Opening Session
Session 1
MULTIME POLLl ABAT
Session 2
SOJii> iSTE POLLUTION AE
on
sponsore y
US EPA.
INDUSTR! AL E N V j R O N M E P 1 1 ' AL R E S E A R ( !
Research, triangle Park, NC
.-.-HfrJKIA
Washington, DC
tER
The William n
-------
The work described in this paper was not funded by the IJ,S. Environmental
Protection Agency. The contents do not necessarily reflect the yiews p.f
the Agency and no official endoresement should be inferred.
CHROMIUM REMOVAL MECHANISMS FOR WASTE MANAGEMENT OF A SPECIALTY STEEL
VIRGINIA M. CUSANO
and
ROBERT H. WILLS, JR,
CRUCIBLE INC,, SYRACUSE, N.Y.
INTRODUCTION
The Resource Conservation and Recovery Act along with recent rounds
of State Pollution Discharge Elimination System permits reflecting the
application of Best Available Technology Economically Achievable haye
significantly increased costs for transportation, treatment and disposal
of both solid and liquid wastes, In addition, the New York State Super^
fund places an assessment on certain wastes.
Crucible Specialty Metals Division in Syracuse, New York has been
involved in various resource recoyery projects since 1975, These projects
include:
1. Investigation of the effectiveness of an abandoned
alkaline wastebed to absorb and retain Chrome and
other heavy metals present in steel mill wastes.
2. A method of detoxification of hazardous air pollu~
tion baghouse dusts by transferring soluble chrome
from the solid phase to the liquid phase for recycle
or treatment.
3. Reuse of waste acid, from our pickling operations,
as the primary coagulant at Crucible's Industrial
wastewater treatment plant.
-------
Crucible Specialty Metals is a specialty steel mill located on a 75
acre site in Syracuse, New York. One thousand (1000) tons of stainless,
tool, high speed, valve, alloy and super alloy are produced per week. The
main markets for the products are the aerospace, automotive and machining
industries.
Melting operations utilize EAF's (electric arc furnaces). In 1906
Crucible started the first and only electric arc furnace in the Western
hemisphere. A twenty two (22) and thirty five (35) ton EAF are now in
use. High grade scrap and alloying materials are charged into the fur-
naces along with slagging agents. When the initial melt is done, the
molten steel is transferred to a thirty five (35) ton Argon Oxygen De-
carburization vessel (ADD) for refining.
Crucible also has a Vacuum Melting department consisting of Vacuum
ArcRemelting (VAR) and Crucible Particle Metallurgy (CPM) processes. The
VAR furnace uses a metal electrode (poured in the Melt Shop) and remelts
the steel using high direct current and vacuum to remove impurities. The
CPM process uses an induction furnace to remelt high speed steel. The
molten steel is then poured through an atomizer where water or nitrogen
quenches it, forming micro ingots. The micro ingots are heated to compac-
tion in an autoclave.
Cogging operations consist of a 2000 Ton Davy press and 26" two-high
reversing mill which reduces the large ingots into billets.
The Rod and Bar Mill consists of a 18" two high reversing mill for
roughing, 3 stands of 14" Mill for intermediate reduction, 2 stands of
12" Mill for bar finishing and a 10" vertical stand and 7 stands of 9"
Mill for wire products. The North Rolling Mills consist of 14", 12-2 and
9" jobbing hand rolling mills.
The North Hammer Shop uses 2,000, 4,000 and 8,000 pound steam hammers
to provide a good hot finished surface on the steel.
Conditioning consists of grinding and pickling operations. Mechanical
grinders and hand operated swing frame grinders use abrasive wheels to re-
move the ingot .skin and any other major surface defects. Pickling is util-
ized to chemically remove mill scale and to impart a satisfactory finish
on the product. Pickling solutions include muriatic acid, sulfuric acid,
nitric hydrofluoric acid, potassium permanganate, molten caustic, suspended
lime and various coating solutions.
-------
Bar Finish operations include straightening, grit blasting, lathe
turning, grinding, machining, cold draw, sawing, inspection and shipping.
Waste streams from these various operations are similar in volume
and basic pollutants to any other steel mill, however, since Crucible
makes 400 grades of specialty steel, more "specialty" pollutants (e.g.
Cr, Ni, W, etc.) and less "poor quality" scrap pollutants (e.g. Pb, Zn,
etc.) are seen. The 400 grades produced also lead to variability within
the waste streams.
Typical waste stream analysis for solid wastes are presented in
Table 1, for leachables from solid waste in Table 2, and for liquid
wastes entering the waste water treatment plant in Table 3.
Of specific concern is the parameter Cr (Hexavalent chromium) due
to its toxic and soluble nature. Cr is present in EAF baghouse dust
and AOD baghouse dust, as a result of the high temperatures and oxida-
tion that occurs during melting and refining. Cr is also generated
in the molten caustic tanks. The chrome from the mill scale and the
product are oxidized into the hexavalent state. The Cr+" is not a
problem per se while in the solid waste form, but, becomes a major
problem when water is leached through the material.
The solid wastes generated at the mill are landfilled on top of an
abandoned alkaline waste bed. The waste bed contains 60 feet of waste
material from the Solvay Process production of soda ash (Na£ CO^). The
alkaline nature of the waste bed, along with its homogeneity and adsorp-
tive capacity make it a unique repository for metallic wastes. Ground
water analysis from various levels both in ancLunder the waste bed show
non detectable levels of metals (including Cr ) from the steel making
wastes which have been deposited there over the past 9 years. Metals
which are leached by rain water are adsorbed and retained in the top
layer of the Solvay Process waste.
EAF and AOD dusts have not been deposited at our landfill since
March 3, 1982 and are now being recycled for recovery of nickle and
chrome units.
Prior to recycling, a method for detoxification of EAF and AOD dusts
was developed (See figure 1). The basic philosophy is to transfer the
hazardous component (Cr~*"°) from the solid phase to the liquid phase with
subsequent treatment or reuse. The resulting wet dust would not contain
any hazardous leachable components (as defined by RCRA) and could be dis-
posed of in a sanitary landfill.
The detoxification process involves a leach tank and high shear
mixer to thoroughly wet the dust. The dilution ratio is 5 (total mixture
weight) to 1 (solids weight).
The solvent used will be dependent on the anticipated final use of
the liquid:
-------
TABLE 1
COMPOSITION OF CRUCIBLE HASTES
ing/Kg or ppm
Wastes
Parameters
ALK Al
As Cd CN Cr(T) Cu
Pb Hg
N1
Zn
A) Non-Hazardous Wastes
Coolant Swarf (Krasney)
WWTP Sludge
Slag
Mill Scale
Boiler House Ashes
Grinding Dust
Coolant Swarf, South
Wire Hill
B) Hazardous Wastes
Air Pollution and
AOD Oust
Waste Caustic Solids
62 590 3.0 100 0.04 50,000 5520 136,000 10 0.04 NA 19
930 2,000 5.0 48 0.04 10,800 660 92,000 MO 0.2 3,340 460
400 13,400 2.0 38 0.15 6,400 89 29,000 30 0.04 2.560 40
12 210 10 70 0.04 1.380 1840 58,600 2.0 0.04 16,000 18
48 2,000 2.0 1.2 0.04 47 33 111,000 2.0 0.04 48 17
20 290 7.5 8.1 0.004 60.800 3360 500,000 18 0.04 25,000 21
94
1600
6000
120 5.0 71 0.04 26.000 47 166,000 4 0.04 11,800 50
4000 7.0 74 0.40
120 7.0 8.0 2.75
6,600 840 100,000 2600 0.6 4,200 5000
7.200 75 8.600 2 0.04 336 7.0
-------
TABLK 2
FP TOXICITY TEST LEACHATE RESULTS
Wastes
Parameters (mg/1 )
Arsenic
Detection Limits
A) Non-Hazardous Wastes
WWTP Sludge
Slag
Mill Scale
Boiler House Ashes
Grinding Dust
Coolant Swarf, South
Hire Mill
Excavated Soil Material
B) Hazardous Wastes
Air Pollution Dusts
ADD Dust
Waste Caustic Solids
LT
LT
LT
LT
LT
LT
LT
LT
LT
0.002
.002
.002
.002
.002
.002
.002
.002
.002
.002
0.007
Barium
LT
LT
LT
LT
LT
LT
LT
LT
LT
LT
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
Cadmium
0.01
0.05
0.01
0.20
0.01
0.11
0.04
0.10
1.59
0.33
0.04
Total
Chromium
0.01
LT 0.01
0.02
0.15
0.30
0.33
1.70
0.09
56
LT 28
LT126
Lead
0.02
0.1B
0.16
0.08
0.08
0.02
0.10
0.48
0.08
0.16
0.26
Mercury
0.002
LT .002
LT .002
LT .002
LT .002
LT .002
LT .002
LT .002
0.03
LT .002
.002
Selenium
0.002
LT .002
0.022
LT .002
LT .002
LT .002
LT .002
LT .002
0.012
0.024
0.008
Silver
0
LT
LT
LT
LT
LT
LT
0
0
LT
LT
.01
.01
.01
.01
.01
.01
.01
.03
.01
.01
.01
RCRA Limiting Con-
centrations
5.0
100
1.0
5.0 5.0
0.2
1.0
5.0
-------
SUMMMARY OF
TABLE 3
AVERAGE DAILY
CONTAMINANT POUNDAGES
TANK A INFLUENT
(POUNDS PER 4-HOUR PERIOD)
Parameter
TSS
TOC
Cyanides-Total
Cyanides-Oxid.
Nitrate
Sulfate
Chloride
Cadmium-Total
Cadmium-Soluble
Chromium- Hex.
Chromium-Total
Chromium-Soluble
Copper-Total
Copper-Soluble
Iron-Total
Iron-Soluble
Zinc-Total
Zinc-Soluble
Fluoride
Average Value
4- days
606
112
1.42
0.087
132
3803
665
ND
ND
1.09
26
5.4
3.6
0.6
172.5
25.6
4.0
1.86
90
Highest
Daily Average
Value
1451
226
2.60
0.21
175
5647
811
ND
ND
2.18
66.0
17.0
6.6
1.12
495
100
7.1
3.5
111
Lowest
Daily Average
Value
94
27
0.79
0.007
7.0
2104
401
ND
ND
0.024
1.3
0.1
0.4
0.08
20.0
0.4
1.84
0.6
63.0
Ratio
Variability
15.4
8.4
3.3
30.0
2.5
2.7
2.0
0
0
90.8
50.8
170
16.5
14.0
24.7
250
3.9
5.8
1.8
-------
-__—pH AND REOOX POTENTIAUMonitor/Riordir)
-Dial Lrah Tor*
Solid! Tfanilir Pump
FIGURE 1
-------
1. If the process is used for recyling, water will be the
solvent. The resulting chromate solution has potential
for use as a corrosion inhibitor, for sale to a chrome
plater, or for chrome recovery via electro winning or
other processes.
2. If the process is used strictly for detoxifying the dusts,
a water and acid mixture will be the solvent used to increase
the leaching and initiate treatment of the chromate. The
actual chromate reduction is discussed later in this report.
The resultant wet dust can be pumped to a sludge and vacuum filter
or be disposed of directly, depending on the solids content.
Chromate can enter the waste water at pickle lines by drag out,
carryover, or during normal rinsing and quenching operations. The
chromate enters the waste stream in slugs due to the batch type oper-
ation. This accounts for the large variability ratio as shown in Table
3.
Crucible's industrial waste water (4.5 MGD) is collected at three
main pumpstations ranging in flow capacity from 200 gpm to 3800 gpm (See
Figure 2). The waste water is pumped to the aeration tank where:
1. It is blended
2. Gross pH adjustment occurs
3. FeS04 (ferrous sulfate) is added
The waste water flows by gravity to a 120 foot diameter primary clarifier
where:
1. pH is controlled at 7.6
2. .6 ppm of a highly anionic polymer is added
3. FeS04 is added as the primary coagulant and
to insure proper chromate reduction
87% (3.9 MGD) of the clarified water flows into a wet well and is
recycled back to the mill for reuse. The remaining 13% (.6 MGD) flows
to a secondary clarifier where:
1. pH is adjusted to 8.2
2. .6 ppm of a highly anionic polymer is added
Sludge from both clarifiers is pumped to a sludge thickener and then to
two 8 foot diameter vacuum filters.
Hexavalent Chrome may be removed from waste water by any of the
four processes:
1. electrochemical reduction followed by
chemical precipitation
2. Ion exchange
3. Dissolved air flotation
-------
14"
Mill
9"
Mill
-Q
Pump
Station
"4
Pickling
J
—a—
Pump
Station
"I
.-
Furnace
Straight Line
Pickling
26"
Mill
Needle
Wire
Copper
Plating
Pump
Station
Recycle
Wet Well
Boiler
House
I 261 Mill I
[ Furnaces _J
Melt 1
Shop |
1 , s*
^\
Ji Resevolr 1
Solids
•—I Contact
Tank
Effluent
Treatment
Effluent Dltchorge
To Tributary 5A
FIGURE 2
-------
4. chemical reduction followed by chemical
precipitation
Electrochemical reduction processes rely on electrolytic production
of ferrous iron to initiate hexavalent chrome reduction. Since a ready
supply of ferrous iron is available in waste pickling acids, electro-
chemical reduction processes are unnecessary and would be expensive.
Ion-exchange is a hex-chrome removal technique used for waste streams
containing no metal species other than chromium, such as cooling water blow-
down. Presence of other metals in the Crucible waste stream makes use of
Ion-exchange impossible.
Dissolved-air ion flotation is a removal process whereby a cationic
surfactant is added to hex-chrome bearing water forming a surface active
complex which exists as a collodial suspension. This suspension is floated
to the surface by air bubbles, forming a froth which is aggregated by
polymer additions and drawn off for disposal.
Both Ion-exchange and dissolved-air Ion flotation remove hex-chrome
without reduction and would leave a disposal problem of uncertain magni-
tude and severity.
Hexavalent chrome may be removed from the waste stream by chemical
reduction of the hex-chrome to trivalent chrome and eventual precipitation
of the trivalent chrome as the hydroxide.
Reduction may be carried out using S02 or the Sulfite/bisulfate so-
dium salts, or with ferrous iron (Fe+^). Chemical precipitation of tri-
avalent chrome as the hydroxide may be done with caustic soda or lime.
Since Crucible has an available supply of ferrous iron in the form of
waste pickling acids and presently uses caustic soda, the ferrous reduction/
caustic precipitation method is the most cost effective and logical treat-
ment scheme.
Hexavalent chromium is actually present in waste waters in the form of
oxyanions chromate (CrO,~2) or dichromate (C^Oy"^). xhe proposed redox
reaction with ferrous iron would proceed according to the equation:
Cr04~2 + 8H++3Fe+^»- Cr+3 + 4H20 -1- 3Fe+3 and has a redox potential
of +-56 volts indicating an energetically favorable couple with three moles
of ferrous iron reducing one mole of chromium.
Laboratory testing demonstrated that a 5:1 molar ratio of Fe+^: Cr+6
resulted in optimum chromium reduction.
Crucible's Pickle lines generate both waste sulfuric (^804) and waste
muriatic acids'(HC1). Both acids were studied to determine their capacity
for hex-chrome reduction, and both proved equally effective. However,
spent sulfuric acids constitute a larger volume and a more readily available
-------
ferrous source, and therefore are more often used as the reducing
agent.
Testing also demonstrated that the ferrous/chromate redox
reaction is essentially complete in one minute with both H2S04 and
HC1 waste pickle liquors, (see Table 4, 1 through 6). The redox
reaction also appears to proceed independent of pH, with chrome re-
duction evidenced from pH 3 to pH 10.
Tests 7 through 9 were done to evaluate the stability of the
reduced and precipitated chromium to wide pH fluctuations and are
also reported in Table 4. The precipitate was digested with NaOH
resulting in final pH of 10.9 and no increase in soluble chrome.
Digestion of the precipitate with HC1 to a final pH of 2.0 also pro-
duced no increase in soluble hex-chrome, but did show considerable
soluble trivalent chrome. Precipitation of chrome with caustic under
automatic pH control guarantees low pH resolubilization of trivalent
chrome.
Waste Water Treatment Plant influent was sampled and analyzed at
the aeration tank in an effort to quantify mean hex-chrome concentrations.
Analysis of data showed a range of hex-chrome from .006nig/:L to 4.65mg/l
with a mean concentration of 1.39mg/l, Estimates to compute spent acid
requirements used an average volume of 4.5 mgd of water being processed
and a ferrous iron concentration of 5 x 1.4 mg/1 = 7.0mg/l required to
reduce the chrome. Using spent H2S04 with a ferrous iron strength measured
at 63,400 mg/1 waste sulfuric acid requirements are calculated to be 500
gallons /day maximum. This level of usage is equivalent to the upper limit
established for Ferric Chloride, the primary coagulant originally used at
the WWTP. Ferrous iron (Fe+2) oxidized to ferric iron (Fe+3) in the chrome
reduction process may now be used as a primary coagulant as well as a Hex-
chrome reducing agent.
A Ferrous Sulfate solution of approximately 3% Fe+^ is metered to the
aeration tank.
Pumping rates are established according to approximate flow rates and
an assumed Cr+6 influx of 2.0 mg/1. Ferrous Sulfate is metered to main-
tain a 5:1 molar ratio of Fe+2: Cr+ . Automatic pH control establishes
required pumping rates for caustic soda with a target pH of 7.2 as a goal
before discharge to the primary clarifier (SCTA).
pH is automatically adjusted in SCTA to 7.6 using ferrous sulfate and
caustic soda as the reagents. A high molecular weight organic polyelec-
trolyte is also added to SCTA to assist in the flocculation of the waste
water. This polymer is anionic in nature, and appears to work synergis-
tically with the ferrous sulfate resulting in improved coagulation.
Approximately 87% of the effluent from SCTA is recycled back to the
mill and requires no further processing. The remaining 13% of the flow
is routed to the secondary clarifier (SCTB). 15% l^SO^ acid, anionic
-------
TABLE 4
LABORATORY RESULTS FOR HEX-CHROME REDUCTIONS
REA
Test t CAUSTIC RtNS
1
2
3
4
5
6
7
8
Cr-tot
(mg/I) "
50.5
50.5
50.5
50.5
50.5
50.5
50.5
£r_
(nig/
50.5
50.5
50.5
50.5
50.5
50.5
50.5
CTANTS
E ACID PI(
Fe+2
1) (n.R/1.:
H2S04
H2S04
H2S04
HC1
HC1
HCL
HCl
:KLE
)
25
25
25
42
42
42
42
- (ppt. filtered
LIQUOR
,000
,000
,000
,000
,000
,000
,000
MIXTURE
MOLAR RATIO
Fe*2:
5:
5:
5:
5:
5:
5:
5:
and digested
Cr+6
1
1
1
1
1
1
1
.01N
PH
ADJUSTMENTS
PRODUCTS
Fe-tot Cr-tot Cr+6
(mg/1) (mg/1) (mg/1)
4.0 -
4.0 -
5.0 -
9.4 -
9.4 -
9.8 -
9.4 -
10.9 -
1 min. contact
15 min. contact
90 rain, contact
1 min. contact
15 min. contact
90 min. contact
no adjustment
no adjustment
3.0
4.0
.22
.01
.01
.02
.60 .01
1.77 .35
0.1
0.1
o.i
0.1
0.1
0.1
0.1
0.1
NAOH)
(ppt. filtered and digested .01N 2.0 - no adjustement
HCl)
46
11.7 0.1
-------
polymer, and caustic solution are metered and final effluent is automati-
cally maintained at a pH of 8.2. This base adjustment in pH is insurance
for complete removal of trivalent chromium as a metal hydroxide.
This reduction/precipitation removal mechanism for chromium is character-
ized by indisputable economic benefits:
1. Elimination of transportation and disposal charges for spent acid
2. Reduction of tax liabilities (State Superfund on disposal of
wastes)
3. Reduction of chemical costs at the Waste Water Treatment plant
(complete elimination of FeCl3 as the primary coagulant)
4. Improved overall treatment
5. A decrease in future, capital spending for additional pollution
abatement devices.
The use of spent acids have the additional capacity of reducing toxic
hexavalent chromium from EAF and ADD dusts, thus diminishing a potential
solid waste problem should recovery of metals from these dusts prove un-
economical.
-------
DISCLAIMER
The work described in this paper was not funded by the U.S. Environ-
mental Protection Agency. The contents do not necessarily reflect the views
of the Agency and no official endorsement should be inferred.
GROUNDWATER MONITORING STRATEGIES
FOR STEEL INDUSTRY RESIDUE DISPOSAL OPERATIONS .
by: Andrew P. Pajak and David W. Hupe
Michael Baker, Jr., Inc.
4301 Dutch Ridge Road
Beaver, PA 15009
ABSTRACT
Regulations issued by the U.S. Environmental Protection Agency and
state regulatory agencies require that owners or operators of existing and
new surface impoundment, landfill, or land treatment hazardous waste manage-
ment facilities implement a groundwater monitoring program. Moreover,
evolving non-hazardous waste regulations are including groundwater monitor-
ing and protection requirements. RCRA hazardous waste regulations published
on July 26, 1982 are significantly different than previously imposed regula-
tions. This paper highlights the new regulatory approach and differences
between the existing and new federal RCRA regulations, especially in the
area of flexibility and variances. Design of an acceptable monitoring
network and protocol are discussed. Problems experienced in monitoring
network design and interpretation of resulting data are illustrated.
Concise flowcharts of actions required by the regulations and an outline for
designing a groundwater monitoring program are provided.
INTRODUCTION
Most owners or operators of hazardous waste treatment, storage, or
disposal facilities with interim status have, as required by Resource
Conservation and Recovery Act (RCRA) regulations promulgated on May 19,
1980, initiated groundwater monitoring programs. In fact, Baker Engineers
has assisted numerous steel industry clients in designing and implementing
these programs. However, on July 26, 1982, new interim final regulations
(Subtitle C, Part 264, Subpart F) were published specifying a substantially
revised groundwater monitoring and response program that must be applied at
new and existing surface impoundments, landfills, waste piles, and land
treatment facilities. Rather than utilize previous projects as case studies
to describe groundwater monitoring approaches, this paper focuses on several
key areas of the new regulations; especially, those areas that allow the
regulated entity to use technical judgments in complying with the objectives
of the regulations.
An overview of the groundwater monitoring requirements of the July 26,
1982 regulations is presented. Steps in designing and implementing a
technically acceptable program then are outlined. Problems experienced by
-------
Baker Engineers on previous projects are identified. The remainder of this
paper highlights areas of the regulations.allowing flexibility (but requir-
ing a sound technical basis to gain variances) and briefly addresses a risk
analysis approach to formulating the technical basis.
OVERVIEW - GROUNDWATER MONITORING AND RESPONSE PROGRAM
In the interim final RCRA regulations of July 26, 1982, three types of
groundwater monitoring programs are specified: detection monitoring,
compliance'monitoring, and corrective action. The objectives are to:
1. detect leakage from the regulated unit (any impoundment, landfill,
pile or land treatment waste management unit that receives hazard-
ous waste after January 26, 1983)
2. determine if contaminant concentrations in the groundwater exceed
levels that protect human health and the environment, and
3. undertake and measure the effectiveness of corrective actions if
these concentrations have been exceeded.
The programs must be conducted throughout the active life of the unit
as well as a post-closure care period. They dictate increased response to
groundwater contamination as evidence of contamination increases. There-
fore, the programs may be implemented sequentially, or it may be necessary
to shift between the compliance and corrective programs when hazardous
constituent levels in the groundwater fluctuate above and below constituent
limits, or it may never be necessary to go beyond detection monitoring.
However, only one program can be in operation at a time. Figure 1 presents
an overview of the groundwater monitoring and protection approach. Fig-
ures 2, 3, and 4 illustrate more detailed flowcharts of the actions required
in each of the three programs.
GROUNDWATER MONITORING PROGRAM DESIGN
As might be expected, the level of effort required to design and
implement the groundwater monitoring program is site specific. Availability
of background data, complexity of site history and uses, variability of
subsurface conditions, site accessibility, proximity to other potential
sources of contaminants and potential water users, as well as other factors
play significant controlling roles in the final effort required. A detailed
explanation of a recommended sequence for design of a monitoring program to
comply with the new hazardous waste regulations cannot be provided within
the extent of this paper. Therefore, an outline is provided in Table 1. It
can serve as a reference planning tool during formulation of early monitor-
ing (detection type) programs. This outline recounts most of the actual
steps followed during a recent investigation by Baker Engineers for a major
steel industry client.
Several problems have been encountered during the monitoring projects
which are worth noting as they may relate to steel industry monitoring
-------
Owner Applies
for Permanent
Hazardous Waste
Facility Permit
Site is New
Site Existing and
Permitted Under
Interim Status
Regulations
Administrator May Provide
Provision in Permit to
Implement Detection and
Compliance Monitoring
Programs and Corrective
Action Program in Response
to Certain Conditions
i
Continue Operation
Without Groundwater
Monitoring
Design/Operation
Fails (e.g. Liner
Leaks)
• NO
YES
Site is Exempted From Moni-
toring Due to Design or
Operation (e.g. Double
Lined Surface Impoundment)
Implement Detection
Monitoring Program or
Correct Failure (e.g.
repair liner)
Statistical Analysis of
Indicator Parameters of
Interim Program Reveal
a Potential Contamina-
tion Problem at Down-
gradient ("Compliance")
Point(s)
Administrator Requires
Implementation of Detection
Monitoring Program and
May Provide Provision in
Permit to Implement
Compliance Monitoring
Program and Corrective
Action Program in Response
to Certain Conditions
\
• YESI
Figure 1. Standards for Owners and Operators of Hazardous Waste
Treatment, Storage and Disposal Facilities. Basic
Groundwater Monitoring and Response Program Elements.
(40 CFR, Part 264, Subpart F) Effective January 26, 1983.
Administrator Requires Imple-
mentation of Compliance Moni-
toring Program and May Provide
Provision in Permit for Correc-
tive Action Program in Response
to Certain Conditions
\
\
-------
Establish:
— Sulficient Network of
Wells to Monitor
Upgradient and Down-
gradient Conditions
— Procedures & Techniques
° Sample collection
0 Preservation & shipment
0 Analytical procedures
o chain ol cuslody
\
Determine Background
Quality at Upgradienl
WeH(s) by Quarterly
Sampling lor 1 Year
(Min. 4 samples each
lime)
Determine Groundwater
Elev. each Sampling
Run
\
Not Required Bui Should Also Con-
tinue Regular Sampling of Upgradiont
Well(s)
Site is Existing
And Permitted
Under Interim
Status Regs.
(Background Data
Available)
— Sample Downgradient
Wells at a Minimum
Semi-annually (4
replicate samples
at each well)
— Determine Groundwaler
Elev. each Sampling
Run
Annually (minimum)
Determine Groundwaler
Flow Rate and Direction
If Modification to De-
tection Program (e.g
new well) is Necessary
Submit Application to
Administrator Requesting
Required Change Within
90 Days
Each Sampling Run Deter-
mine i| Statistically
Significant Change In
Quality Has Occurred at
Downgradient ("Compliance")
Points By Comparison with
Initial Background Values
Owner Believes Change is
Attributable to Another
Source or Error in
Sampling, Analysis or
Evaluation
Well or Wells
Seemingly Affected
Are Resampled and
Analyses Repeated
- Notify Administrator Within 7
days Regarding the Change
Delected
- Indicate That Owner Intends to
Prove That Supposed Contamina-
tion Irom Facility is Unfounded
- Submit a Report Containing
Proof of Findings Within 90
Days to the Administrator
- Submit Within 90 Days a Plan
it Required of Changes to the
Detection Program
\
Note: "Actions" specified above and the respective
time periods tor completion begin in response to
the confirmation of significant change in groundwater
quality. The various times are not cumulative.
