
3.91
3.03
5.10
3.99
2.98
mg/1
mg/1
mg/1


Average
Effluent Composition (d)
NH3
11±18
9±16
4±3
12±22
7±8
14+25
7+10
3±2
SCN
mg/1 CNF
5±6 <0.1
11±22 <0.1
11110 <0.1
13±19 <0.1
9±10 <0.1
11±19 <0.1
6±5 <0.1
29±46 <0.1
<()OH
Ug/1
20±20
17±16
17+11
25128
31±47
22±26
68±90
79+58
(a)   Basis - suspended solids




(b)   Feed contains 350 mg/1 SCN




(c)   Feed contains 210 mg/1 NH3



(d)   Negative values observed due to variability in the actual numbers.
                                     406

-------
              Figure  6.  Prediction of Transient Conditions Using

                          Equation 2.
0
                                                                          129
   120
                            Predicted SCN = 134 mg/1
   100
c
0)
o

o
u


01
60
O

4-1
•H
z


-------
The thiocyanate concentrations can be predicted by an equation similar to
Equation 2, where the thiocyanate concentrations are substituted for nitrogen.
Thus, for the situation shown in Figure 6, the degradation rate kgcN is
75.4 rag/ I/ day and

                  S - Si - 75.4 T - Be-T/T                         (7)

                  S = 134 - 28 e~°'A5T                            (7A)

From Equation 7A, the new steady state concentration of thiocyanate would be
134 mg/1 which compares well with those observed as shown in Figure 6.

In comparing the predicted with those of the observed, one notices a marked
difference, especially in the early stages of the transition period.  At best
the mathematically formulations presented will provide an estimate of effects
resulting from changes.  It must be recognized that biological treatment
systems are dynamic systems and during transition periods such as that shown
in Figure 6, there is a potential for the number of nitrifying organisms tn
increase resulting in the lower observed values.  However, if the imposed
change results in events occurring at a rate greater than the growth ratey
of the organisms then there will be an accumulation of materials in the
reactor.  These materials which accumulate in the reactor in turn could exert
an inhibitory effect on the micro-organisms once certain tolerance levels are
exceeded.  Alteration of activity by inhibitory materials could result in
severe reductions in oxidation rates.  The result would be observed rate
values greater than predicted.

In the course of biological treatment, certain quantities of excess sludge
are produced.  This sludge production can be predicted by


                            U - b


where 9C   = sludge retention time, days
      b    = microblal maintenance energy coefficient, day~l
      U    * specific substrate utilization rate, day~l
             maximum sludge yield coefficient
From Equation 8, a plot of Qc~^ against U will produce values for ymax and
b.  Figure 7 presents a plot of operating data taken from the control reactors
during the study.  From this plot, the value of yraax is 0.7 mg SS formed per
mg of nitrogen oxidized and b is 0.004 day~l.  With this information the
potential sludge production can be estimated from
                  ye    Vmax

From Equation 9, the amount of sludge produced, Sp, can be determined from the
amount of nitrogen oxidized and  the wastewater  flow rate according to

                  Sp - Q  (ZNln - ZNout)ye                         (10)
                                     408

-------
      10
       -2
       -4
                           I
                          8         12

                               u x 103
16
20
                                           b « 0.004 day"1
                                             mQm7jaJL
                                                            u - b
                                                   ,ngNoxld
Figure 7.  Sludge Production in the Single-Stage Phenol-Nitrification
           Process for Coke Plant Wastewaters.
                               409

-------
where Sp - sludge production rate, mg SS/l/day
      Q  * wastewater flow rate, I/day
      IN - concentration of oxidizable N, mg/1

EFFECTS OF REACTOR CONDITIONS

The reactor operating conditions of primary concern are:  1) mixed liquor
solids concentration, 2) hydraulic residence time, 3) pH and alkalinity,
A) dissolved oxygen concentration and 5) temperature.

Mixed Liquor Solids and Hydraulic Residence Time

In designing a biological treatment system there are essentially two main
factors which will affect the sizfe of the system.  These are the mixed liquor
solids concentration and the hydraulic residence time.  The mixed liquor
solids concentration is related to the reactions rates shown in Figure 5.
However, these rates can be influenced by other factors such as pH, tempera-
ture, and dissolved oxygen.  The reaction kinetics presented earlier can be
manipulated to show the relationship between hydraulic residence time and
mixed liquor solids as follows:

                      TVS - £Ni".T2£NOU.t                          (11)


Thus, for a given wastewater stream and desired effluent quality, the size of
the aeration basin can be related to the mixed liquor solids concentration
using Equation 11.  This equation indicates that the higher the mixed liquor
sludge concentration the smaller the volume of the aeration basin.

pH and Alkalinity

Two of the experimental pilot reactors were operated over a range of pH con-
ditions to determine the effect(s) of pH on nitrification, the results of
which are summarized in Figure 8.  It is, noted that the range shown in Figure
8 is somewhat greater than the 7.0-7.5 range mentioned in the literature as
being the optimum environmental condition for the maintenance of the nitrifi-
cation reaction.(2)  it is also noted that the sludge content of the two
reactors was  different and the reactor with the greater mixed liquor sludge
concentration was used in examining the higher pH region.  From Figure 5 it
would be expected that the reactor with the higher sludge content would have
a higher oxidation capacity.  This greater oxidative capacity coupled with a
long hydraulic residence time may have counteracted any negative effect at pH
levels greater than 7.5.  The result could be the higher efficiencies observed
at pH 7.7 and 7.9 when compared to optimum efficiencies discussed in the
available literature.

The data points at pH 8.1 and 8.3 on Figure 8 are also significant in that
steady.state conditions were not achieved during the experimental work-
suggesting that the real efficiency could have been lower than shown.  The
detailed data for these two data points are presented in Figure 9.  From
Figure 9, it can be seen that with a pH increase from 7.9 to 8.1, there was
a slight increase in thiocyanate concentration.  In addition, further
                                      410

-------
               Figure 8.    Effect of pH on Nitrification of Coke Plant
                           Wastewaters.
  100 -
   90
c
o

-------
    0.3
g  0.2
•O
 0)
ir
 a.
 o
2
4J
c

-------
 increasing the pH to 8.3 resulted in a substantial increase in thiocyanate
 concentration.   Decreasing the reactor pH from 8.3 to 7.2 resulted in an
 immediate reduction in the reactor thiocyanate concentration.   In effect,
 high  pH levels,  >8.1,  appear to adversely  affect SON degradation.

 The mixed liquor thiocyanate concentration,  possibly along with the synergis-
 tlc effect of  pH,  appears to adversely affect the nitrification reaction.
 Deterioration  of nitrification appeared at a thiocyanate  concentration about
 100 mg/1 and pH  of 8.3.

 The effect of  pH on the biological treatment process is further illustrated
 in Figure 10.   Figure  10 represents the performance of a  reactor receiving
 ammonia still  waste as produced,  i.e., the raw wastewater composition varies.
 Further,  in this test  sequence the feed flow rate also varied,  as shown.   The
 data  in Figure 10  suggest  the following:   1)  both high and low pH  conditions
 affect  nitrification,  2)  the adverse effect  of low or high pH  can be counter-
 acted by decreasing the wastewater flow rate and  3)  high  pH conditions affect
 SCN degradation.   In spite of the observed fluctuations,  phenol treatment  was
 99.9% 4- effective, with effluent  concentrations less than 200  vig/1.

 The ammonia-nitrogen oxidation reaction produces  acid according to  the follow-
 ing equation:

                       NH4 + 1.502 -»• N02" + 2H+ +  H20             (12)

 The formed acid  tends  to decrease the pH of  the reaction  medium,  and in order
 to maintain optimum reaction efficiency, alkalinity  must  be  added to the
 system  to neutralize the acid produced.  Stoichiometrically, 7.14 units of
 CaC03 alkalinity are required to  neutralize  the acid generated  from  the oxida-
 tion  of one unit weight  of ammonia-nitrogen.   A series of tests was conducted
 to determine the alkalinity requirements.  Figure 11 presents the results of
 those tests and  the alkalinity requirement A,  can be estimated  from

                       A  - 4.46N -  517

where A - CaCC>3 alkalinity required,  mg/1
      N « nitrogen oxidized,  mg/1

 The correlation shows  that 4.5 units  of  CaC03  alkalinity  are required  for every
 unit  weight of nitrogen  oxidized.

 Dissolved Oxygen

 In the  course of the study,  project efforts were  taken to maintain all reactor
 dissolved oxygen levels  greater than  1.0 mg/1.  There was  no deliberate
 affort  to  determine the  effects of dissolved oxygen  concentrations.

Temperature

The temperature studies covered the range  from 60°F  to 95°F.  From the temper-
ature evaluation it  was determined that, 1) optimum  reactor operating  temper-
ature is between 70°F and  80°F, 2) sludge oxidation  occurs at 95°F, 3) normal
                                       413

-------
                 Figure  10.  Performance of a Phenol-Nitrification Reactor Receiving As-Produced ASW.
g  .6



QJ  ^4




B  .2

AJ

3   o






    8




a   7
a

    6



    5







   40
  30
ed

fc 20
8
                                                            Legend


                                                           O  Ammonia
         2   46   8  10 12  14  16   18 20  22  24  26 28  30 32  34 36  38  40 42  44  46  48 50 52  54 56  58 60



                                              Days  of Operation

-------
en
         • 6"
       en
       u
.4






 0






 8



 7


 6


 5






40
         30
       c
       o
       •H
       4-1
       R3
       G  20
       QJ


       §
          10    -
                        figure 10.  Performance of a Phenol-Nitrification  Reactor  Receiving As-Produced

                                    ASW.  (continued)
Legend



  O Ammonia



  D Thiocyanate
                 62  64  66  68  70  72  74  76 78  80  82  84  86   88  90   92   94   96  98 100  102 104


                                                    Days of Operation

-------
             Figure H.  Alkalinity  Requirement  for the Control of pH in
                         the Nitrification  of Coke Plant Wastewaters.
  4000
  3000
o
CO

r-l
<


O


u
   2000
   1000
                  200
                                I
                                              y - 4.46X - 517
                                                Reactor D.O 1.0-5.0 mg/1
                                            I
I
                               400         600         800

                                   Nitrogen Oxidized, tag/1
           1000
                                      416

-------
operating residence times will have to be increased at reactor temperatures
below 70°F to maintain nitritication and 4) phenol oxidizing bacteria may be
adaptable to operating temperatures less than 70°F.  Figure 12 shows the
nitrogen oxidation rates for the temperatures of 95°F, 80°F, 70°F and 60°F.
Optimum rate of nitrification was observed to be between 70°F and 80°F.
Significant decreases in the oxidation rate occur  outside this boundary.
Extrapolation of the data line from the data point at 60°F into lower temper-
ature ranges and observing all data points up to a temperature of 80°F indi-
cates that the rate of nitrification increases with temperature throughout the
range.  However, when the reactor temperature was allowed to increase beyond
the optimum range, the nitrification rate decreases possibly due to bacteria
cellular destruction.

Figure 13 is a graphic presentation of the reactor concentrations of
ammonia, phenol and thiocyanate at the evaluated temperatures.  Although all
data points are connected by straight lines, it is noted that system adjust-
ments occurred between temperature changes.  It was necessary to replenish
the mixed liquor suspended solids in the reactor after the 95°F evaluation.

Only a short evaluation period for a temperature of 80°F was performed as
prior experience during the initial phases of the project had shown that 80°F
was near an optimum operating temperature.  The intent of the temperature
evaluation was to determine the stresses, if any, on the bio-process when
the temperature was allowed to vary from the established optimum temperature.

Two evaluations were made at the 95°F operating temperature, although only
one set of data is shown in Figure 13.  In the first experiment it was noted
that a significant loss of mixed liquor suspended solids occurred, although
no deliberate sludge wasting was being practiced.  After twenty operating
days, the reactor concentrations of ammonia, phenol and thiocyanate
had increased significantly indicating that the biological functions had been
severely inhibited.  The obvious cause was the decrease in the mixed liquor
suspended solids or a lack of micro-organisms for contaminant oxidation.  The
pilpt reactor'was replenished with sludge from the sludge bank to a mixed
liquor suspended solids concentration corresponding to the concentration when
the experiment originally began.  Figure 14 shows volatile suspended solids
concentration versus time and the reactor residual ammonia, phenol and thio-
cyanate concentration at an operating temperature of 95°F.  An immediate
decrease in the volatile suspended solids concentration was observed which
continued for 20 operating days.  Correspondingly, there was a sharp decrease
in thiocyanate removal, followed by a decrease in ammonia and phenol oxidation.

It was concluded that at an operating temperature of 95°F (and above) the
activated sludge experiences combustion (oxidation or cellular destruction)
to a degree that renders the activated sludge system useless even though
other parameters such as hydraulic residence time, pH and DO were held con-
stant resulting in no external stresses to the bio-system.

In general, the mechanisms affecting the nitrogen oxidation rates below and
above the optimum rates appear to be different.  At temperatures below the
optimum rate, bacterial action is reduced by lower metabolism rates.  At
temperatures greater than optimum, nitrification is affected by sludge
                                      417

-------
   70
Figure 12.  Nitrogen Oxidation Rate Versus
            Reactor Operating Temperature.

            pH             =7.3
            Retention Times 5.3 Days
            MLSS           =~2.3 g/1
   60
 l^»
5
.-1

 »50
 0)
|
4J
•3 40
I
 £
 M
 8
   30
   20
   10
                                1
               I
                   60
  70          80
    Temperature °F
              418
90
100

-------
(II
4J
id

§
^
u
o
•rl
 g

 rt
 •H rH
 c o
 O C
 0
M
 C
 
-------
  1600

&1400
 •k

31200
o
CO
•giooo
•a
t 800
9
u 600
iH
•H
,5 400
o

   200

  >100


    90


    80


    70

I  60
    50
s
I
    30

    20

    10
                                      Figure 14.  Residual Reactor Concentrations of
                                                  Ammonia, Phenol, and Thiocyanate
                                                  Corresponding to Volatile Suspended
                                                  Solids Concentrations, Reactor
                                                  Temp. - 95UF and Mo Sludge Wasting.
                     Reactor Thiocyanate
                     Concentration ppm
                                                           Reactor Ammonia
                                                           Concentration ppm
                                                      1
                         10
                                  15        20        25
                                      Operating Days
30
35
                                      420

-------
oxidation reducing the quantity of bacteria available for contaminant removal.

To increase the oxidation rates at operating temperatures lower than optimum,
the residence time could be increased.  Increased residence times will simply
allow for more contact time between the contaminants in the feed water and
the bacteria operating at reduced metabolic rates.

One exception to the effect of temperature at lower temperatures (<70°F), is
the apparent ability of the phenol oxidation organisms to adapt or acclimate
to the lower temperatures as can be observed in Figure 13.

EFFECTS OF DIFFERENT COMPONENTS IN RAW WASTEWATER

Different components in coke plant wastewater were examined for their effects
on the biological phenol-nitrification treatment process.  The components
examined were:  1) ammonia, 2) thiocyanate, 3) cyanide (free and complex),
4) light oil (by-product BTX), 5) sulfide and 6) phenol.  The objective of
examining the effects of these materials was to understand the biological
process such that in the event of a shock loading upset condition, proper
corrective measures could be implemented.

Ammonia

The effects of ammonia were examined by incrementally increasing the concen-
tration of ammonia in the feed wastewater.  Two pilot reactors, nos. 7
and 8, were used to (a) provide duplication of the observations and (b) to
observe the influence of mixed liquor sludge on those effects.

Figures 15 and 16 present the chronological observations on ammonia and thio-
cyanate concentrations.  Other pertinent data such as mixed liquor solids,
hydraulic residence time, feed ammonia and INj_, mean pH and N-oxidation rate,
k, are presented.  In Figure 15 it can be seen that with the reactor operating
at a HRT "5.2 days the feed ammonia concentration was increased from 210 mg/1
to 510 mg/1 in a 30 day period without any adverse effect on the effluent
ammonia and thiocyanate concentration.  It is also noted that k^ also increased
from 56.0 mg/l/day to 104.5 mg/l/day.  For the mixed liquor sludge concentra-
tion, the 56.0 mg/l/day oxidation rate was expected; however, the virtual
doubling of the oxidation rate was not expected.  Apparently, by increasing
the ammonia content of the feed wastewater, a population shift in the sludge
occurred, i.e., there was an increase in the number of nitrifying organisms.

Another interesting facet of the data shown in Figure 15, is the apparent
failure.  Up to the 35th day, the reactor functioned effectively, i.e., there
were low concentrations of ammonia and thiocyanate in the treated wastewater.
For the succeeding period, 36th to 50th day, the wastewater loading to the
reactor was increased, hydraulic residence time of 4.15 days, resulting in a
gradual increase in both ammonia and thiocyanate concentrations and the bio-
process appeared to be headed toward failure.

However, with a decrease in the reactor loading, hydraulic residence time of
5.5 days, for the final 10 days of the test, the ammonia concentration appears
to be stabilized at 40 mg/1 and tending toward even lower concentrations; the
                                      421

-------
                  Figure 15.  Effect of Increasing  the  EN Load on Reactor No. 7.
   160
IS> T-l
fS» 4J
  «
  H
    80
W
                Feed NH-j
                Feed EN±
                     HRT
                      PH
                Mean TVS
                     SRT
                             210 mg/1      310
                             285 mg/1      385
                             5.0 days      5.2
                             55 mg/I/day
                             7.2           7.3
                             3434 mg/1
                             230-287 days
460
535
5.24
103
7.3
510
585
5.4
104
7.3
510
585
4.15

7.4
5.5
93.5
7.4
                                                    Legend

                                                      O NH3
                                                      D SCN
                         10
                                                      Days  of  Operation

-------
OJ
     i
g
-
-
-
M
U
C«
<1)
-
=
:
U
         200
         160
         120
          80
     —
        _  Feed NH3
            Feed EN
                 HRT
          Reactor pH
                  kA
               MLTVS
                 SRT
                 <<) OH
                             Figure 16.  Effect of Increased £N Loading on Reactor No. 8.
260 mg/1          310
335 mg/1          385
5.0 days          4.3
7.2               7.4
67 mg/l/day        79
8200 mg/1
280-470 mg/1
>200 pg/1
360
435
5.2
7.2
 85
560
635
5.2
7.3
120
650
725
4.7
7.4
154
750
825
4.7
7.4
176
                                                  Legend

                                                  O NH3
                                                  D SCN
                              10
                                            20
                                       30
                            40
                                    50
                                                        Days of Operation

-------
Figure 16.  Effect of Increased ZN Loading on  Reactor No.  8.  (continued)
750
825
4.1
7.5
201
     70
     80                 90

Days of Operation
100
                                                                                                9.0
                                                                                                    P.
                                                                                                8.0
                                                                                 =
                                                                                 jj
                                                                             7.0

-------
thiocyanate concentration while high also, appeared  to have peaked and began
to decline.  During  the  final 10 days,  the average k was 93.5 ing/I/day which
compares well with the 104 mg/l/day observed prior to the 35th day.  In
effect the micro-organisms were effectively removing the substrate material
both in the feed and  that which had accumulated in the reactor.  The reactor
operational adjustment of reducing the  loading rate  (increasing the hydraulic
residence time) was an effective way of achieving recovery from the upset.

In the duplicate experiment, chronological data shown in Figure 16, the mixed
liquor sludge was about  8200 mg/1 TVS.  With this reactor sludge concentration,
an estimate of k^ is  about 120 mg/l/day from Figure  5.  Thus, the effective
treatment during the  initial 40 days of the experiment was not unexpected.
However, with the gradual increase in nitrogen loading to the reactor, feed
ammonia of 750 mg/1 and  EN of 825 mg/1, kA increased to about 200 mg/l/day;
again almost a two fold  increase.  On the 78th day,  the reactor loading was
again increased this  time both in feed concentration, ZN of 975 mg/1, and
hydraulic residence time of 3.9 days.  The reactor responded with an immediate
increase in ammonia concentration which gradually decreased with time.
At this point, the reactor appeared to be operating  in a stabilized manner.
On the 99th day, the  reactor loading was again increased by decreasing the
hydraulic residence time to 3.2 days.  This increase in loading resulted in
an increase in the reactor thiocyanate concentration and an eventual increase
in ammonia concentration.  The residual concentration of contaminants in the
reactor increased to  what appeared to be a failure of the system for nitrogen
removal.

From the two series of experiments shown in Figures  15 and 16, it can be con-
cluded that the wastewater feed ammonia concentration per se had little effect
on the effectiveness  of  the treatment reactors.  However, other factors such
as loading rates, hydraulic residence time, reactor  pH and sludge concentra-
tion greatly influence the treatability of feeds with high concentrations of
ammonia.

Thiocyanate

The effects of thiocyanate were examined in two reactors, no. 3 and no. 4.
In reactor no. 4, the feed thiocyanate concentration was incrementally
increased from the normal 300 mg/1 to 500 mg/1.  This reactor was observed
for a period of 55 days  and there appeared to be no  adverse effect on the
performance of the reactor as a result of the increased concentration of
thiocyanate in the feed wastewater.  Table 3 presents the performance data
for this test, and the average effluent quality for  the test was as follows:
1) ammonia concentration = 5 ± 5 mg/1, 2) thiocyanate concentration «• 9 ±
16 mg/1 and 3) phenol concentration «= 77 ± 219 yg/1.   The reactor operating
conditions were:  1) hydraulic residence time = 3.9 ± 1.0 days, 2)  mixed
liquor sludge = 3465 mg/1 TVS, 3) kA - 63 mg/l/day, 4) pH - 7.2 ± 0.2 and
5) temperature - 80-90°F.

Reactor no. 3 was used to examine the effect of direct spiked addition of
thiocyanate on the performance of the bio-process.  Table 4 presents the per-
formance data for this series of tests and the indications are that spiked
concentrations of thiocyanate up to 40 mg/1 did not have any adverse effect
on the reactor operating at the conditions shown.
                                      425

-------
        Table  3.    EFFECT OF FEED SCN CONCENTRATION ON PERFORMANCE
                   OF A REACTOR WITH 3465 rag/I TVS
Reactor Concentration^3) Feed^3)
Day of
Operation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15'
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55

Reactor
PH
7.2
7.0
7.1
7.0
7.2
7.0
7.1
7.1
7.1





7.0
7.3
7.1
7.2
6.9
7.1
7.3
7.2
7.2
7.3
7.0
7.3
7.6
7.3
7.6
7.5
7.4
7.2
7.2
7.0
7.1
7.0
7.0
7.4
7.2
7.3
7.4
7.3
7.0
7.3
7.5
7.2 ± 0.2
HRT,
Days
6.2
5.3
4.7
4.3
4.2
4.0
4.5
4.8
4.5





6.7
4.2
3.6
3.7
3.6
3.7
4.0
3.6
3.8
3.9
4.9
5.6
5.9
3.4
3.3
3.2
3.4
3.4
2.7
2.8
3.0
3.4
3.2
3.4
4.6
3.5
2.8
2.7
3.3
2.4
2.7
3.9 ± 1.0
NH

6
8
5
3
9


6
6
7
5
3


4
4
6
9

2
4
0
0
0


0
2
0
0
2


16
4
15
10
0


1
1
25
2
2
5 ±
3 SCN
mg/1


8




12

3

5


4


0

3

2

2


8

12

73


6

7

7


4

5

6
5 9 ± 16
4>OH
pg/1


62

20


30

0

0


16


7

3

148

10


16

10

966


16

3

2


16

2

143
77 ± 219
SCN
mg/1
300
300
300
300
300
300
300
350
350
350
350
350
350
350
400
400
400
400
400
400
400
400
400
500
500
500
500
500
500
500
500
500
500
500
500
500
500
300
300
300
300
300
300
300
300

(a)  Other feed components NH*
(b)  Negative numbers observed
-150 mg/1; $OH - 500 mg/1, Alkalinity
due to variability in actual numbers.

        426
60 mg/1

-------
   Table 4.   EFFECT OF SPIKE ADDITIONS OF SON ON REACTOR PERFORMANCE <
                                    Reactor Concentration
           Day of    Reactor  HRT,   NH3    SCN    <|>OH       SCN
          Operation    pH     Days   _ mg/1 _    yg/1      mg/1

              0        7.2     4.8    3
              1        7.2     4.6    6      7      45
              2        7.2     4.1    4
              3        7.4     4.0    9             34
              4        7.1     4.4
              5        7.2     4.5
              6        7.2     4.0   16      1      16        10
              7        7.2     4.3    9                       10
              8                      13      8      11        10
              9                       6                       20
             10                       8      6      10        20
             11
             12
             13        7.0     3.7    8      3       0        20
             14        7.3     3.9    7                       30
             15        7.2     3.3    7
             16        7.2     3.8    9
             17        7.0     4.1                            40
             18        7.1            02
             19        7.6     4.0    5
(a) Reactor operated with 3790 mg/1 TVS mixed liquor solids; temperature
    74-86°F.

(b) SCN added to reactor to give concentrations shown.
                                     427

-------
In the test series shown in Figure 15 (ammonia concentration evaluation),
where the reactor was operated at a pH ~7.4, the thiocyanate accumulated up
to about 160 mg/1 without any apparent adverse effects on the nitrification
reaction.  On the other hand, the data in Figure 9 (pH evaluation) show
nitrification was adversely affected at a thiocyanate concentration about 90
mg/1 at a pH of 8.3, indicating a possible aynergism between thiocyanate con-
centration and pH.  This possibility was further reinforced by the observations
shown in Figure 16 (ammonia concentration evaluation) where the eventual
reactor failure could be attributed to the combined effect of high pH and
high thiocyanate concentration.

Free Cyanide

The coke plant wastewater used in this study program contained less than 0.1
mg/1 of free cyanide, CN^, and to determine the effect of this material, the
feed water was spiked to 40 mg/1 CNf.  The response of the biological treat-
ment system to the free cyanide spiked feed is shown in Figure 17.  Because
of the generally low level of free cyanide in the wastewater, the micro-
organisms were not fully prepared to respond to the new feed.  The result was
an accumulation of free cyanide in the reactor.  The presence of free cyanide
adversely affected the nitrification reaction resulting in the ammonia concen-
tration increasing to about 120 mg/1.  The most surprising feature of this
experiment was the ability of the nitrifying organisms to acclimate or tolerate
the high free cyanide concentration, 12 mg/1.  This acclimation or tolerance
resulted in the rapid reduction of the reactor ammonia concentration.
Throughout this test, the reactor thiocyanate concentration was about 2 mg/1
and the phenol concentration was less than 0.2 mg/1.  This observation was
surprising because it was previously reported'^' that 0.5 mg/1 free cyanide
completely inhibited the nitrification reaction and 3.0 mg/1 inhibited the
thiocyanate reaction.  However, it does demonstrate that the nitrifying
organisms can tolerate and adapt to high concentrations of free cyanide.

Complex Cyanide

The complex cyanide content of the wastewater used varied from 6 to 108 mg/1.
Furthermore, it was noticed that these compounds tended to pass through  the
biological reactor unaltered and without exerting any adverse effect on  the
reactor performance.  In spite of these observations, a test reactor was
assembled to monitor the effect of complex cyanide.  The raw wastewater was
spiked with potassium ferri-cyanide, K.3Fe(CN)g, to produce a complex cyanide
concentration of 84 mg/1.  Figure 18 shows the complex cyanide concentrations
observed during the operation of the test reactor and compares these observa-
tions with a predicted profile.  The basis for the prediction profile are:
1) no alteration of complex cyanide, 2) operating conditions of the reactor,
i.e., hydraulic residence time of 3.9 days and 3) an initial complex cyanide
concentration of 25 mg/1 in the reactor.  This prediction profile is mathema-
tically described as:

                            C = Ct - 59e~°'256T                    (13)

There is no real reason for the difference between the prediction and observed
values, and the magnitude of these differences does raise some questions re-
garding the ineffectiveness of biologically degrading the complex ferro-
cyanides.  In the course of this test, the following are the reactor conditions:
1) ammonia concentration - 3.7 ± 2.6 mg/1, 2) thiocyanate concentration -


                                     428

-------
       Figure  17.  Effect  of  Free Cyanide on Nitrification.
  0.4
•:•
  0.2
 :
a
  100
   75
CO
iJ  50
ai

SJ  25
                       Start  Feed  w/CNF = 40 rag/1
 Legend

O  NH3

   Free CN

                     5              10

                         Days  of  Operation
               15
20
                              429

-------
                          Figure 18.  Complex Cyanide Profile Through Reactor No. 7.
tA>
o
                80
                60
             g  40
             iH


             I
             U
                20
                          No Loss of CN,
                                 - 59e
                                      -0.256t
                                                  I
                                                  10              15


                                                      Days  of  Operation
20
25

-------
3.6 ± 2.6 mg/1, 3) phenol concentration = 32 ± 26 yg/1, 4) hydraulic residence
time - 3.9 days, 5) pH = 7.3 ± 0.2, 6) temperature = 80-90°F and 7) reactor
dissolved oxygen concentration = >1.0 mg/1.

Light Oil

In the by-product operation of coke plants, light oils are produced.  These
oils are mixtures of benzene, toluene and xylene (BTX).  In the course of
these product operations, there is always the possibility that some of these
light oils could reach the wastewater treatment system.  With this possibility
in mind, the effects of these light oils on the performance of the' biological
treatment process were examined.  Because these light oils are nearly immiscible
in the wastewater, it was believed that they would enter the biological treat-
ment system as a "slug."  Thus the procedure for this evaluation was to add
certain quantities of the light oil directly to the reactor.

The performance data for the test reactor are presented in Table 5.  Despite
the two ammonia peaks at days 2 and 25, the overall performance data appear to
be very similar to data obtained from control reactors maintained during the
study.  In effect the light oil does not appear to have any direct adverse
effect on the nitrification reaction.

There is a potential  for an indirect adverse effect, i.e., reduction of the
oxygen transfer and mixing capabilities of the aeration equipment.  Light oils
entering the aeration basin could change the surface chemistry characteristics
of the mixed liquor; the most likely characteristic which could be altered is
the surface tension of the liquor.  A severely decreased surface tension could
adversely affect both mixing and oxygen transfer.

Sulfide

In most coke plants where vacuum-carbonate desulfvirization is practiced, the
blowdown from this unit is processed through the ammonia still with the waste
ammonia liquor.  In the free leg still, almost all of the acid gases including
H2S are removed.  The ammonia still waste used in the course of this study
contained relatively low concentrations of sulfide, about 25 mg/1.  However,
there is always the potential of high sulfide containing wastewaters being
inadvertently routed directly to the wastewater treatment system.

The effects of a wastewater containing high concentrations of sulfide were
examined by progressively increasing the sulfide content of the feed waste-
water from the normal 25 mg/1 to 500 mg/1.  Figure 19 summarizes the results
of this evaluation showing the feed wastewater sulfide concentration had
virtually no effect on the reactor ammonia, thiocyanate and phenol concentra-
tions.  It was, however, observed that the wastewater sulfide tended to
suppress the pH of the reactor requiring more frequent measurement and con-
trol.

Direct addition of sulfide to concentrations up to 40 mg/1 to another test
reactor also had no effect on the treatment process.

Throughout the course of this study program the reactor effluent sulfide con-
centrations were generally less than 0.5 mg/1.
                                     431

-------
   Table 5.    EFFECT OF LIGHT OILS ON THE PERFORMANCE OF A SINGLE-STAGE
              PHENOL-NITRIFICATION REACTOR(a)

                                 Reactor Concentration
Day of
Operation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Reactor
PH
7.1
7.1
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.1
7.0
7.3

7.2
7.2
7.2
7.3
7.3
7.2
7.3
7.5
7.6









7.2
7.1
7.2
7.1
7.0
7.4
7.6
HRT,
Days
4.3
4.0
4.3
4.2
4.7
4.3
4.2
4.2
3.9
3.9
4.1
4.0
3.8
4.7
4.3
4.1
4.2
4.0
3.6
4.1
3.9
3.7
3.9
4.2
4.0
3.8









2.9
4.0
3.5
3.5
3.2
3.1
3.7
NH3
mg/1
6
21
5
2
4
—
-
10
5
6
7
6
—
—
2
8
6
10
7
-
-
3
3
6
22
8
-
-
6
5
6
10
4
8
4
10

6
8



SCN

2
4
1
0



0

5

0


0

4

0


3

0

5


5

9

2


4

0




4>OH
yg/1
0
0
0
0



41

20

34


10

20

3


92

20

30


45

45

13


0

13




Oil Addition
mR/1
10


10


20
20
50


100


200



500




1000




1500







2000





                                 7±5   3±3   23 ±24

(a)  Standard Feed:  NH3 - 150 mg/1; SCN - 300 mg/1; $OH - 500 mg/1; ..Reactor
    Mixed Liquor Solids - 4330 mg/1 VSS; Temp. ~74-84°F; Reactor DO >1.0 mg/1
(b)  By-product Light Oil added directly to reactor in amounts to yield con-
    centration shown.

                                     432

-------
              Figure  19.  Effect of Sulfide on Nitrification of

                          Coke Plant Wastewaters.
 e
 o
 o
o
 o
 4-1


 w
0)

8P
M


I
   30
   20
   10
O  Ammonia Cone. - mg/1



D  Thiocyanate Cone. - mg/1



A  Phenol Cone. - jjg/1
                  100         200          300          400


                            H2S Concentration  in Feed
                   500
                                      433

-------
Phenol

Phenol is a major constituent of coke plant wastewaters.  The concentration
of phenol in the wastewater depends on the type of coal being processed, the
coking operation and the by-product recovery practice.  The phenol concentra-
tion of coke plant wastewaters could be as high as 2000 mg/1.  The ammonia
still waste used in this program contained an average of about 120 mg/1.
This concentration was adjusted to 500 mg/1 for the standard feed.  The
effect of phenol on the performance of a phenol-nitrification reactor was
examined from two directions; 1) spike additions to the reactor and 2) pro-
gressive increases in the feed phenol content.

     Spike Additions.  Phenol was added directly to the reactor in amounts
which would produce a predetermined instantaneous reactor concentration.  The
additions were made in progressive increments starting  from 0.5 mg/1 up to
20 mg/1.  Table 6 presents the data for this series of  tests and  the indica-
tions are that the nitrifying organisms were capable of acquiring tolerances
up to about 30.0 mg/1 without adversely affecting nitrification.

     Increases in Feed Phenol.  The effect of increased feed wastewater phenol
content was examined in reactor nos. 5 and 6.  The mixed liquor solids were
the major difference between the two reactors.  Figure  20 presents the per-
formance profile for reactor no. 5.  Throughout the 40 days of operation with
the high phenol feed there was effective nitrification and phenol removal.
Similar performance was observed in reactor no. 6, data from which are shown
in Figure 21.  However, it is noted that after the 30th day both  the thio-
cyanate and phenol concentrations gradually increased.  While it  is evident
that the phenol concentration was increasing it must be recognized that the
concentrations were still relatively low, about 200 yg/1.  The thiocyanate
concentrations, on the other hand, were significantly high, about 150 mg/1.

The precise cause for the increase  in phenol and thiocyanate concentrations
after the 30th day is not evident.  However, the inability to account for 81%
of the wastewater nitrogen, alludes to the possibility  that denitrification might
have occurred during this period.  Denitrification would occur if there were
low dissolved oxygen levels in  the  reactor.  If this was the situation, then
there could have been competition between the nitrifiers and the  thiocyanate
oxidizing organism for the available oxygen.  The results  tend to show  that
the nitrifiers prevailed.  It was unfortunate, due to dissolved oxygen
measuring equipment problems,  that  the reactor dissolved oxygen levels which
are generally taken during a test series were not taken during this evaluation.
However,  the few readings taken during the  final stages of this evaluation
somewhat  confirms the supposition presented  above, where,  at the  low  dissolved
oxygen level, "0.5 mg/1,  there was  a definite trend  for both the  phenol and
thiocyanate concentration levels  to increase.

In  the operation of reactor nos. 5  and 6, there was  no  deliberate sludge
wasting.  The only sludge lost was  that in  the effluent, about 80-100 mg/1.
This mode of operation insured  the  retention of the  nitrifying organisms  in
 the  reactor and  is quantitatively expressed  by the extremely long sludge
 retention times  of 179-325 days.  Under a more traditional mode of  operation,
 sludge retention times of  30-50 days,  the nitrifiers and  thiocyanate  organisms
                                       434

-------
     Table 6.   EFFECT OF SPIKE ADDITIONS OF PHENOL  ON THE PERFORMANCE
                OF A PHENOL-NITRIFICATION REACTOR(a)
       Day of
      Operation

          0
          1
          2
          3
          4
          5
          6
          7
          8
          9
         10
         11
         12
         13
         14
         15
         16
         17
         18
         19
         21
         22
         23
         24
         25
Reactor
  PH

  7.4
  7.5
  7.6
  7.4
  7.4
  7.4
  7.4
  7.3
  7.5
  7.6
  7.6
  7.5
  7.2
  7.6
  7.4
  7.3
  7.6
  7.4
  7.5
  7.8
HRT,
Days

 4.0
 3.9
 4.3
 4.0
 4.3
 4.4
 3.1
                                Reactor Concentration
NH3
  SCN
(mg/1)
OH
 0
 0
 0
 0
 0
 0
 5
 0
  47

  38
3.6
4.3
4.3
4.3
4.3
4.3
4.1
4.2
3.2
4.1
4.6
4.8
4.9
0
0
0
0

24
25
26
0

0

0
               *

               *
 0.05

 0.12


 0.05

 0.02

14.75


 2.32

20.60


17.00

29.50
  *

  0

21.21
         OH Addition^)
             (mg/1)
            0.5
            0.5
            0.5
            1.0
            1.0
                                 10
                                 10
                                 20
(a) Reactor operation conditions:
    •Average mixed liquor solids - 3620 mg/1 TVS
    •Temperature
    •All sludge produce was retained in the reactor, except for that lost
     in effluent ~100 mg/1 SS.  Approximate SRT "134 days.
(b) Phenol added directly to reactor in quantities which would produce con-
    centrations shown.  Additions were made on days shown only.

(c) Analytical result was unrealistic and questionable.
(d) Reactor DO was ~0.2 mg/1.  Generally the DO was >1.0 mg/1.
                                   435

-------
                 Figure  20.   Effect of High Feed Phenol Concentration on Nitrification - Reactor No. 5,
                     Feed  Concentration
                       NH3
                       SCN
                    150 mg/1
                    300 mg/1
                    1250 mg/1
                                          1500 mg/1     3500 mg/1     4000 mg/1
                                                  200 mg/1
                                                  300 mg/1
                                                 4000 mg/1
-£»
(A>
CTi
      t-i
      f.
      a
      tt
      
-------
                     Figure  21.   Effect of High Feed Phenol Concentration on Nitrification in Reactor No.  6.
    0.4
    0.2
 a)
 .
P. 6
•e- 0j

  to
  c
  o
 a
 OH - 4000 mg/1
                     D.O = 3.7
                 OH « 5000 mg/1
                                   .4>OH
                       10
                                        20
                                                                                81% Denitrification    D.O ~0.5
30              40



 Days of Operation

-------
would be washed from the reactor and in all likelihood an adverse effect would
have been observed.

From the observations on reactor nos. 5 and 6, it would appear that the key
factors maintaining nitrification in situations of high phenol wastewater con-
centrations are long sludge retention times and adequate mixed liquor dis-
solved oxygen.  Throughout the entire study program, phenol removal efficiencies
greater than 99.9% were achieved under an extremely wide range of conditions.

PROCESS ENHANCEMENT

From the preceding section, it is evident that the reaction rates for the
nitrification reaction are relatively slow.  Thus, any economical means of
enhancing these reaction rates would be of significant value.  In this program
three approaches were examined as possible modes of enhancing the phenol-
nitrification process.  These are:  1) addition of activated carbon to the
reactor, 2) the application of mutant strains of bacteria and 3) carbonate
nutrient supplement.

Activated Carbon Addition

There are numerous reports on the ability of activated carbon to extend or
enhance the capacity of the activated sludge treatment system when added to
the mixed liquor.(3, 4, 5)  Further, with impending BAT effluent limitations
on priority organic compounds there was the potential of the added benefit
of controlling these pollutants by simply adding controlled amounts of
activated carbon to the aeration basin of the activated sludge process.  With
the prospects of these benefits in mind, a preliminary evaluation of activated
carbon addition to the phenol-nitrification process was undertaken.

