United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
EPA-600/9-82-021
December 1982 V
Research and Development
Proceedings:
Symposium on
Iron and Steel
Pollution Abatement
Technology for 1981
-------
EPA-600/9-82-021
December 1982
Proceedings: Symposium on
Iron and Steel Pollution Abatement
Technology for 1981
(Chicago, IL, 10/6-10/8/81)
Franklin A. Ayer, Compiler
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, NC 27709
Contract No. 68-02-3152
Task No. 8
Program Element No. 1BB610
EPA Project Officer: John S. Ruppersberger
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
-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.
11
-------
PREFACE
These proceedings for the "Symposium on Iron and Steel Pollution
Abatement for 1981" constitute the final report submitted to the
Industrial Environmental Research Laboratory, U.S. Environmental
Protection Agency (IERL-EPA), Research Triangle Park, NC. The symposium
was conducted at the McCormick Inn, Chicago, IL, October 6-8, 1981.
The opening session included a keynote address, presentations on the
who, what, when, and where of environmental research in the iron and steel
industry, and the air pollution of a steel mill from an environmentalist's
viewpoint. Other sessions were conducted on air pollution abatement,
covering inhalable particulates, fugitive emission control, coke plant
emission control, innovative air pollution technology, and iron and steel-
making emission control; solid waste pollution abatement including a panel
discussion on destruction of hazardous waste in iron and steel furnaces;
and water pollution abatement, covering recycle/reuse of water, coke plant
wastewater treatment, and new developments in wastewater treatment.
John S. Ruppersberger, Environmental Health Engineer, Industrial
Environmental Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC, was Project Officer and General Chairman for
the symposium.
Franklin A. Ayer, Manager, Conference Planning Office, Center for
Technology Applications, Research Triangle Institute, Research Triangle
Park, NC, was symposium coordinator and compiler of the proceedings.
111
-------
TABLE OF CONTENTS
Page
OPENING SESSION 1
Richard D. Stern, Chairman
Opening Remarks 2
Richard D. Stern
Statement of Symposium Objectives 3
John S. Ruppersberger
Keynote Address 5
Carl J. Schafer
Whither Research? 8
E. F. Young, Jr.
The Unfinished Agenda: An Environmentalist's View of
Steel Mill Air Pollution 17
Kevin Greene
Session 1: AIR POLLUTION ABATEMENT 23
Philip X. Masciantonio, Chairman
INHALABLE PARTICULATES
Inhalable Particulate Matter Sampling Program for
Iron and Steel: An Overview Progress Report 24
R. C. McCrillis
FUGITIVE EMISSION CONTROL
Cost Effectiveness Evaluation of Road Dust Controls 39
Chatten Cowherd, Jr.,* Thomas A. Cuscino, Jr.,
and Mark Small
Blast Furnace Casthouse Control Technology - Fall
1981 Update 52
Thomas J. Maslany* and Peter D. Spawn
COKE PLANT EMISSION CONTROL
"'Denotes Speaker
iv
-------
Page
Coke Quench Tower Emissions and Emissions Control 69
Dennis Wallace,* Naum T. Georgieff, Dana Peckworth,
and Sandra Stell
Current Regulations and Control Performance for
Visible Emissions from Wet-Coal Charging, Door Leaks,
and Topside Leaks 87
Marvin R. Branscome* and William L. MacDowell
A Review of Shed and Gas Cleaning Systems for Controlling
Coke Pushing Emissions from Coke Plants 103
Jack Shaughnessy* and Dilip Parikh
INNOVATIVE AIR POLLUTION TECHNOLOGY
Armco's Experience with Application of the Bubble Concept . . . .111
B. A. Steiner
Engineering Study of Roof Mounted Electrostatic Precipitator
(REP) for Fugitive Emission Control on Two Basic Oxygen
Furnaces of 300 Ton Capacity 120
Richard Jablin and David W. Coy*
Demonstration of the Use of Charged Fog in Controlling
Fugitive Dust from a Coke Screening Operation at a
Steel Mill 139
Edward T. Brookman,* Kevin J. Kelley,
and Robert C. McCrillis
IRON AND STEELMAKING EMISSION CONTROL
Performance of BOF Emission Control Systems 154
Leonard J. Goldman,* David W. Coy,
James H. Turner, and John 0. Copeland
Investigation of Opacity and Particulate Mass Concentrations
from Hot Metal Operations 174
David S. Ensor
Retrofitting Emission Controls on Electric Furnaces
at a Steel Mill 187
Michael P. Barkdoll* and Donald E. Baker
The Present and the Future for the Industrial Treatment of
Fumes in the French Steel Industry 202
Jacques Antoine,* Alain Milhau, and Jean Raguin
Modeling of Hood Control of Blast Furnace Casting Emissions . . .216
S. F. Fields,* C. K. Krishnakumar, and J. B. Koh
"'Denotes Speaker
-------
Session 2: SOLID WASTE POLLUTION ABATEMENT 231
Thomas M. Barnes, Chairman
Robert S. Kaplan, Cochairman
RCRA Regulatory Changes and the Steel Industry 232
Stephen A. Lingle* and William J. Kline
Characterization, Recovery and Recycling of Electric
Arc Furnace Dust 246
N. H. Keyser,* J. R. Porter, A. J. Valentino,
M. P. Harmer, and J. I. Goldstein
Treatment of Carbon Steel Electric Furnace Flue Dust
by Sulfation 261
Pinakin C. Chaubal, Thomas J. O'Keefe,
and Arthur E. Morris*
PANEL: DESTRUCTION OF HAZARDOUS WASTE IN IRON AND STEEL
FURNACES 279
Engineering Requirements for Thermal Destruction
of Hazardous Waste in High Temperature Industrial
Processes 279
E. Timothy Oppelt
Suitability of Open Hearth Furnaces for Destruction of
Hazardous Waste 279
William F. Kemner
Suitability of Blast Furnaces for Destruction of
Hazardous Waste 280
George R. St. Pierre
Session 3: WATER POLLUTION ABATEMENT 281
Terry N. Oda, Chairman
RECYCLE/REUSE OF WATER
Minimizing Recycled Water Slowdown from Blast Furnace
Gas Cleaning Systems 282
Richard L. Nemeth* and Leonard D. Wisniewski
Minimizing Water Slowdowns from Selected Steel
Plant Processes 300
Harold J. Kohlmann* and Harold Hofstein
^Denotes speaker
VI
-------
Zinc Control in a Blast Furnace Gas Washwater
Recirculation System 307
R. Gregory Elder"" and Roy Littlewood
Investigation of Reverse Osmosis for the Treatment
of Recycled Blast-Furnace Scrubber Water .320
M. E. Terril* and R. D. Neufeld
Review of Water Usage in the Iron and Steel Industry:
Blast Furnace and Hot Forming Subcategories 344
Albert P. Becker, III, and Thomas M. Lachajczyk*
Control of Scale Formation in Steelplant Water
Recycle Systems 362
G. R. St. Pierre* and A. H. Khan
Reduction of Wastes Discharged from Steel Mills in
Metropolitan Chicago through Local Ordinance Enforcement 373
Richard Lanyon* and Cecil Lue-Hing
A Mass Balance Model for Rinsewater in a Continuous
Strip Halogen Electrolytic Tinning Operation for Use
in Evaluating Wastewater Treatment and Recovery
Alternatives 388
James Skubak* and Ronald D. Neufeld
COKE PLANT WASTEWATER TREATMENT
Investigation of the Solid-Liquid Phase Separation '
of Preheated and Pipeline Charged Coke Battery
Charging Liquor 403
S. R. Balajee,* A. I. Aktay, and R. R. Landreth
Assessment of the Biological Treatment of Coke-Plant
Wastewaters with Addition of Powdered Activated Carbon
(PAC) 424
Leon W. Wilson, Jr.,* Bernard A. Bucchianeri,
and Kenneth D. Tracy
Biological Treatment of By-Product Coke Plant Wastewater
for the Control of BAT Parameters 446
G. M. Wong-Chong* and S. C. Caruso
Two-Stage Biological Fluidized Bed Treatment of
Coke Plant Wastewater for Nitrogen Control 460
S. G. Nutt, H. Melcer,* and J. H. Pries
NEW DEVELOPMENTS IN WASTEWATER TREATMENT
"'Denotes speaker
vii
-------
Page
Trace Metal Removal from Steel Plant Wastewaters Using
Lime and Ferrate 476
J. A. FitzPatrick,* J. Wang, and K. Davis
Pilot Evaluation of Alkaline Chlorination Alternatives
for Blast Furnace Blowdown Treatment 493
Stephen A. Hall,* Karl A. Brantner, John W. Kubarewicz,
and Michael D. Sullivan
Closing Remarks 508
John S. Ruppersberger
UNPRESENTED PAPERS 509
Application of Second Generation Chemical
Solidification/Fixation Processes to Iron
and Steel Hazardous Waste Problems 510
Jesse R. Conner
Optimizing Existing Wastewater Treatment Facilities
in Preparation of Meeting BAT/BCT Regulations in the
Iron and Steel Industry 525
Meint Olthof
Development of a Deoiling Process for Recycling Millscale . . . .536
Derek S. Harold
Recycling of Tar Decanter Sludge 543
R. B. Howchin and M. S. Greenfield
APPENDIX: ATTENDEES 546
^Denotes speaker
viii
-------
OPENING SESSION
Chairman: Richard D. Stern
Industrial Environmental Researeh^taboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
OPENING REMARKS
Richard D. Stern
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC
Good morning and welcome to the third symposium on Iron and Steel
Pollution Abatement Technology. My name is Richard Stern and I am Chief
of the Industrial Processes Branch at EPA's Industrial Environmental
Research Laboratory at Research Triangle Park, North Carolina.
I am particularly pleased to welcome you all this morning since I
am relatively new in this industry. About six months ago the Laboratory
underwent a reorganization and the ferrous metallurgical industry,
•including iron and steel, was transferred into my area of responsibil-
ity. Norm Plaks, who headed up the branch previously, was transferred
into Chief of the Particulate Technology Branch and is sorry he could
not be here. However, he sends his regards to the friends and col-
leagues that he has acquired over the thirteen years he has worked with
you in this industry.
Although I have only been involved with the industry a short time,
I have been impressed and encouraged by the cooperation I have seen
between us in the EPA's research and development program and you in the
iron and steel industry. I have met a number of you these last few
months, and I am looking forward to meeting many more of you during this
symposium and possibly working with some of you on potential future
cooperative projects. Regarding those projects, close coordination to
select and focus on key areas of effort will be imperative. In view of
declining resources, we will all have to zero in on very select, high-
priority projects. I am looking forward to meeting many of you in the
next few days and working with you.
Now I would like to introduce your Symposium Chairman, John
Ruppersberger.
-------
STATEMENT OF SYMPOSIUM OBJECTIVES
John S. Ruppersberger
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC
I would like to add my welcome to this, the third symposium on Iron
and Steel Pollution Abatement, sponsored by EPA's Industrial Environ-
mental Research Laboratory at the Research Triangle Park. The symposium
provides a forum for the exchange of information on iron and steel
multi-media pollution abatement technology. This symposium continues as
a high priority activity due primarily to the opportunity it provides
for learning and the cooperation encouraged by gathering members of
industry, government, research and engineering organizations, universi-
ties, associations, control equipment designers and vendors, and others
interested in pollution abatement for the iron and steel industry. The
symposium provides an opportunity to learn, not only from what is pre-
sented by the speakers, but from questions and other discussions. We
need to learn from each other. The results can be better solutions to
pollution abatement problems. These solutions can take the form of more
efficient and cost-effective technology, more productive 'research and
engineering, and improved equipment design and operation.
This is an EPA office of research and development symposium, only
to the extent that we have sponsored it, and we have worked to make it
happen. This year only about one-fourth of the papers are from work
funded by EPA-ORD. Our prime function is to serve as a catalyst. We
seek to provide better information; that is, not just more information,
but rather more accurate, more pertinent, and more practical information
in a well-engineered sense. It is good to have greater participation in
this symposium from others, especially from industry, due to the simple
fact that any real achievements in pollution abatement are accomplished
there.
After the opening session, the papers are grouped by media--air,
solid waste, and water. Solid waste is getting more emphasis this year
to reflect its increased importance. Many of the papers are actually
multi-media, but were placed in the media of principal concern. Several
of these papers contain the broader perspective of multi-media impacts
that is essential in achieving effective pollution abatement technology.
This is an open forum--we encourage your active participation and a
variety of viewports. Time for questions has been scheduled at the end
of each paper. However, since this is a technical symposium, please
-------
save any non-technical questions and discussions for breaks and other
times. The papers represent the views of their authors. They do not
represent EPA policy. Please ask your questions—they contribute much
to the symposium. We are here to learn from each other. I welcome you
all and hope you benefit from and enjoy this symposium.
-------
KEYNOTE ADDRESS
Carl J. Schafer
Industrial and Extractive Processes Division
U.S. Environmental Protection Agency
Washington, DC
As befits one from Washington, DC, nowadays, I think I should pro-
bably keep my own remarks brief in deference to Earl and Kevin, who I
think probably have more to say than I do. The fact that this is the
third symposium on Iron and Steel Pollution Abatement Technology, that we
have been working together that many years, that we have got that much
research going on to report on, is significant in its own right. It
occurred to me to look back at some of the people and some of the projects
that we have been involved in as I was thinking about the theme that we
might adopt for this third symposium on pollution abatement in the iron
and steel industry.
Our cooperative efforts began more than a decade ago. Several pro-
jects were initiated in the late 60"s; for example, combined treatment of
steel mill and municipal wastewater which is, of course, now common prac-
tice in several areas of the United States and, especially, right here in
Chicago, was studied with National Steel Corporation with the help of Bill
Smith of that company. Countercurrent rinsing on the halogen tin plate
line at National Steel Corporation's Weirton Plant demonstrated the large
blowdown reduction and accompanying economic benefit derived by recovery
of the tin previously lost to the system. More recent cooperative efforts
include several projects to encourage increased process water reuse/re-
cycle. One study is of low cost modifications for blowdown reductions.
The efforts of Don Lang at Inland Steel and Bob Peterson at U.S. Steel,
Baytown, are stated in this study. Other work, of course, involves our
mobile wastewater treatment trailer which is currently at Republic Steel
in Cleveland on blast furnaces 5 and 6. Dick Nemeth and Len Wisniewski of
Republic Steel will present a paper on this project later on in the ses-
sion about the opportunities to significantly reduce blowdown. The air
portion of our program with the iron and steel industry has also relied on
cooperative projects, beginning in 1970 with the initiation of the Jones
and Laughlin Pittsburgh work, P-4 battery, coke oven charging demonstra-
tion. In this project we worked, of course, with Earle Young. The pro-
ject served as a proving ground for many innovative technologies, some of
which are in common use today.
Another more recent EPA/AISI project coordinated with the AISI Coke
Practice Committee under Don Gregg concerned the development and demon-
stration of improved coke oven door seals. Major field work was done with
Jim McCord at Bethlehem's Lackawanna works and with Lloyd Hoopes at
-------
Republic, Youngstown. Most current projects depend on industry cooper-
ation in the form of providing access to facilities for the purpose of
emission quantification and to support these activities through install-
ation of sampling facilities, etc. A good example here is the work Murray
Greenfield of Dominion Foundries and Steel Company on inhalable particu-
late emissions. The work there was in a cast house and the coke quench
tower. Armco's support of similar tests of open source EOF and electric
arc furnace emissions at three of their plants is another example of
EPA/industry cooperation. Bruce Steiner of Armco will be presenting a
paper on this work this afternoon as well.
A significant area of cooperation between EPA and AISI is through our
activities involving the American Iron and Steel Institute. In the past
we have enjoyed an excellent relationship through Bill Benzer and we are
continuing that relationship today in the same spirit with Steve Schwartz.
The AISI organizes several committees composed of members from corpor-
ations that belong to AISI to meet with the EPA personnel to coordinate
major program areas and work together on specific projects. The AISI and
the EPA have cosponsored several research projects at various universi-
ties. Particularly satisfying is the support that we have received from
AISI on our program with other nations, especially the Soviet Union where
we have an active exchange program on iron and steel technology. Over the
.past seven years, we have been fortunate to have industry experts such as
Bill Benzer and Len Wisniewski as members of the team during visits to the
Soviet Union. AISI and many member companies, including U.S. Steel,
National Steel, and Inland Steel, have served as gracious hosts during
visits of Soviet specialists to the United States. The information ex-
change was mutually beneficial to both sides.
There are many opportunities now to develop better solutions to the
problems that we are all aware of. The need for more secure handling and
disposal of hazardous waste has spurred renewed emphasis on prevention,
recovery, or other more cost-effective alternatives to disposal. Often
these alternatives to disposal include more effective utilization of
resources or resource recovery. These alternatives seek to be more cost
effective. They may produce a profit. These alternatives are not only
good environmental solutions, they are good business. The AISI Task Force
on Solid Waste, as well as ourselves, have listed several projects of
mutual interest, but also projects which promise mutual benefit. Many of
these projects are in the area of solid waste. Especially, I would like
to note that the solid waste session of this symposium does include a
paper on electric arc furnace dust, one of the AISI Task Force initiated
projects. This project is of a special interest because the AISI working
through its member companies, through agencies of the U.S. Government,
surveyed the range of areas available and identified this particular
project as being the one of particularly high potential for a payoff that
would benefit the companies and, if we make it work, would benefit the
environment through the alleviation of this solid waste disposal problem.
Of course, the other media--air and water—also include opportunities for
developing mutually beneficial cost-effective technology. These include
the charged fogger paper which you will hear in the air session and the
water reuse/recycle papers in the water session. One of our recently
-------
completed wastewater studies involved meetings with foreign steel plants
and regulatory personnel of ten foreign governments from all over the
world. One thing noted by our participants in these meetings was a spirit
of cooperation between the industry and the government. The industry
seemed willing to openly discuss its pollution problems and the government
seemed willing to work toward national, well-engineered solutions to these
problems. This spirit of cooperation between industry and government is
one that we have enjoyed and that we wish to continue to engender through
symposia such as this and other cooperative programs. There is a lot yet
to be done; we have many opportunities. Technical people working together
to achieve technical solutions can achieve the critical mass necessary to
solve these problems reasonably balancing environmental concerns with our
economic resources. The name of the game is, as it has been, cooperation.
Reviewing the history, then, we have seen plenty of that cooperation.
This cooperation now is not so much an institutional, dry, bureaucratic
cooperation as it is the cooperation of people working together. Just in
naming a few of the people with the steel companies that have been working
on these projects with our own people brings this sharply to mind. The
conference is a forum for those people who are working on individual
projects who may not have the opportunity to know of the other projects,
to crossfertilize each other. The forum is one to engender increased
cooperation, increased technical excellence, increased technological
solutions to problems. Let us keynote for this symposium then the cooper-
ation and working together between EPA and a revitalized iron and steel
industry as we go ahead.
Thank you very much.
-------
WHITHER RESEARCH?
by: E. F. Young, Jr.
Vice President, Energy and Environment
American Iron and Steel Institute
1000 Sixteenth Street, N.W.
Washington, DC 20036
ABSTRACT
This paper reviews the WHO, WHAT, WHEN, and WHERE of environmental
research in the iron and steel industry, and makes projections and sugges-
tions as to future conduct of research'in the field.
It is a real pleasure to be one of the lead-off speakers at the third
annual symposium on Iron and Steel Pollution Abatement Technology. I
began, at the first of these symposia, with a review of the many different
lines of technology development—research, development, and demonstra-
tion--on pollution control in the iron and steel industry. Now that the -
third symposium is here, I would like to take an overall look at what we
have done, where we are, where we are going, and where we should be going.
And when I say "we," I mean all of us, EPA, industry, academia, and inde-
pendent consultants and researchers. We are all working toward the common
goal of an improved environment. And we all have to recognize that achiev-
ing that improved environment is a matter not just of technology but also
of dollars and cents. With that thought in mind, I would like to review
what has been covered in past symposia and is being covered in this one. I
would like to look for trends and to project where we should be going in
the future.
As a first step in trying to analyze where we have been and where we
are going, I took a look at the subject matter (Table i) of the papers from
these first three symposia. In 1979 50% of the papers dealt with air,
-------
TABLE 1. SUBJECT MATTER
Air
Water
Solid Waste
Multimedia
Percentage
1979
50
25
21
4
1980
42
39
13
6
1981
38
33
17
12
25% with water, 21% with solid waste, and a mere 4% with what I call multi-
media assessment. In 1980 papers on air pollution control dropped off to
42%, water pollution control was up to 39%, solid waste dropped to 13%, and
multimedia studies were up a little at 6%. This year's program involves
38% on air pollution, 33% on water pollution, 17% on solid waste, and 12%
on what I consider multimedia analysis. One conclusion that could be drawn
from this table is that it simply represents what papers were available for
presentation in a given year. Another would be that the planners of the
program had attempted to organize it a little differently. But I think
there is something a little deeper to be drawn from this analysis, and I
hope a guideline for the future; that is, the growth of multimedia papers.
We tend all too often to look at narrow problems in isolation and yet, as
time goes on, we realize more and more that, in cleaning up the water, we
tend to produce solid wastes which must be disposed of. And in disposing
of solid wastes, we run into the danger of contaminating both air and
water. Since the real goal of our efforts is not just to achieve clean air
and not just to purify water but to assure an improved overall environment
for man, it seems to me that we need more analyses which consider all
aspects of the actions that we take. We need more work which considers not
a single medium but the overall impact of various control actions on all
aspects of the environment. I think the little trend chart we have here
shows that we have started giving more consideration to overall environ-
mental impacts, and I hope that this presages more analyses of this type
and more looking at the overall environment in the future.
The next question I asked myself was, who has been doing the work?
So, again, I reviewed the three symposia to date (Table 2) and I found that
the first was very much dominated by EPA and its contractors. Between
-------
TABLE 2. SOURCE OF PAPERS
EPA
EPA Contractor
Steel Company
Joint
Other
Percentage
1979
9
58
22
4
9
1980
10
41
26
10
13
1981
2
37
31
9
21
them, they gave 67% of the papers. That dropped off in subsequent years to
51% at the second and 39% at the third. The steel companies themselves
presented 22% of the papers at the first, 26% at the second, and 31% at the
third. Joint research between EPA and the industry accounted for only 4%
of the papers at the first, 10% at the second, and 9% at the third. Papers
from other sources went from 9% at the first to 21% in the current sympo-
sium. What is significant here? EPA domination of the first symposium was
to be anticipated, I think, since EPA set the symposium up, and pulling the
first one together was a big job. So, they had to rely on themselves and
their contractors. That first meeting was a success, so papers from other
sources have poured in, and EPA has proportionately cut back its domina-
tion. Industry has increased the number of papers it is presenting here.
I do not think that this is an indication of increasing work on the part of
the companies, but rather increasing recognition that this symposium is an
important place to present the work. Jointly-sponsored work between in-
dustry and EPA has increased but still represents a rather modest propor-
tion of the papers presented. And others have increased significantly.
Who are these others? Some are guests from other countries, and I think
this is a very sound and good development: the problems of pollution
control are problems of the world, not just of this country; it is highly
desirable to have an international flavor to these symposia. And others
represent work by contractors who have seen opportunities and markets in
the steel industry and are using this as an opportunity to present the sort
of technical information that can permit their work to be evaluated.
There is one thing these particular statistics do not show, however,
that I think is important. These statistics show only a small proportion
of the papers represent joint work between EPA and the steel industry.
In reality, however, there is a great deal of cooperation between the
industry and the Agency. I think this cooperation is essential and should
be stressed to a greater extent.
10
-------
When EPA selects a contractor and puts him to work on a study of some
aspect of our industry, that study cannot properly be done in isolation.
EPA recognizes this and in almost every case, once a contractor has been
selected for a significant study of the steel industry, EPA has approached
the American Iron and Steel Institute and asked for our technical assist-
ance in carrying out the project. This is normally followed up by a meet-
ing among EPA, the contractor, and AISI technical people to discuss the
project prior to its inception. During the conduct of the project, occa-
sionally there are progress meetings. Then, as reports are prepared, AISI
will work with the contractor and EPA on the assembly of a final report.
AISI's advice and comments are, of course, not binding on EPA or its con-
tractor, nor are they meant to be. This is a case of technical cooperation
between the industry and the Agency.
I believe that this voluntary technical cooperation is of great ad-
vantage to all parties involved. The industry is much more likely to agree
with results of the studies when we have had an opportunity to make input
as that study develops. The contractor is much less likely to start down
unproductive lines when he has the direct technical input of the industry
experts most familiar with the operations. The Agency then gets a better
work product as a result.
My next effort at analyzing where we have been consisted of classify-
ing the papers by types (Table 3). I think I should emphasize at this
TABLE 3. TYPES OF PAPERS
Regulatory Analyses
Technical Review
Quantification of Emissions & Effluents
Laboratory Conceptual Studies
Pilot Studies
Demonstration
Environmental Assessments
Cost/Effectiveness/Cost/Benefit Analyses
Percentage
1979
12
12
22
8
17
17
8
4
1980
6
13
10
30
16
19
6
1981
11
14
8
31
19
17
11
-------
point that this is a totally arbitrary breakdown based solely on my think-
ing. You might not agree with the categories I have established; you might
not agree with what categories I have placed some papers in. Nonetheless,
I hope I have been able to come up with some useful analysis.
I set up a number of categories: Regulatory Analysis, Technical
Review, Quantification of Emissions and Effluents, Laboratory and Concep-
tual Studies, Pilot Studies, Demonstrations, Environmental Assessments, and
Cost/Effectiveness/Cost/Benefit Analyses. Before I discuss what the sta-
tistics or trends are in these categories, perhaps I should explain what is
meant by each category and maybe give you some examples of papers in each.
The first is Regulatory Analyses. Since the nation's regulatory
programs determine so much of what is required in the area of pollution
control, an essential part of the background of any thorough look at re-
search is an analysis of the regulations and what is required under them.
Papers fitting in this category would be typified by one of the first
papers in the first symposium, Don Goodwin's presentation, "Air Pollution
Emission Standards." A more recent example would be the paper being given
tomorrow on "Impact of RCRA.on Solid Waste from the Steel Industry." I
think these papers are invaluable in that it is the regulatory programs
that guide the nation's environmental control program. Thus, the technol-
ogy needed is a function of these regulatory analyses.
The next category I set up was Technical Review. Again, this is,
hopefully, a pretty straightforward category. It comprises those papers
which look around and see what technologies are available and in use today.
Papers in this category range from "Review of Foreign Air Pollution Control
Technology for BOF Fugitive Emissions" presented at the first symposium up
to "An Air Pollution Control Equipment Inventory of the U.S. Steel Indus-
try" which will be presented tomorrow. These reviews are also an important
part of symposia on technology. The last thing we need is reinvention of
the wheel or a failure to recognize that someone already has solutions to
particular problems.
The third category I have selected is Quantification of Emissions and
Effluents. I think this is an extremely important type of paper in that a
knowledge of what pollution is emitted is extremely important in estab-
lishing regulatory programs and in establishing the needs for technology
and the effectiveness of technology. I think the classification is clear
enough; I do not need to point out examples. There have been a number in
each of the three symposia.
My next category is Laboratory and Conceptual Studies. This is an
important area because ideas have to start somewhere. There has to be some
preliminary analysis to determine whether a process or technology has any
potential for solving one of the environmental problems that have been
defined. Clearly we need basic science and we need basic thinking. The
types of papers that I have included here range from "Formation and Struc-
ture of Water Formed Scales" presented at the first symposium to the panel
discussion at the current symposium on "Destruction of Hazardous Wastes in
Iron and Steel Furnaces." As I say, these studies are quite important
12
-------
because it is clear that there is much room for new technology in the
environmental control area. On the other hand, they generally represent
technology which is a long way from utilization on an industrial scale, and
a long way from providing the basis for realistic regulations applicable to
the industry.
My next category is Pilot Studies. This is another pretty obvious
title. It covers the next step from successful laboratory or conceptual
work to trials on small scale to demonstrate the technical feasibility of
the ideas. Again, the titles of the papers make these things obvious, so I
will not cite any examples.
The next category is one that is dear to my heart: Demonstration. I
believe that the most difficult part of the development of technology for
environmental control in the steel industry is making things work. Because
of the scale and difficult conditions encountered in the steel industry, no
technology can be considered acceptable until it has been demonstrated and
made to work on full-scale plant installations. I think that the major
technical effort that has been expended by the steel industry over the
years in environmental control has been in this area. Papers that would
clearly be covered range from the "Coke Oven Door Seal Demonstration" in
the first symposium to the paper, "Minimizing Recycled Water Slowdown from
Blast Furnace Gas Cleaning Systems," being presented at this symposium.
From the standpoint of disseminating knowledge on what can and cannot be
done, and on how to do it and what it costs to do it, I think perhaps this
is the most important category of papers that can be presented at a sympo-
sium such as this.
My next category is Environmental Assessments. An environmental
assessment in my opinion is a study which looks at all environmental as-
pects of a process and, out of this, perhaps, develops an idea of what
should and what should not be done in the way of regulation of specific
environmental aspects of the process. There were a couple of papers at the
first symposium clearly defined as environmental assessments. In last
year's symposium, there was an "Environmental Appraisal of Reclamation
Processes for Steel Industry Ironbearing Solid Wastes." I think the key
factor of an environmental assessment is the multimedia approach; that it
looks, not just to what are the emissions from a process, or what are its
effluents, or what are its solid wastes, but it attempts to look at all
media, all environmental effects, and in this way put the process in per-
spective.
My final category is one that I feel is extremely important today;
that is, Cost/Effectiveness/Cost/Benefit Analyses. The terms, I think are
self-defining. The importance of them I will talk of later. And the
example I will present is one paper at the first symposium, "Coke Battery
Environmental Control Cost Effectiveness."
Now that I have described my categorization of the various types of
papers, I would like to look at how many papers fit into each of these
categories. I would like to emphasize at this point that these categories
are arbitrary, selected by me, and the assignment of papers to various
13
-------
categories involved a judgment in many cases. Nonetheless, I hope I have
developed a reasonable breakdown of what has been presented.
Regulatory analyses started as a pretty significant part of the pro-
gram, then dropped off, then increased again. Technical Reviews have
remained at about the same level for the three years. Quantification of
Emissions and Effluents started as a very significant part of the sympo-
sium, but has been dropping off ever since. I think that makes a great
deal of sense. I think maybe it indicates that we have learned what the
emissions and effluents from our processes are, so emphasis has turned from
finding out what they are to doing something about them. Laboratory and
Conceptual Studies started at a pretty low level, and then jumped to be a
major part of the symposia. Perhaps this ties in with the previous item,
that, now that we have got some problems well defined, it is time to come
up with ideas for solving them. Pilot studies have remained relatively
constant as a significant part of the symposium. Demonstrations have re-
mained pretty much constant as a significant part of the symposium. The
last two items, however, are different. Environmental assessments have
dropped off. I think this represents a bad trend. I think it reflects
thinking which has become compartmentalized. I think we are looking at too
many problems in isolation. I think each regulator has been looking at
narrow problems. And I think this narrow viewpoint is a mistake. Finally,
Cost/Effectiveness/Cost/Benefit Analysis started at a very low level and
has just disappeared from more recent symposia.
That pretty much concludes my analysis of where we have been of the
who, what,, where, to some extent the why of where we are today. Now I
would like to look to the future within this same framework and project
where I think we ought to go, and maybe give you an idea of what I think
future symposia might best involve.
There are a number of things that we have to consider in projecting
where we should go. One is the status of the nation's accomplishments at
this point. We are now at the point where primary standards for air have
been met in most areas, and are projected to be met in almost all areas
within the next couple years. We are now at the point where essentially
all of industry has installed Best Practicable Technology for water pollu-
tion control, and will have installed within a few years Best Available
Technology for water pollution control. The recently issued report by
Arthur D. Little for AISI, "Environmental Policy for the 1980's: Impact on
the American Steel Industry," shows that the steel industry has installed
equipment for control of 95% of its particulate emissions and will shortly
have installed equipment for control of 96% of its emissions. Facilities
installed or under construction will remove 91% of our water pollutants and
within a few years will be removing better than 98% of our water pollut-
ants. I think these factors must be considered in looking at future dis-
cussions of pollution control technology.
The Clean Air Act and Clean Water Act are both due to reauthorization
and perhaps some fairly extensive revision in the near future. Changes in
these laws, of whatever nature, will lead to regulatory changes and will
lead to different requirements on the technology for the industry.
14
-------
The ambient air standards, and specifically the standard for particu-
late matter, are under review right now and we can anticipate a major
change in the near future. It appears that the controlled particulate will
be changed from total suspended particulate to some finer particulate
fraction. This type of change in standards will be accompanied by major
upheaval in the type of regulations enforced, and possibly in the technol-
ogy required to achieve the new standards.
We have in this country today a new administration dedicated to regu-
latory reform. Much of the regulatory reform will be directed at more
cost/effective, cost/beneficial regulations. As these regulatory reforms
are implemented, and I am sure they will be, it will be essential that we
know more about the cost/effectiveness and cost/benefit of regulations on
the books and of forthcoming regulations.
I have covered very quickly a whole series of factors which I think
change the way we will be looking at iron and steel pollution abatement
technology in the future. In light of all these factors, I would like to
go back to a couple of my tables and say, here is what I think the emphasis
should be in future symposia.
First of all, as to subject matter (Table 1): I anticipate there will
be a change in the ambient air standard which will require changed emphasis
in our approach to air pollution control for the industry. On water pollu-
tion control, I think we are approaching today the definition of Best
Available Technology and when we have that defined, there will be less need
and less interest in water pollution control technology developments. As
the programs for solid waste control under RCRA become better defined, we
will need to continue a significant emphasis on solid waste, but the big
area of interest for the future I feel has to be on multimedia pollution
control, on the impact of one type of pollution control on another, and on
improving our overall environment, so I am looking for a major increase in
multimedia papers.
As to types of papers (Table 3), I anticipate major changes in the
regulatory patterns in this country for three reasons:
1. Because we are already approaching many of our goals.
2. Because I anticipate significant regulatory reform.
3. Because there will be changes in the Air and Water Acts.
Therefore, there will be a continuing need for further regulatory analyses.
I think there will be less need for technical reviews as time pro-
gresses. Progress is not so fast and changes are not so frequent that
technical reviews are needed on a continuing basis. Quantification of
emissions and effluents will change. We have developed a good idea of what
our emissions and effluents are considering present standards and present
knowledge. However, as the direction of regulatory control changes — for
example, if we go to control of total thoracic particulate rather than
15
-------
total suspended particulates--we will need to know what are the emissions
from the industry. So, there will be continuation but a change of emphasis
in the area of quantification. As to laboratory and conceptual studies, I
think we have, basically, the technology today to control pollution from
the iron and steel industry. New ideas may come along and be developed,
but we do not need the intensive search for methods we have had in the
past. So, we may hear less of these in the future. There may be less on
pilot studies for the same reason. Our problems will continue to be the
application of technology.
Demonstration will continue to be the major issue. The application of
technology will continue to be the principal technical problem of the
industry and, therefore, I think the emphasis on demonstration must and
should—and will—continue. As to my last categories, environmental
assessments and cost/effectiveness/cost/benefit analyses, I think the time
has come when these should be the major emphasis of our looking to the
future. I think with a framework of air, water, and solid waste laws, we
will be in a better position to assess the overall environmental impacts of
any process change or any pollution control directed at one medium, and I
think we will need more of this to achieve an adequate balancing of the
overall needs of the environment.
As to cost/effectiveness/cost/benefit analyses, I think the time has
come. In a time when the country is plagued with problems of energy,
unemployment, inflation, and other new problems, we cannot continue to look
on the environment as a goal set aside from all others. We must look at it
in the perspective of social needs. The way to do this is by making sure
when we do things that we are getting the biggest bang for the buck, and
that the bucks we are spending are achieving something worthwhile. Now is
the time to devote our intelligence and pur energy to realistic cost/effec-
tiveness/cost/benefit analyses in the environmental area.
One last point I would like to make is the question of who is to do
the work in the future. We have talked about EPA, their contractors, steel
companies, and others doing work, and we discussed the need for joint,
cooperative work between EPA and the industry. I think there has been good
cooperation in the past and I discussed this earlier. But I think there is
one fault with what has been done in the past and one area where we must do
better in the future. In the past, EPA has set up their programs and
defined their objectives, and then sought assistance from AISI in carrying
out these programs. We were very glad to give that assistance. But, back
in school, in teaching us the scientific method, they always started with,
first, define the problem. And I think that is one area where joint effort
between industry and EPA can and should guide the programs of the future.
I would like to see more joint input to research planning, in the analysis
of what needs to be studied, and in the analysis of how to study those
demnstrated needs. I think I can tell EPA that AISI stands ready to work
with them to develop more effective programs to solve the technology prob-
lems of pollution control in the industry, not just from the Agency view-
point, not just from the industry viewpoint, but, hopefully, from the
nation's viewpoint.
16
-------
THE UNFINISHED AGENDA: AN ENVIRONMENTALIST'S
VIEW OF STEEL MILL AIR POLLUTION
Kevin Greene
Research Associate
Citizens for a Better Environment
59 E. Van Buren
Chicago, Illinois 60605
ABSTRACT
The steel industry has made substantial progress in controlling air
pollution. Particularly noteworthy are efforts to control coke oven
emissions in a comprehensive manner. In addition, several companies are
developing non-capture systems to control fugitive emissions from iron and
steelmaking facilities.
But the job is not done. Total suspended particulate levels still
exceed the primary standards in major steelmaking areas of the country by
a significant margin. Therefore, additional controls on process sources
will be necessary. Attainment of the standards will also depend upon
achieving good work practices. To make sure such practices become part
of the daily operating routine, self-monitoring programs should be
instituted at steel mills.
17
-------
THE UNFINISHED AGENDA: AN ENVIRONMENTALIST'S
VIEW OF STEEL MILL AIR POLLUTION
If you had picked up the newsletter of a local environmental group as
recently as just 2 or 3 years ago, you would probably have found an article
complaining about large visible clouds of smoke and fumes emanating from
a nearby steel mill. The owners of the mill would have been severely
criticized for failing to address these pollution problems in a timely manner.
Today that situation has changed to a certain extent. Though unre-
solved issues still exist, most notably the listing of coke oven emissions
as a hazardous air pollutant, our group recognizes that the steel industry
has made substantial progress in the overall control of particulate
emissions. Further, we realize that the solutions have not always been
simple because emission sources are numerous and quite varied.
But the job is not done. Total suspended particulate (TSP) levels
still exceed the primary standards in major steelmaking areas of the
country by.a significant margin. To reach attainment, additional controls
on process emission sources and the achievement of good work practices will
be necessary.-
Before I discuss these points,! would like to focus on control
activities that have been taking place in the industry. Particularly
noteworthy are efforts to control coke oven emissions in a comprehensive
manner. One example is U.S. Steel's program for minimizing door leaks at
Clairton Works where the company has upgraded existing door technology
by using more temperature-resistant seals and plunger springs. In addition,
guide/stop blocks have been added to each door to l) prevent excessive
stress on the sealing edge,and 2) produce better door sealing by repeatedly
positioning the door in the same place. Finally, operating and maintenance
(O&M) practices have been expanded at Clairton to provide more thorough
inspection, cleaning and repair of doors.(l, 2)
Locally, a noticeable improvement in control performance has been
made by Interlake at its two-battery coke plant. This accomplishment can
be credited, in part, to a major overhaul of the facility. The larry. cars
have also been retrofitted with extra piping to help retain particulates
in the coke plant's enclosed by-product systems. Finally, Interlake has
installed a travelling hood/fixed scrubber system to reduce pushing
emissions.
Recent developments indicate that more efficient pollution control
technologies will be available to the industry in the near future.
Several companies have been developing alternative approaches to control
systems which first capture emissions and then transport them to air-cleaning
18
-------
devices. The incentive to explore such alternatives has "been strong because
of the costs associated with systems that totally enclose a "building such as
a blast furnace cast house and then evacuate it to a baghouse fabric f"1"1'1
One development which has received a great deal of att^ircion in the
last few months is a non-capture system developed by Jones & Laughlin (J&L)
to control cast house emissions. While little information about this system
is available to the public because J&L has applied for a patent, I understand
that it essentially reduces the amount of participates produced in the
first place by minimizing contact of hot slag and metal with the cast house
atmosphere, thus suppressing the generation of iron oxide fumes. No
capture hoods or air-cleaning devices are required.
U.S. Steel has also been investigating the feasibility of applying
this concept at its mills. At Geneva Works,a recent test showed that cast
house emissions could be suppressed to very low levels by using a flame to
purge oxygen from the hot metal ladle before and during the cast. At South
Works, U.S. Steel has devised a non-capture system at its electric arc shop
consisting of a ring of steam jets attached to the top of the hot metal
ladle. During the tap, steam is injected into the ladle to suppress the
generation of fumes. This system has performed effectively.
These developments are very encouraging, particularly since they
come at a time when some steel companies have shown a reluctance to further
control fugitive process emissions. Instead, they have been focusing more
attention on open dust sources. Though unpaved roads and raw material
stock piles can be significant sources of air pollution, additional controls
on process emission sources will also be necessary to achieve the primary
standards.
At this point, I would like to briefly mention several studies which
were commissioned by the U.S. Environmental Protection Agency for the
purpose of supplementing local modeling efforts in heavily industrialized
urban areas of the Midwest.
In one study,the role of open dust sources was investigated on days
when the 2U-hour primary standard for TSP was exceeded at a monitoring
station located close to three steel mills in southeast Chicago. It was
hypothesized that if violations of the 2U-hour standard were primarily due
to open dust, then few violations should occur on days when there was
precipitation cover on the ground or when wind speeds were below critical
thresholds (12 m.p.h.) for the entrainment of particulate matter from
ground or from raw material storage piles.
During the two year study period, 228 days of TSP readings were made.
Of these, Uo were greater than the 2U-hour standard (260 micrograms per
cubic meter). During these ho days there was precipitation either greater
than 0.01 inches on the sampling day or 0.1 inches the day before sampling
on nine occasions. In other words, 22.5% of the 2^-hour exceedences
occured when the impact of open dust sources was minimal. Additionally,
29 days or 72.5% of ihe exceedences occured on days in which the resultant
wind speed was less than 12 miles per hour.(3)
19
-------
Microscopic analyses of particles found on high volume filters from
monitoring sites near steel mills in northvest Indiana and Cleveland have
also "been informative. These analyses have identified, in general terms,
the contributions of various source categories to TSP levels. Particles
were identified by means of their morphology (structure and form), size,
transparency, color and other physical properties. In addition, they were
grouped into categories such as minerals, combustion products, biological
matter and miscellaneous, with further subcatagorization where possible.
The concentrations on the filters frequently exceeded the primary
and secondary 2U-hour TSP standards. Analyses showed that no single source
category was the sole cause of TSP standard excursion on any date or from
date to date. For example, in one study open dust from" raw materials
handling activities contributed a large portion of the particle types on
one sampling date, while on another date emissions from a steel melting
process were indicated as the primary particle type.'^' In another study,
emissions from iron melting as well as open dust from raw materials handling
were found to be the major components of TSP at one monitoring site, while
at a nearby site steel melting emissions were the primary component .( 5 )
These analyses are limited in capability. Both reference samples
from the sources in question arid meteorological data would be needed in
most cases to accurately identify the particular operations in the steel
mill which are responsible for the particles found on the filters.
Nonetheless, they do point out the complexities of air quality problems
in major steelmaking areas and the need for comprehensive control programs.
Even with a complete program directed at process and open dust
sources, attainment of the primary standards for TSP will also depend upon
the achievement of good work practices. This is obviously true in the case
of controlling coke oven emissions where strict adherance to specific O&M
practices is critcal, but the same can apply to other steel mill sources.
For example, some companies have modified operating procedures at basic
oxygen process shops to help reduce charging emissions. At J&I/s Indiana
Harbor Works, hot metal is poured into the vessel as slowly as possible
to reduce agitation of the bath and splashing. In addition, the vessel is
tilted as little as possible, . thereby keeping its mouth close to the main
collection hood in order to improve capture efficiency.
In some cases , manpower will be the key to achieving good O&M
practices. Data on offtake leaks at Kaiser Steel^s Fontana Works show
that a substantial improvement in control of leaks occured after one
additional employee was provided to help with luting topside leaks and tend
ing the
To ensure that good work practices are established and that they
become part of the daily operating routine, self-monitoring programs should
be instituted at steel mills. Control performance at coke batteries, for
example, could be determined by smoke readers recording visible emission
observations. Steel companies in Pennsylvania have already instituted
this practice in order to comply with state law.
20
-------
One steelmaker has found self-monitoring programs to be very beneficial.
According to CF&I, self-monitoring "helps to identify the cause rather than
the symptom of a problem so that the trouble area or condition can be elimi-
nated rather than just treated. It also monitors progress, obtains operator
involvement, reinforces productive efforts, directs maintenance efforts, and
documents improvements."T6)
An example of an instance in which self-monitoring could make a differ-
ence in control performance involves one of our local steel mills. Though
several coke batteries at this mill have been modified for stage charging,
visible emissions still range from 1000 to 1200 seconds per five consecutive
charges. Even though a debate over what represents a "reasonable" level
of control for charging emissions has been taking place, I believe there is
room for considerable improvement at these batteries.
In addition to CF&I, I am aware of the self-monitoring program
instituted by U.S. Steel - again at Clairton Works. The fact that both
U.S. Steel and CF&I have achieved a great deal of success in controlling
coke oven emissions illustrates the importance of self-monitoring for
optimizing- emission control. We hope more steelmakers will develop
similar programs.
In conclusion, we recognize the special needs of the steel industry
and we applaud the efforts which have been made to deal with them. However,
violation of the primary standards still persist in major steelmaking areas.
Additional time has been provided by the steel "stretch-out" bill to
address pollution requirements and to modernize. Where strong efforts are
being made to solve air quality problems by improving existing control
systems and developing new technologies, the industry will receive our
support.
21
-------
REFERENCES
1. U.S. EPA. Coke Oven Emissions from By-Product Coke Oven Charging, Door
Leaks, and Topside Leaks in Wet-Coal Charged Batteries - Background
Information for Proposed Standards. Draft EIS. July 1980.
2. Spawn, P., et al. Assessment of Air Emissions from and Controls for Iron
and Steelmaking Sources - U.S. Steel. Corporation Clairton Works Coke
Plant. Draft Final Report. U.S. EPA Contract No. 68-01-UlU3.
3. GEOMET, Inc. Study of Ambient Air Quality in Vicinity of Major Steel .
Facilities. Appendix A. Analysis of Impact of Open Dust Source TSP
Potential on Urban 2U-hour Standard Exceedance Rate. Prepared for
U.S. EPA. October 1979-
h. Graf, J. and R. Draftz. Report on Analysis of High Volume Samples from
Gary, Indiana. IIT Research Institute. Project No. C6U53.
December 1979.
5. Graf, J. and R. Draftz. Total Suspended Particulates Non-Attainment
Study for" Cleveland, Ohio. IIT Research Institute. Project No. C61*53.
August 1979.
6. Oliver, J. and J. Lane.' Control of Visible Emissions at CF&I's Coke
Plant - Pueblo, Colorado. Journal of the Air Pollution Control
Association. 2£ (9): September 1979, p. 920-925.
22
-------
Session 1: AIR POLLUTION ABATEMENT
Chairman: Philip X. Masciantonio
U.S. Steel Corp.
Pittsburgh, PA
23
-------
INHALABLE PARTICULATE MATTER SAMPLING PROGRAM FOR IRON AND STEEL
AN OVERVIEW PROGRESS REPORT
by: R. C. McCrillis
Industrial Processes Branch (MD-63)
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park
North Carolina 27711
ABSTRACT
EPA's Office of Research and Development has entered into a major,
program to develop inhalable particulate (IP) matter emission factors,
where IP is defined as airborne particles of <15 pm aerodynamic equivalent
diameter. The Industrial Processes Branch of EPA's IERL-RTP is respon-
sible for the ferrous metallurgical industry segment of this program.
Efforts to date for the iron and steel category are summarized in this
paper. IP requirements are meshed with those of other EPA sampling
programs whenever possible, thus reducing overall cost to EPA and
minimizing inconvenience to the host plants.
A thorough literature review and compilation of existing data revealed
the existence of particle size data for several of the major iron and
steel sources. However, none of these data were obtained using current
IP measurement technology; most of them do not cover the full IP size
range and, in many cases, there is insufficient documentation to completely
determine test procedures followed and to fully define the process operation
during the tests. The current field test program is designed to augment
this existing data base by directing resources toward those sources with
the combination of high priority and low existing data quality. Both
process sources and open sources are included. To date, processes tested
are basic oxygen furnace (EOF) charging and tapping, hot metal desulfuriza-
tion, blast furnace cast house (both building evacuation and local control
technologies), sinter discharge, and EOF main stack (limited combustion
system after scrubber). Open sources tested are paved and unpaved roads
and coal storage pile maintenance, all both with and without controls,
and an uncontrolled open area.- Several additional tests are underway or
scheduled. All data gathered will be summarized and published in a
single report early next year.
24
-------
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 consid-
eration is being given to a size-specific particulate standard. EPA has
initiated an extensive program to compile and review existing data and,
where necessary, conduct field sampling programs from which reliable
size-specific emission factors can then be determined.
EPA's Office of Research and Development is responsible for developing
these inhalable particulate emission factors. A major part of this
effort is directed toward the steel industry. In this paper, the rationale
and approach being followed to select test sites are discussed. This
paper also discusses briefly the review of existing iron and steel source
particle size data, the field sampling program being undertaken, and
results obtained to date.
DISCUSSION
Source Selection
At the commencement of the size-specific emission factors development
program, the aerodynamic equivalent particle diameter upper cut point
was set at 15 ym. This was defined as inhalable particulate (IP)^-*-'.
The Clean Air Science Advisory Committee recently reviewed the basis for
a size-specific standard; based on this review, the upper cut point
diameter may be reduced to 10 ym. This change will not affect the
field sampling protocols'^-5) which require, for example, the use of
cascade impactors with a 15 ym precutter cyclone for ducted emissions.
The precutter is still required to remove large particles which, if
allowed to enter the impactor, would tend to bounce from one stage to
another, thus giving erroneous stage weights. Lowering the upper cut
point would expand the data base since most of the existing data were
gathered with devices having an upper cut point of about 10 ym.
At the outset of the iron and steel IP sampling program, the decision
was made to proceed with field test site selection on a voluntary source
basis, rather than through the application of Clean Air Act, Art. 114.
Industry was contacted through the American Iron and Steel Institute
(AISI) which 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 several
field test programs. Armco, Inc.'s Middletown Works was selected for
measurment of open source emissions and their BOF stack. Inland Steel
Co.'s sinter discharge and Armco, Inc.'s Kansas City electric furnace
25
-------
shop will be tested during October. J&L Steel's full combustion EOF
stack at Aliquippa will be tested in November as will U.S. Steel's
Fairfield Q/BOP.
In addition to the selection procedure coordinated with AISI, every
effort has been made to combine measurement of IP with other EPA sampling
programs. Not only does this reduce EPA expenditures, it also reduces
inconvenience to the host companies. In this category, tests have been
completed at Kaiser Steel Co.'s new EOF shop (charging and tapping
fugitives) and hot metal desulfurization. Dominion Foundries and Steel
Co. was the site of a blast furnace cast house test; measurements at
their No. 1 coke quench tower are now in progress. Finally, Bethlehem
Steel Co.'s "L" blast furnace cast house at Sparrows Point was tested in
April.
The initial source priority ranking, shown in Table 1, was developed
based on estimated controlled total particulate emissions from each
source on a nationwide basis. This prioritization represented an average
o.f emission factors developed under separate efforts: one represented
factors from specific short term emission tests *•"'; the other, presented
values which might be termed typical for long term operation '''. Although
this procedure was only qualitative, it did nevertheless provide a rational
approach for initial source selection.
Review of Existing Data
The initial source selection priority list was based on total particu-
late data due to the paucity of known particle size data. At the outset
of this program 2 years ago, only six particle size data sets from .iron
and steel sources were contained in EPA's Fine Particle Emissions Informa-
tion System (FPEIS)^°). These data sets, consisting of three open hearth
furnace stack tests, two electric arc furnace tests, and a coke oven
pushing shed test, are summarized in Figures 1-3, respectively. These
data are judged to be good, although consideration must be given to when
they were obtained (1974-77) and the advancements made in particle size
sampling technology since then. Ideally, these three sources should be
tested again but not before other high priority sources with no existing
data are tested.
A thorough review of both published and unpublished literature was
recently completed by GCA/Technology Division in a concerted effort to
ferret out all existing particle size data'H'. This review produced
over 30 unpublished test reports containing particle size data. None of
these data were obtained using the current IP measurement technology.
It is apparent, however, that many of these data are of sufficient quality
to warrant delaying new tests of these sources until high priority sources
with little or no data are tested.
The initial source test priority has been revised to reflect the
discovery of these size-specific data. Table 1 also indicates the revised
rank of each source and also the amount and type of data gathered. In
26
-------
TABLE 1. IRON AND STEEL SOURCE TEST PRIORITY RANKING (CONTROLLED EMISSIONS)
Rank
Initial
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
Revised
1
2
3
4
5
27
6
28
29
30
7
14
31
10
8
9
11
12
13
15
16
17
18
19
32
20
21
22
23
24
25
26
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
Reheat Furnaces
Blast Furnace Combustion
OH Roof Monitor
Coal Charging
Open Area
Machine Scarfing
BOF HMT
OH Misc. Fugitives
Soaking Pits
EAF Misc. Fugitives
OH - HMT
Industry Total
Part icu late
Emissions, Mg/yr'8'
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
2,000
2,000
2,000
1,800
1,100
670
650
640
570
540
190
Data Sets
Discovered^"'
2(b), 2(d)
l(d)
-
6(f)
7(f)
3(b), 4(d), l(e)
2(b), 2(c)
5(b), l(d), l(e)
2(d)
7(b), l(d), l(e)
-
-
Kb), 3(d)
-
Kb), Kd)
Kb)
-
-
-
-
-
-
-
-
2(d)
-
-
-
-
-
-
"
IP Tests
Planned
X
X
X
X
X
X
X
X
X
(a) Megagrams per year (• Metric tons per year).
(b) Particle Sizing Method: a - follows IP protocol, b - impactors, good process data taken;
c - SASS train; d - Coulter counter, sieve, or microscopic;
e - test methods unknown; f - exposure profile/wind tunnel
with impactor.
27
-------
99.990
99.950
99.90
99.80
99.50
99
98
95
90
UJ
o
-------
99.990
99.950
99.90
99.80
99.50
99
98
95
90
UJ
O
or
UJ
0.
UJ
80
70
60
50
40
30
20
10
5
2
I
0.5
0.2
0.15
O.I
0.0
10"
PROCESS AVERAGE
O MELT
Q TAP -MELT
Figure 2.
.0° .O1
PARTICLE DIAMETER, micrometers
(9)
Average size distribution - Marathon
LeTourneau electric arc facility.
10'
29
-------
99.990
99.950
99.90
99.80
99.50
99
98
95
90
Ul
O
-------
addition those sources which have been or will be tested under this IP
program are also shown in Table 1.
Results to Date
The following briefly summarizes the test results available to date.
Armco Inc., Middletown Works
The extensive source fugitive emissions program undertaken for EPA
by Midwest Research Institute at Armco, Inc.'s Middletown Works will be
addressed in detail by the next speaker^-^) . Later this afternoon,
Bruce Steiner will address the specific control strategies instituted
under Armco1 s bubble application^-*'.
In summary, emissions from paved and unpaved roads were measured
before and after the initiation of emission reduction procedures. For
paved roads, this procedure consisted of water flushing and/or sweeping-
vacuuming at regular intervals. Berms were treated with Coherex®. Unpaved
roads were treated either with Coherex® or water, also at regular intervals.
Tests of the Coherex® treated road were conducted on the second and third
days after suppressant application. Additional tests would be required to
determine the long term control efficiency decay. Windblown emissions
from a coal pile during pile maintenance operations were also measured.
The Armco/Middletown open source tests are summarized in Table 2.
The Middletown limited combustion BOF main stack was tested for IP in
July 1980'1-^. Although the IP protocol calls for measurements before and
after the control device, the nature of the limited combustion BOF operation
at Middletown precluded measurements before the scrubber. Measurements
after the scrubber included total particulate and particle size as per the
protocol. The results, shown in Table 3, are presented for two production
rates, normal and intermediate. The ratios of IP to total particulate are
69 and 57 percent, respectively, for the controlled emission.
Kaiser Steel Corporation
The first two sources tested under the iron and steel IP 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 and cost-shared by
EPA's Office of Enforcement through EPA's Region V office, consisted of
total particulate by EPA Method 5 and particle size before and after the
control device (in both cases, a baghouse). EPA's contractor, Acurex
Corporation, conducted this extensive source test program during March-May
1980.
Results of the HMDS tests reported at last year's EPA Iron and Steel
Symposium(16) have been revised due to recent improvements in the data
reduction procedure. The revisions resulted in a lowering of the total
particulate emision factors accompanied by an increase in the IP fraction.
31
-------
TABLE 2. SUMMARY OF ARMCO, INC., MIDDLETOWN OPEN SOURCE FUGITIVE EMISSION MEASUREMENTS
GO
no
Source
Paved Roads
Overall Ave .
Test Site A
Test Site A
Test Site D
Test Site D
Unpaved Roads
Test Site B
Test Site B
Test Site C
Test Site E
Test Site E
Test Site E
Coal Storage
Pile Maintenance
Vehicle
Type
Mix
Mix
Mix
Mix
Mix
Light Duty
Light Duty
Heavy Duty
Heavy Duty
Heavy Duty
Heavy Duty
Dozer
Control
Measure
Uncontrolled
Uncontrolled
Vacuum Sweeping
Uncontrolled
Flushing
Uncontrolled
Coherex®
Coherex®
Uncontrolled
Water(b)
Water(c)
Uncontrolled
Particulate
Total
599
478
164
1070
680
3320
252
1530
34075
2400
8300
845
Emission Factor f
< 15 ym
158
114
92 .
298
336
864
57
321
4382
237
3410
305«>
, / VKT^ a'
<2.5 ym
40
30
25
69
96
269
20
43
1070
59
610
405(d)
(a) Grams per vehicle kilometer traveled unless otherwise indicated.
(b) Immediately after application.
(c) 75 min. after application.
(d) Grams per minute of dozer activity.
-------
TABLE 3.
SUMMARY OF LIMITED COMBUSTION EOF MAIN STACK PARTICULATE EMISSION FACTORS
(15)
Date Run Nos.
7/11/80 PSD-1 (b)=-
through PSD-6
7/14/80
7/14/80 PSD-7 (c)
and PSD-8
Cumulative Emission Factor, kg/Mg (1'b/ton) SteelVa^
<2.5 ym
0.010
(0.020)
0.007
(0.013)
< 10 ym
0.010
(0.021)
0.007
(0.014)
< 15 ym
0.011
(0.022)
0.008
(0.015)
Total
0.016
(0.031)
0.014
(0.027)
Low Carbon Steel
Produced, Mg (tons)
196
(216)
152
(168)
CO
CO
(a) Steel produced.
(b) Results are the average of the first three heats and are considered to represent emissions
during normal production rates.
(c) Results are for the last heat tested and represent emissions for intermediate production
rates.
-------
The revised emission factor for the uncontrolled fume generated is 0.54 kg/Mg
of hot metal desulfurized (range: 0.23 to 0.77 kg/Mg). The average IP
fraction was 32.2 percent. The revised baghouse outlet emission factor is
0.0015 kg/Mg of hot metal desulfurized (range: 0.0011 to 0.0017 kg/Mg).
The revised average IP fraction was 81.1 percent. Baghouse collection
efficiency averaged 99.7 and 99.3 percent for total and IP particulate,
respectively.
Separate measurements of hot metal charging and tapping emissions
were made on the No. 6 EOF. These emissions are collected by hoods in the
furnace enclosure (doghouse) and conveyed to the baghouse. Uncontrolled
emissions were measured simultaneously in the two main fugitive emission
ducts serving No. 6. This site was chosen for its accessibility and
noninterference from other emission sources ducted to the baghouse.
Controlled emissions were measured in 2 of the 12 baghouse exit stacks
only during periods when No. 6 was charging or tapping and other sources
ducted to the baghouse were not generating emissions.
Dominion Foundries and Steel Company (DOFASCo)
DOFASCo first installed cast house emissions control several years
ago. The combined control system for cast houses No. 2 and 3 was started
up in November 1978. These cast houses are evacuated by a common fan and
baghouse sized to control emissions in either one by isolation valves.
The IP emission tests were run during the week of November 10, 1980,
on the combined control system serving cast houses No. 2 and 3; however,
measurements were made only when furnace No. 3 was casting. Emission
tests followed the protocol for ducted sources. Measurements of the fume
generated were made for EPA/IERL-RTP and EPA's Office of Enforcement through
EPA's Region III office by GCA/Technology Division in the duct upstream of
the baghouse. A detailed discussion of the preliminary results was
presented at the Third Symposium on the Transfer and Utilization of Parti-
culate Control Technology'17'. Uncontrolled total particulate emissions
averaged 0.2 kg/Mg iron cast. Average IP fraction was reported as 62
percent; revisions and corrections to the IP data reduction scheme
currently being implemented may change this value. In this case, only the
IP fraction value will be affected by these revisions.
Bethlehem Steel Corporation, Sparrows Point
CCA/Technology Division conducted emission tests for EPA/IERL-RTP and
EPA's Office of Enforcement through EPA's Region III office at the new "L"
furnace cast house at Sparrows Point during April. This is a large modern
furnace employing close-fitting hoods and covers over the trough, iron
runners, and spouts, a practice pioneered in Japan. These and other
fugitive emissions are ducted to a large baghouse. Emissions, following
the ducted source protocol, were measured in the duct upstream of the
baghouse. Since this baghouse also controls emissions from many other
fugitive sources, the discharge emissions would not be representative of
casting emissions. Therefore, the open monitor discharge was not sampled.
34
-------
Preliminary data analysis indicates an average total uncontrolled
particulate emission factor of 0.65 kg/Mg hot metal cast. The particle
size distribution is being recalculated in light of revisions to the data
reduction program. Initial indications are that the IP fraction is some-
what less than was measured at DOFASCo. These results, compared to the
DOFASCo results, seem reasonable since local evacuation should also capture
the large particles which would have settled out in the DOFASCo cast house
and would, therefore, not have been captured by the roof monitor exhaust.
Future Tests
Five additional IP field sampling tests are either underway or sched-
uled. The No. 1 quench tower at DOFASCo is now being tested; field work
will be completed by the end of October. This is a complex project with
partial funding by EPA/IERL-RTP, EPA's Office of Air Quality Planning and
Standards, EPA's Division of Stationary Source Enforcement, and DOFASCo.
Emission measurements are being made above and below the baffles. Separate
tests will be run at two quench water qualities: one using coke plant
wastewater as makeup to the sump; and the other, Lake Erie water. Three
particle sizing techniques are being used: Andersen Mark III with a 15 ym
precutter cyclone, Southern Research Institute's two-cyclone IP train, and
the EPA/Southern Research Institute dilution train.
Inland Steel Co.'s sinter plant discharge emissions will be sampled
in October. Measurements will be taken before and after the baghouse.
The J&L Aliquippa EOF main stack ESP is scheduled for tests in
November. IP measurements will be made before and after the ESP. This
will be a jointly funded test involving EPA/IERL-RTP, EPA's Office of Air
Quality Planning and Standards, and EPA's Office of Enforcement through
EPA's Region III office.
U.S. Steel's Fairfield Q/BOP main stack is also scheduled for tests
in November. IP measurements will only be made after the scrubber due to
the difficulty of installing ports prior to the quencher.
Armco, Inc's. Kansas City electric furnace shop will be tested
separately for main furnace emissions and for fugitive emissions captured
by the canopy hood. Measurements, to be made before and after the control
devices, are tentatively scheduled for November.
Once the above field tests are completed and each test report reviewed
by EPA and the host plant, all data, new and old, will be summarized and
published in one source category document giving emission factors versus
particle size for all sources in the iron and steel industry for which
data are available. In addition to the source test data, emission factors
will be summarized in formats appropriate for AP-42'-'-°'. It is anticipated
that testing will be completed in November and the iron and steel category
report published in March 1982. These data will then be available to the
states for the preparation of revised state implementation plans if a
size-specific ambient air particulate standard is adopted.
35
-------
ACKNOWLEDGEMENTS
The three, sampling contractors for EPA1s IP program are GCA/Technology
Division, Midwest Research Institute, and Acurex Corporation. The AISI
and its member companies (in particular Armco, Inc., Bethlehem Steel Cor-
poration, Dominion Foundries and Steel, Ltd., Kaiser Steel Corporation,
J&L Steel Co., and U.S. Steel Corporation) have been most helpful.
EPA's Office of Enforcement and EPA's Office of Air Quality Planning and
Standards were instrumental in developing and implementing the jointly
funded tests. The cooperation of Environment Canada and the Ontario
Ministry of Environment is also appreciated.
36
-------
ENDNOTES
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 K. 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. Cuscino, T. A., "Particulate Emission Factors Applicable to the
Iron and Steel Industry," EPA-450/4-79-028 (NTIS PB 81-145914),
Midwest Research Institute, September 1979.
7. Barber, W. C., Particulate Emissions from Iron and Steel Mills,
EPA/OAQPS Internal Memorandum, dated November 6, 1978.
8. 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.
9. 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.
10. Gronberg, S., Program Summary for Characterization of Inhalable
Particulate Matter Emissions (Draft), EPA Contract 68-02-3157,
TD 5, GCA/Technology Division, September 1980.
11. Farino, W., and S. A. Beaton, Review of Iron and Steel Particle
Size Data (Draft), EPA Contract 68-02-2687, TD 16, GCA/Technology
Division, August 1981.
37
-------
12. Cowherd, C., and M. Small, "Open Source Fugitive Emissions,"
Symposium on Iron and Steel Pollution Abatement Technology for
1981" (October 1981, Chicago, IL).
13. Steiner, B. A., "Armco's Experience with Applications of the Bubble
Concept," Symposium on Iron and Steel Abatement Technology for 1981
(October 1981, Chicago, IL).
14. Small, M. , Iron and Steel Plant Open Source Fugitive Emission Control
Evaluation - Field Testing (Draft), EPA Contract 68-02-3177,
Assignment 4, Midwest Reserch Institute, June 1981.
15. Inhalable Particulate Emission Characterization Report: Armco
Steel's No. 16 Basic Oxygen Furnace, Middletown, OH (Draft),
Prepared by PEDCo Environmental, Inc., under subcontract, EPA
Contract 68-02-3158, TD 6, Midwest Research Institute.
16.. Steiner, J. , and D. Bodnaruk, "Particulate and S02 Emission Factors
for Hot Metal Desulfurization," in Proceedings: Symposium on Iron
and Steel Pollution Abatement Technology for 1980 (Philadelphia, PA,
11/18-11/20/80), EPA-600/9-81-017 (NTIS PB 81-244808), Research
Triangle Institute, March 1981.
17. Spawn, P. D., R. M. Bradway, and S. Gronberg, "Inhalable Particulate
Emission Factors for Blast Furnace Casthouses in the Iron and Steel
Industry," Third Symposium on the Transfer and Utilization of
Particulate Control Technology (March 1981, Orlando, FL).
18. Compilation of Air Pollutant Emissions Factors Third Edition,
No. AP-42 (NTIS PB 275525), U.S. Environmental Protection Agency,
August 1977.
38
-------
COST EFFECTIVENESS EVALUATION OF ROAD DUST CONTROLS
by: Chatten Cowherd, Jr., Thomas A. Cuscino, Jr., and Mark Small
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
ABSTRACT
Preliminary analysis of control measures for fugitive particulate emis-
sion sources in the iron and steel industry indicates that control of open
dust sources has a highly favorable cost effectiveness ratio in comparison
with control of process sources of fugitive particulate emissions. However,
rigorous cost effectiveness evaluation must await accurate and well docu-
mented information on control performance and cost. This paper addresses
the analytical and practical considerations involved in acquiring meaningful
cost effectiveness data for the major open dust sources in the iron and
steel industry—vehicular traffic on unpaved and paved roads.
Results of extensive performance testing of road dust controls are pre-
sented. The control measures tested were watering and chemical treatment of
unpaved roads and vacuuming, flushing, and broom sweeping of paved roads.
The mean efficiencies of control measures tested, except for vacuum sweeping
of paved roads, were found to be independent of particle size. The mean
control efficiency of freshly applied Coherex® to unpaved roads was higher
than the efficiencies of the other measures tested. An analytical framework
for control cost effectiveness analysis is proposed, and control cost data
for road dust controls at two steel plants are given.
INTRODUCTION
Previous studies of fugitive particulate emissions from integrated iron
and steel plants have provided strong evidence that open dust sources (spe-
cifically vehicular traffic on unpaved and paved roads and storage pile ac-
tivities) should occupy a prime position in control strategy development.1'2
These conclusions were based on comparability between industry-wide uncon-
trolled emissions from open dust sources and typically controlled fugitive
Paper presented at EPA Symposium on Iron and Steel Pollution Abatement
Technology for 1981, held at the McCormick Inn in Chicago, Illinois, on
October 6-8, 1981.
39
-------
emissions from major process sources such as steel-making furnaces and sin-
ter machines. Moreover, preliminary cost-effectiveness analysis of promis-
ing control options for open dust sources indicated that control of open
dust sources might result in significantly improved air quality at a lower
cost in relation to control of process sources. These preliminary conclu-
sions warranted the gathering of more definitive data on control performance
and costs for open dust sources in the steel industry.
As a result, a field sampling program was undertaken to obtain data on
the performance of control measures for vehicular traffic on unpaved and
paved roads within integrated iron and steel plants. A 1978 emissions in-
ventory of open dust sources within the steel industry had shown that ve-
hicular traffic accounts for approximately 70% of suspended particulate
emissions on an uncontrolled basis.2 The selection of control measures to
be tested was based on the extent of use within the industry and the results
of the preliminary evaluation of cost effectiveness of control options for
unpaved and paved roads. Control measures selected for unpaved roads were
watering and chemical treatment; and control measures selected for paved
roads were vacuum sweeping, flushing, and a combination of broom sweeping
and flushing.
The study design was developed to provide the following data needs for
cost-effectiveness evaluation of each control measure tested:
1. Emission factors for three particle size fractions: total particu-
late (TP); inhalable particulate (IP)--particles smaller than 15 |Jm equiva-
lent aerodynamic diameter (EAD); and fine particulate (FP)--particles
smaller than 2.5 urn EAD, before and after application of the control mea-
sure.
2. Emission factor correction parameters for uncontrolled and con-
trolled test road segments, providing adequate characterization of road sur-
face and traffic conditions.
3. Control application parameters: intensity, frequency, and time af-
ter application.
4. Investment and operating costs of control.
TEST METHODS
Table 1 lists the equipment that was used to sample particulate emis-
sions from traffic on unpaved roads and paved roads. Equipment locations
and intake heights are specified. The primary tool for quantification of
emission rate was the exposure profiler, operated in the moving point
source mode.
The exposure profiler consisted of a portable tower supporting an array
of five sampling heads. Each sampling head was operated as an isokinetic
exposure sampler directing passage of the flow stream through a settling
chamber (trapping particles larger than about 50 urn in diameter) and then
40
-------
upward through a standard 8 by 10 in. glass fiber filter positioned hori-
zontally. Sampling intakes were pointed into the wind, and sampling veloc-
ity of each intake was adjusted to match the local mean wind speed, as de-
termined prior to each test. Throughout each test, wind speed was monitored
by recording anemometers at two heights, and the vertical profile of wind
speed was determined by assuming a logarithmic distribution. Normally, the
exposure profiler was positioned at a distance of 5 m from the downwind edge
of the road.
TABLE 1. SAMPLING EQUIPMENT FOR EXPOSURE PROFILING
Distance Intake
from source (m) Equipment height (m)
Upwind 10 1 Standard Hi-Vol 2.0
2 Hi-Vols with 15 jjm Inlets 1.0
3.0
1 Continuous Wind Monitor 4.0
Downwind 5 1 MRI Exposure Profiler 1.0
with 5 Sampling Heads 2.0
3.0
4.0
5.0
2 Hi-Vol Parallel-Slot Cascade 1.0
Impactors with Cyclone 3.0
Precollectors
2 Warm-Wire Anemometers 1.0
3.0
Particle size distribution in the dust emission plume was measured us-
ing a high-volume parallel-slot cascade impactor preceded by a cyclone pre-
separator. This provided for direct isokinetic measurement of the total
particle size distribution but required extrapolation from the cyclone cut-
point (11 [Jm EAD) to determine IP concentrations. Particle sizing samplers
were operated along side of the exposure profiler at two heights.
Also, a high-volume sampler (Hi-Vol) with a size-selective inlet (SSI)
was operated at the upwind monitoring station to determine the IP fraction
of the background particulate. For tests of controlled emissions, a second
Hi-Vol/SSI was operated at a higher elevation to determine the change of
background IP concentration with height. Conventional high-volume samplers
were operated at one height both upwind and downwind of the source.
41
-------
In addition to the measurements of wind speed obtained at two heights
on the profiling tower, a meteorological instrument was also located at the
background monitoring station. Continuous measurements of wind speed and
direction at a height of 4m were recorded at the upwind site.
In order to determine the properties of aggregate materials being dis-
turbed by the action of machinery or wind, representative samples of the ma-
terials were obtained for analysis in the laboratory. Unpaved and paved
roads were sampled by vacuuming and broom sweeping to remove loose material
from lateral strips of road surface extending across the traveled portion.
Throughout a test of traffic generated emissions, a vehicle count was
maintained by a pneumatic-tube traffic counter. Periodically (e.g., dur-
ing 15 min of each hour) vehicle mix was determined by compiling a log of
vehicles passing the test point segregated by vehicle type (number of axles
and wheels). Vehicle speeds were measured with a radar gun. Data on vehi-
cle weight were obtained from plant personnel.
At the end of each run, the collected samples of dust emissions were
carefully transferred to protective containers within the MRI instrument
'van. Glass fiber filters from the MRI exposure profiler and from standard
Hi-Vol units and impaction substrates were folded and placed in individual
envelopes. Dust that collected on the interior surfaces of each exposure
probe was rinsed with distilled water into separate glass jars. Dust was
transferred from the cyclone precollectors in a similar manner.
Dust samples from the field tests were returned to MRI and analyzed
gravimetrically in the laboratory. Glass fiber filters and impaction sub-
strates were conditioned at constant temperature and relative humidity for
24 hr prior to weighing, the same conditioning procedure used before taring.
Water washes from the exposure profiler intakes and the cyclone precollec-
tors were filtered after which the tared filters were dried, conditioned at
constant humidity, and reweighed.
After the gross samples of surface particulate were taken to the labo-
ratory, they were prepared for moisture and silt analysis. The first step
consisted of reducing the sample to a workable size. A riffle sample
splitter was used for this purpose, following the principles of ASTM Method
D2013-72, as appropriate.
The reduced samples of surface particulate were dried to determine
moisture content and screened to determine the weight fraction passing a
200 mesh screen, which gives the silt content. A conventional shaker was
used for this purpose. The procedures for moisture and silt analysis were
patterned after ASTM Method C136-76.
TEST DESCRIPTION AND RESULTS
The field sampling of emissions, with and without control application,
was performed at Armco's integrated iron and steel plants in Middletown,
Ohio and Houston, Texas. At the Middletown works, testing was conducted in
42
-------
July 1980 and in October and November 1980; in August 1980, Armco imple-
mented an extensive control program for open dust sources at that plant.
The control measures tested at the Middletown works included treatment of
unpaved roads with Coherex® and with water and vacuuming of paved roads. At
the Houston works, testing was conducted in June and July of 1981. The con-
trol measures tested in Houston were flushing of paved roads and a combina-
tion of flushing and broom sweeping of paved roads.
Testing of uncontrolled emissions was conducted either at sites where
no control measures had been used or where control measures had been sus-
pended allowing the source to return to its uncontrolled state. Whenever
possible, a control.measure was introduced and tested at the same location
where the uncontrolled tests had been performed.
Emission factors and control efficiencies were obtained for TP, IP, and
FP matter, as shown in Tables 2, 3, and 4, respectively. It should be noted
that the mean efficiencies of the control measures tested, except for vac-
uuming of paved roads, were nearly independent of particle size.
Much of the observed variation in the controlled emission factors re-
flects differences in time between the control application and emissions
testing. Except for the treatment of unpaved roads with Coherex®, the mean
control efficiencies presented are thought to represent average controlled
conditions. The control efficiencies for the application of Coherex® to un-
paved roads reflect only the early stage in the lifetime of that control
measure.
In the case of Coherex® application to unpaved roads traveled by light-
duty vehicles, a well-defined initial decay in efficiency was observed for
TP, IP, and FP (Figure 1). This is thought to reflect the effect of buildup
of surface dust loading on the treated road because of vehicular tracking of
material from adjacent unpaved areas. The curve for TSP in Figure 1, which
was obtained in a previous study,2 reflected the accentuated tracking ef-
fects resulting from the short length of the test strip.
CONTROL COSTS
Cost data for the control measures tested were obtained from the steel
plants where testing was performed. These data included: (a) annualized
costs of equipment purchase and installation, and (b) annual operating
costs. The annualized investment costs took into account the initial costs,
the lifetime of the equipment, interest, and taxes. To calculate the total
annualized cost, the average annual cost of operation was added to the prod-
uct of the initial capital investment and the capital recovery factor. The
capital recovery factor is the percentage of the initial investment which
would be paid yearly on a loan of mortgage. Table 5 presents a summary of
the cost data obtained.
43
-------
TABLE 2. TEST RESULTS SUMMARY FOR TOTAL PARTICULATE EMISSIONS
Emission factor (Ib/VMT)
Test type
Paved Road
Unpaved road
(light-duty
traffic)
Unpaved road
(heavy-duty
traffic)
Number
Control of
measure tests
Uncontrolled
Middletown
Houston
Vacuum sweeping
Flushing
Flushing and
broom sweep-
ing
Uncontrolled
Coherex®
Uncontrolled
Coherex®
Watering
7
4
4
3
4
4
5
4
4
2
Range
0.29-4.6
2.9-3.7
0.24-1.2
1.0-1.6
0.51-1.8
10-14
0.089-1.3
99-130
3.4-8.5
8.3-29
Mean ± Control
standard efficiency
deviation (%)
2.3 ± 1.5
3.1 ± 0.41
0.69 ± 0.38 70, 65*
1.2 ± 0.29 61, 58-
1.0 ± 0.55 67
12 ± 1.9
0.88 ± 0.50 92
120 ±16
5.4 ± 2.2 96
19 ± 15 50- »
Based only on uncontrolled testing at the same site.
Based on an 8 hr watering cycle.
-------
TABLE 3. TEST RESULTS SUMMARY FOR INHALABLE PARTICULATE EMISSIONS
Emission factor (Ib/VMT)
Number
Test type
Paved Road
Unpaved road
(light-duty
traffic)
Unpaved road
(heavy-duty
traffic)
Control
measure
Uncontrolled
Middletown
Houston
Vacuum sweeping
Flushing
Flushing and
broom sweep-
ing
Uncontrolled
Coherex®
Uncontrolled
Coherex®
Watering
of
tests
7
4
4
3
4
4
5
4
4
2
Range
0.14-1.4
0.55-1.2
0.15-0.69
0.27-0.58
0.095-0.44
1.0-4.2
0.061-0.38
26-34
1.2-2.0
0.99-4.7
Mean ±
standard
deviation
0.66 ± 0.42
0.95 ± 0.28
0.37 ± 0.24
0.41 ± 0.15
0.23 ± 0.15
3.0 ± 1.4
0.27 ± 0.14
30 ± 4.2
1.5 ± 0.37
2.8 ± 2.6
Control
efficiency
(%)
-
-
44, 47*
56, 62*
76
-
91
-
95
63**
Based only on uncontrolled testing at the same site.
Based on an 8 hr watering cycle.
-------
TABLE 4. TEST RESULTS SUMMARY FOR FINE PARTICULATE EMISSIONS
Emission factor (Ib/VMT)
Test type
Paved Road
Unpaved road
(light-duty
traffic)
Unpaved road
(heavy-duty
traffic)
Number
Control of
measure tests
Uncontrolled
Middletown
Houston
Vacuum Sweeping
Flushing
Flushing and
broom sweep-
ing
Uncontrolled
Cone rex®
Uncontrolled
Coherex®
Watering
7
4
4
3
4
4
5
4
4
2
Range
0.050-0.33
0.15-0.43
0.039-0.20
0.060-0.076
0.0042-0.11
0.24-1-.3
0.032-0.094
5.5-8.8
0.27-0.59
0.22-1.0
Mean ±
standard
deviation
0.18 ± 0.091
0.28 ± 0.13
0.12 ± 0.077
0.068 ± 0.0079
0.074 ± 0.026
0.86 ± 0.44
0.067 ± 0.023
7.6 ± 1.5
0.46 ± 0.14
0.62 ± 0.57
Control
efficiency
(%)
-
-
36, 35*
75, 65*
73
-
92
-
94
67**
Based only on uncontrolled testing at the same site.
Based on an 8 hr watering cycle.
-------
100
90
u
0)
o
u
TSP (Ref. 2)
80
I
I
I
I
200 400 600 800 1000
Vehicle Passes After Treatment
With Coherex®
1200
1400
Figure 1. Initial Decay Curve for Control Efficiency of
Coherex®Applied to Unpaved Roads.
47
-------
TABLE 5. SUMMARY OF OPEN DUST CONTROL COST DATA
co
Plant
Middletown
works
Houston
works
Source
Paved
roads
Unpaved
roads
Paved
roads
Actual
source
extent
(miles) Control
16.9 '2 Vacuum
sweepers
Flusher
7.1 Coherex®, dis-
tribution
truck, and
storage tanks
14.6 Broom sweeper
No. 1
Broom sweeper
No. 2
Flusher
Purchase
and
installation
cost ($)
144,000
68,000
100,000
18,000
20,000
34,000
Year of
purchase
1980
1976
1980
1978
1980
1978
Estimated
lifetime
(yr)
5
10
7
5
5
7
1980
operating
and
maintenance
costs ($)
214,000
57,000
287,000
65,100
57,000
52,300
1980
treated
source
extent
(miles)
2,020
5,080
1,630
888
888
1,776
-------
Care must be taken in compiling figures for the items presented in Ta-
ble 5 so that all of the costs are properly accounted for. Purchase and in-
stallation costs must include cost for freight, tax, and loan interest. The
lifetime of the control application equipment is difficult to estimate be-
cause of lack of experience in using this equipment. The operating and
maintenance costs should reflect increasing frequency of repair as the
equipment ages, along with increasing costs for parts, energy, and labor.
The treated source extent, which is the product of the mileage of roads
treated and the annual number of treatments, requires accurate recordkeeping
on the use of control application equipment. Figure 2 depicts the varia-
tions in costs over the lifetime of the application equipment.
Equipment, Installation, Freight, Tax, and Interest
O
—~ Depreciation Tax Deduction
LIFE OF EQUIPMENT
Scrap
Value
Figure 2. Graphical Presentation of Open Dust Control Costs.
The selection of one road dust control measure over another cannot ra-
tionally be determined on the basis of raw costs alone. The proper basis
for comparison is the cost effectiveness which is defined as follows:
where:
CE =
D_
ER
CE = Cost Effectiveness ($/mass of emissions reduced)
D = Total Annual Expenditure ($/yr)
ER = Annual Emission Reduction (mass of emissions reduced/yr)
49
-------
In the selection process, each candidate control measure should be analyzed
to determine application parameters corresponding to its optimum cost effec-
tiveness for the particular road segment(s) being considered. Finally, the
control measure selected for implementation should be the one with the most
favorable optimized cost effectiveness.
For example, the determination of optimum use of vacuuming of a par-
ticular paved road might include the following reasoning. The efficiency of
vacuuming decays from the value at the time of application to zero as the
surface dust loading builds up to its uncontrolled (equilibrium) value. Al-
though the highest control efficiency would be achieved by continuous use of
the available vacuum trucks, operating costs would be at a maximum. If the
decay of control efficiency to zero consumed a period of 3 days, a single
vacuuming of the road each day would produce about two-thirds of the maximum
emission reduction at a fraction of the cost.
Clearly, the optimum cost effectiveness of vacuuming is a complex func-
tion of the mileage of road to be cleaned, the time it takes to pass over
the road with the vacuum truck, and the rate at which the control efficiency
•decays after vacuuming. Mathematical cost effectiveness functions are being
developed for each of the control measures tested, as part of this research
program.
An approximate cost effectiveness value for a particular control mea-
sure may be estimated for one plant based on the value rigorously derived
for another plant. The scaling procedure used for this purpose assumes
that the optimal application frequency for a given control measure is inde-
pendent of the application site. For example, the total cost of control
per mile of road, as given in Table 5, may be used to estimate the cost of
control of a road of specified length located at another plant by a simple
multiplication.
CONCLUSIONS
The mean efficiencies of the control measures tested, except for vacuum
sweeping of paved roads, were found to be independent of particle size. The
fractional efficiency of vacuum sweeping decreased with decreasing particle
size.
The mean control efficiency of Coherex® application to unpaved roads,
which exceeded 90%, was substantially higher than the efficiencies of the
other control measures tested. However, this efficiency reflected only the
early stage in the expected lifetime of that control measure.
In the case of Coherex® application to unpaved roads traveled by light-
duty vehicles, a well-defined initial decay in efficiency was observed.
This is thought to reflect the buildup of surface dust loading on the
treated road because of vehicular tracking of material from adjacent unpaved
areas.
50
-------
In the selection of the optimum control measure for a particular appli-
cation, the proper basis for comparison is the cost effectiveness, which is
defined as the total annual expenditure divided by the annual emission re-
duction. For each candidate control measure, cost information should be
developed for several application scenarios, each with its associated con-
trol efficiency, and the optimum scenario identified. Finally the control
measure selected for implementation should be the one with the most favor-
able optimized cost effectiveness.
ACKNOWLEDGMENT
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/vehicle mile
Ib
mile
0.282
0.454
1.61
kg/vehicle km
kg
km
REFERENCES
1. Bohn, R., T. Cuscino, Jr., and C. Cowherd, Jr. Fugitive Emissions from
Integrated Iron and Steel Plants. EPA-600/2-78-050, 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.
51
-------
BLAST FURNACE CASTHOUSE CONTROL TECHNOLOGY
FALL 1981 UPDATE
Prepared by:
Thomas J. Maslany
U.S. Environmental Protection Agency
Sixth and Walnut Streets
Philadelphia, Pennsylvania 19106
V
Peter D. Spawn
GCA/Technology Division
213 Burlington Road
Bedford, Massachusetts 01730
ABSTRACT
- This paper describes blast furnace casthouse information that has become
available in the past six months. Results of total mass and inhalable particu-
late testing on the baghouse inlet at B'ethlehem Steel's Sparrows Point L furnace
are presented. Baghouse outlet test data are also presented for Inland Steel's
No. 7 furnace and Wheeling-Pittsburgh's Monessen No. 3 furnace. The status of
United States and Canadian casthouse control systems as of September 1981 is
provided along with a listing of world-wide controlled casthouses, excluding
Japan. Cost data for installed systems are also given.
52
-------
BLAST FURNACE CASTHOUSE CONTROL TECHNOLOGY -
FALL 1981 UPDATE
INTRODUCTION
The number of blast furnace casthouse emissions control systems in the
U.S. equalled nine by the Fall of 1981. Two of the controlled furnaces are
new, modern, large furnaces (Bethlehem Steel, Sparrows Point "L" and Inland
Steel No. 7), but the remaining seven systems are retrofits to older, smaller
furnaces. Previously published papers have described existing systems and
presented visible and mass emissions data.1'2 Since publication of these
two earlier papers, additional information has become available and will be
presented herein.
The recent information includes total mass and inhalable particulate tests
of ducted, baghouse inlet emissions at the Sparrows Point L furnace; baghouse
outlet mass emissions tests at Inland Steel's No. 7 and Wheeling-Pittsburgh's
Monessen No. 3 furnace;; additional data describing U.S. Steel's efforts at
the Edgar-Thomson Nos. 1 and 2 furnaces, and their experience with emissions
suppression technology. Additionally, Dofasco has converted their Nos. 2 and 3
furnaces from total evacuation to local hood systems. Also summarized are cost
data for each control option, indicating capital costs for installed systems
and operating and maintenance costs.
The current trends in casthouse emissions control technology are focusing
on local hoods and non-capture shrouding techniques. At present, only one
blast furnace casthouse in North America is controlled by a total evacuation
system - the No. 1 furnace at Dofasco.
CASTHOUSE CONTROL STATUS - UNITED STATES AND WORLD
Table 1 shows the status of casthouse control systems in the United States.
By 1981, the American steel industry had made committments to government agen-
cies to bring 128 (71%) of the approximately 180 standing blast furnaces into
compliance by 1982. Of these 128 committments, 70 represent actual installation
of control equipment. The remaining 58 furnaces are currently shutdown, or
planned for shutdown before the end of 1982. These shutdowns are due to the
retiring of older, smaller, less productive furnaces. As mentioned in previous
papers, partial replacement of lost production will be accomplished by increas-
ing the output of existing furnaces through techniques such as external
desulfurization, modest hearth diameter increases during reline, burden and
fuel improvements, installation of second tapholes, and increasing wind quan-
tity and quality.
In summary, the American steel industry has made committments to install
casthouse emissions control systems of 57 percent of the 122 blast furnaces
currently planned for operation after 1982.
Blast furnace casthouse emissions were first controlled in the United
States at a ferromanganese furnace at the Bethlehem Steel Corporation's
53
-------
TABLE 1. U.S. CASTHOUSE CONTROL SYSTEMS
INSTALLED AS OF FALL 1981
Plant
BSC/Johnstown
BSC/Bethlehem
B
C
D
E
USSC/Edgar Thomson
1
2
Wheeling-Pittsburgh
Monessen No. 3 (Jane)
J&L Steel ,
A-4d
H-3
C-3
BSC/Sparrows Point, L
Inland Steel No. 7
Furnace
hearth
diameter,
ft-in.
-
30-0
27-11
30-0
24-0
28-10
28-10
28-0
29-0
32-0
30-6
44-6
45-0
Capture
system
TE
PE
PE
PE
PE
LH/NCS
NCS
LH
NCS
NCS
NCS
LH
LH
Exhaust
f lowrate
per fee. ,
acfm
400,000
330,000°
per fee.
140,000
None
130,000
None
None
None
300,000
250,000
Retrofit or
new, single
or multiple
tapholes
R/S
R/S
R/S
R/S
R/S
R/S
R/S
R/S
R/S
R/M
R/M
N/M
N/M
TE = total evacuation; PE = partial evacuation; LH
NCS = non-capture shrouding.
Out of service since 1977.
°Total flow to baghouse is 660,000 acfm.
Demonstration systems being refurbished; see text.
local hoods;
54
-------
Johnstown Works in 1973. However, the first systems in the world were in-
stalled with the construction of new furnaces in Japan in the mid-sixties.
The Japanese steel industry reported to EPA in 1976 that all furnaces at that
time were either built with or retrofitted with various degrees of casthouse
controls. Most Japanese retrofit systems are a combination of local hoods with
the newer furnaces having canopy hood auxiliary systems. The cleaning systems
are predominantly baghouses, but scrubbers and electrostatic precipitators
have also been used.
In the last few years the Japanese introduced another control option for
blast furnace casting control. At the Kakogawa Works of Kobe Steel, one new
furnace and two existing furnaces were fitted with roof monitor electrostatic
precipitators. This system has no large fans or duct work and uses the natural
bouyancy of the plume to move emissions through a wide plate electrostatic
precipitator.
In Canada, both Stelco and Dofasco have experience with blast furnace
casthouse control. Stelco's present experience is with a local hood control
system on their new furnace at their Lake Erie Works and experiments with local
hooding at the Hilton Works. Stelco plans to retrofit all four furnaces at the
Hilton Works with local hooding.
Controls were retrofit to several West German blast furnaces between 1975
and 1980. Blast furnace casthouse controls have also been reported at facili-
ties in Canada, Britain, Italy, France, Sweden, and the U.S.S.R. Table 2
summarizes available data for world-wide casthouse control systems. A partial
listing of Japanese systems appears in Reference 3 which reports controlled
furnaces operated by two Japanese steel companies.
MASS AND INHAIABLE PARTICULATE TESTS AT THE BETHLEHEM STEEL L FURNACE
GCA measured total mass emissions and the particle size distribution in
the ductwork serving the baghouse inlet at the L blast furnace at Bethlehem
Steel's Sparrows Point plant in April 1981. The testing was sponsored by
EPA's Industrial Environmental Research Laboratory as part of EPA's inhalable
particulate (IP) measurement program.
CONTROL SYSTEM DESCRIPTION
Figure 1 illustrates one of the four identical emissions capture systems
installed on each taphole and runner system at the L furnace. The iron trough
and iron and slag runners are covered and evacuated at the following four
points:
• notch-area local hood, 5 to 6 feet above the taphole,
• area encompassing the iron pool and dam (trough),
• area over the tilting iron runner, and
• area over the slag spoon.
55
-------
TABLE 2. PARTIAL LIST OF CONSTRUCTED BLAST FURNACE
CASTHOUSE CONTROL SYSTEMS IN THE WORLD3
Furnace
hearth
_, diameter,
Country Plant f
Canada Stelco
Lake Erie
Hamilton
Dofasco
No. 1
No. 2
No. 3
No. 4
West . Krupp/Rheinhausen
Thyssen/Hamborn No . 4
Thyssen/Schwelgern No. 1
Thyssen/Huttenbetieh
No. 5
Thyssen/Ruhrort Nc .6
Mannesmann "A"
Mannesmann "B"
•
NAh
NA
20-9
20-9
22-3
28-0
37-9
37
45
NA
NA
34
34
Exhaust
f lowrate
Capture per fee. ,
system acfm
LHC
LH
TEfc
LH
LH
LH
LH
LH
LH
LH
LH
LH
LH
NA
NA
325,000
200,000
200,000
NA
183,000
341,000
412,000
NA
NA
421,000
NA
Retrofit or
new, single
or multiple
tapholes
N/M
R/S
R/S
R/S
R/S
R/M
R/M
R/M
R/M
NA
NA
R/M
NA
Sweden
Norrbottens Jarnverk
AB. Luleae
NA
Great British Steel Corp.
LH
88,000
Japanese systems not included; see Reference 3.
NA = not available.
LH = local hoods, TE = total evacuation
Under construction.
NA
Britain
France
Italy
USSR
South Teeside No. 1
Llanwern
Usinor/Dunkirque No. 4
Italsider/Taranto
No. 5
Krivoy Roy/
Krivorozhstal No. 9
Krivoy Roy
Cherepovets
Magnitogorsk
45-10
45-10
46-8
45-10
45
NA
NA
NA
LH 338,000
LH NA
LH 530,000
LH 353,000
LH
LH
LH
LH
N/M
R/M
R/M
N/M
N/M
N/M
N/M
N/M
"Reference 5
56
-------
in
— EAST CASTHOUSE —
MOVABLE TBOUGMHOOD (TYP. )
'EN SE*1 (TYP.)
— WEST CASTHOUSE —
Figure 1. Schematic of emission control system
for "L" blast furnace at BSC/Sparrows
Point (illlustrated runner is typical
of all four runners).
Figure 2. Approximate sketch of No. 7 blast
furnace casthouse layout. (Not
to scale).
-------
Total system exhaust flowrate measured during the tests averaged 320,000 acfm
at 184° F. The flowrate applied to each exhaust point is not available at this
time. The first section of the iron trough hood, adjacent to the taphole, is
removed for drilling and plugging. All other sections of the exhaust system
remain in place during casting with the exception of several small access doors
that may be opened by workers.
Iron is alternately cast from two diagonally opposed tapholes for about
one week before the off-line tapholes (and runners) are returned to service.
Nos . 1 and 3 tapholes were operating during these tests. The exhaust system
is sometimes used to exhaust notch-area hoods on the off-line runners to cool
maintenance workers, although this only occurred for one 10-minute period
during testing. Several casts overlapped during the field tests, i.e. two
tapholes cast simultaneously. The overlap usually lasted for only 15 minutes,
and testing was halted during overlap periods. Bethlehem stated that when
casting two tapholes simultaneously, the ducts on both systems are fully opened
and the exhaust capacity is approximately 50 percent on each.
Captured emissions are controlled by a Wheelabrator-Frye positive pressure
baghouse, with a design capacity of 312,000 acfm at 140° F. The five chamber
baghouse operates with one chamber always in a cleaning cycle (shaker) , result-
ing in a net air/cloth ratio of 2.5:1. Compartment APs ranged from 2 to 6
inches H20 during the evaluation.
MASS, INHALABLE AND VISIBLE EMISSIONS DATA
Four EPA Method 5 test runs were conducted on 28 and 29 April 1981 during
casting (iron and slag flow) . Three to four Andersen impactor runs were
conducted during each Method 5 test. Table 3 summarizes the test results.
Testing did not include drilling and plugging emissions and also excluded
clean-up operations that occur after taphole plugging. Testing was halted
for a few periods of casting overlap when the exhaust flow was divided between
two tapholes. Individual Andersen impactor test runs are not shown in Table 3
since the IP test protocol recommends averaging of all size data for a more
accurate, single number. Also shown in Table 3 are available process data
describing casting conditions during test periods.
Method 9 opacity observations of visible emissions escaping capture and
exiting casthouse roof monitors during testing showed the following results
for 8.5 hours of observation:
• 7.1% of Method 9 observations were >_20% opacity
• 1.1% " " " " >40%
• 0.5% .... .. "
One additional hour of observations are not included in the above summary be-
cause interference from another process precluded observation of the roof moni-
tor. Table 4 compares roof monitor visible emissions with process operations.
All roof monitor emissions greater than 10 percent opacity resulted from dril-
ling, taphole lancing, plugging, and overlapping casting. During single taphole
operations, most roof monitor emissions (excluding drilling and plugging)
58
-------
TABLE 3. SUMMARY OF MASS AND IP TESTS AT THE L FURNACE
Run Total Mass Emissions Mass <15ym, Mass <2.5Mra,
Number (Baghouse Inlet),3 Ib/ton Ib/ton
Ib/ton iron
1
2
3 '
4
Average
1.21
1.36
1.40
1.21
1.3
(Individual impactor runs
not reported-see text)
-
0.27 0.
17
Hot metal sulfur, average 0.024, wt. percent
Hot metal temperature, average maximum - 2740°F
average at slag over - 2692°F
dFinal results.
Preliminary results.
TABLE 4. COMPARISON OF VISIBLE EMISSIONS ESCAPING THE CASTHOUSE
TO PROCESS OPERATIONS, "L" BLAST FURNACE, 27-29 APRIL 1981
1
Operation
Drilling and 02
Lancing
Single cast
(full evacuation)
Overlapping cast '
(partial evacuation)
Plugging
Visible emissions, % opacity
Typical
15-20
0-5
10-15
20-40
Range
0-60
0-10
0-30
5-90
Typical duration of
operation (min)
2-3
100-115
10-15
5-7
Q
Excluding drilling and plugging emissions.
Emissions shown are for whichever casthouse showed the highest emissions
during simultaneous casts.
59
-------
originated from an opening in the slag runner cover used for slag working.
The small access door on this opening had apparently been removed in order for
workers to gain access to the slag surface.
INSHOP EMISSIONS OBSERVATIONS AND PROCESS DATA
The inshop process observer assessed the origin and magnitude of emissions
escaping capture using the following criteria:
• Light emissions - those appearing to be in the 5 to 25 percent
opacity range as viewed by the inshop observer,
• Moderate emissions - those appearing to be in the 30 to 60 percent
opacity range, and
• Heavy emissions - those appearing to be of 65 to 100 percent opacity.
Table 5 briefly summarizes inshop emissions observations, as recorded by the
process observer stationed on the casthouse floor.
The inshop observer estimated that the notch hood captured approximately
90 to 95 percent of drilling emissions and 70 to 80 percent of oxygen lancing
emissions. Uncaptured lancing emissions observed inside the shop were generally
in the moderate to heavy range, while drilling emissions were usually light.
Once the iron trough hood was in place, emissions escape from the system
was usually zero, except during simultaneous casting. When slag flow began.
light emissions escaped the slag spout hood opening as mentioned previously.
During plugging, the notch-area hood appeared to capture roughly 40 to 50 per-
cent of the emissions.
SUMMARY OF OTHER RECENTLY AVAILABLE INFORMATION
Since preparation of the two previous papers, additional information has
become available for several existing control systems. The following
paragraphs update previously published information, and present some new data.
Refer to the earlier papers for a complete technology overview and background
information for systems discussed below. '
BAGHOUSE OUTLET TESTS AND PROCESS OBSERVATIONS AT INLAND STEEL'S NO. 7 FURNACE
Compliance tests of the baghouse outlet stack serving the casthouse control
system at Inland Steel's new No. 7 blast furnace were conducted on 27 March
1981. The three test runs reported to EPA showed outlet grain loadings of
0.0042, 0.0059, and 0.0069 gr/dscf for a test average of 0.0057 gr/dscf. Two
60
-------
TABLE 5. INSIDE CASTHOUSE EMISSIONS ASSESSMENTS,
"L" BLAST FURNACE, 28-29 APRIL 198la
Emissions
Operation
Drilling
0« lancing
Notch
area
L-M
M-H
Iron
trough
L-M
M-H
Dam
hood
0
0
magnitude
Iron Spout
hood
0
0
Slag Spout
hoodb
0
0
Single cast
(full evacuation)
L-M
Overlapping cast
(partial evacuation)
Plugging6
L
H .
0-L
H
L-M
0
0-L
0
L
L
Assessment technique using L (Light), M (Moderate), and H (Heavy); see
text.
Emissions only when slag was running (^50% of total cast time). '
"Three of six drill periods observed were oxygen lanced.
Data for casthouse showing highest emissions.
g
Partial notch-area evacuation due to simultaneous casting occurred for two
of six plugs included here.
TABLE 6. ROOF MONITOR VISIBLE EMISSIONS DATA FOR
WHEELING-PITTSBURGH'S MONESSEN NO. 3
CASTHOUSE
Test Period No. of Casts Total Method 9 % of Total Method 9
Observed Observations Observations
EPA Demonstration 50
Company data after
improvements 61
12,837
N/Aa
_, >40%, >60%
7.251.781.04
2.8 N/A <0.1
Not available.
61
-------
tests were apparently conducted during overlapping casts (two tapholes cast
simultaneously). The system exhaust flowrate varied about five percent
between test runs, averaging 255,000 acfm at 115° F.