Immediate Action
— Notify Administrator Within 7 Days
of Change Detected and Details
Determine Concentrations of All
that Are Present in Groundwater
(1 sample from each well)
— Sample Upgradient Well(s) Quarterly
or more often to Determine Back-
ground Values of the Appendix VIII
Points During Screening (min. 4
samples each lime)
\
ve
e to
undwater
\
\
Action Wllhln 3 Months (90 Dayi)
Submit an Application for a Permit
Modification to Establish a Corn-
Administrator did not include a
including:
of any Appendix VIII Constituents
found at Downgradienl Wells During
— A Proposed Concentration Limit
Including either the Groundwater
Prelection Limils in Table 1, the
Background Concentralion of the
Constituent, or a Notice of Intent to
For Each Hazardous Constituent
Found at the Downgradienl Wells
— Description of any Proposed
Changes to the Monitoring
System, Sampling Frequency,
Fulfill Compliance Monitoring
Program Requirements
\
\
X,
Variance Sought For
Concentrations of
Any Constltuent(s)
\ vis
Action Wllhln 6 MOB. (1
Submit All Data to Just
Variance Sought
\ f
Owner Desires to Seek
(ration Variance For Eve
Hazardous Constituent
Wells Which Significant
Diflers From the Backg
Level or Appropriate Gr
Protection Limit in Sec
Table 1
\ 1
| _[
&-
All Hazardous Constituents
Detected al Downgradient
Wells Are Listed in Section
264 94. Table 1 anj Their
Concentrations Do Not Exceed
Ihe Respective Limils
\
\
Acllon Wllhln 6 Mo». (180 Dayi)
Submit Engineering Feasibility
Plan lor Correction Activities
Figure 2. Detection Monitoring Program Requirements.
(40 CFR, Part 264, Subpart F, Section 264.98)
Implement Compliance Monitoring
Al Direction of Administrator
and/or Other Required Action
-------
— Sample Downgradient Wetls
At a Minimum Quarterly
(4 Replicate Samples a!
Each Well)
— Analyze (or Hazardous
Constituents Specified
— Determine Groundwator
Elev. Each Sampling Run
Not Required But Should Also
Regularly Monitor Upgradi-
ent Wetl(s)
- Annually (Minimum) Analyze
Snmples Irom Downgradienl
Wells (or AM Appendix VIM Consli-
luonls (I sample from each well)
Report Concentration ol Nfw \-\f\7-
firdous Constituents lo Ailmmis
Irator Wilhrn 7 Days Admmislra-
lor May Add These Constituents to
Monitoring Program
Annually (mitiimgm) Determine
Gtoundwater Flow flnle and
Direction
It Modification lo Compliance Pro-
gram (e g new well) is Npcrssary
Submit Application to Administra-
tor Requesting Required Changs
Within 90 Days
\
Each Quarterly Sampling
Determine if a Statisti-
cally Significant
Detrimental Change in
Hazardous Constituent
Concentrations Has Occurred
at Downgradient Points
Single Point in Time Com-
parison of Upgradtent and
Downgradienl Quality where
There is a Demonstrated
High. Long-Term Correlation
Betwen Quality at Upgradi-
enl and Downgradienl Welts
I
Significant Change
Detected
Well or Wells Seem-
ingly Affected Are
Resampled and Analyses
Repeated
!
~\
Significant Change
Confirmed
Owner Believes Change is
Attributable to Another
Source or Error in Sampling,
Analysis or Evaluation
\
- Notify Administrator Within
7 Days Regarding the Change
Dectected
- Indicate That Owner Intends
to Prove that Supposed Con-
tamination from Facility is
Unfounded
- Submit a Report Containing
Proof of Findings Within
90 Days to Administrator
- Submit Within 90 Days A
Plan of Any Changes
Required to Compliance
Monitoring Program
\
— Notify Administrator of
Findings Within 7 Days
— Submit Application lo
Administrator Within 180
Days for Permit Modifica-
tion to Establish a
Corrective Action Program
(if the Administrator did
not include a provision in
original permit) Including
at a Minimum:
• detailed description of
proposed corrective action
• proposed plan for ground-
water monitoring to
demonstrate effectiveness
— Submit Above Application
Within 90 Days If An
Engineering Feasibility
Plan for Correction Acti-
vities Was Submitted During
Detection Monitoring Program
\
Implement Corrective Action
Program At Direction ot
Administrator and/or other
Required Action
\
_\
Figure 3. Compliance Monitoring Program Requirements.
(40 CFR, Part 264, Subpart F, Section 264.99)
-------
Initiate Approved Correction Program
to Remove or Treat In Place
Hazardous Constituents
Occurring at the Compliance
Point (Downgradient well
locations) Which Exceed the
Permit Concentration Limlt(s)
i
Monitor Groundwaler Quality
To Determine If the Hazardous
Constituents Exceed Specified
Concentrations at the Com-
pliance Point. (The Moni-
toring Program Must Be As
Good or Better Than The
Compliance Monitoring
Program)
1°
Monitoring and Corrective
Program Requirements
Fulfilled For Specified
Compliance period
Determine If Hazardous
Constituent Concentrations
at Compliance Boundary Have
Been Reduced to or Are Lower
Than Permit Concentration
Limit Due to Correction
Activities
— Continue/Reinstate Corrective
Action and Continue Monitoring
— Submit Application For Permit
Modification within 90 Days If
Change Is Required In Corrective
Action
Monitor Groundwater To Determine
If Hazardous Constituents Are
Present Between The Compliance
Point And the Downgradient
Facility Property Boundary
Owner's Responsibilities End
1
Monitoring and Corrective
Program Requirements Fulfilled
For Specified Compliance Period
Determine If Hazardous
Constituent Concentrations
Between Compliance Point
And Downgrndient Facility
Boundary are in Excess o(
Limit Specified in The
Permit
Implement Corrective Program
To Remove or Treat In Place
Hazardous Constituents and
Continue Monitoring
\
\
\
Determine if Hazardous
Constituent Concentrations
Between Compliance Point
Facility Property Boundary
Have Been Reduced to or Are
(ration Limit Due to Correc-
tion Activities
Monitoring Program
Requirements Fulfilled
For Specified Compliance
Period
X
V
•M
\
5
r
Report Effectiveness of the
Correction Program Semi-
Annualty
f
Discontinue Corrective Action
But Continue Monitoring
L_
— Continue/Reinstate Correction
Program and Continue Monitor-
ing
— Submit Application for Permit
Modification Within 90 Days
If Change is Required in Cor-
rective Action
\
Figure 4. Corrective Action Program Requirements.
(40 CFR, Part 264, Subpart F, Section 264.100)
-------
TABLE 1. OUTLINE FOR GROUNDWATER MONITORING PROGRAM DESIGN
I. Conduct Initial Background Investigations
A. Research published literature
1. Regional geology, soils, and probable aquifers
2. Regional groundwater quality and water well completion
records
3. Potential background sources of contamination
B. Obtain background data from site personnel/owner
1. Local drilling logs/well logs
a. underlying natural liners or permeable zones
b. approximate depth to groundwater
c. groundwater quality
2. Site history
a. historic uses
b. natural soils in-place or disturbed
c. other wastes previously disposed on or near site and
still present or removed
3. Base mapping and aerial photos
4. Anticipated disposal area extent and configuration
5. Waste characteristics (physical, chemical)
C. Field reconnaissance and mapping in immediate vicinity of disposal
site
1. Locate outcrops for geologic correlations
2. Locate existing wells and springs as potential monitoring
points
3. Map pertinent site features and identify other potential
contamination sources
4. Identify possible monitoring well locations
II. Develop Conceptual Monitoring Plan
A. Compile general geologic profile for site by correlating available
site subsurface and outcrop information
B. Identify appropriate aquifer(s) to be monitored
C. Select preliminary monitoring well locations and appropriate
depths (consider including significant springs or substituting
them for wells)
D. Identify additional data needed
E. Review with site personnel/owner
III. Conduct Subsurface Investigation(s) to Supplement Existing Data
A. Exploratory drilling
1. Drill appropriate number of exploratory borings around
existing site as deep or deeper than the uppermost aquifer
and potentially interconnected underlying aquifers and
minimum 10 feet into groundwater. For new site drill a few
similar borings on-site
2. Conduct pump tests or laboratory permeability tests on
aquifer materials and determine permeability and porosity
(required for annual evaluations of rate of groundwater flow)
3. Case exploratory borings to maintain open holes
-------
4. Fully characterize site soils through laboratory classifica-
tions
B. Determine direction of groundwater movement
1. Allow groundwater levels in exploratory borings to equalize
2. Measure groundwater levels in all exploratory borings,
existing wells, and determine surface elevation of all local
lakes, etc.
3. Contour data to define direction(s) of local groundwater
movement
C. Geophysical surveying as possible supplement or substitute for
drilling
IV. Design Permanent Monitoring Plan
A. Design and install permanent monitoring system
1. Revise preliminary geologic profile using supplementary
subsurface data
2. Revise conceptual well depths and locations as dictated by
new profile and groundwater flow pattern
3. Convert exploratory borings to permanent monitoring wells,
install additional wells, and/or develop springs to ade-
quately monitor background and downgradient conditions.
Grout exploratory holes not used for monitoring.
B. Formulate sampling and analysis procedures
1. Sampling stations (drilled monitoring wells, existing wells,
springs, related surface water samples)
2. Select parameters to be analyzed
a. general indicators of contamination
b. expected specific parameters related to waste type and
potential reaction products
3. Identify sampling frequency (at least semi-annually required)
4. Well evacuation and .sampling methods and equipment to assure
representative samples are obtained
5. Sample volumes, container types, and sample handling/
preservation to assure sample integrity
6. Field versus laboratory analyses
7. Documentation procedures for field activities
8. Sample custody control and documentation to establish a
record of proper sample handling
9. Analytical methods for desired detection level
10. Quality control
11. Resampling and possible sample splitting
C. Formulate data management procedures
1. Computer storage of data and assist in calculating statisti-
cal difference from background using student T-test
2. Graphical summaries for trend or single-point-in-time com-
parisons
3. Reporting procedures and format
V. Implement Monitoring Plan
-------
situations. Foremost is the situation where the disposal site being moni-
tored is situated on or near other waste disposal sites. In these circum-
stances, it may be nearly impossible to determine conclusively if the
disposal site being monitored is the source of detected contaminants, espe-
cially if similar waste constituents are present in both areas. Much more
work than usual may be necessary at such sites to adequately characterize
background influences so that these contributions are not wrongly attributed
to the disposal site.
A second significant problem relates to selection of monitoring well
locations and long-term protection of the wells. Downgradient, detection
type monitoring wells must be situated as close as possible to the waste
management unit to provide early warning of groundwater contamination.
Additionally, the path of probable downgradient flows from a relatively
small waste unit usually is very limited in width. Similarly, proper
location of upgradient wells may be very crucial in some instances to
properly define upgradient or background quality, especially when affected
by another localized source. As a result, suitable areas for location of
both downgradient and upgradient wells to accomplish the necessary monitor-
ing usually are very limited.
Repeatedly, owners of waste disposal sites fail to recognize the lack
of freedom in monitoring well placement without significantly compromising
the monitoring objectives. Owners and their plant personnel must dedicate
the necessary locations for monitoring points and endeavor to protect these
locations. Although monitoring wells routinely are protected at the ground
surface using steel casing, wells have been lost in the past due to inad-
vertent damage by heavy equipment, burial, and inundation.
Another subject for concern is the overseeing (quality control) of
continuing monitoring activities. The best designed monitoring program can
be severely compromised if proper evacuation of wells before sampling,
prevention of cross contamination of wells, proper sampling and preservation
of samples, laboratory quality control, and consistency are not given proper
attention. The ultimate monitoring results may be erroneous and confusing
as well as indefensible if questioned by regulatory agencies.
AREAS OF FLEXIBILITY OR NEGOTIATION
Within each of the three monitoring programs there are several areas
where the interim final regulations specify that actions must be taken, but
leave the details to be resolved by the regulated entity and EPA Regional
Administrator. Moreover, partial or entire exclusions from groundwater
monitoring can be obtained for specific types of units (piles). designs
(landfills and impoundments with double liners), operations (hazardous
constituents of waste have been effectively treated), or site locations
(hydrogeologic setting precluding potential for hazardous constituent
migration to uppermost aquifer during active life and post-closure care
period; or, possibly in rare circumstances, where the uppermost aquifer is
severely contaminated and cannot be used provided that this aquifer is
isolated from other aquifers or surface waters).
-------
In the detection and compliance monitoring programs, resolution of the
following major issues are based upon situation-specific information:
• selection of the parameters to be monitored based upon types and
quantities of wastes managed at the unit and waste constituent
concentrations, leachate reactions with underlying soils, detec-
tion procedures and accuracies, and variability of background
groundwater quality
• development of a monitoring network yielding representative
samples; the number of monitoring wells at the compliance point
are not specified and continued sampling of upgradient wells is
not mandated although it will be necessary to effectively accom-
plish the program
• specification of groundwater'protection standards comprised of a
list of hazardous constituents, concentration limits, point of
compliance, and compliance period; the numerical concentration
limit may be based upon the background concentration of a con-
stituent, Maximum Concentration Limits (MCL) for the constituents
in the Groundwater Protection Standards, or an Alternate Concen-
tration Limit (ACL) based upon a variance demonstrating no adverse
affect on human health and the environment
• method for determining compliance; based either upon single-
point-in-time comparison of upgradient and downgradient samples or
statistical comparison of downgradient samples against a set of
background samples pooled over time
• actions which can be taken when a statistically significant
detrimental change from initial background values has been de-
tected (verification of result, proof of contribution by another
source).
In the corrective action program, the regulations do not specify the
remedial measures should the groundwater protection standards be exceeded.
A performance specification rather than a design specification is imposed.
Actions to reduce hazardous constituent concentration to meet the concentra-
tion limits at the compliance point and clean-up of a contamination plume
between the compliance point and facility property boundary are required.
All of these variances or site-specific issues will be addressed and
resolved as part of the site permitting process which may be an ongoing
activity. Because final specification of requirements lies with the EPA
Regional Administrator, it is anticipated that philosophies and policies are
likely to differ in the various regional offices. This could be an advan-
tage or disadvantage. However, the regulatory approach does provide for
site specific information to be used to formulate the groundwater protection
program rather than attempting to mandate a program which effectively
addresses all likelihoods in a topic as broad ranging as hazardous waste
management. The performance specification approach allows for greater use
of technical knowledge possessed by the facility operator and acknowledges
-------
the degree-of-risk issue. It, however, must be recognized that in states
having primacy for writing facility permits the flexibilities may differ or
may not exist at all.
RISK ASSESSMENT
Operators now have a greater opportunity to influence permit conditions
for their facilities. To effectively exercise this opportunity, it is
necessary to be familiar not only with the regulatory language and tech-
niques for accomplishing groundwater monitoring, but also with methods for
developing a technical basis for permit requirements. One approach to
developing such a technical basis is performance of a risk assessment.
Assessing risk involves determining the probability of an occurrence and the
consequences of that occurrence. As it pertains to groundwater monitoring,
the focus is more towards examining the consequences once a statistically
significant difference from background levels has occurred.
Assessments can vary widely in degree of complexity and confidence of
results. The degree of refinement necessary will vary depending upon a
number of site-specific factors. In general, to explore the consequences of
detection of contamination in a monitoring well the following closely
interrelated elements must be examined:
• reactions between migrating contaminants and soils in the area
• contaminant migration direction, extent, rates, and
transformations
• hazard of migration to public health and the environment in the
pathway of migrating contaminants.
The scope of such an examination eould include the following activities:
1. A literature review to identify leachate transformation or attenua-
tion reported in similar soils for the constituents of concern.
2. A review of public agency records to identify water resources and
uses which could be affected by migrating leachate.
3. Laboratory bench scale column or sequential leaching tests or
field scale lysimeter studies to determine site specific leachate
formation and migration parameters.
4. A site water balance to estimate the quantity of leachate produced
and potential areas to control leachate production.
5. Explorations characterizing the subsurface and the aquifer and
providing the data base necessary to predict leachate movement.
6. Mathematical modeling of leachate transport and physical, chemical,
or biological transformations such as attenuation (adsorption, ion
exchange, filtration, precipitation), dispersion, and dilution.
-------
7. Modeling of transport and dispersion in surface waterways pre-
dicted to be impacted by leachate migration.
8. Literature review or laboratory testing to determine allowable
constituent concentrations in the affected water bodies that would
not preclude the intended beneficial uses.
For a new site, many of these activities should be undertaken at the
time of site design in order to arrive at a technically suitable, environ-
mentally protective, and cost-effective design.
Having assessed the potential impact on beneficial water uses and
public health, a determination can be made with regard to the implications
of detection of contamination in the monitoring wells. Risk assessment is a
powerful tool which can be used effectively to develop reasoned, technically
sound basis for permit requirements for residue disposal operations.
-------
The work described in this paper was not funded by the USEPA. The
contents do not necessarily reflect the views of the agency and no official
endorsement should be inferred.
"LANDFILLING SOLID WASTES FROM SPECIALTY STEEL PRODUCTION
ON A SOLVAY PROCESS WASTEBED"
J. A. Hagannan, Ph.D., "R. W. Klippel, P.E.
Calocerinos & Spina, Consulting Engineers
and R. H. Wills, Jr.r Crucible, Incorporated
INTRODUCTION
State and Federal regulations governing waste disposal have required
the Iron and Steel Industry to determine the hazardous nature of all
wastes, to keep close account of hazardous wastes from source to ultimate
disposal and to treat or dispose of hazardous wastes in an environmentally
safe manner. Implementation of these regulations leaves industry account-
able for past disposal practices and will make industry responsible for
adverse impacts from present or future waste disposal.
In the case of Crucible Incorporated, Specialty Metals Division in
Syracuse, New York, Department of Environmental Conservation regulations
require permitting of landfills for disposal of both hazardous and non-
hazardous solid wastes. The New York State regulations equal or exceed
Federal regulations for disposal of hazardous wastes.
Crucible has been landfilling solid wastes on an abandoned Solvay
Process wastebed adjacent to Onondaga Lake for the past nine years. As
a result of a two-year study of the chemistry, hydrogeology and structural
stability of the landfill, the continued operation of this site has been
permitted by the State as a non-hazardous solid waste management facility.
SOLVAY PROCESS WASTE
Solvay Process began producing soda ash in the Syracuse area in 1884
"because of the abundant supply of raw materials including limestone, salt
and process water. Waste solids from the production were landfilled from
1916 to 1950 in a swampy lowland adjacent to the western shore of Onondaga
Lake. The lakeside Solvay Process wastebed eventually comprised 364 acres
fil'led to a maximum height of 60 feet. To construct the wastebeds, a
waste slurry was transported to the area by pipeline, confined in sections
by hand-dug dikes and allowed to drain. Segments of the total area were
abandoned at several elevations resulting in a terraced structure sur-
rounding the central plateau which is the site of the Crucible landfill.
-------
Solvay Process waste is composed of a mix of insoluble oxides and
silicates, alkaline hydroxides which buffer the groundwater at pH 11, car-
bonates and soluble chloride salts which produce high salinity in the
wastebed groundwater.* The silty wastes, when drained and settled, assume
a moist, cake-like consistency with numerous cemented layers containing
flyash. Salt content of the wastebed surface decreases by leaching several
years after placement allowing development of a sparse but diverse vegeta-
tion including alkali-tolerant weeds, grasses and woody species, predom-
inantly willows and poplars. The waste remains nutrient-deficient, largely
in phosphorus. Fertilization results in marked increases in density of
vegetative cover.
CRUCIBLE SOLID WASTES
The Crucible landfill presently occupies 17 acres of the lakeside
Solvay Process wastebed with a maximum fill height of 5 feet. Types of
waste landfilled, average annual volumes and results of Extraction Pro-
cedure Toxicity testing are shown in Table 1.
•
TABLE 1. CRUCIBLE SOLID WASTES
Specific Waste
Annual
Quantity
cu. yd. Classification
Hazardous
Parameters
Current
Disposition
Currently Landfilled Wastes
Slag
Construction Debris &
. Refractories
Boiler House Ashes
Coolant Swan5
Mill Scale
WWTP Sludge
Total
Previously Landfilled Wastes
Grinding Dusts
- Air Pollution &AOD Dusts
Waste Caustic Solids
Acid Pickling Solids
Total •
6.290
4,104
1,437
1.375
1,121
165
14,492
688
1.176
200
50^
2,114
Non-Hazardous
Non-Hazardous
Hazardous
Hazardous
Hazardous
None
None
Chromium &
Cadmium
Chromium
Chromium
Landfill
Recycled & Reused
Recycled & Reused
CECOS Int'l
On-Site Treatment
C.S. Grove, Jr., et al, The Rehabilitation of Solvay Process Wastebeds,
Bureau of Solid Waste Management, Environmental Control Administration,
Public Health Service, Department of Health, Education and Welfare,
Grant Number 5-R01-U1-00537-02, Dec. 1, 1969.
-------
Slag, boiler house ash, swarf, scale and wastewater treatment plant sludge
are non-hazardous wastes which comprise 63% of the total landfilling vol-
ume. Slag is a mineral residue generated during melting of scrap metal
and added alloys in electric arc furnaces. Boiler house ash is a mixture
of flyash and bottom ash produced by combustion of coal for generation of
steam. Metal finishing processes, such as grinding, 'produce a swarf con-
sisting of abrasive and metallic particles and waste coolant. Scale is a
metallic oxide coating which develops on metal surfaces during heating and
cooling and is removed by machining. Wastewater treatment plant sludge is
coagulated and dewatered solids removed from wastewater by a two-stage,
in-plant treatment process. Air pollution dust is collected by baghouses
from both the Electric Arc Furnaces (EAF) and an Argon-Oxygen Decarburi-
zation (AOD) Vessel. This material is hazardous due largely to the con-
tent of soluble, toxic hexavalent Chromium. In addition to testing by
the Extraction Procedure Toxicity Test, leaching studies by both column
and successive batch procedures give a hexavalent Chromium content of
approximately 430 mg. per kg., dry dust. Because of the toxic leaching
potential of air pollution dusts landfilled from 1973 in the case of EAF
dust and from 1977 for AOD dust, a geotechnical and chemical study of the
Crucible landfill was initiated. The general objective of this study was
to determine the impact of the Crucible landfill on groundwater quality
in the area. Specifically, the fate of soluble, toxic hexavalent Chromium
contained in air pollution dusts was to be determined. In addition, land-
filling of air pollution dust was discontinued in March of 1982. Dusts
are presently being recycled for recovery of Chromium and Nickel.
GEOTECHNICAL AND CHEMICAL STUDIES
The geotechnical study was performed by Thomsen Associates and their
affiliate Empire Soils Investigations, Incorporated. The objectives of
this study were to:
sample deposits in the area including landfilled Crucible wastes,
Solvay Process waste and underlying natural deposits,
- define subsurface conditions and the hydrogeolic regime at the
site,
- install a groundwater monitoring system to determine impact of
the Crucible landfill on groundwater quality and
- evaluate stability and settlement characteristics of the lake-
side wastebed relative to the Crucible landfilling activities.
In order to accomplish these objectives, 43 borings were advanced in-
to the wastebed using a machine-driven 3-1/4 inch hollow-stem auger. Soil
samples .were retrieved from 25 of these borings using a split-barrel sam-
pler. Soil samples were photographed and retained for classification arid
chemical analysis. While the majority of these borings penetrated Solvay
Process wastes, three were advanced into natural deposits under the waste-
bed with a maximum depth of 167 feet. Deposits underlying Solvay Process
waste generally consist of swamp deposits, lacustrine silts and alluvio-
-------
deltaic deposits over,glaciolacustrine silts and sand. Depths and types
of deposits vary with location, but beneath the Crucible site organic de-
posits average 11 to 15 feet in thickness.
Boreholes were converted into multiple piezometers* 2 inch monitoring
wells or lysimeters to obtain groundwater surface measurements and water
quality samples. Detail of groundwater monitors is shown on Figure 1.
0 -
10 -
20 -
30 -
40-
50 _
60 -
Well Lysimeters Piezometers
3V*" Borehole
%" PVC Risers
1- PVC Riser
Lysimeter
Sandpack
2" PVC Riser
Bentonite Seal
Well Screen
Swamp and Lacustrine Deposits
Crucible Waste
Solvay Process
Waste
Water
' Table"
Solvay Process
Waste
Piezometer
FIGURE 1. DETAIL OF GROUNDWATER MONITORS
-------
Locations of these monitors are shown in Figure 2.
Onondaga
FIGURE 2. GROUNDWATER MONITORING LOCATIONS
-------
A total of 62 monitors were utilized to investigate the groundwater flow
regime. A large groundwater mound was found to exist in the Solvay Process
waste beneath the Crucible landfill. Results of a series of 30 field and
laboratory hydraulic conductivity measurements on the various geologic
units at the site are consistent with a pattern of slow vertical movement
of .groundwater through Solvay Process waste and the swamp/lacustrine layer
into underlying deposits. In underlying deposits, downward movement is
restricted and groundwater flows laterally to surface waters of Nine Mile
Creek and Onondaga Lake. Flow rates in the Solvay Process waste are cal-
culated to be approximately 30 ft./yr., in the swamp/lacustrine deposits
less than 10 ft./yr. and in underlying deposits as high as 310 ft./yr.
A three-dimensional mathematical model of the groundwater flow regime
at the site was developed using a United States Geological Survey computer
program. The model simulates groundwater recharge and movement using
estimates of infiltration from rainfall and snowmelt and hydraulic con-
ductivity from field and laboratory measurements. The calibrated model
closely correlates with groundwater surface data obtained from the monitor
system and predicts 23 inches of infiltration annually through the Cru-
cible waste and 9 inches through the Solvay waste. These infiltration
amounts produce a total annual leachate volume for the Crucible waste of
approximately 10 million gallons. This leachate is diluted by approxi-
mately 30 million gallons of groundwater in underlying alluvial and
glaciolacustrine deposits prior to discharge to surface waters.
Analyses of groundwater samples from the 41 monitoring wells (Figure
2) over the past two years show high pH (11) and dissolved solids (mostly
Sodium, Calcium and Chloride) within the Solvay Process waste due to the
alkaline and saline nature of the waste. The Chromium concentration in
groundwater samples is consistently below the detection limit of the Atomic
Absorption technique (.01 mg./L.) which is below State groundwater stand-
ards (.05 mg./L.). The only groundwater influence attributible to the
Crucible landfill is a low level of Iron and Manganese typically below the
State groundwater standard of .3 mg./L.
During the late winter/early spring snowmelt period, numerous seeps
develop near the base of the Solvay Process wastebed as the groundwater
mound rises due to high recharge. A series of five seepage galleries were
installed in areas of heavy seep flows as part of the geotechnical project.
Location of these monitors is shown in Figure 3. Analyses of these five
seeps, as well as other smaller seepage areas, also show Chromium concen-
trations below the detection limit of .01 mg./L. From these analyses of
well and seep samples, it is evident that the soluble hexavalent Chromium
contained in landfilled air pollution dusts has not reached the groundwater
regime in detectible concentrations in the nine years of operation of the
landfill.
In order to determine the fate of this potential pollutant, a series
of four lysimeters were installed in an area of the Crucible landfill known
to contain fresh (-^ 1 year) air pollution dust. This location is shown
as Ly 201 on Figure 2. The lysimeters were installed at six-foot intervals
-------
FIGURE 3. SEEPAGE GALLERY LOCATIONS
in the unsaturated zone of the. Solvay Process waste directly beneath the
fresh air pollution dust. A vacuum is placed on the sampler overnight to
draw pore water into the sampler. The sample is then retrived by pressure
the following day. Analyses of samples from this monitor system showed high
concentrations of hexavalent Chromium (approximately 1,000 mg./L.) in the
surface lysimeter directly under the Crucible landfill but Chromium below
detection limits in deeper lysimeters.
A series of twelve test pits were also dug with a backhoe around the
edge of the Crucible landfill both in areas of air pollution dust deposits
and in areas with no dusts. The location of these pits is shown in Figure 4.
-------
CRUCIBLE
FILL
FIGURE 4. TEST PIT LOCATIONS
In each case where air pollution dust was landfilled, a yellow stain was
observed in the surface of the Solvay Process waste immediately under the
dust. Analyses of this yellow stain showed high levels (to 500 ppm) of
hexavalent Chromium adsorbed to the Solvay Process waste. Comparison of
the extent of migration of the stain to the age of the dust deposits re-
veals that the soluble hexavalent Chromium is leached from the air pollu-
tion dust within one year of landfilling and is contained in the top six
feet of Solvay Process waste within three years. No further migration of
the Chromium is evident up to five years after landfilling. The complete
monitoring system installed in the fresh air pollution dust deposits is
shown in Figure 5.