This preliminary evaluation was designed to test the effects of a one time
addition of carbon on the performance of a phenol-nitrification reactor.
Sufficient finely grounded activated carbon was added to the reactor to
increase the suspended solids level by 1400 mg/1.  This reactor was operated
with a standard feed wastewater and the performance characteristics are pre-
sented in Table 7.  The test period can be divided into four segments -

                   Segment A - Before carbon addition
                   Segment B - Carbon addition
                   Segment C - Acclimation
                   Segment D - True evaluation

In segment A, before the carbon addition, the reactor performed at a k^ rate
of 78 mg/l/day.  For the mixed liquor sludge used, a k^ rate of 80 mg/l/day
is predicted, (see Figure 5).  In segment B, immediately following the car-
bon addition, the observed oxidation rate was 85 mg/l/day, well within the
limits of that predicted and previously observed.  However, after seventeen
days of operation, about day 23, there was a gradual increase in the reactor
thlocyanate concentration; an event which covered about 15 days.  Prom the data
in Table 7 there appears to be no reason for this excursion in the thiocyanate
reaction.  Also, it appeared as if the effort to push the reactor, days 20
through 26, by increasing the feed flow could have resulted in the increased
                                      430

-------
 Table 7.   PERFORMANCE OF REACTOR NO. 5 WITH 1400 mg/1 ACTIVATED CARBON
            ADDED TO THE MIXED LIQUOR^)
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
  20
  21
  22
\ 23
  24
  25
  26
  27
  28
  29
  30
  31
  32
  33
  34
  35
  36
  37
  38
  39

£H
7.3
7.1
7.3
7.0
7.0
7.0
7.1

7.6









7.3
7.1
7.6
7.5
7.4
7.6

7.5
7.5
7.4
7.2
7.5
7.4
7.5
7.4
7.2
7.2
7.3
7.2
7.2
7.2
7.2
7.2
HRT,
Days
3.3
2.9
2.3
2.6
2.7
2.8
2.5
2.7 ± 0.3
2.7
2.8
3.2
2.3
2.4
2.1
4.3
2.8
3.4
2.8
2.6
2.4
1.8
1.6
1.8
2.6
2.6 ± 0.7
2.9
1.7
1.7
4.3
1.9
5.1
5.5
5.8
—
11.6
5.7
5.4
5.0
4.3
4.4
4.2
DO
mg/1
3
4
4
4
2
3
3

3









6
6
3
3
5
5

3
3
3
2
6
4
4
4
3
5
5
5
5
5
3
5
NH3
mg/1
12
-
-
0
32
3
18
13 ± 11
0
6
0
0
14
0
0
0
—
1
0
1
1
9
-
-

44
20
6
9
12
-
3
2
4
8
8
-
-
6
3
10
SCN
mg/1
4


4

2

3
2
2
2
1
2
2

0

2

3

2



20

25

12


103
67

81


81
47
0
cfrOH
yg/1
22


3

0

8
0
0
0
0
0
4

0

0

20

13



16

41

34


13
56

7


0
71
52
                                                         Activated Carbon Addition
                                                         to 1400 mg/1 SS on day 7
k • 85 mg/l/day
                                    439

-------
Table 7.   PERFORMANCE OF REACTOR NO. 5 WITH 1400 mg/1 ACTIVATED CARBON
           ADDED TO THE MIXED LIQUOR^8) (CONTINUED)
Reactor Conditions and Concentrations
Day of
Operation
/ 40
/ 41
' 42
43
44
\ *5
\ 46
I 47
48
49
50
51
52
53
54
/ 55
/ 56
/ 57
58
59
60
61
62
63
\ 64
\ 65
V 66

pH
7.2
7.1
7.2
7.3
7.5
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.1
7.2
7.2

7.2
7.2
7.2
7.2
7.2
7.3
7.3
7.3
7.3
HRT,
Days
4.3
4.4
4.6
2.6
4.1
3.9
3.9
5.1
4.6
4.9
4.4
4.8
4.3
4.0
4.3
4.4
6.5
7.9
4.8
4.1
3.3
3.4
2.9
3.6
3.8
2.8
2.4
DO
mg/1
5
4
5
5
5
4
4
4
4
4
4
3
4
2
5
5
5

5
3
5
5
5
5
4
5
4
NH3
mg/1
7
-
—
10
0
0
0
2
_
3
11
0
2
1
—
—
7
9
9
9
-
-
3
9
9
9
3
SCN
mg/1
0


9
3
5
5
3

0
0
0
0
0



7

5


1

0

3
OH
yg/1
34


41
80
10
0
0

0
0
0
0




84

30


128

41

37
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
7.4
7.1
7.3
6.9
6.9
7.6
8.0
7.9
7.9
7.8
7.4
7.3
7.2
6.8
6.6
7.1
7.0
1.5
1.9
1.7
1.6
1.6
^ 1.4
1.7
1.6
1.5
-
-
1.9
1.7
1.4
1.5
1.5
1.6
6
5
2
2
5
1
1
2
2
5
4
3
4
4
3
4
0.6
9

2
12
13
0
2
4
11
0
1
7
0

4
5
4
0

0

3
4
19
92
125
5

6


0
21

0

10

3
10
27
34
10
13

0


23
23

                                                          high pH
                                   440

-------
  Table 7.   PERFORMANCE OF REACTOR NO. 5 WITH  1400 mg/1 ACTIVATED  CARBON
             ADDED TO THE MIXED LIQUOR*a>  (CONTINUED)
 Day of
Operation
Reactor Conditions and Concentrations	
     HRT,      DO     NH3      SCN    <|>OH
              mg/1    mg/1     mg/1   yg/1
7.4
7.6
7.5
7.7
6.5
6.8
6.5
6.5
6.5
6.9
7.3
7.5
7.4
7.4
6.6
6.6
7.1
7.1

1.6
1.7
1.7
1.7
—
1.5
1.3
1.4
1.4
1.2
0.9
1.9
1.3
1.4
1.8
1.9
1.3
1.5
1.6 ± 0.2
0.8
0.6
0.6
0.6
2.2
0.8
0.4
6
4
3
4
5
4
5
5
3
5
5

                                    6
                        3
                       62
                       78
                       99
                       15
                       35
                       54
                       31
                       26
                        0
                        0
                        0
                        0
                        0
                       13
                       16
                        3
                        0
                       ±
(a)  Reactor temperature  79-88°F
     Feed to reactor:   -NH3 = 150 mg/1
                        •SCN = 300 mg/1
                        •OH - 500 mg/1
   0

   0
                                                 7
                                                 1
                                                 7
 0



 0

16

 0


23
                                                             low DO
                                                             low DO
                                                            k  «  136  mg/l/day
   5      0
4 ± 5(c,d)
(b)  Values on days 85-88 and 89 and 90  omitted.

(c)  Values on days 73  to 75 omitted.

(d)  Negative numbers observed due to variability of actual numbers.
                                      441

-------
ammonia concentrations.  A possible explanation is that the carbon might have
gleaned out trace quantities of toxic or inhibitory materials from the waste-
water  thus creating an environment of relatively high levels of these toxic/
inhibitory materials.  The results were ammonia and thiocyanate excursions, and
the need to re-acclimate the sludge.

Following this acclimation period, the reactor was once again pushed and in
segment D it can be seen that high levels of treatment were achieved.  There
were a few accountable excursions.  Nevertheless, the oxidation rate was
much higher than that previously observed, RA of 136 nig/I/day.  It is not
conclusive that this increased oxidation rate was due to the carbon because
no effort was made in the course of this experiment to determine the amount
of new sludge formed.  Further, in  previous experiments it was observed that
the biological sludges used can be pushed to produce much higher oxidation
rates than those predicted by the correlation shown in Figure 5 (k^ =15.2 TVS
mg/l/day).

From the above data, the ability of the activated carbon to enhance the bio-
logical reaction when added to the mixed liquor remains unconfirmed and should
be re-examined.

The addition of activated carbon to the experimental reactors did produce
other positive effects, these include: 1) the elimination of foaming in the
reactor, 2) improved settleability of the sludge and 3) significant improve-
ment in the color of the effluent from a definite brown to a pale straw color.
Analysis for priority organic pollutant control showed that the activated
carbon addition to the mixed liquor did not produce any better control than
other reactors which operated without carbon.  This observation must be
qualified by the facts that  (a)  carbon  addition  to  the reactor was a one-time
addition and (b) the sample analyzed was taken after 1055 1 of wastewater had
been treated over an operating period of 103 days.  These conditions could
have been well beyond the breakthrough point for the carbon used.   The impact
of the priority pollutants regulations and the potential of this mode of oper-
ation as a control technology are more reasons for another examination of
carbon addition to the mixed liquor.

Application of Commercially Available Mutant Bacteria

In the preliminary studies which lead up to the present program, the experi-
ence after an upset was a lengthy time consuming effort to re-establish nitri-
fication in the test reactor.  One of the goals of this program is to develop
an understanding of the process such that more effective corrective measures
could be taken in the event of an upset.  However, it was also felt that other
means of accelerating the recovery would be beneficial.  Certain suppliers of
mutant strains of bacteria claim their products are capable of accelerating the
recovery of biological treatment systems that are experiencing problems.  Thus
it was decided to test two mutant strains — a "hydrocarbon degrader" and an
"ammonia recoverer."

These organisms were tested in a combined dose at rates recommended by the
supplier.  The mixture was added to a reactor in which nitrification activity
was disrupted and the data in Figure 22 show the chronology of events.   On the
basis of the present understanding of the biological process involved in the
                                     442

-------
a
 o

 U
               Figure  22.   Effect of Mutated Bacteria on Ammonia,  Thiocyanate

                            and Phenol Degradation in a Pilot Reactor Experiencing
                            Loss of Nitrification and Phenol Oxidation.
      7




      6
    300
oc
g
-e-
4-1
a
QJ

x
01
en
C
o
n)
QJ
O


O
U

(-1
O
4J
U
CO
01
   200
   100
          (a)
      Legend

(a)  Added 1 gm Phenobac

    + 1 gm Nitrobac

(b)  Added 1 gm Phenobac

(c)  Added 5 gms Phenobac

    + 5 gms Nitrobac
  <}iOH
                                10          15

                                Days of Operation


                                      443
    20
25

-------
single stage phenol oxidation-nitrification of coke plant wastewater, it
appears that the initial failure  (loss of nitrification) of the reactor was
due to the synergistic effect of  the high thiocyanate concentration  (>300
mg/1) and high pH  (~8.6).  The three additions of the bacteria show no
immediate effect on phenol or ammonia.  It appears that the thiocyanate
response was due to the increase  in activity as a result of the change in pH,
and the return of  nitrification activity was in response to the inhibitory
stress of high thiocyanate and pH being relieved.  The results of the experi-
ment, the data in  Figure 22, appear to show no real clear cut benefit from the
bacteria addition.  Thus it was decided to repeat the experiment under more
controlled conditions.

Figures 23 and 24  show the chronology of events for the test reactor and a
control reactor.   In this experiment, there were rapid responses to high phenol
and thiocyanate concentrations in both test and control reactors. However, there
was a more rapid nitrification response in the test reactor; but in earlier
experiments there  were also observations of unexpectedly high nitrification rates
when the reactor received high ammonia loadings.  Thus, the indications are
still not clear that the commercially available mutant bacteria will be
beneficial in producing immediate recovery from an upset episode.  However,
the results are sufficiently interesting that this possibility could be applied
on a hit or miss understanding or a more detailed and controlled experimental
program be conducted to delineate the true potential.

Carbonate Nutrient Supplementation

The organisms responsible for thiocyanate and sulfide oxidation, and nitrifi-
cation are autotrophs.  They utilize carbonate as their carbon source in cell
synthesis.  It is  generally believed that carbonate supplied by the activity
of the heterotrophs and that solubilized during aeration would satisfy the
carbonate needs of the autotrophs.  However, there was no supporting evidence
in the reports reviewed.  Consequently, a test reactor was operated with
supplemental carbonate in the feed wastewater to test the premise that the
autotrophic reactions were not carbonate limited.

Sodium carbonate was used as the  carbonate source and not as a source of
alkalinity.  The feed to the test reactor was supplemented with 1000 mg/1 of
Na2CC>3.  This carbonate addition  did result in an increase in both pH and
alkalinity.  However, appropriate corrections were made to maintain the
standard feed alkalinity of about 1000 mg/1 and a reactor pH of 7.2 by acid
addition.  It is recognized that  acid addition would liberate some of the
carbonates as C02«  However, both the feed pH and the pH of the reactors are
maintained sufficiently high such that the bicarbonate pH end-point of 6.5
is not exceeded.   Thus, while some carbonate was lost, sufficient remained
to test the premise.

The performance data for this experiment is shown as that for reactor no. 4
in Table 2.  From  a comparison of the data for reactor no. 4 and the other
reactors it can be concluded that sufficient carbonate is supplied by the
activity of the heterotrophs and  through aeration to prevent a carbonate
limited condition.  In effect, carbonate addition did not enhance autotrophic
activity.

-------
•a
ll
o
u
O
a
   300
                  2.9 ppm
                      Legend

                      (A)   Added
                      (B)   Added
                      (C)   Added
                      (D)   Added
                                        1 Gram Phenabac & Nitrobac
                                        1 Gram Phenabac
                                        1 Gram Nitrobac
                                        1 Gram Phenabac
                t
               (A)
                  ttt
                (B)(C)(D)
15
20
                           Days  of  Operation
25
Figure 23.
    Effect of Mutated  Bacteria on Ammonia,  Phenol and Thiocyanate De-
    gradation at  a  pH  of  8.6.
                                  445

-------
        Wi
        o
           6 *"
        >300

         300
      oc
      a.
      g.  200


      I

      2
        100
      O
      o
                               10
15
20
25
                                                                    30
                                   Days of Operation

Figure 24.  Effect of Mutated Bacteria on Ammonia, Phenol and Thiocyanate De-

            gradation Corresponding to Control for Mutant Bacteria Experiment
            Shown on Figure 23.
                                      446

-------
CONTROL OF PRIORITY ORGANIC POLLUTANTS

Thirteen grab samples were taken for gas chromatograph/mass spectrograph, GC/MS,
analysis in an effort to determine the fate of priority organic pollutants in
the course of the treatment of coke plant wastewaters.  The treatment train
under consideration is shown in Figure 25.  Each sample was prepared and
handled in accordance with the EPA specified protocol.  The analysis examined
78 organic compounds covering the range of purgables  (volatiles), base/neutrals
and acid extractables.  This discussion will be directed at only those compounds
which were detected.

The thirteen samples analyzed covered two treatment schemes:

          •  A conventional phenol biological reactor (Weirton
             Steel's Brown's Island Coke Plant), and

          1  An advanced single-stage phenol-nitrification bio-
             logical treatment.

Table 8 presents the analytical data for the conventional phenol biological
reactor scenario.  This data strongly suggest a significant reduction in the
purgables as the wastewater traversed the ammonia still.  However, higher
concentrations of purgables appear in the bio-plant effluent.  Overall there
were lower concentrations of priority organic compounds in the bio-plant
effluent than the ammonia still waste entering the plant, especially the base
neutrals and acid extractables.

Table 9 presents the data for the advanced single stage scenario,  more pre-
cisely treatment across a phenol-nitrification activated sludge reactor.  This
data also show an overall decrease in the concentrations of the priority
organic compounds.  On the average, the concentration of the compound detect-
ed was less than 10 yg/1.  There was only one exception, methylene chloride.
Further, the concentrations observed in the effluent from the experimental
bio-reactors were much lower than that from the Brown's Island plant.

In evaluating the data in Table 9 it is apparent that the effluent qualities
from the different reactors are all about equivalent.  However, it must be
recognized that reactor no. 5 dated November 13, 1979 was operated with
activiated carbon added to the mixed liquor.  It appears that the carbon
addition did not improve the quality of the treated effluent relative to
controlling the priority organic pollutants.  This observation should not
be viewed as being totally negative because it must be noted that  the carbon
addition was a one-time addition and the effluent sample for GC/MS analysis
was taken after about 1055 liters of wastewater had been treated over a 103
day period. - It is very likely that after this extended service that the
adsorptive capacity of the carbon might have been exhausted.

In a comparison of the phenol values determined by the GC/MS procedure and the
Standard Methods wet chemistry procedure, there were significant discrepancies,
as shown in Table 10.  Three major factors are believed to be responsible for
these discrepancies.  The first factor is the fact that several phenols, other
                                      447

-------
Flushing Liquor,
Light Oil Inter-
ceptor Sump,
Barometric Conden-
ser,- Desulfurizer
Slowdown
Ammonia
 Still
                                          ASU
                            Bio-Plant
                               or
                            Experimental
                             Reactor
                                                                                  Bio-Plant/Reactor
                                                                                       Effluent
 Figure 25.  Treatment Train for By-Product Coke PJant Wastewater under
              Consideration for Effect on Priority Organic Pollutants.

-------
      Table 8.    PRIORITY ORGANIC POLLUTANT PROFILE THROUGH THE WEIRTON STEEL
                 BROWN'S ISLAND COKE PLANT WASTEWATER TREATMENT SYSTEM
                                                     ASW
                             11/7/79
                   Bio Plant
                  Effluent^'
                   9/14/79
9/15/79    11/7/79
  (all concentrations in yg/1)
                                ND
12.34
               ND
3.04
           Bio Plant
          Effluent(c)
           11/12/79
210.73
ND(^)

2.26
185.21
4,507.34
960.89
ND
6,356.25
449.34
ND
31.59
ND
ND
0.53
0.24

7.81
3.92
2.35
ND
125.64
30.24
38.52
41.40
64.11
23.76
48.30
48.52
ND
•
ND
ND
257*88
ND
ND
14.51
ND
ND
ND
ND
ND
0.41
0.16

11.48
3.78
0.25
ND
ND
1.00
2.21
ND
5.39
0.34
2.95
310.22
ND

5.82
302.82
17.01
7.00
ND
ND
10.56
ND
ND
ND
ND
                                       ND
Sample Source

Date
Compound

Purgables
methylene chloride
1,1-dichloroethane
chloroform
1,1,1-trichloroethane
benzene
toluene

Base Neutrals
bis(2-chloroethoxy)methane
naphthalene
acenaphthalene
acenaphthene
diethylphthalate
1,2-diphenylhydrazine
N-nitrosodiphenylamine
phenanthrene
anthracene
di-n-butylphthalate
fluoranthene
pyrene
benzo(a)anthracene\
chrysene          J
butylbenzylphthalate
benzo (b) fluoranthene*)
benzo(k)fluoranthene)
benzo(a)pyrene
Acid Extractables
phenol
2,4-dimethylphenol
(a) All wastewaters (flushing liquor,  light oil interceptor scrap,  barometric con-
    denser, and desulfurize blowdown)  collected together for processing  through  an
    ammonia still.
(b) Feed to bio-plant, normal operation of ammonia still (consistent with  present
    discharge permits).
(c) Brown's Island operating conditions: MLSS '1000 mg/1;  HRT "2.2  days; SRT "5-10
    days.
(d) Analyzed for but not detected.
ND
324.70
312.38
ND
58.22
ND
ND
12,073.50
ND
632.88
261.33
418.20
1,591.80
ND
ND
ND
97,834.80
302.35
ND
1,038.00
837.76
525.70
ND
236.32
441.90
21,897.44
ND
23.63
13.29
28.06
85.54
ND
ND
ND
ND
0.01
ND
210.71
210.57
421.74
ND
276.48
744.19
1.02
ND
                                         449

-------
      Table 9.   PRIORITY ORGANIC POLLUTANT PROFILE THROUGH PHENOL-NITRIFICATION BENCH SCALE ACTIVATED
                 SLUDGE REACTORS TREATING COKE PLANT WASTEWATER
No. 2
Feed
8/23/79
ASw(a)
11/7/79
No. 1&5
FeedO>>
9/14/79
No. 1
Effluent
9/19/79
No. 5
Effluent
9/19/79
(all concentrations in
5.14
ND^d)
8.54
2.17
0.37
ND
ND
ND
93.14
66.13
ND
213.44
ND
502.16
467.19
527.00
1,310.23
ND
ND
ND
ND
60,378.00
174.91
5.78
ND
1.92
8.67
18.78
ND
ND
ND
43.38
ND
ND
ND
ND
ND
528.66
462.84
560.58
ND
242.59
1,716.34
ND
32,249.32
255.04
1.91
ND
0.43
2.64
1.01
ND
467.76
ND
37.26
161.46
5.99
ND
31.95
869.52
152.55
162.69
457.02
W>
ND
ND
ND
65,820.00
256.34
71.28
ND
1.18
2.81
0.17
ND
ND
ND
ND
0.28
ND
0.35
ND
2.47
0.22
0.33
ND
ND
ND
ND
N&
0.56
ND
75.75
ND
0.43
1.48
0.17
ND
ND
Z.-14
4.O9
0.53
ND
ND
ND
8.98
1.74
1.64
ND
ND
ND
ND
0.32
0.08
ND
No. 2
Effluent
11/12/79
Pg/D
417.18
ND
6.38
ND
8.42
ND
ND
0.19
0.23
ND
ND
0.98
ND
1.15
35.18
ND
2.70
1.03
2.28
10.16
ND
0.84
ND
No. 8
Effluent
11/12/79

5.43
ND
3.28
11.00
6.60
ND
ND
ND
0.57
ND
ND
ND
ND
0.87
3.48
3.07
6.65
ND
3.86
7.67
ND
0.15
ND
No. 5
Effluent
11/13/79

4.21
ND
1.00
17.90
10.70
ND
ND
ND
0.22
ND
ND
ND
ND
0.41
3.46
1.95
5.58
ND
0.87
7.80
ND
0.23
ND
Sample Source
Date

Compounds
Purgablea
methylene chloride
1,1-dichloroethane
chloroform
1,1,1-trichloroethane
benzene
toluene
Base Neutrals
bis(2-chloroethoxy)methane
naphthalene
acenaphthalene
diethylphthalate
1,2-diphenylhydrazine
N-nitrosodiphenylamine
phenanthrene
anthracene
di-n-butylphthalate
fluoranthene
pyrene            ^
benzo (a)anthracene£
chrysene          J
butylbenzylphthalate
benzo(b)fluoranthene\
benzo (k )fluorantheneJ
benzo(a)pyrene
acenaphthese
Acid Extractables
phenol
2,4-dimethylphenol

(a) Ammonia still operated to produce low ammonia concentration, < 150 ing/1.  This ASW subsequently amended with
    ammonia, phenol, thiocyanate and other components as required; this amended ASW was used as feed to experi-
    mental reactors.
(b) Same as (a) but used in the preparation of feed to reactors nos. 1 and 5.
(c) Reactor 5 operated with activated carbon added to the mixed liquor;
(d) Analyzed for but not detected.

-------
   Table 10.   COMPARISON OF TOTAL PHENOL VALUES DETERMINED BY GC/MS AND
               STANDARD METHODS PROCEDURES
                                           Total Phenol Concentration, yg/1
                Sample                          GC/MS     Standard Methods

ASW Feed to BI-Bio Plant, 9/14/79              98,000         280,700

WAL + (Charge to Ammonia Still), 11/7/79       12,244         207,400

ASW Feed to BI-Bio Plant, 11/7/79              21,897         176,600

ASW (Still Operated for Low NH3), 11/7/79      32,504         193,000

BI-Bio Plant Effluent, 9/14/79                   0.13              34

BI-Bio Plant Effluent, 11/12/79                  1.02              52

ASW* (Still Operated for Low NH3)              66,000         268,000
Feed for Reactor Noa. 1 & 5 Before
Alteration

Feed ASW to Reactor No. 2, 8/23/79             61,000         500,000

Effluent from Reactor No. 1, 9/19/79             0.66          ND^)

Effluent from Reactor No. 5, 9/19/79             2.44          ND

Effluent from Reactor No. 2                      0.84          ND

Effluent from Reactor No. 8, 11/12/79            0.15          ND

Effluent from Reactor No. 5, 11/13/79            0.25          ND



(a)  Not detectable.
                                     451

-------
than those measured by GC/MS, are present in the wastewater.  The second
more important factor is the relatively poor solvent extraction efficiencies
for such materials when applying the EPA protocol.  The third factor, which
is probably the most troubling, is the observation of "crossover" of extract-
ables, i.e., acid extractables compounds were found in the base neutral
extraction and vice versa.  The severity of this "crossover" appeared to vary
from sample to sample but was most severe with the high concentration samples.
The problem of extraction efficiency is not limited to the phenols only.
Stamoudix et al(&) demonstrated poor extraction efficiencies for several com-
pounds in single component systems using the EPA extraction protocol; data
reproduced in Table 11.

     Table 11.  RECOVERY EFFICIENCIES FOR DIFFERENT ORGANIC COMPOUNDS
                EXTRACTED FROM SPIKED DISTILLED WATER (DATA OF
                STAMOUDIS ET AL)

                      Compound Name    % Recovery

                      0-xylene              42
                      3-octanone            71
                      1-heptanol            70
                      n-butylbenzene        32
                      phenol                61
                      cresol                75
                      o-ethylphenol        105
                      d^Q-anthracene       133

SUMMARY AND CONCLUSIONS

It is anticipated that the Federal government will issue Best Available Tech-
nology Economically Achievable (BAT) limitations that will severely limit the
discharge of ammonia, sulfides, cyanides, phenol and priority pollutants in
coke plant wastewater discharges.  Preliminary indications are that the tech-
nology to meet the limitations will be staged biological treatment followed
by alkaline chlorination and filtration or activated carbon adsorption followed
by alkaline chlorination and filtration.  A study of a single stage phenol-
nitrification process for the treatment of coke plant wastewaters was under-
taken to evaluate its potential as an alternative to either stage biological
treatment or activiated carbon technology.  Objectives of the single stage
phenol-nitrification process study were:

1.  To determine the operating conditions necessary to achieve an effluent of
10 ppm or less of ammonia and measure the corresponding concentration of
other pollutants;

2.  To determine the effect (inhibitory) of certain constituent compounds and
ions in coke plant waters;

3.  To conduct preliminary examination of methods for enhancing the operation
of the process;
                                      452

-------
4.  To determine  the  effect  of  the  process on priority organic pollutants in
coke plant waters  and

5.  To develop a better  understanding of  the different reactions and inter-
actions, operation of and performance of  the process.

The study was essentially a  laboratory investigation in which actual coke
plant wastewaters  were examined.  In addition, in evaluating the biological
degradation of priority  organic pollutants, a set of samples was taken from
an existing industrial facility whiph represented a current Best Practical
Treatment (BPT) facility.  Water used in  the laboratory work was an actual
coke plant wastewater that was chemically adjusted as needed for the study.

The specific conclusions from the study are:

1.  The single stage  phenol-nitrification process has the potential of pro-
ducing an effluent concentration of NH-j = 10 mg/1, total cyanides = 10-
110 mg/1; phenols  = <200 yg/1; and sulfide =0.5 mg/1 from ammonia stripped
undiluted coke plant  wastewater.

2.  Typical required  operating conditions:

            a.  Detention Time        - 3 days
            b.  Biomass  Concentration - 2-3 grams per liter
            c.  Dissolved Oxygen      - Above 1.5 mg/1
            d.  Temperature           - 80°F
            e.  pH                   - 7.0 to 7.7

3.  The process is effective in the degradation of organic priority pollutants,
Major factors that influence the degree of priority pollutant degradation are
hydraulic and sludge  residence time.  Longer residence time improves removal.

4.  The process is very effective upon free cyanides but ineffective upon
complex cyanides.

5.  The rate of nitrogen oxidation observed in the control reactors was in
agreement with the correlation

                          kA = 15.2 TVS, mg/I/day

where TVS = mixed  liquor total volatile solids,  g/1.   Sludge growth was
determined,  Yraax of 0.7 mg SS/mg N oxidized; biological maintenance energy
utilization rate, b, of 0.004 day1.

6.  Optimum pH condition for the phenol-nitrification process is in the range
of 7.0-7.7 with the temperature range being 80-90°F.

7.  Alkalinity requirements are 4-5 mg as CaC03  per mg of nitrogen oxidized.

8.  Sudden nitrogen loading increases to the reactor can be tolerated for
reasonable periods of time provided reactor conditions,  especially pH and DO,
                                      453

-------
are maintained at near optimum conditions without severe disruption of biolo-
gical activity.  It must be noted that the effluent quality may change, i.e.,
ammonia concentrations may increase.

9.  A 67% increase in thiocyanate loading did not affect nitrification.

10. Thiocyanate concentrations up to 150 mg/1 in the reactor did not disrupt
the nitrification process but a synergistic effect between thiocyanate con-
centration and pH greater than 8.0 did produce inhibition at lower concentra-
tions .

11. Reactor pH's greater than 8.0 noticeably inhibited thiocyanate degrada-
tion.

12. Free cyanide severely inhibited nitrification, but the nitrifying organisms
were capable of acclimating to concentrations up to 12.0 mg/1 free cyanide.

13. Free cyanide concentration up to 12 mg/1 did not affect the thiocyanate
organisms.

14. Complex cyanide concentrations between 10-110 mg/1 were found to pass
through the biological process unaltered.

15. The direct addition of by-product light oil to the reactor in an amount up
to 2000 mg/1 did not appear to have any effect on the biological reactions.
In practice this occurrence might affect the oxygen transfer and mixing
ability of the aeration equipment.

16. Both direct addition of sulfide to the reactor to concentrations up to
40 mg/1 and progressively Increasing the sulfide loading rates to test
reactors produced no adverse effects.

17. The nitrifying organisms were capable of acclimating to phenol concentra-
tions in the reactor up to 30 mg/1.

18. Wastewater with phenol concentrations up to 5000 mg/1 were effectively
treated.  Throughout the study, treated wastewater phenol concentrations were
consistently less than the 0.5 mg/1 BAT limit proposed (Alternate 1).

19. Process enhancement by the addition of activated carbon to the mixed
liquor was not conclusive.  Additions of activated carbon aesthetically im-
proved the appearance of the effluent.  It also appeared to enhance the
settleability of suspended solids in the effluent.

20. In the treatment of coke plant wastewater, the autotrophic reactions are
not carbonate limited.

21. The process was found to require close operator attention, close control
of the environment within the reactor and reasonably constant loading rates.
                                      454

-------
REFERENCES

1.  Wong-Chong, G. M., and S. C. Caruso, "Biological Oxidation of Coke Plant
Wastewaters for the Control of Nitrogen Compounds in a Single Stage Reactor,"
Proc. of the Biological Nitrification/Denitrification of Industrial Wastes
Workshop, Wastewater Technology Center, Canada Center for Inland Water, Bar-
lington, Ontario, Canada, 1977.

2.  Wong-Chong, G. M., "The Kinetics of Microbial Nitrification as Applied to
the Treatment of Animal Wastes," Ph.D. Thesist Cornell University, Ithaca,
N.Y., 1974.

3.  "duPont PACT Process" Bulletin published by E. I. duPont de Nemours and
Company, Wilmington, Delaware.

4.  Robertaccio, F. L., "Powdered Activated Carbon Addition to Biological
Reactors," Proc. 6th Mid-Atlantic Industrial Waste Treatment Conference,
U. of Del., November 15, 1972.

5.  Adams, A. D., "Improving Activated Sludge Treatment with Powdered
Activated Carbon," Proc. 28th Annual Purdue Industrial Waste Conf., Purdue
University, May 1-3, 1973.

6.  Stamoudis, V. C., R. G. Luthy and W. Harrison, "Removal of Organic Con-
stituents in a Coal Gasification Process Wastewater by Activated Sludge
Treatment," Argonne Nat'l Lab, Energy and Environmental Systems Division
Report ANL/WA-79-1.
                                     455

-------
         NITROGEN AND CONTAMINANT CONTROL OF COKE PLANT EFFLUENTS
                     IN AN UPGRADED BIOLOGICAL SYSTEM

              T.R. Bridle, H. Melcer, W.K. Bedford, B.E. Jank

      Wastewater Technology Centre, Environmental Protection Service
                  Environment Canada, Burlington, Ontario
ABSTRACT

          Bench scale treatability studies were conducted to evaluate the
performance of the single sludge pre-denitrification nitrification process
configuration for nitrogen and contaminant control of coke plant effluents.

          Complete nitrogen control was achieved provided wastewater dilu-
tion was practised or low levels of powdered activated carbon (PAC) were
added to the bioreactors.  The minimum aerobic SRT required to achieve
complete nitrification at 20-24°C was 22 days.  Operation at high system
SRT (60 d) did not facilitate nitrification of undiluted wastewater.
However at anoxic and aerobic HRT's of 0.5-1 and 1-3 days, respectively,
complete nitrogen control and high levels of contaminant removal were
effected in undiluted wastewater by maintaining a PAC level of 500 mg.L"1
in the reactors.  The equivalent PAC feed concentrations ranged from 20
to 50 mg.L"1.  The addition of PAC overcame Nitrobacter inhibition,
which was evident in the treatment of both diluted and undiluted wastewater.

          The organic carbon in the wastewater was used as the energy
source for denitrification and no supplemental organic carbon was required
to achieve complete denitrification provided the feed FOC/TKN ratio >3.5.

          The fate of trace organics was monitored using GC/MS methodology.
Enhanced trace organic removal was effected through PAC addition.

          Parallel units operated with and without calcium indicated that
up to 3000 mg-L"1 of dissolved calcium in the wastewater was not detrimental
to biological activity.  Analysis confirmed that calcium phosphate tetra-
baslc was precipitated discretely in the reactors.  Phosphoric acid require-
ments increased 10-fold when calcium was present.
                                  457

-------
       NITROGEN AND  CONTAMINANT CONTROL  OF  COKE PLANT EFFLUENTS
                   IN AN  UPGRADED  BIOLOGICAL  SYSTEM

INTRODUCTION
          The  complete mix activated  sludge process is generally  used  for  the
treatment of coke plant wastewaters in North  America.   The  oxidation of  phen-
olics, cyanide, thiocyanate  and sulphides is  achieved  but few full-scale
facilities achieve nitrification and  no  bioplants  currently practise complete
nitrogen control.  This study has  sought to determine  the process conditions
by which nitrogen control can be achieved economically.  Inherently, this
requires an optimum  balance  between the  nitrification  and denitrlflcatlon
processes.  Evaluation of full-scale  industrial experience^ and cost analysis
data" ied to  the adoption of a single sludge pre-denitrificatlon nitrifi-
cation process configuration for further study.  The cost saving  advantages
offered by this process are  twofold;  firstly,  the  supplemental carbon  required
for denitrification  is minimised or eliminated by  the  presence of raw  waste-
water organic  carbon and  secondly, alkalinity requirements  are greatly
reduced by coupling  the nitrification and denitrification processes.
          Accordingly, bench-scale treatability studies were  initiated at  the
Wastewater Technology Centre, Burlington, Ontario,  to  produce a non-acutely
lethal (to rainbow trout) effluent low In nitrogen  concentration.  Reactors
were operated from October 1978  to April 1980.  This period may be conven-
iently divided into  two phases,  the first being concerned with startup
and acclimation and  the second,  with  the determination of process  conditions
required to achieve  high  levels  of nitrogen removal  from a  full .strength coke
plant wastewater.
          Wastewater was  provided by Dominion Foundry  and Steel Limited
(Dofasco), Hamilton, Ontario.  This comprised a mixture of  limed  ammonia still
effluent and a light oil  interceptor  sump wastewater.
          A U.S. EPA survey  of  steel industry effluents10 reported the presence
of 73 of the 129 priority pollutants.  The  level to which this technology was
successful in removing trace contaminants was evaluated by  GC/MS  analysis.
This analysis was extended beyond the EPA priority pollutant  list  to include
those trace contaminants  that are indigenous  to coke plant wastewaters.

                                   458

-------
EXPERIMENTAL  PROCEDURES
          Three  identical process trains, A,  B and C, were operated in para-
llel.  Figure 1  depicts the process sequence  for each train; a complete mix
anoxic reactor (Dl),  a complete mix aerobic reactor (D2), and an upflow clari-
fier.
          Operating procedures for reactor control have been detailed
previously^.   Briefly, feed rates of 5  to 15  L«d~^ were selected, allowing
ranges of nominal anoxic hydraulic retention  time (HRT) from 0.5 to 1.0 day
and nominal aerobic HRT's from 1.1 to 3.0 days to be effected.  Mixed liquor
was recycled  from the clarifier to the  anoxic reactor at a ratio of 8:1.
Strict SRT control was maintained.  Aerobic reactor pH and DO were controlled
at 7.0 and 3.0 rag-lT1 respectively.  Temperature in the reactors varied
from 20  to 24°C.  An effluent phosphorus residual was maintained by control-
led phosphorus addition to the anoxic reactor.
                                 PHOSPHORIC  CONTROLLED
                                 ACID      Na,C03
                               -Q* —    .-.ADDITION
ANOXIC
REACTOR
(P1)
                                      H
                                       TO INDICATING
                                       CONTROLLERS
                                             X. CONTROLLED
                                     REACTOR
                                     
-------
RESULTS AND DISCUSSION

Phase I
          The seven-month initial phase incorporated startup and sludge accli-
mation, the objective being to establish fully equilibrated nitrifying sys-
tems.  SRT was initially maintained at 60 to 70 days.  Removal of 80% of the
filtered organic carbon (FOG) and 99% phenol were quickly achieved but the
systems were unstable with regard to nitrification and to oxidation of cya-
nide and thiocyanate.  This behaviour was attributed to the wide variations
in feed characteristics.
          Pseudo-equilibrium conditions were subsequently achieved by control-
ling feed characteristics; raw feed was diluted 2:1 to 4:1 with tapwater and
respiked with phenol, thiocyanate and methanol to achieve initial feed para-
meter concentrations.  Values of equalized feed parameters are listed in
Table 1.  SRT values were then allowed to decline to a mean of approximately
30 days.  Typical system operating parameters are summarized in Table 2 and
typical effluent quality in Table 3.
          Oxidation of carbonaceous material improved to 94% FOG removal and
almost total phenol removal.  The organic carbon requirements for the
pre-denitrification systems were approximately twice the theoretical require-
ments confirming Sutton^al's observations11.  Maintenance of the FOC/TKN
level above a minimum of 3.5 ensured complete denitrlfication.  This para-
meter is intimately associated with ammonia still operation.  Excessively
high ammonia levels in the still effluent will exceed the organic carbon
availability in the wastewater and require carbon supplementation to maintain
denitrification.  This study showed that nitrification was sensitive to
ammonia variability.  In the early part of Phase I, this variation was
four-fold before subsequent equalization reduced it to 1.6.  Data from
a full-scale plant treating a high-strength organic chemical wastewater^ sup-
ports these findings; nitrification was achieved consistently provided the
TKNjaajj/TKNj^au ratio <2.0.  Equalization is not normally feasible in a steel-
works environment since there is a constraint on land availability.  However,
efficient still operation can produce an effluent with approximately
100 mg'L"1 ammonia and thereby derive a three-fold cost saving since
a greater amount of ammonia is recovered and neither carbon supplementation
nor  equalization are required.

                                   460

-------
Table 1,   FEED CHARACTERISTICS - STEADY-STATE OPERATION, PHASE I
Parameter
FOC
Phenol
TKN
NH3-N
TCN
CNS
Median
(mg-IT1)
470
215
130
75
1.0
170
95%*
(mg.IT1)
570
245
185
120
1.6
180
Variability**
1.21
1.14
1.42
1.60
1.60
1.06
*  95% of values were equal to or less than this value.
** 95% value divided by median.
Table 2,   TYPICAL OPERATING CONDITIONS - PHASE I
Parameter
HRT (d)
Temp (°C)
DO (mg-IT1)
pH
TSS (mg-L"1)
VSS (mg.L"1)
OUR (d"1)
SVI (mL.g"1)
Anoxic Reactor (Dl)
1
20- 24
0.3
7.5-8
2000 - 2800
1200 - 1500
• -
—
Aerobic Reactor (D2)
3
22 - 24
3.0
7.0
2000 - 2800
1200 - 1500
0.15 - 0.3
50 - 100
Clarifier
Effluent


-
-
30
20
-
—
Table 3,   TYPICAL AEROBIC REACTOR EFFLUENT QUALITY - PHASE I
           (SRT >30d at 20 to 24°C)

                     FOC  Phenol  TKN  NH3-N  N02~N  N03-N  TCN  CNS  TN*

 Effluent (mg.L"1)    30   0.050    5    <1     15     0     <1   <1   20

 % Removal            94   >99.9   96   >99      -     -    ~50  >99   85

* Total Nitrogen
                                  461

-------
          The minimum SRT required to maintain nitrification (as measured by
TKN disappearance) was identified as 30 days which equates to a minimum aero-
bic SRT of 22 days.  All the oxidized nitrogen was completely denitrified.
The absence of effluent NO-j-N indicated the inhibition of Nltrobacter by
some specific trace contaminants in the coke plant wastewater.  This phen-
omenon has been observed by other investigators2^.  Nevertheless, the 4:1
dilution of raw wastewater permitted nitrification and denitrification to pro-
ceed but was insufficient to prevent Nitrobacter inhibition.
          Theoretical relationships indicate that 7.07 units of alkalinity as
CaC03 are consumed per unit of NH3~N nitrified and 3.57 units of alkalinity
are generated per unit of N02~N or N03-N denitrified12.  The stoichlometric
net alkalinity requirement for the combined process can be estimated from the
degree of nitrification and denitrification achieved.  Approximately 50 to 60%
of the alkalinity required in this system configuration was supplied via
Na2C03 addition (for pH control), the remaining requirement being satisfied
by the residual alkalinity in the feed wastewater.  On the basis of feed and
effluent data in Tables 1 and 3, the alkalinity requirement was 2.38 g
CaC03/gTN removed.
          Fish bloassay tests were carried out to assess the toxicity of the
effluent.  The tests involved 96-h static bioassays using juvenile rainbow
trout (Salmo gairdneri).  Results indicated that the effluent was non-lethal,
with zero mortality.

Phase II
          The mode of equalization used in Phase I restored mean concentrations
of the major pollutants, FOG, phenol and thiocyanate, by respiking, but dilu-
ted the level of trace contaminants.  Ammonia variability was reduced which
allowed nitrification to proceed as far as nitrite formation.  Oxidation to
nitrate did not occur since Nitrobacter growth was inhibited probably by the
presence of the trace contaminants, albeit at low concentrations.  Thus
nitrogen control was achieved but the discharge of nitrite is environmentally
unacceptable.  In practice, to exercise nitrogen control by dilution and then
discharge the nitrite would be an inappropriate procedure.  Thus the main
thrust of the treatability work was directed to defining the process condi-
tions required to achieve high levels of nitrogen removal from a full
strength coke plant wastewater.

                                  462

-------
Part 1 - Effect of Full-Strength Wastewater
          Raw wastewater from Dofasco was used as received except that metha-
nol was added as required to maintain the FOC/TKN ratio >3.5.  Table 4 sum-
marizes feed characteristics.

Table 4,   FEED CHARACTERISTICS - PHASE II (PART 1)
Parameter
FOC
Phenol
TKN
NH3-N
TCN
CNS
Median
(mg.L-1)
535
185
155
80
4.4
210
95%*
(rag. IT1)
640
269
214
88
4.8
237
Variability**
1.20
1.45
1.38
1.10
1.09
1.13
 * 95% of values were equal to or less than this value.
** 95% value divided by median

          At the start of Phase II, all three systems were operating at equil-
ibrium SRT values between 30 and 35 days.  Nitrification deteriorated signifi-
cantly  in all systems within two weeks of restoring the full strength feed;
effluent ammonia concentration varied from 30 to 100 mg-L"*.  This was not a
satisfactory mode of operation.