Roof monitor visible emissions observations conducted during daylight
portions of the tests showed that 0.3 percent of the Method 9 observations
were greater than or equal to 20 percent opacity. Simultaneous inshop obser-
vations found that all roof monitor emissions occurring during casting were
caused by taphole plugging.
The No. 7 furnace is very similar to the L furnace at Sparrows Point with
respect to both furnace design and emissions control system layout (see
Figure 2). However, the notch-area hood at Inland consists of hooding above
and along both sides of the taphole, while the notch-area hood at the L furnace
consists of a single, horizontal hood above the taphole. Another similarity
between Inland's No. 7 and Bethlehem's L furnace is that when two tapholes
are cast simultaneously, the available exhaust flow is split nearly equally
between each taphole system.
Observations of four casts inside the No. 7 furnace during the stack tests
noted that during main cast periods (including slagging), only minor puffs
escaped capture, quickly dissipating to zero percent opacity within the cast-
house. Approximately 95 percent of drilling emissions and most plugging (and
iron trough hood removal) emissions were estimated to be captured by the notch-
area hoods. Observation of two-taphole casting from inside the No. 7 cast-
house found that emission escape from hoods and runner systems was essentially
unchanged from single cast operations.
UPDATE OF CASTHOUSE CONTROL EFFORTS AT DOFASCO
As reported previously, Dofasco converted the total evacuation systems
on their Nos. 2 and 3 blast furnaces to local hoods in late 1980 and 1981.
Both furnaces now have two tapholes. By the summer of 1981, notch-area hoods,
currently exhausting at about 150,000 acfm (per furnace), were installed on
both furnaces. Local hoods were also installed on iron spouts on the No. 3
furnace, exhausting about 50,000 acfm from the ladle in use. Dofasco-designed
runner covers were installed on the iron and slag runners of both casthouses.
Dofasco plans to complete the two systems by mid-1982 with addition of local
hoods on iron ladle spouts at the No. 2 casthouse, and installation of a new
"third-generation" runner cover design on both casthouses. The runner covers
are based on the non-capture, supression concept. The Nos. 2 and 3 furnaces
can now be cast simultaneously with controls, since total exhaust flow rate
under simultaneous casting is 400,000 acfm, the capacity of the existing bag-
house. With the old total evacuation system sized at 400,000 acfm per cast-
house, both furnaces could not be controlled simultaneously.
Local hoods were at one time being considered for the totally-evacuated
No. 1 casthouse. However, an improved, stainless steel seal between the fur-
nace shell and the bustle pipe was installed during the summer 1981 reline.
The company reports that the new seal is effective in controlling emissions
escape from this critical area, and Dofasco plans to continue total evacuation
62
-------
control of the No. 1 furnace.
The local hood and runner cover system that is partially complete on
Dofasco's No. 4 furnace is scheduled for completion by August of 1983.
WHEELING-PITTSBURGH MONESSEN NO. 3 UPDATE
The Wheeling-Pittsburgh system consists of an iron trough hood and two
iron spout hoods, evacuated to a baghouse. The trough hood is approximately
.27 feet long, extending four feet past the dam.. The trough hood is constructed
in two sections so the 15 foot long forward (taphole) section can be removed by
the shop crane to facilitate work around the trough area. The height of the
hood bottom, above the casthouse floor, is about five to six feet. The two
iron spout hoods measure about five feet square and have a steel pipe curtain
around the hood .perimeter, extending to the top of the hot metal ladle (U.S.
Patent No. 4,245,820).
During casting, the trough hood and one iron spout hood are evacuated at
a design flowrate of 140,000 acfm (total system). When the first ladle is
full, iron flow and hood exhaust flow are diverted to the second iron^spout.
The exact flowrate for each hood is not available at this time. Figures 3
and 4 illustrate the layout of the Monessen system.
The demonstration test of roof monitor visible emissions from the Wheeling-
Pittsburgh Monessen No. 3 local hood system was completed in April 1981.
Comparison of inside and outside observations indicated that most emissions
leading to roof monitor visible emissions >20 percent opacity came from the
uncontrolled slag runners and slag spout, and to a lesser extent, the iron
runners. Since the test period, Wheeling-Pittsburgh has been refining their
system and investigating additional techniques for reducing emissions. Speci-
fic areas of improvement to date are.-: ij changing slag runner material, 2)
installing additional cross-wind barriers, and 3) lowering the bottom edge of
the iron trough hood by one or two feet. Table 6 summarizes the results of
the EPA-sponsored demonstration test of 50 casts, and the more recent company-
supplied data showing results of the improvements made since the EPA tests.
Method 5 tests were recently conducted on one of the 10 stub stacks
serving the 10 compartment baghouse on the Monessen No. 3 furnace. Three test
runs showed outlet grain loadings of 0.0016, 0.0017, and 0.0023 gr/dscf, for an
average of 0.0019 gr/dscf. Since only one compartment was tested, the total
control system flow rate could not be measured.
U.S. STEEL, EDGAR THOMSON UPDATE
Preliminary evaluation of the Edgar Thomson No. 1 system was conducted in
July 1981. Final evaluation of the system by U.S. Steel and EPA will not be
completed for several months.
The Edgar Thomson No. 1 furnace is fitted with a local hood above the
iron trough, measuring.about 16 feet in length from the taphole. The hood
covers the first 60 percent of the trough area, and is evacuated at 140,000
•acfm through two ducts, located behind and in front of the bustle pipe. The
S3
-------
TO BAQHOUSE
EXHAUST DUCT
TROUGH HOOD
Figure 3. Isometric View of Trough Hood at
Wheeling-Pittsburgh, Monessen No. 3,
(Courtesy of JACA Corporation)
OFFTAKE TO
BAQHOUSE
PIPE
CURTAIN
Figure 4. Isometric View of Hot Metal Ladle Hood
(Pipes Enlarged to Show Detail), at
Wheeling-Pittsburgh, Monessen No. 3.
(Courtesy of JACA Corporation)
64
-------
lower edge of the hood lies 11 to 12 feet above the casthouse floor. The hood
exhausts to a baghouse.
An inverted U-shaped air curtain, fitted with a number of horizontal air
jets along the U-shaped perimeter, is located about 10 to 15 feet from the dam.
The air curtain device is designed to push skimmer and trough area emissions
into the hood, a so-called "push-pull" system. Rated at 4700 acfm, the air
curtain can be throttled down to provide the proper push/pull effect.
The remainder of the No. 1 casthouse is fitted with J&L-type non-capture
suppression technology. The No. 2 furnace at the Edgar Thomson Works is en-
tirely fitted with J&L-type technology (no fans or baghouse). As in the case
with the No. 1 system, final company and EPA evaluation of the No. 2 system is
not yet complete.
J&L'S NON-CAPTURE TECHNOLOGY UPDATE
The three J&L prototype systems used for last year's EPA demonstration
tests are not presently in service. J&L is constructing permanent versions
of these systems and modifying the control systems, furnaces, and casthouses
to improve performance. J&L reports that the permanent systems will enter
service on furnace C-3 at Cleveland, A-4 at Aliquippa, and H-3 at Indiana
Harbor. The H-3 furnace was not used in the demonstration test, but will be
fitted with the permanent system because the furnace is currently down and can
be more easily retrofit than the (operating) H-4.
U.S. STEEL'S NON-CAPTURE, EMISSIONS SUPPRESSION EXPERIMENTS
U.S. Steel is experimenting with non-capture suppression techniques that
differ from the J&L technology. Development work is underway at the No. 2
blast furnace at the Geneva, Utah plant. No formal evaluation of the system
has been conducted by EPA. Additional details are not available because U.S.
Steel considers this a proprietary system and has exerted a confidentiality
claim on their work.
AVAILABLE COST DATA
The cost of controlling blast furnace casthouse emissions will vary
depending on such factors as furnace production size, ease of retrofitting
the control system, type of control system selected (amount of exhaust air to
be moved), emissions characteristics of the furnace, and the overall effective-
ness of the system after installation. The cost figure can be divided into two
basic categories - capital cost, and annual utility (operating) plus mainten-
ance costs (O&M). These two components can-vary for two identical furnaces
with the same control option located at separate plants based on retrofit
difficulty, number of common systems (shared baghouses), and the local cost of
power.
Table 7 summarizes cost data for installed systems, as provided by each
company. The origin and basis of each cost is discussed below.
65
-------
TABLE 7. SUMMARY OF AVAILABLE COST DATA3
Company
Dofasco
Bethlehem,
Bethlehem
Wheeling-Pittsburgh,
Monessen
J&L Steel
Blast
Furnace
Nos. 2&3
No. 1
B.C.D.E
No. 3
2500 tpd
furnaces
Capture
System
Total
Evacuation
Total
Evacuation
Partial
Evacuation
Local
Hoods
Non-
Capture
Capital
Cost,xl06
8.3°
-
11. 4d
1.9
0.75-1.15g
O&M Cost, per
year, xlO3
-
144
V56
100-150f
130s
All cost data in 1981 American dollars.
All systems use baghouses for gas cleaning, except J&L technology.
c
Cost for two furnaces, with single baghouse.
Cost for four furnaces, with a single baghouse. Includes costs incurred
during original study on E furnace; part of equipment salvaged for final
construction.
g
Based on annual operation of two and one-half furnaces.
Company rough estimates vs. actual costs for all other data.
8J&L's estimate for 2500 tpd furnace.
66
-------
DOFASCO - TOTAL CASTHOUSE EVACUATION
Dofasco reported that the total building evacuation system installed on
their Nos. 2 and 3 blast furnaces at the Hamilton Works cost $1,100,000 for
curtain walls, $2,800,000 for duct work, and $2,500,000 for the baghouse and
fan.** This total of $6,400,000 (Canadian dollars) was calculated to be equi-
valent to $8,270,000 in March 1981 dollars, using the Chemical Engineering
Plant Cost Index. Using 18 percent interest amortized for 20 years, the annual
captial cost is approximately $1,500,000.
For the similar total building evacuation system on Dofasco's No. 1
furnace, annual operating costs of $74,000 for electrical power and $22,000
for maintenance were reported. For that furnace, the annual operating cost
would be $96,000 per year (1978 Canadian dollars). Assuming that costs in-
creased 30 percent from this period and assuming the same U.S./Canadian con-
version rate of 15 percent, Table 7 shows 1981 O&M cost estimates.
BETHLEHEM STEEL - PARTIAL EVACUATION
Bethlehem Steel Corporation (BSC) indicated that the capital cost of the
four partial evacuation systems installed in their Bethlehem Works in 1980 was
$10,400,000 (including costs incurred during original study on E furnace). BSC
also stated that the annual operating and maintenance costs when two and one-
half furnaces are operating is $375,000 (averaged over the year). Updated to
March 1981 dollars, the installed capital cost is approximately $11,400,000.
It should be noted that this is a high flow, partial evacuation system with all
four furnaces sharing a common baghouse. BSC reported that the capital cost
figure includes the spacing problem for the single baghouse and the large dia-
meter, long run ductwork with substantial supporting.
WHEELING-PITTSBURGH - LOCAL HOODS
The recent installation of a local hood system at the Wheeling-Pittsburgh
Monessen Works No. 3 blast furnace was reported to cost approximately $1,900,000
in 1981 dollars. Although O&M costs have not yet been reported, the company's
rough estimate is in the range of $100,000 to $150,000 per year. Amortizing
this capital cost over 20 years and assuming a mid-range O&M cost shows an
annual cost of $342,000 for capital and $125,000 for O&M. Wheeling-Pittsburgh
plans to use this baghouse with additional modules when they install local
hooding systems on the other furnaces, to help reduce baghouse capital costs.
J&L STEEL - NON-CAPTURE SHROUDING
The system that shows the greatest potential to date for reducing the
cost of blast furnace casthouse controls is the non-capture shrouding tech-
nology demonstrated by J&L Steel. J&L markets this technology and requires
licensing fees. The fees quoted by J&L are $50,000 to review the system at a
J&L plant, $150,000 to receive the complete engineering package, technical
know-how, and license for one taphole, and finally, $50,000 for the license for
each additional taphole. J&L's current estimate is $750,000 in capital cost
for a single taphole furnace with the slag spouts outside the arcade, and
$1,150,000 for a single taphole furnace with the slag spouts inside the arcade.
67
-------
J&L's estimated operating cost is $0.15 per ton of hot metal which trans-
lates to about $130,000 per year for O&M for a 2500 tpd, single-taphole fur-
nace. J&L's current installations are on furnaces of 2,200, 2,750, and 3,000
tpd. J&L also reports that these three furnaces, used for the EPA demonstra-
tion tests, experienced a one percent increase in hot metal yield as a result
of using the non-capture emissions control system. Also experienced was an
increase in the temperature of hot metal delivered to the EOF shop.
REFERENCES
1. Spawn, P.D. and T.J. Maslany. Blast Furnace Casthouse Control Technology
and Recent Emissions Test Data. Presented at the 1981 APCA Specialty
Conference on Air Pollution Control in the Iron and Steel Industry.
Chicago, Illinois. April 1981.
2. Spawn, P.O., T.J. .Maslany, and R. Craig. Status of Blast Furnace Cast-
house Control Technology in the United States, Canada, and West Germany
in 1980. Presented at the Symposium on Iron and Steel Pollution Abate-
ment Technology for 1980.
3. May, William P. Blast Furnace Casthouse Emission Control Technology
Assessment. Betz Environmental Engineers, Inc. EPA Publication No. 600/
2-77-231. November 1977.
4. Samson, D.H. Dominion Foundries and Steel, Limited. Hamilton, Ontario.
Blast Furnace Casthouse Emission Control., Dofasco, Hamilton, Ontario.
July 1979.
5. Telephone conversation. Chuck Wilbur, American Air Filter, Louisville,
Kentucky to Peter Spawn, GCA/Technology Division.
68
-------
COKE QUENCH TOWER EMISSIONS AND EMISSIONS CONTROL
Dennis Wallace, Associate Environmental Scientist, Midwest Research Insti-
tute, Kansas City, Missouri
Naum T. Georgieff, Environmental Engineer, Emission Standards and Engineer-
ing Division, U.S. EPA, Durham, North Carolina
Dana Peckworth, Senior Environmental Engineer, Midwest Research Institute,
Raleigh, North Carolina
Sandra Stell, Research Assistant, Midwest Research Institute, Kansas City,
Missouri
ABSTRACT
Available test data indicate that the mass of particulate emissions
from a coke quench tower may be affected by quench water quality (based on
dissolved or total solids in the quench water), and the design and location
of baffles in the tower. This paper examines the impacts of the water qual-
ity and baffle design on quench tower emissions and emissions control. Emis-
sions test data and engineering models are used to estimate the relationship
of these parameters to emissions. Six possible control schemes are defined
and the impacts of the controls are evaluated.
INTRODUCTION
Over the past decade emissions from coke quench towers and the develop-
ment and evaluation of control measures for these emissions have drawn in-
creased attention from both air pollution control agencies and industry per-
sonnel. Prior to 1976, concerns about quench tower emissions centered on
the compliance of specific towers with state and local regulations. In
1976, the U.S. Environmental Protection Agency (EPA) conducted extensive
tests at the U.S. Steel Corporation's Lorain, Ohio plant. The results of
these tests brought added attention to the problems of quench tower emis-
sions. Over the past 3 years, interest in quench tower emissions has height-
ened as different offices within EPA have: (a) considered a new source per-
formance standard for wet coke quenching; (b) issued Reasonably Available
Control Technology (RACT) guidelines for coke quench towers; and (c) evalu-
ated coke quench towers as one of the major sources of inhalable particulate
emissions in the iron and steel industry. These activities have resulted
in a number of emissions tests and comprehensive data analyses that have
69
-------
significantly increased our knowledge of quench tower emissions. Many of
the results were summarized in a recent paper by Bloom and Jeffery1 and are
described in more detail in an earlier study by Gorman, et al.,2 and in a
recent EPA study.3
While our understanding of quench tower emissions and emissions control
have increased greatly in the last 10 years, questions are still unanswered.
In their paper, Bloom and Jeffery1 identify a number of factors related to
tower and baffle design, water quality, and operating characteristics which
might affect quench tower emissions and emissions control. In addition,
they identify existing data needs related to back-half particulate composi-
tion (i.e., the composition of particulate captured in the impingers of an
EPA Method 5 sampling train), the relationship of quench water and makeup
water quality, particle size distribution, and particulate generation mech-
anisms .
This paper will address the impacts of two of these parameters, water
quality and baffle design, on quench tower emissions and emissions control.
Bloom and Jeffery1 found that for a large number of tests under a variety
of process conditions the front-half particulate emissions (those emissions
capture on or prior to the filter in an EPA Method 5 sampling train), are
proportional to the dissolved solids in the quench water lost up the stack.
Although these results (see Figure 1) are strongly influenced by the tests
from U.S. Steel-Lorain, they found that a regression line through the data
from U.S. Steel-Gary and DOFASCO had almost the same slope. These data sug-
gest that emissions are related to the product of the spray water solids
concentrations and the quantity of water emitted from the tower. Data fur-
ther suggest that the quantity of water lost up the stack is dependent on
baffle design.
Both of these parameters have potentially significant impacts on quench
tower emissions, and control of the parameters is feasible. This paper re-
views the current state of knowledge and identifies major data gaps related
to the relationship of water quality to quench tower emissions and the per-
formance of different baffle systems, control schemes involving water qual-
ity control and baffling, and potential secondary impacts of these control
schemes.
IMPACT OF WATER QUALITY ON QUENCH TOWER EMISSIONS
RELATIONSHIP BETWEEN QUENCH WATER SOLIDS AND PARTICULATE EMISSIONS '•
The nature of the quenching operation, given the quantities of water
vapor and droplets lost up the stack, strongly suggests that the quantity
of particulate emissions from quench towers will be related to the quantity
of solids in the water sprayed onto the coke. This relationship was first
demonstrated in the analysis of test data from U.S. Steel-Lorain, Ohio, by
70
-------
GARY « 3 CLEAN
GARY # 3 DIRTY
GARY *5 CLEAN
GARY *5 DIRTY
LORAIN CLEAN
LORAIN DIRTY
DOFASCO RECYCLED
BAFFLE SPRAY
DOFASCO RECYCLED
DOFASCO BAY WATER
95% CONFIDENCE
LEVEL
1.0 20 3.0 40 50 60 70 BO 90
QUENCH TDS FACTOn. Ib/lon COAL
10.0
HO
120
110
140
is.q
Figure 1. Front half particulate, Ib/ton coal.
-------
Edlund, et al. Those initial tests suggested a relationship between par-
ticulate emissions and total dissolved solids (IDS) in the quench water
demonstrated by the regression equation:
E = 0.18 DW x 10"6 + 1.0 Eq. 1
where E = particulate emissions (kg/Mg of coke)
D = dissolved solids in quench water (mg/£)
W = water lost up the tower during quenching (£/Mg)
This relationship was substantiated by Bloom and Jeffery1 who found that
for a number of different towers (see Figure 1) the emissions are related
to TDS by the equation:
E = 0.19 DW x 10"6 + 0.40 Eq. 2
where the variables are defined as above.
Note that, in both of the above analyses, emissions from the quench
tower were compared to a quench water solids factor that combined the con-
centration of TDS in the quench water and the quantity of water lost up the
stack. In order to isolate the impact of solids concentration, these same
tests were used to compare particulate emissions to the concentration of
total solids (both dissolved and suspended). Total solids (TS) was chosen
as a parameter because it is likely that the mechanisms which lead to the
emission of TDS will also lead to the emission of suspended solids. Hence,
total solids concentration may better indicate the impact of water quality
on emissions.
The results of the analysis are shown in Figures 2 and 3. Figure 2
represents the results for quench towers having single row baffles. For
both sets of test data (U.S. Steel-Lorain and U.S. Steel-Gary, Tower No. 5)
the results suggest a linear relationship between particulate emissions and
the concentration of total solids in the quench water. Both slopes were
found to be statistically significant at the 5% level. Using the average
data from tests at these two towers, emissions were found to be related to
total solids concentration by the equation:
E = 1.46 x 10"4(TS) + 0.43 Eq. 3
where E = emissions (kg/Mg)
TS = total solids concentration in the quench water (mg/£)
Figure 3 presents the results for three towers with multiple row baf-
fles. The slope of this line is quite small, suggesting that little rela-
tionship may exist between total solids concentration and emissions over
the range of solids concentration tested (300 to 2,000 mg/2). However, the
slope of the line was not found to be statistically significant, and the
results are for a narrow range of total solids concentrations. Further data
are needed to evaluate the impact over a wide range of solids concentrations.
72
-------
Jl
o
U 3
O)
I
" g 2
« 1
•• U.S. Steel Lorain
•A U.S. Steel Gary
Average
E = 1.46xlO-4 (TS)+0.43
E = 0.84xlO-4 (TS) + 1.0
E = 1.73x10-4 (TSJ+0.17
J_
J_
1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 11,000 12,000 13,000
Concentration of Total Solids (mg/l)
Figure 2. Relationship of particulate emissions to quench water solids
concentrations for towers with single row baffles.
-------
I
4r-
s 2
n ^
• U.S. Steel Gary
A Dofasco
• Donner Hanna
U.S. Steel Only
All Tests
E =4.02x10-5 (TS)+0.227-
(TS)+0.202
1000 2000
Total Solids Concentration in Quench Water (mg/l)
3000
Figure 3. Relationship of Particulate Emissions to quench water solids
concentrations for towers with multiple row baffles.
-------
In summary the test data suggest that, for towers with single row baf-
fles, particulate emissions can be reduced by controlling the quantity of
solids in the quench water. However, note that with respect to suspended
solids the conclusion is weak, in that for all available tests the quantity
of suspended solids was minor in comparison to the quantity of dissolved
solids. Over the range of total solids concentrations (300 to 2,000 mg/£),
no statistically significant relationship was found between particulate emis-
sions and quench water solids for composite data for towers with multiple
row baffles. Future testing may provide a clearer understanding of this
relationship.
CONTROL OF QUENCH WATER SOLIDS
In the typical quenching process, water is pumped from a sump to a hold-
ing tank and then supplied to the spr.jy nozzles by gravity flow. As the
water is sprayed onto the hot coke, about 25 to 30% is "removed" from the
quench tower by one of three paths: (1) the water is evaporated and carried
to the atmosphere as vapor; (2) droplets are captured in the quench tower
draft and exhausted to the atmosphere; and (3) the water is carried out in
the quench car as moisture in the coke. Most water is lost via the first
two pathways. The water that is not Lost is captured and recirculated to
the holding tank through the sump. Makeup water is added to the sump to
replace that which is lost from the sump.
Solids can be added to the quencci water at two points in the cycle:
from the coke to the sump return water or with the makeup water. Data in-
dicate that few added solids enter the system with the return water. In
fact, in most cases the concentration of total solids in the return water
is less than that in the quench water. Control of these solids beyond that
typically practiced is unlikely in that dissolved solids in the return water
cannot reasonably be controlled, and most added suspended solids are now
controlled by settling in the sump.
The primary source of solids in the quench water and the source which
is most amenable to added control is the makeup water to the sump. Potential
sources of makeup water for the quench tower include the various effluent
streams from the by-products plant, blowdown from process cooling waters,
blowdown from wet scrubbers used for charging or pushing control, and local
surface or ground waters. The solids concentrations of these streams and
the quantities of water available for makeup are shown in Table 1. For pro-
cess waters, the table also shows the potential incremental increase in emis-
sions that can be expected from the use of process waters rather than natural
water sources. These estimates are based on Equation 1 and the assumptions
that all makeup water is eventually emitted from the tower and that total
solids in natural water sources average 300 mg/£. The data in Table 1 sug-
gest that the sources of makeup water which potentially have a major impact
on quench tower particulate emissions are the excess ammonia liquor stream
(sometimes called flushing liquor or weak ammonia liquor) and blowdown from
coke plant scrubber systems.
75
-------
TABLE 1. SOURCE OF QUENCH TOWER MAKEUP WATER
Water source
Excess ammonia liquor
Firinl cooler uluwdowii:
Once through
Recirculate
Benzol plant:
Once through
Recirculate
Barometric condenser:
Once through
Recirculate
Scrubber blowdown:
Charging
Pushing
Other noncontact cooling:
Water blowdown
Natural water sources:
Lake Erie
Lake Michigan
Ohio/Mahoning River (avg.)
Allegheny/Monongahela
River (avg.)
Typical industrial (avg.)
Quantity of water
available
for quenching
SL H20/Mg coke3
79-430 (158.2)
46-625 (378)1
8.3-42 (38.3)
625
63-500 (71.6)
83-625 (625)1
12-42 (46)
198f
625 f
625
625
625
625
625
625
TDS
(mg/2)
!5,875b
262. 2d
840e
K
1,054°
l,054b
NA
NA
450j
450e
510e
1708
160?
ISO8
3738 h
42-435 (171)"
TSS
(mg/Jd)
59C
40C
40C
67c
67C
NA
NA
3,202;!
3,202
32e
NA
NA
NA
NA
NA
Potential1^
Incremental Emissions
Contributions
(kg/Ng)
0.50
< 0.001
0.004
0.10
0.012
-
-
0.13
0.42
0.030
-
|
-
-
"
Unless otherwise stated, ranges were obtained
from Reference 5, pp. 37-41. Averages shown in
parenthesis are based on the data from ques-
tionnaire responses, Reference 5, p. 46.
Maximum water use is assumed to be 625 £/Mg
,of coke.
Unpublished data received from E. Dulaney,
U.S. EPA, March 20, 1980.
^Reference 6. pp. 42-45.
"Reference 7.
'^Reference 8.
'Based on average of all plants,
Reference 5, p. 46.
^Reference 9.
Based on TDS levels in public water
supply for 20 cities in which coke plants
.are located.
Average of high and low values from
.pp. 37-41, Reference 5.
^Assume equal to pushing controls.
Assumes 20% of total solids are emitted
as particulate and 300 mg/£ total solidf
for natural water source.
NA - Data not available.
-------
IMPACT OF BAFFLES ON QUENCH TOWER EMISSIONS
Historically, the method most frequently used to reduce particle and
droplet emissions and subsequent fallout around the quench towers has been
the installation of baffles (sometimes called mist eliminators) in the tower
above the sprays. The baffles remove particulate and droplets from the emis-
sions stream primarily by impaction. As described in References 3 and 4, a
wide variety of baffle designs have been used in domestic quench towers to
enhance this removal.
Few emissions data are available to directly evaluate the impact of
such parameters as baffle configuration (e.g., number of rows, angle of baf-
fles, spacing between baffles, and size of baffles) and the vertical loca-
tion of the baffles in the tower on baffle performance. Because of this
data gap, an engineering model has been used to estimate the performance of
baffles. '10 The paragraphs below briefly describe the limited data, pre-
sent the model that has been developed, and evaluate baffle performance
based on that model.
No domestic test data which were obtained by isokinetic or near-
isokinetic conditions and taken upstream and downstream of baffles or before
and after baffle installation have been identified. The only such data that
we have identified were obtained by Jackson and Waple11 in Scotland in the
late 1950's. In a series of tests using a suction sampling system and a
1-1/2 in. cyclone for collection, emission measurements were taken under
five sets of baffle conditions: (a) no baffles; (b) wooden chevron baffles
located in the top of the tower; (c) wooden chevron baffles located imme-
diately above the sprays; (d) corrugated asbestos baffles near the top of
the tower; and (e) corrugated asbestos baffles located immediately above
the sprays. The results indicated that for the last three conditions the
baffles were quite ineffective, providing only about 20% control. However,
the chevron baffles installed in the top of the tower reduced particulate
emissions by about 60%. While these results provide some insight to the
performance of baffles, they should be viewed with some skepticism in that
little is known about process parameters, specific test results, or the ac-
curacy and precision of the sampling method.
In the absence of test data to definitively evaluate the performance
of quench tower baffle systems, theoretical models that estimate perfor-
mance have been examined. The model which we have chosen is described by
Gorman, et al.,3 and Ertel.10 This model is based on the works of Calvert,
et al.,12'13 on zigzag baffle arrangements for mist eliminators in wet scrub-
bers. The choice of the model that represents zigzag baffles is appropriate
in that almost all baffles currently employed in domestic quench towers are
either zigzag baffle arrangements or are arrangements that can be approxi-
mated as a sequence of zigzag baffles.
Based on Calvert's model, the following expression for baffle effici-
ency as a function of baffle design characteristics and emission stream
properties was developed:3
77
-------
- d2 p U n8
E = - P
(5157)M
o
where E = collection efficiency, fraction
d = particle diameter, cm
pP = particle density, g/cm3 (this value is assumed to be
" 1.4 since most particles are either coke or water)
U = superficial gas velocity, cm/sec
n = number of rows of baffles
8 = angle of baffle to flow path, degrees
|j = gas viscosity, poise (assumed to be 2 x 104)
b8 = horizontal spacing between two consecutive baffles
in same row, cm
Note that the design parameters influencing baffle efficiency are the number
of rows (or bends) in the system, the angle of each row or bend with respect
to direction of gas flow, and the horizontal spacing between two baffles on
the same row.
Although no quench tower data are available to compare measured effici-
ency with the estimated efficiency predicted by Equation 4, Calvert found
that the theoretical collection efficiency compared well with the measured
efficiency on tests of water droplet removal by a zigzag baffle mist elimi-
nator. Further, these measured efficiencies were not greatly altered by
entrained particulate in the water droplets. Both the physical character-
istics of the quench tower emissions stream and available test data suggest
that most particulate emissions are entrained in water droplets. Hence,
Calvert1 s zigzag baffle model provides a reasonable model for quench tower
baffle performance.
As shown :in Figure 4, Equation 4 was used to develop performance curves
for various baffle arrangements. These curves are plots of baffle penetra-
tion (one minus efficiency) versus particle (or droplet) size for three
velocities which span the range of typical quench tower operating conditions.
These curves show the dependency of performance on both particle size and
gas velocity. They indicate that most arrangements are not very effective
for particles (or droplets) less than 10 (Jm in diameter and are highly effi-
cient for particles greater than 150 jjm in diameter. Hence, the particle
sizes of greatest concern when comparing different baffle arrangements are
in the 10- to 150-(Jm range.
CONTROL SCHEMES AND THEIR IMPACTS
For any individual quench tower a range of controls is available, in-
cluding control of quench water solids concentrations over a continuum of
about 300 to 12,000 mg/£ and a variety of specific baffle configurations.
78
-------
5 10 SO 100 200 300
PortlcU Olomtcr. /tm
(a) Single row, 20*
(b) Double row, 20C
'I 3 10 30 100 200 300
Portlcb Dlmtor. urn
(c) Double row, 45
I 3 10 » 100 200 300
Poitlcb Olomnr. /i«
(d) Triple row, 45C
Figure 4.
(e) Carl Still
Performance curves for various baffle designs.
79
-------
For our analysis of the impacts of water quality and baffle design control
measures, we h.ive chosen six control schemes which represent the range of
available controls. These six control schemes are defined in Table 2.
The definitions of these control schemes and the subsequent discussions
of emissions and secondary impacts are based on our analysis of currently
available data. These data are limited, particularly with respect to the
impact of multiple row battles on emissions. The U.S. EFA is currently
conducting extensive tests at DOFASCO to address these limitations. The
results of those tests may modify the analysis and the conclusions that
are developed in the following sections.
EMISSION IMPACT
As indicated earlier, for towers with single row baffles, the test data
suggest a linear relationship between particulate emissions and the concen-
tration of tot.il solids in the quench water. Using Equation 3 and the best
estimate of solids concentration from Table 2, average particulate emission
factors for the first three control schemes are eslimated to be:
TABLE 2. QUENCH TOWER CONTROL SCHEMES
Total Solids
Concentration (mg/JJQ
Control
Scheme Water Control
Range
Current
"Best" Baffle
Estimate Type
1
2
No Control
Eliminate excess ammonia
5,000 -
1,000 -
10,000
2,000
5,500
1,800
Single Row
Single Row
liquor and scrubber blow-
down
3 Local surface or ground 200 - 600
water or equivalent
4 No Control 5,000 - 10,000
5 Eliminate excess ammonia 1,000 - 2,000
liquor and scrubber blow-
down
6 Local surface or ground 200 - 600
water or equivalent
500 Single Row
5,500 Multiple Row3
1,800 Multiple Row
500 Multiple Row
Multiple row baffles are those which are "most efficient" as defined by
Equation 4. Our data include a double row of 20° baffles (DOFASCO), a
double row of 45° baffles (Donner Hanna), a triple row of 45° baffles
(Armco-Houston), and Carl Still baffles (U.S. Steel-Gary, Tower No. 3).
80
-------
Control Scheme 1 1.2 kg/Mg
Control Scheme 2 0.69 kg/Mg
Control Scheme 3 0.50 kg/Mg
Hence, Control Scheme 2 results in an average emissions reduction of 42%
and Control Scheme 3 a reduction of 53% from Control Scheme 1.
The data presented earlier also suggest that no relationship may exist
between emissions and quench water total solids concentration for towers
having multiple row baffles and solids concentrations of less than 2,000 rag/
liter. This fact is suggested by the emissions data shown in Figure 5.
Since no such relationship could be found, emissions for Control Schemes 5
and 6 are estimated to be 0.25 kg/Mg, the average of the five tests shown
in Figure 5. This represents a 79% reduction from Control Scheme 1.
No test data are available to estimate the emissions rate for Control
Scheme 4. Hence, the emissions rate for Control Scheme 1, the particle size
data from tests at U.S. Steel-Lorain,14 and Equation 4 were used to estimate
emission rates for Control Scheme 4.
For Control Scheme 1, the particle size data from Lorain are used to
divide the emissions into two size ranges, those less than 10 |Jtn and those
greater than 10 |jm in diameter. This split results in the following size
distribution:
Emission Factor (kg/Mg)
< 10 (Jm > 10 (Jm Total
Control Scheme 1 0.70 0.50 1.2
To calculate the emissions levels, it was assumed that for the particu-
late less than 10 pm in diameter, single row baffles had no control, while
multiple row baffles had about 10 to 25% efficiency. These values were based
on the efficiency curves shown in Figure 4. Data were insufficient to deter-
mine particle size distribution of the particles >10 [an in diameter. Since
Equation 4 is highly sensitive to size differences over this range, a sensi-
tivity analysis was performed using estimates of mean particle size of 20
and 150 pm for the particles greater than 10 |Jm. Given the amount of fairly
large droplets containing particulate in the quench tower emissions stream,
20 |Jm appears to be a reasonable minimum for the mean size for the +10 |jm
particles. The maximum mean size of 150 |jm was calculated by Laube,
et al.,14 as the largest mean size that can be expected based on settling
of larger droplets. The results obtained from applying equations to the
data above are:
81
-------
H-
TO
i-l
fD
Emissions, Kg/Mg
00
ro
C en
M H-
ft O
i-j
O
p-
O
O
3
rt
ft
3
fD
tn
(X
ro
O)
Oi
O
0
« 5.
C •*•
» o O
00
Z7 3
-I 3
1
US Steel-Gary
Tower No. 3
Donner
Hanna
L.
C-3
DDFASCO
US Steel-Gary
Tower No. 3
Armco
Houston
* *
-------
Emission Factor (kg/Mg)
< 10 pro > 10 \im Total
Control Scheme 4 0.52 - 0.63 < 0.0001 - 0.28 0.52 - 0.91
These data suggest a particulate reduction of 24 to 57% for Option 4 over
Control Scheme 1.
In order to validate the conclusions, the same analysis was performed
for Control Schemes 5 and 6 using the emissions estimates for Control
Schemes 2 and 3 and assumptions on particle size from the U.S. Steel-Lorain
data. We found that the model predicted emissions of 0.38 to 0.56 kg/Mg
for Control Scheme 5 and 0.32 to 0.43 kg/Mg for Control Scheme 6.
While the above analyses provide an indication of the performance of
multiple row baffles, the results should be viewed with some skepticism in
that: (a) no test data are available on the accuracy of Calvert's model in
predicting quench tower baffle performance; and (b) the available particle
size data are quite limited. We anticipate that the results from the current
DOFASCO tests will provide added data on baffle performance and particle
size. These data should enable a more certain evaluation of performance
for Control Scheme 4.
SECONDARY IMPACTS
Each of the control schemes described above has costs and secondary
environmental impacts. These impacts influence the choice of control for a
particular tower. The magnitude of the impacts is an important factor in
the need for more knowledge of the performance of Control Scheme 4.
The use of multiple row rather than single row baffles has minimal cost
and no secondary environmental impacts. Data from quench tower manufacturers
indicate that the installation of multiple row baffles will increase the
capital and annualized cost of baffling about 40%. There are no secondary
environmental impacts since the particulate captured by the baffles is re-
turned to the sump and eventually settles out as sludge. The sludge, which
primarily is comprised of coke breeze, is periodically removed from the sump
and is used in the coke oven or as a fuel for sintering.
On the other hand, the control of quench water quality may have signifi-
cant cost and secondary environmental impacts. These impacts are related
to the disposal of wastewaters that are excluded from use in the quench tower.
Before these waters can be discharged, they must be treated to meet BAT (Best
Available Technology Economically Achievable) or pretreatment standards.
The capital cost of equipment necessary to treat the effluent is quite high.
The annualized cost of treatment of water replaced in the quench tower is
estimated to be in the range of 75 to 200% of the annualized cost of a new
quench tower (depending upon the quantity of water which must be treated).
83
-------
In addition to the cost impact, the control of water also has secondary
environmental impacts. Any coke plant wastewaters not used in the quench
water are added to the plant effluent. These waters will be treated to meet
BAT standards. While this treatment will remove most of the organic constit-
uents, oil and grease, and suspended solids from the water, it will not affect
the dissolved solids, primarily chloride salts. These salts may adversely
affect surface waters near these plants.
Treatment of the wastewater will also produce a sludge which must be
disposed of. These sludges will include biological sludge and lime or caus-
tic soda from treatment of excess ammonia liquor in lime stills. Because
of their chemical composition, the liaie still sludges have been classified
as a hazardous waste, increasing disposal problems for the waste sludges.
CONCLUSIONS
1. For towers with single row baffles, particulate emissions
are linearly related to the solids concentration in the
quench water over the range of solids concentrations found
in domestic quench towers.
2. No relationship is found between particulate emissions and
solids concentrations in the range of 300 to 2,000 mg/£ for
quench towers with multiple row baffles.
3. More emission test data are needed to define the per-
formance of multiple row baffles when used in quench towers.
4. Because of the large cost and secondary environmental im-
pacts associated with the control of quench water solids,
knowledge of multiple row baffle performance under
high solids conditions is essential.
84
-------
REFERENCES
1. Bloom, B., and J. Jeffery. A Critical Review of Coke Plant Quench
Tower Particulate Emission Rates (1950-1980), What We Know and What We
Don't Know. Proceedings: Air Pollution Control In The Iron and Steel
Industry, Specialty Conference. Air Pollution Control Association.
1981.
2. Gorman, P. G., L. J. Shannon, D. Wallace, and F. Hopkins. Engineering
Analysis of Emission Controls for Wet Quench Towers. U.S. Environmental
Protection Agency, Contract No. 69-02-2609, Assignment 7. January 1979.
3. Coke Wet Quenching - Background Information for Proposed Standards Draft
EIS, Preliminary Draft. U.S. Environmental Protection Agency. March
1981.
4. Edlund, C., A. H. Laube, and J. Jeffery. Effects of Water Quality on
Coke Quench Tower Emissions. Paper 77-6.3, presented at the 70th Annual
Meeting of the Air Pollution Control Association. June 1977.
5. Development Document for Proposed Effluent Limitations Guidelines, and
Standards for the Iron and Steel Manufacturing Point Source Category,
Vol. II By-Product and Beehive Cokemaking Subcategory-Draft. EPA
440/l-79/024a. U.S. Environmental Protection Agency, Washington, D.C.
October 1979.
6. Laube, A. H., and B. A. Drummond. Coke Quench Tower Emission Test Pro-
gram, EPA 600/2-79-082. U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. April 1979.
7. Letter. H. Withers, Alabama.By-Products Corp., to D. Goodwin, U.S.
Environmental Protection Agency. May 22, 1979.
8. Letter. C. G. Cramer, Armco Steel Corp., to E. Cohen, U.S. Environ-
mental Protection Agency. December 28, 1978.
9. Lohr, E. W., and S. K. Love. The Industrial Utility of Public Water
Supplies in the United States, 1952. U.S. Government Printing Office,
Washington, D.C. 1954.
10. Ertle, G. L. Quench Tower Particulate Emissions. Journal of the Air
Pollution Control Association, 29(9) pp. 913-919. September 1979.
11. Jackson, R., and E. R. Waple. The Elimination of Dust and Drizzle from
Quench Towers. The Gas World-Coking, pp. 75-84. May 7, 1960.
12. Calvert, S., I. Jashnani, S. Yung, and S. Stallerg. Entrainment Separa-
tors for Scrubbers—Initial Report, EPA 650/2-74-119a, PB-241 289.
U.S. Environmental Protection Agency, Research Triangle Park, NC.
October 1974.
85
-------
13. Calvert, S., S. Yung, and J. Leung. Entrainment Separators for Scrub-
bers—Final Report, EPA 650/2-74-119b. U.S. Environmental Protection
Agency, Washington, D.C. August 1975.
14. Laube, A. H., J. Jeffery, and D. Sommerer. Evaluation of Quench Tower
Emissions, Part II, Draft Report. U.S. Environmental Protection Agency.
Contract No. 68-01-4138, Task 3. December 1977.
86
-------
CURRENT REGULATIONS AND CONTROL PERFORMANCE FOR VISIBLE EMISSIONS
FROM WET-COAL CHARGING, DOOR LEAKS, AND TOPSIDE LEAKS
by: Marvin R. Branscome (7-3)
Research Triangle Institute
Research Triangle Park, North Carolina 27709
(919) 541-6956
William L. MacDowell (5AHAP)
U.S. Environmental Protection Agency
Chicago, Illinois 60604
(312) 886-6043
ABSTRACT
The Research Triangle Institute, under contract to the U.S.
Environmental Protection Agency (EPA), compiled background information on
emissions from wet-coal charging, door leaks, and topside leaks from coke
ovens. The study was undertaken as part of EPA's effort to examine the
need for coke oven regulations. This paper summarizes a portion of the
background information which was compiled. Current regulations from State
Implementation Plans, consent decrees, and Occupational Health and Safety
Administration (OSHA) requirements are summarized. Emission test results
and performance data in terms of visible emission control are presented.
1.0 INTRODUCTION
Nearly all of the metallurgical and foundry coke produced in the
United States is produced in slot-type, by-product recovery coke ovens.
There are approximately 13,000 such ovens located in 199 batteries at 58
plants. Ninety-three percent of coke-making capacity is owned by integrated
steel companies and the remainder is owned by independent merchant coke
producers. Coke oven batteries are located in 17 states, with 55 percent
of the capacity in Pennsylvania, Ohio, and Indiana.
The emission points that were examined include the wet-coal charging
operation, door leaks, and topside leaks. In a wet-coal charging system,
gases and particulate can be emitted from the oven charging ports throughout
the 3 to 8 minute charging period. Charging emissions may be visible at
the point where the charging car drop sleeves meet the charging ports, or
from the top of the charging car if emissions have escaped up through the
drop sleeves and coal hopper. During the coking period (usually 16 hours
or more), emissions may leak from the doors, charge port lids, and offtake
87
-------
piping system on the oven. In addition, leaks occasionally occur in the
collecting main(s) which carries the off-gases to the by-product recovery
plant.
Since 1975, methods have been developed by EPA and State agencies to
measure visible emissions from coke oven charging, door leaks, and topside
leaks. The method for charging most widely used involves reading not
opacity, but cumulative seconds of visible emissions during the charging
period. When coal begins flowing from a charging car hopper into the oven,
the observer starts recording on an accumulative stopwatch the time during
which he sees any visible emissions from the charge ports, or from the tops
of the charging car hoppers. The stopwatch may be started and stopped
several times during a charge. The observation ends when the last charge
port lid is replaced. After several consecutive charges are observed, a
sum or average of the number of seconds of visible emissions associated
with charging is determined.
Visible emissions from door leaks are measured by a walk around the
battery with the observer pausing at each oven to observe leakage from the
doors and door areas out to the buckstays. Any visible emission qualifies
the door as leaking. When machinery blocks the observer's view of a door,
he may return later as long as the total time to read all doors does not
exceed 45 minutes. The percent of the total number of doors on operating
ovens that are leaking is then determined and the result is recorded as
"percent leaking doors" or "PLD."
Topside leaks are measured in the same way. Leaks from charge ports
and offtake systems are recorded separately and a "percent leaking" value
is determined for each. The abbreviations are PLL for percent leaking lids
(i.e., charge port lids) and PLO for percent leaking offtakes.
2.0 CURRENT CONTROL REQUIREMENTS
Coke oven batteries are currently subject to the control requirements
of State implementation plans (SIP's) and Occupational Safety and Health
Administration (OSHA) standards. In addition, some batteries must meet the
requirements of consent decrees that have been negotiated on a plant-by-
plant basis. SIP's and consent decrees contain emission limits for
charging, door leaks, and topside leaks. The OSHA standards regulate
worker exposure and specify certain required engineering and work practice
controls.
Table 1 presents a summary of current SIP emission limits. These SIP
limits are at different stages of the federal approval process. Charging
limits range from 60 seconds as a sum of 4 consecutive charges to 170
seconds for 5 charges. Door leaks range from 10 to 16 PLD, charge port
lids from 2 to 5 PLL, and offtake systems from 5 to 10 PLO. Measurement
methods vary somewhat, but most are very similar to what have been described
above. EPA has issued Reasonably Available Control Technology (RACT)
guidance for iron and steel particulate emission sources including coke
88
-------
ovens (45 FR 59198). That guidance calls for 25 seconds per charge averaged
over 4 to 7 charges (with an optional exclusion of the highest reading in
20 observations), 10 PLD, 5 PLL, and 10 PLO. Based on current regulations,
68 to 100 percent of coking capacity would be at the RACT guidance level or
below. A summary of current regulations and the number of affected
batteries is given in Table 2. The effect of excluding one charge in 20
observations has not been considered in this analysis.
TABLE 1. STATE REGULATIONS FOR COKE OVEN EMISSIONS
State
Charging
(seconds per charge)
Percent maximum leaking
Doors Lids Offtakes
Alabama
California
Illinois
Indiana
Maryland
Michigan
Missouri
New York
Ohio
Pennsylvania
Wisconsin
a
60/4
170/5D
125/5°
160/5
80/4
120/6
150/5C
170/5
75/4
h
15
10
IOH
10d
10 e f
10-12 '
15,
iof
16
108
10
5
3
5
3
3
4
2
2
5
2
5
10
10
10
10
10
4
10
5
10
5
10
.Visible emissions £ 20 percent opacity except for £ 3 min/hr.
200/5 for existing 5-meter, 3-hole batteries.
jMay exclude one in 20 charging observations.
Excludes four doors.
fTen percent for short batteries and 12 percent for tall batteries.
Chuck doors are counted as separate doors.
"Excludes two doors, counts all door area leaks.
Equipment and work practice requirements.
Consent decrees have been signed for batteries representing about
one-third of coke-making capacity. These decrees were generally negotiated
as settlements of outstanding SIP violations and may specify required
equipment and work practices as well as visible emission limits. All but a
few consent decree limits are equal to or more stringent than the RACT
guidance level. The lowest emission limits currently in effect apply to
new or rebuilt batteries. Several consent decrees for these new batteries
have limits of 55 seconds for 5 charges, 5 PLD, 2 PLL, and 5 PLO.
In 1976, OSHA promulgated standards designed to protect workers from
exposure to coke oven emissions. One part of the rule sets a limit on
worker exposure which is to be monitored quarterly by the company and
during OSHA inspections by OSHA personnel. Another part specifies
89
-------
engineering and work practice controls and includes general prohibitions on
fugitive emissions. Because specific visible emission limits are not given
in the rule, direct comparison of OSHA requirements with SIP and consent
decree limits is not possible.
TABLE 2. SUMMARY OF BASELINE CONTROL REQUIREMENTS
Charging limit (avg s/chg)
No limit
30-40
25
19-20
15
11-12
Percent leaking doors
20
15-16
VL2
10
8
4-5
Percent leaking lids
5
4
3
2
1
Percent leaking offtakes
15
10
6
5
4
Number of
batteries
CHARGING
23
44
32
62
7
15
DOOR LEAKS
1
51
76
46
3
6
LID LEAKS
58
9
40
71
5
OFFTAKE LEAKS
1
109
3
56
14
Percent of
capacity
7.4
24.7
21.7
33.4
2.9
10.0
0.5
21.5
48.2
24.4
1.7
3.7
26.4
7.3
21.8
41.7
2.7
0.5
55.4
1.7
32.3
10.0
90
-------
3.0 EMISSION ESTIMATES
Coke oven emissions consist of a yellow-brown gas which contains over
10,000 compounds as gases, condensible vapors, and particulates. The
components of primary concern include the known or suspected carcinogens
belonging to a class of compounds termed polycyclic organic matter (POM).
POM, which condenses on fine particulates at ambient temperatures, consists
of compounds with two or more fused rings. There are thousands of POM
compounds which vary widely in physical and chemical characteristics.
These pollutants are sometimes reported as benzene soluble organics (BSO)
or as a quantity of a specific surrogate compound, such as benzo(a)pyrene
(BaP). BSO is composed of many compounds, some of which are not POM.
The emission of pollutants is generally characterized by both the
concentration and the flow rate of the pollutant stream. However, these
characteristics are difficult to apply to emissions from coke oven doors,
lids, offtakes, and charging. The rates of emissions are highly dependent
on the time into the coking cycle, the gap size, the number of gaps, and
oven pressure. The concentration of pollutants also varies with time, and
there may be a variation in the concentration of BSO from battery to battery
caused by operating conditions and the coal type or blending practices.
Even if the leaks were well characterized in terms of the size and length
of the gap, there would be potential difficulties in assessing the flow
rate of the pollutant. Measurement of the concentration of the BSO above
the coke battery is difficult because the concentrations are transient.
The monitored particulate concentration is a function of the location of
the sampler, the existing wind conditions, and interference from other
emissions.
The collection and measurement of fugitive coke oven emissions is
further complicated by the fact that the gases which are emitted from the
oven condense on metal surfaces present in the sampling system. These tars
even condense on the hot oven jambs. This condensation can lead to
erroneous results when the gases are carried through ducts before they
reach the sampling device.
3,1 CHARGING EMISSIONS
Particles emitted during the charging cycle have been identified as
coke balls, pyrolitic carbon, high-temperature coke, char, coal, mineral
matter, and fly ash.1 2 For this type of fugitive emissions source,
collection of representative emission samples is extremely difficult and,
consequently, very little data on mass rates are available. Estimated
emission factors in the literature vary by at least one order of magnitude,
and the accuracy of the emission factor should be judged according to this
variation. The following emission estimates .are given in terms of g BSO/kg
coal and g particulate/kg coal.
With the assistance of the American Iron and Steel Institute, EPA has
compiled and analyzed data on particulate emissions from iron and steel
91
-------
mills to assist in the definition of particulate emission factors for each
process. This study suggests an emission factor for uncontrolled charging
of .25 to .75 grams of particulate per kilogram (0.5 to 1.5 Ib/ton) of coal
charged.3 This range is based on measurements of particulate deposition on
greased plates and from the solids collected from tests of a scrubber to
control charging emissions. The mid-range value yields an estimate of
about 0.5 g particulate/ kg coal.
In addition to the AISI estimate, a test was conducted by EPA at J&L
Steel, Pittsburgh, to compare the American Iron and Steel Institute (AISI)
larry car with a conventional Wilputte larry car.4 Samples were collected
by putting enclosures around the Wilputte larry car drop sleeves and
evacuating the emissions through a stack where they could be sampled.
Isokinetic conditions could not be maintained because of a high variability
in the flow of emissions. The reported particulate measurements represented
composite samples from different emission points. The particulate catch
averaged 815 g/charge, or about 0.05 gram of particulate per kilogram of
coal (0.1 Ib/ton) from tests of 10 charges with an average sampling time of
3.5 min. The average amount of BSO measured (excluding the impinger catch)
was 57 percent. The impingers averaged 96 percent of the mass collected in
the front of the sampling trains and contained an average of 60 percent
BSO.4 The particulate emission factors were combined with these results to
calculate a BSO emission factor of 0.055 g BSO/Kg coal. The particulate
emissions are based on particulate captured by the filter and do not reflect
the BSO collected in the impingers. The resulting mass emission estimates
for poorly-controlled charging is 0.05 to 0.5 g particulate/kg coal, or
roughly 0.055 to 0.55 g BSO/kg coal.
Emissions from uncontrolled or poorly-controlled charges generally
appear as dense clouds. In contrast, during observation of charges
controlled by the stage charging operating procedure, EPA observers noticed
that during good charges (a small duration of visible emissions), the
emissions were generally small wisps or puffs which drifted from around the
drop sleeves on the larry car. For charges where the duration was longer,
the emissions changed to clouds of smoke which escaped to the atmosphere
with higher velocities. Generally, the longer the duration, the more large
clouds and fewer wisps were seen. A series of inspections were conducted
in which one inspector recorded seconds of emissions greater than 20 percent
opacity, and another inspector recorded seconds of any visible emissions.
For charges with 25 seconds or greater of any visible emissions, there was
very little difference in the two methods. For charges with less than 25
seconds of emissions, the difference in the two methods is significant.
For example, the variance in the two methods is 21 percent for 15 seconds
of emissions, and the variance increases to 47 percent for 5 seconds of
emissions. These results imply that the concentration of pollutants is
less for a short duration of visible emissions which are often composed of
wisps with an opacity less than 20 percent.
92
-------
3.2 DOOR LEAK EMISSIONS
The fugitive nature of door leaks has posed the same collection and
sampling problems that were previously discussed. Probably the most
reliable coke oven door data are those gathered on BSO emissions from
cokeside sheds. A cokeside shed is a large hood which extends over the
entire coke side of the battery to capture both pushing and cokeside door
emissions. Available cokeside shed test results are generally
representative of high levels of percent leaking doors (i.e., levels greater
than 30 PLD). Emission factors for door leaks are given as g/kg coal and
as an average rate for a leaking door.
In May 1977, EPA conducted four 10-hour tests of Wisconsin Steel's
shed that covered the coke side of 45 5-meter ovens.5 Sampling was
discontinued during pushing so that the data would only reflect emissions
from doors. BSO emissions (from the full sampling train) during the test
averaged 6 kg/hr from an average of 31 leaking doors (70 PLD) with a
resulting emission factor of about 0.25 g BSO/kg coal (or about 0.2 kg
BSO/hr for a leaking door). Particulate emissions (front half) ranged from
0.17 to 0.21 g/kg coal, or 65 to 85 percent of the BSO.
A similar test was conducted at Armco, Inc. in Houston, Texas, in
October 1979.6 7 The Armco shed encloses the coke side of 62 4-meter
ovens. Three tests conducted during nonpush periods measured 6.8 to 13
kilograms of BSO per hour from 10 to 24 leaking doors (16 to 39 PLD).6 The
resulting emission rate is about 0.6 kg BSO/hr for a leaking door or 0.4 to
0.8 g BSO/kg coal.
Bethlehem Steel sampled emissions from its Burns Harbor shed on Battery
I.5 The shed covers the coke side of 82 6-meter ovens. During these
tests, BSO emissions during nonpush periods averaged 3.9 kg/hr; the number
of doors leaking was not reported. Bethlehem Steel also sampled emissions
from temporary stacks mounted on pusherside doors at their Burns Harbor
plant. A total of 14 samples were collected at four doors that were
completely enclosed between buckstays. Toluene soluble organics averaged
0.22 kg/hr for each door.
The shed test data, summarized in Table 3, reveal a range of BSO
emissions of approximately 0.2 to 0.7 kilogram of BSO per hour per leaking
door for the tests where the number of leaking doors was recorded. A range
of 0.25 to 0.8 g BSO/kg coal was derived from these tests. Based on the
Wisconsin Steel tests, particulate emissions (front half) are estimated as
0.16 to 0.68 g/kg coal.
93
-------
TABLE 3. COKESIDE SHED TEST RESULTS
Test
Kilograms of Number of Kilograms of BSO/
BSD/hour leaking doors PLD hour/leaking door
Wisconsin Steel Shed
Average
ARMCO, Inc. Shed
Average
7.0
5.9
5.4
6.0
6.1
6.8
11
13
10.3
33
35
27
31
32
10
19
24
18
73
78
60
69
70
16
31
39
29
0.21
0.17
0.20
0.19
0.19
0.68
0.59
0.55
0.58
3.3 TOPSIDE EMISSIONS
An emission test was conducted at U.S. Steel's, Clairton Battery 1 by
EPA in August 1978 to measure topside leak emissions.8 During the second
hour of coking, samples were collected from a vent pipe which had been
installed on a charging port lid. The leak rate was adjusted to yield
small leaks with a 0.3-meter (1-foot) visible plume and large leaks with a
1- to 2-meter (3- to 6-foot) visible plume. The results, listed in Table
4, show a range of 0.0017 to 0.0053 kg/hr ,for a small leak, with an average
rate of 0.003 kg/hr. Emissions from the large leak ranged from 0.012 to
0.035 kg/hr, with an average rate of 0.021 kg/hr. The analysis for BaP
showed that 1.4 to 1.8 percent of the BSO was BaP.
TABLE 4. TOPSIDE LEAK EMISSION TEST8
Leak size
BSO(kg/hr)
Large (1-2 m)
Average
Small (0.3 m)
Average
0.017
0.035
0.012
0.021
0.0017
0.0029
0.0053
0.0033
These emission rates can be used to derive an emission factor in g/kg
of coal by assuming a typical battery on an 18 hour cycle time with 62
ovens, 3 lids and 2 offtakes per oven, and 16 Mg coal per oven. Also
assume that current control performance is 3 PLL and 10 PLO which would
94
-------
yield 18 topside leaks. The emission factor for this case would be 0.001 g
BSO/kg coal if all of the leaks are small, and 0.01 g BSO/kg coal if all of
the leaks have large plumes. Particulate emissions were not measured
during the topside leak test, but they are expected to be similar to door
leak emissions (65 to 85 percent of the BSD).
4.0 CONTROL TECHNOLOGY AND PERFORMANCE
The current control techniques used by the industry were reviewed to
identify equipment and procedures which have been demonstrated for control
of charging, door leaks, and topside leak emissions. The performance data
were collected by EPA personnel or their contractors during official EPA
inspections.
4.1 CONTROLS FOR CHARGING
Current regulations require stage charging and the associated equipment
modifications. Stage charging is a systematic procedure for introducing
pulverized coal into a coke oven so that an open passage is constantly
maintained for the exit of gases to the collecting main. This procedure
allows gases and other matter that evolve during charging to be effectively
contained within the oven while they are being drawn into the collecting
main by steam aspiration and then exhausted through the regular gas handling
equipment to the by-product recovery plant. Containment and removal of
pollutant-laden gases occur with minimal losses to the atmosphere.
The requirements for good stage charging include:
The stage charging operating procedure.
Battery modifications, such as repaving the battery top or
modifying the coal bunker on some batteries.
New or modified larry cars with increased capacity in the outer
hoppers, independently operated drop sleeves, and independent
hopper control of coal flow.
Double drafting with either a second collecting main or jumper
pipe.
Adequate steam aspiration.
Control of coal bulk density.
A smoke boot on the leveler bar.
Gooseneck cleaning.
Decarbonizing equipment.
Training program for employees.
Battery top workers and the operating procedures they follow perform
an equally important role in emission control. Detailed observation of
charging practices revealed that even the best controlled batteries
experienced occasional lapses in work practices or equipment malfunctions
which resulted in higher emission levels. A few of the worker job functions
that are critical to control of emissions from charging include inspection
and cleaning of goosenecks, prompt lid replacement, turning the aspiration
95
-------
system on and off, observing the position of drop sleeves, and spotting the
larry car.
4.2 PERFORMANCE DATA
The U.S. Steel Clairton Works has developed a stage charging system.
In addition to the physical ability to stage charge, emission control was
optimized through extensive training, observation, and monitoring programs.
Topside manpower was increased by creating a utilityman position with the
prime responsibility of providing assistance to the larry car operator for
gooseneck cleaning and other environmental responsibilities. Process
observers were also employed to monitor emission performance and to identify
when corrective action might be needed. Supervisory personnel are informed
of equipment problems and deviations from prescribed work and operating
practices.9
CF&I also reported a program to optimize control of charging emissions.
The company studied the charging procedure and established priorities for
improving equipment and work practices which caused the greatest amount of
emissions. Equipment modifications were made and detailed operating and
inspection procedures were developed. The company also reported that
inspections and recordkeeping are important to identify the cause of control
problems so that the trouble area or condition could be eliminated. The
recordkeeping also monitors progress, obtains operator involvement,
reinforces productive efforts, directs maintenance efforts, and documents
improvements.10
Visible emission data on the CF&I and U.S. Steel Clairton charging
operation were compiled and analyzed. These data included observations
from 92 charges at CF&I and 16 to 65 charges per battery at Clairton. A
statistical analysis of the charging data was performed to obtain confidence
intervals that described individual battery performance. The data were not
characterized by a normal distribution. To allow the application of normal
statistics, various data transformation techniques were investigated to
determine an appropriate transformation that would yield a normal
distribution. A transformation of log (S + 1), where S is equal to seconds
of visible emissions, provided a normal distribution for independent groups
of five or more consecutive observations.
The variance components used in calculating the confidence levels
include the variance between observers within charges, between charges
within days, and between days. Based on 10 observations and the variance
components, the 15 batteries listed in Table 5 had log averages of 0.5 to
11 seconds per charge with a range of 95-percent confidence levels of 1 to
16 seconds per charge. The lowest levels of visible emissions were observed
on Batteries 16 and 17 which were using run-of-the-mine coal. The highest
level was a log average of 11 seconds per charge with a 95-percent
confidence level of 16 seconds per charge.
96
-------
TABLE 5. DATA FOR OPTIMIZED STAGE CHARGING
a
Company
U.S. Steel, Clairton
CF&I, Pueblo
U.S. Steel, Clairton
Arithmetic
Battery average
16
17
B
1
22
10
19
3
7
21
9
2
8
20
11
1.0
1.2
5.4
8.6
6.6
8.8
6.8
7.8
7.6
13.1
8.4
9.0
8.9
12.0
11.6
Log b
average
0.5
0.8
4.5
4.8
5.1
6.0
6.1
6.9
7.3
7.3
7.5
7.7
8.1
8.1
11.1
95-percent
level
1.2
1.7
7.4
7.9
8.3
9.8
9.9 .
11.2
11.6
11.7
12.0
12.4
13.0
13.0
16.1
All results are seconds of emissions per charge.
Log average = e -1, where Y = (InCsj + 1) + In(s2 + 1) + .
In(s10
TABLE 6. ADDITIONAL STAGE CHARGING DATA
Company
Battery
Arithmetic
average
Log
average
95-percent
level
U.S. Steel, Fairfield
Shenango, Neville Island
J&L, Pittsburgh
Lone Star Steel
U.S. Steel, Gary
9
6
3
4
P2
P4
C
1
5.4
10.0
6.1
6.6
7.3
6.2
7.4
13.0
4.0
6.8
5.1
6.2
5.6
5.6
6.6
8.4
6.7
10.9
8.4
10.0
9.
9.
12.0
15.0
.1
.1
97
-------
Other batteries which used the stage charging procedure were also
examined, and although these plants have not reported the extensive
optimization of Clairton and CF&I, the control performance is similar. The
data in Table 6 represent a range of battery types, including single and
double collecting mains, short and tall batteries, different battery ages,
variety of coal sources, and a variety of equipment and construction
features.
4.3 DOOR LEAK CONTROLS
Most batteries control door leaks with doors and seals that are called
"self-sealing." This means that the small gaps between the metal seal and
door jamb are sealed by the condensation of tar from the coke oven gas.
Current regulatory requirements include replacement of damaged doors,
seals, and jambs for some batteries and a basic door leak control program
for all batteries. The door leaks control program involves:
Routine inspection and cleaning.
Proper door placement and seal adjustment.
Maintaining door seals within specifications ("blueprinting").
Prompt replacement or repair of damaged doors, seals, and jambs.
Improvements to the door repair shop.
Metal seals are most effective when they are new, properly adjusted,
and used on relatively clean, straight jambs. Effective sealing can be
inhibited by several factors such as distortion and damage to jambs, doors,
sealing strips, and adjusting hardware. Most of the components of the
oven's doors assembly are tightly constrained; consequently, when the
assembly is heated, gross distortions are prevented. Thermal cycling under
these constrained conditions causes thermal warping and damage to the metal
components. Inspection, maintenance, repair or replacement, blueprinting,
and better materials of construction are the elements that would be
necessary to overcome these causes of door leaks. The basic door leak
control program also includes cleaning and adjustment. Cleaning removes
encrustations which can cause gaps between the sealing edge and jamb.
Proper door placement and adjustment of the seal in place are also important
aspects of effective door leak control.
The batteries at CF&I and U.S. Steel, Clairton have been equipped with
modified door seals and springs constructed with more temperature resistant,
durable alloys. The modified seals, coupled with the basic door leak
control program, have provided consistent control of door leak emissions.
4.4 DOOR LEAK PERFORMANCE
The modified seal technology coupled with a leak control program of
routine inspection, cleaning, and repair has been implemented on the
batteries at CF&I in Pueblo, Colorado, and U.S. Steel in Clairton,
98
-------
Pennsylvania. Other batteries have demonstrated a control performance that
is equivalent to the performance of the CF&I and U.S. Steel, Clairton
batteries without implementing the modified seal technology. However, the
technology for door leak control has been defined to explain one method
that has been used by the industry to achieve effective emission control.
The 19 battteries in the data base are listed in Table 7 and include a
range of battery types with original construction dates that range from
1951 to newly rebuilt.
TABLE 7.- SUMMARY OF DOOR LEAK DATA
Company
Battery
Average PLD
Range PLD
U.S. Steel, Clairton
CF&I
U.S. Steel, Fairfield
3
1
2
7
21
22
19
20
16
10
8
17
11
9
C
B
D
9
2
0.6
0.9
2.2
2.9
3.0
3.4
3.9
4.5
5.0
5.8
6.1
6.2
8.6
9.0
3.2
4.3
5.9
3.9
7.2
0
0
0
0
1
2.3
3.4
2.3
4.2
4.8
5.5
4.9
4.7
3.9
1.1
0.8
1.1
2.7
3.0
- 1.6
- 2.3
- 5.5
- 5.5
- 5.7
- 5.2
- 4.6
- 7.5
- 6.7
- 7.1
- 6.3
- 8.2
- 14.1
- 16.4
- 6.4
- 9.2
- 16.1
- 4.8
- 13.0
The door leak data were characterized by a Poisson distribution.
Analysis of the data revealed that 12 of the batteries averaged less than 5
PLD, and all 19 averaged less than 9 PLD. The 95-percent confidence level
of the highest average, 9 PLD for Battery 9 at U.S. Steel, Clairton, is 12
PLD for an average of three inspections.
4.5 TOPSIDE LEAK CONTROL
Topside leaks occur around the rim of charging port and standpipe
lids; standpipes can also leak at thier bases or through other cracks.
These leaks are primarily controlled by proper maintenance and operating
procedures which include:
99
-------
Replacement of warped lids,
Cleaning carbon deposits or other obstructions from the mating
surfaces of lids or their seals,
Patching or replacing cracked standpipes,
Sealing lids after a charge or whenever necessary with a slurry
mixture of clay, coal, and other materials (commonly called
lute), and
Sealing cracks at the base of a standpipe with the same slurry
mixture.
Because there are many places where leaks can develop, keeping all
charging lid and standpipe leaks sealed is a continuous job. In essence,
success in controlling these emissions is directly related to the amount of
manpower, dedication of the employees, and the priorities of management.
Some equipment designs may reduce the effort required to keep leaks
sealed. Heavier lids or better sealing edges may reduce leaks. Automatic
lid lifters can rotate charging-hole lids after they are seated and provide
a better seal. Even with such equipment, manual effort will still be
required to seal leaks.
4.6 TOPSIDE LEAK CONTROL PERFORMANCE
The technology for controlling lid and offtake leaks is luting
manpower, improved luting mixtures, modification or replacement of offtakes
on some batteries, and the conscientious sealing (luting) of leaks when
they are observed. This technology has been implemented on the batteries
at U.S. Steel, Clairton. The conscientious luting of leaks was observed
during EPA inspections of these batteries.
The topside leak data base is large and is not reproduced here because
of space limitations. The 23 batteries in the lid leak data base include
all of the U.S. Steel batteries at Clairton. Fourteen batteries averaged
0.2 PLL or less during at least one inspection, and the highest average of
all 23 batteries was 1.8 PLL. The 95-percent confidence level associated
with this average is 3 PLL when averaged over three runs. In the offtake
leak data base, 11 batteries at U.S. Steel, Clairton averaged 0.7 to 3.4
PLO. The 95-percent confidence level associated with the average of 3.4
PLO is 6 PLO when averaged over three runs.
These levels of control performance have been achieved through proper
equipment design and maintenance and through consistent application of
luting materials to seal leaks. For many batteries, attaining these levels
of control may require the addition of one worker to the topside area. In
addition, it may be necessary to repair, modify, or replace damaged or
poorly designed offtakes on some batteries.
100
-------
5.0 SUMMARY
Current regulations for visible emissions from wet-coal charging, door
leaks, and topside leaks range from 11 to 40 seconds per charge, 4 to 16
PLD, 1 to 5 PLL, and 4 to 10 PLO. More than 50 percent of the batteries
have current emission limits that are less than or equal to 25 seconds per
charge, 12 PLD, 3 PLL, and 10 PLO. The batteries at U.S. Steel's Clairton
Works and at CF&I have optimized the stage charging procedure. Inspections
at these plants showed varying control performance which ranged from a log
average of 0.5 to 11 seconds per charge with a 95-percent confidence level
of 1 to 16 seconds per charge. Several other batteries have demonstrated a
control performance within this range. The average control performance for
door and topside leaks at U.S. Steel's Clairton Works ranged up to 9 PLD,
1.8 PLL, and 3.4 PLO. The 95-percent confidence levels associated with
these averages are 12 PLD, 3 PLL, and 6 PLO when averaged over three runs.
6.0 ACKNOWLEDGMENTS
This paper represents a portion of the information gathered under
contract to EPA's Office of Air Quality Planning and Standards (OAQPS).
The authors wish to acknowledge the cooperative effort and contributions of
many individuals within OAQPS, EPA's regional offices, State agencies, the
American Iron and Steel Institute, and the American Coke and Coal Chemicals
Institute. These individuals provided guidance, data, and review of the
background information on coke oven emissions.
7.0 REFERENCES
1. Herrick, R. A. and L. G. Benedict. A Microscopic Classification of
Settled Particles Found in the Vicinity of a Coke-Making Operation.
JAPCA. 19:325-328. May 1969.
2. Stolz, J. H. Coke Charging Pollution Control Demonstration. AISI and
USEPA. Publication No. EPA-650/2-74-022. March 1974. 302 p.
3. Cuscino, T. A. Particulate Emission Factors Applicable to the Iron
and Steel Industry. Publication No. EPA-450/4-79-028. September
1979. p. 54.
4. Bee, R. W., et al. Coke Oven Charging Emission Control Test Program--
Volume I & II. Publication No. EPA-650/2-74-062. July 1974.
5. Trenholm, A. R., L. L. Beck, and R. V. Hendriks. Hazardous Organic
Emissions From Slot Type Coke Ovens. USEPA. 1978.
6. Benzene Soluble Organics Study - Coke Oven Door Leaks (Draft). Clayton
Environmental Consultants, Inc. EPA Contract No. 68-02-2817. EMB
Report No. 79-CKO-22. December 1979.
101
-------
7. Coke Oven Emission Testing - Armco Steel Corporation, Houston, Texas
(Draft). TRW Environmental Engineering Division. EPA Contract No.
68-02-2812. (Test conducted October 1979).
8. Hartman, M. W. Emission Test Report - U.S. Steel Corporation.
Clairton, Pennsylvania. TRW Environmental Engineering Division. EMB
Report No. 78-CKO-13. July 1980. p. 3.
9. Perry, A. C. Effective Stage Charging Through Training, Observation,
Monitoring. Technical Conference on Control of Air Emissions from
Coke Plants, Pittsburgh, Pennsylvania. April 1979.
10. Lane, J. T. and J. F. Oliver. Control of Visible Emissions at CF&I's
Coke Plant. Technical Conference on Control of Air Emissions from
Coke Plants, Pittsburgh, Pennsylvania. April 1979.
102
-------
A REVIEW OF SHED AND GAS CLEANING SYSTEMS FOR
CONTROLLING COKE PUSHING EMISSIONS FROM COKE PLANTS
By
Jack Shaughnessy and
Dilip Parikh
MikroPul Corporation
Summit, New Jersey U.S.A.
ABSTRACT
The body of information in this paper is directed to coke producers and
their management, the environmental control agencies, and labor organizations
interested in further protection of their members.
There are a number of different areas of concern in pollution control for
coke ovens. We are directing our efforts to properly control all the emissions
on the coke side of the oven.
In the United States in the past seven to ten years, there have been a
number of concepts used by the steel industry and coke producers to control
pushing emissions and door leakage on coke side of ovens. These concepts
included various types of quench cars, land based systems, hooded hot cars, and
sheds.
The purpose of this paper is to review the evolution of the shed concept
up to the present and the new concepts presently being offered. We will also
review and compare the gas cleaning devices used on the early sheds which were
either high energy scrubbers or wet electrostatic precipitators versus the
baghouse which is today's accepted technology on coke pushing emissions.
INTRODUCTION
A coke side shed is a special designed structure that is erected on the
coke delivery side of a battery of coke ovens. Its purpose is to capture all
the emissions developed on that side of the ovens from the following sources:
1. Removal of doors
2. Actual pushing
3. Spillage
4. Door leakage
103
-------
The coke side shed properly designed and installed can control all of these
sources.
Once of the first sheds was developed and installed by the Great Lakes
Carbon Corporation at their St. Louis, Missouri, coke plant in the early
1970's. Figure 1 shows a cross section of this concept which was patented by
the Great Lakes Carbon Corporation under U. S. Patent 3,844,901. This type
shed was installed at four (4) steel companies in the United States and Canada
in the middle 1970's and have been reasonably successful. This type shed is
presently being installed on eleven (11) ovens at four (4) locations at other
steel mills in the United States. The shed is a totally passive system with no
moving parts. One it is installed and working, it is just there.
On the first sheds installed in the early to middle 1970's, the gas
cleaning devices that were installed on the Great Lakes Carbon type sheds were
flat plate continuously flushed electrostatic precipitators. The reasons for
using the wet precipitator as the gas cleaning device were as follows:
1. Both the particulate and the condensable hydrocarbons were required
to be collected. The condensables could only be collected by a wet
device. The gaseous hydrocarbons that could condense at low temper-
atures in an EPA train would pass right through a dry baghouse.
2. There was concern by both the steel and coke producing industries and
the baghouse manufacturers that the hydrocarbons could cause
plugging of the fabric resulting in extremely short bag life or
inability to meet the guarantees.
3. The pilot testing of the wet. precipitator showed it could achieve
extremely low levels of outlet grain loading or high efficiency when
both the front half and back half were considered. Outlet loadings
as low as .003 grs/SCF were achieved with inlet loadings as low as
.03 grs/SCF giving an overall efficiency of 90%. These outlet
loadings were also achieved during the pushing operations. In tests
at four (4) coke plants at steel companies in the United States and
Canada, these efficiencies could be achieved with the wet electro-
static precipitator.
Like any pollution control system installed in a large process plant, the
combination MikroPul shed and wet precipitator had advantages and disadvan-
tages. Some of the main pluses for the system were:
1. A totally passive system — no moving parts.
2. Workers under the shed not exposed to elements.
3. Workers not exposed to any increase in coal tar pitch volatile
levels.
4. Door leakage positively controlled.
5. All pollution sources under the shed controlled.
104
-------
CONVENTIONAL ROOF
304 STAINLESS
STEEL
F DUCT TO REMOTE
BAGHOUSE
Figure 1. Mikropul - G.L.C. Coke Side Shed (See Figure 3 for Legend)
105
-------
6. No loss of coke production.
7. Existing hot car equipment can be used.
8. No modification to battery or track required.
Some of the main disadvantages were as follows:
1. Solve 'the air pollution problem but create a water treatment
problem. The wet precipitator required four (4) to seven (7) gallons
per 1,000 CFM to operate using a blowdown rate of about 10%. A
200,000 CFM precipitator would require pumping approximately 1,000
to 1,500 GPM in its liquid circuit and a blowdown of 100 to 150 GPM
requiring final clean-up. Water treatment was every expensive.
2. The wet precipitator, while being an excellent collecting device
with very high efficiencies, was very expensive from the capital
investment standpoint.
3. The shed allowed the large particulate to fall out onto workers under
the shed during pushing. This was large particulate not considered
inhalable but could get into shirt collars, etc., and was considered
tolerable but annoying.
4. Particulate also settled out on the heat shield shown in Figure 1 and
this had to be manually vacuumed or cleaned off every few months.
5. During periods of windy or breezy weather depending on wind direc-
tion, a wind tunnel effect could be created in the shed causing end
blowout or spillover causing a visible emission during this period
while pushing.
6. The wet precipitators suffered severe corrosion problems after
several years of operation. This was due generally to lack of atten-
tion to the operation of the water treatment system associated with
the unit. The precipitators were all constructed of mild steel with
protective coatings but pH had to be controlled to keep the units
from corroding away.
In general, the MikroPul-Great Lakes Carbon shed systems have proven very
satisfactory over a period of several years of operation. They have proven to
be very reliable, have not interfered with coke production, have been accepted
by the various pollution enforcement agencies, and have had no objections from
the unions working with them.
One major steel corporation, after trying almost every other type of
system available at least once or twice and making what they considered a
complete evaluation, has decided to install eleven (11) sheds at four (4)
different plant locations.
On the shed systems presently being installed, changes have been incor-
porated into the design to overcome most of the objections mentioned previ-
106
-------
ously. These changes are as follows:
1. Pulse-jet type baghouses or fabric filters are being used as the
particulate collection device instead of the wet precipitator. This
eliminates the water treatment and corrosion problems and reduces
the capital cost of the system considerably. I would like to point
out that only the front half particulate is being guaranteed
regarding outlet emissions from the baghouse on the sheds. In recent
years we have installed several baghouses on coke pushing emission
systems with only having to guarantee the front half. Evidently the
steel companies and coke producers, in working out their various
agreements, have decided not to include the back half or condensables
as particulate.
2. Secondary sheds or tunnels are being installed between the end of the
primary shed and the quench tower to eliminate end blowout during
windy weather.
3. Higher air volumes are being used to increase face velocities, give
better capture, and quicker evacuation of the shed.
4. Vacuum systems are being installed to ease the clean-up job on the
heat shield as shown in Figure 1.
The above changes help solve most of the problems encountered with the
first MikroPul shed and precipitator systems.
MikroPul installed and has in operation three (3) pulse-jet baghouse
systems on coke pushing operations in the United States. These systems have
been in operation since July 1979, or over two years each, and have been
working well. No bag changes to date and no substantial operating problems.
As I mentioned earlier in this paper, it was thought by both the steel
industry and the baghouse manufacturers that the soot, tars, hydrocarbons,
etc., emitted from "green" pushes would be detrimental to the operation of a
baghouse. However, the capital cost of precipitators plus the very serious
objections and cost of water treatment led MikroPul to try pulse-jet fabric
filters on this application. We had considerable experience in dry scrubbing
in the aluminum industry where hydrocarbons and light loadings of tar particles
plus other gaseous condensables were involved. By properly coating or condi-
tioning of the filter bags and continuing to do so, these elements were never
allowed to impinge on the fabric but only on the dust coat itself allowing a
pulse-jet fabric filter to be used successfully. Having over 12,000,000 CFM of
successful operating experience in the aluminum industry, this same basic pre-
coat technology was applied successfully to coke pushing operations.
The first MikroPul systems that were installed on coke pushing had a
complete pre-coat and recycle system to continuously inject a pre-coat mate-
rial to the baghouse. We have recommended a 200 mesh agriculatural limestone
be used as a pre-coat. Subsequent experience has shown that continuous pre-
coat is not necessary. The larger particulate that is captured by the pushing
emissions, that is the coke fines, fly ash, soot, etc., is sufficient to act as
107
-------
a pre-coat and no addition is necessary. MikroPul's experience in applying
pulse-jet baghouses to coke pushing applications is to apply an initial pre-
coat to the virgin fabric filter bags of 200 mesh agricultural limestone.
After this initial pre-coat, no further pre-coat is necessary. The collected
dust from the pushing operation will act as the pre-coat. However, the
cleaning of the pulse-jet collector must be carefully controlled so as to not
overclean the unit.
Particulate testing run by third party independent testing companies show
that outlet loadings of .01 or less can be achieved by the pulse-jet units.
The recommended air-to-cloth ratio is about 6:1 (6 CFM per square foot of
fabric). The fabric filter will operate at between 3" and 6" w.g. across the
fabric. Bag life to date is over two (2) years and we would expect an addi-
tional two (2) years on this type of operation.
A third generation of shed has now been developed which combines the bag-
house and shed into a single integral unit. The baghouse, which on previous
installations had to be located remote from the shed requiring considerable
runs of very large ductwork and extensive support towers and foundations, has
now been made an integral part of the shed, eliminating all the ductwork,
foundations, towers, etc. This is illustrated in Figure 2.
This concept, where applicable, would reduce the capital investment
considerably and definitely reduces the horsepower consumption drastically.
As an example, a typical pair of batteries operating under a shed would have
600 to 700 feet of ductwork inside the shed. Depending on the baghouse loca-
tion there would be an additional 200 or 300 feet of ductwork outside the shed
for an overall duct length of close to 1,000 feet. The fans sized for this type
of system would normally be set up to operate at 18" to 20" w.g. On the
MikroPul-Patton integral-type system, the total system static would be
approximately 8" w.g.
CONCLUSION
It is evident even at first glance that this new technology will
considerably reduce the capital cost of coke side sheds and also the operating
costs. We feel this is a step in the right direction.
108
-------
M MODULAR DOME ROOF.
304 STAINLESS STEEL
Figure 2. Mikropul-Patton Coke Side Shed (See Figure 3 for Legend)
109
-------
THERMAL WALL PANEL
B THERMAL HEAT DEFLECTOR
C THROAT OR STRUCTURE ORIFICE
D CAPTURE AND EXPANSION ZONE
E EVACUATION DUCT
p FILTER OR CLEANING DEVICE
G FALLOUT CAPTURE AREA
H END WALLS
I OVEN BATTERY
J COKE GUIDE
K QUENCH CAR '
L LIGHTING INSIDE OF THE SHED
M MODULAR DOME ROOF OR
CONVENTIONAL ROOF
N BENCH AREA
Figure 3.
110
-------
ARMCO 'S EXPERIENCE WITH APPLICATION OF THE BUBBLE CONCEPT
by: B. A. Steiner
Armco Ihc .
P.O. Box 600
Middle town, Ohio
ABSTRACT
Armco 's early efforts to demonstrate the advantages of the Bubble Con-
cept led to the implementation of a comprehensive fugitive dust control pro-
gram at its Middletown Works. The program was fully operational by August
of 1980 and led to an acceptable alternative control plan approved by U.S.
EPA in March of 1 981 . The prototype program has shown that significant im-
provements in ambient air concentrations of total suspended particulate are
possible with a comprehensive fugitive dust control program and that im-
provements in smaller particle size fractions are realized as well.
BACKGROUND
In 1977 following the bitter battle and public rhetoric over amendments
to the Clean Air Act, relationships between the U.S. EPA and the steel
industry reached an ebb. The opportunity to improve upon this poor relation-
ship came with the publication in December of 1977 of the Solomon Report, a
governmental study of the many economic ills of the steel industry (1).
Among those factors recognized as having negative impacts was the burden of
environmental control costs to the industry. The report stated "...it may
be possible to achieve our goal of a cleaner environment at a reduced eco-
nomic cost if there were certain changes in the regulatory process... The
EPA agrees and is willing to investigate certain areas to see if this is
possible and appropriate."
. In subsequent meetings between government and steel industry officials,
seven areas were identified in which EPA had discretionary authority to
minimize the economic burden on the industry without compromising environ-
mental goals. Among these was the principle of total plant compliance, a
concept which would allow consideration of a total plant's emissions and
would provide for the flexibility to control emissions from multiple sources
within a plant with the most cost-effective mix of control techniques. In
1978, this principle came to be known as the Bubble Concept, reflecting the
idea that an imaginary bubble could be placed over an entire plant complex
to allow consideration of the complex as a single source.
Ill
-------
In late 1977 and early 1978» Armco had begun to assemble data and in-
formation which supported the position that controlling fugitive dust
sources would be more effective and less costly than controlling process
fugitive emission sources, which were receiving the most regulatory atten-
tion. Methods of estimating emissions of fugitive dust sources, such as
traffic on paved and unpaved roads, material handling, and storage piles,
had been developed by Midwest Research Institute under contract from EPA (2).
A comprehensive emission inventory of Armco's Middletown Works revealed that
over 6
-------
6. Surface treatment of storage piles with fixed or mobile spray
equipment to minimize windblown emissions.
7. Installation of an ambient air monitoring network to gauge progress
and effects of the program.
Details of this program have been reported previously (3).
The monitoring system was made operational in August of 1979* A total
of 16 monitoring sites, 6 of which were Armco-operated, were collecting data
for all or part of the year preceding the complete implementation of the
fugitive dust control program (Figure 1). The network provided an important
data base that allowed comparison of data before and after control and that
also could be used for calibration purposes for subsequent air modeling.
Following publication of the proposed Bubble Policy, Armco waged an
extensive campaign to convince the U.S. EPA that open dust sources could not
and should not be excluded from Bubble Concept considerations if the concept
was to be useful to the steel industry. The major effort in this campaign
was, of course, the commitment to proceed with the Middletown program
described above at a cost of some $6,000,000, while having no guarantee of
success or acceptance. However, the campaign also included efforts to con-
vince employees, shareholders, journalists, legislators, other industrial
groups, the public, and environmental groups that the concept was important
to the industry and needed their support. The U.S. EPA was also kept advised
of the progress of Armco's Middletown project. The final Bubble Policy was
published in the Federal Register on December 11, 1979* It no longer ex-
cluded open dust sources, but it did set some rather limiting requirements
for making demonstrations in cases involving such sources.
The Ohio EPA had already adopted regulations providing for Bubble Con-
cept variances, and such a variance was requested by Armco in February of
1980. The request included a description of the plan, emission inventories,
a comparative analysis of the alternative plans, ambient air data, necessary
supporting information and studies, and some simplistic point source model-
ing to show the negligible impacts of controlling the three process fugitive
sources (blast furnace cast house, open hearth, and basic oxygen shop) for
which the fugitive dust controls were an alternative. The Ohio EPA asked
the U.S. EPA for assistance in evaluating Armco's plan.
In the ensuing months, but prior to full implementation of the program,
Armco submitted additional information to show preliminary and projected
improvements with the program. More extensive modeling had been done using
a CDM model and historical ambient air data for model calibration. However,
U.S. EPA insisted that acceptance of the program would have to be based on a
post-control analysis of the air data together with an historical trend
analysis and a modeling demonstration using the more recent air data for
calibration.
The Armco fugitive dust control plan was fully implemented in August of
'1980.
113
-------
N
Legend
^Temporary Sites
^Permanent Sites
ARMCO
PLANT
Figure 1. Middletown Area TSP Monitors
114
-------
In October of 1980, Armco presented the historical trend analysis and
the first two months of post-control air data. Also included were presenta-
tions of the representativeness of production and meteorology of the pre-
and post-control periods. U.S. EPA was sufficiently convinced of the success
of the demonstration that they announced tentative approval of the plan, con-
tingent upon completion of the modeling demonstration and execution of all
procedural requirements to convert the entire program into an approvable SIP
revision.
On December 23, 1980,-Armco submitted its .request for a SIP revision to
the Ohio EPA. It included four months of post-control air data. Several
supplemental submittals were made and numerous meetings were held over the
next few months to exchange and clarify data and to draft permits for all
affected sources. Proposed approval of the SIP revision was published by
U.S. EPA January 27, 1981, and final approval was published March 31> 1981.
Table 1 chronicles the events leading to this approval.
TABLE 1. CHRONOLOGY OF EVENTS
August 1977
December 1977
February 1978
May 1978
June 1978
January 1979
March 1979
August 1979
September 1979
December 1979
February 1980
August 1980
October 1980
December 1980
January 1981
February 1981
March 1981
Clean Air Act Amendments signed
Solomon Report issued
Total Plant Compliance concept born
Midwest Research Report published
Bubble Concept terminology coined
Proposed USEPA Bubble Policy Statement
Armco plan committed
Armco monitoring system operational
Ohio EPA Bubble Concept adopted
Final USEPA Bubble Policy Statement
Variance request to Ohio EPA
Armco plan operational
Tentative USEPA approval
SIP Revision request submitted
Proposed USEPA SIP approval
Ohio EPA Public Hearing
Final USEPA SIP approval
An important part of the demonstration was a modeling exercise which
was required to show that annual ambient air standards could be met under
more adverse meteorological conditions and when operating at higher produc-
tion rates than prevailed during the post-control period. The standard U.S.
EPA model CDM was used for this purpose. Key factors in the successful
application of this technique included a valid and comprehensive inventory of
point source and open dust source emissions, use of an area source grid, and
a successful calibration of the model through use of valid meteorological
data, time-period specific emissions estimates, and a sufficient number of
properly located monitors. Details of this modeling effort have been
described previously
115
-------
RESULTS
Table 2 summarizes the total suspended particulate (TSP) ambient air
data for both the year preceding the implementation of the program and the
year since for the permanent stations in the Middletown area. Reductions
are seen at all Middletovm area stations, particularly those nearby and
downwind of the plant. These reductions were accomplished at a signifi-
cantly higher plant production level and under comparable meteorological
conditions.
TABLE 2. CONTROL PROGRAM RESULTS
Station
TSF Avg. Monthly Geom. Mean (ug/m )
Before Control
(Aug 79-Jul 80)
After Control
(Aug 80-Jul 81)
Change% Change
1.
2.
3.
u.
5.
6.
7.
8.
9.
Reeds Yard
Lefferson
Oxford State
Main Gate
Coil Paint
SOS
Hook Field
Verity School
Srepco
Avg. Production
1U3
77
62
Ik
66*
54
70
61
89*
160,780
91
62
52
61
63
kl
59
5U
75
231,777
-$2
-15
-10
-13
-3
-7
-11
-7
-1U
_
-36
-19
-16
-18
-5
-13
-16
-11
-16
(T/Month)
*Data available only for Mar 80-Jul 80
Two of Armco's monitoring stations are also equipped with size
fractionating ambient air samplers (5). The SOS station is predominantly
upwind of the plant, and the Lefferson station is predominantly downwind
of the plant (but not necessarily at the point of its maximum impact).
Table 3 shows effects of the control program for various size fractions
measured at these two stations.
At the upwind station, there was relatively little change in TSP and
virtually none in the fine (<2.5 urn) or inhalable (<15 urn) fractions. How-
ever, at the downwind station, there were significant reductions in TSP and
the coarse fractions £>2.5 urn) and lesser reductions in the fine fractions.
It may be concluded from this data that fugitive dust control programs can
also have impacts on inhalable and fine fractions of particulate matter as
well as on total suspended particulate matter.
116
-------
TABLE 3. CONTROL PROGRAM EFFECTS FOR VARIOUS SIZE FRACTIONS
Station
Size Fraction*
LEFFERSON
TSP
<2.5 urn
2.5-15 urn
Total IP
>1515
-------
spray system is designed only to provide moisture to the pile surface to
prevent windblown emissions, there have been no operating problems created
by excessively wet coal. The .coal pile system is operated and maintained by
coke plant personnel.
Maintenance problems to date have been frequent, as with most mechanical
and mobile equipment, but they have not been so excessive as to cause serious
operating difficulties. When one sweeper is out of service, the other must
be operated more hours to cover the required road surfaces. However, the
spray trucks are sufficient in number and flexible enough to provide adequate
coverage when a unit is out of service.
The air monitoring network is serviced by the plant's Engineering
personnel. .This includes minor repairs, normal filter replacement, calibra-
tion, and data compilation. Analytical, quality assurance, audit, and
troubleshooting support is supplied by the plant and Corporate Research
laboratories. Corporate Environmental Engineering provided assistance for
site selection and equipment acquisition and assists in the ongoing manage-
ment and interpretation of air data.
SUMMARY AND CONCLUSIONS
Armco's efforts to demonstrate the benefits of the Bubble Concept as
applied to iron and steel plant fugitive dust sources have spanned several
years, including that period of development of the U.S. EPA's policy on the
concept. The•prototype program at Armco's Middletown Works has shown that
significant improvements in ambient air quality are possible with a compre-
hensive fugitive dust control program. The improvements can be obtained
with less capital and operating costs, with less energy and in less time
than process fugitive emission controls. Moreover, improvements are not
limited solely to total suspended particulates but also include reductions
at smaller particulate size fractions.
118
-------
HHiH'HiHHTWfi'l'IS
1. A.M. Solomon, "Report to the President-A Comprehensive Program for the
Steel Industry," December 6, 1977.
2. R. Bohn, T. Cuscino and C. Cowherd, "Fugitive Emissions from Integrated
Iron and Steel Plants," U.S. Environmental Protection Agency Publication
No. EPA-600/2-78-050, Research Triangle Park, North Carolina, March,
1978.
3- B. A. Steiner and S. L. Francis, "A Steel Works Fugitive Dust Control
Program," Paper No. 80-17*3» Air Pollution Control Association Annual
Meeting, Montreal, Quebec, Canada, June 22-27, 1980.
k> J. A. Grantz, "Use of CDM for an Integrated Iron and Steelmaking
Complex Including Open Dust Sources," APCA Specialty Conference on
Dispersion Modeling of Complex Sources, St. Louis, Missouri, April 7-9>
1981.
£. J. A. Grantz, "Inhalable Participates in the Vicinity of an Integrated
Iron and Steelmaking Complex," Paper No. 81-5-U, Air Pollution Control
Association Annual Meeting, Philadelphia, Pennsylvania, June 21-26,
1981.
119
-------
ENGINEERING STUDY OF ROOF MOUNTED ELECTROSTATIC PRECIPITATOR (REP) FOR
FUGITIVE EMISSION CONTROL ON TWO BASIC OXYGEN FURNACES OF 300 TON CAPACITY
by: Richard Jablin
Richard Jablin & Associates
2511 Woodrow Street
Durham, North Carolina 27705
(919) 286-4693
David W. Coy (7-3)
Research Triangle Institute
Research Triangle Park, North Carolina 27709
(919) 541-6940
ABSTRACT
A number of alternatives are available for controlling fugitive dust
emissions from basic oxygen furnace operations. Use of local hoods and
partial building evacuation are the common means for capturing emissions.
An alternative not used in the United States,, but successfully used in
Japan, is roof-mounted electrostatic precipitators (REP). A study examining
the feasibility of REP installation has been performed for a BOF shop with
two 225 metric ton (280 ton) vessels. The purposes of the study were to
determine the applicability of the devices to the BOF fugitive dust sources,
examine the changes needed to existing plant facilities to interface the
new equipment with the old, estimate the costs for modification of existing
plant facilities and the addition of new facilities, and examine the
expected performance of the proposed REPs.
In performing the study, the following steps were taken:
1. A preliminary quote for the REP was obtained from Sumitomo Heavy
Industries, Ltd.
2. The requirements for electrical power supply and spray washing
system were investigated.
3. Structural reinforcement to the existing BOF building was
investigated.
4. An estimate of cost to install and to operate the REP was
generated.
5. The quantity of fugitive furnace emissions escaping the building
with the REPs in place was estimated.
The results of this study are presented in this paper.
120
-------
1.0 INTRODUCTION
The conventional technique for controlling secondary emissions from
existing BOF facilities is to provide local hoods at each source. This
paper examines the alternative of installing roof-mounted electrostatic
precipitators (REP). It presents a study which examined the feasibility of
such an REP installation for an existing plant. The purposes of the study
were to determine the applicability of the devices to the BOF fugitive dust
sources, examine the changes needed to existing plant facilities to inter-
face the new equipment with the old, estimate the costs for modification of
existing plant facilities and the addition of new facilities, and examine
the expected performance of the proposed REPs.
In performing the study, the following steps were taken:
1. Obtain a preliminary quote for the REP from Sumitomo Heavy
Industries, Ltd.
2. Investigate requirements for electrical power supply and spray
washing system.
3. Investigate structural reinforcement to the existing BOF building.
4. Generate an estimate of cost to install and to operate the REP on
the roof of an existing BOF building.
5. Estimate the quantity of fugitive furnace emissions escaping the
building with the REPs in place.
The results of this study are presented in subsequent sections of this
paper.
2.0 SUMMARY
One method of controlling secondary emissions from the operation of an
existing BOF is to provide roof-mounted electrostatic precipitators (REP).
This method of controlling secondary emissions has been successfully applied
in Japan.
The REP is of the type which is manufactured by Sumitomo Heavy
Industries, Ltd. For a typical facility which contains two BOF furnaces of
273 metric ton (300 ton) capacity each, the estimated cost, delivered and
erected on site is $3,020,000 or 60 percent of the estimated project cost.
The remaining costs are adsorbed primarily by structural reinforcement and
modification of the existing BOF building. The total estimated project cost
is $5,010,000. Estimated annual operating costs are $679,000. All costs
are escalated to the third quarter of 1982.
Electrical power for the REP is supplied from the existing 4.16 KV
switchgear and amounts to 189 KW. An additional 40 KW is required for the
motor which drives the pump in the spray washing system. Power for the
motor is obtained from the control room on the scrubber platform.
121
-------
The removal of dust from the REP is accomplished by automatic washing
of the collector plates. Clean water for washing is pumped from the
existing separator cooling towers of the primary gas scrubbing system.
Contaminated wash water drains by gravity to the existing scrubber water
system, which is assumed to have the capacity of absorbing the moderate
additional hydraulic loading.
The existing EOF building will require extensive bracing and reinforce-
ment to sustain the loads imposed by the precipitators, primarily against
wind loads at right angles to the building aisles. A computer analysis was
performed on the existing EOF building. At crosswind speeds of 100 mph,
drift of the structure may be a problem which requires further structural
analysis.
The REP specifications provided by Sumitomo Heavy Industries offer a
design efficiency of 91.5 percent. The critical phase of furnace operation
with respect to REP performance is hot metal charging. Using emission
factors and several assumptions, the inlet concentration during hot metal
charging is estimated as 0.96 gram/acm (0.42 grain/acf) giving an estimated
outlet concentration of 0.082 gram/acm (0.036 grain/acf) where only one
precipitator receives the fumes. The range of concentrations estimated to
produce 20 percent opacity at the REP discharge is 0.112 to 0.222 gram/acm
(0.049 to 0.097 grain/acf). Therefore, it appears the REP installation is
capable of achieving discharges of less than 20 percent opacity during hot
metal charging. A number of assumptions were necessary to make these
performance estimates. It is important to verify the validity of some of
these assumptions prior to proceeding with an installation.
3.0 APPLICABILITY OF REPs TO STEEL FURNACE FUGITIVE EMISSIONS
The applicability of roof-mounted electrostatic precipitators to
fugitive steel furnace emissions has been demonstrated, in general, by
steel plant operators in Japan. Sumitomo Heavy Industries, Ltd. has many
operating REP installations, some in the following plants on the indicated
process sources:
1. Kawasaki Steel Corporation, Chiba Works, No. 3 Q-BOP, 2 vessels,
230 metric tons
2. Kawasaki Steel Corporation, Mizushima Works, EOF, 3 vessels, 250
metric tons
3. Sumitomo Metal Industries, Wakayama Works, BOF, 3 vessels, 160
metric tons
4. Kobe Steel, Kakogawa Works, Blast furnace cast house
REPs are also manufactured by Sumitomo Metal Mining Company. They have
several operating REP installations with one at Kishiwada Steel Company on
a 43 metric ton electric arc furnace shop.
122
-------
The process furnaces at both Kawasaki Steel plants are closed hood
design. At Chiba Works they are OG design and at Mizushima Works the
Irsid-Cafl system is in use. The Chiba Works was constructed with a full
furnace enclosure while the Mizushima Works has a partial enclosure i.e.,
tapping side and top enclosed with charging side open.
In the case of Mizushima Works, the REPs were retrofit to improve
furnace fugitive emission control. The REPs supplement partial furnace
fugitive controls (charging1 and tapping hoods in the furnace enclosures)
that apparently were insufficient to meet local pollution control
regulations.
The Chiba Works REPs were constructed as part of the initial plant
design. These REPs also supplement local fume capture in the enclosure.
Charging hoods alone apparently were believed insufficient for the more
difficult fugitive fume capture from bottom blown furnaces. Very limited
observations of the REPs at both Kawasaki plants indicated no emissions in
excess of 10 percent opacity (individual readings) during a furnace cycle
of operation.1
At Kishiwada Steel the primary electric arc furnace emissions are
captured by a direct shell evacuation system. Control of charging and
tapping emissions is achieved entirely by REPs. Some limited visible
emissions data (Method 9) taken at this plant are available in USEPA Region
III files.
4.0 PRELIMINARY QUOTATION FOR THE REP
Preliminary quotations were requested from both Sumitomo companies,
however, only one responded in sufficient detail to satisfy the study
needs. Sumitomo Heavy Industries, Ltd. supplied a quotation for a REP
installation to serve a BOF facility which meets the following criteria:
Heat size 260 metric tons (285 tons)
273 metric tons maximum (300 tons)
Hot metal charge 200 metric tons (220 tons)
Production 30 to 33 heats per day or
8,500 metric tons (9,400 tons) per
day
Blow rate, oxygen 570 to 600 acmm (20,000 to 21,000
scfm) normal
650 acmm (23,000 scfm) maximum
Number of furnaces 2
Field measurements at such a facility indicated that the plume from
charging and tapping the furnace rose at the rate of approximately 2.7
meters/ second (9 ft/sec).
123
-------
At the present time, many EOF facilities operate one furnace at a
time. It is, however, possible to increase steel production by relatively
minor modifications which would permit two furnace operation. For this
reason, it was decided to arrange the REP so that it could accommodate the
simultaneous operation of both furnaces. The cost of such accommodation,
if provided in the initial design, was nominal and therefore desirable.
Sumitomo Heavy Industries furnished the specification for the REP. It
includes two precipitator units that have the capability of handling a
total volume 27,000 acmm (953,000 cfm). Included with the units is a set
of electrical equipment including high voltage power supplies (45 KV) and
water sprays for washing the plates. Spraying takes place automatically
for 10 minutes per section after taking sections of the REP out of service.