-------
0 -
10 -
20 -
30 -
40-
50 J
CO-
' 1
— '• ' •
I
• _ -
•: • •'•;' f1
.',:'•''. $jjj$$l&Mig£
TT ' J
u Extent of
Test Pit {
'
.
Crucible Waste
Solvay Process
Wasle
-Conlaminent
Enclave
Water
Table
Splvay Process
Waste
I
Swamp and Lacustrine Deposits
FIGURE 5. FRESH AIR POLLUTION DUST MONITORING SYSTEM
The partitioning of hexavalent Chromium between groundwater and Solvay
Process waste has been studied in our laboratory by measuring the adsorption
isotherm for Solvay Process waste samples taken from the surface of the
wastebed, the bottom of a test pit and at a depth of 23 feet. In all cases,
a linear isotherm was obtained with distribution coefficients of 2.9 ml./g.
for the surface sample, 5.2 ml./g. for the test pit sample and 7.4 ml./g.
for the sample from 23 feet in the wastebed. The adsorption maximum of
5,300 mg./kg. for the deep Solvay Process waste may be compared to values
of 670 mg./kg. for montmorillonite clay and 190 mg./kg. for kaolinite clay
as determined by Griff en in similar tests.* The stains found in the test
Griff en et al, Effect of pH on Adsorption of Chromium From Landfill-
Leachate by Clay Minerals, J. Environ. Sci. Health, A12(8), 431-449 (1977)
-------
pits are examples of "contaminant enclaves" which are stable with finite
pollutant sources such as the air pollution dust deposits. Solvay Process
waste immobilizes the hexavalent Chromium much more efficiently than an
equivalent layer of clay typically used as a landfill liner.
Analysis of metal content in samples of Solvay Process waste taken from
borings at the landfill reveals low levels of reduced, precipitated non-
toxic Chromium in the vicinity of the groundwater table beneath the Crucible
landfill. No Chromium was found in samples taken from borings off the
Crucible landfill. The samples containing Chromium also showed high levels
of Iron relative to the rest of the Solvay Process waste samples taken both
under and off the Crucible landfill. Beyond the contaminant enclaves iden-
tified in test pits an extensive zone of very low concentrations of soluble
hexavalent Chromium could slowly transfer very low levels of Chromium from
the enclaves to the water table. This transfer would be accentuated during
late winter and early spring when high recharge of the aquifer occurs. At
the water table, very low concentrations of hexavalent Chromium could be re-
duced and precipitated by redox coupling with soluble Iron in the ground-
water.
The impact of leachate from the Crucible landfill on the waters of
Onondaga Lake has been estimated by calculation of daily groundwater loads
of trace metals. As a percentage of total loads to the Lake, groundwater
loads from the Crucible landfill are negligible.
Analysis of the strength and structural stability of the Solvay Process
wastebed relative to the Crucible landfill included in situ vane shear and
torvane tests, laboratory unconfined compression tests and consolidation
tests. Slope stability analyses indicate an acceptible factor of safety for
a 10 foot lift of Crucible waste with an edge setback of 100 feet. Maximum
settlement with a 10 foot Crucible lift was estimated to be 5 feet resulting
in a dish-shaped surface which will promote containment of surface runoff
and downward percolatioa of infiltrated water.
PRESENT STATUS"AND FUTURE OPERATING PLANS
Crucible is presently operating the landfill under a five-year permit
from the New York State Department of Environmental Conservation as a non-
hazardous solid waste disposal facility. No air pollution dust or other
hazardous wastes will be landfilled at the site.
As part of the operation permit, a quarterly landfill inspection and
"groundwater sampling and analysis program have been developed. Water sam-
ples from selected wells, seeps and lysimeters will be closely monitored
to'insure that hexavalent Chromium concentrations at the water table do not
exceed State groundwater standards and to further document the secondary
redox control of soluble Chromium by Iron. Neither daily nor intermediate
cover will be required prior to closure of the present site. A maximum
lift of ten feet of waste is permitted, producing a lifetime for the present
landfill of ten years. Details of closure of the present landfill and re-
quirements for liner, leachate collection and cover for a future expansion
area are presently under deliberation by the State.
-------
MILL SCALE DE-OILING BY CRITICAL-FLUID EXTRACTION
By: Thomas J. Cody, Jr.
Richard P. de Filippi
Christopher P. Eppig
CRITICAL FLUID SYSTEMS, INC.
An Arthur D. Little Company
25 Acorn Park
Cambridge, Massachusetts
Presented at the Symposium on Iron and Steel Pollution Abatement Technology
NOVEMBER 16, 1982
-------
ABSTRACT
A process for de-oiling mill scale with critical-fluid solvents (lique-
fied gases or supercritical fluids) has been successful in, bench-scale
tests. A process design and evaluation has been completed, and preliminary
economic estimates for both capital and operating costs appear promising.
Additional testing for final process definition is expected to .be concluded
in several months.
-------
The work described in this paper was not funded by the U. S. Environ-
mental Protection Agency. The contents do not necessarily reflect the
views of the Agency and no official endorsement should be inferred.
MILL SCALE DE-OILING BY CRITICAL-FLUID EXTRACTION
MILL-SCALE DE-OILING
INCENTIVES FOR DE-OILING
In steel mills, mill scale amounts to 4% to 5% of the total raw steel
production. Its oil content is varied, but can run as high as several
percent for the composite scale for a total mill, and much higher for certain
specific operations. Chemically, the solids are largely iron oxides, and
the oils can be mineral oils or fats and fatty oils typically used as lubri-
cants in rolling mills.
A 1978 study showed that about half the mill scale produced is re-
cycled to sinter plants for return to the blast furnace;.the other half is
stockpiled. The same study showed that about 11% of all lubricants purchased
by steel mills ultimately becomes the oil fraction of mill scale.
(2)
Coarse mill scale (+5mm) has a low oil content , and may be recycled
as a direct metallic charge to the blast furnace. Mill scale fines (-5mm),
which are 80% to 85% of the total mill scale, have an oil content sufficiently
high to adversely effect sinter plant operations. The oil is volatilized
during sintering and recondenses in the sinter-plant off-gas. Without treat-
ment, the gaseous effluent can produce-a visible plume; where bag houses are
used, the oil can condense and impair the operation.
(2)
Tests have shown that the maximum tolerable content of oily mill
scale in the sinter-plant charge is about 7%. In facilities where de-oiling
of mill-scale fines is carried out, the sinter-plant charge can accept at
least 25-30% of this material as feed. Therefore, major gains in mill-scale
recovery can be made with de-oiling, coupled with the elimination of very
substantial quantities of solid wastes.
DE-OILING METHODS
Several de-oiling methods have been tested, and some have been put into
commercial use.
Water washing using hot alkaline solution can remove the majority of
oil on mill-scale fines. The method has the advantage of low cost. However,
at best, 5% to 10% of the initial oil remains on the mill scale.
-------
Thermal incineration methods have been tested, and a direct-fired kiln
has been reported in commercial use(2). The principal advantage of incin-
eration is high percentage removal of oil. The major disadvantage is high
fuel cost, which includes treatment of the kiln off-gas by after-burning.
Consideration has been given to the use of liquid solvent washing, in
which a chlorinated hydrocarbon solvent is used to extract the oil from mill-
scale fines, and then distilled to recover the solvent. While no details are
available, this method would seem to have the advantage of recovering the oil
fraction for use as fuel, or even for recycle as a lubricant. Energy costs
for solvent distillation may be high, and residual solvent on the de-oiled
fines would have to be evaluated.
A new solvent washing process is in the final stages of development,
using condensed gases (critical fluids) as the solvent. This process appears
to offer the advantage of oil recovery, low-energy costs, and solvent-free
de-oiled mill-scale fines. Process information is given in this paper.
CRITICAL-FLUID EXTRACTION
Critical-fluid extraction refers to the use of fluids which are gases
at ambient temperature and pressure, but which become good solvents when
compressed to liquids (i.e., at pressures below the critical pressure) or
supercritical fluids (at pressures above the critical pressure). Those
regimes are shown in Figure 1, which is a generalized reduced pressure-
temperature-density map. Supercritical fluids (SCF) are in the range of
reduced pressures (actual pressured divided by critical pressure) of 1.0 to
5.0, and reduced temperatures of 1.0 to about 1.4. Much above these pressures,
costs become uneconomical because of equipment and compression requirements;
at temperatures above this range, densities become low, and attendant solu-
bilities of solutes in supercritical fluids decrease drastically.
A near-critical liquid (NCL) refers to conditions near but below the
critical temperature, and a range of pressures around the critical pressure.
Generally, there is a continuity of properties within the complete super-
critical and near-critical liquid regimes, with no abrupt changes occurring
across the critical pressure or temperature.
In these regimes, fluids possess high solubilities much like liquid
solvents because of high densities (specific gravities of 0.2-0.7). In
addition, kinetic properties such as diffusion coefficient and viscosity are
quite favorable: diffusion coefficients are higher and viscosities are lower
than for ordinary liquids. Some generalized properties are summarized in
Table I.
Some fluids which have been described in the literature as effective
critical-fluid solvents are listed in Table II. Carbon dioxide has been
considered in many cases because of its low cost and non-hazardous chemical
-------
0.1
Reduced Density
Figure 1. Reduced Pressure-Density Diagram. Supercritical
Fluid (SCF) and Near Critical Liquid (NCL) Regions,
as Indicated
-------
TABLE I. COMPARISON OF CRITICAL-FLUID PROPERTIES
LIQUID SUPERCRITICAL GAS
DENSITY, g/cra3 1.0 0.2-0.7 0.001
VISCOSITY, CENTIPOISES 0.5-1.0 0.05-0.10 0.01
DIFFUSIVITY, cm2/sec 10~5 10~4 - 10~3 lO"1
-------
TABLE II. CRITICAL PROPERTIES FOR SELECTED FLUIDS
FLUID
Ethane
Propane
Butane
Pentane
Ethylene
Ammonia
Carbon Dioxide
Sulfur Dioxide
Water
T °C
32.3
96.9
152.0
296.7
9.9
132.4
31.1
157.6
374.3
P ,atm
48.8
42.6
38.0
33.8
51.2
112.8
73.8
78.8
221.1
3
P . ,g/cm
t—
0.203
0.220
0.228
0.232
0.227
0.235
0.468
0.525
0.326
-------
nature. Hydrocarbons have also been considered, often because of superior
solubilities for organic materials. Another group, not shown, includes the
halocarbons, such as chlorofluoro methanes and ethanes.
EXTRACTION OF OILY SOLID WASTE
BENCH-SCALE EVALUATIONS
A series of bench-scale experimental evaluations has been carried out
on several different oily solid wastes. In each case, the objective was to
extract petroleum-based oils or triglycerides from mineral solids, to pro-
duce an essentially oil-free solid and to recover the oil phase.
The bench-scale equipment used for testing extractions is shown in
Figure 2. Oil-laden solids are charged to an extraction vessel, typically
holding 0.25 to 0.5 Ibs of material. Liquefied gas or supercritical fluid
is pumped up to extraction pressures, which range from 200 psia to 2500 psia
depending on the fluid employed. Temperatures are typically near ambient.
In the experimental unit, solvent flow is once through, and the extract
leaving the extraction vessel is let down in pressure to atmospheric, at which
point the oil separates out as a liquid phase. Quantity of gas solvent and
oil are measured to determine extract concentrations. Flow rate, temperature,
and pressure can be varied to assess the extraction kinetics.
Table III summarizes data from two other separations of oils from mineral
solids: triglycerides from clay solids, and diesel oil from oil-well drill
cuttings. Both feeds were extracted with halocarbon and hydrocarbon solvents,
and with carbon dioxide.
Triglycerides are much more readily extracted with the halocarbon/hydro-
carbon solvents than with C02. High C02 pressures are required for it to be
an effective solvent, and relatively large solvent quantities are required.
Diesel oil is more easily extracted with each solvent, compared to triglycer-
ides.
Table IV gives a summary of a series of runs made from de-oiling mill
scales. Two different mill-scale samples were obtained from stockpiled
material: about 3.7% oil, and 1.0% oil. The respective water contents were
approximately 0.8% and 1.3%.
Results are shown for hydrocarbon and halocarbon solvents, which gave
similar extractibility, and for carbon dioxide. The hydrocarbon and halo-
carbon extractions were carried out at about 700 psia; this high pressure
was maintained to assure that the solvent would remain condensed in
the experimental system, where pressure fluctuations occur due to pump
cycling, and where the solution is sometimes heated so that there will be no
oil precipitation in the small-diameter lines and valves in the laboratory
set-up. Actual extractions may be carried but at pressures on the order of
-------
L J
Gas Supply
Heater
Q
Compressor
1
Extraction
Column
Expansion
Valve
Flow
Meter
Collector
Vent
0
Dry Test
Meter
Figure 2. Benchscale Test Equipment
-------
TABLE III. DE-OILING OF MINERAL SOLIDS
FEED
WEIGHT %
OIL
SOLVENT
PRESSURE, TEMP, SOLVENT/FEED
Psia °F RATIO
OIL EXTRACTED,
WT % OF FEED
Triglycerides.
from clay
39.6 Halocarbon/
Hydrocarbon
150
74
8.4
40.4
Triglycerides
from clay
31.4
Carbon Dioxide 2000
68
164
27.7
Diesel oil from
crushed drill
cuttings
10.8 Halocarbon/
Hydrocarbon
850
140
2.8
10.6
Diesel oil from
crushed drill
cuttings
9.9
Carbon Dioxide 2150
99
3.7
9.9
-------
TABLE IV. MILL-SCALE DE-OILING BENCH-SCALE TESTS
SOLVENT
PRESSURE,
Psia
TEMP,
op
SOLVENT/FEED
RATIO
SAMPLE WT %
OIL WATER
% OF FEED
EXTRACTED AS:
OIL WATER
Hydrocarbon/
Halocarbon
700
75
0.32
3.7 0.8
3.7 0.04
Hydrocarbon/
Halocarbon
700
75
1.4
1.0 1.3
1.0 0.12
Carbon Dioxide 2200
120
17
1.0 1.3
0.9 1.3
-------
200 psia. Solvent-to-feed weight ratio between 0.3 and 1.4 were used, and
in the case of both mill scale samples, essentially complete oil removal was
achieved. Because of the low solubility of water in the solvent, water re-
moval was incomplete, and the residual mill scale retained some moisture.
Carbon dioxide was run at 2200 psi and 120 F, with a solvent/feed ratio
of 17. Even at this high solvent use, there was incomplete oil removal,
although the sample was completely dried of water. The comparative oil ex-
tractions are typical, in that carbon dioxide generally has a lower solubility
fpr triglycerides and mineral oils than the hydrocarbon and halocarbon
solvents.
PROCESS DESIGN AND ECONOMICS
Figure 3 shows an overall flow diagram for the process. The process
system contains two parallel extractor columns, which are used alternately:
one being in the solvent circulation loop for mill-scale de-oiling, and the
other being discharged of de-oiled solid, and recharged with oily solids.
The solvent circulation is continuous, while the solids are handled in
batches alternating- between the two vessels.
Solids may be charged to the extractors in a slurry with additional oil,
most of which is displaced from the extractor with the initial introduction
of solvent. The slurry oil is taken from a portion of the extracted oil and
recycled to a slurry tank.
Solvent is separated from oil and recycled to the extractor. Separation
is carried out by vapor-recompression evaporation, which greatly reduces the
energy requirement for the system, and is facilitated by the very high vola-
tility of the solvent. In this way, oil can be stripped to very low levels
of residual solvent, and the extracted solids are solvent-free.
Table V summarizes estimated process economics. The base case evaluated
is for a system handling 25,000 Ibs/hr (11.4 long tons/hr) of oily mill scale,
with feed oil content up to 5%. The principal operating cost is labor, since
energy and solvent make-up costs are low. The estimated total processing
cost, including capital-related costs is about $6/long ton.
CONCLUSIONS
Bench-scale experimental evaluations and engineering and economic studies
have shown that mill scale can be de-oiled with a critical-fluid solvent
extraction process. Mill scales with oil contents from 1-4% have been tested,
and operating economics appear favorable.
Further bench-scale testing is currently in progress, which we expect
to conclude by January, 1983. We anticipate that process design and economics
will be firmed up shortly thereafter.
-------
SOLVENT RECYCLE
MILL-SCALE FINES
DE-OILED
MILL-SCALE
FINES
SLURRY-OIL
TRANSFER
PUMP
FLOW DIAGRAM
Figure 3. C.F. s. SOLVENT EXTRACTION SYSTEM
— -
MILL-SCALE DE-OILING
CRITICAL FLUID SYSTEMS, INC.
-------
TABLE V. MILL-SCALE DE-OILING WITH CRITICAL-FLUID EXTEACTION
PRELIMINARY ESTIMATES OF PROCESS COSTS
CAPACITY: 11.4 long tons/hr (25,000 Ibs/hr)
8000 hrs/year
CAPITAL COST: $1,300,000
$/TON
OPERATING COSTS:
POWER: 300 HP 0.99
MAKEUP SOLVENT: 7.5 Ibs/hr 0.46
LABOR + OVERHEAD: 1 man/shift 1.76
3.21
CAPITAL RELATED:
DEPRECIATION: 10% of capital 1.43
MAINTENANCE, TAXES, INSURANCE: 1.43
10% of capital
2.86
TOTAL 6.07
-------
REFERENCES
Serne, J. C., and K. Wilson, "The Use and Fate of Lubricants,
Oils, Greases, and Hydraulic Fluids in the Iron and Steel
Industry", EPA Report No. EPA-600/2-78-101, May, 1978.
^Balajee, S.R., "De-.Oiling and Utilization of Mill Scale",
presented at the First EPA Symposium on Iron and Steel
Pollution Abatement Technology, Chicago, Nov., 1979.
-------
Although the research described in this paper has been funded wholly or in
part by the U.S. Environmental Protection Agency, it has not been subjected to
Agency review and therefore does not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
RECOVERY OF METALLIC VALUES FROM
ELECTRIC ARC FURNACE STEELMAKING DUSTS
by: E. Radha Krishnan and William F. Kemner
PEDCo Environmental, Inc.
11499 Chester Road
Cincinnati, Ohio
for presentation to
Symposium on Iron and Steel Pollution Abatement Technology for 1982
Pittsburgh, Pennsylvania
November 16-18, 1982
ABSTRACT
Dust generated from electric arc furnaces employed in steel plants is
currently listed as a hazardous waste. Disposal of the waste at a controlled
landfill is becoming costly as disposal sites become scarcer and more distant
from the point of origin. The dust, however, represents a potential source of
metals such as iron, zinc, lead, chromium, nickel and molybdenum; recovery or
recycling of these metals appears to be a logical alternative to the disposal
problem. This paper presents a technical and economic assessment of various
technologies available for recovery of metallic values from electric arc
furnace (EAF) steelmaking dust.
The type of recovery process which can be employed depends on the chem-
ical composition of the EAF dust, which in turn is dependent on the type of
steel product being made. Carbon and low alloy steel EAF dusts are rich in
zinc and lead, while dusts from stainless and specialty alloy steels contain
significant quantities of chromium, nickel, and molybdenum. Five recovery
processes are discussed in this paper: 1) a caustic leach electrolytic zinc
recovery process studied in a pilot plant by AMAX Base Metals Research and
Development, Inc., in Carteret, New Jersey; 2) the Waelz kiln practice for
processing EAF dusts which is being commercially employed by the New Jersey
Zinc Company. Inc., in Palmerton, Pennsylvania; 3) an electrothermic shaft
furnace process using plasma heat piloted by SKF Steel Engineering in Sweeden
and currently being commercialized; 4) a commercial pyrometallurgical smelting
technique developed by Inmetco in Ellwood City, Pennsylvania; and 5) a re-
cycling process perfected by the U.S. Bureau of Mines in coordination with
Joslyn Stainless Steel Division in Fort Wayne, Indiana.
The AMAX and New Jersey Zinc processes are applicable to recovery of
heavy metals, such as zinc and lead, from carbon steel EAF dusts; the SKF
process can be applied to both carbon steel and specialty steel dusts to
-------
recover metals such as iron, zinc, lead, chromium and nickel; the Inmetco and
Bureau of Mines processes are suited to recovery of iron and alloying elements
from stainless steel EAF dusts.
INTRODUCTION
Each year the steelmaking industry produces more than 1.8 x 10° metric
tons of dusts containing approximately 1 x 106 metric tons of iron, 90,000
metric tons of zinc, 9,000 metric tons of lead, 2000 metric tons of chromium,
500 metric tons of nickel, and 100 metric tons of molybdenum,1'2 These
dusts originate from different types of steel furnaces such as open hearths
(OH) , basic oxygen furnaces (EOF), and electric arc furnaces (EAF) .• Although
the amount of dust generated from the furnaces at any one steel plant is
relatively small, the cumulative quantity is large and represents a signifi-
cant loss of heavy metals and scarce alloying elements.
In the electric arc furnace, approximately 1 to 2 percent of each charge
is converted to dust or fume, which is collected in scrubbers or baghouses.
The chemical composition of the dust depends on the type of steel product
being made, as shown in Table 1. Carbon steel EAF dusts tend to be richer in
zinc and lead because of the greater use of galvanized and other coated
products in the melt.1* On the other hand, alloying elements such as chromium,
nickel and molybdenum are prominent in dusts from stainless steel and
specialty alloy furnaces. Electric furnaces account for about 25 percent of
domestic steel production, and the total zinc in EAF dust generated in the
United States is estimated to be around 75,000 metric tons per year.5
EAF dust is currently listed as a hazardous waste (Federal Register, Vol.
45, No. 98, Section 261.32, page 33124) because of the leachability of its
toxic constituents (Pb, Cd, and Cr 6). Its EPA-assigned hazardous waste
number is K061. As a hazardous waste, the dust must be encapsulated or
transported to a controlled landfill. Disposal is, however, becoming a costly
affair as disposal sites become scarcer and more distant from the point of
origin. Disposal costs of $50 to $100 per metric ton of dust are quite
common. This has led the steelmaking industry to look to other viable options
for handling the dust.
This dust represents a potential source of valuable elements, the re-
covery of which appears to be a logical alternative to the disposal problem.
This can also augment our domestic supplies of metals such as zinc, chromium,
and nickel, which are largely imported at this time. Figure 1 presents an
outline of different recovery options for EAF dust. Recycling to the furnaces
for possible additional iron recovery could be practiced if the zinc and lead
contents of the dust are fairly low. Zinc and lead concentrations are some-
what high, however, especially in carbon steel EAF dusts, and their levels
could build up in the furnace. At high concentrations, these metals (as
oxides or ferrites) have been found to attack refractories and cause rapid
deterioration of linings in blast furnaces5; however, this effect has not been
proven in electric arc furnaces. This problem does not arise for stainless
steel EAF dusts because of the lower concentrations of zinc and lead.
-------
TABLE 1. CHEMICAL COMPOSITION OF DUST FROM ELECTRIC ARC FURNACES1*
Carbon and low-alloy steels
Element
Al
Ca
Cd
Cr
Cu
K
Mg
Mn
Mo
Na
Ni
Pb
Zn
Fe (total)
Fe+3
Fe+2
Fe°
Cr+6
Si
Cl
F
P
so"
Wt, %*
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
0.70
Range , %
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
0.01-0.88
Stainless steel and specialty alloys
Element
Al
Ca
Cd
Cr
Cu
K
Mg
Mn
Mo
Na
Ni
Pb
Zn
Fe (total)
Fe+2
Cr+6
Si
Cl
F
P
Wt, %**
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
*
Arithmetic
averages
averages
of dust from
of dust from
seven plants.
four plants
-------
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
L
o Electrowinning
o Blast Furnace
o Slag Fuming
o Electrothermal
•
Non-Ferrous
Recovery
Zn, Pb, Cr
Concentrate
1
Zn, Pb, Cr
Product
Figure 1. Options for recovery of resources from electric arc furnace (EAF) dust,
4
-------
The recovery of nonferrous elements, e.g., zinc, lead, chromium, from EAF
dust is another possible alternative. Such recovery could result in a residue
that would either have been detoxified for disposal in a landfill or recycl-
able to the furnace for recovery of iron. When the zinc and lead levels in
the dust are not high enough to be of interest, recycling to the furnaces
could possibly concentrate the nonferrous elements to a sufficient degree to
make the dust attractive for subsequent recovery of these heavy metals. Zinc
contents of greater than 15 percent in the dust are generally considered
viable for zinc recovery.
The primary operations in nonferrous metals production are ore concentra-
tion and smelting. Ore concentration techniques such as flotation are not
feasible for treating EAF dust because of the limitations posed by the fine
particle size of the dust. Individual particles are generally in the range of
0.1 urn to 1.0 ym.1* For carbon steel EAF dusts, smelting for zinc recovery is
possible with lead being recovered as a byproduct of the smelting operation.
The presence of tramp elements such as iron, cadmium, and chlorine in the EAF
dust, however, can interfere with the operations encountered in conventional
smelting processes. Hence, recovery of nonferrous values from EAF dust re-
quires a separate or add-on smelting facility dedicated to recovery of metals
from steelmaking dusts and similar materials. Because limited quantities of
steelmaking dust may be generated at any particular plant, the economics will
generally require that the recovery process be used on a mixture of dusts from
various sources.
This paper discusses five promising technologies that appear to be
feasible for recovery of metallic values from EAF dusts. Two of the tech-
nologies are applicable to carbon steel dusts, two are suited to stainless and
specialty alloy dusts, and one can be applied to dusts produced in the manu-
facture of both types of steels. Figure 2 presents an overview of the proc-
esses.
ASSESSMENT OF PROMISING RECOVERY TECHNOLOGIES
The developers of the five promising technologies (shown in Figure 2)
have had some involvement (either at the pilot plant or commercial scale) in
recovery of metals from EAF dust or materials similar in composition to the
dust. AMAX Base Metals Research and Development, Inc. has investigated a
caustic leach electrolytic zinc recovery process in a pilot plant in Carteret,
New Jersey; the Waelz kiln practice for processing EAF dusts is being used
commercially by the New Jersey Zinc Company, Inc., in Palmerton, Pennsylvania;
whereas an electrothermic shaft furnace process using plasma heat has been
piloted by SKF Steel Engineering in Sweden, and they are scheduled to place a
commercial plant into operation in Landskrona, Sweden, by late 1984. The
Inmetco process is currently being applied commercially in the Ellwood City,
Pennsylvania, plant on wastes from stainless steel mills; whereas the Bureau
of Mines process has been successfully demonstrated in large-scale industrial
tests conducted in coordination with Joslyn Stainless Steels in Fort Wayne,
Indiana.
-------
CARBON STEEL
EAF DUST
CARBON STEEL
EAF DUST
AMAX
PROCESS
ZINC
LEAD
t
IRON RESIDUE (NONHAZARDOUS)
IRON
NEW JERSEY
ZINC PROCESS
IMPURE ZINC OXIDE
IRON RESIDUE (NONHAZARDOUS)
CARBON STEEL OR
SPECIALTY STEEL
EAF DUST
SLAG
SKF
PROCESS
ZINC
LEAD
ALLOYED HOT METAL
4.
5.
SPECIALTY STEEL
EAF DUST
SPECIALTY STEEL
EAF DUST
INMETCO
PROCESS
SLAG
BUREAU OF
MINES
PROCESS
SLAG
NICKEL-CHROMIUM-IRON METAL
STAINLESS STEEL
Figure 2. Overview of promising recovery processes for EAF dust.
-------
AMAX CAUSTIC LEACH-ELECTROWIN ZINC PROCESS
AMAX initially studied the caustic leach-electrowin zinc process in a
pilot plant to determine the feasibility of recovering zinc from a high-iron
zinc ore.7 With minor modifications, this process can be used to recover zinc
from complex zinc oxide-bearing flue dusts such as EAF dust. AMAX has run
bench-scale tests on various steel plant residues containing >20 percent zinc
and obtained very encouraging results.8 Because of the limited solubility of
iron in caustic, the caustic leach-electrowin zinc process is especially
advantageous for recovering zinc from carbon steel EAF dust; the conventional
sulfate electrolytic zinc smelting process would be fouled by the iron in the
dust. The final product of the process is high-grade zinc (>99.9% Zn) in
particulate form.