Part 2 - Effect of PAC Addition
          In order to re-establish nitrification, powdered activated carbon
(PAC) was added to systems B and C at rates equivalent to 33 and 50 mg.lT1 in
the feed.  System A did not receive PAC, remaining as the "control".  The
resulting equilibrium reactor conditions are summarized in Table 5.
          The feed characteristics for the remainder of the study are pre-
sented in Table 7.  With the exception of organic carbon including phenol,
the feed characteristics are very similar to those recorded earlier in the
study.  The feed organic carbon content increased steadily over the period
that these data were averaged.  Also*improved operation of the ammonia still
reduced ammonia variability considerably below that experienced in the early
part of Phase I.
          Two weeks after PAC addition, System A remained unchanged whereas
nitrification was reestablished in B and C.  Nitrobacter inhibition had

                                   463

-------
    Table 5,   EQUILIBRIUM OPERATING CONDITIONS - PAC AND SRT EVALUATION
2
Average Reactor Solids
(ing L~l)

Reactor
System
Phase II
Part 3 •<
Phase II f
Part 2 <
A
A
B
C
HRT (d)
Anoxlc Aerobic
1 3
1 3
1 3
1 3
SRT Biological
(d) MLVSS
40 1780
60 2330
30 1920
40 2090
MLSS PAC
2580 0
3880 0
3000 250
2750 500
Clarifier
Equivalent Effluent
PAC Feed VSS
* «
(mg'L"1) (mg'L'1)
0 6
0 7
33 9
50 8



Table 6, MEAN REACTOR EFFLUENT QUALITY - PAC AND SRT EVALUATION

Reactor
System
Phase H
Part 3
Phase II
Part 2 '
»
*
**
***
* A
_A
B
C
Anoxlc
FOC
Effluent***
Z Removal -
Effluent 76
Z Removal -
Effluent 66
Z Removal -
Effluent 59
Z Removal -
Reactor
NOJ.-N* FOC
- 31
- 94
0.2 35
- 94
1.4 31
- 95
2.8 28
- 96

Phenol
0.034
>99.9
0.041
99.98
0.040
99.98
0.031
99.99

TKN
72
54
97
42
6
96.5
7
95.7

NH3-N
68
91
0.6
99.2
0.2
99.8
Aerobic Reactor
N02-N N03-N TCN
4.0 0.0 3.4
23
4.1 0.3 4.4
47
4.7 6.9 3.9
52
1.5 10.7 3.9
51

CNS
1.1
99.5
1.2
99.4
1.3
99.4
0.9
99.6

ON
5.5
78
2.8
89
4.8
81

TN**
76
51
101.4
41
17.6
90
19.2
89
Total oxidized nitrogen.
Total nitrogen, TO 4- NOT-N.
Expressed in mg.L~l.

-------
ceased since oxidation to nitrate had occurred.  This trend continued to be
observed over a 2-month period.  Typical reactor effluent quality data are
reported in Table 6.

Table 7,   FEED CHARACTERISTICS - PHASE II (PARTS 2-5)
Parameter
FOC
Phenol
TKN
NH3-N
TCN
CNS
ON***
PH
Median
680
300
180
88
8
240
25
9.3
95%*
810
460
235
120
24
355
62
10.5
Variability**
1.19
1.53
1.30
1.36
3.00
1.48
2.48
1.13
  * 95% of the values were equal to or less than this value.
 ** 95% value divided by median.
*** Organic nitrogen, calculated as TKN-(NH3-N)-(CNS-N)-(TCN-N).

          Two sets of data are shown for system A in both Tables 5 and 6:
the effect of high SRT was being examined concurrently in system A and so
pertinent data is included in these tables although the effect  is addressed
in part 3.  The data demonstrating the effect of PAC compares the three
systems A, B, and C, during the second month of operation in this mode, under
well-equilibrated conditions.
          Effluent distribution data for TKN in Figure 2 shows  that  system A,
the conventional pre-denitrification nitrification, could  not  nitrify full
strength Dofasco wastewater confirming the observations in Part 1.   Low
levels of PAC addition in system B, enabled nitrification to proceed almost
to completion.  A higher nitrite concentration than in system C indicated
that some minor inhibition of Nitrobacter had occurred.  The higher  level of
PAC addition used In system C eliminated this Inhibition and complete nitrifi-
cation to nitrate was evident.  Inhibition may have been due to the  presence
of one or more trace contaminants that were possibly adsorbed on to  the PAC
in systems B and C.  Thiocyanate oxidation was achieved to the  same  degree In
A, B and C (as was the oxidation of phenol and FOC) indicating  that, in the
case of coke plant wastewater, the nitrification process appears to  be the
most sensitive to inhibitory trace contaminants.
                                  465

-------
                     5   1015203040506070808590  95  98
                      PERCENT OF OBSERVATIONS & STATED VALUE
            Figure 2,   Effluent TO  data;  PAC  evaluation.

Part 3 - Effect of High SRT
          Control data from system A  during the FAC evaluation trials showed
that nitrification of full strength wastewater  could not be achieved at a sys-
tem SRT of 40 days (Table 6).  The system SRT was  Increased to an equilibrium
level of 60 days to determine whether nitrification could be sustained at a
higher SRT without PAC addition.  A comparison  of  the effluent quality data
for system A for both SRT values shows little change Indicating that the
single sludge pre-denitrification nitrification system was not capable of
nitrifying full strength coke plant wastewater  at  elevated SRT values.
Removal of FOC, phenol and thiocyanate remained unimpaired.
Part 4 - Effect of Calcium Precipitation
          Most steel mills use  combined free/fixed leg ammonia stills for the
removal of ammonia.  The most common  alkali  used  for  pH elevation in the fixed
leg of the still is calcium hydroxide.  As a result,  the calcium content of
limed weak ammonia liquor ranges  from 1500 to 3000 mg-L"1.   Observations
made in Phase I* revealed precipitation of calcium in the reactors,  with the
result that the mixed liquor volatile fraction was reduced  to less than 50%.
                                    466

-------
Other work1*** has attributed process  Instability  to the formation of an
inactive sludge caused by the precipitation  of  calcium carbonate.  Conse-
quently, parallel studies were conducted  in  Phase II to define the effect of
precipitated calcium salts on the  nitrification performance of the single
sludge process configuration.  To  this  end,  two reactor systems were operated
in parallel.  System  A  was fed wastewater  as  received from Dofasco, con-
taining between 1500 and -2500 rag-IT1 of  calcium.  The feed to sys-
tem  C  was carbonated at elevated pH to  reduce the calcium content to levels
always lower than 100 mg-lT1.  Both systems  received PAC at a rate sufficient
to maintain a reactor concentration of  500 mg-L"1.  The reactor systems were
operated at steady  state conditions (Table 8)  for a four week period.  The
mean reactor effluent quality data for  this  period are shown in Table 9.
They show  no significant differences in nitrification performance and process
stability  (Figure 3).  There was,  however, an increase in effluent soluble
organics  in system   C  .  This could be  attributed to the lower adsorptive
capacity  of the lower PAC  loading in the influent to system  C .
                   20-
                 O)
                 U.
                 111
                    'g
                    ?
                    6
                    5
                        SYSTEM  FEED Ca~ 
-------
    Table 8,   STEADY STATE OPERATING  CONDITIONS -  CALCIUM EFFECT
00
Reactor
System
A
C
Table 9
Reactor
System
A
C
Feed
Calcium HR1
(mg-L"1-) Anoxic
1500 - 2500 1
<100 1
Average Reactor Solids Clarifler
(mg'L""1) Equivalent Effluent
C (d) SRT Biological
Aerobic (d) MLVSS
3 40 2680
2 40 2900
MLSS
7230
4100
PAC Feed VSS
PAC (mg-L"1) (mg-L"1)
500 50 12
500 33 10

, MEAN REACTOR EFFLUENT QUALITY - CALCIUM EFFECT
Anoxic
FOC
Effluent* 83
% Removal -
Effluent 81
% Removal -
Reactor
NOT-N FOC Phenol TKN
4.7 38 0.076 10
- 94.6 99.98 94.3
5.7 47 0.086 9
- 93.4 99.98 95.1
Aerobic Reactor
NH3-N
0.9
99.0
0.7
99.1
N02-N N03-N TCN CNS ON
2.5 21.2 4.5 1.2 4.4
49 99.5 82
2.4 13.3 6.3 2.1 4.2
27 99.1 83
TN
33.7
81
24.7
86
    *  Expressed in mg

-------
          Mixed  liquor  solids  data In Table  8  Indicate  that  the volatile  frac-
 tion  Increased from 37  to  70%  when the  feed  was  pretreated for  calcium
 removal.  Phosphorus requirements  for system G were  reduced  markedly;  rough
 balances  indicated  a tenfold reduction  In  P  requirement.  This  observation
 tended  to indicate  the  Inorganic material  precipitating in system A,  was  a
 calcium phosphate.   To  confirm this, and to  define the  structure of  the pre-
 cipitate, sludge samples were  analyzed  by  x-ray  diffraction  (XRD)  and scan-
 ning  electron microscopy/energy dispersive x-ray spectroscopy (SEM/EDS)
 analysis.  XRD analysis indicated  the Inorganic  phase in  sludge A to  be cal-
 cium  phosphate tetrabasic  (4 CaO^Os)  However,  EDS.  analysis  would  tend to
 indicate  that the material is  octa calcium phosphate.
          In general, observation  of sludge  A  showed that the calcium phos-
 phate did not coat  the  floe.   Rather, it was discretely precipitated  through-
 out the floe suspension and could  not be responsible for any reduction in per-
 formance.  This  data complements the process data which showed  no  difference
 in performance between  systems A and C.

 Part  5  -  Effect  of  Hydraulic Retention  Time
          In Phase  I, studies were  conducted at  fixed anoxic and aerobic HRT's
 of one  and three days,  respectively.  These  values were chosen  based on both
 theoretical and  practical  considerations,  and Included a safety  factor of two.
 Since the HRT Impacts severely  on  system capital  costs, studies were conducted
 to define the limiting  hydraulic requirements capable of effecting consistent
 nitrification.   This was accomplished by operating the  three systems in para-
 llel  at varying  HRT's;  0.5, 0.67, and 1.0  day for the anoxic reactors and
 1.1,  2  and 3 days for the  aerobic  reactors.  The  experimental conditions eval-
 uated,  and steady state reactor conditions achieved,  are shown  in Table 10.
Mean  effluent quality achieved at  these operating conditions is  summarized in
Table 11.
          A general  deterioration  in effluent quality with decreasing aerobic
HRT was observed with increasing effluent SS, FOG and phenol and an increas-
ingly unstable tiltrificatloti process as shown in Figure 4.  These effects may
 have  been compounded by the decreasing adsorptive capacity of the decreasing
PAC loading at the lower HRT's.  An increase in the level of effluent organic
nitrogen with deereasing.HRT may have been due to the presence of heterocyclic
nitrogenous compounds that were not adsorbed by the PAC and which may, there-
fore,  have contributed  to  the inhibition of nitrification.

                                   469

-------
 Table  10,   STEADY  STATE  OPERATING CONDITIONS - HRT EFFECT
Average Reactor Solids Clarlfier
(mg.L*1) Equivalent Effluent
Reactor
System
A
B
C
HRT (d)
Anoxic Aerobic
0.67
1
0.5
Table 11, MEAN
Reactor
System
A
B
C

Effluent*
% Removal
Effluent
% Removal
Effluent
% Removal
2
3
1.1
REACTOR
Anoxic
FOC
119
114
106
SRT
Anoxic
10
10
15
EFFLUENT
Reactor
NO-jr-H
0.5
0.7
0.8
(d)
Aerobic
30
30
30
QUALITY
Biological
MLVSS
5 500
3 630
6 920
- HRT EFFECT
MLSS
16 200
9 550
13 840

PAC Feed VSS
PAC (mg-L"1) (mg-L"1)
500
500
500

33 18
50 10
18 49


Aerobic Reactor
FOC
41
94.5
38
95.0
60
91.9
Phenol TKN
0.084 7
99.8 96.0
0.072 7
99.8 96.0
0.098 27
99.7 86.5
NH3-N
1.7
98.0
1.1
98.7
19.8
78.7
N02-N
3.0
9.7
8.3
N03-N TCN CNS ON
18.7 8 1.1 2.2
21 99.5 91
8.3 8.1 1.1 1.2
21 99.5 95
3.4 8.3 1.4 4.1
18 99.4 84
TN
28.7
84.4
25.0
86.0
38.7
80.0
*  Expressed in mg-L""l.

-------
           Denitrification did not appear to have been affected by HRT over
the  range examined  (Figure 5) indicating that  the full denitrification capacity
of  the system had not been exploited.
                           90  96
   5   «  20 30 «0  50 «0 70  80
    PERCENT OF OBSERVOTIONS * STATED WLUE
                                                       TEM ANOXIC HRT (d)
                                                     • A      0.67
                                                     o B      1.0
                                                     A C      0.5
                                              0.1
                                                2   5  10  20 30 40 5060 70 80  90 95  98

                                                 PERCENT OF OBSERVATIONS S STATED VALUE
Figure  4,   Effluent NH  -N Data;

            HRT effect.
Figure  5,    Denitrification performance;

             HRT effect.
                                      471

-------
TRACE ORGANICS REMOVAL
          Coke production generates a great variety of poly-nuclear aromatlcs
(PNA's) many of which are not identified In the list of priority pollutants.
The U.S. EPA have shown7*13 that adsorption was the main removal mechanism
for PNA' s in sewage treatment plants.  The analysis of trace organics in this
study was conducted to Identify and quantify these compounds.
          Aliquots of feed, final effluent and waste activated sludge from
systems A and C were analyzed by GC/MS and HPLC for the presence of trace
organics.  These samples were taken during the period when the effect of cal-
cium on nitrification performance was being evaluated, and thus the reactor
conditions and effluent quality depicted in Tables 8 and 9 are representative.
Both systems were operated at 40-day SRT, with 500 mg«L~l PAC.  However, the
aerobic HRT's were different - 3 days in system A and 2 days in system C.
Some selected data is shown in Table 12.  It must be mentioned that only one
replicate of each sample was analyzed, and while the limited quality assur-
ance program indicated good recoveries of spiked compounds (58 to 108%), the
absolute values reported must be treated with caution.
          The data Indicated the presence of 16 base-neutral and acid extrac-
table priority pollutants.  In addition, 32 non-priority pollutants, primar-
ily heterocyclic nitrogenous compounds, were Identified.  These compounds
could not be quantified due to lack of standards.  The units expressing trace
organic concentrations have been selected to facilitate data comparisons on a
similar numerical basis.  Concentrations are expressed at a ppb level in both
the liquid streams and the sludge so that the fate of the trace organics may
be traced more easily through the treatment process.  Operation at a 3-day
aerobic HRT generally produced a 'cleaner* effluent, with marked improvement
in removal of PNA's.  Not all PNA's were accumulated in the sludge.  Of the
PNA's in the feed only indeno-pyrene, naphthalene, pyrene and benzo-a-anthra-
cene were adsorbed on the sludge.  Although significant quantities of
phthalates and naphthalene accumulated in the sludge, mass balances indicate
that more than 90% of those compounds were biologically degraded.  In con-
trast, the indeno-pyrene and pyrene remained adsorbed on the sludge.
          A substantial concentration of organics can occur which illustrates
the role of the sludge as a "sink" for some organics.  Benzo-a-anthracene,
for example, was not detected In the feed but was present at 360 ng/g In the
sludge.  Mass balance calculations show that, assuming no losses to other
removal mechanisms, a concentration factor of approximately 4000 was In effect.
                                  472

-------
    Table 12,   SELECTED TRACE ORGANICS DATA
U»
System 'A*
I
Anthracene
Benzo-a-pyrene
Chrysene
Diethylphthalate
Bis (2 ethyl hexyl) phthalate
Fltioranthene
Fluor ene
Indeno pyrene
Naphthalene
Phenanthrene
Pyrene
Benzo-a-anthracene
Cl Pyrldine
Quinoline
1H tndole
9H Carbazole
9H Anthracene Carbonitrile
P henanthri dine
Phenanthridinone
Indolizine
Feed
Cug'L"1)
1.0
0.5
4.0
300
Trace
2.6
4.0
0.7
760
3.5
2.2
ND
-H-
4+
++
•H-
+
+
+
•f
Effluent
0.3
0.4
1.7
100
Trace
1.0
1.5
0.2
ND
0.75
0.7
ND
ND
ND
ND
ND
ND
ND
ND
ND
Sludge
(ng'g"1) '
ND
ND
ND
9 100
11 000
ND
ND
9 500
7 900
ND
740
360
+
+
+
++
•H-
ND
+
ND
System 'C* Detection
Feed
(\ig-L~1)
1.0
ND
18
1 310
150
4.1
3.0
1.2
3 800
Trace
ND
ND
-H-
•H-
-H-
•H-
+
+
+
+
Effluent
1.4
1.5
4.0
110
Trace
2.0
3.0
ND
ND
0.5
2.0
ND
ND
ND
ND
ND
ND
ND
ND
ND
Sludge
(ng.g"1)
ND
ND
ND
6 600
20 000
ND
ND
1 800
2 600
ND
ND
160
•f
+
•f
•H-
++
-H-
-H-
-H-
Limit
0.25
0.1
0.4
10.0
10.0
0.1
1.0
0.2
10.0
0.4
0.6
0.6








    Note:  Sludge values reported  on  a dry  sludge  basis.
    ND * not detected.
    +  = a minor peak.
    -H- =» a major peak.

-------
The feed concentration of benzo-a-anthracene would then have been
0.09 yg'lT1 which is below the detection limit of 0.6 pg-lT1.  It could be
concluded that other organics are undetected in the feed but will be adsorbed
to significant levels in the sludge.
          None of the heterocyclic nitrogenous compounds identified in the
feed could be detected in effluents  A  or  C .  However, compounds such as
9H carbazole, 9H anthracene carbonitrile, indolizine, phenanthridine and
phenanthridinone were accumulated in the sludge.
          The degree to which PAC addition aided in adsorption of PNA's and
heterocyclic nitrogenous compounds could not be determined.  The literature1*.
does, however, indicate that many PNA's are readily adsorbed by PAC.  No
conclusive evidence has been generated to identify the mechanism by which PAC
addition prevents inhibition of nitrification but it is possible that this
occurs by the adsorption of the heterocyclic nitrogenous compounds.
                                    474

-------
CONCLUSIONS
          1.   Complete nitrogen control of Dofasco coke plant wastewater can
               be achieved In a single sludge pre-denltrlfIcatlon nitrifi-
               cation system only by the addition of low levels of PAC (at
               approximately 50 rag-IT1).
          2.   Over the temperature range, 20-24°C, nitrification was stable
               at aerobic HRT's in excess of two days; denitrification is
               stable at anoxic HRT's of 0.5 day or greater.
          3.   The organic carbon in the wastewater can be utilized as the
               energy source during denitrification.  At FOC/TKN ratios >3.5,
               carbon supplementation is not required.
          4.   The presence of high levels of calcium in the wastewater does
               not affect nitrification.
          5.   The efficient operation of the ammonia still is critical to
               the maintenance of a stable nitrification process.  It is
               essential to minimize variation in wastewater ammonia levels:
               TKNjnax/TKN^an should be should be I2-0.
          6.   GC/MS analysis identified 32 heterocyclic nitrogenous compounds
               and 16 priority pollutant organics in the feed.  The
               pre-denitrification nitrification system, with PAC addition,
               operated at a high SRT and HRT, is capable of effecting good
               removal for most of these organics.  Both adsorption and
               biodegradation are major removal mechanisms.

ACKNOWLEDGEMENTS
The authors wish to express their appreciation to Dofasco for their cooper-
ation in providing the wastewater used in this study.  In addition, appreci-
ation is expressed to Dr. Derek Houghton of McMaster University for his
invaluable assistance in generating the SEM/EDS results.
                                  475

-------
REFERENCES
 1. Ashmore, A.G., «£ a!l,  "The Biological  Treatment of Carbonization
    Effluents, I", Water Research,  1_,  pp.  605-624, 1967.

 2. Beccari, M.,  et^ aL,  "Results and Perspectives of  Coke Oven Wastewater
    Treatment Documentary",  Cebedeau.  412,  pp.  145-150, 1978.

 3. Bridle, T.R., et^ al^ "Operation of a Full Scale Nitrification
    Denitrification Industrial Waste Treatment  Plant", Water Pollution
    Control Federation Journal, 51, 1, pp.  127-139, 1979.

 4. Bridle, T.R., e£ al^ "Biological Nitrogen Control of Coke Plant Waste-
    waters", Proceedings of 10th IAWPR Conference, Toronto, Ontario, June, 1980.

 5. Bridle, T.R., et^ al, "Biological Treatment  of  Coke Plant Wastewaters
    for Control of Nitrogen and Trace  Organics", Presented  at 53rd Annual
    Water Pollution Control Federation Conference, Las Vegas, 1980.

 6  Catchpole, J.R. and R.L. Cooper, "The Biological  Treatment  of  Carbon-
    ization Effluents, III. Water Research. £,  pp. 1459-1474, 1972.

 7. Convery, J.J., e£ al^,  "Occurrence  and Removal  of  Toxics in  Municipal
    Wastewater Treatment Facilities",  Presented at 7th Joint U.S./Japan
    Conference, Tokyo, Japan, May,  1980.

 8. Cooper, R.L.  and J.R. Catchpole, "The Biological Treatment  of Carbon-
    ization Effluents,  IV", Water  Research, 7_,  pp. 1137-1153,  1973.

 9. Environmental Protection Service,  "Meat and Poultry Products Plant
    Liquid Effluent Guidelines", Report EPS l-WP-77-2, Ottawa,  July,
    1977.

 10. Robertson, J.H.,  et a^,  "Water Pollution Control", Chem. Eng, 87_ (13),
    pp.  102-119,  June 30, 1980.
                                    476

-------
11. Button, P.M., e£ al^ "Single Sludge Nitrogen Removal Systems",
    Canada-Ontario Agreement on Great Lakes Water Quality Research
    Report No. 88, 1979.

12. U.S. EPA, "Process Design Manual for Nitrogen Control", EPA 625/l-7l-002a,
    October, 1973.

13. U.S. EPA, "rate of Priority Pollutants in Publicly Owned Treatment
    Works", EPA-440/1-79-300, October,  1979.

14. U.S. EPA, "Carbon Adsorption Isotherms for Toxic Organics",
    EPA 600/8-800-023, April, 1980.

15. Wilson, R.W. e£ al^, "Design and  Cost Comparison of Biological Nitrogen
    Removal Systems",  Presented at 51st Annual Conference Water Pollution
    Control Federation, Anaheim,  California,  1978.
                                  477

-------
HYDROTECHNIC CORPORATION
            AN  INVESTIGATION  OF  FOREIGN  BY-PRODUCT COKE PLANT AND

                  BLAST  FURNACE WASTEWATER CONTROL TECHNOLOGY




                     HAROLD HOFSTEIN  AND  HAROLD J.KOHLMANN

                           HYDROTECHNIC   CORPORATION


                                   ABSTRACT
             A study  was  made  to determine if  more advanced processes
       for  the treatment  of  by-product  coke plant and blast furnace gas
       cleaning wastewaters  were used in foreign plants than in domestic
       ones.   Some  unusual techniques for the  treatment of blast furnace
       gas  cleaning wastewaters  were  found. Aeration of gas cleaning
       wastewater prior to clarification improved settling and resulted
       in a greater rate  of  recirculation.   Filtering the wastewater
       through slag or flue  dust removed cyanide although the removal
       mechanisms is  not  known.

             Treatment of by-product  coke plant and blast furnace gas
       cleaning wastewater is, generally, not  more advanced in foreign
       plants than  in the United States.   However, blast furnace gas
       cleaning water in  foreign plants is generally recycled to a greater
       degree.

             Discussions  were  held with plant  and corporate personnel at
       26 plants in 14 countries and  with regulatory agencies in 10 of
       the  14 countries,  to  determine  the regulations imposed upon the
       plants,  the  incentives  provided  to reduce pollution loads to re-
       ceiving waters and to investigate treatment technology.

             Recommendations for research projects are made as there ap-
       pears  to be  promising areas for  improvement of wastewater treatment
       techniques.
                                    479

-------
INTRODUCTION





      In its continuing effort to make information available on



the most advanced and efficient methods of reducing water pollu-



tion from iron and steel production, the U.S. EPA Industrial En-



vironmental Research Laboratory, Research Triangle Park, NC, con-



tracted with Hydrotechnic Corporation to perform an engeneering



study of foreign steel plants.  This was to determine if there were



water pollution control practices being employed for by-product



coke plant and blast furnace wastewaters that were superior to



those used in the United States.  In fulfillment of this contract



Hydrotechnic visited 25 plants in 14 countries.  Plants in the



United States, Canada and Eastern Bloc nations were not included.



The plants visited account for over 23 percent of the steel pro-



duced outside of the three areas mentioned.  One of the plants



visited was a by-product coke plant only and one plant consisted



of a single blast furnace.  Neither of these plants had other



production facilities normally associated with steel plants.





      Three factors were considered in determining the selection



of the plants to be evaluated:



          Based on published literature and personal correspond-



          ence, the likelihood of the plants utilizing exemplary



          or innovative treatment technology.





          Based on prior investigation, the relative abundance



          or lack of water in the plant area.
                              480

-------
          Based on published information, the degree of environ-
          mental concern in the countries where the plants are
          located.

      Of the 25 plants visited, 23 provided information that was
useable to permit evaluation of their wastewater treatment systems.
These 23 plants are listed below by country.
      ARGENTINA
      AUSTRALIA

      BELGIUM
      ENGLAND

      FRANCE
      ITALY
      JAPAN
      MEXICO
      NETHERLANDS
      SOUTH AFRICA
      SWEDEN

      TAIWAN
      WEST GERMANY
Plant requested anonymity
Broken Hill Proprietary - Newcastle Works
Australia Iron & Steel - Hoskins Kembla Works
SIDMAR
British Steel - Scunthorpe Works
              - Orgreave Works
Pont-a-Mousson
Italsider - Taranto Works
Nippon Kokan KK - Ogishima Works
Sumitomo Metal Ind. - Kashima Works
Kobe Steel Ltd. - Kakogawa Works
Kawasaki Steel - Chiba Works
Altos Hornos de Mexico
Hoogovens
ISCOR - Pretoria Works
      - Newcastle Works
      - Vanderbijlpark Works
Svenskt Stal - Norrbottens Jarnverk
Surhammars Bruks - Spannarhyttan
China Steel
Roechling Burbach
Thyssen
Hoesch Huttenwerke

-------
      In addition to visiting plants and corporate engineering
staffs, and observing wastewater treatment operations at the pro-
duction facilities, nine government agencies were consulted.  In
a tenth country a trade association was consulted.  The agencies
provided information on how regulations affected the degrees of
treatment and on the incentives provided for increasing recircula-
tion of water within the production facilities.

      The nine governments were:
          Argentina
          Australia (New South Wales)
          Japan (two agencies)
          Mexico
          Netherlands (two agencies)
          South Africa
          Sweden
          Taiwan

      At the meeting with the trade association,VDEh representing
the West German Iron and Steel Industry, a representative from
the local West German water and waste agency was present.

SUMMARY
1.    By-Product Coke Plants
      The volume of waste ammonia liquor produced at foreign by-
product coke plants ranged from 0.14 to 0.24 m /Mg (34 to 178 gpt)
                              482

-------
These volumes  are  higher  than those encountered  in  the United
States by-product  coke plants evaluated.

      The treatment of by-product coke plant wastes at foreign
plants is basically similar to that practiced in the United
States.  Single  stage biological treatment is used  at 14 of the
by-product coke  plants visited.  Nine of these fourteen plants
add dilution water to reduce high ammonia concentrations in the
wastewater to  levels not  toxic to the organisms.  At one plant
in Japan salt  water is used.  All plants utilizing  biological
treatment add  nutrients,  usually in the form of  phosphoric acid.
Two of the plants  pretreat the wastes by filtering  the wastewater
through a coarse coke bed.  This procedure removes  tar that may
be detrimental to  the biological oxidation process.  Two other
plants further treat their effluent by sand filtration and, follow-
ing, by activated  carbon  adsorption.

      Of the 23  by-product coke plants for which some data was
available, fourteen plants discharged their biologically treated
wastewater to  public waters, five plants treated their wastewater
in free ammonia  stills and then discharged them, one plant treated
its wastewater in  both free and fixed ammonia stills prior to
discharge, one plant utilized a free ammonia still  and a dephe-
nolizer prior  to discharge, one plant provides no treatment at all
prior to discharge and one plant uses the raw waste ammonia liquor
to irrigate a  grass crop  that is used for animal feed, reportedly
with no ill effects to the animals.

-------
2.    Blast Furnaces
      Blast furnace gas cleaning systems were used at all of the
plants visited.  The gas washer wastewater application rate varied
depending on the wet type of gas cleaning system used.  The rates
varied from 2.1 to 28 m /Mg  (507 to 6715 gpt) of iron produced.
The weighted average application rate was 6.09 m /Mg  (1460 gpt).

      All but one of the plants visited treat their gas washer
wastewater for solids removal prior to reuse or discharge.  This
plant is under Government directive to provide treatment within
the next two years. Of the 23 blast furnace installations studied,
three do not recycle their wastewaters.  The remaining 20 plants
have recycle rates ranging from 27.4 to 99.2 percent with a weighted
average rate of 92.4 percent.

      Two of the plants provide treatment of their blowdowns for
.cyanide removal.  One uses alkaline chlorination and one uses
Caro's Acid  (H2S05)•  Three  other plants reported unexpected cyanide
reductions which are not due to planned treatment.  One of these
plants reported that the cyanide reduction is a result of seepage
of water through the accumulated sludge in its flue dust ponds; one
reported cyanide reduction due to sparging of steam in its clarifier
to prevent freezing, and the third reported cyanide reduction when
the gas washer wastewater blowdown was used to quench slag.
                              484

-------
                                                             -  7  -
 3.    Regulatory Agencies
      Regulatory agencies of nine foreign governments were visited
 to gain insights into the regulatory climate and the relationship
 that these agencies have with industry.  This information provides
 a better understanding of the individual plant pollution control
 practices.  Countries that are members of the European Economic
 Community  (EEC) have been issued a policy directive with regard to
 control of water pollution in the community.  To date the regula-
 tions of the individual countries have taken precedence over the
 EEC directive.

      In addition to the regulatory agencies, VDEh, a West German
 trade association which represents the iron and steel industry,
 was visited.  In attendance at the meeting with VDEh were repre-
 sentatives from several steel corporations and a representative
 of a local West German Federal Government authority.

      Only two of the ten countries from which regulations were
obtained have or will have regulations specific to the iron and
 steel industry.  All others have regulations which pertain to the
quality of water discharged to,  or the effect of the discharge on,
the receiving body.  The regulations are based upon the use that
is made of the receiving body:  i.e., potable water, fishing,  re-
creation,  etc.
                              485

-------
      It was reported that input from outside of industry had
little effect in the establishment of regulations.  Generally,
the bases for regulations are: preservation of public health,
minimizing environmental effects, aesthetic considerations and
water conservation.  The economic impact of the regulations on
the individual plant, the industry and the country is considered.

      All agencies reported that the industries or individual
companies to be affected by proposed regulations are conferred
with prior to the establishment of the regulations.

      In all of the countries variances to the regulations are
subject to negotiation both prior to and subsequent to promulga-
tion.  They may be based upon available technology and/or economic
conditions.  The final regulations as they apply to the individual
plants may be referred to differently in each countries, e.g., in
England, they are called "consent conditions" and in South Africa,
"relaxed standards."

4.    Comparison between Foreign and United States Treatment

      A comparison of foreign and United States by-product coke
plants and blast furnace wastewater treatment systems reveals
that:
          in general, the treatment applied to these wastewaters
          in foreign plants is similar to that used in United
          States plants;
                              486

-------
          effluents  from  foreign plants are not monitored  for
          pollutant  content to the same degree that United States
          plants are, i.e., more parameters are monitored  in the
          United States than in foreign countries;
          foreign plants  generally recirculate blast furnace gas
          washer water to greater degrees than do United States
          plants.

      Table 1 shows  the comparative compliance with United States
effluent guidelines  limitations as presented in the "Draft Dev-
elopment Document for Proposed Effluent Limitations and Standards
for the Iron and Steel Manufacturing Point Source Category" (EPA
400/l-79/024a, October 1979) for the foreign plants observed and
the plants for which detailed data was available in the United
States.
      An indication of water use efficiency can be obtained by
comparing the degrees of EPA compliance to mass limitations.   A
larger portion of the foreign blast furnace treatment systems
that meet the guidelines limitations at BAT levels with respect
to concentrations also meet the guidelines limitations with re-
spect  to mass discharges for suspended solids and cyanide. This
indicates that less water is being discharged per unit of produc-
tion resulting in the lower mass discharges.
                              '4B7

-------
TABLE T.   COMPARATIVE COMPLIANCE OF  FOREIGN  AND.U.S.  BY-PRODUCT
COKE PLANT AND BLAST FURNACE NASTEHATER TREATMENT  FACILITIES
  WITH U.S.  EPA  DRAFT EFFLUENT  GUIDELINES FOR BPT  AND BAT
AREA
FOREIGN
U.S.
FOREIGN
U.S.
FOREIGN
U.S.
FOREIGN
U.S.
COKE
PLANT
OR BLAST
FURNACE
BY
PRODUCT
COKE PL ANT
ii
a
n
BLAST
FURNACE
it
it
ii
LEVEL
OF
TREATMENT
BPT
«
BAT
n
BPT
n
BAT
tl
PARAMETER (EXCLUDING PRIORITY POLLUTANTS)
SUSPENDED
SOLIDS
5
100%
100%
5
60%
6O%
5
80%
80%
5
40%
60%
13
100%
69%
6
100%
•00%
13
8%
15%
6
17%
0%
CN
10
100%
100%
5
100%
80%
10
70%
50%
5
0% I 0%
14
100%
93%
6
83%
67%
J4
29%
29%
6
33%
17%
CNS
NL
NL
, f
2
0%
0%
5
0%
0%
NL
NL
NL
NL
OIL AND
GREASE
2
I007o
IOO%
4
50%
75%
2
100%
100%
4
50%
75%
NL
NL
NL
NL
PHENOL
II
73%
45%
4
100%
100%
10
10%
10%
4
50%
50%
2
100%
too%
6
100%
83%
2
50%
50%
6
50%
33%
PHENOLICS
NR
—
—
5
67%
67%
NR
—
—
2
50%
50%
NL
N


NL
NL
AMMONIA
9
11%
22%
5
40%
60%
9
11%
11%
5
40%
40%
5
100%
60%
6
100%
83%
5
20%
0%
6
33%
0%
SULFIDE
NL
'NL
1
0%
0%
4
25%
50%
NR
— .
—
6
83%
67%
NR
—
—
6
40%
0%
FLUORIDE
NL
NL
NL
NL
5
40%
80%
5
100%
80%
5
2O%| 0%


j
0%
N
-------
5.    Other Observations

      While visiting Australia the opportunity to visit the John
Lysaght  (Aust.) Ltd. organization was taken to discuss the hot and
cold mill water systems at its Westernport Bay facility.  The hot
strip mill operates with the lowest blowdown of any such facility
in the world and features four recirculating water systems.  One
is a completely closed non-contact cooling water systems for the
reheat furnace skid cooling.  The other three systems have the
water cascaded with the makeup water consisting of a mixture of a
purchased supply and collected storm water.  The makeup is applied
to the area where highest quality water is required.  Blowdown is
cascaded from high water quality systems to facilities which may
tolerate lower quality.  The contact cooling water is filtered,
cooled and recirculated.  Blowdown from the system discharges to
Westernport Bay via the plant's cold mill effluent lagoon.  The
plant reports that the total discharge from the mill is 0.2 m /Mg
(48 gpt) with mass discharges of 0.002 kg/Mg (Ib/lOOOlb) each of
suspended solids and oil.
      Their cold mill complex consisting of a hydrochloric acid
pickler, a five stand cold reduction mill, a coating line and a
paint line is also an excellent example of conservation and reuse
which also results in significant pollution control.  The key to
minimizing plant water use is the segregation of water systems.
All non-contact cooling water is collected, cooled and reused in
a separate system.  Sanitary sewage is collected and treated
                              489

-------
separately.  Waste pickle liquor is regenerated in a hydrochloric



acid regeneration plant.





      The process water is treated in two separate systems: one



is the industrial water treatment system in which the relatively



clean wastewater from stands 1 and 5 of the cold mill and the



picler process water are treated, cooled, combined with tertiary



treated sanitary wastes and returned to the mill for reuse.  The



second wastewater treatment system receives the cold mill rolling



solution blowdown and dumps, the pickle liquor regeneration plant



excess rinse water, galvanizer alkali dumps, and the industrial



water treatment plant blowdown.  These wastes are treated for



discharge to receiving waters.







INNOVATIVE TECHNOLOGY




1.    Blast Furnaces




      A unit operation, not known to be practiced in the United



States, was observed at two foreign plants, August Thyssen in



West Germany and Chiba Works of the Kawasaki Steel Corporation.



It is the aeration of gas washer water prior to settling in



clarifiers or thickeners.  A portion of the settled sludge is



recirculated back to the aeration basin to act as a seed for



precipitation of carbonates.  The purpose of this operation is



to increase the cycles of concentration while not increasing the



likelihood of scale formation in the recirculation system.
                              490

-------
      Four methods of cyanide removal other than alkaline chlori-
nation from gas washer wastewater were noted.  Three of these
methods were not utilized as intentional unit operation, i.e.,the
purpose of the operation was not for the specific purpose of cya-
nide removal although removal was noted.  These operations are:

          Sparging steam through the waste.  At one plant in
          Sweden (Spannarhyttan) cyanide reduction was noted
          after steam sparging.  Steam was utilized to prevent
          freezing of water in the clarifier and apparently
          resulted in cyanide reduction from an influent con-
          centration of 30 mg/1 to 2.4 mg/1.

          Filtration of blast furnace wastewater through flue
          dust.  Two plants owned by Hoesch Estel in West Germany
          utilize sludge disposal as the means of blast furnace
          gas washer water blowdown.  The sludge is discharged
          to flue dust ponds and the excess water seeps through,
          is collected in an underdrain pipe, and discharges to
          a river.   Alkalinity is added at both plants, at one
          in the form of cold mill sludge and at the other in the
          form of caustic (sodium hydroxide).  It was noted that
          the cyanide concentration in the liquid phase of the
          sludge was 0.2 mg/1 and the cyanide concentration of
          the underdrain flow was 0.1 mg/1.  The plant has theor-
          ized that the reduction is due to metallo-cyanide
                              491

-------
          complexes being formed and being adsorbed on the flue
          dust.   No work has been done to confirm this hypothesis.

          Use of gas washer wastewater for slag quenching.  One
          plant, ISCOR's Pretoria Works in South Africa,  reported
          that when a portion of the gas washer wastewater blow-
          down is used for slag quenching the leachata from the
          slag pile is free of cyanide.  The plant has not re-
          ported the cyanide content of the raw water but stated
          that they believe that the reason for the lack of
          cyanide in the leachata is due to biological activity
          in the slag pile. No work has been done to verify this
          hypothesis.

          Pont-a-Mousson in France uses Caro's Acid (H2SO5) for
          cyanide destruction.  The plant discharges a quantity
          of gas washer water from the flue dust settling pond
          on a batch basis to chemical treatment tanks where
          Caro's Acid is added.  It reacts with and oxidizes the
          cyanide.  In the process some phenol reduction is also
          observed.

2.    Coke Plants
      Of the 23 by-product coke plants observed, 14 utilize bio-
logical methods for treatment of their wastewater.  At China
                             492

-------
Steel the by-product coke plant wastes are pretreated by fil-
tration through a bed of coke to remove excess tars that might
interfere with the biological process.  After the filtration
step, sanitary wastes from the entire plant are combined with
the coke plant wastes and treated in an activated sludge process.

      One plant combines untreated coke plant wastewater with
blast furnace gas washer water blowdown and uses the combined
wastes for irrigation of grass fields.  The grass crop is used
for cattle fodder.  No ill effects to the cattle have been re-
ported.  This method cannot be considered as innovative treatment
but rather as an innovative means of disposal.

OTHER OBSERVATIONS
      Plant and corporate managements are intimately familiar
with wastewater treatment practiced at the individual plants
and are usually apprised of potential problems before they ac-
tually occur.  Operators, in many cases, are familiar with the
theoretical as well as the practical aspects of the treatment
plant operations.
      Generally, housekeeping was observed to be of a high order.
Water was not running where it was not needed.  In plants where
space permitted green areas were set aside both to enhance the
appearance of the physical plant and to reduce noise in the plant
environs.
                              493

-------
      In one blast furnace cast house all runners were covered
with hoods and a vaccum applied.  This resulted in a noticeable
lack of fugitive emissions.

CONCLUSIONS

      Based on observation at 25 foreign plants visited that
operate either by-product coke plants or blast furnaces, or both,
it is concluded that the wastewater treatment practiced in for-
ign plants is basically similar to that practiced in the United
States.  Generally, blast furnace gas washer water is recircu-
lated to a greater degree than at United States plants.

      Two plants in Japan reported that the by-product coke
plant wastewater passed through a tertiary treatment phase, i.e.,
sand filtration followed by activated carbon adsorption.  Of all
the plants observed or reported, these were the only plants that
apparently addressed the problem of priority pollutants; however,
no data with regard to the efficiency of removal of priority pol-
lutants or effluent levels was provided when it was requested.

      Foreign effluent quality regulations are usually negotiated
between government and industry on a case by case basis.  The
economic impacts of the regulations are a major concern.
                              494

-------
 RECOMMENDATIONS





       Research should  be  conducted  to quantify  the  pollutant re-



 ductions  attainable  and to  ascertain the mechanisms by which the



 reduction of  cyanide occurred  for two of the methods observed.



 These  methods are:



          Sparging of  steam through wastewater.  Research on this



          method  should also include the effects on air quality



          and energy requirements.





          Filtering  the wastewater  through flue dust.  The re-



          search  on  this  method should also include the possible



          effects on the  air quality at sinter plants or briquet-



          ting plants  if  the cyanide containing flue dust is



          used as a  feed  stock.





       Research should  also  be  conducted to determine the effect



 of increased  recirculation  at  blast furnace gas washer opera-



 tions.  Specifically,  the method of increasing recirculation by



 aerating  the  solids  laden gas  washer water prior to settling



 should be investigated.