The specification for the REP is similar to the units which were
supplied by Sumitomo to the Chiba Works of the Kawasaki Steel Corporation.
Included with the specification was a general arrangement drawing (Figure
1) that shows the REP in position on the roof of the assumed EOF facility.
5.0 ELECTRICAL POWER SUPPLY
The continuous electrical consumption for the REP, based on both
precipitators operating simultaneously, is as follows:
1. High voltage supply 170 KW - 310 KVA
2. Insulating fans 14 KW - 18 KVA
3. Area and indoor lighting 5 KW - 5 KVA
189 KW - 312 KVA
The power for the REP is proposed to be connected in the existing
plant to the spare housing of the existing 4.16 KV, metal-clad switchgear
assembly. The existing 4.16 KV switchgear obtains its power through a
5,000 KVA transformer. Since the present load on this transformer has a
reactive power factor of 0.95 the present inductive reactance would be
essentially balanced by the capacitive reactance of the new load due to the
REP, thereby bringing the resultant power factor to approximate unity.
Because of these considerations, it was decided that the present 4.16 KV
switchgear and transformer would accept the new electrical load from the
REP without causing overload conditions.
The control room of the REP is of sheet metal construction with a
concrete floor. It is located indoors, in the furnace aisle, at elevation
254 meters (835 feet), immediately above the storage bins for furnace
additives. The room is provided with ventilation, but without air
filtration equipment.
The new electrical equipment in the control room includes a 4.16
KV/480V, 500 KVA silicon transformer, low voltage switchgear and
precipitator controls. The precipitator controls are furnished with the
124
-------
ro
en
Figure 1. Proposed installation of roof mounted precipitator on EOF shop.
-------
REP and include control units for operating the water valves for the
automatic spray washing of the,collector plates in the precipitator.
The only other electrical load imposed by the REP installation comes
from the new pumps for the spray washing system. There are two pumps, one
operating and one stand-by, each being driven by a 50 HP, 440V, 3 phase
motor. The pumps are located on the existing scrubber platform. Power for
the pumps is supplied from the existing electrical units in the control
room which is also on the scrubber platform.
The existing 4.16 KV substation is located on the ground floor,
immediately adjacent to the EOF building. The electrical connection between
this substation and the new control room for the REP is provided by means
of 4 cables, 5 KV, EPR insulated, installed in a 12.7 centimeters (5 inch)
diameter galvanized conduit.
In the new control room, cable duct is proposed for connecting the
switchgear to the control units which are supplied with the REP. There is
electrical and key interlocking between the new 1,200 amp circuit breaker
in the existing cubicle No. 7 and the new 600 amp load-break disconnect
switch at the primary of the new 500 KVA transformer. The control for the
new pumps has push buttons in the control center and at the motors with
lock-out switches at the motors.
The lighting for the new control room and for the operating areas of
the precipitators will be supplied directly from the power bus of the low
voltage switchgear. The voltage of the lighting will be 277 VAC. Lighting
contactors are provided in the switchgear and low voltage push buttons in
appropriate locations. Indoor lighting will be low-bay mercury vapor and
outdoor lighting will be sodium vapor units.
6.0 SPRAY WASHING SYSTEM
The continuous water usage in the spray washing system for the REP,
based on both, precipitators operating simultaneously is as follows:
1. For sprays 1.2 m3/min (317 gpm) continuous
2. For sprinklers .3 m3/min ( 79 gpm) 180 min/day
Total 1.5 m3/min (396 gpm)
The supply water to the precipitators is required to meet the following
criteria:
Suspended solids 50 PPM, max.
Particle size 20 mesh, max.
PH 7-8
SO 4" 100 PPM, max.
126
-------
Cl 100 PPM, max.
Pressure at elevation 4 kg/cm2G (131 ft.)
258 meters (847')
The water for spray washing and sprinklers is provided from an existing
piping header in the present scrubber water system. Since the EOF has a
scrubber on its primary gas cleaning system, the water quality from the
clarifier is assumed to be-generally adequate for the new service.
In the existing system, the flow rate of the water is 18.2 m3/min
(4,800 gpm). This is the design flow rate when both primary gas cleaning
systems are in operation, corresponding to the simultaneous operation of
both furnaces. The additional flow for the new REP system is 1.5 m3/min
(396 gpm) which is an increase of 8.3 percent over the design range for the
two-furnace operation.
Thus for two-furnace operation, there is a moderate increase in flow
which should be easily accommodated in the existing water system.
The spray water in the new REP is used for washing dust from the
collecting plates of the precipitators. After receiving the dust from the
washing operation, the water flows to the existing callow cone of the
present scrubber system. The following calculations indicate the effect on
the present system which results from the new contaminated water:
1. Amount of dust collected in the REP
0.35 kg/metric ton (0.7 Ib/ton) of dust from charging and
tapping x 8,500 metric tons/day (9,400 tpd) of steel produced =
2,970 kg (6,580 Ibs) per day. (This calculation is based on the
emission factor from reference 2.)
At a guaranteed collection efficiency of "91.5 percent, the
amount of dust transferred to the water is:
.915 x 2,970 kg/day = 2,720 kg/day (6,020 Ibs/day)
2. Quality of spray water after washing (one furnace operating)
2.720 kg/day 1 x 10e + 50 PPM
1,440 min/day 1.5 m3/min x 1,000 kg/m3
= 1,310 PPM of particulates
3. Flow increase to existing water system
The design flow to the callow cone for two-furnace operation
is 10.0 m3/min (2,650 gpm). For one-furnace operation, the flow
is reduced to 5.6 m3/min (1,470 gpm). Considering the new added
flow of 1.5 m3/min (396 gpm) from the REP, under two-furnace
127
-------
operation, the flow will be 15 percent above design and there
will be an addition of 330 PPM to the suspended solids in the
inlet water. Under one-furnace operation, the flow will be 30
percent less than design and there will be an addition of 270 PPM
to the suspended solids.
Under one-furnace operation, the scrubber water system is
able to accept the new added flow from the REP without exceeding
the design parameters of the system. Under two-furnace operation,
if such takes place in the future, the system may require
modification, primarily to the callow cone, unless there is
sufficient excess capacity in that unit to accept the small over-
load which would then occur.
The addition of water from the REP to the existing scrubber
water system imposes a moderate increase in water flow and in the
quantity of suspended solids. Under one-furnace operation, there
is enough reserve capacity in the existing system to accept the
additional loads. If two-furnace operation takes place in the
future, minor additions to the existing scrubber water system may
be required at that time.
The water supply and return system for the REP is shown in
Figure 2. It has two-pumps, one operating and one stand-by, each
rated 1.1 m3/min (300 gpm) at 100 meters (330 feet) tdh. Because
of the relatively low flow and high head, it is necessary to
operate the pumps at approximately 2,600 rpm and to drive them
through V-belting. The pumps are located in the clean water
supply system, the return of contaminated water being by gravity
flow. In order to insure long life under continuous service and
high rotating speed, it is recommended that the pumps be heavy-
duty, Refrax lined.
The water supply system includes motor-operated valves to
isolate the flow to each precipitator. These valves are
controlled from a remote location in the EOF building, thereby
facilitating individual operation of the REPs if desired.
7.0 STRUCTURAL REINFORCEMENT OF EOF BUILDING
The most critical consideration in the proposal to install two roof-
mounted precipitators on the existing EOF building is the ability of the
building to safely carry the new loads.
Figures 3, and 4 show some of the structural modifications and
additions which are required to sustain the new loads. Several conditions
of particular importance are as follows:
1. The load of the precipitators is carried by new steelwork into
the columns on either side of the furnace aisle, per Figure 3.
128
-------
ISi
vo
LOW CONE
NEW PUMPS FROM EXIST.
VENTURI SUPPLY MAIN
EXIST. SCRUBBER PLATFORM
Figure 2. Water supply and return flow diagram.
-------
ELEV. 200'
TO GRADE
NEW ELECTROSTATIC
PRECIPITATOR
NEW SUPPORTING STEEL
i_50K
NEW' SIDING
SIDING IN ROOF TRUSS
TO CHANNEL FUMES
Figure 3. Structural reinforcement of existing EOF building.
130
-------
co
NEW ROOF-MOUNTED ELECTROSTATIC
PRECIPITATORS
27,000 M3/MIN. (954,000CFM)
91.5% EFFICIENCY (
170 KW, 45 KV.
NEW SHEET METAL EACH
SIDE OF BOF
CHARGING EMISSIONS
EXISTING BOF PRIMARY
FUME DUCT
REMOVE EXISTING ROOFING
NEW SHEET METAL PARTITION
TAPPING EMISSIONS
Figure 4. Cross section of BOF.
-------
Although each precipitator spans 3 columns in the north-south
direction, the stringers under the precipitators are arranged to
carry the loads into the end columns only. The center columns do
not run to the ground but span the furnaces. It was decided,
therefore, not to impose any additional loads on these center
columns.
2. The dead load of each precipitator is 138.6 metric tons, thereby
adding a new load to the building of 277.2 metric tons. This
load may be accommodated by the existing steelwork without serious
problems.
3. The most serious structural considerations are the horizontal
loads which result from a design wind loading of 161 km/hour (100
mph). The horizontal loads in the north-south direction require
the provision of new X bracing between column lines Nos. 10 and
12. The bracing chosen for the estimate is somewhat overdesigned
and further detail analysis may result in reduction of the
sections. The new structure immediately under the precipitators
has horizontal and vertical wind bracing in each direction.
4. The horizontal wind loads in the east-west direction are of
greatest concern, particularly in the existing steelwork above
elevation 228 meters (749 feet). Below that elevation the columns
are of heavy section and there is substantial wind bracing.
Above that elevation, the columns are drastically reduced in size
and there is relatively little wind bracing.
It should be noted that the EOF building is unsymmetrical in cross-
section. Because of this condition, because the wind-loads on the new
precipitators are high, and because the supporting columns have unusual
configuration and size, computer "stress" runs were made using four loading
cases and assuming fixed and pinned bottom connections:
1. Existing structure -- wind to the east
2. Existing structure — wind to the west
3. Structure with precipitators -- wind to the east
4. Structure with precipitators -- wind to the west
The analysis showed that the structure below elevation 228 meters (749
feet) was adequate; however, there appeared to be excessive lateral drift
at the roof line of the furnace aisle.
Because of the computer analysis, there is serious concern about the
ability of the structure to accommodate the new wind loads in the east-west
direction. This concern is somewhat tempered by the fact that the BOF
building has extensive horizontal bracing which unifies the structure and
transfers load from one column to the next. The computer analysis did not
allow for this distribution of wind load on the precipitator.
132
-------
For the purpose of the preliminary design and cost estimate, it was
decided to reinforce the columns under the precipitators down to elevation
228 meters (749 feet). Also K-bracing is proposed above elevation 255
meters (835 feet) in order to increase lateral stiffness. It is anticipated
that these measures will provide adequate reinforcement for the existing
building. Nevertheless, further investigation and analysis will be required
to definitely prove this contention. In making the analysis it will be
necessary to re-examine certain critical structural connections to insure
that they are adequate to meet the new loadings.
Figures 3 and 4 show modifications to the sheet metal siding. The
purpose of the modifications is to direct the secondary emissions from the
furnace into the bottom of the new precipitators. The modifications include
removal of roofing in the way of the precipitators and the provision of new
siding in three general areas. These are:
1. From the roof of the charging aisle to the east side of the
precipitator.
2. In the furnace aisle on the west side of the lance crane.
3. In the roof trusses of the charging aisle at column lines Nos. 7,
and 15.
8.0 ESTIMATE OF COSTS
The estimates of costs for installing and operating the REP are given
in Tables 1 through 5. In all cases, the costs are based on the third
quarter of 1982. Costs for this period were obtained by first preparing
estimates for the third quarter of 1980 and then applying escalation
factors. The factors are assumed to be 15 percent per year resulting in a
two-year value of 32 percent.
The estimates for electrical, process piping and structural
reinforcement are given in Tables 1 through 3. In preparing them, the
costs were calculated using the technical information that appears on the
appropriate drawings. Each estimate includes a contingency factor of 15
percent.
The estimate for providing and erecting the precipitators is derived
from cost data which is included in the specification by Sumitomo Heavy
Industries. These costs were developed for the Chiba Works of the Kawasaki
Steel Corporation, an REP facility which is very similar to the one under
consideration. The REP costs that appear in the cost summary, Table 4,
include escalation and also a 10 percent contingency factor.
In order to erect the precipitator and its supporting steelwork,
access to the existing EOF building is proposed from the east side of the
building. Preliminary investigations indicate that there is room at this
location to position a crane for direct lifting and mounting of fabrications
.and components. There will be some interference with railroad movements on
133
-------
TABLE 1. ELECTRICAL INSTALLATION COSTS FOR ROOF MOUNTED
ELECTROSTATIC PRECIPITATOR
1. Air circuit breaker, including controls
2. Equipment in electrical equipment room
3. Electrical cable
4. Conduit and fittings
5. Starters and controls
6. Lighting
7. Labor to install switchgear, items 1 and 2
8. Labor to install conduit, cable and controls
9. Engineering
10. Contingency, 10 percent
TOTAL
$ 22,000
67,000
10,500
16,000
3,500
9,000
67,000
90,000
5,000
30,000
$320.000
TABLE 2. PROCESS PIPING FOR ROOF MOUNTED ELECTROSTATIC PRECIPITATOR
1. 2 Refrax-Lined pumps, 200 GPM @ 300 Ft. TDK $31,700
2. 6" Main piping 15,000
3. 4" Branch piping 9,800
4. Foundation & Installation of Pumps 5,000
5. Crane Service 8,500
6. Spares 2 percent 1,400
7. Engineering 10 percent 7,000
8. Contingency 15 percent 11,600
TOTAL
TABLE 3. STRUCTURAL REINFORCEMENT COSTS FOR ROOF MOUNTED
ELECTROSTATIC PRECIPITATOR
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Precipitator support framing, 70 tons
Column reinforcing, 150 tons
Connection reinforcing
Wind bracing in BOF building, 40 tons
Miscellaneous framing, 20 tons
Purlins, 35 tons
Siding, 50,000 sq. ft.
Electrical control room
Remove sign
Remove roofing
Crane rental
Contingency, 15 percent
Engineering
TOTAL
$ 185,000
495,000
150,000
90,000
46,000
70,000
116,000
33,000
35,000
20,000
100,000
200,000
80,000
$1,620,000
134
-------
TABLE 4. INSTALLATION OF REP—SUMMARY OF CAPITAL COSTS
1. Roof mounted electrostatic precipitator materials, $2,090,000
fabrication and engineering2 3
2. Shipping from Japan 40,000
3. Field erection of REP 890,000
4. Electric power supply 320,000
5. Process piping 90,000
6. Structural reinforcement of EOF building 1,620,000
and local hooding
TOTAL INSTALLED COST
Notes:
1. Costs given for the 4th quarter 1982. Calculated using January
1981 costs and adding 32 percent for inflation.
2. Cost based on information from Sumitomo Heavy Industries, Ltd.
3. Costs include spare parts.
TABLE 5. INSTALLATION OF REP—OPERATING COSTS
1.
2.
3.
4.
5.
6.
Maintenance, 2 percent of installed cost
Operators, 1 man per shift @ $40,000*
Electrical power at $0.05 per KWH
Service water at $.25 per 1000 gal.
Depreciation, 20 year straight line
Insurance, 1 percent of investment
TOTAL ANNUAL OPERATING COST
$100,000
160,000
83,000
36,000
250,000
50,000
$679.000
*Labor cost includes 25 percent fringe benefits, escalated to 3rd quarter
1982.
the tracks at ground level, but this was not deemed to be an insurmountable
problem.
In general the costs were calculated using conservative methods of
costing as well as conservative values for escalation and for contingencies.
Thus, it is reasonable to utilize the cost values of $5,050,000 for
installation and $679,000 for annual operation.
9.0 EXPECTED PERFORMANCE OF REPs
With respect to performance of the REPs, the critical furnace operation
is expected to be hot metal charging. Depending on scrap quality and its
impurities, hot metal charging is expected to provide the most severe test
of REP performance. Tapping typically generates less intensive emissions
spread over a longer time than charging, but can be as high as charging on
occasion. Emissions during scrap charge, turndown (for sampling and
135
-------
temperature measurement), puffing (during blowing), and vessel deslagging
are relatively low. Provided these latter emissions are directed (by means
of partitions and baffles) to the REPs, they should easily be collected.
The REP specifications received from Sumitomo Heavy Industries provides
a guarantee of 91.5 percent efficiency when the inlet concentration exceeds
0.4 gram/Nm3 dry (0.175 grain/scf dry). At or below the design inlet
concentration of 0.4 gram/Nm3 the guaranteed outlet concentration is not to
exceed 0.034 gram/Nm3 dry (0.015 grain/scf dry).
Relating the above performance guarantee to the needs of the BOF being
considered is complicated by the fact that regulations and consent decrees
applying to BOF fugitive emissions are written in terms of visible emissions
from the roof monitor. A theoretical basis does exist for relating emission
concentration to plume opacity and that relation will be used in this
examination of expected performance. However, it is important to keep in
mind the conversion of concentrations to visible emissions is not a well
defined science.
To estimate the performance of the REPs as applied to this plant it is
necessary to make several assumptions. Should a decision to proceed with a
particular REP installation be made, some additional design data to verify
or alter these assumptions will have to be obtained.
1) No measurements of uncontrolled furnace fugitive emission rates
were available for this study; therefore, typical emission factors
for charging were assumed.
2) No measurements of gas flow rate discharged from the roof monitor
were available for this study. For the purposes of these
calculations, the gas rate specified for the precipitator design
(based on Japanese plant experience) was assumed to be the rate
of gas discharge during hot metal charging operations.
3) The upper building (within the partition walls) will act as a
reservoir for the sudden fume emissions from hot metal charging.
It is desirable to maintain a minimum time for hot metal charging,
assumed to be two minutes. When coupled with the reservoir
effect, it is assumed the hot metal charging emissions will pass
through the precipitators over a three-minute period.
4) It is assumed that all the hot metal charging emissions from
charging one furnace will pass through one precipitator. (As
will be shown later it is desirable to divert, actively or
passively, some portion to the other precipitator to reduce the
dust load on a single precipitator.)
Using the above assumptions, the following calculations 2stimate the
performance of the REP installation during hot metal charging (anticipated
worst case). The emission factor for hot metal charging is 0.2 kg/metric
ton (0.4 Ib/ton) of hot metal poured.2 For a hot metal charge of 195 kg
(215 tons) the particulate charging emissions are given as:
136
-------
195 tonnes (0.2 kg/tonne) = 39 kg
The concentration of emissions entering the single REP is given as:
39 kg/chg (1,000 grams/kg) . n, . ,. . . . . ...
,„ cXr» — a - • / u -\ = 0.96 gram/acm (0.42 grain/acf)
13,500 acms (3 min/chg) 6 6
Assuming an efficiency of 91.5 percent gives an average outlet concentration
during the hot metal charge of:
(1.0 - .915) (0.92) = 0.082 gram/acm (0.036 grain/acf)
The concentration at which the plume opacity will equal 20 percent is
estimated as follows:3
-pK
where:
W = concentration (grams/m3)
L = light path length through plume (meters)
p = density of particles (grams/cm3)
K = specific particulate volume/extinction coefficient ratio
(cm3/m2)
I/I = light transmittance
For this case the variables having the following values:
L =10 meters (distance across top of REP west to east)
p =5 grams/cm3 (iron oxide)
K = 1 to 2 cm3/m2 (iron oxide, mass mean radius of 5 microns,
geometric standard deviation of 3 to 4)
I/I =0.8 transmittance for 20 percent opacity
Therefore,
„ _ -5 (1 to 2) In (0.8)
W ~ 10
= 0.112 to 0.223 gram/acm
= 0.049 to 0.097 grain/acf
137
-------
The estimated outlet concentration based on 91.5 percent efficiency is
0.082 gram/acm (0.036 grain/acf). On a theoretical basis the REP
installation appears to be able to achieve a discharge opacity less than 20
percent during hot metal charging. The cautions with respect to the
estimating methodology must he recalled at this point. However, the hot
metal charge represents the most difficult collection problem, and the
other portions of the furnace cycle should be much better in terms of
visible emissions.
An additional essential point of discussion with respect to performance
is that Sumitomo Heavy Industries recommends the use of a local secondary
hood (probably a charging hood) to reduce the total load of particulate
going to the REP during charging, if it is desired to maintain 0.034
gram/Nm3 (0.015 grain/scf). The two EOF installations of Kawasaki Steel do
have this feature. The Japanese also state that lack of a local secondary
hood will cause dirtier in-plant conditions than in their plants due to
dust fallout. However, one would not expect the REPs to substantially
alter the in-plant conditions as compared to present operations, which are
assumed to have no secondary control. Should lower discharge concentrations
be desired during hot metal charging two alternatives are possible. As was
suggested previously in this section, provisions can be made to divert a
portion of the charge emissions to the adjacent REP by means of partitions
and baffling. A fan assist might even be tried. The second alternative is
to provide larger REPs with higher efficiency than those in the
specifications. This latter alternative is consistent with the Japanese
recommendation of larger REPs if no local secondary hoods are provided.
10. ACKNOWLEDGEMENT
This report was prepared under contract funds provided by U.S.
Environmental Protection Agency Contract No. 68-02-2651 entitled "Evaluation
and Applicability of Foreign Steel Industry Pollution Control Technology."
Mr. Robert C. McCrillis of U.S. Environmental Protection Agency, Industrial
Environmental Research Laboratory, Research Triangle Park, North Carolina
27711, is the project officer.
11. REFERENCES
1. Trip Report. Kawasaki Steel Corporation. Chiba Work, Mizushima Works.
Research Triangle Institute. September 25, 26, 1979.
2. Cuscino, T. A., Jr. Particulate Emission Factors Applicable to the
Iron and Steel Industry. Midwest Research Institute, Kansas City,
Missouri. EPA-450/4-79-028. September 1979. p. 27-31.
3. Ensor, D. S., and M. J. Pilat. Calculation of Smoke Plume Opacity
from Particulate Air Pollutant Properties. Journal of the Air
Pollution Control'Association. Volume 21, Number 8. August 1971. pp.
496-501.
138
-------
DEMONSTRATION OF THE USE OF CHARGED FOG
IN CONTROLLING FUGITIVE DUST FROM A
COKE SCREENING OPERATION AT A STEEL MILL
by: Edward T. Brookman and Kevin J. Kelley
TRC-Environmental Consultants, Inc.
Wethersfield, Connecticut 06109
Robert C. McCrillis
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
ABSTRACT
TRC-Environmental Consultants, Inc. (TRC) has been contracted by the
Industrial Environmental Research Laboratory of the Environmental Protection
Agency at Research Triangle Park, North, Carolina (EPA/IERL-RTP), to test a
commercially available electrostatically charged fogger on several
large-scale industrial sources within the iron and steel and sand and gravel
industries. This paper discusses tests conducted at a coke screening
operation at a steel mill. Tests were run with no fog, uncharged fog,
negatively charged fog, and positively charged fog. Data analysis indicates
a doubling in total suspended particulate control efficiency when a
positively charged fog was used relative to uncharged fog. For the same
case, removal efficiency for particles less than 16 urn improved 2.5 times.
INTRODUCTION
Although the charged fog concept of dust control has been applied to
industrial sources of fugitive dust, little data are available on fogger
control efficiency. To obtain such data, EPA/IERL-RTP contracted TRC to
conduct a full-scale demonstration of a charged fogger on several industrial
fugitive emission sources. In particular, EPA/IERL-RTP was interested in
testing the largest fogger, "Fogger IV," manufactured by the Ritten Corpor-
ation,* on several sources within the iron and steel and sand and gravel
industries.
* The Ritten Corporation, 40 Rittenhouse Place, Ardmore, PA 19003
139
-------
To date, four sources have been tested. The results of the testing at
the first three sources (primary rock crusher, secondary rock crusher, and
cast house molten iron spout hole) have been reported in previous papers*
and will not be discussed herein. The intent of this paper is to present
and discuss the results of the fourth fogger field test which was conducted
at a coke screening operation located at Stelco's Hilton Works in Hamilton,
Ontario, Canada. The field test was performed during the period of May 1-7,
1981, with a total of 51 test runs conducted. Detailed results of all four
fogger field tests are contained in the draft Interim Report.**
DESCRIPTIONS OF SITE AND PROCESS
As part of the overall steelmaking process, coal is converted to coke in
order to obtain a fuel which can be used in a blast furnace to provide the
high temperature reducing atmosphere necessary to smelt the iron out of the
ore. To accomplish this, coal is placed into large ovens and heated to
drive off volatiles. The resulting product, known as coke, is then removed
from the ovens and transferred via railcar to the next step of the process.
One of the subsequent steps in the process is to segregate the still
warm coke into two different size categories. The coke is transferred from
a conveyor belt onto an inclined vibrating screen. Pieces of coke that are
larger than the pore size of the screen travel down its face and are deposi-
ted into a hopper at its end. Pieces of coke that are smaller than the pore
size pass through the screen into a different hopper. Conveyor belts then
transport the separated materials to the next steps in the process. At
Stelco's Hilton Works, coke arrives at the screen in runs which generally
last 2 to 6 minutes. The runs are usually separated by 3 to 10 minutes.
The discharge end of the conveyor belt, the shaker screen, and the hop-
per inlets are all located in one room. The screening operation takes place
on two different levels within this room. The conveyor belt and top of the
screen are on the upper level. The hoppers and bottom of the screen are on
the lower level. A catwalk runs around the perimeter of the screen on the
upper level. Figure 1 is a sketch of the room which illustrates these fea-
tures. Figures 2 and 3 are top and side views, respectively, that provide
dimensions of the important features.
* Brookman, E., Demonstration of the Use of Charged Fog in Controlling
Fugitive Dust from Large-Scale Industrial Sources. Presented at the
Symposium on Iron and Steel Pollution Abatement Technology for 1980,
Philadelphia, PA, November 1980.
Brookman, E., et al., Demonstration of the Use of Charged Fog in Control-
ling Fugitive Dust from Large-Scale Industrial Sources. Presented at the
Third Symposium on the Transfer and Utilization of Particulate Control
Technology, Orlando, FL, March 1981.
** Brookman, E.T. and K.J. Kelley, Demonstration of the Use of Charged Fog
in Controlling Fugitive Dust from Large-Scale Industrial Sources;
Interim Source Test Report (Draft), EPA Contract 68-02-3115, TD 109,
TRC-Environmental Consultants, Inc. Undated.
140
-------
INPUT
CONVEYOR
OVERSIZE
HOPPER
UNDERSIZE
HOPPER
Figure 1. Coke screening operation.
-------
[RUBBER
if FLAP i!
Figure 2. Top view of coke screen operation.
-------
CONVEYOR
o
O
O
3.0m
t—CATWALK (UPPER
LEVEL
UNDERSIZE
HOPPER
LOWER
LEVEL
-6.7m-
Figure 3. Side view of coke screen operation.
143
-------
While the coke is being screened, emissions of coke dust rise up into
the room from the screen and the hoppers. The majority of this dust exits
the room through a large opening in the wall at the end of the screen on the
second level. The rest of the dust either settles out into the room or
exits the room via roof monitors or doorways.
TEST EQUIPMENT
The equipment used for the field test included two charged fog devices,
five high-volume (hi-vol) particulate samplers, two size-selective inlets,
and one cascade impactor. Each of these items is described below.
CHARGED FOGGERS
Two identical foggers were specially designed and fabricated for the
project by the Ritten Corporation. Ritten's standard Fogger III was modi-
fied and upgraded in order to allow for variations of its operating condi-
tions. The final configuration, designated "Fogger IV," is shown schematic-
ally in Figure 4.
In the generation of charged fog by the Fogger IV, water is atomized as
it is ejected from a nozzle via a compressed air supply. 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 centrifugal fan, pro-
jects the fog toward the dust source. A control panel, located on the back
of the fogger, allows for fogger operation and parameter variability.
The requirements for and capabilities of the operational parameters are:
o 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
1.5 kW (2 hp) compressor. The air flow through the nozzle is
variable from 0 to 11.3 m3/hr (0 to 400 scfh).
o 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 to 151 1/hr (0 to 40 gph).
o Power - The foggers require a power supply of 230 V, single phase, 60
Hz. The current requirements do not exceed 35 A. The power at the
induction ring is 12,500 V.
o Centrifugal fan - The fan, driven by a 3.7 kW (5 hp) explosion-proof
motor, operates at a maximum of 79 m3/min (2800 scfm). The maximum
output air velocity is approximately 3048 m/min (10,000 fpm). The
fan flow rate is variable from 0 to 100 percent of capacity.
o 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 10* for 75 1/hr (20 gph) water
flow. This can also be expressed as a charge/mass ratio of 0.11
PC/g.
144
-------
:BELT GUARD
^JUNCTION BOX (MOTOR)
15.2 cm DIAMETER INDUCTION RING
AIR ATOMIZING NOZZLE
NOSECONE
WATER LINE
22.9 cm
••-1—3.7 kW MOTOR
BELT DRIVEN CENTRIFUGAL FAN
-48.3 cm
WEATHER PROOF CONTROL
PANEL ENCLOSURE
CONTROL PANEL
:ONTROL BOX
AIR AND WATER INPUT
CONNECTION PORTS
230 VAC RECEPTACLE
230 VAC MAIN
DISCONNECT SWITCH
CONTROL CABINET
^-LIFTING EYE FOR SKID JACK
Figure 4. Schematic of the Ritten Corporation's Fogger IV
-------
SAMPLING EQUIPMENT
Hi-vol air samplers were used for the particulate measurements. The
hi-vols were manufactured by Misco Scientific of Berkeley, CA, and were
equipped with automatic flow control. This enabled the mass flow rate to be
held constant irrespective of filter loading, atmospheric conditions, or
line voltage changes. The hi-vols were operated at a nominal flow of 1.1
ms/min (40 cfm) which corresponds to a design particulate size cutoff of
approximately 30 um.
Two of the hi-vols were fitted with size selective inlets (SSI's) manu-
factured by Sierra Instruments of Carmel Valley, CA. These inlets are
designed to remove all particulates larger than 15 pm from the sampled air
before filtering the remaining particulates onto a standard hi-vol filter.
A Sierra Instruments Series 230 four-stage cascade impactor (CI) was
also used during several of the tests. When used in conjunction with an
SSI, the four stages separate the collected particles into size ranges of:
stage 1 - 7.2 to 15 vm, stage 2 - 3.0 to 7.2 ym, stage 3 - 1.5 to 3.0
Vim, and stage 4 - 0.95 to 1.5 ym. The remaining submicron particles are
collected on a backup hi-vol filter.
EQUIPMENT PLACEMENT
The equipment used for the majority of the coke screen test runs in-
cluded five hi-vols (two with SSI's) and the two foggers. The five hi-vols
were placed on the upper level catwalk in front of the doorway since the
plume was observed to travel across this area. The two foggers had to be
placed on the same side of the screen due to space limitations. One fogger
was placed on the upper level and aimed down and across the screen. The
other fogger was placed on the lower level about 2.7 m from the hopper. The
front end of this fogger was slightly elevated so that it aimed across and
above the hopper area. Figure 5 shows the positions and serial numbers of
the equipment.
The equipment positions remained constant for all of the test runs; not
all of the samplers were used for every run. All five samplers were used
for the first 31 runs. For the next 16 runs, only four samplers were
operated (standard hi-vol 7094 was eliminated) in order to allow more test
runs to be conducted. The last four test runs were conducted using only one
hi-vol (number 7092) fitted with both an SSI and a four-stage cascade
impactor. This sampler was moved to the center of the doorway for these
runs.
TEST PROGRAM AND PROCEDURE
The test program consisted of 51 runs during 6 days of testing. In-
cluded in the 51 runs were 13 uncontrolled, 8 fan only, 16 uncharged fog, 7
positive fog, and 7 negative fog. Specific test conditions are presented in
Table 1.
146
-------
GE
8031318
OWER LEVEL
FOGGER
803019
(UPPER LEVEL)
Figure 5. Equipment positions for coke screen.
147
-------
TABLE 1. TEST CONDITIONS - COKE SCREENING OPERATION
CO
Foqger 803018
Run
No.
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
Date
5-1-81
5-1-81
5-1-81
5-1-81
5-1-81
5-1-81
5-1-81
5-1-81
5-1-81
5-1-81
5-1-81
5-4-81
5-4-81
5- -81
5 81
5- -81
5- -81
5- -81
5- -81
5-4-81
5-4-81
5-5-81
5-5-81
5-5-81
5-5-81
5-5-81
5-5-81
5-5-81
5-6-81
5-6-81
5-6-81
5-6-81
5-6-81
5-6-81
5-6-81
5-6-81
5-6-81
5-6-81
5-6-81
Start
Time
1040
1100
1116
1210
1230
1247
1300
1318
1337
1402
1414
1300
1315
1330
1345
1400
1440
1450
1507
1520
1535
1400
1422
1443
1500
1511
1522
1536
0732
0812
0832
0837
0846
0856
0905
0915
0925
1003
1015
Duration
Of Run
(min.)
5.0
3.5
4.1
4.2
4.3
4.6
3.7
3.5
4.8
5.4
3.4
'3.2
6.3
3.1
7.3
2.9
5.1
3.8
3.9
4.9
4.4
3.3
5.4
3.0
4.1
4.6
3.8
4.8
1.2
2.9
2.5
3.1
3.6
5.8
3.5
2.7
3.2
3.1
3.0
Hater
Equipment Flow
Type of Test Positions (1/hr)
Uncontrolled
Uncontrolled
Uncontrolled
Uncharged Fog
Uncharged Fog
Uncontrolled
Uncontrolled
Uncharged Fog
Uncharged Fog
Uncontrolled
Uncontrolled
Uncontrolled
Uncharged Fog
Negative Fog
Uncontrolled
Uncharged Fog
Negative Fog
Uncontrolled
Uncharged Fog
Positive Fog
Uncontrolled
Uncontrolled
Uncharged Fog
Positive Fog
Uncontrolled
Uncharged Fog
Positive Fog
Fan Only
Fan Only
Uncharged Fog
Negative Fog
Fan Only
Uncharged Fog
Negative Fog
Fan Only
Uncharged Fog
Negative Fog
Fan Only
57
53
91
83
61
61
83
79
61
79
76
76
68
79
83
83
*
* 83
* 76
*
* 83
* 83
*
Uncharged Fog * 76
Mr
Flow
(m'/hr)
2.5
3.1
3.4
4.2
4.5
4.5
3.4
3.4
2.B
2.5
2.8
3.7
2.8
2.8
4.0
3.3
3.4
3.1
3.7
4.5
3.7
Fan
Speed
(%)
60
60
50
50
50
50
50
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
Fogger 803019
Water
Flow
(1/hr)
57
68
76
79
49
49
76
72
53
53
61
68
91
91
83
76
79
83
76
76
91
Air
Flow
(itiYtir)
2.5
2.8
2.8
2.5
3.1
3.1
2.1
1.4
2.B
2.3
2.8
2.4
2.5
2.3
3.5
3.4
3.7
3.7
3.7
3.7
3.1
Fan
Speed
(%)
50
50
50
50
50
50
50
40
40
40
40
40
40
30
40
40
40
40
40
40
40
40
40
40
40
40
NOTE: Refer to Figure 5
* Five samplers - 3 standard, 2 SSI
**Four samplers - 2 standard, 2 SSI (7094 eliminated)
-------
TABLE 1. TEST CONDITIONS - COKE SCREENING OPERATION (Continued)
10
Fogger 803016
Run
No.
40
41
42
43
44
45
46
47
48
49
50
51
Date
5-6-81
5-6-81
5-6-81
5-6-81
5-6-81
5-6-81
5-6-81
5-6-71
5-7-81
5-7-81
5-7-81
5-7-81
Start
Time
1025
1040
1055
1110
1225
1238
1255
1315
0922
0940
0956
1015
Duration
OC Run
(min.)
3.1
2.4
2.6
6.4
2.6
1.8
1.4
4.6
2.7
2.7
5.6
2.2
Water
Equipment Flow
Type of Test Positions (l/hr)
Positive
Fan Only
Uncharged
Positive
Fan Only
Uncharged
Positive
Negative
Fan Only
Uncharged
Negative
Positive
Fog
Fog
Fog
Fog
Fog
Fog
Fog
Fog
Fog
79
76
83
79
76
76
*
* 87
* 79
* 79
Air
Flow
(m'/nr)
4.2
3.1
3.7
3.7
4.8
4.0
3.4
3.1
3.4
Fan
Speed
(»)
40
40
40
40
40
40
40
40
40
40
40
40
Fogger 803019
Water
Flow
(l/hr)
87
91
76
79
76
87
83
79
76
Air
Flow
(m'/hr)
3.1
3.7
3.7
5.0
4.5
4.0
4.0
3.7
4.2
Fan
Speed
(»)
40
40
40
40
40
40
40
40
40
40
40
40
NOTE: Refer to Figure 5
** Four samplers - 2 standard, 2 SSI (7094 eliminated)
***One sampler - standard hi-vbl with SSI and CI
-------
The procedure was the same for each of the test runs. Pre-weighed
hi-vol filters were placed in the samplers between coke runs. The samplers
were simultaneously started once coke began to fall from the conveyor onto
the screen and simultaneously turned off at the end of the coke run. During
tests in which the foggers were used, they were started and adjusted to the
proper settings prior to the start of the coke run. The hi-vol filters were
immediately removed from the samplers at the end of each run and placed in
envelopes. At the completion of the field test, all of the filters were
returned to TRC's chemistry laboratory, desiccated, and weighed. The
resulting filter loadings were then used in conjunction with the sampler
flow rates to calculate particulate concentrations. These concentrations
were then used directly to calculate fogger efficiency.
TEST RESULTS
Table 2 summarizes the calculated concentrations for each of the five
test conditions (uncontrolled, fan only, uncharged fog, negative fog, and
positive fog). The values presented were obtained by calculating the geo-
metric mean of the data set for each hi-vol. Geometric means were preferred
over arithmetic means so that the effect of outliers in the data sets was
reduced. Also included in the table are the average TSP levels of the
s.tandard hi-vols and the hi-vols with SSI's. Hi-vol 7094 was not used in
the standard hi-vol averages because it was not operated during all of the
test runs and would thus bias some of the results.
Table 3 presents the fogger efficiencies that were calculated using the
previously described geometric means for the average of the two standard
hi-vols and the average of the two hi-vols with SSI's. In calculating the
efficiencies, the fan-only particulate matter concentrations were used as
the baseline. This was because the fans create an artificial wind effect
that is constant for all conditions except the uncontrolled one. The fan
air tends to redirect and, to some extent, reentrain some of the dust due to
the limitations imposed by the test apparatus positioning. The particulate
matter readings are spot readings only and, as noted, augmented by the fan
air. They are not typical of the general work exposure levels in the
facility. This phenomenon would probably not be present at a permanent
installation since the fog nozzles would most probably be positioned above
the source and aimed down at it. This arrangement is not possible with the
experimental test equipment.
As shown in Table 3, there was a slight reduction (15 to 25 percent) in
particulate matter concentrations as a result of the application of an un-
charged water fog on the dust emissions at the coke screen operation. When
a negative charge was applied to this water fog, the concentrations were
reduced only slightly further (approximately 10 percent). When a positive
charge was applied to the water fog, the concentrations were reduced an
additional 30 - 35 percent. This indicates that the dust plume was primar-
ily composed of negatively charged particles. The positively charged fog
produced by the two Fogger IV's reduced the concentrations at the coke
screen operation 40 to 50 percent. This level is consistent with observa-
tions which indicated that more than two foggers would be necessary to con-
trol the dust emissions from the operation.
150
-------
TABLE 2. RESULTS OF FOGGER TESTING AT COKE SCREEN OPERATION:
GEOMETRIC MEAN PARTICIPATE MATTER CONCENTRATIONS (pg/m3)
Hi-Vol Designation
Run
Condition
Uncontrolled
Fan Only
Uncharged
Fog
Negative
Fog
Positive
Fog
Standard
7112
73022
163859
147814
117689
100042
SSIa
7105
35800
76426
66128
56403
46459
Standard
7101
42699
143659
80755
71733
51372
SSI3
7092
26371
50514
40763
44190
29545
Standard
7094
45609
61799
56602
77376
37304
Avg. of
7112
and
7101
58300
157765
117098
101413
76598
Avg. of
7105
and
7092
31249
63764
53953
50658
38117
a <16 vim fraction.
TABLE 3. RESULTS OF FOGGER TESTING AT COKE SCREEN OPERATION:
FOGGER EFFICIENCIES (%)
Percent Reduction
Formula Used In
Calculation*
Standard
Hi-Vols
Hi-Vols
With SSI's
Fan Only - Uncharged Fog
Fan Only
Uncharged - Negative Fog
Uncharged
Uncharged - Positive Fog
Uncharged
Fan Only - Negative Fog
Fan Only
Fan Only - Positive Fog
Fan Only
x 100
x 100
x 100
x 100
x 100
26
13
35
36
51
15
29
21
40
* Inputs to formula are the geometric mean particulate matter concentrations,
151
-------
The last four test runs (48-51) were conducted using a hi-vol with an
SSI and a four-stage cascade impactor operated at 20 cfm. The purpose of
these four tests was to obtain particle size distributions. Only four runs
were conducted due to the length of time required for sample recovery.
Process and plume variations would rule out determining collection
efficiency reliably for this small number of tests. The results, shown in
Table 4, indicate the size distribution is bimodal and that it is relatively
unaffected by fogger conditions. Of the total particulate mass <16 ym,
22-36 percent falls in the <1.3 pm range. Referring back to Table 2,
the results of runs 1-47 showed that the <16 pm fraction represents
about 50 percent of the TSP, regardless of test conditions. Detailed
results of this test, the earlier tests, and four additional tests now
underway will be contained in the final report to be published in April
1982. The four tests now underway are co-funded with Armco, Inc.
CONCLUSIONS AND RECOMMENDATIONS
The results of the tests reported in this paper substantiate the
improvement of control efficiency expected when a charged fog is used to
suppress emissions from a coke screening operation. Although limited by fog
generating capacity and restricted as to placement of the foggers, these
tests achieved approximately a 50 percent reduction in particulate
emissions. This reduction occurred across the whole particle size
distribution generated by this process. It is felt that with increased fog
generating capacity and optimized fogger location the level of control could
be significantly improved.
Tests of charged fog for suppression of particulate should be evaluated
on other fugitive emission sources 'to gain further experience with this
technique.
152
-------
TABLE 4. RESULTS OF FOGGER TESTING AT COKE SCREEN OPERATION: CASCADE IMPACTOR DATA
/
on
GO
Measured Concentrations (vg/m* )
Test Stage 1 Stage 2 Stage 3 Stage 4
Run No. Type <10.2-L6um) (4 .2-10.2um) (2.1-4.2gm) (1 .3-2 .lum)
48 Fan Only 12313 20746 6194 4328
49 Uncharged 8424 10727 4545 3394
Fog
50 Negative 11036 16022 4268 2381
Fog
5.1 Positive 14088 17737 5766 3431
Fog
Backup Filter Total
(0-1. 3iim) (0-16um)
16866 60448
15333 42424
9496 43193
13066 5408B
-------
PERFORMANCE OF EOF EMISSION CONTROL SYSTEMS
By: Leonard J. Goldman, David W. Coy, and
James H. Turner
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, NC 27709
John 0. Copeland
U.S. Environmental Protection Agency
MD-13
Emission Standards and Engineering Division
Research Triangle Park, NC 27711
ABSTRACT
Many of the recently constructed EOF facilities worldwide have in-
corporated both primary and secondary emission control systems in initial
construction programs. Domestically, as a result of consent decrees and
state air regulations, some older plants have retrofitted secondary emis-
sion control systems and upgraded their primary emission control systems.
The effectiveness of these recent air pollution control systems varies
widely. The performance of most primary control systems has been good.
The largest variation in performance domestically appears among secondary
control systems. This paper discusses the technology in use at some plants
and reports the performance as well. Some data reported in this paper have
been obtained in the process of developing background information for New
Source Performance Standards for the United States Environmental Protection
Agency.
154
-------
PERFORMANCE OF BOF EMISSION CONTROL SYSTEMS
by: Leonard J. Goldman
David W. Coy
James H. Turner
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, N.C. 27709
and
John 0. Copeland
U.S. Environmental Protection Agency (MD-13)
Emission Standards and Engineering Division
Research Triangle Park, N.C. 27711
INTRODUCTION
The main goal in designing and installing a basic oxygen furnace (BOF)
emission control system is to limit the particulate emissions entering the
atmosphere. The degree to which this goal has been met is attested to, in
part, by the decline between 1970 and 1979 in both total nationwide particu-
late emissions and iron and steel related particulate emissions. Nationwide
estimates for various categories of emission sources are shown in Table 1.
The industrial process category, which includes the iron and steel industry,
has had a decline in emissions from the 1970 level of 10.2 teragrams (Tg)
(11.2 megatons (ton x 106)) to the 1979 level of 4.3 Tg (4.7 ton x 106) for
a decrease of 58 percent. During that time, nationwide emissions declined
from 21.0 Tg (23.2 ton x io6) to 9.5 Tg (10.5 ton x 106) for a 55-percent
decrease. At the same time, the iron and steel industries' emissions de-
creased from 1.25 Tg (1.38 ton x IO6) to 0.47 Tg (0.52 ton x IO6) (Table 2)
for,a 62-percent decrease, slightly better than the national average.1 The
BOF contribution to these emissions was 64.9 Gg (72 ton x IO3) in 1979,
which was roughly 14 percent of the iron and steel total (Table 3).2
BOF emissions may be divided into two categories: primary emissions
and secondary emissions. Primary emissions are those generated during the
oxygen blow phase of the BOF production cycle that are captured by the
primary hood system. Fugitive emissions generated during charging, primary
hood puffing, turndown for sampling, tapping, and deslagging are classified
as secondary emissions. Virtually all BOFs presently operating in the
United States are equipped with primary emission control systems. These
systems are of two types: closed hood equipped with wet scrubbers and open
hood equipped with either wet scrubbers or electrostatic precipitators
(ESP). One open hood shop in this country has used a baghouse for primary
emission control.
155
-------
TABLE 1. NATIONWIDE ESTIMATES OF PARTICULATE EMISSIONS
FOR 1970 AND 1979 *
Emissions
1970
Source
Transportation
Stationary source
combustion
Industrial processes
(includes iron
and steel industry)
Solid waste disposal
Miscellaneous
Total
Tg (ton x ios)
1.3
7.3
10.2
1.1
1.1
21.0
(1.4)
(8.1)
(11.2)
(1.2)
(1.2)
(23 '.2)
1979
Tg (ton x ios)
1.4
2.5
4.3
0.4
0.9
9.5
(1.5)
(2.3)
(4.7)
(0.4)
(1.0)
(10.5)
change
•cO. 08
-66.0
-58.0
-64.0
-18.0
-55.0
TABLE 2. PARTICULATE EMISSIONS
FROM THE IRON AND STEEL
INDUSTRY1
Emissions
Year Tg (ton x
1970 1.250-(1.38)
1979 0.470 (0.52)
TABLE 3. PARTICULATE EMISSIONS
FROM THE IRON AND STEEL
INDUSTRY—19792
Emissions
Source Gg (ton x 103)
Coke 131.5 (145.0)
Blast furnace 23.6 (26.0)
Sintering 42.9 (47.3)
Open hearth 26.5 (29.2)
Basic oxygen furnaces 64.9 (72.0)
Electric arc furnaces 95.7 (105.5)
Other 89.6 (98.8)
156
-------
PRIMARY EMISSION CONTROL SYSTEM
CLOSED HOOD SYSTEMS
In the closed hood system, a movable hood skirt is lowered to fit close
to the furnace mouth to restrict the inflow of combustion air. Since the
gases remain in a highly combustible state it is necessary to limit the
amount of air infiltration downstream of the hood. Potential leakage points
such as the lance port and flux chutes must be sealed and the entire system
must be nitrogen purged before use. Gas flow through these systems is
approximately 0.5 dscms/Mg (290 dscfm/ton) of heat size for top blown fur-
naces and 0.17 dscms/Mg (325 dscfm/ton) for bottom blown furnaces.
Cooling of the gas leaving the furnace initially occurs in the water-
cooled hood and continues in a sparkbox or quencher where grit and coarse
particles resulting from refractory wear and chunks of slag or metal are
separated from the gas stream. Quenchers reduce the gas temperature to less
than 93° C (200° F) and saturate the gas with water vapor.
From the quencher the waste gas flows to a high-energy scrubbing device,
where removal of the fine particles occurs. ESPs are not suitable for
closed hood installations because of the explosion potential inherent in the
combustible waste gas stream. The most common scrubber type is a venturi
with an adjustable throat. The venturi throat is opened or closed to increase
or decrease gas velocity, thereby controlling pressure drop through the
throat. Typical pressure drops are about 16.17 kPa (65 in. wg). The system
may have multiple venturi throats, but draft is provided by a single fan.
An integral part of the scrubbing unit is a moisture-separating device
to knock out drops of water carried out of the throat. This may be a series
of baffles or a centrifugal chamber in which the gas rotates, causing the
drops to impinge on the chamber walls. In addition, an after-cooling chamber
is sometimes used where cooling water is sprayed to reduce gas temperature
further. At cooler temperatures, moisture condenses, thus reducing the
volume of gas to be handled by the fan. The gas-cleaning facilities are not
shared between adjacent furnaces.
At present all closed hood systems in the United States flare the
carbon monoxide-rich waste gas. As an alternative to flaring, the gas may
be recovered in a gas-holding device and used as fuel gas for other plant
operations.
Draft control in closed hood systems is critical for proper control of
the combustion reaction. The systems typically, include hood pressure sen-
sors to alter draft via the adjustable venturi throat. Because the hood
draft is carefully limited to maintain near-atmospheric pressure, there is a
tendency for these primary systems to emit puffs of emissions at the hood-
furnace interface.
157
-------
TABLE 4. CLOSED HOOD PRIMARY SYSTEM EMISSION DATA--EPA REFERENCE METHOD 5
Emissions —
process weignt
basis
Test
Plant date
(TOP BLOWN)
Kaiser Staei
: ontana . CA4
30F 5 12/16/73
12/18/73
12/13/78
30F 6 12/14/78
12/14/78
12/16/73
steel
, S/Mg
produced
(Ib/ton)
9.
9,
3,
7
5
6
.30
.09
.35
.26
.22
.97
(0.0136)
(0.0182)
(0.0177)
(0.0145)
(0.0105)
(0.0140)
Slow
time
(min
31.
29.
32.
31.
30.
30.
)
31
64
25
92
28
97
dumber
of
cycles
;
2
2
2
2
•>
Emissions —
outlet
concentration
j/dscm
Ur/dscf)
0
0
0
0
0
0
.043 (0
.046 (0
.041 (0
.034 (0
.030 (0
.039 (0
.021)a
.020)1
.013)a
.01S)a
,013)a
.017)a
Scrubber
pressure drop
cm
(in.
175
173
135
226
^04
'? ** 1
H.:O
H,0)
(69)
^70)
u3)
'(39)
(38)
(37)
Armco Steel
Middletowa. OH-2
(. BOTTOM 3LOWN)
U.S. Steel.
10/20/71
10/21/71
10/23/71
5.77 (0.0116)
7.01 (0.0141)
7. 11 (0.0143)
123.75
117.30
124.L4
0.043 (0.021)3
0.071 (0.031)a
0.046 (0.020)*
-152 (-60)
-152 (-60)
-152 (-60)
r airfield. AL3 °
'"Z" Furnace 10/17/78
10/18/78
10/19/73
"X" Furtiaca 12/13/73
12/14/78
12/15/73
7.37
8.32
8.13
7.24
10.23
7.43
(0
(0
(0
(0
(0
(0
.0148)
.0167)
.0164)
.0145)
.0205)
.0150)
78.
71.
64.
69.
30.
71.
50
10
30
33
15
95
6
5
5
5
s
5
0.048
0.043
0.053
0 . 043
0.055
0.043
(0.
(0,
(0.
(0
(0
(0
.021)
.021)
.023)
.019)
.024)
.021)
165
163
170
173
168
173
(65)
(66)
(67)
(63)
(66)
C63)
Adjusted to oxygen blow time only.
The performance data base for closed hood systems is divided into two
subsets. The first subset, presented in Table 4, is made up of tests con-
ducted in strict accordance with EPA Reference Method 5.3 4 5 6 Outlet
concentration data for the top blown furnaces were obtained by sampling
during both blowing and nonblowing periods of the furnace cycle. For
Table 4, the concentrations were adjusted by calculation on the basis of
assuming all particulate mass was emitted during oxygen blowing only. This
assumption tends to overestimate the actual concentration that would have
been measured during the oxygen blow. The three-test averages for both
Kaiser Steel furnaces were below the present NSPS of 0.050 g/dscm (0.022 gr/
dscf) even with the adjustment.4 The three-test average for Armco Steel, on
an adjusted basis, is 0.055 g/dscm (0.024 gr/ dscf) which is above the
present NSPS.3
158
-------
TABLE 5. SUPPLEMENTARY CLOSED HOOD PRIMARY SYSTEM EMISSION DATA
Plant
Test
date
Saissions —
process weight
basis, g/ttg
steel produced
Qb/ton)
31ow
tvme
(rain)
Humber
01
cycles
Imissioas--
ouciet
concentration
g/iisan
(gr/dsct)
scrubber
oressurs droa
cm iioO
(in. H20)
(TOP BLOWN^
u.S. Steel
11/16/71
11/17/71
11/18/71
11/26/72
11/27/72
11/27/72
0.97 (0.0020)
3.93 (0.0079)
(0.0035)
(0.0021)
(0.0014)
1.73
1.04
0.71
2.40 (0.0048)
97.20
102.20
109.10.
102.80°
102.80°
102.80
i BOTTOM BLOW
U.S. Steel,
cast
West
Republic Steel,
South Chicago10
4/17/75
4/21/75
4/24/75
4/22/75
4/23/75
4/23/75
3/4/77
3/6/77
13.77=
14.52C
10.49
5.42=
2.21=
5.96C
(0.0276)
(0.0291)
(0.0210)
(0.0109)
(0.0044)
(0.0119)
73,
77,
00
00
6.67C (0.0134)
6.91C (0.0129)
73.00
67.00
68.00
72.00
43.00
43.00
0.007 (0.003)3
0.030 (0.013)3
0.011 (Q.005)3
0.007 (0.003)a
0.005 (0.002)d
0.018 (0.008)3
0.021 (0.009)3
0.023 (0.010)a
0.013 (0.008)3
0.011 (0.005)3
0.005 (O.OC2)a
0.011 (O.OOS)1
0.053 (0.023)
0.050 (0.022)
>i7S (>TO)
>173 (>70)
160 (63)
153 (62)
173 (63)
160 (63)
158 (62)
163 (64)
203 (SO)'
203 (30)c
L'.S. Steel,
Fairiield11 li
•rU" Furnace 11/6/74
11/7/74
11/7/74
"C" Furnace 9/3/73
9/9/73
9/9/78
4
5
5
7
10
1
.76C (0.0095)
.57C (0.0112)
.42C (0.0109)
.79 (0.0156)
.68 (0.0214)
.91 (0.0159)
75.
64.
59.
57.
55.
70.
,00
,00
00
25
,20
,67
5
4
4
5
4
5
0
0
0
0
0
0
.030
.032
.034
.055
.059
.050
(0
(0
(0,
(0,
(0
(0,
.013)
.014)
.015)
.024)
.026)
.022)
ISO
175
173
170
160
173
(71)
(69)
(58)
(67)
(63)
(70)
Adjusted to oxygen blow time only.
Average oxygen blow based on earlier tests.
Based on 131 Mg/heac (200 tons/heat), nominal production.
uesign value, Reference 25.
Data for the bottom blown closed hood systems at the U.S. Steel Fair-
field plant were obtained in tests in which emissions were measured during
the blow period only (Table 4). As shown in the table, the average of the
three runs for each furnace was nearly 0.050 g/dscm (0.022 gr/dscf).5 6
Data from other tests conducted at this plant are presented in Table 5.
The second data base subset, presented in Table 5, is composed of a
series of tests, which, for one reason or another, could not be verified as
159
-------
being in strict compliance with EPA Reference Method 5. This supplementary
data base supports the data in Table 4. Of particular interest are the low
outlet loading values reported for the U.S. Steel plants at Lorain, Ohio,7 8
and Gary, Indiana.9 These adjusted values are nearly an order of magnitude
less than values reported for the Kaiser and Armco plants (Table 4). The
origin of these differences is not readily apparent from the available
information.
With one exception, the top blown furnace closed hood primary system
emission tests presented in Tables 4 and 5 demonstrate that, for those
plants, the current NSPS emissions level [0.050 g/dscm (0.022 gr/dscf)] can
be met, even in the worst-case situation of having all the primary emissions
attributed to the oxygen blowing portion of the production cycle. Even
plants that predate the current standard are able to meet it. The one top
blown closed hood system that, on the adjusted basis, exceeded the NSPS
.limit (Armco Steel, Middletown, Ohio) did so on only one of three test runs.
On an unadjusted basis, i.e., with continuous sampling from the start of the
blow to the beginning of tapping, the plant complied with the current NSPS
limit.
OPEN HOOD SYSTEMS
An open hood scrubber control system is basically the same as that
described for closed hoods, except that the position of its hood skirt is
fixed instead of movable. Since all the combustible gases are burned in the
hood, no precautions against leakage into the system are necessary. Control
systems may be shared between furnaces and multiple fans operating in a
parallel flow arrangement can be used'if desired. Gas volumes are typically
0.78 dscms/Mg (1,500 dscfm/ton) of heat size for top blown furnaces.
When a precipitator is used, gas cooling downstream from the hood skirt
is achieved by use of water sprays located in the upper part of the hood.
These sprays are generally controlled by time and/or temperature to turn on
and off at various points in the operating cycle. The intent is to limit
the gas temperature reaching the precipitator and to condition the gases
with moisture for better precipitation. Maximum temperature of gases enter-
ing the precipitator is usually kept under 343° C (650° F). Flaring of
carbon monoxide-rich gas practiced in closed hood systems is not necessary
in open hood systems since the carbon monoxide is burned within the hood.
Because the gas temperature is relatively low during the early minutes
of a blow, some plants use steam injection at the hood or spark box to
achieve the desired conditioning of gases. Water sprays do not evaporate
sufficiently under the low temperature conditions, and puffs of iron oxide
fume are typically observed from the stack during this period. Steam injec-
tion both at the beginning and end of the blow can reduce these emission
puffs.
Downstream of the sparkbox the gases are carried to an inlet plenum
that distributes them to multichambered precipitators. On the outlet side
of the precipitator there is usually a manifold arrangement that distributes
160
-------
the gases among multiple fans. The precipitators may or may not have spare
capacity in terms of an extra chamber or extra collection field in the
direction of gas flow. At least one spare fan commonly is available.
Performance data from the period 1975 to 1979 for open hood control
systems are available. The data were obtained primarily during compliance
tests and show that all the plants were within limits set by the present
NSPS for BOFs (Table 6). All the tests were conducted during oxygen blowing
only.
The open hood systems had no difficulty meeting the NSPS emission
limits due, in part, to the dilution that results from the high evacuation
rates required in these complete combustion systems. When emissions are
expressed on a process weight basis, open hood systems in general do not
perform as well as closed hood systems (Table 4, 5, and 6). For a given
sized plant, a closed hood system allows fewer emissions into the atmosphere
than an open hood plant in spite of the lower outlet concentrations associ-
ated with open hood systems.
SECONDARY EMISSION CONTROL SYSTEMS
A variety of systems are used to capture secondary BOF emissions. Many
systems have been retrofitted to existing shops and are therefore unique,
while others were designed as original equipment in new shops. The systems
discussed are furnace enclosures (doghouses) with local hoods (Kaiser Steel,
Fontana, California; and Republic Steel, South Chicago, Illinois), partial
furnace enclosures with local hoods supplemented by partial building evacua-
tion (Inland Steel number 2 BOF shop), and open primary hoods modified to
facilitate secondary emission capture (Bethlehem Steel, Bethlehem, Pennsyl-
vania; and Republic Steel, Gadsden, Alabama).
Full Furnace Enclosures
Kaiser Steel (Closed Hood, Top Blown)--
The Kaiser Steel secondary emission control system at the Fontana,
California, plant controls furnace emissions (charging, tapping, puffing of
the primary, and turndown), and emissions from hot metal transfer and hot
metal skimming.20
The system has two fans each rated at 149 m3/s (315,000 acfm) at 50 cm
(20 in.) of water column and 230° C (450° F).21 Both fans operate to provide
the baghouse design flow of 283 m3/s (600,000 acfm). Dampers are used to re-
duce gas flow and energy consumption when full system flow is not required.
Air flow is divided among the various secondary hoods according to each opera-
tion' s needs. Operations permitted to occur simultaneously depend on whether
one or both furnace vessels are being used. Based on design information, hot
metal charging requires the largest air flow, about three-quarters of system
capacity. The Kaiser system does not permit hot metal transfer, hot metal
skimming, or hot metal charging to the other vessel while one vessel is being
charged. However, the system does permit oxygen blow, turndown, tapping, or
deslagging of one vessel during charging of the other. Hot metal transfer
161
-------
TABLE 6. OPEN HOOD SYSTEM PERFORMANCE DATA
Plant
U.S. Steel South Works,
Chicago, IL1-* - Scrubber
CFSI Steel
Pueblo, CO1* - (ESP)
Republic Steel,
Buffalo, NY1S - (ESP)
Wisconsin Steel,
Chicago, ILlb - (ES?)
Jones & Laughlin,
Aliquippa, PA17 - (ESP)
Test
date
6/27/77
6/29/77
7/1/77
4/10/73
4/11/78
4/12/78
10/20/75
10/21/75
10/22/75
11/10/76
11/12/76
11/16/76
11/18/76
8/10/76
8/11/76
Emissions —
process
weight basis
g/Mg steel
produced
(Ib/ton)
1.95 (0.0039)3
2.92 (0.0059)
3.45 (0.0011)3
2.94 (0.0059)3
5.64 (0.0113)a
5.14 (0.0130)
3.39 (0.0068)a
3.86 (0.0077)3
5.74 (0.0115)
6.04 (0.0121)a
4.41 (0.0088)3
4.10 (0.0082)3
33.38 (0.0669)3
13.30 (0.0267)a
18.71 (0.037S)a
13.77 (0.0267)3
12.55 (0.0251)3
-17.97 (-0.0360)
-28.44 (-0.0570)
-17.96 (-0.0360)
-20.96 (-0.0420)
6.29 (0.0126)
Emissions —
outlet
concentration
g/dscm
(gr/dscf)
0.0089 (0.0039)
0.0086 (0.0033)
0.0103 (0.0045)
0.0087 (0.0038)
0.0141 (0.0062)
0.0127 (0.0056)
0.0112 (0.0049)
0.0101 (0.0044)
0.0167 (0.0073)
0.0165 (0.0072)
0.0118 (0.0052)
O.C119 (0.0052)
0.0503 (0.0220)
0.0224 (0.0098)
0.0263 (0.0115)
0.0222 (0.0097)
0.0214 (0.0094)
0.0275 (0.0120)
0.0297 (0.0130)
0.0275 (0.0120)
0.0211 (0.0092)
0.0302 (0.0132)
0.0216 (0.0094)
0.0263 (0.0115)
0.0092 (0.0040)
0.0031 (0.0014)
Youngstown S&T,
Indiana Harbor, IN18 -
(ESP)
(Now J&L Steel)
Crucible Steel
Midland, PA"
(Baghouse)c
6/12/78
6/13/78
6/14/78
6/11/80
6/11/80
6/12/80
6/12/80
18.64 (0.0373)^
10.98 (0.0220)°
13.62 (0.0330)
9.40 (0.0188);
16.65 (0.0333)
14.35 (0.0237)"
0.030 (0.013)
0.018 (0.008)
0.027 (0.012)
0.0048 (0.0021)
0.0044 (0.0019)
0.0073 (0.0032)
0.0064 (0.0023)
Based on 181-Mg (200-ton) nominal capacity.
Based on 254-Mg (230-ton) nominal capacity.
Based on 95-Mg (105-ton) nominal capacity, and 20 minute blow.
162
-------
or hot metal skimming require about one-third system flow capacity and may
occur at any time, providing that neither furnace is being charged.
Performance of the secondary emission control facility during single
vessel operation at Kaiser was evaluated in April 198022 using visible
emissions measurements of roof monitor discharges (EPA Reference Method 9).
These measurements have been analyzed by two methods. Table 7 presents
the results analyzed according to EPA Method 9, i.e., 6-minute average
opacities based on observations made every 15 seconds. The table shows the
cumulative frequency distribution for 6-minute averages taken each day. On
the worst day none of the averages exceeded 15 percent opacity, while on the
best day none of the averages exceeded 5 percent opacity.
Table 8 presents data analyzed on the basis of average number of indi-
vidual opacity observations equal to or exceeding 20 percent opacity for
each segment of 21 production cycles. It is evident from the table that
turndown for sampling produced the greatest number of opacity excursions at
Kaiser Steel.
Republic Steel, Chicago (Closed Hood, Bottom Blown)—
Only three plants in the United States presently have bottom blown
furnaces (Q-BOP). The two furnace vessels in the Republic Steel plant, East
Chicago, Illinois, have a capacity of 205 Mg (225 tons). The secondary emis-
sions system at this plant includes full-furnace enclosures with charging
hoods at the front of each enclosure. There are no tapping hoods, and neither
hot metal transfer emissions nor hot metal skimming emissions are ducted to
this system.
Operations of the Q-BOP during charging and turndown require gas (either
nitrogen or oxygen) to be blown through the tuyeres to prevent liquid metal,
slag, or solids from entering and clogging the tuyeres. This purging makes
capture of the secondary emissions more difficult than with top blown furnaces.
Draft for the charging hood at the Republic plant is obtained from the
primary fume control system. Each furnace has its own primary gas cleaning
system; however, a crossover duct between the two furnaces permits the
system of the nonoperating furnace to be used for secondary emission control
on the operating furnace. With both gas cleaning system fans drafting the
charging hood, the flow rate is about 176 m3/s (373,000 acfm) at 93° C
(200° F) during hot metal charging.23 During charging, turndown, and tapping,
the charging hood is drafted by the scrubbing system of the nonoperating
vessel. Fume capture during these operations is assisted by drafting the
primary hood. Fumes captured in the secondary (charging) hood bypass the
quencher and pass directly to the venturi in the scrubbing system. The
design pressure drop of the venturi during furnace charging is 218 cm (86 in.)
water column.
In general, the performance of the secondary emission control system at
Republic Steel was poorer than the best performing top blown furnace secondary
systems. Data are available for two test series, June 1979 and June 1980.23 24
163
-------
TABLE 7. KAISER STEEL, FONTANA, CALIFORNIA22
Date/
opacity
4/7/80
4/8/80
4/9/80
4/10/80
4/11/80
Cumulative percent of 6-rainute averages Number of
less than or equal to given opaciiv individual
0
85.0
60.1
87.3
39.1
97.9
3
93.5
93.2
100
99.3
98.7
10 15
99.8 100
39.4 100
100
100
observations
1,633
2,310
2,151
1,533
1,549
TABLE 8. AVERAGE NUMBER OF TIMES PER HEAT THAT OPACITIES
WERE EQUAL TO OR GREATER THAN TWENTY PERCENT
Plant
Kaiser Steel22
Font.ana, CA
Republic Steel 197923
S. Chicago
(Q-BOP)
19802*
Inland Steel26
02 BOF Shop
Bethlehem Steel29
Bethlehem, PA
Republic Sterl30
Number of
heats
tested
21
22
5
6
11
42
Average number of observations >20 percent opacity
Charging 02 Blow
0.23
2.0 1 . 64
4.20 1.0
2.7 5.5
0.09 0
0.48 4.57
Turndown
1.86
3.45
3.40
1.7
0.18
0
Tapping
0.05
7.40
3.36
0.17
0
0.02
Gadsilen, AI.
164
-------
TABLE 9. REPUBLIC STEEL, CHICAGO, ILLINOIS23 24
Dace/
opacity
6/18/79
6/19/79
6/20/79
6/21/79
6/22/79
6/2/80
6/3/80
6/4/80
6/5/30.
6/6/80
Cumulative percent of
0
10.3
54.6
54.4
42.2
13.5
61.8
21.1
64.1
52.7
24.4
5
33.5
94.3
95.9
90.0
67.3
96.6
53.3
96.8
79.7
30.3
10
53.1
99.0
100
93.1
81.8
97.2
75.6
99.6
86.8
95.1
6 -minute averages less than or equal to Riven opacity
15
73.7
100
97.2
90.3
98.2
86.1
100
91.3
99.3
20 25 30 35 40 45
38.6 96.4 98.3 100
97.9 98.7 98.9 99.3 100
96.6 97.3 98.8 100
99.2 99.7 100
95.8 98.3 99.4 100
94.1 97.2 98.9 99.5 99.8 100
100
Number
of
individ-
ual
observa-
tions
1,363
1,319
1,901
1,670
588
1,241
1,765
2,448
5,675
1,174
These measurements have been analyzed by two nethods. Table 9 presents
results of data analyzed according to EPA Method 9, i.e., 6-minute average
opacities based on observations made every 15 seconds. The table shows the
cumulative frequency distribution for 6-minute averages taken each day.
No 6-minute average opacity exceeded 45 percent during the tests at
Republic Steel. However, only 3 of the 10 days during which observations
were made had no average opacities over 15 percent, as opposed to Kaiser
where all 5 days had no average opacities over 15 percent.
Table 8 presents the same data analyzed on the basis of the average
number of individual opacity observations equal to or exceeding 20 percent
opacity for each segment of 27 production cycles.
From examining this plant's performance on the basis of individual
cycles, it is evident that performance is characterized by extremes. One
group of cycles had 0 to 6 excursions equal to or greater than 20 percent,
and a second group of cycles had 19 to 36 excursions. No cycles fell between
these two groups. Specific causes for this large variation have not been
identified. The data also show that in many of the production cycles with a
high number of total excursions, 'each of the cycle segments, rather than
just one segment, contributes to the overall high emissions. Tapping,
however, appears to produce more excursions than the other cycle segments
(Table 8).
165
-------
During the 1979 tests it was noted that leakage occurring on the tapping
side of the furnace enclosure contributed to fugitive tapping emissions.
The addition of a separate tapping side hood to the enclosure might improve
overall system performance.
Partial Furnace Enclosure
Inland Steel, East Chicago, Indiana (Closed Hood, Top Blown)--
The No. 2 BOPF shop at Inland Steel's East Chicago plant contains two
195-Mg (215-ton) capacity top blown furnaces. The primary gas-cleaning sys-
tem is a closed hood type with venturi scrubbers.
There are two principal secondary emission control systems in this
plant. One system is composed of local hoods located in the partial furnace
enclosure. Fugitive emissions in the charging aisle are captured by a
partial building evacuation system, which also receives emissions from local
hoods at the hot metal transfer and hot metal skimming stations.
Local hoods within the partial furnace enclosure include a charging
hood, tapping hood, and a wraparound hood (at the sides of the furnace) to
capture puffing emissions during the oxygen blow. During charging, only the
charging hood is drafted; during tapping, the tapping hood and wraparound
hoods are drafted. During the oxygen blow, all three hoods are in service.
Draft for the furnace enclosure secondary emission control system is induced
through a venturi scrubber by a fan rated for 62 m3/s (131,000 acfm) at 21° C
(70° F).25 -Overall system pressure drop is 130 cm (51 in.) water column.
This evacuation rate is not sufficient to capture all charging and furnace
deslagging emissions, some of which escape to the partial building evacuation
system.
The partial building evacuation system is limited to the furnace charg-
ing aisle. There is a curtain wall which prevents the charging emissions
from escaping into the uncontrolled furnace aisle. There are two duct
takeoffs in the charging aisle roof, one centered above each furnace. A
damper is provided in each takeoff to open or close it as necessary. During
hot metal charging and furnace deslagging the damper is opened to maximize
the evacuation rate above the affected furnace.
Total air flow capacity for the partial building evacuation-hot metal
handling secondary emissions system is 189 m3/s (400,000 acfm) at 135° C
(275° F). Flow is divided between partial building evacuation and hot metal
handling, with 130 m3/s (275,000 acfm) allotted to the roof ventilation sys-
tem and 59 m3/s (125,000 acfm) to the hot metal handling station. The avail-
able system pressure drop is 38 cm (15 in.) water column and gas cleaning is
provided by a baghouse.25
Roof monitor visible emissions observations were made at this plant
during May 1980.26 These measurements were analyzed on the same basis as
the Kaiser Steel data; i.e., 6-minute average opacities were calculated
according to Method 9 procedures and the number of excursions equal to or
greater than 20-percent opacity was determined. Table 10 presents the
166
-------
TABLE 10. INLAND STEEL, EAST CHICAGO, INDIANA26
Number
of
individ-
ual
Date/
opacity
5/12/80
5/13/80
5/14/80
5/15/80
5/16/80
Cumulative percent of 6-minute averages less than or equal to given opacity
0
UH.
83.
iO.
22,
97,
,1
.6
.0
o
.2
5 10 15 20 25 30 35 40 45
79.2 89,6 98.6 100
100
88.4 100
66.2 83.6 89.6 94.7 97.3 98.4 98.9 99.9 100
100
observa-
tions
479
1,373
1,504
1,460
6-minute average opacities for 5 days and Table 8 presents the average
number of excursions per cycle segment for six production cycles.
The data in Table 10 show that on 3 of the 5 days the 6-minute average
opacities were as good as those observed during the Kaiser Steel tests. On
1 of the remaining 2 days there were several readings in the 15- to 20-percent
range, but none over 20 percent. On May 15, 6-minute averages in the 35- to
40-*percent range were observed. No correlation was found between these high
readings and events which transpired in the shop on that day.
Modified Primary Hoods
Bethlehem Steel, Bethlehem, Pennsylvania—
This EOF shop contains two 272-Mg (300-ton) furnaces equipped with open
hoods ducted to an ESP. Each furnace is partially enclosed by side walls,
with no enclosure on the charging or tapping sides. The only modification
to the primary hood system is an awning-like structure constructed on the
tapping side between the side enclosures that extends toward the teeming
aisle. This awning acts like a flanged extension that helps direct tapping
fumes into the primary hood.
During hot metal charging operations the gas evacuation rate for the
primary hood is 236 m3/s (500,000 acfm) at about 82° C (180° F).23 The
initial portion of the hot metal charge is performed with the furnace mouth
tipped only partially out from under the hood.. As the charge nears comple-
tion the furnace is tipped further, bringing the entire mouth out from under
the hood. Fume escape is worst at the end of the charge. During the oxygen
blow the primary hood evacuation rate is increased to 353 m3/s (750,000
acfm) at a temperature of 210° C (420° F). When the vessel is turned down
for tapping or other reasons the evacuation rate is reduced to the same
level as for charging.27 28
167
-------
TABLE 11. BETHLEHEM STEEL, BETHLEHEM,
PENNSYLVANIA29
Dace/
opacity
6/23/80
6/24/30
6/25/80
6/26/80
Cumulative
percent of
6-ninute averages
less than or equal
to given ooacity
0 5
80.6 100
93.9 100
96.3 100
81.4 100
Number of
individual
observations
1,413
1,920
1,920
1,811
Roof monitor visible emissions observations were performed at this
plant in June 1980.29 Table 11 presents the results analyzed according to
EPA Method 9, i.e., 6-minute average opacities based on observations made
every 15 seconds. The table shows that no 6-minute average opacity exceeded
5 percent on the 4 test days.
Table 8 presents the data analyzed on the basis of the average number
of individual opacity observations equal to or exceeding 20-percent opacity
for eleven production cycles. The data are broken down into segments of the
production cycle.
A significant portion of performance achieved at the Bethlehem plant
must be attributed to good operating practice and skillful crane and furnace
maneuvering. However, techniques employed at Bethlehem are not applicable
to BOF shops with closed hoods that have insufficient draft for good second-
ary emissions capture.
Republic Steel, Gadsden, Alabama—
This BOF shop contains two 136-Mg (150-ton) furnace vessels with open
hood primary gas-cleaning facilities. The secondary emission control system
consists of Gaw dampers installed under the face of the primary hood of each
furnace. During hot metal charging the Gaw damper closes off about 50 per-
cent of the primary hood face area. Increased gas velocity at the front of
the hood face improves capture efficiency of the primary hood during charg-
ing operations. The damper does not serve a similar function during tapping.
The gas evacuation rate during hot metal charging is about 283 m3/s (600,000
acfm) at a temperature of about 77° C (170° F). A reduced evacuation rate,
about 165 m3/s (350,000 at 77° C), is used during tapping, deslagging, and
other turndowns.30
Visible emissions data were gathered at this plant in August 1979.29
Roof monitor opacities were read during 42 furnace production cycles. These
data were analyzed on the same basis as previously described.
168
-------
TABLE 12. REPUBLIC STEEL, GADSDEN, ALABAMA30
Number
of
individ-
ual
Date/
opacity
6/20/79
3/21/79
3/22/79
3/23/79
Cumula;
0
60
52.
40,
48.
.5
.6
.9
.9
:ive
percent
5
94,
90.
84
94
.4
.0
]-
.3
of
10
97
96
90
97
.2
1
.0
.1
6-minute averages less than or eaual to given opacitv
15
98,
98
95
99
.2
.3
.1
.3
20 25 30 35 4C 45
99.1 100
99.5 99.7 100
97.9 98.8 98.9 99.1 99.2 100
100
observa-
tions
1
2
2
1
,992
,002
.344
,360
The data in Table 12 show that 6-minute average opacities were not as
low as those encountered at Kaiser Steel and Bethlehem Steel. However, the
data analysis presented in Table 8 shows that the main problem with emissions
occurred during the oxygen blow rather than during hot metal charging,
turndown, or tapping. Process observations made during the tests indicate
that these emissions were due to leaks through the oxygen lance hole, leaks
in the primary hood, and a primary system gas evacuation rate lower (by
10 percent) for one furnace than the other.30 The referenced report concludes
that the Gaw damper effectively controlled hot metal charging emissions.
Performance Analysis
The performance of secondary emission control systems varies greatly
between shops and very often, for any single system, varies on a day-to-day
basis and sometimes on a heat-to-heat basis. The most consistent performance
was seen at Bethlehem Steel, Bethlehem, Pennsylvania, where during 4 days of
testing no 6-minute average opacity exceeded 5 percent (Table 7). In compari-
son, performance of the system at Republic Steel's South Chicago Q-BOP shop
was not as good. On the best test day (6/20/79) approximately 4 percent of
the 6-minute average opacities exceeded 5 percent while on the worst day
(6/5/80) over 20 percent of the averages exceeded 5 percent (Table 9) with
the highest exceeding 40 percent. The Kaiser system is analogous to the
Republic Q-BOP system in that both have full furnace enclosures with local
charging hoods. The Kaiser system which also has a tapping hood, has demon-
strated superior performance with no 6-minute average opacity ever exceeding
15 percent and on the best day no 6-minute average exceeding 5 percent.
This system's good performance is a result of the high evacuation rate and
proper placement of the charging hood.
When comparing the Kaiser system to the Republic Q-BOP system, it
should be kept in mind that tuyeres in bottom blown furnaces must be purged
during all phases of the production cycle and that purge gases tend to
propel emissions away from the mouth of the furnace making their capture by
169
-------
local hoods difficult. Such remedies as increasing the charging hood evacua-
tion rate or installing a partial building evacuation system have not been
tried in any domestic Q-BOP installation. Attention should be drawn to the
fact that at one new Q-BOP installation in Japan the fugitive emission
problem has apparently been solved through the use of both local hoods and a
roof mounted ESP.31
The data displayed in Table 8 have been analyzed to determine the
effectiveness of these systems during different segments of the production
cycle. It is evident that the systems are not uniformly effective through-
out the cycle. The Kaiser system allowed greater opacity emissions to
escape during turndown than during charging or tapping. The Inland system,
however, allowed more emissions during charging than during turndown or
tapping. Inland's high opacity readings during oxygen blow have not been
accounted for although it seems unlikely that they were due to any defi-
ciencies in the secondary emission control system. A similar situation
existed at the Republic Steel, Gadsden, plant, with the exception that high
opacity readings during the oxygen blow were attributed to leaks in the
primary system.31
During the 1980 test at Republic Steel, South Chicago,25 there were no
great differences in the number of opacity observations 20 percent or greater
during charging, turndown, or tapping.25 During the 1979 test, however,
tapping created the greatest number of high readings.24
In general, the secondary emission control systems discussed above are
capable of maintaining roof monitor opacities below 20 percent, with some
exception for"bottom blown furnaces. The effectiveness of secondary emission
control systems is not uniform from one shop to another, and may not be uni-
form from one heat to another within a given shop.
REFERENCES
1. National Air Pollutant Emission Estimates, 1970-1979. EPA-450/4-81-010.
March 1981.
2. Unpublished data. National Air Data Branch, U.S. Environmental Protec-
tion Agency, Research Triangle Park, North Carolina. August 1981.
3. Midwest Research Institute. Source Testing--EPA Task No. 2. Armco
Steel Corporation, Middletown, Ohio. EPA Contract No. 69-02-0223 (MRI
Project No. 3585C). February 7, 1972.
4. Engineering Science. Report on Source Tests, Visible Emissions and
Plant Observations. Kaiser Steel Corporation, Fontana, California.
U.S. Environmental Protection Agency, Region IX, Contract No. 68-01-
4146, Task Order 50/TSA 2. February 1979.
5. CH2M Hill. Particulate Emission Measurement on Q-BOP "C" at United
States Steel Corporation, Fairfield, Alabama. MG63302.80. November
1978.
170
-------
6. CH2M Hill. Particulate Emission Measurement on Q-BOP "X" at United
States Steel Corporation, Fairfield, Alabama. Project No. Mg63302.80.
December 1978.
7. Engineering Science, Inc. U.S. Environmental Protection Agency Test.
United States Steel Corporation, Lorain, Ohio. December 21, 1971.
8. Engineering Science, Inc. U.S. Environmental Protection Agency Test
No. 72-MM-02. Basic Oxygen Furnace. United States Steel Corporation,
Lorain, Ohio. June 27, 1972.
9. United States Steel, Gary Works. Stack Emission Tests, No. 2 Q-BOP,
18-E-005(003). 1975.
10. Mostardi-Platt Associates, Inc. Particulate Emission Studies. Republic
Steel Corporation, Chicago, Illinois. U.S. Environmental Protection
Agency, Chicago, Illinois. P.O. 600-792434-621, MPA 70620-71.
August 4-7, 1977.
11. United States Steel Corporation. Q-BOP Emission Tests at United States
Steel Corporation, Fairfield, Alabama. December 12, 1974.
12. CH2M Hill. Particulate Emission Measurement on Q-BOP "C" at United
States Steel Corporation, Fairfield, Alabama. MG40.40. October 4,
1978.
13. Engineering Services Division, Department of Environmental Control,
City of Chicago. Stack Test Particulates B.O.F. Scrubbers, U.S. Steel
South Works. Report No. SS-258. July 1977.
14. York Research Corporation. Performance Testing of Basic Oxygen Furnace
Electrostatic Precipitators. Report No. 7-9651. CF&I Steel Corpora-
tion, Pueblo, Colorado. May 8, 1979.
15. Republic Steel, Buffalo, New York. Basic Oxygen Furnace Stack Emission
Tests. November 1975.
1,6. Engineering Services Division, Department of Environmental Control,
City of Chicago. Stack Test Particulates Basic Oxygen Furnace. Wiscon-
sin Steel Works, Chicago, Illinois. Report No. SS-255. November 1976.
17. Commonwealth of Pennsylvania, Department of Environmental Resources,
Bureau of Air Quality and Noise Control. Stack Emission Tests. Jones
and Laughlin Steel Corporation, Aliquippa, Pennsylvania. August 10 and
11, 1976.
18. Acurex Corporation. Particulate Matter Emission Rates for EOF Opera-
tions at Youngstown Sheet and Tube, East Chicago, Indiana. U.S. Envi-
ronmental Protection Agency, Chicago, Illinois. EPA Contract No. 68-
01-4142, Task 9. Acurex Report TR-78123, Volumes 1 and 2. August
1978.
171
-------
19. Crucible Steel Company, Midland, Pennsylvania. Emissions Testing of a
Positive Pressure Baghouse Control Device for a T.O.C. Test Report sub-
mitted by W.F.I. Sciences Company. June 30, 1980.
20. Trip Report. Kaiser Steel. Fontana, California. Research Triangle
Institute. December 2, 1979.
21. Letter and attachments from Martzloff, J. A. Kaiser Steel Corporation,
to Goldman, L. J., Research Triangle Institute, December 13, 1979.
Response to BOPF Questionnaire.
22. Clayton Environmental Consultants. Steel Processing Fugitive Emissions--
Emission Test Report Kaiser Steel Corporation, Fontana, California.
U.S. Environmental Protection Agency. Research Triangle Park, NC. EMB
Report 80-BOF-3. August 1980.
23. GCA Corporation. Assessment of Air Emissions From Steel Operations,
Republic Steel Corporation, Chicago District Q-BOP Shop Emission Evalu-
ation. U.S. Environmental Protection Agency, Washington, D.C. Contract
No. 68-01-4143, Task No. 58 Report. September 1979.
24. Clayton Environmental Consultants. Steel Processing Fugitive Emissions--
Emission Test Report Republic Steel Company, South Chicago, Illinois.
U.S. Environmental Protection Agency, Research Triangle Park, N.C. EMB
Report 80-BOF-7. September 1980.
25. Letter with attachments from Lang, D. C., Inland Steel Company to
Goldman, L. J., Research Triangle Institute. May 12, 1980. Response
to BOPF Questionnaire.
26. York Research Corporation. Inland Steel Plant No. 2, Indiana Harbor
Works, East Chicago, Indiana, Visible Emissions Observations Measure-
ments Program. U.S. Environmental Protection Agency. Research Tri-
angle Park, N.C. EMB Report 80-BOF-6, 1980.
27. Trip Report. Bethlehem Steel, Bethlehem, Pennsylvania. Research
Triangle Institute. May 21, 1980.
28. Letter and attachments from Ricketts, A. T. Bethlehem Steel Corpora-
tion, to McGrogan, J. E. Pennsylvania Department of Environmental
Resources, July 14, 1977. Test report on Bethlehem Pennsylvania BOPF
shop.
29. Clayton Environmental Consultants. Steel Processing Fugitive Emissions--
Emission Test Report, Bethlehem Steel Corporation, Bethlehem, Pennsylvania.
U.S. Environmental Protection Agency, Research Triangle Park, N.C. EMB
Report 80-BOF-9.
30. GCA Corporation. Assessment of Air Emissions From Steel Plant Opera-
tions, Republic Steel Corporation, Gadsden, Alabama. U.S. Environmen-
tal Protection Agency. Washington, D.C. Contract No. 68-01-4143, Task
Order 58 Report. March 1980.
172
-------
31. Coy, D. W. and R. Jablin. Review of Foreign Air Pollution Control
Technology for EOF Fugitive Emissions. In: First Symposium
on Iron and Steel Pollution Abatement Technology, Ayer, F. A. (ed.)
U.S. Environmental Protection Agency. Research Triangle Park, N.C.
EPA-600/9-80-012. February 1980. p. 233-251.
173
-------
INVESTIGATION OF OPACITY AND PARTICULATE MASS CONCENTRATIONS
FROM HOT METAL OPERATIONS
By: David S. Ensor
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, North Carolina 27709
ABSTRACT
The objective of this study was the investigation of possible relation-
ships between plume opacity, mass concentration, and sub-10-ym mass concentra-
tion for three hot metal processes. Size distribution data from a blast
furnace casthouse, BOF shop, and hot metal desulfurization were used to
compute mass light extinction coefficients. A wide range of materials (iron
oxide, carbon, and glass) were assumed to model the emissions. The results
suggest that opacity-to-mass concentration should be insensitive to composi-
tion, and a good correlation of mass-to-plume opacity is expected. The
emissions studied were of similar optical activity as those reported in the
literature in other industries. The results of the analysis also imply that
the sub-10-ym particles should be well correlated to mass.
It is recommended that future field tests to measure particle size dis-
tribution should include instrumental measurements of opacity. Thus, opac-
ity, mass concentration, and sub-10-ym particle concentration could be
subjected to a correlation analysis. In addition, the refractive index and
particle density of particulate material should be measured on bulk material
to aid in explanation of the test results.
INTRODUCTION
OBJECTIVES
The interrelationship of various emission limits for metal casting or
pouring is of considerable importance in the determination of how to best
regulate emissions. The purpose of this paper is to compare opacity, total
thoracic particulate concentrations (particles with aerodynamic diameters
less than 10 ym), and mass emissions. The present work is not intended to be
definitive but indicative of the possibilities in emission data
interpretation.
174
-------
BACKGROUND
Opacity limitations are of considerable value in regulating the iron and
steel industry. In particular, fugitive emissions common in iron and steel
processes cannot easily be quantified in any other way.
Recently, test data were acquired and analyzed for emissions from cer-
tain processes involving operations with hot metal. The data included, in
addition to process parameters, particle size distributions measured by cy-
clones and cascade impactors. A new computer program, which was recently
written for the TRS-SO™ microcomputer by Cowen, Ensor, and Sparks (1981) to
allow computation of opacity from emission aerosol properties, facilitated
the analysis.
THEORY
OPACITY AND TRANSMITTANCE
Opacity and transmittance are interrelated by the following equation:
Op - 100 (1 - T) (1)
where Op is the opacity (percent) and T is the fractional transmittance. The
Beer-Lambert law relates transmittance to intensive properties of an emission
aerosol:
T = exp (- bL) (2)
where b is the extinction coefficient (I/length) and L is the distance the
light travels. In the present case, L is the diameter of the stack or plume.
The quantity of interest is the mass concentration, M, at actual condi-
tions of temperature, pressure, and moisture. Equation 2 can be rewritten to
include mass concentration:
T = exp [- (b/M) ML]. (3)
The ratio of extinction coefficient to mass concentration, Sm = b/M, is
widely used in light transmission studies. It has also been called the mass
extinction coefficient. Ensor and Pilat (1971) used the symbol "Sp." Conner,
Knapp, and Nader (1979) termed it the quantity "A." In the present study, we
will use "Sin." In computational work it is advantageous to work with aerosol
volume concentration rather than mass concentration. The theoretical calcu-
lations, if performed with aerosol volume concentration for a given particle
refractive index, can be scaled to any particle specific gravity.
The parameters based on specific aerosol volume are:
Sv = b/M p (4)
where p is the specific gravity or density of the particulate matter. Ensor
and Pilat (1971) defined a parameter "K" which is the reciprocal of Sv. The
parameter K is very useful in relating mass to opacity (the current study is
involved in relating opacity to an observed mass concentration). The equa-
tion using K to relate mass concentration to opacity is:
175
-------
M = - In [(100 - Op)/100] Kp/L. (5)
In comparison to Equation 5, the equation relating opacity to mass concentra-
tion is:
Op = 100 [1 - exp (- Sm M L)]. (6)
Equation 6 will be used in the present study.
The effort in this study was directed at computing Sm as a parameter to
characterize the emission. Sm is a qualitative measure of the "optical ac-
tivity" of the emission and indicates the mass concentration required to meet
an opacity limitation. The principle data reported from the field test were
particle size distribution determined by cascade impactors. The other data
required to compute Sm directly, such as opacity measured by a transmissom-
eter, were not taken.
The extinction coefficient, b, from Equation 3, is computed by integra-
ting over the size distribution using the Mie equations describing the inter-
action of electromagnetic radiation with dielectric spheres. The mass
concentration, M, is computed by integrating the aerosol volume over size
distribution and multiplying by particle density. The computational approach
is described by Cowen, Ensor, and Sparks (1981).
APPLICATION TO A METAL FUME
QUALITATIVE DESCRIPTION OF THE EMISSIONS
The emissions resulting from the pouring of hot metal are a complex mix-
ture of materials. The bulk of these materials is probably iron oxide. The
remainder of the materials could be kish (a carbonaceous material), vaporized
refractory material, or dust entrained by the hooding system. An additional
complication in determining this mixture is the possible change of the emis-
sion properties during the process cycle, Nicola (1979). Cascade impactor
samples have been reported (Gronberg, Piper, and Reicher 1981) for casthouse
emissions to appear reddish in the lower stages and blackish in the large
particle stages. A similar observation was reported by PEDCo (1981) for EOF
shop emissions. The red material could be iron oxide in the hematite form.
The black material could be iron oxide in magnetite form, or graphitic car-
bon. Thus, the evaluation of the mass-to-opacity relationship with various
materials is an important part of the present study.
PHYSICAL PROPERTIES
The rather limited information available on particle properties of emis-
sions from operations involving hot metal dictated a survey of a handbook for
refractive index and specific gravity. The possible range in refractive in-
dex (transparent, 1.55, to absorbing, 3-li) and specific gravity (1 to
5.3 g/cm3) is quite large. Thus, one of the objectives of the analysis was
to determine the sensitivity of the computed Sm to refractive and particle
density.
176
-------
DEVELOPMENT OF CASE STUDIES
The case studies were selected using a combination of measured particle
size distributions and realistic physical properties. The following materi-
als were assumed:
1. Iron oxide.
2. Graphitic carbon.
3. Glass.
The particle size distribution case studies were:
1. Blast furnace casthouse emissions during casting.
2. Blast furnace casthouse emissions between casts.
3. EOF shop emissions.
4. Hot metal desulfurization.
The blast furnace casthouse emission particle size distributions were
reported by Gronberg et al. (1981) from tests conducted at DOFASCO in
Hamilton, Ontario. The tests were made in the total building evacuation con-
trol system at the inlet of the baghouse. The emissions should be reasonably
representative of uncontrolled emissions from the process.
The BOF emission particle size distributions were reported by PEDCo
(1981) from tests conducted at the Armco Steel No. 16 basic oxygen furnace at
Middletown, Ohio. The tests were done in the stack of the emission control
system downstream of a high-energy venturi scrubber. The BOF is top-blown
with a closed hood for emission control. Tests were conducted only during
oxygen blowing.
The hot metal desulfurization emission particle size distributions were
reported by Pope and Steiner (1980) from tests conducted at Kaiser Steel,
Fontana, California. Predetermined amounts of calcium carbide and calcium
carbonate were blown through a lance with nitrogen at 30 to 40 psi into a
torpedo car containing hot metal. Emissions escaping the stopper around the
lance were captured by a hood and ducted to a baghouse. Tests were conducted
at both the inlet and outlet of the baghouse. Only the inlet particle size
distributions were used in the present analysis.
Thus, a total of 12 different computer simulations were conducted be-
cause three different materials were modeled for four different size
distributions.
PARTICLE SIZE DISTRIBUTION
The particle size distributions measured by various EPA contractors were
used in this study. From a general examination of the data, all of the dis-
tributions are not lognormal. Therefore, an incremental integration of the
size distributions was conducted. The program used, called HISTOGRM/BAS, is
described in detail by Cowen, Ensor, and Sparks (1981).
177
-------
The reports used for the source of the size distribution data all are
in draft form and should be considered preliminary.
RESULTS
EXTINCTION COEFFICIENT TO MASS CONCENTRATION FOR EACH PROCESS
The results of the computations for each particle property case and pro-
cess are reported in Table 1. A surprising result for each process is the
relative insensitivity of the calculation to the particle properties.
(Graphitic carbon yields the highest Sm.) In addition, all the hot metal op-
erations investigated have similar Sm's.
The insensitivity of calculated opacity to refractive index and density
is explained by the shape of the particle size distributions emitted from
these processes. The size distributions are larger than 1 ym aerodynamic
diameter and very polydisperse. Thus, the extinction coefficient is averaged
over a wide range of particle diameters resulting in a value of Sm insensi-
tive to both composition and mean diameter fluctuations.
The opacity-to-mass concentration relationship should also be consistent
from the modified Beer-Lambert law in Equation 6. In other words, opacity
should be well correlated to mass concentration. A transmissometer would
provide a reasonably accurate measure of mass concentration.
EXTINCTION COEFFICIENT AS A FUNCTION OF PARTICLE DIAMETER
Additional examination of the particle size distribution effects is
shown in Figures 1 through 4. The cumulative percentage "less than" is plot-
ted as a function of aerodynamic particle diameter for each process. In all
cases the size distribution is very broad, as pointed out in the previous
section.
The EOF emission appears to be bimodal. One.distribution is less than
2 vim, while the other is larger than 10 ym. The EOF size distributions were
the only case determined downstream of a control device (high-energy venturi
scrubber). Thus, the bimodal nature of the size distribution may be due to
the scrubber and not the particulate formation process. The high sub-
2-ym particle concentrations may results from low efficiency of the scrubber
in that particle size region, while the large particles may result from en-
trainment from the scrubber. However, the reported test results were inade-
quate to explain the observed distribution.
A significant observation is the extinction coefficient contribution by
the material less than 10 ym in all processes. One obvious implication is
control of opacity will require control of sub-10-ym particles (a restatement
of common knowledge). Opacity is also a good indicator of sub-10-ym particle
concentration.
Each case of assumed particle properties was plotted for each process.
As shown in the figures, the emission properties do not affect the outcome to
any great extent.
178
-------
COMPOSITE EMISSION PROPERTIES
The emissions for these processes appear to be complex mixtures of com-
pounds and a realistic simulation should consider a combination of the case
study materials. For the purpose of comparison to other industries, a hypo-
thetical mixture of components was used to compute an average Sm for the
process shown in Table 2. It is understood that in reality the mixture of
components may change during process cycles and may be substantially differ-
ent than assumed here. In a very detailed study, the composition of the
emissions would need to be'carefully characterized. However, for the size
distribution case studies investigated in this paper, the exact proportions
are not critical.
COMPARISON TO OTHER INDUSTRIES
The literature was reviewed for opacity-to-mass data for other indus-
tries. These data are reported in Table 3. Note that the larger the Sm, the
greater the resulting opacity for a given particle mass concentration or the
greater the optical activity. One interesting result is that emissions from
hot metal processes are not exceedingly optically active. Although pulver-
ized coal-fired utility boiler emissions have about one-half the optical ac-
tivity as hot metal processes, other industries, such as Portland cement
manufacturers, have similar optical activity.
OPTICAL ACTIVITY OF TOTAL THORACIC PARTICLES
The analysis also suggests an important possibility for relating total
thoracic particle concentrations and plume opacity. As demonstrated by the
present theoretical calculationsj the extinction coefficient results almost
entirely from the total thoracic particles (sub-10 ym diameter). This sug-
gests two possibilities:
1. Plume opacity would be highly correlated with the total thor-
acic particulate concentration. The correlation would be even
better than that found for total mass concentration measured
with EPA Method 5. This idea is analogous to results reported
by Charlson et al. (1978) for visibility in the atmosphere.
The correlation of visibility to sub-2.5-ym particle mass was
observed to be much higher (r = 0.82 to 0.95) than to total
mass concentration (r = 0.3 to 0.92). A total thoracic Sm
(Smt) is defined as the ratio of the extinction coefficient to
the total thoracic mass concentration.
2. Suit'8 might be quite similar in magnitude for a wide range of
industrial processes. The basis of this theory is that if
only a narrow range of particle size (sub-10 ym) is considered
and particle refractive index and density are secondary fac-
tors, the mass extinction coefficient would be relatively
insensitive to the process.
These ideas are evaluated in Table A. Smt was computed by dividing Sm by the
fraction of particles less than 10 yn aerodynamic diameter. The correlation
of opacity to total thoracic mass concentration could not be directly tested
by the rather limited nature of the data for the hot metal processes. (The
179
-------
opacity was not instrumentally measured, for example.) The potential of good
correlation of the total thoracic mass concentration to opacity is suggested
by some of the literature power plant data because the range of Smt's is
smaller than S^ for the same source tests.
The contention that the Smt's for industrial emissions should be similar
is borne out to some degree by the interindustrial comparison. The range of
the Sjnt's is from 0.4 to 4 m2/g. A systematic rereduction of the literature
data combined with the objectives of determining the Smt's for various in-
dustries might narrow the range of values. At a minimum, the range of Smt's
for each industry could be established in addition to determination of the
correlation of opacity to total thoracic mass concentration.
In general, Table 4 demonstrates that these concepts have merit; however,
insufficient data exist to develop firm conclusions.
CONCLUSIONS AND RECOMMENDATIONS
CONCLUSIONS
Evaluation of the emissions expected for three hot metal processes in
four test cases revealed that optical activity is fairly insensitive to the
properties of the particles studied—iron oxide, carbon, and glass. Of these
three materials, carbon was consistently the most optically active. Although
the optical properties of the processes were quite similar, the EOF shop had
the greatest optical activity (largest Sm). The insensitivity of the com-
puted results to the particle properties is due to the averaging by the very
polydisperse sized emissions of the extinction coefficient over a wide range
of particle sizes.
In all cases the sub-10-ym particles were responsible for the light ex-
tinction. This observation suggests the hypothesis that various industries
may have similar magnitudes of the total thoracic particulate Smt (the ex-
tinction coefficient divided by the mass concentration of particles less than
10 ym aerodynamic diameter). It is also believed that opacity should be well
correlated with total thoracic particulate concentration. However, the ex-
isting data base is too limited to prove the concept.
Comparisons of the computed Sm's and literature Sm's from other indus-
tries indicated similar magnitudes. Portland cement emission optical activ-
ity was almost identical to the computed results for hot metal activities.
Pulverized coal-fired utility boiler emissions had about one-half the optical
activity as the studied processes. However, comparisons of opacity from
these two industries should consider the effects of stack diameter. The
utility industry, with larger stacks than the iron and steel industry, may
have similar opacity-to-mass concentration relationships.
RECOMMENDATIONS
Future test programs for particle size dependent emission factors should
consider the use of optical transmissometers to measure opacity directly over
a known pathlength. The opacity-to-mass relationships suggested by the
180
-------
computations in this paper should be verified experimentally. The light
scattering computations described in this paper could be used to provide
quality control as well as investigating the ramifications of the data.
Additional laboratory work should be conducted on the collected emis-
sions from the hot metal processes to determine the particle physical prop-
erties. In particular, the particle density and the refractive index should
be measured. The particles may have to be sorted into various particle
populations as part of the investigation.
REFERENCES
Brennan, R. J., and R. Dennis. (1980) Review of Concurrent Mass Emissions
and Opacity Measurements for Coal-Burning Utility and Industrial
Boilers. EPA-600/7-80-062, U. S. Environmental Protection Agency.
Charlson, R. J., A. P. Waggoner, and J. F. Thielke. (1978) Visibility Pro-
tection for Class I Areas: The Technical Basis. PB-288 842, The
Council on Environmental Quality, Washington, D. C.