Figure 3 presents a schematic of the process that AMAX used in their
pilot plant to recover zinc from low-grade zinc concentrates. The pilot plant
can produce 1360 kg of zinc per day by caustic leaching and subsequent elec-
trolysis of a calcine obtained from a low-grade zinc concentrate with a zinc
content greater than 40 percent and an iron content of about 17 percent. In
this process, the dead-roasted calcine is first retreated by a kiln-type
reduction roast to decompose the zinc ferrite formed during dead roasting and
thereby improve the solubility of the zinc. The reduced calcine is then
leached in a 30 percent caustic solution; the leach dissolves more than 90
percent of the zinc and lead present and most of the cadmium, but only a very
small amount of iron (less than 0.5 percent). The zinc solubilization step
can be represented as:
ZnO + 2NaOH -»• Zn(ONa)2 + H20
sodium
zincate
The leach residue obtained after filtration is treated in an electric
furnace (lined with burned magnesite brick9) to recover copper and precious
metals and to produce an oxide dust containing zinc and lead, which is re-
turned to the leaching circuit. The filtrate containing sodium zincate is
purified in two steps: 1) aeration to precipitate iron as ferric hydroxide
[Fe(OH)3], and 2) cementation with the zinc powder product to remove lead,
copper, cadmium, and other metallic impurities that are more noble than zinc.
The cake remaining after cementation is separated and processed for lead
recovery, whereas the purified solution is electrolyzed to recover pure zinc
and to regenerate sodium hydroxide. The electrowinning step can be described
by the reaction:
Zn(ONa)2 + H20 •* Zn + 2NaOH + 1/2 02
Power consumption is approximately 1.2 x 10 J per kg of zinc produced.
Makeup caustic required is about 20 kg per metric ton of zinc.9
-------
The process variation suggested for treating steelmaking dusts is de-
picted in Figure 4. For most steel dusts, the reduction roast step is be-
lieved to be unnecessary; however, for many carbon steel EAF dusts in which a
significant portion of the zinc is tied up as zinc ferrites, reduction roast-
ing with carbon in a neutral atmosphere is essential prior to caustic leach
treatment.9 The iron residue from the leaching circuit is reported to be
nonhazardous, and can be landfilled. If desired, metallization of the residue
can be carried out in a direct reduction kiln, and the iron product obtained
can be recycled to the steel plant. The process can recover over 98 percent
of the zinc in the dust.
The AMAX process presents a technically feasible alternative for recovery
of metals such as zinc and lead from carbon steel EAF dusts. AMAX is cur-
rently performing bench-scale tests on dusts from domestic and European steel
companies to determine the applicability of their process to the particular
dusts.9 They believe that economic commercial operation would require a plant
capable of producing about 20,000 metric tons of zinc per year.8 Different
types of steelmaking dusts could be combined and processed in a commercial
plant.
NEW JERSEY ZINC WAELZ KILN PROCESS
Over the past two years, The New Jersey Zinc Company has intermittently
processed carbon steel EAF dusts from various domestic steel plants in their
Waelz kilns at Palmerton, Pennsylvania.10 Using the "American process," the
plant also separately processes a zinc silicate ore (franklinite) for zinc
oxide production. Approximately 22,000 metric tons of EAF dusts were treated
in 1980. The final product obtained by processing the dusts is an impure zinc
oxide containing about 50-55 percent zinc and 6 to 7 percent lead. This
product is currently stockpiled and will be sold to another smelter for
recovery of its metallic value.
The Waelz kiln operation at the New Jersey Zinc smelting facility is
similar to the practice at the Berzelius plant in Duisburg, West Germany. The
Waelz unit at the latter plant has been processing pelletized EAF dust at a
rate of 50,000-55,000 metric tons per year on a full scale since the end of
1977; the process is stated to be economical for zinc contents in the dust of
greater than 15 percent.11 The impure zinc oxide sinter obtained at the plant
is processed in an Imperial Smelting shaft furnace for simultaneous recovery
of zinc and lead.
A simplified flow diagram of the process used to treat carbon steel EAF
dusts at the Palmerton plant is shown in Figure 5, whereas a detailed flow
sheet for the "Waelzing" operation is presented in Figure 6. The EAF dusts
arrive at the plant by trucks or rail. The facility has a pelletizer for
treating dusts that arrive in nonpelletized form. The New Jersey Zinc Company
has obtained a permit for storing the material. The pelletized dusts, lime-
stone, and anthracite coal constitute the feed to the kilns. In the kilns,
gas flow is countercurrent to solids flow, and the zinc oxide fume and kiln
gases are withdrawn at the feed end. The zinc oxide fume, known as Waelz
oxide, is cooled by air dilution and collected in baghouses. The Waelz oxide,
which typically contains 50 to 55 percent zinc and 6 to 7 percent lead, is the
-------
CONCENTRATE
_L
*- TO ACID PLANT
CAUSTIC
Pb RESIDUE
*
T0 Pb SMELT"
ZINC
PRODUCT
Figure 3. AMAX process for treating low grade zinc concentrates.7
*
Fe RESIDUE
I
TO Fe METALLIZATION STEP
*
Fe TO STEEL PLANT
r
RESIDUE
(Cu.Pb, ETC.)
CAUSTIC
ZINC PRODUCT
Figure A. AMAX process for recovery of zinc from steelmaking dusts.
ANTHRACITE COAL »-
EAF DUST ••
LIMESTONE »•
UAELZ KILN
ZnO
FUME
DUST COLLECTION
SYSTEM
»- ZnO PRODUCT
(WAELZ OXIDE)
i
IRON
RESIDUE
Figure 5. Simplified flow diagram of the New Jersey Zinc Waelz kiln process.
-------
GANTRY CRANE
WAELZ PROCESS OXIDE
•— rRQM SCRUBCR WATER
TREATMENT SYSTEM
STORAGE
BINS
LIMESTONE COAL
BiTUM INDUS
OIL.NAT. GAS I fCE. GAS
BURNERS
5VPULVEPIZER
GANGUE
TO STOCK PILE
TRANSFER CAR TO
AIR AND GASES
MECHANICAL
BAG ROOM
WAELZ OXIDE
(IMPURE ZINC
OXIDE PRODUCT)
OUT -•
BELT CONVEYOR
Figure 6. Detailed flow diagram of Waelz kiln operation at the New Jersey Zinc Company,
Palmerton, Pennsylvania.
-------
final product and is being sold to another company for recovery of zinc and
lead. The residue from the kilns contains only 30 to 40 percent iron and
cannot be recycled directly to the steel plant furnaces. Tests conducted by
The New Jersey Zinc Company have, however, found the residue to be nonhazard-
ous; therefore, it can be disposed of in a sanitary landfill.
The fee charged by the Palmerton zinc plant for processing of carbon
steel EAF dusts from steel plants is a function of the zinc content of the
dust. Normally, dusts with zinc contents greater than 20 percent are pre-
ferred.
SKF PLASMADUST PROCESS
The PLASMADUST process, which has been developed by SKF Steel Engineering
in Sweden, uses a special type of electrothermic shaft furnace to recover
zinc, lead, chromium, nickel, molybdenum and iron from EAF steelmaking dusts.
The process can be applied to dusts from both carbon steel and specialty steel
manufacture. Unlike the Waelz kiln process, which renders an impure zinc
oxide concentrate for further reduction to metallic form by thermal or elec-
trolytic methods, the PLASMADUST process produces a zinc vapor that can be
directly condensed to liquid metal.12 The liquid iron tapped from the furnace
is alloyed with chromium, nickel and molybdenum when stainless steel dusts are
processed.
A flow diagram of the PLASMADUST process is presented in Figure 7. A
feed consisting of untreated flue dust, coal powder, and process gas that has
been preheated in a plasma arc to about 3500°C is fed into the tuyere level of
the shaft furnace. Coke is added from the top to fill the shaft. An endo-
thermic reaction occurs at the bottom of the column between the oxide dust and
the reducing agent (coal/coke). The plasma gas, which has a heat content of
about 2.0 x 10 J/Nm3, supplies the heat for the reaction since no oxygen or
air is admitted to the furnace. In the reducing atmosphere created by excess
coke, carbon monoxide and hydrogen are formed, which rise in the shaft. Most
of the suspended particles are caught on the coke surface and eventually
returned to the reaction zone. Crude iron (containing the alloying elements)
and slag are tapped separately from the bottom of the furnace. Zinc and lead
vapors rise with the furnace gases and are condensed and separated in a splash
condenser. Since formation of carbon dioxide and water vapor is eliminated
because of the reducing atmosphere, the reoxidation risk of zinc and lead
vapors is minimized.
The PLASMADUST process has been piloted by SKF Steel Engineering in
Sweden. The pilot plant, which is capable of processing about 300 kg/h of
flue dust, has shown an iron yield of 99 percent and zinc and lead yields of
greater than 97 percent on a typical carbon steel EAF dust containing 50
percent Fe203, 25 percent ZnO, and 2 percent PbO. Theoretically, waste
oxides with only a few percent zinc are equally well suited for the PLASMADUST
process, the main technical requirement being that they be sufficiently dry
and fine to allow for pneumatic feeding. The economic lower limit, however,
is estimated to be about 10 percent zinc.12 Carbon steel EAF dusts, which
typically have zinc levels of 15 to 25 percent and lead levels of 1 to 4
percent, are especially attractive raw materials for treatment by this proc-
ess.
11
-------
A commercial plant using the SKF process and having a capacity of 70,000
metric tons of dust per year is scheduled to be placed into operation at
Landskrona in southwest Sweden by late 1984.13 Baghouse dust from steel mills
in Sweden and other Scandinavian countries will form the raw material to the
plant; metals recovered will include iron, zinc, lead, chromium, nickel and
molybdenum. Heat recovered from the plant will be used by the Landskrona
municipality for district heating and other purposes.
SKF Steel Engineering is currently looking into the possibility of
constructing commercial plants in the United States. The promoters of the SKF
process (SiAR, Inc., Cambridge, Massachusetts), believe that Pittsburgh or
Chicago would be good locations for such plants because of the concentration
of steelworks in those districts.14 A supply of at least 50,000 metric tons
of dust per year from neighboring sources would be required for economic
operation of a plant.
INMETCO PROCESS
The Inmetco process as applied in the Ellwood City, Pennsylvania facility
treats approximately 110 metric tons of wastes per day from stainless and
specialty steel mills to produce 60 metric tons per day of a nickel-chromium-
iron pig which is sold back to the steel mills as a melting feedstock.15 Over
three years of operating experience has proven that the Inmetco process is
viable and profitable. The Inmetco plant which serves the majority of the
specialty steel mills in the eastern part of the U.S. receives three major
types of wastes: EAF dust, mill scale, and grinding swarf. Typical composi-
tions of the raw materials are shown in Table 2. Typically, 50 percent of the
feed is EAF dust.
TABLE 2. CHEMICAL COMPOSITION OF WASTES RECEIVED AT THE INMETCO PLANT15
Type of waste
EAF dust
Mill scale
Grinding swarf
Range , wt . %
Cr
2.2-19.1
7.8-13.1
13.4-20.0
Ni
1.0-5.4
1.3-4.1
4.8-20.1
Mo
0.1-1.0
0.1-0.4
0.3-1.6
Zn
0.7-2.0
Cu
0.2-1.7
0.1-1.4
0.2-1.4
The major steps in the Inmetco process are 1) feed preparation, blending
and pelletizing, 2) reduction, and 3) smelting and casting. Figure 8 is a
simplified flow diagram of the Inmetco process.
The EAF dust and the other wastes arrive at the Inmetco facility by truck
tankers, and are assigned to storage silos based on their nickel and chromium
contents. The materials in the various silos are carefully blended to produce
a mix with satisfactory processing characteristics and uniform alloy content.
Any deficiency in either nickel or chromium content is made up by adding
nickel alloy or ferrochrome. The wastes are combined with coal/coke, lime
and water in a pug mill, and the resulting mixture pelletized before under-
going partial reduction in a rotary hearth furnace. A portion of the zinc,
lead and halogens contained in the flue dust are exhausted into the off-gas
treatment system. The hot, partially metallic, sintered pellets (containing
12
-------
unreduced chromium oxide) are fed to an electric arc smelting furnace where
the pellets are melted and the chromium oxide is reduced by the residual
carbon in the pellet. Lime and other siliceous materials form a liquid slag
which is tapped separately from the alloyed hot metal. The metal is cast into
pigs before being sold. The product typically analyzes 14.0 wt. percent Ni,
14.0 wt. percent Cr and 64.0 wt. percent Fe. Typical metallic recoveries are
as follows: Ni-99 percent, Cu-94 percent, Mo-92 percent, Fe-98 percent. The
slag is essentially inert and is used locally as fill and road ballast. A wet
gas scrubbing system is used to treat the gas emitted from the furnace. The
scrubber filter cake which contains ~10 wt. percent Zn is disposed of through
a secondary zinc producer.
The operations at Inmetco demonstrate the commercial viability of the
process for recovering valuable alloying elements from stainless steel EAF
dusts. The composition of the waste materials and the proximity of the waste
generator to the Inmetco facility are the factors which can influence the
profitability of the process.
BUREAU OF MINES PROCESS
The U.S. Bureau of Mines has developed a recycling process for recovery
of alloying elements from stainless steel EAF dusts.16 The Bureau of Mines
technique is similar to the Inmetco process in that different specialty
steelmaking wastes are blended, agglomerated on a pelletizer with an addition
of a small amount of carbonaceous reducing agent, and charged into a furnace.
After initial laboratory tests indicated that the process had good
potential for recovery of metals, a series of larger industrial-scale tests
were conducted at Joslyn Stainless Steels in Fort Wayne, Indiana. Figure 9 is
a schematic of the Bureau of Mines process. The raw materials which comprised
stainless steel EAF dust, argon-oxygen decarburization dust, grinding swarf
and mill scale were agglomerated to produce pellets 3/8" to 1" in diameter,
analyzing 39.2 wt. percent Fe, 8.9 wt. percent Cr, 3.7 wt. percent Ni and 0.5
wt. percent Mo. The pellets were manufactured by an outside contractor.
During the tests, 14 to 19 percent of the charge to the 18-ton electric arc
furnaces consisted of pelletized wastes generated at the plant. Stainless
steel scrap constituted 80 to 90 percent of the total charge. All the trials
produced successful commercial heats of Type 316 stainless steel. Results
showed chromium recoveries up to 90 percent, and nickel, iron and molybdenum
recoveries of up to 99 percent.17 Moreover, the heats were routinely pro-
duced, and no additional energy or manpower was required.
The Bureau of Mines process may be attractive to specialty steelmaking
companies who would wish to practice in-plant recycling.
ECONOMIC EVALUATION
Economics is the governing factor in deciding if any of the metal re-
covery alternatives are attractive to the steelmaker. Obviously, if the cost
of disposing the EAF dust at an off-site hazardous waste landfill is greater
than any associated transportation and handling costs incurred in shipping the
dust to a recovery plant, recycling/recovery would be worth pursuing by the
steelmaker.
13
-------
COKE
FURNACE
GAS
SLAG HOT METAL
HIGH Btu GAS _
Figure 7. Simplified flow diagram of the SKF PLASMADUST process
12
STEEL HASTES *•
COAL m-
LIME *•
UATER »-
PUG
MILL
PELLETIZER
ROTARY
HEARTH
FURNACE
ELECTRIC
ARC
FURNACE
^. HOT METAL
PRODUCT
SLAG
Figure 8. Simplified flow diagram of the Inmetco process.
OVERSIZE
MILL SCAL
COKE STAINLESS
BREEZE CEMENT WATER STEEL SCRAP
STAINLESS STEEL If I t '
EAF DUST ~-
GRINDING SWARF •_
ADD DUST —
FINE MILL SCALE »-
BLENDER
__^^^
PF1 1 FTI7FR
' uL.L.L.1 1 i. L 1\
E
SLAG
CONDITIONERS
1
ELECTRIC
ADP
nKL
FURNACE
SLAG
Figure 9.
STAINLESS
STEEL MELT
(TO ARGON OXYGEN
DECARBURIZATION)
Schematic of Bureau of Mines process for recycling stainless
steel wastes.
14
-------
The recovery facility owner may wish to pay for transportation of the
dust, provided the steel plant is within a reasonable distance from the
recovery plant. For example, the New Jersey Zinc Company has previously paid
for transportation from sources within 300 miles of their plant. Transporta-
tion costs do not arise in the case of the Bureau of Mines process which
involves in-plant recycling of the dust.
AMAX and SKF have reported that the capital investment required for a
recovery plant capable of treating 70,000 to 100,000 metric tons per year of
steelmaking dusts/wastes would be about $20 million.8'12 New Jersey Zinc
also estimates the cost of replacing their Waelz plant to be about $20 mil-
lion.10 Table 3 presents our economic evaluation of these three carbon steel
EAF dust recovery options for an assumed processing capacity of 90,000 metric
tons per year. The evaluation is based on an average dust composition of 30
wt. percent Fe, 24 wt. percent Zn and 2 wt. percent Pb. Our budget estimates
show that all the three recovery alternatives could be profitable.
TABLE 3. ECONOMIC EVALUATION OF RECOVERY
PROCESSES FOR CARBON STEEL EAF DUSTS*
(Basis: 90,000 metric tons per year capacity plant)
Item
AMAX
Electrowin
zinc process
New Jersey Zinc
Waelz kiln
process
SKF
PLASMADUST
process
Direct Operating costs,
$/metric ton of dust
Indirect operating costs,
$/metric ton of dust
Transportation costs,**
$/metric ton of dust
Total annualized costs,
$/metric ton of dust
Product credits,
$/metric ton of dust
Profit,
$/metric ton of dust
67
42
20
129
211
82
72
42
20
134
160
26
120
42
20
182
251
69
Based on 1981 cost data.
Assumed.
No cost data were available for the Inmetco process, although their
operation is reported to be profitable. An economic evaluation of the Bureau
of Mines process has been conducted; a profit of about $130 per metric ton
of agglomerated waste is estimated based on a plant designed to smelt 5 metric
tons per day of pelletized wastes.
15
-------
CONCLUSIONS
Several promising technologies have been developed at either pilot or
commercial scales which offer technically feasible routes for recovery of
metallic values from electric arc furnace steelmaking dusts. With the excep-
tion of the Bureau of Mines process which allows in-plant recycling of the EAF
dust, all the other processes require that EAF dusts and other compatible
wastes from different steel companies be shipped to a centralized processing
facility. Economic operation of the facility requires that the steel plants
be in close proximity to the recovery facility. The New Jersey Zinc Waelz
kiln plant in Palmerton, Pennsylvania, and the Inmetco plant in Ellwood City,
Pennsylvania are the two commercial recovery facilities currently in opera-
tion.
Recovery of metallic values from EAF dust offers several benefits.
Firstly, it eliminates the disposal costs involved in landfilling the hazard-
ous waste. In-plant recycling has the added advantage of reducing the steel
producer's cost by not wasting expensive materials. Further, the recovery
technologies can help reduce the nation's demand for imports of chromium, zinc
and other critical metals.
ACKNOWLEDGMENTS
The authors would like to thank the following people for providing
invaluable information for this study: N. H. Keyser, Consultant, Hinsdale,
Illinois; W. R. Opie and H. P. Rajcevic, AMAX Base Metals Research and Develop-
ment, Inc., Carteret, New Jersey; P. L. Kern and M. R. Silvestris, The New
Jersey Zinc Company, Inc., Palmerton, Pennsylvania; P- Baumann, SiAR, Inc.,
Cambridge, Massachusetts.
REFERENCES
1. Barnard, P. G., A. G. Starliper, W. M. Dressel, and M. M. Fine. "Recycl-
ing of Steelmaking Dusts," U.S. Department of the Interior, Bureau of
Mines, TPR 52, February 1972.
2. Powell, H. E., W. M. Dressel, and R. L. Crosby. "Converting Stainless
Steel Furnace Flue Dusts and Wastes to a Recyclable Alloy," U.S. Depart-
ment of the Interior, Bureau of Mines, Report of Investigations 8039,
1975.
3. Higley, L. W., and H. H. Fukubayashi. "Method for Recovery of Zinc and
Lead from Electric Furnace Steelmaking Dusts," U.S. Department of the
Interior, Bureau of Mines, a paper in the Proceedings of the Fourth
Mineral Waste Utilization Symposium, Chicago, Illinois, May 7-8, 1974.
4. Keyser, N. H., J. R. Porter, A. J. Valentino, M. P. Harmer, and J. I.
Goldstein. "Characterization, Recovery and Recycling of Electric Arc
Furnace Dust," presented at the Symposium of Iron and Steel Pollution
Abatement Technology for 1981, Chicago, Illinois, October 6-8, 1981.
5. Metal Market 1981. Red Book of Annual Statistics (1980 Data).
16
-------
6. Higley, L. W., and M. M. Fine. "Electric Furnace Steelmaking Dusts - A
Zinc Raw Material," U.S. Department of the Interior, Bureau of Mines,
Report of Investigations 8209, 1977.
7. Anderson, W. W., H. P. Rajcevic, and W. R. Opie. "Pilot Plant Operation
of the Caustic Leach-Electrowin Zinc Process," IMS Paper Selection,
A81-52, The Metallurgical Society of AIME, Warrendale, Pennsylvania
15086.
8. Opie, W. R., Director, AMAX Base Metals Research and Development, Inc.,
Carteret, New Jersey. December 14, 1981, personal communication.
9. Rajcevic, H. P., AMAX Base Metals Research and Development, Inc.,
Carteret, New Jersey. March 24, 1982, personal communication.
10. Kern, P. L., Director of Manufacturing and Technical Services, and M. R.
Silvestris, Environmental Control Engineer, The New Jersey Zinc Company,
Inc., Palmerton, Pennsylvania. December 15, 1981, personal communica-
tion.
11. Maczek, H., and R. Kola. "Recovery of Zinc and Lead from Electric-
Furnace Steelmaking Dust at Berzelius," Journal of Metals, p. 53-58,
January 1980.
12. Herlitz, H. G., SKF Steel Engineering AB, Hofors, Sweden. "The
PLASMADUST Process for Recovery of Metals from Waste Oxides," presented
at the Symposium on Resource Recovery and Environmental Issues of
Industrial Solid Wastes, Gatlinburg, Tennessee, October 28-30, 1981.
13. Herlitz, H. G. "SKF Steel Engineering Bulletin," published by SKF Steel
Engineering AB, Hofors, Sweden, June 1982.
14. Baumann, P., SiAR, Inc., Cambridge, Massachusetts. November 30, 1981,
personal communication.
15. Hanewald, R. H., J. K, Pargeter, and J. A. MacDougall. "Agglomeration
and the Inmetco Process," paper provided by Inmetco, Ellwood City,
Pennsylvania. 1982.
16. Higley, Jr., L. W., L. A. Neumeier, M. M. Fine, and J. C. Hartman.
"Stainless Steel Waste Recovery System Perfected by Bureau of Mines
Research," 33 Metal Producing, November 1979.
17. Bureau of Mines, U.S. Department of the Interior. "In-Plant Recycling of
Stainless and Specialty Steel Wastes," Technology News, No. 91, January
1981.
18. Barnard, P. G., W. M. Dressel, and M. M. Fine. "Arc Furnace Recycling of
Chromium-Nickel from Stainless Steel Wastes," U.S. Department of the
Interior, Bureau of Mines, Report of Investigations 8218, 1977.
17
-------
Session 4
WATER POLLUTION ABATEMENT
Symposium
on
Iron and Steel
Pollution Abatement Technology
For 1982
Sponsored by
US EPA
INDUSTRIAL ENVIRONMENTAL RESEARCH
LABORATORY
Research, Triangle rark, NC
and
AMERICAN IRON AND STEEL INSTITUTE
Washington, DC
NOVEMBER 16, 17, AND 18, 1982
The William Penn Hotel
Pittsburgh, PA
-------
BLAST FURNACE - RECYCLE SYSTEM
SIDE STREAM SOFTENING
SUMMARY OF FULL SCALE TRIAL
BY
LEONARD D. WISNIEWSKI
ASSISTANT DIRECTOR - WATER
ENVIRONMENTAL CONTROL
AND
RICHARD L. NEMETH
ASSISTANT SUPERINTENDENT
ENVIRONMENTAL CONTROL
REPUBLIC STEEL CORPORATION
CLEVELAND, OHIO 44101
NOVEMBER, 1982
-------
The work described in this paper was not funded by the U.S.
Environmental Protection Agency. The contents do not necessarily reflect
the views of the Agency and no official endorsement should be inferred.
ABSTRACT
Republic Steel Corporation's Cleveland District has undertaken a
program to reduce blast furnace recycle system blowdown to volume levels
that can either achieve regulatory requirements or be disposed of via
quenching and evaporation on blast furnace slag. This is a potential
alternate to the treatment of blast furnace blowdown using alkaline-
chlorination treatment. It will be necessary to remove scale forming
constituents from the recirculating waters to prevent harmful buildup
throughout the recycle system as the cycling is increased. Republic is
attempting to accomplish this by investigating the effectiveness of "side-
stream" softening a portion of the total recycle volume.
This paper details the results of the work conducted to date on a trial
softening program at its 5 & 6 Blast Furnace Recycle System at the Cleveland
District steelmaking facility. Included is a description of the modified
treatment system and a summary of the water chemistry softening results. A
summary which compares the system's discharge loads to the recently
promulgated Best Available Technology (BAT) effluent limitation guidelines
is also included. This paper is a followup to a paper presentation made by
the same authors last year at the U.S. EPA Symposium on Iron and Steel
Pollution Abatement Technology for 1981, which dealt with the efforts to
minimize Cleveland's 5 & 6 Blast Furnace recycle system blowdown rate to
below BAT levels.
BACKGROUND
BLOWDOWN MINIMIZATION
Recycle water systems for gas cleaning and cooling were installed at
both Cleveland District Blast Furnace complexes to conform with Best
Practicable Technology (BPT) requirements. These systems were designed to
operate at blowdown rates of 1,000-1,200 gpm, enabling them to meet these
guidelines. It was apparent that these design criteria (the high blowdown
rates) could not achieve compliance with proposed Best Available Technology
(BAT) limitations. The options available dictated the necessity to reduce
blowdown to a more manageable rate, thereby minimizing future capital
expenditure and operating costs to meet these BAT regulations.
The immediate concern in efforts to reduce the blowdown rates centered
around elimination of extraneous water infiltration to the recycle systems.
These sources were the cause of hydraulic imbalance which required the
elimination of water from the systems for hydraulic rather than water
chemistry reasons. The fact that these were retrofit systems to existing
facilties complicated the task. However, results of identifying and
systematically eliminating sources of extraneous water culminated in the
-1-
-------
ability to hydraulically operate one of these systems, the 5 & 6 Blast
Furnace Recycle System, with a sustained very low blowdown rate. Althoug**
unresolved problems with hydraulic balance of this system remain, the
sources have been identified and can managed be either physically or
adminis tratively.
Figure 1 illustrates the results of blowdown reduction, tracking the
system from January, 1980 to September, 1982. The apparent increase after
November, 1981 was due to a decrease in iron production which reflected an
increase in the blowdown rate (gallons per ton) while the actual blowdown
rate (gallons per minute) remained relatively constant. The extraneous
water resulting in the blowdown volume was related to the blast furnace in
operation and should be resolved when this furnace is taken down for
rebuild.
REPUBLIC STEEL CORP
[CLEVELAND DISTRICT
586 BL FCE RECYCLE
|M M|J|S N|J|M|M|J|S|N|J M M|J
FAJAODFAJAODFAJA
1980
I 981
1982
Figure 1. Republic Steel Cleveland District - Nos. 5 & 6 Blast Furnace
Recycle System - Blowdown Flow by Month.
Another alternative to meet BAT limitations was to not discharge
blowdown from the recycle system. No discharge, of course, would eliminate
the concern for constituent limitations. No discharge does not mean zero
blowdown. If blowdown rates are minimal, methods for blowdown disposal
become very viable alternates to discharge treatment. Water evaporation
from blast furnace slag quenching or pelletizing could then provide adequate
disposal for blowdown from a very tight system.