      Treatment of by-product  coke  plant wastes by  biological



means is  a generally accepted  and proven procedure.  However,



 the authors believe  that  coke  plant wastewater can  be combined



with blast furnace gas cleaning blowdown water prior to treat-



ment.  During discussions with steel plant personnel both in the
                              495

-------
United States and abroad, this concept was raised.  The only
concrete objection raised was that the heavy metals present in
the blast furnace wastewater would be toxic to the biological
systems.  However, fluoride which is present to some degree in
blast furnace gas washer wastewater, is a limited U.S. guide-
line parameter and lime precipitation will be required to remove
the fluoride.  When lime is added to precipitate calcium fluo-
ride, hydroxyl ions will, at the proper pH values, form metal
hydroxide precipitates.  The removal of these precipitates
should reduce heavy metals to varying degrees to levels that
would be well below those toxic to biological systems. Therefore,
when lime is added, two benefits are realized:  (1) the fluorides
are reduced to acceptable levels, and  (2) toxic metals are re-
duced to permit discharge to a biological system where the
regulated biodegradable contaminants can be oxidized.  Confirma-
tion of this concept should be proven by a research program.

      Further research should be performed to verify a second
stage biological process to nitrify the ammonia in the combined
wastewater streams.
                              496

-------
                                                         SETEC-CE-80-044
                       FACTORS  INFLUENCING BIOLOGICAL

                                 NITRIFICATION

                        OF  STEEL  INDUSTRY WASTEWATERS
                                        by
                               Ronald D.  Neufeld

                Associate Professor of  Civil Engineering
                          University of  Pittsburgh
                                   ABSTRACT

       Laboratory experiments  were conducted on the rate  of ammonia
bio-oxidation by an autotrophic culture of strict nitrifiers.  The quantitative
influence of pH, un-ionized  (free) ammonia, phenol and  elevated temperatures
on Michaelis-Menten type nitrification biokinetics was  evaluated.

       Total ammonia and pH  act via a "substrate inhibition" mechanism
to nitrification.  The maximum specific rate of nitrification decreases
proportionally to the square root of ambient phenol concentrations.  Temperatures
in excess of 30°C decrease the maximum specific rate of nitrification, decrease
nitrifier yield coefficients,  nad increase the "Michaelis-Menten" constant
leading to an overall decrease in rate kinetics and potential process instabilities
at such elevated temperatures.  Conclusions based on engineering calculations
are presented to illustrate  design and operational considerations for the
biological removal of wastewater ammonia.

KEYWORDS;

nitrification;  biological treatment;  activated sludge; biokinetics;
phenol; nitrosomonas;  free ammonia;  toxic  inhibition; temperature; coke plant;
steel industry;
ACKNOWLEDGEMENTS:
This project  is jointly supported by the AISI,  and the
U.S. EPA-IERL.  The author sincerely thanks  Mr.  John Ruppersberger,
technical project officer from the EPA for his  in-depth review
of this manuscript.
                                         4*7

-------
     "FACTORS INFLUENCING BIOLOGICAL NITRIFICATION OF STEEL INDUSTRY

      WASTEWATERS"

      Ronald D. Neufeld,    University of Pittsburgh,   Dept. Civil Engineering
INTRODUCTION

The overall objective of this research is to conduct basic studies into
possible causes of biological nitrification process instability as currently
observed in many industrial wastewater operations, and in the longer term,
to propose rational and pragmatic process operational alternatives for
the biological oxidation of ammonia.

Research activities to date have centered on quantification of the
influence of key reproducible parameters on the biokinetics of nitrification,
and calculations to illustrate the effect of changed biokinetics on
nitrification process design and operational strategies (1,2).

      Theoretical Considerations

Although several genera of autotrophic bacteria have been identified as
capable of causing nitrification, the genera nitrosomonas and nitrobacter
are considered responsible for most naturally occurring nitrification as:
      2NH+   +    302  nitrosomonas
                                     and
      2NO-   +     02  nitrobacter - >  2NQ-                          (2)

The key to a rational utilization of any biological process to
non-stereotyped applications is an understanding of the appropriate
biokinetics and defining bio-relationships.  For nitrification applications,
it has been found that appropriate utilization of Michael is-Menten kinetic
theories can serve in most cases to describe the system with reasonable
clarity.  Accordingly, the specific experimental goal of this research effort
is to develop, from a deterministic base,  modifications of nitrification
bio-kinetic relationships described by Michaelis-Mentel kinetics to account
for apparent industrial process instability observations.

Nitrification process instability can be caused by the interactions of
organic "external agent" via a toxic inhibition mechanism, and/or
via a substrate inhibition mechanism with high enough levels of un-ionized
ammonia as governed by the aqueous ammonium-ammonia equilibria.

In addition, all biological organisms exhibit optimum temperature ranges
for substrate utilization.  At lower temperatures, classical Arrhenius theory
predicts a decrease in reaction rates t  while at elevated temperature,
simultaneous enzyme-protein denatuation predicts a similar decrease in rate
biokinetics.
                                        498

-------
        Biokinetics

  The defining  equations for  biological  reactions may  be written  by
  equations  3 and  4 as:


                     v  -  Vmax  S  '  (Km  + S)                              (3)

        where;
               v -  specific nitrification rate  ( Ib NH, used/lb VSS-day)
               V     = maximum specific utilization rate
               max
               K     =* "Michaelis-Menten" constant  (mg NH./L)
               n                                        j
               S     »  substrate level (mg NH3/L)

                                       and

                    dX/dt - a(dS/dt) - bX                                 (4)

       where;
             X  *  biomass concentration (mg  VSS/L)
             t  = time
             a = yield coefficient (Ib VSS grown/lb NH3 utilized)
             b  = decay coefficient ( 1/tlme)

 The above equation is often approximated by a one-constant equation of
 the form:

                      AX  = (a)Qb AS                                    (5)

 for conditions of steady state  continuous  culture  performance.   The
 observed net yield coefficient, (a)ob»  is  tne measured harvest of biomass
 per unit ammonia removed.

       Substrate  Inhibition Models

 Many substances  act as  nutrients  at  low concentration levels, and serve to
 inhibit biokinetics at  higher levels.

 Mathematically,  the relationship  for  substrate inhibition  may be modified
 from equation  (1)  as  :

            v  =  Vmax /  { 1 +   (Km/S)   +   (S/Ki)n  }                     (6)

            where  Ki -  inhibition constant
                    n •  order of inhibition

 This  type of relationship  has been found to best describe nitrification
 biokinetics as will be  shown below.

 EXPERIMENTAL APPROACH

A culture of nitrifying organisms was developed in our laboratory two years
prior to undertaking this effort.  The organisms were acclimated  to a
                                       499

-------
synthetic ammonia waste, with sufficient inorganic carbon (sodium bicar-
bonate) and trace nutrients in open semi-continuous systems with hydraulic
detention times of about 2 days and sludge ages varied in the range of 5
to 20 days.  At no time during the course of this research were the organisms
fed any carbon source nutrient other than Inorganic carbon.  The organisms
exhibited a red color found to be typical of pure nitrification bio-systems.

The experimental approach to determining biokinetics under a variety of
conditions is that of batch respirometric evaluations.  The philosophy
behind this approach is to measure specific oxygen utilization in a constant
temperature chamber fitted with adissolved oxygen probe so that the initial
sample would not be destroyed upon monitoring and analysis.  It was noted
consistently during the course of these experiments, that ammonia was
oxidized to nitrite, and at no time during testing was nitrite oxidized to
nitrate.  Thus for purposes of this research, oxygen utilization data could
be directly correlated with specific rates of ammonia oxidation in direct res-
pirometric evaluat ions .

A calibrated Orion specific ion probe with Orion model 407 analyzer was used
for all ammonia analysis.  Alkalinity, volatile suspended solids, nitrite,
nitrate, and other trace material analysis were conducted in accord with
Standard Methods (3) .

     Un -Ionized Ammonia Kinetic Considerations

Aqueous ammonia is thought to undergo the following reactions in water;


            H20 + NH3   -  N^total  *  K"  +  OH~

            where H+  +  OH~  -  H20

The total ammonia in an aqueous solution is the sum of the un-ionized and
ionized forms of ammonia, for which all practical purposes is


            ^3 total  '  ^3  +  ml                                   <8>
                          un-ionized and ionized

From the equilibrium expression for the dissociation of un-ionized ammonia

                                                                        (9)
                     " (N!W
and         K   -  (H+)   (OH~)                                           (10)
             w

the following may be derived for the ratio of un-ionized ammonia to total
aqueous ammonia as:
                                       500

-------
                  NH-     .   .   ,
                    3 un-ionized
                                                                         (11)
                       total                  (K)  (H+)
                                                w
 Thus, the ratio of un- ionized ammonia to total ammonia in a stream is a
 function of pH and temperature, with temperature influencing the numerical
 values of K. and K .
            1      w
 Figure 1 is a summary plot of specific ammonia utilization rate (v - gNH_
 used/gvss-Day) as a function of aqueous un-ionized ammonia level at pH=8.0.

 The equation of the smooth curve fit through data points is:

             V-1-2 / (1+          +
 It should be noted that in these equations,  the order  of  inhibition  (n)  is one.

 Table 1 is a summary of results of  correlation  of all  laboratory data  in accord
 with the substrate inhibition equation  #6.

                                    TABLE  1

                 Summary of Nitrification  Biokinetic Parameters
PH
7.0
8.0
9.0
V
max
g NH3/g VSS-DAY
0.7
1.2
0.95
K
m
mg/L NH3
0.184
0.184
0.184
mg/L
500
500
200
n
1
1
1
      Application  to Design and Operation

 Using values of yield coefficient  "a" of 0.13 g VSS/g NH  and decay co-
 efficient  "b" of 0.04 DAY as also found in this research, an overall equation
 relating sludge age (9) to effluent ammonia from a one stage activated sludge
 reactor may be derived as:

          _i—  =  a (v)  -  b                                         (13)
            9

 Figure 2 is a series of calculated curves predicting the influence of pH
and sludge age on effluent total ammonia (ionized + un-ionized).   This curve
was calculated from laboratory data presented  on Table  1 and figure 2  coupled
with the ammonia-ammonium/pH equilibrium relationship of equation 9.   Also
                                       501

-------
outlined on this figure is a proposed Pennsylvania discharge standard
of 10 mg NH-/L illustrating the influence of pH and sludge age to meeting
this standard.

INFLUENCE OF ELEVATED TEMPERATURES ON NITRIFICATION

While much Municipal oriented research has been conducted on the effect
of low temperatures on nitrification, little research to date has been
published on the influence of elevated temperatures on nitrification
bio kinetics.

There is some controversy in the literature as to the optimum temperature
for nitrification.  Buswell et al., (4) cite 30°C to 36°C as the optimum
for Nitrosomonas.  Painter and Loveless (5) reported 34°C to 35°C as an
optimum for NiEFobacter, while Laudelot and VanTichelen (6) found 42°C was
the best for the same organisms.  Gibbs (7) reported that 53°C to 55°C
inactivated nitrifiers.  Sawyer and Bradney (8) in their BOD work showed
that pasteurization at 55°C proved very effective in the inactivation of
nitrifiers.  Shammas, in his doctoral dissertation (9), indicated no optimum
temperature for nitrification with a constant "activity" in the temperature
range of 15 C to 35°C with 50 percent of the nitrification activity (rate)
occurring at around 12 C.

It has been assumed by many that the cell synthesis coefficient remains
constant and independent of temperature.  Sayigh and Malina (10), Zononi (11),
and Sawyer and Rohlich (12) however, have shown this assumption not to be true.
Sayigh and Malina (10) observed cell synthesis coefficients of 1.67, 1.35, and
1.52 (Ib MLVSS synthesized/lb soluble COD removed) at 4 C, 10°C, and 20°C
respectively.  At 31 C, they showed the cell synthesis coefficient decreasing
to 0.62.

       Experimental

For this series of experiments, the nitrifiers were acclimated to specific
temperatures by equipping the complete mix semi-continuous reactors with
inexpensive "fish tank heaters".  The sludge was held at the desired temper-
ature for a period of at least two sludge ages prior to data gathering.
Data was collected in a manner similar to the  the above except that a specially
designed reactor with water bath was employed.  The steady-state data presented
below are for nitrifiers acclimated to the temperatures indicated at pH =8.0 .

Figure 3 is a plot of maximum specific ammonia utilization as a function of
temperature illustrating that at a temperature range of 22 C to 30°C.
Via = 1.26 Ib NH3/lb VSS-day, and for temperatures in the range .of 30 C to 45°C,

                        Vm - 3.78 - 0.084 (T)                            (14)

Figure 4 is a plot of Km with temperature illustrating an apparent slope
reversal at temperatures on either side of 30°C.    Figure 4, is a plot
of observed yield coefficient (defined in equation 3)  as a function of
temperature.
                                       502

-------
      Application to Design and Operation

For any temperature, values of "Vin" and "Km" may be obtained.  These
values may be substituted into equation #3 for evaluation of "v" as
a function of effluent ammonia (ammonia level surrounding the biota) .
This relationship may be substituted into equation 13 to develop a
continuous relationship of sludge age (9) vs. effluent ammonia at any
given temperature.  Figure 6 is a family of such curves of effluent
ammonia vs sludge age as functions of wastewater temperature.

In order to best interpret the  concepts developed by this figure,
a cross plot of sludge age vs. temperature was developed for a family
of effluent ammonia levels.  This was done by placing horizontal lines
across the curves of figure 6 at various effluent ammonia levels, and
plotting the points of intersection on figure 7.    The cross plots were
done for effluent ammonia levels of 5 mg/L to 50 mg/L.  As can be best
seen from figure 7, the joint effects of temperature on Vm, Km, and
yield coefficient as shown on figures 3, 4, and 5 respectively serve to
cause the nitrification process to become unstable at elevated temperatures.
The relative flatness of the curve at temperatures below 30 C indicates
a process stability for nitrification, which is also in accord with some
steel industry observations.

From a biochemical viewpoint, the overall decrease in rate kinetics at
temperatures observed in excess of 30 C may be anticipated due to progres-
sive denaturation of enzyme proteins.  Two simultaneous reactions are occurring
at the same time;  an increase in rate of biological nitrification (the
foward reaction) , coupled with an increased rate of chemical protein
denaturation (an analogy to the reverse rate) resulting as an apparent
optimum in overall reaction rate at a temperatureof about 30 C (86 F)
followed by a decrease in overall nitrification rate as the temperature
exceeds 30 C.  It should be noted that extrapolation of kinetic parameter
and yield coefficient data appears to show that an upper limit for sustained
(but slow) biological nitrification for these mesophilic organisms is
about 45°C ( 113 F) ; a value which closely agrees with textbook information
on the microbiology of nitrosomonas organisms.

INFLUENCE OF TRACE ORGANICS -PHENOL

In summary of a third research effort in this overall project, phenol
was found to effect only the Vinax term of equation #3 in a manner proportional
to the square root of the phenol concentration(l) .  An overall rate
expression for the influence of phenol on nitrification at pH-8.5 is:

             4.14 (N)
             {4.4 + /TK2.78 + N)

      An operational model of the form of equation #13 may be written as:
           m  [p 13[  Q.92 {NH3-N> - 3-0.04]  day'1          (16)

                    {(1 + /P" )(2.78 +  NH.-N)}
                          4~4             J
                                       503

-------
SUMMARY OF RESULTS AND SPECIFIC CONCLUSIONS

Based on laboratory results obtained to date coupled with theoretical
considerations for biological systems we find the following:

         1)   PH acts in conjunction with total ammonia level to
               cause the "un-ionized" (or free)  ammonia concentration
               in solution to act as a master variable influencing
               nitrification biokinetics.  Free ammonia level acts
               as a substrate inhibitor to nitrification biokinetics,
               with a maxima in specific nitrification removal rates
               existing at free ammonia levels of about 10 mg/L.

         2)   Un-ionized ammonia begins to inhibit nitrification at con-
               centrations greater than 10 mg/L.

         3)   Suggested sludge ages to maintain effective nitrification
               when un-ionized ammonia is neither limiting or inhibitory
               (considering a safety factor of about 2)  is

                  pH 7.0:  15-18 days
                  pH 8.0:   9-12 days
                  pH 9.0:  12-15 days

         4)   Nitrification biokinetics appear very sensitive to elevated
               temperatures with rates of nitrification increasing to an
               apparent maxima at 30°C, beyond which the overall  rate
               decreases.  This is found to be caused by a decrease in
               the maximum rate of nitrification (Vmax), a decrease in the
               observed yield coefficient, and Increase in Michaelis-
               Menten constant (Km)  at temperatures in excess of  30°C.

         5)   Based on theoretical calculations coupled with laboratory
               experiments, it is suggested that wastewater temperatures
               be kept below 30°C to assure stable nitrification  producing
               effluents of 10 mg/L ammonia or less.

         6)   Trace organics have been found to act as toxic inhibitors
               to nitrification biokinetics.  As one example, phenol was
               found to inhibit the maximum rate of specific ammonia
               utilization (Vmax)  in a manner proportional to the square
               root of the phenol concentration.

         7)   Work is continuing in the area of evaluation of the  influences
               of SCN, ethyl pyridine, and other trace substances on the
               biokinetics  of nitrification.  It is anticipated  that an
               overall model, or linkages of models may be developed for a
               better understanding and design of "one-stage" systems for
               carbonaceous and ammonia removals from coke plant  and other
               phenolic based wastewaters.
                                      504

-------
                          REFERENCES
 (1)  Neufeld, R.D., Hill, A.J., Adekoya, D.O.,
      "Influence of Uh-ionized Ammonia and Phenol on Nitrification
      Biokinetics for Steel Industry Wastewaters"  Annual Progress
      Report for the year ending August 1, 1979, AISI Project #78-395

 (2)  Neufeld, R.D., Rieder, C.B., Greenfield, J. H., "Influence
      of Elevated Temperatures on Nitrification Biokinetics as
      Applied to Steel Industry Wastewaters" Annual Progress Report
      for the year ending August 1, 1980, AISI Project #78-395

 (3)  APHA, AWWA, WPCF, "Standard Methods for the Examination of
      Water and Wastewater"  14th Edition, 1975

 (4)  Buswell, Shiota, Lawrence, and Meter, "Laboratory Studies on
      the Kinetics of the Growth of Nitrosomonas, with the Relation
      to the Nitrification Phase of the BOD Test,"  Applied Micro-
      biology, 2 (1954)

 (5)  Painter, H.A. and Loveless, J.E., "The Influence of Metal Ion
      Concentration and pH Values on the Growth of a Nitrosomonas
      Strain Isolated from Activated Sludge," Journal of General
      Microbiology, 52, (1968)

 (6)  Laudelout, H. and Van Tichelen, L., "Kinetics of the Nitrite
      Oxidation by Nitrobacter Winogradski." Journal of Bacte-
      riology, 79, pp.  392-442.  (1960)

 (7)  Gibbs, W.M., "The Isolation and Study of the Nitrifying
      Bacteria," Soil Science, Vol. 8, No. 6, P. 427. (1920)

 (8)  Sawyer, C.N. and Bradney, L., "Modernization of the BOD Test
      for Determining the Efficiency of Sewage Treatment Processes,"
      Science of Sewage Works, 18, p. 1113.  (1946)

 (9)  Shammas, Nazih Kheirallah "Optimization of Biological Nitri-
      fication" PhD Dissertation, Dept Civil Engineering, University
      of Michigan, 1971

(10)  Sayigh  B.A., Malina, J.F., "Temperature Effects on the Activated
      Sludge Process" JWPCF 50 no.4, 678-687, April 1978

(11)  Zononi, A.E., "Secondary Effluent Deoxygenation at Different
      Temperatures"  J.W.P.C.F. 41, 640, (1969)

(12)  Sawer, C.N., Rohlich, G.A.  "The Influence of Temperature upon
      the- Rate of Oxygen Utilization by Activated Sludges" Sewage
      Works Journal Vol II, No. 6, P 946-964  (1939)
                               505

-------
                                      FIGURE 1
       >   .11
                                  UMMZBD MMGMA (NH*> fmgj.)
  1000
I
                         FIGURE 2
                 AMMONIA vs,  SLUDGE  AGE
                                                     1.4
                                                     1.2
                                                  I
 J» 08
 _»

    as

 i
O  04


    O2
                                                                  Vnwa  vs.  Temperature

                                                                     ii       i-p
            FIGURE  3
ft- 378-j084{T«C)
tor TJfcKTC
                                                      20     »    2B     32     36     40
                                                                     TEMP   (*C)
               SludB. Ag» fa«y»>
                                          506

-------
       Km  vs.   Temperature
                                                                  Observed Yield Coefficient
IUU
60
60
40
30
20
8
3
2

'• -
FIGURE H i -
I
1 -
i
tog Km- -18829 +.06228(1) j
for T i 30» C \~ » /
tog Km • 153-0.030(7)
for 2Z«C i T i 30"C
10 2O 30 40 S
TEMP CO

at
.07
(0)ob feVSS/«N»g
> & 8 & i & 8
a
0
Temperature
i i • • * •
' «*•<-» . JiSTILr
e "X FIGURE 5
D $S.
(a)oB>OJ935-aO043(T*C) .N
tar 3C**T*48*C ™ NS
N
1 1 1 1 • IV
D 24 26 32 36 40 44
TEMP CO
Calculated Required Sudan Age
to meet Indicated NH. Effluents
                                                                              vs.
                                                                         Temperature
       Effluent NH, Concentration
                  vs.
              Sludge Age
                                                       90
                                                       eo
                                                       70
                                                  I   «,
                                                       so
                                                  I   «
                                                  I   30
                                                  "   20
                                                       10
                                                         20
24
  28      32
 TEMP   CO

FIGURE 7
                                                                                         36
                                 40
020   3040906070809000  110  BO
           SLUDGE  AGE  (day*)

              FIGURE 6
                                               507

-------
                       FLOTATION OF IRON-CYANIDE COMPLEXES

                      FROM IRON AND COKE PLANT WASTE WATERS
                         R. 0.  Bucsh
                         Research Engineer
                         International Nickel,  Inc.
                         Sterling Forest
                         Suffern, NY  10901

                         G. W.  Lower
                         Professor of  Metallurgical  Engineering
                         Michigan Technological University
                         Houghton,  MI  49931

                         D. J.  Spottiswood
                         Associate  Professor of Metallurgical Engineering
                         Colorado School of Mines
                         Golden,  CO 80401
ABSTRACT

    Distribution curves indicate that long chain quaternary amines will complex
cyanide, ferrocyanide and ferricyanide.  The neutral organometallic complexes
formed are hydrophobic because of the long chain alkyl groups of the amine
and therefore should be capable of attachment to an air bubble and concentrated
in the froth product.  This concept was tested on synthetic ferricyanide
solutions and to a limited extent on ferrocyanide solutions in a small con-
tinuous flotation column.  Ferricyanide removal was found to be a function of
retention time, initial ferricyanide concentration, mole ratio of amine to
ferricyanide, and chloride interference.  The air flow rate and solution pH
in the range of 4 to 10 had relatively minor effects.  Under optimum conditions
ferricyanide removals of approximately 80% were achieved in a single stage
flotation.
                                      509

-------
                     FLOTATION OF IRON-CYANIDE COMPLEXES

                    FROM IRON AND COKE PLANT WASTE WATERS


     Soluble cyanide species exist in by-product coke plant effluents in concen-
trations of 7.0 to 110 mg/H of cyanide.   The cyanide is present as metallic
cyanide complexes, principally ferrocyanide and ferricyani.de, as well as free
cyanide.  The U.S. Environmental Protection Agency (EPA) has issued regulations
limiting the discharge of cyanide as given in the BATEA guidelines to 0.25 mg/£.
These regulations are scheduled to take effect on July 1, 1985.

     Economic removal of the relatively low levels of cyanide in these effluents
requires a process capable of high volume treatment at low cost.  Flotation
processes such as used in the mineral industry meet these requirements.  However,
in contrast to flotation of fine mineral particles, cyanide removal requires the
flotation of an ionic species or microparticulate particles.  One possible method
of achieving this ionic flotation is to complex the ionic cyanide species with a
hydrophobic complexing agent to produce a neutral ion-pair complex capable of
attaching itself to an air bubble.  Grieves and Bhattacharyya2 have demonstrated
that a cationic surfactant, ethylhexadecylammonium bromide allows removal of
ferrocyanide in a batch foam separation process.  They also demonstrated that
microparticulate iron-ferrocyanide would respond to this reagent.

     In the present study the ion flotation of ferricyanide was studied using a
quarternary amine surfactant, tricaprylmetnylammonium chloride (Aliquat 336,
General Mills), in a single stage column flotation process.

COMPLEX FORMATION

     In order to determine the ability of the amine to complex cyanide species,
distribution tests were run using a chloroform solution containing 1% amine and
aqueous solutions of 2.0 mg/fc FeCCN)^"1, 20.0 mg/Jl Fe(CN)g3, and 20.0 mg/Jl CN~
respectively.  As shown in Figure 1 the amine extracts all of these ions from
the aqueous phase indicating complex formation and therefore the possibility of
their removal by flotation.

FLOTATION OF FERRICYANIDE

     All flotation tests were carried out in a continuous flotation column using
synthetic solutions of K3Fe(CN)6 containing 36.8 mg/fc Fe(CN)63.  The column had
an inside diameter of 4.7 cm and an effective length of 47 cm.  A fritted glass
disc with a pore size of 25-50 micrometers was used as a gas dispenser.  A diagram
of the apparatus is shown in Figure 2.  The general flotation procedure was to
disperse the amine in a small volume of feed solution and add this dispersion - to
the bulk of the feed in the conditioning tank.  After a period of conditioning,
usually 10 to 12 minutes, the feed solution was pumped through the column counter-
current to the air flow and the froth product removed.  Unless otherwise stated
an air flow rate of 0.16 Jl/min/cm2 and a flotation time of 1.74 minutes was used.

     The effectiveness of flotation was determined by the fraction removed, Fr,
and by the removal factor, RF, as defined below:

              Fr - (Af Vf - AU Vu)/Af Vf                                     (1)

              RF - 1 - Au/Af                                                 (2)

                                      510

-------
   4.0
   3.0
I
Ol
c
o

u 2.O
O

L.
 y
.'c
 a
 O)
 i_
O
   1.O
      0            0.1             0.2

           Aqueous-Ferrocyanide (tig/ml)
                                                80
                                                70
             60
                                              ai
                                                50
                                              o

                                             .y  40
                                              u

                                             £
                                             -« 30
                                             a
                                             E1
                                             o
                                                20
                                                1 0
                                                 0
                      0.1     O.2    0.3     0.4

                     Agueous-Ferricyanide
                                                                                     0.5
                                                                                            80




                                                                                            7O




                                                                                            60



                                                                                          E
                                                                                          01 50
                                                        § 4O
                                                        >>
                                                        U
                                                        I
                                                        y


                                                        |30

                                                        6
                                                          20
                                                          1 0
                                                           O
0.5     1.0     1.5

   Aqueous -Cyanide
                                                                                         2.0
                                                                                                                                  2.5
Figure 1.  Distribution equilibria  for
                                                                                      .  apeciec.

-------
                              To Vacuum
en
H->
ro
   Pump
                Constant
               Head Tank
                           IU1
             Conditioner
                          — Air Supply
r
Overflow
  Trap
Air-flow
 Meter
                              Anti-syphon
                               Underflow
                                 Outlet
                            Underflow
                            Solution
                              Tank
                                                      Air Inlet and
                                                     Trap Assembly
                              Figure 2. Flotation apparatus.

-------
 where Af and Ay are the cyanide concentrations in the feed  and  underflow respec-
 tively,  and Vf and Vu the volumes of the feed and underflow.  The  fraction removed
 gives the fraction of total cyanide in the overflow and  is  a  function of the amount
 of solution carried over in the froth product.  The removal factor assumes the
 overflow solution volume is zero.

 Amine Concentration

      The overall reaction for complex formation is shown in Equation  3 below:

           3R4N Cfc + FeCCN)63 - Fe(CN)6 (RAN) 3 + 3 CJT1                          (3)
 Therefore  the  theoretical  quantity  of amine required  to form a neutral  complex
 would be at  an amine to  ferricyanide mole  ratio of  3  to 1.   Since  this  above  reac-
 tion is an equilibrium reaction,  some excess amine  would be  required  to assure
 essentially  complete reaction.  Maximum recovery was  achieved at an amine- ferri-
 cyanide ratio  of  3.75 to 1 as shown in Figure  3.  The neutral complex formed
 agglomerates into small  wax-like  particles and is recovered  in the froth product.
 This amine usage  is  somewhat higher than the theoretical requirements as might be
 expected based on equilibrium considerations.  At lower ratios the recovery decreases
 because of insufficient  amine.  At  higher  ratios the  recovery also decreases.  At
 the  high amine ratios, emulsification occurs with & resulting decrease  in recovery.
 This emulsification  probably results from  adsorption  of excess amine  on the wax
 particles  with the hydrophilic end  of the  amine outward resulting  in  particle
 repulsion.

     Many  coke plant waste waters contain  relatively  high concentrations of
 chloride.  Since  the amine is in  the chloride  form, higher amine concentrations
 should be  required in chloride solutions in order to  drive the reaction to the
 right as indicated in Equation 3.   Flotation from solutions  containing  5000 mg/Jl
 chloride indicate that this occurs  as shown in Figure 4.  However, in these solu-
 tions the  recovery does  not reach a maximum but continues to increase with increas-
 ing  amine  addition.   Emulsification does not occur  in these  chloride-containing
 solutions  at these higher  amine ratios.

     Sutherland and  Wark3  have shown that  the  presence of chloride ion  decreases
 the  critical micelle concentration  of amines.  The  polar groups of the  amines
 forming the  micelle  would  be at the surface of the  micelle and thus react with
 the  ferricyanide.  In addition, by  being in this micelle formation, the hydrocarbon
 end  of the amine  is  not  available for adsorption on the wax  particles.  Consequently
 no emulsification occurs.

 Flotation  Variables

     Recovery was  found  to increase rapidly with increasing  conditioning time until
 a steady state  was reached at approximately thirteen minutes as shown in Figure 4.
 It is possible  that  the  conditioning time  could be  decreased by dissolving the
 amine in a water  soluble solvent  such as ethanol prior to addition.

     Recovery also increases with increasing flotation time as shown  in Figure 5.
As indicated the  recovery  increased from 75% at a flotation  time of 1.07 minutes
 to 83% at  a  flotation time  of 2.3 minutes.

     Increasing the  air  flow rate over a range of 0.08 to 0.22 fc/min/cm2 had only a
minor effect on recovery as shown in Figure 6.  The sharp decrease in the removal
 factor at  the high air flow rate resulted  from excessive carryover of solution
into the overflow product.

                                      513

-------
   0-9r
   0-8-
   0-7
b
a
LL
   0-6
   0-5
o
o 04
&
   0-3
    02
    0-1
o Without  Chloride
• With Chloride
     0
      0
     1-5     30      4-5     6-0      7-5     9-0
    Mole  Ratio - Amine  to Ferricyanide  ^
    Figure 3.  Recovery as a function of amine addition.
                               514

-------
   0-9
   0-8
u
C9
cc
   0-6
                        Removal  Factor
   0-5
      0       5       10      15      20      25

                   Conditioning Time  (min)


         Figure 4.  The effect of conditioning time on recovery.
30
                           515

-------
o
u
cs>
ce
   0-9
   0-8
   0-7
     0-5
           o Fraction  Removed

           • Removal Factor
 1-0       1-5      2-0


Flotation Time  (min)
2-5
  Figure 5.  The effect of flotation time on recovery.
                 516

-------
   0-8
   0-7
o
u
   06
  05
            o Fraction  Removed

            • Removal  Factor
    •05      -10      -15       20      -25


          Air  Flow (liters/min-cm )



        Figure 6.  Recovery as a function of air rate.
                 517

-------
Solution Variables

     Recovery remained essentially constant at about 82%  over a ferricyanide con-
centration range of 30 to 70 ing/A as shown in Figure 7.  Below this range the
recovery decreased to about 65% at 5.0 mg/fc ferricyanide.

     The solution pH had little effect on recovery over a pH range of 4 to 7 as
shown in Figure 8.  The slight decrease in recovery above pH 7 probably results
from increased competition for the amine by the hydroxyl ions.

FEKROCYANIDE FLOTATION

     A limited amount of work on ferrocyanide flotation gave recoveries comparable
to that of ferricyanide under comparable conditions.  However, complex formation
was slower with ferrocyanide.  Cyanide ion does not float well at all probably
because of ion hydration.  However, free cyanide can be converted to ferrocyanide
by addition of ferrous sulfate at neutral or slightly acidic pH values.  Conversion
of free cyanide to ferricyanide is much more difficult.  Present work is being
directed toward ferrocyanide flotation.

ACKNOWLEDGEMENTS

     This work was carried out at Michigan Technological University with the
support of the U.S. Environmental Protection Agency and the American Iron and
Steel Institute.

REFERENCES

1.  U.S. Environmental Protection Agency, "Development Document for Effluent
    Limitations Guidelines and New Source Performance Standards for the Steel
    Making Segment of the Iron and Steel Manufacturing Point Source Category,"
    EPA-440/l-74-024-a, p. 371, Table 67 (1974).

2.  Grieves, R. B., and D. Bhattacharyya, "Foam Separation of Complexed Cyanide:
    Studies of Rate and of Pulsed Addition of Surfactant," J. Appl. Chem., Vol. 19,
    p. 115 (April, 1969).

2.  Grieves, R. B., and D. Bhattacharyya,  "Precipitate Flotation of Complexed
    Cyanide," Separation Science, p. 301 (August, 1969).

2.  Grieves, R. B., and D. Bhattacharyya, "Foam Separation of Cyanide Complexed
    by Iron," Separation Science, p. 185 (April, 1968).

3.  Sutherland, K. L., and Wark, I. W., "Principles of Flotation," Aus* I.M.M.,
    Australia, p. 244 (1955).
                                      518

-------
       1.0
       0.9
01
>0.8

o»

o
u
(*

*  0.7
       0.6
       0.5
           0       10
                                                o  Fraction Removed

                                                •  Removal Factor
                       20      30      40     50      60

                         Ferricyanide Concentration  (ppm)
70     80
                         Figure 7.  Tar. o.ffoct of fo.rricyanica concentration on recovery.

-------
01
fo
o
u
&
o:
       0.9
       0.8
        0.7
               •o-


               • ••
                            o   Fraction  Removed

                            •   Removal Factor
                                               7

                                              PH
                                                   8
10
11
                                 Figure 8.  Recovery as a function of pH.

-------
Session 3: SOLID WASTE POLLUTION ABATEMENT

Chairman:   John S. Ruppersberger
           Industrial Environmental Research Laboratory
           U.S. Environmental Protection Agency
           Research Triangle Park, NC
                     521

-------
IMPACT OF THE RESOURCE CONSERVATION AND RECOVERY

        ACT (RCRA) ON THE STEEL INDUSTRY
                Penelope Hansen
                      and
                William J. Kline

     Hazardous & Industrial Waste Division
             Office of Solid Waste
      U.S. Environmental Protection Agency
                          523

-------
                   RCRA and the Steel Industry

     The Resource Conservation and Recovery Act (RCRA) is
structured to ensure that our society will view its waste manage-
ment responsibilities seriously.  The failure to realize or predict
the consequences of improper waste management in the past have and
probably will continue to manifest itself with often tragic inci-
dents.  RCRA will have a major effect on how all wastes, whether
hazardous or nonhazardous,  will be handled in the future.

     Nearly 60 million metric tons of solid waste are currently
generated each year by the steel industry in the U.S.  Approxi-
mately 80-90% of this total quantity is probably non-hazardous.
Due to commercial sale and/or in plant recovery of over 60% of
these solid wastes, approximately 19 million metric tons (excluding
rubble) of non-hazardous solid wastes remains to be disposed each
year.  Table 1 shows the estimated types and quantities of non-
hazardous solid waste annually generated.

     The steel industry also generates hazardous wastes.  Table 2
is a list of some wastes which EPA believes to be hazardous.

     This paper gives a description of the regulatory requirements
for hazardous.and non-hazardous wastes and their expected impact on
the steel industry.

Regulation of Hazardous Wastes

     EPA's new hazardous waste managment system may not eliminate
all of the dangerous sites and problems resulting from our past
complacency regarding proper hazardous waste disposal, but the
regulations do initiate the establishment of a system which will
lead to the proper management of hazardous waste and prevent the
creation of new catastrophic situations.

     This system is based upon the concept of "cradle-to-grave"
management, i.e., tracking the waste from its point of generation
to its point of disposal.  All generators, transporters, and
owner/operators of storage, treatment, and disposal facilities
for hazardous waste have responsibilities within this sytem.

     Under Subtitle C of the Resource Conservation and Recovery
Act, EPA has promulgated six regulations:
                                 524

-------
  RCRA
 Section
      Subject of  Regulation
  Final  Regulation
   3001


   3002


   3003


   3004
   3005


   3006


   3010
 Identification and Listing of
 Hazardous Waste

 Standards for Generators of
 Hazardous Waste

 Standards for Transporters of
 Hazardous Waste

 Standards for Hazardous Waste
 Facilities:
 Phase 1 - Preliminary Facility
          Standards
 Phase 2 - Technical Design
          Standards

 Permits for Treatment, Storage,
or Disposal Facilities

Guidelines for Development of
State Hazardous Waste Programs

Notification Process
   May  19,  1980


February  26,  1980


February  26,  1980




   May  19,  1980

    Fall, 1980


   May 19,1980


   May 19,1980

February 26, 1980
        Identification  and  Listing of Hazardous Waste  (3001)

     The first major area  of the regulations  is the identification
and listing of hazardous wastes.  RCRA defines a hazardous waste
essentially as a solid waste that may cause substantial hazard to
health or the environment  when improperly managed.  The Act also
instructs EPA to list known hazardous wastes  and to establish
criteria for the testing of all wastes to determine whether or not
they are hazardous.

niaracteristies of Hazardous Waste

     Hazardous wastes are  identified on the basis of measurable
characteristics for which  standardized tests  are available.  The
principal characteristics  that make a waste hazardous are specified
levels of:

          ignitability - posing a fire hazard during routine
          management
                                 525

-------
                                                 Table 1

                               Estimated Annual Quantities of Non-Hazardous
                               Solid Wastes Generated by the Steel Industry
                        (Based on Raw Steel Production of 114 Million Metric Tons) (Ref.1,6)
        Waste
     Slags
  Blast Furnace
  Basic Oxygen Furnace
  Electric Furnace
  Open Hearth Furnace

     Scales
  Soaking Pit
  Primary Mill
  Continuous Casting
  Rolling Mills

     Sludges
01 Blast Furnace
en Basic Oxygen Furnace1
  Rolling Mills
  Plating,  Galvanizing

     Dusts
  Blast Furnace
  Basic Oxygen Furnace2
  Open Hearth Furnace

     Miscellaneous
  Fly Ash/Bottom Ash3
  Rubble
Quantity Generated

  Thousand Metric
  Tons (dry wgt.)

      20,850
      10,200
       3,000
     	4,400
     Waste Dispositions
      38,450

       1,000
       4,600
         100
         700
       6,400

       1,800
         800
         275
         100
       2,975

       1,200
         400
         250
       1,850

         250
       3,400
       3,050
   Percent
Recycled/Sold

    95%
    50%
    20%
    25%
    69%
    80%
    80%
    80%
    67%

    75%
    25%

   100%
    56%

    85%
    25%
    15%
    55%
                                                  0%
 Percent
 Disposed

    5%
   50%
   80%
   75%
   31%

 100%
   20%
   20%
   20%
   33%

   25%
   75%
 100%

   44%

   15%
   75%
   85%
   45%

 100%
 100%
" 100%
 Quantity Landfilled

       Thousand
      Metric Tons

        1,040
        5,100
        2,400
        3,300
  1  64%  of BOF's utilize wet emission controls, 35% utilize dry controls.
    will be generated whichever device is utlized.
       11,840

        1,000
          920
           20
          140
        2,080

          450
          600
          275

        1,325

          200
          300
          200
          700

          250
        3,400
        3,650

Dust in the form of kish
  2  Since 90% of electric furnaces utilize dry controls, assume dry controls are used solely.
  3  Assume (a)  2.5 million metric tons of coal consumed (b) ash content of coal is 12%.

-------
                                 Table  2

                           AnnuaJ
               Solid Wastes
Estimated Annual Quantities of Hazardous
lid Wastes Generated by the Steel Industry(Ref•2)
                                      Quantity               Percent
        Waste                         Generated              Disposed

Decanter Tank Tar Sludge          65,000 metric tons            55%

Ammonia Still Lime Sludge        870,000 metric tons            85%

Electric Furnace Dust/Sludge     340,000 metric tons          100%

Spent Pickle Liquor               1-4 billion gallons           40%

Sludge from Lime Treatment       5 million metric tons        100%
  of Spent Pickle Liquor1

Spent Halogenated Solvents
  and Recovery Sludges/Still        20,000 metric tons        100%
  Bottoms
  Assumes  treatment of all spent pickle liquor.
                                     527

-------
          corrosivity - ability to corrode standard containers,
          or to dissolve toxic components of other wastes

          reactivity - tendency to explode under normal manage-
          ment conditions, to react violently when mixed with
          water, or to generate toxic gases

     -    EP toxieity - (as determined by a specific extraction
          procedure) - presence of certain toxic materials at
          levels greater than specified in the regulation.

List of Hazardous Wastes

     Wastes that possess any of the four hazardous waste charac-
teristics or that meet the criteria for general toxieity have
been included in the hazardous waste listing.  The waste listing
is composed of several sections:  specific wastes, waste sources,
and waste processes.