Chemical Rubber Corporation. (1959) Handbook of Chemistry and Physics.
Conner, W. D. (1974) Measurement of the Opacity and Mass Concentration of
Particulate Emissions by Transmissometry. EPA-650/2-74-128, U. S.
Environmental Protection Agency.
Conner, W. D., K. T. Knapp, and J. S. Nader. (1979) Applicability of Trans-
missometers to Opacity Measurements of Emissions. EPA-600/2-79-188,
U. S. Environmental Protection Agency.
Conner, W. D., and N. White. (1981) Correlation between Light Attenuation
and Particulate Concentration of a Coal-Fired Power Plant Emission.
Atmospheric Environment, 15:939-944.
Cowen, S. J., D. S. Ensor, and L. E. Sparks. (1981) In-Stack Opacity Com-
puter Programs: User and Programmer Manual. Draft report, U. S. En-
vironmental Protection Agency Cooperative Agreement R806718010.
Ensor, D. S., and M. J. Pilat. (1971) Calculation of Smoke Plume Opacity
from Particulate Air Pollutant Properties. JAPCA, 21:496-501.
Ensor, D. S., S. Cowen, R. Hooper, and G. Markowski. (1979) Evaluation of
the George Neal No. 3 Electrostatic Precipitator. EPRI FP-1145, Elec-
tric Power Research Institute, Palo Alto, Calfornia.
Gronberg, S., S. Piper, and E. Reicher. (1981) Blast Furnace Casthouse
Emission Factor Development DOFASCO, Hamilton, Ontario. Draft report,
U. S. Environmental Protection Agency Contract Nos. 68-01-4143,
68-02-2687, and 68-02-3157.
Nicola, A. G. (1979) Blast Furnace Casthouse Emission Control. Iron and
Steel Engineer, 56:33-39.
181
-------
PEDCo Environmental, Inc. (undated, presumed 1981) Inhalable Particulate
Emission Characterization Report, ARMCO STEEL's No. 16 Basic Oxygen
Furnace, Middletown, Ohio. Draft report, U. S. Environmental Protection
Agency Contract No. 68^-02-3158.
Pope, R., and J. Steiner. (1980) Particulate Mass and Particle Size Mea-
surements for the Hot Metal Desulfurization Plant at Kaiser Steel,
Fontana, California. Draft report, U. S. Environmental Protection
Agency Contract No. 68-02-3159.
TABLE 1. CALCULATED MASS EXTINCTION COEFFICIENT S (m2/g)
m
Material:
Refractive index:
Density (g/cm3) :
Process:
Blast furnace casthouse
During casts
Between casts
BOF shop
Hot metal desulfurization
Iron oxide*
3.0-l.Oi
5.3
0.926
0.629
1.87
0.681
Graphitic carbon**
1.96-0.661
2.0
1.68
1.11
3.74
1.31
Glass**
1.55
2.4
1.06
0.766
1.67
0.942
Wavelength of light - 0.55.
* Refreactive index and density estimated from The Handbook of Chemistry
and Physics (1959). The imaginary or absorbing part of the refractive
index was assumed to be 1.0.
** The refractive index and density from Ensor and Pilat (1971).
182
-------
TABLE 2. ESTIMATION OF COMPOSITE EMISSION MASS
EXTINCTION COEFFICIENT S (m2/g)
Fraction
Component of
component
Iron oxide 0.60
Graphitic carbon 0.35
Glass O.OS
Total
Blast Furnace Casthouse Hot metal
Durina casts Between casts BOF shop dtBuiruilisalluii
Pure Fraction Pure Fraction Pure Fraction Pure Fraction
0.926 0.556 0.629 0.377 1.87 1.12 0.681 0.409
1.68 0.585 1.11 0.390 3.74 1.31 1.31 0.459
1.06 0.053 0.766 0.0383 1.67 0.084 0.942 0.0471
1.20 0.805 2.52 1.11
TABLE 3. COMPARISON OF THE CALCULATED MASS EXTINCTION
COEFFICIENT TO OTHER SOURCES
Source
(m2/g)
Reference
Pulverized coal-fired boiler
Coal stoker smoke
Portland cement (wet process)
Portland cement (dry process)
Oil combustion
Casthouse
During casts
Between casts
BOF shop
Hot metal desulfurization
0.30 to 0.49
0.78
0.58
0.4 to 0.92
0.1 to 0.5
6.1
1.55
0.92
0.20 to 0.43
1.2
0.81
2.5
1.11
Conner (1974)
Ensor and Pilat (1971)
Conner (1981)
Ensor et al. (1979)
Brennan and Dennis (1980)
Ensor and Pilat (1971)
Conner et al. (1979)
Conner et al. (1979)
Conner at al. (1979)
This study
This study
This study
This study
183
-------
TABLE 4. ESTIMATION OF S
mt
Process
Total Sm Fraction less than Smt
(m2/g) . 10 \m diameter (m2/g)
Reference
Pulverized coal-fired boiler
0.58
C.91
0.64
Conner (1981)
Portland Cement
Wet process
Dry process
011-fired boiler
Blast furnace casthouse
During casts
Between casts
BOF shop
Hot metal desulfurization
1.55
0.92
0.20
0.43
1.2
0.8
2.5
1.1
0.85
0.75
0.60
'0.90
0.53
0.39
0.68
0.73
1.82
1.23
0.33
• 0.48
2.2
2.1
3.7
1.5
Conner et al .
Conner et al .
Conner et al .
This study
This study
This study
This study
(1979)
(1979)
(1979)
mt Fraction mass less
than 10 inn
184
-------
<
z 100
UJ
O
£ 90
o.
£ 80
UJ
E *>
u.
O 60
u
| 50
I «°
5 „
uj 30
8" 20
10
°0
EXTINCTION COEFFICIENT DISTRIBUTION
Refractive Index Density g/cm3
• 3-1i 5.3
Al.96-0.66i 2.0
• 1.55 2.4
O SIZE DISTRIBUTION BY MASS
1.0 10.0
AERODYNAMIC PARTICLE DIAMETEa cm
100.0
Figure 1. Cumulative percentage of mass and extinction
coefficient for casthouse emissions during casts.
EXTINCTION COEFFICIENT DISTRIBUTION
Refractive Index Density g/cm3
• 3-1i 5.3
A 1.96-0.661 2.0
• 1.55 2.4
O SIZE DISTRIBUTION BY MASS
1.0 10.0
AERODYNAMIC PARTICLE DIAMETER, pm
100.0
Figure 2. Cumulative percentage of mass and extinction
coefficient for casthouse emission between casts.
185
-------
EXTINCTION COEFFICIENT DISTRIBUTION
Refractive Index Density g/cm3
• 3 11 5.3
A 1.96 0.66i 2.0
• 1.55 2.4
O SIZE DISTRIBUTION BY MASS
1.0 10.0
AERODYNAMIC PARTICLE DIAMETER, jim
100.0
Figure 3. Cumulative percentage of mass and extinction
coefficient for EOF shop emissions.
<
100
u 90
cc
uj
a- 80
UJ 70
o
it 60
UJ
O
o 50
o
I 30
X
UJ
. 20
I 10
EXTINCTION COEFFICIENT DISTRIBUTION
Refractive Index Density g/cm3
• 3-1i 5.3
A 1.96-0.661 2.0
• 1.55 2.4
O SIZE DISTRIBUTION BY MASS
0.1
1.0 10.0
AERODYNAMIC PARTICLE DIAMETER, jim
100.0
Figure 4. Cumulative percentage of mass and extinction
coefficient for hot metal desulfurization emissions,
186
-------
RETROFITTING EMISSION CONTROLS ON ELECTRIC FURNACES AT A STEEL MILL
by: Michael P. Barkdoll, P.E.
Enviro-Measure, Inc.
Knoxville, Tennessee 37901
Donald E. Baker
Rockwood Iron & Metal Company
Rockwood, Tennessee 37854
ABSTRACT
The body of information presented in this paper is directed to design
engineers, electric arc furnace owners, and others who are interested in air
pollution emission control technology. This paper presents the methodology
and results of a two year project at the Knoxville Iron Company. An exten-
sive field measurement program was conducted to quantify plume generation
rates from two-30 ton electric arc furnaces melting scrap. Plume measurements
were made for charging, tapping, and melting. Measurements were also made for
tundish lancing and billet cutting. Maintenance and operational problems
of the first generation side draft-hood and baghouse system were inventoried.
Based on the source characterization and system performance, an integrated
air pollution control system was designed and installed. Installed system
capacity was 300,000 ACFM utilizing two shaker-type baghouses with flow
switching and continuous limestone injection. Capital and installation costs
are presented by major category. In addition, pertinent design parameters
and system performance data are presented.
PRODUCTION FACILITIES
The shop has one 12 ft. diameter furnace rated at 12.5 MVA and one 12.5
ft. diameter furnace rated at 15 MVA. Both furnaces are floor mounted, and
top charged with scrap bucket. Liquid steel is tapped into a ladle which
is positioned in the tapping pit. Nominal weight of each tap is 30 tons,
with approximate tap-to-tap times of 2 to 2^ hours. The melt shop operates
continuously.
Charge materials consist of shredded automotive scrap, borings, turnings,
and other miscellaneous scrap iron, lime, and ferroalloys. The product from
the furnaces is low carbon steel used for reinforcing bar production. Oxygen
lancing through the slag door is used for carbon content control.
Molten metal from each furnace is transferred by crane to the casting
platform. Molten metal is withdrawn from the bottom of the ladle. It is
split into three liquid streams by means of a refractory lined manifold
(Tundish). If the openings in the bottom of the tundish become clogged or
flow of the molten metal is not continuous, the tundish is removed from the
casting machine. The openings are unclogged by blowing oxygen (tundish
187
-------
lancing) through them. A small diameter pipe with oxygen at 50 to 100 psi is
used. The injected oxygen burns away the hot metal which has solidified in
the bottom of the tundish. Each liquid stream is continuously cast into a 3"
to 4" square strand. Each strand is automatically cut to a predetermined
billet length. Cutting is performed with oxy-natural gas torches. The hot
billets are transferred to a scrap preheater and are then stockpiled for
future use in the rolling mill, where they are reheated and rolled into
reinforcing bar.
HISTORICAL PERSPECTIVE
Air pollution control equipment was installed on Furnace No. 1 in 1972.
The system consisted of an eight compartment shaker type pressure baghouse.
Design flow was 128,700 ACFM @ 250° Fahrenheit with a gross air-to-cloth ratio
of 3.25, and a net air-to-cloth ratio of 3.71. The fume collection system
consisted of a 70" diameter duct connected to a 4' x 6' side draft hood
mounted on the furnace. In addition a 70" duct was installed in the building
roof above the furnace to collect fumes from charging and tapping.
In July 1975 a second furnace was installed and fitted with a similar
emission control system. The system consisted of a ten compartment shaker
type pressure baghouse. Design flow was 168,000 ACFM @ 250° Fahrenheit with
a gross air-to-cloth ratio of 3.11 and a net air-to-cloth ratio of 3.46. A
side draft hood and an 82" overhead duct were also installed. (See Figure 1.)
During the period of 1975 to 1977, various portions of the emission
control system became inoperative. Chronic maintenance problems were ex-
perienced at the shaker mechanism, isolation dampers, overhead switching
dampers, side draft hoods, bags, and dust handling systems. By 1977, the
total system had degenerated to a state where it could not adequately control
emissions despite major maintenance efforts. In 1977 the company was formally
notified by the local regulatory agency that corrective measures would have
to be instituted.
PHASE I. FEASIBILITY AND CONCEPTUAL DESIGN
The feasibility and conceptual design phase was started in September of
1977. Phase I consisted of two major areas, 1) Source characterization and
2) Existing baghouse system performance.
I. Source Characterization
A. Furnace Operations
1. Melting
2. Tapping
3. Charging
II. Existing Baghouse System Performance
Concerning source characterization, various methods were used to quantify
plume generation rates.
MELT OPERATIONS
Pitot tube traverses were taken in the ductwork that services the side
draft hoods. In addition, static pressures (draft) near the mouth of the
hoods and fan amps in conjunction with total pressures across the fans were
measured during normal melt operations. Visual observations as to the degree
188
-------
OVERHEAD ROOF DUCT
10 COMPARTMENT BAQHOUSE
FIGURE 1. OLD EMISSION CONTROL SYSTEM FURNACE NO.2
189
-------
of capture during these measurements were also taken. From these data, the
necessary flow to provide adequate capture at the side draft hoods was
determined.
TAPPING/CHARGING AND TUNDISH LANCE OPERATIONS
The volumetric flow rate of the plume generated during the furnace tap-
ping operation was measured by two methods. Method 1 consisted of timing
the centerline ascent of the plume over a known distance in conjunction with
the taking of a photograph. The photograph was subsequently analyzed to
determine the entrainment angle and the cross sectional area, see Figure 2.
A recent investigation (Reference 1) of hot, buoyant plumes indicate
that plumes generated from a point source rise and spread according to some
definite physical relationships. It was found that unobstructed plumes of
this nature spread at an angle of approximately 18° (entrainment angle) and
that the average rise velocity for the total cross section is approximately
50% of the maximum rise velocity observed at the plume centerline or core.
These relationships were used to calculate the volumetric flow rates. Method
2 consisted of allowing the plume to impact on the roof of the building,
spread to the total width of the building and then be transported along the
roof trusses by overhead drafts. The depth of the confined plume below the
roof was observed to determine the cross sectional area and the total length
of the plume was also observed. The above measurements yielded a total plume
volume in cubic feet and the volumetric flow rate was calculated based on
the measured time of plume generation.
TORCH CUTTING OPERATIONS
The volumetric flow rates above the three strand torch cutting operation
were measured with a vane type anemometer. Velocity measurements were taken
at 24 points approximately 5 ft. above the actual cutting location. Photo-
graphs and visual observations were made to determine the cross sectional
area of the hot bouyant plume.
RESULTS OF SOURCE CHARACTERIZATION
Detailed results of the source characterizations are presented in
Reference 2. Based on the results of the field measurements, the following
design rates were determined:
1. Side Draft Hood 90,000 ACFM each
2. Charging/Tapping Hoods 200,000 ACFM each
3. Tundish Lancing Hood 110,000 ACFM
4. Torch Cutting 30,000 ACFM
EXISTING BAGHOUSE EVALUATION
A detailed analysis of the baghouse system was undertaken. The analyses
consisted of:
1. Total system flow rates
2. Static pressure profiles throughout the entire system
3. Pressure drop across individual compartments
4. Review of maintenance records
190
-------
LADLE
EXISTING ROOF
DUCT
LEGEND
ACTUAL PLUME OUTLINE FROM PHOTO
>:•:•,•' CONCEPTUAL OUTLINE OF PROBABLE
$* PLUME SHAPE
\ ! GRAPHICAL INTERPRETATION OF PLUME
\! BOUNDARIES
GRAPHICAL INTERPRETATION OF STREAM
LINES
SIDE
DRAFT
HOOD
PLUME No. -
-------
5. Extensive interviews with maintenance personnel
6. Inspection of failed parts
Based on field measurements, the dust collection system was operating
at about one-half of its design capacity (i.e., a, 55,000 ACFM at baghouse #1
and ^ 75,000 ACFM at baghouse #2). Pressure drop measurements showed from
13 to 14 inches pressure drop (back pressure) across each bag compartment.
Air-to-cloth ratios were measured at 3.7 to 5.1:1 on functioning compartments.
These conditions were detrimental to baghouse performance and expecially bag
life. The bags did not last long at these high air;cloth ratios. Baghouse
compartments which had recently been rebagged quickly blinded because of the
high air-to-cloth ratios and heavy dust loadings. Total available static
pressure across the fans was determined to be ^ 15 inches (w.g.) at 150
Fahrenheit. Both baghouses were operating at the peak of the fan curves with
87-93 percent of the resistance in the system due to the filter bags.
Normally, the system should have operated with 50-60% of the resistance due
to the bags and 40-50% due to the ductwork resistance.
The primary reason for the reduced system performance was due to the
number of bag compartments which were not functional due to old blinded bags.
Bag blinding occurs when fine particulate matter fills the interstices of the
fabric so completely that there are few remaining pathways for gases to travel
through the fabric, hence the bags presented a high resistance to air flow.
Bag blinding has been exaggerated by the "sticky" nature of the dust which
was not easily removed from the bags.
The decline of the system performance was not tracked to any single
cause, but rather a series of events. Of prime importance were isolation
damper malfunctions and of secondary importance was fabric selection. The
isolation dampers were equipped with small electric actuators which could not
adequately open and close the dampers., The problem was exaggerated because
the dampers were located on the dirty air side of the bags. The bags became
heavily loaded with dust and overloaded the shaker mechanisms which failed
repeatedly. Some early fabric selections had been toward rather heavy bags
with poor cake release properties. It was impossible for the maintenance
personnel to keep up with rebagging, repairing shaker mechanisms, and main-
taining isolation dampers. Hopper bridging and maintenance on the lead
section of the side draft hoods was also a continual problem. The large
dampers which were designed to switch the flow from the side draft hoods to
the overhead duct were inaccessible except by crane, and had become inoperable.
DETAILED DESIGN
TORCH CUT
Further investigation into the torch cut emission problem revealed that
an oversized torch tip was being used. The torch manufacturers were con-
tacted. They indicated that a smaller torch tip should be adequate. Aside
from the physical size, the old torch tip used considerably more oxygen and
gas. New torch tips were installed and tested. The new torch tips adequate-
ly cut the billets and visible emissions were virtually eliminated. In
conjunction with the new torch tips it was also learned that operators had a
tendency to set the oxygen regulator up to 140 psi. This was done when the
body of a torch became misaligned, and rather than correcting misalignment,
oxygen pressure was increased to compensate. Any over supply of oxygen
192
-------
again produced some visible emissions.
As a consequence, the oxygen regulators were locked at 100 psi. With these
two minor corrections, a serious source of visible emissions was completely
eliminated.
MODIFIED SYSTEM ARRANGEMENT
It was determined that if minor adjustments were made to the process
schedule, the existing baghouse system had sufficient capacity to control
emissions. Maximum fan capacity and flow switching would be required to
maintain design volumes at each emission point.
Shown in Figure 3 is a plain view of the new system arrangement. Bag-
house No. 1 was redesigned to handle 130,000 ACFM at 150° Fahrenheit, and
baghouse No. 2 was redesigned to handle 160,000 ACFM at 150° Fahrenheit. In
order to insure maximum flow rates, four additional compartments were added
to each baghouse. This lowered the gross air-to-cloth ratio to 2.1 to 1.
Compartments were added with minimal ductwork modifications.
The existing side draft hoods were judged to be quite adequate, except
for the persistent maintenance problems associated with the water cooled
section. A new inlet section was designed using 340 stainless V corrugated
plate. The corrugated material was used to minimize stresses due to thermal
expansion and contraction. The overhead roof ducts were judged to be totally
inadequate, and a new overhead hood system was designed.
The new overhead hood system was designed in conjunction with the .;side
draft hoods to provide the following design volumes.
Side Draft Hood 80,000 CFM (adjustable)
Overhead Tapping/Charging 210,000 CFM
. Tundish Lancing 130,000 CFM (On demand)
: Process limitations were that only one f-ua;p.ace could be tapped or charged
at a time, and tundish lancing only when both furnaces are under normal melt
conditions. Flow distribution is accomplished- by sensing static pressure on
the side draft hood leg. The measured static pressure is maintained by a feed
back loop connected to the modulating damper in the overhead hood^system. In
this manner, excess capacity from each baghouse (that which is in excess of
80,000 CFM) is used to control tapping/charging, or tundish lancing.
So that the design flows would be used most efficiently to capture
emissions, new hoods were designed for above each furnace and the tundish
area. Shown in Figure 4 is a drawing of the charging/tapping hoods. Each
hood is internally partitioned, and appropriately dampered to direct the
flow to the charging/or tapping area. Each hood has overall dimensions of
44' x 24* at the face. Each partitioned area is 32' x 24' at the face, which
produces an average face velocity of 275 fpm. A flat flanged area, 5 ft. wide
was installed around the perimeter of the hoods. The flange helps to increase
the net effective area of the hoods. The hoods are located 37 ft. above the
top of the ladle during tapping, and 30 ft. above the lip of each furnace.
The tundish hood is 18' x 22' at the face, with an attached 5 ft. wide flange,
and is located 15 feet above the tundish. The sides on all hoods have a mini-
mum slope of 45° from horizontal. The upper portions of the hoods and duct-
work were fabricated out of 10 gage mild steel, and the lower section
utilized 20 gage corrugated sheeting.
193
-------
10
MODULATING DAMPERS
EJ-j-S] - OLD EQUIPMENT
ijlji - NEW EQUIPMENT
I I - NEW DUCTWORK
FIGURE 3. GENERAL ARRANGEMENT OF NEW
EMISSION CONTROL SYSTEM.
LIMESTONE SILO
BAGHOUSE-NO. 2
-------
26-6"
ISOLATION DAMPER
vo
en
TAPPING DAMPER
^INTERNAL BAFFLESJ
-SIDE DRAFT HOOD
-PERIMETER FLANGE
FIGURE 4. NEW OVERHEAD FURNACE HOOD
-------
BAGHOUSE
In addition to the placement of four new compartments on each baghouse,
extensive work was done on the old compartments. The bag material finally
selected was a 5.7 oz/sq. yd. Nomex, with a smooth surface for good cake
release. Although many arc furnace systems use polyester bags, it was felt
that since this system had to operate at maximum capacity, the addition of
cooling dilution air would have adversely affected the long term performance.
A detailed examination of all of the isolation dampers revealed that
aside from dust build up on the seating surfaces, the dampers were operable.
It became evident that the electric actuators were not adequate. For these
reasons, new compressed air actuators were installed on all of the dampers.
All dampers except the fan inlet dampers and the two overhead modulating
dampers were retrofitted with pneumatic actuators.
. A limestone injection system was also installed. The system consists
of a 50 ton pressurized storage silo, connected to two variable speed feeders.
Pulverized limestone is injected at a rate of 80 to 120 Ibs/hr. to each bag-
house. Feed rate is adjusted according to type of scrap being melted, with
the higher rate corresponding to the melting of oily scrap. The limestone is
used primarily to slow the process of bag blinding, and to aid in cake release.
Also of prime importance is the fact that the limestone effectively changes
the size distribution of the dust, and tends to build a more "porous cake",
hence a lower pressure drop across the filter. The limestone lessens the
phenomena of bleeding by extending the effective filtration depth. The lime-
stone injection system is used to condition new bags. The injection system
is completely automatic, and is refilled from a pressurized tank truck.
A centralized control room was built to .house the compressed air system,
motor control center, and control panel. Considerable effort went into the
design of the control panel, which is shown in Figure 5. The main objective
of the control panel design was to assist maintenance personnel in system
operation, and to enable them to be able to detect short term malfunctions,
and long term problems. Aside from the usual manditory control and monitor
functions such as fan performance, and compartment cleaning, a series of
other functions are monitored or controlled. Every damper in the system was
equipped with a magnetic type position detector, and damper position is
indicated on the panel with color coded display lights. Every individual
compartment function is also monitored with indicator lights. Functions
include shaker motors, rotary air-lock drive, screw conveyor motors, and shake
and isolate time. Inlet gas temperature to each baghouse is monitored on
strip chart recorders. The set point controllers and monitors for side draft
hood static pressure are displayed in conjunction with modulating damper
position. Miscellaneous functions such as limestone injection feeders, lime-
stone silo fill conditions, and compressed air pressures are displayed.
System pressure drop across each baghouse is also monitored and continuously
recorded on strip charts. Shown in Figure 6 is an example of the trace from
baghouse No.2. System pressure recording allows the operators to adjust
cleaning cycle times to optimize equipment life. Pressure drop monitoring
also allows the operators to spot long term problems if they arise. Most
systems that do not record pressure drop only allow the operators to re-
spond to short term acute problems. Hence, no potential to develop a long
term maintenance program which is vital to successful baghouse operation.
"'196
-------
FIGURE 5o SYSTEM CONTROL PANEL
197
-------
10
00
^tfJtoxw ^(-WJWft, -sklisiwXt j, ^W^Wg-xja^jw^^^-ii^^.lfeAtoj^.
'-r' *-.?>
FIGURE 6. TYPICAL STRIP CHART RECORDING OF PRESSURE DROP ACROSS BAGHOUSE.
-------
Extensive walkways were added so that every damper in the system could
be easily serviced by maintenance personnel. The dust handling system was
rebuilt. New rotary locks and screw conveyors were added. In addition, a
common vented dust storage area for each baghouse was constructed. Pressure
taps were installed on each compartment, and a panel of pressure gages was
mounted in the control room. The entire system was completely rewired and
painted.
SCHEDULING
Shown in Figure 7 is an aerial view of the completed project. Baghouse
No. 1 is on the right, with the limestone storage silo, and the new control
room is located between the two baghouses. The overhead ductwork and hoods
are shown. The long duct run in the background is to the tundish area. The
torch cut area is at the base of the tall stack.
Construction and installation of the system was begun in 1978 and com-
pleted in April 1979. Critical path method scheduling was employed and the
project was completed on time. No production time was lost due to the in-
stallation. Two week long highly manpower intensive periods were employed
to install critical equipment. These periods were scheduled to coincide with
scheduled furnace maintenance periods. Work completed during these periods
consisted of removing all old bags and placement of new bags, completely
rebuilding shaker mechanisms, replaceing rotary locks and screw conveyors, and
completing final electrical hook up. Work was completed by a 30 to 40 man
crew. Upon completion of the installation, the system was started, and the
new bags were conditioned for 16 hours with limestone dust.
The entire project had been cost estimated during the feasibility study
period at approximately 1.2 million dollars. The project was completed with-
in the budget allocation. Approximately 50% of the total budget was for
materials, 40% for labor, and 10% for engineering and project management.
Performance and compliance tests have been done on the system. The
results show that baghouse No. 1 is running at 99% of design capacity of
130,000 CFM and baghouse No. 2 is running at 116% of its design capacity of
160,000 CFM. Visible emission observations by the local inspectors have
consistently shown that visible emissions from the mill have been eliminated.
Outlet grain loadings averaged 0.003 gr/DSCF for six tests, and mass emission
rates were 6.9 Ibs/hr. total from both baghouses.
SYSTEM UPDATE
Since the completed installation in 1979, several system/operations
modifications have been made. In September 1980, a scrap preheat system was
installed. The scrap preheater is used to preheat approximately 70% of the
charge material. Heat for the scrap preheater is supplied by the hot billets.
The scrap is preheated to a sufficiently high temperature to drive off oil
and rubber fumes. The exhaust gases from the scrap preheater are cooled and
then pass through a knockout chamber and .small filter section to remove con-
densed oil and rubber aerosols. The exhaust gases are then vented to bag-
house No. 1. The reduced oil and rubber fumes from the furnace melting
operations have decreased substantially the "stickiness" of the dust emissions
to the baghouses. Because of this, it is now possible to operate the system
without the limestone injection system.
199
-------
ro
o
o
FIGURE 7. AERIAL VIEW OF THE COMPLETED PROJECT.
-------
The absence of limestone in the baghouse dust has helped with dust dis-
posal. The dust is currently being shipped to a recycling operation where
zinc is recovered. To aid with dust handling, a centralized pelletizing
operation has also been installed. Dust from both baghouses is pelletized.
Current estimates of bag life are 16 months. Bags are currently being
replaced on a 16 month cycle even though bag failures are not severe for this
duration. Bags are being replaced at this interval to insure good system
operation and availability.
REFERENCES
1. Bender, Manfred, "Fume Hoods—Open Canopy Type. Their Ability to
Capture Pollutants in Various Environments," Air Pollution Control
Association Meeting, Toronto, Canada, June 1977.
2. Baker, Donald, and Barkdoll, Michael, "Retrofitting Emission Controls
on the Electric Arc Furnace Facility at Knoxville Iron Company," Iron
and Steel Engineer, Volume 58, No. 8, August 1981.
201
-------
THE PRESENT AND THE FUTURE FOR THE INDUSTRIAL
TREATMENT OF FUMES IN THE FRENCH STEEL INDUSTRY
Prepared by:
Jacques ANTOINE
Laboratoire d'Etude et de Controle
de 1'Environnement Siderurgique
METZ (F)
Alain MILHAU
Ministere de 1'Environnement
NEUILLY/Seine (F)
and
Jean RAGUIN
Chambre Syndicale de la Siderurgie Franchise
PARIS (F)
ABSTRACT
Over the last few years, the French Steel Industry has commit-
ted large sums and made an unprecedented technological effort to improve the
environment around its plants.
Thus, during the period between 1976-1980, 880 million francs
were spent in anti-pollution investments in spite of the unfavorable economic
situation.
Two thirds of these investments were devoted to the fight
against air pollution and, in several cases, new technologies were used.
This report covers the present situation in seven projects :
two coking plants ( Sollac and Pont-a-Mousson ), three oxygen steel-making
shops ( Fos, Mondeville and Neuves-Maisons ) and two electric steel-making
shops ( Firminy and Les Dunes ).
In each case, the authors analyse the problems to be solved on
a technical and regulatory level, the solution applied, especially when tech-
nologically unusual, and the results obtained.
From the present situation we can anticipate an evolution of
pollution standards and of anti-pollution technology.
202
-------
THE PRESENT AND THE FUTURE FOR THE INDUSTRIAL
TREATMENT OF FUMES IN THE FRENCH STEEL INDUSTRY
INTRODUCTION
In France, the Steel Industry is one of the key industries,
employing 120 000 people with an annual turnover of about 50 billion francs
for a steel production of 23.4 million tons in 1979.
A considerable effort of modernization is under way. This is
illustrated in Table 1 concerning the investments made between 1976 and 1980.
This modernization is especially related to the steel-making shops and continu-
ous castings. Many new installations have been started over the last few years.
TABLE 1. INVESTMENTS OF FRENCH STEEL INDUSTRY FOR
PRODUCTION OR FOR POLLUTION CONTROL
( 1976 - 1980 )
in 10 F For
Production
Coking plants )
Power plants )
Blast furnaces )
Charge preparation )
Steelmaking
Continuous Castings )
Rolling mills )
Other
TOTAL :
890
1 525
2 485
4 670
1 130
10 700
Pollution
Control
1 15
102
483
157
23
880
In spite of the problems encountered in investments because of
the present economic situation, the French Steel Industry has not sacrificed
environmental protection in and around its plants. Thus, during the 1976-1980
period, the fight against pollution represented 8.2 % of investment spending
for the entire profession, although the plan optimistically developed before
the energy crisis only allowed for 6.8 %.
Priority is above all given to the control of air pollution
( 65.5 % ), then water pollution ( 32.8 % ) and noise and wastes ( 1.7 % ). The
main sectors concerned are the steelshops, the rolling mills and the coking
plants.
To illustrate the exceptionnel effort to install air cleaning
equipment, a new integrated steelplant of 3 million tons/yr operates 25 dust-
separators with a flowrate of 5 million m3/h of fumes which needs an installed
power of 10 MV. In this report seven installations for fumes collection and
dust removal have been chosen in the three following sectors :
- Coking plants of SOLLAC ( Seremange ) and of PONT-A-MOUSSON
- Oxygen steelshops of SOLMER ( Fos ), USINOR ( Neuves-Maisons)
203
-------
and of METALLURGIQUE de NORMANDIE ( Mondeville ).
- Electric steelshop of CREUSOT LOIRE ( Firminy ) and CFAS
( Bunkerque-Les Dunes ).
THE SOLLAC COKING PLANT
SOLLAC started in 1979 at Seremange a battery of 64 coke ovens
of the " Compound " type which are capable of producing 600 000 t of coke per
year. The ovens are 6.2 m high, 13.7 m long and 0.43 m wide. They are charged
by gravity. The coke is unloaded by a mobil coke-car.
The Authorities imposed the following limitations :
- Fumes emitted during charging of the coal and unloading of
the coke must be collected and dedusted,
- The dust content in both cases must not exceed 10 mg/m N,
- Maintenance and surveillance of the installations must be
carefully handled in order to prevent any visible diffused emissions.
In order to respect these conditions, SOLLAC made an agreement
of cooperation with the Japanese company NKK and entrusted the study and the
construction of dust cleaning equipment to the French company NEU.
During charging, the fumes are collected by two independant
circuits : a convention exhaust circuit thanks to ammoniacal water injected
under high pressure ( 42 bar, flowrate,19 m3/h ) and a special exhaust circuit
near the charging machine connected to a fixed collector. This one was built
all along the battery on the unloading side. When an oven is charged, a teles-
copic pipe is positionned in front of the oven concerned which has a suction
hole in order to collect the fumes created by over pressure and by partial
distillation. Fumes are cleaned by a wet ventury-scrubber ( Table 2-A ).
During unloading, fumes are collected by a mobil hood located
at the top of the coke guide. Then they are channelled into a stationary collec-
tor and dust is removed in a first separator and in a suction baghouse. The
bags are woven with polyester fibers with a small proportion of stainless steel
fibers as a precaution against electrostatic charges ( Table 2-B )
The Authorities have requested an exhaust as air tight as possi-
ble for both cases. This goal has been obtained during charging, where fumes
are rarely visible. During unloading, the results are spectacular since nothing
escapes from the hood, given its dimensions and its 300 000 m3/h exhaust rate.
As far as dust content is concerned, the standard of 100 mg/m3 after cleaning
is respected with 80 mg/m3 for charging and 1 to 2 mg/m3 for unloading.
Over a period of 26 months, the installation was stopped only
one day due to an incident on the coke guide. Besides this, none of the bags
have been pierced and no condensation have been observed.
Since the results obtained were judged sufficiently successful,
204
-------
the Ministry of Environment discerned its annual prize on the SOLLAC coking
plant in 1979 in order to acknowledge their effort made against pollution.
TABLE 2 - FUMES TREATMENT AT THE SOLLAC COKING PLANT
A./CHARGING
Fumes collection
Charging cycle
Fume emission
Fumes dedusting
Flowrate
Initial dust concentration
Water flow in the saturator
Water flow in the scrubber
Pressure drop
Fumes exhaust
Centrifugal ventilator
Nominal flowrate
Motor power
Nominal pressure drop
7 min 25 s
3 min
21 000 m3N/h
5-15 g/ms
25 m3/h
36 m3/h
1 940 mm of Water
36 000 m'/h ( 530° C )
300 kW
2 480 mm of Water
B./ UNLOADING
Fumes dedusting
Fumes emission
Exhausted flowrate
Initial dust concentration
Pressure drop in the filter
Filtering surface
Filter type
Fumes exhaust
Centrifugal ventilator
Nominal flowrate
Motor power
2 min
300 000 in'/h
up to 10 mg/m2
80 mm of Water
5 150 m
polyester 320 g7m2
300 000 mz/h ( 100° C
300 000 m2/h ( 100° C
1 200 kW
THE PONT-A-MOUSSON COKE PLANT
The PONT-A-MOUSSON Company put a battery of 24 " Underjet "
coke-ovens into service in 1981. The furnaces are 4.5 m high. The coke is
unloaded and quenched by the process developed by the German company ERIN.
This new plant has to respect the following conditions :
- Collection of fumes during charging and unloading,
- Dust content of the fumes rejected to be less than 50 mg/m3,
- During coke quenching, efficient retention of the particles
205
-------
( less than 200 g/t of coke ),
- Careful maintenance and control of the installations in order
to prevent or to eliminate visible emissions.
The hot coke is gathered in a cubic tank which stays in a fixed
position during unloading. In order to avoid any notable pollution, the coke
guide is hooded and an 8 000 m3/h ventilator sucks up the fumes.
The dust is removed from the exhaust fumes by a cyclone and a
centrifugal scrubber with a water consumption of 150 1/t of coke. These sepa-
rators are put aboard the coke-car.
After unloading, the ERIN type case, without its lid, is taken
to a quenching stand. It is then covered by a second lid equipped with 360
water injection nozzles. Steam and fumes given off are sucked into the bottom
of the wagon which has a double bottom ( figure 1 ). Fumes are cleaned then
by cyclones.
Later on, when the capacity of the coking plant is doubled, the
installation will be equiped for the recovery of energy in the form of steam
at 330° C.
This new coke-plant has been operating' since the beginning of
1981. Controls have been made on the installation and it is still too soon to
tell, especially as certain modifications are being made, but the results ob-
tained at quenching already meet the stipulated limits.
THE SOLMER LD STEEL SHOP. HOT METAL TRANSFER
The SOLMER integrated steel plant has a steel-shop with two LD
converters of 280 tons.The supply of hot metal is assured by 450 ton torpedo
ladles. The content of these ladles is tapped into 270 ton ladles handled by
crane to charge the converters.
During the rapid pouring of the hot metal from one ladle to
another, colored fumes are given off. The Authorities have asked that these
fumes be collected and the dust removed in order to eliminate their charac-
teristic color. Besides this, the dust content of smoke given off into the
air was fixed at 120 mg/m3N.
The SOLMER company has conceived of and realized a removable
hood which closes off the area of smoke emission as tightly as possible at
the impact of the hot metal in the ladle. This hood overhangs the pit where
the ladle to be filled is located. It is made up of four removable parts :
two parts, on each side of the rails, which swing up to form the top of the
hood and two side-panels which are let down to close the hood. Fumes collec-
ted during the transfer at the rate of 200 000 m3N/h are cleaned first by
cyclone to eliminate the large incandescent particles, and then by filter
bags, furnished by the french company AIR INDUSTRIE.
The depression in the hood is such that no smoke can escape
206
-------
ro
o
Water
tank
»
Hi 10
Regulation j|
Heat Exchanger
( Steam production )
Cleaned
gases
FIGURE 1. THE " ERIN " COKE QUENCHING PROCESS AT PONT-A-MOUSSON
-------
during pouring of the hot metal. This is also true even when a side panel is
accidentally not in place. Since Fumes are cleaned by filtering, the dust con-
tent limit of 120 mg/m3 is respected with no problem. Adjustments had to be
made on the opening system for the largest piece of the hood, but now the ins-
tallation is quite satisfactory.
THE USINOR Q-BOP STEELSHOP AT NEUVES MAISONS
The USINOR company started a new steel plant with two 125 ton
converters in July 1979 at Neuves-Maisons. These converters use the Q-BOP
process and refine phosphorous hot metal. The Authorities enforced applica-
tion of the French regulations concerning red fumes in order to obtain color-
less fumes at all times and to respect the dust content limitation of
120 mg/m3N on the average during the blow.
All of the installation was built by LURGI.
Collecting converter gases during blowing is now a classic
operation. But it was done at Neuves-Maisons with partial combustion of the
gases (A= 0.3 ), an operation with suppressed combustion is foreseen in or-
der to recover and valorize these converter gases in the plant.
Neuves-Maisons is the first steel shop in the world to operate
dry electrostatic precipitatora with fumes from bottom-blown converters.
The three-field precipitators are installed on a platform outside. Each one
is a 16 m long cylinder having an inside diameter of 10.8 m ( figure 2 ). The
gas flow was studied to avoid turbulence and. thus a mixing of the different
flows alternately rich in air or in CO and H2 especially at the beginning
and at the end of each blowing cycle.
TABLE 3- CONTROL OF CONVERTER FUMES AT NEUVES MAISONS
Electrostatic precipitators
Number of fields 3
Secondary voltage 60 kV for field Nr. 1
35 kV for fields Nr.2 and 3
Secondary intensity 400mA
Fumes extraction
Nominal flow 340 000 m3/h
Nominal temperature 150° C
Motor power 250 kW
Since no installations can be totally protected from an acci-
dental explosion, the electrofilters are designed to resist an over pressure
of 2 bais and they are supplied with safety valves on the entry and exit cones.
However, the gases are continuously analysed and the safety systems are set
off when there is
CO > 6 % 02 > 4 % H2/CO > 0.14
or when CO > 9 % 02 > 4 %
The extraction ventilator is axial in order to avoid any risk
208
-------
26000
65000
r\>
o
29 500
FIGURE 2. THE NEW USINOR Q-BOP STEELSHOP AT NEUVES-MAISONS
1. Q-BOP converters of 125t 12. Electrostatic precipitators
-------
of destruction by allowing an easy displacement of a possible shock wave co-
ming from an explosion in the precipitators.
Till now, the results obtained for both collection and dust
control satisfy the standards set by Administration.
THE LD-AC STEELSHOP OF SMN-HOT METAL CHARGING EMISSION CONTROL
The new steelshop of the METALLURGIQUE de NORMANDIE (SMN) at
Mondeville has two 100 ton converters which were started in 1977. They are
top blown converters and treat high phosphorous hot metal. They are equipped
for primary fumes with a full combustion exhaust and a wet dedusting system.
The Authorities have imposed collection and cleaning of fumes
emitted during the charging of the converters with the following double goal :
residual dust content less than 120 mg/m3N and permanent absence of colored
fumes released into the atmosphere. This last constraint has imposed, in fact,
to respect a maximum dust content of 50 mg/m3.
When liquid hot metal is poured into a hot converter charged
with scrap, impressive red fumes come out of its mouth. Two reactions general-
ly come into play : on one hand, oxidation of the metal and emission of gra-
phite, as during the transfer of hot metal from one ladle to another; on the
other hand, the quasi-explosive combustion of part of the oils and greases
contained in the scrap already charged. At Mondeville, refining of high phos-
phorous hot metal is made in two phases and the end slag is recycled. It stays
in the converter between two cycles and, as it is highly oxidized, it reacts
with the carbon of the hot metal with increased fumes emission during charging.
These local conditions explain why it is particularly necessary in this shop
to collect and dedust these fumes.
To this end, the converter hood is doubled with a secondary
suction circuit which is used during charging. Flames and fumes are collected
by a frontal hood extended by chains. The actual exhaust rate of this hood is
about 25 m/s with a dynamic depression of 30 milimeters of water. The frontal
hood and the secondary duct are connected to the primary fumes circuit of each
converter which is highly favorable since collection during blowing is made
with full combustion ( figure 3 ).
All of the observations made show that the fumes are almost
completely collected. The intake of the frontal hood is such that any small
emission rising in front of the chain curtain is drawn through it. When the
reaction in the converter becomes too violent, the crane operators can reduce
the hot metal flow, which is normally 35 t/min in this shop.
Then, these fumes are dedusted with the " Kinpactor " venturi-
scrubber, which is also used for the primary fumes. This equipment was made by
AMERICAN AIR FILTER.
After adjustement of the ventilator and the scrubber, especial-
ly by reduction of the flow of water injected, the performances demanded by
the Authorities have been respected with a good reliability since four years.
210
-------
ro
FIGURE 3. SECONDARY FUMES COLLECTION SYSTEM AT MONDEVILLE STEELSHOP
A. Charging hood
B. Primary fumes hood
C. Tapping hood
-------
With a pressure drop of 1700-1800 mm of water in^the scrubber, the measured
dust contents are now in the range of 20-40 mg/m .
THE CREUSOT LOIRE ELECTRIC STEELSHOP AT FIRMINY
The Firminy steelshop operates 3 arc furnaces. The 35t furnace
which specialized in stainless steelmaking, was equipped with a dust con-
trol system in 1978. This furnace has the following characteristics :
- inner vesser diameter = 4.2 m
- transformer power = 17 MW
- 5 heats per day.
The following standards have been set by local Administration :
- efficiency of fumes collected so that, except on rare occa-
sions, no fumes are visible and a 90% minimal efficiency is required for fumes
collected in the hood;
- residual dust content of less than 10 mg/m% on the average
for a cycle, not taking into account the possible dilution in the cleaning
system for the whole steelshop. The dust content between 10 and 20 mg/m^N is
tolerated for only 28 days per year.
In any furnace producing special steels, it is necessary to
reduce the risks of oxidation of the charge by air intake. Because of this,
the operators gave up the idea of a fourth hole and chose instead a collection
by a canopy hood divided iitothree parts : charging, melting-refining and tap-
ping hoods are separate since, in contrast with the usual practice, the vessel
moves along the floor for charging ( figure 4 ).
Several measurements have been made by LEGES, the French orga-
nization specializing in pollution studies in the Steel Industry. It has been
shown that the 90% collection efficiency in the hood demanded by the Adminis-
tration has been respected and that it is even possible to save energy by mo-
dulating the exhaust rate depending on the fumes output ( 200,000 to
400,000 m /h ) . Table 4 shows, as an example, the results obtained on a heat
by controlling the dust balance.
TABLE 4. DUST BALANCE OF THE 35 TON ARC FURNACE AT FIRMINY .
Phase
Melting
Refining
Tapping
Collected
in baghouse
108 kg
300 kg
20 kg
Settling
in bay
2.5 kg
0.2 kg
1.0 kg
Released
outside
10.2 kg
18.5 kg
0.6 kg
Collection
yield
90 %
94 %
93 %
TOTAL
428 kg
3.7 kg
29.3 kg
93 %
Dust separator is a baghouse using the .FLAKT technique of
" superpulse " ( felt filters and pressure jet cleaning system ). The operators
212
-------
IX)
CO
FIGURE A. FUMES CLEANING INSTALLATION IN THE 35t EAF AT FIRMINY
1. Charging hood
2. Melting hood
3. Tapping hood
4. Fumes duct
5. Bag filter plant
6. Exhaust fan
7. Stack
8. Dampers
-------
had no trouble respecting the limit of 20 mg/rn^N. Simulated accidents have
shown that with two bags taken off, the dust content varies between 2 and
15 mg/m3. The installation is very satisfactory and the only adjustment con-
cerned the reduction of electrical consumption and noise insulation of the fu-
me s tack.
THE ELECTRIC STEELSHOP OF CFAS - LES DUNES
The works in Les Dunes near Dunkirk has installed a 80 ton UHP
arc furnace in 1978. It is located in an individual bay which is completely
closed. This furnace has the following characteristics :
- vessel diameter 5.8 m
- power 48 MVA
- nominal capacity 80 t
- 12 heats per day.
The local standards are the same as in Firminy ( no visible
emission, dust content of rejected fumes less than 20 mg/m3N.
The furnace is equipped with a classical fourth hole and with
canopy hoods. The roof being completely closed above the furnace, no fumes
can escape in the atmosphere. The results are thus excellent as far as the en-
vironment is concerned, but under certain unfavorable conditions fumes stagna-
te in the bay ( difficulty of exploitation of the regulation of the installa-
tion ) .
Primary and secondary fumes are mixed before cleaning. Dust
control is done by filtering on felt-bags using the FLAKT technique of " su-
perpulse " already presented for the Firminy baghouse ( Table 5 ). The fumes
are extracted by a 400,000 m3/h ventilator placed downhill from the baghouse
and they are released into the atmosphere through a stack which has an opaci-
meter installed on it for continuous control.
TABLE 5 - FUMES CLEANING INSTALLATION IN THE ELECTRIC
SHOPS AT FIRMINY OR LES DUNES
Exhaust
Nominal output
Power of motors
Cleaning equipment
Bag type
Bag dimensions
Filtering surface
Pressure drop in filter
Bag cleaning system
400,000 / 480,000 m3 /h ( 135" C )
2 x 400 kW ( Firminy )
4 x 250 kW ( Les dunes )
polyester-felt bags ( 0.5 kg/m2 )
0 0.3 m h = 5 m
2 690 m2
130 - 240 mm of water
sonic cleaning at low pressure ( 1.7 bar )
Periodicaly dust contents measurements and opacimetric controls
are made and they show that the standard of 20 mg/m3N has been strictly res-
pected except for incidents in the baghouse.
214
-------
The installation for the treatment and cleaning of fumes works
under harsh conditions since the UHP furnace has a high productivity. Twice,
violent reactions in the charge have caused the destruction of many of the
bags.
CONCLUSIONS AND PROSPECTS
We have just presented seven installations for air pollution
control now in operation in the French Steel Industry. After the necessary
operating adjustments are made, their results meet the standards set by the
Administration.
A considerable effort has been made in the last few years. HOT
wever as far as environmental protection is concerned, the seven examples given
are not truly representative of the entire Frernh Steel Industry, as
the list presented is not exhaustive. There also remain plants with much less
satisfactory pollution control equipment.
It must also be noted that we have focussed our attention on
the difficult problems linked to the collection and cleaning of fumes.
Generally, the governments will be more and more demanding in
the future and will impose strict controls of emissions into the atmosphere,
not only for dust but also for certain pollutants such as heavy metals, hydro-
carbons, gases and so on.
Other constraints than those imposed by the environment will be
more and more often imposed, for better health and working conditions or for
saving of raw materials and energy.
In expanding industries, it will be to the interest of the mana-
gers to change their methods of production and to adopt what has been called
11 clean technologies ", while associating automation with it. Such an evolu-
tion will probably be slower in the Steel Industry, given the present economic
situation and the large investments needed for construction of production
units.
For many years yet, a large part of production will thus be
assured by plants of the type which presently exist and the French examples,
which we have given, show that it is possible to successfully protect the en-
vironment .
215
-------
MODELING OF HOOD CONTROL OF BLAST FURNACE CASTING EMISSIONS
S. F. Fields, Manager, Mechanics Research
C. K. Krishnakumar, Research Engineer
J. B. Koh, Research Engineer
GARD, INC.
7449 N. Natchez Avenue
Niles, Illinois 60648
ABSTRACT
A laboratory scale model test technique has been developed to simulate
the thermal and flow characteristics of blast furnace casting emissions and
to evaluate the performance of potential hood collection devices. The test
system utilizes fresh water as the source fluid to model casting emissions
and concentrated sodium chloride solution to model the denser surrounding
environment. Excellent flow visualization is obtained through the generation
of light reflecting hydrogen bubbles in the source fluid stream by electrol-
ysis. Application of the technique to qualitative evaluation of hood
performance is illustrated by an example. In- addition, use of the technique
to generate numerical values of hood collection efficiencies is described.
INTRODUCTION
During tapping of molten iron from blast furnaces, emissions are
released due to exposure to the atmosphere. The magnitude of these emissions,
which are primarily fugitive in nature, can vary considerably from blast
furnace to blast furnace or between casts from the same furnace. Significant
efforts to curtail these emissions were initiated in the U.S. during the last
3-4 years. (For a review of the status of cast house control technology in
the U.S., Japan and Western Europe, see References 1, 2, 3).
Techniques that have been developed to control blast furnace casting
emissions include total evacuation systems, non-capture shrouding systems and
local hood systems. Total evacuation systems effecting 60 or more cast house
volume air changes per hour are expected to be effective. Due to the large
gas volumes handled, these systems are highly energy intensive. The non-
capture shrouding method is a proprietary technology developed by the Jones
and Laugh!in Steel Corporation in the fall of 1980. A system that has been
successfully used by the Japanese over the years consists of local hoods over
the primary emission sources (such as tap holes, troughs and spouts) and dome
shaped covers over runners. This primary capture system might be supple-
mented by a secondary capture system that controls emissions from the tap
hole operations. The refractory-lined hoods and covers are designed so as to
216
-------
permit easy removal by overhead cranes for maintenance and repairs. These
systems require large areas for the storage of spare parts and maintenance
equipment, laying of take-off ducts and the movement of cranes. The frequency
and the extent of maintenance required on the close fitting hoods and runner
covers will have to be established under U.S. furnace operating conditions
before implementing this technology in this country. Further, more than 90%
of the existing blast furnaces in the U.S. are single tap hole furnaces as
opposed to the multiple tap hole furnaces in Japan. This reduces the time
available for removal or repositioning of close fitting covers and hoods in
the area of the trough and tap hole between casts. The solution to this may
be the application of hoods with retractable curtains for capture of emis-
sions from this area. However, the effectiveness of such a system must be
established by tests before final design.
Hoods at or near the truss level have been attempted in a few North
American blast furnaces (References 1, 3). Performance of these hoods in
general have been unsatisfactory. If, with proper design and location of
these hoods and the use of deflectors or guide plates, satisfactory capture
of emissions from the major sources (tap hole, trough, skimmer and spouts)
can be achieved, then this system could prove to be the optimal choice for
retrofitting to many of the existing cast houses in the U.S. and possibly for
installation in some of the new cast houses. However, it should be noted
that the capture efficiency of hoods positioned several feet above the
surface of the molten metal could be very sensitive to their location and
design features such as hood skirting. Scale model tests that predict hood
performance and optimize hood design can provide valuable information
required to make decisions in cases of this sort.
The present paper describes a scale model test system based on the
principles of dimensional analysis to characterize casting emissions and to
evaluate the effectiveness of different hood capture systems. \
THE MODEL TEST SYSTEM AND TEST PROCEDURE
Model development was based on the primary scaling parameters applicable
to buoyant turbulent plumes. These are discussed in the following section.
Scaling considerations mandated the use of a liquid model as opposed to gas
models. Tap water with a trace of anhydrous sodium sulphate was selected as
the source fluid to model the casting emissions and concentrated brine solu-
tion was chosen to model the denser surrounding medium. Excellent flow
visualization is obtained by the generation of tiny hydrogen bubbles (less
than one hundredth of an inch in diameter) in the source stream by electrol-
ysis. The hydrogen bubbles reflect light beams from high intensity lamps so
that the plume flow can be observed and photographed.
Figure 1 shows a schematic of the laboratory test system. Source fluid
is pumped through the air eliminator and the hydrogen bubble generator into
.the lower chamber of the blast furnace model. As the fluid flows through
the bubble generator, hydrogen bubbles are generated by electrolysis. The
bubble-water mixture rises through controlled openings in the trough and
runner generating the model plumes. By properly scaling the source emission
• flow rates and the densities of the source and surrounding fluids, and by
217
-------
TO POSITIVE-DISPLACEMENT PUMP
t=J
_ROTAHETER
^SOURCE PUMP
FRESH WATER
SOURCE SOLUTION
FOR EMISSIONS
SIMULATION
200-GALLON TANK CONTAINING
BRINE SOLUTION
3-HAY BALL VALVE
-VENTURI METER
"^POSITIVE-DISPLACEMENT PUHP
Figure 1. Schematic of Laboratory Test System
adjusting the opening areas of the runners and the trough, the model plumes
were tuned to simulate the important characteristics of the full-scale
emissions. (During the course of the experiments, the specific gravity of
the surrounding fluid was constantly checked and corrected when necessary.)
To evaluate the performance of a hood, the variable speed P.O. pump was set
at the scaled volume flow rate evacuating fluid through the hood and dis-
charging to the brine reservoir or to drain. For each case, a qualitative
evaluation of the collection efficiency of a hood could be made by observing
and comparing the plume rise/spread phenomena with no hood and then with the
hood in position and the evacuation pump on and then off.
SCALING CONSIDERATIONS
The relevant parameters for modeling buoyant plumes are Rrashof number
(Gr), Reynolds number (Re) and Froude number (Fr). These are defined as
Gr =
Re =
g x 6 x
-------
p ~ p
oo S
Most plume problems are treated in terms of the densimetric Froude number,
For values of ps significantly different from pro, a more accurate expression
for Froude number is
U
Fr = . s _ _ , Reference 4
'g x D x ^ " PS
Scaling in terms of the densimetric Froude number will insure basic similarity
between plumes. In addition, it is necessary to ensure that model plume
Reynolds numbers are large enough to be in the turbulent regime since the
full-scale plumes are always turbulent.
FIELD OBSERVATIONS AND DERIVATION OF FULL-SCALE DATA
Full-scale data for this study was obtained by making field observations
of actual emissions from a local cast house (the No. 3 cast house at the East
Chicago plant of J & L Steel Corporation). Emissions from 4 casts were
observed over a 3-day period. The casts differed significantly in duration
and were performed under significantly different ambient conditions, so that
the observations covered a large range of possible emission characteristics.
The tasks involved in obtaining field data included motion photography of the
dust plume generated at the notch and along the runners, visual observations
Symbols.
D = Diameter or characteristic length; g = Acceleration due to gravity;
Q = Source emission rate or hood evacuation rate;
T = Absolute temperature; U = Flow velocity;
p = Specific gravity of fluid; v.= Kinematic viscosity;
B = Coefficient of volume expansion
Subscripts.
f •> full-scale; m -> model; s ->• source^fluid:
0° -> undisturbed surrounding fluid; s,f -*• source, full-scale;
s,m -> source, model; oo,m -> surrounding fluid, model
219
-------
of typical operating practice, and documentation of tons poured, start and
finish times, atmospheric conditions, existing cross-drafts during casting,
and other pertinent observations. The most significant task was motion
photography of the generated plume. The intent of this photography was to
document the expansion of the plume as it rises and the rate of rise and
bouyancy of the plume.
The motion photography was also supplemented by some still photography
which among other things, was used to document the nature and location of
building openings permitting air infiltration into the cast house. These
openings along with the other important geometric details of the cast house
were later incorporated in an approximate sense into the laboratory model
which was fabricated to a length scale of 1:45 (model: full-scale) using
clear Plexiglas sheets.
Emissions modeled in the present, study included those during initiation
of iron flow to No. 1 ladle, dumping (draining) of the trough to No. 1 ladle
and the lancing operation. Average values of fume rise velocities during
initiation of iron flow to ladle No. 1 and trough dumping were estimated to
be between 600 and 700 fpm from the analysis of the movie films and the field
observations. Rise velocities of lancing fumes were approximately half of
these values. Mean values of the effective source areas for emissions of the
former two events were also determined by analysis of the movie films and the
dimensions of the trough and runners obtained from shop drawings. Using
these data, the mean values of source emission rates for these two events
were calculated as 9000 acfm and 18,000 acfm respectively. Determination of
the effective source area for the emissions during lancing was more diffi-
cult. However, this was also roughly estimated by analyzing the movie films,
from which the emission rate for lancing was calculated as 4000 acfm. To
ensure that this value was conservative, a lower bound for the emissions was
also determined by an indirect calculation based on the oxygen consumption
during the lancing operation. This was obtained as ~ 2250 acfm. The value
of 4000 acfm calculated as explained above was therefore considered conserva-
tive and accurate enough for modeling purposes.
SELECTION AND DESIGN OF MODEL HOODS
From drawings of the full-scale cast house, full-scale hood dimensions
were determined such that the hoods appeared to be compatible with the
existing structure while providing (potentially) reasonable capture of the
emissions being considered. Locations of the three hoods inside the full-
scale cast house, for which model testing was performed, are illustrated in
Figures 2 through 5.
The large hood at truss level spans, longitudinally, two bays between
the steel trusses. It covers the major source emission areas associated with
trough dumping and iron flow to the No. 1 ladle. This hood was expected to
be efficient in capturing fumes from the main trough, the No. 1 ladle runner,
the runner used for trough dumping to the No. 1 ladle, and possibly the iron
notch. Since the emissions during trough dumping and iron flow to the ladles
peak only for very short durations, it was envisioned that this hood, with
its large plenum, would contain the fumes and provide satisfactory performance
220
-------
Figure 2. Position of Large Hood at Truss Level -Elevation View
Figure 3. Position of Large Hood at Truss Level - Plan View
221
-------
Figure 4. Positions of Small Hood at Truss Level - Elevation View
Figure 5. Positions of Low-Level Hood - Elevation View
222
-------
even at face velocities somewhat lower than conventional values.
The small hood at truss level has the same lateral dimension as the
large hood, but its length is equivalent to only one bay between adjacent
steel trusses. This hood was of interest because, with proper positioning
over the various emission sources and somewhat high face velocities, it was
expected to perform satisfactorily. This hood has a face area approximately
half that of the larger hood. It was tested in two locations, as shown in
Figure 4..
Dimensions of the low-level hood (at the level of the bustle pipe) are
severely constrained by the area available at this level near the furnace.
Without compromising too much on its location (which was considered critical),
the proper design appeared to be one with a face area of 13 feet x 13 feet.
This hood was tested in two positions, one right against the bustle pipe and
the other slightly offset from the bustle pipe (see Figure 5).
Take-off ducts for the model hoods were sized such that the scaled flow
velocities in the full-scale ducts fall within the conventional range, 3500
to 7000 fpm. The geometric configurations for the hood take-offs were not
otherwise modeled.
MODEL SYSTEM PARAMETERS
First, the maximum value of full-scale hood flow rate to be modeled was
determined by taking the product of the plenum area of the largest hood and a
reasonable value of face velocity (125 fpm) at the plenum. This was calcu-
lated as ~ 150,000 acfm. The corresponding value of model hood flow rate was
calculated as 18.8 gpm using a volume flow rate scale of 1:60,000 (model:
full-scale). The length scale of the model having been selected as 1:45 (as
cited earlier), the remaining model parameters were calculated as illustrated
in the following.
Velocity scale
U
s,m
Js,f
s,m
(a)
•s,m
Fpr equality of densimetric Froude numbers,
s,m
Ds,m x
Ds,f x
D ~ D
oo ^
_^+ Ps.
~ p - p *
oo S
u -r + PS -
m
f
iUs,f
900
(b)
Assuming ideal gas behavior,
P.-
(c)
223
-------
Calculation of Liquid Densities for Modeling Initiation of Iron Flow to No. 1
Ladle and Dumping of the Trough
Substituting reasonable values for TS and TOT (1460°R and 560°R
respectively) into equation (c), °°
= 0.698
2~ + "s J
From equation (b),
oo ^
L 2
-L
x x .698 = 0.035
m
Specific gravity of the model source fluid was determined as 1.005 for proper
concentration of sodium sulphate to give trouble free generation of hydrogen
bubbles. With this value of PSjm the specific gravity of the brine solution
in the tank p1 m was calculated'as 1.059.
roo,rn
Calculation of Liquid Densities for Modeling Lancing
Values of volume rate scale and length scale for lancing were kept the
same as for the other events. Therefore the velocity scale also remained
unchanged.
Km]
lancing
lancing
trough dumping
and
" P. - Ps'
L£ + °sJ
p ~ p
OO- C:
Pocr
m, lancing
m. trouah
trough dumping
lancing
= TT , from field data
trough dumping
With p = 1.005 and (p )
s'm
one obtains (p
)
'm trough dumping
= 1.018
= 1.059
m
'lancing
Calculation of Source Emission Rates
Table 1 gives values of full-scale source emission rates (as cited
earlier) and the corresponding values of model source flow rates based upon
the volume rate scale of 1:60,000.
224
-------
TABLE 1. SOURCE EMISSION FLOW RATES
Event
Lancing to open notch
Initiation of iron flow to ladle No. 1
Dumping of trough to ladle No. 1
Estimated Full -Scale
Source Emission Rate
acfm
4,000
9,000
18,000
Calculated Model
Source Flow Rate
gpm
0.5
1.2
2.2
Check for Similarity of Flow Regimes
For similarity of flow regimes, it is important to ensure that the model
plume Reynolds numbers close to the source are well above the critical value
of 300-600 (Reference 5). Approximate values of model plume Reynolds numbers
can be obtained as follows.
Us
x Velocity scale x D
g f x Length scale Ug f
1
Kinematic viscosity of model source fluid
s,m
Taking a conservative value of 3 feet for the characteristic length dimension
Ds,f
em, trough dumping
Rem, lancing * 105°
These values establish that the model plumes are well into the turbulent
regime.
Time Scale and Filming Rate
The time scale of modeling is defined as the ratio of the average time
of fume rise for the model to that of the full-scale system.
Time scale =
height of roof monitor
average fume rise velocity
height of roof monitor
average fume rise velocity
m
1.5
This implies that, for corresponding events of the model and full-scale
systems, fume rise velocities of the model would appear to be faster than
those of the model plumes by a factor of 3 to 2. Therefore, for visual
comparisons of the model and field movies, the model films should be
projected at two-thirds of the speed at which they were shot. (In this
study, model films were shot at 36 frames per second thereby requiring a
projection speed of 24 frames per second for visual comparisons.)
225
-------
TEST RESULTS
A summary of the model tests (which were also documented in a captioned
movie film) is given in Table 2.
OBSERVATIONS REGARDING MODEL HOOD PERFORMANCE
A brief summary of the observations for each of the hoods in its differ-
ent locations is given below (selected photographs taken during testing of
the hoods are shown in Figures 6 arid 7).
Low-level Hood, Offset From Bustle Pipe
This hood was intended for capture of lancing fumes. Even for the
relatively high face velocity used, performance of this hood in the offset
position is poor. Plume fluid excapes through the gap between the hood and
the bustle pipe and also around the bottom of the bustle pipe.
Low-level Hood, Against the Bustle Pipe
In this position, hood collection is much better than in the offset
position. Plume fluid reaching the immediate vicinity of the hood is
captured whereas that hitting the bustle pipe and spreading out underneath
it still escapes.
Small Hood at Truss Level, Located Close to Furnace
Performance of this hood for lancing is somewhat inferior to that of the
low-level hood in the position against the bustle pipe. For the other
events, large fractions of the major emissions miss the hood. However,
emissions from the main trough are captured.
Small Hood at Truss Level, Offset From the Furnace
This hood, in the offset position, was not filmed for collection of
lancing fumes as it was very ineffective in this case. During initiation of
iron flow to the No. 1 ladle, most of the emissions from the No. 1 ladle and
the runner leading to it are collected. However, emissions from the trough
area close to the furnace are not collected. During dumping of the trough to
the No. 1 ladle, a fair amount of the total emission is captured by the hood.
However, spill-over at the edges of the hood can be observed, indicating that
the hood area is insufficient to handle such large emission rates.
Large Hood at Truss Level
Performance of this hood during lancing is comparable to that of the
small hood at truss level, when located close to the furnace. Both for
initiation of iron flow to the No. 1 ladle and dumping of the trough to the
No. 1 ladle, collection efficiency appears to be good. However, a tendency
for the emissions to overflow the edges of the hood can be noticed at times.
This observation, together with the fact that the plume fluid beneath the
hood overflows and clouds the area in immediate response to the loss of hood
226
-------
TABLE 2. SUMMARY OF MODEL TESTS
Test
No.
Model Source
Flow Rate
gpm
Full-Scale Source
Emission Rate
acfm
Model Hood
Flow Rate
gpm
Full-Scale Hood Flow
Rate & Face Velocity
a c fm : fpm
0.5
LANCING TO OPEN NOTCH
No Canopy Hood Collection
4,000 0
0 : 0
0.5
Low-Level Hood
Offset from Bustle Pipe
4,000 6.5
52,000 : 308
0.5
Low-Level Hood
Against Bustle Pipe
4.000 6.5
52,000 : 308
0.5
Small Hood at Truss Level
Located Close to Furnace
4,000
18.4
147,200 : 229
0.5
Large Hood at Truss Level
4,000 18.4
147,200 : 122
1.2
IRON RUNNING TO NO. 1 LADLE
• No Canopy Hood Collection
9,600 0
0 : 0
1.2
Small Hood at Truss Level
Located Close to Furnace
9,600
18:4
147,200 : 229
1.2
Small Hood at Truss Level
Offset from Furnace
9,600 18.4
147,200 : 229
1.2
Large Hood at Truss Level
9,600 18.4
147,200 : 122
10
2.2
TROUGH DUMPED TO NO. 1 LADLE
• No Canopy Hood Collection
17,600 0
0 : 0
11
2.2
Small Hood at Truss Level
Located Close to Furnace
17,600
18.4
147,200 : 229
12
2.2
Small Hood at Truss Level
Offset from Furnace
17,600
18.4
147,200 : 229
13
2.2
Large Hood at Truss Level
17,600 18.4
147,200 : 122
227
-------
Figure 6. Large Hood at Truss Level, Iron Running to No. 1 Ladle
Simulated Hood Evacuation Rate: 147,000 acfm
Figure 7. Small Hood at Truss Level, Trough Dumped to No. 1 Ladle
Simulated Hood Evacuation Rate: 147,000 acfm
Hood Offset from Furnace
228
-------
evacuation, suggests that the hood plenum is probably filled to capacity in
these cases and a somewhat higher evacuation rate would therefore be desir-
able.
CONCLUSIONS AND RECOMMENDATIONS
The two truss-level hoods were evacuated at the same volume flow rate
(full-scale equivalent: 147,000 acfm) for purposes of comparison. Corre-
sponding full-scale hood face velocities were 122 fpm for the larger hood and
229 fpm for the smaller. The larger hood, in spite of its lower face veloc-
ity, gave the most promising overall performance in terms of control of all
emissions under consideration. The low-level hood at bustle-pipe level (for
control of lancing and main trough emissions only) did not perform entirely
satisfactorily for any of the cases tested and was found to be highly sensi-
tive to horizontal positioning relative to the bustle pipe. However, it does
offer the possibility of some control with a small hood and a small volume
flow rate. Its performance could most likely be improved by the addition of
hood skirting. In view of the severe restrictions on the dimensions and
location of this hood, it was tested at the relatively high full-scale equiva-
lent face velocity of 308 fpm (50,000 acfm). This value of the hood velocity
was arrived at from a set of preliminary tests. The preliminary tests showed
that the performance of this hood improved significantly when the hood face
velocity was raised from about 180 fpm to 308 fpm (full-scale equivalent
values). .However, only marginal improvement was observed when the hood face
velocity was further increased to the full-scale equivalent of 500 fpm. (In
these model tests, the gap between the bustle pipe and the furnace wall was
blocked off, thereby cutting off the passage of fumes through this area.
This would be necessary in the full-scale situation to achieve reasonably
efficient hood capture of lancing emissions and to a somewhat lesser extent
emissions from the main trough.)
The modeling technique described in this study can be used to select
optimal designs of hoods, their locations and evacuation rates. Although in
the present series of tests only qualitative analysis of hood performance was
done, the authors have formulated a test procedure to yield numerical values
of hood efficiencies using the same test setup. This involves the additional
determination of sodium sulphate concentration in the hood flow by chemical
analysis. This data can then be used, in conjunction with the known values
of source emission and hood volume flow rates and the sodium sulphate concen-
tration in the source fluid (the surrounding fluid contains essentially no
sodium sulphate), to determine numerical values for hood collection efficien-
cies, thus allowing a quantitative evaluation of hood performance. Chemical
analysis of the fluid collected by the hood can be done either by the gravi-
metric method (separating and weighing) or by nephelometry (measurement of
light scattering). Either method will yield sodium sulphate concentrations
with accuracies sufficient to differentiate hood efficiency variations of
less than 5%. With this capability, one can also accurately evaluate the
effects on hood efficiency of other factors such as hood skirting and hood
take-off configurations for a finer degree of system design optimization.
The present study did not take into account the effects of cross-drafts
induced by winds. However, outside air infiltration patterns into the
229
-------
building under calm conditions were roughly simulated by incorporating open-
ings in the cast house model. In cases, where significant cross-drafts are
present, their effects on .hood performance may also have to be evaluated by
model tests before finalization of system design. This can be done by
examining hood performance while recirculating the brine solution through wall
openings in the model cast house at flow rates that scale the air volume flow
rates through the full-scale openings.
\,
The laboratory model has at this point been fully developed and suffi-
ciently demonstrated in a sample application to justify its intended use in
the evaluation of potential designs for hooding systems to control blast
furnace casting emissions; In general, for each application the following
steps would be required:
1) Site visit and selective filming of full-scale emissions.
2) Derivation of full-scale data by analysis of the film, shop drawings
and other information.
3) Selection of scales and fabrication of model cast house.
4) Determination of potential full-scale hood dimensions and locations.
5) Design and fabrication of model hoods.
6) Calibration check of model system for proper simulation of emissions.
7) Performance of model tests to select optimal hood design, location and
evacuation rate.
8) Evaluation of the effect of cross-drafts on hood performance and
incorporation of design modifications if necessary.
ACKNOWLEDGEMENT
This scale model study was performed under subcontract 492-S with Pacific
Environmental Services, Inc. for Region 5 of the U.S. Environmental Protection
Agency. Mr. Larry Kertcher of Region 5 served as task manager for this
project.
REFERENCES
1. "Blast Furnace Cast House Control Technology and Recent Emissions Data",
P.O. Spawn, paper presented at the APCA specialty conference on Air
Pollution Control in the Iron & Steel Industry, Chicago, April 21-23, 1981.
2. "Basic Iron Blast Furnace Casthouse Emissions and Emission Control",
R. McCrillis, paper presented at the 71st annual meeting of the Air
Pollution Control Association at Houston, June 25-30, 1978.
3. "Blast Furnace Casthouse Control Technology Assessment", Environmental
Protection Technology Series No. PB 276999, November, 1977.
4. "On the Motion of Turbulent Thermals", Escudier and Maxworthy, J. Fluid
Mechanics, V. 61 - Pt. 3, 1973, pp. 541-552.
5. "Turbulent Plume in a Laminar Cross Flow", D. P. Hoult and J. C. Weil,
Atmospheric Environment, Pergamon Press, 1972, Vol. 6, pp. 513-531.
230
-------
Session 2: SOLID WASTE POLLUTION ABATEMENT
Chairman:
Thomas M. Barnes
Heckett
Butler, PA
Cochairman: Robert S. Kaplan
U.S. Bureau of Mines
Washington, DC
231
-------
RCRA REGULATORY CHANGES AND THE STEEL INDUSTRY
By: Stephen A. Lingle and William J. Kline
Hazardous and Industrial Waste Division
Office of Solid Waste
U.S. Environmental Protection Agency
Washington, DC 20460
ABSTRACT
The objectives of the Resource Conservation and Recovery Act (RCRA) in
the management of solid and hazardous wastes are to promote the protection
of human health and the environment and to conserve valuable material and
energy resources. This paper presents an update of the actions being taken
by the Environmental Protection Agency (EPA) in fulfilling the objectives
of RCRA with respect to hazardous waste management. It discusses: (1) the
status of regulations promulgated to manage hazardous wastes, and (2) RCRA
activities specifically associated with steel industry hazardous wastes.
Standards were published in May 1980, establishing a foundation for
the management of hazardous waste, and additional technical standards
necessary to permit treatment, storage, and disposal facilities were pub-
lished in January and February 1981.
However, some wastes listed as hazardous in May of 1980 have subse-
quently been delisted, and the effective dates of some of the January and
February standards have been deferred.
EPA has also been involved in the investigation of means to alleviate
the problem of spent pickle liquor generated by the steel industry via
increased reuse/recovery. With regard to the latter, EPA has recently
promulgated a conditional exemption from RCRA management standards for
spent pickle liquor use and is also currently considering a Section 6002
procurement guideline for the specific use of spent pickle liquor in waste-
water treatment as a phosphorous removal agent.
An overview of each of these EPA activities under RCRA is provided in
this paper.
232
-------
RCRA AND THE STEEL INDUSTRY
The objectives of the Resource Conservation and Recovery Act (RCRA)
in the management of solid and hazardous wastes are to promote the protec-
tion of human health and the environment and to conserve valuable material
and energy resources. This paper presents an update of the actions being
taken by the Environmental Protection Agency (EPA) in fulfilling the objec-
tives of RCRA with respect to hazardous waste management. It discusses:
(1) the status of regulations promulgated to manage hazardous wastes, and
(2) RCRA activities specifically associated with steel industry hazardous
wastes.
Structure of Subtitle C
Under Subtitle C of RCRA, EPA is required to establish a Federal
hazardous waste management system that involves "cradle-to-grave" control
of hazardous waste. Section 3001 of Subtitle C defines criteria for the
identification and listing of hazardous waste. Section 3002 and 3003 mandate
standards for generators and transporters, including a manifest system which
is designed to track the movement of hazardous waste from generators through
transporters to hazardous waste treatment storage and disposal facilities.
Section 3004 specifies that owners and operators of treatment, storage and
disposal facilities comply with standards that "may be necessary to protect
human health and the environment." All standards become effective six
months after their promulgation. Subtitle C requires in Section 3005 that
these standards be implemented through permits, issued by EPA or authorized
states. Recognizing that not all permits would be issued within six months
of promulgation of the Section 3004 standards, Congress created "interim
status" in Section 3005(e) of RCRA. Owners and operators of hazardous
233
-------
waste treatment, storage, and disposal facilities who qualify for interim
status will be treated as having been Issued a permit until EPA takes
final administrative action on their permit application. Finally, Section
3006 requires that the Agency establish guidelines for authorizing states
to carry out the RCRA Subtitle C program.
The Subtitle C mandates and associated regulatory subjects are sum-
marized below.
RCRA
Section
3001
3002
3003
3004
3005
3006
3010
Regula-
tions in CFR
Part 261
Part 262
Part 263
Part 264,
265,266,
and 267
Part 122
124, 125
Part 123
Subject of Regulation
Identification and Listing of Hazardous Wastes
Standards for Generators of Hazardous Wastes
Standards for Transportation of Hazardous Wastes
Standards for Owners and Operators of Hazardous
Waste - Treatment, Storage, and Disposal Facil-
ities
Permits for Treatment, Storage, and Disposal of
Hazardous Waste; Consolidated Permits
Guidelines for Authorized State Hazardous Waste
Program
Preliminary Notification of Hazardous Waste
Activity
Status of Subtitle C Rulemaking
In May of 1981 the Agency published Phase I of the Subtitle C regulations.
The Phase I regulations included: identification and listing of hazardous
waste (Part 261); a manifest system and other standards for generators and
transporters; (Parts 262 and 263); interim status standards and some general
administrative standards for treatment, storage and disposal facilities
(Part 264 and 265); permitting procedures (Parts 122 & 124); and guidelines
for authorized state hazardous waste programs (Part 123). The Phase I
234
-------
regulations established the foundation of the RCRA Subtitle C regulatory
program. However, they did not include the important technical standards
necessary to permit treatment, storage and disposal facilities.
The latter standards, representing Phase II of the Subtitle C regula-
i
tory program, were published in January and February of this year. They
are known as general standards and are codified in Parts 264 and 267. These
regulations are all currently undergoing an intensive regulatory review and
consequently the effective dates of some of these standards have been deferred.
More specifically, the status of the Phase II standards is as shown in Table 1.
The status presented in Table 1 is a general, rather than a detailed
or legal, statement of the status of the regulations, and is Intended only
as a general reference. It does not include roughly a dozen technical
amendments which have clarified or modified certain parts of the regulations.
It also does not include response to petitions for "delistlng" wastes, and
it does not reflect the aspects of these regulations which have been chal-
lenged in litigation proceedings, and which therefore could undergo change
through negotiated settlements or court action.
In addition to the regulatory development and review efforts, the
Agency is moving as rapidly as possible to authorize States to assume the
RCRA regulatory program. To date, 25 States have been authorized for Phase I
of the RCRA program.
Specific RCRA Activities Affecting the Iron and Steel Industry
Since the publication of the Initial hazardous waste management
regulations under RCRA in May 1980, the Agency has been involved in at
least two activities related to steel industry hazardous wastes:
1) the ongoing review of the need to list or delist specific steel
industry waste streams
235
-------
TABLE 1
Status of Phase II RCRA Regulations
Type of Facilities
Storage and treatment in
containers, tanks, and piles
(Part 264; Subparts I, J, L)
Incinerators
(Part 264; Subpart 0)
Storage and Treatment
in surface impoundments
(Part 264; Subpart K)
Land disposal in landfills,
surface impoundments, and
underground injection wells;
land treatment; groundwater
monitoring (Part 267)
Financial requirements
(Part 264; Subpart H)
Closure & Post Closure Plans
Date Published
Jan. 12, 1981
Jan. 23, 1981
Jan. 12, 1981
Feb. 13, 1981
Jan. 12, 1981
May 19, 1981
Effective Date
July 13, 1981
July 22, 1981-
(New Facilities)
July 22, 1981
(Existing Facil-
ities) *
July 13, 1981
(New Facilities)
July 13, 1981
(Existing Facil-
ities) J-
Aug. 13, 1981 2
April 13, 1982 3
May 19, 1982
On July 24, 1981 .the Agency announced its intention to initiate rulemaking
to suspend the effective date of these standards pending further evalua-
tion of their appropriateness for existing facilities.
These standards apply to new facilities only.
The Agency announced in the Federal Register on September 30, 1981 a
deferral of the effective date from October 13, 1981 to April 13, 1982.
236
-------
2) the investigation of means by which to alleviate the problem of
spent pickle liquor disposal via increased reuse/recovery
A. Steel Industry Hazardous Wastes Listing
In the May 19, and July 16, 1980 Federal Registers, six wastes asso-
ciated with the steel industry were listed as hazardous wastes by EPA.
The six specific wastes listed and their hazardous characteristic(s) are:
1) Ammonia still lime sludge from coking operations, (toxicity);
2) Decanter tank tar sludge from coking operations, (toxicity);
3) Emission control dust/sludge from the primary production of steel
in electric furnaces, (toxicity);
4) Sludge from the lime treatment of spent pickle liquor from steel
finishing operations, (toxicity);
5) Spent pickle liquor from steel finishing operations, (corrosivity,
toxicity).
6) Dewatered Air Pollution Control Scrubber Sludges from Coke Ovens
and Blast Furnaces, (toxicity).
A number of other wastes (from non-specific sources) which may be
associated with some steel industry operations were also listed as hazard-
ous. These include-such .wastes as. spent halogenated and non-hologenated
solvents, still bottoms from the recovery of these solvents, and various
sludges generated from metal heat treating and electroplating operations.
Based on further investigation of the characteristics of these wastes
and review of comments submitted regarding their hazardous waste listing,
EPA has made the following revisions to the original listing of hazardous
wastes associated with the steel industry:
- "Sludge from lime treatment of Spent Pickle Liquor from Steel
Finishing Operation - (K063)
EPA has decided to revise the regulatory approach for this waste by
deleting it from the hazardous waste list but still regulating it as
237 -
-------
hazardous since it is generated from the treatment of another listed
hazardous waste - spent pickle liquor - (K062). Delisting petitions will
be considered when it can be demonstrated that the concentration of lead
and chromium in EP extracts are significantly less than the maximum concen-
tration levels of 100 X Interim Primary Drinking Water Standards. The
Agency is also in the process of evaluating the feasibility of an Industry-
wide delisting petition for the waste submitted by the American Iron and
Steel Institute.
Spent Pickle Liquor from Steel Finishing Operations (K062)
EPA has exempted from the hazardous waste regulations any spent pickle
liquor which is intended for use in NPDES permitted wastewater treatment
facilities. Notice of this exemption appreared in the Federal Register of
September 8, 1981. Aside from this specific use of spent pickle liquor,
the disposal of this waste remains subject to the hazardous waste regulations.
Spent Non-Halogenated Solvents (F005)
In the listing of spent non-halogenated solvents and still bottoms
from the recovery of these solvents, t'vj solvents rujtVumol and methyl
isobutyl ketone were listed due to both toxic and ignitable characteristics.
Based upon further review, the Agency Ls no loa^ec listing either solvent
as a toxic waste, although both will continue to be listed as ignitable
wastes under Hazardous Waste No. F003 since both are highly ignitable.
- Wastewater Treatment Sludges from Electroplating Operations
(F006)"
EPA has received comments suggesting that a number of these processes
would not generate a hazardous waste because hazardous constituents, such
as cadmium, chromium, cyanides, and nickel, are not used, and thus would
not be expected to be present in the treatment sludges. Accepting these
238
-------
arguments, the Agency has modified the listing to exclude wastes generated
by several electroplating processes, including:
1) tin plating on carbon steel
2) zinc plating (segregated basis) on carbon steel
3) aluminum or zinc-aluminum plating on carbon steel
4) all cleaning/stripping associated with tin,
zinc, and aluminum plating on carbon steel
Emission Control Dust/Sludge from Electric Furnace Steel
Production (K061)
Due to some confusion as to the applicability of this listing, EPA
affirmed that the intent of the listing was to include only wastes from
primary steel production. Foundry electric furnace emission control dusts
and sludges are not included in this specific listing but are being
evaluated separately.
- Dewatered Air Pollution Control Scrubber Sludges from Coke Ovens
and Blast Furnaces (F016) ~
After assessing comments, the Agency believes that it overestimated
the amount and migratory potential of coraplexed cyanides in these wastes.
Consequently, this waste has been deleted from the hazardous waste list.
In addition to the above mentioned changes to the listings of steel
Industry hazardous waste, EPA is also currently undertaking a study to
comprehensively assess the characteristics of wastes associated with coke-
making processes. This study will enable EPA to determine the hazardous-
ness of these waste streams.
B. Promotion of Spent Pickle Liquor Recovery/Reuse
One of the goals of RCRA is to increase resource recovery and conser-
vation. Therefore, in addition to assuring environmentally sound management
practices for the disposal of solid wastes, it is also EPA's Intent to pro-
239
-------
mote the recovery/reuse of solid waste.
We believe that the environmentally safe recovery and reuse of spent
pickle liquor has been demonstrated and will be feasible in numerous situa-
tions. Specifically, in many cases the use of spent pickle liquor in waste-
water treatment operations for the removal of phosphorous represents an
acceptable waste management technique.
Although almost 1 billion gallons of spent pickle liquor are generated
each year, only perhaps 5% - 10% of this amount is reused or recovered for
this use. The reasons for this low recovery rate include: (1) previously
inexpensive disposal options, (2) the difficulty of marketing a wastu
product, (3) unwillingness of many wastewater treatment plant operators to
try a new product, (4) inability of municipalities to enter into long term
contracts, and (5) only relatively recent stringent requirements for phos-
phorous removal.
The treatment and disposal of spent pickle liquor has always been a
problem for the steel industry. In the past, various dumping alternatives
allowed for a relatively low cost means of spent pickle liquor disposal.
As a result, development and utilization of recovery and reuse techniques
have not been strongly pursued. Coramoa methods for management of the
spent pickle liquor included: untreated disposal (61%), on-site neutral-
ization (21%), contract hauler (11%), reuse or recovery (7%). Untreated
disposal includes (1) direct discharge into waterways, (2) deep well injec-
tion, and (3) disposal on a slag pile or similar dumping. Under RCRA,
these untreated disposal practices will be virtually halted, thereby neces-
sitating the development of an acceptable alternate management option for
the disposal of a major portion of spent pickle liquor generated.
240
-------
Neutralization, although costly, may provide a cost effective short term
alternative. Although acid recovery (in which the dissolved iron salts
are removed from the spent acid, which can then be reused) has been shown
to be the most favorable spent pickle liquor management alternative based
on resource recovery and environmental protection goals, its implementation
has been held back by the relatively large capital costs needed to construct
recovery facilities. This impediment should gradually be alleviated as a
result of continuing development of less expensive acid recovery units
adaptable to a wide range of pickling facilities.
Due to regulatory requirements for the disposal of hazardous wastes,
it is expected that the cost of disposing of spent pickle liquor will
increase significantly not only due to additional treatment requirements
but also to the decreasing number of acceptable disposal sites. Until
such time that acid recovery techniques are wholly feasible, EPA believes
that the direct use of spent pickle liquor for phosphorous removal in
wastewater treatment facilities will serve to:
(a) provide an effective, inexpensive alternative phosphorous
removal agent;
(b) serve as an appropriate waste management technique for spent
pickle liquor; and
(c) assist in the conservation of our natural resources by reducing
the need for those raw materials used in or as commercial chemi-
cals for phosphorous removal purposes.
241
-------
Spent Pickle Liquor for Phosphorous Removal
The removal of phosphorous from wastewaters Is considered essential
for the control of eutrophication* and the prevention of water quality
deterioration in many receiving streams. Perhaps the most outstanding
example of this need for phosphorous control is the Great Lakes Region
where the discharge of phosphorous into Lakes, is or will be, restricted
to 0.5 to 1.0 milligrams per liter in order to preclude further deteriora-
tion of the waterways and hopefully reverse the eutrophication condition.
Of the methods available for phosphorous removal, chemical precipita-
tion (coagulation) using aluminum, ferric iron, ferrous iron, and lime is
considered to be the most effective and economical. These chemical precip-
itation methods depend on the use of soluble salts of the metal to coagulate
the phosphorous. Iron, both ferrous and ferric, and aluminum are employed
as the sulfate and chloride salts. Spent pickle liquor, has the Iron For
phosphorous removal present as ferrous sulfate or ferrous chloride (de-
pending on whether the pickling operation uses sulfuric or hydrochloric
acid). This being the case, spent pickle liquor can in many cases substi-
tute for commercial chemicals.
Examples of the successful use of spent pickle liquor include its
applications in the wastewater treatment faci.li.tLes at Washington, DC,
Baltimore, Milwaukee, Detroit, and Roanoke, VA. These examples of spent
pickle liquor use in wastewater treatment facilities are but a sampling of
similar applications which demonstrate the effectiveness of waste pickle
* Eutrophication designates that state of a body of water in which the in-
crease of mineral and organic nutrients has rertu-;-.vl the dissolved oxygen,
thereby resulting in an environment that favors plant over animal life.
242
-------
liquor in removing phosphorous in municipal wastewater treatment facilities
that have traditionally utilized higher-priced commercial chemicals for
that purpose.
The feasibility of increased utilization of spent pickle liquor led
the Agency to initiate two specific actions to foster its increased use.
One of these actions is a conditional exemption for spent pickle liquor
from the management standards that apply to hazardous wastes. This exemp-
tion applies to facilities which reuse spent pickle liquor, to generators
and brokers who accumulate spent pickle liquor, and to those who transport
spent pickle liquor, where the spent pickle liquor is to be reused in
wastewater treatment at a facility holding a National Pollution Discharge
Elimination System (NPDES) permit.
The Subtitle C regulations previously exempted the actual reuse appli-
cation from regulation, because all ligitiraate reuse and recovery facilities
were exempted in Section 261.6 of the regulations. However, for listed
wastes, such as spent pickle liquor, a manifest and other generator and
transporter standards applied, and the storage at the reuse facility was
regulated. This discouraged the use of spent pickle liquor because waste-
water treatment facilities were reluctant to be publically identified as a
hazardous waste management facility. The exemption removes all of these
standards.
EPA believes that this exemption will not pose a substantial hazard
to human health or the environment because in this specific use of spent
pickle liquor the objectives of the hazardous waste regulations appear to
be achieved by current in-place controls. For instance, the spent pickle
liquor must be shipped in compliance with Department of Transportation
243
-------
hazardous materials regulations, and when received by the user facility,
will be stored in tanks comparable to those in which similarly corrosive
commercial chemicals (ferric chloride, alum) are kept.
In making the exemption, the Agency will thus reduce the regulatory
burden to those individuals who reuse spent pickle liquor in wastewater
treatment, encouraging its use, but without increasing the risk of harm
to human health and the environment.
In addition to making an exemption from the hazardous waste regulations
for that spent pickle liquor which is utilized at NPDES permitted wastewater
treatment facilities, EPA is also specifically promoting the increased use
of spent pickle liquor in such facilities as a phosphorous removing agent.
The particular avenue within RCRA for promoting this increased use of spent
pickle liquor is RCRA Section 6002.
Section 6002 of the Solid Waste Disposal Act, as amended by RCRA, 42
U.S.C. 6962, requires Federal, State, and local procuring agencies, using
appropriated Federal funds, to purchase items composed of the highest per-
centage of recovered materials practicable, given that reasonable levels of
competition, cost, availability, and technical performance are maintained.
EPA is required to prepare guidelines to assist procuring agencies in com-
plying with the requirements of Section 6002.
Although the potential for use of spent pickle liquor is somewhat
limited at Federal facilities, EPA believes that its use at these facilities
would offer a cost savings to the government with respect to the purchasing
of treatment chemicals, and would additionally provide increased opportuni-
ties for demonstration of the effectiveness of spent pickle liquor for
phosphorous removal. Hopefully, a "spinoff effect" of this spent pickle
244
-------
liquor utilization will be an increased awareness and willingness on the
part of non-Federal wastewater treatment facility operators to consider
the use of spent pickle liquor for phosphorous removal.
A RCRA Section 6002 guideline would designate spent pickle liquor used
in wastewater treatment operations for phosphorous removal as a product
for which procuring agencies must exercise affirmative procurement under
Section 6002 of RCRA and presents recommendations for carrying out the
requirements of Section 6002 regarding the use of spent pickle liquor in
wastewater treatment operations.
The requirements of Section 6002 apply to "procuring agencies". The
term "procuring agency" is defined in RCRA as "any Federal agency, or any
State agency or agency of a political subdivision of a State which is
using appropriated Federal funds for such procurement, or any person con-
tracting with any such agency with respect to work performed under such
contract".
One approach being considered for the guideline is to recommend that
any purchases of chemicals for phosphorous removal made with Federal funds,
either directly or indirectly, allow for spent pickle liquor to be bid as
an alternate material, unless it can be shown that the use of spent pickle
liquor is technically inappropriate for a particular wastewater treatment
application.
Whatever approach is chosen, it is EPA's hope that it would assist in
achieving a significant increase in the reuse and recovery of spent pickle
liquor, thus lessening the environmental and economic burdens its disposal
now poses.
245
-------
CHARACTERIZATION. RECOVERY AND RECYCLING
OF ELECTRIC ARC FURNACE DUST
by: N. H. Keyser, Consultant, 122 W. Walnut St.,
Hinsdale, IL 60521
J. R. Porter, A. J. Valentino, M. P. Harmer and
J. I. Goldstein, Lehigh University, Department
of Metallurgy and Materials Engineering,
Whitaker Lab #5, Bethlehem, PA 18015
ABSTRACT
Electric arc furnace dust samples have been obtained from a number of
steel companies chosen to represent the broad spectrum of steelmaking
practice. The samples have been chemically and structurally analyzed in the
bulk and individual particles have been characterized using analytical elec-
tron microscopy techniques. Procedures to extract elements from the dust
have been investigated by both physical and chemical methods. Magnetic sepa-
ration procedures show some promise for recovering zinc from high zinc bear-
ing dust. Other options for resource recovery are discussed including the
technologies which exist for processing high grade dust.
INTRODUCTION
Dusts and fumes from steelmaking in electric arc furnaces contain impu-
rities that are undesirable for recycling by the normal methods. The major
undesirable elements in the dust are zinc, lead, sulfur, phosphorus, sodium,
potassium and their compounds. Recycling or disposal of electric arc fur-
nace dust is further complicated by the fact that the dust is extremely fine
and difficult to handle. Moreover, the problem of disposal became more
acute when the dusts were declared a hazardous waste. Poor characterization
of electric furnace dust in the past has made the selection of recovery,
recycling or disposal options difficult. In addition the high cost of dis-
posal has generated new interest in methods for recovery or recycling.
The U.S. Department of Commerce through the Economic Development Admin-
istration has provided a grant to Lehigh University to characterize electric-
arc furnace dusts, reviewing the recovery/disposal options, and to estimate
the available resources represented in these dusts. This paper presents
some of the findings of the Lehigh study.
* Most particles analyzed were complex iron rich spinel oxides with varying
amounts of other elements in solid solution.
246
-------
CHARACTERIZATION OF ELECTRIC FURNACE DUSTS
INTRODUCTION
Thirty-three samples of electric-arc furnace dust were supplied by
twenty-five steel plants representing a wide variety of steel products. The
dust samples were carefully homogenized and split. Bulk chemical analyses
including the EP Toxicity Test were undertaken by the U.S. Bureau of Mines,
Avondale Research Center. The electron microscopy characterization of indi-
vidual particles of dust, by size and chemistry was undertaken at Lehigh
University. In addition x-ray diffraction and MHssbauer spectroscopy tech-
niques were used to establish the mineralogy of the samples.
One large representative sample of carbon steel dust and one large
representative sample of stainless steel dust were given a series of chemi-
cal dissolution or leaching tests by the U.S. Bureau of Mines, Rolla Research
Center, and physical separation tests by the U.S. Bureau of Mines, Twin
Cities Research Center.
Dust from stainless and high alloy steels differs from carbon steel
dust. Thus each type of dust represents a different set of problems rela-
tive to the disposal and recovery options. Table 1 illustrates the differ-
ences in bulk chemistry. As expected the alloying elements chromium, nickel
and molybdenum show up prominently in the stainless steel and specialty alloy
dust. The amounts of these elements are sufficient to make recovery a viable
activity. The copper level is high in the alloy dust because monel type
alloys are produced in one of the selected furnaces. Cadmium is observed in
large amounts as a result of reclaiming nickel from nickel-cadmium batteries.
High fluorine contents probably result from greater use of spar for slag
conditioning. Carbon steel dusts are richer in zinc and lead because of the
greater use of galvanized and other coated products in the melt. The quan-
tities of zinc involved in electric furnace dust, if recoverable, would rep-
resent a significant increase in the United States resources of zinc.
Most of the dusts failed the EP leachate toxicity tests primarily be-
cause of the high levels of cadmium, lead and hexavalent chromium. However
only stainless and high alloy steel dust contained excessive hexavalent
chromium. In general the cadmium and lead were highest when the amount of
purchased steel scrap was high. In the case of cadmium, 40 to 100 percent
of the cadmium in the dust was found in the leachant. For lead only 2 to 17
percent was transferred to the leachant. When chromium was present in sig-
nificant amounts, a relatively small percentage ended up in the leachant,
although the amount still exceeded the specified limits.
MINERAL COMPOSITION
Temperatures near the electric arc are sufficiently high to vaporize
various constituents of the charge including both volatile and non-volatile
metals. The processes of vaporization, condensation and oxidation gives rise
to the formation of complex fumes. The dominant mineral phases in all sam-
ples are oxides with the spinel structure and lattice parameters indicating
compositions close to magnetite ^6304.) or zinc ferrite (ZnFe204). Individual
247
-------
TABLE 1. CHEMICAL COMPOSITION OF DUST FROM ELECTRIC FURNACES
Carbon and Low-Alloy Steels
Element
Al
Ca
Cd
Cr
Cu
K
Mg
Mn
Mo
Na
Ni
Pb
Zn
FeTotal
Fe+3
Fe+2
Fe°
Cr4*
Si
Cl
F
P
wt7ow
0.25
4.19
0.05
0.22
0.23
0.66
1.68
3.29
0.02
0.99
0.04
2.02
18.3
31.3
29.7
1.46
0.09
<0.01
1.81
1.11
0.41
0.03
Range 7»
0.09-0.53
1.85-10.0
0.03-0.15
0.06-0.58
0.06-0.32
0.06-1.12
0.77-2.93
2.46-4.60
<0. 02-0. 08
0.29-2.31
0.01-0.12
1.09-3.81
11.12-26.9
24.9-46.9
20.5-42.8
<0. 01-3. 96
<0. 02-0. 34
<0. 01-0. 02
1.35-2.49
0.51-2.36
0.01-0.88
0.01-0.08
Stainless Steel and Specialty Alloys
Element
Al
Ca
Cd
Cr
Cu
K
Mg
Mn
Mo
Na
Ni
Pb
Zn
FeTotal
Fe+2
Cr"*
Si
Cl
F
P
wt7o*''<'
0.40
3.91
0.46
5.88
0.62
2.07
3.78
3.72
1.08
2.12
1.69
0.52
4.58
27.0
4.47
0.10
3.38
0.81
2.48
0.02
Range %
0.20-0.60
1.76-6.93
0.006-1.79
2.01-10.1
0.09-1.26
0.80-5.07
1.70-4.74
2.36-4.59
0.37-1.46
0.47-4.60
0.15-3.34
0.23-0.78
1.77-6.22
22.2-35.5
0.53-5.90
<0. 01-0. 17
2.54-3.92
0.47-1.17
1.36-4.83
0.01-0.04
*arithmetic averages of dust from seven plants.
**arithmetic averages of dust from four plants.
248
-------
particle characterization, however, demonstrated that many elements can sub-
stitute for the iron or zinc, in particular manganese and chromium.
"Calcium oxide was present as free oxide in many of the bulk dust samples.
Zinc oxide was present in samples containing large amounts of zinc. In all
cases zinc is present in the spinel phase. Small amounts of hematite (a-
FeZ&j) were found in most samples. An expanded lattice parameter and micro-
analysis indicated the presence of manganese in the hematite structure. Lead
compounds were not observed by x-ray diffraction but lead-rich particles were
seen in the SEM and TEM. Dissolution studies indicate that lead is present
primarily as the oxide and secondarily as a silicate and that cadmium was
probably present as an oxide.
The presence of calcium difluoride, potassium chloride and sodium chlo-
ride was indicated by x-ray diffraction. Minor amounts of metallic iron and
possibly magnesium oxide were also observed in most samples. In addition
small amounts of several unidentified structures were noted. Possibly these
unidentified structures were complex silicates. Minor amounts of graphite
were detected in most samples. Those samples having the most intense graph-
ite peaks also had a higher free carbon content as shown by chemical analysis.
The graphite particles were probably blasted from the graphite electrodes by
the arc and may have some role in the nucleation of fume.
PARTICLE CHARACTERIZATION
Individual particles of dust from all the thirty-three samples have been
examined using a scanning electron microscope (SEM) equipped with energy
dispersive x-ray analysis (EDS). In addition fractions of the dust, size
separated by centrifugation techniques and magnetically separated have also
been prepared and examined. The technique allows the size and shape of in-
dividual particles to be correlated with the chemical elements present in
the dust, with the restriction that low atomic number elements cannot be
detected. A further restriction is that sodium in the presence of zinc can-
not be detected and sulfur, molybdenum and lead have overlapping x-ray peaks.
Specimens were prepared for scanning electron microscopy by ultrasonic dis-
persal in isopropyl alcohol followed by deposition on (0.2 ^m pore size)
nuclepore filters. The filters were viewed directly after carbon deposition.
For transmission electron microscopy (TEM), the filter material was dissolved
in chloroform, leaving the dust particles attached to a carbon film. More
detailed studies have been performed on a small number of the samples using
a scanning electron microscope with an automatic image analysis system* and
a scanning transmission electron microscope (STEM) which allows for more
accurate chemical analysis.
A typical specimen of dust particles is shown in Figure 1. Chemical
analysis of individual particles by EDS showed most particles to be very
similar with iron being the major detectable constituent with lesser amounts
of manganese and calcium present. A few particles contained traces of sili-
* The authors wish to acknowledge the assistance of R. Lee of U.S. Steel
Research Labs in performing this work.
249
-------
Figure 1: Dust particles supported Figure 2: Large particle with outer
on nuclepore filter for analysis. shell iron rich.
Figure 3: Broken shell of large Figure 4: Agglomerated dust parti-
particle revealing inner structure. cles- A linear agglomerate is
arrowed.
250
-------
con and zinc. Of the particles analyzed, only one (arrowed) differed and
that particle was calcium rich, with iron a strong secondary peak. The bulk
chemical analysis for this sample showed only calcium (8.9%), manganese (5.37,)
and zinc (2.9%) to be present above the 2% level. No measurements were made
for silicon. X-ray diffraction of the bulk dust sample showed the major phase
to be spinel with a lattice parameter consistent with magnetite having zinc
and manganese substituting for iron. Microanalysis of small submicron parti-
cles confirmed that zinc and manganese were dissolved in the spinel lattice.
Individual particles tend to be in the range of 0.1 ^m to 1 ^m. Occa-
sionally larger particles are observed, as in Figure 2, which shows a parti-
cle 11 ^m in diameter, with smaller, sub-micron, particles attached to it.
Only the presence of Fe was detected for this particle. The number of very
large particles is small; however they may well constitute a major fraction
of the dust by weight. One explanation for the formation of such large par-
ticles is that after initial formation, smaller particles agglomerate. If
these particles are not swept out of the region of the furnace, further mate-
rial is deposited onto them during the melt. There is evidence to suggest
that volatile elements, such as zinc, are incorporated into the dust early in
a melt and that later on a higher proportion of the dust is iron. One expla-
nation for the high iron content of the large particle is that the outer layer
which was analyzed was not deposited until near the end of the melt cycle.