Water chemistry became a factor of significant impact, once having
resolved the hydraulic needs of the system. As demostrated by deposit
analyses shown in Figure 2, continued operation of this system with very low
blowdown showed a marked increase in the calcium carbonate concentration.
Applying patented chemical treatment to cope with this potential did provide
for a degree of success, however, sustained operation with this scheme was
still unproven. For this reason a study in the use of side stream water
softening to control system hardness was undertaken in 1981.
-2-
-------
40
P
Q.
UJ
O
U.
£
o* I0
o
ro
8 n
4
F. J
\1
Co
6
I
J
-
nil
. _
\
I
n
1
I
n
JZ
1
AUG
1978
MAY
1980
JUNE
1961
DEC
1981
DEPOSIT ANALYSIS HISTORY, 586 BL FCE RECYCLE SYSTEM
Figure 2. Republic Steel Cleveland District - Nos. 5 & 6 Blast Furnace
Recycle System - Deposit Analysis History.
PILOT STUDY
The U.S. EPA Mobile Pilot Treatment Plant was utilized to investigate
various softening modes on a 5-6 gpm side stream to determine the
feasibility and cost effectiveness of side stream softening to control
system hardness. This method of controlling calcium carbonate scaling
potential was a technology transfer from closed, recirculating cooling
systems to a blast furnace recycle system. Results of the trial affirmed
this potential. Trial results demonstrated the ability to sucessfully
soften the blast furnace recycle water by several means, from classic lime-
soda ash softening to simply elevating the pH by caustic soda addition. The
pilot studies also affirmed that remixing of softened water with unsoftened
recycle water generated continued softening reactions and sedimentation in
the mixed stream, a situation which must be considered in the future design
of a side-stream softening system.
Questions concerning practicality, operating costs and overall effect
of side stream softening on the entire recycle system had to be resolved.
The decision was made to expand the investigations to a full scale trial to
provide answers to these questions.
FULL SCALE TRIAL
OBJECTIVES
A major concern in the full scale softening trial was the control and
monitoring of calcium and total hardness, these being the most impactive
constituents for system problems. Softening potential of this water had
already been established, so the overall effect of continuous side stream
softening to maintain a lowered system hardness was of significant interest.
It was necessary to observe what effects this scheme had on equipment at
various stages of the system, and if other constituents in the water would
indicate additional problems.
-3-
-------
The cost of side stream softening had to be more precisely determined
than those costs projected by the pilot plant study. These determiantions
had to be established within the constraints of existing equipment at the
water treatment plant. A staff from Republic's Research Center was utilized
to operate, monitor and evaluate the softening trial.
PHYSICAL DESCRIPTION
One of the two 90' diameter clarifiers was modified for use as a
softener. No. 2 Clarifier was selected because of it's proximity to the
influent stream, which allowed for remixing of the softened water with the
influent stream ahead of the other clarifier. Any precipitation due to
continued post-mix softening would be settled out in either clarifier and
removed along with the settled solids at the vacuum filters. Clarifier
discharge piping had to be modified to redirect softener effluent from the
hot well of the recycle system cooling tower to the influent trench as shown
in Figure 3. This was accomplished by adding a second discharge pipe from
the overflow launder of the softener-clarifier to the influent trench, and
blanking off the overflow launder discharge to the hot well as shown in
Figure 4. These modifications were made so No. 2 Clarifier could be quickly
returned to a full stream clarifier if it became necessary to abort the
trial due to a malfunction of No. 1 Clarifier.
REPUBLIC STEEL-CLEVELAND
NOS 586 BLAST FURNACES
RECIRCULATING SYSTEM
NO 2 COKE
PLANT
LAGOON
SOURCES
NO Z
POWER
HOUSE
NO 6
BLAST
FURNACE
NO 5
BLAST
FURNACE
-37 '/.i ^
y
EVAP a
DRIFT
Figure 3. Republic Steel Cleveland District - Nos. 5 & 6 Blast Furnace
Recycle System - Flow Schematic for Side Stream Softening Trial.
-4-
-------
REMOVABLE TEMPORARY
EXISTING CLARIFIER
OVERFLOW TO COOLING
TOWER HOT WELL
MODIFIED OVERFLOW
TO
INFLJENT
('TRENCH
I
NO 2 CLARIFIER
MODIFICATION
Figure 4. Republic Steel Cleveland District - Nos. 5 & 6 Blast Furnace
Recycle System - No. 2 Clarifier Modification for Side Stream
Softening Trial.
The next concern was to develop a method to measure the side stream
flow which was to be regulated by positioning No. 2 clarifier shut off gate.
A temporary flume, similar to a Palmer-Bowles type, was constructed and
inserted into No. 2 Clarifier influent launder. This flume, although crude,
was an effective primary flow measuring device. The method for softening
was simply pH elevation by addition of caustic soda (sodium hydroxide). The
method was selected for simplicity, as liquid caustic was readily available
in bulk, and feed could be easily controlled and measured. As shown in
Figure 5, caustic was fed directly from a bulk tank truck through a
regulated value to provide the desired feed rate. The feed rate was
measured manually at the feed point in the clarifier influent launder.
CAUSTIC
FEED LINE
SIDESTREAM
TO NO 2
CLARIFIER
•»TO NO I
CLARIFIER
PRESSURE
RELIEF VALVE 1
REGULATOR
PLANT
AIR SUPPLY
CAUTION CAUSTIC SODA
Figure 5. Republic Steel Cleveland District - Nos. 5 & 6 Blast Furnace
Recycle System - Caustic Soda Feed System for Side Stream
Softening Trial.
-5-
-------
CONSTRAINTS
Numerous constraints had to be reconciled both prior to and during the
trial period. Logic dictated not to attempt this experiment during the
winter, considering the physical characteristic of 50% caustic soda and its
solution freeze point of 54 F. During these severe economic times, fundin»
of such a project can also be restrictive. Credit district and corporate
management for authorizing funds for this program. The modifications
required for the trial interfered with scheduled clarifier maintenance and
imposed a time constraint.
During this trial, only No. 5 blast furnace was in operation. Of the
two furnaces, No. 5 caused the most severe hydraulic imbalance and unsteady
operation since its general condition was less than ideal as it concludes
its current campaign. Therefore, general system stability (water flows,
blowdown, water chemistry) was somewhat erratic prior to, and during the
trial. Coupled to the normal operating problems inherent to any blast
furnace and water treatment system, considerable effort was necessary to
maintain system tightness sufficiently to proceed.
The modifications to perform the trial, though adequate, were far from
ideal. Flow control of the side stream and the caustic feed required close
monitoring and constant attention. Calculated caustic feeds were soon
discarded in favor of caustic feed adjustment to obtain desired pH in the
softener-clarifier. Monitoring of the trial under these constraints
required many man hours and the efforts of several people.
Chemical constraints of the system were present. It was unknown at
what pH the entire recycle system would stabilize while the softening
clarfier was maintained at the pH level necessary to sustain the softening
reaction. The potential existed for acid to be fed to the recycled water
for pH adjustment, because of discharge requirements and for cyanide
control. The initial softening effect of pH elevation in the system had to
be addressed carefully so as not to cause deposition throughout the system.
TRIAL PROCEDURE
Previous data while operating the system at very low blowdown rates
prior to softening would be used as the base of comparison. System hardness
was reduced by increasing blowdown to minimize the effect of initial
softening and to reduce both the time and the caustic required. The
blowdown was then reduced as the trial began to allow concentrating of the
hardness constituents. The trial proceeded by softening a 250 gpm influent
side stream at a pH of 10.5, then at a pH of 9.5, then back to a pH of 10.5.
Next, a clarified and cooled water side stream of 150 gpm was softened at a
pH of 10.5. Without sophisticated pH control equipment, pH values varied
somewhat throughout each trial phase. However, control was adequate enough
to demonstrate the principle and observe the results.
-6-
-------
TRIAL MONITORING
It was first necessary to monitor certain physical parameters to keep
on track with the trial procedures. Calibration of the influent flume fo~
flow measurement allowed for monitoring and adjusting flow to maintain a
constant side stream flow. Measuring caustic feed indicated consumption
rates and helped to regulate softener pH. Desired softener pH determined
caustic feed. Continuous monitoring of system flow, make-up flow and
blowdown flow indicated the degree of system tightness.
The next monitoring requirement dealt with water chemistry, which
included monitoring for several constituents at many points throughout the
system. Temperature, conductivity, pH, alkalinity, calcium hardness, total
hardness, chloride concentration and turbidity were monitored daily at five
points throughout the system. This data would indicated the degree of
success or failure of the project objectives.
Last came observations of the overall effect on the physical system.
This was accomplished by frequent inspections of strategically located
deposition coupons. Before and after inspections of the gas cooler were
conducted to Determine if there were any major indications of potential
problems.
RESULTS
The trial began on June 8, 1982 and proceeded in four modes until
August 2, 1982. System history as shown in Figure 6 demonstrates that at
the comparable BAT blowdown rate of 70 gallons per ton, the system total
hardness was approximately 750 mg/1. Data collected during the softening
trial showed a hardness level of 500 mg/1 could be maintained at a blowdown
range of between 3.5-7.0 gallons per ton. Historical data indicated that at
this very low blowdown rate, total hardness without softening was over 1,000
mg/1.
1500
o
:iooo
O
U
V)
OT
O
K
500
SOFTENING
REPUBLIC STEEL CORP
CLEVELAND DISTRICT
5 a 6 BL FCE RECYCLE
100 200 300 400
BLOWDOWN RATE, GAL/TON
500
Figure 6.
Republic Steel Cleveland District - Nos. 5 & 6 Blast Furnace
Recycle System - Total Hardness as a Function of Blowdown Rate
(January, 1981 - August, 1982), including Side Stream
Softening Trial.
-7-
-------
During the trial period, system pH remained at 8.0 to 8.2 with no
significant changes in pH occuring at any time. Inspection of deposition
coupons throughout the system showed that no significant deposits occurred
during the trial. The overflow weirs of the softener-clarifier did show a
white scale deposit that did not pose a problem during the trial.
Concurrently, vacuum filter operation appeared more effective, with media
life doubling. It has been theorized that high system hardness may cause
reduced filter-bag life because of premature blinding of the media fibers.
A post-trial inspection of the blast furnace gas cooler showed no signs
of increased depostion, and general system cleanliness remained very good.
The net reduction of total system hardness was optimum when the 250 gpm side
stream was softened by pH elevation to 10.5, when system hardness was
maintained at the 300-400 mg/1 range. Operating the side stream at a pH of
9.5 caused a slight increase in system hardness to the 400-500 mg/1 range.
Reduction of the side stream to 150 gpm and operating at a pH of 10.5
resulted in system hardness of 500-600 mg/1. Calcium hardness in the
recycle system throughout the trial for the most part remained in a 130-230
mg/1 range.
CONCLUSION
This full scale trial has conclusively indicated that side stream
softening can maintain reduced system hardness in this blast furnace system
when operated with blowdown levels significantly lower than BAT rates.
Hardness levels in the 300-500 mg/1 range have been demonstrated to be
manageable with conventional water treatment schemes. This method of side
stream softening for hardness control can be viable with a properly designed
system. However, the levels of nonprecipitating ions such as sodium and
chloride may continue to concentrate to high levels which could subsequently
lead to potential corrosive conditions. To date, chloride concentrations up
to 2,000 mg/1 have been observed without indications of abnormally high
corrosion.
FUTURE WORK
There still remains unanswered questions with regards to this trial.
The time constraints previously mentioned limited verification of long term
effects. A trial running for several months would provide more insight and
allow for additional related studies of overall system effects such as
deposition, corrosion, scrubber and cooler efficiencies, and effects on
other BAT related water constituents such as ammonia, cyanide, phenol, lead,
and zinc.
-8-
-------
PROJECT IMPACT
COSTS
The cost of chemical treatment for this side stream softening trial are
shown in Figure 7 as compared to current BPT treatment costs and projected
alkaline chlorination costs. The current chemical costs are a base line
figure which includes clarification and deposit control agent currently in
use. The alkaline chlorination costs reflect single and two stage, and air
stripping with two stage, as outlined in Pilot Evaluation of Alkaline
Chlorination Alternatives for Blast Furnace Slowdown Treatment, Metcalf &
Eddy, presented at the U.S. EPA Symposium on Iron and Steel Pollution
Abatement Technology for 1981. Costs were calculated in Dollars per Million
Gallons Treated (Slowdown or Side Stream) and Dollars per Million Gallons
Recirculated, to put them into perspective with BPT treatment costs. These
figures indicate that chemical treatment for side stream softening by
caustic addition are in the same range as projected costs for alkaline
chlorination treatment of blowdown water to attain BAT limitations. Pilot
plant studies conducted in 1981 indicate a lower cost for softening by lime
and soda ash, which would inhance the chemical cost advantage of side stream
softening over alkaline chlorination.
TREATMENT SCHEME
CURRENT BPT
ALKALINE- CHLORINATION
SINGLE-STAGE
TWO-STAGE
AIR-STRIP W/TWO-STAGE
SIDE-STREAM SOFT
pH 9.5
pH 10.5
pH 10.5
SYSTEM
FLOW
4SOO
9000
9000
9000
4500
4500
4500
TREATMENT
STREAM FLOW
4500
300
300
300
250
250
150
COSTS
TREATED
28
1207
1253
936
B 25
1028
1000
$/MG
RECIRCULATED
28
39
41
30
46
57
33
Figure 7.
Republic Steel Cleveland District - Nos. 5 and 6 Blast Furnace
Recycle System - Treatment Scheme Cost Comparison.
Detailed figures for the capital costs or other operating costs of this
side stream softening method of treatment are not available. The total
costs of modifications to this recycle system to attain very low blowdown
rates were less than $0.2 million, and the costs for the modifications
required to run this trial were less than $50,000. The capital costs of
alkaline chlorination treatment systems have been estimated by Metcalf &
Eddy in their presentation to be $1.2-$1.3 million.
-9-
-------
BAT
The 5 & 6 Blast Furnace Recycle System effluent quality approaches the
recently promulgated BAT limitations by means of very low blowdown flow
without the requirement for additional blowdown treatment. Figures 8-12
show the general long-term trends for the regulated parameters on both a
gross concentration and loading basis. Note the loading curves are
asymptotically approaching zero for all of the parameters, despite the
corresponding increase in respective concentrations. Listed in Table 1 are
the BAT Effluent Limitations as promulgated, on May 27, 1982 for this
facility.
REPUBLIC STEEL CORP
CLEVELAND DISTRICT
5 8 6 BL FCE RECYCLE
I !
CC
UJ
O
en
en
§
(9
-TOTAL DISSOLVED
X
•^
^ ^ Hi
'...^
,
30,000
1
20,000
10,000
O
3
7S-T6 77
79
YEAR
si'"
82
,?2,
in
o
5
NOTES:
(I) EXCLUDES JAN 1*81 DATA
121 DATA FOR JAN -AU« l»2
Figure 8. Republic Steel Cleveland District - Nos. 5 & 6 Blast Furnace
Recycle System - Total Dissolved Solids gross concentration
and gross loading discharge by year.
100
1000
E
O
UJ
O
O
o
en
en
o
K
o
REPUBLIC STEEL CORP
CLEVELAND DISTRICT
586 BL FCE RECYCLE
40 -
20,
75-76 77
80 8I1"
NOTES:
(I) EXCLUDES JAN 1981 DATA
2) DATA FOR JAN-AUG 1982
Figure 9. Republic Steel Cleveland District - Nos. 5 & 6 Blast Furnace
Recycle System - Total Suspended Solids gross concentration
and gross loading discharge by year.
-10-
-------
r 100
E
O
E
Ul
o
(fl
NOTES:
-------
0.8 i
o>
1 1 1
REPUBLIC STEEL CORP
CLEVELAND DISTRICT
586 8L FCE RECYCLE
8.0
75-76 77
78
79
YEAR
80 8I">
NOTES:
(I) EXCLUDES JAN 1911 DATA
(2) DATA FOH JAN-AUO 198!
Figure 12. Republic Steel Cleveland District - Nos. 5 & 6 Blast Furnace
Recycle System - Phenol gross concentration and gross
loading discharge by year.
TABLE 1.
BAT EFFLUENT LIMITATIONS
Monthly Daily
Constituent Average* Maximum*
Suspended Solids 132 398
Ammonia(N) 15 44
Cyanide(T) 1.5 2.9
Phenol (4AAP) 0.15 0.29
Lead (T) 0.37 1.11
Zinc (T) 0.44 1.34
*Gross loads (Kg/day)
Comparing the monthly average effluent limitations as shown in Table 1 to
the long-term (yearly) average loading data as shown in Figures 9-12, the
system should achieve the BAT limitations for suspended solids and phenols
and approach these limitations for ammonia and cyanide. Lead and zinc
loading comparison to BAT could not be made since this data had not been
collected. It is possible that BAT could be achieved with the correction of
the existing hydraulic problem in the system. However, blowdown levels
would have to be controlled below 20 gallons per ton to assure compliance of
the BAT regulations based upon alkaline-chlorination. A method to ensure
this blowdown level will be installed at No. 6 Blast Furnace by utilizing
recycle system water as make-up to a slag pelletizing operation. Republic
expects that 30 gallons per ton can be evaporated by this operation and thus
eliminate the recycle system discharge.
-12-
-------
ACKNOWLEDGEMENTS
The authors wish to express their appreciation to Messrs. Mark Vanek, Bob
Phillips and Bill Mozes of Republic Steel Cleveland District for their work
and ingenuity in setting up and maintaining this trial. We would also like
to thank Messrs. John Lefelhocz and George Sheppard of Republic Steel's
Research Center for their efforts in coordinating and monitoring the trial
procedures; and express appreciation for the graphic and preparation support
of Messrs. Al Crespo and Tom Medvin and Ms. Bettye Lavender. Lastly, we
would like to acknowledge Mr. Mitch Kassouf of Betz Laboratories for his
help in providing the many services of his company during this trial.
-13-
-------
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of
the Agency and no official endorsement should be inferred.
COSTS OF PRE-TREATMENT OF COKE-PLANT EFFLUENTS
by: J. Ganczarczyk, Professor of Civil Engineering
University of Toronto, Toronto, Ont. M5S 1A4
(416) 978-3141, and
M. Kelleher, Consulting Engineer,
MacLaren Engineers Inc., Toronto, Ont. M5E 1E7
(416) 365-7337
ABSTRACT
One of the most common strategies for pre-treatment of coke-plant
effluents consists of wastewater equalization, followed by ammonia
stripping. Dissolved air flotation is a compact pretreatment method
which is worthy of consideration for some applications. The capital
and operating costs, and some technological aspects of each of these
three operations are :examined:_'in-.-this .paper.
Equalization, which is practiced at many plants and is always strongly
advised prior to biological treatment, is associated with capital cost
which varies linearly with the chosen detention period. The operating
costs are minimal. It is expected that adequate equalization periods
permit some cyanide conversion, good oil and tar removal, and other
aging phenomena which could make the wastewater more amenable to
biological treatment. Where equalization precedes ammonia stripping,
oil removal improves the effectiveness of the free leg of the still.
Dissolved air flotation removes most of the oil and tar and a portion
of the COD bearing material from coke plant wastewater. Although it does
not perform the function of equalization, dissolved air flotation can be
considered for effective oil and tar removal if the space required for
adequate equalization is not available. It has relativeiyllow.capital-.;.
costs, and low operating costs.
In comparison to the other pre-treatment methods discussed, ammonia
stripping is associated with relatively high capital and operating costs.
Removal of lime solids prior to the fixed leg of a conventional still
considerably reduces operating costs and improves the efficiency of the
unit. Many new ammonia stripping operations use caustic soda for
alkaline pH adjustment because of the numerous operating problems
associated with the use of lime. However, the operating costs of modified
lime systems are less than those of caustic systems for comparable ammonia
removal results.
-------
PRE-TREATMENT OF COKE-PLANT EFFLUENTS
The major objective of the pre-treatment of coke-plant effluents is
to make the wastewater more amenable to subsequent biological treatment and
- sometimes - advanced physico-chemical treatment. The pre-treatment of
coke-plant effluents is unusually complex; there are many pre-treatment
methods available, and numerous pre-treatment strategies are practised.
Generally, such pre-treatment includes wastewater stream separation or
effluent equalization with pre-sedimentation and optional flotation, phenol
extraction or sorption, and ammonia removal. Pre-treatment technologies
for cyanide removal are less common. The pre-treatment of coke-plant
effluents in North America is based on equalization and storage, followed
by free and fixed ammonia stripping. This technology is suggested as
Alternate 2 of the "best practicable control technology currently
available" (BPCTCA), as described by the Effluent Limitation Guidelines
for the Iron and Steel Manufacturing Segment of the Steel Industry.*
COST ESTIMATION METHODOLOGY
Costs presented in this paper are in 1981 Canadian dollars.
Equalization tank costs were based on average rates for average conditions
provided in "Yardsticks for Costing 1981", published by the Canadian
Institute of Quality Surveyors. The capital costs of flotation units
were budgetary quotations provided by three different manufacturers. The
information on the capital and operating costs of ammonia stripping stills
was obtained from a number of North American steel companies, manufacturers
of ammonia stills, and some literature references on the subject.
EQUALIZATION
Because coking is a batch process, the collection and equalization of
coke-plant effluents is a major pre-treatment step. Equalization periods
ranging from 0.5 to 7 days are common practice.**
During the equalization/retention operation some conversion of
cyanide to thiocyanate takes place. Effective oil and tar removal can also
be achieved. These technological features are dealt with in detail else-
where. In this work, attention was primarily directed to the cost of the
required tanks. Their design was based on the following:
* U.S. Environmental Protection Agency. Development Document of Effluent
Limitation Guidelines and New Source Performance Standards for the Steel
Making Segment of the Iron and Steel Industry, EPA-440/l-74-024a, June
1974.
** Ganczarczyk, J.J., Pre-Treatment of Coke Plant Effluents. Chapter 9
Waste Treatment and Utilization. Edited by Moo-Young, M. and Farquhar,
C.J., Pergamon Press, Oxford, 1980.
-------
(i) A flow of 750 m3/d (198,000 U.S. gal per day). The equalization
basins were sized for detention periods varying from one to ten days.
(ii) A single circular reinforced concrete structure was used for each
detention period. The use of two structures would be more practical,
but would be more expensive for the same equalization capacity,
and therefore was not analyzed here.
(iii) Designs were carried out for side water depths of 3.6 m (11 ft.) and
4.6 (15 ft.) for comparative purposes. A freeboard of 0.6 m (2 ft.)
was included in each case.
(iv) The basins would be below grade, with the top 1.2 m (4 ft.) above
grade, and backfilled with excavated material. If rock or poor soil
conditions requiring extensive dewatering were encountered, a buried
basin would not be practical, hence the cost of the basin structure
has been separated out for discussion purposes.
(v) It was assumed that the wastewater contains 1-2% flotable oil that
would flot to the surface of the basin, and that the wastewater
contains very little suspended, settleable material, hence
assumulations of sludge at the bottom of the basin would not be
significant. It was also assumed that skimming equipment might be
considered and its costs have been evaluated.
(vi) Pre-aeration or mixing of the contents of the equalization basin
were not considered appropriate for this type of wastewater. Both
of these operations are used in the pre-treatment of other
industrial wastewaters, where the mixing of different wastewater
streams is important to simplify subsequent treatment operations.
Either operation would interfere with oil and tar removal, which was
considered as an important function of the equalization basin.
(vii) For the sake of simplicity, any addition of chemicals to the waste-
water to enhance oil and tar removal and/or cyanide conversion was
not covered by this analysis.
The estimated costs of equalization basin and skimming equipment are
presented as a function of both basin Yolume and the wastewater detention
time in Figure 1. In Figure 2 the estimated costs of the equalization
basin and skimming equipment are shown as a function of the basin diameter.
A linear regression of these data indicates that for a basin side water
depth (s.w.d.) of 3.6 m (11 ft.), the cost of construction of a partially
buried equalization basin can be approximated as:
$29,540 + $30 (V)
(where V is the volume of the equalization basin in cubic metres.)
In general, the cost of excavation was between 15 and 20% of the total
cost of the equalization basin. The cost of the concrete structure varied
approximately .linearly with detention time, and can be approximated by the
equation:
$47,560 + $24.70 (v) -
-------
Costs estimated for basins with the same capacity, but with a side water
depth of 4.6 m (15 ft.) helped to determine the effect of basin depth on
costs. The surface area of the deeper basins was somewhat less, but the
depths of excavation and wall construction were obviously greater. Up to
a 5-day detention period, the cost of basins of equal capacity, with side
water depths of 3.6 m and 4.6 m are almost identical. The deeper basin has
a cost advantage at longer detention times and is approximately $20,000
less expensive than the shallower basin for a detention time of 10 days.
It has the added advantage of requiring slightly less space.
40
Co
c o : Ba
w1H
Ski
Eqiipm:nt • 369,000 +
$43.
40
y
20
to?
10
Eqiali2
ati(
n Bi sin
Skltmnir
g EC uipienC
atitn filain
witjh Sljimming
0 1000 2000 3000 4000 5000 6000 7000 8000
Basin Volume (cubic metres)
4 6
DecenCion Time (days)
10
Figure 1 - Capital Cost of Equalization Basins and Skimming Equipment
as a function of Basin Volume and Detention Time.
-------
40
30
20
D.
s
10
3.6
f
V
» Equalization Basin
" SWD
Skiiming £q'
Equ ilizatio
with Skimmihg
ipment
Basin
Equip dent
(3.6 m SrfD)
10 20 30 40 50
Basin Diameter (m)
60
Figure 2 - Costs of Equalization Basin and Skimming Equipment as a
Function of Basin Diameter
The cost of the concrete for the base slab of a basin accounts for
40-50% of the total cost of the structure. Other costs vary less
significantly with respect to varying depths, hence the extra cost of
deeper walls is more than offset by savings in the size of the base slab.
In general, deeper equalization basins should be used unless there is some
reason why they are not favoured.
The estimated costs of circular skimmers include a drive head, bridge,
walkway and handrails, centre column, drive and bridge drive cage,
feedwell, two-arm skimmer, scum trough, beach plate, V-notch weir plates
and scum baffle plates. These costs increase approximately linearly with
detention time. The relationship between cost and basin diameter is
approximately linear with a shallow slope ($1910 per metre increase in
diameter) for diameters of 36 m or less, and is linear with a steeper
slope ($3810 per metre increase in diameter) for diameters greater than
40 m. Cost increases more steeply for larger diameter skimmers, because
heavier structures are required for support of the equipment.
-------
The estimated cost of a circular equalization basin equipped with
skimmers can be approximated by the formula:
$69,000 + $43.40 (V).
The cost of the skimming equipment accounts for between 35% and 50% of the
total cost of the basins, hence it is a significant portion of the total
capital expenditure on this unit operation. Therefore, under normal
circumstances, skimming equipment should not be installed on equalization
basins until it is proven necessary.
The capital costs of equalization range from $46,800 for a basin of
volume 750 HP* (1-day retention) to $251,200 for a basin of volume 7500 m3
(10-days retention). This is equivalent to $62.40/m-^ for the small basin
and $33.50/m3 for the larger basin. The operating costs associated with
equalization are minimal. These would consist of $250-$750 per year
power costs, if skimming equipment were installed (which is rather uncommon),
and some allowance for sludge disposal. If the oily sludge is used as fuel
in the coke ovens, sludge disposal could be assumed as cost free. As the
technological advantages of wastewater detention times in excess of those
required for adequate flow and quality equalization are not yet quantified
in detail, a cost-benefit analysis cannot be carried out to compare the
technological improvements of longer retention with the increased capital
cost of providing this retention.
DISSOLVED AIR FLOTATION
The main purpose of dissolved air flotation would .be the removal of
oil and tar from weak ammonia liquor prior to both ammonia stripping, and
the biological "treatment plant. It is well known that this removal would
improve the effectiveness of ammonia stripping but the extent of "this
improvement has not yet been quantified to date. It is also well known
that oil and tar removal prior to -biological treatment -would improve the.-
performance of the biological p-l-ant. '.However, it is difficult to quantify
the effects of the extent of oil and tar removal on performance of
wastewater biological treatment. It was assumed in this study that coke
plant wastewaters contain adequate quantities of polymers which would aid
dissolved air flotation and also that the oily float would be recoverable.