     General toxieity is defined as characteristic of waste which
contain one or more constituents that have been found to have toxic
effects on humans or other life forms.  EPA can also consider
other factors to determine if the waste may cause or potentially
cause "substantial" hazard to human health or the environment.
The other factors which EPA may consider are:

          the degree of toxieity of the toxic constituents of the
          waste;

          the concentration of these constituents in the waste;

          the potential for these constituents or their by-products
          to migrate from the waste into the environment;

          the persistence and degradation potential of the con-
          stituents or their toxic by-products in the environment

          the potential for the constituents or their toxic by-
          products to bioaccumulate in ecosystems;

          the plausible and possible types of improper management
          to which the waste may be subjected;

          the quantities of the waste generated;

          the record of human health and environmental damage
          that has resulted from past improper management of
          wastes containing the same toxic constituents.

     It is possible for the generator to get an exemption from
regulation even if the waste is listed in the regulation.  The
regulation includes delisting procedures for generators to follow
if they believe their facility's individual waste is fundamentally
                                  528

-------
different  from the waste listed.  The generator must demonstrate,
or reference test data that demonstrate, that the specific waste
does not meet the criteria which caused the Agency to list the
waste.  This provision allows flexibility recognizing that indi-
vidual waste streams vary depending upon raw materials, industrial
processes, and other factors.

Excluded Wastes

     Certain wastes are not subject to RCRA Subtitle C hazardous
waste controls.  Some of the wastes excluded and applicable to
the steel  industry include:

     *     industrial wastewater discharges that are point source
          discharges subject to regulation under Section 402 of
          the Clean Water Act, as amended;

     •    wastes that are reused or recycled, except for the
          storage and transportation of sludges and listed
          wastes;

     0    fly ash, FGD sludge, bottom ash from combustion of
          coal or other fossil fuels.

Generator  (3002):

     The regulations (40 CPR Part 262) issued under section 3002
of RCRA require a generator of hazardous waste to determine if
his waste is hazardous.

     This determination may be made via one of the following means:

     (1)  a waste may be listed by EPA as being hazardous;

     (2)  if not listed,  the waste may be tested by the generator
          against the characteristics for determining hazardousness;

     (3)  the generator may declare the waste to be hazardous
          based upon his knowledge of the materials or processes
          used in generating the waste.

     Additionally a generator is required to:

          obtain an EPA identification number

          obtain a facility permit if the waste is accumulated
          on the generator's property more than 90 days

          use appropriate containers and label them properly for
          shipment .

          prepare a manifest for tracking hazardous  waste
                                  529

-------
          assure, through the manifest system, that the waste
          arrives at the designated facility

     -    submit an annual summary of activities.

     Once a solid waste is determined to be hazardous, it is
subject to all of the controls under Subtitle C and the generator
transporters, storers, treaters, and disposers of the waste are
deemed responsible to meet the applicable requirements.

Manifest

     The major mechanism for tracking and controlling hazardous
waste is the manifest system.  A generator of hazardous waste is
responsible for preparation of a manifest containing:

          name and address of the generator;

     -    names of all transporters;

          name and address of the permitted facility designated
          to receive the waste.  (An alternate facility may be
          designated if an emergency prevents use of the first
          facility);

          EPA identification numbers of all who handle the waste;

     -    U.S. Department of Transportation (DOT) description
          of the waste;

     -    quantity of waste and number of containers;

          the generator's signature certifying that the waste
          has been properly labeled, marked, and packaged in
          accordance with DOT and EPA regulations.

The owner/operator of the facility receiving the waste is respon-
sible for verifying delivery of the waste and returning a copy of
the manifest to the generator.

Transporters (3003)

     The regulation for transporters of hazardous waste (40 CPR
Part 263) was developed jointly by EPA and the U.S. Department of
Transportation (DOT).  The EPA regulation on transporters incor-
porates by reference pertinent parts of DOT's rules on labeling,
marking, packaging, placarding, and other requirements for
reporting hazardous discharges or spills during transportation.
DOT, in turn, is amending its regulations on transportation of
hazardous materials to include EPA*s requirements.
                                  530

-------
      The  regulation (Part  263)  requires  a transporter of hazardous
waste to:

           obtain  an EPA identification number;

           comply  with  the  manifest  system for tracking hazardous
           waste;

           deliver the  entire quantity of hazardous waste to the
           facility  designated by the generator on the manifest

           retain  a  copy of the  manifest  for 3 years;

           comply  with  DOT  regulations pertaining to reporting
           of discharges or spills;

           clean up  any hazardous waste discharged during
           transportation.

Facility  Standards  (30041

      Owners and operators  of facilities  that treat, store, or
dispose of hazardous waste must comply with the standards promul-
gated under section 3004 of RCRA (40 CFR Parts 264 or 266).  The
regulations under this  section, which set standards for hazardous
waste facilities, serve a  threefold purpose:

           to establish  proper treatment, storage, and disposal
           practices;

           to provide States with minimum standards in order to
           receive EPA  approval  (required under section 3006 of
           RCRA) of  their hazardous waste programs;

           to provide the technical basis for EPA-issued facility
           permits (required under section 3005 of RCRA) in States
           that do not operate a RCRA program.

      EPA  is promulgating standards for hazardous waste facilities
in two phases.  Phase  I  -  Interim Status Standards (in the spring
of 1980) provide  facilities with temporary authority to continue
their operations, upon  notifying EPA and obtaining an identifica-
tion  number.  This  temporary authority will be effective until
promulgation of the Permanent Status Standards.

      "Interim status" gives hazardous waste facilities temporary
authority  to continue operations pending final administrative
action on  facility permit  applications (required under RCRA
Section 3005).  Of course,  until the permit decision is made,  all
hazardous waste facilities must meet the conditions of the Interim
Status Standards  as stated in the May 19, 1980 Federal Register -
Final Hazardous Waste Regulations.   Facility owners and operators
who qualify for interim  status are treated as having a permit
                                  531

-------
during this period.  To qualify for interim status, a facility
must have been in existence (in operation or under construction)
on November 19th.

     Interim Status Standards are designed to ensure adequate
operating practices and closure and post closure activities.
Requirements for interim status are largely managerial—they do
not include financial assurance or design and operating standards
required for a facility permit.  These standards will soon be
promulgated as Phase II.

Interim Status Standards (Part 265) for nonpermitted facilities,
promulgated as Phase I in spring 1980, include:

1.   Administrative and nontechnical requirements:

     -    General

          —   waste analysis:  detailed chemical and physical
               analyses, waste analysis plan,  specific require-
               ments for each facility type

               security:  artificial or natural barrier with
               controlled entry or 24-hour surveillance, and
               warning signs

          —   inspection:   inspection plan and log;  remedy of
               any deterioration,  malfunction,  or imminent
               hazard

          Personnel training

          —   classroom or on-the-job training, annual review of
               initial training,  records on personnel training

          Preparedness and  prevention

               alarm system and emergency equipment;  access to same

               arrangements with local emergency authorities

          Contingency plan, emergency procedures, and emergency
          coord inatbr

          Manifest system procedures

          Operation records of activities required by the regu-
          lation,  such as manifest information,  waste analyses
          records,  testing  and analytical data,  and demonstration
          reports  for variances

          Reporting requirements,  such as annual reports and
          unmanifested waste reports.
                                 532

-------
 2.   General facility requirements (Phase I):

           General operation requirements

           Special requirements for ignitable, reactive,  and
           incompatible wastes

           Ground-water monitoring (monitoring system to be in
           operation by November 1981)

           Closure and postclosure plans:  estimate of costs and
           description of how facility will be closed,  notice of
           facility closure,  and postclosure monitoring and main-
           tenance

 3.   Specific facility requirements:

           Disposal of liquids in landfills or containers

           Control of runoff  from waste piles,  land treatment,
           and landfills (controls to be in operation by
           November 1981)

           Land treatment  facilities monitoring and restrictions
           on growing foodchain crops

           Incinerators and treatment  facilities

      -     Underground injection

 Permanent  Status  Standards (Part 264)  for  permitted facilities,
 will  soon  be promulgated  as  Phase II.

      General facility requirements will  include technical,
 monitoring,  closure and postclosure, and financial  requirements.
 The facility permit regulation under section  3005  of RCRA  becomes
 effective  and processing  of  permit applications begins at  this
 time.

 Further technical  requirements will be promulgated  by EPA  inter-
 mittently  over  a period of years.  These will  include resolution
 of complex technical  issues  and reproposal and  promulgation of
 more  definitive Phase II  standards, for  example, specific  design
 or operating standards  for landfills.   The technical refinements
nu_y also include standards for specific  industries  and wastes
which require tailored  standards.

 Facility Permits  (3005)

     The regulation promulgated under section  3005  of RCRA (40
CPR Parts  122 and 124)  requires that any person who owns,  operates,
or proposes  to  own or operate  a facility that treats, stores, or
disposes of hazardous waste receive a permit from EPA or a State
                                  533

-------
authorized to conduct its own hazardous waste program.  Most
requirements in the regulation are only applicable where EPA issues
permits; selected portions apply to authorized State programs.

Consolidated Permit Program

     EPA issues permits for controlling environmental problems
under a number of laws.  To facilitate and streamline the regu-
latory process, EPA has consolidated procedures and requirements
for the hazardous waste management program with four other programs
it administers:

          the Underground Injection Control (UIC) program under
          the Safe Drinking Water Act

          the National Pollutant Discharge Elimination system
          (NPDES) under the Clean Water Act

          the Dredge or Fill (section 404) Program under CWA

     -    the Prevention of Significant Deterioriation (PSD)
          Program under the Clean Air Act where this program is
          operated by EPA.

     A facility seeking more than one permit is encouraged to
consolidate application.

Exclusions

     Certain facilities handling hazardous waste do not require
a RCRA permit:

          generators who accumulate hazardous waste on-site for
          less than 90 days

     -    persons who own or operate facilities solely for the
          treatment, storage, or disposal of certain hazardous
          waste excluded from regulation.

Applying for a Permit

     Any person who now owns or operates a hazardous waste facility,
or who plans to in the future,  must apply for a permit.  This re-
quirement applies to all existing facilities, facilities in opera-
tion or an intermittent basis,  or facilities which commenced
construction as late as yesterday, November 19.  The application
is in two parts:

          Part A, which defines the processes to be used; the
          design capability; and the hazardous waste to be
          handled.  For existing facilities, Part A must be
          submitted within 6 months of promulgation of the
          regulation identifying hazardous waste.
                                 534

-------
           Part B,  which contains more detailed information
           Intended to establish that the facility can meet the
           technical standards promulgated under RCRA section
           3004.   For existing facilities,  Parti B must be sub-
           mitted at a date set by the Regional Administrator.

      For proposed  new facilities,  both Part  A and Part B must be
 submitted at least 180 days before physical  construction is
 scheduled to begin.

      If notification is filed with EPA and Part A of the permit
 application  is submitted on time,  an existing facility achieves
 interim status and is considered to have a permit until Part B is
 acted upon.   Obviously,  EPA will not be able to issue thousands of
 hazardous waste  permits in less than several years.   If approved
 by  EPA,  a permit with a term of not more than 10 years will be
 granted.   Meanwhile,  the facility must comply with Interim Status
 Standards promulgated under RCRA section 3004.

 State Programs (3006)

      Congress clearly prefers that States  assume responsibility
 for controlling  hazardous  wastes within their borders.   Federal
 financial assistance  is  available  from EPA to States  for developing
 their programs.  Section 3006 of the  Act specifically provides
 for States to operate  their own hazardous  waste programs in lieu
 of  the  Federal program,  after the  authorization by EPA.   In States
 whose programs do  not  meet the minimum requirements under  RCRA,  or
 who do  not apply for  authorization,  EPA must  administer  the program.

      The  regulation issued under section 3006  of RCRA (40  CFR
 Part  113)  establishes  minimum requirements for  State hazardous
 waste programs in  order  to receive EPA approval.   The  regulation
 is  designed  to assure  consistency  in hazardous  waste management
 from  State to  State.
                                 i
      RCRA generally directs  that to receive EPA "final"  approval
 State hazardous  waste  programs must be  "equivalent to and con-
 sistent with" the  Federal program.  "Equivalent"  is interpreted
 to mean "equal in  effect."   Thus,  the regulations provide minimum
 requirements, with the States  allowed to set more stringent  stan-
 dards.  Another  important  element  is that States may not impose
 any requirement  that might  interfere with the free movement of
hazardous wastes across State boundaries to treatment, storage,
c - disposal  facilities holding a RCRA permit.

     State programs that are  "substantially equivalent" to the
Federal program may receive  interim authorization, then be
gradually upgraded until they qualify for "final" or "full"
authorization.
                                 535

-------
Notification Process (3010)

     The control system starts when those engaged in generating,
transporting, treating, storing, or disposing of hazardous waste
notify EPA as required by section 3010 of RCRA.  After receiving
notification, EPA assigns an identification number to the notifier.
Anyone engaged in transporting, treating, storing, or disposing
of hazardous waste who does not notify EPA during the 90-day period
following the promulgation of the regulation identifying hazardous
waste may not begin or continue operation after the effective date
of the regulation without obtaining an EPA identification number.

     The regulation issued under section 3010 of RCRA requires
that:

          anyone who generates or transports hazardous waste or
          owns or operates a facility that treats, stores, or
          disposes of hazardous waste must notify EPA.

     -    a new generator or transporter must apply to EPA for an
          identification number before any hazardous waste can be
          transported.   Application for an identification number
          must be made on the notification form.

          an owner/operator of a site that conducts more than one
          hazardous waste activity (for example, generation and
          disposal) may file a single form to cover all activities
          at that site.

          an owner/operator of more than one site must file a
          form for each site.

     This hazardous waste management system is a new and complex
plan which can and will work only if everyone involved with
hazardous waste assumes his responsibilities.

Steel Industry Hazardous Waste

     As of today, there are four specific source and a number of
non-specific source types of wastes associated with the steel
industry which have been determined by EPA to be hazardous wastes
(as listed in the Federal Register of May 19 and July 16, 1980).

     The five specific source wastes are:

     (1)  Ammonia still lime sludge from coking operations;

     (2)  Decanter tank tar sludge from coking operations?

     (3)  Spent pickle liquor from steel finishing operations;

     (4)  Emission control dust/sludge from the primary produc-
          tion of steel in electric furnaces.
                                   536

-------
      The above wastes were listed as hazardous due to their toxic
 and/or corrosive characteristics.  A summary of the reasons for
 listing each of these wastes follows:2

 A.    Coking

     1.  Ammonia still lime sludge from coking operations (T)  5/19/80

           a.   These sludges contain the  hazardous constituents
                cyanide,  naphthalene,  phenolic compounds,  and
                arsenic which adhere to the  lime floes and solids
                in significant concentrations.

           b.   Cyanide and phenol leached in significant concen-
                trations  from an  ammonia still lime sludge waste
                sample which was  tested by a distilled water
                extraction procedure.   Although no  leachate data
                is currently available for naphthalene and arsenic,
                the Agency strongly believes that,  based  on con-
                stituent  solubilities,  the high concentration of
                these constituents in  the  wastes, and  the  physical
                nature of the waste,  these two  constituents are
                likely to leach from the wastes in  harmful concen-
                trations  when the wastes are improperly managed.

           c.   It is estimated that a very  large quantity,  963,000
                tons  (l),  of ammonia still lime sludge  (5%  solids
                by weight)  is currently generated annually,  and
                that  this  quantity will  gradually increase  to 1.45
                million tons (5%  solids  by weight)  per  year  as the
                remaining  coke  plants  add  fixed ammonia removal
                capability to comply with  BPT limitations.
                There is thus the  likelihood of large-scale  con-
                tamination of the  environment if these wastes are
                not managed  properly.

          d.    Coke  plant operators generally  dispose of these
                sludges on-site in  unlined sludge lagoons or in
                unsecured  landfill  operations.  These management
                methods may  be  inadequate to impede leachate
                migration.
* Although no data on the corrosivity of ammonia still lime sludge
  are currently available, the Agency believes that these sludges
  may have a pH greater than 12.5 and may,  therefore,  be corrosive.
  Under §262.11, generators of this waste stream are responsible
  for testing their waste in order to determine whether their
  waste is corrosive or would meet any of the other hazardous
  waste characteristics.
                                 537

-------
     2.   Decanter tank tar sludge from coking operations (T) 7/16/80

          a.   The tank tar-sludge contains significant concen-
               trations of phenol and naphthalene.  Phenol is
               highly toxic.  Naphthalene is also toxic and is
               a demonstrated neoplastic substance in experiments
               done on laboratory animals.

          b.   Phenol has leached in significant concentration
               from a waste sample tested in a distilled water
               extraction procedure.  The Agency believes that,
               due to the presence of naphthalene in the tar in
               high concentrations and due to its relative
               solubility, napthalene also may leach from the
               waste in harmful concentrations if the waste is
               improperly managed.

          c.   These tar-sludges are often land disposed in on-
               slte landfills or dumped In the open.   These
               methods may be inadequate to impede leachate
               migration and resulting groundwater contamination.

B.   Steel Finishing

     1.   Spent Pickle Liquor (C, T) 5/19/80

          a.   Spent pickle liquor is corrosive (has been shown
               to have pH less than 2), and contains significant
               concentrations of the toxic heavy metals lead and
               chromium.

          b.   The toxic heavy metals in spent pickle liquor are
               present in highly mobile form,  since it is an
               acidic solution.  Therefore,  these hazardous con-
               stituents are readily available to migrate from
               the waste in harmful concentrations,  causing harm
               to the environment.

          c.   Current waste management practices of untreated
               spent pickle liquor consists  primarily of land
               disposal either in unlined landfills or unlined
               lagoons which may be inadequate to prevent the
               migration of lead and chromium to underground
               drinking sources.   Treatment  of the spent pickle
               liquor by neutralization is also commonly practiced
               by the industry in which case,  a lime treatment
               sludge is generated.

          d.   A very large quantity (approximately 1.4 billion
               gallons of spent pickle liquor)  annually.   Thus,
               there is greater likelihood of large-scale con-
               tamination of the environment if these wastes are
               not managed properly.
                                 538

-------
          e.   Damage  incidents have been reported that are
               attributable to the improper disposal of poorly
               treated spent pickle liquor.

     2.   Spent Pickle Liquor - Lime Treatment Sludge

          It should be noted that the waste "sludge from lime
     treatment of spent pickle .liquor from steel finishing opera-
     tion" has been removed from the list of hazardous wastes.
     Several comments  indicate that this waste may not be hazar
     dous, particularly if the lime treatment process is conducted
     effectively.  At the same time, however, insufficient data
     was submitted to warrant a conclusion that these wastes will
     typically and frequently not be hazardous.  Our concern is
     that these wastes derive from a hazardous waste {spent pickle
     liquor from steel finishing) which may contain high concen-
     trations of lead and chromium.  These heavy metals not only
     will be present in the sludge, but will be found there in
     even more concentrated form.

          Under these circumstances, we have decided that these
     waste sludges still should be regulated as hazardous,  but
     to delete these wastes from the hazardous waste list,  and
     instead to rely in the provisions of §261.3 to bring these
     wastes within the hazardous waste management system.  Since
     these lime treatment sludges are generated from the treat-
     ment of a listed hazardous waste (spent pickle liquor),
     they are considered to be hazardous wastes (§261.3(c)(2)).
     Further, they remain hazardous wastes until they no longer
     meet any of the characteristics of hazardous waste and are
     de-listed (§261.3(d)(2)).

C.   Electric Furnace Production of Steel

     1.   Emission Control dust/sludge from the primary production
          o£ steel in electric furnace (T) 5/19/80

          a.    The emission control dusts/sludges contain signi-
               ficant concentrations of the toxic metals chromium,
               lead,  and cadmium.

          b.    Lead,  chromium and cadmium have been shown to
               leach in harmful concentrations from waste samples
               subjected to both a distilled water extraction
               procedure and the extraction procedure described
               in the Subtitle C regulations.

          c.    A large quantity of these wastes (a combined total
               of approximately 337,000 metric tons)  is generated
               annually and is available for disposal.   There is
               thus likelihood of large scale contamination of
               the environment if these wastes are mismanaged.
                                 539

-------
          d.   The wastes typically are disposed of by being
               dumped in the open, either on-site or off-site,
               thus posing a realistic possibility of migration
               of lead,  cadmium, and chromium to underground
               drinking water sources.  These metals persist
               virtually indefinitely, presenting the serious
               threat of long-term contamination.

          e.   Off-site disposal of these wastes will increase
               the risk of mismanagement during transport.

     In addition to the specific source wastes listed above, a
number of other wastes (from non-specific sources) which may be
associated with some steel industry operations have also been
listed as hazardous wastes.  These wastes are:

     -    The following spent halogenated solvents used in de-
          greasing: tatrachloroethylene, trichloroethlene, methylene
          chloride, 1,1,1-trichloroethane, carbon tetrachloride,
          and chlorinated fluorocarbon; and sludges from the
          recovery of these solvents in degreasing operation.

     -    The following spent halogenated solvents:  tetrachloro-
          ethylene, methylene chloride, trichloroethylne, 1,1,1-
          trichloroethane, chlorobenzene, 1,1,2-trichloro-l,2,2-
          trifluoroethane, o-dichlorobenzene, and trychlorofluoro-
          methane; and the still bottoms from the recovery of
          these solvents.

     -    The following spent non-halogenated solvents:  xylene,
          acetone, ethyl acetate, ethyl benzene, ethyl ether,
          methyl isobutyl ketone, n-butyl alcohol, cyclohexanone
          and methanol;  and the still bottoms from the recovery
          of these solvents.

     -    The following spent non-halogenated solvents:  cresols
          and cresylic acid and nitrobenzene; and the still
          bottoms from the recovery of these solvents.

          The following spent non-halogenated solvents: toluene,
          methyl ethyl ketone, carbon disulfide, isobutanol and
          pyridine; and the still bottoms from the recovery of
          these solvents.

          Quenching bath sludge from oil baths from metal heat
          treating operations.

          Spent solutions from salt bath pot cleaning from metal
          heat treating operations.

          Quenching wastewater treatment sludges from metal heat
          treating operations.
                                  540

-------
     These wastes were  listed as hazardous due to toxicity,

ignitibility, or reactivity characteristics.  A brief listing

summary^ follows:

A.   WastesFrom Usage  of Halogenated Hydrocarbon Solvents in
     Degreasing Operations

     1.   The following spent halogenated solvents used in
          degreasing: tetrachloroethylene, methylene chloride,
          trichloroethylene, 1,1,1-trichloroethane, carbon tetra-
          chloride, and chlorinated fluorocarbons; and sludges
          resulting from the recovery of these solvents in de-
          greasing operations (T) - 5/19/80

     For all of the listed waste solvents, the listing is based
on the following considerations:

     a.   The chlorinated waste hydrocarbons are toxic and, in
          some cases, genetically harmful, while chlorofluoro-
          carbons may remove the ozone layer following environ-
          mental release.

     b.   Many steel facilities dispersed throughout the country
          use halogenated solvents and generate these wastes.
          Halogenated hydrocarbons from these facilities are
          either disposed of annually in landfills or by open-
          ground dumping, either as crude spent solvents or as
          sludges.   Current waste management practices have
          resulted in environmental damage.  Damage incidents
          serve to illustrate that the mismanagement of these
          wastes does occur and can result in substantial
          environmental and health hazards.

     c.   Since a large majority of the spent solvents and sludges
          are in liquid form and are highly soluble,  the potential
          for these wastes to migrate from land disposal facilities
          is high.   Spent halogenated solvents can leach from the
          waste to effect adversely human health and the environ-
          ment through the resulting contamination of groundwater.

B.   Wastes From Usage of Organic Solvents

     1.   The following spent halogenated solvents:  tetrachloro-
          ethylene,  methylene chloride,  trichloroethylene,
          1,1,i-trichloroethane, chlorobenzene,  1,1,2-trichloro-
          1,2,2-trifluoroethane, o-dichlorobenzene and trichloro-
          fluoromethane? and the still bottoms from the recovery
          of these  solvents 
-------
          methyl isobutyl ketone, n-butyl alcohol/ cyclohexanone
          and raethanol; and the still bottoms from the recovery
          of these solvents (I) - 5/19/80

     3.   The following spent non-halogeriated solvents:  cresols
          and cresyllc acid and nitrobenzene, and the still bottoms
          from the recovery of these solvents (T) - 5/19/80

     4.   The following spent non-halogenated solvents:  toluene,
          methyl ethyl Icetone, carbon disUlfide, isobutanol and
          pyridine; and the still bottoms from the rebovery of
          these solvents (I, T)

Wastes resulting from usage of organic solvents typically contain
significant concentrations of the solvent.  The basis for listing^
the above wastes as hazardous is:*

     a.   Each solvent exhibits one or more properties (i.e.,
          ignitability and/or toxicity) which pose a potential
          hazard.

     b.   The nine spent solvents listed for meeting only the
          ignitability characteristic all have a flash point
          below 60°C (140°F) and are thus considered hazardous.

               The solvents listed as either toxic or toxic and
          ignitable pose a further hazard to human health and the
          environment.  If improperly managed, these solvents
          could migrate from the disposal site into ground and
          surface waters and persist in the environment for
          extended periods of time.

               The two fluorocarbons, 1,1,2-trichloro-l,2,2-
          trifluoroethane and trichlorofluoromethanes present a
          different type of hazard.  Due to their high volatiliy,
          these two organics can rise into the stratosphere and
          deplete the ozone, leading to adverse health and environ-
          mental effects.

     c.   Damage incidents resulting from the mismanagement of
          waste solvents have been reported.  These damage
          incidents are of three types:
* The Agency is presently aware that these solvents may contain
  concentrations of additional toxic constituents listed in
  Appendix VIII of the regulations.  For purposes of this listing,
  however, the Agency is only listing those wastes for the presence
  of the halogenated and non-halogenated solvents.  The Agency
  expects to study these listings further to determine whether
  the waste solvent and still bottom listings should be amended.
                                  542

-------
     (1)  Fire/explosion damage resulting from ignition of
          the solvents;

     (2)  Contamination of wells in the vicinity of inade-
          quate waste storage or disposal (with resulting
          illness in at least one instance); and

     (3)  Direct entry of solvent into a waterway/ resulting
          in fish kills.

     These damage incidents show that mismanagement occurs
     and that substantial hazard to human health and the
     environment may result.

Spent Waste Cyanide Solutions and Sludges

1.   Quenching bath sludge from bath pot cleaning from metal
     heat treating operations (R, T) - 5/19/80

2.   Spent solutions from salt bath pot cleaning from metal
     treating operations (R, T) - 5/19/80

3.   Quenching wastewater treatment sludges from metal heat
     treating operations (T) 5/19/80

These wastes are considered hazardous based on the following:

a.   Each of the wastes generated exhibits either reactive
     or toxic properties or both due to their cyanide content,

b.   The land disposal of cyanide wastes containing high
     concentrations of cyanide is widespread throughout the
     United States.

c.   Cyanides can migrate from the wastes to adversely affect
     human health and the environment by the following path-
     ways/ all of which have occured in actual management
     practices:

     (1)  generation of cyanide gas resulting from the
          reactive nature of cyanide salts when mixed with
          acid wastes;

     (2)  contamination of soil and surface waters in the
          vicinity of improper waste disposal resulting in
          destruction of livestock/  wildlife, stream-dwelling
          organisms/  and local vegetation; and

     (3)  contamination of private wells and community
          drinking water supplies in the vicinity of improper
          waste  disposal.
                            543

-------
     These, then are the wastes, applicable to the steel industry,
which have been specifically listed by EPA as being hazardous.
This is not to say that these wastes are the only hazardous wastes
generated by the steel industry.  As was stated earlier, if the
generator thinks or knows that he has other hazardous waste, he
is required to report it to EPA for an identification number,
performing testing, if necessary, to make the determination of
hazardousness.

     These wastes, due to their being classified as hazardous, are
subject to the interim status provisions, as previously discussed.

Non-Hazardous Wastes

     Most of the solid waste generated by the steel industry will
probably be non-hazardous.  Wastes such as blast furnace slag,
EOF sludge, and blast furnace dust will likely fall into this
category.

     The types and quantities of non-hazardous wastes estimated
to be generated by the industry were shown in Table 1.  Over 50
million metric tons of non-hazardous waste are generated per
year.  After commercial sale and/or in plant recovery, about 19
million metric tons (approximately equals 40%) of non-hazardous
waste remain to be disposed.  The most common disposal practice
is to dump or landfill these wastes.  EPA promulgated criteria
for classifying non-hazardous waste disposal sites on September 13,
1979.  These regulations establish eight Federal criteria for
determining whether a disposal site is a "sanitary landfill" or
an "open dump".  The criteria include certain restrictions on
siting in floodplains,  contamination of surface and ground water,
land-spreading of wastes, and open burning.  Protection of
endangered species, protection against disease vectors and
explosive gases,  and bird hazards to aircraft are also addressed.

     If a facility is determined to be an open dump,  by virtue of
failing any one of the eight criteria, it must either be closed or
upgraded to the status of a sanitary landfill within five years.

     RCRA recognizes that prime responsibility for environmentally
sound disposal and resource recovery rests with state and local
government.  Each state will evaluate the individual disposal
sites,  establishing its own priorities for listing open dumps.
The states are required to develop a plan to identify a general
strategy for protecting public health and the environment from
adverse effects associated with solid waste disposal,  for
encouraging resource recovery and conservation,  and for providing
adequate disposal capacity in the state.   The Federal criteria
are the minimum requirements to be used in determining compliance.

     Thus, the requirements imposed by each of the states for
steel industry waste disposal facilities  will be the controlling
factor in determining whether a facility  will be permitted.   An
                                  544

-------
approved state program could be more stringent than the Federal
criteria but cannot be less stringent.  It would be wise to
ascertain the specific requirements of the states in which your
facilities are or will be located.

     Section 7002 of RCRA provides for citizen suits against the
operator of an open dump, which may be enforced in Federal district
court.  EPA does not specifically have the authority to take
legal action against parties that may violate the open dumping
prohibition.

Impact on Steel Industry

     The steel industry must dispose of over 6 million metric
tons of hazardous waste and over 50 million metric tons of non-
hazardous waste per year.

     Although the steel industry presently recycles/recovers a
significant portion of their wastes, a large quantity of wastes
still remains to be properly disposed of.  Up to this time, many
of the wastes have been improperly disposed.  A concerted effort
will be required by the steel industry to reverse this trend.

     Much of the steel industry is located in areas where land
for disposal is scarce and if available - quite costly.  Proper
management of the wastes will certainly dictate that the waste
management facility be designed and operated in the most effective
manner to meet both environmental and economic concerns.

     In a preliminary final economic impact analysis of Subtitle C
interim status hazardous waste regulations prepared in April of
this year,  the impact on the steel industry was determined.3
Tables 3 and 4 below illustrate the estimated impact.

     The estimated impact on the steel industry was made in light
of a number of assumptions/  some of which were:

     1.   The EIA was based on the 10 x drinking water standard
          (DWS)  rather than the promulgated 100 x DWS.

     2.   Analysis was made requiring much stricter closure
          requirements for landfills and surface impoundments
          than the promulgated regs.

     3.   Underground injection was not addressed in the Analysis.
                                 545

-------
                             Table 3
              INCREMENTAL COST TO THE STEEL INDUSTRY
                 OF RCRA INTERIM STATUS STANDARDS

         Cost of Waste Management at the Generator Plant
  Capital
($ Million)

    9.33
  1st Year
  Expenses
($ Million)

    4.56
  Annual
 Operating
($ Million)

   7.02
  Cost of
  Off-Site
  Transpor-
   tation
 (? Million)

    0.19
        Total
        Annual
         Cost
     ($ Million)

        11 02
          1978 Value
            Added
         ($ Million)

           46,000
             Annual Cost
             as a Percent
            of Value Added
              (Percent)

                 0.02
                             Table 4

              ECONOMIC IMPACT ON THE STEEL INDUSTRY
  Number of
   Existing
Plants in 1978

     152
      Plant
     Closures
    Job
   Losses
    negligible    negligible
   U.S.
Production
 Cutbacks

negligible
        Price
      Increases

      negligible
            U.S.
           Demand
          Reduct ion

          negligible
             Balance of
              Payments
               Effects

             negligible
                                  546

-------
      With regard non-hazardous  wastes, an  assessment  was made  (as
 presented In the EIS of  December  1979) on  the  cost  to the  steel
 industry to meet the "RCRA Section 4004  Criteria  for  Classification
 of  Solid Waste  Disposal  Facilities and Practices."4  A number  of
 assumptions,  including the following were  made:

      1.    50% of steel industry solid waste  is disposed onsite,
           50% off-site.

      2.    All non-hazardous steel  industry wastes will be
           disposed  in a  manner  which meets the criteria
           requirements.

      3.    Disposal  facilities will not be  relocated as a result
           of the criteria.

      4.    The groundwater  and floodplain criteria will probably
           have  the  greatest impact on the  industry.

      It  was estimated that the  total current annual disposal cost
 for  steel industry  non-hazardous wastes  is slightly over $50 mil-
 lion.  The annual cost of  complying with the criteria  was  estimated
 at approximately $36  million, a resultant  increase  of  about 72% in
 disposal costs.   Groundwater protection  accounts  for  the major
 portion  of these increased costs.

      It  must  be  remembered that a  significant  number  of the states
 in which steel  industry  disposal facilities are located already
 have regulations comparable to  the 4004  criteria.5  This is an
 important point  since implementation and enforcement of the dis-
 posal criteria  is the responsibility of  the individual  states.
 In fact,  probably 80% of the estimated expected $36 million
 increase for  non-hazardous waste disposal  can  be  attributed to
 comparable State regulations.   The remaining 20% of the increased
 costs can be  assumed  as  being induced by the Federal 4004  criteria.

     The overshadowing goal of  RCRA is resource recovery and
 conservation.   I am  sure that all  of you Tcnow  that the United
 States recovers  far  less of its industrial wastes than many other
 industrialized nations.  One of the major  contributing  factors
 to these low  recovery statistics is the  availability of cheap
 disposal options.   Proper  disposal  will  inevitably raise the
 coat of  disposal, thus increasing  the viability of recovering
 incremental amounts of waste materials.  The steel  industry is
 t  be applauded  on  its already high rate of (60-70%) waste re-
 cycling.   However, much  remains to be done in  this respect.
 Significant quantities of  solid waste that are presently disposed
of could be recycled, thereby not  only preventing the wasting of
valuable  natural  resources but  also extending  the life of  landfills
by decreasing the quantity of waste to be disposed and eliminating
potential  contamination  threats to our nations surface and ground
waters.
                                  547

-------
     EPA firmly believes that resource recovery should be imple-
mented whenever possible as an alternative to>land disposal of
solid wastes.  We wholeheartedly support and encourage studies
regarding the potential recovery/recycle of steel industry wastes.

     Examples of two such studies currently underway by EPA's
Office of Research and Development (ORD) are:

     l.   "Investigation of Toxic Substances during Recovery and
          Recycle of Steel Industry Iron Bearing Solid Waste"; and

     2.   "Uses for Ferrous Sulfate Heptahydrate from Steelmalcing
          Spent Pickle Liquor".

These studies will be discussed in detail this morning.  Although
the Office of Solid Waste is not presently sponsoring any studies
of the steel industry, we will be evaluating the work of others
and may issue industry specific waste recovery and disposal
guidelines in the future.

     In this era of environmental concerns and economic struggles,
it is hoped that industry and government will develop a cooperative
working relationship towards solving the problems faced by the
steel industry.
                                  548

-------
                             References
1.   Kline, William J.,  Solid Waste Generated by the  Iron  and
     Steel Industry, E.P.A. Office of Solid Waste,  December  1978.

2.   E.P.A. 3001 Listing Background Documents, May  1980.

3.   Preliminary Final Economic Impact Analysis (Regulatory
     Analysis Supplement) for Subtitle C, Resource  Conservation
     and Recovery Act of 1976 (RCRA), prepared by A.D.  Little,  Inc.
     for E.P.A. Office of Solid Waste, April 1980.

4.   EIS, Criteria for Classification of Solid Waste  Disposal
     Facilities and Practices, assessment on steel  industry by
     William J. Kline, May 1979.

5.   EIS, Criteria for Classfication of Solid Waste Disposal
     Facilities and Practices, JRB Assoc. estimates,  December 1979.

6.   E.P.A.,  Environmental and Resource Conservation  Considerations
     of Steel Industry Solid Waste, Research Triangle Institute,
     April 1979.
                                 549

-------
                  SPENT SULFURIC PICKLE LIQUOR RECOVERY

                    ALTERNATIVES AND BY-PRODUCT USES
                           Wayne C.  MIcheletti
                             Senior  Engineer
                             Peter  A.  Nassos
                             Staff  Scientist
                           Koren T.  Sherrill
                           Senior  Economist
                          Radian Corporation
                         8501 Mo-Pac Boulevard
                          Austin, Texas 78759
Each year, sulfuric acid pickling of steel produces approximately
600 million gallons of spent sulfuric pickle liquor (SSPL).
Currently, contract hauling is the most prevalent SSPL disposal
method.  Only a small portion of the SSPL is processed for sulfuric
acid recovery.  Resource recovery and environmental protection
objectives favor recovery instead of disposal.  Commercial recovery
processes involve separation of iron salts resulting in a relatively
pure ferrous sulfate by-product.   This paper summarizes a recent
study of alternate SSPL recovery technologies, the identification of
valid end uses for the ferrous sulfate by-product, and process economics,
                                  551

-------
                     SPENT SULFURIC PICKLE LIQUOR RECOVERY
                       ALTERNATIVES AND BY-PRODUCT USES
               W. C. Micheletti, P. A. Nassos, and K. T. Sherrill
INTRODUCTION
     Before steel products can be given a final surface coating or finish, the
surface must be cleaned of any scale and rust that may have formed due to
exposure to the atmosphere.  The process most commonly used in removing surface
scale and rust from iron and steel products is acid pickling.  The removal is
accomplished by immersing the scaled steel in tanks of hot dilute acid, such as
hydrochloric or sulfuric acid.  As the scale dissolves in the acid, iron salts
are formed and the pickling solution becomes ineffective.  Recovery or
disposal of the waste pickle liquor poses a potential environmental problem.
     Of the estimated 91 million tonnes (100 million tons) of steel shipped
in 1979, approximately 22 million tonnes (25 million tons) were pickled by
sulfuric acid.  This pickling operation produced an estimated 2.3 billion
liters (600 million gallons) of spent sulfuric acid pickle liquor.  Currently,
the acid is either recovered or disposed of by contract hauling, deep-well
injection,  neutralization/ponding,  discharge to a waterway or discharge to
a publicly-owned water treatment facility.   Spent pickle liquor is also in
limited direct use as a water treatment chemical.  Of all these methods,
acid recovery is very favorable based on resource recovery and environmental
protection objectives.
     Acid recovery processes involve the removal of iron as ferrous salt
crystals to regenerate the acid solution.  Recovered sulfuric acid can be
                                     552

-------
 returned to the pickling process  as  a  somewhat  dilute makeup.   The water-



 soluble ferrous salt by-product can  either be sold  for  its chemical value or




 disposed of in a secured landfill.   Although sale of the by-product is defi-




 nitely  the preferential  course of action, it depends primarily  upon the



 market  demand for this material including the manufacture of  products  such




 as  colored pigments  and  magnetic  tapes.  Furthermore, since acid recovery is




 not widely used at this  time, implementation of this process by a majority of




 the steel industry could produce  an  oversupply of the by-product.  Therefore,




 before  acid recovery can be encouraged as an environmentally acceptable method



 of  handling spent sulfuric acid pickle liquor, adequate by-product end use




 markets  must be identified and evaluated.




     The obj ectives  of this program were to evaluate the commercially avail-




 able processes  for recovering spent  sulfuric acid pickle liquor, estimate the




 quantity and quality of  by-product generated by these processes, and identify




 the current and potential end use markets for such by-products.   Information



 for this study  was obtained by thoroughly reviewing the available technical




literature,  visiting  steel mills with operating acid recovery processes,  meet-




 ing with representatives of the steel industry,  and contacting process vendors,




by-product  end  users, and other knowledgeable steel industry  personnel.






SULFURIC ACID PICKLING



     Sulfuric acid pickling of steel can be  either a batch or continuous




operation.   In  the case of wire or rod, the  steel is suspended from overhead




conveying racks and dipped into a  single vat of  pickle liquor for a designated




period of time.  For rolled sheet  steel,  the metal is continuously passed




through  a series of acid  vats, each successive tank containing a slightly morti




acidic pickle liquor.





                                     553

-------
     In sulfuric acid pickling, the scale is not only removed by acid dissolution



but also by the-effervescent action ef hydrogen gas'formation under the scale.




The sulfuric acid reacts with the iron base on the surface of the steel to




produce hydrogen gas and ferrous sulfate according to the following reaction:




                         Fe + H2SOi» •*• HZ+ + FeSOi*




The scale which is removed from the surface then slowly dissolves in the




pickling tank to produce additional ferrous sulfate as shown in the following




reaction:




                          FeO + H2SOi» .-»• FeSOit + H20




Increasing amounts of ferrous sulfate from continued use eventually consume




enough sulfuric acid to render the liquor ineffective for further pickling.




     Besides the presence of acid in the pickling tank, other chemicals may




be added to improve the quality of the pickled product.  Various organic




inhibitors are frequently used in pickling to inhibit acid attack on the base




metal, while permitting preferential attack on the scale.  In addition, wetting




agents may be used to improve the effective contact of the acid solution with




the metal surface.




     After pickling, the steel product is rinsed with water to remove any




adhering acid.  In some cases, the pickled product may then be coated with




lime, oil or some other material to protect it from exposure to the atmosphere.




     A recent EPA survey estimates that in 1979 sulfuric acid pickling accounted




for about 42.9% of all the steel products pickled in the United States.1




Pickling with hydrochloric acid and mixed acids accounted for the remaining




45.8% and 11.3%, respectively.  According to this same survey, approximately




105 of 133 major steel picklers use sulfuric acid in batch systems while 44
                                     554

-------
use sulfuric acid  in continuous systems.  Hydrochloric acid is used by 6



plants in batch systems and 41 plants in continuous systems.1




     In the 1960's, the trend in pickling was to switch from sulfuric acid




to hydrochloric acid.  This conversion occurred for several reasons.  At




the time, hydrochloric acid was readily available and relatively inexpensive.