The internal structure of one such large particle is shown in Figure 3.
Microanalysis of the outer shell indicated iron with a subordinate, but high
chromium peak. Analysis of several internal particles showed some to be iron
rich, some calcium rich and some silicon rich. Only a small number of the
analyzed samples contained such large particles. No correlation is observed
between electric furnace dusts which contain large particles and the type of
steel produced, scrap used, and/or furnace size or dust collection system.
A number of particle agglomerates are shown in Figure 4. Typically,
agglomerates are made up of submicron particles, which, in general are
spherical, although faceted particles are also frequently observed. The
smaller agglomerates are often linear and presumably the particles are mag-
netically aligned. Some examples of linear agglomerates of particles are
shown, arrowed in Figure 4.
A TEM photograph is given in Figure 5 and shows that some of the spheri-
cal particles have internal faceted crystalline cores. Other faceted parti-
cles are also observed and electron diffraction techniques show these to be
spinel particles. Each of the particles in Figure 6, from a stainless steel
furnace, are chrome rich and contain a subordinate amount of iron. Also
present at lower levels were magnesium, zinc and silicon with trace amounts
of vanadium, aluminum, manganese, calcium, copper, nickel and potassium.
One of the major results of the microanalysis experiments is that a large
number of elements are present in individual dust particles. In many cases
these elements are present within a single spinel phase.
In many dust samples zinc oxide particles are very small and faceted,
indicating formation by homogeneous nucleation from the vapor (see Figure 7).
Typically such particles have up to 15% zinc substituted by such elements as
251
-------
Figure 5: TEM image revealing
internal structure of particles.
Figure 6: Faceted crystalline
region within particle.
Figure 7: Faceted zinc oxide
particles.
Figure 8: Backscattered image.
Light regions, arrowed, are lead
rich.
252
-------
iron, aluminum, chromium, copper, nickel and possibly silicon. Often, the
zinc oxide particles were part of a larger agglomerate, but still identifi-
able by diffraction and x-ray microanalysis. Some of the zinc oxide, how-
ever, formed into larger spherical particles, indistinguishable visually from
the spinel and calcium rich particles. Lead, like zinc, often occurs in
small particles within an agglomerate, or as small regions within a larger
particle. Backscattered electron images can be used to pinpoint lead within
an agglomerate (see Figure 8). The lighter regions in the figure are high
atomic number and are lead-rich.
Dust samples can be characterized by grouping together those individual
particles with the same one or two elements dominant in the EDS spectrum.
Figure 9 shows such a characterization for two of the dust samples, Figure 9a
representing a carbon steel dust and 9b a stainless steel dust. Each parti-
cle was assigned to one of the six classes or an unassigned class. For par-
ticles assigned to a single element class, the x-ray intensity detected for
the major element in the particle had to be greater than 2% times the inten-
sity of any other element. To be assigned to a two element class, the second
most intense x-ray peak observed had to exceed 40% of the intensity of the
major peak.
Particles in the single element, iron class could be from either the
magnetite or hematite phase. However, it is expected that hematite parti-
cles would fall into this class, since the solubility of other elements in
hematite is small. Particles in the single element zinc class were shown to
be zinc oxide and usually, only very low levels of solvent elements are de-
tected in such particles. The number of particles assigned to the single .,
element zinc class is probably too low, because the zinc oxide particles are
often too small to generate adequate x-ray peaks in the SEM. Their presence
was demonstrated more convincingly by STEM analysis.
The major difference in particle character between a carbon steel dust
and a stainless steel dust is in the two element categories. The carbon steel
dust had a large number of iron-zinc containing particles whereas the stain-
less steel dust revealed iron-silicon, iron-manganese and iron-chromium
associations. The iron-zinc particles are presumably spinel phase close to
zinc ferrite composition. Iron-manganese and iron-chromium could be either
the spinel phase or hematite phase, and probably both exist. The structure
of particles containing iron and silicon is complex and undetermined. In all
cases minor amounts of other elements exist in the particles. For example,
a typical particle in the iron-zinc category of the sample represented in
Figure 9a could also contain lesser amounts of manganese, calcium, silicon
and aluminum. Similarly, a particle assigned to the iron-chromium category
in Figure 9b could also show the presence of manganese, silicon, magnesium
and nickel. For both samples, the majority of the unassigned particles were
calcium rich, with some silicon rich and magnesium rich particles included.
The SEM automated image analysis technique was a valuable complement to
the characterization procedure since many particles could be analyzed in a
'short time, with the restriction that agglomerates were, in general, recog-
nized as single particles.' ' The results from one such analysis are shown
in Figure 10, for a carbon steel dust sample. The results are similar to
253
-------
50
o
-I—.
2.25
«4—
O
0
50
o
Fe Zn Fe Fe Fe Fe Other
Zn Si Mn Cr
0
Fe Zn Fe Fe Fe Fe Other
Zn Si Mn Cr .
a b
Figure 9: Histogram of classes of elemental associations into which parti-
cles fall, a) from a carbon steel dust, b) from a stainless
steel dust.
100 r
o
Q.
50
0
T
Fe Zn Fe.-Zn Fe:Mn Co Fe.-Ca Other
particle category
Figure 10: Histogram of classes of elemental associations for particles
in a carbon steel dust: results are from SEM automated image
analysis.
254
-------
Figure 9a, although a larger percentage of particles exist in the iron-zinc
category.
MAGNETIC SEPARATION STUDIES
The principle phases present in the dust, classified according to their
magnetic character, * are shown in Table 2. This table indicates that mag-
netic separation can, in principle, separate the dust into two categories; a
non-magnetic zinc rich fraction suitable for economic recovery and an iron
rich magnetic fraction suitable for recycling. Ideally, the most efficient
separation of zinc will occur when the zinc is present as either ZnO or as a
spinel of the type Fe/^_x^ Znx Fe2C>4 for 0.7
-------
TABLE 2. CLASSIFICATION OF PRINCIPAL DUST PHASES ACCORDING
TO THEIR MAGNETIC BEHAVIOR
Magnetic
Non Magnetic
Magnetite Fe^O,
Mixed ferrite FeQ_x\
0
-------
solid solution. Chemical analysis of the non-magnetic fraction showed an in-
crease in the total zinc content from 21.6 to 26.5 wt%.
It can be concluded that magnetic separation shows some promise as a
process step in the recovery and recycling of electric furnace dust.
RECOVERY/RECYCLE OPTIONS
Most iron and steel plant iron and alloy bearing wastes are recycled
through the sinter-plant blast-furnace complex. Steelmaking dusts from an
electric-arc furnace are an exception because they contain undesirable impu-
rities. Several interesting processes for recovering the inherent values
have been under study.'^) in many cases the proposed economic practicability
include credits by eliminating the high cost of disposal in commercial, regu-
lated, or monitored landfills.
A schematic of the options for treating electric-arc furnace dust is
shown in Figure 13. The following material summarizes some of the alterna-
tives available today and some possible new options.
Changes in furnace operation may inhibit the formation of complex zinc
containing spinels or the growth of particles which are not easily handled
once they form in the fume. Such changes in procedures may facilitate sepa-
ration techniques. Zinc and other volatile elements tend to be concentrated
in the dust from the early part of the heat. Changing the method of fume
collection could selectively remove a zinc concentrate. In addition magnetic
separation as discussed earlier might further concentrate the zinc from the
iron bearing dust. These options would require modification of the fume col-
lection system.
There is a strong possibility that recycle processes such as injection
or greenballing will concentrate the zinc/ lead, and cadmium to a sufficient
degree (over 20%) to make them attractive for thermal recovery. Specific
experiments are, needed to substantiate this possibility.
Arc furnace dusts from stainless and specialty Steelmaking are best
treated by recycle techniques such as in-plant greenballing or briquetting
with other alloy waste and recycling to the furnace,(3>4) or sending to a
centralized processing plant.^ ' The use of a better grade of scrap in the
manufacture of these highly alloyed products provides much less potential for
recovering the zinc, lead and cadmium.
Only in a few instances are the dusts, as produced, rich enough in the
non-ferrous metals zinc, lead and cadmium to justify thermal recovery. It
appears that in most dusts the recoverable elements need to be concentrated
or purified before they can be shipped and treated in a recovery unit. There
are limited opportunities for practical application of chemical solution
treatments of arc furnace dusts from a technical, economical, and environ-
mental standpoint. In carbon steel dusts, lead appears to be particularly
troublesome. None of the leach treatments attempted in this study were ef-
fective in selectively removing lead to a safe level. Cadmium is relatively
easy to remove to low levels. Of the chemical treatments available, only
257
-------
Scrap and
Charge Materials
o Injection
o Greenballing
Electric Furnace
o Scrap Treatment
o Scrap Separation
o Changes in Operation
Fume Collection
Fume
o Leaching
o Physical Separation
Techniques
Concentration/
Purification
o Direct Reduction Kiln
o Reduction Roast
o Waelz
o Plasma
Thermal Recovery
o Electrowinning
o Blast Furnace
o Slag Fuming
o
I
Fe
•
Non-Ferrous
Recovery
Zn, Pb, Cr
Concentrate
1
Electrothermal
Zn, Pb, Cr
Product
Figure 13: Schematic for recycling and recovery options.
258
-------
one or two appear to have any practical potential and none of these methods
are ready for commercial applications. Various standard physical methods of
separation were also attempted in this study on large quantities of furnace
dust, but none of the techniques proved very successful. The concentration/
purification option for furnace fume treatment still remains to be developed.
Chemical methods to extract zinc and lead from electric arc furnace
dusts have emphasized the thermal recovery step. In most cases the tramp
elements react with a suitable agent and transform volatile compounds which
separate from the solid phases via the gas phase and are subsequently col-
lected. The vaporization generally involves formation of volatile chlorides,
volatile metal or oxides. Most of the processes that have progressed beyond
the pilot stage consist of a direct reduction of the dust in a rotary kiln.
These processes are carried out in the temperature range of between 950 and
1100°C. The gas flow rate is an important factor in carrying the vapors out
of the bed into the gas stream where they are reoxidized. The product of
these processes is an enriched oxide product. At lower temperatures the
reduction and vaporization rates are too slow to be practical. At higher
temperature physical-mechanical factors such as ringing in the kiln are
limiting. In other processes, the temperature is raised to the range that
the iron product and the gangue or slag are molten. This has the advantage
of clean separation of slag from metal with the metal in a readily useable
form. However the high temperatures require a large amount of energy.
Selective reduction and vaporization of zinc and lead in certain gaseous
atmospheres which avoid reduction of iron oxides to metallic iron have been
proposed. Depending on the composition of the dust over half the fuel re-
quirement may be saved. A promising approach in today's technology is the
reduction of iron in one zone of the kiln then reoxidizing it in another zone
of the kiln in such a way that the heat of reduction is recovered in the
process.
Serious efforts have been made by zinc smelters to recover zinc, lead
and cadmium from steelmaking dusts. The contained zinc represents a signifi-
cant potential domestic source of the metal. A German Company regularly
operates a Waelz Kiln on electric arc furnace dust.(°' The Waelz Kiln is
used to concentrate zinc, lead and cadmium from iron bearing zinc ores. The
Wa"elz Kiln practice has been given a successful commercial trial in this
country by the New Jersey Zinc Company. The product of the Waelz Kiln must
be processed further to recover saleable products. A Japanese Company treats
electric arc furnace dust in an electrothermic zinc furnace to produce a
saleable zinc oxide product.''' A new type of electrothermic furnace using
plasma heat has been piloted in Sweden. Lead, zinc and iron were recovered
in metallic form electric furnace dust. A pilot electrolytic zinc plant had
promising results on high iron ores.(°' In the latter case, the dust was
first given a reduction roast to break up the zinc ferrites.
The supply of zinc, lead, and cadmium in useable concentration from dust
is variable. It would appear that the construction of new facilities to re-
cover these non-ferrous elements may be justified if performed in conjunction
with a direct reduction process to recover iron from a collection of iron
bearing wastes. Unfortunately, impurities such as arsenic, antimony, germa-
259
-------
nium, nickel, iron, copper, cobalt and selenium reduce current efficiency
and cause poor plating in electrolytic refining of zinc. Therefore removal
of these elements further complicates the recovery of resources from electric
arc furnace dust.
In summary there are a number of treatment options which can potentially
concentrate iron and non-ferrous zinc, lead and chromium. It appears that
zinc and magnetic concentration during fume collection and recycle options
such as injection and greenballing are possible options that need further
investigation.
ACKNOWLEDGEMENTS
The authors would like to thank the following people for valuable dis-
cussion and providing unpublished information: T. H. Weidner, T. M. Barnes,
P. L. Kern, J. A. Motto, R. 0. Wigger and P. Baumann. The authors would also
like to acknowledge the essential contributions to the project of S. L. Law,
L. A. Neumeier, D. M. Hopstock and J. A. Snyder of the U.S. Bureau of Mines,
C. B. Sclar and G. W. Simmons of Lehigh University and R. J. Lee and J. F.
Kelly of U.S. Steel Research Labs.
REFERENCES
1. R. J. Lee and J. F. Kelly, "Overview of SEM-based automated image
analysis," Scanning Electron Microscopy, pp. 303-310 (1980).
2. D. R. Fosnacht, "Recycling of Ferrous Steel Plant Fines, State-of-the-
Art," I & SM, pp. 22-28, April (1981).
3. "In-Plant Recycling of Stainless and Specialty Steel Wastes," Technology
News from the Bureau of Mines, No. 91, January (1981).
4. L. W. Higley, L. A. Neumeier, M. M. Fine and J. C. Hartman, "Stainless
Steel Waste Recovery System Perfected by Bureau of Mines Research,"
33 Metal Producing, pp. 57-59, November (1979).
5. J. K. Pargeter, "Operating Experience with the INMETCO Process for the
Recovery of Steelmaking Wastes," Proc. of the Seventh Mineral Waste
Utilization Symposium, October 20-21, pp. 118-124 (1980).
6. H. Maczek and R. Kola, "Recovery of Zinc and Lead from Electric-Furnace
Steelmaking Dust at Berzelius," J. Metals, 32,, 53 (1980).
7. T. Miyashita, "Recovery of Zinc from Steelmaking Dust," World Min. &
Met. Tech. Proc., Jt. MMIG-AIME Meeting, 2, 622 (1976).
8. W. W. Anderson, H. P. Rajcevic and W. R. Opie, "Pilot Plant Operation
of the Caustic Leads-Electrowin Zinc Process," TMS Paper Selection
A81-52 (1981).
260
-------
TREATMENT OF CARBON STEEL ELECTRIC FURNACE FLUE DUST BY SULFATION
by: Pinakin C. Chaubal
Former Graduate Student, Department of Metallurgical
Engineering, University of Missouri-Rolla
Rolla, Missouri 65401
Now a Graduate Student, Department of Metallurgy
and Metallurgical Engineering, University of Utah
Salt Lake City, Utah 85112
Thomas J. O'Keefe
Professor of Metallurgical Engineering and
Senior Research Investigator, Graduate Center
for Materials Research
University of Missouri-Rolla
Rolla, Missouri 65401
Arthur E. Morris
Professor of Metallurgical Engineering and
Research Associate in Materials Research
University of Missouri-Rolla
Rolla, Missouri 65401
ABSTRACT
The zinc content of electric furnace flue dust can be sulfated by the
action of iron sulfate in briquetted samples at 600-650°C. Up to 96% of the
contained zinc is converted to water leachable zinc sulfate under optimum
conditions. The other product is a relatively low zinc iron oxide. A
Plackett-Burman statistical design was used to quantify the factors affecting
the sulfation process. The mechanism of the process is explained by the use
of thermochemical diagrams for the Fe-Zn-S-0 system at different temperatures.
The results suggest a process for rendering the electric furnace flue dust
suitable for recycling into the steel plant flowsheet. The process may also
be effective in detoxifying dust which does not otherwise pass the EP toxi-
city test.
INTRODUCTION
Large quantities of oxide dust are generated during the making of steel.
The quantity of certain impurities in the dust, such as zinc, is dependent on
the fraction of galvanized scrap used in the charge. In some cases, contents
of zinc as high as 20% have been noted. The total zinc content of steel-
making furnace dusts has been calculated to equal 10% of the total mined zinc
261
-------
production in the U.S.1
Dust from iron and steelmaking operations has been normally dumped. Ap-
propriate storage sites are becoming scarcer and more distant from the point
of origin. Untreated dumps can lead to water pollution problems, especially
if the dust is classified as a toxic substance. The zinc and iron values of
these wastes remain unrecovered. This coupled with the high cost of dumping,
has led industry to look at recycling for relief.1'2'3
The desirability of recovering iron units from the dust makes the blast
furnace a logical point for recycling. However, zinc has been considered as
the cause for many troubles in blast furnace operations, including refractory
failure and scaffolding. More dramatic events like breakage of coolers, fail-
ure of the iron shell, complete filling of the gas offtakes and blocking of
the bell can occur.1*'5 Zinc recovery from the dust therefore has two advan-
tages. First, the zinc can be used to meet part of the requirements of in-
dustry and second, zinc removal may make the flue dust amenable, after
agglomeration, to recycling into the blast furnace.
Many methods have been proposed for flue dust recycling but only those
involving zinc removal are mentioned here. There appears to be no dearth of
published articles dealing with removal of zinc prior to recycling. Review
articles6'7 have described pyrometallurgical processes involving carbothermic
reduction and chloridization; hydrometallurgical methods involving selective .
leaching with NHi,Cl, HaOa, HaSOi, or organic reagents and some less successful
physical separation methods.
Heins8 has developed a process whereby zinc rich flue dust can be used
as an absorbent to reduce SOa emissions from industrial and public utility
furnace flue gases, the zinc oxide and the ferrite being the most active ab-
sorbing agents. Umetsu and Suzuki9 and Lenchev have shown that a mixture
of SOa and Oa in the presence of a catalyst can effectively sulfate ZnO and
(zinc ferrite). Gaprindashvili et^ jil_.ll have shown that ZnO and
can be almost completely sulfated when mixed with solid FeSOi, at
about 550°C.
In this paper a process is described for the removal of zinc from EFFD
(electric furnace flue dust) which incorporates a steelmaking waste product
derived from pickle liquor residue as the main ingredient. The process in-
volves low temperature sulfation of the zinc by the action of either ferric
or ferrous sulfate to produce soluble ZnSOi, and a relatively low zinc-iron
oxide residue. If need be, pickle liquor waste (FeSOi») can be oxidized to
Fe2(SOi,)3 prior to use, by a process described by Tiwari e_t al.12 Two steel
plant wastes, both difficult to dispose of, are thus converted into two
potentially useful products.
THEORY OF SULFATION
The sulfation process involves a reaction between zinc compounds in the
flue dust and the iron sulfates to produce ZnSOi* and FeaOs. In order for the
above to occur, ZnSOit and FeaOa should be able to coexist in equilibrium.
The Fe-Zn-S-0 thermochemical system can show whether a co-stability region
262
-------
exists for the two compounds. The diagram can also help in determining equil-
ibrium gas compositions. It has been shown that in the sulfation of zinc
oxide and ferrite (the predominant zinc phases in flue dusts8'13), S03 is the
sulfating species.9'11* The composition of the sulfation gas is hence impor-
tant in calculating the quantity of iron sulfate required to sulfate a given
amount of zinc.
Thermochemical diagrams for the Fe-Zn-S-0 system at 627°C (900 K) and
727°C (1000 K) are presented in Fig. 1. The data and reactions required for
their construction are given in Table 1.
TABLE 1. THERMODYNAMIC DATA FOR SULFATION REACTIONS AT 900 AND 1000 K.
NO.
1
2
3
4
5
6
7
8
(1
10
11
12
13
Fe2 (SO
Fe2 (SO
2FeSO,,
2FeSOu
I-ezO,
3Fe20,
SZnSO,.
3ZnSOu
REACTION
i.) 3 - Fe20j * 3SOi
,) 3 .. Fe20j + 3S02 + Ili02
- 0. 321Fe2(SO.,) 3 + 0.679Fe2Oj
+ S02 * 0.0369S03 + O.l>000402
+ S02 + 02 - Fc2(SOJ 3
* 2S02 * 3/202 - 2FeSOk
. 2Fe30, + >i02
- ZnO-2ZnSO. * SO,
- ZnO-2ZnS04 » S02 * !j02
3/2ZnFc2Ou * S02 * %02 - >sZnO-2ZnSO.
3ZnO *
3ZnO- F
ZnO +
ZnO- Fe
+ 3/2|:C20,
l-'ez (SO*) 3 - 3ZnSO^ * Fe203
e205 + Fez (SO.) 3 - 3ZnSOu
* 4 Fez 0 3
2l:eSOi. - ZnSO,. * I:c20j * S02
20, * 21-eSO. - ZnSO. * 2.|:e20-i
* S02
TEMP
T °K
900
1000
900
1000
900
1000
900
1000
900
1000
900
1000
900
1000
900
1000
900
1000
DOO
1000
' 900
1000
900
1000
900
1000
log Krx
-4
-1
-7
-2
-0
0
4
1
2
-0
-6
-4
-3
-1
-3
-2
4
3
6
5
5
5
2
2
1
2
.689
,465
. 164
.236
. 1896
.6747
.646
.928
.518
.463
.28
.837
.168
.8627
.978
.1197
.9796
. 1919
.501
.807
.94
.402
.087
.64
.9
-495
AH°rx at , DECOMPOSITION PRESSURE
T(K.I) (ATM) WHERE APPROPRIATE
560.
558.
854.
851.
150.
108.
-425.
-424.
-428.
-426.
252,
"244.
224.
224.
514.
517.
321.
361.
- 122.
-114.
-100.
- 84.
102.
104,
HO.
114.
15
85
22
7
71
06
72
83
S
87
76
56
82
82
7
67
24
83
9
18
21
49
79
91
36
81
pSO,
pSOj
pS02
pS02
pS02
pSOj
pS03
p02 =
p02 =
pSOj
pso3
pS02
pSOj
= 0
= 0
= 0
= 0
= 0
= 0
= 5
0
.0274
.325
.0322 P02
.401 P02
.741 p02
.0274
.03 p02
.1856
= 0.0161
= 0.2005
= 0.00003
= 0.002
7x10-' '
2.
= 0
- 0
= 0
= 0
1x10- ' 3
.00068
.0137
.0028 p02
.0487 p02
= 0.0014
= 0.02435
The data for constructing the diagrams are from the following sources:
ZnO, ZnS04 and ZnO-2ZnSOi, from an article by Kellogg and Ingraham, ! 5 ZnFe2Oi,
from Barin and Knacke16 and iron compounds from JANAF.17
DECOMPOSITION OF IRON SULFATES
According to the the rmo chemical diagram, Fe2 (80^)3 should decompose in
inert atmospheres or vacuum, to FeaOa, S0a» SOs and 02, the gas composition
being labeled "a" in the diagram. The ratio of S:0 in the gas is 1:3, which
makes it completely useful in the sulfation of the zinc oxide phases. The
decomposition temperature (temperature at which total decomposition gas press-
ure is one atmosphere) of Fe2(SOO3 was calculated to be 729°C (1002 K) by
plotting the total decomposition pressure as a function of temperature.
FeSOi, will initially decompose to Fe2(S04)3, to give an equilibrium gas
263
-------
-2
CO
CL
O -3
O
-4
(ZnS04+
FeS04)
ZnS04 +
Fe2(S04)3
Fe-Zn-S-0 SYSTEM AT 1000 K
phase boundary pgg (atm)
-9 -8 -7 -6
-4 -3 -2
-I
-2
M
O
CO
a.
O -3
O
-4
ZnFe204+
Fe-Zn-S-0 SYSTEM AT 900K
phase boundary Pso3 tatm)
-2 -I
LOG p02
Figure 1. Thermochemical Diagrams for the Fe-Zn-S-0 System
at 900 and 1000 K.
264
-------
composed mostly of S02, at the invariant point "b". S02 is of little use in
sulfating zinc oxide compounds. The decomposition temperature was calcula-
ted to be 639°C (912 K) . Once all the FeSOu has decomposed, the Fe2(SOi»)3
formed will decompose in the manner indicated previously. The secondary de-
composition gases in the above case arise from the decomposition of Fe2(SOO 3
and are therefore completely useful in the sulfation process.
SULFATION OF ZnO
In order to sulfate ZnO, the reaction gases must contain S02, 02 and SO3
at pressures in excess of that required for the thermal decomposition of
ZnSOi*. The first step in ZnSOi* decomposition is the formation of basic sul-
fate15 which is indicated by point "c" on the thermochemical diagrams (Fig. 1)
The decomposition temperature for ZnS04 was calculated to be 853°C (1126 K).
The diagram shows that at 627°C (900 K) the decomposition pressure for ZnSOi,
is over an order of magnitude less than the decomposition pressure for either
of the iron sulfates. Equations 7 and 8 of Table 1 show that the decomposi-
tion gases of ZnSOi* have an S:0 ratio of 1:3, which means that the most
effective sulfating gas must also have the same S:0 ratio.
From the previous discussion it would follow that when the system has
zero gas volume, one mole of Fe2(SOit)3 will sulfate three mols of ZnO/ZnFe2Oi4.
On the other hand two mols of FeSO^ will first decompose to produce 0.037
mols of SO3 gas capable of sulfating 0.037 mols of ZnSOi,. Subsequent decom-
position of 0.321 mols of Fe2 (SOiJa will sulfate 0.963 mols of ZnO/ZnFe^.
Therefore, two mols of FeSOi^ are capable of sulfating only 1 mole of ZnO/
ZnFe2Oi4. The major sulfating action of FeSOi, thus comes from the secondary
Fe2(SOi,)3 decomposition.
The thermochemical diagrams indicate a moderately sized region of co-
stability for ZnSOi, and Fe20s at oxygen pressures greater than 10~5 atm. At
727°C (1000 K) and oxygen pressures of 0.21 atm. such as those found when op-
erating in air, the S02 pressures required for ZnSOi* stability are about 0.01
atmosphere or more. The S02 pressure is an order of magnitude lower at 627°C
(900 K) than at 727°C (1000 K) for ZnSOi, stability and hence easier to main-
tain. ZnSOij stability increases at lower temperatures; the equilibrium SO3
pressures for the iron sulfates decrease and therefore the amount of sulfat-
ing species (S03) present in the gases is also lower. Thus a competition
exists between more favorable thermodynamics at lower temperatures, and higher
reaction 'rates and SO3 content at higher temperatures. The result is an
optimum sulfation temperature which must be determined by experiment. There-
fore, one of the objectives of this work was to determine the effect of temp-
erature on zinc sulfation (and recovery). However, other variables will also
affect zinc sulfation; temperature cannot be considered independent of these.
This situation requires care in experimental design, as will be discussed.
The presence of carbon or CO gas in the system would reduce the oxygen
partial pressure, meaning that a higher S02 partial pressure must be devel-
oped to keep the system in the Fe203 + ZnSOi, region. Thus the presence of
carbon is detrimental to the sulfation reaction and also aids in ZnSOi, de-
composition.
265
-------
The sulfation of ZnFe204 and ZnO is shown as equations 10 and 11 in
Table 1. AH° for both reactions is negative; therefore energy has to be
supplied only to heat the reactants to the reaction temperature and not for
the reaction itself. If FeSOi» is used, the primary decomposition to Fe2(SOi»)3
is very endothermic (AH° = +452.14kJ) as per reaction 3 of Table 1 at 627°C
(900 K) . The overall sulfation reaction (primary and secondary decomposition
of FeS04 and sulfation of zinc compounds) is exothermic.
The discussion has shown that when conditions at 627°C (900 .K) are com-
pared to those at 727°C (1000 K) , the lower temperature results in lower S03
pressures for co-stability of Fe203 and ZnSOi*. When Fe2(SOu)3 is used as a
sulfating agent, the best system is one in which the gas volume approaches
zero. A pellet made up of a mixture of flue dust and the sulfate is the
closest practical approximation to this condition.
If FeS04 is used, a good way to carry out the sulfation might be to heat
the mixture in a loose bed in air so that oxygen is available. Oxygen assists
the sulfation by reacting with S02 to form the sulfating species SO 3. The use
of a loose bed, however, results in some loss of S02 to the surroundings. Ex-
periments are thus necessary to determine if a pellet or a loose bed is better
for sulfation using
The discussion to this point indicates that Fe2 (SOOs has certain theor-
etical advantages over FeSOi^ as a sulfating agent. The source of Fe2 (S0i«) 3
should make no difference. It can be added as the solid sulfate, or be formed
in situ by adding sulfuric acid to the dust.
OTHER SULFATION REACTIONS
Many metal oxides other than zinc are expected to be sulfated by iron
sulfate. Thermodynamic calculations similar to the foregoing indicate that
the oxides of the following metals should be sulfated by iron sulfate: Cu,
Pb, Cd, Mn, Mg, Ca, K, Na, and Ni. Thus a calculation of the sulfation "de-
mand" by the dust involves the sulfation of these elements, as well as the
zinc.
EXPERIMENTATION
An industrially generated electric furnace flue dust (EFFD) sample was
used in this study. Ferrous sulfate in the form FeSOt,'7H20, obtained from a
pickle liquor plant reclamation process, and ferric sulfate as laboratory
grade Fe2 (SOO 3*xH20 were used as sulfating reagents.
SAMPLE CHARACTERIZATION
Characterization involved x-ray, surface area, DTA and TGA studies.
Leaching tests were used to determine the concentration of the zinc phases
in the EFFD. The gases from thermal decomposition of iron sulfates were
analyzed by an absorption method.
To determine the total zinc content, the sample was digested in 30%
H3POit solution and the resulting solution was analyzed for zinc on a Perkin-
266
-------
Elmer atomic absorption unit. The zinc as ZnO was determined by first leach-
ing the EFFD in dilute H2S04 (pH = 1.1), after which the solution was filtered
and the filtrate analyzed by AA.
DTA and TGA graphs were obtained up to 950°C in air and nitrogen on the
EFFD sample, ZnSQi,, and the two iron sulfates, using a Mettler apparatus. The
phases produced during decomposition of the iron sulfates after heating to
various temperatures for about half an hour were determined by x-ray analysis
of the products.
The sulfur content of EFFD was determined with a Leco analyzer. Other
elements were analyzed by the supplier of the dust sample.
EXPERIMENTAL STRATEGY
The factors affecting the sulfation process were studied with the help
of a Plackett-Burman screening design. The P-B screening design is a statis-
tical experimental design which aids in an organized approach towards the
collection and analysis of information.19
The values of the independent variables (factors) are set so that the
entire experimental region of interest is investigated. A large number of
experimental values is obtained by performing tests throughout the experi-
mental region, i.e., at high, low, and mean levels for each independent var-
iable. The P-B design permits a relatively large amount of information to
be gathered from a limited number of experiments since all variables are
changed simultaneously.
The data obtained from the tests are used to calculate the main effect
estimate which represents the change in response to the levels. The precision
of each main effect estimate is generally stated in the form of a confidence
interval, calculated at a fixed confidence level. A level of 90% is most
commonly accepted for industrial experiments and was the value used in this
study. An effect is said to be significant if the confidence interval does
not include zero. Only the significant effects are used in calculating an
expression for the response in terms of the factors studied.
In this study an eight run screening design was used.19 A total of five
factors was studied, which allowed for two dummy variables (to detect inter-
action). Three center point experiments were carried out. Each of the eight
design runs was duplicated to give an idea of the reproducibility of the meas-
ured response in percent zinc recovery. This made a total of nineteen exper-
imental runs.
EXPERIMENTAL PROCEDURE
The influence of certain variables thought to affect the sulfation pro-
cess was evaluated in a number of preliminary laboratory tests. Results of
these tests were used to select the P-B screening test conditions as shown
in Table 2.
267
-------
TABLE 2. INDEPENDENT VARIABLES AND THEIR RANGES
Temperature
Time
Stoichiometry
Low
600°C
30 min
2:1
High
650°C
50 min
3:1
Center Point
625°C
40 min
2.5:1
(mols of iron sulfate
per mol of zinc)
Pellet diameter 1.27 cm 2.54 cm 1.91 cm
Type of sulfate FeSOi, Fe2(SOi,)3 50:50*
*50:50 indicates that half the "SOi," added is contributed by
FeSOi, and the other half by Fe2(SOO3.
Iron sulfates were dried in air at 300°C before use. The dehydrated sul-
fates and the EFFD were crushed and the -65 mesh fraction was used in this
study. Required amounts of EFFD and the appropriate sulfate were weighed out
and mixed with a mortar and pestle. Enough mixture to form a 2.54 cm high
pellet, after pressing, was weighed and pressed at 155 MPa (22,500 psi) in a
Carver laboratory press. The pellets were then reacted in a tube furnace
under the required conditions. The time allowed for the reactions was meas-
ured from the instant the pellet was introduced into the furnace. A thermo-
couple was introduced into the center of a dummy pellet (1.27 cm diameter)
and the pellet at room temperature was introduced into the furnace. Four
minutes were required for the center of the sample to reach within 10° of the
furnace temperature (650°C). An air flow of 47.2 cm3/sec was maintained
through the furnace during the sulfation.
On the completion of the run, the sample was immediately air-cooled. The
cooled sample was weighed, crushed to -65 mesh, and leached overnight in dis-
tilled water. The solution was then filtered and the residue washed and dried
in air. The leach solution was diluted to give an expected zinc concentration
in the range of 2-4 ppm and analyzed for zinc by AA. The sulfur content of
leach residues was also determined.
The study of zinc sulfation as a function of time was performed under the
following conditions.: temperature, 650°C; Stoichiometry, 3:1; sulfate, Fe2
(SOt,) 3; and size, 1.27 cm diameter. Duplicate samples were prepared, sul-
fated and analyzed in the same manner mentioned earlier. Zinc recoveries
were measured after 5, 10, 20, 25, 30 and 45 minutes at 650°C.
EP TOXICITY TESTS ON AS-RECEIVED AND SULFATED DUST
Various carbon-steel EF dusts have been reported as exceeding allowable
limits of soluble lead and cadmium when tested according to published EP tox-
icity test procedures. However, in theory, the sulfated and water-leached
dusts should give much lower results for lead and cadmium than the as received
dusts for two reasons: cadmium sulfate is water-soluble and should tend to
follow the zinc, which is extensively removed; and lead sulfate is relatively
268
-------
insoluble and should not be leached. In order to test these hypotheses, EP
toxicity tests were run on two different EF dust samples. For purposes of
comparison, both dusts were tested for toxicity in the as-received condition,
and after sulfation and water leaching. The dusts were obtained from two
different steel producers, and both contained over 14% zinc.
Toxicity tests of the as-received and sulfated/water leached dust were
performed on 100 g samples, according to the EP toxicity test procedure listed
in Appendix II of the Federal Register, vol. 45, #98, May 19, 1980, p 33127.
In summary, the EP test procedure involves adding water and acetic acid to the
solid waste until the pH reaches 5, and agitating for 24 hours. The mixture
is filtered through an 0.45 micron filter, and the filtrate analyzed by AA.
If the solid waste has a pH of 5 or less after adding water, no acetic acid
is added.
Sulfation conditions were selected to provide maximum recovery of zinc
as ZnSOii, and were as follows: 600°C, 30 min, 1.91 cm pellet diameter, and 3
mols of SOi, added (as Fe2(SOt,)3). Following the sulfation process, the pell-
ets were crushed to -35 mesh, and leached in water overnight. The filtrate
was analyzed for Zn, so as to determine the %Zn recovery. The filter cake was
dried at 150°C, and crushed to -35 mesh (actually, the filter cake practically
falls apart in the mortar and pestle to a rather fine powder and a few small
lumps). The filter cake is then subjected to the EP toxicity test.
RESULTS AND DISCUSSION
CHARACTERIZATION OF THE EFFD
The chemical and x-ray analysis of the EFFD sample are presented in
Table 3. The major phases indicated are shown in Part B of the table.
TABLE.3._ CHARACTERIZATION OF EF DUST
A) CHEMICAL ANALYSIS
Total Fe CaO MgO
32.8% 3.3% 3.8%
K,0 S
1.2% 0.3% 0.18%
B) X-RAY ANALYSIS
Major phases: ZnO, FeaOa, FesOi,, possibly
C) ACID LEACHING TEST
Total Zinc 16.4%
Zn leached out in H2SOit 14.2
% Zn as ZnO 86.6
D) SURFACE AREA
As-received EFFD <0.26 m2/g
Dried at 125°C, 1 hr, air 2.94 m2/g
269
-------
The presence of ZnFeaOi, is difficult to detect with x-rays since the
pattern coincides with that of Fe30i». Dilute H2SOi, leaching tests indicate
that about 85% of the zinc is as ZnO. The rest of the zinc is probably pres-
ent as the ferrite. This is in agreement with the findings of Heins8 and
Fosnacht.l3
Surface area measurements on as received EFFD show a specific surface
area of less than 0.26 m2/gm. A larger surface area (2.94 m2/g) was obtained
after the sample was heated. This indicates the presence of capillary type
porosity -(which is sealed up when moisture is present).
THERMAL ANALYSIS AND DECOMPOSITION STUDIES
The DTA and TGA curves of EFFD showed that the flue dust was relatively
inert to air roasting. The absence of any appreciable weight loss indicated
that the C or other organics which could reduce the stability of ZnSOt, are not
present.
Three distinct regions are observed on the DTA curves of iron and zinc
gulfate, as presented in Fig. 2.
-120
-100
9 -80
jl
a -so
u.
LL.
LJ
I-
S -40
LU
a
-20
DTA FOR SULFATES
ZnS04 THjO, FeS04-7H20 and Fez(S04)3-
Conditions: Air flow 2.5 l/hr.
Heating rate IO°C/min.
Sample size 100mg.
Figure 2. Differential thermal analysis in air for iron
sulfates and zinc sulfate.
270
-------
First, from 100 to 300°C, the water of hydration is evolved, the last water
molecule being driven off around 300°C. In the second region ZnS04 and Fez
(S0i,)3 shown no changes, whereas FeSOi, gives a small exothermic peak around
500°C. In the third region all the sulfates decompose. The iron sulfates
decompose to Fe203 at a maximum rate near.700°C, although the effects are seen
to initiate at about 600°C. The ZnSO., shows an a-B transformation at 750°C
and a decomposition to ZnO-2ZnSOi» around 800°C. The DTA curves show that
FeSOi, decomposition depends on the gas medium used, but Fe2(SO.»)3 or ZnSO., do
not. The DTA curves indicate that the region of interest for sulfation is
between 600° and 750°C, where the iron sulfates decompose readily but ZnSOi,
is relatively stable. The preliminary tests to set design limits for P-B
screening were conducted in this range.
The cooled reaction products produced by decomposing the iron sulfates
at various temperatures were identified by x-rays. The phases detected helped
determine the decomposition sequence of the iron sulfates which is given in
Table 4.
TABLE 4. DECOMPOSITION SEQUENCE FOR IRON SULFATES
REACTION
1) FeSOi»-7H20 - AIR:
FeSOi»-7H20
4FeSO., + 1/202
(SO.,)
NITROGEN:
FeS04-7H20
2FeS04
3Fe202SOi,
Fe2(S04)3
FeSOu
Fe2(S04)3
+ Fe203 +
S02
Fe203 + 3S02**
+ 3/202
FeSO.,
Fe202SO.»*** + S02
Fe2(S04)3
+ 2Fe203
Fe203 + 3S02**
+ 3/202
TEMPERATURE
150°C
300°C
550°C
600°C
150°C
300 °C
300-500°C
550°C
600°C
2)
Fe2(S04)3*xH20:
Fe2(S04)3-xH20
Fe2(SOi,)3-H20
Fe2(SOO3
•> Fe2(SOO3'H20
-»• Fe2(S04)3
-»• Fe203 + 3S02**
+ 3/202
150°C
300° C
600°C
*Some S03 is present but in very small amounts.
**Moderate amounts of S03 are also expected to be present.
***This is a nonequilibrium compound and was identified by Gallagher et al.
The x-ray data to determine this phase were obtained from Skeaff and
2 1
Espelund.
20
271
-------
The DTA curves indicate that the thermal decomposition peak for Fe2(80^)3
produced from FeSOi, occurs at slightly lower temperatures than the reagent
grade Fe2(SOi»)3. This is expected due to the more reactive nature and smaller
particle size of nascent Fe2(SOi,)3, as compared to laboratory grade Fe^CSOOs.
S02 EMISSIONS DURING DECOMPOSITION
The results from S02 emission measurements are presented in Fig. 3 and 4.
Fig. 3 shows iron sulfate and EFFD/Fe2 (SOi,) 3 mixture decomposition curves in
air, and Fig. 4 represents iron sulfate decomposition in Ar. The air decom-
position curves for iron sulfates with and without EFFD are comparable, which
indicates that the presence of EFFD does not appreciably change the S02 emis-
sions from the sulfates.
The iron sulfate decomposition curves in air show a sharp decrease in S02
before leveling off. This drop is probably caused by conversion of S02 toSOs
by air oxidation. Freshly formed Fe20s is known to be a catalyst for SO3 for-
mation22 and hence results in a lowering of the S02 pressure. This is con-
firmed by similar curves in Ar, which show a higher S02 content and no sharp
decline prior to leveling off.
0.7
007 -
CONDITIONS'
Temperature '627 *C
Air flow rale' O.65 Ipm
Sample size: FeS04-7HzO • Ig
CONDITIONS'
Temperature' 627 °C
Argon flow rate' 0.65 Ipm
Sample size' Ig
Condition' Loose powder
2l5g EFFD +
2.95g Fe2(S04)j
10 20
TIME (mini
10 20
TIME (mint
30 Figure 4.
Argon decomposition curves
for iron sulfate.
Figure 3. Air decomposition curves
for iron sulfates and mix-
ture of EFFD and Fe2(SOi«)3.
x H20
272
-------
The effect of lack of 02 is more pronounced in the case of FeSOi,. In Ar,
* gives off much larger amounts of 862 during decomposition than in air;
the difference in the case of Fe2(SOil)3 is minimal. This shows that in the
presence of air the S02 was substantially oxidized to SO 3 thus increasing the
sulfation potential of FeSOij. If all the S02 produced can be oxidized then
the amount of zinc that sulfates could be doubled.
PLACKETT-BURMAN SCREENING TESTS AND RESULTS
The P-B screening test design is presented in Table 5. Table 6 contains
the results and calculations corresponding to the P-B tests.
TABLE 5. COMPLETE DESIGN WITH CENTER POINTS OF THE RUN
RUN NO. AFTER
RANDOM ORDERING 1
AND INCLUDING TEMP.
CENTER POINTS (° C)
I 625
II 600
III 650
IV 650
V 600
VI 625
VI I 600
VIII 650
IX 650
X 600
XI 625
' + ' refers
X2
TIME
(MINS)
40
50
30
50
50
40
30
30
-f
50
30
40
to the high
x .
* 4 x
3 STOICH. OF 5
TYPE OF SULFATE SIZE 6 7 ORIGINAL
SULFATE ADDED (INCHES) (DUMMY) RUN NO.
50:50 2.5:1 3/8 Center Point
FeS04 3:1 1/4 3
Fe2(S04)3 2:1 1/4 6
FeS04 2:1 1/4 4
Fe2(S04)3 2:1 1/2 7
50:50 2T5:1 3/8 Center Point
Fe2(S04)3 3:1 1/4 5
FeS04 3:1 1/2 2
+ + + + +
Fe2(S04)3 3:1 1/2 8
+ + +
FeS04 2:1 1/2 1
50:50 2.5:1 3/8 Center Point
level of the factors, and '-' the low level.
The confidence limit interval results show that at the 90% level, all
factors are significant. The dummy variable x7 is also significant, which
means that interaction between variables (like X!X2, or xsxi^) is a significant
possibility. More extensive experimental work is needed to construct a second
order model that takes into consideration the various interaction possibili-
ties.19
Standard statistical model-fitting procedures were used to construct an
equation relating zinc recovery to the experimental variables. The variables
used in the equation are listed on Table 7, and the resulting equation pre-
273
-------
sented at the bottom of the table. The value of S in the table and the model
represents the fraction of sulfur in the mixture of sulfates that is provided
by Fe2(SOi,)3. Although the % Zn recovered is probably significant only to two
significant figures, more figures are retained in the equation to improve in-
ternal consistency.
TABLE 6. P-B SCREENING TEST RESULTS
A) ZINC RECOVERIES
Run No. I II III
Zn Recovery, % 74.4 88.5 81.5
Run No. VIII IX X
Zn Recovery, % 96.6 93.4 71.9
B) CONFIDENCE INTERVAL CALCULATION
IV
58.8
XI
75.0
V
86.7
VI
73.8
VII
94.7
VARIABLE
xi (temperature)
X2 (time)
xa(type of sulfate)
xi» (stoichioiaetry)
xs(size)
Xe (dummy)
Xy (dummy)
MAIN EFFECT
ESTIMATE
-2.88
-4.33
10.11
18.56
6.28
-0.36
-8.13
CONFIDENCE INTERVAL
AT 90%
-1.89 to -3.88
-3.34 to -5.33
11.11 to 9.12
19.56 to 17.57
7.28 to 5.29
0.6325 to -1.3575
-7.1425 to -9.1325
TABLE 7. VARIABLES USED IN FIRST ORDER MODEL
VARIABLE
Temperature
Time
Sulfate
Stoichiometry
Size
NO. IN
DESIGN
Xl
x2
x3
x5
SYMBOL USED
IN MODEL
T
t
S
X
r
UNITS
°C
min
0 = FeSOi* 1 = Fe2(SO,,)3
Zn Recovery, % = 67.9 - 0.0578T -
pellet radius, cm
0.217t + 10.IS + 18.56X + 9.91r
The model predicts that maximum zinc recovery is obtained by operating
at lower temperatures, shorter times, larger diameter pellets, and use of
larger amounts of ferric sulfate. To test the accuracy of the model, the cal-
culated and experimental results were plotted as shown in Figure 5. The
center points give an idea of the amount of curvature, or deviation, from a
linear model. The linear model seems to fit the system well, since the center
point results are about 10% off, and the other points no more than 6% off.
274
-------
100
90
o>
I 80
IT
UJ
§ 70
UJ
CE
c
N
60
55
O3ZK
CENTER
POINTS
orz
55 -60 70 80 90
% Zn RECOVERY (calculated values)
100
Figure 5. Experimental values versus calculated values
from linear model.
The results generally conform to what would be expected from thermo-
dynamic and kinetic considerations. The stoichiometry of ferric sulfate
favors it as a sulfating agent over ferrous sulfate. The difference in de-
composition pressures between ferric sulfate and zinc sulfate is greater at
lower temperatures. Also, the shorter times and larger pellets tend to min-
imize thermal decomposition of zinc sulfate. Thus, the results of a series
of statistically-designed zinc sulfation and leaching tests on a typical EF
dust were shown to be amenable to a mathematical anaylsis which showed good
agreement with a linear model.
The sulfur content of the leached residues averaged 2.5%, and varied from
1.6 to 4.2%. The sulfur level was affected most by the type of sulfate, am-
ount of sulfate and size of the pellet. The effect of temperature was some-
what less significant. Further studies are required to minimize sulfur
levels in leach residue and to identify the sulfur-containing phases.
275
-------
TEST OF ZINC RECOVERY MODEL ON ANOTHER SAMPLE OF EF DUST
The model developed above was based entirely on experiments with the dust
described in Table 3. In order to test the general applicability of the model,
sulfation and zinc recovery tests were made on a second sample from the same
company. The two dust samples differed mainly in their zinc and calcium com-
position (first dust: 3.3% CaO, 16.4% Zn; second dust: 9.6% CaO, 14.7% Zn),
although most other elements showed some significant differences as well.
Pellets of both dusts were prepared according to the center point conditions
listed in Table 2 and sulfated together in the furnace. The results of the
tests are shown in Table 8. In addition, a larger scale center point sulfa-
tion test was conducted by placing 24 pellets in a combustion tube, and slowly
moving the tube through the hot zone of the furnace. The experiment was per-
formed twice. The resultant pellets were divided into four equal portions and
leached in a stagewise manner, so as to obtain a more concentrated solution
for zinc electrowinning tests. The results are shown in Table 9.
TABLE 8. CENTER POINT SULFATION TESTS OF TWO EF DUST SAMPLES
Dust #1
Sample No.
Zn Recovery, %
1
81.8
2
84.4
3
86.2
4
85.6
1
83.2
Dust #2
2
85.4
3
84.3
4
86.5
TABLE 9. LARGE SCALE CENTER POINT LEACHING TEST RESULTS WITH DUST #2
FOUR-STAGE COUNTER CURRENT LEACHING PROCEDURE
Batch No. 1 2 34
Zn Recovery, % 78.7 78.3 66.0 50.0
According to the model equation given at the bottom of Table 7, 84% zinc
recovery should have been obtained from all tests. The results in Table 8,
using the standard leach procedure, show that two different dusts, with sig-
nificant composition differences, behave about the same and are in good agree-
ment with the model. The results in Table 9 show that fairly good agreement
with the model is attained in the first two leach stages, but extraction in
subsequent stages with the same leach solution give successively poorer re-
sults. This is probably caused by the buildup of zinc in the filtrate, which
left an increasing amount of zinc in the filter cake (which was about 70%
solids).
Another larger-scale test was made on dust #2, for the following sulfa-
ting conditions: 600°C for 30 min, 3:1 stoichiometry with ferric sulfate and
1.91 cm diameter pellets. 95% of the zinc was recovered, as compared to 102%
predicted from the model equation. The agreement with experimental and pre-
dicted results is again within about 5%, as was found at other conditions.
276
-------
EP TOXICITY TESTS
The results of the EP toxicity test are expressed as ppm in the filtrate.
The filtrate was analyzed by AA, according to two methods: method of standard
addition (SA), and calibration curve method (CC). The results were usually
not the same. Both results are presented in the table below. Note that the
EP toxicity test does not specify a toxic limit for zinc, but we have analy-
zed the filtrate for zinc anyway, and reported the results in the table.
TABLE 10. EP TOXICITY TEST RESULTS ON AS-RECEIVED
AND SULFATED EF DUST SAMPLES
Sample
Ohio plant, as-rec'd.
Ohio plant, sulfated*
Illinois plant, as-rec'd.
Illinois plant, sulfated**
SA
2800
15
1590
36
Zn
CC
2500
14
1340
29
Pb
SA
84.3
3.8
103
3.7
CC
87.9
3.6
104
3.9
Cd
SA CC
6.8 6.5
0.16 0.16
11.6 11.8
3.5 3.7
*91% Zn recovery
**95% Zn recovery
The results of the as-received vs. sulfated dust samples show that the
sulfation procedure tends to fix lead as the relatively insoluble PbSO4, while
cadmium is generally sulfated to the relatively soluble CdS04, and tends to
follow zinc in that regard.
CONCLUSIONS
Zinc compounds in electric furnace flue dust can be readily sulfated by
iron sulfates in the temperature range 600° to 650°C. Almost 95% of the con-
tained zinc can be converted into water leachable ZnSOi*. A statistical
analysis of the results show that temperature, time of reaction, size of
pellet and amount and type of sulfate are all important in affecting the sul-
fation process. After agglomeration, the low-zinc residues may be recycled
to the blast furnace. If recycling is not feasible, the residue may be land-
filled under a less toxic classification than in the original condition.
ACKNOWLEDGEMENTS
The authors are grateful to Inland Steel Corporation for providing
chemical analyses, and to the National Steel Corporation for a Fellowship
and financial assistance. We also wish to acknowledge financial assistance
provided by a State Mining Institute grant allotment.
277
-------
REFERENCES
1. L.W. Higley and H.H. Fukubayashi, "Method for the Recovery of Zinc
and Lead from Electric Furnace Steelmaking Dusts," Jt. USBM-IITRI Fourth Min-
eral Waste Utilization Symposium, May 7-8, 1974, Chicago, Illinois.
2. W.M. Dressel, P.G. Barnard, and M.M. Fine, "Removal of Lead and Zinc
and the Production of Pre-reduced Pellets from Iron and Steelmaking Wastes,"
USDI, Bureau of Mines, R17927, 1974, Washington, D.C., U.S. Department of
Interior.
3. J.C. hogan, "Physical and Chemical Characterization of Refining Fur-
nace Flue Dusts," SECSI1, February 18-21, 1974, Japan.
4. R. Nicolle and W.K. Lu, "A review of the Behaviour of Zinc in Blast
Furnaces and Zinc Removal in the Preparatory Processes," Waste Oxides Re-
cycling in Steel Plants - Symposium Proceedings, 1974, McMaster University,
Ontario, Canada.
5. P. Crespin and J.M. Steiler, "Influence of Zinc and Alkali Metals on
Blast Furnace Operations," EUR No. 6275, 1979, Commission of the European
Communities, Boite Postale 1003, Luxembourg.
6. D.G. Brinn, "A Survey of the Published Literature Dealing with Steel
Industry In-Plant Fines and Their Recycling," British Steel Corp. Report PB-
236 359-6ST, August 1974.
7. A. Kitera, "Recycling and Re-use of By-Products," SECS12, June 11-13,
1979, Chicago, Illinois.
8. S.M. Heins, "Method for the Dry Removal of Sulfur Dioxide from Fur-
nace Flue, Coal and other Gases," U.S. Patent 3,983,218, Sept. 28, 1976.
9. Y. Umetsu and S. Suzuki, J. Min. Inst. Jpn., 1952, 68, 529-532.
10. S. Lenchev, Metallurgia (Sofia), 1976, 31(12), 21-3.
11. V.N. Gaprindashvili, R.M. Duduchava, and N.D. Mamageishvili, "Decom-
position of Zinc Ferrite," USSR Patent 382,716, May 23, 1973.
12. B.L. Tiwari, J. Kolbe, and H.W. Hayden, Jr., Met. Trans. B. 1979,
10B, 607-612.
13. D.R. Fosnacht, "Characterization and Utilization of Steel Plant
Fines," EPA Symposium—Iron and Steel Pollution Abatement Technology, October
30-November 1, 1979, Chicago, Illinois.
14. B.K. Dhindaw and S.C. Sircar, "Kinetics and Mechanism of Sulfation
of Zinc Oxide," Trans. TMS-AIME, Aug. 1968, 242, 1761.
15. H.H. Kellogg and T.R. Ingraham, Trans. TMS-AIME, Dec. 1963, 227,
1419-25.
16. I. Barin and 0. Knacke, "Thermochemical Properties of Inorganic
Substances," Verlag Stahleisen m.b.h. Dusseldorf, pp. 887.
17. D.R. Stull and H. Prophet, "JANAF Thermochemical Tables," Second
Edition, 1971, U.S. Department of Commerce, NSRDS-NBS 37.
18. S. Furumura and Y. Nishiziki, J. Min. Inst. Jpn., 1969, 85, 643-44.
19. T.D. Murphy, Jr., Chem. Eng., June 6, 1977, 168-182.
20. J.M. Skeaff and A.W. Espelund, Can. Metall. Q. 1973, Vol. 12(4),
445-54.
21. P. Gallagher, D. Johnson, and F. Schrey, J. Am. Ceram. Soc., 1970,
53, (12), 666-670.
22. V.A. Vanukov, A.V. Vanukov, and T.G. Toropova, Izv. Vyssh. Uchebn,
Zaved., Tsvetn. - Met., 1958, 3, 66-70.
278
-------
PANEL: DESTRUCTION OF HAZARDOUS WASTE IN IRON AND STEEL FURNACES
ENGINEERING REQUIREMENTS FOR THERMAL DESTRUCTION OF
HAZARDOUS WASTE IN HIGH TEMPERATURE INDUSTRIAL PROCESSES
By: E. Timothy Oppelt
U.S. Environmental Protection Agency
Incineration Research Branch
Cincinnati, OH 45268
ABSTRACT
A number of high temperature industrial processes offer temperature,
mixing and a gaseous residence time condition which may be sufficient to
thermally destroy hazardous wastes of various types. Industrial boilers
and cement kilns are most often considered for this purpose. Incineration
in lime kilns, steel making furnaces and other processes may also be tech-
nically feasible and economically attractive. Consideration must be given,
however, to the physical form of wastes, firing methods, residue handling,
air pollution control, "incinerability" of the waste and compatibility of
the waste material with the industrial process of interest.
SUITABILITY OF OPEN HEARTH FURNACES FOR
DESTRUCTION OF HAZARDOUS WASTE
By: William F. Kemner
PEDCo Environmental, Inc.
Cincinnati, OH 45246
ABSTRACT
The high temperatures and retention times achieved in open hearth
furnaces suggest their consideration as candidates for destruction of
hazardous wastes. Conceptually, liquid wastes could be co-fired with fuel
in the burners and semisolid and solid wastes could be charged into the
furnace. A particularly difficult problem in present hazardous waste
destruction is the disposal of contaminated drums. Complete cleaning is
expensive and time consuming. Destruction of contaminated drums and other
solid wastes is an opportunity not offered in many other destruction
schemes.
Serious complications coexist with the advantages of this concept.
Control of fugitive emissions, explosion hazards, and potential worker
exposure are among the problems to be evaluated. It appears unlikely that
hazardous waste destruction could be accommodated in an ongoing steel
279
-------
making operation. An alternative is the conversion of an abandoned furnace
to serve primarily as a destruction operation with the production of low-
grade steel as a by-product. Economics will dictate the ultimate viability
of such an approach as site cleanup activities proceed over the next five
years under the Superfund program.
SUITABILITY OF BLAST FURNACES FOR DESTRUCTION
OF HAZARDOUS WASTES
By: George R. St. Pierre
Ohio State University
Columbus, OH 43210
ABSTRACT
The potential for the injection of hazardous wastes in the tuyeres of
a blast furnace is discussed in terms of the thermochemical conditions that
exist in the raceways and smelting zone of an operating blast furnace. The
temperature in the tuyere zone of a blast furnace is approximately 2121°C
(3850°F). As the preheated air with additives enters the furnace through
the tuyeres, the chemical potential of oxygen rapidly decreases because of
the excess supply of hot coke. Hence, the conditions are characterized by
high-temperature (>2000°C), strong reducing potential PQ^ <10"20 atm), and
moderate pressure (<10 atm). Under such conditions, many hazardous ma-
terials may be dissociated into simple molecules. Currently, tars, oils,
and pulverized coals are injected in the tuyeres. Hence, a technology for
tuyere injection is well developed. During the presentation, the fate of
injected hazardous wastes, e.g., P.C.B., and the potential effects on blast
furnace operations are discussed.
280
-------
Session 3: WATER POLLUTION ABATEMENT
Chairman: Terry N. Oda
Region III
U.S. Environmental Protection Agency
Philadelphia, PA
281
-------
MINIMIZING RECYCLED WATER SLOWDOWN
FROM BLAST FURNACE GAS CLEANING SYSTEMS
By: Richard L. Nemeth and Leonard D. Wisniewski
Republic Steel Corporation
Cleveland, OH 44101
ABSTRACT
Retrofit water recycle systems were put in operation late in 1976 on the
gas cleaning systems for two blast furnace complexes located at Republic Steel
Corporation's Cleveland District steel making facility. Since the start up of
these systems, efforts have been made to reduce blowdown by identification and
e'limination of extraneous water sources. The overall effect of these efforts
has caused significant changes in constituent loadings from these systems to
the receiving stream, the Cuyahoga River. Recently, the goal of blowdown
reduction has shifted to the elimination of discharge to the river, with
efforts directed toward continued operation of a very "tight" recycle system.
Long-term effects are yet to be resolved.
The paper maps the development of a blast furnace treatment facility from
"once through" to "recycled" with a low blowdown rate. Current work
concerning the continuous operation of a tightly closed recycle system is
presented. This includes maintaining hydraulic balance, monitoring water
chemistry and feeding different treatment chemicals to control potential
problems associated with highly cycled water. It also presents a pictorial
display and schematic representation of the system, and a graphic analysis of
key parameters and water quality trends.
Finally, the paper focuses on studies conducted jointly by Republic
Steel and USEPA on one of the recycle water systems. This was accomplished
with the use of the USEPA Mobile Treatment System - Trailer No. 1 during the
spring and summer of 1981. Initial work included testing on a pilot plant
scale (5-6 gpm) of softening a portion of the recycle stream. Secondary work
investigated proposed BAT/BCT Alternate #4 _ (Alkaline-Chlorination) for
blowdown treatment from a very highly cycled blast furnace recycle system.
BACKGROUND
Republic Steel Corporation's Cleveland District is an integrated steel
making facility located along the Cuyahoga River in Cleveland, Ohio. Its
principal products are bar, hot band and cold roll strip. The facility
includes four blast furnaces, two coke plants, a two vessel B.O.F. shop,
billet mill, slabbing mill, bar mills, and a hot and cold rolling mill
282
-------
complex. Blast Furnace Nos. 5 and 6 are one of two separate basic iron making
complexes, the other being Blast Furnace Nos. 1 and 4.
Blast Furnace No. 5 was originally constructed as part of the Defense
Plant Corporation and was put in operation in 1943. No. 6 Blast Furnace was
added in 1952 as a sister furnace to complete the complex. Both furnaces have
similar statistics, with hearth diameters of 29'-6" and working volumes of
slightly over 56,000 cubic feet. Both produce basic iron and are burdened
primarily with pellets, some ore and roll scale depending on stock avail-
ability and clean water quenched coke. They each have a maximum wind rate of
125,000 SCFM, with combined production capability of approximately 5,600 tons
per day.
Original equipment gas cleaning systems were replaced and modified so
that each system now consists of a dry dust catcher, primary scrubber,
primary separator, secondary scrubber, flooded elbow and spray type gas
cooler. The original water treatment facility, built in 1943, consisted of
one 120' diameter thickener with related pumping and filtering equipment.
This thickener system was replaced in 1972 with the present water treatment
plant. The "once through" treatment plant was designed to treat water from
the gas cleaning system with the effluent discharged to the Cuyahoga River.
The treatment plant consists of a scalping pit, influent sump with three
low lift vertical slurry pumps, rapid mixing tank, two 90" diameter reactor-
clarifiers, three disc type vacuum filters for sludge dewatering, chemical
feed systems for lime and ferric chloride, and other ancillary pumping
systems. This plant effectively functioned as a once through system until
November, 1976 when the recycle portion of the facility was completed and
added to the operation.
The recycle system includes a pH adjustment tank, hot well with four
vertical slurry pumps, five cell spray type cooling tower, cold well with four
return water supply pumps, make-up sump with two vertical pumps, three
chemical feed systems, control and recorder conso-le, and other auxiliary
systems. At the blast furnaces, the recycled water supply was tied into the
existing gas cleaning system supply, replacing once used non-contact cooling
water (Cuyahoga River water) as the medium for gas cleaning and cooling. As
shown in Figures 1 and 2, the treated recycle water is fed to each gas cooler,
with a portion of the gas cooler effluent recycled to the primary and secon-
dary scrubbers. This contact water is then collected and directed through a
dirty water return trench to the treatment plant for solids removal, cooling
and recycle back to the gas coolers (Figure 3).
Each furnace has adjacent dual section slag pits. Slag is primarily air
cooled, with spray systems used for final quench prior to solidified slag
removal by a contractor. The quench system on No. 6 furnace is a
recirculating type with runoff collected in a sump, supplemented with make-
up, then resprayed on the hot slag (Figure 4). The final runoff volume at
conclusion of the quenching operation is blown down to the treatment plant.
Various effluents from No. 2 Powerhouse, which services this blast
furnace complex, are also directed to the treatment plant. These effluents
283
-------
GAS
TO
^TREATMENT
PLANT
Figure 1. Republic Steel Cleveland District - No. 5 Blast Furnace Gas
Cleaning System Flow Schematic.
GAS
PRIMARY
SEPARATOR
STOVES 8
SEALS
GAS
NO 2 COKE PLANT
LAGOON
TO
»TREATMENT
PLANT
NO 2 PH
ASH SYSTEM
SEALS
+RECYCLE
WATER
Figure 2. Republic Steel Cleveland District - No. 6 Blast Furnace Gas
Cleaning System Flow Schematic.
284
-------
I
I
I
SLUDGE
J
1
k.
w
EVAP a
DRIFT
Figure 3. Republic Steel Cleveland District - Flow Schematic for Nos. 5
and 6 Blast Furnace Gas Cleaning Water Recycle System.
TO
^TREATMENT
PLANT
RECYCLE
" WATER
Figure 4. Republic Steel Cleveland District
slag pit recirculation system.
- Typical blast furnace
285
-------
consist of miscellaneous seals and boiler drains and discharge from the boiler
fly ash collection and removal system.
SYSTEM HYDRAULICS
The recycle system supplies approximately 8,200 gpm to the blast furnace
gas cleaning systems. No. 5 gas cooler receives about 3,600 gpm and No. 6 gas
cooler approximately 4,000 gpm, with the balance pumped to primary separator
sprays and miscellaneous seals. These flows vary depending on furnace
operations, wind rate and recycle water temperature. Initial design of the
recycle system targeted a blowdown rate of 1,000-1,200 gpm, to be controlled
by conductivity of the system. In order to better control blowdown flows and
discharge loads, the conductivity control was converted to flow control of
blowdown with continuous monitoring of system conductivity.
It became apparent that system blowdown rate was dictated by hydraulic
imbalance rather than high conductivity. This condition, of hydraulic
imbalance in a recycle system is not at all uncommon, especially in retrofit
systems such as those on Republic's two blast furnace complexes. Applying the
methodology of a flow balance, where incoming and discharge flows are
compared, was the approach used to solve this hydraulic problem. Major or
obvious sources of incoming extraneous water were identified and plans were
developed to eliminate them. As investigations continued, identification of
other sources became more difficult, as did the solutions.
REDUCTION OF BLOWDOWN
Establishing the flow balance of the system became the tool for
illuminating various problems. Progress of these efforts was monitored by
comparing blowdown flow to the controlled make-up flow at the recycle system.
Blowdown flow was continously reduced to minimize the need for controlled
make-up.
In most cases, there was no easy way to identify the sources. As with
most operating units, the years tend to complicate an already intricate system
of piping and valving with cross connections, contingency back up and
surreptitious routing of piping throughout. Precautions were taken on the
piping system evaluation by monitoring line pressure to assure direction of
water flow at varous tie-in points. The use of dye, both visually and in
conjunction with a fluorimeter proved to be a valuable aid.
NO. 2 POWERHOUSE
The single major source of extraneous water into the recycle system was
from the fly ash removal and sluicing system for the boilers at No. 2
Powerhouse. This system had been discharging to the treatment plant for
suspended solids removal since the treatment plant was built. To eliminate
this as an extraneous water source, the ash sluice system was made a part of
the entire recycle loop by replacing service water supply with recycle water
supply (Figure 5). This approach of converting extraneous water sources with
286
-------
recycle water when they cannot be diverted from the treatment system became
the most important step in solving hydraulic imbalance problems.
When the ash sluicing system was tied into the recycle loop, recycle
water became available in the powerhouse. Various gas seals which drained to
the recycle system were converted to recycle water supply, as was a sump
eductor that had been supplied with city water. Tie-ins to existing
powerhouse piping became potential sources of water loss from the recycle
system. This problem was corrected by valve identification and tagging.
Boiler blowdown and zeolite softener backwash and rinse water, both high in
TO -
SEWER
MISC
SEALS
ASH SLUICE
SYSTEM
T
BOILER &
MISC DRAINS
H P SPRAYS
TO
BLAST FCES
j
4
? 1
« C*3— i
Q—
W '»
TREATMENT
PLANT
t
-
i 1
1 • fSPl
•> -i * ^A
RECYCLE
WATER
!f
1
SERVICE
WATER
Figure 5. Republic Steel Cleveland District
Schematic.
- No. 2 Powerhouse Flow
287
-------
dissolved solids, were diverted from the recycle system.
NO. 5 BLAST FURNACE
Service water ties to back up pumps for the recycle water supply were
identified and shut. Some cast house floor drains were also found to dis-
charge to the recycle system. Although this problem has not yet been
resolved, it is known to be an extraneous source. Recently, with the
deteriorating physical condition of this furnace as it concludes its present
campaign, turn down tuyere water to these drains has increased the hydraulic
imbalance in the system.
NO. 6 BLAST FURNACE
When this furnace was down for lining repairs in 1979, identified
extraneous water sources were eliminated. The alternate service water con-
nection to the recycle water supply was routed to the cast house area,
identified, tagged and shut. Indirect cooling water from a section of a
furnace cooling circuit was discharging to a drain at the cast house. This
drain was rerouted from the recycle system to the indirect cooling water
discharge sewer. The slag pit quench system became another extraneous source
when shutdown and drained to the recycle system. Service water make-up for
slag quench was replaced with recycle water to eliminate this source.
BLAST FURNACE STOVE VALVE COOLING
In the past, Republic Steel's Cleveland District Blast Furnaces have
experienced problems with valve cooling on the stoves due to the occasional
high silt loadings of Cuyahoga River water during and after rain storms.
Looking at the consistently high quality of treatment plant effluent con-
vinced the blast furnace operators that application of recycled water for
stove cooling would solve this silt problem. Unfortunately, simultaneous
tightening up of the recycle system changed the water chemistry by dramati-
cally increasing calcium carbonate scaling on the heat transfer surfaces.
After several piping changes and unsuccessful combinations of recycle and
service water on the stoves, with service water causing hydraulic imbalance,
the decision was made to return to service water for stove valve cooling water
discharge sewer.
TREATMENT PLANT
Several small sources of extraneous water were identified at the treat-
ment plant, all of which involved the use of city water for bearing lubri-
cation, packing gland seals or water ring compressor seals and cooling.
Consideration has been given to further treat recycle water as a city water
replacement for these applications. A potential water loss was the sump
overflow for the high pressure flushing pump which discharges to the river.
This accounted for occasional discrepancies in the hydraulic balance of the
system. The source of water for this sump is clarifier effluent, so casual
control of the sump level allowed for a recycle water loss at the overflow.
The permanent solution is relocation of the high pressure pump to the cooling
tower cold well. Until this is completed, monitoring of sump level while the
288
-------
pump is in use prevents the overflow and subsequent water loss.
The make-up butterfly control valve was found to be leaking due to a torn
seat. This problem was easily corrected by changing the seat material from
viton to teflon.
INSTRUMENTATION EFFECTS
The entire process of monitoring the hydraulic balance, essential in
reducing blowdown to any desired level, was dependent on information provided
by system instrumentation. The accuracy and reliability of flow monitoring
and control loops must be stressed. An incorrect flow or level indication can
cause false conclusions to be drawn about system status.
Flexibility of instrumentation provided tighter control and more
accurate analysis of the system hydraulics. Original blowdown control based
on conductivity measurement was changed to blowdown flow control to better
track the system hydraulics. Later, this flow control was "biased" by the
addition of cold well high level control which allowed for blowdown of only
extraneous water from the system. By set point adjustment of blowdown flow,
make-up control and "bias" blowdown based on cold well level, minimum make-up,
minimum blowdown and maximum surge capacity in the cooling tower wells were
established.
HYDRAULIC BALANCE
Other factors considered in establishing the hydraulic balance for this
system were water pick-up in gas cleaning and cooling from furnace operation,
water loss at the cooling tower and the concept that the recycle system is not
a terminal treatment plant.
Water pick-up at the furnaces is due to burden moisture, blast moisture,
reaction water and possible water infiltration due to leaks in furnace cooling
members. This water is condensed as the gas is cooled from furnace top
temperature (300-400° F) to gas cooler temperature (80-90° F), and at this
blast furnace complex is estimated to be 70-75 gpm per furnace.
Cooling tower losses due to evaporation and drift also vary due to
operating and atmospheric conditions and cooling tower design. Applying an
estimated cooling tower evaporation rate of 1% of recirculation rate per 10 F
of temperature differential, an estimated average of 250 gpm was established
for this system. The net effect of these factors yielded the ability of the
system to eliminate about 100 gpm of extraneous water.
Applying the rational that the recycle system is not a terminal treatment
plant, a potential extraneous source from the adjacent No. 2 Coke Plant was
averted. Blowdown from the gas scrubber system on a recently installed one
spot coke quench car was to be directed to the treatment plant for suspended
solids removal. By maintaining solids removal capability at the quench
'station, this blowdown was directed to coke quench.
As seen in Figure 6, blowdown flow has steadily decreased since initial
289
-------
SLOWDOWN FLOW, GAL/TON
4000T 1 1 1
REPUBLIC STEEL CORP
CLEVELAND DISTRICT
5 8 6 BL FCE RECYCLE
3000 -
2000
1000
500
400
300
200
100
I
75-76 77 78 79 80 79
YEAR
81
Figure 6. Republic Steel Cleveland District - Nos. 5 and 6 Blast Furnace
Recycle System - Slowdown flow by year.
C03
40
30
20
10
n
, Co, Fe (% OF
V ',
t
,
n^
16
1
1
P
1
1
1
1
Fll
Dl
58
i
1
1
I
P
1
|
EPOS IT
n /
j
)
50
1
1
1
y/.
'%
•f'
i
1
1
CZl CARBONATE (C03)
- ^CALCIUM (Co)
^3 IRON (Fe)
i
i
!
i
&
1
I
i
AUG
1978
JULY
1979
MAY
1980
MAY
1981
JUNE
1981
Figure 7. Republic Steel Cleveland District - Nos. 5 and 6 Blast Furnace
Recycle System - Deposit analyses history.
290
-------
start up of the recycle system in late 1976. Recent efforts of operating
hydraulically at zero blowdown have been achieved only for short periods of
time because of known extraneous water from No. 5 Blast Furnace which includes
turn down water and external shell spray cooling. For this sytem, a sustained
very low blowdown rate is hydraulically achievable and has been demonstrated.
However, long-term operation at a low blowdown rate s'eems possible only with
optimum conditions throughout the entire system.
CONTINUOUS OPERATION AT LOW BLOWDOWN FLOW
WATER CHEMISTRY CHANGES
As blowdown rates were decreased, system data supported predicted
trends. Figure 8 shows the increase in concentration of total dissolved
solids as the system cycled up. Each point represents a yearly average of
weekly samples composited over a 24-hour period. The 1975-76 average is for
once through treatment, prior to recycle. The upward trend graphically
portrays the cycling up of various dissolved constituents as system hydraulic
balance improved and subsequent blowdown rate decreased.
The result of a more pronounced decrease in blowdown from November, 1979
to July, 1981 is demonstrated by increasing trends of total dissolved solids,
total hardness and total alkalinity as shown in Figure 9. Each trend curve
represents a best fit of data statistically significant with 99% confidence.
During this period, the monthly average of blowdown rate was as low as 8
gallons per ton, and for several short periods, there was zero system
blowdown. While at near zero or zero blowdown, system water conductivity
increased to the 7,000-8,000 micromho range, with total hardness
concentration of 1,000-1,400 mg/1.
The nature of deposits throughout the system also showed marked change in
characteristics. Results of deposit analyses over a period of time (Figure 7)
illuminate the change from material high in iron to a deposit high in calcium
carbonate.
TREATMENT CHANGES
Initial water treatment of the once through system, prior to recycle,
concentrated primarily on water clarification for suspended solids removal.
A combination of cationic polymer for particle charge neutralization, anionic
polymer for suspended solids removal and ferric chloride for clarity was used.
On start up of the recycle system, basic water clarification treatment was
supplemented with a combination dispersant product, with a surfactant for
particle wetting, polymer for solids carryover and lignin as an iron
dispersant. As the recycle system was tightened up, deposit analyses
indicated a need for improvement in deposit control, focusing on iron
deposition was well as calcium carbonate scale. The use of ferric chloride
was discontinued to reduce the input of dissolved iron into the system.
Additionally, reduced blowdown flows minimized the need for very high clarity
water. The high conductivity of the system water indicated a higher particle
charge, necessitating changes in both cationic and anionic polymers.
291
-------
GROSS CONCENTRATION.mg/l NET LOADING, kg/DAY
1200
800
400
0'
75
TOT/
— SOL
i
V
-
\L DISS
OS
N
OLVEO
/
/
-76 77 78 79 8
6000
(
4000
2000
0
o 75-
YEAR
RE
— CL
s e
N
-\
—
76 7
78
79
80
Figure 8. Republic Steel Cleveland District - Nos. 5 and 6 Blast Furnace
Recycle System - Total Dissolved Solids gross concentration
and net loading to discharge by year.
GROSS CONCENTRATION, mg/l
REPUBLIC STEEL CORP
CLEVELAND DISTRICT
5 8 6 BL FCE RECYCLE
100 200 300 400
SLOWDOWN RATE, GAL/TON
50O
Figure 9. Republic Steel Cleveland District - Nos. 5 and 6 Blast Furnace
Recycle System - Total dissolved solids, total hardness and
total alkalinity concentration as a function of blowdown rate
(November 1979-July 1981).
292
-------
Most recently, deposit analyses and water chemistry trends have
magnified the need for a change from an anti-foulant to a product which was a
more specific inhibitor for calcium carbonate deposition. Throughout this
process of tightening up the recycle system, constant emphasis was placed on
the need for controlling the potential for deposition which could lead to
operating problems at the treatment plant and the blast furnaces. The water
treatment supplier directed their efforts for minimizing this deposition
potential. Their responsiblity was to keep the system out of trouble as near
zero or zero blowdown was hydraulically approached.
Increased activity with sampling and monitoring followed. Procedures
were set up for daily analyses by the plant operators for conductivity, total
hardness, calcium hardness, alkalinity and chloride concentrations in the
recycle and make-up water. Deposition and corrosion coupons were installed at
key locations throughout the system, including recycle supply line, blast
furnace cooler supply lines, inside the blast furnace coolers and in the
scrubber supply lines.
INSPECTIONS
The frequency of internal gas cooler inspections was increased to
supplement the ongoing sampling and monitoring. The primary objective was to
detect and track changes inside the gas coolers, which provided a full scale
demonstration of water chemistry and treatment effectiveness. The most
obvious observation was the plating out of a thin calcium carbonate layer over
the entire wetted portion of the upper section of the gas coolers. The
appearance was as if the inside had been spray painted. This area had
previously shown signs of corrosion, due to the depressed pH of the water
caused by carbon dioxide absorption. The carbonate coating had covered
previously detected areas of corrosion on side walls, grating, beams and
nozzle piping. Inspection of internal cooling checkers did not reveal
evidence of significant scaling to cause restriction in gas or water flow.
Inspection of deposit coupons in the system has shown no evidence of
major deposit problem. There are indications that corrosion and/or erosion
continues to be a potential problem, with some coolers showing "pitting"
corrosion in internal members. Based upon inspections, it seems possible to
have both calcium carbonate deposition and corrosion occurring simultaneously
inside the coolers.
MAINTENANCE OF HYDRAULIC BALANCE
Daily maintenance of the system hydraulic balance continues to be a
difficult task at extremely low blowdown rates. The large number of potential
sources for gaining and losing water throughout the system requires a thorough
knowledge of its physical layout. Preferably, the water treatment plant
supervisor should be familiar with blast furnace, coke plant and powerhouse
areas to locate and take action to correct hydraulic imbalance expeditiously.
More occurrences of hydraulic imbalance have surfaced due to the
declining physical condition of No. 5 Blast Furnace than had occurred since
recycle system start up. The operation of a blast furnace changes throughout
293
-------
its campaign, cycling every 4 to 5 years. The opportunity to make major
modifications to correct hydraulic imbalance usually occurs only once every
cycle, when the furnace is down for reline. Problems must be correctly
identified and solutions applied at these times.
There were several indicators for monitoring system tightness, including
water sampling and lab analysis. But correction of hydraulic imbalance to
maintain very low blowdown had to be quickly remedied and, therefore, readily
detected. This very tight system will lose more water through evaporation and
cooling tower drift than it picks up from the blast furnace process.
Therefore, make-up is required. Extraneous water sources in the system became
uncontrolled sources of make-up. As long as these sources did not exceed the
volume of water loss, the system remained in hydraulic balance. If these
sources were less then system water loss, additional make-up had to be
provided at controlled rate to maintain system water volume.
Monitoring controlled make-up flow indicated when the system was picking
up extraneous water. At any given blowdown flow, if make-up flow decreased,
the system was checked for extraneous water infiltration. If make-up flow
increased, a check was made for water loss. In either case, chemical
equilibrium of the system changed and was detected as a gradual change in
conductivity.
At zero blowdown, monitoring of the system hydraulics became even more
critical. For this recycle system extraneous water showed an increase in cold
well level. • When surge volume was consumed, this extra water was
automatically blown down.
WATER CONSTITUENT TRENDS
Effects of cycling up the recycle system are graphically portrayed in the
analyses of three prime constituents; ammonia, cyanide and phenol. Data
collected on these constituents tracks the gross concentration and net
loading annually from 1975 to 1980. More recent data, (November, 1979 to
July, 1981) reveals the increaed concentrating effect of decreased blowdown.
These constituent trends displayed a 99% confidence when statistically
analyzed.
Ammonia Concentration
As seen in Figure 10, the increase in ammonia concentration, although
not steady, was significant. The increased concentration was offset by
decreased blowdown flow, yielding an overall net decrease in the ammonia load
discharged. Analysis of recent data (Figure 13) shows a very sharp increase
in ammonia concentrations as very low blowdown rates were achieved.
The concentration of ammonia is dependent on several other factors, with
the type of water used for coke quench being very significant. The majority
of coke used in these blast furnaces has been quenched with river water or
treated process effluent from a physical chemical treatment plant at the No. 1
Coke Plant. Stripping of ammonia throughout the system (clarifiers, launders
and cooling tower) is also to be considered, as is system pH which determines
294
-------
GROSS_CONCENTRATION,mg/l NEJ LOADING, kg/DAY
75-76 77
400
r—\ 300
200
100
1
REPUBLIC
CLEVELA
586 BL
\
-\
~~ AM
^,
—c
MONIA l
1
STEEL CORP
YD DISTRICT
FCE RECYCLE
S
X<
&S (N)
> - -c
75-76 77
78
79
80
Figure 10. Republic Steel
Cleveland District Nos.
5 and 6 Blast Furnace
Recycle System - Ammonia
gross concentration and
net loading to discharge
by year.
YEAR
GROSS_CONCENTRATION,mg/l
LOADING, kg/DAY
3.0
2.5
2.0
0.5
0
7
\
\
\
TOTAL
—
CYANIC
\,
6 /'
/
/
5-76 77 78 79 8
120
100
80'
60
j
40
20
0
0 75-
REP
JBLIC S
CLEVELAND
sa
\
\
\
\
5 BL FC
TOTAL
t^^^—i
TEEL C<
DISTRI
E RECY
CYANI
•>
3RP
CT
CLE
>E
76 77 78 79 8
YEAR
Figure 11. Republic Steel
Cleveland District Nos.
5 and 6 Blast Furnace
Recycle System - Total
Cyanide gross concentra-
tion and net loading to
discharge by year.
GROSS CONCENTRATION, mg/l
0.8
LOAD, kg/DAY
75-76 77
Figure 12. Republic Steel
Cleveland District Nos.
5 and 6 Blast Furnace
Recycle System - Phenol
gross concentration and
net loading to discharge
by year.
YEAR
295
-------
the absorptive capacity of the water for the gas. The net effect is that
ammonia concentration has reached relatively high levels (100-200 mg/1) when
the blowdown was reduced to very low levels.
Cyanide and Phenol Concentration
The results of decreasing blowdown do not indicate consistent long-term
trends for cyanide and phenol concentrations (Figures 11 and 12). These
constituents apparently are more process related. The net loading does show a
drastic reduction, again due to the reduced blowdown rate.
However, analysis of data during significant blowdown reduction
(Figures 14 and 15) display the same concentrating trend at very low blowdown
rates. The effect of cyanide stripping in the system becomes part of this net
effect.
ONGOING CONSIDERATIONS
Efforts toward blowdown reduction were initiated to improve the overall
efficiency of the recycle system and to decrease the net loading discharge of
various constituents. The January 7, 1981 proposed effluent standards for
BAT/BCT technology indicate two alternatives of treatment, zero discharge or
additional treatment of blowdown prior to discharge. For zero discharge,
hydraulic balance is absolutely essential to eliminate the need to
hydraulically dispose of water. For additional treatment of the blowdown
discharged, capital and operating expenditure for a treatment scheme can be
reduced by minimizing the flow rate.
SIDE STREAM SOFTENING
Confirmed by physical inspections, and as seen graphically in Figures 7
and 9, calcuim carbonate scaling is a potential problem as the system
approaches zero blowdown. Continued performance or cost effectiveness of
chemical treatment to cope with this potential is untried and unknown. The
decision was made to examine the possibility of transferring side stream
softening technology to the blast furnace recycle system.
With the aid of the USEPA Mobile Pilot Treatment Plant, a comprehensive
study of the use of water softening chemistry was undertaken. The intent was
to investigate various chemical modes on a 5-6 gpm side stream to determine
the feasibility and cost projections of side stream softening to control
system hardness. The modes investigate were the use of lime, caustic, lime-
soda ash and caustic-soda ash. Also investigated were the effects of
recombination of the softened side stream with the unsoftened system water.
The pilot plant equipment included a rapid mix tank, flocculator and
clarifier. A final filter was added to determine carryover of solids. The
softening operation was determined by jar testing, then pilot plant results
were applied to jar test data to verify the use of those tests in projecting
softening reactions. The initial study was conducted on clarifier effluent.
296
-------
GROSS CONCENTRATION, mg/l
REPUBLIC STEEL CORP -
CLEVELAND DISTRICT
5 8 6 BL FCE RECYCLE
100
80
60
40
20
0 100 200 300 400
SLOWDOWN RATE, GAL/TON
GROSS CONCENTRATION, mg/l
500
REPUBLIC STEEL CORP
CLEVELAND DISTRICT
5 8 6 BL FCE RECYCLE
0 100 200 300 400
SLOWDOWN RATE, GAL/TON
GROSS CONCENTRATION, Alg/1
iocr
500
REPUBLIC STEEL CORP
CLEVELAND DISTRICT
5 86 BL FCE RECYCLE
100 200 300 400
SLOWDOWN RATE, GAL/TON
500
Figure 13. Republic Steel
Cleveland District - Nos.
5 and 6 Blast Furnace Re-
cycle System - Ammonia
gross concentration as a
function of blowdown rate
(November 1979-July 1981).
Figure 14. Republic Steel
Cleveland District - Nos.
5 and 6 Blast Furnace Re-
cycle System - Total
Cyanide gross concentration
as a function of blowdown
rate (November 1979-
July 1981).
Figure 15. Republic Steel
Cleveland District - Nos.
5 and 6 Blast Furnace Re-
cycle System - Phenol gross
concentration as a function
of blowdown rate (November
1979-July 1981).
297
-------
Lime Softening
With the use of lime only for softening, there was a slight decrease in
total hardness. At elevated pH values, magnesium hardness was significantly
reduced, only to be replaced by calcium hardness, lime being the calcium
donor. Thus, the net total hardness decrease was not significant.
Caustic Soda Softening
By elevating the pH with the use of caustic, dramatic decrease in both
magnesium and calcium hardness occurred. This reduction in total hardness
optimized in a range of 10.5-11.0.- There apparently was sufficient carbonate
present in the water to precipitate out most of the calcium. This was
accompanied by a slight increase in total alkalinity, but only significant at
pH values over 10.0.
Lime-Soda Ash Softening
The increased calcium hardness with the use of lime only was eliminated
with the addition of soda ash at elevated pH values. The carbonate donor soda
ash significantly reduced, total hardness and precipitated calcium carbonate.
As total hardness reduction optimized, total alkalinity started increasing
due to the then excess carbonate. This represented a somewhat classic lime-
soda ash softening process.
Caustic Soda Ash Softening
At higher pH values, the addition of soda ash served no value, as
sufficient carbonate was already present in the water for precipitation of
calcium. Adding soda ash had little effect on total hardness reduction, and
only increased total alkalinity.
Remixing of Softened With Unsoftened Water
Of particular concern with side stream as opposed to full stream
softening was the effect of remixing these streams. If the mixed water
continued the softening reaction, precipitation of solids could create a
sludge problem in areas where removal is difficult if not impossible without a
major outage. Remixing of softened water from each softening process with
unsoftened water indicated that calcium hardness decreased as the percent of
softened water increased. Precipitation of solids did occur.
Other Softening Trials
Further tests were conducted to check softening reactions at points of
the system other than the clarifier effluent. Trials of cold well effluent
and dirty water influent parallelled results from the previous tests.
The pilot plant was operated with each softening method and the results
compared to the jar test studies. Pilot plant lime softening reflected the
jar test data. Caustic and caustic-soda ash softening in the pilot plant
showed better results than did the jar tests. However, pilot plant lime-soda
298
-------
ash softening results were not as good as jar test indications. The reasons
for this have not yet been determined, but should be investigated further.
Full Scale Trial
Based on results of these tests, costs of softening this water were
projected for a full scale trial. These costs varied with the chemical used
and ranged from $260 to $2,600 per million gallons treated, and are based on
initial softening of the entire system. However, costs of maintaining the
softening reaction can be put into perspective by comparing them to existing
chemical treatment costs' for clarification and deposit control at
approximately $50 per million gallons treated.
Potential calcium carbonate scaling was overwhelmingly evidenced if
efforts were to continue toward operating this recycle system with very low
blowdown. Because side stream softening appeared as a viable means of cont-
rolling calcium carbonate scaling, the decision was made to proceed with a
full scale trial. Plans for temporary modification of one of the two
clarifiers to side stream soften treatment plant influent are underway as of
this writing. The trial will commence with the use of caustic as the
softening chemical. This method, although expensive, provides a less
cumbersome approach during the trial to evaluate the total system effect. The
objective is to determine if system hardness (calcium carbonate scaling
potential) can be controlled by side stream softening, and to provide more
detailed operating costs.
ALKALINE-CHLORINATION TREATMENT OF BLOWDOWN
Because of the ability of this blast furnace recycle system to achieve
very low blowdown rates, it was decided to utilize the USEPA Pilot Plant
facilities to further investigate the cost of blowdown for cyanide removal.
As seen in Figures 13, 14 and 15, concentrations of ammonia, cyanide and
phenol show a definite increasing trend as blowdown rates decreased.
Specifically, the presence of ammonia can cause significant increase in
chlorine demand before cyanide destruction can occur.
Preliminary results confirm that ammonia concentration is a major
consideration in the alkaline-chlorination treatment scheme. Additional
testing is underway as of this writing to further investigate air stripping as
a supplemental means of ammonia reduction combined with alkaline-'
chlorination.
Acknowledgments
The authors wish to express their appreciation to Messrs. Gary Skerl and
Kenneth Siegel of Republic Steel for their help in compiling and statistically
analyzing the data. Also, we would like to thank Messrs. A. Crespo, G. Main
and T. Medvin of Republic Steel's Research Center for the efforts they
expended in making the figures and slides used for the speech and in the text
of the paper. Finally, we would like to acknowledge Mr. M. Raymond Matuza of
Betz Laboratories for his help in coordinating the many services provided by
his company during this test.
299
-------
MINIMIZING WATER SLOWDOWNS FROM SELECTED STEEL PLANT PROCESSES
By: Harold J. Kohlmann and Harold Hofstein
Hydrotechnic Corp.
New York, NY 10001
ABSTRACT
The objective of the project was to attempt to minimize the volumes of
water blown down from:
Blast Furnaces
EOF
Continuous Casters
Hot Forming Mills
Sinter Plants
Electric Arc Furnaces
Vacuum Degassers
while meeting the following constraints:
existing treatment facilities to be used
no major capital equipment to be used
the reduction would have no adverse impact on the produc-
tion facilities.
The methodology followed to perform the task was to:
A. Review existing literature to determine which plants have
extensive recirculation systems installed and discharge
volumes that are close to BAT volumes (October 1979 pro-
posed guidelines).
B. Contact candidate plants, arrange to visit and present the
proposed program and receive agreement to cooperate.
C. Receive plants' data, perform paper studies, visit the
plants to establish gauging and sampling points.
300
-------
D. Gauge flows and sample at selected points and, where ap-
plicable, perform treatability studies.
E. Interview operating personnel to determine water volumes
and qualities required for efficient plant operation and
criteria for blowing down water.
F. Establish modified operating procedures and/or design
system modifications and minor additions. Prepare a pre-
liminary report.
G. Discuss the report with plants and obtain their concur-
rence.
H. Plants make modifications.
I. Operate modified systems for a period of sufficient length
to prove that systems can be operated with reduced blowdown
and without adverse effects.
J. Prepare a final report for each plant.
N
The paper presented includes the status of the project up to the time
of presentation.
OBJECTIVES
In September 1980 Hydrotechnic Corporation received a contract from
the U.S. EPA to provide technical services to assist in demonstrating at
steel plant sites that wastewater blowdown volumes from steel plant pro-
cesses can be reduced to minimum levels with carefully controlled recycle
systems. The steel plant processes selected by EPA were:
Basic Oxygen Furnace
Blast Furnace
Continuous Casting
Vacuum Degassing
Electric Arc Furnace
Hot Rolling Mill
Sinter Plant
The facilities to be selected for study were to be those that had, in
place and operating, recycle systems that were not combined with other
production process recycle systems. The selected system has to presently
be discharging relatively low volumes of wastewater as blowdown. Addi-
tional constraints were that:
1. Recirculation and treatment facilities already in place were
to form the basis for proposed discharge reductions.
301
-------
2. There is to be no requirement for the addition of major
capital equipment to accomplish the project goals.
3. Any reduction in flows discharged or any increase in recir-
culation rate would have no adverse impact upon the mill
equipment or product quality.
METHODOLOGY
The first step in performance of the project was to select appropriate
plants. A detailed review was made of the Draft Development Document for
Proposed Guidelines (October 1979) and the data published in those docu-
ments were compared for accuracy against the original 308 data provided to
EPA by the individual steel plants. In addition, trip reports written by
the Development Document contractor were provided by EPA, in confidence,
where additional data were available. Lists of candidate plants were then
prepared. These lists were further condensed by eliminating individual
plants that did not have multiple facilities of interest.
After the lists were prepared, a meeting was held between the con-
tractor, EPA (Office of Research and Development, Research Triangle Park,
NC), and the AISI. The purpose of this meeting, which was attended by
various steel corporation representatives, was to explain the project
objectives and to request the cooperation of the industry. The list of
plants selected for the study was presented to the attendees during the
presentation of the project description. If corporate representatives of
the listed plants were present, their cooperation was requested.
Soon after the meeting, the appropriate companies were officially
contacted and two plants with a total of six of the manufacturing facili-
ties of interest (four at one plant and two at another) indicated a will-
ingness to cooperate provided that they were given a more in-depth descrip-
tion of the work to be done and what was expected of their plant personnel.
Subsequently, engineers from Hydrotechnic visited each of the plants to
further explain the objectives of the project and define the plants' re-
sponsibility during the performance of the study. Both plants, after these
presentations, agreed to host the study.
At each facility, existing data consisting of process flows within the
system, treatment facilities in place and mode of operation, blowdown and
makeup flows, qualities of water, production data and drawings (both flow
and piping) were provided.
After study of the data, sampling and gauging locations were selected.
A second trip was then made to each facility and, in the company of know-
ledgeable plant personnel, each sampling and/or gauging point was physical-
ly located and marked.
Most of the points were on pressure lines; therefore, Pitot tubes were
used for flow measurement. Open channel flows were gauged with a Gurley
current meter. The plants provided taps and valves on each of the lines of
interest and, where necessary, uncovered points where flow measurements
302
-------
were to be made in open channels. Each plant also provided laboratory
space where Hydrotechnic personnel performed routine analyses.
The survey consisted of gauging and/or sampling at each selected point
at least two times.
After the surveys were performed, the data was analyzed and reduced
and schematic flow diagrams were prepared for each production facility
studied. Flow balances were made around each facility and the water qual-
ity deterioration, if any, was evaluated. Based on the data, water system
modifications are to be developed and presented to the plants for their
review. The purpose of this plant review is to check for accuracy of flow
paths and to advise of any difficulties that might be encountered in imple-
menting the suggested system modifications.
After consultation with the host plants, designs for implementation of
the system modifications will be prepared and transmitted to the plant for
installation of the required equipment and instrumentation. After instal-
lation, the system will be operated under the new conditions and monitored
by Hydrotechnic engineers. Corrosion and deposition coupons will be in-
serted in the recirculation lines and measured periodically to determine
the effects, if any, on the water systems. This discussion presumes that
water application rates within the individual production processes are not
changed; only the degree of recycle. It is not the function of this proj-
ect to modify production operations which require specified amounts of
water. However, if by making the mill personnel more water conscious, they
may voluntarily reduce the quantities of water that are presently used. If
this occurs, then it may be possible to reduce the volumes blown down by
even greater amounts.
PRELIMINARY RESULTS OF THE STUDY
At one host plant, four different production processes water systems
were studied: Blast furnaces; a B.O.F. shop; a continuous casting shop;
and a secondary hot forming mill. The continuous casting shop has an
electric furnace shop associated with it which utilizes dry gas cleaning.
The host plant is presently reviewing the preliminary recommendations for
the reduction of discharges that are based on this study.
In general it was found that the volume of blowdown in the systems
studied is controlled by the amount of excess water that enters the systems
at various points within the mills and shops. Operating personnel reported
that no problems are experienced with the recirculated water quality re-
turned for reuse from the respective treatment systems. At two mills
operating personnel reported that the reason for adding makeup water at the
mill was due to the inability of the recirculation system to return to the
mill the volumes required for efficient operation.
A specific example of excess water entering a system is at the blast
furnace water system. There are two operating blast furnaces with gas
cooling and gas cleaning systems. Blowdown from the gas cooling system is
cascaded to the gas cleaning system and the combined blowdown discharges
from the gas cleaning system.
303
-------
During the survey, gauging of the following was performed (Figure 1):
1. The flow from the gas cooling cold well.
2. The flow from the gas coolers to a settling basin.
3. The flow from the gas cooling hot well to the cooling tower.
4. Slowdown from the gas cooling system to the gas cleaning
system.
5. Flow from the gas cleaning cold well.
6. Flow to the gas cleaning thickeners.
7. Flow from the gas cleaning hot well to the cooling tower.
8. Slowdown from the gas cleaning system.
9. Service water makeup to the gas cooling cold well.
The total system blowdown flow varied from 370 to 1000 gpm. Cooling
tower losses were calculated to be 125 gpm and sludge hauling records
indicated that the average daily sludge discharges contained approximately
55 gpm of water for a total system loss of from 550 to 1180 gpm.
Additions to the system of condensate from the blast furnace gas,
estimated to be a maximum of approximately !65 gpm, and fresh water makeup,
added to the gas cooling system cold well, amount to approximately 440 gpm.
Therefore, it can be seen that there is frequently more water blown
down from the total system than can be accounted for by additions to the
system.
We were informed that water from underground sumps, that accumulate
ground water infiltration, is pumped into one or both of the water systems.
It may, therefore, be possible to reduce the total blowdown caused by
groundwater infiltration by reducing or eliminating the service water
makeup and use this infiltration water as -a total or partial source of
makeup. Of course,, during dry weather periods, when infiltration is re-
duced, service water may have to be used for all makeup.
The positive Langelier Index calculated for the water in the gas
cleaning system indicates a scaling tendency. However, no problems with
scaling have been reported. Therefore, it is believed that by increasing
the cycles of concentration, the system blowdown can be reduced. It is not
known, nor is there a convenient way of measuring, the volume of seepage
that enters the system; therefore, a guided trial-and-error procedure is
being recommended for which blowdowns will be controlled by water level
control in the gas cleaning system and by the effects on corrosion/deposi-
tion coupons, which would be inserted in the recirculating lines. These
coupons will be inspected periodically. Based on this on-site, empirical
304
-------
GAS CLEANING SYSTEM
GAS COOLING SYSTEM
GO
O
cn
COOLING
TOWER
(TYP)
THICKENERS
Figure 1. Flow Diagram Blast Furnace Water System
-------
operation of the water systems over a period of 2-4 months, the actual
amount of blowdown reduction can be ascertained. After this initial peri-
od, periodic checks are recommended to assure that long-time effects do not
cause the water quality to further deteriorate to a point where an adverse
effect is experienced either by the mill equipment or the product.
During this initial trial period and further test period, close coor-
dination should be maintained with concerned plant personnel so that they
are completely aware of what is happening; thus, when the test periods are
completed, they are assured that the systems modifications as proposed do
indeed produce lower blowdown volumes without affecting equipment or prod-
uct and do not result in impractical cost increases.
CONCLUSIONS
Some reduction of discharges from production facilities can be accom-
plished by judicious water management, if operating personnel are kept
aware of the fact that the quantities of wastewater discharged are of
concern to plant and corporate management, not only with respect to meeting
governmental regulations, but also with respect to the operating costs
incurred when treating large volumes of wastewater. Wherever possible,
recirculated water should be used. Introduction of extraneous water should
be minimized. If this is not possible, as in the example cited, service or
fresh water makeup should be reduced to take full advantage of the extran-
eous water as a source of makeup.
Mill operators should make it an ongoing effort to reduce water vol-
umes to absolute minimums. In this manner it would be possible to reduce
discharges with virtually no major capital expenditures. However, some
in-mill repiping might permit further recirculation or reuse of water
before treatment or cooling. Before these goals can be accomplished, plant
personnel associated with the water systems must be convinced that the
reductions are necessary and do not adversely affect the equipment or
product.
306
-------
ZINC CONTROL IN A BLAST FURNACE GAS WASHWATER RECIRCULATION SYSTEM
by: R. Gregory Elder, Engineer, Utilities Department and
Roy Littlewood, Senior Technical Adviser, Research & Development Dept
Stelco Inc.
Hamilton, Ontario
L8N 3T1 Canada
ABSTRACT
In general, it is not possible to operate blast-furnace-gas-washwater
circuits at a high degree of recirculation in a steel plant which also
recycles a large proportion of zinc-containing scrap and solid wastes.
Certain undesirable elements, particularly zinc, are not removed during iron-
and steel-making, and will accumulate to dangerous levels under high recycle
conditions.
The most serious danger is to blast furnace operations. Zinc threatens
these operations by:
- endangering refractories (which can lead to blowout, collapse or
premature shutdown of the furnace)
- causing the descending charge to stick (leading to operating
irregularities, decreased furnace efficiency, and hazardous 'slips'), and
depositing solids in water recirculating systems (which can block pipes
and cause the collapse of cooling towers).
To avoid these problems, all steel companies restrict the amount of zinc
in materials fed to blast furnaces. A typical maximum permissible level is
0.5 kg/tonne of iron produced - equivalent to about 0.02 percent of zinc in
the raw materials fed to the furnace.
Stelco's Hilton Works in Hamilton, Ontario has an established practice of
recycling solid wastes, including all blast furnace dust from both wet and dry
collection systems. Severe operating problems were caused when a
recirculation system was commissioned in 1978 for three blast furnaces which
had previously operated with a once-through gas-washing and solids-separation
system. Subsequently, it was found possible to operate, provided blast
furnace dusts (up to 40 000 tonne/year) were dumped.