If addition ... of a coagulant were required, it has been assumed that
pickling liquor, which contains ferrous salts, could be added to the
effluent prior to flotation to enhance oil removal and achieve some
cyanide precipitation at little or no cost. Stripping of volatile sub-
stances is expected to be of minor significance.
The capital costs of flotation units were provided by three
different manufacturers. A linear regression analysis of the data obtained
for the least expensive-, unit indicates that the installed costs can be
approximated as:
$25,903 + $1,811 (A)
(where A is the surface area of the unit in square metres) .
-------
In general, the capital cost of dissolved air flotation units is relatively
low, ranging from $30,000 for a unit of area 4.6 sq.m (50 sq. ft.) and the
capacity to treat 225-450 L/min (60-180 gpm), to $70,000 for a unit of
surface area 28 sq.m (300 sq. ft.) which could treat 1350-2700 L/min.
(360-720 gpm), depending on the design criteria. Based on a detailed
study on application of dissolved air flotation for clarification,* it was
calculated that an installed dissolved air flotation unit which could treat
750 cu.m. of wastewater over a 24-hour period, would cost only $33,000.
This is lower than the cost of providing one days equalization capacity
for the same flow, which was estimated at $47,000. However, dissolved air
flotation cannot provide the many advantages of equalization. For this
reason, dissolved air flotation should not be considered as a replacement
for equalization, but could be used to enhance some of the effects of
equalization if space for extended equalization is severely limited. If
possible, at least one days equalization capacity should be provided ahead
of the dissolved air flotation unit.
If it is assumed that dissolved air flotation requires little or no
labour input, that the addition of coagulants is not necessary for
effective flotation because of the nature of the wastewater and that the
scum and sludge from the operation can be disposed of on site at no
charge, then the operating costs of a dissolved air flotation unit consist
of (a) power costs and (b) maintenance costs, which manufacturers estimate
at 1% of capital cost annually. These costs are presented as a function
of the capacity of a unit in Figure 3 and can be approximated by the
linear regression equation.
Annual cost of power and maintenance = $468 + $2(C) (where C is
the capacity of the unit in litres per minute).
* Henry, J.G., and Gehr, R., Dissolved Air Flotation for Primary and
Secondary Clarification. Canada Mortgage and Housing Corporation.
Report SCAT-9. 1981.
-------
10
Operati ig Cost
• !468+ZZ(C)
900
1800 2700
Capacity C (L/min)
3600
Figure 3 - Operating Costs of a Dissolved Air Flotation Unit as a
Function of Design Capacity
AMMONIA STRIPPING
The basic technologies used for ammonia removal from coke-plant
effluents are steam stripping and ion exchange. The first approach is
broadly applied, but the second one is still at the development stage.
Coke-plant effluents contain free and combined ammonia. The free ammonia
may be stripped out by the application of steam, but to strip combined
ammonia alkalization of the wastewater by the addition of caustic soda
or lime is required.
The design of stills is based on: (a) the hydraulic flow to the
still, which determines the surface area of the unit, and (b) the degree
of ammonia removal required, which determines the number of plates or
trays in the still, and hence its overall height. The quantity of steam
used is a function of the influent temperature and the degree of ammonia
removal required. The quantity of steam required for a given percentage
of ammonia removal will decrease as the height of the still is increased.
-------
Because steam consumption accounts for a large portion of the operating
costs of ammonia stills, careful design balances the height of the still
and the steam requirements to achieve a least cost solution. The cost
of steam can vary from one location to another and, because of the high
capital costs of ammonia stills, construction of a high still to reduce
steam consumption is not always warranted.
There are many variations of the basic technology of steam ammonia
stripping. Some of these are: (a) free leg only, (b) conventional free-
and-fixed leg still, with application of lime or caustic, (c) the-
Bethlehem Steel pre-liming system and its modification. Detailed
discussion of the characteristics of each of these variations can be found
elsewhere.
The data collected on capital costs of ammonia stills are presented
as a function of the still design flow rate on Figure 4. There is a wide
variation in the costs quoted from different sources. Plants 1, 2 and 3
all operate at a wastewater flow of 163 cu. m/d, but the cost of the stills
vary by almost 100%. There are numerous reasons why this can occur. The
total installed costs of a still will include for lime handling equipment
in some cases, and not in others, where lime handling equipment was already
available on site. Also, construction costs will vary from one location to
another, and will depend to a certain extent on economic conditions at
the time of construction. If the location of the ammonia still is remote
from manufacturing centres, then this will be reflected in a higher
installation cost. Piping lengths, instrumentation and pump requirements
will vary from one installation to another, causing a variation in price.
Also, the operating flow of a still may differ from its design flow rate,
which would give a higher than expected cost.
It appears that the material for the still structure does not account
for a major portion of the total capital cost. The data labelled 7, 8, 9,
10 and 11 on Figure 5 were obtained from an ammonia still manufacturer and
are the manufacturer's estimates of the material cost for the still
structure, including the internal trays and the dephlegmator, but do not
account for the costs of labour, or any pumps, piping, chemical handling
equipment or instrumentation. In some cases, this can account for a
significant portion of the total capital cost of a still installation,
particularly if extensive piping, (which is very labour intensive) is
involved. Plant 8, which has an estimated cost of $750,000 for the
construction material for the still structure only, has the same capacity
as Plant 5, for which a total capital cost of $6M was quoted. In this
case, the material used in the still structure accounted for 12.5% of the
total capital cost of the installation.
-------
AL COpT
OF AMMONIA BllLLlS
V8 DESIGN [FLCJW pATlE
500 1000 1500
Design Flow Race C (L/min)
-Figure 4 - Capital Cost of Ammonia Stills as a Function of Design Flow
Rate
The total cost of installation of the ammonia still at Plant A in
1971 was $1,000,000. The designer of the still quoted the cost of
materials for the still structure itself, including the internals and the
dephlegmetor as $80,000. The balance of the cost included engineering fees,
labour and the cost of installation of the lime handling system, and all
other equipment, pumps, controls and piping related to operation of the
still. In this case, the material cost of the still structure accounted
for only 8% of the total capital cost quoted for the installation. The
data labelled W, X and Y give lower costs for a complete ammonia stripping
installation that the cost which were provided by a manufacturer for
construction materials for the still structure and the dephlegmator only,
hence these costs are discounted from this analysis.
A linear regression was carried out on the available data which
indicated that the cost of materials for the still structure itself can be
approximated by the equation:
Material Cost = $101,160 + $542 (C)
(where C is the design flow in litres per minutes).
-------
Other costs presented in Figure 5 include those quoted in a
Bethlehem Steel study*, and those of a plant running a caustic system.
The Bethlehem Steel System consists of overliming the still feed, which is
settled in a clarifier prior to introduction to the fixed leg of the still.
In this design, the free leg of the still is no longer necessary thus it
results in a capital cost saving. All costs have been converted to
December 1981 Canadian dollars for comparative purposes.
The costs of both the Plant A and Plant B caustic systems are
approximately $2.80 per cu.m of wastewater treated and that of the current
Plant A system, are slightly less, at $2.52 per cu.m. The costs for the
conventional lime system at Plant A were developed based on information
received from the plant operator, and include the increase in clean-out
costs associated with using a conventional lime system at this plant, but do
not take possible increases in steam requirements into account. The cost
of treatment with a conventional lime system is estimated at $3.88/cu.m of
wastewater treated. When Bethlehem Steel introduced the pre-liming system,
they quoted the cost of conventional lime treatment at $2.92 US/cu.m and
the cost of the pre-liming system as $1.68 US/cu.m. Converting these figures
to December 1981 Canadian dollars gives high costs for both systems.
However, a rate of inflation of 1% per month has been assumed during the
period June 1977 to December 1981, which may be unreasonably high, as the
costs associated with lime treatment may not have risen by 54% for that
period. The result of converstion to December 1981 Canadian dollars is that
the Bethlehem Steel figures for both a conventional lime system and a pre-
liming system are considerably higher than the Plant A figures for similar
treatment systems. The differences in cost can be partially explained by
the fact that some of the steam used in Plant A is free, whereas Bethlehem
Steel charged for all steam in its study. Also, the Bethlehem Steel
conventional lime system used a steam rate of 240-360 g/L of wastewater
treated, whereas the Plant A conventional lime system cost was based on a
usage of 200 g/L of wastewater. No information was available on steam
usage at Plant A prior to the installation of the clarifier ahead of the
fixed leg. It is possible that steam usage was higher with a conventional
lime system, however, this was not taken into account in the estimated
treatment cost which asssumed current steam usage rates. The major cost of
the conventional lime system in the Bethlehem Steel report is the cost of
steam, which accounts for 80% of the total operating costs.
The clean-out costs cited by Bethlehem Steel for the conventional
lime system are much lower than those incurred by Plant A, when it used for
a conventional lime system, and are slightly lower than Plant A current
clean-out costs after a conversion to the modified Bethlehem system, however,
* Rudzki, E.M., _e_t _§_!., An Improved Process for the Removal of Ammonia
from Coke Plant Weak Ammonia Liquor. Iron and Steelmake, June 1977.
-------
these costs are a small percentage of total operational costs. The
information in the. Bethlehem paper did not include for- the cost of operators
for the plant, or for the extra costs of disposing of the increased quantity
of lime sludge generated by the system which uses more lime than a
conventional system, hence the operating costs quoted in their paper do not
take into account some of the items covered in the Plant A cost breakdown.
The major advantage of the Bethlehem Steel system is the saving
in steam costs, as the study found that steam consumption was reduced by
55% by the design changes described previously. The Bethlehem paper states
that a steam rate of 130 g/L is adequate for their modified system, whereas
Plant A, which operates a similar system, has found that a steam rate of
200 g/L is required to achieve an effluent containing less than 100 mg/L of
ammonia. The difference in the designs of the fixed leg of both units may
account for the lower steam requirements.
It should be stressed that the costs of both chemicals and steam will
vary from one plant to another, so that it is sometimes difficult to compare
quoted operating costs directly.
CONCLUSIONS
(1) A certain level of equalization of coke plant effluents is absolutely
necessary, and can be considered either before and/or after ammonia
stripping. It is most effective before both ammonia stripping and
biological oxidation, as it would provide an influent of reasonably
constant quality to both of these units. In this work, the advantages
of extended equalization/retention periods have been discussed but
not quantified. If space is not readily available at a site, the
cost of large equalization/retention periods in terms of land usage
may not be justified. In general, the capital cost of
equalization/retention is relatively high.
(2) Dissolved air flotation is a compact system with reasonable capital
and operating costs, which can be used effectively with or without"
addition of chemicals to remove oil, tar and a portion of the COD
from coke plant effluents. This method of pretreatment would
improve the performance of both the ammonia still and the biological
system, but would not be justified after reasonably long equalization
periods. It is most appropriate at plants where space is severely
limited, and it can act as a supplement to equalization basins"with
low detention periods.
(3) At a daily wastewater flow of 750 cu.m the cost of dissolved air
flotation is equivalent to the cost of 1.5 days equalization
capacity, excluding land costs. At higher wastewater flows, the
cost of dissolved air flotation is less significant in terms of
equalization capacity. Therefore, the decision to consider
dissolved air flotation will vary with the size of plant in question.
-------
(4) At present, there does not appear to be a technological substitute
for steam ammonia stripping, which is the most feasible method for
removing large quantities of ammonia from wastewater. This method
has high capital and operating costs. Steam costs are the greatest
component of the operating costs of an ammonia still, therefore
research should be directed at optimizing still design to minimize
steam usage.
(5) The operating costs of modified lime ammonia stills are much lower
than those of conventional lime systems and appear to be slightly
lower than those of a caustic system. ,For this reason, if a
conventional lime system is in place at a plant, it is advisable
to install a clarifier to reduce the solids loading on the fixed leg
of the still and thus greatly reduce .the numberof cleanout operations
required. The investment required to modify the conventional lime
system would soon be repaid by the savings in operating costs.
However, at slightly higher operating costs, a caustic still may
prove to be a more reliable and efficient system than a modified
lime still.
-------
The work described in this paper was not funded by the U. S. Environmental
protection Agency. The contents do not necessarily reflect the views of
the Agency and no official endorsement should be inferred.
DEVELOPMENT OF A ROTATING-BIOLOGICAL-CONTACTOR FOR THE
REMOVAL OF AMMONIA FROM BLAST-FURNACE RECYCLE-WATER SLOWDOWN
by: K. C. Krupinski and P. H. Damon
United States Steel Corporation
Research laboratory
125 Jamison Lane
Monroeville, PA 15146
ABSTRACT
Proposed Federal regulations have established effluent contaminant
limitations on discharges from blast-furnace gas-cleaning recycle-water
systems. To meet the proposed limits for ammonia, the major contaminant,
the blowdown from these systems may require treatment. As a part of the
development of alternative processes for the discharge-water treatment, a
supported-growth biological method for the oxidation of ammonia was inves-
tigated. An extensive program with a pilot rotating-biological contactor
was conducted for about one year at the U. S. Steel Lorain-Cuyahoga Works
in Lorain, Ohio. The pilot unit contained commercial-sized 12-foot-
diameter disks and was operated over a wide range of influent water condi-
tions, with ammonia-nitrogen concentrations averaging 329 milligrams per
litre (mg/A). Reductions in ammonia-nitrogen concentrations to 1 mg/Jl were
achieved with the process demonstrating excellent resistance to simulated
failure conditions along with the potential for low maintenance and a low
level of operator attention.
BACKGROUND AND INTRODUCTION
During the manufacture of iron in the blast furnace, large quantities
of flammable gas are produced. The gas has a heating value of about 80 Btu
per cubic foot and contains about 25 volume percent carbon monoxide, 3 per-
cent hydrogen, 12 percent carbon dioxide, and 60 percent nitrogen. ' The
gas also contains entrained dust that must be removed before the gas can be
used as a fuel. The wet scrubbing used to clean the gas results in the
introduction of various contaminants, such as ammonia, into the process
water.
The Clean Water Act of 1972 mandated a two-phase reduction in the
amount of contaminants discharged with the blast-furnace gas-cleaning
process water. The iron and steel industry complied with the requirements
See References.
-------
of the first phase by recycling the blast-furnace gas-scrubber and cooling
water, thereby substantially reducing the amount of water and contaminants
discharged. The next phase in the contaminant-reduction program must be
implemented by 1984 and will require treatment of the blowdown from the
blast-furnace recycle-water systems. The gas-scrubbing and recycle-water
system is illustrated in Figure 1 .
Currently the Environmental Protection Agency recommends the use of
alkaline chlorination for the treatment of blast-furnace recycle-water
blowdown. However, alkaline chlorination is a costly and complex process
which has potential chlorine-handling problems. A much simpler and less
expensive alternative treatment would be the biological oxidation of the
contaminants, but the standard suspended growth process cannot be used
because the ammonia-converting organisms are slow growing and would be lost
from the treatment tank. In the present investigation this problem was
solved with the rotating-biological contactor (RBC) unit which provides a
support for the organisms.
An RBC consists of a series of closely spaced circular polyethylene
disks (media) fixed to a horizontal shaft mounted across a tank with about
40 percent of each disk submerged in the wastewater. The shaft is continu-
ously rotated at about 1.5 revolutions per minute and a biological slime
(biomass) grows on the media by using the contaminants present in the
wastewater as food. Aeration is provided by the rotating action, which
repeatedly exposes the disks to the air after contacting them with the
wastewater. Excess biomass is sheared off in the tank where the rotating
action of the media maintains the solids in suspension so they can be
transported from the RBC unit.
Although this process has been developed primarily for the oxidation
of ammonia, the major contaminant in blast-furnace recycle water, other
biologically oxidizable contaminants such as cyanate, cyanide, and phenol
can also be removed by the RBC process. However, contaminants which are
not biologically oxidizable, such as heavy metals, may require a separate
chemical treatment.
The bacteria responsible for the oxidation of ammonia-nitrogen (NH^-N)
to nitrate-nitrogen (NO3-N), collectively called nitrifiers, consist of the
genera Nitrosomonas and Nitrobacter . ' The oxidation of ammonia to nitrate
is a two-step sequential reaction as follows:
+ 4H+ + 20
2NO + 00 2
£. £, -J
Nitrosomonas oxidizes ammonia to nitrite (NC^") and Nitrobacter oxidizes
NO,' to nitrate (NO3~). The overall oxidation reaction can be written
-2-
-------
— • - — Blast Furnace Off-Gas
Recycled Scrubber Water
Solids -* ---
Hot Welt
Cold Well
i
t
i
i
t
Slowdown to Treatment
Figure 1 Blast-Furnace Gas Scrubber Recycle Water System
Clean Gas
For Fuel
Make-Up
-3-
-------
NH + + 202 ^N03~ + 2H + H2°
Assuming that the empirical formula for the bacterial cells is
, the overall synthesis and oxidation reaction is '
NH4+ + 1.83 02 + 1.98 H003~ -0.021 C^NO +
1.041 HO + 0.98 NO ~ + 1.88 H CO
From this equation it can be calculated that for each pound (Ib) of NH3-N
oxidized the following are required:
1) 7.14 Ib of alkalinity, as calcium carbonate (CaCOo), to
neutralize the acid produced.
2) 0.72 Ib of alkalinity, as CaCO,, to provide a carbon source
for bacteria growth.
3) 4.18 Ib of oxygen for the oxidation reaction (the oxygen is
obtained from the air via the rotating disk).
To develop design data and evaluate the costs of the RBC system, U. S.
Steel Research operated a pilot-scale RBC at the Lorain-Cuyahoga Works,
Lorain, Ohio for about 1 year. The results of the study are presented in
this paper.
MATERIALS AND EQUIPMENT
A single full-sized RBC unit, for which the specifications are summa-
rized in Table 1, was used in the study. The pilot unit contains the same
type of 12-foot-diameter polyethylene disks that are used in a full-scale
RBC. The only major difference between the pilot- and full-scale units is
the amount of surface area available for biomass growth.
The disk system used in the pilot unit is shown in Figure 2 before
installation in the 2800-gallon tank, and the fully installed RBC is
pictured in Figure 3 along with the RBC effluent clarifiers. The disks are
shown uncovered in Figure 3 but during portions of the pilot study the
disks were covered with polyethylene sheeting to prevent heat and water
loss due to evaporation. The RBC tank is divided into four cells, each
holding 700 gallons of water, which can be operated in a series of stages
or in parallel depending on the position of the valves located on the feed-
line manifold. The flow to the unit was controlled by metering pumps.
A typical flow pattern through the RBC is shown in Figure 4. The
first three cells were usually operated in series and the fourth ceil was
-4-
-------
TABLE 1. PILOT RBC SPECIFICATIONS
Media (disks): Diameter - 12 foot
Material - Polyethylene with carbon black
UV inhibitor
Cells: 4
Tank Volume: 2,800 gallons; 700 gallons/cell
Disk Rotation: 1.5 rpm
TABLE 2. RBC ANALYSES - FREQUENCY AND METHOD
Once a day
a. pH - Orion Model 701 Digital pH meter
b. Dissolved oxygen - Orion Model 97-06 Oxygen Electrode
c. Temperature - Tel-Tru Dial Thermometer, Range 0 to 180°F
d. Conductivity - Markson Model 10 Portable Conductivity Meter
e. Alkalinity - Titration with 0.03N HCl using brom cresol green - methyl
red indicator
f. Free cyanide*
g. Ammonia-nitrogen**
Twice a week**
a. Nitrite - nitrogen
b. Nitrate - nitrogen
c. Phosphate
d. Thiocyanate
e. Chemical oxygen demand
f. Cyanate
g. Cyanide, total
h. Phenol
i. Total suspended solids
j. Volatile suspended solids
Once a week
a. Hardness - Titration with ethylenediaminetetraacetic acid
•American Iron and Steel Institute Method.
"Methods for Chemical Analysis of Water and Wastes, U. S.
Environmental Protection Agency, EPA-6000/4-79-020, March 1979.
-5-
-------
Di* AMMnbly Died
in Pilot Rotating-
Biological Contactor
Figure 2
Pilot Rotating-
Biological Contactor
A - Drive Motor
B - Effluent Trough (Treated Water)
C - Clarifiars (Modified 55-gallon Drums)
D - Manifold Flow-Control Piping
Figures
- 6 -
-------
Blast-Furnace
Recycle-Water
System
Water Cell NC-.
Sodium Car- 1
bonate & 1
»^PhosphateJ
^S^^^X^ Stages
i Cell
pH Controller I- *•
,,
1
I
1st
•
I
t_
f"
2
1 1
2nd
A
_J
niarifia
~f
3
1 1
3rd
i
1
1 '
r
\
i
\
,
r
1
1
,
r
*-To Sewer
Figure 4 Schematic of RBC Flow Pattern
NH,-N CONCENTRATION, mg/t
500
400
300
200
100
I I l i I
NH,-N in Influent (to First Stage)
NH3-N in RBC Effluent (from Third Stage)
20 40 60
ELAPSED TIME, days
80
Figure 5 RBC Demonstration Test-NH3-N Concentration in RBC
Influent and Effluent (3-Stage Operation)
-7-
-------
operated independently. The pH of each cell was independently controlled
at 7.2 by means of a pH controller and a sodium carbonate solution pump.
Phosphoric acid was added to the sodium carbonate solution to provide
phosphorus, a required nutrient.
START-UP
To initiate the growth of the biomass, the RBC was seeded with sludge
from a coke-plant biological-treatment facility. The unit was filled with
plant-service water to which was added 10 milligrams per litre (mg/Jl) of
NH3-N as ammonium hydroxide and 100 gallons of the biological sludge. A
mixture of blast-furnace recycle-water blowdown and river water was then
metered to the RBC. The initial concentration of blowdown in the service
water was 10 percent, and as the biomass developed, the percentage of
blowdown was gradually increased to 100 percent over a three-week period.
During the first 15 days of start-up, a portion of the RBC effluent was
recycled to return any nitrifiers lost in the effluent and to enhance the
seeding process.
ANALYSES
Analyses of the RBC influent and effluents were performed on grab
samples except for influent NH-j-N, which was determined on a 24-hour
continuously composited sample. The only parameter continuously recorded
was pH; all other analyses were performed once a day, twice a week, or once
a week, as indicated in Table 2.
DEMONSTRATION TEST
An 87-day demonstration test was conducted during the pilot testing
program. The purpose of the test was to duplicate, as closely as possible,
the conditions that would be expected to exist in a full-scale RBC
facility. The RBC influent was taken from the cooling-tower hot well;
cells 1, 2, and 3 were operated in series; the disk assembly was covered
and insulated; and the influent flow was maintained at a fairly constant
rate. The NHo-N concentrations in the influent to stage 1 cell and
effluent from stage 3 cell are plotted in Figure 5, and the averaged test
data are summarized in Table 3.
During the demonstration test the NH3~N concentration in the RBC
influent varied between 190 and 520 mg/£ and the average NH3~N content of
the RBC effluent was 1.1 mg/Jl. No equalization of the RBC influent was
provided during the pilot study, but a tank providing concentration equali-
zation would be a requirement in a full-scale facility.
As shown in Table 3 the NH-j-N, thiocyanate, cyanate, and free cyanide
were almost completely converted to nitrate. The alkalinity of the
influent was consumed, and additional alkalinity as ^2(203 was added during
the demonstration test to neutralize the acid produced. The hardness did
not decrease, indicating no scaling or calcium precipitation, and the
dissolved solids increased because of the addition of NaCO. The
-8-
-------
TABLE 3. RBC DEMONSTRATION-TEST-DATA SUMMARY*
THREE-STAGE SERIES FLOW
Water Temperature, *F
pH, SU
Alkalinity, mg/Jl Ca003
Hardness, rag/i CaCO,
... umhos-3
Conductivity, •e= :
cm
Dissolved Oxygen, mg/t
Ammonia-nitrogen mg/1
Nitrite-nitrogen, mg/t
Nitrate-nitrogen, mg/t
Thiocyanate , mg/t
Cyanate, mg/t
Cyanide, Refractory, mg/t
Cyanide, Free, mg/t
Cyanide, Total, mg/t
Chemical Oxygen Demand, mg/t
Phosphate, mg/t
Phenol, mg/t
Total Suspended Solids, mg/t
Volatile Suspended Solids, mg/t
Total Dissolved Solids, rag/t
RBC**
Influent
101.
7.48
554.
1250.
6320.
1.6
329.
0.6
0.6
2.2
32.
0.5
0.5
1.
110.
0.4
0.06
76.
13.
3400.
RBC
Effluent
82.
7.21
35.
1270.
6840.
5.0
1 . 1
1.2
342.
0.01
0.5
0.2
0.01
0.2
56.
4.6
<0.006
14.
7.
5700.
*Values averaged over the 87-day test period.
•From cooling-tower hot well of blast-furnace recycle-water
system.
TABLE 4. RBC DEMONSTRATION TEST DISTRIBUTION OF AMMONIA-,
NITRATE-, AND NITRATE-NITROGEN IN THE RBC STAGES
Concentration, mg/t
NH3-N
N02-N
N03-N
Influent Stage 1 Stage 2
329. 138. 22.
0.6 46. 35.
0.6 157. 279.
Stage 3
(Effluent)
1.1
1.2
342.
-9-
-------
phosphate content increased because phosphoric acid was added to the Na2C03
solution to provide phosphorus.
The distribution of NH-j-N, NO2-N, and NO-j-N in the three stages is
shown in Table 4. Note the decrease in NH-j-N and NO2~N and the increase in
N03-N in stages 1 through 3. The higher concentration of NH^-N in stages 1
and 2 inhibited. the conversion of NC>2~ to NO3~ by Nitrobacter and NO2
accumulated. ' The distribution of the various forms of nitrogen in the
RBC influent and effluent is shown in Table 5. The data in the table show
that 96 percent of the nitrogen present in the RBC influent is derived from
NH3 (3% from cyanate and 1% from nitrite, nitrate, and cyanide), whereas
97 percent of the nitrogen in the RBC effluent is present in the form of
nitrate (2% of the nitrogen is incorporated into the biomass).
EFFECT OF MEDIA SPACING
Two types of plastic media are available from an RBC manufacturer—
standard-density with about 7/8-inch spacing between the disks and high-
density or nitrifying media with about 1/2-inch spacing. The standard-
density media are normally used in applications in which a thick biomass is
developed such as in the conversion of carbonaceous materials in a
municipal-sewage treatment plant. Because nitrifying bacteria do not
develop as thick a biomass as organisms that convert carbonaceous mate-
rials, the closer-spaced media can be used for nitrification. The nitri-
fying media provide 50 percent more surface area per given shaft length
than standard-density media; that is, a 26-foot-long shaft with 12-foot-
diameter disks will provide 100,000 and 150,000 square feet (ft2) of
surface area with standard-density and nitrifying media, respectively.
The pilot RBC came equipped with standard-density media and these were
used in cells 1, 2, and 3 during the entire test program. To determine
whether the nitrifying media are compatible with blast-furnace recycle
water, the media in cell 4 were replaced with nitrifying media about half-
way through the pilot study. The nitrification rates obtained with the
nitrifying media were comparable to rates achieved with the standard-
density media. At the end of the pilot study, sheets of standard-density
and nitrifying media were removed, and both were found to have a uniform
coating of biomass over the entire surface of the sheet. Therefore, the
nitrifying media performed satisfactorily and are recommended for use in a
full-scale facility.
SHAFT-WEIGHT INCREASE AND POWER CONSUMPTION
The shaft weight was measured about twice a month by using load cells.