In addition, pickling with HC1 is somewhat faster than using sulfuric acid




and the pickling solution retains its effectiveness longer.  Furthermore,




hydrochloric acid  is considered to give a "brighter" finish.  Recently,




however, the conversion from sulfuric acid to hydrochloric acid has




apparently ended for two reasons.  First, the price of hydrochloric acid



has increased such that it is no longer economically attractive.   And




second, fumes generated during HC1 pickling are a difficult and expensive




air pollution control problem.  Therefore, the percentage of steel products




being pickled by sulfuric acid will probably remain the same in the near term.






PICKLING WASTES



     Waste pickle liquor is generated when the pickling solution becomes




saturated with iron salts.  The pickle liquor is no longer effective in re-




moving scale and, consequently, must be replaced.  In addition to free sul-




furic acid and iron salts (typically 8.0 wt % each), the spent pickle liquor




may contain varying amounts of lubricants, suspended solids, heavy metals,




and additives such as inhibitors.  Table 1 presents the overall range of




compositions determined from the analyses of several waste pickle liquor




samples.1  Since a variety of chemical additives may be used in pickling




operations,  the quantity and composition of these additives in waste pickle




liquor is not reflected in Table 1.   However, the presence of these additives




is also of environmental concern in the handling of waste pickle liquor.




                                     555

-------
Table 1.  Analyses of sulfuric acid pickling wastes (ng/1) * '




Parameter	Spent Pickle Liquor    Rinse Water	Scrubber Blowdown
Dissolved Iron
Oil and Grease
Suspended Solids
Arsenic
Cadmium
Chromium
Copper
Cyanide
Lead
Nickel
Silver
Zinc
38,750-66,500
8-35
236-2363
<0.1
<0.6
6-269
£4.7
ND
£10
6.8-27
0.28-0.59
0.7-244
36-2900
£22
£750
£0.33
£0.13
£3.8
£10.4
£0.1
<2.0
<£.6
£0.85
£59
£305
2-30
2-200









ND -  indicates that the component was not detected.






     The amount of spent pickle liquor generated will depend on the original




quality of the pickle liquor and the amount of surface scale on the steel




product.  Based on a 1.0% iron loss, pickling 0.9 tonnes (1.0 ton) of steel




will generate 95 liters (25 gallons) of spent pickle liquor.  Assuming 22



million tonnes (25 million tons) of steel were pickled by sulfuric acid in




1979, then the resulting volume of waste pickle liquor was approximately



2.3 billion liters (600 million gallons) containing 220,000 tonnes (250,000 tons)




each of free acid and dissolved iron.3  These values illustrate the signi-



ficant potential for sulfuric acid recovery and by-product manufacture.




     In addition to spent pickle liquor, steel pickling processes may generate




two other wastewater streams:  acidified rinse water and scrubber blowdown.




Acidified rinse water results from washing the pickled product to remove any




adhering pickle liquor.  Methods of rinsing may vary from a single-stage immer-




sion to a'multi-stage system.  Many steel companies are switching from rinse




systems that flush the steel with large quantities of water to systems that




use fine sprays.  Depending on the overall water balance for the pickling
                                     556

-------
 operation,  rinse water can frequently be  used as makeup  for  the  pickling
 tanks.   Table 1 presents  the  overall range of compositions determined  from
 the analyses of several rinse water samples.
      Scrubber blowdown results  from emission control  equipment to collect
 and absorb  acid fumes  and mists emitted from the pickling tanks.  A wet
 scrubber typically recylces water with a  small purge  or  blowdown stream to
 control  the levels of  sulfuric  acid.  Under efficient operation, the
 scrubber may achieve less than  3% blowdown, which can be used as makeup to
 either  the  rinse system or the  pickling tanks.  Table 1  presents the over-
 all range of compositions determined from the analyses of several scrubber
 blowdown samples.

 WASTE PICKLE LIQUOR DISPOSAL  TECHNIQUES
      Currently,  waste  pickle  liquor disposal includes
      •   contract hauling,
      •   deep-well  injection,
      •   neutralization/ponding,
      •   discharge  to a waterway,
      •   discharge  to a publicly-owned water treatment facility,
      •   direct use as a water treatment chemical,  and
      •   acid recovery.

The three most commonly used disposal  techniques  are contract hauling,  deep-
well  injection, and neutralization.
     Contract hauling has  long been  a  favorable means of  handling waste
pickle liquor because of moderate operating  cost and little or no capital
cost to the steel industry. However,  the  growing  cost of energy  for
                                    557

-------
transportation is being reflected by an increased operating cost for dis-




posal by contract hauling.  Furthermore, the operating cost for contract




hauling can be expected to increase even more as regulatory agencies enact




stricter controls on disposal.  From 1970 to 1980, the cost of contract




hauling has risen approximately 340%, from 0.9C/liter (3.5£/gallon) to




3.2c/liter (12c/gallon).*'5




     A second popular method of waste pickle liquor disposal is deep-well




injection.  However, this disposal method is limited to favorable subsurface




geological formations that will protect local groundwater from contamination.




Current estimates indicate that only a dozen or so wells in the U.S. are used




for the disposal of waste pickle liquor, with a majority of these wells being




located in North Central Illinois and Northwestern Indiana.6'7  Although




deep-well injection has been used for many years,  concern over  the




potential for groundwater contamination may mean stricter regula-




tory control and possibly an end to this method of disposal.




     Neutralization of waste pickle liquor with lime, soda ash or caustic




soda has been an established practice for some time.  Addition of these




chemicals increases the initially acidic pH of the pickle liquor to a neutral




level.  The increasing pH causes the iron to precipitate as a gelatinous iron




hydroxide sludge which settles very slowly.  Hence, the neutralized mixture




is placed in a pond where it can be-contained indefinitely.   The cost of




waste pickle liquor disposal by neutralization/ponding has been steadily




increasing over the past several years.  The increasing cost primarily reflects




the rapidly rising cost of chemicals, particularly lime, and the inherent value




of the land required for lagooning.
                                    558

-------
     Each of  the  three previously described disposal methods for waste pickle




liquor are currently in widespread use.  None of these disposal techniques




makes any attempt to recover the acid or dissolved iron.  In some in-




stances, these methods may be trading one type of environmental problem for




another.  On  the  other hand, the use of acid recovery units results in the



recovery of  sulfuric acid and an iron salt by-product,.and virtually




eliminates water  pollution associated with sulfuric acid pickling operations.







ACID RECOVERY PROCESSES




     Two types of acid recovery processes are commercially available: high




temperature and low temperature.  High temperature processes heat the waste




pickle liquor to  about 93°C (200°F) and produce a ferrous sulfate monohydrate




(FSM) precipitate.  These processes are not widely used because of the




associated high energy costs and the problems of ferrous sulfate monohy-




drate scaling.  At present, two such acid recovery processes are in




operation in  the  U.S.:   A Pureco unit at Wilson Steel and Wire in Chicago  .




and a Sulfex unit at Metal Processing Company in Maple Heights, Ohio.




     Low temperature acid recovery processes cool the waste pickle liquor to




about 15°C (45°F) and produce a ferrous sulfate heptahydrate (FSH)  precipi-




tate.  Commercially, three low temperature processes are available in the




U.S.:  the Kerachemie,  the Crown,  and the KSF processes.   The Crown and KSF




processes are the most  widely accepted of the low temperature acid recovery




processes.   One Kerachemie unit has been installed in the United States




(Fitzsimmons Steel Company,  Youngstown,  Ohio),  but it is  no longer in use




at the plant.
                                    559

-------
      Both the Crown and the KSF acid recovery processes are modular batch-




 type units.  Each unit consists of a chiller/crystallizer, a slurry separator,




 and separate storage for recovered acid and by-product crystals.  Waste




 pickle liquor is fed to the crystallizer and cooled by submerged chilling




 coils.  The Crown process lowers the temperature of the waste pickle liquor




 to about 2-10°C (35-50°F) by circulating freon refrigerant through Teflon




 cooling coils.  The KSF process uses chilled water to achieve a temperature




 of about 7-10°C (45-50°F).  As the temperature of the waste pickle liquor




 decreases,  ferrous sulfate heptahydrate (FeSOij'THaO)  crystals begin to form.




 In both processes, a motor-driven agitator stirs the acid/crystal slurry to




 maintain a  uniform temperature and to prevent the crystals from agglomerat-




 ing into large chunks.




     The slurry is pumped from the chiller/crystallizer to a separation unit




for removal of the ferrous sulfate heptahydrate crystals.  The recovered acid




can either be recycled to the chiller for additional processing or stored for




future acid makeup to the pickling operations.  The by-product FSH crystals




can be washed to remove any sulfuric acid adhering to the surface.  The wash




water is recycled to the chiller for additional processing.  As .indicated by




Table 2, the by-product crystals are relatively pure ferrous sulfate hepta-




hydrate and can be sold as a source of iron for several uses.5  Since the FSH




crystals are water-soluble, its storage should be protected from the




weather.
                                     560

-------
Table 2.  Analysis* of ferrous sulfate heptahydrate by-product (wt.  %).5»
                                                                       5 8

FeS
-------
 in operation in West Germany for several years.  Although this method of

 ownership reduces the amount of capital any one company must invest in an

 acid recovery system, the approach also has disadvantages.


 Table 3.   Economic comparison of waste pickle liquor treatment alternatives.1*
                            (Total Annual Costs, $1,000)

Item         	Acid Recovery  Neutralization   Contract Hauling
Investment
Salaries & Wages
Operators
Foremen
Utilities
Steam
Process Water
Electricity
Raw Materials
H2SO.»
CaO
Shipping & Hauling Costs
Crystals
Sludge
Pickle Liquor
Maintenance
General Plant Overhead
Wastewater Costs
Sewer Fees
pH Adjustment
Taxes and Insurance
Depreciation
FSH By-Ptoduct Credit
Total Annual Costs
630.0

12.5
1.5

35.4
(8.1)
19.0

(50.0)
0

28.8
0
0
37.8
36.0

0
0
3.2
63.0
(52.0)
126.1
770.0

12.5
3.8

4.2
-
4.0

-
68.7

0
110.0
0
46.2
41.7

9.2
0
3.9
77.0
0
381.2
0

6.2
0

_
—
-

-
0

0
0
350.0
-
. 7.8

10.0
62.5
0
0
0
: 436.5
Rinse Water included in process water for acid recovery plants at 25 gal/ton
  pickled steel.
Basis:  100,000 tons/yrof steel pickled; 1% iron loss; spent pickle liquor
        composition 8% dissolved iron, 8% HaSOi* all figures-in thousands
        of dollars per year (1976 base).
                                     562

-------
      One of the major disadvantages associated with regional acid recovery




 systems is transportation of the waste pickle liquor to  the  facility.




 Currently, waste pickle liquor is classified  as a  hazardous  material.   There-




 fore,  its transport may be highly limited  or  even  restricted in  certain




 areas  and would be subject to RCRA regulations with attendant costs.  Also,




 the rising cost of fuel means that transporting large quantities  of waste




 pickle could become fairly expensive.   This factor is already being re-




 flected in the increasing cost of contract hauling.   Transportation and




 storage of waste pickle liquor could be further complicated  by premature




 precipitation of ferrous sulfate heptahydrate  under  cold climatic  condi-




 tions.   Heated or insulated  vehicles and storage vessels would eliminate




 this potential problem,  but  would also  increase associated costs.




     Another  problem that may  be encountered with  regional facilities is




 the need to segregate waste  pickle liquor by source.  Since many steel




 picklers use  proprietary additives in the pickling operations, the waste




 pickle  liquor  from  each plant will require separate processing.  The use of




 such additives makes  it essential  that a plant receive acid recovered from




 its own  pickle liquor because some additives may cause adverse effects in




 a different pickling operation.  Therefore, adequate pickle liquor and




 recovered acid storage governed by a strict accounting procedure tracking




 source and final destination will be required.  In  order  to maintain




 flexibility for dumping ineffective batches,  it will also be  necessary for




steel picklers to have on-site storage for  waste pickle liquor and makeup




acid.
                                    563

-------
     Individually, none of these difficulties is insurmountable; collectively




the disadvantages of relying on a regional acid recovery facility appear to




overshadow the advantage of reduced capital cost to the pickling plants.






BY-PRODUCT MARKET




     Although the economics of acid recovery is not strongly dependent on the




sale of by-product ferrous sulfate heptahydrate (FSH), proper use of the




water-soluble crystals is essential for acid recovery to be considered an




environmentally acceptable method of handling waste pickle liquor.  There-




fore, a major focus of this study was to investigate the sources of FSH pro-




duction and to identify and evaluate current and potential end use markets




for FSH.




      Comprehensive data for annual production,  consumption,  and prices for




 FSH are not available.  However, rough estimates  can be made using informa-




 tion from various sources.  The current annual U.S. capacity for FSH pro-




 duction is about 363,000 tonnes (407,000 tons).  Of this amount, approxi-




 mately 296,000 tonnes (332,000 tons) is manufactured by eight chemical




 companies which act as commercial producers for the U.S. market.9  Yet,




 as much as 98,000 tonnes (110,000 tons) of this capacity may soon be removed




 from service for various reasons.10  At present,  acid recovery processes




 for pickling operations have the capacity to produce 67,000  tonnes




 (75,000 tons) of FSH, which represents only 18% of the total U.S.  capacity.11



 In addition to current domestic FSH production capacity, an  estimated 93,000




 tonnes (104,000 tons) is imported from West Germany,  Japan and Mexico.10




 These figures indicate that the U.S. market for FSH is strong enough for




 imports to absorb the cost of transportation and  still be sold at a profit.
                                     564

-------
     The market price for ferrous sulfate heptahydrate can vary considerably.




The current market price for moist  (having some surface water) FSH is about




$19 per tonne  ($17 per ton).10  However, in the Midwest, the market price




for moist FSH by-product from acid recovery processes ranges from $2 to $50




per tonne ($2 to $45 per ton).12  Imports from West Germany are usually




transported up the Mississippi River by barge and marketed in the Midwest for




approximately $34 per tonne ($30 per ton) . l °  Dry ferrous sulfate heptahydrate




has had the surface moisture removed by a moderate heating process and cur-




rently sells for $101 per tonne ($90 per ton).10  Ferrous sulfate heptahydrate




can also be converted to the monohydrate form by heating.  Ferrous sulfate




monohydrate currently sells for $190 to $224 per tonne ($170 to $200 per ton).10




     At present, FSH is used almost exclusively as a source of synthetic iron




oxide for the manufacture of pigments,  ferrites and magnetic tapes,  fertilizers




and animal feed, and catalysts and for water and sewage treatment.  As Table 4




indicates, the two major end uses are colored pigments and magnetic  tapes,




accounting for 45% and 35%,  respectively, of the total .FSH consumption in
Table 4.  Ferrous sulfate consumption by end use, 1978 and 1972. 13s11*

Iron oxide pigments
Magnetic tapes and ferrites
Fertilizers and stockfeed
Water and sewage treatment
Catalysts
Miscellaneous
Total
1978
45%
35%
8%
5%
3%
4%
100%
1972
45%
30%
12%
5%
3%
5%
100%
                                     565

-------
     In recent years the pigment industry has been a large consumer of ferrous




sulfate heptahydrate as a source of iron oxides.  In 1974, 121,000 tonnes




(135,000 tons) of iron oxides were consumed in the production of colored




pigments.15  Of this amount, slightly more than half (53.3%) was supplied




by synthetic (by-product) oxides as opposed to natural oxides derived from




pulverized iron ore and pyrite cinders.  Although the cost of synthetic oxides




is approximately three to four times greater than natural oxides, the syn-




thetic oxides are preferred because they provide a wider range of colors and




brillance.  Furthermore, synthetic oxides function well in water-based paints




and many natural oxides do not.  This factor is important in that a growing




trend to water-based paints as a means of reducing atmospheric solvent emis-




sions will probably mean an increased demand for synthetic iron oxides.




     The use of ferrous sulfate heptahydrate as a source of iron oxides for




magnetic tape manufacturing has been steadily increasing.  Currently, only a




few companies produce synthetic iron oxide for magnetic recording.  Yet, the




future demand for ferrous sulfate as a raw material in magnetic tape manufac-




turing should be strong.  This prediction is based on the fact that the demand




for magnetic tapes is closely associated with the high technology electronics




industry, which has been experiencing consistently rapid growth during the




last two decades.  Similarly, the demand for hard and soft ferrites in'




electronics should parallel the growth of the industry.  Iron oxides recovered




from the by-products of waste pickle liquor processing have been used as a




raw material in producing hard ferrites which are used in permanent magnets.15




     Although water and sewage treatment have typically only accounted for 5%




of the annual FSH consumption, this particular market represents the greatest
                                     566

-------
area of potential use  in  the near future.  Ferrous sulfate heptahydrate is




used as a coagulant in the treatment of drinking water, ass an additive for




sludge fixation, and as an agent for phosphorus removal in municipal waste-




water treatment.  The  key potential use for FSH is as an agent for phosphorus




removal.




      Phosphorus control  is considered  critical  in  some bays,  coastal  areas,




and drainage basins of lakes.   It is especially critical  in  the  drainage




basins of  the North American Great Lakes, which contain approximately 20% of




the world's  supply of  surface  fresh water.  The International Joint Commission




 (IJC) of Canada and the United States  established  a  program  in 1978 to mini-




mize  eutrophication problems in the Great Lakes by reducing  phosphorus inputs.




Consequently, increasing demand for chemicals used in phosphorus removal




can be expected in the eight states along the Great Lakes.  Six of these




states are also among  the top ten steel producing states.   Hence, pickling




plants in this area which use acid recovery could possibly have a fairly




substantial local market for the by-product ferrous sulfate heptahydrate.




     Currently, several chemicals are being used to remove phosphorus  from




municipal wastewater,   including aluminum sulfate (alum),  sodium aluminate,




ferric chloride, ferrous chloride,  ferric sulfate,  and lime.   Of all these




chemicals,  alum and ferric chloride are the most widely used.  In the lower




Great Lakes basins,  iron salts and aluminum salts equally share about 99%  of




the chemical market for phosphorus  control.16   Data concerning the types of




iron salts (e.g., ferric chloride,  ferrous chloride,  ferrous  sulfate)  used




are not  available.
                                     567

-------
     Although alum and ferric chloride are widely available and well known




to wastewater treatment plant designers and operators, ferrous sulfate can



compete both technically and economically with both of these chemicals.




Technically, ferrous sulfate can reduce effluent phosphorus as effectively




as commonly used chemicals.  Based on FSH chemical analyses (Table 2) intro-




duction of other pollutants should be no more of a problem than with other




chemicals.  Other concerns, such as pH changes and metal leakage, are common




to most phosphorus precipitating chemicals and will depend in some measure




on the proper operation of the treatment facility.  Furthermore, storage, feed




and treatment equipment should not vary significantly for different chemicals,



so that minimum modifications will be required to switch from alum or ferric




chloride to ferrous sulfate.




     Economically, ferrous sulfate is a very attractive alternative to alum




and ferric chloride.  Since utilities (electricity), operator time and other




operating costs are typically low relative to raw material costs, the economics




of phosphorus removal is basically governed by the cost of the treatment




chemical.  Table 5 presents an economic comparison of phosphorus removal by




different chemicals (exclusive of transportation costs).  This comparison




indicates that use of ferrous sulfate heptahydrate has a reasonable economic




advantage over the use of ferric chloride and a significant economic advantage




over the use of alum.  However, transportation may add significantly to the




cost of chemicals, and thus the shipping distance may be the primary factor




in chemical selection.  For states bordering the Great Lakes,  transportation




costs for ferrous sulfate heptahydrate from acid recovery processes should



be minimal due to the proximity of the pickling operations.
                                     568

-------
 Table 5.  Economic comparison of  phosphorus removal by different  chemicals.



                          Requirement per Unit ot Phosphorus Removed  Chemical Cost  Phosphorus Removal CTSL
Chemical
Alum - A12(SOJ3'14
Dry - 9 wt % Al
Liquid - 4.4 wt
Sodium Aluminate ~
Dry - 45 wt Z Al
Liquid - 26 wt X
kg/kg or Ib/lb liter/kg (gal/lb) S/ tonne ($/ton) $/k?
.3HjO
ion
X Al ion
Na2Al2Oi.
203
A1203

22.2 119
34.2 (4.1) 100

8.4 816
10.0 (1.2) 272

(120)
(110)

(900)
(300)

2
5

8
4

.93
.51

.33
.96
'$/lb;

(1
(2

(3
(2

.33,
.50)

.78)
.25)
Ferric Chloride - Feds
Liquid - 40 wt %
Ferric Sulfate - Fe
Dry - 19.5 wt X
FeCl3
2(SOi,)3'7H20
Fe3*
Ferrous Sulfate Heptahydrate - FeSO\'7H
2+
Dry - 20 wt X Fe
10.0 (1.2) 96

10.3 106
:2o
10.0 68
(106)

(117)
(75)
1

1
0
.68

.32
.84
(0

(0
(0
.76)

.60)
.38)
Basis: 1) Metal ion to phosphorus removed weight ratio of 2.0 to 1.0 for all metal ions
     2) 10.0 mg/1 phosphorus inlet concentration
     3) <1.0 mg/1 phosphorus effluent concentration
     4) Phosphorus removal with secondary treatment




      Estimating  the  potential market  for FSH by-product  in the municipal


wastewater treatment sector is difficult.  Based on  a wastewater  flow of


380 liters (100  gallons)  per capita  per day and a  phosphorus removal of


9.5 mg/1, then the potential market  for ferrous sulfate  heptahydrate in the


Great Lakes States is estimated to be 1.0 million  tonnes (1.12 million tons)


in 1985.  If  all U.S.  sulfuric acid  pickling operations  practiced  acid recovery


in 1985, the  estimated FSH by-product generated would be almost 740,000 tonnes


(830,000 tons).   Therefore, the FSH by-product generated by all acid recovery


sulfuric acid pickling operations in  the U.S. (ignoring  other FSH  end use


demands) would only  be able to supply 73.2% of the total demand for


municipal wastewater treatment in the Great Lakes  area.   If ferrous  sulfate


heptahydrate  replaced alum and ferric chloride as  the most popular chemical


agent for controlling phosphorus  in wastewater effluent, the demand  for


FSH could easily exceed the supply.
                                        569

-------
CONCLUSIONS




     Acid recovery is economically competitive with contract hauling and




neutralization.  A major portion of the overall annual cost for acid re-




covery is due to capital investment.  Using regional acid recovery




facilities to treat waste pickle liquor from several local plants is a




potential means of reducing the capital investment for any of the partici-




pating plants.  However, these regional facilities do not appear practical




for three reasons.  First, transportation costs will be excessive because




of the large volumes involved and the potential for premature iron salt




precipitation.  Second, processing costs will be increased by the need to




segregate waste  pickle liquor and recovered acid by company in order to




prevent recovered acid contamination from different proprietary chemical




additives.  Finally, the cost to each steel mill is increased by the need




to have waste pickle liquor and recovered acid storage facilities on-site




in order to maintain flexibility with regard to spent acid dumping.




     The demand for by-product FSH as a raw material for the production




of iron oxide pigment could increase more dramatically than historically




indicated.  The cause of this sudden potential increase is two-fold.  First,




almost 30% of the current U.S. FSH production capacity will be removed from




service for various reasons in the near future.  Second, the need to reduce




fugitive emissions from painting operations will mean an increase in the




use of water-based paints which require pigments produced from synthetic




sources such as by-product FSH.  Combined, these two factors indicate a




much stronger demand for by-product FSH in future iron oxide pigment




production.
                                    570

-------
     The potential impact of the municipal wastewater treatment market




as an end use for FSH is difficu.il to assess.  Rough csLlmatcs I.nd lc:,-i Lc




this potential to be much greater than any of the current demands for KSH.




For instance, if all U.S. sulfuric acid pickling operations practiced acid




recovery in 1985, the estimated amount of by-product FSH generated would be




740,000 tonnes (828,000 tons).   If all eight of the Great Lake states used




strictly FSH instead of alum or ferric chloride for phosphate removal,  the




total estimated FSH required in 1985 would be almost 1,000,000 tonnes




(1,130,000 tons).  Therefore, FSH from acid recovery would only be able to




supply 73.2% of the total demand for municipal wastewater treatment in that




area alone.  Clearly, the potential market for by-product FSH from waste




pickle liquor recovery is considerable.






RECOMMENDATIONS




     Although acid recovery represents the only method of recovering and




reusing the chemical constituents found in waste pickle liquor, only 10%




of the total spent sulfuric acid pickle liquor generated each year is




treated by acid recovery.  Since pickle liquor recovery is currently not a




widespread practice,  additional studies should be undertaken to evaluate




the influence of spent pickle liquor composition on the quality of FSH




by-product and recovered acid.   Experience to date indicates that product




quality has not usually been a  problem.   However,  each pickling operation




is a unique case and  it may be  that under certain conditions,  recovered




acid may be unsuitable for reuse,  and/or FSH by-product may be unsuitable




for sale.   Therefore,  better correlation of data pertaining to pickle liquor




composition,  recovery process operation, and acid and by-product  quality




is needed.






                                    571

-------
    The reuse of by-product FSH is a major factor in assessing the




environmental advantages of using acid recovery.  Currently, the demand for




FSH exceeds supply.  Widespread use of acid recovery by sulfuric acid




picklers could reverse this situation.  However, the use of FSH as a




chemical agent for controlling phosphorus levels in municipal wastewater




appears to be a largely untapped market.  Although limited use of FSH in




wastewater treatment indicates very satisfactory removal of phosphorus, use




at other wastewater treatment plants will probably involve evaluation on a




case-by-case basis to assess the true removal effectiveness under a number




of different conditions.  This is an important factor in determining if FSH




can displace alum and ferric chloride as the primary chemical for phosphorus




control.  Such an investigation might begin with a comparison of data from




previous applications under similar conditions and continue with bench-scale




laboratory studies or on-site pilot plant testing.
                                    572

-------
 REFERENCES

 1.  U.S.  Environmental Protection Agency.  "Development Document for Proposed
     Effluent Limitations, Guidelines, and Standards for Lhe Iron and Steel
     Manufacturing Point Source Category."  Draft, Volume VIII, EPA-440/1-79/
     024a, October 1979.

 2.  U.S.  Environmental Protection Agency,  Interagency Memo from
     L. G. Twidwell to J. S. Ruppersberger.  June 11, 1979.

 3.  Knook, P. R.  "Analysis of the Use of Waste Pickle Liquor for Phosphorus
     Removal."  Whitman, Rezuardt, and Associates Engineers.  Baltimore, MD,
     September 1978.

 4.  EPA Technology Transfer Capsule Report.  "Recovery of Spent Sulfuric Acid
     from  Steel Pickling Operations."  EPA-625/2-78-017, 1978.

 5.  Personal Communication with J. C. Petterson.  Crown Chemical.  July 7,
     1980.

 6.  Lackner, R. J.  "Acid Recycling Systems for Pickling Lines."  Presented
     to the Association of Iron and Steel Engineers Youngstown District
     Section.  Givard, Ohio, May 6, 1974.

 7.  Bayazeed, A. F., and E. C. Donaldson.  "Sub-Surface Disposal of Pickle
     Liquor."  R. I.  7804, U.S. Bureau of Mines, Washington, B.C., 1973.

 8.  Personal Communication with R. J. Lackner.   Wean United, Inc.

 9.  SRI International.  Directory of Chemical Producers, 1980, United States
     of America.  SRI International, Menlo Park, California, 1980.

10.  Personal Communication with Don Gordon,  Quality Chemical,  Ltd.,
     September 1980.

11.  Bhattacharyya, S.  Steel Industry Pickling Waste Ferrous Sulfate
     Heptahydrate and Its Impact on Environment.  Unpublished report  prepared
     for U.S. Environmental Protection Agency,  Office of Research and
     Development, Washington, D.C., 1979.

12.  Telephone survey conducted by K.  T.  Sherrill,  Radian Corporation,
     Austin,  Texas, September 1980.

13.  Chemical Marketing Reporter.   Profile:   Ferrous Sulfate.   January 1,
     1979.

14.  Chemical Marketing Reporter.   Profile:   Ferrous Sulfate.   October 23,
     1972.
                                     573

-------
REFERENCES (continued)

15.  Jones, Thomas D.   Iron Oxide Pigments,  Part I.   U.S.  Department of  the
     Interior, Bureau of Mines.   U.S.  Government Printing  Office,  Washington,
     B.C., 1978.

16.  De Pinto, Joseph V., et al.   Phosphorus Removal in Lower Great Lakes
     Municipal Treatment Plants.   Unpublished report.   U.  S.  Environmental
     Protection Agency,  Municipal Environmental Research Laboratory,
     Cincinnati, Ohio, n.d.
                                    574

-------
ACKNOWLEDGEMENT
The authors of this paper would like to acknowledge Dr.  S.  Bhattacharyya for




his unpublished work in the area of spent pickle liquor  recovery processes.




Dr. Bhattacharyya?s previous work served as the starting point and provided




the direction for the work presented in this paper.
                                   575

-------
             ENVIRONMENTAL APPRAISAL OF RECLAMATION PROCESSES
               FOR STEEL  INDUSTRY,  IRON-BEARING SOLID WASTE
                A. 0. Hoffman, E. J. Mezey, J. Varga, Jr.
                   W. G. Steedman, and R. D. Tenaglia
                                BATTELLE
                          Columbus Laboratories
                          Columbus, Ohio  43201
                                ABSTRACT

          The objective of this study was to investigate existing and
emerging processes for the reclamation of the three largest quantities of
iron-bearing solid wastes being landfilled by the steel industry—oily
mill scale, steelmaking dust, and blast furnace dust and sludge.  The processes
considered are designed to remove contaminants to the degree that resource
recovery (recycle) can be practiced.  This paper summarizes the results of
this study in terms of process identification and description, appraisal
of the environmental aspects of reclamation including the appraisal of the
potential for new environmental problems,  and considers the economics and
energy requirements of the various processes.
                                   577

-------
             ENVIRONMENTAL APPRAISAL OF RECLAMATION PROCESSES
               FOR STEEL INDUSTRY, IRON-BEARING SOLID WASTE
                                    by
                 A. 0. Hoffman, E. J. Mezey, J. Varga, Jr.
                   W. 6. Steedman, and R. D. Tenaglia
                      Battelle Columbus Laboratories
                           Columbus, Ohio  43201
           In September of 1980, the Steel Tripartite Committee  ' report
was released on Technological Research and Development in the Steel Industry.
Pertinent to this study are a recommendation and a statement in the
Tripartite report.  The recommendation was that "government support should
be given to the development of technologies and plant practices for
recycling hazardous wastes produced in steel manufacturing".  The statement
was that "environmental and occupational safety and health issues should
be considered as an integral part of technological research and development
in the steel industry..."
           This paper deals with an environmental appraisal of reclamation
processes for steel industry, iron-bearing solid waste.  This appraisal
study was funded by the EPA's Industrial Environmental Research Laboratory
(IERL-RTP) and as an appraisal it is in line with both stated sentiments
of the R & D Group of the Tripartite Committee.

TECHNICAL OBJECTIVE AND SCOPE OF THE STUDY

           The overall objective of this study was to investigate existing
and emerging processes for the reclamation of steel industry iron-
bearing wastes being landfilled.  These reclamation processes should
be capable of extracting and/or eliminating undesirable contaminants to
                                     578

-------
the degree that the processed material is acceptable for recycle.  If this
were possible in operations that did not introduce new pollution problems,
it would result in a conservation of resources and a decrease in problems
pertaining to waste management.  The term "investigate" in this instance
included:
           (1)  Identification and description of processes including
                capability and mass and energy balances, where available
           (2)  Appraisal of the potential for new environmental problems
                that may be part of these existing and emerging processes
           (3)  Appraisal of the overall economics of reclaiming and/or
                using approved landfilling for contaminated solid wastes
           (4)  General ranking of the processes including identification
                of any remaining environmental problems.

Definitions and Listing of the Types of Wastes of Interest in This Study

           The U.S. steel industry routinely recycles about 80 percent
                    (2\
of its solid wastes.     The remaining 20 percent is not recycled because
the material is either (a) nearly worthless, e.g., trash, rubble, and some
slags, or (b) although valuable, the material is contaminated with troublesome
elements from either an operational or product viewpoint.
           One viewpoint of metallurgical processing is that it is the
science/art of separating (by many routes) the desired element(s) from
the undesirable elements or compounds that are almost always present in
the starting materials, i.e., the undesirable elements in ores, coal,
scrap, etc.  It is self-evident that there must be  an "outlet" for the
contaminating materials from an operational viewpoint.  In .the past, this
"outlet", generally speaking, was either simple landfilling (dumping) or
storage, i.e., landfilling awaiting a suitable, profitable reclamation
process.
           In this report, a reclamation process is defined as a method
of rescuing a material from an undesirable state.  For contaminated materials,
an ideal reclamation would' economically separate for use both the contaminant(s)
and the purified basic material for recycle (e.g. to obtain oil and oil-free
mill scale from oily mill scale).
           This study is focused on reclamation processes for steel industry,
iron-bearing wastes (now being dumped)  which offer the best potential
                                     579

-------
 for  the  greatest recovery of iron and possible valuable by-products  (former
 contaminants).  Included is iron-bearing electric-arc furnace steelmaking
 dust that are now listed as being hazardous (leachable) when placed  in
 ordinary landfills/    The waste materials of interest in this study
 are listed in Table 1.

 Introduction to the Reclamation Processes Appraised in This Study

          This paper presents information and appraisals on four specific
 reclamation processes and one general type of a pyrometallurgical reclamation
 process that has many variations.  The format for presenting information
 is according to the iron-bearing waste name or type.  Two processes, one
 used at Inmetco of Ellwood City, Pennsylvania and the other at Huron Valley
 Steel of Toledo, Ohio, were not included because they were learned of
 too late for inclusion in this study.
           A process is regarded as "existing" if it is, or has been,
 in operation on a plant scale.   The designation of "an emerging process"
 is more arbitrary.   If a process has reached the pilot plant stage or has
been tested even briefly on a pilot or plant scale, and there are technical
 reasons for believing that the desired technical results can be achieved,
 it was considered an emerging process.
           Almost all of the individual processes are the development of
 a single organization or company.  For proprietary processes the commercial
 interests of the owners often limited the amount of information that was
 released.

 RECLAMATION PROCESSES FOR THE DEOILING OF MILL SCALE

           Scale is the oxidized surface layer that forms on semi-finished
 steel during the heating and hot-working operations in rolling mills.
 During the hot-forming operations (multiple steps) the mill scale is
 periodically broken away from the steel shape by breaker rolls and/or
 water and steam jets.
                                     580

-------
    TABLE 1.   STEEL INDUSTRY,  IRON-BEARING SOLID  WASTES
                 OF INTEREST  IN THIS  STUDY
                                                                     (2)
 Waace

 Mill Scale
 (672 of total
  is recycled)
 Steelmaking
   Dusts and
   Sludge*
 (20% of total
  is recycled)
 Blast Furnace
   Dust and
   Sludges**
 (782 is recycled)
Amounts of Waste
Landfllled per
125 Million Tonnes
of Steel Produced
(millions of tonnes)

        1.7
  Range of
Iron Content,
  percent

  58 to 70
        1.8
  45 to 60
        0.75
                               10 to 40
Reasons For
Landfill ing

Oil content too
high for trouble-
free recycling
via sintering
Zinc and/or
alkali content
too high for
recycling to
blast furnace

Contaminated with
oil and non-
ferrous compounds
 * - This represents the amount of landfilled dust  and sludges from the pollution
     control equipment on all U.S. Steelmaking processes.  The total dust
     from electric-arc furnace Steelmaking is about 350,000 tonnes or about
     20 percent  of  the total weight of Steelmaking  dust collected.
     Electric-arc furnace dust is listed as hazardous because of the leachable
     lead, cadmium, and chromium content.

** - This represents the 22 percent of the total dust and sludge collected at blast
     furnaces which is not recycled.
                                       581

-------
           As the scale is broken away from the hot steel It falls through
the roll tables into a water flume.   Also entering the flume are variable
quantities of lubricating greases and oils from the rolling machinery.
In the carryout of scale, it passes a series of traps or basins which
provide an automatic size classification  system.      The large particles
with a low oil content are collected at the beginning of the system;  on
the other end of the system, there is an oil sludge in the clarifiers
which has a small amount of mill scale fines.
           Mill scale is partly oxidized steel and, therefore, contains
a high  percentage   of iron (72 to 75 percent), and no or low tramp
element contamination.   Without the oil content that accumulates, and
looking only at the chemical composition, mill scale is a much better
quality iron source than iron ore pellets.
          Generally, mill scale is screened and the coarse fraction is
used for direct recycle to blast furnaces.  Mill scale fines  (usually
less than 4.76 mm (3/16 in.) were almost  always in the past recycled
to the sinter machines for agglomeration with other materials, followed
by direct return to blast furnaces.  In present practice, as indicated
              (2)
in one survey,    about 67 percent of the mill scale generated is recycled
and about 33 percent or 1.7 million tonnes (about 1.9 million short tons)
is landfilled annually in the U.S. because it is too oily for recycling.
This oil mill scale, if charged to sintering operations, would result
in air pollution and operational problems.
An Existing Reclamation Process For
0ily Mill Scale—Thermal Deoiling
          The only commercialized thermal method known for deoiling
mill scale involves use of the direct-fired rotary kiln operated by
the Luria Company for the Inland Steel Company.  A schematic diagram
                                     (4)
of this process is shown in Figure 1.
           The rotary kiln deoiler is a counter-current reactor in which
air is drawn from the mill scale exit end.  The kiln is fired with natural
gas and in passage through the kiln the mill scale is effectively deoiled
                                    582

-------
 -4.76mm INLAND MILL SCALE FINES
           r
      VIBRATING SCALPING
      SCREEN 16mmx50mm
                                      I
                                      I
                                      I +16mm
                                  OVERSIZE STORED
            (100 WT.%~0.4% OIL,-4.3% H2O)

             103WT.5S
     DIRECT FIRED DEOILING
         KILN 3mx20m
                                 "57. AS STEAM
                                  AND HYDROCARBONS IN
                                  KILN OFF-GAS (-315°C)
     —2mm
     »3 WT.%
— BREECH
  MATERIAL
 •co.oir. OIL
  0.00% H2O
  — 4.7 6mm
   -93 WT.%
KILN PRODUCT <0.01% OIL
  55 Mfl/h     0.00% H2O
                   -0.6mm ~3 WT. %
                   .WET SCRUBBER,
                   •— SLUDGE  ^	
                          <0.01% OIL
            FINAL DEOILED
            MILL SCALE TO
             SINTER PLANT
              
-------
from 0.4 to less than 0.01 percent residual oil.  Two afterburners
operating at 650 C  (1200 F) and higher are used to assure complete
combustion of any hydrocarbon vapors unburned in the kiln.  The off
gas from the afterburners is then scrubbed in a venturi scrubber to
remove particulates.  The wet scrubber sludge is added to the deoiled
kiln product for return to the steel plant.
Energy Requirement—
          The mill  scale is received both wet and oily.  The total
energy usage as fuel is 0.22 G cal/tonne (810,000 Btu/short ton) of
product.  About 40  percent of the total fuel consumption is used in
the afterburners for environmental protection.

Processing Costs--
          No purchase cost or credits are taken for the Incdming mill
scale.  No toll or  royalty charges are considered because this infor-
mation is proprietary.
          For a production rate of 100,000 tonnes per year (sized for a
typical steel plant), with new equipment, it is estimated that processing
costs with steel plant labor rates are $35.93/tonne or product ($31.69/net
ton).  This is not  an estimate of the processing costs for the Luria
operation at Inland Steel which is about 5 times larger in capacity.
The economics of scale are very pronounced in rotary kiln operations.
Environmental Appraisal Summary—
          No stack  data on emissions from this process were available
for this study.  However, the emissions can parallel those from other
hydrocarbon-rich sources.  If the afterburners on this process are
closely controlled, it is expected that emission control will be accept-
able and that no new environmental problems will be created.
An Emerging Reclamation Process for Oily Mill
Scale—Solvent Washing

          Colerapa  Industries (Ravenna, Ohio) has a solvent washing
pilot plant in operation for oily mill scale and mill scale sludge which
employs the Duval-Pritchard process.   Oily mill scale is "dry cleaned"
at a rate of 1.8 tonnes/hour (2.0 ton/hour) using methylene chloride
(CH^Cl.) as solvent.  This solvent boils at 40 C (104 F)  and has a density
of 1.35 g/ml.  The process flow sheet for this operation is shown in
Figure 2.  This process is claimed to handle mill scale of almost any
particle size, oil  level, or water content, with oil sludge containing
                                 584

-------
BLOCK DIAGRAM -. MILL SCALE DEQIL1NG  PROCESS
OILY HOPPER EXPOSURE P-l MIXING P-2 VIBRATING MIXING P-3 VIBRATING VI
SOLIDS TANK TANK SCREEN TANK SCREEN
OR OR
HYDROCONE HYDROCONE
Q »|» H-l -** ET-1 —,$]-* MT-1 _-4D7-»* S-l -P* MT-2 ^^^ S-2
1
en
oo
tn
i


\
EXPENDED 	
SOLVENT T-
(SOLVENT RETURN* 1 T SOLVENT RETURN 1

^""~ EXPANDED FRESH
1 SOLVENT SOLVE
FINE SETTT-TNG STORA<
TANK TANK
1


STUKAuE • '
TANK
-H^
NT
GE

BRATING DRYER
SCREEN
S-3
(I

T-2

1




7^
C ^\TT C
SOLVI
^h-


D-l
S
;NT
i

C-l
j
1
E-l
SCREEN
CONVEYOR
sc-i
/WV\|
1
OIL FREE
SOLIDS
CONDENSER
STILL
T
OIL TO
STORAGE
 Figure 2.  Block Diagram of Mill Scale
            Deoiling in  Duval-Prttchard
            Process

-------
small amounts of mill scale being particularly valuable because of the
level of recoverable oil in the feed material.
          The path of the mill scale through the process is counterflow
to that of the solvent.  As in ordinary dry cleaning of clothing, with
discontinuance of agitation following treatment there is a density separa-
tion of the solvent, water, and oil at the end of the process.  To collect
final products (dry mill scale, solvent-free water, and solvent-free oil)
each raw segregation segment is withdrawn and subjected to a heating
operation to strip out the low-boiling solvent.  Contaminated water is sub-
jected to a stripping operation for solvent recovery, and solvent-contaminated
oil is treated in a still to recover solvent and particle-free oils.
Material Balance—
          Process literature indicates a solvent makeup of 0.6 pound of
solvent per ton of product mill scale.  If this loss can be maintained,
the operation could be considered to have an emission rate less than
that of a well-controlled dry cleaning establishment.
Energy Requirements—
          The process developer takes energy credit for the oil recovered
and the amount of water displaced and not evaporated.  Reported energy
purchased is 3.3 kwhr/ton for electrical energy and 0.07 G cal/tonne of
feed (255,000 Btu/ton) for the fuel required for solvent evaporation and
distillation.
          With the assumption that the incoming scale can average 12 per-
cent oil and 14 percent water, the net energy gain (mainly in recovered
oil) is 1.7 G cal/tonne (6.2 million Btu/ton) of mill scale product.
Processing Costs—
          Costs for removing oils and greases from mill scale by washing
with a solvent were estimated to be $16.84/tonne or $15.28/ton.  This
assumes a credit for about 30 liters of oil (about 8 gallons) per ton of
deoiled mill scale produced.
Environmental Appraisal Summary—
          No emission data were available for this process so an evaluation
of the potential gaseous, liquid, and solid wastes had to be made.
          From the viewpoint of an environmental engineer, solvent vapor
emissions from this process could be reduced to a'very low level if
special design considerations are given to solvent emission control in
terms of equipment design (tightness) and equipment types.  For example,
                                  586

-------
efficient condensation of the low-boiling vapors may require refrigerated
condensers.  All exhaust pipes from the system may require activated-carbon
filters.  Favorable to the process is the fact that the listed toxic con-
centration for the selected solvent in air is higher than that for most
commercial solvents that are routinely used.   However,  the solvent is a
listed priority pollutant for water and any application of the process
should include consideration of the environmental significance of residual
methylene chloride in the water and losses to air and solids.