307
-------
To avoid this disposal problem, a method of generating a zinc-laden
blow-down stream has been devised, which enables zinc to be selectively purged
from the system. This technique allows the water recirculation system to be
operated, while still recycling most of the blast furnace dusts.
ABOUT STELCO
For the benefit of those unfamiliar with Stelco, the company is the
largest steel producer in Canada. With an annual production capacity of over
6 million tonnes, Stelco produces about 35% of the nation's steel. Its
operations are integrated from the mining of coal, iron ore, and limestone to
the manufacture and distribution of iron, steel and steel products.
Stelco?s manufacturing facilities comprise nineteen plants, located
across Canada.
The main steel plant, Hilton Works, occupies 440 hectares, with harbor
facilities in Hamilton, Ontario and is considered to be one of the most
efficient integrated steel plants in North America. This high efficiency
results in part from a long-standing emphasis on energy saving and the
utilization of wastes.
RECYCLING OF WASTES
The steel industry has a strong tradition of recycling not only the
wastes it generates itself, but also waste steel generated by the community.
Scrapped steel products constitute a la'rge proportion of the charge of
steelmaking furnaces, from about 30 percent for EOF furnaces up to 100 percent
for electric arc furnaces.
This tradition of waste recycling also extends to other solid wastes,
especially those containing a high iron content, such as the iron oxide dusts
collected by gas-cleaning equipment and the iron oxide scale formed during the
heating and rolling of steel. These oxides are usually recycled to the blast
furnace after agglomeration in a sinter plant.
But there are practical limits to the amounts which can be recycled.
Some materials, including blast furnace flue dust and sludges, contain
so-called "tramp" elements which are not removed in iron- and steel-making
operations and which have undesirable effects on processes and products.
Recycling of wastes containing such tramp elements leads to a successive
build-up of the tramp elements in the processing system, causing deleterious
effects on the process, or the production of unacceptable output.
Zinc is one tramp element of serious concern in steel manufacture, and it
is usually associated with lead. Zinc and lead behave similarly, but zinc is
usually present at higher levels than lead and so attention is normally
directed to controlling it. The most serious effects of zinc and lead are on
blast furnace operation.
308
-------
EFFECTS OF ZINC AND LEAD ON BLAST FURNACE OPERATION
Serious operating problems result if zinc or lead is present in the raw
materials charged to the blast furnace. The problems include:
- endangering the refractories (which can lead to blow-out, collapse or
premature shutdown of the furnace);
sticking of the descending charge (which can cause operating
irregularities, decreased furnace efficiency, and dangerous "slips"); and
- deposition of zinc or lead compounds in water recirculating systems
(which can block pipes and cause cooling towers to collapse).
Restriction of zinc and lead in materials fed to blast furnaces is
practiced by steel companies throughout the world. A typical maximum
permissible level is 0.5 kg of zinc/tonne of hot metal produced (equivalent to
roughly 0.02 percent of zinc in the raw materials fed to the furnace).
Recycling of all blast furnace dusts from both wet and dry collection
systems is a long-established feature of Stelco's operations. Although it is
common practice in many steelplants to dump all or part of these wastes, at
Stelco we recycle them all via a sinter plant to the blast furnaces. Control
of zinc levels in the blast furnaces is therefore a continuing concern to our
blast furnace operators.
EFFECTS OF RECIRCULATION OF BLAST FURNACE WASHWATER
Prior to 1978, three blast furnaces "B", "C" and "D", at Hilton Works in
Hamilton operated on a "once through" washwater system. Water was pumped from
Hamilton Bay to the gas coolers, scrubbers and wet precipitators. The water
was then cleaned in thickeners, and returned to Hamilton Bay. All blast
furnace flue dust and sludge filter-cake were recycled to the sinter plant
feeding the blast furnaces.
Figure 1 schematically represents the zinc paths in the system under
these once-through conditions. Zinc is present in the materials fed to the
blast furnaces, notably in steelmaking slag (another recycled waste) and also
in ore, fluxes and coke. Zinc is also present in the feeds to the sinter
plant, most of which are recycled wastes such as blast furnace flue dust and
filter cake, millscale, steelmaking fines and calcite-dolomite screenings.
These zinc inputs to the system are balanced by zinc outlets from the system.
The balance between zinc inputs and outputs determines the ultimate zinc level
in the blast furnaces.
309
-------
PELLET I ZED ORE
FLUXES (LIMESTONE, DOLOMITE)
COKE
SIEELHAKING SLAG
BLOMUOWN
MILL SCALE
STEELMAKING FINES
PURCHASED ORE
CALCITE-DOLOMIIE SCREENINGS
Figure 1: Zinc Paths in the Blast Furnace/Sinter Plant System
- Once-Through Water Flow
Zinc outlets from the system include the small amounts contained in the
molten iron and slag tapped from the blast furnaces, and the low
concentrations contained in thickener overflows returned to Hamilton Bay. The
zinc concentration in this water was about 10 mg/L.
When we commissioned the recirculation systems for gas scrubber water at
"B", "C" and "D" blast furnaces, zinc levels in the blast furnaces quickly
rose above the critical levels. Recirculation of the washwater had to be
discontinued to avoid damaging the blast furnaces. There were also other
problems in the water system, where heavy, gelatinous precipitates deposited
on the cooling tower, parts of which collapsed.
Figure 2 schematically represents the zinc flows in the system under
recirculation conditions. The volume of thickener overflow rejected to the
Bay was reduced under this mode of operation. Zinc outputs from the system
were less .than under once-through conditions, and as a result, zinc levels in
the blast furnaces were higher at the balance point.
310
-------
rri.iEiizto ORE
FLUXES (IIMF.SIONE, DOLOMITE)
COKE
SIFELMAKIMG SLAG
SLAG
HOT HE1AI.
BLOHDOHN
MILL SCALE
STEELMAKING FINES
PURCHASED ORE
CALCITE-DOLOMITE SCREENINGS
Figure 2: Zinc Paths in the Blast Furnace/Sinter Plant System
- Recirculated Water
It was subsequently found possible to operate the blast furnaces at
acceptable zinc levels under recirculation conditions provided all blast
furnace dusts were dumped, at an estimated rate of about 40 000 tonne/year,
rather than recycling them (figure 3). This solution to our zinc problem
would replace a water environmental problem with a solid waste disposal
problem - a situation not acceptable in our operations, where wastes recycling
is a strong tradition. A better alternative was needed for keeping zinc at an
acceptable level for blast furnace operation, while recycling as much as
possible of the flue dust and filter cake. A method of bleeding zinc from
from the system was necessary, which would give the same zinc-purging action
as the old once-through washwater system.
311
-------
PELLET I ZED ORE
FLUXES (LIMESTONE, DOLOMITE)
COKE
STEELHAKIN6 SLAG
MILL SCALE
STEELHAKING FINES
PURCHASED ORE
CALCITE-DOLOMITE SCREENINGS
Figure 3: Zinc Paths in the Blast Furnace/Sinter Plant System
- Recirculated Water, Dumped Blast Furnace Dusts
METHODS OF, ZINC PURGING
A state-of-the-art review revealed that, although the conventional
practices of rejecting zinc in water effluent or in solid wastes are no longer
environmentally satisfactory, most steel companies in the world still practice
one of these two disposal methods. Nevertheless, the state-of-the-art review
suggested that there might be several other possibilities.
REDUCTION PROCESSES
A few companies in Japan and Germany treat the contaminated wastes in a
separate kiln process to remove the zinc. The non-ferrous metals are
collected as a flue dust by-product, while the purified iron oxides, reduced
to sponge iron in the process, are usable in blast furnace or steelmaking,
since they are freed of zinc contamination.
312
-------
This route is expensive, in both capital and operating costs. One plant
manufacturer estimates (optimistically) that the economic break-even size for
a plant is 400 000 tonne/yr (about ten times the amount of blast furnace dusts
normally recycled by Stelco in a year). In Germany and Japan, there are
examples of centralized processing facilities serving several companies, which
by processing other zinc-contaminated wastes (eg steelmaking dusts) as well,
achieve economic operation. One Japanese steel company recently closed down
its waste-processing kiln because the process was too costly.
Several reduction processes have been mooted for application in North
America, but none have reached the stage of full-scale commercial exploitation
for zinc removal from steel plant wastes, primarily for economic reasons.
Processes which are feasible in principle include the SL/RN process,
developed for iron ore reduction; the Wa'lz process for the processing of
zinc-containing wastes; the Kowa-Seiko process of chloride reduction-roasting
for zinc recovery; and the Inmetco process, used for the recovery of electric
furnace waste oxides.
LEACHING
Substantial amounts of zinc contamination can be dissolved from iron
oxide wastes with acid or alkali. If sufficient purification were achievable,
the leached oxides could be recycled through a sinter plant. However, a
zinc-rich residue would still require disposal.
Such processes have been widely studied in the laboratory. Several
processes have been patented and pilot plants are known to have operated in
the United Kingdom, Luxembourg and Sweden. None have become commercial yet
because of a combination of operating and economic problems. Our laboratory
studies of the leaching concept indicate that only about half the zinc in
steelplant iron oxide wastes is readily removed by acid or alkali - a severe
technical limitation.
PHYSICAL METHODS
Zinc tends to be associated more strongly with the smaller particles in
blast furnace dusts. Separation of the smaller particles is therefore a way
of preparing a zinc-rich fraction of the waste.
1 2
The principle of the method was worked out in France and in Japan .
Separation systems based on rejecting the zinc-rich, fine particles in a wet
system have been reported operating in Czechoslovakia and Japan . The
rejected zinc-rich fraction can be processed to recover the zinc if commercial
facilities are available or else it can be dumped. In either case, the
volume of residue is a fraction - perhaps one-quarter - of the unprocessed
'waste which would otherwise have to be rejected.
313
-------
PRECIPITATION FROM WATER
In principle, substantial amounts of. zinc can be removed from the water
circuits by a combination of precipitation and settling. For example,
water-softening techniques might be contrived to remove dissolved zinc from
the water along with other undesirable contaminants.
Our test work indicated such techniques were feasible in principle.
Zinc-rich suspended matter and precipitated zinc hydroxide were successfully
removed in a small reactor clarifier.
ZINC PURGING CONCEPT
Experimental studies of all the alternatives listed in the previous section
showed each had some potential for solving our zinc problem. We focused
attention on choosing a method which could be incorporated into our existing
operations with minimum equipment and process changes.
The chosen strategy was to purge zinc from a point in the blast
furnace/sinter plant system, so that zinc concentration in the blast furnace
dusts remained low enough for direct recycling to the sinter plant.
Measurements indicated that, under our conditions, we would probably need to
purge about 20 to 40 kg/h of zinc from the system to maintain acceptable
furnace operations at low blowdown rates. However, this data was
indeterminate-and the first concern was to obtain a reliable estimate of thit
important design parameter.
There were many possible locations in the system from which zinc might be
purged. The combined water-washing circits of the blast furnaces at Hilton
Works form a highly complex network containing several miles of piping.
Comprehensive analysis of the various process streams revealed that a
recirculation stream - designated the "Return Pit Flow" - was enriched in
zinc. This stream was diverted from the system to a treatment lagoon for a
test period of 10 weeks, thus purging zinc continuously, as shown in figure 4.
314
-------
OAT WATER
' MAKE UP
E FURNACE SLOWDOWN TO
WEST SIDE TRANSFER SYSTEM
(WSTJ)
B7C » 0 FURNACES BtOWDOWN
TO WEST SIDE OPEN CUT
FILTER-CAKE
TO SINTER STRAND
RECYCLED TO SPUTTER
SOX IN NORMAL
DURATION
^^ _f ?_["•*•*• OVEB-FLOW
Figure 4: Flow Diagram - BCD and 'E1 Furnace Recirculation
System - Slurry Recovery System - Arrangement during
Return Pit Trial
SUCCESSFUL OPERATION AT LOW BLOW-DOWN
The continuous purge of zinc enabled the water-recirculation system to be
successfully operated at low blowdown rates while recycling all flue-dust and
filter-cake to the sinter plant. In previous attempts to operate under these
conditions, without a zinc purge, zinc levels in the blast furnace began to
approach dangerous levels after only a few days of operation at a blowdown
rate of 30 percent of the total flow. Using a zinc purge rate averaging 27
kg/h, we were able to operate at successively lower blowdown rates, while
keeping zinc levels well within the acceptable range for blast furnace
operation (figure 5). Ultimately, the system operated at the lowest blowdown
rate achievable hydraulically (13 to 16 percent)1 for a period of 30 days.
315
-------
12
.10
.08
.06
.02
9LOMXXVN RATE
•— air. •;•
10
20
3O
50
6O
DAY
Figure 5: Zinc Content of Sinter
Water stability conditions in the recirculation system were easily
controllable. No deposition, scaling or corrosion problems arose during the
trial period, even though dissolved solids content of the water rose by a
factor of four.
DETAILS OF THE PURGE METHOD
The Return Pit Flow, diverted from the system to a treatment lagoon,
comprised contributions from three sources.
1. Filtrate from the sludge drum-filters,
2. Overflow from a small thickener in the Filter Building, fed with
thickener sludge,
3. Intermittent overflow from the sludge drum-filters.
Separating these contributing flows for trial purposes was not possible
without complex engineering work, so we simply diverted the whole of the
Return Pit Flow, recognizing this method might not be the most desirable way
to get the maximum zinc purging effect.
The Return Pit Flow was purged at about 2000 L/min for ten weeks, before
cold-weather effects on equipment and detrimental effects on the behaviour of
the treatment lagoon forced the trial to be terminated. The ten-week trial
ended with a 30-day period at the lowest blow-down rates achievable
hydraulically.
The amount of zinc purged from the system via the Return Pit Flow
fluctuated widely from day to day, but there was no apparent upward trend as
blowdown rate was reduced (figure 6).
316
-------
SLOWDOWN RATE
5O
-* 3O
ID
z
Q
4
3 20
10
X X 14 38
DAY
44 50 54
Figure 6: Zinc Loading at Return Pit
These wide swings are an operating characteristic of our system.
Periodic flushing of zinc via the top gases is typical of blast furnace
operation. Fluctuations in suspended solids concentrations also arise from
our slurry handling practice.
FUTURE PLANS
The next stage in the development is the design and construction of a
permanent zinc-purging system. The system will use the principle of
separating a zinc-containing stream. The stream finally selected will depend
on several economic and technical factors.
An obvious candidate is the Return Pit Flow diverted during the trial,
over 80 percent of which is overflow from the Filter Building thickener.
Table 1 shows that this thickener acts as a particle size classifier, which
results in zinc enrichment of the overflow solids. This characteristic is an
attractive feature, since discarding a zinc-enriched stream will require
disposal of smaller amounts of solid waste. An economic assessment of capital
and operating costs of several alternatives is being made.
317
-------
TABLE I: ZINC CONTENT AND SIEVE ANALYSIS OF
SOLIDS IN THE FILTER-BUILDING THICKENER
Process
Stream
Content
Zinc
of
Solids
Sieve Analysis*
+45)jm
Fraction
of Solids
-45 pn
Fraction
Inflow 3.4
Overflow 4.8
32
11
68
89
* portion retained on, or passing through No 325
sieve
So far as we are aware, a thickener has not
been used as a
zinc-enrichment device previously. In Japan , a hydrocyclone has been used to
achieve zinc enrichment in a similar way to that achieved in our existing
Filter Building thickener. In Czechoslovakia , the fine particle fraction
collected by a venturi washer in the gas-cleaning system was treated
separately from the coarser dust collected in a wet scrubber. The fine
particle fraction was directed to a separate thickener and discarded.
Using the Return Pit Flow as the the zinc purging stream will still
present disposal problems, as our experience indicates it has deleterious
effects when fed directly to a treatment lagoon and filtration plant.
Clarification will be necessary to remove the suspended solids. These
zinc-contaminated wastes still present disposal problems, but at least the
amount to be disposed of has been reduced to about 10 000 tonne/year. The
filtrate is of similar composition to water blowdown from the recirculation
system and we envisage it will be given the same treatment.
CONCLUDING REMARKS
We have some way to go, then, before we bring into permanent operation a
satisfactory system for zinc control. In this connection, it is worth noting
that generally in the steel industry, it takes about 15 years to introduce a
major process change from the time of initial development to full-scale
application in the developing company. It looks as if development of a
satisfactory blast furnace water recirculation and zinc control system will
fit into a similar time scale.
This symposium's stated purpose of providing a forum for the exchange of
information on technology problems related to environmental control encouraged
us to present this paper at a stage when a solution to our .zinc control
problem is still incomplete. Our task is a good illustration of the depth and
complexity of the technical problems often faced by industry in conforming to
the environmental demands placed on it.
318
-------
The problem of adequate zinc control in blast furnaces with a recycled
waste load is a complex situation. In this paper, we have demonstrated that a
requirement to reduce volumes and levels of contamination of water discharges
is not easily met, even though it may appear simple and straightforward to an
unitiated outsider. Such a requirement has wide ramifications on the
operation of the steelplant, with difficulties which were not at first
recognized either by the environmental authorities or by the steelplant
operators.
The topic of zinc control in blast furnaces is also an example of the
growing conflict between requirements for air and water quality control and
reduction in the volume of solid wastes. Too often, improvements in the
quality of air and water discharges are demanded which result, unthinkingly,
in increased generation of solid wastes. The interaction between these
conflicting environmental factors needs to be examined to see whether
worthwhile trade-offs can be made and to decide what developments are needed
to dispose of satisfactorily the increasing amounts of solid residues from air
and water quality control activities.
We
topics.
^•4- VJVACL.L-i.U9 V^.VSJ.J.Wl-W-l. C*^.W-l-V-i-U..J-»—WJ»
hope this paper has directed your thoughts to the importance of these
REFERENCES
1. J. M. Cases et al, "Recovery of lead and zinc contained in blast furnace
gas-cleaning sludges". Proceedings of an Information Meeting on Wastes
and Residual Materials in the Iron and Steel Industry, Commission of the
European Communities, Luxembourg, June 1978, pp 67-85.
2. H. Toda et al, "Separation of Nonferrous Metals from Blast Furnace Flue
Dust by Hydrpcyclone", Nippon Steel Technical Report No 13 (1979), pp
73-79.
3. J. Vesely, "Method of partial separation of zinc from blast furnace flue
dust", Hutn. Listy, 1979, 34 (12), pp 876-77.
4._ S. Uno et al, "Dezincing Equipment and Operation Based on Wet
Classification of Wet Cleaned BF Dust", Nippon Steel Technical Report No
13 (1979), pp 80-85.
319
-------
INVESTIGATION OF REVERSE OSMOSIS FOR THE TREATMENT
OF RECYCLED BLAST-FURNACE SCRUBBER WATER*
by: M. E. Terril
Research Engineer, U. S. Steel Corporation
Research Laboratory, Monroeville, Pennsylvania 15146, and
R. D. Neufeld
Associate Professor of Engineering
University of Pittsburgh, Pittsburgh, Pennsylvania, 15213
ABSTRACT
Entrained dust in blast-furnace off-gas must be removed before the gas
can be used as fuel. Wet scrubbing is the standard method for cleaning blast-
furnace gas. Scrubber water, in addition to removing dust, also dissolves
contaminants including ammonia, phenol, and cyanide from the blast-furnace
gas. The scrubber water is contained in a recycle system and a sidestream
must be discharged to prevent scaling in the recycle system.
The objective of this research was to evaluate reverse osmosis (RO) as a
candidate technique to minimize the quantity of wastewater discharged from the
gas-scrubber recycle system. The RO permeate would be recovered as make-up
water for the recycle system. The concentrate stream would be discharged for
further treatment or possible evaporation via slag quench. Samples of blast-
furnace recycle water were obtained from a local facility, and tests were con-
ducted with a spiral-wound cellulose acetate RO module at a pH of about 5.
Operating pressures investigated ranged from 350 to 450 pounds per square
inch, gage, and feed operating temperatures were varied between 74 to 86°F.
Permeate flux rates (gallons/day/square foot) were measured as a function of
water-volume recovery level. Recovery levels ranged from 10 percent to over
80 percent.
Low membrane rejections were obtained for phenol, free cyanide, thiocya-
nate, and sulfide, indicating that these substances would be returned via the
permeate to the recycle loop. Consequently, significant reductions in the
discharge loadings of these materials could be achieved in the concentrate
stream. There are indications that these substances may not be conserved in
the recycle system and, therefore, may not increase in concentration within
* This paper is based on an M.S. Thesis submitted by M. E. Terril in
partial fulfillment of the requirements for the degree of Master of
Science in Civil Engineering at the University of Pittsburgh.
320
-------
the system. Aqueous-discharge reductions of 70 percent to 80 percent appear
to be possible.
Evaluation procedures for the RO system are discussed, applicability
advantages and disadvantages are reviewed, and conceptual flow diagrams
showing further treatment and disposal options for the reject stream are pre-
sented.
INTRODUCTION
Entrained dust in blast-furnace off-gas must be removed before the gas
can be used as fuel. Wet scrubbing is the standard method for cleaning blast-
furnace gas. As illustrated in Figure 1 the scrubber water is contained in a
recycle system. A sidestream (blowdown) is discharged to prevent scaling or
corrosion in the recycle system.
BLAST FURNACE OFF-GAS
RECYCLED SCRUBBER WATER
-«._--/ CLARIFIER I »
CLEAN GAS
" FOR FUEL
BLOWDOWN TO
DISCHARGE
Figure 1. BLAST-FURNACE GAS SCRUBBER RECYCLED WATER SYSTEM
Presently,-wastewater discharged from the gas-scrubber recycle system is
subject to Best Practical Control Technology Currently Avi'ilable (BPT) limi-
tations. Compliance with Best Available Technology Economically Achievable
(BAT) and Best Available Conventional Control Technology (BCT) limits will be
required by July 1984. These limits have been proposed by the U. S. Environ-.
mental Protection Agency (EPA) and are shown in Table I. Additional treatment
321
-------
will be required to achieve BAT limits. Alkaline chlorination is, currently,
the treatment method recommended by the EPA. For those systems that have
blast-furnace-slag quenching facilities, the gas-scrubber recycle-system blow-
down may be disposed of by evaporation on hot slag. Prior research was
conducted by Osantowski and Geinopolos1> who investigated ozonation, alkaline
chlorination, and reverse osmosis (RO) with polyamide membranes followed by
alkaline chlorination or ozonation of the RO concentrate as treatment
approaches for gas-scrubber blowdown water. Our study was based on the use of
RO for treatment of waters discharged from cooling-water recycle systems 2'3^
and advanced the work of Osantowski and Geinopolos by investigating spiral-
wound cellulose acetate membranes. The use of conventional spiral-wound
cellulose acetate membrances represents significant potential capital cost
savings. Figure 2 is a sketch of a typical spiral-wound membrane. A second
reason for selecting cellulose acetate was that it displays low rejections of
cyanide and phenol at pH values below 7.
PERMEATE
CONCENTRATE
MESH SPACER
FEED
PERMEATE CARRIER
MEMBRANE
ADHESIVE
Figure 2. SPIRAL WOUND MEMBRANE CONFIGURATION
Consequently, these contaminants should enter the RO permeate, which would be
recovered for reuse within the gas-scrubber recycle system. This concept is
shown in Figure 3. The EPA4* has indicated that gas-scrubber recycle systems
reduce the discharge of contaminants over a once-through system. This would
indicate that some mechanism, perhaps biological, exists to limit contaminant
(cyanide and phenol) loadings. Return of these contaminants to the recycle
system via an RO permeate stream would be expected to cause no increase in the
recycle-system loadings.
Cellulose acetate rejects calcium, magnesium, and sulfate, which are
scaling components, and chloride, a corrosive component. It is expected that
these substances would be removed via the concentrate stream where they are
discharged.
322
-------
BLAST FURNACE OFF-GAS
RECYCLED SCRUBBER WATER
ACID
CONCENTRATE
TO TREATMENT
OR DISPOSAL
Figure 3. REVERSE OSMOSIS TREATMENT OF RECYCLED SCRUBBER WATER SLOWDOWN
Table 2 is a partial listing of the contaminants found in the blast-
furnance gas-scrubber recycle-system water used in this study.
EXPERIMENTAL METHOD
The test device utilized in this research was an OSMO-1960-SS97PES pilot
scale unit manufactured by Osmonics Inc. It is rated for a permeate flow of
12.4 gallons per hour (gph) and a concentrate flow of 87.2 gph at a 600 pounds
per square inch, gage (psig) operating pressure with Osmonics SEPA-97 mem-
brane. This membrane was used in the evaluation and its characteristics are
detailed in Table 3.
323
-------
Scrubber water samples (55 gallons each) were obtained from a local
blast-furnace facility. The pH was adjusted to 5.0 ±0.2 units with sulfuric
acid prior to testing to destroy alkalinity, and to shift the dissociation
equilibrium for cyanide and phenol to their unionized state. Operating pres-
sures investigated ranged from 350 to 450 psig, and feed temperatures were
varied between 74 to 86°F. Permeate flux rates (gallons per day per square
foot, gpd/ft^) were measured as a function of water-volume recovery level. A
flow diagram of the experimental set-up is shown in Figure 4. Recovery levels
REVERSE
OSMOSIS
UNIT
PERMEATE
TO DRAIN
FEED
CONCENTRATE
WATER RECOVERY, % =
H -H.
100
H, = HEIGHT AT TIME t
FEED PUMP
55 GALLON
SAMPLE
CONTAINER
Figure 4. EXPERIMENTAL SET UP FOR WATER RECOVERV TESTING
ranged from 10 percent to approximately 80 percent. Two control runs with
distilled water were made before each experimental run. The first control was
run at an operating pressure of 400 psig and a feed temperature of 74°F to
detect changes in membrane performance. The second control was run at the
experimental conditions to determine the scrubber-water osmotic pressure. A
series of six tests was conducted to determine membrane rejection values of
the contaminants listed in Table 1. All the test work was conducted in a
batch mode. Data were analyzed according to the solution-diffusion model
describing RO water and solute transport5'*>) as outlined below.
EXPERIMENTAL AND THEORETICAL APPROACH AND RESULTS
WATER TRANSPORT
Membrane water flux rate is a function of the effective pressure applied
across the membrane described by the following equation:
F = W (AP - AIT)
w p
(1)
324
-------
where
2
FW = membrane water flux rate, gpd/ft
W = water permeability coefficient, gpd/ft2/psig
AP = differential applied pressure across membrane, psig
ATT = differential osmotic pressure across membrane, psig
Membrane water flux rates declined with increasing water recovery levels due
to increasing solution osmotic pressure caused by membrane solute rejection.
Observed flux declines with water recovery are presented in Figures 5 and 6.
Operating pressures were varied at 350, 400, and 450 psig levels with feed-
water temperatures ranging from 74°F to 86°F.
10
x
oc
Ul
S
UJ
S
OPERATING FEED
PRESSURE. TEMPERATURE
RUN NO. psig °F
350
350
400
400
74
86
76
76
10
20
80
90
30 40 50 60 70
WATER RECOVERY. %
Figure 5. MEMBRANE WATER FLUX DECLINE WITH WATER RECOVERY
325
-------
11
10
!
x
f
z
c
CD
Z 7
OPERATING FEED
PRESSURE, TEMPERATURE.
RUN NO. psig _°F
400
450
450
450
74
74
83
84
10
20
30
40 50 60
WATER RECOVERY. %
70
80
90
Figure 6. MEMBRANE WATER FLUX DECLINE WITH WATER RECOVERY
Water flux rates are also influenced by temperature because of the depen-
dence of the water permeability coefficient on this parameter. Water perme-
ability coefficients for the scrubber-water samples were determined by
plotting water flux rates against effective applied pressure in accordance
with Equation (1). These results are shown in Figure 7. Water permeability
coefficient values determined in this manner are listed in Table 4.
326
-------
t
u.
X
Fw = W
(.&/
( I RUN NO.
200 300 400 450
EFFECTIVE APPLIED PRESSURE. (aP -*i:|. psig
Figure 7. DETERMINATION OF THE WATER PERMEABILITY COEFFICIENT. Wp
Good agreement was obtained between values of each control run at
400 psig and 74°F, indicating no significant decline in membrane water flux
rates over the test period. Agreement was also obtained between values of the
control run at experimental conditions and the experimental runs. This indi-
cates that differences observed between each of the experimental runs were due
to temperature effects and not to differences in the individual samples. This
temperature dependence is displayed in Figure 8. Using linear regression, the
magnitude of the water permeability coefficient temperature dependence was
determined as 5.4 x 10~4 qpd/ft2.°K for the control runs at the experimental
conditions, and 6.2 x 10"^ gpd/ft^.°K for the gas-scrubber water samples.
These values agree within experimental error.
327
-------
111
8
m
DC
UJ
a.
<
DISTILLED WATER CONTROL
AT EXPERIMENTAL CONDITIONS
296 297 298 299 300 301 302
ABSOLUTE TEMPERATURE. "K
303
304
Figure 8. TEMPERATURE DEPENDENCE OF THE WATER PERMEABILITY COEFFICIENT. Wp
Water flux rates for a model full-scale RO unit were estimated by deter-
mining median flux rates from each of the flux decline runs. These rates were
corrected for the observed temperature'range (60°F to 90°F) of the gas-
scrubber recycle system using the relationship derived above and are presented
in Table 5.
SOLUTE TRANSPORT
Membrane solute flux rate is a function of the solute concentration
differential across the membrane as shown by the following equation:
F. = k (C - C )
i p c p
(2)
where
Fi = solute flux, g/day/ft2
k = solute permeability coefficient, I/day/ft''
CG = concentrate solute concentration, g/1
C = permeate solute concentration, g/1
328
-------
Total dissolved solids (TDS) measurements were made to determine solute flux
rates across the membrane in accordance with Equation (2). TDS flux rates
were calculated in accordance with Equation (3).
(3)
w,p
where
FW = membrane water flux, g/day/ft
C = permeate solute concentration, mg/1
GW = permeate pure water concentration, assumed to equal 10 mg/1
TDS permeability coefficients were calculated by plotting TDS flux rates
against the membrane concentration difference as shown in Figures 9 and 10.
The permeability coefficients obtained in this manner ranged from 1.10 litres
per day per square foot (I/day/ft2) to 1.83 I/day/ft2.
24
20
16
2
to
4 8 12 16
MEMBRANE TDS CONCENTRATION DIFFERENCE.
(Ce-Cp),«/«
Figure 9. DETERMINATION OF THE TDS PERMEABILITY COEFFICIENT. Kp
329
-------
28
24
20
1
|
uT 16
x"
12
IU
Fi-KplC -Cp)
RUN NO.
4
5
6
7
8
KP.
^/day/ft*
1.11
1.53
NO DATA
1.68
1.25
0 4 8 12 16 20
MEMBRANE IDS CONCENTRATION DIFFERENCE, (Cc - Cp),g/C
Figure 10. DETERMINATION OF THE TDS PERMEABILITY COEFFICIENT. Kp
Tests to determine rejection values for various wastewater contaminants
were also conducted. These results are presented in Table 6. Negative
rejections were obtained for cyanide/ phenol, and sulfide, indicating
preferential sorption of these materials across the membrane. Thiocyanate
also displayed low rejection levels. Rejection values obtained for other
contaminants examined were found to compare favorably with reported ranges in
the literature.
PERMEATE WATER QUALITY
Permeate solute concentration increases directly with feed solute concen-
tration and water recovery level. Figures 11 and 12 displays these relation-
ships for the gas-scrubber recycle-water samples. Initial TDS concentrations
of the samples ranged from approximately 2800 mg/1 to 6100 mg/1 resulting in
permeate concentrations from 300 mg/1 to 600 mg/1 at a 70 percent
330
-------
10
2 8
X
1
o
c
t-
z
UJ
O
5
cc
RUN NO.
INITIAL FEEDTDS
CONCENTRATION, mg/V
10 20 30 40 50 60
WATER RECOVERY. %
70
80
90
Figure 11. PERMEATE WATER QUALITY DECLINE WITH WATER RECOVERY
10
I
a
4
UJ
X
INITIAL FEEDTDS
RUN NO. CONCENTRATION,
102030405060708090
WATER RECOVERY. X
Figure 12. PERMEATE WATER QUALITY DECLINE WITH WATER RECOVERY
331
-------
water-recovery level. For water-recovery levels above about 70 percent, per
meate concentrations increased dramatically.
CONCENTRATION POLARIZATION AND FOULING EFFECTS
Concentration polarization is defined as the ratio of solute concentra-
tion at the membrane surface to solute concentration in the bulk solution.
This may be expressed mathematically by
B= (4)
B
where
B = concentration polarization
Cm = solute concentration at membrane surface
Cg = average bulk solute concentration
It is caused by concentrate solute concentration building up at the membrane
surface exceeding the bulk solution concentration. When this occurs the
solute tends to back-diffuse away from the membrane into the bulk solution.
Concentration polarization increases the local osmotic pressure at the mem-
brane surface, reduces membrane water flux rates, increases permeate solute
concentration, and allows the concentration increase of sparingly soluble
solutes, possibly causing precipitation, onto the membrane. It is enhanced by
high recovery levels but may be minimized by recirculation of the concentrate
stream. The magnitude of concentration polarization cannot be determined
directly because measurement of the membrane surface solute concentration is
required. However, the presence of concentration polarization or fouling may
be detected experimentally. As the solute concentration at the membrane
surface increases so does the local osmotic pressure. Osmotic pressure
increases more rapidly as a result of concentration polarization or fouling
than would be expected simply from increases in concentrate solute concentra-
tion .
A plot of osmotic pressure versus concentrate solute concentration may be
used to detect concentration polarization or fouling, as shown in Figure 13.
Figure 14 is a similar plot for each of the scrubber-water samples. Linear
relationships were established for each of the samples, indicating that con-
centration polarization and fouling effects were negligible.
332
-------
Ill
IT
O
O
CONCENTRATION
POLARIZATION
OR FOULING
NO CONCENTRATION
POLARIZATION OR
FOULING
CONCENTRATE SOLUTE CONCENTRATION •
Figure 13. DETECTION OF CONCENTRATION POLARIZATION OR FOULING
UJ
a.
160
140
120
100
80
60
40
(8)
( ) RUN NO.
(2
6 8 10 12 14 16
CONCENTRATE TDS CONCENTRATION, g/t
18
20
Figure 14. EVALUATION FOR THE PRESENCE OF CONCENTRATION POLARIZATION OR FOULING
APPLICATION OF RESULTS TO HYPOTHETICAL DESIGN
Calculations were prepared of the permeate and concentrate water quality
from a model full-scale RO unit treating 200 gallons per minute (gpm) of gas-
scrubber blowdown water. The rejection data contained in Table 6 were used
instead of determining median rejection values since it was desired to obtain a
conservative estimate of the contaminant loadings in the concentrate stream.
These calculations are based on the average concentration values shown in Table 2.
333
-------
Calculated contaminant loadings are presented in Table 7. Using conser-
vative rejection values we find that phenol and total cyanide discharge
loadings are still within proposed BAT limits while zinc and ammonia loadings
exceed these limits.
Calculated reductions in contaminant discharge loadings, for a single
pass through the RD unit and various water recovery levels, are presented in
Table 8.
SUGGESTIONS FOR THE TREATMENT OF REVERSE OSMOSIS CONCENTRATE
Disposal of a portion of the RO concentrate may be accomplished by evapo-
ration via quenching hot blast-furnace slag.
Treatment by alkaline chlorination and ozonation has been investigated by
Osantowski and Geinopolos.1) Both processes were found capable of achieving
proposed BAT limitations and are outlined in Figure 15. Ammonia was found to
BLAST FURNACE OFF-GAS
RECYCLED SCRUBBER WATER
Figure 15. TREATMENT OF REVERSE OSMOSIS CONCENTRATE
334
-------
be the critical parameter for both processes such that if ammonia concentra-
tions were reduced below BAT limits, all other oxidizable pollutants were also
reduced below their respective limitations.
RESULTS AND CONCLUSIONS; SUMMARY
Based on the results of this research we may conclude that:
1. Reverse osmosis with cellulose acetate membranes appears capable of
effecting significant reductions in discharge volume and contaminant
loadings from a blast furnace gas-scrubber recycled-water system.
Cellulose acetate membranes displayed preferential sorption of phenol
(-14%), free cyanide (-3%), and sulfide (-14%) with low rejection of
thiocyanate(8%).
2. Hypothetical plant calculations indicated that phenol and cyanide
discharge loadings were below BAT limits.
3. High rejections of zinc (>99%) and ammonia (93%) were obtained, indi-
cating the need for additional treatment for their removal prior to
discharge. It should be noted that zinc concentrations in the samples
collected were unusually high because of the composition of sinter and
scrap used as blast-furnace charge. Rejections of calcium, magnesium,
sulfate (>99%) and chloride (94%) were also high, assuring the reusa-
bility of the permeate stream in the recycle system.
4. Water recoveries ranging from 70 percent to 80 percent appear to be
possible, although permeate water quality decreases dramatically for
water recoveries exceeding 70 percent.
5. Water flux rates appeared to be adequate (5.9 to 10.4 gpd/ft at
400 psig) although because of the temperature dependence of these flux
rates, sizing of a full-scale unit should carefully take into account
temperature variations.
6. Concentration polarization and membrane fouling did not appear to be
significant. However, calcium sulfate scaling could be a major opera-
tional problem. Appropriate pre-treatment procedures, such as disper-
sant and anti-precipitant addition, may be required to prevent fouling
of the RO membrane.
7. Studies to determine membrane permeate flux and rejection decline were
not conducted because of sample limitations and the validity of
performing these studies in a batch mode. These studies would have to
be made in order to determine the economic feasibility of RO in this
application.
8. Because of differing physical and operational characteristics of
blast-furnace gas-scrubber recycle systems, additional work.would be
required to establish the viability of RO in this application.
Recycle-water chemistry is not fully understood nor are the
335
-------
operational parameters which affect it. Variabilities of contaminant
concentrations would have to be determined to properly size an RD unit
and to provide protection against membrane fouling.
Q)*
Additional data for this study may be found in the work of Terril.
REFERENCES
1. Proceedings; First Symposium on Iron and Steel Pollution Abatement
Technology, Chicago, October 30 - November 1, 1979, "Physical-Chemical
Treatment of Steel Plant Wastewaters Using Mobile Pilot Units, by Richard
Osantowski and Anthony Geinopolos" (Research Triangle Park: U. S.
Environmental Protection Agency, EPA-600/9-80-12, February 1980),
pp. 325-340.
2. Chian, E. S. K., and Fang, H. H. P., "RO Treatment of Power Plant Cooling
Tower Slowdown for Reuse," AICHE Symposium Series-Water, Vol. 71, No. 151,
(1975), pp. 82-86.
3. Kosarek, L. J., "Significantly Increased Water Recovery From Cooling Tower
Slowdown Using Reverse Osmosis," AICHE Symposium Series-Water, Vol. 75,
No. 190 (1979), pp. 148-155.
4. Development Document for Effluent Limitations Guidlines and Standards for
the Iron and Steel Manufacturing Point Source Category, Vol. 11
(Proposed), by D. M. Costle et al., U. S. Environmental Protection Agency,
EPA 440/1-80-024b, December 1980.
5. Merten, U., ed., Desalination by Reverse Osmosis, (Cambridge: M.I.T.
Press, 1966), pp. 93-160.
6. Weber, Jr., W. J., Physicochemical Processes For Water Quality Control,
(New York: Wiley-Interscience, 1972), pp. 307-329.
7. Spatz, D., Industrial Waste Processing With Reverse Osmosis, (Hopkins,
Minn.: Osmonics, Inc., 1971), unpublished.
8. Lonsdale, H. K., Merten, U., and Tagami, M., "Phenol Transport in
Cellulose Acetate Membranes," Journal of Applied Polymer Science, Vol. 11,
No. 9, (September, 1967), pp. 1807-1820.
9. Terril, M. E., The Applicability of Reverse Osmosis to the Treatment of
Blast Furnace Gas Cleaning Water Recycle System Slowdown, M. S. Thesis,
University of Pittsburgh, Pittsburgh, PA, December 1980.
* It is understood that the information in this paper is intended for general
information only and should not be used in relation to any specific applica-
tion without independent examination and verification of its applicability and
suitability by professionally qualified personnel. Those making use thereof
assume all risk and liability arising from such use or reliance.
336
-------
TABLE 1. GAS-SCRUBBER RECYCLE-SYSTEM PROPOSED DISCHARGE LIMITATIONS
lb/1000 Ib IRON PRODUCED (mg/£)*
Total Suspended
Solids
Cyanide (Total)
Phenol
Ammonia
As Nitrogen
Oil & Grease
Zinc
Lead
BPT
0.026 (50)
0.0078 (15)
0.0021 (4)
0.0535 (103)
BAT
BCT
0.000292 (1)
0.0000292 (0.1)
0.000292 (1)
0.0000876 (0.3)
0.0000730 (0.25)
0.00438 (15)
0.00292 (10 max)
* Concentration calculated based on 125 gallons per ton (gpt) iron
produced for BPT limits/ and 70 gpt iron produced for BAT and
BCT limits.
337
-------
TABLE 2. CONTAMINANT CHARACTERIZATION OF BLAST-FURNANCE
GAS-SCRUBBER RECYCLE SYSTEM
Average
Standard
Concentration Deviation
Contaminant
Calcium
as CaC03
Magnesium
Zinc
Conductivity
umhos/cm^
Total Cyanide
Free Cyanide
Thiocyanate
Phenol
Ammonia
as Nitrogen
Fluoride
Chloride
Sulfate
Sulfide
mg/1
358
47
27
4650
3.4
3.4
0.4
0.8
128
25
1125
554
0.2
mg/1
76
10
24
1330
2.8
3.2
0.6
0.8
31
6
278
348
0.2
Maximum
mg/1
728
60
79
8200
9.0
9.1
0.8
2.3
158
38
1621
1265
0.3
Minimum
mg/1
190
34
5
1880
0.42
0.33
<0.01
0.08
55
19
815
186
<0.05
Numbe r of
Observations
91
9
9
91
9
7
2
9
9
9
9
7
2
338
-------
TABLE 3. OSMONIC SEPA-97 MEMBRANE CHARACTERISTICS
Membrane Material
NaCl Rejection
f\
Permeate Flux, gpd/ft
(Tap Water)
Maximum Operating Pressure
Normal Operating Pressure
pH Range
Surface Area
Cellulose Acetate
94-97%
10-14@ 400 psig
800 psig
400-500 psig
2-8
19 ft2
TABLE 4. WATER PERMEABILITY COEFFICIENTS,
Wp, FOR CONTROL AND EXPERIMENTAL RUNS
Run
Number
1
2
3
4
5
6
7
8
Distilled
Water;
74°F, 400 psig
(gpd/ft2/psig)
0.0251
0.0258
0.0267
0.0263
0.0255
0.0264
0.0257
0.0261
Distilled Water,
Experimental
Condition
(gpd/ft2/psig)
0.0251
0.0281
0.0267
0.0263
0.0255
0.0264
0.0289
0.0301
Gas-Scrubber Water,
Experimental Conditions
(gpd/ft2/psig
0.0251
0.0286
0.0268
0.0263
0.0256
0.0258
0.0295
0.0300
339
-------
TABLE 5. ESTIMATED WATER FLUX RATES FOR A MODEL FULL-SCALE RO UNIT
Operating Water Flux
Pressure, Rate,
psig gpd/ft2
60°F 90°F
350 5.1 - 9.0
400 5.9 - 10.4
450 6.8 - 12.6
* Osmotic pressure of the gas-scrubber water samples ranged from
40 psig to 90 psig which is accounted for in the stated water
flux ranges.
340
-------
TABLE 6. OBSERVED CONTAMINANT MEMBRANE REJECTION
Solute
Calcium
As CaCO,
Magnesium
Zinc
Total Dissolved
Solids
Total Cyanide
Free Cyanide
Thiocyanate
Phenol
Ammonia
Fluoride
Chloride
Sulfate
Sulfide**
Average
Rejection, %
>99
>99
>99
97
-1
-3
8
-14
93
91
94
>99
-14
Observed
Rejection
Range , %
>99
>99
>99
96 to 98
-12 to +12
-7 to +6
-10 to +24
-18 to -10
91 to 94
89 to 92
92 to 96
>99
_
Reported
Rejection
Range / %
96.3 to 99.76)*
93 to 99.9 6>
89 to 996)
-
O for pH <77)
-
-20 to -108)
77 to 956)
88 to 986)
86 to 976>
99 to 1006)
_
* Low.concentrations responsible for broad rejection range
** Only one observation was above maximum
analytical sensitivity limits (0.05 mg/1)
*** See References
341
-------
TABLE 7. CALCULATED DATA: APPLICATION OF RO TO
BLAST-FURNACE GAS-SCRUBBER RECYCLE-SYSTEM SLOWDOWN TREATMENT
oo
-Pi.
r\>
Contaminant
Water Recovery
Calcium
As CaCO3
Magnesium
Zinc
Total Cyanide
Free Cyanide
Thiocyanate
Phenol
Ammonia
As Nitrogen
Fluoride
Chloride
Sulfate
Sulfide
Permeate
Contaminant
Mass Flow Rate
Ib/day
70%
19
2.6
1.5
5.7
5.8
0.66
1.4
43
10
332
30
0.3
80%
33
4.3
2.5
6.6
6.6
0.76
1.6
67
16
523
51
0.4
90%
71
9.3
5.4
7.4
7.4
0.86
1.8
119
27
948
110
0.4
Concentrate
Contaminant
Mass Flow Rate
Ib/day
70%
841
110
63
2.5
2.4
0.30
0.5
265
50
2372
1302
0.2
80%
827
108
62
1.6
1.6
0.20
0.3
241
44
2181
1281
0.1
90%
789
103
59
0.8
0.8
0.10
0.1
189
33
1756
1222
0.1
Contaminant
Mass Discharged BAT Limitations
1bs./1000 Ibs. lbs./1000 Ibs.
Hot Metal x 10~4 Hot Metal x 10~4
70%
837
110
63
2.5
2.4
0.9
0.5
264
50
2361
1296
0.2
80%
824
108
62
1.6
1.6
0.9
0.3
240
44
2171
1275
0.1
90%
786
103
59
1.3
1.3
0.9
0.1
188
33
1748
1216
0.1
-
-
0.876
2.92
-
-
2.92
2.92
-
-
-
w
-------
TABLE 8. CALCULATED CONTAMINANT DISCHARGE LOADING REDUCTIONS,
PERCENT (SINGLE PASS)
Contaminant
Calcium As
CaC03
Magnesium
Zinc
Total Cyanide
Free Cyanide
Thiocyanate
Phenol
Ammonia As
Nitrogen
Fluoride
Chloride
Sulfate
Sulfide
Water Recovery Level
70%
2.3
2.3
2.3
70
70
69
74
14
17
12
2.3
60
80%
3.8
3.8
3.8
80
80
79
84
22
27
19
3.8
80
90%
8.2
8.2
8.2
90
90
90
95
39
45
35
8.2
80
343
-------
REVIEW OF WATER USAGE IN THE IRON
AND STEEL INDUSTRY: BLAST FURNACE
AND HOT FORMING SUBCATEGORIES
by: Albert P. Becker, III and
Thomas M. Lachajczyk
Envirodyne Engineers, Inc.
12161 Lackland Road
St. Louis, MO 63141
ABSTRACT
This paper presents the results of a study of water usage and recircula-
tion techniques in the iron and steel industry blast furnace and hot
forming subcategories. The project was conducted for USEPA's Industrial
Environmental Research Laboratory, Research Triangle Park, North Carolina.
Initially, a list of zero and low discharge plants in the subject sub-
categories was developed from information provided in the "Draft
Development Document for Proposed Effluent Limitations Guidelines for the
Steel Industry". USEPA wished to verify the water application and dis-
charge rates presented in the Development.Document and update this informa-
tion if necessary. The work also included a determination of specific
process water quality requirements in blast furnace and hot forming
operations, an identification of factors which limit process water re-
cyclability, a survey of techniques used to implement zero or low discharge
systems, and a study of .the feasibility of implementation of these
techniques at existing plants. The study included telephone and written
communications with plant personnel, American Iron and Steel Institute
representatives, EPA Regional and Effluent Guidelines staff members, and
equipment manufacturers, a review of technical literature, and plant
visits.
INTRODUCTION
The United States Environmental Protection Agency (USEPA) had ob-
tained information during the Effluent Guideline Development process which
indicated that a number of plants in several major industrial sub-
categories of the iron and steel industry were operating at zero or near-
zero water discharge. However, Effluent Limitation Guideline flow levels
for these subcategories were set above zero discharge because the zero
discharge claims had not been evaluated and substantiated by USEPA. In
many cases, it was suspected that claims of zero discharge were due to
inaccurate reporting or unique system operation or design.
344
-------
In order to further investigate water usage in the iron and steel
industry, USEPA's Industrial Environmental Research Laboratory requested
Envirodyne Engineers, Inc., to conduct a project entitled "Verification
and Documentation of Low Water Discharge Steel Processes". Since the
available budget to fund this work was limited, the scope of work was
restricted to a study of the blast furnace and hot forming categories
in order to allow a thorough review. Investigation of these particular
industry subcategories received high priority because of the relatively
large number of plants at which these operations take place and high
volumes of water utilized.
The investigation of zero and low water usage was designed to
identify existing systems which are truly exemplary and which might
serve as a model to improve Best Available Technology (BAT) Guidelines.
Initially, information presented in the Draft Development Document for
Proposed Effluent Limitations Guidelines for the Steel Industry was used
to draw up a list of plants conducting blast furnace and/or hot forming
operations which reported a process wastewater discharge rate less than
or equal to the recommended BAT model flow rate. The Development
Document recommends a BAT discharge rate of no more than 50 gallons per
ton (gpt) for the Blast Furnace Subcategory. The Hot Forming Subcategory
is further subdivided into primary (with and without scarfing), section
(carbon and specialty steels), flat (strip and sheet, carbon plate, and
specialty plate) and pipe and tube segments. Recommended BAT flow rates
for hot forming operations range from 60 to 260 gpt, depending on the sub-
category. The initial list of all plants with a flow less than or equal
to the recommended BAT level (as indicated in the Development Document)
was quite lengthy, and was shortened at the suggestion of the project
officer to include only those plants which reported a discharge flow of
at most 50 percent of the recommended BAT discharge rate. This list
provided a numerically-manageable group of plants which had apparently
achieved exemplary water use and conservation.
In order to supplement the list and identify additional plants which
might be exemplary contact was made with the Cyrus W. Rice Division of
NUS Corporation (which has previous experience as a. contractor to USEPA
in the iron and steel industry). The American Iron and Steel Institute
and USEPA personnel were also contacted to identify exemplary plants and
discuss water conservation techniques. These contacts allowed the
study of several plants which had made improvements since the Draft
Development Document was published.
In addition to contacts with individuals mentioned above, several
additional approaches were utilized to verify and update water usage
reported in the Development Document. Analysis of individual Section 308
responses helped clarify some of the information presented. Information
published recently in technical literature updated information relative
to specific plants. Telephone and written communication with representa-
tives of each of the plants believed to be exemplary was initiated in
order to confirm and update information obtained from other sources,
1 discuss the application of treatment technologies, and inquire about the
345
-------
possibility of a voluntary plant visit. Direct contact with plant
representatives proved to be the most useful and cost-effective means of
updating information concerning individual facilities. During the course
of the study, almost every plant on the original list of supposedly
exemplary plants was contacted directly by phone and/or in writing.
As a culmination of the study, several plant visits were conducted.
The plants selected for visits represented a wide range of sizes, ages,
and status with respect to water usage. The telephone, written, and in-
person visits with plant personnel provided an excellent opportunity to
discuss factors which affect process water recyclability, process water
quality requirements, technology useful for reduction of process water
usage, and the feasibility of further reductions in water usage and
discharge.
The following sections summarize results of the study and provide
conclusions and recommendations to both USEPA and industry concerning
rule-making and water conservation techniques.
REVISED STATUS OF WATER USAGE AT INDIVIDUAL PLANTS
In Table 1, the water usages presented in the Development Document
are compared with revised water usages calculated from updated production
figures and flow rates, supplied by plant representatives during the
plant surveys. It can be seen that in many cases the revised figures
are considerably higher than those originally reported. Several factors
were found to.contribute to this situation, including:
1) In some cases, there was a misinterpretation of the Section 308
Questionnaire and subsequent inaccurate responses by plant
respondants.
2) Calculation of water usage in gallons per ton was based on capacity
rather than actual production. With many plants operating at only
50-75 percent of capacity recently, but with no change in
water application rates, calculated water use per ton of pro-
duction increased accordingly.
3) Water losses through holding pond seepage or other leaks were
not considered or included in water use calculations.
4) There have been changes in operations at some plants since the
308 response was submitted.
5) Several once-through systems were mistakenly representated as
zero discharge.
6) In some cases represented as zero discharge, process wastes
were actually cascaded to other operations and subsequently
treated and discharged.
346
-------
TABLE 1. UPDATED HATER DISCHARGES AND APPLICATION RATES
Water Discharge
Rate Presented in
Plant/Code
Blast Furnaces
732A
920B
856Q
860H
860B
448A
Hot Forming-Primary
00601
0940A
0440A
0492A
684H
088D
248B
0060D
176
0612-01
860H-02
Hot Forming-Section
460A
672A
316B
396D
684H-01
684H-03
684H-06
684H-07
0068B
0060K-01
860H-03
00601-01
00601-02
672B-01
672B-02
684H-02
088D
60F-05
384A-06
468B
612-01
612-04
432-A
256N
316A
248B
Flat-Strip and Sheet
248B
112D
Pipe and Tube-Hot Worked
0060R-01
0060R-02
684H
728
240B-05
856-C
(gpt)
0
?
0
0
?
0
0
0
0?
96
41
96
4
36
78
65
0
0
0
0
0
0
0
0
0
35
3.5
0
0
0
0
0
73
59
79
89
84
84
?
0
0?
92
0? U
0
0
0
82
0-231?
100
Updated
Water Discharge
(gpt)
Updated
Water Application
(gpt)
4 to 28l ' 3,800
20 1,200
0(unknown amount) 4,230
=0 8=1,620 11=2,760
10=3,100 12=3,530
0 (144)
(b)
0 (131,99)
(b)
2,200
2,600 3,300
,000
700 5,
No information could be found
0,_> 124,000
4,850(C>
100
11.,.
88(d)
*
178
71-120
108
0
-------
.TABLE 1. UPDATED WATER DISCHARGES AND APPLICATION RATES
(Continued)
NOTES: (a)
4 gallons per ton reported by T. Oda, based on preliminary
plant studies; 28 gallons per ton based on review of Data
Collection Portfolios. Both values refer to discharge
resulting from gas cooling and cleaning system only.
Value in parenthesis is amount of evaporative blowdown off
of slag.
(c)
Once-through system.
Average value for all alloy mills which include a blooming
mill with scarfing, bar mill, merchant mill and billet mill.
(e)
Suspected seepage problem.
Cascading recycle system for which plant personnel could not
estimate a system blowdown.
(q)
These are cold worked-pipe and tube mills.
(*)
A letter has been sent to plant personnel requesting updated
information and no response has been received.
(?)
Indicates unavailable or could not be ascertained.
7) In general, due to complicated water systems, it is usually
difficult to establish exact water usage per production,
and these figures are at best estimates subject to change.
The survey pointed out the relationship between site-specific production
considerations and water application and discharge rates at individual plants.
For example, water usage in section mills is dependent upon the desired cross-
sectional area reduction of the original blooms, slabs or billets. If two
mills begin with identical billets but produce products of different
diameters, the mill producing the smaller diameter product would use more
water per ton of steel, because more passes are required to achieve this
smaller diameter. This fact may not have been fully considered in development
of BAT models.
Similarly, in the production of hot formed pipe and tube, the larger
the pipe diameter, the greater the water usage for direct contact cooling.
If two mills start with the same 'raw material, a mill producing large
diameter pipes would use more water per ton of product than one that
produces small diameter pipes.
348
-------
WATER QUALITY REQUIREMENTS
BLAST FURNACE SUBCATEGORY
Table 2 illustrates the average water quality characteristics of
raw and treated process wastewater in this subcategory. This information
was derived from analytical data collected during the original guidelines
survey and supporting sampling and analysis program. Some typical values
for these parameters in an exemplary blast furnace recycle system are
also given to provide additional water quality characteristics.
As this data indicates, reduction of the suspended solids concentra-
tion is essential to the gas cleaning scrubber operation. A total sus-
pended solids concentration of 50 milligrams per liter or less is optimal.
The scrubbing of the dirty blast furnace gases results in a marked
increase in the alkalinity, hardness, total dissolved solids, total sus-
pended solids, chloride and sulfate concentrations of the process water.
HOT FORMING SUBCATEGORY
The average water quality characteristics of raw and treated process
wastewater as determined from analytical data collected during the original
guidelines development and supporting sampling and analysis program are
presented in Table 3.
This data indicates that in the treated process wastewater of the
hot forming subcategory, total suspended solids concentrations of 40
milligrams per liter or less and grease and oil concentrations of 15 milli-
grams per liter or less are typical. Primary, section and plate mills
usually have higher grease and oil concentrations in their process
waters than do hot strip mills or pipe and tube mills.
Plant representatives indicated that the following water quality
requirements were important in their recycle systems:
1) Constant water quality is important if a consistently good pro-
duct is to be produced.
2) Removal of water soluble oils is important because they can foul
heat exchangers and cooling towers as well as inhibit the roll's
ability to grip the hot metal.
3) Total suspended solids concentrations must be kept to a minimum
as they can foul bearings. Larger solids can clog spray
headers.
4) Some plants chemically treat to keep dissolved solids concentra-
tions at acceptable levels.
5) Water temperatures ranging from ambient to 180°F have been used
successfully.
349
-------
TABLE 2. BLAST FURNACE-SUBCATEGORY WATER QUALITY REQUIREMENTS
GO
on
O
Parameter
(a)
pH (units)
Total suspended solids
Total dissolved solids
Hardness CaCC>3
Alkalinity CaCO3
Phenolphthalein
Methyl orange
Chloride
Sulfates
Ammonia
CN~
Phenols
Fluoride
Average Values from Original
Guidelines and Toxic Pollutant
Survey
Plant 448-A
Raw Process
Wastewater
6.4-10.2
1,038
7^
15
1.8
33
Treated Process
Wastewater
6.7-10.9
51
66
9
1.4
32.5
Blast Furnace
Make-up Water
7.3
39
550
200
0
50
170
100
Blast Furnace
Recycle Water
7.1
50-300
3,000
900
220
1,100
370
NOTES: (a)
All results in mg/1 except as noted.
-------
TABLE 3. HOT FORKING-SUBCATEGORY WATER QUALITY CHARACTERISTICS
Parameter
Total Suspended
Solids
Oil and Grease
Primary
T(b) fb)
I 0
83 27
41 9.3
Hot Strip
Section and sheet
10 10
67.8 35 47 23.9
28.3 8.8 12 12
Plate
I
75
39
0
14
16
Pipe and
Tube
I 0
89 40
9.1 5.4
NOTES: (a) nn ^. . /n
All concentrations in mg/1.
(b)
I = Raw Wastewater input to treatment system
O = Effluent from treatment plant to be recycled
6) Bacterial problems have been noted and remedied by chemical
addition in some closed loop systems.
7) In some plants, raw water supplies must be softened before use.
Representatives of Plant 248B, who operate a very tight recycle loop
about their hot forming mills, indicated that the following water quality
requirements are important.
1) Manufacturers have recommended that they operate, their hot strip
mill with water that has a total suspended solids concentration
of 20 ppm or less, but they have successfully operated their
mill with total suspended solids ranging from 30 to 200 ppm.
2) Total dissolved solids concentrations in their recycle loop have
stabilized, and generally range from 700 to 900 ppm. They
have experienced no fouling problems.
3) Water temperatures of less than 120°F have been specified by
manufacturers but they have operated at temperatures as high
as 180°F without any problems.
4) Grease and oil concentrations are kept at 5 ppm or less.
A summary of water quality requirements specified by an equipment
manufacturer and by six iron and steel plants for the hot forming
category is provided in Table 4. The most closely controlled operating
parameters seem to be total suspended solids, total dissolved solids,
grease and oil, chlorides, and sulfates. Toxic metals do not seem to
present any major problems to final product quality.
351
-------
TABLE 4. HOT FORMING--SUBCATEGORY WATER QUALITY REQUIREMENTS
Parameter
Phenol
NH, (total)
ctr
Chromium (total)
Copper
Iron (dissolved)
(total)
Lead
Nickel
Zinc
Sodium
BOD5
Nitrates
584-F 868-ft
Hot Strip Hot Forming 948-C
Manufacturer Run-out . Mills Run-out Hot Rolling Mills 448-B
Recommendations Table Spray Table Spray Hot Strip Cooling Hot Rolling Mills
7.0 7.7
<25 37 25 40 50
175 450 300
120
70
10 21
<100 61 50
<300 78 143 70
0.015 0.03
0.8 3.5
0.037 0.11
0.04
0.63
11.4
0.06
0.20
20
3.0
0.7
248B German
Hot Strip Mill Iron & steel Plant
Rolling Mills Hot Rolling Mills
120-180 86
6.5-9.5
30-200 1.0
700-900
5 5
1.0
4.0
3.0
no limit
5.0
5.0
NOTES: (a)
All values in mg/1 excpet as noted.
-------
Process water temperatures vary widely from plant to plant depending
on individual recycle rates per ton of hot formed product. All that is
required is that the combination of water temperature and velocity be
adequate to cool the rolls and in some cases the hot formed product.
FACTORS WHCH LIMIT RECYCLABILITY
This section contains a discussion of the general requirements that
must be met to implement any recycle system, followed by some particular
points that are specific to the blast furnace and hot forming subcategories.
GENERAL CONSIDERATIONS
Plant Layout
It is essential that process wastewaters, non-contact wastewaters,
and storm water systems remain separate if high recycle ratios are to be
attained. Segregation and separate treatment of incompatible waste streams
also enhances recyclability.
Economics of Water Supply Versus Treatment
Depending on the source of raw water, it may be more economical to
operate a once-through water system rather than treat and reuse.
Chemistry of Raw Water Supply
The blowdown necessary to avoid scaling, corrosion and fouling is
dependent on the minerals present in the raw water supply.
Hydraulic Balance
The maintenance of a hydraulic balance in^ various portions^ of a
closed loop system is necessary to avoid overloading the system. The
system must have the capacity to handle sudden surges resulting from ab-
normal, necessary introductions of water to the system. If this capacity
is not provided for, excessive blowdowns may be required to avoid flooding.
In general, systems which are physically spread out over many acres and
thereby provide for settling, cooling, and evaporation tend to reduce
blowdown and maintain a high recycle rate.
Manpower Requirements
The plant must have the ability and resources (in terms of manpower)
to constantly seek out and correct the sources of hydraulic leaks into the
closed loop system. Otherwise, the recycle system may eventually be over-
loaded, resulting in the degradation of water quality or the necessity
for significant blowdown.
353
-------
Water Infiltration
Set procedures must be established to avoid the intentional intro-
duction of new water into the closed loop system. For example, if an acci-
dent occurs that requires water for cooling, it should be pumped from
within the system rather than utilizing fresh water. Plant workers must
be fully instructed as to the proper procedures necessary for the main-
tenance of the closed loop.
Water Losses
If an earthen lagoon is used for solids settling, cooling or water
storage, it must be properly designed to avoid possible losses which can
result through seepage and/or overflow during periods of heavy rain.
Foreign Materials
The operation of the recirculation system is susceptible to the intro-
duction of foreign materials (hardhats, cans, light bulbs, etc.). These
have the potential to block flow and upset the system.
SPECIFIC CONSIDERATIONS
Blast Furnace Subcategory
The major process wastewater source in the blast furnace operation is
the water used.in cleaning dirty blast furnace gases. In blast furnace gas
cleaning recycle systems, the following factors limit the ability of plant
personnel to recycle a higher percentage of their wastewater.
Scaling is a major problem in most blast furnace gas cleaning systems.
Scaling has been known to plug spray nozzles, reduce the effective area
in venturi throats, and clog supply pipes. The precipitating material
that causes this scaling is usually calcium carbonate. The tendency of
the recycle system towards calcium carbonate scaling can be determined
through the use of the Langelier Saturation Index (LSI). In the LSI
system, a zero value indicates that the system is in equilibrium in re-
gards to calcium carbonate. A positive index value indicates that thsre
is an abundance of calcium carbonate in solution; to achieve equilibrium,
the calcium carbonate will precipitate out, resulting in scaling. If the
LSI is negative, there is a shortage of calcium carbonate in solution
resulting in corrosion of pipes as the system tries to reach equilibrium.
The contribution of calcium and alkalinity from the limestone in the flux
and the carry-over of calcium chloride results in the precipitation of
calcium carbonate whenever the solubility limits are exceeded in the
blast furnace gas cleaning system. In many blast furnaces, the LSI is kept
stable by a system blowdown and the addition of acid or anti-scaling agents.
The Ryznar Stability Index (RSI) has also been used at plants to
monitor corrosion and scaling tendencies. Calculation of the RSI is based
on five basic water quality parameters: alkalinity, salinity, calcium
concentration, pH, and temperature. Water with an RSI value greater than six
354
-------
is considered corrosive, while scaling will occur when the RSI is less than
six. Thus, through regular monitoring, plant operators can make periodic
adjustments to the water treatment system and maintain system control.
The burden and flux materials charged to the blast furnace determine
the types and concentrations of pollutants found in the dirty blast furnace
gas and ultimately in the scrubber wastewater. These concentrations affect
the recycle ratio.
The heat load placed upon the process water affects the recycle ratio.
Evaporative cooling is often used to dissipate excess thermal energy but
water is lost and pollutants concentrated. If raw water makeup is not of
sufficient quality to dilute pollutant concentrations, scaling can pose a
problem because dissolved solids concentrations would increase. A system
blowdown must be discharged, cascaded, or evaporated from hot slag or
coke in order to maintain pollutant concentrations.
Hot Forming Subcategory
In the hot forming subcategory, process wastewater results in many
cases from the direct application of water to the hot formed product.
Representative uses include: product and roll cooling, descaling, flume
flushing, and shear spray cooling water. In primary mills, if hot scarfing
is employed, a waste stream results from the hot scarfer spray flushing
and cooling system and the wet gas cleaning system. Factors which limit
the ability of plant personnel to recycle a higher percentage of the
wastewater from hot forming operations are discussed below.
Much of the process water used in hot forming operations is applied
directly to the hot formed product. If product quality requirements are
to be met the recycle must be of consistently good quality. This can
dictate the need for a system blowdown to keep pollutant concentrations
at acceptable levels. In some hot forming finishing operations, city
water must be applied to attain the required quality product.
The quality of the raw water used as make up to the system to replace
water lost to evaporation, to sludge disposal, and to system blowdown,
limits the amount of process wastewater the plant is able to recycle. If
the raw water contains a significant concentration of calcium carbonate,
the gradual accumulation of calcium carbonate due to evaporative losses
unstabilizes the LSI; scaling can result. Thus, plants with better raw
water quality may be able to attain a higher recycle ratio than those that
do not.
Some plants indicated that suspended solids limited their ability to
recycle. Spray headers used for roll cooling can tolerate particles up
to about one inch in diameter before clogging. In primary mills that
employ hot machine scarfers, scrubbers are required to clean the dirty
combusion gases. Plugged spray nozzles and venturi throats can cause a
problem if suspended solids concentrations become too great'.
355
-------
Normality of production also affects the recycle ratio. Highest re-
cycle ratios are achieved when production is high and all mills are
operating continuously. If the demand for steel declines, this may result
in down-time for a mill; if it is for any considerable length of time, the
system must be flushed to prevent clogging and corrosion of lines. The
system must be blown down again before production can resume .
TECHNOLOGY USEFUL FOR REDUCTION OF WATER USAGE/DISCHARGE
BLAST FURNACE SUBCATEGORY
Wastewater generated in the blast furnace subcategory results
primarily from furnace top gas cooling and cleaning. Treatment of this
wastewater can consist of the following.
The initial step in the treatment of blast furnace wastewaters is the
removal of suspended solids. Most plants use a thickener or similar
gravity sedimentation component for suspended solids removal. Lamella
thickeners work very effectively on dirty blast furnace wastewater be-
cause high incoming solids concentrations are conducive to development
of a thick Lamella sludge blanket, which results in better solids removal.
The slurry from the thickener underflow is usually dewatered by
vacuum filtration. The filtrate is recycled to the thickener and the sludge
containing toxic pollutants and other pollutants is landfilled. In other
cases, a blowdown stream is taken off the bottom of the Lamellas and
evaporated off of hot slag. The water granulates the hot slag and also
traps air inside it. This granulated product can be used as insulation,
road bed, and even lightweight concrete if mixed with cement. Another
alternative is to send the hot underflow stream in slurry form to a
sinter plant. Here, the slurry would be thickened and the hot solids
used as sinter strand for their iron and thermal content. The water from
the slurry would be used as makeup water in the sinter plant.
To improve solids removal performance, thickeners and coagulant aids
such as polymers and ferric chloride are added to the wastewater stream
at the thickener inlet. These coagulant aids enhance solids removal by
a l d T i-n-r -i >-i -I-V»^ •f i"»vm a -f- i /-M-» /~\f la V/T^V mir\vO T-oa/^ i "I
J ". - \> Ii" " -"__ — —
A certain amount of makeup water will have to be added to the
thickener/clarif ier overflow before this stream is recycled to the gas
cooling and cleaning system. The LSI or RSI of the recycle stream should
be closely monitored and acid and descaling agents added as needed to
prevent scaling.
If burden materials charged to the blast furnace result in the
formation of unacceptable levels of cyanide, ammonia or phenols, alkaline
chlorination may be considered. Alkaline chlorination involves addition
of chlorine or sodium hypochlorite to wastewaters that have an alkaline
pH. Alkaline chlorination destroys cyanides by oxidizing them to carbon
dioxide and nitrogen; ammonia by oxidizing it to nitrogen and water; and
phenols by oxidizing them to carbon dioxide. Care must be taken to
356
-------
closely monitor the input of chlorine because an excessive amount could
cause chlorine cracking of metal surfaces and could result in the possible
formation of chlorinated hydrocarbons.
Activated carbon may be a recommended polishing step if quantities
of water are to be blown down to a municipal treatment facility. Activated
carbon reduces COD, BOD, and some organic pollutant concentrations.
Open trenches may be used whenever possible to cut down on corrosion
problems inherent in metal piping.
HOT FORMING SUBCATEGORY
Wastewater generated in the hot forming subcategory results primarily
from roll cooling, descaling, flume flushing, shear cooling, and hot
machine scarfing. As noted previously, the wastewaters generated from all
hot forming operations have similar characteristics and can therefore be
treated together. The following basic treatment units are generally
employed: primary sedimentation, surface oil removal, secondary settling
or filtration, and recycle. The following technologies are useful for
reduction of water in these units.
If the scale pit is located immediately adjacent to the mill stands,
the volume of flume flush water necessary to wash the scale into the pits
may be reduced. Open trenches should be used as often as possible to cut
down on corrosion problems. Scale pits must be easily accessible because
scale must be frequently removed using heavy equipment.
Slotted tube, rope or belt-type oil skimmers are used for grease and
oil removal. Grease and oil concentrations must be kept to a minimum
because they can foul heat exchangers and cooling towers as well as
inhibit the roll's ability to grip the hot metal.
A portion of the wastewater exiting the primary scale pit is
recycled to the mill for flume flushing.
The Wastewater leaving the primary scale pit is discharged to a
clarifier or thickener for additional suspended solids removal. The
removal of solids aids in the cooling of the wastewater. The sludge is
vacuum filtered with the filtrate returned to the thickener/clarifier and
the hot solids landfilled.
Many plants with the available land requirements use a settling
lagoon to achieve secondary settling. This greatly enhances the cooling
process but solids are not removed immediately and serve as a heat source.
The disadvantages of using settling lagoons are seepage and overflow
problems, which can be controlled through proper design.
Many plants use cooling towers to reduce the thermal load placed
upon the roll cooling water. The wastewater must occasionally be treated.
with scale inhibitors and acid added to prevent scaling. Some plants
357
-------
must soften their raw water supply to improve recyclability.
Again, it should be stressed that separation of contact from non-
contact wastewater is advantageous in achieving a low to zero discharge
recycle system. By separately treating the contact wastewater the burden
placed upon the treatment system is reduced. The non-contact wastewater
must undergo only minimal treatment before it can be recycled.
Water application rates should be reassessed and reduced if this will
not interfere with plant operation. Unnecessary water usage should be
eliminated.
FEASIBILITY OF FURTHER REDUCTION OF WATER APPLICATION
AND DISCHARGE RATES
The technology is available for the iron and steel industry to
significantly reduce water application and discharge rates associated with
blast furnace and hot forming operations. Many plants have not installed
BAT technology. Lack of segregated process, non-contact, and storm water
systems will severely hamper efforts to make significant improvements.
In some cases, once-through systems have survived rather than being re-
placed by cascading water use or closed loop recirculation systems due
to economic considerations. Water is often needlessly circulated through
unit operations regardless of their operating status. Heated, enclosed
buildings and improved management practices would eliminate this
situation. The volumes of water used for direct contact cooling are often
not adjusted on the basis of the water temperature. Water loss through
seepage or overflow at settling/cooling ponds is a problem at some
plants.
Water discharges resulting from blast furnace gas cleaning can be
reduced by the methods described previously. The pollutant loads in
discharges from this operation are dependent upon the burden charged to
the blast furnace. Today, many plants are faced with the decision of
whether to install improved water treatment technology in their existing
blast furnace operation or to replace their iron making facilities with
electric arc furnaces.
Evaluation of the potential for reduction of water application and
discharge must be made on an individual plant-by-plant basis. Specific
designs for systems cannot usually be transferred directly from one plant
to another. The degree of recycle which can be economically achieved is
dependent on the availability and chemical characteristics of fresh water
supplies, the existing layout of the plant, and the economics of the situa-
tion. If reductions in water use and discharge to levels approaching the
BAT recommended flow cannot be achieved, it is usually because of economic
rather then engineering constraints. Further flow reductions to approach
zero discharge will require extensive commitments from the corporation to
site-specific detailed design, proper start-up, expert management, constant
maintenance, laboratory support, and the education and cooperation of all
plant personnel. Through these types of combined efforts, it has been
358
-------
demonstrated that essentially zero discharge hot forming operations can be
operated over extended periods of time.
CONCLUSIONS
A review of zero and low discharge plants in the blast furnace and hot
forming subcategories as selected from the Development Document for Proposed
Effluent Limitations Guidelines for the Iron and Steel Industry revealed that,
in many cases, the water usage/discharge reported was too low.
In the blast furnace subcategory, three of six plants reviewed had a zero
discharge; all three plants accomplished this by means of an evaporative blow-
down onto slag. The remaining three plants have recycle systems with dis-
charges of less than 30 gpt of.iron produced.
One of nine mills reviewed in the hot forming-primary category had a zero
discharge with three other mills discharging less than the BAT recommended
flow. Seven of twenty-seven mills reviewed in the hot forming-section cate-
gory had zero wastewater discharge with an additional eleven mills discharging
less than the BAT recommended flow. One of the two hot strip mills reviewed
was a zero discharge facility. Of the six pipe and tube mills studied, only
four were verified hot-worked facilities and none of these had a zero dis-
charge. Two of these mills discharged at rates less than the BAT recommended
flow.
Seven of the nine mills that have achieved zero discharge recycle systems
in the hot forming subcategory are mini-mills. This achievement can be
attributed to: their relative simplicity; a design that incorporates segre-
gation of process, non-contact, and stormwater systems; the compatibility of
their waste streams for treatment purposes; relatively simple treatment
requirements; and the ability to maintain a hydraulic balance.
Some key elements that must be present before a zero or low discharge
system is attainable are: segregation of contact, non-contact and storm
water systems; proper design and operation of a hydraulically balanced system;
sufficient manpower to locate and eliminate sources of water infiltration or
loss; and control of water chemistry to eliminate scaling, corrosion, fouling
and foaming.
In the blast furnace subcategory, reduction of the total suspended solids
concentration in the gas cleaning wastewater is essential if a low to zero
discharge recycle system is to be attained. Other parameters which must be
controlled include alkalinity, hardness, total dissolved solids, chloride and
sulfate concentrations. Scaling is a major problem in most blast furnace gas
cleaning systems. Scaling has been known to plug spray nozzles, reduce the
effective area in venturi throats, and clog supply pipes. The precipitating
material that cuases this scaling is usually calcium carbonate. Proper
control of the Langelier Saturation Index (LSI) or Ryznar Stability Index
by addition of acid or anti-scaling agents and an adequate system blowdown
can reduce the chances of scaling. BAT Model No. 3 - 100 percent recycle -
appears impractical because dissolved solids concentrations and calcium
carbonate concentrations would increase. Some sort of blowdown is necessary,
359
-------
whether it be an evaporative blowdown off of slag or a cascade to a sinter
plant, to stabilize the LSI.
The types and quantities of pollutants found in the gas cleaning waste-
water are dependent upon the burden and flux materials charged to the blast
furnace, furnace operating conditions, and raw water characteristics. The
technology is available to implement a very tight recycle system, but indi-
vidual plant conditions will determine whether a low discharge recycle loop
is economically achievable.
Wastewater characterization of hot forming-primary, section, pipe and
tube, and flat mills indicates that their wastewater is similar and can be
treated together. Key parameters to be controlled in hot forming wastewaters
before a zero to low discharge recycle system can be implemented are total
suspended solids, total dissolved solids, grease and oil concentrations, and
temperature. Operating experience at individual plants indicates that water
quality requirements with respect to these parameters are not particularly
stringent. The technology is available to control these parameters, but
achievement of a zero to low discharge recycle loop will vary from plant
to plant depending on current plant layout, operating conditions, raw water
characteristics, and economic considerations.
RECOMMENDATIONS
It is recommended that the revised water discharge rates obtained during
this study be considered before any changes are made in regulations.
The results of this study suggest the need for an update of the status
of other industry subcategories. Regularly scheduled updates in the form of
annual DCP supplements would be useful to assist the USEPA in maintaining
an accuate data base.
Direct financial incentives to the plants may provide the motivation
to make the in-plant changes needed in water use and recirculation. These
incentives might take the form of tax credits, fines or other penalities, or
demonstration grants. Without these incentives, some plant operators appear
complacement about maintaining a "status quo."
The economic impacts of segregation of process, non-contact, and storm
water systems, and treatment of fresh water supplies to remove hardness, are
highly plant-specific. In some cases, these economic impacts would prohibit
the installation or improvement of water recirculation systems. Economics
should therefore be considered if revisions to standards are contemplated.
In some cases, the size and shape of raw materials and intermediate or
finished products affect water application rates as expressed in gallons per
tons. These factors should be considered in development of standards.
Development of the technology to adjust flows based on mill throughput
should be investigated. Plant operators should stop the application of water
during non-productive periods wherever possible.
360
-------
Plants should experiment with water quality parameters in order to
identify plant-specific minimum requirements. Many plants impose unneces-
sarily stringent constraints for total suspended solids, grease and oil,
dissolved solids, temperature, and flow. If greater toleration for these
parameters is proven, on a site-specific basis, to have no adverse impact
on operations and product quality, significant expenses associated with
operation of recirculation systems may be eliminated.
Based on the relatively high amount of recirculation achieved by mini-
mills, the USEPA may wish to consider separate subcategorization and regula-
tion of this segment of the industry.
361
-------
CONTROL OF SCALE FORMATION IN STEELPLANT WATER RECYCLE SYSTEMS
by: Dr. G. R. St.Pierre, Professor, and A. H. Khan
Department of Metallurgical Engineering
Ohio State University
Columbus, Ohio 43210
ABSTRACT
Several options for scale formation control are discussed with parti-
cular reference to the recycling of scrubber water for blast furnace top
gas. Experiments on carbonate scale formation are described and several
important operating variables are discussed. Several types of sensors that
might be used to aid in the avoidance of scaling are reviewed.
INTRODUCTION
The concentration of dissolved solids in industrial waters tends to
increase with the extent of recycling. In the absence of removal treatments,
the steady-state concentration of a dissolved solid entering the water
system by process contact increases inversely with decreased blowdown.
As systems are tightened extensively, they may become supersaturated with
respect to one or more scaling reactions. In this communication, some
experiments on the formation and growth of carbonate scales are described,
some procedures for the prevention of scaling are outlined, and the feasi-
bility of sensors for the prediction of instability are discussed. The
complexities of scaling reactions and treatment procedures are such that
no universal solutions are available^ '. In fact, it may be said that each
industrial system has unique features that require special considerations.
However, there are some concepts that may be applied to all such systems.
Three general options are available for the prevention of undesirable
scaling. They may be summarized as follows:
Use of pH control, inhibitors, and blowdowns to maintain the steady-
state dissolved solid concentrations in tolerable ranges.
Removal of dissolved (and suspended solids) by pH control, additions,
precipitation, settling, clarification, and related treatments to
prevent critical supersaturation.
Inducement of scale formation in localized regions where it can be
tolerated and conveniently removed periodically.
Experiments on scale structure and growth rate and the development of
stability sensors are essential in the advance of all three options.
362
-------
EXPERIMENTS ON CARBONATE SCALES
During the past two years, experiments have been performed to determine
the effects of several important variables on the formation and growth of
carbonate scales in flowing systems. Figures 1 and 2 are schematic diagrams
of the water flow system and the test cell wherein scaling was induced by
controlled temperature gradients. (*•' At flow rates of about 10 1/m, the
growth of carbonate scales on several rings of stainless steel (T302) were
studied. A series of heaters and thermocouple-controller circuits were used
to maintain the desired temperature gradients. The composition of the
circulating water was adjusted by means of chemical additives for pH and ion
concentration control, additions of suspended solids, and bubbling of
COa/ Argon mixtures to equilibrate the water with particular values of the
partial pressure of carbon dioxide. Only several representative features of
the experimental work will be discussed.
The regions of predominance of calcite, aragonite, and vaterite are
shown in Fig. 3 from the work of Roques and Girou.' ' These data are of
particular importance in analysing the results of the thermally-induced
scales formed in the present experimental work. Figs. 4 and 5 are repre-
sentative photomicrographs taken in a JEOL-JXA-35 scanning electron micro-
scope equipped with EDAX-9100 analytical equipment. Structural analyses
were also performed by conventional X-ray diffraction. In Fig. 4, the
formation of aragonite crystals (a) on a layer of iron chloride and (b) on
the stainless steel substrate from water to which fine particles (minus 320
mesh) were added. In general, the presence of an iron chloride layer
inhibits carbonate scale formation and blunts the aragonite needles. The
presence of suspended iron oxides caused a shift from predominantly hexa-
gonal calcite to predominantly orthorhombic aragonite deposits. In Fig. 5,
representative mixtures of calcite and aragonite in the absence of an iron
chloride layer are shown. The principal conclusions from these structural
studies are summarized as follows:
At substrate temperatures less than 50 C, calcite tends to be the
predominant form; at temperatures greater than 60°, aragonite tends
to be predominant; between 40° and 60°C, transient vaterite can be
observed occasionally.
The presence of Mg ions promotes the formation of aragonite.
The presence of Fe ions tends to suppress carbonate crystal forma-
tion, in part through the formation of an adherent iron chloride
layer.
Addition of iron oxide powder has little effect on either the
structure or growth rate of carbonate crystals.
A number of other interesting effects have been observed in the forma-
tion of carbonate crystals. As expected, high flow rates and high substrate
temperatures favor aragonite formation. At higher Mg"1"1" concentrations,
dolomitic carbonates and hydrated carbonates were observed. The specific
effects of some of the operating variables on growth rates are described in
the following section.
For illustration, the rate at which thin scale deposits form at times
up to about 15 hours are shown in Figs. 6 and 7. These data are presented
363
-------
Thermocouple
Leads
Plexiglass Tubing
Water Circulating Pump
Figure 1. Schematic Diagram of Flow System
Quartz Tubing
(Electrical Element
Inside, not shown )^. Thermocouples^
Plexiglass Tubing
(5I.4I.D. x64.20.D.)
1100-
Figure 2. Schematic Diagram of Test Cell (Dimensions in mm)
364
-------
A: Argonite
CCalcite
V: Vaterite
rC02 Atm
CO
CT>
cn
io"
.62
io3
T f
IO°C
Calcite
Equilibrium
— Curve
Calcite
j i
Atm
T I I [ I
50°C
Argonite
Equilibrium
- Curve
i y i
0 40 80 120 160 200 0 40 80 120 160 200 0 40 80 120 160 200
mg/t (Ca**)
mg/t (Co )
mg/t (Cat+)
Fig. 3. Regions of Predominance of Structural Forms of Calcium Carbonate.
-------
Fig. 4. Scanning electron micrographs of carbonate crystals. Left: Aragonite crystals on
iron chloride substrate at 3000X; Right: Aragonite and calcite crystals on stainless
steel with particles of iron oxide at 1000X.
-------
00
Fig. 5. Scanning electron micrographs of aragonite needles and calcite crystals with
iron oxide powder on polished stainless steel substrate at 1000X.
-------
as graphs of weight gain per unit area versus the square of time because it
has been found that the data correlate well with a "parabolic growth model."
After a variable incubation time, the rate of growth is inversely propor-
tional to current scale thickness. The conditions for Figs. 6 and 7 were
approximately identical except that in the case of the data of Fig. 7, one
half of the Ca"1"1" ions were replaced with Mg ions. In Fig. 8, the parabolic
growth rate constant is shown as a function of C03= concentration expressed
as the difference between the carbonate concentration at the interface and
the carbonate concentration in the bulk solution. It was found that this
driving force was most influential in determining the scale growth rate.
AVOIDANCE OF SCALE FORMATION
The avoidance of carbonate scale formation in recycled blast furnace
top gas scrubber water serves as an excellent example of one type of scale
avoidance problem. The problem has been discussed in this symposium by
Nemeth and Wisniewski.(4) Qne approach to the scaling problem that has been
suggested by Brower, Luther, and Ryckman^' is to lower pH of the scrubber
water by the addition of waste acids from pickling solutions. If the acid
addition is made prior to the clarifier several benefits might occur. In
addition to the removal of excess basic (Ca, Mg) carbonates, the iron content
of the waste acids may promote flocculation and settling in the thickeners
and also lower the cyanide content by the formation of iron cyanide complexes.
Inherent in this type of control procedure is the need to avoid serious
corrosion problems as well. Close control of water chemistry established
by the use of reliable sensors is a necessary feature in the recycling of
contact process waters.
SENSOR DEVELOPMENT
There are many approaches that may be taken in the development of
sensors for process control. A few examples are given as follows:
Specific ion electrodes
Thermal gradient devices
Partial pressure sensors for COa, Oa, etc.
Surface property of coupon devices
Accumulated mass devices
Direct saturation index devices.
At present, specific ion electrodes including pH measurement and standard
coupons are the only methods to be used extensively other than direct
periodic sampling and analysis. In the experiments described briefly in this
report, it has become clear that devices based on thermal gradient measure-
ment and those based on COa partial pressure measurement might be used
successfully in relatively clean, e.g. laboratory, conditions. If these
devices can be applied to industrial water systems, it would be possible to
introduce automatic equipment for treatment additions and blowdowns.
368
-------
^ 3
M
E
1
Ca " ions only
pH=9.5~9.7
Double Distilled Water
c
0)
c
1'
o
O
O
CO
Slope = 0.439 gms /m2- hr1
Timel/2,(hrs)'
Fig. 6. Scale Formation on Stainless Steel Substrate
„_
C°2
= 0.36 atm; Ca""" = 22 ppm; Bulk Temp = A7C;
Surf. Temp = 70 C.
369
-------
o>
I
d"
o>
<2
"c
.9?
o
o
CO
Co "and Mg"ions
pH=9.5-9.9
Double Distilled Water
I0f
Slope = 1.61 gms/m-hr
0
2 3
1/2 /, \l/2
Timel/2,(hrs)'
Fig. 7. The Influence of Mg++ on Scale Formation.
370
-------
"E 0.7
3f 0.6
0.5
g 0.4
0.3
0.2
a,
8
to
»
o
o
o
o
O.I
\ I I I I I I I
Co*ions only
pH=9.5-9.7
Distilled Water
4 6 8 10 12 14 16
C03 (i )-C03 (b) (g moles/1 iter)
18
Fig. 8. The Influence of Carbonate Concentration on Parabolic Rate
Constant.
371
-------
REFERENCES
(1) J. C. Cowan and D. J. Weintritt: "Water-Formed Scale Deposits", Gulf
Publishing Co., Houston, 1976, p. 250.
(2) A. H. Khan: "The Effect of Operating and Metallurgical Variables on
Water-Formed Calcium Carbonate Scales", M.S. Thesis, Ohio State
University, 1980.
(3) H. Roques and A. Girou: Water Research, 1974, Vol. 8, pp 907-20.
(4) R. L. Nemeth and L. D. Wisniewski: Symposium on Iron and Steel
Pollution Abatement Technology, U.S.E.P.A., Chicago, 1981.
(5) G. R. Brower, P. A. Luther, and S. J. Ryckman: Proc. Indtl. Waste
Conference, 1977, Vol. 32, pp 549-57.
372
-------
REDUCTION OF WASTES DISCHARGED FROM STEEL MILLS IN
METROPOLITAN CHICAGO THROUGH LOCAL ORDINANCE ENFORCEMENT
by: Richard Lanyon
Cecil Lue-Hing
The Metropolitan Sanitary District of Greater Chicago
100 East Erie Street
Chicago, Illinois 60611
INTRODUCTION
The Metropolitan Sanitary District of Greater Chicago (District) was
organized in 1889 and under its current charter, has the responsibility of
providing sewage collection and treatment service for an area of approxi-
mately 860 square miles, including the City of Chicago and approximately 130
surrounding communities. The District serves a connected domestic population
of about 5.5 million, plus an industrial sector with a waste load equivalent
to about 4.5 million population. The District operates and maintains seven
wastewater treatment plants that treat approximately 1.5 billion gallons per
day of combined industrial and domestic wastewaters. All of the plants are
biological activated sludge systems, with some including biological nitrifi-
cation and dual media filtration.
These plants receive flow from over 500 miles of intercepting sewers and
5,000 miles of local sewers. The District is also responsible for enforcing
applicable State and Federal water quality regulations along 72 miles of
navigable inland waterways, over 200 miles of small rivers and streams, and
36 miles of Lake Michigan shoreline.
INDUSTRIAL ACTIVITIES WITHIN DISTRICT
A substantial amount of industrial activity is conducted within the Dis-
trict area. This activity covers a wide spectrum, from iron and steel manu-
facturing to household appliances. Out of approximately 12,000 industries
within the District area, 6,000 are classified as wet industries, that is,
discharging some process wastewater and/or cooling water.
Cook County, which is approximately the same area as served by the Dis-
trict, is the most productive county in the United States in terms of manu-
facturing activity. Total shipments of manufactured goods out of Cook County
in 1979 amounted to $56,086,000,000 from 5,654 plants with 20 or more employ-
ees. Some of the largest manufacturing industries in Cook County include;
organic chemical, miscellaneous plastic products, blast furnaces and steel
mills, paints and allied products, telephone and telegraph apparatus, metal
cans, motor vehicles and passenger car bodies, radio and television receiv-
ing sets, and confectionery products among others.
373
-------
INDUSTRIAL WASTE CONTROL PROGRAM
The District is one of the few municipal sewage treatment agencies with a
long and effective record of enforcement regarding discharge of industrial
wastes to the public sewer system and to the waterways. On July 13, 1962, the
board of Trustees of the District adopted an Industrial Waste Ordinance that
set forth certain limiting conditions for the discharge of liquid industrial
wastes into the sanitary sewer system. Contained as part of these conditions
were the limits for pH not to be lower than 4.5 or higher than 10.0 and fats,
oils, and greases not to exceed 100 mg/1.
The Illinois Sanitary Water Board, the predecessor of both the Illinois
Pollution Control Board and the Illinois Environmental Protection Agency, on
June 28, 1967, adopted regulations for water and effluent quality for dis-
charge to waters in the State of Illinois. On January 28, 1968, the United
States Department of the Interior approved the Illinois standards and they be-
came law in the State of Illinois on April 1, 1968.
Subsequent to the passage of these State regulations, the District deter-
mined that a new industrial waste ordinance was needed to further control the
discharge of industrial waste into the sewer system. The purpose of this
ordinance was to insure that the quality of effluent from the District treat-
ment plants would meet the new standards for discharge to the waters of the
State. An Advisory Committee, consisting of representatives from industry,
consulting engineering firms, the academic community, and the District, was
appointed to develop a new industrial waste ordinance. The charge given to
the Advisory Committee was to develop an ordinance that would insure that the
effluent from the District's treatment plants would meet the new regulations
set by the Illinois Sanitary Water Board, protect the treatment plants from
upsets, and allow industry ample latitude with regard to the type of pretreat-
ment systems necessary to control the quality of industrial wastes discharged
to the sewer system.
On September 18, 1969, the Board of Trustees of the District adopted the
Sewage and Waste Control Ordinance which set specific industrial discharge
limits for 13 contaminants and 9 limiting conditions on discharges to sewers
and limits on 25 contaminants and 2 limiting conditions on discharges to
waterways.
Enforcement of the Ordinance has been delegated to the Industrial Waste
Division (Division) of the District which is responsible for monitoring the
discharge from over 6,000 industries within the District's jurisdiction. Two
groups in the Division, a field unit consisting of 28 Pollution Control Offi-
cers, 2 Sanitary Engineers, and 34 Water Samplers; and an enforcement section
consisting of 4 Pollution Control Officers and 2 Sanitary Engineers, adminis-
ter the Ordinance. Since pollutants can be most effectively controlled at the
point source of generation, continued inspection, monitoring and sampling of
industrial effluent discharge is required before industrial wastes enter the
sewer system. The field unit performs the physical inspections and sampling
of industrial facilities while the enforcement section handles the administra-
tive processes required when a company fails to meet compliance with the spec-
ified discharge limits. A laboratory group consisting of 3 Sanitary Chemists
374
-------
and 16 Laboratory Technicians provide analytical support for the enforcement
responsibility.
Field surveillance and monitoring activities include complete inspections
of industrial facilities to determine: the nature of the industrial opera-
tions; the types of wastes generated; waste pretreatment processes employed;
type, volume and storage of sludge and/or process residue generated for dis-
posal, method and frequency of removal; identification of both hauler and dis-
posal facility for sludge and/or process residue disposal; water consumption
information; and plant layout of all floor drains, sewers and sampling loca-
tions. With this information, along with knowledge of the production and
clean-up hours of the facility, a sampling program is designed to obtain rep-
resentative samples of the industry's sewer discharge. An industrial facility
can be sampled by: (1) 24-hour sampling trailer utilizing flow measuring
equipment, if necessary; (2) automatic sampling by either battery or vacuum
operated samples on an 8- to 24-hour basis; (3) manual sampling of 1- to 8-
hour composite or grab sampling; or (4) single grab sampling.
INDUSTRIAL WASTE ENFORCEMENT
Using the results of field inspections and samplings, an on-going review
is conducted of the status of compliance with the Ordinance of the numerous
industrial facilities and other dischargers to the waterways and sewers in the
District's jurisdiction. When a company is in violation, representatives ap-
pear for a conciliation meeting. The company is required to submit a plan and
schedule for compliance with the Ordinance. 'The company is encouraged to use
good housekeeping techniques, procedural and chemical changes, etc., to reduce
its pollutant discharge to the sanitary sewerage system. If necessary, pre-
treatment equipment must be installed. Conciliation continues so long as the
company is making a good faith effort toward compliance.
When conciliation is no longer productive, Show Cause proceedings are
recommended. Show Cause is and adversary proceeding with parties represented
by counsel. The offending discharger is ordered to comply with the Ordinance
by a specific date. If the order is not complied with a lawsuit is filed,
seeking compliance through injunction and penalties through fines.
CONTROL OF STEEL MILL POLLUTION
Included in the District's jurisdiction are five major steel making fa-
cilities, all located within a 10-mile reach of the Calumet River. The facil-
ities have a combined total production capacity of 7,800,000 tons of steel per
year. One of the District's major water pollution enforcement activities is
in the area of steel mill wastes. Enforcement work with the steel mills began
in 1969, at which time these five facilities were discharging industrial
wastes into the Calumet River and Lake Michigan from 45 separate outfalls.
See Figure 1.
CONDITIONS IN 1969
The five steel mills had a total of 45 outfalls discharging to the Calu-
met River system in 1969. The number of outfalls from each mill is as follows:
375
-------
Steel Mills
BCD E Total
Number of outfalls 9 6 5 7 18 45
The combined total quantities of flow and contaminants from these five
mills in 1969 are expressed as a daily average and shown as follows:
Parameters Quantity
Liquid Volume 546,000,000 gallons
Suspended Solids 166,000 pounds
Iron 25,000 pounds
Ammonia 5,150 pounds
Oil and Grease 39,500 pounds
Phenols 268 pounds
Cyanide 376 pounds
Detailed breakdown of the above quantities for each of the five mills is
shown on Table 1.
The Calumet River has a net flow away from Lake Michigan. Fluctuations
in the water level of Lake Michigan cause short term reversals in flow, espe-
cially toward the lakeward end of the Calumet River. The O'Brien Lock, as
shown on Figure 1, is operated in such a fashion so as to cause a relatively
small, but steady flow away from Lake Michigan.
Compared to the volume of waste discharged by the steel mills, the river's
net flow is incapable of preventing industrial waste from these steel mills
from flowing into the lake. The comprehensive revision and amending of the
District's Sewage and Waste Control Ordinance in 1969 incorporated the State
of Illinois waterway effluent standards. These standards gave the District
the authority to initiate action to bring these steel mills into compliance
with Illinois effluent quality standards.
ENFORCEMENT ACTION AND PROGRESS
The District initiated enforcement in 1969 following sampling of all five
steel mills. Samples were taken each hour for 24 hours each day over a ten
day period. The hourly grab samples were composited into daily composite sam-
ples for each outfall. These samples were taken to the District laboratory
for analysis. The analysis indicated that the discharges for the steel making
facilities were in violation of the District Sewage and Waste Control Ordi-
nance. Enforcement procedures were initiated against the steel mills.
In the matter of each of these five steel mills, conciliation proceed-
ings eventually broke down for various reasons and Show Cause action was
initiated. The Show Cause hearings resulted in the Board of Commissioners
passing orders requiring all five steel mills to have their industrial waste
discharges in compliance with the Sewage and Waste Control Ordinance by spe-
cific dates. It became apparent to the District during the conciliation and
show cause hearings, that only a lawsuit would cause the corporate management
of the various steel mills to commit the necessary capital and operating funds
to this type of program. The steel mills did not comply with the Order of the
Board of Commissioners and as a result separate lawsuits were filed against
each steel mill in the Circuit Court of Cook County. In each of the five
376
-------
lawsuits, the court proceedings and conferences resulted in a consent decree
and a compliance schedule for each steel mill for the reduction of the pollu-
tion loading to the Calumet River.
The nature of the wastes being discharged by mills to the Calumet River
system dictated the following requirements as being desirable in achieving
compliance.
(1) Separate process waste from cooling water so that cooling water could
be returned to the river as noncontact.
(2) Reduce or eliminate the liquid volume of process waste being dis-
charged to the river.
(3) Provide process waste treatment systems to reduce the constituent con-
centrations to within the waterway effluent standards for waterway dis-
charge or within the sewer discharge limits for disposal to the sewer.
Separation of cooling water from process waste was a reasonable and prac-
ticable request. This would achieve a large reduction in the liquid volume of
process waste being discharged to the river. To further reduce the liquid
volume of process waste, and to achieve a measure of water conservation, a
significant amount of recycling should occur.
A major problem facing the technical personnel of the various steel mills
involved in the pollution control programs was the age and condition of much
of the mill facilities. At the time, most of the mill facilities were in ex-
cess of 30 years in age, and as a result there was insufficient information
on the location of underground drainage, pipelines, and other facilities. In
addition, the age and condition of the production facilities was such that a
major expenditure of funds for pollution control and the need to install new
equipment in old facilities warranted some modernization of the production
facilities.
In 1969, the steel mills were found to employ the basic treatment proc-
esses of skimming for oil removal and clarification for solids removal. How-
ever, due to the age of facilities and an apparent lack of maintenance, these
facilities were found to be ineffective and often not capable of efficient op-
eration. Breakdowns, outages, and overloading were common operational problems.
The principal means of improved treatment for wastes from the steel mills
would consist of skimming for oil and grease removal, clarification for solids
removal, and filtration to achieve a final high degree of removal of solids
and oil and greases to facilitate recycling. In a few cases, separate proc-
ess treatment for other constituents, principally cyanide, needed to be em-
ployed, in order, to meet the limits for discharge to the sanitary sewer sys-
tem or waterway. For some constituents, such as ammonia and suspended solids,
specific reduction was not required if discharged to the sewer. The reason
for this is that District treatment plants presently have, or are scheduled
to have, the capability to treat and remove suspended solids and ammonia.
.ABATEMENT PLANS FOR EACH STEEL MILL FACILITY
Steel Mill A. The major production facilities at Steel Mill A that generate
industrial waste and the waste treatment systems installed are:
377
-------
(1) Five blast furnaces from which the wastewater streams are directed to
clarifiers, cooling towers and then recycled. The blowdown from this
recycle system is used as quenching water for slag.
(2) Plate mills and structural mills from which the wastewater flows
through scale pits, oil separators and then to a central wastewater
treatment facility which consists of three high rate reactor clarifi-
ers, cooling towers, sludge thickener, and a vacuum filter. The
treated effluent water is then recycled back to these mills for reuse.
(3) Wastewaters from four electric furnaces and the basic oxygen process
shop pass through a first stage clairifier and then flow to the cen-
tral wastewater treatment facility for processing and recycle.
(4) Sintering plant wastewater which also flows to the central wastewater
treatment facility for processing and recycle.
Steel Mill A has stopped the discharge of process wastewater to any
waterway and is now in a 100 percent water recycle and reuse-mode. From 4.5
to 5.0 MGD of blowdown now goes to the District for treatment.
Steel Mill B. The major production facilities at Steel Mill B that generate
industrial waste and the waste treatment systems installed are:
(1) A coke battery that consists of 45 ovens. The wastewaterwater from
these facilities is recycled; the blowdown from the recycle system
passes through a settling basin, an oil skimmer, and then discharges
to the- sanitary sewer.
(2) The basic oxygen process shop, consisting of two 120-ton vessels.
Wastewater flows are recycled; the blowdown passes through settling
basins and then to the District.