The load cells were placed under the two shaft bearings, and the shaft
weight was measured with the disks rotating and static. The cumulative
shaft-weight over an 11-month period is shown in Figure 6; both rotating
and static weights are plotted. The weight of the disks and shaft
increased steadily from 5200 to 6800 pounds over the 11-month period. For
about half the period, cell 4 was operated with standard-density media,
then with high-density nitrifying media which contain 50 percent more
-10-
-------
TABLE 5. RBC DEMONSTRATION TEST DISTRIBUTION OF NITROGEN
(weight percent)
NH3-N
NO2-N
NO3-N
SCN-N
CNO-N
CN-N
Biomass
Influent
96.20
0.16
0.19
0.16
3.10
0.19
"
Effluent
0.32
0.36
97.33
<0.02
0.05
<0.02
1.94
TABLE 6. CHEMICAL ANALYSIS OF BIOMASS
(Dry Basis)
C, wt % 25.7
H 3.9
N 2.2
P 1.7
Fe 14.8
Ca 2.9
Na 0.7
Volatile Matter 49.4
TABLE 7. COMPARISON OF ELEMENTAL ANALYSIS OF BIOMASS
TO THEORETICAL COMPOSITION
Composition of Theoretical
Volatile Matter Composition
wt t wt t
C 52.0 51.7
H 7.8 6.0
N 4.4 12.1
P 3.4 2.6
0 (by difference) 32.4 27.6
-11-
-------
6800
6400
SHAFT
WEIGHT, 6000
Ib
5600
5200
Rotating
Disks
Nitrifying Media
Installed in Cell 4
l l l l I I l l i I
0 24 6 8 10 12
ELAPSED TIME, months
Figure 6 Increase in Pilot RBC-Shaft Weight
SHAFT WEIGHT, Ib
7000
6600
6200
5800
l
40 80 120
ELAPSED TIME, days
AVERAGE BIOMASS
THICKNESS, mm
160
0.51
0.40
0.28
0.19
200
Figure 7 Shaft Weight With Nitrifying Media in Cell 4 (Weight Measured
While Disks Rotating)
-12-
-------
sheets of plastic per given length of shaft than standard-density media.
The shaft weight increased about 180 pounds because of the extra sheets of
media, and this increase is shown in Figure 6. The shaft weight with
nitrifying media is plotted in Figure 7 and a line of best fit, which was
derived by using least-squares linear regression, is shown.
Also indicated on Figure 7 is the average biomass thickness calculated
from the weight increase and surface area of the media. Cell 1 media had
the thickest biomass, estimated to be about 1 millimetre (mm), but only an
average thickness can be calculated from shaft weight because all four
cells are on the same shaft.
The curves in Figure 6 show a steady increase in biomass growth during
the 11-month period, with no sign of a leveling off in weight gain.
Reports in the literature, ' indicate a maximum nitrifier thickness of
about 2 mm when the biomass growth rate and biomass sloughing rate are
balanced. Therefore, at the growth rates shown in Figure 7, the shaft
weight would show a steady increase for an additional two or more years. A
steady increase in the nitrification rate would not be expected because
photomicrographs of the biomass showed that the viable organisms are mostly
at the surface of the biomass (see section on biomass characterization).
The power consumption of the electric motor that rotated the shaft was
measured daily with a voltmeter and an ammeter. The average power consump-
tion remained fairly constant at about 1500 watts and showed no tendency to
increase, even though the shaft weight increased about 1600 Ib.
CHARACTERIZATION OF BIOMASS
Samples of the biomass were scraped off the disks and the chemical
composition was determined, Table 6. The biomass contained 91 percent
water, and 49 percent of the dry biomass was volatile. In Table 7 the
carbon, hydrogen, nitrogen, phosphous, and oxygen contents of the volatile
portion of the biomass are compared with the theoretical composition of
nitrifying bacteria; except for the nitrogen content the agreement is good.
Samples of the biomass were also sent to the University of Alabama in
Birmingham, Department of Biology, for characterization by electron micros-
copy. ' Electron photomicrographs of the bacteria responsible for
oxidizing NHo to NO2~, Nitrosomonas, and the bacteria that oxidize N02~ to
N0j~, Nitrobacter, are shown in Figures 8 and 9, respectively; note the
difference in membrane structures between Nitrosomonas and Nitrobacter.
A small section of media and biomass (about 0.5 square inch) was cut
from a disk and electron micrographic cross-sections (3750X) of the biomass
were taken at various depths. Figure 10 represents the edge of the biomass
farthest from the plastic media, and Figure 11 represents the biofilm
closest to the media, intermediate areas were also examined. An overview
of all the sections indicates that the population density of healthy,
intact bacterial cells appears highest on the outer edge of the biofilm,
Figure 10. The amount of cellular debris and degree of cellular disruption
-13-
-------
FipmS
Figures
Electron Micrograph
of RBC BiomMS
Nrtrotomonas Europaea
Electron Micrograph
of RBC I
Nil
0.1 jrni
0.1 KIN
, »
; '
.-
•'^ "-i ' '
Figure 11
Electron Micrograph of
RBC Bkmass Section
Farthest from Media
Electron Micrograph of
RBC Braman Section
CkMMtto Media
ion
- 14 -
-------
Figures
Figure 9
Electron Micrograph
of RBC BJomws
Nitrosomonas Europaaa
Electron Micrograph
of RBC Bwman
Nrtrobacter
0.1
0.1 Jim
v-
«-. •*.
'i
I
> ,*
-
I -
• %-*e*
* " •'"*
,i* t
' '
Figure 11
Electron Micrograph of
RBC Biomass Section
Farthest from Media
Electron Micrograph of
RBC Bmmass Section
CkMKt to Media
- 14 -
-------
appears to increase inward toward the disk surface. This is probably due
to the lack of availability of nutrients and oxygen on the inside of the
film.
CLARIFICATION OF RBC EFFLUENT
Over the period of the pilot study, the blast-furnace recycle-water
blowdown contained about 100 mg/& of total suspended solids (TSS) and about
15 mg/£ of volatile suspended solids (VSS), whereas the effluent from the
third stage contained about 26 and 12 mg/Jl of TSS and VSS, respectively.
In the event that the RBC effluent should require further clarification,
clarifier and pressure-filtration tests were conducted.
Because the suspended-solids content of the RBC effluent is variable,
a clarification test was conducted over a six-week period. During this
period, the effluent from cell 3 was diverted to a small clarifier constructed
from a 55-gallon drum. During this period cells 1, 2, and 3 were operated in
series (influent •*• cell 1 ->• cell 2 > cell 3 •*• clarifier ->• sewer) and after
steady-state operation was achieved, samples of the RBC influent and
effluent from each cell and the clarifier overflow were analyzed for TSS
and VSS. The results of this test are summarized in Table 8. The average
TSS concentration in the RBC influent was 170 mg/A, the TSS in the RBC
effluent (stage 3) was 13 mg/Jl, and the clarifier overflow contained 6 mg/A
of TSS; no flocculating polymer was added during this test.
Pressure filtration was also investigated; the tests were conducted in
a 4-litre Dicalite bomb filter press with and without filter aids. The
results of the pressure-filtration tests are summarized in Table 9. The
tests demonstrated that a filtrate containing less than 5 mg/JJ. of TSS can
be obtained at a high filtration rate of about 1.4 gallon per minute per
square foot for an 8-hour cycle.
It was found that the RBC effluent can be easily clarified. The
choice of a specific clarification method will be a function of the TSS
discharge limit and the type of sludge desired.
EQUIPMENT-FAILURE SIMULATION
During the pilot study, the RBC and ancillary equipment were carefully
maintained and operated; consequently there were no equipment failures of
long duration. In a full-scale facility with 10 or more RBC units,
failures of various systems lasting 12 to 24 hours may be possible.
Possible areas of failure are (1) loss of the sodium carbonate solution
pumps, (2) loss of disk rotation, and (3) loss of influent flow. These
possible conditions were simulated to determine their effect on the biomass
and nitrification, especially how the biomass recovered from the shock
caused by the failure.
The sodium carbonate solution pump was shut off for 24 hours to one
cell of the RBC, while influent flow and disk rotation were maintained.
The pH of the water in the RBC cell dropped from the normal 7.2 to 6.4 and
-15-
-------
TABLE 8. CLARIFICATION OF RBC EFFLUENT
Average Solids Concentration, mg/t
RBC
RBC Effluent Clarifier
Influent Stage 1 Stage 2 (Stage 3) Overflow
TSS* VSS** TSS VSS TSS VSS TSS VSS TSS VSS
170 16 30 12 16 5 13 6 6 2
•TSS - Total Suspended Solids
**VSS - Volatile Suspended Solids
TABLE 9. PRESSURE-FILTRATION TESTS ON RBC EFFLUENT*
Solids Concentration,
rag/ 1
Test Conditions
Test
No.
1
2
3
4
5
6
Pressure ,
psig
5
5
5
5
5
5
Temp . ,
OF
58
74
70
75
75
58
Filter
Aid**
Superaid
Superaid
Speedplus
Speedplus
5000
5000
Before After
Filtration Filtration
TSS
20
18
10
14
41
20
VSS
9
9
6
8
11
9
TSS
8
11
<0.5
2
6
4
VSS
2
7
<0.5
2
4
4
Filtration Rate***
4-hr
Cycle,
gpm/ft2
0.41
0.80
1.35
2.19
1.89
0.62
8-hr
Cycle,
gpm/ft2
0.24
0.152
0.80
1.37
1.07
0.35
•Using 4-litre Dicalite filter press.
••Filteraids (Dicalite, Grefco, Inc.)!
Superaid - low clarity, relative-flow-rate rating 1 (fast)
Speedplus - medium clarity, relative-flow-rate rating 7
5000 - high clarity, relative-flow-rate rating 20.5 (slow)
•••Extrapolated to 50 psig operation.
-16-
-------
the nitrification rate decreased about 50 percent during the outage. After
the pump was restarted, about 72 hours were required for the nitrification
rate to return to normal. The 24-hour outage was probably a "worst case"
simulation because the pH dropped significantly within one hour of the pump
failure and in a full-scale facility a pH alarm would alert the operator to
the problem.
The disk rotation was stopped for 23 hours, and about every 6 hours
the disks were rotated for four revolutions to keep the biomass moist (the
biomass will die if allowed to dry out). During the stoppage, the alka-
linity, pH, and NH-j-N all increased because of loss of nitrification, but
the effect was temporary and nitrification returned to normal after disk
rotation was resumed.
The flow to one of the cells was stopped for 36 hours, while normal
rotation and pH control were maintained. Once residual NH^ in the cell was
converted, nitrification ceased but resumed immediately when the flow was
restarted. Therefore, depriving the biomass of any food for 36 hours does
not appear to harm the bacteria. During the demonstration test, the final
stage received very little NHo for over a week with no apparent damage to
the biomass, which confirms the results of this failure-simulation test.
CONCLUSIONS
1. The pilot RBC successfully treated blowdown from the lorain-Cuyahoga
Works blast-furnace recycle-water system. During an 87-day demon-
stration test, the NH^-N concentration was reduced from 329 to 1 mg/£.
2. Both standard-density media and closely spaced high-density nitrifying
media were tested and both were satisfactory.
3. The shaft weight increased 1400 Ib due to biomass growth, but the
power requirement to rotate the disks did not increase.
4. The RBC effluent contained about 25 mg/£ of TSS and could be further
clarified by filtration or sedimentation.
5. Various equipment failures were simulated (loss of alkalinity control,
disk rotation, and influent flow) with no lasting effects on the
biomass.
-17-
-------
REFERENCES
1. The Making, Shaping, and Treating of Steel 9th edition, H. E. McGannon,
editor, United States Steel Corp., Pittsburgh, 1971.
2. L. D. Benefield and C. W. Randall, Biological Process Design for Waste-
water Treatment, Prentice-Hall, Inc., Englewood Cliffs, New Jersey,
1980, p. 218.
3. Process Design for Nitrogen Control, United States Environmental
Protection Agency, October 1975, pp. 3-2, 3-3, and 4-49.
4. A. C. Anthonisen, R. C. Loehr, T. B. S. Prakasam, and E. G. Srinath,
"Inhibition of Nitrification by Ammonia and Nitrous Acid," Journal
Water Pollution Control Federation, Vol. 48, No. 5, May 1976,
pp. 835-852.
5. C. P. C. Poon, ^t al_., "Upgrading with Rotating Biological Ctmtactors
for Ammonia Nitrogen Removal," Journal Water Pollution Control
Federation, Vol. 53, No. 7, July 1981, p. 1162.
6. K. Ito and T. Matsuo, "The Effect of Organic Loading on Nitrification
in RBC Wastewater Treatment Processes," Proceedings, First National
Symposium/Workshop on Rotating Biological Contactor Technology
(Champion, Pennsylvania, February 1980), Vol. II, University of
Pittsburgh, June 1980, p. 1169.
7. J. J. Gauthier and D. D. Jones, University of Alabama in Birmingham,
Department of Biology, private communication, 1981.
The material in this paper is intended for general information only. Any use
of this material in relation to any specific application should be based on
independent examination and verification of its unrestricted availability for
such use, and a determination of suitability for the application by profes-
sionally qualified personnel. No license under any United States Steel
Corporation patents or other proprietary interest is implied by the publi-
cation of this paper. Those making use of or relying upon the material assume
all risks and liability arising from such use or reliance.
-18-
-------
Although the research described in this paper has been funded wholly or
in part by the U.S. Environmental Protection Agency, it has not been sub-
jected to Agency review and therefore does not necessarily reflect the views
of the Agency and no official endorsement should be inferred.
EFFECTIVE OPERATION AND MAINTENANCE PRACTICES
FOR WASTEWATER TREATMENT SYSTEMS IN THE
IRON AND STEEL INDUSTRY
by: William F. Kemner - Assistant Director, Engineering Division
Richard T. Price - Environmental Engineer
PEDCo Environmental, Inc.
11499 Chester Road
Cincinnati, Ohio 45246
ABSTRACT
This paper presents some of the results of a current study of operating
and maintenance (O&M) practices for wastewater treatment systems in the iron
and steel industry. The purpose of the study is to document effective O&M
practices and how their application might affect permit parameters and
operational upsets. The ultimate goal of the study, which is funded by the
Industrial Environmental Research Laboratory (Research Triangle Park, North
Carolina) of the Environmental Protection Agency, is to produce an O&M
report that will assist both agency inspectors and plant personnel.
The wastewater generating processes selected for study were byproduct
cokemaking, ironmaking (blast furnaces), steelmaking (basic oxygen furnaces),
hot forming, and acid pickling. Selections were based on pollutant load-
ings, system complexity, control costs, and commonality among the majority
of integrated steel mills.
Information was gathered from several sources: a literature search for
wastewater-related O&M practices in the iron and steel industry, discussions
with state and regional agency personnel to identify major areas of concern
and effective O&M practices, review of agency files, discussions with waste-
water treatment equipment vendors, and discussions with industry representa-
tives .
Various steel mills were visited, and treatment plant operators and
environmental staff members were interviewed. The information produced by
these visits includes typically encountered problems and their solutions,
troubleshooting efforts, extent of operator training, efforts to minimize
the effect of operational upsets, and preventive maintenance practices.
The study will culminate in the preparation of a report intended to
provide a better understanding of wastewater problems in the iron and steel
industry, to help agency inspectors to be more effective in evaluating the
effect of O&M practices on wastewater treatment performance, and to provide
information that will assist plant personnel in practical and cost-effective
fine-tuning of their systems.
-------
INTRODUCTION
Because the depressed production levels in the iron and steel industry
made it more difficult than expected to schedule plant visits, the study is
not yet complete. Therefore, this paper describes some of the findings to
date and is limited to discussion of effective O&M management practices,
some problems with effective O&M practices for example treatment system
components, and sampling and self-monitoring data.
In general, the authors were favorably impressed with the operation and
maintenance of the 12 treatment systems examined to date. No system was
without its problems (varying in nature and severity) but none of the sys-
tems seemed inadequate.
Most of the effective O&M practices described in this paper are exam-
ples of actual practices observed at several steel mills. The information
was collected by interviewing treatment system operators and maintenance
personnel. The plant's environmental department made provisions for us to
visit the plants and coordinated the visits with various O&M supervisors and
operators. Every plant was most cooperative.
EFFECTIVE OPERATION AND MAINTENANCE MANAGEMENT PRACTICES -
TREATMENT SYSTEM GENERAL
Effective O&M practices at a wastewater treatment system are those that
can keep the system operating at its optimum (hopefully design) performance
with minimal downtime.
It is management's responsibility to initiate and maintain an O&M plan.
Although most of the following O&M practices (i.e., communication, operator
training, staffing, recordkeeping, preventive maintenance, and treatment
system auditing/evaluation) are probably implemented in some degree at all
plants, their effectiveness depends on the degree to which they are prac-
ticed, how carefully they are practiced, and whether they are fine-tuned
periodically (if necessary) to achieve optimum effectiveness.
Our plant visits have shown that the management organization for water
pollution control has much to do with successful communications and with
obtaining the cooperation necessary to solve problems expeditiously. Gener-
ally, the management structure at the plants visited was as follows:
0 The utilities (power and fuels) department is responsible for
operating the wastewater treatment systems.
0 The process department, which is responsible for the steel mill
process operations, assumes at least part of the responsibility
for delivering wastewater to the treatment systems.
0 The environmental department keeps all pertinent personnel informed
of current and expected water regulations that will have an impact
on the plant. This department serves as a go-between for the
plant's pollution control section of the utilities department and
the regulatory agencies and is responsible for all correspondence
with the agencies.
-------
Table I shows a breakdown of which departments actually operate or
maintain the process wastewater treatment systems at the plants visited to
date.
TABLE 1. DEPARTMENTS THAT OPERATE AND MAINTAIN WASTEWATER
TREATMENT SYSTEMS AT VISITED PLANTS
Utilities Process Environmental
department department department
Blast furnace
Coke plant:
Biological treatment
Ammonia still
Hot strip mill:
Scale pits
Clarification/sedimentation
BOF
0
2
1
1
4
3
11
M
1
1
1
1
4
0
1
2
3
6
M
1
1
2
4
3
2
13
0 M
0 0
Observation of the workings of these departments (particularly process
and utilities) indicates that pollution control needs are met most effective-
ly when the utilities department has both a maintenance staff and an oper-
ating staff. It also appears to be advantageous for some equipment that has
typically been operated by the process department (e.g., scale pits) to be
operated instead by the utilities department, or that a strong spirit of
cooperation exist between process and utilities departments. This is espe-
cially important when the operation of the control equipment is greatly
affected by the operation of the process equipment (e.g., coke byproduct
plants). For example, an effective situation at one coke plant involves a
byproducts engineer and a process supervisor (responsible for the biological
treatment system) working hand-in-hand. This arrangement is effective
because both are in the same department and have a common goal—the optimum
performance of the byproducts plant/biological treatment system.
When a utilities department does not have its own maintenance staff and
has to rely on maintenance help from production departments, there is compe-
tition for maintenance resources. If production needs are greater, chances
are good that the pollution control department will have to wait until
maintenance personnel are available.
An alternative to the utilities department having its own maintenance
staff is to have operators perform some of the day-to-day maintenance, such
as routine equipment oiling/lubrication and minor pump repairs. This ap-
proach seemed to work at one plant. With the burden of routine maintenance
taken off the maintenance department, they are able to make time for larger
repair jobs requested by the utilities department.
-------
COMMUNICATION
Proper communication is essential to effective O&M. Figure 1 illus-
trates effective and ineffective communication structures, both of which
were noted during the plant visits. In the effective system, all personnel
connected in some way to the generation of wastewater and its treatment have
good communication lines. In the plants visisted, most of the treatment
systems were operated by the utilities department, so the utilities super-
visor is a key person in making the communication system work.
In the example shown, strong two-way communication lines are required
between the utilities supervisor and his operators and maintenance personnel.
Frequent contact must exist between the supervisor and operators to keep the
supervisor posted on the day-to-day operation of the treatment systems. We
noted direct communication between operators and supervisors (face to face
or by telephone) anywhere from once an hour to once a shift. Indirect
communication (use of logs, records, computer terminals) occurred at about
the same frequency.
Effective communication allows the utilities supervisor to stay on top
of things and to make decisions more quickly when a problem arises. Frequent
communication seems to make the operators realize the importance of their
jobs. Only through effective communication ties between operators, super-
visor, and maintenance personnel can problems be brought to the attention of
the necessary parties. These open communication lines also allow discussion
and consideration of alternate solutions to problems.
The utilities supervisor also must maintain frequent and effective
communication with the lab to be kept abreast of analytical results. These
results are indicators of treatment system performance that can be communi-
cated up or down the management ladder as required.
Good communication lines also should be established between the utili-
ties supervisor and the environmental supervisor to keep the latter informed
regarding how well the plant's treatment systems are operating and whether
problems have arisen that have caused or could cause outfall excursions.
The environmental supervisor must also have an open line to the lab super-
visor and to the utilities supervisor. The latter needs to know how treat-
ment system problems can affect an outfall, how the regulatory agencies may
react to certain situations, and what new regulatory developments are occur-
ring that will affect him.
Another important communication line is that between the utilities
supervisor (or his operators) and the process supervisor (or his operators)
so that process problems or significant changes affecting water quality and
quantity can be relayed to the treatment system operating department as
quickly as possible. Intermediate action often can be taken at the treat-
ment plant to reduce the possibility of a treatment system upset (e.g.,
diversion of spilled material or slug to a holding basin).
-------
AGENCY
ENVIRONMENTAL
SUPERVISOR
PROCESS rC
SUPERVISOR
AGENCY
SUPERVISORx'l
LAB
RESULTS
PROCESS |-^»
SUPERVISOR
ENVIRONMENTAL |- -I
'sv
I
1 *'
M'
UTILITIES
j? ^y SUPERVISOR
POLYMER
REPRESENTATIVE
LAB
RESULTS
MAINTENANCE
PROCESS DEPARTMENT
OPERATOR
TREATMENT PLANT
INEFFECTIVE
MAINTENANCE OPERATOR
TREATMENT PLANT
EFFECTIVE
Figure 1. O&M communication structure.
-------
OPERATOR TRAINING
Proper and adequate training of operators is very important for effec-
tive operation of a wastewater treatment system. The operators need to know
0 How their systems operate and how to handle routine daily problems.
0 What pollutants are being removed and where they are destined to
go (sinter plant, landfill, recycle).
0 Where the treated water goes.
0 How the steel mill depends on them (directly for return of needed
water, or indirectly from potential regulatory action).
0 What daily routines must be followed.
0 What to do during emergency situations.
An effective operational practice at one plant was a personalized
examination prepared by the operating supervisor and given to the operators.
Although this test was not required by the plant or given by other operating
supervisors within the plant, the supervisor we interviewed was convinced
that it was necessary for his own assurance that the operators knew how to
handle essentially all conceivable situations. It should be mentioned that
requiring operators to take exams is not popular with the union; however,
there are ways to test operators within contract guidelines.
After talking to various operators, we were convinced that those who
were present at system startup and had the benefit of receiving thorough
training by the equipment manufacturers were good, well-trained operators.
Subsequent generations of operators seemed to have received less formalized
training or practically none at all.
An effective operational practice, therefore, is to ensure that new
operators receive the same level of training as the "startup" operators. If
this is not possible, a practice found to be effective is to have a process
supervisor (for the process whose wastewater is going to treatment) show the
operators the source(s) of wastewater and explain the process variations
that can affect the flow rate and the quality of the wastewater. Also, a
representative of the plant's environmental department should explain the
importance of effective operation and the role of the environmental depart-
ment.
Another useful aid that some operators had developed involved simpli-
fied schematics of their equipment and/or overall treatment scheme. These
diagrams were usually much less complex than those in the O&M manuals pro-
vided by the design engineers.
LOGS/RECORDKEEPING
Good logs and recordkeeping are effective O&M tools. Our plant visits
indicated that current practice varies from maintaining essentially no
-------
records to recording numerous parameters every two hours. It would be
difficult, if not impossible, to recommend a standard record sheet and
logging frequency for a particular treatment system because operations vary
greatly from plant to plant and system to system. The variables depend on
management philosophy, who the logs benefit, the necessity of the recorded
data, the degree of instrumentation for a given system, etc.
Nevertheless, it can be said that at those plants observed to have
well-operated systems the operators made a practice of logging in key oper-
ating parameters on a periodic basis. In addition, the operating super-
visors at these plants review the logs for abnormal conditions and initiate
followup action. In other words, the logs were not taken to lie idle.
Examples of log sheets (blanks) that were used effectively at two plants are
shown in Figures 2 and 3.
Collected data also are useful for plotting certain parameters versus
time to show trends. This can be very effective in showing how process
changes affect water quality of the influent to a control system, how con-
trol system changes or a new polymer affect outlet water quality, or how a
new equipment installation affects outfall water quality. It is frustrating
to search through reams of logs and records (Figure 4) to try to understand
what effect a change had on something.
The use of records to plot trends did not appear to be a common prac-
tice at any of the plants visited (even though the systems were well oper-
ated for the most part); however, such plots can be very effective in
illustrating a point to process, utilities, and environmental supervisors
and to regulatory agency personnel.
A process supervisor at one plant was able to use data that he had
plotted to convince plant management there would be fewer permit excursions
and less process equipment downtime if a preventive maintenance (PM) sche-
dule were followed. Figure 5 shows a section of the plotted data. For each
peak (excursion above control limit), the process supervisor reviewed main-
tenance work orders and was able to write down the cause of the excursion.
In doing this over an approximate two-year period, he saw a trend in equip-
ment breakdowns or pluggages vs. the excursions. This information enabled
him to prepare a preventive maintenance schedule (Figure 6), which was
implemented by management and has proved to be very effective.
Most plants have log sheets they have developed for their systems.
Although none was seen in the plants visited, another effective tool that
the authors have seen in use outside the iron and steel industry is the
inclusion of a control chart with maximum and/or minimum parameter values
shown on the log sheet so the operator can circle exceeded values. This
forces the operator to know what the acceptable operating range is; it also
provides his supervisor a quick visual aid for ascertaining where deviations
from the normal range have occurred. Taking this further, a deviation or
exception report can be prepared periodically (e.g., weekly or monthly) to
show the supervisor how the system has been operating.
-------
FILTER QPEMTQK
I—
r
11-T
•11
TIME
IOUIMENT
M
MKVtCI
•*t«ATINO O • T »
i!
E .
»ILT«« C««C QUALITY
8 ;
II
SUMP
•eCMCNTS CHAMOCD
MODS CHANCED
fTART OF TUHN
END or TURN
•
e
U
Figure 2. Example of log sheet - vacuum filter.
-------
o
A
T
I
I*
11
»
13
14
n
<•
IT
»
It
»
>1
n
n
14
n
M
IT
n
2t
JO
31
WASTE LIQUOR ANALYSIS M.IM • mv «M j
BIOLOGICAL OXIDATION PLANT
FEED
LKMJOft
*H
MM,
al
1MB-
CV«t
»J
torn
AERATION TANKS
SUSPENDED
SOLIDS PPM
SW
SE
NW
NE
SET.
TEST
%
SVI
VOLATILES
WEIGHT %
SW
SE
NW
NE
CLANIFIEN
OVERFLOW
pH
HH3
1-,.,.-.
IPP«"I
a)
lpp-»l
TMo.
CTMI
Ippm
1.1.
Wml
,
nix-
•KM
(••ml
Wsls^it
l«l
•.«T
IO«T*I
COMMENTS
Figure 3. Example log sheet - biological treatment system.
-------
WRONG
Figure 4. Effective use of records.
TREATMENT PLANT BULLETIN BOARD
RIGHT
-------
500
g 400
t-i
i
§ 300
S
o
f 200
100
CONTROL
LIMIT
PM PROGRAM INITIATED
10
15
20 25
WEEKS
30
35
40
45
Figure 5. Ammonia still performance.
-------
ITEM
TRAY NO. 1
TRAY NO. 2
LIME LEGS
SEPARATOR
SLAKER WATER
VALVE
CONTROL VALVES
NO. 1 pH
NO. 2 pH
RETURN LIQUOR
NO. 1 LEVEL
NO. 2 LEVEL
BIO. PLANT
LIQUOR TO
CLARIFIER
UNDERFLOW
OVERFLOW BOX
JAN.
A
A
FEB.
A
A
A
A
A
MAR.
A
4
A
A
A
A
APR.
A
i
A
A
A
A
MAY
A
A
A
t
A
JUNE
A
A
A
A
i
A
JULY
A
A
AUG.
A
A
A
A
A
A
A
SEPT.
i
A
A
A
OCT.
k
A
A
A
A
NOV.
A
A
A
DEC.
A
A
A
A
Figure 6. Example of annual preventive maintenance schedule.