RECLAMATION PROCESS FOR STEELMAKING DUSTS

          For U.S. steel production at the rate of 125 million tonnes/year
(138 million tons), the amount of dust collected annually from the exhaust
gases of the three types of steelmaking processes totals an estimated 2.2
million tonnes.^  Of this quantity, only about 20 percent is recycled and
about 1.8 million tonnes (2.0 million tons) is landfilled.  This estimate
includes landfilling of about 350,000 tonnes of electric-arc furnace steel-
making dusts and sludge that have been listed by the EPA as being hazardous.^ '
          The majority of the steelmaking dust collected is not recycled
because it contains contaminants that (a) are not removed during any agglo-
meration process, and (b) if recycled to blast furnaces (following agglo-
meration), would cause operating problems.  With regard to blast furnace
operations, the contaminants of concern are mostly zinc and alkalies.  Of
concern to the EPA are the lead, cadmium, and chromium contents in the
steelmaking dusts.  The source of these contaminants is the steel and iron
scrap used in every steelmaking process, and the level of hazardous elements
in steelmaking dust depends on the amount and type of scrap used in a parti-
cular steelmaking operation.  Practically speaking, electric-arc steelmaking
furnaces use 100 percent scrap charges and the steelmaking dusts collected
at these furnaces can contain hazardous contaminants to levels of 4 percent
lead and 0.05 percent cadmium.  High chromium-content levels in steelmaking
dusts are usually restricted to dust collected in alloy and stainless steel
processing.
                                     587

-------
"Greenballing"—The Reclamation of Steelmaking
Dust by Recycling to Steelmaking Furnaces

           The literature in the 70's described a reclamation/recycle
approach for contaminated Steelmaking dusts which consisted of pelletizing
the finely divided dust and recycling the pellets back to any Steelmaking
furnace.  The term "greenballing" was used because the pellets had not been hardened
but were "green" or freshly made.
           In a Steelmaking furnace charge, the iron oxides in green pellets
are used as a substitute for iron-ore pellets and are utilized in slag
formation and melt cooling.  The zinc, lead, cadmium, and perhaps some
alkalies in the scrap plus the reducible nonferrous oxides in the
green pellets are reduced and some are vaporized, and collected in the standard
dust collection equipment.  Assuming complete collection of the volatile nonferrous
oxides and no losses to the slag or molten steel, the nonferrous content
in the Steelmaking dusts should rise steadily during green pellet
recycling.  In theory, after a number of recycles, the nonferrous content
in the Steelmaking dust could be high enough to warrant periodic diversion
of this dust to a nonferrous smelter.  The continual buildup in zinc content
in the collected dust during recycling has not occurred, indicating that
there is a bleed from the system.
          Discussions with personnel at various steel companies indicate
that (a) the failure of the zinc content in the dust to steadily increase
during recycling remains a mystery, and (b) all companies contacted
have curtailed or discontinued reclamation via greenballing.  Some
companies indicate that greenballing has been curtailed because of the
present low level of steel production and because of the increase in sulfur
content in the steel upon recycling Steelmaking dusts.
          It is assumed that greenballing is being technically examined
by steel companies for a more complete understanding and/or improvement.
This process was therefore included in this study.  The flow sheet of
Bethlehem Steel's greenballing process is given in Figure 3.  '   As indicated
in the foregoing text, the material balance on greenballing is being
investigated.  There is no information of the total energy requirement
for this approach.
                                     588

-------
" HIGH-ENERGY SCRUBBERS
                                                                     ELECTROSTATIC PRECIPITATOR  "
                                 FIGURE 3.  FLOW DIAGRAM OF. BETHLEHEM STEEL CORPORATION'S
                                            GREENBALLING  PROCESS FOR RECLAIMING STEEL-
                                            MAKING DUSTS.  FIGURE WAS DRAWN BY BATTELLE.
                                                     589

-------
Processing Costs—
          An estimate of processing costs for greenballing is about $36.50/
tonne.  If a credit is taken for the iron content in the green pellets as
compared to that in commercial oxide pellets, the production costs nearly
break even with the value of the greenballs.
Environmental Appraisal Summary—
          Inasmuch as lead and cadmium normally vaporize with zinc when
reduced with carbon or carbon monoxide, a balance of these elements must be
obtained to determine the environmental impact of any losses or bleeds.
          The fact that very finely divided compounds may be very difficult
to completely recover from a large volume of exhaust gas causes concern.

RECLAMATION PROCESS FOR BLAST FURNACE SLUDGE

           The top gas stream from almost all blast furnaces is passed
through a series of dust collectors.  In a  first stage, mainly coarse
dust  is removed in a simple dust catcher (expansion chamber) and the
dry dust consists mainly of fine particles  of ore, flux, and coke.  If
the alkali content is not too high, this coarse dust is recycled through
the sinter plant.  The carbon content in the dust serves as a sintering
fuel.
           The second and third stages of gas cleaning consist of wet
collection of fine particles.  These are collected as a slurry which is
subsequently dewatered to a sludge holding  20 to 25 percent water.  Sixty
to seventy-five percent of this dust is 44  microns or less in particle-size
diameter.
           In some blast furnaces, about 75 percent of the zinc that: enters
 the blast furnaces in ore or in sinter reports to the blast furnace sludge.
 Continuous recycling of this sludge would  result in an overload of zinc
 in a furnace, and operating problems would develop.  Steel plants are
 landfilling this sludge when the zinc and  alkali contents are considered
 too high for recycling.
                                     590

-------
           About  22 percent of the blast furnace dust and sludge collected
in the United States is not recycled to in-plant sintering operations.
At least some of  this dumping occurs because some steel plants do not have
sintering operations.

Hydrgclassification^j'or DeZincing Blast Furnace Sludge

           In Japan, the lead and zinc content of some blast furnaces sludge is
higher than that reported in the United States.  The lead content can
reach a level of 0.7 percent and the zinc content can reach 7 percent.
           In 1974, Nippon Steel Corporation and Kowa Seiko Co. (a nonferrous
processing company) began a joint study on utilizing nonferrous metal-
bearing blast furnace sludge in processes "other than pyrometallurgical
reduction".     The zinc and lead contents were found to be concentrated
in the finer particle-size portion of the sludge.  For example, about 80
to 90 percent of the lead and zinc compounds are concentrated in the
portion of the sludge that is smaller than 44 microns in particle-size
diameter.  No plant data are available on the particle-size distribution
of the alkali  metals.
           A wet classification system was then developed to separate
out the nonferrous-bearing portion of the sludge to permit recycling
of the "cleaned" portion.  Laboratory and field tests indicated that the
desired separation or beneficiation could be accomplished by means of
hydroclones.  The action of hydroclone equipment on a slurry is analogous
to passing a gas and dust mixture through a dust-collection cyclone.  In
both instances, classification by centrifugal forces occurs.
           By  1977,  the  wet-classification reclamation method for  blast  furnace
 sludge was in plant operation at the Kamaishi  Works in  Japan.   The  flow sheet
 of this process is shown in Figure 4.      At the Kamaishi  Works,  the
 slurry from the dust cleaning systems  on 2 blast furnaces  is adjusted
                                     591

-------
                                    Hydronegaclone >>6 units
                                         Settling tank for
                                         overflow slurrv
' Water for gas cleaning
I and dual collection
I
i
                                                    Siphonic pressure
                                                    control valve

                                                               Coagulant
.
[
ckL
»
[


.
E
PI
r


Jt
1
^

                                                                                                  jiCompratcd
                                                                                                    air
                                                       Vacuum drum
                                                       filler          Filter pr
                                                                                               Catchment
                                                                                               tank
Underflow tank
                       Clean water
Low zinc dust High-zinc dint

     i
  Sintering
      Figure  4.   Flow Sheet  of  the  Blast Furnace Sludge  Processing  Method
                     at  the  Kamaishi Works  in  Japan
                                                  592

-------
with water to have a 5 to 15 percent dust concentration.  This slurry
is passed through 6 parallel, upgraded versions of hydroclones.  Each
hydroclone  has 75-ram (3-in.) inside diameter and the flow rate through
the hydroclones is in the range of 20 to 100 I/minute (5 to 25 gal/minute).
          The overflow slurry contains about 75 percent of the zinc in
the feed material (and presumably most of the lead) and totals about
25 percent of the feed weight.  The underflow contains about 83 percent
of the original contained iron and 77 percent of the contained carbon.
The underflow stream is filtered and is granulated into mini-pellets
for use as a sinter strand raw material.  The use of these high-carbon
pellets reportedly resulted in a decrease of 2 kg (4.4 Ib)/tonne in the
use of coke breeze in the sintering operation.
          The overflow material is allowed to settle and is then dewatered
in a filter press.  This nonferrous-bearing portion contains about 18
percent iron, 13 percent zinc, and 23 percent carbon.  No information was
obtained on the subsequent processing and the subsequent sale or  disposal
of this material.
          Nippon Steel Company has indicated that while they are successful
in reclaiming blast furnace sludge by wet beneficiation, they have not
as yet been successful in using this approach to separate nonferrous
compounds from steelmaking dusts.
Material Balance and Energy Requirement—
          This classification operation is purported to be a very simple
process with essentially 100 percent recovery of the original wet-collected
blast furnace sludge.  Relative to the other reclamation processes, the
energy requirements for this process are considered negligible.
Processing Costs—
          No attempt was made to estimate the low processing costs of this
reclamation operation.  The Japanese author labeled the development "an
epoch-making resources utilization technique".  From what is known of the
process, it appears to be a "real winner".   It can and is being used to
recover iron and nonferrous values from landfilled blast furnace sludge
storage.  It would appear that this physical-separation reclamation process
has the potential of being the starting point for eliminating the need to
landfill about 750,000 tonnes of sludge annually in the United States.
                                    593

-------
Environmental Appraisal Summary—
          It is anticipated that the bleed streams of waste water would
have to be treated to remove the cyanide and water soluble organics.  Such
treatment technology is currently practiced by the industry.
          It was judged that the emissions from this process appear con-
trollable and there is no reason to believe that the process would intro-
duce any new environmental problems.
PYROMETALLURGICAL, COREDUCTION RECLAMATION PROCESSES FOR RECLAIMING
MILL SCALE, STEELMAKING DUSTS, AND BLAST FURNACE DUSTS
          In Japan, there are seven or more rotary kilns in operation
that reclaim  steel industry, contaminated iron-bearing wastes.  In
this paper these processes are called coreduction operations because the
contained iron oxides are reduced to value-added, direct-reduced iron
(metallized iron) while some of the nonferrous oxide contaminants are
also reduced.  In this instance the reducible and volatile nonferrous
elements are vaporized and are collected by means of the exhaust-gas
dust collection system.
          Within the various rotary kiln operations in Japan there are
five different process variations of only minor importance to this
study.  Only the Kawasaki process is described and appraised here because
it is the newest operation (1977) and presumably has the latest in
dust collection equipment.
          In the United States, the rotary kiln, coreduction approach
for the reclamation of various iron-bearing wastes was examined and
tested by Inland Steel and Heckett Engineering in tjhe late 60 *s and early
70's.     To minimize future operating costs, the planning in this
instance called for the use of a very large kiln for reclaiming wastes
from four different steel plants in the Chicago area.  No plant was
ever built because the economics did not appear to be favorable and
cooperation between competing steel companies was difficult to obtain.
However, domestic interest in kiln reduction/reclamation continues.
                                     594

-------
           Over  the  past decade,  several process developers  in  the  United
 States  have been  active in  testing and promoting  their approaches  to  cold
 bonding waste materials for recycling to blast furnaces.  These  approaches
 are  substitutes for sinter  agglomeration and differ  from normal  sintering
 or iron-ore pellet  induration  in that they use only  a small amount of
 process heat.   The  most active cold-bonding technique developers are  the
 PelleTech  and Reclasource Corporations.  PelleTech uses the Michigan
 Technological University (MTU) hydrothermal process.  In this  instance,
 the  binding reactants are hydrated lime and silica that are activated
 for  bond formation  by autoclaving pellets for 1 to 2 hours at  2068 KPa
 (300 psig) steam  pressure.  Reclasource uses a pitch or asphalt  binder
 addition and the  agglomerates  (briquets) are cured at about 260  C  (550 F)
 to form a  carbonaceous bond.
           The early thrust  of the PelleTech and Reclasource efforts was
 to (a)  agglomerate  mainly valuable mill scale, and (b) begin testing
 of their recycle  agglomerates in blast furnaces to prove the strength of
 their agglomerates  to potential  buyers.  This study, is concentrated
 on reclamation  and  not on recycling methods.  However, by extension,
 the  cold-bonding  approaches show promise of agglomerating contaminated
 waste materials (such as steelmaking dusts) for input into some  pyrometallurgical
 reclamation process.  Reduction  of cold-bonded agglomerates of contaminated waste
 materials  in kilns,  rotary hearths, or shaft furnaces would be akin to
 the  coreduction practices developed by Inland Steel, Kawasaki Steel,
 and  others, i.e.,  the products would also be direct-reduced iron  (or smelted
 iron in a  cupola  operation) and  by-product nonferrous oxide dusts.
           The advantage claimed  by both cold bonding organizations is
 that they  are in  a  position to include (and retain) carbon inside of  their
 agglomerates (during cold bonding).  This will result in a much faster coreduction
 of the  agglomerates upon heating.  Faster, that is, than any coreduction
 using coke or char  external to the pellets to produce the necessary
 carbon  monoxide reducing gas.   Technically speaking,  the "high-speed"
 advantage claimed by the cold bonders has been well established.   This
 technical advantage has yet to be translated into an economic advantage,
but  the odds appear favorable that it can be done.  However, process speed
only affects pyrometallurgical coreduction reclamation in the area of
processing cost.  All other appraisal factors for pyrometallurgical
reclamation processes are about the same.   Therefore, only the Kawasaki

                                     595

-------
process  and the environmental  appraisal common to  all coreduction
processes  are discussed here.

Kawasaki Rotary Kiln, Coreduction Reclamation

           The newest "dust-reducing" plant installed by Kawasaki Steel
(1977) is  at the No. 2 plant in  Chiba, Japan.  The installation is
rated at 1000 tonnes/day of material feed.  This feed consists of pellets  of
combined blast furnace dust  and  sludge, oxygen  steelmaking dust, and
sinter dust.   A generalized flowsheet for this plant is shown in
Figure 5.
                                             WIT
                                            OUSTS

                                            fILTIB
                      COAL OR
                     COKE BREEZE
PELLETIZER |
   r
                                              ZINC COLLECTION
                                 PREHEATER
              RECYCLE
                                   ±
      J
                          LUMP -^REDUCTION KILN V-
                          e(Wi  »———-i^——I
                          COKE
                __J OPP-OAS
                                            WATER
                                  PRODUCT
            Figure 5.  Generalized Flowsheet of Kawasaki Steel's
                       Coreduction Process
                                       596

-------
           Kawasaki  Steel avoids  the need  for  drying  sludge by  blending dry
 and wet materials  (and by using  supplementary, dry,  fine  iron  ore when nec-
 essary).   No  information is available on  the  quantities and  range in  quantities
 of each waste material introduced into the process.  Normally  no binder
 is used.   A grate preheater (downdraft) is positioned ahead  of the kiln
 to dry and heat-harden the pellets of waste oxides.  Pellet  heating is
 done by means of the exhaust gases from the rotary kiln.Recent information
 indicates  that  the  grate kiln heats the pellets  to 1000 C (1,830 F) before
 they enter the  kiln.  The variation and level  of  the  carbon content in the
 preheated  pellets are not available.
           The new  Kawasaki kiln  is  4 meters  in diameter  (16.5  feet) and
 55 meters  long  (180 feet) .  The auxiliary heat is supplied with a heavy-
 oil burner positioned on  the discharge end of the kiln.   This  burner  is
 capable of firing  3,000 I/hour  (793 U.S.  gal/hr). Product recovery is
 about  600  tonnes/day.  The feed  to  the kiln  is a mixture  of  pellets and
 coke breeze.  The  coke breeze is used as  a fuel  and  a source of reducing
 gas.   The  consumption rate of coke breeze is  about 305 kg/tonne of product
 (610 Ib/net ton of  product).  The operating  temperature in the kiln
 is in  the  range of  1,100 to 1,200 C (2,000 to 2,192  F).   No  information
 was attainable  on  the gas velocity in the freeboard  zone  of  the kiln  interior.
 (See Environmental  Appraisal.)
           During passage through the kiln, the  iron burden  is metallized
 to the range of 90  to 95 percent.  Also during coreduction,  about 95
 percent of the  contained zinc is  reduced  (and blown out).  Lead elimination
 is about 95 percent and elimination,of alkalies  is about  50  percent
 (also blown out).  The metallized-iron product contains about 0.3 percent sulfur
 and 15 percent  gangue,  which minimizes  the possibility of using the
 product in steelmaking.
           The exhaust gases from the kiln pass  through the  grate
preheater,  through a water atomizing tower (to lower the gas  temperature),
and then through a huge electrostatic precipitator.  It  was not possible
 to obtain many details  on the operation of this gas cleaning  system.
           According to  one source,     the typical composition of the
kiln exhaust dust of the Kawasaki coreduction kiln is as  follows:
                                     597

-------
                    Element              Weight Percent
                    Zinc                       14
                    Lead                        4
                    Iron                       26
                    Carbon                     18
                    Sulfur                      1
                    Sodium                      1.7
                    Potassium                   1.9

Kawasaki Steel states that this dust has little or no value to a buyer
and is disposed of to the nonferrous industry.  The above powder can be
considered to be a low-grade zinc concentrate contaminated with iron.  This
problem of iron contamination of the nonferrous byproduct collected was
also reported by Holowaty in 1971.
Material Balance—
          Kawasaki Steel Corporation has not published material balance
data on their coreduction rotary kiln process.  Of interest in a material
balance would be the volume and composition of the off gases, as well
as the dust loading in the off gas.
Energy Requirements—                                              ;
          The reported total fuel requirement per tonne of product is in
the range of 3.5 to 4.2 G cal/tonne.  This is equivalent to 12.6 to 15.1
million Btu/net ton of product or about 18.6 to 22.3 million Btu/net
ton of metallic iron in the product.  About one-third of the total fuel
requirement is in the form of heavy fuel oil, and the remainder is coke
breeze.
          The zinc content of the pellets charged to the Kawasaki kiln
is less than 1 percent and the fuel requirement for the reduction of the
contained zinc and lead oxides is therefore negligible, i.e., the listed
fuel requirement is almost entirely for the reduction of the iron oxides
in the waste materials.
Process Costs—
          To a major steel plant, the total reclamation costs using a core-
duction process would be made up of the elements in the following equation:
                         VALUE OF               PROCESSING         TOTAL COST
SAVINGS       +          PRODUCTS      -        COST          *    OR PROFIT
(Avoiding                (Metallized iron +
    landfilling)         Nonferrous dust)
                                     598

-------
Dealing with each element of cost in the preceeding equation, estimates
were developed as shown in the following paragraphs.
          Obtaining estimates for the cost of regulated landfilling proved
to be a problem.  Firm data on this topic may not as yet have been developed.
Oral statements by steel company personnel place the 1980 landfill costs in
the area of $94 to $100/tonne ($85 to $90/net ton), with multiplying in-
creases expected in subsequent years.  This cost level was somewhat confirmed
by inquiries made to commercial landfill companies who are quoting a minimum
price for regulated disposal at 4$ per pound, delivered.
          The maximum possible theoretical value of the iron content in
metallized iron is equal to the selling price of the iron in high-quality,
direct-reduced iron.  Direct-reduced iron is listed for sale at $130/tonne or
$116/net ton.  Discounting for the lower iron content in the coreduced
product brings the theoretical value (based on iron content only)
to about $97/tonne ($88/net ton).  However, metallized iron containing
high gangue and sulfur is not suited for steelmaking operations and can
only be recycled to blast furnaces. While it is a fact that high-quality
direct-reduced iron charged to blast furnaces both decreases the specific
amount of coke required per ton of product and increases the furnace
production rate; the gangue in the coreduced product has heat requirements
that act counter to these gains.  As a judgment, the value of high-
gangue, metallized pellets to a blast furnace is expected to be about
$70 to $80/tonne or $63.50 to $72.50/net ton.
          The by-productnonferrous oxides collected during any coreduction
operation are not expected to have any value because of the high iron
content (and related low zinc content) and objections to the presence of
alkalies, chlorides, and other contaminants.  Because disposal of this
exhaust fume is also a problem, procedures (probably hydrometallurglcal)
will have to be developed to upgrade this material.  Based on this expectation,
no negative value was taken for the mixed nonferrous oxides.
          Processing costs are estimated to be $126/tonne or $115/net ton.
This is for a plant producing 100,000 annual tonnes of metallized product, ,
starting with a total of 133,000 tonnes of waste materials.  All unit costs
are taken from published information, including the average employment cost
of labor in the steel industry at $19.11/mahhour. Labor requirements are
taken from Kawasaki information.
                                    599

-------
Total Coats or Profit
SAVINGS
Repeating the basic equation:
               VALUE OF
     +         PRODUCTS
PROCESSING
COSTS
   TOTAL COST
   OR PROFIT
and filling in the derived numbers:
(1.33 x $94)
                 $80
  $127
$78/tonne of
   product,profit
$707ton of productt
   profit.
The element that causes the result to show a profit is the savings of
$125/tonne of product in not landfilling the wastes.  Stated another way,
if there were no substitute for the coreduction processes and the assumed
savings in avoiding the cost of regulated landfilling are factual, the
processing cost of a coreduction reclamation process could reach the
level of $205/tonne of product and theoretically "break even".
          The data base for the above presentation of costs is weak and
conditions did not permit any detailed cost analyses.  However, the
expected increased cost of landfilling will act as an incentive toward
developing suitable reclamation processes.
Environmental Appraisal of Coreduction Operations
          Information from Kawasaki Steel indicates that the flue gas re-
leased after the electrostatic precipitation contains between 0.001 and
         3
0.03 g/Nm  of dust.  Lead oxides, cadmium oxides (and presumably zinc
oxide), and organic particulates (hydrocarbons) are below "identification
limits".
          As effective as the collection of particulates in the Kawasaki
process appears to be, it may or may not be sufficiently effective to meet
the U.S. primary ambient air quality standards for lead and particulate
                            3                    3
emissions of 1.5 microgram/m  and 75 micrograms/m  respectively—as measured
at the plant property line.
                                     600

-------
          The nonferrous metal emission  from a coreduction process
is likened more to that of secondary lead (or zinc) recovery processes
in the United States.  In these recovery processes the amount of stack
gas being treated is about one-tenth the flow rate being treated in
the Kawasaki process.  This fact and the probable low concentrations
of dusts in the effluent , as well as, the 200 C temperature of the gases
entering the electrostatic precipitator, suggest a potential for poor
collection efficiency of the nonferrous metal dusts (98 percent recovery
claimed).
          The fuel used in the Kawasaki process is heavy oil of unknown
sulfur and ash content.  Because of the high temperatures in the kiln
and grate preheater,  hydrocarbons in the exhaust gas from the fuel oil
burner and volatile matter in the coke are expected to be consumed.
          Insufficient information was available to make environmental
assessments of coreduction operations.  However, it was judged that
these coreduction operations have the potential for unacceptably high
nonferrous metal emissions.  It is believed that this potential exists
because of (a) the extreme fineness of the non-ferrous fume, and (b) the necessity
to remove a small amount of fume from a very large volume of exhaust
gases.  The stack gases from coreduction operations should be measured
to determine environmental acceptability for the United States.
          The judgment on the environmental aspects of coreduction
operations holds for all coreduction operations in which lead,  cadmium,
zinc, and alkalies are vaporized from the charge.  Neither the method
of waste agglomeration nor the method of heating the agglomerates has
any bearing on this vaporization.   The same vaporization would occur
in rotary hearth furnaces and in shaft furnaces,  including smelting
of the agglomerates in cupolas.
                                     601

-------
PROCESS RANKING AND CONCLUSIONS

          The reclamation processes considered and appraised in this
study are ranked based on a criteria listing that includes:
          •   Waste-processing capability
          9   Consideration of possible environmental problems
          e   Overall reclamation economics
          •   Energy considerations
Following the outline of this paper, processes are classified by the
type or types of waste that they can reclaim.  The general ranking of
these processes in qualitative  terms is given in Table 2.

Conclusions

        (1)   Within the reclamation processes that can only reclaim
              oily mill scale, the emerging solvent-washing process may
              have apparent advantages In terms of:
              (a)   Being able to process materials having a wider range
                    of oil-to-scale ratios
              (b)   Having a lower estimated processing cost primarily
                    because it recovers oil instead of requiring oil
                    or natural gas as fuel.
              Because of the low boiling temperature of the solvent used
              in the solvent-washing process for mill scale, it may be
              necessary to include refrigerated condensers and traps
              to hold solvent emissions to an acceptable level.
              Given controlled operation of the afterburners in the
              rotary kiln, mill scale deoiling method; there is
              no concern about acceptable emission control or the
              development of any new environmental problems.
        (2)   Based on the limited information that is available on the
              Japanese physical classification process for blast furnace
              sludge, this process appears to have the attributes of a
              winning process in terms of:
                                    602

-------
                                           TABLE 2.  GENERAL RANKING OF RECLAMATION PROCESSES FOR OILY MILL SCALE,
                                                     STEELMAKING DUSTS, BLAST FURNACE SLUDGE
Ranking Criteria
Was te-Procesaing
Capability
Process Type
Status
Mill Scale
Only
Rotary
Kiln
(Pyro)
Exist-
ing
Solvent
Hashing
Emerging
Steel making
Dusts Only
Greenballing
(Pyro)
Once Existing
Now in Abeyance
Blast Furnace
Sludge Only
Physical Separa-
tion
(Wet Classification)
Existing in Japan
All Wastes (Mill Scale,
Blast Furnace, and Steel-
making Dusts and Sludges
Pyro Metallurgical
Coreductlon
Existing in Japan.
May be eaerging in U
                        Environmental
                        ConsIdera tions

                        Economic
                        Considerations
(1)
(2)
O
CO
        Moderate  Low
        Cost      Cost
                        Energy Requirements
        Moderate
        (Fuel
        Used)
Very Low
(Oil
Recovered)
                                    101
             Moderate
Low
                                         E+l
                      Very Low Cost
Neglible
                                                 with cold bonding
                                                 agglomeration.

                                                        [0]
High Cost, Profitable
only by avoiding expensive
landfilling

High
                        (1)   A [+]  rating—From the Information available about  the process and  the  technology available  to  control  related
                                           emissions,  the process is judged favorable.
                             A [0]  rating—Limited data on process and emission  composition.  It is  judged  that  insufficient information  Is
                                           available to make environmental appraisals without further data.
                        (2)   At a stated minimum cost of  future landfilling of  $100/tonne of waste, or more, all of  the above  processes
                             are profitable on an overall basis when expensive  landfilling  is avoided.  The qualitative judgment listed
                             refers to processing costs.

-------
      (a)    Eliminating the need for landfilling

      (b)    Being a true reclamation process in the recovery  of
            iron and carbon units in one stream and concentrating
            (for further treatment)  the nonferrous contaminants
            in the overflow stream
      (c)    Introducing no  environmental problems
      (d)    Having a low processing  cost and low energy requirement
      (e)    Being retrofittable to the dust and sludge processing
            systems of  existing blast furnaces.
(3)    Unfortunately,  there  is no known emerging process(es)
      geared to the simple  reclamation of steelmaklng dusts.
      The  simple approach of recycling pelletized steelmaking
      dusts to steelmaking  furnaces  (greenballing) apparently
      requires further  development because this method is
      not  being practiced at this time.
      Research and Development is definitely needed for reclamation
      methods that will suitably process only steelmaking dusts.
      Particular emphasis should be  given to electric-arc
      furnace (EAF) steelmaking dusts—the one type of steelmaking
      dust that is listed as being hazardous.  While EAF dusts
      are  high in recoverable nonferrous resources, they unfortunately
      are  low in iron content and very high in near valueless and
      relatively inert  gangue content.  With these characteristics,
      these dusts do not necessarily represent good feed material
      for  any coreduction operation.  The answer to the resource
      recovery from EAF dust is expected to be some new in-plant
      process that is economical on  a small scale and is capable  of
      recovering zinc compounds and  removing or rendering
      harmless the hazardous nonferrous components.  Favorable
      to any processing cost will be the savings in avoiding both
      landfilling and possible shipping costs.  To the best
      of our knowledge  no reclamation process for steelmaking
      dusts that is based on physical classification has been
      successful.  However, further  research efforts in this
      direction are suggested.
                            604

-------
(A)    There are existing  and  emerging processes  that can process
      all of the iron-bearing wastes of  interest to this study  (mill
      scale, some steelmaking dusts, and blast furnace dust and
      sludges.)
      All of these processes  are  pyrometallurgical coreduction
      operations.   The  existing processes  use rotary kilns
      and the emerging  processes  are also  considering shaft furnace
      smelting in cupolas and shaft furnace  solid-state reduction
      (direct reduction). All of these  pyrometallurgical processes
      reclaim contaminated wastes by vaporizing  some of the
      contaminating nonferrous metals in wastes. Vaporization  is
      followed by burning the volatilized elements and then stripping the
      finely divided fume (mainly oxides)  from a large volume of  exhaust
      gas.  Until stack sampling  data become available, it is
      necessary to be concerned about whether pyrometallurgical
      coreduction processes can meet the present standards for
      lead emissions and  the  future standards  for cadmium emissions.
      Coreduction processes might eliminate  the  potential of
      a water pollution problem but could  need further emission-control
      developments to avoid an air pollution problem.

(5)    It would appear that  for future reclamation processes  for
      steel Industry, iron-bearing wastes; methods other than
      pyrometallurgical reductions have  greater  appeal in terms
      of lower processing costs,  fewer  (if any)  environmental
      problems, lower energy  requirements, and possibly less
      sensitivity to the  economics of scale. With the advent
      of low-cost, near-ambient temperature  reclamation processes
      for mill scale and  blast furnace  sludge,   a simple reclamation
      process for steelmaking dusts is becoming  a definite need.
                            605

-------
REFERENCES
(1)   Steel Tripartite Committee, "Report of the Working Group on
        Technological Research and Development, September 1980.

(2)   Baldwin, V. H., et al., "Environmental and Resource Conservation
        Considerations of Steel Industry Solid Wastes"  EPA-600/2-79-074,
        April, 1979.

(3)   Federal Register. Volume 45, No. 98, May 19, 1980.

(4)   Balajee, S. R., "Deo11ing and Utilization of Mill Scale", First
        Symposium on Iron and Steel Pollution Abatement Technology, 1979.
        EPA-600/9-80-012.

(5)   Olsen, J. U. and Wheeler, J. E., "Green Balling- The Recovery of
        Iron Units in Waste Metallurgical Steelmaking Fume", Steelmaking
        Proceedings, ISS-AIME, 1978.

(6)   Toda, H.t et al., "Separation of Nonferrous Metals From Blast-
        Furnace Flue Dust "by Hydroclone.  Nippon Steel Technical Report
        No. 13., June 1979.

(7)   Holowaty, M. 0., "A Process for Recycling of Zinc-Bearing Steelmaking
        Dusts", Regional Technical Meeting, American Iron and Steel
        Institute, October 14, 1971.

(8)   Harris, M. M., "The Use of Steel Mill Waste Solids in Iron and
        Steelmaking", Technical Session of AISI, May, 1978.

(9)   Lemmon, W. A., and Haliburton, D., "An Overview of Controls in Primary
        Lead and Zinc".  Proceedings of Symposium on the Control of Particulate
        Emissions in Primary Nonferrous Metals Industries, Monterey, CA.
        March 18-21, 1979, page 135.
                                      606

-------
            HANDLING AND DEOILING OF ROLLING MILL SCALE

            AND SLUDGE—A PROFIT CENTER FROM A PROBLEM
                           L. A. Ouval

              President, Colerapa Industries,  Inc.

                           Ravenna, Ohio
     "Hazardous Waste" is a classification which  has  recently been
expanded to include significantly more steel mill waste streams.
Oil laden rolling mill scale and sludge is one  type of waste which
has been included in this now broader category.   Management of  these
oily wastes has always been a difficult problem for the steel indus-
try, but the EPA's new regulations make the problem even more com-
plex.

     Disposal costs have been soaring and will  continue to do so
as currently used sites are exhausted and more  remote locations
must be secured.  Future water quality standards  will probably
result in a further increase in waste disposal  costs.

     The high iron content of oily wastes has long been recognized
as valuable, but increasingly stringent air quality regulations
have all but eliminated their reuse in sintering  facilities.
Agglomeration techniques have also been generally unsuccessful
because of the oil content of these sludges.  The many obstacles
to their reuse /combined with the costs and complexities of
disposal, create a complex problem for the steelmaker.

     Colerapa Industries has developed a technique for dealing
with waste sludges which affords the steelmaker a most attractive
alternative to an otherwise bleak situation.  The Colerapa system
starts with oily waste handling at the point of initial collection,
using proprietary equipment to hydraulically excavate and trans-
port these sludges.  A process system is then utilized which
separates the iron units from the hydrocarbon contamination.

     The complete process provides the steelmaker with five
specific benefits!
                              607

-------
l)  Greatly improved solids collection efficiency.

2)  Elimination of the problems normally associated with
    sludge excavation and transportation*

3)  Significant, and in some instances, total, waste volume
    reduction.

4)  Production of a high quality iron source, suitable for
    use in any agglomeration operation.

5)  Recovery of oil from the reclaimed mill scale for reuse
    as fuel.
                          608

-------
                          INTRODUCTION
     The subject of this paper is the patented Duval-Pritchard
handling and processing technology as it relates to the recovery
of iron values produced during Rolling operations at Steel Mills.

     The description given in this paper is of a Tuo Ton Per Hour
Pilot Plant Facility used to develop the necessary design criteria
for full scale operating plants to recover the iron values.  These
values vary in size and are mixed uith oil and uater.  The mate-
rial recovered is in the form of clean, dry iron oxide suitable
for recycling back to an agglomeration operation and oil suitable
for use as a fuel.
                             609

-------
HANDLING AND DEOILING OF ROLLING HILL SCALE AND SLUDGE


     All steel mills put steel slabs through a hot form rolling
process.  On a daily basis, valuable iron units are lost in every
facility in the world.

THE PROBLEM

     Throughout the hot strip rolling process, the slab, sheet,
bloom, billet, or bar is being oxidized, cooled and uashed uith
a high pressure uater spray*  Uhen the hot steel is exposed to
oxygen in the air, as uell as the wash uater, a layer of iron
oxide is formed on the surface of the steel being rolled.  This
layer of oxides is called "mill scale" or simply, "scale".

     As steel is rolled, this layer of scale is broken away and
replaced by a neu layer.  Generation of this new scale occurs
each and every time the size or shape of hot steel is changed.

     As scale breaks auay from the steel, it falls through the
roll tables into a flume, or seuer, through uhich high velocity
uater is flouing.  In addition to the scale and uater, a large
amount of lubrication greases and oils from the rolling machinery,
along uith other mill debris, find their way into the flume.  The
larger scale pieces become coated uith oil while finer particles,
uater, grease and oil combine to form a sludge.

     These combined materials pose a serious uater pollution
problem if they are discharged into a uateruay.  In order to con-
trol this pollution, settling pits and basins are used to collect
these sludges and prepare the uater for reuse or discharge*  Much
has been done during recent years in the design and construction
of these collection facilities to increase collection efficiency.
In addition, terminal treatment facilities have been constructed
and installed in an effort to upgrade the quality of uater at the
discharge from collection pits.

     These terminal facilities include lagoons, and filtration
units.  Houever, as uith any collection facility, efficiency remains
high only as long as the system is relieved of uhat it collects.
Once collected, the sludge presents a double problem; hou to relieve
the collection facility, and uhat to do uith the removed material.
                              610

-------
 EARLIER  PRACTICE

      Traditionally* scale drags, ejectors,  and  mobile  cranes
 equipped uith  clamshell  buckets have  been used  to  remove  materials
 from  the collection facility.  Uhen these practices  are used,  the
 larger solids  are removed and  trucked  to the  mill  for  reuse in the
 iron-making  operation.   The smaller and more  concentrated particles
 (uhich are more difficult to handle),  found in  the terminal lagoon
 and filter backuash, are disposed  of  by dumping in land fills.

      The required installation of  air  pollution control systems
 at agglomerating facilities in the iron-making  operation, has
 posed neu problems involving the hydrocarbon  or oil  carryover.
 Thus, more of  the total  scale  is unacceptable for  reuse.

      The basic concept used in the design of  settling  pits and
 basins is to create a quiescent body  of uater that slows  the
 highly turbulent seuer flou, allouing  the waterborne solids to
 settle, uhile  permitting the lighter  oil in the uater  to  rise  to
 the surface.   Problems have arisen in  the use of the traditional
 systems in that the quiescent conditions are disturbed during.
 the excavation of the scale and sludge.  The conventional approach
 to managing these sludges, then, is not a satisfactory solution
 to the problem.

 HYDRAULIC EXCAVATION OF  THE MILL SCALE

      The development of  Hydraulic  Excavation Technology involved
 engineering equipment specifically designed for the variety of
 sizes and shapes of scale pits.  The  Hydraulic  Excavator  provides
 the quiescent conditions in the scale pit so necessary to promote
 the settling of solids as uell as  enabling oil  to  rise to the
 uater surface for^skimming.  The louer turbulence  resulting from
 hydraulic excavation increases the efficiency of the scale pit  and
 reduces the load on the terminal uater treatment facility.  The
 result of this technology is reduced capital expenditure  in scale
 pits and/or filters at the terminal treatment facility.   The
 excavated material, uhich contains louer amounts of oil due to
 the reduced contact between the oil and the particles, is separated
 at the pit site* by  use of hydroclones and classifiers, uhile the
 uater used in transportation of the solids is returned to the
 influent end of the scale pit.

 DEOILING THE MILL SCALE

     Development of this process of deoiling mill  scale started
 years before the present emphasis on air emission  and hazardous
uaste management began.  It uas evident that materials handled
 hydraulically uere  more free of oil than those  handled uith
clamshells and drags.   A uater uashing method uas developed uhich
extended the period  that the mill scale uas exposed to uater.   This
 resulted in the continued attrition of the oil  from the particle and
                              611

-------
                 FIGURE 1
BLOCK DIAGRAM - MILL SCALE DEOILING PROCESS
OILY HOPPER EXPOSURE P-l MIXING P-2 VIBRATING MIXING P-3 VIBRATING VIBRATING DRYER SCREEN
SOLIDS TANK TANK SCREEN TANK SCREEN SCREEN CONVEYOR
OR OR SC-1
HYDROCONE HYDROCONE
^_y — ^|^ H-I —^ ET-I ^57~^ MT~1 i»3 f~"^ s-1 ~~^ MT~2
! i


I
EXPENDED ""'•
SOLVENT T-
T SOLVENT RETURN! 1 1

1 SOLVENT
_. FTNK SKTTTiTNG
TANK
i


TANK
.— -j j- _^ _^.
— T3£ * 2 ° 3 1 b 1 AAAA|
:| < ^OILS 5 1
SOLVENT RETURN 1 1 SOLVENT •
OIL FREE
1 SOLIDS
SOLVENT T-2 ^ - C-l
qTnw»nR CONDENSER
TANK , i


T
OIL TO
STORAGE

-------
the production of a solid product with substantially lower oil
content.  Oils removed by this water wash process contain sub-
stantial amounts of water and very fine particles and have
limited use without additional treatment.

     Further improvements were made to the system through the use
of detergents and alkaline solvent solutions; however, great care
in reclamation of the water was required not only from the stand-
point of the cost of these additives but also because of the
carryover effects on the mill water.

     The ultimate improvement was the development of the Duval-
Pritchard process to convert hazardous hydrocarbon laden steel
mill wastes to oil for use as a fuel or to be recycled, and to
high grade iron concentrates for reuse in steelmaking.