(3) The blast furnace plant. Wastewater passes through clarifiers, an
alkaline chlorination cyanide destruct process, and deep bed sand
filters. This flow is then recycled as noncontact cooling water for
the blast furnace plant or the rolling mills and then discharged to
the Calumet River.
(4) Blooming and rolling mills. Process wastewater passes through scale
pits, oil separators, and deep bed sand filters and is recycled as
noncontact cooling water and eventually discharges to the Calumet
River.
Steel Mill B has reduced the number of outfalls to the Calumet River from
six in 1969, to one outfall discharging treated process water which has been
used for cooling. In 1980, this mill was closed due to business reasons.
Steel Mill C. The major production facilities at Steel Mill C that generate
an industrial waste and the waste treatment systems installed are:
(1) The coke plant, consisting of 100 ovens. Wastewater flows are
378
-------
partially recycled, with the blowdown going through a clarifier, oil
separator, cyanide stripper and then discharging to the sanitary
sewer. Noncontact cooling water is pumped from and discharged to the
Calumet River.
(2) Blast furnace and sintering plant. Wastewater flows through clarifi-
ers, cooling towers, and are then used for gas scrubbing. Blowdown
from this system goes to the sanitary sewer. This mill has reduced
the number of outfalls from five in 1969 to three for discharging
noncontact cooling water in 1980.
Steel Mill D. The major production facilities at Steel Mill D that generate
an industrial waste and the waste treatment systems installed are:
(1) One blast furnace. Process waste treatment consists of oil skimmers,
clarifier with chemical treatment, and a vacuum filter. This treated
flow is then diverted to cooling towers and recycled. Blowdown is
discharged to the sanitary sewer.
(2) Coke plant, consisting of 75 ovens. Process wastes pass through a
waste treatment system consisting of oil skimmers, clarifier with
chemical treatment, and a vacuum filter. The blowdown is used for
quenching push.coke in a closed loop system.
(3) By-product plant. Process wastes are diverted tu che sanitary sewer
system. Noncontact cooling water is pumped from and discharged to
the Calumet River.
(A) Basic oxygen furnace shop, rolling mills, tube mills, and wire mills.
Process wastes are treated through deep bed filters and recycled.
The filter backwash is treated in a waste treatment system. This
mill has reduced the number of outfalls to the Calumet River from
seven in 1969, to one in 1980.
Steel Mill E. The major production facilities at Steel Mill E that generate
an industrial waste and the waste treatment systems installed are:
(1) The basic oxygen process shop, consisting of two, 100-ton units. Pro-
cess wastewaters are diverted to clarifiers, cooling towers and re-
cycled in a closed loop system. Blowdown is used for slag quenching
with no discharges.
(2) Primary mills. Wastes are discharged to three settling pits which
prior to 1970 discharged to the Calumet River. This is now a closed
loop system with the discharge being returned to the mill for recycl-
ing use. The blowdown from the mill goes to a pump station and then
to a sanitary sewer. Noncontact cooling water is pumped from and
discharged to the Calumet River.
(3) The hot strip mill. Process waste goes to a sand filtration system,
with treated water recycled back to the mill for reuse.
379
-------
(4) All pickling wastes are treated through a waste treatment system
with the final effluent going to the sanitary sewer.
(5) Cold mill. Wastes are treated through an oil emulsion system.
Wastes: f rom the blowdown tanks are sent back to the pickling waste
treatment system. Oil is scavenged out.
(6) Galvanizing and electroplating wastes are treated through the pick-
ling waste treatment system.
Steel Mill E has reduced the number of outfalls to the Calumet River from
eighteen in 1969 to four noncontact cooling water outfalls in 1980.
IMPACT ON THE WATERWAY
For six parameters (suspended solids, iron, ammonia, oil and grease,
phenols, cyanide), reduction in the pollutants discharged to the waterways
has averaged 100 percent from the five steel mills discharging to the Calu-
met River and Lake Michigan. Discharges that have been diverted to the Dis-
trict sanitary sewer systems are frequently monitored and are at present in
compliance with the District Sewage and Waste Control Ordinance.
A summary of total reduction in discharge to the waterways from these
five steel mills from 1969 to 1980 is shown in the following table. Quanti-
ties are expressed as daily averages.
Percent
Parameter Units 1969 1980 Reduction
Liquid Volume MG 546 0 100
Suspended Solids Ibs. 166,000 0 100
Iron Ibs. 25,000 0 100
Ammonia Ibs. 5,150 0 100
Oil & Grease Ibs. 39,500 0 100
Phenols Ibs. 268 0 100
Cyanide Ibs. 376 0 100
Number of outfalls 45 9 80
Examination of water quality data for the eleven-year period from 1970
through 1980 has been made to determine the impact of these dramatic reduc-
tions on the quality of the receiving waterway.
The District performs routine monitoring of the entire waterway system
under its control. Two of the stations which are routinely monitored are
within the reach of waterways which are influenced by four of the five steel
mills. These stations are at the Ewing Avenue and 130th Street Bridges over
the Calumet River as shown on Figure 1. The District's routine monitoring of
the waterways consists of monthly grab samples taken at these bridges, analy-
sis of these samples and reporting of analytical results.
In the reach of the Calumet River between its mouth at Calumet Harbor and
the O'Brien Lock, the steel mills comprise the principal inflow. There are
few combined sewer gravity overflows and only one combined sewer storm pumping
380
-------
station in this reach. Their frequency of discharge to the river is low and
their impact on receiving water quality is minor compared to the output of
the steel mills.
As to the impact of Steel Mill E on the Little Calumet River, this is
more difficult to judge by receiving waterway data, as the reach of the river
to which this steel mill discharges also receives discharges from numerous
combined sewer outfalls, a combined sewer storm pumping station and a major
municipal sewage treatment plant.
As shown on Table 1, discharge to the Calumet River of process flow and
a number of chemical constituents has ceased. As indicated earlier in the
discussion, the desirable objectives of the enforcement action included sepa-
ration of process water and cooling water flows, a reduction of process flows
and treatment of process waste flows. These have been accomplished. As a
measure of overall achievement of the enforcement program and pollution abate-
ment plans implemented by the steel mills, Figures 2 and ^ show significant
improvement in both dissolved oxygen and temperature. Over the eleven-year
period, the annual average dissolved oxygen increased from 6.9 mg/1 in 1970 to
11.2 mg/1 in 1980 at Ewing Avenue and from 6.4 mg/1 to 10.0 mg/1 at 130th
Street. Maximum and minimum values for each year have also shown improvement.
The waterway standard for dissolved oxygen as established by the Illinois Pol-
lution Control Board (IPCB) is not less than 2.0 mg/1 prior to December 31,
1977, and not less than 4.0 mg/1 after January 1, 1978. Nearly all values
were in compliance with this standard. The minimum values in each of 1970 and
1978 fell below the standard during some months of both summers.
Similarly, the annual average temperature decreased from 18.2°C in 1970
to 10.7°C in 1980 at Ewing Avenue and from 17.5°C to 10.8°C at 130th Street.
Annual averages within the period varied somewhat beyond these values, but
there is, nevertheless, an overall trend toward lower average annual tempera-
tures. Maximum annual temperatures varied considerably at both stations,
ranging between 30.0°C and 17.0°C. No trend in annual maximum temperatures
is evident. However, minimum annual temperatures show a definite trend toward
lower values. The IPCB standard for temperature is not to exceed 34°C. No
values at either Ewing Avenue or 130th Street exceeded this limit.
As shown by Figure 4, total suspended solids do not evidence a trend
toward lower values, despite the reductions in suspended solids discharged to
the waterway by the steel mills. This suggests that total suspended solids
in the waterway are at a background level and were not significantly affected
by reductions in discharge from the steel mills. Annual averages of total
suspended solids varied between 20.8 mg/1 and 10.8 mg/1 at Ewing Avenue and
between 27.3 mg/1 and 15..2 mg/1 at 130th Street. Annual maximum and minimum
values at each station vary considerably without any apparent trend. The
IPCB standard for total suspended solids is not to exceed 25 mg/1. All annual
average values were below this standard, except for 1977 at 130th Street. At
Ewing Avenue, many values in 1970, 1972, 1973, 1977, and 1980 were above this
limit, whereas at 130th Street, the standard was exceeded in all years except
1980.
For ammonia nitrogen, the IPCB has established separate standards for sum-
mer and winter. For the months of April through October, the standard is not
381
-------
to exceed 2.5 mg/1, while for the months of November through March, the stand-
ard is not to exceed 4.0 mg/1. As shown by Figure 5, there were values which
exceeded the summer standard in 1972 and 1973 at Ewing Avenue and in 1973 and
1978 at 130th Street. The annual average and minimum values at 130th Street
and the annual minimum values at Ewing Avenue show a decreasing trend in the
ammonia nitrogen concentration. The higher concentrations of ammonia nitrogen
at both stations in the later years of the 1970 to 1980 period are not explain-
able by any records of discharge by the steel mills.
Elevated ammonia nitrogen concentrations in the latter part of the period
are also evident in the winter data as shown on Figure 6. However, all values
at both stations are well below the standard of 4.0 mg/1. Data for the winter
period are less variable and suggest a trend toward lower ammonia nitrogen
concentrations.
Fats, oils, and greases, over the eleven-year period do not evidence any
trend. While the annual averages for both Ewing Avenue and 130th Street for
the 1970 through 1979 period show a trend toward increasing concentrations,
concentrations in 1980 are much lower than the several preceding years. As
shown on Figure 7, the annual averages at both stations are quite close in
magnitude and vary between a high of 25.3 mg/1 and a low of 5.3 mg/1. The an-
nual averages for the year 1976 through 1979 exceed the IPCB standard of 15
mg/1 at both stations. The annual average at 130th Street in 1975 is also
above the standard.
Some of the values for fats, oils, and greases at Ewing Avenue exceeded
the limit in each of the eleven years. At 130th Street, all years except 1972
and 1980 had some values which exceed the standard. Effective November 23,
1977, the IPCB adopted a change in which further defined this standard. The
change provides that this constituent "...shall be analytically separated into
polar and non-polar components if the total concentration exceeds 15 mg/1. In
no case shall either of the components exceed 15 mg/1 (i.e., 15 mg/1 polar
materials and 15 mg/1 non-polar materials)." The District analyzes all efflu-
ent samples that exceed 15 mg/1 of total fats, oils, and greases for each of
the components. However, this is not done routinely for waterway monitoring
samples.
Total cyanide concentrations in the Calumet River at Ewing Avenue show a
definite trend of improvement, whereas no trend is evident at 130th Street, as
shown by Figure 8. The annual average total cyanide varied from 0.028 mg/1 in
1970 and 0.04 mg/1 in 1971 down to 0.004 mg/1 in 1980 at Ewing Avenue. At
130th Street, these annual averages varied between a high of 0.023 mg/1 and a
low 0.005 mg/1 over the period of eleven years. During the last four years,
the annual average steadily decreased at 130th Street. The IPCB standard for
total cyanide was 0.025 mg/1 up until September 6, 1978. During this period
of time, there were numerous values which exceeded the standard. Effective
September 7, 1978, the standard for total cyanide was 0.1 mg/1. No values
exceeded the standard subsequent to this change.
The record of improvement in the water quality of the Calumet River dem-
onstrates that pollution control programs and expenditure of private and public
monies do produce results. As shown by the data presented herein, the local
enforcement program of the Metropolitan Sanitary District has caused five
382
-------
major steel making facilities to reduce their discharge of pollutants to the
waterway and has caused improved water quality of the Calumet River.
TABLE 1. QUANTITIES OF FLOW AND CONTAMINANTS FROM FIVE STEEL MILLS IN 1969
Steel Mills
Parameter
Liquid Volume
Suspended Solids
Iron
Ammonia
Oil and Grease
Phenols
Cyanide
Units
MG
Ibs.
Ibs.
Ibs.
Ibs.
Ibs.
Ibs.
A
330
69,760
6,800
1,600
25,400
86
150
B
38
13,800
1,590
650
684
48
11
C
43
14,300
1,020
1,070
3,180
36
42
D
86
48,200
14,600
1,080
3,600
57
67
E
49
20,200
1,010
745
6,500
13
3
Notes: Steel Mill B was closed in 1980 due to business reasons.
Quantities are expressed as daily averages.
Discharge of process flow and contaminants had ceased by 1980 for
all steel mills. Therefore, all steel mills had achieved a 100
percent reduction in discharge of process flow and contaminants
by 1980,
Figure 1. Steel Mills Located Along Calumet River and Little Calumet River
383
-------
14.0 «i
13.0-
12.0-
11.0-
^ 10.0-
| 9.0-
S 8.0-
>
X
0 7.0-
Q
£ 6.0-
0 5.0-
* ,
1
1
1
1
l ;
MAX.
I
1
|
I AVG.
< —
1
4---"
1
|
| MIN.
|MAX.
&$•-
i
1
L
J
1
I
1
1
1
1
1
i^^
|
1
1
1 1
1 MIN.
1
1
1
1
_L
STANDARD PRIOR TO DECEMBER 31,1977
t .
.
• ^^
^^^
/\ ''
1
I
1
AFTER
JANUARY 1,1978
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 I960
YEAR
Figure 2. Calumet River Dissolved Oxygen Concentrations - 1970 thru 1980
35-1
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980
Figure 3. Calumet River Temperature - 1970 thru 1980
384
-------
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980
YEAR
Figure 4. Calumet River Total Suspended Solids Concentrations - 1970 thru 1980
LEGEND:
Ewing Avenue
130th Street
STANDARD i
1970 1971 1972 1973 1974 1975 1976 1977. 1978 1979 1980
Figure 5. Calumet River Ammonia Nitrogen Concentrations - April thru October
1970 thru 1980
385
-------
LEGEND:
Ewing Avenue -1
130th Street
'70-'71 '7l-'72 '72-'73 '73-'74 '74-'75 '75-76 '76-77 '77-78 '78-'79 '79-'80
YEAR
Figure 6. Calumet River Ammonia Nitrogen Concentrations - November thru March
19-70 thru 1980
70.0-,
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980
0.0
Figure 7. Calumet River Fats, Oils and Greases Concentrations - 1970 thru 1980
386
-------
0.16-
0.14-
0.12-
E 0.10-
UJ
Q
< 0.08-
u
< 0.06-
K
0.04-
0.02-
LEGEND:
Ewing Avenue
130th Street
(0.324)
'I
l~~~-
•)MAX.
MAX.
^
1 MIN.
1
1MIN.
.-STANDARD
lj_ |.
STANDARD
AFTER SEPTEMBER 7, 1978
R£^
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980
YEAR
Figure 8. Calumet River Total Cyanide Concentrations - 1970 thru 1980
387
-------
A MASS BALANCE MODEL FOR RINSEWATER IN A CONTINUOUS
STRIP HALOGEN ELECTROLYTIC TINNING OPERATION FOR USE IN
EVALUATING WASTEWATER TREATMENT AND RECOVERY ALTERNATIVES
by: James Skubak
Energy & Environmental Control Department
Jones & Laughlin Steel Corporation
Box 490, Aliquippa, Pa. 15001
Ronald D. Neufeld, Ph.D
Associate Professor of Civil Engineering
University of Pittsburgh
949 Benedum Hall
ABSTRACT
The operation and rinsewater application of J&L's electrolytic tinning
operation at Aliquippa are studied in order to specify a plating chemical
recovery technology for wastewaters. The existing treatment method for
wastewaters, hydroxide precipitation, is adequate to meet discharge limits,
but economics favor a change that uses a form of recovery
A mathematical mass balance model -is developed to predict rinsewater
concentration changes as flowrates are varied. The relationships described
by the model are tested and it is shown that the model for tin is followed
within a narrow margin of error. The significance of this is that waste-
water flow (and mass rate of tin in the wastewater) can be measurably
reduced with no apparent product defects.
The model is refined to reflect the installation of rinsewater controls
that optimize rinsewater use, and various recovery techniques are applied to
this refined model. A comparison of these recovery methods indicates that
several alternatives exist that can recover plating chemicals successfully
and have capital recovery periods of less that one year. In summary, the
importance of modeling in this type of application is paramount to
successful recovery.
INTRODUCTION
The electrodeposition of tin onto steel is a simple concept to under-
stand, but a difficult manufacturing process to effect in a competitive
market. Numerous variables impact on the successful commercial production
of tin plated steel or tin plate. Some basic variables are chemical
compatibility of all raw material used, knowledge by the production force
of the process and even simpler concepts such as market supply and demand,
388
-------
The objective of this research is to develop and demonstrate a fundamental
sound approach to the proper management of a process wastewater.
The tin plate manufacturing process requires a careful application of
water for cleaning and rinsing, but this water must then be removed from
the process to insure product quality within customer specifications. In
most cases, the water is not reusable within the tinning process.
The Aliquippa tinning operation of the Jones & Laughlin Steel
Corporation, presently transfers tin line wastewater to a treatment
facility to remove chemicals for acceptable discharge of the water to
the Ohio River. This facility, referred to as the Chemical Rinse
Treatment Plant, was constructed in 1971. This facility treats a combined
flow of 25 different waste streams from various operations including
wastewaters from the tinning operation. The primary constituents in the
waste stream from the tinning operation are tin and fluoride. It was
found that tin and fluoride were entering the plant at rates of 770 and
1090 Ibs./day respectively, and being wasted to sludge. At an ever-
increasing price for tin, it is clear that an economic recovery evaluation
of the tin line waste streams is warranted.
ALIQUIPPA'S TIN PLATING PROCESS
The Jones & Laughlin Steel Corporation's Aliquippa Works operates
two tinning lines for the manufacture of tin plate for use in the canning
industry. Tin lines, numbered 3 and 4, operate using the proprietary
halogen electrolytic tinning processes.(1)(2) The electroplating of tin
onto steel strip is shown on Figure 1. One side of the strip is tinned by
1
Moving /""^Ny
steel f \
strip V )
\^^
0
Induced
electric
current
e
Cathode: Negatively charged contact
rolls impart negative charge
to steel strip
~ ~ ~ ~ -4 Sn Sn Sn Sn Sn Sn Sn Sn Sn
SnF,~2 or. SnFc"'* J AF~ or 6F~
4 O
\s— Induced current
"O
corrodes tin
/ electrolyte carries it to the
Sn+2 /
t
Steel strip
coated on
_ one side
with tin
from anode and
cathodic strip
Figure 1. The Halogen Electrolytic Tinning Process
389
-------
rolling the strip into an electroplating bath which has solid tin bars or
anodes submerged at the bottom of the bath. The composition of the bath
and an induced electric current allow the tin to corrode from the anode and
deposit on the steel strip cathode in a thickness dictated by the amount of
electric current and the speed of the steel strip.
In addition to tin and fluoride, sodium chloride, chelating and
brightening agents, and hydrochloric acid are required in the bath for
proper electrodeposition, but the economic consequence of their use is
small compared to the use of tin and fluoride.
At various locations along the plating bath,usually near a contact
roll, the strip tends to rise to the surface, making it susceptible to
drying. To prevent drying a water hose is mounted over the strip and
applied as needed. This practice not only serves as a deterrent to strip
drying, but indirectly is a make-up for water leakage and evaporative
losses.
WASTEWATER CHARACTERIZATION
The electrolyte solution is rinsed from the strip in the recovery rinse
tank after plating. The recovery rinse tank is an immersion rinse to which
deionized water is continuously added. Figure 2 shows this rinse.
Spray
rinse
Water
addition
f\l
V)
Sod]
bifluc
.urn
>ride
/
Immersion^rinse Q J/
^
Recovery
stream
oho strip
T^S ivy _ „„ fr.
<
I
J {] Qj final
^ — rinse
returns to
rinse tank
Wastewater
overflow
Moving steel strip with
bath drag-out
Figure 2. The Recovery Rinse Tank
The recovery rinse is so named because the intent of this rinse is to
recover plating chemicals by rerouting a portion of this rinse back to the
plating bath. The point at which the recovery water is returned to the bath
is on one side'of the strip as it leaves the bath. This serves to rinse
some of the electrolyte from the strip, resulting in less "drag out" from
the tinning bath. Also, dry sodium bifluoride salt is added to the recovery
rinse tank. This material is added on an as-needed basis by operators to
prevent the recovery tank solution from becoming milky, a phenomenon
associated with a defective product. It is thought that the cause of this
phenomenon is a precipitation reaction resulting from a deficiency of
fluoride in the recovery tank. After the strip emerges from the recovery
390
-------
tank, a spray rinse is applied which drains back into the recovery tank.
EXPERIMENTAL APPROACH
Out of all of the mass inputs into the recovery tank, only a small
portion is recovered in the recovery stream. Most of the tin and fluoride
that exit this immersion rinse are discharged in wastewater streams. The
recovery tank overflow contains over 95% of the tin, fluoride and other
electrolyte chemicals contained in rinsewaters.
In order to obtain a statistically adequate data base, a continuous
flow recording device and a time actuated composite sampler were used to
characterize the overflow from the #3 _line recovery tank over a total of
160 hours of operation. The results showed that the recovery tank overflow
averaged 59 ± 16 gallons per minute. The tin content was calculated to be
320 ± 99 milligrams per liter. Fluoride was 1426 ± 783 milligrams per
liter. These concentrations represent average mass rates of 226 pounds per
day and 1010 pounds per day for tin and fluoride, respectively. It was
also observed that tin and fluoride concentrations in the recovery tank
overflow varied inversely with overflow rate. It is inferred from this
observation that the overflow rate (or water use rate) is an operator
controlled variable that affects tin, fluoride and water losses. Continuous
chart recordings of the overflow rate indicate that short and long term
flow variations are caused by three major factors. These factors are:
(A) Line speed; As strip passes through the recovery tank, the strip's
shear force on the tank water tends to increase the depth of the discharge
side of the tank causing a greater overflow. This does not occur, however,
on #4 Tin Line which incorporates a "surge line" to convey rinsewater
back to the entry end of the tank.
(B) Hot deionized water pressure; Depending on the level of activity of
other associated Tin Plate Department operations, line pressure at the
discharge points of spray rinses can vary by as much as 100 percent.
Since there are no flow controls (or automatic pressure controlling
devices) on any of the rinsing operations, flow rates will vary as a
function of line pressure.
(C) Different operators will adjust throttling valves on rinsewater lines
as perceived necessary for adequate rinsing.
All three of these factors indicate that the use of water for rinsing
is dictated by physical consequences and not necessarily by the solubility
and chemical characteristics that normally should be used to prescribe a
rinsing program. It is presumed that with a conscientiously applied program
of rinsewater use, the average amount of rinsewater required would be some-
what lower than that level found in the characterization survey.
Other significant losses of tin and fluoride are housekeeping losses,
such as leakage and spillage collected in the basement sump under each tin
line. These losses were found to contain an average of 96 Ibs./day of
tin and 126 Ibs./day of fluoride for each tin line.
The summary of the sampling survey indicates that from 115,500 to
391
-------
201,750 Ibs. of tin are discharged each year from both tin lines in rinse-
waters, leakage and spillage. The volume flow which carries this tin
ranges from 150 to 190 gallons per minute during operations. This flow
also carries from 328,750 to 510,000 Ibs. per year of fluoride.
A RINSING MODEL
In order to effectively understand the rinsewater system, as well as
predicted cause/effect relationships within the system, a mathematical
model based on material balance considerations was developed.
MODEL ASSUMPTIONS
The basis for the model development is shown in Figure 3.
See Figure 5 y f—Q? /—See Figure 6 r~Q9
Q8C8
Q0c0-
(Total
drag-
out)
Q6C6(Application of Q7
returns to bath
as Q6)
QHCII
Recovery Q2C2
See Figure 4
QlOC10
Electrolyte from
circulating tank
FOR ALL CASES: C2 = C3 = C5
Qs = Qe = Qu
To Circulating Tank
Figure 3. Tin Line Rinsing Schematic For Mass Balance Claculations
Before the mathematics of the rinsing process can be developed, certain
assumptions must be made. These include:
(A) Finite Volume: The strip carries with it (at a certain line speed) a
finite coating of fluid whenever it leaves a bath or spray.
(B) Continuity: The electrolytic chemical concentration of the finite
coating is the same as that of the bath in which it was immersed. In the
case of a spray rinse, the rinsewater completely mixes with the finite
coating to form a coating of different concentration.
(C) Wringer Efficiency: All of the wringer rolls between rinses will be
assumed to have the same efficiency. This implies as in Figure 3 that
Qs = Qe = Qll- (In reality, these are all different at different times
depending on operator adjustment and frequency of maintenance replacement
of the wringer rolls).
(D) No Evaporation: It will be assumed that between the bath and Qu there
is no significant evaporation.
(E) Uniformity: The tin and fluoride concentrations throughout the recovery
rinse tank are uniform, i.e., C\ = €2 = GS = 03. (The strip averages almost
392
-------
1100 feet per minute through the tank, and observed eddy currents caused by
the strip's shear force serve to strengthen this, assumption.)
(F) Fixed Recovery: The rate of recovery water which is sprayed on the
strip as it leaves the bath will be fixed at 3 gallons per minute (average
of field measurements).
(G) Assumed and Measured Values: Using values found in the characterization
survey and operating parameters the concentrations and flow that will be
used in the development of this model are in Table 1.
TABLE 1. VALUES USED IN MODEL DEVELOPMENT
Recovery tank tin concentration
Recovery tank fluoride concentration
Recovery tank overflow
Recovery tin concentration
Bath fluoride concentration
320 mg/1
1426 mg/1
59 gal./min.
17 gm/1
35 gm/1
(H) Constant Constituent Ratio: For any concentration of tin in the
recovery rinse, the ratio of fluoride to tin must remain greater than or
equal to (1426 mg/1 F)/(320 mg/1 Sn) = 4.46. (This is an assumption based
upon the necessity of sodium bifluoride additions to the recovery tank.)
(I) Steady State: The tin line has no line stoppages, and a steady state
condition of operation is experienced. This steady state condition also
implies that the line speed is constant and directly proportional to the
dragout.
MODEL THEORY
To formulate a model to describe this rinsing operation, it is
necessary to state all of the known relationships for each individual
application of rinsewater. In each of these relationships, there are known
and unknown quantities. Through algebraic manipulation among the relation-
ships the unknown quantities can be narrowed down to several solvable
quadratic equations. To begin the modeling, the tin dragout from the bath
will be quantified. A simplified portion of Figure 3 is shown in Figure' 4
as the location where the strip leaves the plating bath and is sprayed on
one side with recovery water.
f Total resultant drag-out from bath to
recovery tank,Q0C0=QDCD+QcCc
(Resultant inside drag-out, QDCD=Q2C2-K}ACA-QDCD (QD=QA+V)
Q2C9, from recovery tank
Resultant to bath, QBCB=Q2C2+QACA-QDCD
-Wringer rolls
Outside. 4
drag-out 1
(Resultan
^"^••^^
t QACA
h
Deflector
roll
r
Figure 4. Recovery Rinsewater Application.
393
-------
Since the deflector roll is grooved and serves merely to change the direction
of the strip, it will be assumed that QQ = QA, i.e., both sides of the strip
have equal volumes of water per lineal foot. (Although the deflector roll
does "wring" the strip to a certain extent, it will be assumed that during
the movement between P and P', some liquid is lost tangentially from the
outside of the strip, balancing the amounts on both sides at point P!.)
Also, it will be assumed that out of the 3 gpm, Q2, that is sprayed on the
strip, only a certain volume, V, stays on the strip. The fraction of Q2
that flows back to the plating bath, QB, equals (3-V). In addition the
total volume of liquid on the inside of the strip (after the application of
Q2), QD, is the sum of V and QA. In summary the volumes of QD and QB can
be represented as
QD = QA + V and Qfi = 3-V (la),(lb)
In order to quantify the total dragout from the bath, QQ CQ, refer to
Figure 3 and note:
Qo C0 = Q2 C2 + Q3 C3 (2)
which is actually a rough mass balance around the recovery tank. This
equation excludes the quantity Qe CG on the left side of the equation and
Qs GS on the right side. These terms can be considered equal if the dragout
from the recovery tank is assumed negligible. Substituting known values
for Q2, C2, Q3 and C3:
QQ C0 = 19840* (3)
QO CQ can also be represented exactly by
QO C0 = Qc Cc + QD CQ (4)
which is the summation of the dragout on both sides of the strip after the
application of Q2. Substituting values from equation (3) and from Table 1,
equation (4) becomes
19840 = 17,000 Qc + QD CD (5)
Referring again to Figure 4, the complete mass balance of the spraying of
recovery water can be represented as
Q2 C2 + QA CA = QB CB + QD CD (6)
Substitution of known values from Table 1 and equations (la) and (Ib), and
using CD = CB (Assumption B) yields
960 = 17,000 QA = (3-V)CB + (QA + V)CB ' (7)
* Units for Cx and Qy are mg/1 and gpm respectively, but will be excluded
for ease of mathematical manipulation.
394
-------
Similar substitutions into equation (5) and recalling that QA = Qc yields
19840 = 17,000 QA + (QA + V)Cg (8)
Equations (7) and (8) can be combined to form the quadratic equation
QA2 + -945QA + .50AV + .028V = 1.751 = 0 (9)
which has the solutions QA = -71 gpm, V = 1.5 gpm when equations (1), (2),
(3), and (4) are used as constraints. Substituting these values into
equation (1) yields QQ = 2.'21 gpm and Qg = 1.5 gpm. Also QC = QA = •?1 8Pm-
Making similar substitutions into equation (4) yields CD = 3512 mg/1.
The exact mass balance around the' recovery tank can be represented
[as compared to equation (2)] as
Qc Cc + QD CD + Q4 CL> + Q5 C6= Q2 C2 + Q3 C3 + Q5 C5 (10)
This mass balance is illustrated in Figure 5.
Outside
^rag-out,
Inside
drag-out1, '
VD
Resultant to spray
rinse, Q C,
Water
addition,
QiP/.
S4i
Vi
Immersion rinse
Recovery,
Q2c2
QCCC + QDCD
Return from spray
' rinse, Q Q
b o
Overflow
to waste,
Q3c3
Q6C6 = Q2C2
Q5C5
Figure 5. Mass Balance In Equation (10)
Making all known substitutions into equation (10) and using Qg = Q7 = 20 gpm
(from field measurements) reduces (10) to the two variable equation
Q6 = (32005 + 8.48)/20
(11)
Also, the complete mix of the next rinse that drains back into the recovery
tank can be represented as
Q7 C7 = Q6 C6 + Q8 C8
(.12)
This mass balance represented by equation (12) is illustrated in Figure 6.
Using assumptions B and C, Qs = Q8 and C8 = GS, and substituting in to
equation (12) reduces to the two variable equation
32005 - 20C6 = C6 Q5
(13)
395
-------
Resultant
spray rinse
Q5c5
/
to
'01
-o
Spray
rinse
"\
u
Results
final
Q8C
Q
o
nt to
rinse
Resultant to
8 final rinse
iW
sultant returns
to rinse tank, Q,C
6 6
Q5c5 + Q7c? = Q6c6 + Q8c8
Figure 6. Mass Balance in Equation (12)
Equations (11) and (13) can be combined to form
Q52 + .0265Q5 - .5300 = 0
(14)
which has the solution Q5 = .71. Substitution of this result in (11) yields
Cg = 11.8 mg/1. Finally using Qg = 30 gpm obtained from field measurements
in the final rinse complete mix produces C^Q = GH = .27 mg/1, This result
indicates that essentially no aqueous tin remains on the strip after the
last rinse. A similar analysis can be used to calculate fluoride flows
around the recovery. All results are summarized in a mass balance
illustrated in Figure 7.
40 gpm
20 gpm
Q0 = 2.92 gpm ]
238 Ib./d tin L.
506 Ib./d fluoride]
,- 555 Ib./d fluoride
2X I (as NaF-HF)
\ * V7 * /
V-sTin = 320 mg/1 r}/'
Fluoride = 1426 mg/1
[Q2 = 3 gpm
J 111.5 Ib./d tin
*^=Jpl.3 Ib./d fluoride
/
J
lo *
-u t
J^^
IT
-Pr
227
|_1010
1 Q1 . IQ^
JJJj [:;
59 gpm
Ib./d tin
Ib./d fluoride
= .71 gpm
Ib./d tin
.4 Ib./d fluoride
I Ib./d = pounds per day
KEY- 8Pm = gallons per minute
mg/1 = milligrams per liter
Figure 7. Mass Balance of the Recovery Tank Using the Model.
USE OF THE MODEL FOR IN-PLANT WATER RECOVERY
Using the framework presented by the model theory, it is possible to
project the effects of measured charges in rinsing practices. Any rinsing
change discussed, here, is presumed to have a beneficial effect on the
overall tin and fluoride mass balances, i.e., changes that result in less
396
-------
tin and fluoride being wasted to treatment via rinsewaters.
One simple change in operating practice is to use recovery tank water,
instead of fresh water, for strip wetting in the plating bath. This reduces
the use of fresh water and allows for re-use of rinsewater.
Another means of increasing recovery (or reducing waste) is to use the
mass balance model to prescribe a change in the rinsing program. As shown
in Figure 7, the amounts of tin and fluoride recovered is limited, because
the recovery flow cannot exceed 3 gpm. Thus, the only other way to increase
recovery is to increase the concentrations in the recovery tank. This
increase in concentration can be effected by reducing the volume of water
used in the recovery rinse. As shown in Figure 7, Q^ and Qy are the only
sources of fresh water input, and either can be reduced to raise the
recovery tank concentrations.
Both Qi+ and Qy are used to dilute the rinse tank, but Qy has the added
purpose of a spray rinse. This use of Qy is an example of counter-current
rinsing, which is described in various literature as a more efficient use
of water than any once-through use. (3) (4) Thus, Qif should be eliminated
and any deficiency caused by its elimination can be made up using Qy, This
allows Qy to be the single control parameter which can be used to regulate
recovery tank concentration and overflow rate.
The overflow rate, Q3, is the dependent variable used to represent the
effects of varying Qy, since it contains the mass rate Q3C3, the most
significant loss of tin and fluoride from the system. Thus Q3 is used
as the indicator of net effects of changes made on the rinsing system.
This analysis will proceed with the intention of representing the
recovery tank concentration as a function of the recovery tank overflow,
Q3. To begin, the mass balance around the recovery tank [equation (10)]
will be examined. From Figure 5,
Qc Cc + QD C D+ Q6 C6 + Q4 C4 = Q2 C2 + Q3 C3 + Q5 C5 (10)
The values Cp, Qg» Cg, C2, C3, C$ and Q3 will vary as the inflow to recovery
tank is changed. The values C2, C3 and C5 are all equal and will be con-
sidered simply as C, the recovery .tank concentration. The values Qc, GC> Q5
are known from the model development. Qi^ = 0 and is eliminated. Therefore,
the unknown values in this equation are Cp, Qg Cg, Q3 and C.
The value C can be simplified using equation (6), and recalling that
r = r
CD S
CD=
Q2 C + QA CA
All of the values in equation (15) are known, except C.
The value Qg can be reduced by using the volume balance around the recovery
tank.
397
-------
Qc + QD + Qi, + Q6 = Q2 + Q3 + Q5 (16)
Substituting all known values reduces this equation to
Q6 = Q3 + .79 (17)
Also the complete mix of the application of Q7 [equation (12)] reduces to
Q5 C5
C6 = - (where C7 = 0, C7Q7 = 0) (18)
Q6 + %
Substituting equation (17) into (18) with other known values reduces to
• 71C
Substituting equations (15), (17), and (19) into equation (10) produces
Q2C + Q, C .71C
Q- Cr + Q_ - — - + (Q3 + -79) - = C(Q3 + Q2 + Q5) (20)
C C D QB + QD Q3 + 1-5
Substituting all known values into equation (20) yields
19260Q3 + 28890
C = - (2i)
Q32 + 2.71Q3 + 2.32
This function is plotted in Figure 8. Also in Figure 8 is the prediction
for fluoride using the constituent ratio assumption.
The projections in Figure 8 should be considered only as physical mass
balance functions which were calculated without considering product quality.
Any recovery tank concentration that results in greater masses of tin and
fluoride on the strip as it leaves the application of Q7, will require
greater amounts of rinsewater in the final rinse, Qg. This necessity may
outweigh the benefits of the higher recovery tank concentration. Also, the
fluoride to tin ratio of 4,46 is an assumption based on sampling results.
At this writing there is no known equilibrium relationship between tin and
fluoride in the recovery tank. In actuality, this ratio may change as the
concentration of tin and fluoride change.
TESTING THE MODEL
In order to evaluate the performance of the rinsing model, a test was
conducted whereby in-plant flows of rinsewater were reduced to increase the
recovery tank concentrations. Over a period of three weeks the use of
rinsewater was systematically re'duced so that the overflow from the recovery
tank was lowered to less than 20 gallons per minute. Also accomplished
was the use of 2 gpm of recovery tank water for strip wetting. Results
of this testing are illustrated in Figure 9.
398
-------
z
8
5000
4000
3000
2000
1000
500
400
300
FLUORIDE
46 C
tin
~ 5000
"g 4000
z 3000
M
1 2000
1
u
z
8 1000
^
I 500
o 400
u
S 300
Figure. 8.
10 20 30 40 50 60
RECOVERY TANK OVERFLOW Qj (gpm)
Rinsing Model Recovery
Tank Concentration
Vs. Overflow Rate
* The model prediction Is modified In this
figure to reflect the additional 2 gpm
used for strip wetting
N = number of data points at
specific Qj
19260(Q3 + 2) + 28890
2.71(Q3 + 2) + 2.32V
10 20 30 40 50 60
RECOVERY TANK OVERFLOW Q3 (gpm)
Figure 9. Results of Model Testing
The data points for tin follow the general path of the predicted
model. Fluoride analyses remained statistically unchanged and therefore
were excluded from further analysis. Testing below a Q3 of 17 gpm was
prohibited by physical constraints associated with the rinsing tank. (7)
Although the model predictions for tin were not followed exactly and
although fluoride modeling was excluded for further analysis, the desired
trend of reducing mass rates of discharges was followed Using the
beginning and endpoints in Figure 9 it can be calculated that a 40%
reduction in tin mass rate via 0.3 was achieved. Since the fluoride con-
centration remained unchanged, it was calculated that a 70% reduction in
fluoride mass rate, via Qs, was achieved due simply to a reduction in the
flow rate of 0,3. These reductions do not represent those possible. Further
testing, at Qs less than 17 gpm, could be performed if the physical limi-
tations preventing this were eliminated. These limitations could be
eliminated by a rinsewater control system which maximizes recovery and
minimizes losses.
A MODIFIED RINSING MODEL
Assuming that an engineered set of rinsing controls are installed at
the J&L operation and that the functional relationship of the model is
accurate, a new mass flow condition is constructed to represent steady
state rinsing. Two extremes of this steady state rinsing model can be
constructed.' Phase I represents steady state at 0,3 = 17 gpm, the endpoint
of the model test. The mass balance associated with Phase I is shown in
Figure 10. Also shown in Figure 10 are the conditions of steady state
Phase II which represent the ultimate endpoint of Q3 = 0. (The flow rates
in parentheses are the average daily flow rates assuming a 12% delay' rate.
Without parentheses is flow during operation, i.e., during delays flows
are off.) Phase II may not be amenable to a quality product, and is only
399
-------
an extrapolation used for comparison purposes,
DRAGOUT
I 2.92 (2.57) gpm
221 Ib/d tin
II 2.92 (2.57) gpm
312 Ib/d tin
IN RECOVERY TANK
I 942 mg/1 tin
II 5742 mg/1 tin
r?
RECOVERY
15 (4.40) gpm
50 Ib/d tin
II 5 (4.40) gpm
303 Ib/d tin
I 20 (17.6) gpm
II 2.79 (2.46) gpm
•f' •'
ITT 1
.71 (.66) gpm
<1 Ib/d tin
'II .71 (.66) gpm
9 Ib/d tin
OVERFLOW
I 17.2 (15.14) gpm
171 Ib/d tin
II
0.0 gpm
0 Ib/d tin
Figure 10. The Modified Mass Balance
RECOVERY ANALYSIS USING THE MODIFIED MODEL
There are several recovery technologies which can be used to recover
tin (and in some cases, fluoride) from process wastewater. These tech-
nologies include electrolytic recovery, chemical precipitation, evaporation,
electrodialysis and reverse osmosis. (3)(4)(5)(6) The mass rates in
Figure 10 can be scaled up to reflect both of J&L's tin lines' wastewater,
including leakage and spillage, and all of the wastewaters generated in
each of the Phases can be applied to each of the technologies.(7) A
comparative summary of these applications is in Table 2.
TABLE 2. COMPARISON OF THE MODIFIED RINSING MODEL APPLIED TO RECOVERY ALTERNATIVES
(Cost Unit/Year)
Alternative
Phase I & Electrolytic Recovery*****
Phase II & Electrolytic Recovery*****
Phase I & Electrodialysis
Phase II & Electrodialysis
Phase I & Reverse Osmosis****
Phase II & Reverse Osmosis****
Phase I & Chemical Precipitation
Phase II & Chemical Precipitation
Phase I & Evaporation
Phase II & Evaporation
No Recovery
Net Value Of*
Recovered Tin
11.4
18.3
11.4
8.4
11.4
8.4
5.3
5.6
11.5
8.3
-12.9
Cost Of**
Alternative
4.2
4.2
6.0
2.9
7.3
3.8
4.2
2.7
23.4
11.1
0
Net Value***
Of Alternative
7.2
14.1
5.4
5.5
4.1
4.6
1.1
2.9
-11.9
-2.8
-12.9
Rank
2
1
4
3
6
5
8
7
10
9
11
* This value is calculated as AB-CD where: A is the amount of tin recovered; B is the
unit value of recovered tin (different for different recovery methods); C is the
amount of tin lost to wastewater; and D is a reasonable value of the lost tin.
** Capital recovery and operation.
*** Net value of recovered tin minus cost of alternative.
**** Reverse osmosis may not be a viable alternative because the process has not been
tried in this application.
***** Electrolytic recovery is not associated with the tin plating production facility,
but is a separate treatment method that uses similar electrochemical principles.
400
-------
Electrolytic recovery is shown to have the best economic potential.
This is due to the fact that recovered tin can be recycled directly in a
tin smelter with no intermediate refining steps. Electrodialysis and
reverse osmosis are somewhat lesser economic choices, but are still better
than the "No Recovery" option. Chemical precipitation has a lesser benefit
because the tin would have to be extracted from a hydroxide sludge.
Evaporation is not attractive in this application due to high capital
and energy costs associated with this option. In each case, the Phase II
application has a higher economic benefit compared to the Phase I
application, and economic recoveries can be interpolated for rinsewater
reductions between the values defined by these two extremes.
CONCLUSIONS AND RECOMMENDATIONS
(1) A model for tin line wastewater application was developed, tested
and shown to be a useful representation of the rinsing system,
(2) Although J&L can continue to produce a high quality tin plating
effluent from the Chemical Rinse Treatment Plant that will meet all
applicable regulations, there appears to be several tin recovery
alternatives that are economically favorable.
(3) This study has shown that a more careful application of rinsewater
can reduce the mass rates of tin and fluoride discharge and it is recommended
that improvements to the rinsing system continue.
(4) The requirements and effects of fluoride were not adequately
represented by this study, and it is recommended -that a better under-
standing of the disposition of fluoride be pursued.
(5) It is recommended that a wastewater monitoring program be
instituted for a better understanding of rinsing requirements.
(6) It is recommended that a training program for associated operating
personnel be developed, which would include operating parameters to be
maintained and goals which stress maximization of recovery and minimization
of waste.
REFERENCES
1. Tin Electrodepositing Composition and Process, E. W. Schweikher,
assignor to E. I. DuPont deNemours and Company, U.S. Patent #2,402,184,
Patented June 18, 1946.
2. Electrodeposition of Tin, E. W. Schweikher, assignor to E. I. DuPont
deNemours and Company, U.S. Patent #2,407,579, Patended September 10, 1946.
3. Electroplating: Fundamentals of Surface^ Finishing, F. A. Lowenheim,
McGraw-Hill Book Co., 1978.
401
-------
4. Environmental Pollution Control Alternatives; Economics of Wastewater
Treatment Alternatives for the Electroplating Industry, U.S. Environmental
Protection Agency, Technology Transfer - Industrial Environmental Research
Laboratory, Cincinnati, OH, June, 1979 (EPA 625/5-79-016).
5. Treatment of Waste Water from the Tinning Lines at Hoogovens Ijmuiden
B.U., D.C. Heijwegen, P. VanLanghout, Presented at the Second Symposium on
Environmental Control in the Steel Industry, Chicago, June 11-13, 1979.
6. Summary Rejort: Control Technology for the Metal Finishing Industry-
Evaporators , U.S. Environmental Protection Agency, Technology Transfer -
Industrial Environmental Research Laboratory, Cincinnati, OH, June, 1979
(EPA 625/8-79-002).
7. A Mass Balance Model for Rinsewater in a Continuous Strip Halogen
Electrolytic Tinning Operation for use in Evaluating Wastewater Treatment
and Recovery Alternatives, J. Skubak, April, 1981, Graduate Project,
University of Pittsburgh.
402
-------
INVESTIGATION OF THE SOLID-LIQUID PHASE SEPARATION
OF PREHEATED AND PIPELINE CHARGED COKE BATTERY CHARGING LIQUOR
by: S. R. Balajee, A. I. Aktay, and R. R. Landreth
Senior Research Engineers, Inland Steel Company
East Chicago, Indiana 46312
ABSTRACT
Inland's preheat and pipeline charged "C" Battery was built with a
solid-liquid phase separation system containing a series of tanks to separate
the coal fines collected during the charging process from the charging liquor.
This phase separation system did not remove enough coal fines to prevent
plugging of the charging liquor spray nozzles and a build up of material in
the charging main when the liquor was recycled. Consequently, several phase
separation methods were investigated on the laboratory, pilot plant, and plant
scale to establish their ability to decrease the solids content of the "C"
Battery charging liquor to a concentration suitable for recycling. These
methods were: gravity settling, horizontal belt filtration with and without
precoat, hydrocyclone separation, spiral rake classification, conventional
flotation, and dissolved air flotation^.
All of the phase separation methods were able to separate coal fines to
some degree, but three methods, namely, gravity settling, dissolved air
flotation, and horizontal belt filtration with precoat gave the best results.
Gravity settling is easy to operate, but requires a large amount of land area
for installation of a pit and the use of chemical additives to accelerate the
settling of fines. Dissolved air flotation requires chemical additives and a
liquor with a high solids content for best operation. Horizontal belt
filtration requires a precoat for efficient operation. In view of the
variable solids and tar concentrations in the charging liquor, the gravity
settling method was selected and settling pits were installed to separate coal
fines from charging liquor at Inland's "C" and No. 11 preheated and pipeline
charged coke batteries.
INTRODUCTION
In 1974 and(£978, Inland Steel commenced operation of "C" Battery^ and
No. 11 Battery , respectively. Both /batteries are equipped with coal
preheating and pipeline charging systems^ '. In this system, wet coal is
preheated by hot gases to 220-260 C1' to remove moisture from the coal
prior to charging. The preheated coal is conveyed into the ovens by steam in
two pipelines. Coal fines are carried into the charging main during oven
403
-------
charging where they are removed by recycled charging liquor sprays. The
liquor is diverted to the coal fines recovery system where it is cleaned of
solids and recycled. However, during the initial phase of "C" Battery's
operation, the coal fines recovery system did not function efficiently and
large amounts of coal fines in the recycled charging liquor caused plugging of
the spray nozzles in the charging main. Therefore, a study, was initiated to
investigate methods to improve the solid-liquid phase separation in the coal
fines recovery system.
In this paper, the testing of various devices on the laboratory, pilot
plant, and plant scale are discussed in terms of improvement in the
solid-liquid phase separation of the "C" Battery charging liquor. On the
basis of this investigation, conclusions were drawn concerning the phase
separation devices tested, which led to the selection of the most plausible
phase separation method - the gravity settling pit. The performance of the
plant scale settling pit operation was evaluated.
LITERATURE REVIEW
In comparison to conventional wet charged coke batteries, preheated coke
batteries can utilize relatively high percentages of,«weakly coking coals in
the blend, and reduce airborne particulate emissions. However, the capture
of these particulates by various liquor streams has transformed air pollution
into a water based environmental problem.
Because the technology of utilizing a preheated and pipeline charged coke
battery for manufacturing blast furnace coke is relatively new, the literature
on the solid-liquid phase separation for charging liquor from these batteries
is rather limited. There appears to.be only two literature sourceSr oji\ this
subject - the Rrdtish Carbonization Reserach Association (BCRAr " and
Degremont-Laing. '
The BCRA work on the phase separation of charging liquor included pilot
scale testing of several mechanical phase separation devices in a combined
flowsheet which consisted of a decantation tank, froth flotation cells, and
rotary vacuum disc filter. In the test, the final filter cake product
contained less than 50% solids. The filtrate was reported to be essentially
free of ?§li
-------
;d by a preconcentration step (sink/float separation or floc/cujlation-
ion step) and the liquor tar concentration was 15% or less/ ' ' In
thickened
decantation
addition, another dewatering device, called a squeezer, was tried on
prethickened charging liquor. Cationic and anionic polyelectrolytes were
first added to flocculate the thickened slurry, then it passed into a drainage
section over a filtering belt, and finally the sludge entered a pressing
section where air at a pressure of up to 70 kPa was applied to both sides of
the filter cloth to squeeze the sludge. A solid mate/Mai containing about 20%
moisture was produced at the rate of about 3.3 kg/m s dry solids. The cake
thickness varied from 4 to 7 mm. The squeezer type filter press was tested in
collaboration with Degremont-Laing. The squeezer press appeared to perform
well even with the addition of preheater washer fines to the charging liquor.
It was also indicated that the dry solid product from the saueezor press was
recycled at the rate of 2% of the preheater throughput rate. '
The Degremont-Laing v ' phase separation system for the charging liquor
included: (1) a fines wetting tank for the addition of two chemicals, ferrous
sulfate and a polymer, and (2) a slurry sedimentation tank. An additional
stage of prolonged heating of the slurry was necessary when the tar content of
the slurry exceeded 10%. The cleaned liquor was of such a quality that it
could be recycled in the charging main; however, the solids concentration of
the treated clean liquor was not specified. The initial solids content of the
charging liquor was reported to be 3000-4000 mg/kg, and the size consist was
about 80% -0.15 mm. It was stated that the phase separation system should
process liquor at a temperature high enough to keep the tar in a fluid stage.
In addition, the flow of liquor should be by gravity in order to minimize
pumping requirements and pump related problems, and the collection of the
concentrated slurry at the clarifier underflow should be as straight forward
as possible. Operating data for a commencval Degremont-Laing charging liquor
phase separation system was not reported/ '
This literature review includes information on the solid-liquid phase
separation of charging liquor published after the experimental test program
was planned.
EXPERIMENTAL TEST PROGRAM
The solid-liquid phase separation of preheated and pipeline charged "C"
Battery charging liquor was investigated by the following methods:
LABORATORY SCALE
Jar test
Leaf filtration
PILOT PLANT SCALE
Horizontal belt filter test with and 'without precoat
Hydrocyclone test
Spiral rake classifier test
Conventional froth flotation test
405
-------
Dissolved air flotation test
PLANT SCALE
Conical Tank Test
EXPERIMENTAL PROCEDURE AND EQUIPMENT
The plant scale conical tank test with the charging liquor was conducted
before any of .the other tests. Modifications made to the original conical
tank system affected all the laboratory and pilot scale tests and, therefore,
the conical tank test is discussed first. "C" Battery was converted from side
charging to top charging in an attempt to reduce the amount of solids going to
the coal fines recovery system. During side charging, coal enters the oven
at an angle from the pusher side of the oven. Under top charging, coal enters
the oven from the top through one of the decarbonization holes. The plant
scale conical tank test was made only during side charging conditions, whereas
most other phase separation tests were conducted under both side and top
charging conditions at "C" Battery.
PLANT SCALE TEST
Cojiical Tank
In the original design of the coal fines recovery system, the charging
liquor containing coal fines, tar, and other compounds (Table I) flowed by
gravity from the charging main to a conical tank for solid-liquid separation.
Later, a vertical and a horizontal tank were added to the system in series to
improve the solid-liquid phase separation of the conical tank underflow, but
the additional tanks did not solve the problem of recycled fines. The flow
diagram of the original charging liquor system is given in Figure 1. A 1/10
scale model of the original conical tank was built in order to determine the
mechanism of the solid-liquid separation. Subsequently, an aeration ring and
two whirl jet nozzles were placed in the conical tank in order to introduce
fine uniformly distributed air bubbles which float the coal fines to the top
of the tank. Based on the phase separation tests with the model conical tank,
two Heyl and Patterson nozzles were placed in the conical section of the
existing -tank at "C" Battery. The modified conical tank, which has about
20,000 dm liquor .capacity, is shown in Figure 2.00)
LABORATORY TESTS
Jar Test
In order to determine the feasibility of using gravity settling for
separating coal fines from charging liquor, laboratory jar tests were
performed. These tests were conducted in 500 ml and 1000 ml jars to determine
the optimum dosage and synergistic effect of cationic and anionic polymers on
the settling rate of coal fines in the charging liquor. Two liquor samples
from the conical tank intake and the horizontal tank (Figure 1) were tested at
a temperature of about 60 C. Betz Cationic 3390 and Betz Anionic 3330, as
well as Nalco Cationic 7763 and Nalco Anionic 7132 polymers, were used. The
406
-------
CHARGING MAIN
CHARGING
LIQUOR
CONICAL TANK
OVERFLOW
i— (CONICAL TANiCl^
|T^ CTD
1
HH STRHintK
BACKWASH
UNDERFLOW
n A
REJECT
, , | | 1 | | J .
'T 1
CHARGING L
LIOLtnR
FLOTATION
CELL
-T
FROTH FROM
SCRUBBER
FROTH
FLOTATION CELL
T 1) DISC FILTER
STRAINER
FILTER CAKE
.ORIGINAL SYSTEM
•LATER ADDITIONS
— 1
aj»
1 1 *
I— !
1
1
1
1 1
VERTICAL
TANK
V
(HORIZONTAL'
i TAW* I
TANK
.SLUDGE
Figure 1 Schematic Flow Diagram of the Original Charging
Liquor System at C Battery
TYPICAL CHEMCAl MALTSIS OF CHAR61II6 LIQUOR
Solid! (11
Fixed Carbon
Volatile utter
Ash
Sulfur
Oil 1 Tar (Chloroform Solubles)
Energy
Liquid (m/kQ)
Free tanonla (as N?)
Fixed »B»on1j (as N2>
Phenol
Cyanide
Thlocyanate
Sulfate
Sulflde
011
Chlorides
Suspended Solids
Total Dissolved Solids
62.10
-32.W
5.50
0.85
7.66
30 Hi/kg
250
2950
540
11
550
950
12
60
9725
2000
15448
. . LIQUOR
Figure 2
UNDERFLOW
Modified Conical Tank with
Heyl and Patterson
Nozzles
pH (units)
407
-------
cationic polymer dosage ranged from 5 to 45 mg/kg, whereas, 0.25 to 10 mg/kg
of anionic polymer was used. The cationic polymer was first added to the
liquor with rapid mixing for 60 seconds followed by the addition of the
anionic polymer and mixing for another 30 seconds. The settling rate was
determined by observing the solid-liquid interface as a function of time..
Samples of the supernatant liquor were taken to determine the solids
concentration.
Leaf Filtration
A 108 mm diameter EIMCO leaf filter with a filter area of 0.00929m2 was
employed to evaluate filter cloths and to study the filtration characteristics
of "C" Battery charging liquor at 25 C to 90°C. Three filter cloths, nylon
(NY-518F), polyethylene (PO-^OIHF), and polypropylene (POPR-853F), were
tested. Approximately 3.79 dm of charging liquor was filtered using a vacuum
pump with a suction capacity of 101.3 kPa. The filter time, suction, and
filtrate volume were recorded, and the wet and dry weights of the cake were
determined-.
PILOT SCALE TESTS
Horizontal Belt Filter
An Eimco Model 112 horizontal belt filter was used to separate the solids
from the charging liquor. The belt filter was 3.7 m long by 0.3 m wide, had
the abililty to generate a suction of 6.2 kPa .and ran at a speed of 0.043 m/s.
Two different filter cloths were used during the testing program, nylon
(NY-518F) and polypropylene (POPR-853F). The nylon cloth had finer pores than
the polypropylene cloth. Liquor was fed from various sources to the unit
through a feed box at a rate of 0.63 to 3.15 dm /s. The liquor was pulled
through the filter belt into a drain trough with a suction that was usually
5.0 kPa, but always greater than 2.5 kPa. The solids formed a cake and
remained on the belt until it reached the end, where the cake fell onto a
conveyor. Samples were taken of the feed, filtrate, and filter cake from the
unit for solids content analyses.
Horizontal Belt Filter with Precoat
In order to increase the filtration efficiency of the horizontal belt
filter and overcome the problem of filter cloth plugging due to the tar
content of the charging liquor, a precoat filtration system was employed. The
same Eimco horizontal belt filter was used for the precoat filtration study.
The belt filter ran at a speed of 0.03 m/s for all tests. Two different
filter cloths were used during the pre-coat filtration runs; nylon (NY-518F),
and polypropylene (POPR-853F). The froth from the scrubber flotation cell was
fed to the belt filter as a precoating material, since the solids
concentration is higher and the tar content is lower. The charging liquor was
fed onto the filter cake which was previously formed on the belt. The flow of
preheater scrubber effluent varied from 0.21 to 3.79 din /s, whereas, the
charging liquor flow rate varied from 0.16 to 1.89 dm /s. The belt was
maintained at a suction which was usually about 5.0 kPa, but always greater
408
-------
than 2.5 kPa. Samples of charging liquor feed, scrubber liquor feed, filtrate
and filter cake were taken for solids content analyses.
Hydrocyclones
Two WEMCO cyclones, 76.2 mm and 152.4 mm in diameter, were employed in the
solid-liquid phase separation test of C Battery charging liquor. Two cyclone
products were produced in these tests; the cyclone overflow liquor containing
mostly -0.045 mm coal particles, and the cyclone underflow liquor which
contains coarser coal particles.
The 76.2 mm diameter cyclone had three apexes with diameters of 6.35 mm,
9.5 mm, and 12.7 mm. Feed rates of 0.88, 1.32, and 1.7 dm /s were obtained by
operating the 76.2 mm cyclone at an inlet pressure of 69, 138, and 207 kPa,
respectively. An overflow/underflow volume ratio in the range of 2:1 to 10:1
was obtained by changing the inlet cyclone pressure or the apex diameter.
The 152.4 mm diameter WEMCO cyclone was lined with nitrile rubber in order
to prevent corrosion and erosion by materials contained in the charging
liquor. This cyclone contained a 50.8 mm diameter hydraulically adjustable
hinged apex operating at pressures up to 621 kPa. Tests were made at apex
diameters of 15.9 mm, 25.4 mm, 38.1 mm, and 50.8 mm. At operating cyclone
inlet pressures of 69, 138, and 207 kPa, corresponding feed rates of 7.26,
11.36, and 13.25 dm /s were obtained. An overflow/underflow volume ratio of
10:1 was observed with the 152.4 mm diameter cyclone having an apex of
38.1 mm. The phase separation tests were made using the 152.4 mm diameter
cyclone only under top charging conditions.
The cyclone feed, overflow and underflow liquor streams were sampled for
suspended solids. The flow rates of the cyclone streams were measured.
Spiral Rake Classifier
A 0.46 m diameter x 3.5 m long spiral SM classifier was employed in the
phase separation tests. This WEMCO 150 Series classifier had a full flare
tank and a rated capacity of 44.3 dm /s for 0.15 mm separation, of solid
particles as a classifier sand. The rotation of the spiral rakes was set at
0.12 revVs. The flow rate of the charging liquor feed ranged from 1.89 to
4.40 dm /s. Coal solids must accumulate in the bottom of the tank and
underneath the spiral rake housing tank before a steady state condition is
reached and classified sand is produced. Two classifier products, a
classifier sand and an overflow liquor, were sampled along with classifier
liquor feed and analyzed for solids concentration.
Conventional Flotation Cell
3
A 227.4 dm capacity Heyl and Patterson (H+P) flotation cell was used in
separating solids as a froth from the charging liquor. The flotation cell
operates at a feed rate of 1.20 to 1.89 dm./s which gives a retention time of
2 to 3 minutes. One third of the liquor by volume is recirculated at an
agitation pressure of 13.5 kPa. Air bubbles are generated when the recycled
liquor comes in contact with a curtain of air in the 6 vortex chambers. The
fine air bubbles are attached to the solids which carry them to the top of the
409
-------
cell in the form of a froth. Each v9rtex chamber is equipped with a 9.5 mm
diameter ceramic spray nozzle. An air pressure of 34.5 kPa is required to
maintain a desired air/liquid flow ratio of 3.74.
Samples of froth, underflow, and feed were taken for solids concentration
analyses. The H+P flotation tests were made with and without the addition of
200 mg/kg of alum (aluminum sulfate) as a flocculation agent.
Dissolved Air Flotation Unit .
A Carborundum Model 50 Dissolved Air Flotation Unit (DAF) was employed for
separation tests with the charging liquor. The DAF unit had a surface^area of
2.63 m , a 4410 dm flotation tank, and a flow capacity of 3.78 dm /s; all
tests were performed at 3.78 dm /s. The charging liquor enters the unit
through a mixing tank into which flocculants can be added. The liquor flows
through a pump where air is injected, and into a tank which is pressurized to
between 138 and 276 kPa; most tests were performed at 138 kPa. Under
pressure, the air dissolves, and, upon release of the pressure in the
flotation tank, micron size bubbles are formed which are enveloped by the
suspended solids in the charging liquor. These air bubbles contain coal fines
and rise to the surface where they are removed as froth by skimmers. Any
large particles which settle to the bottom of the tank are removed by a
scraper arm. The clean liquor exits through side ports.
Alum and an anionic polymer (Carborundum CF500A) were used separately and
together as flocculating agents. The alum and polymer concentrations were 178
mg/kg and 4 mg/kg, respectively. The DAF tests were made only under top
charging conditions. Samples were taken of the froth, feed, and underflow
from the unit for solids concentration analyses.
RESULTS AND DISCUSSION
The chemical analyses of a typical charging liquor and the contained
solids age presented in Table I. The liquor is hot with a temperature of
about 90 c. The solids concentration was variable, fluctuating from 400 to
20,000 mg/kg. The solids are high and variable in oil and tar (3-10%). The
typical size analysis of the solids indicated that solids containing
relatively coarse coal particles are obtained in the charging liquor during
side charging than top charging (Table II). The difference in the size of the
charging liquor solids may affect the separation efficiency of phase
separation devices tested under both charging conditions. The apparent size
of the solids may be coarser because of agglomeration of fines that may occur
during transport. The agglomeration of coal fines also depends upon the
initial solids concentration, and oil and tar concentration, which are known
to fluctuate during the charging cycle. Thus, variable solids concentration
and the tar content of the solids could have affected the performance of the
various phase separation devices that were tested in this investigation.^ '
410
-------
TABLE II
TYPICAL SIZE ANALYSIS OF CHARGING LIQUOR COAL FINES
Cumulative Weight % Passing
Screen Size C Battery
(mm) (Side Charging)
"As is" Coal Fines
— — — — —_
0.150 30
0.075 ND
0.045 0.6
0.038 ND
£h]_0£0fprm_k[ashed C_oaJ_Fj[nes* C Battery
(Top Charging)
0.150 72
0.075 58
0.045 43
0.038 39
0.020 30
0.010 16
ND: Not Determined
* Oil and tar free coal fines
Wet Screening done above .038 mm; Coulter Counter Analysis for
-0.038 mm fines.
TABLE III
EFFECT OF AIR NOZZLES ON THE CONICAL TANK SEPARATION EFFICIENCY .
Average Solids concentration
x +_ a (mg/kg)
No. of Nozzles* Charging Strainer Underflow Solids Removal
in the Conical Liquor Backwash from Conical Efficiency
Tank Feed Feed Tank (%) j a
0 6273+1582 731 + 697 5526 +_ 956 28+6.6
2 2761 + 1421 354 + 228 987 + 704 71 + 13.1
*Heyl and Patterson Nozzles
411
-------
PLANT SCALE TEST
Conical Tank
Solids concentration in the feed to the conical tank fluctuated during the
charging cycle. Typically, there were 400 mg/kg of solids in the charging
liquor at the beginning of the side charging cycle. A peak value in the range
of 19500 mg/kg was reached approximately 5 minutes into the cycle. In some
cases, a second peak was observed.
Based on a constant "C" Battery charging liquor flow rate of 113.6 dm /s
and an average charging cycle of 11 minutes, the total coal fines carryover
during side charging for four oven charges were calculated to be 323, 463,
480, and 654 kg/charge, far greater than the design capacity of 45 to
91 kg/charge. Coal fines carryover increased as much as 50% when two oven
charges overlapped. The solids removal efficiency was increased from
71% when fine air bubbles were introduced in the modified conical tank
(Table III and Figure 2). Even with the improved solids-liquid separation
efficiency, the solids concentration of the conical tank underflow was still
well over 200 mg/kg, the desired level for recycling.
It appears that possible oxidation of the coal particles and changes in
the surface characteristics of the coal particles due to tar and oil in the
charging liquor may have caused many coal particles to sink rather than float
in the conical tank.
LABORATORY TESTS
Jar Test
The results of the laboratory jar tests indicated that a combination of
cationic and anionic polymer addition gave the best settling rate of the coal
fines in the liquor. Betz polymers gave an equivalent settling rate of about
150 mm/min of coal fines at lower polymer concentrations as compared with
Nalco polymers (Table IV). The clear liquor, after the settling of floc-
culated coal fines for less than 3 minutes, averaged less than 10 mg/kg of
suspended solids, thereby suggesting that the settling of coal fines in a pit
is a satisfactory method for obtaining a clean charging liquor for recycling
to the charging main.
Leaf Filtration
The leaf filtration tests with "C" Battery charging liquor were performed
to determine the filterability of the liquor. The results are summarized in
Table V. A solids removal efficiency of up to 99.9% was achieved. The clean
liquor averaged less than 50 mg/kg of suspended solids. The filtration rate
increased with the initial suspended solids concentration in the liquor. The
usage of chemical additives did not increase the filtration rate. The Nylon
filter cloth NY-518F and the polypropylene cloth POPR-853F appeared to
withstand the attack of the various coal chemicals contained in the charging
liquor, whereas the polyethylene cloth PO-801HF hardened after filtration.
412
-------
TABLE IV
LABORATORY JAR TESTS FOR C BATTERY CHARGING LIQUOR
Polymer Combination and Dosage Rate for Settling of Coal Fines*
Cationic Polymer
+
Anionic Polymer
Betz Nalco
Type mg/kg Type mg/kg
3390 20 7763 40
3330 0.5 7132 6
B. £harg_i_ng_L_[quqr .frpmjfertica± TankJJriderf l£W
Cationic Polymer
+
Anionic Polymer
Betz
Nalco
Type mg/kg Type mg/kg
3390 20 7763 20
3330 0.25 7132 2
tquivalent settling rate of about 150 mm/min.
TABLE V
LABORATORY LEAF FILTRATION OF CHARGING LIQUOR
Filter Liquor Filtering
Cloth Temp. Vacuum
TO (kPa)
NY-518F
25
25
25 (a)
25(b)
44
70
80
80(a)
80 (b)
POPR-853F 25
25
Tpj> £harg_i_ng.
NY-518F 25
88
95
95
95
70
88
95
91
85
88
95
Solids
Concentration*
(mg/kg)
3506
6077
3313
3329
5028
5753
16517
14565
11286
6024
1321
853
Moisture'
In Filter
Cake
32
42
46
71
39
32
41
40
57
44
63
44
Solids
Removal
Efficiency
(*)
99.4
99.0
99.9
98.9
99.2
99.0
97.4
97.6
97.4
98.8
99.9
99.9
Filtration
Rate
(kg/mz-s)
0.03
0.11
0.05
0.06
0.05
0.08
0.26
0.23
0.19
0.02
0.07
0.01
(a) Betz polymer added.
(b) Nalco polymer added.
* Charging liquor
413
-------
2
Filtration rates of 0.05 to 0.2 kg/m s were obtained. The suction during the
test ranged from 70 to 95 kPa. The filtration of "C" Battery charging liquor
was sensitive to the presence of tar and the solids concentration.
A preconcentration step may be required to obtain a feed with a uniform
solids concentration in excess of 1000 mg/kg, which is needed for proper unit
operating so that the tar in the liquor does not blind the pores of the
cloth/ ' The blinding is a result of (1) the absorption of tar on the
particle surfaces of coal, and (2) the formation of a sticky thin film of
tarry material on the cloth.
PILOT SCALE TESTS
Horizontal Belt Filter With or Without Precoat
Typical results of the horizontal belt filter tests conducted at
"C" Battery using a nylon filter cloth (NY-518F) are presented in Table VI.
It was found that there is no correlation between the solids removal
efficiency and solids concentration in the filtrate. However, it was noted
that the lower limit of acceptable solids concentration in the feed was about
1000 mg/kg, below which the belt filter solid removal efficiency is
drastically reduced (Figure 3). Thus, the horizontal belt filter appears to
be -able to clean charging liquor to recyclable levels (less than 200 mg/kg
solids concentration) when the solids concentration in the feed is high.
Horizontal belt filter tests with a precoat of the concentrated scrubber
effluent indicated a strong correlation between the feed rate and filtration
rate of solids (Figure 4). The test data on precoat belt filter operation
showed that the solids concentration of the scrubber effluent and the charging
liquor fluctuated from 3721 to 62883 mg/kg and 24 to 5668 mg/kg, respectively.
Average solids concentration of the filtrate was 122 mg/kg and ranged from 26
to 315 mg/kg. Average thickness of the filter cake was 6.6 mm. The average
filter cake moisture content was 29.8% and ranged from 23.0 to 42.7% (Table
VI). Precoating the filter cloth with a $9lid allowed efficient belt filter
operation even with charging liquor containing low solids or high tar content.
The precoat acts as a porous media which captures .fine coal particles and tar
present in the charging liquor, and allows the liquor to pass through without
blinding the filter cloth.
Hydrocyclones
The tests on the solids-liquid phase separation of charging liquor using
the 76.2 mm and 152.4 mm diameter WEMCO cyclones showed that typically during
side charging conditions, the 76.2 mm cyclone was effective in producing an
overflow containing 100 to 300 mg/kg of suspended solids. However, under top
charging conditions, both cyclones failed to provide an overflow with a
concentration of suspended solids below 200 mg/kg. The addition of alum (200
mg/kg) did not aid in the cyclone operation.
«
As shown in Figure 5, the 76.2 mm cyclone during side charging yielded a
95% solids removal efficiency and a concentration of 65-110 mg/kg solids in
414
-------
1000 20^0 30)
4000 5000
SOLIDS CONCENTRATION (mg/tgf
Figure 3 Typical Plot of Solids Removal
Efficiency Versus Solids Concentration
in the Feed for the Horizontal Belt
Filter
40 -
Figure 4 Filtration Rate Versus
Feed Rate of Solids in
the Combined Charging
Liquor and Precoat
Scrubber Effluent
1000 1500 2000
FEED RATE (g/min)
2500 3000
415
-------
TABLE VI
TYPICAL HORIZONTAL BELT FILTER OPERATING DATA FOR CHARGING LIQUOR
Filter
Cloth
Average
Charging Liquor
Solids Cone.
(mg/kg _+ o)
Feed Filtrate
Average
Filter Cake
Moisture
(%±<0
Solids
Removal
Efficiency
(%)
Filtration
Rate
(kg/mz. s)
Average
Scrubber Effluent
Solids Concentration
(mg/kg + o)
jn_de_Cha£gj[ng_a_nd_No £recoa_t
NY-518F
973 +_ 1535 33 +_ 19
1966 + 2714 48 + 13
16 +_ 2
20 + 5
96.6
97.6
.0007
.0015
T_O£ £hjirg_i_ng_a_ncl_No
NY-518F 2949 + 2048 82 +20 41 + 10
97.2
.0050
with
NY-518F
1245
2015
5668
+ 1189
+ 1804
+ 3132
124
222
54
+ 90
+ 150
T. 21
34.8
42.7
.23.0
99.
98.
99.
4
4
8
0
0
0
.028
.012
.036
5101
34898
48693
+ 28885
+ 16572
+ 35633
*Scrubber effluent was not used as a precoat.
100
u 80
z
uj
U
tt 60
O 40
f.
a 20
1000 2000 3000 • 4000 5000
SOLIDS CONCENTRATION (mg/kg)
125
Z
O
100 <
u
75Z
252
>
O
— 100
1600 3200 4800 6400 8000
SOLIDS CONCENTRATION (mg/kg)
Figure 5 Solids Removal Efficiency
and Overflow Solids
Concentration Vs Solids
'Concentration in the
Feed for the 76.2 mm
Cyclone During Side
Charging
Figure 6 Solids Removal Efficiency
and Overflow Solids
Concentration Vs Solids
Concentration in the
Feed for the 76.2 mm
Cyclone During Top
Charging
416
-------
the cyclone overflow at a solids concentration in the charging liquor feed
above 1500 mg/kg. Under top charging conditions, when over 1500 mg/kg solids
are present in the feed, a solids removal efficiency of about 50 to 80% and
500-800 mg/kg of solids concentration in the cyclone overflow were obtained
(Figure 6).
The 152.4 mm cyclone tests were conducted only under top charging
conditions. During the tests, the solids concentration in the cyclone feed
liquor was in the range of 100 to 800 mg/kg. The maximum solids removal
efficiency was 50%.
The cyclone size and other related specifications were selected to obtain
solids separation at 0.045 mm in the cyclone overflow. The 76.2 mm cyclone
gave excellent cleaning when treating charging liquor during side charging
conditions. Poor results were obtained with both the 76.2 mm and 152.4 mm
cyclones in the phase separation of liquor during top charging conditions.
This may be a result of smaller size of the solids contained in the charging
liquor during top charging in comparison to side charging (Table II).
However, it was very difficult to obtain a true size distribution of solids in
the liquor because of agglomeration caused by tar and other surface active
substances contained in the charging liquor.
Spiral Rake Classifier
The classifier test data is presented in Table VII. On the average, 55%
of the solids in the feed were removed as classifier sand during side charging
and 43% during top charging. The solids in the classifier overflow contained
1000 mg/kg, a value much higher than can be recycled to the charging main. The
classifier sand had a moisture content of 40 to 50%* The classifier could not
by itself clean the liquor, but might be used as a pretreatment system to
supply a liquor containing uniform solid concentration to other phase
separation devices.
Conventional Flotation Cell
The Heyl and Patterson (H+P) flotation test data obtained during both side
and top charging are presented in Table VIII. The underflow from the
flotation cell had a moderate solids concentration, except for the test
conducted with high tar concentration in the charging liquor and the test when
alum was used as an additive. A minimum solids concentration (1000 mg/kg) in
the feed is needed before the flotation cell will operate with a nigh
efficiency (Figure 7). The conventional H+P flotation cell operation was not
improved with the use of the alum additive and, in fact, alum was detrimental
in floating the solids from the liquor. One possible explanation may be the
breakage of the alum flocculated particles during the recirculation of about a
third of the liquor entering the flotation cell. Poor flotation results could
also be obtained in the case of the high tar content liquor, if the tar laden
solid particles were too coarse to float and did not have the affinity for the
air bubbles required for good flotation response (Figure 7).
417
-------
TABLE VII
SPIRAL RAKE CLASSIFIER DATA
Moisture
in Sand
Solids Concentration (mg/kg^o) (X+ a)
Solids
Removal Charging
Efficiency Condition
Feed
2874 ^ 3249
1334 +1556
Overflow
1470 +. 2060 48.5 +_ 8.6 55.3 side
1004 + 1236 40.5 + 5.8 43.9 top
Feed to the conical tank was the source of the charging liquor used in these
tests (Figure 1).
TABLE VIII
CONVENTIONAL FLOTATION CELL OPERATING DATA
So'lids Concentration
(mg/kg +_o)
Feed
5932
6344
1540
2271
280
389
255
+_ 10098
1 10033(a)
i 295
+. 1226(b)
i 470
+_ 454
i 277
Underlow
111
739
398
1329
70
243
203
± 91
^412
i274
+. 680
i 46
+_ 299
i 202
Solids ,,,
Removal Feed1 ' Charging
Efficiency Source Conditions
Froth
1027
6944
16534
10460
269
829
2485
+_ 11273
1 8310
i 14559
+_ 20999
^280
+_ 837
^4432
84.1
55.9
73.0
49.3
57.5
40.0
46.5
FCT
FCT
FCT
FCT
COF
COF
3CO
side
side
top
top
side
top
top
(1) FCT = Feed to the conical tank
COF = Classifier overflow
3CO = 76.2 mm (3 in) cyclone overflow
a) Excessive tar in the charging liquor feed
b) 200 mg/kg alum used as an additive
418
-------
o.
. NORMAL TAR CONTENT
A HIGH TAR CONTENT
J
0 1000 2000 3000 4000
SOLIDS CONCENTRATION (ing/kg)
5000
Figure 7 Solids Removal Efficiency
Vs Solids Concentration
in the Feed for the
Conventional Flotation
Cell During Side
Charging
1000 2000 3000 4000 5000 6000
SOLIDS CONCENTRATION (mg/kg)
7000
Figure 8 Solids Removal Efficiency Vs
Solids Concentration in the
Feed for the Dissolved Air
Flotation Unit with Alum
Addition
TABLE IX
DISSOLVED AIR FLOTATION OPERATING DATA*
Solids Solids Solids
Concentration Concentration Concentration
in the feed in the Underflow in the Froth
(mg/kg ^o) (mg/kg ^ o ) (g/kg+_o)
1844 i 1178 812 +_ 723
2060 +_ 3042 257 +_ 98
631 +_ 103 1110 +_ 308
1536 +_ 889 1580 ^ 507
1277 +_ 106
126 +_ 89
51 i 45
20 +_ 21
Chemical
Additions
None
178 mg/kg alum
178 mg/kg alum
4 mg/kg polymer
178 mg/kg alum
4 mg/kg polymer
All samples were taken under top charging conditions.
419
-------
Dissolved Air Flotation Unit
Three sets of additions were used in testing the dissolved air flotation
unit (DAF): no additives, 178 mg/kg alum, and 178 mg/kg alum plus 4 mg/kg
anionic polymer. The DAF unit with no additives produced a liquor which
averaged 812 mg/kg solids, as compared to 257 and 1374 mg/kg for the liquor
when alum and alum plus polymer were added to the system, respectively. Table
IX presents the average values for all of the measured parameters. For the
DAF tests with no additions, analysis of the percent removal of solids in the
feed parameters indicated that there was a lower limit of solids concentration
in the feed below which the DAF solids removal efficiency was drastically
reduced. No numerical value could be set for the DAF tests with no alum
additions due to the small number of experimental data points. For the tests
run with alum and alum plus polymer additions, the value of the lower limit
was 700 mg/kg and 1200 mg/kg of solids in the feed, respectively. The solids
removal efficiency above the limiting point was greater than 85% when alum was
added to the system, but only 20% when both alum and polymer were added. The
plot of solids removal efficiency versus solids concentration in the feed for
the DAF with alum addition is presented in Figure 8.
The DAF operated best at 138 kPa, a value somewhat lower than anticipated
from previous work. The reason for this probably is due to the method by
which the pressure was increased in the pressurizing tank. This was
accomplished by restricting the flow into the flotation chamber. At lower
pressures, higher flow rates were permitted. It appears that a high flow rate
must be maintained for the DAF unit to function properly. On a full scale
unit, higher pressures can be used while maintaining a high liquor flow rate.
This increased pressure should enhance the unit's cleaning efficiency and
broaden the range of solids concentration that the DAF can efficiently clean.
The DAF produced the lowest solids loading in the clean liquor when alum
was added to the system. The alum acted as a flocculating agent which aids in
the floating of fine solids. The addition of anionic polymer to the unit
offset the effect of the alum addition. Alum is positively charged and the
solids in the charging liquor are negatively charged. The alum agglomerates
the solids into larger, easier-to-float particles which are still negatively
charged. The anionic polymer probably interacts with the agglomerates and
disperses them, since both are negatively charged. Large amounts of alum
would make the polymer addition useful, but small amounts of additives is the
goal in order to be cost effective. A cationic polymer may be of value in
flocculating solids, since it would perform in a manner similar to alum and
might form even larger agglomerates than alum. Note that alum is beneficial
in the DAF but detrimental in conventional flotation.
SETTLING PITS
This investigation has indicated that the following three solid-liquid
phase separation methods were superior to the others tested:
1. Gravity settling with polymer addition.
2. Dissolved air flotation with alum addition.
3. Horizontal belt filtration with precoat.
420
-------
Of the three methods, gravity settling appears suitable for a liquor with
widely varying concentrations of solids and tar. Dissolved air flotation
functions well for the varying solids loading in the liquor, if the solids
content is above 700 mg/kg, but the effect of tar on the unit's operation,
maintenance, and cleaning efficiency has not been established. The horizontal
belt filter requires a precoat in order to operate efficiently. Both of the
latter two require large amounts of maintenance plus operation by skilled
personnel. Considering the above factors, and based on the satisfactory
operating experience of settling pits of the preheat coke battery at Jones &
Laughlin Steel's Aliquippa Works, two settling pits (7.6 x 7.6 x 7.6 m) were
installed at Inland's "C" Battery to clean the charging liquor (Figure 9).^ '
For the same reason, No. 11 Battery also has two settling pits (9.1 x 9.1 x
7.5 m) to process its charging liquor prior to recycling to the charging main.
At any given point in time, only one pit is in operation, and settled coal
fines sludge is removed from the other pit by clamshell operation.
CHARGING LIQUOR HEADERS
CHARGING MAIN
POLYMER
ADDITIONS
EXCESS TO
STORAGE
NORTH
SETTLING
PIT
COAL FINES
SOUTH
SETTLING
PIT
BACKWASH
NORTH
CLEAR
WELL
SOUTH
CLEAR
WELL
COAL FINES
Figure 9 Schematic Diagram of the Settling Pit
Solid-Liquid Phase Separation System (4)
The settling pit operation of the "C" Battery cannot be realistically
evaluated at the present time because of the possible presence of some
extraneous water flows. Consequently, only the settling pit data for No. 11
Coke Battery will be discussed in some detail.
Based on monthly average data, the clean overflow charging liquor from No.
11 Battery settling pit operation contains as low as 300 mg/kg solids
concentration. However, the solids loading in the treated liquor, which is
recycled into the charging main, generally fluctuated from 300 - 1000 mg/kg
with an average loading of about 600 mg/kg solids. This fluctuation in the
421
-------
solids concentration of the treated liquor might be due to the specific coal
blend composition, coal grind, and the type, as well as the dosage, of
polymers added to the charging liquor in order to accelerate the settling rate
of coal fines in the pit. The retention time of the coal fines in the
settling pit was calculated to vary between 20-50 minutes, depending on the
level of settled fines in the pit. The polymers used are: Betz Cationic 3390
at 7-10 mg/kg, and Betz Anionic 3325L at 2 mg/kg. It was necessary to use
polymers in order to achieve solid-liquid phase separation in the pit. The
operating experience indicated that the higher the dosage of the polymer, the
lower was the solids loading in the treated recyclable charging liquor.
The coal fines sludge material recovered from the settling pits by a
clamshell type operation has a very high moisture content (50-55%), which
causes severe handling problems .during transport and recycl/i^ig^\ This sludge
material can be recycled as a boilergfuel in a power station^ ' ' or by mixing
with coal blends for cokemaking. ' In order to solve or minimize the
handling problem of the settling pit sludge materials, various dewatering
methods, such as selective agglomeration of charging liquor coal fines,
compacting of the "as is" pit sludge, and the briquetting of pit sludge mixed
with-other coal fines are currently being investigated.
SUMMARY AND CONCLUSIONS
Various solid-liquid phase separation devices were tested in order to
produce a clean recyclable charging liquor at the preheated and pipeline
charged "C" Coke' Battery. These included: laboratory jar settling tests and
leaf filtration tests; pilot scale horizontal belt filter, hydrocyclone,
spiral rake classifier, conventional and dissolved air flotation cell tests;
and plant scale holding tank and aeration tests. On the basis of the results
of this investigation, the following conclusions can be made:
1. The solids concentration of the charging liquor fluctuates during the
charging cycle; the tar concentration of the liquor also varies.
2. A minimum solids concentration of about 1000 mg/kg is necessary for
an effective solid-liquid separation by the various devices tested,
which indicates a need for a preconcentration step.
3. Gravity settling with polymer addition, dissolved air flotation with
alum addition, and a horizontal belt filter with precoat gave the
best results in achieving high solids removal efficiency (85-95%) and
in producing a clean recyclable liquor containing a solids
concentration of about 200 mg/kg.
4. In view of the variable solids and tar concentration in the charging
liquor, the gravity settling method was selected for solid-liquid
phase separation, and settling pits were installed at Inland's "C"
and No. 11 Batteries.
422
-------
5. The processing of charging liquor in the No. 11 Battery settling pits
indicated that the settling pit overflow, i.e., the clean recyclable
charging liquor, contained a solids concentration as low as 300
mg/kg.
References
1. McMorris, C. E., "Inland's Preheat-Pipel ine Charged Coke Oven
Battery," Ironmaking Proceedings, ISS/AIME, Vol. 34, Toronto, 1975,
pp. 330-338.
2. Sorenson, M. E., "Inland's Greenfield Site - No. 11 Battery,"
Ironmaking Proceedings, ISS/AIME, Vol. 37, Chicago, 1978, pp. 24-36.
3.
Davis, R. F., "The Coaltek System - Pipeline Charging Coke Ovens,"
Ironmaking Proceedings, ISS/AIME, Vol. 35, St. Louis, 1976, pp.
479-486.
4. Holowaty, M. 0., and Robins, N. A., "Coal Preheating and Pipeline
Charging at Indiana Harbor," Iron and Steel International, Vol. 51,
1978, pp. 285-290.
5. Graham, J. P., and Pater, V. J., "Recent Work on Preheating at BCRA,"
Agglomeration 77, AIME, Edited by K. S. V. Sastry, Vol. 2, pp.
1057-1070.
6. "An Investigation of Various Methods of Recovering Solids from
Solids-Bearing Effluents Produced by the Preparation and Charging of
Preheated Coal to the 6.5 m Oven at BCRA," Carbonization Research
Report 62, The British Carbonization Research Association, December,
1968, pp. 1-19.
7. Graham, J. P., and Pater V. J., "The Influence of Different Charging
Techniques in the Carbonization of Preheated Coal," Ironmaking
Proceedings, ISS/AIME, Vol. 38, 1979, pp. 448-449.
8. "Further Investigation into the Treatment of Carryover Slurries from
Coal Preheating and Pipeline Charging," Carbonization Research Report
92, The British Carbonization Reserach Association, December, 1980,
pp. 1-13.
9. Anon., "Treatment of Effluents from Coal Preheating and Predrying
Plant," Effluent and Water Treatment Journal, 1979 (10), pp. 535-536.
10. Aktay, A. I., Bodnaruk, B. J., and Holowaty, M. 0., "Apparatus and
Method for Separating a Mixture of Liquid and Coal Fines," U. S.
Patent 4,065,385, December 27, 1977.
423
-------
ASSESSMENT OF THE BIOLOGICAL TREATMENT OF COKE-PLANT
WASTEWATERS WITH ADDITION OF POWDERED ACTIVATED CARBON (PAC)
by: Leon W. Wilson, Jr.
Senior Research Engineer-Research Laboratory
United States Steel Corporation
Monroeville, Pennsylvania 15146
by: Bernard A. Bucchianeri
Manager, Technical Service-Clairton Works
United States Steel Corporation
Clairton, Pennsylvania
by: Kenneth D. Tracy, Principal
Environmental Dynamics Incorporated
Greensboro, North Carolina 27407
ABSTRACT
The United States Steel Corporation, in cooperation with the United
States Environmental Protection Agency, has conducted an extensive experi-
mental biological-treatment program to develop data relative to Best
Available Technology Economically Achievable (BATEA) for coke-plant waste-
waters. This program has included testing at both a bench-scale level and a
pilot-scale level. One of the features of this program was an assessment of
the benefits of addition of powdered activated carbon (PAC) to the biologi-
cal reactor. In bench-scale reactors gross removal efficiencies, as
measured by chemical oxygen demand (COD) and total organic carbon (TOG),
were determined with and without PAC. Comparisons were made on the basis of
the effects of solids-retention time and on PAC-dose levels. A dual-train
pilot-scale study was made at optimum conditions established during the
bench-scale research to confirm the bench-scale comparisons. The results
indicated that PAC did not significantly enhance effluent quality when
compared with the effluent quality of a biological reactor operated at opti-
mum conditions. There was also evidence to indicate that PAC-supplemented
biological reactors treating coke-plant wastewater may undergo desorption of
a substrate toxic to the thiocyanate-degrading organisms.
INTRODUCTION
Clairton Works of the United States Steel Corporation is one of the
world's largest producers of metallurgical coke. In addition, Clairton
Works also has a fully integrated system for the recovery and refining of
coal chemicals from the coke-oven off-gases. Naturally occurring ammonia is
recovered as an anhydrous liquid product through use of U. S. Steel's
424
-------
proprietary PHOSAM* process. Light-oil fractions are separated from the
coke-oven gas by a cryogenic-regenerator system, which also yields an ultra-
pure, hydrogen-rich gas for consumption in a synthetic-ammonia plant.
Complementary systems are operated to recover benzene, toluene, and xylene
as well as a complete line of tar-based fractions.
Clairton Works generates somewhat in excess of 9,500 cubic meters per
day (m3/d) [2.5 million gallons per day (MM gpd)] of contaminated water.
About 45 percent (%) of this water is generated directly from the coking
operation whereas the remaining 55% is attributed to the chemical opera-
tions. The typical composition of this contaminated water is shown in
Table 1.
The principal elements of the Clairton Works contaminated water treat-
ment facilities are shown schematically in Figure 1. Following gravity
Free Ammonia
Still _
, PHOSAM Abiorber
Figure 1. CONTAMINATED-WATER TREATMENT PLANT, CLAIRTON WORKS - UNITED STATES STEEL CORPORATION
separation of both solids and suspended oils in the settling tanks, the
contaminated water is processed through the USS CYAM* process also developed
by U. S. Steel. .Here the water is steam-stripped of so-called "free
ammonia", pH-adjusted by the addition of lime to liberate "fixed ammonia",
and further steam-stripped to yield a biological plant feed stream of
desired ammonia content. In addition to ammonia removal, the CYAM system
also removes more than 99% of the free cyanide as well as other acid
* PHOSAM and CYAM are trademarks.
425
-------
gases. Following ammonia removal, the water is cooled and clarified in a
conventional center-well, peripheral overflow clarifier prior to biological
treatment.
The biological-treatment system is a single-stage process consisting of
two independent aeration basins operating in parallel. The total system
volume is in excess of 25,000 m3 (6.5 million gallons) with aeration and
mixing being provided by low-speed mechanical surface aerators. After
biological treatment and clarification, the wastewater passes through dual-
media filters before it is discharged into the Monongahela River. The CYAM
system and the biological treatment plant have been in operation since
1976.
During the third quarter of 1979, as part of the Mon Valley Consent
Decree, the United States Environmental Protection Agency (USEPA) and United
States Steel Corporation (USSC) agreed to conduct an extensive experimental
program to develop data relative to the application of Best Available
Technology Economically Achievable (BATEA) for the Clairton coke-plant
wastewaters. Environmental Dynamics Incorporated (EDI) of Greenville, South
Carolina was the consultant selected to carry but this program in conjunc-
tion with USSC Research and Clairton personnel in cooperation with the
USEPA. The program involved extensive testing with bench-scale reactors
[0.028 m3 (7.5 gallons)] and pilot-scale reactors [3.5 m3 (935 gallons)].
One of the goals of the program was to evaluate the impact of the addition
of powdered activated carbon (PAC) to the aeration chamber of a biological
system treating coke-plant wastewater. This presentation describes the
results of that investigation.
SELECTION OF POWDERED ACTIVATED CARBON
On the basis of previous experience with the use of PAC in petroleum-
industry wastewater treatment, EDI selected four commercially available PACs
for isotherm testing in order to determine the best PAC for use in the
bench- and pilot-scale experiments. The four PACs chosen for isotherm
testing on effluent from the Clairton activated-sludge system to determine
the absorptivity of the biologically refractory substances in the wastewater
were Westvaco Nuchar SA and Nuchar SA-15, and ICI Darco HDC and Darco KB.
Because there were no specific, readily measurable substances which could be
used as parameters in these isotherm tests, the gross parameters of chemical
oxygen demand (COD) and total organic carbon (TOC) were used. In addition,
because the wastewater exhibited a high ultraviolet absorbance at 400 nano-
meters (nm), a normalized absorbance was defined as the ratio of the equi-
librium absorbance to the initial absorbance at 400 nm to provide a third
isotherm parameter.
Freundlich isotherms1* for these three measured parameters are
presented in Figures 2, 3, and 4. The COD isotherm (Figure 2) indicated no
large difference among the four PACs but showed Nuchar SA-15 marginally
better than the other three PACs. Also, the fact that the isotherms were
* See References.
426
-------
gCOD/gPAC
1.00
0.40
0.30
0.20
0.10
0.05
0.03
0.02
0.01
_ NUCHAR SA-15
NUCHARSA
Darco HOC
Darco KB
10 20 30 50 100 200 300 500 1000
EQUILIBRIUM COD, mg/l
Figure 2. COD ISOTHERMS
gTOC/gPAC
1.00
0.50
0.30
0.20
0.10
0.05
0.03
0.02
0.01
NUCHAR
SA-
HAR II /
15 . I / / Darco KB
Darco HDC
NUCHAR SA
10 20 30 50 100 200 300 500 1000
EQUILIBRIUM TOC, mg/l
Figure 3. TOC ISOTHERMS
427
-------
10.0
9.0
8.0
7.0
6.0
5.0
ABSORBANCE REMOVED 4.0
PER MASS OF PAC
[1-A/A0/M] x 103 3Q
2.0
1.0
Darco KB
NUCHAR SA
Darco HDC
Q = Absorbance in Control
A = Absorbance After Contact with PAC
M = Weight of PAC, mg
0.1
0.2
0.3 0.4 0.5 0.6 0.8 1.0
EQUILIBRIUM ABSORBANCE (A/AQ)
Figure 4. ISOTHERMS BASED ON ULTRAVIOLET
ABSORBANCE AT 400 NANOMETERS
nearly vertical was an indication that PAC addition might not significantly
enhance the quality of the Clairton wastewater as measured by COD.
The TOC iostherms (Figure 3) showed a somewhat wider variation among
the PACs, with the Darco HDC showing the greatest effect on TOC equilibrium
with changes in PAC concentration; however, additional TOC reduction, even
at the high PAC dosage, was small.
All four PACs gave substantial reductions in the absorbance at 400 nm
(Figure 4), indicating a high degree of removal of the substance or sub-
stances responsible for this absorbance. No attempt was made to determine
these substances or to determine their level of concentration in the
wastewater. Therefore, the carbons may have been removing a substance
present in only minor concentrations that would not affect biological oxida-
tions. Nuchar SA-15 and Darco HDC were marginally better than Nuchar SA and
Darco KD.
On the basis of the three gross parameters tested in these isotherms,
Nuchar SA-15 was selected for use in the remainder of the program; however,
it was agreed by all concerned with the project that there was very little
detectable difference among any of the candidate carbons.
EFFECTS OF PAC ADDITION TO BENCH-SCALE REACTORS
Previous investigations ' have shown that the effectiveness and
economic viability of PAC addition for upgrading activated- sludge
428
-------
performance is a function of both carbon-addition rate and solid-retention
time (SRT). To explore both effects, carbon doses were varied to provide
aeration-basin carbon (Nuchar SA-15) concentrations of 1500, 3000, and 6000
milligram per litre (mg/1) at 20- and 60-day SRTs. Control reactors at 20-
and 60-day SRTs without carbon were also maintained. All reactors were fed
from the same 0.208 m^ (55 gallon) drum, which was continuously fed a
slipstream from the influent line to the Clairton aeration basin. The water
volume was maintained at a constant 0.189 m^ (50 gallon) by a float and
valve mechanism controlling the slipstream flow. The feed drum was also
constantly agitated to obtain concentration equilization.
Analytical data for the reactors are presented in Tables 2 and 3.
These data represent the average values obtained during the last week of a
six-week period of operation. Examination of the data at both a 20-day SRT
and a 60-day SRT indicated that better performance was obtained at the
1500 mg/1 PAC level than at the higher PAC levels. The only exception to
this was sludge- volume index (SVI), which showed better settling character-
istics with increasing PAC levels, as would be anticipated. The poor
performance of the higher level PAC reactors at both SRTs in terms of COD,
TOC, and five-day biochemical oxygen demand (BODg) was attributed to the
loss of thiocyanate metabolism in these reactors. It was believed that a
desorption from the carbon of some substance toxic or inhibitory to the
thiocyanate-oxidizing organism had occurred. Such desorption phenomena have
been observed in other studies of the application of PAC to industrial
wasterwaters. ' The loss of thiocyanate metabolism did not occur in either
of the non-PAC control reactors, which was taken as an indication that there
was no inhibitory level of a toxic substance in the feed at the time the PAC
reactors lost thiocyanate metabolism. In PAC units, it is theorized that
substances present in the feed at sub-inhibitory concentrations can accumu-
late on the carbon. If desorption then occurs, the material can be eluted
at much higher than normal concentrations, resulting in inhibition of the
organisms.
A comparison of performance of the non-PAC reactors with the best PAC
reactors (1500 mg/1 carbon dose) is presented in Table 4. It was evident
that the 6.0-day SRT non-PAC reactor performed better than the 20-day SRT
non-PAC reactor. It was also evident that there was not much improvement in
the 1500 mg/1 PAC reactor at a 60-day SRT over the 20-day SRT. Overall, it
was concluded from the data that there was not sufficient improvement in
performance at a 60-day SRT to justify the use of PAC.
The ammonia-removal performance as measured by the specific ammonia-
removal rate is presented in Table 5. The data indicate that there was a
definite benefit from carbon addition when the reactors were maintained at a
20-day SRT. It should also be noted, however, that there was no apparent
additional benefit in the use of higher PAC concentrations at the 20-day
SRT. A possible explanation for this was that only a small amount of carbon
was required to absorb a low level toxic material inhibitory to nitrifica-
tion. Thus, no benefit was obtained from the remainder of the carbon that
was added. The data at a 60-day SRT indicate that carbon addition was of no
benefit with respect to ammonia removal as measured by the specific ammonia-
removal rate. Also, the 60-day SRT non-PAC control reactor exhibited better
429
-------
ammonia removal than the 20-day SRT non-PAC control reactor, which would
indicate that a 60-day SRT was closer to optimal than a 20-day SRT.
The data presented in Table 6 show the results of the specific organic
analysis of the influent and effluents of the reactors by gas chromato-
graphic/mass spectrographic (GC/MS) procedures. Twenty-four hour composites
were taken at the conclusion of the bench-scale test for these analyses.
The results indicated an excellent reduction of most of the priority pollut-
ants through biological treatment. The results also indicated that carbon
addition was of no significant benefit in terms of providing additional
reductions in the priority pollutants.
Static bioassays, with PIMEPHALES PROMELAS (fat-head minnows) as test
organisms, were performed on the influent and effluents of each reactor at
the conclusion of the experimental run. Effluents were collected over a
period of 48 hours. The results, presented in Table 7, indicated that PAC
was beneficial when the reactors were controlled at a 20-day SRT, with the
LC 50 (lethal concentration at which 50 percent of the test organisms die
during a 96 hour test) getting higher with increasing levels of PAC to the
point where an LC 50 was not reached at 6000 mg/1 PAC concentration.
However, when the SRT was maintained at 60 days, PAC was of no benefit. The
LC 50 results at a PAC level of 3000 and 6000 mg/1 at an SRT of 60 days were
probably due to the suspected desorption from the carbon .that had occurred
in these reactors.
EFFECTS OF PAC ADDITION TO A PILOT-SCALE REACTOR
A 20-week pilot study was undertaken at the conclusion of the bench-
scale experiments, which consisted of a 9-week acclimation period followed
by an 11-week period of intensive testing. Two pilot units were utilized,
one serving as a control while the other contained PAC. Each unit consisted
of a 3.5 m (935 gallon) aeration chamber separated from an integral
clarifier by a sliding baffle. Fine-bubble aeration through dacron-covered
diffusers provided mixing and oxygen transfer. Optimum operating condi-
tions, established in the bench-scale studies, were maintained in both
units. Each unit was fed influent from a common feed tank, which was
supplied by a side stream off the main influent feed line to the Clairton
aeration basin. Based on the bench-scale results, an SRT of 60 days was
maintained in both reactors and a PAC concentration of 1500 mg/1 was
maintained in one of the reactors.
The average performance results covering the 11 weeks of intensive
testing are presented in Table 8. The results paralleled the" bench-scale
results in that carbon addition did not measurably improve effluent
quality. The only possible exception would be with ammonia-removal
performance, where the PAC reactor exhibited a better specific ammonia-
removal rate and a lower ammonia-N average effluent concentration, as shown
in Table 9. However, as shown in Figure 5, the PAC reactor, near the end of
the testing period, lost thiocyanate metabolism, so that the ammonia-N load
on the nitrifiers was reduced with the overall result that effluent
ammonia-N concentration was lower. The loss of thiocyanate metabolism in
the pilot PAC reactor gave additional support to the PAC desorption theory
expressed in the discussion of the bench-scale results. One additional
430
-------
200
150
EFFLUENT
THIOCYANATE 100
CONCENTRATION,
mg/l
60
PAC Reactor
/\
Non-PAC Reactor
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
OCTOBER, 1980
Figures. EFFLUENT THIOCYANATE ANALYSIS
result during the desorption incident that was not recognized or noticed in
the bench-scale reactors was the tendency for the biomass in the carbon
reactor to float and foam. The foam was exceptionally unusual because it
would not break up when anti-foam was added. The noncarbon reactor also was
susceptible to foaming but it was easily controlled with antifoam. No
explanation for this observation has been offered.
During the intensive testing period, four separate composite samples
were collected and analyzed for specific organic compounds (priority pollut-
ants) by GC/MS procedures. The results of these analyses are presented in
Tables 10, 11, 12, and 13, and showed that effluent from the PAC reactor was
no better than effluent from the control reactor. The sample collected
October 30, 1980 was particularly noteworthy, however, because it indicated
that the PAC reactor effluent had several polynuclear aromatic hydrocarbons
in higher concentrations than the influent. This sample was collected after
the PAC reactor had lost thiocyanate metabolism from a suspected carbon
desorption, which has lead to the speculation that these compounds were the
substances desorbed and were the cause of the loss of thiocyanate metabo-
lism.
Two 96-hour flow-through bioassays were performed on the reactor efflu-
ents, and the results are presented in Table 14. The first bioassay was
made at the mid-point of the testing period, and the second bioassay was
made at the end of the testing period. Although an 1C 50 was not
encountered in the first bioassay, the PAC reactor exhibited less effect on
the fat-head minnows than the control. In the second bioassay, the PAC
effluent had an LC 50 of 49 percent while the control effluent did not have
431
-------
an LC 50; in fact, only 15 percent of the test minnows were affected by the
full-strength water. It was assumed that the poor bioassay test with the
PAC effluent was a direct result of the suspected desorption incidents that
had previously occurred.
CONCLUSIONS
The effluent from a properly operated and controlled biological reactor
treating coke-plant wastewater is not significantly improved by the addition
of PAC to the aeration basin of the reactor; in fact, such addition may be
detrimental to effluent quality because of the possibility of desorption of
toxic substances that affect thiocyanate metabolism. Conversely, the
effluent quality of a sub-optimally operated biological reactor treating
coke-plant wastewater might be improved by the addition of PAC to the aera-
tion basins, as long as desorption does not occur.
REFERENCES
1. Weber, W. J., Jr., Physiochemical Processes for Water Quality Control,
pp. 206-211, Wiley-Interscience, New York (1972).
2. Stenstrom, M. K., and Grieves, C. G., "Enhancement of Oil Refinery
Activated 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).
3. Thibault, G. T., Steelman, B. L., and Tracy, K. D., "Enhancement of the
Refinery Activated Sludge Process with Powdered Activated Carbon," Paper
presented at the 6th Annual Industrial Pollution Conference, WWEMA, St.
Louis, MO (April 1978).
4. Crame, L. W., "Pilot Studies on Enhancement of the Refinery Activated
Sludge Process," API Publication 95 (October 1977).
5. Thibault, G. T., Tracy, K. D., and Wilkinson, J. B., "PACT Performance
Evaluated," Hydrocarbon Processing, 78, 143 (May 1977).
It is understood that the material in this paper is intended for general
information only and should not be used in relation to any specific
application without independent examination and verification of its
applicability and suitability by professionally qualified personnel. Those
making use thereof or relying thereon assume all risk and liability arising
from such use or reliance.
432
-------
Table 1. COMPOSITION OF COKE-PLANT CONTAMINATED WATER
Ammonia
Phenol
Thiocyanate
Cyanide (Total)
Oil/Grease (Freon Extractibles)
Total Suspended Solids
PH
Temperature
— 1500-2000 PPM
-- 800-1200 PPM
— 600- 700 PPM
— 200- 400 PPM
— 2000-4000 PPM
— 300-1500 PPM
8-9
-- 130-170°F.
433
-------
Table 2. COMPARISON OF PERFORMANCE AT A SOLIDS-RETENTION TIME OF 20 DAYS
Parameter3
PAC Concentrations 0
COD
TOG
BOD5
Oil and Grease
Phenol
Thiocyanate
Ammonia-N
MLSS
MLVSS
MLVSS ' c
308
91
12
2.4
0.11
0.22
122
7852
5827
-
1500
219
43
19
2.1
0.07
0.34
80
7108
6416
4991
3000
712
95
86
2.5
0.07
320
2
8751
7696
4846
6000 Influentb
521
78
51
2.9
0.06
200
37
11698
11182
5482
2760
817
1089
18
562
370
56
_
-
-
SVI
42
43
28
19
a All values are mg/1 unless otherwise stated.
b After 26% dilution with river water.
c Assumes a 95% volatility for PAC.
434
-------
Table 3. COMPARISON OF PERFORMANCE AT A SOLIDS-RETENTION TIME OF 60 DAYS
Parameter3
PAC Concentrations
COD
TOC
BOD5
Oil and Grease
Phenol
Thiocyanate
Ammonia-N
MLSS
MLVSS
MLVSS ' c
0
182
53
6
2.5
0.08
0.13
91
13350
11350
—
1500
170
50
14
4.7
0.05
0.34
122
12624
11172
9747
3000
208
51
19
4.2
0.07
2.7
86
15327
14130
11280
6000
726
126
90
4.1
0.12
400
3
17932
16298
10598
Influent0
2760
817
1089
18
562
370
56
-
SVI
66
69
58
30
a All values are mg/1.
b After 26% dilution with river water.
c Assumes a 95% volatility for PAC.
435
-------
Table 4. COMPARISON OF PERFORMANCE BETWEEN
BEST PAC REACTORS AND NON-PAC REACTORS
Parameter3
PAC Concentrations
COD
TOC
BOD5
Oil and Grease
Phenol
Thiocyanate
Ammonia-N
SVI
SRT=20
0
308
91
12
2.4
0.11
0.22
122
42
Days
1500
219
43
19
2.1
0.07
0.34
80
43
SRT=60
0
182
53
6
2.5
0.08
0.13
91
66
Days
1500
170
50
14
4.7
0.05
0.34
122
69
Influent15
—
2760
817
1089
18
562
370
56
__
a All values are mg/1.
b After 26% dilution with river water.
436
-------
Table 5. COMPARISON OF AMMONIA-REMOVAL PERFORMANCE
PAC Concentrations
Total Ammonia-Nc
Remaining
Influent Flow, I/day
MLVSS
,d
Specific Ammonia-
Removal Rate x 103
fmg NH -N \
mg VSS day/
SRT = 20 Days
Influent'
145
_0 1500 3000 6000
122
80
79
85
15.26 15.98 14.54 14.69
5827 4991 4846 5484
2.13 7.34 6.99
5.66
SRT = 60 Days
Total Ammonia-Nc
Remaining
Influent Flow, I/day
MLVSS|d
Specific Ammonia-
Removal Rate x 10
/ mg NH -N \
145
91 122
87 100
18.14 14.69 22.32 18.72
11350 9747 11280 10598
3.05 1.23 4.05 2.81
mg VSS day
a All values are mg/1 unless otherwise stated.
b After 26% dilution with river water.
c Includes ammonia from conversion of thiocyanate.
d Assumes a 95% volatility for PAC.
437
-------
Table 6. GC/MS ANALYSES OF INFLUENT AND REACTOR EFFLUENTS FOR PRIORITY POLLUTANTS
CO
oo
(micrograms per litre)
SRT = 20 Days
PAC Concentration mg/1
Volatiles
Acrylonitrile
Methylene chloride
Benzene
Toluene
Ethylbenzene
Trichloroethylene
Base-neutrals
Diethyl phthalate
Butyl benzyl phthalate
Di-2-ethylhexyl phthalate
Di-n-octyl phthalate
Anthracene/phenanthrene
Fluoranthene
Pyrene
Chrysene
Benzo [b] f luoranthene
Benzo [ a ] pyrene
Di-n-butyl phthalate
Benzo [k] f luoranthene
Naphthalene
Indeno [1,2, 3-cd] pyrene
Benzo [ghi] perylene
Dimethyl phthalate
Acenaphthene
Acids
Phenol
2 , 4-Dimethylphenol
Influent
16,000
9
80
36
11
~
17
18
520
17
96
76
56
13
5.6
3.2
-
-
-
-
-
. -
—
~110,000
-
0
_*
30
38
3.4
-
7.4
-
1.6
6.2
0,8
0.4
0.4
1.2
1.4
1.2
2
7.6
-
-
24
8.8
-
—
200
2.6
1500
-
20
22
2.4
-
~
220
2.2
16
6.4
1.8
0.2
0.4
0.4
-
-
14
-
-
-
-
23
—
19
2.6
3000
-
8.2
98
-
-
6.2
260
-
11
0.8
3.0
1.8
1.8
5
9.8
7
40
-
-
-
-
42
0.8
-
-
6000
-
12
52
2
-
4.6
130
2.2
5.8
5.4
1.6
1
1
2.8
6.6
5.4
42
-
4
-
-
20
1
-
-
0
-
13
30
7.4
-
™
170
-
•—
-
2.4
0.2
-
-
-
-
6.8
-
8
—
-
26
~
1.8
-
SRT =
1500
1,200
13
72
6
-
"~
0.8
2.4
110
2
0.4
0.2
0.6
0.4
1
0.8
22
0.2
-
-
-
-
—
6.6
-
60 Days
3000
-
3.2
36
2.8
-
^
260
3.6
38
2.2
2.2
0.6
0.4
0.4
1.4
1.8
19
-
-
—
-
28
•~
-
-
6000
-
15
82
10
1.4
™
190
1.4
12
1.8
2.8
0.6
0.4
0.4
-
-
17
-
-
—
-
30
"•
-
62
2-Chlorophenol
~5.4
* Dash indicates compound is below detection limits.
-------
Table 7. RESULTS OF STATIC BIOASSAY TEST
PAC Concentration Influent SRT f 20 Days SRT = 60 Days
(mg/1) — _0 1500 3000 6000 _0_ 1500 3000 6000
LC 50 1.5 27.0 29.5 63.0 '-* -* -* 57.0 45.0
(Volume Percent)
* LC 50 not encountered when less than 65 percent of the test organisms died
when subjected to undiluted effluent.
439
-------
Table 8. AVERAGE PERFORMANCE RESULTS OF THE PILOT-SCALE REACTORS
Parameter (mg/1) Influent
Non-PAC
Effluent
PAC
Effluent
COD
BOD5
TOG
Oil and Grease
TSS
Phenol
Total CN
Amenable CN
Thiocyanate
Ammonia-N
2680
307
403
1579
616
32.4
57.5
595
3.4
0.40
333
89
22
40
3.2
139
0.05
2.3
0.04
3
82
45
48
2.7
175
0.06
2.1
0.04
18
38
Table 9. AMMONIA-REMOVAL PERFORMANCE OF THE PILOT-SCALE REACTORS
Non-PAC PAC
Influent Reactor Reactor
Total Ammonia-Na/ mg/1
Remaining
169
83
42
Influent Flow, I/day
MLSS, mg/1
MLVSS'b
1714 1714
10428 11640
8941 8793
Specific Ammonia
Removal Rate x 103
Cmg NH -N \
mg MLVSS-day /
4.7
7.0
a Includes ammonia from conversion of thiocyanate.
b Assumes a 96 percent volatility for PAC.
440
-------
Table 10. PRIORITY POLLUTANTS OBSERVED IN PILOT-SCALE REACTORS
SAMPLE DATE: 9/16/80
(CONCENTRATION, pg/1)
Volatiles
Acrylonitrile
Benzene
Methylene chloride
Trichloroethylene
Trichlorofluoromethane
Toluene
Base-neutrals
Anthracene
Benzo[a]anthracene
Benzo[a]pyrene
3,4-Benzofluoranthene
Benzo[g,h,i]perylene
Benzo[k]fluorathene
Di-2-ethylhexyl phthalate
Butyl benzyl phthalate
Chrysene
Dibenzo[a,h]anthracene
3,3-Dichlorobenzidine
Diethyl phthalate
Di-n-butyl phthalate
Di-n-octyl phthalate
Fluoranthene
Indeno[1,2,3-cd]pyrene
Napthalene
N-nitrosodiphenylamine
Phenanthrene
Pyrene
Acid Extractables
2-Chlorophenol
2,4-Dichlorophenol
2,4-Dimethylphenol
P-Chloro-m-cresol
Phenol
2,4,6-Trichlorophenol
Influent
5500
0
1
0
0
1
40
200
30
30
0
20
40
20
0
0
80
0
0
1
30
0
0
0
0
30
0
1
900
0
4500
0
Non-PAC
Effluent
210
0
1
0
0
1
1
70
30
30
1
60
20
1
0
1
0
1
1
0
30
1
1
20
0
30
1
1
1
1
1
1
PAC
Effluent
180
0
1
0
0
1
1
20
1
1
1
20
40
1
0
1
20
1
1
1
1
1
1
0
0
1
1
1
1
1
1
1
NOTE:
A "0" in the table signifies that the compound was not detected in
the sample. A "1" in the table signifies that the level detected
was less than the screening level. Priority pollutants not detected
on any of the sampling dates have been omitted from the table.
441
-------
Table 11. PRIORITY POLLUTANTS OBSERVED IN PILOT-SCALE REACTORS
SAMPLE DATE: 10/1/80
(CONCENTRATION, yg/1)
Influent
5900
0
0
0
0
1
1
1
0
0
0
0
0
0
1
0
0
0
0
0
1
0
0
0
1
1
0
0
0
90
3840
0
Non-PAC
Effluent
120
0
1
0
0
0
0
1
1
1
0
1
1
0
1
0
1
0
0
0
1
0
0
1
0
1
0
0
0
0
0
0
PAC
Effluent
170
0
0
0
0
1
0
1
1
1
0
1
1
0
1
0
1
0
0
0
1
0
0
1
0
1
0
0
0
0
0
0
NOTE:
Volatiles
Acrylonitrile
Benzene
Methylene chloride
Trichloroethylene
Trichlorofluoromethane
Toluene
Base-neutrals
Anthracene
Benzo[a]anthracene
Benzo[a]pyrene
3,4-Benzofluoranthene
Benzo[g,h,i]perylene
Benzo[k]fluorathene
Di-2-ethylhexyl phthalate
Butyl benzyl phthalate
Chrysene
Dibenzo[a,h]anthracene
3,3-Dichlorobenzidine
Diethyl phthalate
Di-n-butyl phthalate
Di-n-octyl phthalate
Fluoranthene
Indeno[1,2,3-cd]pyrene
Napthalene
N-nitrosodiphenylamine
Phenanthrene
Pyrene
Acid Extractables
2-Chlorophenol
2,4-Dichlorophenol
2,4-Dimethylphenol
P-Chloro-m-cresol
Phenol
2,4,6-Trichlorophenol
A "0" in the table signifies that the compound was not detected in
the sample. A "1" in the table signifies that the level detected
was less than the screening level. Priority pollutants not detected
on any of the sampling dates have been omitted from the table.
442
-------
Table 12. PRIORITY POLLUTANTS OBSERVED IN PILOT-SCALE REACTORS
SAMPLE DATE: 10/20/80
(CONCENTRATION, \ig/l)
Influent
1690
1
0
0
0
1
10
20
20
20
0
20
0
0
30
0
20
0
0
0
20
0
0
0
10
10
0
1
2750
0
11810
0
Non-PAC
Effluent
0
0
0
0
0
1
0
0
1
1
0
1
0
0
0
0
1
0
0
0
1
0
0
0
0
1
0
0
0
0
1
0
PAC
Effluent
0
0
0
0
0
1
0
1
1
1
0
1
0
0
1
0
1
0
0.
0
- 1
0
0
0
0
1
0
0
1
0
0
0
NOTE:
Volatiles
Acrylonitrile
Benzene
Methylene chloride
Trichloroethylene
Trichlorofluoromethane
Toluene
Base-neutrals
Anthracene
Benzo[a]anthracene
Benzo[a]pyrene
3,4-Benzofluoranthene
Benzo[g,h,i]perylene
Benzo[k]fluorathene
Di-2-ethylhexyl phthalate
Butyl benzyl phthalate
Chrysene
Dibenzo[a,h]anthracene
3,3-Dichlorobenzidine
Diethyl phthalate
Di-n-butyl phthalate
Di-n-octyl phthalate
Fluoranthene
Indeno[1,2,3-cd]pyrene
Napthalene
N-nitrosodiphenylamine
Phenanthrene
Pyrene
Acid Extractables
2-Chlorophenol
2,4-Dichlorophenol
2,4-Dimethylphenol
P-Chloro-m-cresol
Phenol
2,4,6-Trichlorophenol
A "0" in the table signifies that the compound was not detected in
the sample. A "1" in the table signifies that the level detected
was less than the screening level. Priority pollutants not detected
on any of the sampling dates have been omitted from the table.
443
-------
Table 13. PRIORITY POLLUTANTS OBSERVED IN PILOT-SCALE REACTORS
SAMPLE DATE: 10/30/80
(CONCENTRATION, yg/1)
Influent
1180
1
1
10
1
1
0
1
1
1
0
0
0
0
1
0
0
0
0
0
20
1
0
0
0
1
0
1
1220
0
6680
0
Non-PAC
Effluent
250
1
1
1
1
1
0
1
0
1
0
0
0
0
1
0
0
0
1
0
70
0
0
0
0
1
0
0
0
0
1
0
PAC
Effluent
0
1
1
1
1
1
0
10
20
10
1
0
0
1
20
0
0
0
1
0
820
20
0
0
0
10
0
0
1
0
1
0
NOTE:
Volatiles
Acrylonitrile
Benzene
Methylene chloride
Trichloroethylene
Trichlorofluoromethane
Toluene
Base-neutrals
Anthracene
Benzo[a]anthracene
Benzo[a]pyrene
3,4-Benzofluoranthene
Benzo[g,h,i]perylene
Benzo[k]fluorathene
Di-2-ethylhexyl phthalate
Butyl benzyl phthalate
Chrysene
Dibenzo[a,h]anthracene
3,3-Dichlorobenzidine
Diethyl phthalate
Di-n-butyl phthalate
Di-n-octyl phthalate
Fluoranthene
Indeno[1,2,3-cd]pyrene
Napthalene
N-nitrosodiphenylamine
Phenanthrene
Pyrene
Acid Extractables
2-Chlorophenol
2,4-Dichlorophenol
2,4-Dimethylphenol
P-Chloro-m-cresol
Phenol
2,4,6-Trichlorophenol
A "0" in the table signifies that the compound was not detected in
the sample. A "1" in the table signifies that the level detected
was less than the screening level. Priority pollutants not detected
on any of the sampling dates have been omitted from the table.
444
-------
Table 14. FLOW-THROUGH BIOASSAY RESULTS USING PILOT-REACTOR EFFLUENTS
Percent
Affected Organisms
in Undiluted LC 50*
REACTOR DESIGNATION Effluent (Volume Percent)
BIOASSAY I (9/25/80)
Non-PAC 65 Not encountered
PAC 20 Not encountered
BIOASSAY II (10/25/80)
Non-PAC 15 Not encountered
PAC 100 49
* LC 50 not encountered when less than 65 percent of the test organisms died
when subjected to undiluted effluent.
445
-------
BIOLOGICAL TREATMENT OF BY-PRODUCT COKE PLANT
WASTEWATER FOR THE CONTROL OF BAT PARAMETERS
by: G.M. Wong-Chong
Environmental Research & Technology, Inc.
Pittsburgh, PA 15219.
S.C. Caruso
Mellon Institute, Carnegie-Mellon Univ.
Pittsburgh, PA 15213.
ABSTRACT
This paper presents the findings of an evaluation of the activated
sludge process for the treatment of coke plant wastewaters (CPW) to meet
compliance with BAT limitations for conventional/non-conventional
parameters. The evaluation entailed the operation of batch and continuous
flow laboratory scale reactors, and a 50 gpd pilot plant.
The study produced understandings of (a) the sequence in which the
different components in CPW are removed, (b) the order of the degradation
reaction associated with each component, (c) the potential inhibitory
effects of different CPW components on the different degradation reactions,
and (d) kinetic expressions associated with treatment rates and sludge
production which could be used to design the biological treatment system.
INTRODUCTION
The treatment of wastewaters from coke and coal carbonization plants is
not new. Some of the earliest treatment studies of waste liquors from these
plants were conducted during the early 1900s. In a 1907 report, Frankland
and Silvester(l) concluded, "considerable but not insuperable difficulties
(exists) in the purification of (gas liquor wastes) by bacterial means; on
the other hand, the oxidation of these wastes by chemical means can only be
effected at a prohibitive cost." Today 74 years later, in spite of all of
the work conducted, many of these difficulties persist and they are the
causes of concern regarding compliance with the impending "best available
technology" (BAT) discharge limitations. These limitations will require
controls on several specific pollutants (conventional and priority); this
report addresses only the conventional pollutants — ammonia, cyanide,
phenol, oil & grease, and suspended solids.
Despite the long history of investigations into the biological treatment
of coke plant wastewaters (CPW), treatment facilities have not been designed
446
-------
or operated to produce effluent quality which approaches that required by
the proposed BAT limitations. Without this prior experience and/or a
comprehensive knowledge of the biological reactions involved in this mode of
treatment, the design and operation of treatment systems to consistently
produce compliance with BAT requirements will be difficult.
In order to gain a better understanding of the biological treatment of
coke plant wastewaters this study was initiated. This understanding would
provide the basis for the design and operation of treatment facilities which
may be capable of producing compliance with impending BAT limitations. To
achieve this objective the following information was sought:
• The order of the different reactions,
• The sequence in which the-reactions occur and which
reaction(s) is the limiting step,
• The effects of different wastewater components on different
reactions, and
• The relationship between operating parameters such as mixed
liquor solids, hydraulic and sludge residence time, and
treatment performance.
This study examined the biological treatment of CPW in a single stage
activated sludge treatment process for the control of ammonia, free cyanide,
phenol and thiocyanate.
The study entailed a review of the published literature, laboratory
scale experiments with both batch and continuous flow reactors, and a 50 GPD
pilot plant which was located on-site at a coke plant. This report presents
the findings of this study.
Subsequent sections of this report will present:
• A brief review of the public literature,
• Results of the experimental program, and
• A discussion of the application of the experimental data to
design a BAT biological system for CPW.
LITERATURE REVIEW
Coke plant wastewater (CPW) contain many components which are of concern
from a pollutional standpoint; thus, their discharge will be regulated.
Further, many of these components may exert inhibitory effects on the
different biological reactions in the course of the treatment process.
Thus, in attempting to achieve effective treatment many process modes have
been examined. For example:
• Kostenbader and Flecksteiner(2) found it was necessary to
dilute waste ammonia liquor (WAL) to ammonia concentration
less than 2000 mg/1 to achieve phenol removal,
• Biczysko and SuschkaO) achieved effective phenol removal from
undiluted wastewater which contained ammonia concentrations of
150-2280 mg/1,
447
-------
• Barker and Thompson(4) evaluated the treatment of WAL in a
multi-staged reactor system with limited success,
• Catchpole and Cooper(5) examined "growth factors" as means of
enhancing/accelerating the biological treatment process; their
results were inconclusive, and
• Luthy et al(6) demonstrated the nitrification of coke plant
waste in a single stage reactor; reactor conditions were
40 days SRT and 9 days HRT.
The first two studies were primarily concerned with achieving phenol
removal; the latter three studies were directed at achieving advanced
treatment—control of ammonia, cyanide, phenol, and thiocyanate. For the
composition of the raw wastewater (ammonia about 100 mg/1) used in the
investigation of Luthy et al(6), the 9 day HRT required for achieving
nitrification is economically impractical, and at best this study
illustrates advanced treatment in a single stage reactor is achievable. The
limited success in the reported studies is ample evidence that the status of
the understanding of advanced biological treatment of coke plant wastewaters
is still very primative. Subsequent discussions will attempt to further the
present status.
EXPERIMENTAL BIOTREATMENT OF CPW
MATERIAL AND METHODS
The experimentation entailed the following:
• Treatment of both synthetic and actual CPW,
• Operation of laboratory scale batch and continuous flow
reactors, and
• Operation of a 50 GPD pilot treatment system at a coke plant.
All reactors were operated at temperatures about 20-25OC, pH about 7.0-7.5
and D.O. concentrations -2.0 mg/1.
The synthetic wastewater was formulated with essential components being
ammonia, free cyanide, phenol and thiocyanate. The concentration of these
components were varied as required. The CPW wastewater used was that
produced after ammonia stripping; the composition was ammended as required.
Phosphorus was also added to the CPW as a nutrient. Table 1 presents the
average composition of the wastewaters used in this study.
The sludge used in this study was obtained from an activated sludge
system treating coke plant wastewater. This sludge was allowed to acclimate
to CPW until an effective nitrifying population was established. This
acclimation took about 4 to 6 weeks because the original sludge was not
actively nitrifying. The continuous flow reactor experiments were conducted
for a minimum period of about 10 to 12 weeks under a given set of conditions
before changes to another set were made. While these test periods were
significantly less than three times the sludge retention time (SRT), it was
significantly greater than three times the hydraulic residence time (HRT).
No attempt was made to operate for 3 times SRT because it was felt that the
448
-------
biological populations were sufficiently stable after the initial
development of the microbial population and the acclimation to the test
conditions under consideration.
TABLE 1. AVERAGE COMPOSITION OF THE WASTEWATERS USED IN THE
CONTINUOUS FLOW EXPERIMENTS
Synthetic
Coke Plant
ASW 1
ASW 2
Pilot Plant
(all concentrations mg/1)
Ammonium-N
Free Cyanide, CNp
Phenol, OH
Thiocyanate, SCN~
*ZNin
**pH
205
182
1121
522
429
9.5 & 10.5
93
82
1113
524
264
9.7
249
72
1163
607
434
9.7
100
1 .
500
220
153
9.0
*ZNin = *• NH4+ ~ N)+ °*54 CNF + °*24 SCN"
**The pH of synthetic waste was changed from 9.5 to 10.5 for two sets
of experiments.
BATCH REACTOR EXPERIMENTS
A batch reactor experiment entailed monitoring of a component(s) with
the passage of time. The experiments were prepared by adding an aliquot of
wastewater to an aliquot of acclimated sludge. From these experiments it
was possible to establish the following:
• Reaction order for each component,
• Reaction sequence, and
• Tolerence limits for different components on individual
reactions.
These experiments were facilitated by using synthetic wastewaters. However,
spot tests with CPW were in complete agreement with the observations made
with the synthetic wastewaters.
Types of Reaction and
Reaction Sequence
Figures 1 and 2 show typical depletion profiles for (a) ammonia,
(b) free cyanide, (c) phenol, and (d) thiocyanate. These profiles show the
ammonia, phenol and thiocyanate reactions to be zero order and the cyanide
reaction to be first order. The profiles in Figure 2 illustrate the
449
-------
^ 80
o»
E
60
0>
o
20
(a) Phenol
Zero Order
Reaction
16
O»
e ,
c-12
O
o
Z 8
0)
o
c
o
O 4
O
(b) Cyanide
First Order
Reaction
048
Reaction Time, hrs
4 8 12 24
Reaction Time, hrs
(c) Thiocyanate
Zero Order
Reaction
246
Reaction Time, hrs
8
Figure 1. Depletion profiles for phenol, cyanide and tliiocyanat
by acclimated sludge
450
-------
- 250
01
120-
6 8 10 12
Reaction Time hrs.
14
Figure 2. Reaction sequence for ammonia, cyanide, phenol and thiocyanate in a batch reactor
-------
sequence of the reactions as they occur in a batch biological reactor
treating coke plant wastewater. The sequence shows that ammonium oxidation
is the last reaction to occur. Consequently, for advanced biological
treatment of coke plant wastewater the ammonium oxidation reaction would be
the process limiting reaction and process design criteria based on the
ammonium oxidation reaction would accommodate the degradation of the other
components — cyanide, phenol and thiocyanate.
The ammonium profile in Figure 2 shows a gradual increase during the
early stages of the experiment. This increase corresponds with the ammonia
produced during the degradation of cyanide and thiocyanate.
Reaction Interactions
The profiles in Figure 2 also show that the microorganisms responded to
both phenol and cyanide immediately. This occurrence suggests that the
concentrations of phenol examined were not inhibitory to cyanide degradation
or vice versa. Experiments with initial phenol concentration up to 110 mg/1
and cyanide concentrations of 20 mg/1 shows similar uninhibited responses.
On the other hand, the lag segment of the thiocyanate profile (see
Figure 2), indicate the inhibitory effect of cyanide at concentrations
greater than 3.0 mg/1; the nitrification reaction was inhibited by cyanide
concentrations greater than 0.5 mg/1. The data in Figure 2 does not show
any inhibitory effects for phenol; this observation is due to the rapid rate
at which the phenol was degraded. However, it must be recognized that
phenol can inhibit the nitrification reaction(7).
In other experiments where thiocyanate inhibition of the nitrification
reaction was examined, it was observed that the inhibition was not as
dramatic as that from cyanide. In fact, the nitrifying organisms were
capable of adpating to thiocyanate concentrations as high as 200 mg/1;
Downing et al(7) reported inhibition to nitrification in activated sludge at
<300 mg/1. Consequently, in the operation of a biological treatment system
to achieve effective advanced treatment it will be necessary to accommodate
for the potential inhibitory effects of cyanide, phenol and thiocyanate.
This accommodation can be achieved by developing treatment process design
criteria which considers the nitrification reaction(s) as the process
controlling step.
CONTINUOUS FLOW REACTION EXPERIMENTS
The continuous flow experiments were conducted in completely mixed
reactors with integral clarifiers, Figure 3. These experiments were
designed to establish performance trends associated with variations in feed
flow rate, feed composition and mixed liquor solids concentration. A
reactor's performance was measured by material balances on oxidizable
nitrogen and biological solids.
Oxidizable Nitrogen Oxidation
In the biological treatment of free cyanide and thiocyante, ammonia is
produced; this ammonia production is almost stoichiometric in quantity.
452
-------
FEED
CONTROL
GO
^ELEVATED
FEED RESERVOIR
QUIESCENT ZONE WITH
iVERFLOW TUBE
TREATED
WASTE
Figure 3. Schematic flow diagram of laboratory scale reactors
-------
Thus, in advanced biological treatment of CPW for the control of ammonia,
that produced from cyanide and thiocyanate must also be considered. The
total ammonia or oxidizable nitrogen in the wastewater can be estimated by
IN = (NH4 - N)+ 0.24 SCN + 0.54 CNF ....(1)
The reaction sequence shown in Figure 2 indicates that the nitrifica-
tion reaction is the process controlling step and the reaction is zero order
with respect to oxidizable nitrogen. Thus a material balance for oxidizable
nitrogen around a continuous flow reactor at steady state can be mathemati-
cally expressed as
ZNout = ENin - kAT ---- (2)
where
N^n = feed oxidizable nitrogen concentration,
NQut = reactor oxidizable nitrogen concentration,
k^ = rate of oxidation at a specified mixed
liquor solids concentration,
T = reactor hydraulic residence time.
In all of the experiments conducted, the concentration of phenol, free
cyanide and thiocyanate in the reactor at steady state were less than 0.1,
0.1 and l.O.mg/1 respectively. Thus, the oxidizable nitrogen in the
reactor was essentially the reactor ammonium concentration. Thus
Equation 2 can be rewritten as
. IN. - (NH* - N) ,,,.
k. = in 4 ....(3)
where(NH4 - N) = reactor ammonium-nitrogen concentration.
Several reactor conditions (hydraulic residence time, mixed liquor
solids, raw waste composition and sludge residence times) were examined and
the results of these are presented on Figure 4 where oxidation rate, k^ is
correlated with mixed liquor solids.The regression analysis produced the
following relationship:
kA = 0.015 TVS - 0.0045 ....(4)
The correlation coefficient, r, was 0.96. However, the k^ value of
-0.0045 at zero TVS may be an artifact of the regression analysis, thus
Equation 4 will be truncated to read:
454
-------
200 r-
O Synthetic Coke Wastewater
® Real Coke Wastewater
Pilot Plant
150 _
15.0 TVS
0.96 (correlation
Coefficient)
10.0
12.0
Figure 4.
Mixed Liquor Volatile Solids, TVS, grn/J.
Effect of mixed liquor solids concentration on ammonia oxidation rate
455
-------
kA = 0.015 TVS ---- (5)
where
TVS = mixed liquor volatile solids concentration, mg/1
Biological Sludge Growth
In biological wastewater treatment processes, the microorganisms use a
fraction of the energy produced by the metabolic reactions for cell growth
and reproduction, and the remaining fraction for maintenance of other
cellular functions. Biological sludge growth in the activated sludge
process has been expressed as
= Y U - b ....(6)
6 e
where b = microbial maintenance energy coefficient,
day-1
U = specific substrate utilization rate,
day-1
Ye = observed sludge yield
9 = sludge residence time, days
Figure 5 presents the sludge growth relationships observed in this
study; these are:
•8-1 = 0.019 Ye-- 0.0012 ---- (7)
The value 0.019 day-1 for y compares well with 0.015 day-1 produced by
the correlation of k^ vs TVS (see Figure 4). From equation 7, values for
Ymax and b are estimated at 0.65 and 0.0012 respectively. Consequently
for any given set of operating conditions, the sludge production rate can be
estimated from
Sp = F(ZNin - ZNout)Ye ---- (8)
where Sp = sludge production rate.
DESIGN OF A BAT BIOTREATMENT SYSTEM
FOR COKE PLANT WASTEWATERS
Equations 3, 5, 7, and 8 are the essential relationships for designing a
coke plant wastewater biological treatment system. These equations can be
used to size the aeration basin and to determine the amount of sludge
produced. Other data necessary to complete the design would be (a) oxygen
up-take rates, and (b) sludge settling characteristics; these data are
required for sizing/specifying the aeration device and sludge clarifier.
456
-------
0.010
en
—i
B
O
•H
0)
00
T)
i-l
U)
n)
u
o
M
a,
•H
u
-------
From Equation 5, oxidation rates can be determined for specified mixed
liquor solids; these oxidation rates can then be applied to Equation 3 to
determine aeration basin hydraulic residence times (HRT) for the desired
treatment. In turn, the HRTs can be used to determine the size of the
aeration basin, given the quantity of wastewater to be treated.
From Equation 7 a limiting sludge residence time (SRT) can be
determined; i.e., that at which Ye is equal to Ymax (0.65). Operations
at SRT conditions less then this limiting SRT will result in the loss of the
mixed liquor solids by washout with a concommitant loss in treatment
performance. Equation 7 also predicts that at long SRTs, Ye will be small
which will result in small sludge generation.
CONCLUSIONS
A single stage activated sludge process was examined for the treatment
of CPW for the control of BAT conventional and non-conventional para-
meters — ammonia, free cyanide, phenol and thiocyanate. The study was
conducted by observing the performances of batch and continuous flow
laboratory scale reactors and a 50 gpd pilot plant.
The study produced knowledge of:
• The sequence in which the different components in CPW are
removed,
• The potential inhibitory effects of different CPW components
on the different degradation reaction, and
• Kinetic expressions which define treatment performance and
sludge generation in the process.
In addition, this laboratory scale study demonstrated that if the
treatment system is designed and operated for ammonia oxidation, treatment
of free cyanide, phenol, and thiocyanate are achieved; treated effluent
concentrations for these parameters were observed in this study to be less
than 0.1, 0.1 and 1.0 mg/1 respectively. However, this treatment/design
approach should be tested in a full scale plant operation.
ACKNOWLEDGEMENT
This study was supported by a research grant from the American Iron and
Steel Institute.
REFERENCES
(1) Frankland, P.F. and H.J. Silvester (1907), The Bacterial Purification of
Sewage Containing a Large Proportion of Spent Gas Liquor. J. Soc. Chem.
Ind., 26, 231-237.
(2) Kostenbader, P.D. and J.W. Flecksteiner (1969), Biological Oxidation of
Coke Plant Weak Ammonia Liquor. J. Water Pollution Control Federation,
41, 199-207.
458
-------
(3) Biczysko, J. and J. Suschka (1966), Investigation of Phenolic Waste
Treatment in an Oxidation Ditch. Advances in Water Pollution Research,
Proc. Int. Assn. Water Pollution Research Conf., Munich, 2, 285-308.
(4) Barker, J.E. and R.J. Thompson (1973), Biological Removal of Carbon and
Nitrogen Compounds from Coke Plant Wastes. Environmental Protection
Technology Series. EPA-R2-73-167.
(5) Cooper, R.L. and J.R. Catchpole (1973), The Biological Treatment of
Carbonization Effluents - IV. The Nitrification of Coke-Oven Liquors
and Other Trade Wastes and the Enhancement of Biological Oxidation of
Resistant Organic Compounds by the Addition of Growth Factors to
Activated Sludge. Water Research, 7, 1137-1153.
(6) Luthy, R.G. and L.D. Jones (1981), Biological Oxidation of Coke Plant
Effluent, J. Env. Eng. Div. Am. Soc. of Civil Engineers, 106, 847-851.
(7) Downing, A.L., T.G. Tomlinson and G.A. Truesdale (1964), Effects of
Inhibitors on Nitrification in Activated Sludge Process. J. Inst. Sew.
Purification, 537-554.
459
-------
TWO-STAGE BIOLOGICAL FLUIDIZED BED TREATMENT
OF COKE PLANT WASTEWATER FOR NITROGEN CONTROL
by: S.G. Nutt
Dearborn Environmental Consulting Services
Box 3060, Mississauga, Ontario L5A 3T5
H. Melcer
Wastewater Technology Centre, Environment Canada
Box 5050, Burlington, Ontario L7R 4A6
J.H. Pries
Canviro Consultants Ltd., 178 Louisa Street
Kitchener, Ontario N2H 5M5
This paper was presented at the 54th Annual Water pollution Control
Federation Conference in Detroit, Michigan on October 8, 1981 and has been
submitted to the Journal of the Water Pollution Control Federation for
publication.
INTRODUCTION
Wastewaters from steel mill coking operations contain high concentra-
tions of phenolics, thiocyanates, ammonia, cyanide, sulphides and a variety
of complex hydrocarbons. Of the steel mills in North America which employ
biological systems for oxidation of phenolics, thiocyanates, cyanide and
sulphides, the majority utilize the activated sludge process (1). None of
these systems consistently effect a significant degree of nitrogen control
(2). Researchers (3,4,5) have shown that complete nitrification of coke
plant effluents can be attained under various conditions in either two-stage
or single-stage biological systems.
Bridle et al (6,7) demonstrated that complete nitrogen control was
feasible in a single sludge pre-denitrification-nitrification configuration
provided that wastewater dilution or powdered activated carbon addition (PAC)
was practised in conjunction with strict SRT control. Medwith and Lefelhocz
(8) were able to nitrify coke plant wastewaters in a hybrid suspended growth-
fixed film biological system by the addition of an inert particulate, such as
coke breeze or coal dust to a conventional activated sludge system. It was
found that the inert material facilitated operation at long SRT's (50 to 200
days) and allowed complete nitrification to occur at nominal hydraulic reten-
tion times in the range of 12 to 24 hours for coke plant wastewaters diluted
in the range of 1 to 5 parts wastewater per one river water.
460
-------
The biological fluidized bed process has been demonstrated to be
feasible for carbon removal from municipal (9,10,11) and concentrated
industrial wastes (12), for nitrification (9,13) and for denitrification
(14,15) of a variety of low and high strength wastewaters. The fluidized
bed system, in specific instances, can provide significant advantages
relative to suspended growth systems due to the high biomass concentrations
which can be maintained in the biological reactors. In order to assess
the technical and economic feasibility of treating coke plant wastewaters
in a two-stage fluidized bed system operated in the pre-denitrification-
nitrification mode, a pilot-scale investigation of the process is being
conducted at the Environmental Protection Service's Wastewater Technology
Centre (WTC) in Burlington, Canada.
The primary objective of the initial stages of the investigation has
been to establish the minimum hydraulic retention times required in each
reactor to maintain adequate nitrogen control during treatment of undiluted
coke plant wastewaters. The preliminary results of these on-going investi-
gations are the topic of this paper.
DESCRIPTION OF PILOT-SCALE FLUIDIZED BED SYSTEM
The fluidized bed process is a modification of more conventional fixed
film processes, such as the trickling filter, in which wastewater is passed
upward through a bed of granular support media, typically sand, at a
sufficient velocity to expand or fluidize the media. The granular media
provides a large surface area for the establishment of a biological film.
The pilot-scale system under examination at WTC consists of two fluidized
bed reactors in series, coupled to provide carbon oxidation, nitrification
and denitrification in the pre-denitrification operating mode. In this flow
configuration, the supplemental carbon requirements for denitrification
are minimized as the organic carbon in the raw wastewater provides the
energy source and serves as the electron donor for the denitrification re-
action. Full-scale industrial experience (16) and cost comparisons based
on municipal systems (17) have indicated the potential of the pre-denitri-
fication process for nitrogen control.
The process flowsheet of the coupled, two-stage fluidized bed pilot
plant is .shown in Figure 1. The pilot plant consists of an anoxic denitri-
fication fluidized bed reactor, 150 mm in diameter, and an oxygenic nitri-
fication fluidized bed reactor, 290 mm in diameter. The initial empty bed
reactor volumes were 58.5 litres and 210 litres for the anoxic and oxygenic
beds respectively; however, the bed heights and reactor volumes are
adjustable by relocation of the position of the sand-biomass separation
systems. Pure oxygen is supplied to the nitrification reactor through a
proprietary oxygen transfer device provided by Dorr-Oliver Inc. High
internal recycle rates are necessary for both reactor to maintain the required
fluidization fluxes, to ensure an adequate supply of oxygen to the nitri-
fication process and to return nitrate and nitrite to the denitrification
reactor. The support media in both reactors is quartzite sand with an
effective size of 0.48 mm.
461
-------
-p.
CT>
r>o
SAND-BIOM ASS
SEPARATOR
BIOMASS
ANOXIC
PHOSPHORUS
ALKALINITY
CLARIFIER
COKE PLANT
WASTE WATER
DO/pH
CONTROL
SAND
--
_l
ATION
OR
J
ft"
REC
TAN
0Z
RECYCLE
EFFLUENT
RECYCLE
TANK
OXYGENIC
NITRIFICATION
REACTOR
Figure 1. Two-Stage Biological Fluidizcd Bed Process Flowsheet.
-------
The raw feed to the pilot-scale treatment system consists of limed
ammonia still effluent and light oil interceptor sump wastewater obtained,
on a batch basis, from Dofasco Inc. in Hamilton, Ontario. This is the same
wastewater source used by Bridle et^ a± (6,7) in the evaluation of the
single sludge suspended growth biological treatment process. The feed under-
goes primary clarification prior to entering the denitrification reactor.
Phosphorus is added in the form of phosphoric acid on a continuous basis
to maintain 1 to 2 mg-L~l of soluble phosphorus in the treated effluent. In
addition, there are automatic dissolved oxygen and pH control-systems on the
nitrification reactor. Control of pH is by means of sodium biocarbonate
addition, which also provides supplemental alkalinity for the nitrification
reactions.
EXPERIMENTAL AND ANALYTICAL PROCEDURES
Subsequent to an extended acclimation period to establish viable
populations of micro-organisms in each fluidized bed reactor, the treatment
system was operated at a number of pseudo-steady state loading conditions
to define process performance. Daily samples of treated process effluent
were collected at each pseudo-steady state condition and analyzed for
filterable organic carbon (FOC), ammonia-nitrogen (NH^-N), nitrate (N03~N)
and nitrite (N02-N) nitrogen, total and filterable Kjeldahl nitrogen (TKN),
phosphorus and alkalinity. In addition, treated effluents were analyzed
three times per week for phenol, total and filterable chemical oxygen demand
(COD), total cyanide (TCN), thiocyanate (CNS) and sulphide. Total and
volatile suspended solids analyses were conducted five times per week.
Effluent from the anoxic denitrification reactor was monitored daily
for FOC, N03-N, N02~N and TKN. The quality of the raw feed to the system
was monitored on a weekly basis for FOC, TKN, NH3-N, TCN, CNS, COD, pH and
alkalinity. The concentration of biomass (bed volatile solids, BVS) in each
reactor was measured once per week, based on compositing at least three
individual samples from various positions in the reactors. All analyses were
conducted according to standard procedures (18).
Pseudo-steady state conditions were assumed to have been attained when,
after a step change in the process loading, the final effluent NH-j-N con-
centration had stabilized at a constant level. This typically required
from twenty to forty days, depending on the magnitude of the step change.
Pseudo-steady state performance data were collected for one to three weeks
at each loading condition. These data do not represent true steady state
in terms of the biomass as a constant biomass concentration and equilibrium
SRT were not attained in the reactors.
All pseudo-steady state performance data presented in this paper were
collected from the same batch of coke plant wastewater obtained from Dofasco
in November 1980. The quality of the feed varied slightly during this
operating period due to freezing of the wastewater and some losses of volati-
le compounds, in particular, ammonia. The average quality of the wastewater,
based on the weekly analyses, is summarized in Table 1 and was similar to
that reported by Bridle ^t jil (7). The removal efficiencies reported for the
463
-------
pseudo-steady state experiments are based on the raw feed quality measured
during the specific period of operation under consideration.
TABLE 1. COKE PLANT WASTEWATER CHARACTERISTICS
Median Cone.
Average Cone. per Bridle ^t a.1 (7)
Parameter (mg-L~l)
FOC
Phenol
TKNUF
TKNp
NH3-N
TCN
CNS
CODuF
CODF
824.
560.
214.
207.
84.6
9.0
381.
3395.
3194
680.
300.
—
180.
88.
8.
240.
—
"
RESULTS
START-UP AND ACCLIMATION
The fluidized bed reactors were started up in June 1980 using sand
obtained from a Dorr-Oliver pilot scale fluidized bed which had been opera-
ted for combined carbon removal and nitrification of municipal wastewaters.
The reactors were filled to the 2 m height with seeded sand and hydraulically
expanded by fifty percent for the accliination period. Feed at start-up was
Dofasco coke plant wastewater diluted in the ratio of 19:1 with secondary
effluent from a pilot scale extended aeration plant treating municipal
wastewater.
As shown in Figure 2, the system acclimated rapidly to the coke plant
wastewater. Within twenty days, the biomass concentration in both reactors
had approximately doubled. Significant production of oxidized nitrogen
(nitrite and nitrate) was evident in the oxygenic reactor within five days,
and the nitrification rate gradually increased during the initial twenty
days of operation. During this period, the ratio of dilution water to coke
plant wastewater was gradually decreased from 19:1 to 5:1.
Between the twentieth and fortieth day of the acclimation period, the
degree of dilution was further reduced until, by Day 40, the system was
operating on full strength coke plant wastewater. Some inhibition of
nitrification was evident as the amount of dilution was reduced; however,
the process rapidly recovered and appeared to be completely acclimated to
the coke plant wastewater after forty days of operation.
PSEUDO-STEADY STATE PERFORMANCE
To date, process performance has been assessed under four different
pseudo-steady state operating conditions which are summarized in Table 2.
464
-------
The initial steady state run (Run 1) was conducted at loading conditions
similar to those specified by Bridle et al (7) as necessary to maintain
consistent nitrification in suspended growth systems. The total system
hydraulic retention time (HRT) for Run 1 was approximately 45 hours (1.9
days), where the system HRT is defined as the HRT in the anoxic denitrifi-
cation reactor plus the HRT in the oxygenic nitrification reactor. The
reactor HRT is based on the flow rate of coke plant wastewater and the
empty bed reactor volume. In each subsequent run, the hydraulic and conta-
minant loadings were increased in an effort to define the minimum reactor
HRT requirements.
10
20 30
TIME (days)
Figure 2. Start-up and Acclimation of the Fluidized Bed System.
The results of the pseudo-steady state runs are summarized in Table 3.
It is evident from these data that, at the maximum loading investigated to
date (Run 4), virtually complete oxidation of organic carbon and nitroge-
nous species was maintained by the coupled fluidized bed process. Total
nitrogen removal efficiency during Run 4 was in excess of ninety percent.
The treated effluent qualities for the pseudo-steady state runs are compared,
in terms of FOC and NHg-N concentrations, in Figures 3 and 4 respectively.
465.
-------
TABLE 2. PSEUDO-STEADY STATE OPERATING CONDITIONS
Run
1
1
2
3
4
Anoxic Reactor
HRT
(h)
9.8
4.8
3.2
1.9
SRT
(d)
-
17.5
30.0
5.6
Temp
('O
30-33
26-30
27-31
25-28
PH
7.2
7.3
7.3
7.4
BVS
99.9
0.17
>99.9
0.14
>99.9
TKN
UF
9.0
95.6
13.3
94.6.
9.8
94.0
11.4
92.9
F
7.5
96.3
11.8
95.1
8.8
94.1
10.5
93.3
NH.J-N
1.0
98.8
2.4
97.4
0.4
99.5
0.3
99.5
TO)
4.5
50.0
3.8
57.8
6.2
42.2
7.0
17.7
CNS
1.8
99.6
1.8
99.6
1.6
99.5
2.3
99.4
NOj-N
0.0
0.4
0.0
1.6
NO?-N
2.5
1.0
1.8
1.0
TN
11.5
94.3
14:7
94.0
11.6
92.8
14.0
91. 3
S3
IJi
• t-
* ' ^
iiJ
i;--
cn
* Hydraulic retention time, based on empty bed reactor volume and coke waste feed rate.
1 Median concentration, expressed in mg-L~ , based on 9 consecutive days of pseudo-steady state operation.
2 Median concentration, expressed in mg-L~^, based on 23 consecutive days of pseudo-steady state operation.
3 Median concentration, expressed in mg-L~ , based on 16 consecutive days of pseudo-steady state operation.
4 Median concentration, expressed in mg-lT , based on 21 consecutive days of pseudo-steady state operation.
TN - Total Nitrogen, TKN + NOT"N
UF - Unfiltered
F - Filtered
-------
70-
v^ 60-
>
50-
Q
cr
in
40-
30-
O
O
O 20
2
10-
SYSTEM HRT
• 44.8 h
o 26.5 h
A 15.9 h
—r~
10
—r~
20
—T~
90
T
-1 F 1 1 1 T
12 5 10 20 30 40 50 60 70 80 90 95 98 99
PERCENT OF OBSERVATIONS EQUAL TO OR LESS THAN
STATED VALUE
Figure 3. Effect of HRT on Treated Effluent FOC Content.
Ll 5H
SYSTEM HRT
• 44.8 h
o 26.5 h
A 15.9 h
'TIIIT
30 40 50 60 70
I
80
2 5 10 20 30 40 50 60 70 80 90 95 98 99
PERCENT OF OBSERVATIONS EQUAL TO OR LESS THAN
STATED VALUE
Figure 4. Effect of HRT on Treated Effluent NH3N-Content.
467
-------
The treated effluent FOC concentration has consistently been in the
range of 35 to 70 mg-L~l throughout the experimental program, representing
a removal efficiency approaching 95 percent. The stable performance of the
system in terms of carbon oxidation was unaffected by step changes in the
process loading.
The maximum loading, in terms of the process nitrification capacity,
has not been reached in the four pseudo-steady state runs. Effluent ammonia-
N concentrations of less than 2 mg-L~l were maintained despite the three-
fold reduction in system HRT.
In all cases, inhibition of Nitrobacter was evident based on the incom-
plete conversion of ammonia to nitrate. This inhibition was also noted by
Bridle et^ al (7) and several other researchers (4,5). The apparent dete-
rioration in total cyanide removal efficiency between Run 1 and Run 4 may
be due to changes in the relative quantities of free and combined cyanide
in the raw wastewater as a result of long-term storage.
The effluent has consistently contained in excess of 50 mg-L suspended
solids and has averaged approximately 100 mg-L"^-. Mass balances and micro-
biological tests have not conclusively defined the source of these suspended
solids; however, based on the nature of the biofilm in the two reactors, it
would appear that they represent biomass from the nitrification system. The
biofilms in the nitrification reactor have been less dense and considerably
more dispersed than those in the denitrification reactor. Excessive losses
of biomass from the fluidized bed biofilms have not been identified as a
problem in any other industrial or municipal fluidized bed studies (19,20).
Therefore, it is possible that the effluent quality will improve once true
steady state has been attained in terms,of the biofilm characteristics.
The conclusion of Bridle et al (7) with respect to suspended growth
biological treatment of Dofasco coke plant wastewater was that complete
nitrification and denitrification of full strength waste could only be
achieved if low levels of powdered activated carbon (PAC) were added to the
system. Furthermore, stable nitrification could not be maintained, even
with-PAC addition, at an aerobic HRT of 26 hours (1.1 days). A retention
time of up to 48 hours in the nitrification reactor was necessary to ensure
stable operation. Based on the pseudo-steady state results collected to
date, the coupled fluidized bed process has been shown to provide a similar
degree of treatment without PAC addition at a nitrification reactor HRT of
sixteen hours or less. The ability to effect nitrification at increased
volumetric loadings without PAC addition appears to be related to the high
reactor biomass concentrations and the high SRT's which can be maintained
in the nitrification reactor.
BIOFILM CHARACTERISTICS
As noted by other researchers who have studied fluidized bed denitri-
fication systems (9), biomass concentrations approaching 40 g-L~l could be
readily attained in the denitrification reactor. Typically, the denitri-
fication biofilms, during this study, were in the range of 100 to 200 y
in thickness. The establishment of active biofilms on the sand media allowed
468
-------
reduction in the fluidization flux from approximately 0.9 m-min-1 at start-
up to an operating level of 0.5 m«min~l.
As the biomass concentration in the denitrification reactor increased,
white deposits were noted within the biofilms. At the same time, the
density of the biofilms increased significantly, requiring an increase in
the hydraulic flux to the reactor to maintain adequate fluidization. Ana-
lyses of the biomass showed that calcium and phosphorus were accumulating
within the biofilms. The progressive increase in BVS, calcium and phosphorus
in the upper regions of the danitrification reactor is shown in Figure 5.
At Day 0, the upper region of the denitrification reactor was poorly fluid-
ized. An accumulation of white deposits was removed from this area of the
reactor. Analyses of the biological material indicated high concentrations
of calcium and phosphorus. Clean sand was added to the system to replace
the media removed and the concentration of calcium and phosphorus in the
upper region of the reactor monitored on a regular basis. After about fifty
days of operation, the deposition of inorganic matter had again resulted in
poor fluidization of the denitrification reactor. The calcium and phosphorus
concentrations had reached the high levels noted prior to the addition of
clean sand.
Based on analysis of the biofilms, the mass ratio of calcium to phospho-
rus was approximately 2.8, indicative of tetrabasic calcium phosphate
(4CaO-P205). X-Ray diffraction confirmed the presence of tetrabasic calcium
phosphate in the biomass. The precipitation of tetrabasic calcium phosphate
was also noted by Bridle et al (7) during operation of suspended growth
biological treatment units.
The accumulation of inorganics in the denitrification reactor did not
appear to affect the biological treatment efficiency; however, the poor
fluidization characteristics prevented adequate control of bed expansion
and reactor SRT. Therefore, the biomass (BVS) concentration in the deni-
trification reactor has been maintained at approximately 25 g-L-1 to avoid
excessive accumulation of inorganic matter.
The deposition of inorganic matter has also occurred in the nitrifica-
tion reactor but at a lower level and has not adversely affected reactor
operation; however, biomass concentrations equivalent to those maintained
in the d'enitr if ication reactor have not been established in the nitrifi-
cation system. Previous investigations (19) of fluidized bed nitrification
of municipal secondary effluent showed that establishing high concentrations
of biomass was a limiting factor in attaining high volumetric conversion
rates. A similar problem was identified during this investigation.
Figure 6 shows the concentration of biomass in the nitrification reactor
since start-up in June 1980. After the acclimation period and subsequent
to restabilization of the process following a severe pH shock, the biomass
concentration in the nitrification reactor reached an equilibrium level of
approximately 4.5 g-L~l. The biofilms were extremely thin, in the range of
50 to 60 y. Hydraulic fluxes in the range of 0.9 to 1.0 m-min"! were
necessary to maintain fluidization as the biofilms were not sufficiently
thick to affect significantly the overall particle density. These high
469
-------
hydraulic loadings created excessive turbulence in the reactor, preventing
the establishment of high biomass concentrations.
Figure 5.
70-
60-
O)
50-
40-
UJ
O
o
O
30-
20-
10-
-10
30
40
50
0 10 20
TIME (days)
Accumulation of Biomass, Calcium and Phosphorus in the
Dehitrification Reactor.
The original reactor design included a conical entrance to the column
and up-flow entry of the feed stream. Based on discussions with Dorr-Oliver,
the inlet piping was redesigned such that the feed entered vertically down-
ward into the apex of the cone. As evident from Figure b, there was a
slight increase in the concentration of biological matter in the reactor;
however, the biomass concentration appeared to stabilize at approximately
5.5 g-L-1. Therefore, a perforated distribution plate was designed for
the reactor and the inlet piping restored to the original up-flow configura-
tion. The plate distributor was installed on Day 233 of operation. In
retrospect, according to Figure 6, the down-flow distribution system may
have produced the same result as the perforated plate. The nitrification
reactor BVS concentration increased to approximately 13 g'L"* over the next
hundred days of operation. At this time, the plate was removed since
plugging had become evident. The hydraulic flux on the nitrification reactor
was gradually decreased to an operating level of 0.5 m-min"! and removal
470
-------
of the distribution plate at the reduced turbulence level did not adversely
affect the system.
14-
6)
10-
CC 4-
2-
V
O.
i
CO
CC
CO
CO
8
Q
UJ
Ul
CC
Ul
" • ' •"
•••
i
V
40
80
120
Figure 6.
160 200 240 280 320 360 400
DAYS OF OPERATION
Biomass Concentration in the Nitrification Reactor.
On the 296th day of operation, as shown in Figure 6, the main recycle
pump failed, causing the beds to settle. Approximately six hours of
maintenance was required to remedy the problem, during which there was no
flow through the reactors. At start-up, the turbulence created during
• fluidizat ion sheared some biomass from the nitrification reactor; however,
the losses were minimal and had little effect on the process.
SUPPLEMENTAL CHEMICAL REQUIREMENTS
The pre-denitrification flow configuration reduces the requirement for
supplemental carbon addition by utilizing the raw waste components as the
carbon source for the denitr if ication reactions. Bridle et^ al (6) esta-
blished that complete denitrification of Dofasco coke plant wastewater
required a feed FOC/TN ratio in excess of 3.5. The raw wastewater used
during the pseudo-steady state investigations fulfilled this requirement
and no supplemental carbon addition was necessary.
471
-------
18-
16-
14-
12-
10-
O
P
6-
4-
2-
3.05(ANOT-N)+1.34
00
Figure 7.
0123456
ANOT-N (mg-L~1)
Relationship Between FOC and NOT-N Removal in the Denitrification
Reactor.
The data presented in Figure 7 show the relationship between organic
carbon removed in the denitrif ication reactor and the removal of oxidized
nitrogen species (nitrate and nitrite) . The denitrif ication reactor was
responsible for the removal of approximately 3 mg FOC per mg NOT-N removed.
The pre-denitrif ication operating mode significantly reduces the oxygen
requirements of the oxygen ic nitrification reactor by the removal, under
anoxic conditions, of a considerable fraction of the feed organic carbon
content.
The phosphorus demand of the wastewater ranged from approximately 50 mg
P per litre of coke plant effluent treated to 170 mg P per litre treated.
As discussed, a fraction of the phosphorus demand was related to the
precipitation of calcium phosphate. In the biological treatment of coke
472
-------
plant wastewater from a fixed ammonia still operated using sodium hydroxide
pH adjustment, Medwith and Lefelhocz (8) supplemented the feed with 40 to
70 mg P per litre of wastewater treated.
CONCLUSIONS
Based on the preliminary results of the on-going investigation of two-
stage treatment of coke plant wastewater using the fluidized bed process,
the following conclusions can be drawn:
• The fluidized bed process is capable of achieving complete
nitrification and denitrification of undiluted coke plant
wastewater without the addition of powdered activated carbon
to the system. Total nitrogen removal efficiencies in excess
of 90 percent could be maintained in the fluidized bed process
at a total system HRT of 16 hours. In a suspended growth system
treating a similar wastewater, a system HRT of up to 60 hours
and the addition of low levels of PAC were necessary to attain
the same degree of treatment.
• The pre-denitrification flow configuration significantly
reduces the oxygen requirements of the nitrification reactor.
Mass balances around the denitrification reactor indicated
the anoxic removal of approximately 3 parts of organic carbon
per part of oxidized nitrogen removed.
• Biomass conce'atrations up to 40 g-L~ were achieved in the
denitrification reactor; however, operational concentrations
are limited to approximately 25 g-L~l due to the accumulation
of inorganic deposits of tetrabasic calcium phosphate in the
biofilms.
• Flow distribution appears to be a critical factor in achieving
high biomass concentrations in the nitrification stage of the
coupled fluidized bed process. Once improvements were made
to the oxygenic reactor design, biomass concentrations up
to approximately 15 mg-L~l were achievable. Excessive turbulence
does not appear to be as significant a factor in the establish-
ment of biofilms in the denitrification system.
ACKNOWLEDGEMENTS
The authors wish to express their appreciation to Dofasco Inc. for
their cooperation throughout this study.
REFERENCES
1. Finn, D.R. and J.D. Stockham, "Survey of Biological Treatment in the
Iron and Steel Industry", EPA-600-2-79-010, (1979).
473
-------
2. Gauthier, J.J., D.D. Jones, L.W. Wilson and C.R. Majors, "Combined
Biological Treatment of Coke-Plant Wastewater and Blast-Furnace
Recycle-Water System Slowdown", presented at the 36th Purdue Industrial
Waste Conference, Lafayette, Ind., (1981).
3. Cooper, R.L. and J.R. Catchpole, "The Biological Treatment of Carboniza-
tion Effluents- IV", Water Research, ]_, 1137, (1973).
4. Ganczarczyk, J.J., "Second-Stage Activated Sludge Treatment of Coke
Plant Effluents", Water Research, 13, 337, (1979).
5. Barker, J.E. and R.J. Thompson, "Biological Oxidation of Coke Plant
Waste", presented at the Chicago Regional Technical Meeting of AISI,
(1971).
6. Bridle, T.R., W.K. Bedford and B.E. Jank, "Biological Nitrogen Control
of Coke Plant Wastewaters", Prog. Wat. Tech., 12> 667, (1980).
7. Bridle, T.R., W.K. Bedford and B.E. Jank, "Biological Treatment of Coke
Plant Wastewaters for Control of Nitrogen and Trace Organics", presented
at the 53rd Annual WPCF Conference, Las Vegas, Nevada, (1980).
8. Medwith, B.W. and J.F. Lefelhocz, "Single-Stage Biological Treatment
of Coke Plant Wastewaters with a Hybrid Suspended Growth — Fixed Film
Reactor", presented at the 36th Purdue Industrial Waste Conference,
Lafayette, Ind., (1981).
9. Jeris, J.S-, R.W. Owens, R. Hickey and F. Flood, "Biological Fluidized
Bed Treatment for BOB and Nitrogen Removal", J_. Wat. Pollut. Control
Fed. ^9 (5), 816, (1977).
10. Sutton, P.M., W.K. Shieh, C.P. Woodcock and R.U. Norton, "Oxitron
System Fluidized Bed Uastewater Treatment Process: Development and
Demonstration Studies :> presented at the Joint Annual Conference of the
APCA and PCAO, Toronto., Canada, (1979).
11. Nutt, S.G., J.P. Stephenson and J.H. Pries, "Aerobic Fluidized Bed
Treatment of Municipal Wastewater for Organic Carbon Removal", presented
at the 52nd Annual WPCF Conference, Houston, Texas, (1979).
12. Sutton, P.M., D. Langley and K. Warner, "Oxitron Fluidized Bed Waste-
water Treatment System: Application to High Strength Industrial Waste-
waters", presented at the 34th Industrial Waste Conference, Purdue
University, (1979).
13. Nutt, S.G., J.P. Stephenson and J.H. Pries, "Nitrification Kinetics in
the Biological Fluidized Bed Process", presented at the 53rd Annual
WPCF Conference, Las Vejjas, Nevada, (1980).
14. Stephenson, J.P. and K.L. Murphy, "Kinetics of Biological Fluidized Bed
Wastewater Denitrification", Prog. Wat. Tech., 12, 151, (1980).
474
-------
15. Jeris, J.S. and R.W. Owens, "Pilot-Scale, High-Rate Biological Denitri-
fication", J_. Wat. Pollut. Control Fed., 47 (8), 20A3, (1975).
16. Bridle, T.R., B.C. Climenhage and A. Stelzig, "Operation of a Full-
Scale Nitrification - Denitrification Industrial Waste Treatment Plant",
j;. Wat. Pollut. Control Fed., 51^ (1), 127, (1979).
17. Wilson, R.W., K.L.-Murphy, P.M. Button and S.L. Lackey, "Design and
Cost Comparison of Biological Nitrogen Removal Systems", presented at
the 51st Annual WPCF Conference, Anaheim, California, (1978).
18. Analytical Methods Manual (1976), Wastewater Technology Centre,
Burlington, Ontario (unpublished),
19. Dearborn Environmental Consulting Services, "Pilot-Scale Assessment of
the Biological Fluidized Bed Process for Municipal Wastewater Treatment",
Project Report prepared for Canada Mortgage and Housing Corporation,
Ottawa, Canada, (November 1980).
20. Cooper, P.F. and B. Atkinson, Biological Fluidized Bed Treatment of
Water and Wastewater, Ellis Horwood Limited, Chichester, England, (1981).
475
-------
TRACE METAL REMOVAL
FROM STEEL PLANT WASTEWATERS
USING LIME AND FERRATE
by: J. A. FitzPatrick, J. Wang and K. Davis*
Department of Civil Engineering
Northwestern University
Evanston, Illinois 60201
ABSTRACT
Trace metals (zinc, cadmium and lead) and suspended solids (clays and
biological floes) were removed from simulated and actual Steel Plant waste-
waters (scrubber water blowdown from EOF and blast furnance and coke plant
lime still and secondary biological effluent) using treatment with 1) lime
followed by potassium ferrate (K FeO ), 2) ferrate only and 3) ferric
chloride only. Results show that ferrate can achieve very high cadmium
removal following lime treatment. Ferrate treatment alone on simulated
scrubbing waters can remove Cd and Pb better or poorer than ferric depending
principally on solution pH. Biological solids effectively scavenge metals
such as lead, zinc and cadmium in the treatment of coke plant biological
wastewaters. If settling of biological solids is inefficient, ferrate can
coagulate the suspended solids and associated trace metals, although high
doses are required.
INTRODUCTION
USEPA pretreatment guidelines for industrial waste discharge as well as
wastewater treatment authority ordinances have pushed the burden of pretreat-
ment of industrial wastewaters particularly those containing toxic heavy
metals to treatment at their source. Economic and regulatory motives then
exist for pretreatment of metal bearing wastewaters of a wide variety to
minimize carriage into publicly owned treatment works (POTW). Often, metals
are removable in POTW's but contaminate waste sludges, thereby limiting their
utility as an agricultural resource.
Pretreatment techniques such as precipitation, oxidation and adsorption
have been proposed for a variety of steel industry wastewaters. In this
work, focus is on a novel form of iron (ferrate) as the treatment reagent to
remove trace metals and suspended solids from steel industry wastewaters,
particularly those from blast furnace gas cleaning and coke production. This
^Present address: Engineering Department, Northwestern Industries,
Chicago, IL
476
-------
work attempts to refine and extend earlier effort (1, 2) where ferrate was
reported to have good oxidation/coagulant properties.
The work herein describes some behavior of ferrate in terms of suspended
solids and trace metal removal on both well defined solutions and actual
wastewaters. Combinations of lime pretreatment and ferrate postreatment as
well as ferrate and ferric treatments alone are used. The experimental
system employed utilized bench scale jar tests to examine feasibility of
ferrate treatment as well as obtain possible mechanistic insights into trace
metal removal processes. Equilibrium chemistry model calculations are com-
pared to experiment in an attempt to generalize results.
THEORY
Since this was an exploratory research project, jar tests were employed
as the major experimental tool. Such tests simulate unit operations of rapid
mixing, flocculation, gravity settling and granular filtration if supernatant
is filtered. Major removal mechanisms include adsorption/coprecipitation,
flocculation, settling and possibly filtration and the sequence of occurrence
are schematized on Fig. 1. Whatever the removal mechanisms are, a full scale
process design cannot be directly scaled from far test results, although such
results are widely used as a major element in design.
In the coagulation process, K FeO reacts in water to produce probably
amorphous ferric hydroxide floe particles (Fe(OH)_). A simplified expression
of this reaction is:
Fe+6 +3e~ »- Fe+3 +3 OH~ >• Fe(OH>3 ^
(purple color (yellow or (white
in solution) colorless solution) precipitate)
As shown, the reaction actually has two steps: The first is an oxidation-
reduction reaction in which the Fe from the FeO, ion oxidizes water and/or
other reduced species present; the iron in turn is reduced to Fe . The ,
second reaction is a hydrolysis (ligand exchange) reaction in which the Fe
coordinates effectively with three OH ions to form Fe(OH) solid phase
particles (and other, soluble hydrolysis products).
The slow stirring in the second phase of the jar test promotes inter-
particle collisions. Some fraction of these collisions result in lasting
contact and particle (floe) growth. Eventually, particles large enough to
settle out in the 30 minute settling period of the test are formed.
Leckie et al (3) examined the removal of trace metals by adsorption/
coprecipitation on aquo ferric precipitates. Ferric oxy hydroxide is the
common form of iron generated in either coprecipitation where iron is added
to a trace metal solution or in precipitate adsorption/scavenging where it is
preformed, then added to the trace metal solution. If the trace metal exists
at very low concentrations (insufficient to generate sufficient extensive
colloidal precipitates) the general removal extent for cations by adsorption
on ferric oxy hydroxide is a sigmoidal function of pH. The position of the
477
-------
Full
1) scale
processes
Rapid
Mix
Flocculator
Filtration
2)
Jar
test
analogs
Millipore
filtration
oo
3) Mechanisms
of trace
metal
removal
Adsorption
co-precipitation,
or precipitation
Velocity
gradient
flocculation,
delayed
precipitation
Type I-discrete
or type II-
flocculent
settling
In depth
granular
filtration
Figure 1. Schematic of Sequential Treatment Processes and Removal Mechanisms
-------
sigmoidal adsorption edge (see Fig. 2) is a function of the particular trace
metal concerned, its initial concentration, and to a certain extent the
concentration of competing ligands. It was concluded (3) that coprecipita-
tion or adsorption steps are very fast compared to typical process times and
thus not rate limiting.
1
4J
c
0)
u
1-1
0)
P-I
Increasing
Iron
Solid Phase
PH
FIGURE 2. Trace metal adsorption behavior on iron hydrous oxide.
However, when the trace metal exists, at higher concentrations and can
form a colloidal precipitate, i.e. as a carbonate, hydroxide or mixed hydroxy
carbonate, the removal is likely to be governed by hererocoagulation between
the forming colloidal Fe(OH). and the trace metal colloid, both of whose
surface charges may vary as different functions of pH. Thus, as pH is
lowered, rapidly forming Fe(OH) coagulates trace metal colloids at pH values
where both surface charges diminish, approaching their zero point of charge
(ZPC). However, at high pH, the removal is hindered by the fact that the net
charge on both the iron and trace metal colloids grow more negative leading
to mutual repulsion. Thus only weakly or partially destabilized colloids
aggregate slowly and not extensively enough to permit rapid sedimentation
and separation without filtration.
If lime is used in pretreatment, the trace metal is removed as a carbon-
ate or hydroxide precipitate in the sweep CaCO floe formed. Residual trace
metal not removed in these steps could be removed by decreasing pH and using
post ferrate addition. The post ferrate treatment should be qualitatively
the same as without lime pretreatment if the solution chemistry is similar.
A complete theoretical framework for interpreting results is unavailable
for either model or even actual wastewaters; in modal experiments, initial
concentrations of trace metals is high enough that colloidal rather than
dissolved metals occur and thus coagulation/destabilization rate processes
may govern the removal but are only poorly understood. In actual waste-
waters, the existence of added complexing ligands, surfactants, emulsions,
etc., are not simulated by corresponding theory and agreement between model
and actual wastewaters may only be qualitative at best. Further interpreta-
tion of results presented here is developed elsewhere (4).
479
-------
EXPERIMENTAL
Experiments consisted of a series of model studies with jar tests
followed by treatment of actual steel plant wastewaters. Lime alone, lime
plus ferrate, ferrate alone and ferric chloride alone were employed in sepa-
rate tests to remove initial concentrations of Cr(VI), Pb(II), Cd(II) and
Zn(II) that ranged from 0.2 to 5 mg/1. Solution chemistry was varied in the
model as well as actual wastewater treatment schemes. The major independent
variables were pH and coagulant dose. Alkalinity and sulfate were the major
anionic species controlled in the model tests. Test conditions are divided
into four phases as shown on Table 1 where the order of the phase is not
necessarily chronological.
Sequence of testing included finding the optimum lime dose on waters
that might chemically mimic some of those found in an integrated steel mill
(Phase IA). Then post lime treatment with ferrate would be employed to
"polish" the effluent and remove trace metals to a very low level. This
procedure was repeated on actual steel plant wastewaters in Phase IB. Phase
II provided data on ferrate only treatment of Lake Michigan water spiked with
several trace metals (Cr, Pb and Cd). Phase III compared ferric and ferrate
behavior under independent variation of solution pH, Fe dose and initial
concentration of alkalinity, sulfate and thiocyanide in an attempt to bracket
some background chemical conditions found in steel plant wastewaters. Phase
IV examined utility of ferrate to remove Cd and Zn from actual steel mill
wastewaters. In Phase IB waters were not spiked with any trace metals.
Jar test procedures used for all study phases employed essentially
similar conditions: Rapid mixing at 100 to 120 RPM for 1 minute followed by
slow mixing at 30 to 50 RPM for 20 minutes, followed by quiescent settling
for 30 to 60 minutes. At the conclusion of the test, samples were siphoned
off and some millipore filtered (0.22p,m or 0.45|j,m) and analyzed for residual
trace metal by atomic absorbtion spectrometry or Y~ray spectroscopy when
radioisotopes of trace metals were employed. The schematic diagram of the
jar test experiments is shown in Fig. 3. Exact experimental procedures
employed and individual test equipment is described elsewhere (4, 5, 6).
RESULTS
The presentation of results encompasses the four phases outlined on
table 1. Chronological sequence of phases was: II, IA, IB, III, IV. In
addition, other efforts to better determine the colloid chemical behavior
of ferrate in water, coagulation of kaolinite, bentonite and polystyreee
latex were essentially run in parallel with phase II.
Phase IA results are presented in greater detail elsewhere (5) but
nevertheless show that for independent variation of initial concentrations
of cadmium, sulfate and alkalinity, optimum cadmium removal from solution
occurred at anywhere from 90 to 1500 mg/1 of lime, (as Ca(OH).) as shown
on table 2. These optimum treated waters had the same as initial sulfate
but drastically reduced alkalinity and roughly 60% removal of cadmium.
Large batches of each initial water was lime treated at its optimum, then
adjusted to pH 7, 8, or 9 and post treated with ferrate at fixed dose and
fixed pH. Representative results are in figure 4A and 4B for two conditions:
480
-------
TABLE 1. EXPERIMENTAL CONDITIONS
Test
IA
IB
II
III
IV
Chemical and Dose Range
Lime
Pretreat
X
X
X
X
Ferrate
alone
0.5 - 10
5 - 40
10 -100
post treat
1.5 - 20*
1.5 - 12
Ferric
alone
5-40*
Water
model
X
X
X
X
natural
X
wastewater
X
X
X
Trace Metal
Cd Pb Zn Cr
X
X
XX X
X
XX X
XX X
XX X
X X
-p.
oo
* denotes dose range of Fe compound (rag/1 as Fe).
-------
£
NaOIl Ferrate or
or HNOa "\ / / Lime
) v A • )
stock / su er_ /
solution ^. )ar te/ts SUP"> /
^V natan* ^ ,
« ^^, \
r\> , _ __ . r5* \
stii^laTrd or l "«in«te rapid mixing
20 minutes slow mixing
actual wastewaters 30 minutes quiescent
settling 1
•^
^
\
5\
s >
0 )
rl /
0 /
n /
(
f
filtra-
tion
0.45^ni
^
filtra-
tion
0.45(jgn
• >
atomic
absorp-
tion
y-ray
detectoi
— > effluent Lo sink
° >effluent to stnk
^ naf.lonlc
•xchnnger
waste tank
Figure 3. Schematic Diagram of Jar Test Study
-------
100
80
60
£ 40
47 " 960 ng/1 is SO*
>3 J -18 ng/1 as CO*
I+JJ, • 0.3ng/l as Cd
10
20
30
'.0
50
60
Potassium Ferrate Dose, mg/1 as K_FeO,
FIGURE 4-A. Ferrate Post Treatunt of High Ionic Strength Water
100
Legend*:
O pK - 7
A 8
V 9
: Filtered Data
: Unfittered Data
Solution Conditions:
l?°4 J ' «8 mg/1 as SO~*
[C0j ) - 47 mg/1 as COj2
fCdfIl -0.2 ng/1 as Cd
20
30
SO
60
Potassium Ferrate Dose, mg/1 as K.FeO,
FIGURE 4-B. Ferrate Post Treatment of Low Ionic Strength Water
483
-------
-2 -2
one with high ionic strength (due to SO ) and one with low [SO. ].
These
show significantly different behavior with and without post filtration.
With post filtration, higher removal efficiency is observed with the high
ionic strength case at all doses and pHs8 whereas low ionic strength waters
showed considerable improvement in removal efficiency only at low dose and
pHs8. Monotonic increase in removal is observed for all but the case with
high (5 mg/1) initial cadmium but low sulfate and alkalinity where optima
of 5 mg/1 K2FeO, and 30 mg/1 I^FeO, were observed at pH=8 and 7, respective-
ly. These complex patterns were displayed for other solution conditions and
are discussed later.
TABLE 2. Lime pretreatment solution conditions (Phase IA)
series
1
2
3
4
5
[so4-2]
48
1920
1920
960
48
CCT]*
305 (25)
1220 (44)
1220
610 (18)
305 (47)
.. +2-.
[Ca ]
(10)
(ID
(11)
(10)
Ccd^2 ]
5 (2)
5 (~2)
0.5 (0.32)
5 (0.32)
0.5 (0.2)
Lime mg/1
90
150
1500
740
200
* Bracketed items are concentrations after lime pretreatment.
** Lime dose is "optimum" based on marginal utility of less than
0.17.7. removal increase/mg/1 dose increase
*** All concentrations are mg/1 as the ion, and Lime is as Ca(OH) .
# C is total alkalinity (carbonate form at pH following lime pretreatment)
Phase II results are presented elsewhere (4) due to space limitation
here and the fact that they focused on a natural water. Those results show
that Cr is not appreciably removed unless reduced to the +3 oxidation state
as has been shown by Sorg (7). However, Cd 957o lead removal at as
little as 7 mg/1 K2Fe04 (as Fe) at optimum pH between 7 and 8. For cadmium,
up to 20 mg/1 K^Fe04 (as Fe) can only remove 80-907» of original cadmium at
the optimum pH=9. These along with Phase III results provide a basis for
generalizing the interpretation of Cd, Pb and Zn removal from selected steel
plant wastewaters.
Phase III allows us to gain some insight into how a few solution condi-
tions affect trace metal removal, in well defined laboratory systems. Type
484
-------
and concentration of Fe coagulant, solution pH, ionic strength, and sulfate
and thiocyanide concentrations are independently varied. Thiocyanide rather
than thiocyanate was chosen because cyanide has carbon at a lower oxidation
state than cyanate and ferrate may oxidize it somewhat in the process of re-
moving the heavy metals (8). Ferric iron added as FeCl3 on the other hand
should show no such reaction. Process wastew.aters may be pretreated by lime;
solution chemistry reflecting this was also examined in Phase III. A solu-
_3
tion condition of 4 x 10 eq./l alkalinity, 20 mg/1 sulfate and zero thio-
cyanide was standard. Initial solution pH was equilibrated in batches at
from 6.5 to 9.5 prior to dosing with potassium ferrate or ferric chloride at
from 5-40 mg/1 (as Fe).
Some of the more salient results for cadmium, zinc and lead removal by
ferrate and ferric are shown in Figs. 5, 6 and 7 respectively. More detailed
results are also available elsewhere (6). Cadmium and zinc behave similarly;
both show rapid increase in removal with increasing pH up to 8. Thereafter,
behavior is different for each coagulant; at a high dose (40 mg/1 as Fe),
ferrate induced removal decreases from a 9070 maximum at pH = 8 while ferric
generated removal plateaus at 95% at pH>8. At lower doses ferrate shows a
minimum removal at pH = 8 that is significantly less than for ferric. In both
cases at pH = 9, considerable removal (707, for Zn and 50% for Cd) is attribu-
ted to precipitation induced removal alone. Both ferric and ferrate removed
lead more efficiently at lower doses and especially at lower pH than cadmium
or zinc. As little as 5 mg/1 of Fe (III) removed an average of 85% Pb over
the pH range of 7 to 9 while even 40 mg/1 Fe (VI) would not achieve this un-
less pH<8. .
The three trace metals Pb, Cd and Zn were generally removed better by
FeCl3 than K2Fe04. The difference in removal almost disappears when jar test
supernatatant is filtered. This is most true for zinc and for all three
metals at pH<8. Furthermore, comparable experiments at ionic strength up to
0.05, sulfate to 2000 mg/1 and thiocyanide to 100 mg/1 showed only a slight
diminution (<15%) in removal. This only occurred around 15 mg/1 ferrate (as
Fe) and at no other dose up to 40 mg/1.
Actual wastewaters from a local Chicago steelmaking plant were tested in
Phases IB and IV. Wastewater streams from basic oxygen furnace (EOF), blast
furnace, hot and cold rolling, acid pickling and cokemaking were sampled and
analyzed on three occasions for trace metal and other parameters. These
analyses guided selection of solution conditions for model tests in Phases IA
and III and evaluated the wastewaters most amenable to ferrate treatment.
Details of this are also presented elsewhere (4). Finally, two wastewater
streams with potential and need for treatment to remove trace metals were
chosen: blast furnace scrubber blowdown and coke plant wastewater. The form-
er is a moderate volume waste stream that may be discharged while the latter
requires considerable treatment to remove organics and ammonia before dis-
charge. In both waters, ambient levels of most trace metals including Pb and
Cd were in the 0.05 mg/1 range and were largely not pursued. Zinc existed in
the 1.0 mg/1 range in all blast furnace blowdown samples and was given great-
est attention in terms of direct measurement in laboratory treatment tests.
Wastewaters were treated with both lime (pretreatment) and ferrate (post
treatment) in Phase IB and ferrate only treatment in Phase III. Considerable
lime dose (1500 mg/1) was required to remove zinc in BF blowdown from
485
-------
100
J
< 80
o
60
UM
z
u
o
a:
uj
a.
40
20
8
P
H
10
FIGURE 5. Cadmium removal vs. pH for fixed dose of ferric (III)
and ferrate (VI). Dose and oxidation state, 3 or 6 of coagulant
is indicated as pairs e.g. (40,6) is feriate at 40mg/4 as Fe.
Removal by pH adjustment alone is the (0) curve.
486
-------
100
NC
-I
o
80
60
Z
IU
o
ao
8
H
/O
FIGURE 6: Zinc removal vs. pH for fixed dose of Fe(III) and Fe(VI)
Same comments as figure 5.
487
-------
100
J
0
80
IU ,.
a: 60
I- 40
z
LJ
O
U 20
Q-
7
LEAD
/(o)
8
p
H
10
u
FIGURE 7: Lead removal vs. pH for fixed dose of Fe(III) and Fe(VI),
Same comments as figure 5.
488
-------
1.1 mg/1 .to 0.15 mg/1 but none was required for EOF and cokemaking secondary
treated wastewater with initially 0.06 mg/1 and 0.03 mg/1 zinc, respectively.
With virtually no lime addition these latter wastewater concentrations were
reduced to analytical detection limits. Pretreatment of BF blowdown with
150 mg/1 lime resulted in 0.7 mg/1 residual zinc at pH~8 and only 1.5 mg/1
ferrate (as Fe) post treatment was required to achieve 0.05 mg/1 residual Zn.
An additional set of samples was obtained to measure the matrix effects
of wastewater constituents on trace metal removal. Samples of ammonia lime
still effluent from a coking operation both before (1) and after (2) acti*
vated sludge treatment and a blast furnace scrubber blowdown (3) were spiked
with labelled Zn65 to 1.0 mg/1 and CdiQ9 to 0.5 mg/1. Wastewater samples
were adjusted to pH = 7.5-8.0 and dosed with 0-100 mg/1 (as Fe) ferrate;
results are given in Table 3 in the order of samples 1, 2 and 3. In the
upper part of the table, pH adjustment alone shows large removal of trace
metals. The lower part of the table shows that after pH adjustment residual
Cd and Zn may vary considerably. Then with increasing dose of ferrate at
fixed pH, trace metal removal also monotonically increases. Sample 3 re-
quires least and sample 1 the greatest dose to achieve a 50% removal percent-
age for both Cd and Zn. Samples 1 and 3 also show closely parallel removal
of Cd and Zn with dose suggesting that removal is closely related to the fact
that the added trace metal is associated with a solid phase at time of fer-
rate dosing. For comparison, bracketed values on Table 3 show that post lime
treatment by ferrate at as little as 1.5 mg/1 as Fe would remove 767» of total
zinc in an actual wastewater sample. This compares closely to the steep
dose-response relationship observed for the spiked BF blowdown sample.
DISCUSSION
Experimental results from Phases IA and III were compared to calcula-
tions using REDEQL/II (9), a computer code for determining chemical equilib-
rium in heterogeneous natural or simulated waters. REDEQL provides for pre-
cipitation of solid phases of trace metals and adsorption of soluble trace
metals. Theoretical calculations for Phase IA significantly overpredict ob-
served results at any pH. For Phase III, the qualitative trend of a sig-
moidal adsorption edge (cf. Fig. 2) is predicted for Cd and Zn, but shows
little dependence on ferrate dose. Results presented in Figs. 5 and 6 show
portions of such sigmoidal curves with a strong dependence of position on
dose. The reason for this discrepancy is not clear, but could be that the
process is kinetically controlled. REDEQL/II results for lead agree better
with experiments showing qualitatively that removal should peak at about
pH = 8. Theoretical prediction in that case still underpredicts observed
removal. The overriding observation from all the REDEQL calculations is
that dose effects are more poorly predicted than pH effects. One conclusion
is that more attention to rate processes is required if model calculations
are to have wider utility and predictability.
Model wastewaters (well defined chemistry) and actual steel plant waste-
waters are most similar for blast furnace blowdown and show good correla-
tion, cf. Figs. 5 and 6 and Table 3. When suspended solids exist in the ac-
tual wastewater, high removal of trace metals by ferrate or ferric must be
paralleled by high suspended solids removal. Ferrate reaction at high pH
489
-------
TABLE 3. Percent Removal of Cd and Zn from Three Wastewaters
(1) (2) (3)
SOURCE
Trace metal
Trace metal
initial cone.
Cone, after
pH adjust to
7.5 - 8.0
% removal
by pH adjust
to 7.5 - 8.0
% removal
after pH
adjust and
ferrate dose
given
DOSE
mg/1
as Fe
0
1.5
10
20
40
100
LIME STILL
Cd
0.5
0.075
85
0
7*
13
20
47
Zn
1.0
0.035
96
0
14
29
43
71
LIME STILL
(After secondary
treatment)
Cd
0.5
0.43
15
0
20
35
60
82
Zn
1.0
0.87
13
0
11
24
47
79
B.F. SLOWDOWN
Cd
0.5
0,06
89
0
64
73
86
91
Zn
1.0 (0.7)**
0.29 (0.7)
71 (0)
0
(76)
72
76
90
93
* Percent removal calculation after pH adjustment.
** Bracketed entries are for lime pretreated wastewater.
(1), (2), (3) are sample numbers as mentioned in the text.
490
-------
produces a relatively stable sol of iron hydroxide that requires post fil-
tration to achieve high trace metal removal. Ferric coagulation of trace
metals does not have this liability.
Economic considerations for electrolytic generation of ferrate from
scrap iron at a steel plant suggest that production may be on the same order
as the related oxidant, permanganate (1).
CONCLUSIONS
Based on the work described herein several conclusions are:
1. Ferrate has promise for solids and metal removal in steel plant
wastewaters.
2. Lime pretreatment is effective to reduce ferrate doses but post
filtration may be required
3. Complex dependence on solution chemistry is observed for lime
plus ferrate or ferrate only treatment.
4. Comparison of ferrate and ferric show varying optima dependency on
the metal ion to be removed. Cadmium is better removed by ferrate
at low pH and lead by ferric.
5. Theoretical equilibrium computations show considerable difference
from experiment except for Pb removal.
6. Removal of metals in actual wastewaters were high at low dose when
provision for adequate solid-liquid separation is included.
7. Conjunctive oxidation benefits of ferrate should be explored for
priority pollutant removal in steel plant wastewaters.
ACKNOWLEDGEMENT
We gratefully acknowledge the financial support of The American
Iron & Steel Institute and U.S. Environmental Protection Agency for
partial support of this work under Grants No. 78-397 and R806064-01,
respectively.
491
-------
REFERENCES
Waite, T.D., "Feasibility of Wastewater Treatment with Ferrate,"
J. Envr. Engr. Div. ASCE. 105, No. EE6, pp. 1023 - 1034, 1979.
Waite, T.D. and Gilbert, M., "Oxidative Destruction of Phenol and Other
Organic Residuals by Iron (VI) Ferrate," J. Wat. Poll. Ctrl. Fed^. 50.
pp. 543 - 551, 1978.
Leckie, J.O., et. al. "Adsorption/Coprecipitation of Trace Elements from
Water with Iron Oxyhydroxide," Final Report Project 910-1, EPRI CS-1513,
Electric Power Research Institute, Palo Alto, Calif., Sept. 1980.
Fitzpatrick, J.A., "Trace metal and Suspended Solids Removal from Water
and Wastewater by Potassium Ferrate," Final Report Amer. Iron & Steel
Inst. Grant No. 78-397 and EPA Grant No. R806064-01, 1982.
Wang,j., "Cadmium Removal from Aqueous Solution by Lime and Ferrate,"
M.S. Thesis, Northwestern University, August, 1981.
Davis, K., "Removal of Cadmium, Zinc and Lead from Water and Steel Plant
Wastewater with Ferrate," M0S. Thesis, Northwestern University, June, 1982.
Sorg, T., "Treatment Techniques for the Removal of Inorganic Contaminants
from Drinking Water," in Manual of Treatment Techniques for Meeting the
Interim Primary Drinking Water Regulations, EPA - 600/8-77-005, U.S. EPA,
MERL, WSRD, Cincinatti, Ohio, May, 1977.
Nerezov, V.M. et. al,, "Detoxification of Cyanide-Thiocyanide Containing
Waters," Russian Patent No. 432, 765 (June 25, 1977), Otkrytiya, Izobret,
Prom. Dbraztsy, Tovarnye Znacki, 54. 23, p. 194, 1977.
McDuff, R.E. and Morel, F., "REDEQL, A General Program for the Computation
of Chemical Equilibria in Aqueous Systems," Tech. Rept. EQ-72-01, Keck
Laboratories, Calif. Inst. Tech., Pasadena, Cal. 1972.
492
-------
PILOT EVALUATION OF ALKALINE CHLORINATION
ALTERNATIVES FOR BLAST FURNACE SLOWDOWN TREATMENT
by: Stephen A. Hall. Karl A. Brantner.
John W. Kubarewicz. & Michael D. Sullivan
Industrial Division
Metcalf & Eddy, Inc.
Boston. Massachusetts
INTRODUCTION
Ammonia and cyanide are among the pollutants found in blast furnace blow-down for which Best
Available Technology Economically Achievable (BAT) effluent limitations have been proposed. In order to
meet these limitations, both ammonia and cyanide may have to be removed prior to discharge. Alkaline
chlorination is a potential process for achieving removal of both pollutants from a single wastewater source.
In keeping with the goals of BAT, alternative treatment systems incorporating alkaline chlorination have
been evaluated with respect to performance and cost-effectiveness in removing ammonia and cyanide from
blast furnace blowdown. The three alternative treatment systems were single-stage alkaline chlorination.
two-stage alkaline chlorination and combined air stripping/alkaline chlorination. Pilot plant testing was
conducted to develop performance data, operability and cost estimates for full scale blowdown treatment
incorporating alkaline chlorination.
PROCESS ALTERNATIVES
The three alternative alkaline chlorination treatment systems were tested to determine their effective-
ness for the removal of both ammonia and cyanide. The unit processes that constituted the treatment
systems included breakpoint chlorination and air stripping for ammonia removal, alkaline chlorination for
oxidation of cyanide, dechlorination for reduction of residual chlorine, and flocculation. clarification and
filtration for removal of the suspended solids generated by these chemical treatment processes.
AMMONIA REMOVAL
Ammonia removal from blast furnace blowdown can be accomplished by either breakpoint chlorina-
tion or air stripping. Theoretically, complete ammonia removal ran be achieved by either method. In break-
point chlorination. the degree of ammonia removal by chlorination is a function of pH and chlorine dosage
(1). As chlorine is added to the wastewater, the combined chlorine residual increases, indicating the forma-
tion of chloramines. An increasing dosage of chlorine results in a decrease of combined residual to a
minimum or breakpoint. Further chlorine addition results in a free chlorine residual signifying complete
ammonia oxidation. Ammonia is oxidized to nitrogen gas by the following reaction:
2 NH3 + C12-*"N2 + 6 HC1
493
-------
The optimum pH range for breakpoint chlorination is 6.5 to 8.5. Outside of this range the chlorine dosage
required to reach breakpoint increases and the reaction rate decreases. If insufficient chlorine is present to
achieve breakpoint chlorination. chloramines will remain. Under alkaline conditions monochloramine pre-
dominates.
In air stripping, ammonium ions are converted to the dissolved gaseous phase by raising the pH (2).
The wastewater is then introduced into a stripping tower where air flow is induced in a counter-current
mode to liberate the ammonia gas. The rate of ammonia removal is a function of pH and temperature. As
the pH is raised from neutral pH. ammonium ions are converted to the ammonia (Nr^) form as shown by
this equation:
NH4+ -**NH3 (gas) + H+
At pH 12 only the dissolved ammonia gas is present. Normally, the optimal pH for air stripping is between
10.5 and 11.5. a level that insures that the ammonia is in the dissolved gaseous phase. Greater ammonia
removal is obtained at higher temperatures. In the case of a blast furnace recycle system, the temperature
vear round would be favorable.
Once the ammonia is in the gaseous phase, there are two factors which affect the transfer from liquid
to gas:
1. Mass transfer
2. Ammonia equilibrium
The resistance to mass transfer between liquid and air is at a minimum when water droplets are being
formed and reformed. Normally this formation and reformation of water droplets is most effectively
accomplished in a packed tower as opposed to a tower without packing (c.wling tower). The approach to
ammonia equilibrium in the air around a water droplet will affect the transfer. Therefore, air flow is
usually induced to purge the ammonia. As the air to liquid ratio increases, the ammonia transfer increases.
The potential of employing air stripping as a preliminary step to the alkaline chlorination process was incor-
porated in the pilot plant as an alternative to remove ammonia, thus reducing the chlorine demand.
CYANIDE REMOVAL
Alkaline chlorination for cyanide removal is achieved by raising the pH of the wastewater and dosing
with chlorine. In an alkaline system (pH > 8.5). cyanide is oxidized to cyanate according to the following
reaction:
NaCN -I- 2 NaOH + C12 •*• NaCNO + NaCl + H20
The reaction rate increases with pH. At pH 10 to 11. the reaction time is approximately 5 to 7 minutes.
Cyanate undergoes further oxidation with excess alkali in the presence of free chlorine according to
the following reaction:
2NaCNO + 1 NaOH 4- 3C12 -»2 C02 + 6 NaCl + N2 + 2 H20
494
-------
This reaction proceeds rapidly at pH 8.5 to 9.0. requiring 10 to 15 minutes for completion. However, at pH
10 the reaction is slow and only partial cyanate destruction may be achieved in a practicable detention time.
The complete destruction of cyanide to carbon dioxide and nitrogen theoretically requires 6.82 parts
of chlorine per part of free cyanide and 1.225 parts of hydrated lime per part of chlorine applied (3). In
practice it requires more chlorine to oxidize other chlorine demanding compounds. Lime requirements will
vary depending on the initial alkalinity of the wastewater.
Alkaline chlorination of cyanide wastes has been successfully applied in the' metal plating industry
where high cyanide concentrations (dumped plating baths) of 500 to 1.500 ppm are not uncommon (3.4.5).
The usual treatment scheme employs a two-stage reactor system. The first stage is maintained at a pH of 10
to 11. optimum for oxidation of cyanide to cyanate.. In the second stage reactor, the pH is controlled at
approximately 8.5 for optimum cyanate oxidation. Chlorine may be added in both stages by independent
control systems to insure an adequate supply.
DECHLORINATION
Following alkaline chlorination. dechlorination must be employed to remove residual chlorine prior to
discharge. Dechlorination may be accomplished by addition of sulfur dioxide, sodium sulfite. sodium
bisulfite or sodium metabisulfite (l). Addition of any of these compounds provides a source of bisulfite
(HS03~) either directly or by hydrolysis. The bisulfite reduces hypochlorite to hydrochloric acid according
to the following reaction:
HOC1 + HS03--^ Cl- + S04= + 2H+
This reaction is virtually instantaneous and. thus, complete reduction of hypochlorite is simply a function
of adequate mixing. If chloramines are present from incomplete oxidation of ammonia, they will be reduced
to ammonia and hydrochloric acid according to the following reaction:
NH2C1 + HS03- + H20-»-Cl- + S04= + NH4+ + H+
This reaction is also instantaneous. Thus, any ammonia that was converted to chloramines upon chlorina-
tion will not be removed from the wastewater.
Dechlorination reduces alkalinity by releasing protons. Destruction of high chlorine residuals results
in' a depression of pH. The degree of pH depression depends upon the alkalinity of the wastewater.
Neutralization may be required to adjust the pH to a suitable level for discharge.
INTEGRATED TREATMENT
The unit processes previously discussed were combined to provide three different integrated processes
to be demonstrated for achieving effective removal of ammonia and cyanide. The process flow trains that
were demonstrated were single-stage alkaline chlorination. two-stage alkaline chlorination and air stripping
followed by two-stage alkaline chlorination. as shown in Figure 1. Variations of the two-stage process aimed
at cost reduction were piloted. The objective was to determine whether these variations could effectively be
operated and still achieve both cyanide and ammonia oxidation.
495
-------
1. SINGLE-STAGE ALKALINE CHLORINATION
NaOCI
LIME
SLUDGE
2. TWO-STAGE ALKALINE CHLORINATION
NaOCI NaOCI
LIME
\
©
sl
H2S04
\l •
>
-^
|pHl||QRPl||pH2||ORP2|
PH10.S pH8.S
FILTER
\\
SLUOQE
3. COMBINED AIR STRIPPING/ALKALINE CHLORINATION
Tl
PH11.0
\ll
y
PM-10.I pHM
Figure 1. Alternative Process Flow Trains
496
-------
The two-stage process would normally be employed to achieve both cyanide and ammonia removal by
alkaline chlorination. This system, similar to that used for cyanide plating wastes, is sized to allow sufficient
detention time for all cyanide to be oxidized to cyanale in the first reactor. The second reactor is sized to
allow sufficient detention time for all cyanales and ammonia to undergo complete oxidation. The reactor
effluent is then clarified, filtered and dechlorinaled prior to discharge.
Single-stage alkaline chlorination varies from the normal two-stage alkaline chlorination process in
that the pH is maintained at 10 to 11 in a single reactor and not readjusted throughout the treatment system
until discharge. This system would be sized to allow sufficient time for oxidation of all cyanide to cyanate
and for partial oxidation of cyanate and ammonia in one reactor. The detention time in the flocculator and
clarifier is utilized to provide sufficient lime for oxidation of cyanate and ammonia, since pH 10 to 11 is not
favorable for this reaction to occur in a practicable reactor detention time. Prior to discharge, the effluent is
dechlorinated and may be filtered for polishing.
The combined air stripping and alkaline chlorination alternative reduces the chlorine requirements by
air stripping the ammonia while providing favorable conditions for the complete oxidation of cyanide. The
optimum pH level of 10.5 to 11.5 is maintained through the air stripping lower to provide ammonia
removal. The alkaline chlorination step is sized to allow sufficient time for the cyanide to be oxidized to cya-
nate at pH 10.5 to 11.5 in the first reactor. The second stage reactor is sized to provide adequate detention
time for the cyanates to be oxidized and as a polishing step to oxidize ammonia not removed by air strip-
ping. Filtration is employed to remove residual solids. Following filtration the waste is dechlorinated before
discharging the treated blast furnace blowdown.
TESTING PROGRAM DESCRIPTION
A trailer-mounted physical/chemical treatment pilot plant with facilities for chemical feed, floccula-
tion. clarification, and filtration was employed to demonstrate the alternatives for alkaline chlorination
treatment process. The configuration of the pilot equipment is illustrated in Figure 1.
The mobile pilot plant was operated to demonstrate alkaline chlorination of blast furnace blowdown at
two different steel mills. The source of water at both sites was the blowdown from blast furnace gas scrub-
bing recycle system. Prior to blowing down, the recycled water had undergone clarification for suspended
solids removal and cooling for thermal reduction.
Bench scale tests were performed at the outset of each pilot study to determine the oxidation-reduc-
tion potential (ORP) set points and estimate chemical requirements. Jar tests of varying chlorine dosages
were conducted to simulate pilot plant operating conditions. The pH was adjusted to operating levels using
lime and sulfuric acid as necessary. During these tests ammonia, cyanide. pH. ORP. and total and free
residual chlorine were monitored. These parameters were then defined as functions of chlorine dosage. The
chlorination breakpoint and oxidation potential levels could then be determined.
The importance of the bench tests should not be overlooked because of the influence of water quality
on the treatment process. Chlorine demand and ORP set points vary among wastewaters. ORP is very pH-
dependent and therefore tight pH control must be maintained. Timing is important to allow for time-depen-
dent reactions such as cyanate oxidation to reach completion, and can affect monitoring of both ORP and
residual chlorine.
497
-------
The pilot plant was operated at ft gallons per minute (gpm) through the rapid mixing tank for chemi-
cal treatment and rlarifier for suspended solids removal. A flow rate of 5 gpm was maintained through the
dual media filtration system for [Hilishing. At these flow rates the detention time was 8.3 minutes in the
floeculator and 2 hours in the clarifier. The clarifier overflow rate was 350 gallons per day per square foot
(gpd/ft-) or 0.25 gprn/ft^. The filter loading rate was 5 gpm/ft^. The sludge accumulated in the bottom of
the clarifier was wasted at the rate of production.
All pumps and rapid mixers were controlled from a control panel. There were two pH meters and two
ORP meters located at this panel. Both pH and ORP were indicated and recorded. Associated with these
meters were controllers which were employed to regulate chemical feed based on specified set points and in
response to pH and ORP probes.
SINGLE-STAGE ALKALINE CHLORINATION
Wastewater was introduced into the first reactor with a 10-minute detention time where a slurry of
hydrated lime was applied to maintain pH 10.5. The lime addition was controlled by a pH controller with its
input probe located in the first reactor. Sodium hypochlorite was also applied in the first reactor. An ORP
controller was employed for sodium hypochlorite addition. The ORP setpoint (+700 mV) was determined
in the bench test as the level at which free chlorine residual existed. The input probe was located in a
second reactor with a detention time of 10 minutes to allow additional reaction time before measuring the
potential level. No chemicals were added to the second reactor. The flocculator-clarifier detention time pro-
vided additional r |