     The system is able to treat materials that are:

     a)  stockpiled

     b)  sludge-like and higher in oil and water contents
     c)  smaller sized

DESCRIPTION OF OUVAL-PRITCHARD SOLVENT EXTRACTION SYSTEM

     This slide shows a process schematic of the pilot plant.
The plant contains all planned recycle streams and will produce
a totally deoiled mill scale product, plus a recovered oil product.
Solvent recovery facilities are also included in the pilot plant
design*

     The mill scale is deoiled in two mixing stages of solvent
washing, with a counter-current solvent flow.  The two stages of
solvent washing are followed by a solvent rinse to insure total
deoiling.  Spent solvent is evaporated and recovered for reuse in
the process.

     The mill scale is first fed into a hopper (H-l) with a front-
end loader.   The scale is transferred to an exposure tank via a
screw conveyor*  In the exposure tank the mill scale is slurried
with spent solvent and pumped to the first stage mixing tank (PIT-l),
This mixing tank is an agitated vessel which provides total wetting
of the mill scale with the solvent/oil solution.  The mill scale
slurry is then transferred from this first stage mixing tank to a
second stage mixing tank through a transfer pump (P-2)*

     In order to achieve a good counter-current washing effect,
the solids must be de-wetted between the first and second mixing
stages.  This is achieved in a hydroclone (S-l), i-n which the
solids are de-wetted and sent to the second stage mixing tank.  The
liquids are then returned to the first stage mixing tank along with
tho overflow from tho second otage mixing tank.
                              613

-------
     The second stage mixing tank  (MT—2) is  also agitated  to pro-
vide complete wetting of the mill  scale with a lean solvent solu-
tion.  After mixing, the slurry is transferred to another  hydro-
clone (S-2) uhich de-uets the solids*  The solids then receive a
final rinse on a vibrating screen  (S-3). The liquids from  S-2
flow back to the second stage mixing tank along uith the tinge
liquids from S-3.

     The deoiled solids, uith some entrained fresh solvent, are
transferred from the rinse screen  to a dryer (D-l) uhich has a
steam jacket and steam heated screw.  In this dryer, the solvent
is vaporized and removed from the  mill scale, leaving a dry,
uarm scale uhich is transferred to a product pile through  a screu
conveyor (SC-1),  Vapors from the  dryer are sent to a water-
cooled condenser, uhere the solvent is re-liquefied.  The  liquid
solvent then flous to the fresh solvent storage tank (T-2).

     The oil rich solvent from the first stage mixing tank flous
to the spent solvent fine settler  tank (T-l).  In this tank the
fines settle to the bottom and the uater contained in the  mill
scale floats to the top.  The uater is removed through a side-
mounted drain and the fines are recycled to the process*   The
spent solvent is transferred to a  spent solvent storage tank
(T-3).  The spent solvent storage  tank adds surge capacity to the
system and also gives one final stage for uater separation from
the solvent.

     The spent solvent is recovered as follous:  It is sent, on
a batch basis, to an evaporator (0-1) uhich is steam jacketed;
steam is supplied to the evaporator uhich boils off the solvent,
leaving oil in the liquid phase: solvent vapors are condensed in
the uater cooled condenser (C-1J and sent to the fresh solvent
storage tank.  The oil is pumped to oil storage.

     Each storage tank is sealed in order to minimize evaporation
losses of the solvent.  A package  boiler system provides steam
for the dryer and the evaporator.

     The following is an example of the process energy balance
based upon a typical analysis of materials recovered from  the
rolling mill uaste uater of steel mills.  Iron oxides from 3/8"
size to +400 mesh are typically found in such material.

     The range of analysis for the three major components  of the
material is as fallows:

                                      Content bv weight in %
          Oil and grease                    .5$ -

          Iron oxide particulates           45/6 -

          Uater                              4?5 - 30?£
                               614

-------
     A weighted industry average analysis is as follous:
          Oil and grease
          Iron oxide particulates           74/6
          Water
     If this material is to be used in a Sinter Plant briquetter
     or pelletizer for agglomeration, the ideal specifications
     uould be:
          Oil and grease                     Q%
          Iron oxide particulates          1QO%
          Uater
     Note:  The iron oxides in these uastes have a total Fe value
            of 72.4I&, and an oxygen value of 27.1
     A typical analysis of iron ore concentrates as mined and
processed by the principle iron ore producers in 62% Fe uith 7%
Si02» 2Q% 02, and 3$ other ingredients, including uater..
     The Ouval-Pritchard process makes possible the recovery of
the oil and grease, and the elimination of uater from these
hazardous uastes.  The solids are, therefore, an ideal source of
iron units.
     The follouing is a calculation of the results of processing
the industry average sample of material.
                             RAU FEED
          Oil and grease                   12% »  240#/Ton
          Iron oxide particulates          74$ * 1480#/Ton
          Uater                            14J6 -  280#/Ton
          TOTAL                                  2000#/Ton
                             ENERGY INPUTS
     Electrical energy in BTU's =
     3.3 KU/Ton X 3413 BTU/KU   »            11,300 BTU/Ton
          Heat Required
          a)  Distillation of expended
              liquid solvents =    105,000 BTU
          b)  Evaporation in dryer
              Cycle           -     45.000 BTU
              Total Heat                    150.000 BTU
          TOTAL ENERGY INPUT                161,300 BTU/Ton
                               615

-------
                          ENERGY CONSERVED


          Oil recovered

          240# a 18,500 BTU/#        «            4,440,000  BTU/Ton

          LJater displaced (not evaporated)

          280# ® 1,000 BTU/#         =              280.000

          Total Energy Conserved                  4,720,000  BTU/Ton

                        NET ENERGY PICK-UP

          Total Direct Energy Conserved           4,720,000  BTU/Ton

              Less Process Energy Required        - 161.300
          Net Direct Energy Pick-up               4,558,700  BTU/Ton

Additional energy savings augment the benefits of this process
are as follous:

     1.  Si02 content of ore concentrates must be eliminated in
         the Blast Furnace Slag, which requires heat.  The process
         produces concentrates uith no Si02 content.

     2«  Ulatar content in the use of conventional ore results in
         a sensible heat loss.  No uater is contained in tha
         finished product.

     3.  The elimination of other contaminants found in conven-
         tionally processed ore uses energy.  This extra energy
         is not needed in the process, as there are no other
         contaminants present.

     4.  The use of the iron concentrate from the system reduces
         the amount of ore that must be mined.

     5.  The energy required to transport mined ore is lessened
         as thajt ore requirement is reduced.

     6.  Energy is also conserved in the reduced handling- and
         managing of dumps to uhich these uastes are now being
         committed*

ECONOMICS OF THE PROCESS

     Direct values derived from each ton of rau material (oily
solids) fed to the process.   Contents:  88% solids,, 12% oil.

     Market Values           Solids    $ 27/Ton

                             Oil       S.SO/gallon
                               616

-------
     Values from Solids
                  $27 x 88#                    =      $23.76
     Values from Oil
                  .12 x 20DO#/Ton
                    6.5#/Gallon    x

     Estimated Disposal Cost/Ton

                  Total Values/Ton

     Less Estimated Process Costs/Ton

     Net Direct Benefits of the Process


     The addition of the dollar savings of using the waste instead
of disposing of it is a direct benefit of the process*  Since the
disposal costs vary uith each location, the total benefits will be
reflected by the difference in the actual disposal cost at the
specific location*
                               617

-------
                          CONCLUSION
     This paper concerns a process for steel mill use to recover
the oil and mill scale from environmentally hazardous material
produced in rolling mill operations.

     The system has no negative environmental impact*

     The value of the process is directly proportional to the
magnitude of the problem.  A large quantity of oily waste can be
converted from a costly hazard to a valuable material if the
system is utilized*  The use of hazardous uaste to make valuable
resources is a conversion of a negative to a positive factor in
all rolling mill operations • • • • a problem to a profit.
                              618

-------
SYMPOSIUM SUMMARY: Closing Remarks

Robert V. Hendriks
Symposium General Chairman
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC
                     619

-------
                             Closing  Remarks

                           Robert V.  Hendriks
              Industrial Environmental Research Laboratory
                       Research Triangle Park, NC
During the past few days we have discussed a wide range of environmental
topics relating to the iron and steel industry.  We have heard discussions
of advances in technology  that have been made,  of problems that are re-
maining, and of new problems emerging.

In his keynote address,  EPA Assistant  Administrator, Bill Drayton gave us
a brief inside look at the way regulations are developed and  described some
of the Agency's efforts to develop regulations that will ensure adequate
environmental protection at the  lowest possible  cost.  Flexibility in re-
gulations provides a  positive incentive  for  developing better pollution
control equipment and techniques.

The opening session also  gave us some insight into an area given little
attention during  last year's symposium -  innovative technology.   Mr.
Hollowaty  of  Inland  Steel  described  the  environmental  aspects of  a
continuous coking process and Mr.  Hirschhorn  gave  results of an Office of
technology Assessment  study describing potential new steelmaking tech-
nology  (particularly,  direct reduction), indicating  important  environ-
mental considerations.

In the Air Session, we saw an emphasis on fugitive  emissions, a relatively
unknown area only  a few years ago.  Several years from  now,  we will likely
have papers in areas where there is little work  today,  such as control of
volatile organics  and  development of methods to improve the  reliability of
control equipment.

In the Water  Session,  emphasis was on the major water source - coke plants.
We had  papers  on  theory, design, and  operation of  biological treatment
plants for cokemaking  wastewaters.  The most fruitful research area  for the
future appears to  be in recycle and reuse of the large  quantities of water
used in the steelmaking process.

The Solid Waste Session was the briefest of all, although certainly no less
important.   As the air  and water  problems  get  solved  and as RCRA is
implemented, a greater  emphasis  will  be put on disposing  and using the
materials removed from the waste streams.
                                      621

-------
I am  particularly gratified at  the  significant attendance at  the  sym-
posium,  despite  the  current  poor  economic  climate,  and  the  lively
discussion  that has  taken  place. This  is  a  strong indication  of  the
interest of all of us in improving  environmental  control in  the steel
industry.  The  significant  progress within  the last  year  in  controlling
pollution problems is also an indication of the rewards possible through
cooperative discussions,  planning, and research to improve environmental
control in the steel industry.  Only in this  cooperative spirit will it be
possible  to  develop  the  technology   required to  meet  the  industry's
environmental control needs  in a cost effective manner.
                                      622

-------
APPENDIX



 Attendees
   623

-------
cn
Ackermann
Adams
Alpago
Alton
Aiinantraju
Arent
Armbrust
Arnold
Arora
Ayer
Bartchy
Basinski
Beaton
Bechard
Bertick
Bernhardt
Bessent
Bhattacharyya
Bhattacharyya
Bidez
Billmyre
Blair
Brady
Brooknan
Bucchianeri
Buchko
Burcaw
Burns
Butler, Jr.
Carpenter
Centi
Chadick
Chung
Clark
Cline, Jr.
Clouse
Cochran
Cooper
Cowherd, Jr.
Coy
Craig
Craig
Crawford
Dacey
DeMa rco
Deshpande
Diaz
Dickerson
Drapalik
Draper
Draper
Dray ton, Jr.
Durrant
Kurt
Dick
Robert
Donald
Gopal
David
Robert
David
Ronesh
Franklin
Rod
Ralph
Sandra
Georges
Robert
Donald
Robert
Aniruddha
S.
W.
Richard
Thomas
Dennis
Edward
Bernard
Nicholas
Kenneth
Robert
Janes
John
T.
Bill
Neville
Robert
Raymond
Robert
Lynn
Leah
Chatten
David
Fraser
Richard
David
John
Paul
A run
Arturo
Janes
Glenn
Glenn
Roy
Williaa
John
J.

T.
E.
K.
V.
A.


A.

R.


C.
J.
A.


E.

R.
J.
T.
A.
S.

A.
J.
A.
J.

K.
J.
A.


A.

W.
L.
A.
A.
W.


Rn
H.




H.
3426 East 89th Street, South Works
101 Herritt 7, Air Correction Div.
Bethlehem Plant
35 East Wacker Drive
11499 Chester Road
6th & Walnut Streets
600 Delaware Ave.
6th and Walnut Streets
10 Chatham Road
P. 0. Box 12194
115 Gibraltar Road
900 Agnew Road
213 Burlington Road
1746 Massachusetts Avenue, N. W.
550 Pinetovn Road
3839 W. Burnhan  Street
12161 Lackland Road
Box A South Park Station
10 West 35th Street
31 Inverness Center Parkway
515> S. Harmon Street
P. 0. Box 9948
S-3556 Lake Shore Road
125 Silas Deane  Highway
Clairton Works,  Chesical Operations
200 Neville Road
Homer Research Laboratories
4 Research Place
900 Agnew Road
S. 3556 Lake Shore Road
1910 Cochran Road, C. W. Rice Div.
P. O. Box 96120, Industrial Road
50 Staniford Street
152 Floral Avenue
Wrstoo Way
Butler Works
Highway 259 South
Highway 259 South
425 Volker Blvd.
P. O. Box 12194
Queen Street West
2AIR-AF, 26 Federal Plaza
Allegheny County Airport
1010 Jorie Boulevard
3197 Independence Avenue
WO Bay Street, Air Resources Branch
P. 0. Box 46-A, Lazaro Cardenas
Gary Works, M.S.#188, P. 0. Box 59
One Brown & Root Center
1201 Elm Street, 6 AEAE
1105 North Point Blvd.
401 « Street, S. W., PH-208
One Plymouth Meeting Mall
Chicago                 IL          60617
Norwalk                 CT          06856
Bethlehem               PA          18016
Chicago                 IL          60601
Cincinnati              OH          45246
Philadelphia            PA          19106
Buffalo                 NY          14202
Philadelphia            PA          19106
Summit                  NJ          07901
Research Triangle Park  NC          27709
Horsham                 PA          19044
Pittsburgh              PA          15227
Bedford                 MA          01730
Washington              DC          20036
Fort Washington         PA          19034
Milwaukee               WI          53215
St. Louis               HO          63141
Buffalo                 NY          14220
Chicago                 IL          60616
Birmingham              AL          35243
Indianapolis            IN          46225
Austin                  TX          78766
Buffalo                 WY          14219
Wethersfield            CT          06109
Clairton                PA          15025
Pittsburgh              PA          15225
Bethlehem               PA          18016
Rockville               MD          20850
Pittsburgh              PA          15102
Buffalo                 NY          14219
Pittsburgh              PA          15220
Houston                 TX          77013
Boston                  MA          02114
Hurray Hill             NJ          07974
West Chester            PA          19380
Butler                  PA          16001
Lone Star               TX          75668
Lone Star               TX          75668
Kansas City             MO          64110
Research Triangle Park  NC          27709
Sault Ste.  Marie, Ont.  CANADA      P6A 5P2
New York                NY          10278
West Mifflin            PA          15122
Oak Brook               IL          60521
Cleveland               Oil          44105
Toronto, Ontario        CANADA      H5S IZB
Michoacan               MEXICO
Gary                    TN          46401
Lombard                 IL          60148
Dallas                  TX          75201
Baltimore               MD          21224
Washington              DC          20460
Plymouth Mooting        PA          19462
U.S. Steel Corp.
HOP, Inc.
Bethlehem Steel Corporation
Kaiser Engineers, Inc.
PEDCo Environmental
U.S. EPA, Region III
NY State Dept. of Envir. Conservation
U.S. EPA, Region III
MikroPul Corporation
Research Triangle Institute
ID Conversion Systems,  Fnc.
Jones & T.aughlin Stefl  Corp.
GCA Corporation
Canadian Embassy
JACA Corp.
Babcock. & Wilcox Co.
Envirodyne Engineers,  Inc.
Donner-Hanna Coke Joint Venture
IIT Research Institute
Combustion Engineering
Crown Environmental Control Sys.,  Inc.
Radian Corporation
Bethlehem Steel Corporation
TRC-Environmental Consultants, Inc:
U.S. Steel Corp.
Shenango Incorporated
Bethlehem Steel Corporation
NUS Corporation
Jones & Laughlin Steel Corp.
Bethlehem Steel Corporation
NUS Corporation
Armco Inc.
Metcalf & Eddy, Inc.
Wilputte Corporation
Roy F.  Weston, Inc.
Armco Inc.
Lone Star Steel Co.
Lone Star Kteel Do.
Midwest Research Institute
Research TriaugJc Institute
Algoma Steel Corp., Ltd.
U.S.  EPA,  Region I]
Energy Technology Consultants, Inc.
Aquatechnics, Inc.
Republic Steel Corp.
Ministry of the Environment
Siderurgica Las Truchafs
U.S.  Steel Corp.
Brown & Root, Inc.
U.S.  EPA,  Region VI
PORI International, Inc.
U.S.  EPA
Bfitz-Converse-Hurdock Inc.

-------
en
 Duval            L.            A.   4017 Nanway Boulevard
 Eckstein         G.            E,   119 Walnut Street
 Edwards          Moyer         B.   f. O. Box 10246
 Ehlert           Nick               140 Centennial Parkway North
 Elfstrom         Robert             550 Pinetown Road
 Ellis            Russell       G.   6200 Oak Tree Blvd.
 Ertel            Gerald             P. 0. Box 460
 Evans            Richard       R.   77 Havemeyer Lane
 Finnerty         Edward        J.   275 Broad Hollow Road
 Fitzpatrick      Marjorie           550 Pinetown Road
 Francis          Steve              1801 Crawford Street
 Fredrickson      H.            E.   One Penn Plaza
 Gogola           Gordon             3 Blue Ball Road, P. 0. Box 1100
 Goldman          Leonard       J.   P. 0. Box 12194
 Goldman          Stuart        A.   716 Oxford Valley Road
 Goonan           Thomas        G.   One Penn Plaza
 Gorman           Edmund        J.   401 M Street, S. W., EN-341
 Green            Lois               215 Fremont Street, Enforcement Div.
 Greenfield       Murray             Box 460, 1330 Burlington St., E.
 Griscom          Robert        W.   7777 Bonhomne Ave., Suite 1008
 Gronberg         Stephen            213 Burlington Road, Tech. Div.
 Gula             Robert             10 Chatham Road
 Guseman          J.            R.   6OO Grant Street, Room 1181
 Hagarmaa         James         A.   1020 Worth Seventh Street
 Haines, Jr.      George        F.   Homer Research Laboratories
 Hall             John          D.   3 Springs Drive
 Hamme            Sawny              422 River Road
 Mansen           John               Queen Street
 Hansen           Penelope           401 M Street, S.W.
 Hanson           Dennis        R.   North Point Boulevard
 Harrington       James         T.   55 Hect Monroe
 Harvey           Robert        M.   Bethlehem Plant Office, Rm. 684
 Haskill          Jin           M.   Water Pollution Control Dept.
 Hawthorne        J.            0.   125 Jamison Lane, MS-57
 Heijuegcn        C.            P.
 Kendriks         Robert        V.   HD-62, IERL
 Hirschhorn       Joel          S.   Office of Technology Assessment
 Hoffman          Albert        0.   505 King Avenue
 Hoffman          T.            W.   One NCNB Plaza
Hoffman, Jr.      Charles       F.   Research Laboratory
Nofstein         Harold             1250 Broadway, 33rd Floor
Holmes           Donald        L.   8252 Martin Tower
Holowaty         Michael       0.   30OI East Columbus Drive
Hudiburgh, Jr.   Gary          H.   401 M Street, S. V. (EN336)
Hutten-Czapski   Leon               C. P. 1000,  Usine de Contrecoeur
James            Deborah       A.   Homer Research Laboratories
Jasinski         Michael            213 Burlington Road
Jeffrey          John          D.   213 Burlington Road, Tech.  Oiv.
Josis            Ch.            R.   Rue Ernest Solvey, 11
KammenMyer      R.                 p. 0. Box 460
Kirsbner         Marvin             2400 Ardmore Boulevard
Kluth            Harry         W.   North Point  Boulevard
Koralek          Craig              1725 I Street, N. W.
Ravenna                 OH          44266
Johnstown               FA          15907
Birmingham             At          35202
Stoney Creek, Ontario   CANADA      L8E 3H2
Fort Washington         PA          19034
Cleveland               OH          44131
Hamilton,  Ontario       CANADA      N1R 6Vft
Stamford                CT          06904
Melville                NY          11747
Fort Washington         PA          19034
Middletown             OH          4S043
New York                NY          10119
Elkton                  HD          21921
Research Triangle Park  NC          27709
Yardley                 PA          19067
New York                NY          100!9
Washington             DC          22202
San Francisco           CA          94105
Hamilton,  Ontario       CANADA      L8N 3J5
St. Louis               MO          63105
Bedford                 MA          01730
Summit                  NJ          07901
Pittsburgh             PA          15230
Liverpool               NY          13088
Bethlehem               PA          18016
Weirton                 WV          26062
Conshohocken            PA          19428
Sault Ste. Marie, Ont.  CANADA      P6A 5P2
Washington             DC          20460
Sparrows Point          MD          21219
Chicago                 IL          60603
Bethlehem               PA          18016
Ottawa, Ontario         CANADA      K1A 1C8
ftonroeville             PA          15146
Ymuiden                 HOLLAND
Research Triangle Park  NC          27711
Washington             DC          20510
Columbus                OH          43201
Charlotte               NC          28280
Monroeville             PA          15146
New York                NY          10022
Bethlehem               PA          18016
East Chicago            IK          46312
Washington              DC          20460
Contrecoeur, Quebec     CANADA      JOL ICO
Bethlehem               PA          18016
Bedford                 MA          01730
Bedford                 HA          01730
4000 Liege, Bruxelles   BELGIUM
Hamilton, Ontario       CANADA      L8N 3J5
Pittsburgh              PA          15221
Sparrows Point          HD          21219
Washington              DC          20076
Colerapa  Industries,  Inc.
Bethlehem Steel Corporation
Alabama By-Products Corp,
Environment  Ontario
JACA Corp.
Davy-McKee
DOFASCO
Dorr Oliver  Inc.
United States Filter  Corporation
JACA Corp.
Arroco Inc.
Envirotech Corporation
W, L. Gore & Associates
Research  Triangle Institute
Stanford  Associates
Envirotech Corporation
U.S. EPA
U.S. EPA, Region  IX
DOFASCO
National  Engineers and Associates
GCA Corporation
MikroPul  Corporation
U.S. Steel Corp.
Calocerinas  S Spina
Bethlehem Steel Corporation
National  Steel Corporation
Keystone  Coke Co.
Algoma Steel Corp., Ltd.
U.S. EPA
Bethlehem Steel Corporation
Rooks, Pitts, Fullagar & Poust
Bethlehem Steel Corporation
Environment  Canada
U.S. Steel Corp.
Estel Hoogovens bv
U.S. EPA
U.S. Congress
Batletlp-Co)iimbus Laboratories
Midrex Corporation
U.S. Steel Corp.
Hydrotechnic Corp.
Bethlehem Steel Corporation
Inland Steel Company
U.S. F.PA
SIDBEC-DOSCO
Bethlehem Steel Corporation
GCA Corporation
GCA Corporation
Centre De Reoherches Metallurgiques
DOFASCO
Energy Impact Associates, Inc.
Bethlehem Step) Corporation
Natural  Resonrrrs Defense Council

-------
ro
 Kovacs
 Kozy
 Krocta
 Krzymowski
 Kueser
 Kulberg
 Kunskman
 Lachajczyk
 Lafreniere
 Lander
 Lefelhocz
 Lester
 Lindsay
 Linsky
 Littlewood
 LoBue
 Lower
 Luton
 Hahar
 Hal in
 Mancke
 Manda
 Marteney
 Maslany
 Mazumdar
 McCrillis
 Medwid
 Melcer
 Metzger
 Micheletti
 Middleton
 Miller
 Miller
 Moore
 Moores
 Morganti
 Moss
 Mount
 Mueller
 Mura
 Murphy
 Myers
 Neufeld
 Nicola
 Nuuno
 Oda
 Olthof
 Osantowski
 Ottesen
Parikh
Parker
Patarlis
Patton
Ernest
Michael
Harry
Cezary
Paul
Harold
Peter
Thoaias
A.
Cecil
John
Gary
Michael
Benjamin
Roy
Joseph
George
John
Kevin
Morris
Edgar
John
R.
Thomas
S.
Robert
Crady
Henryk
Daniel
Wayne
Andrew
A. Leslie
Bruce
Ben
.Charles
Ed
Mitch
David
Patrick
William
Samuel
Dennis
Ronald
Arthur
Thomas
Terry
Meint
Richard
John
Dilip
Richard
Thomas
James

J.



A.


J.

F.

F.



W.
W.

S.
B.
G.
E.
J.

C.
J.

J.
C.
C.



W.


R.
G.


H.
D.
G.



A.
A.

D.

D.
1105 North Point Blvd.
1406 Chamber of Commerce
147 E. 2nd Street
1701 First Avenue, D. A. P. C.
Box 1899
6200 Oak Tree Blvd.
Queen Street West
12161 Lackland Road
Wilcox Street
32nd Street & A.V.RR
6801 Brecksville Road
550 Pinetown Road
Tour Echelon Plara
608 Arlington
100 King St., W., Stelco Tower
10 Chatham Road
Dept. of Metallurgical Engineering
P. 0. Box 96120, Industrial Road
Box A South Park Station
P. 0. Box 2063, Fulton Building
P. 0. Box 547
20th & State Streets

6th and Walnut Streets, Curtis Bldg.
10 Chatham Road
IERL, MD-62
913 Bowman Street  (P. 0. Box 247)
Box 5050, Wastewater Tech. Centre
7777 Bonhomie Ave., Suite  1008
8501 Mo-Pac Blvd.
440 College Park Dr.
Koppers Building
345 Courtland Street, M. E.
345 Courtland Street, N. E.
One Broadway
P. 0. Box 460
442 River Road
100 King Street, W.
1129 Be11wood Ave.
4400 Fifth Avenue
U.S. 127 By-Pass South
100 S. Main, APCD
Dept. C. E. , 939 BKII
32nd Street
213 Burlington Roaii
6th and Walnut Streets, Curtis Bldg.
3185 Babcock Boulevard
5103 West Bcloit Ro,,d
f. C. 416, S.  Bedford St r»-et
10 Chatham Road
i90i Morena  Boulevard, Suite 402
4400 Fifth Avenue
10 Chatham Road
Baltimore               MD          21224
Pittsburgh              PA          15219
Mineola                 NY          11501
Maywood                 IL          60153
Pittsburgh              PA          15230
Cleveland               OH          44131
Sault Ste. Marie, Ont.  CANADA      P6A 5P2
St. Louis               MO          63141
Hamilton, Ontario       CANADA      L8N 3T1
Pittsburgh              PA          15201
Independence            OH          44131
Fort Washington         PA          19034
Vobrhees                NJ          08043
Morgantown              WV          20505
Hamilton, Ontario       CANADA      L8N 3T1
Summit                  NJ          07901
Houghton               MI          49931
Houston                 TX          77013
Buffalo                 KY          14220
Harrisburg              PA          17120
Portland               PA          18351
Granite  City            IL          62040
Fairless Hills          PA          19030
Philadelphia            PA          19106
Summit                  NJ          07901
Research Triangle Park  NC          27711
Mansfield               OH          44901
Burlington, Ontario     CANADA      L7R 4A6
St. Louis               MO          63105
Austin                  TX          78758
Monroeville             PA          15146
Pittsburgh              PA          15219
Atlanta                 GA          30067
Atlanta                 GA          30067
Cambridge               MA          02142
Hamilton, Ontario       CANADA      L8N 3J5
Conshohocken            PA          19428
Hamilton, Ontario       CANADA      L8N 3T1
Be11wood                IL          60104
Pittsburgh              PA          15213
Frankfort               KY          40601
Pueblo                  CO          81003
Pittsburgh              I'A          15261
Pittsburgh              I'A          15201
Bedford                 MA          01730
Philadelphia            PA          19106
Pittsburgh              PA          15237
Milwaukee               WI          53214
Burlington              MA          01803
Summit                  NJ          07901
San Diego               CA          92117
Pittsburgh              PA          15213
Summit                  NJ          07901
 PORI  International,  Inc.
 Koppers  Co.,  Inc.
 The Ducon  Co.,  Inc.
 Illinois EPA
 Energy  Impact Associates,  Inc.
 Davy-McKee
 Algoma Steel Corp.,  Ltd.
 Envirodyne Engineers,  Inc.
 Stelco Inc.
 Pennsylvania Engineering Corp.
 Republic Steel  Corp.
 JACA  Corp.
 United Engineers & Constructors
 A Different Air-Skyline
 Stelco Inc.
 MikroPul Corporation
 Michigan Technological University
 Arraco Inc.
 Donner-Hanna  Coke  Joint Venture
 Dept. of Environmental Resources
 Edgar B. Mancke Associates,  Inc.
 Granite  City  Steel
 U.S.  Steel Corp.
 U.S.  EPA,  Region  III
 MikroPul Corporation
 U.S.  EPA
 Empire-Detroit  Steel Division
 Environment Canada
 National Engineers and Associates
 Radian Corporation
 Koppers  Co.,  Inc.
 Koppers  Co.,  Inc.
 U.S.  EPA,  Region  IV
 U.S.  EPA, Region  IV
 Badger America, Inc.
 DOFASCO
 Keystone Coke Co.
 Stelco Inc.
 Faville-LeVally Corp.
 Mellon Institute
 Natural  Resources &  Env. Protection
 Colorado Dept. of Health
 University of Pittsburgh
 Pennsylvania Engineering Corporation
 GCA Corporation
 U.S.  EPA, Region III
 Duncan,  Lagnese nnd Associates, Inc.
 Roxnord  Corporation
 (on Physics Company
MikroPul Corporation
Air Pollution Technology,  Int.
Mellon Institute
MikroPul Corporation

-------
CO
Penrose, Jr.
Peterson
Pike
Piper
Plaks
Plenderleith
Polglase
Porter
Pretti
Price
Puchalski
Radigan
Reggi
Rice
Ridolfi
Riley
Ruppersberger
Saldanha
Schwartz
Shackleton
Shah
Sharpe
Shaughaessy
Shibata
Shi land
Shilton
Shoup
Simons
Sipe
Sokolowski
Soo
Spawn
St. Pierre
Stagias
Stebbins
Steiner
Steiner
Sterner
Stewart
Stouch
Sylvester
Symons
Szuhay
Telford
Thomas
Tonelli
Trembly
Tucker
Tucker, Jr.
Twork
Uffner
Vachon
Vajda
R.
Joseph
Daniel
Steve
Norman
Janes
William
Christopher
11.
William
Walter
Patrick
John
Michael
Janes
William
John
Geoff
Stephen
Michael
Raj
Susan
Jack
Hiroshi
Thomas
David
Stefan
Larry
Janes
Hank
S.
Peter
George
Nicholas
Lloyd
Bruce
Jim
Charles
A.
Janes
Hark
Carl
Laurence
Anton
Jean
F.
Martin
A.
William
John
Julia
Derek
Stephen
G.
C.
E,



L.
H.
G.
D.
J.



P.
J.
S.

n.
A.




W.

P.

P.

L.
D.
R.
J.
n.
A.

j.
D.
C.

8.
A.
N.
G.

C.
L.
B.
V.



900 Agnew Road
515 S. Harmon Street
145 Cedar Lane
213 Burlington Road
IERL, WD-62
P. 0. Box 460
MD-15
Room 2112, FOB, 3001 Miller Road
65 East Elizabeth Avenue
3100 E, 45th Street
4636 Somerton Road
20th & State Street
1911 Warwood Avenue
Hewer Research Laboratory
1875 New Hope Street
Martin Tower
IERL, MD-62
100 King Street, West
1000 16th Street, N.W.
485 Clyde Ave., Energy & Env. Div.
425 Volker Blvd.
IERL, MD-62
10 Chatham Road
345 Park Avenue
6th and Walnut Streets
P. 0. Box 316
3210 Watling Street, 2-110
P. O. Box 1899
Chemicals and Pigments Building
6th and Walnut Streets
144 Mechanical Engineering Bldg.
213 Burlington Road, Technology Div.
2041 N. College Road
154 Floral Avenue
P. 0. Box 8000, Four Echelon Plaza
P. 0. Box 600
485 Clyde Avenue
1465 Martin Tower
503 Queen Street East
570 Realty Road
201 West Preston Street
Homer Research Laboratories
P. 0. Box 6778
2200 Churchill Road
SoMcrton Road
Stelco Tower Phase 2
25089 Center Ridge Road
3001 Dickey Road
P. 0. Box 6778, Republic Bldg.
Hewer Research Lab., Bldg. A/0326
550 Pi netown Road
867 Lakeshore Road, Box 5050
900 Agnew Road
Pittsburgh              PA          15230
Indianapolis            IN          46225
Englewood               NJ          07631
Bedford                 MA          01730
Research Triangle Park  NC          27711
Hamilton, Ontario       CANADA      MR 6V8
Research Triangle Park  NC          27711
Dearborn                HI          48121
Bethlehen               PA          18018
Cleveland               OH          44127
Trevose                 PA          19047
Granite City            It          62040
Wheeling                WV          26003
Bethlehem               PA          18016
Norristowa              PA          19401
Bethlehem               PA          18017
Research Triangle Park  NC          27711
Hamilton, Ontario       CANADA      L8N 3T1
Washington              DC          20036
Mountain View           CA          94042
Kansas City             MO          64110
Research Triangle Park  NC          27711
Summit                  NJ          07901
New York                NY          10154
Philadelphia            PA          19106
Pueblo                  CO          81002
East Chicago            IN          46323
Pittsburgh              PA          15230
Wilmington              DE          19898
Philadelphia            PA          19106
Urbana                  IL          61801
Bedford                 MA          01730
Columbus                OH          43210
Murray Hill             NJ          07974
Voorhees                NJ          08043
Hiddletown              OH          45043
Mountain View           CA          94042
Bethlehem               PA          18016
Sault Ste. Marie, Ont.  CANADA      P6A 5P2
Monroeville             PA          15146
Baltimore               MD          21231
Bethlehem               PA          18016
Cleveland               OH          44101
Springfield             IL          62704
Trevose                 PA          19047
Hamilton, Ontario       CANADA      L8N 3T1
Westlake                OH          44145
East Chicago            IN          46312
Cleveland               OH          44101
Bethlehem               PA          18016
Fort Washington         PA          19034
Burlington, Ontario     CANADA      1.7R 4A6
PittsliuiRh              I'A          15227
Jones & Langhlin Steel Corp.
Crown Environmental Control Sys.,  Inc.
Neptune Air Pol, Inc.
GCA Corporation
U.S. EPA
DOFASCO
U.S. EPA
Ford Motor Company, Steel Division
Wheelabrator-Frye, Inc.
Republic Steel Corp.
Betz Laboratories. Inc.
Granite City Steel
WV Air Pollution Control Commission
Bethlehem Steel Corporation
PA Dept. of Environmental Resources
Bethlehem Steel Corporation
U.S. EPA
Stelco Inc.
American Iron fi Steel Institute
Acurex Corporation
Midwest Research Institute
U.S. EPA
MikroPul Corporation
Nippon Steel U.S.A., Inc.
U.S. F.PA, Region III
CF&I Steel Corp.
Inland Steel Company
Energy Impact Associates, Inc.
E. I. DuPont Company
U.S. EPA, Region III
Univ. of Illinois at tlrbana-Champaign
GCA  Corporation
Ohio State University
Wilputte Corporation
United Engineers & Constructors
Armco1 Inc.
SEA, Divi&ion of Acurex Corp.
Bethlehem Steel Corporation
The  Algoma Steel Corp., Ltd.
GAT  Consultants, Inc.
Maryland Air Quality Program
Bethlehem Steel Corporation
Republic Steel Corp.
Illinois F.PA
Betz Laboratories,  Inc.
Stelco Inc.
U.S. EPA, Region V
Jones S [.aughlin Steel Corp.
Republic Steel Corp.
Bethlehem Steel Corporation
JACA Corp.
Environment Canada
Jones & Laiighlin Steel Corp.

-------
Vakharia         Hikil         X.   P. 0. Box 6778
Van Zuidan       Dick Michael       Emmastraat $1
Voruz            Ted                2901 Butterfield Road
Waechter         Ralph         W.   600 Kossman Building, 100 Forbes Ave.
Wall, Jr.        William       T.   1901 S. Prairie Avenue
Wallace          Anna          W.   P. 0. Box 12194
Wang             Chingwen           Gen. Res. Inst. of Bldg. & Const.
Watson           Ray                4901 Broadway
Watson           Rohert        G.   2901 Butterfield Road
Waugh            John          H.   1930 Bishop Lane
Wear             Myrl          R.   P. 0. Box 600, 24 Ji. Hain
Veiubergei       Jack               20\ Scnuyl^.i.11 Kvenue
Weinzapfel       Robert        B.   10  Chatham Ave.
Westbrook        C.  Wayne           P.  0.  Box 12194
Vhitehead        Martin        F.   628 W. Parklane Tovers
Wilhelmi        A.            R,   Military Road
Wilson          William           900 Agnew Road
Wilson,  Jr.      Leon         W.   125 Jamison  Lane  -  MS 54
Winkler          Howard        A.   P.  0.  Box 6778
Withrow          Villian        '   309 W. Washington,  Suite  300
Wong-Chong       George        M.   700 Fifth Avenue
Xavier          James        F.   Environmental Control Department
Yan             Xingzhong         56  Block, Qing Shan, Hopei  Prov.
Yocom           John         E.    125 Silas Deane Highway
1'oel             W.                 4625  Roanoke Boulevard
Zachritz        W.  Howard         7777  Bonhorarae Ave.,  Suite 1008
Zuikl           James        R.   200 Neville  Road
                                                                               Cleveland               OH          44101
                                                                               Haarlera                 HOLLAND
                                                                               Oak.Brook               IL          60521
                                                                               Pittsburgh              PA          15222
                                                                               Waukesha                WI          53186
                                                                               Research Triangle Park  NC          27709
                                                                               Beijing                 CHINA
                                                                               San Antonio             TX          78209
                                                                               Oak Brook               IL          60521
                                                                               Louisville              KY          40277
                                                                               Middletovn              OH          45043
                                                                               Reading                 PA          19601
                                                                               Summit                 NJ          07901
                                                                               Research Triangle Part  HC          27709
                                                                               Dearborn                til          4S185
                                                                               Rothschild              Wl          54474
                                                                               Pittsburgh              PA           15227
                                                                               Monroeville            PA           15146
                                                                               Cleveland               OH          44101
                                                                               Chicago                 IL          60606
                                                                               Pittsburgh              PA           15219
                                                                               Sparrow Point          HO          21219
                                                                               Wuhan                  CHINA
                                                                               Wetbersfield           CT          06109
                                                                               Kansas City            MO          64112
                                                                               St.  Louis               MO          63105
                                                                               Pittsburgh              PA           15225
ro
to
Republic Steel Corp.
Ministry Nat'l Health & Pollution Cont.
Nalco Chemical Co.
PA Dept. of Environmental Resources
Envirex Inc.
Research. Triangle Institute
Ministry of Metallurgical Industry
RaiLtex
Nalco Chemical Co.
American Air Filter Co.
Armco Inc.
Wagner Associates
MikroPul Corporation
Research Triangle Institute
Ford Motor  Company
Zimpro  Inc.
Jones & Laughlin  Steel Corp.
U.S. Steel  Corp.,
Republic Steel Corp.
Illinois Pollution  Control Board
ERT Inc.
Bethlehem Steel Corporation
Ministry of Metallurgical Industry
TRC-Environ»ental Consultants,  Inc.
The Pritchard Corporation
National Engineers  and Associates
Shenango Incorporated

-------
                                TECHNICAL REPORT DATA
                         (Please read Instructions on the reverse before completing)
 i. REPORT NO.
 EPA-600/9-81-017
                                                      3. RECIPIENT'S ACCESSION NO.
4. T.TLE AND SUBTITLE Proceedings \ Symposium on Iron and
 Steel Pollution Abatement Technology for 1980 (Phila-
 delphia, PA, 11/18-11/20/80)
              5. REPORT DATE
              March 1981
              6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
                                                      8. PERFORMING ORGANIZATION REPORT NO
 Franklin A. Ayer, Compiler
9. PERFORMING ORGANIZATION NAME AND ADDRESS
                                                      10. PROGRAM ELEMENT NO.
 Research Triangle Institute
 P.O.  Box 12194
 Research Triangle Park, North Carolina 27709
              1BB610
              11. CONTRACT/GRANT NO.

               8-02-3152, TaskS
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC 27711
              13. TYPE OF REPORT AND PERIOD COVERED
              Proceedings; 11/80	
              14. SPONSORING AGENCY CODE
               EPA/600/13
is. SUPPLEMENTARY NOTES jERL-RTP project officer is Robert V.  Hendriks, Mail Drop 62,
 919/541-2547.  EPA-600/9-80-012 was the previous proceedings.
16. ABSTRACT rp^e rep0rt documents presentations at the second EPA-sponsored sympo-
 sium on iron and steel pollution abatement technology, in Philadelphia, PA, Novem-
 ber 18-20,  1980. (The first was in Chicago, IL, in October 1979.) The symposium
 provided participants an opportunity to exchange information on technology problems
 related to air, water, and solid waste pollution control in the rion and steel indus-
 try, and included a keynote  address, presentations on the environmental aspects of
 a proposed formcoke demonstration plant, and the future of steel technology and the
 environment. Sessions were conducted on: (1) air pollution abatement,  covering coke
 plant emission control, fugitive emission control, innovative air pollution control
 technology, iron and steelmaking emission control, and inhalable particulates; (2)
 water pollution abatement, covering recycle/reuse of water,  coke plant wastewater
 treatment, and coke plant wastewater new developments; and (3) solid waste pollu-
 tion abatement.
 7.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
 b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATl Field/Group
 Pollution            Aerosols
 Iron and Steel Industry
 Coking              Waste Water
 Processing          Water Treatment
 Leakage             Waste Treatment
 Dust
  Pollution Control
  Stationary Sources
  Formcoke
  Fugitive Emissions
  Particulate
13B
05C,11F
13H

14G
11G
07D
 8. DISTRIBUTION STATEMENT

 Release to Public
 19. SECURITY CLASS (This Rtport)
  Unclassified
21. NO. OF PAGES
      636
                                         20. SECURITY CLASS (TMi page)
                                          Unclassified
                          22. PRICE
EPA farm 2220-1 (9-79)
630

-------