-------
STAFFING
Obviously, proper operation of a treatment system requires a sufficient
number of operators (and assistants and helpers) with clearly defined job
functions.
The systems at all the plants visited under this study were well staffed
In fact, a few of the systems could have been properly operated with one
less operator. At one plant, however, an efficient operator had to devote
all of his time to manually controlling the treatment system (filtration
plant). Had this operator been responsible for areas other than the filtra-
tion plant, the filters would have plugged and the system would have been
damaged and/or solids in the filter effluent would have increased drama-
tically.
TREATMENT SYSTEM AUDITING
Results of this study indicate that most plants have proceeded from
permit negotiation to system design to routine operation without ever having
stepped back to take an overview of their situation. As in any ongoing
operation, the individuals involved are too busy with day-to-day concerns to
take time to assess their overall system. One person is aware of some of
the problems and another person knows of other problems, but they seldom
have the opportunity to exchange notes. A formal O&M audit had not been
performed in any of the plants visited, and none was planned. One company
that we have not visited is currently performing such a study on their own,
however.
We would suggest that audits or self-evaluation programs be implemented
periodically (e.g., annually) by a plant team or an outside consultant.
When the evaluations reveal less-than-optimum conditions, corrective action
should be taken with followup to ensure that approved suggestions have been
carried out.
EXAMPLES OF PROBLEMS AND EFFECTIVE O&M PRACTICES
WITH TREATMENT SYSTEM COMPONENTS
SCALE PITS
Although scale pits are one of the least sophisticated and low-profile
treatment technologies, they play a key role in the performance of hot strip
mill (HSM) treatment plants. The scale pit performs three functions:
1) initial removal of floating oil and scum, 2) equalization of flow and
temperature, and 3) sedimentation of heavy solids.
In every plant, excessive oil from the mill is a common occurrence.
The oil skimming arrangements often have difficulty keeping up. Even after
the skimming, oil containment and handling operations are major activities.
Apparently, the cost of oils, greases, and hydraulic fluids for the mills do
not yet justify the cost of spill and leakage control at the source. Treat-
ment plant operators are resigned to the oil problem and do the best they
can to cope with it. Oil carryover into deep-bed filters, for example,
decreases filter efficiency, increases backwash frequency, and inhibits good
cake formation in the vacuum filter.
13
-------
Equalization is an important function because the demand for water at
hot strip mills is highly variable, changing dramatically each time a slab
is rolled. Scale pit capacity must be large enough to dampen these varia-
tions while still holding as level as possible. If the level gets too high,
the pit may overflow to a blowdown; if the level is too low, the oil skim-
mers may be less effective (depending on design) and settling time is
decreased.
The solids removal efficiency of a scale pit is dependent on the effec-
tive retention time. If the pit fills up with scale, or if level is main-
tained too low, removal efficiency decreases and the solids removal burden
on downstream treatment modules increases. Several plants reported in-
creased solids loadings out of the pit during scale removal activities. It
is advisable to clean the pits during downtimes or have divided pits. Some
plants use flight conveyors, but high maintenance caused one plant to dis-
continue this approach in favor of clam shelling.
Adequate attention to scale pit operation is advisable to decrease the
burden on the remainder of the treatment system. In general, problems in
the front end of a treatment system tend to be magnified and become progres-
sively more difficult to rectify as they proceed through the system.
DEEP-BED FILTERS
On first consideration, deep-bed filters on hot strip mills appear to
be relatively simple and reliable. One such filtering system was observed
for about a day, however, and 22 O&M related problems were observed, as
summarized in Table 2. In fairness to the particular plant, the filter
effluent was crystal clear and was meeting permit parameters. The mislead-
ing part of this example showing 22 problems and few if no permit excursions,
however, was the operator attention required. The operator was constantly
busy observing changing flow conditions and pressure drops across the filters,
and manually controlling flows, backwashing cycles (scrub, high-rate rinse),
valves, and pumps. He also had to constantly monitor scale pit level. In
this system, the absence of a hard-working, above-average operator would
have been disastrous.
CLARIFIERS
One of the most common maintenance problems with a clarifier is a
plugged sludge line as a result of such items as hardhats, tools, and gen-
eral trash. When these items sink, a rake moves them into the sump at the
bottom of the clarifier and they plug the sludge line. Also, sludge that
becomes too dense will cause pluggage problems. The best way to minimize
this problem is to design the sludge lines so they can be cleaned easily.
Some facilities provide a tunnel around the sludge line so the maintenance
man will have access to the entire length of the line. Another technique is
to install two sludge lines so that if one becomes plugged, the other can be
used. At some installations a wire mesh or some other barrier has been
installed across the top or on the sides of the clarifier to prevent large
objects from falling in. Anytime the clarifier is drained for maintenance,
the sludge sump should be inspected thoroughly before the unit is put back
on stream to ensure that no miscellaneous items (e.g., nuts, bolts) were
left there by maintenance personnel.
14
-------
TABLE 2. PROBLEMS OBSERVED ON HSM DEEP-BED FILTRATION
1. Flow meters inoperative
2. One of 12 filters out of service over several months
3. Oil skimmers in scale pit unable to handle all the oil
4. Vacuum filter system inoperative (and no installed spare)
5. Frequent backwashing due to sludge returned to pit
6. Automatic backwash not functional—manual backwash required
7. Little communication between HSM and treatment plant
8. No log book maintained
9. Samples taken once every two weeks
10. Maintenance needs communicated verbally
11. No maintenance logs
12. Separate electrical and mechanical maintenance
13. System highly operator-dependent
14. Excessive oil from HSM into scale pit
15. No inspection of filters conducted
16. Possible for collected oil to overflow back into pit
17. Pit level control dependent on operator attention
18. No preventive maintenance program
19. More frequent plugging of filters closest to inlet—distribution problem
20. Good cake formation hindered by oil in vacuum filter
21. No spare parts inventory available to maintenance—kept by engineering
22. Lack of adequate communication; responsibilities divided among operations,
utilities, engineering, and environmental departments
15
-------
Debris in the clarifier can also jam the rake. Rake failure also can
be caused by mechanical problems such as a section of the rake falling off
or the rake getting out of balance and scraping the bottom of the clarifier.
If the sludge is not pumped away rapidly enough or its composition changes
so that it becomes denser, it can create enough resistance to stop the rake.
When the sludge is not removed, it builds up in the clarifier and eventually
overflows the weir, creating a dirty effluent. Also, the vacuum filter may
not operate well on dilute sludge.
Most clarifier rake drives are equipped with a torque meter set to
sound an alarm at a certain level (e.g., at 75 percent of full torque) and
will automatically shut down at a higher level to prevent the motor from
burning out. One facility installed a microswitch that recorded each time
the rake made a full revolution. This switch was wired to a timer, and if
the rotation time of the rake became excessive, it would automatically stop
the rake drive and sound an alarm. Another safety factor that is sometimes
built in is an automatic rake-lifting device. This lifting action will
sometimes free the rake from what was restricting its motion.
The condition of the clarifier feed can be checked with a simple settling
test. The test performed by most of the plants visited consists of putting
a sample of clarifier feed in a graduated cyclinder or Imhoff cone and
noting the sludge level after a certain time. At two plants, the clarifier
feed was considered to be satisfactory if the sludge level was between 25
and 30 ml. Although this test does not give an absolute measure of settling
rate, it is a good check for relative settling rates and will detect an
upset in the system. If a plant adopts such a test and takes this measure
every few hours, a problem in the clarifier feed can sometimes be detected
before it leads to a major upset.
The turbidity in the clarifier overflow should be checked frequently.
None of the facilities we visited had what they considered to be a satisfac-
tory continuous turbidity measuring device on the clarifier overflow. None
of the instruments tried was reliable. In the absence of such an instrument,
the operator should manually pull a sample of clarifier overflow and do a
lab check on the turbidity. If a satisfactory turbidity meter is found,
such manual checks should still be made occasionally to ensure that the
instrument is operating properly.
A visual inspection of the clarifier will tell the operator if the
sludge level is getting near the overflow weir. The operator should also
check to be sure the effluent flow over the clarifier perimeter weir is
uniform. If the flow over one portion of the weir is greater than over
another portion, the increased velocities can create turbulence.
If all the factors mentioned above have checked out and the clarifier
effluent is dirty, the clarifier may be operating at too high a feed rate.
The actual feed rate should be checked against the design feed rate. If the
feed rate and the area of the clarifier are known, the liquid rise rate can
be calculated. If the feed rate is excessive, there may not be a simple
solution. This condition usually results from the treatment plant having
been underdesigned initially or the occurrence of plant expansions after the
treatment plant was designed. Usually the only solution to this problem is
16
-------
to add another clarifier. If the treatment plant uses multiple clarifiers,
it may be that the flow is not being diverted equally to the clarifiers. In
this case, the reason for the uneven flow must be found and corrected.
A more subtle reason for poor clarifier performance may be a change in
the feed composition or a problem with the polymer being used,. It is apparent
that the use of polymers is both common and effective in many steel wastewater
treatment systems. Most plants leave much of the responsibility for the
polymer program (mixing and addition systems) up to the chemical supplier.
Whereas this practice appears to be satisfactory in most cases, periodic
formal review of the program is recommended, both to control chemical costs
and to review the effectivenes. The flocculation and settling performance
of the clarifier feed can be checked by using standard jar test apparatus.
These test parameters are approximate, and the test is most useful if a
benchmark test is available that was run during a period of good clarifier
operation. A bad polymer can be found by using this test to compare the
polymer being used,in the clarifier with a fresh batch of polymer made up in
the lab. After the settling rate has been determined, some of the liquid
can be siphoned from the jar for a suspended solids analysis. The suspended
solids level should approximate the clarifier overflow, although again this
is most useful as a comparative test against the benchmark test.
If the solids being settled in the clarifier come from a scrubber such
as those used for basic oxygen furnaces, the problem may be an increased
level of fines in the water due to the way the furnace is operating. These
fines, which are scrubbed out and sent to the clarifiers, make settling more
difficult. Settling can sometimes be improved by switching polymers.
Most clarifiers in BOF treatment systems are preceded by a (primary)
separator (e.g., cyclone). If the cyclone malfunctions or is bypassed, it
will increase the solids loading on the clarifier.
AMMONIA STILLS/BIOLOGICAL TREATMENT
Some of the effective O&M practices observed at two plants include the
following:
0 Caustic is now used in the ammonia stills at one plant rather than
lime. The still used to plug up frequently, creating fluctuations
in the ammonia concentration of the wastewater to the treatment
system. With the use of caustic, problems with the ammonia stills
are minor.
0 The lime plugging problems at the other plant were minimized by
improving the lime slaking operation. (The plant experimented
with caustic and found it to be very effective for good still
performance with good ammonia reduction; however, cost of using
caustic vs. lime was several times greater, so the plant decided
to use lime.) They try to maintain an optimum lime slaking tempera-
ture of approximately 95°C, and have a lime crush size of minus
1/4 inch (75-80% through screen). They also try to maintain a
purchased lime quality of >95% calcium oxide. The lime slurry is
continuously cycled from the lime slaker tank to the ammonia
still. A bleed line from the loop to the still supplies the lime
to the stills as needed.
17
-------
0 Large tanks preceding the bio-oxidation basins serve as equaliza-
tion tanks at one plant and as surge tanks at the other plant.
Lime carryover from the still settles in the surge tank and is
periodically pumped out with a vacuum truck. The level in the
equalization tank is recorded continuously on a strip chart.
0 At both plants, conditions at the bio-oxidation basin are monitored
frequently to observe any changes that could cause system upsets.
Phenol, ammonia, pH, thiocyanate, and total suspended solids are
sampled and analyzed at least daily. Both plants maintain their
basin temperatures between 26° and 30°C, not allowing it to vary
more than 2° or 3°C during any season. One plant logs aeration
basin temperatures every two hours throughout each day; the other
plant measures temperature continuously. Temperatures are adjusted
with coolers or live steam as required. Influent flow is also
measured and recorded.
0 Phosphoric acid is used as a nutrient in the aeration basins. The
amount used is adjusted to give a desired phosphate residual.
(The plants reported 1 to 10 ppm.)
0 Both plants add service water in amounts required at the basin
influents to equalize the inlet concentrations to get maximum
performance from the microorganisms.
0 Normal foaming that occurs in the biological treatment systems is
controlled at one plant by adding an antifoam agent at each aerator.
Water sprays are used to break up foam at the center well of the
clarifier (following the bug pond) at the other plant.
0 An amp meter is used effectively at one plant to indicate clari-
fier rake amperage. It is normally 25 to 30 amps and is checked
on a log sheet every two hours. Deviations from this range are
cause for quick action.
0 Sludge recycle rates are monitored closely and solids settling
tests (with Imhoff cones) are performed every two to four hours at
the plants. No sludge is wasted at either plant.
Some conditions that could cause upsets to the biological treatment
system are as follows:
0 Low phosphate residual level due to pump malfunction, inadequate
amount of phosphoric acid added, or frozen lines.
Excessively high ammonia concentrations into the system because of
a breakdown in the ammonia still.
Loss of dilution water. This could result from pump malfunction
or from experimenting in optimizing the amount of dilution water.
Change in temperature above or below optimum range due to fouling
or breakdown of heat exchangers.
18
-------
0 Loss of an aerator. If prolonged, this can cause bugs to die as a
result of low dissolved oxygen. This also causes the water tempera-
ture to rise, as the aerators provide cooling.
Effective O&M practices at the plants have minimized the potential of
upsets; e.g., frequent checks on phosphate residual, good communications
between ammonia still operator and treatment plant operator, constant checks
on temperature, and monitoring of aerator performance.
SELF-MONITORING DATA
Another aspect of the study included analysis of self-monitoring data
for NPDES permits. Data were collected on approximately 35 steel mills, an
average of roughly two years of data at each mill. The excursions and
reported causes were usually in letter form, which had to be reduced to
summary quantitative form. Figure 7 shows the frequency of upsets by pollutant
The data show that the three most troublesome excursions involve oil, solids,
and pH. This is at least partially attributable to the fact that they are
the most commonly monitored. The qualitative reasons for excursions were
categorized based on judgment selections of a reasonable number of categories.
The resultant data are shown in Table 3. The most frequently cited causes
are generally categorized as leaks, housekeeping, and operator error. We
hope to subdivide this category further into more specific causes later in
the project.
Our evaluations of self-monitoring data have led us to conclude that
plants freely report excursions, causes are often not known or reported,
frequency of excursions exceeds that contemplated by guidelines, inadequate
design is rarely cited as a cause, and there is no single dominant cause for
excursions.
Nonrepresentative sampling is given as a cause for permit excursions in
15 percent of the cases examined from noncompliance data at state offices.
TABLE 3. REPEATED CAUSES OF EXCURSIONS, BASED ON SELF-MONITORING DATA
Percent
Leaks/housekeeping/operator error 35
Non-representative sampling 15
Equipment malfunction 11
Maintenance/shutdown 10
Hydraulic overload (mostly rain) 6
pH/polymer control 4
Freezing conditions 4
Sludge dewatering/handling 3
Process variation 3
Design 1
Other 8
This includes samples taken at the wrong location, contaminated samples, and
inadequate or faulty analytical methods. Although this is the second highest
cause of excursions in the self-monitoring data, none of the plants visited
cited any of the above factors as a major problem.
19
-------
<_>
X
JOU
320
280
240
200
160
120
an
40
n
••^
»s=
Ul
CO
0
5
_J
•—I
o
CO
o
1— 1
o
CO
Q.
CYANIDE
»— 1
—
2. """"
-D ar uj UJ
f- o ° Q
Jj QC 3 ^-t •-< Ul
f*i !• • *j ^j ^5 ^3 ^"^ ^™
Figure 7. Frequency of excursions, by pollutant, based on self-monitoring data.
-------
SAMPLING AND MONITORING
Measurement of flow and maintenance of automatic samples, although
traditional problems in the wastewater field, were not cited by the plants
as a major area of concern. Flow monitoring is generally not accurate,
however. At one plant, none of three relatively new flow meters was working,
but this was an exception.
In general, the operation and maintenance of the treatment system take
place at one level and the NPDES monitoring and reporting take place at an
entirely different level, with very limited interaction. In most cases, the
operators have developed their own sampling and monitoring programs and key
indicators for their own use in running the plant. The actual outfall
monitoring and permit matters, however, are conducted by the environmental
department. The two activities go on rather independently, and it is diffi-
cult to relate O&M practices quantitatively to permit upsets.
The high maintenance requirements of pH probes and on-line analytical
monitors seem to be universal. Some plants use frequent (once or more per
shift) grab measurements of pH and do not really rely on the pH instrumenta-
tion. Some attempt is made to maintain, clean, and repair pH probes on a
frequent basis (once/2 weeks), but this is a continuing battle. At one
installation, two pH probes are used and an alarm is activated based on a
differential signal. The probes are then examined to determine which is
faulty.
On-line analyzers, such as those for ammonia, cyanide, and oxygen
reduction potential (ORP), suffer the same shortcomings as the pH systems.
Most were not working, probably because they are such a high maintenance
item and the plants seem to meet permit conditions without them.
Bubbler tubes for level control are used in many locations and apparent-
ly work well with little maintenance.
In general, it appears that sophisticated monitoring and instrumen-
tation equipment does not work reliably without unrealistically high mainte-
nance requirements. On the other hand, most plants have found simple reli-
able monitoring techniques that enable them to meet their permit conditions
without relying on the sophisticated instrumentation.
CONCLUSIONS
The general highlights of the plants visited to date are the high
degree of operator know-how and job interest, the high degree of system
monitoring to maintain performance, apparent concern and interest in pollu-
tion control, good attention to visual inspections, well-staffed depart-
ments, and cooperative attitude and willingness to share information.
The general areas of concern are lack of communication between process
operators and water treatment operators; competition between treatment plant
and production process for maintenance resources; performance of pH monitor-
ing, on-line analyzers, and flow measurement devices; hydraulic balance;
excessive oil from rolling mills; high maintenance required on lime-based
ammonia stills; debris in clarifiers/tanks/open-top containers; and lack of
detailed maintenance records.
21
-------
This report will produce more conclusions and data as it continues for
another two months and one or two more plant visits are made. All of the
current information plus the newly collected data will be evaluated and
included in a final report at the end of January.
ACKNOWLEDGMENTS
Without identifying plant names, as was agreed upon from the onset, the
authors would like to express their sincere appreciation to the corporations
that participated in this study. All have been very cooperative in providing
up to a week of time and usually involving several people. The collected
information is invaluable to the success of the project and will be used to
generate a final report intended to benefit both plant and agency personnel.
22
-------
DISCLAIMER
Although the research described in this paper has been funded wholly or in
part by the U.S. Environmental Protection Agency, it has not been subjected
to Agency review and therefore does not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
NITRIFICATION KINETICS AS INFLUENCED BY COKE PLANT .WASTEWATERS
by: Ronald D. Neufeld, PI
Professor of Civil Engineering
University of Pittsburgh
Pittsburgh, PA 15261
Jeffrey H. Greenfield
Graduate Research Assistant
University of Pittsburgh
ABSTRACT
The overall objective of this research is to conduct basic
studies into possible causes of biological nitrification process
instability as often observed in steel industry wastewaters,
and in a longer term, to propose rational and pragmatic process and
operational alternatives for nitrification in two sludge and one-
sludge combined systems. The experimental approach taken to date
is to evaluate the influence of elevated free ammonia levels, pH,
elevated temperatures, cyanides, and certain trace organics that tend
to pass through the carbonaceous reactor zone on the kinetic parameters
that quantify nitrification. Theoretical calculations based on
laboratory defined parameters are made which yield suggestions as
to allowable concentrations of trace inhibitors, and operational
strategies for stable nitrification of coke plant waste waters.
INTRODUCTION
There are many "advanced wastewater" treatment processes based on
physical-chemical and biological techniques for the removal of wastewater
ammonia, however, usually these processes are not often employed in industry
due to economics, marginal overall! removal efficiencies, and questionable
technological application to specific industrial operations. Of the
processes available for the removal of ammonia from steel industry waste-
waters, engineered biological systems are the most pragmatic. Data from
steel industry and other industrial operations indicate that a properly
designed and operated biological nitrification facility is capable of
reducing ammonia levels to less than 10 mg/L over extended periods of
1
-------
time. Biological nitrification processes, however, are known to exhibit
unaccounted for upsets, and thus are considered in some industrial sectors
as unreliable.
This research investigated possible causes of biological nitrification
process instabilities as currently observed in coke plant and other
industrial plant wastewater treatment facilities. Nitrification process
instability can occur as a result of toxic inhibition as caused by an inter-
action of organics, free ammonia, elevated temperatures, pH, complexed
and free cyanides, phenolics as well as by nonbiodegradable organics that pass
through carbonaceous removal stages.
From laboratory data obtained on studies of nitrification process
instability, conclusions with engineering calculations are presented to
illustrate design and operational considerations enabling industry to meet
EPA effluent restrictions for ammonia-nitrogen and phenolic compounds. The
conditions utilized in this study were aimed at specifically assisting the
coke plant segment of the Iron and Steel Industry.
Table 1 summarizes the BAT limitations, a technology level to be
implemented by July 1, 1984 and the prior BPT effluent limitations for the
Iron and Steel Industry and may thus be considered as the goal to be met
by the industry.
OVERVIEW OF RESULTS
Experimental results to date show that unionized ammonia acts as a
toxic inhibitor at levels in excess of 10 mg/L unionized ammonia. Quantifi-
cation of the influence of unionized ammonia on biokinetics yields an
expression similar to classical substrate inhibition models. Suggestions
are made as to pH controls for systems treating high levels (greater than
250 mg/L) of total ammonia or systems experiencing large variations of
total ammonia. (1).
Nitrification biokinetics are highly sensitive to elevated temperatures
with rates of nitrification increasing as temperature increases to a
maximum at about 30 C, beyond which the overall rate of nitrification
decreases. Suggestions were made as to a pragmatic temerature range to
assure stable nitrification in commercial sized facilities. (2).
Free cyanide (CN ) was the most toxic inhibitor investigated. Con-
centrations of free cyanide greater than 0.11 mg/L must be avoided for
stable operation of biological nitrification processes. (3).
A comparison of the toxic effects of coke plant tar acid phenolics
showed that this substance was of greater toxicity to nitrifiers than
reagent grade phenol on an equal phenol concentration basis. The apparently
more severe toxicity of the coal tar acids may be due to a few substances
-------
TABLE 1
EFFLUENT LIMITATION GUIDELINES FOR THE BY-PRODUCT COKEMAKING
SUBCATEGORY OF THE IRON AND STEEL INDUSTRY(a)
BAT EFFLUENT LIMITATIONS
POLLUTANT
PARAMETER
Ammonia-N
Cyanide
Phenols (4 AA-P)
Benzene
Naphthalene
Benzo(a)pyrene
MAXIMUM FOR ANY 1 DAY
AVERAGE OF DAILY VALUES
FOR 30 CONSECUTIVE DAYS
Discharge Load
(Kg/kkg)
(lbs/1000 Ib)
0.0543
0.00638
0.0000638
0.0000319
0.0000319
0.0000319
Discharge
Cone.
mg/l(b)
65.1
7.6
0.08
0.04
0.04
0.04
Discharge Load
(Kg/kkg)
(lbs/1000 Ib)
0.0160
0.00351
0.0000319
Discharge
Cone .
mg/1
19.2
4.2
0.04
BPT EFFLUENT LIMITATIONS
TSS
Oil and Grease
Ammonia-N
Cyanide
Phenols (4 AA-P)
pH within the range
of 6.0 to 9.0
Discharge Load
(Kg/kkg)
(lbs/1000 Ib)
0.253
0.0327
0.274
0.0657
0.00451
Discharge
Cone.
mg/1 (b)
303.4
39.2
328.5
78.8
5.4
Discharge Load
(Kg/kkg)
(lbs/1000/lb)
0.131
0.0109
0.0912
0.0219
0.00-150
Discharge
Cone.
157.1
13.1
109.4
26.3
1.8
Notes: (a) Federal Register, 1982
(b) Discharge concentration (for comparitive purposes) are based on
• effluent flows of 200 gal/ton of coke for BAT and BPT
-------
which may be in the liquid matrix, or a synergistic approach representing
combinations of large numbers of substances as being toxic to nitri-
fication. (4).
Experimental Approach
Nitrification biokinetics has been shown to follow classical Monod
kinetics of the form:
v = V S/ (K +S) (1)
max m v
where: V = rate of nitrification (t )
S = ammonia level in solution (mg/L)
V , K = system constants
max m
In addition, for biological systems, the following system equation
may be derived:
1/-Q- = av - b (2)
-6- = sludge age (days)
Where a = observed yield coefficient (g/g)
b = endogenous decay coefficient (day )
The approach taken in this research, is to quantify the influence of
inhibitors typical of coke plant wastewaters on the biokinetic parameters
for nitrification. Toxic or biokinetic inhibition tends to alter the shape
of the effluent ammonia vs. sludge age design curve which results from the
combination of the above two equations.
/
The final form of data from this research are summary plots of the
influence of parameters typical of coke plant waste waters on the biokinetic
constants of nitrification, and on the system design curve for process
nitrification. The conditions considered are free (or unionized) ammonia
which acts as a "substrate inhibitor", free and complexed cyanides and
thiocyanate, phenol and tri-methylated phenolics, pyridine, and elevated
temperatures. For comparison purposes, a sample of coal tar acids was obtained
from a local coke plant which, as indicated above, was found to be more toxic
to nitrifiers when compared on an equal phenol (measured by 4-AA) basis.
-------
Results of free ammonia inhibition have been published elsewhere, and
will not be included in this paper (1). Details of trace substance inter-
action, and temperature interactions are included in prior reports to
AISI (2,3,4). These results are summarized below.
Experimental Results of Elevated Temperatures on Nitrification Biokinetics
The experiments conducted were developed to quantify the influence
of elevated temperatures on the defining parameters at a pH of 8.0. It
should be noted that the data developed were for nitrifiers that were
acclimated to the elevated temperature in continuous cultures; the data
does not reflect the shock effects of temperature, but rather does reflect
steady-state bio-kinetics for essentially pure nitrifiers.
Figure 1 is a summary plot of maximum nitrification rate (Vm) as a
function of temperature. This Figure shows that at a pH = 8.0, Vm is
constant at 1.256 gNH /gVSS-Day in the range of 22°C and 30°C and then
decreases to a value approximately equal to zero at T = 45 C. This
decrease in Vm was found to be described by the equation:
Vm = 3.78 - 0.084 (T) (3)
for 30°C > T°C 2. 45°C
The numerical value for the Michaelis-Menten parameter "K " was also
evaluated with temperature as shown on Figure 2. In a qualitative sense,
as the K value decreases, the kinetic curve is pushed to the left indicating
an increased rate of nitrification at a constant design value of ammonia
level in the effluent. As the K value increases, the kinetic curve
is pushed to the right, indicating a relatively lower rate of nitrification,
thus requiring an increased sludge age for stable operations at the same
effluent ammonia level.
From the experiments conducted, the value of K was found to decrease
in the temperature range of 22 C to 30 C in accordance with the equation:
Log (K ) = 1.53 - 0.0315 (T) (4)
and at temperatures greater than 30 C, K was found to follow the
expression:
Log (K ) = -1.8829 + 0.08228 (T) (5)
where: T = temperature, ( C)
K = constant, (rag/L ammonia)
A similar evaluation was made of observed yield coefficient. For the
temperature range of 22°C to 30°C, this coefficient was a constant as:
-------
(a) ob = 0.063 Ib VSS/lb NH3 removed (6)
and for temperatures in excess of 30 C, the yield coefficient follows a
relationship of:
(a)Qb = 0.1935 - 0.0043 (T) (7)
It should be noted that the observed yield coefficient approaches zero
when the wastewater temperature is equal to 45 C.
From the biokinetic and sludge yield parameters obtained, calculated
data of sludge age and effluent quality at pH = 8 were obtained. A
family of such curves for the temperature ranges considered in this study
is shown on Figure 3.
In order to best interpret Figure 3, a cross plot was made of
required sludge age to meet ammonia effluents as listed as a function
of varying temperatures. This cross plot, shown on Figure 4 points out
that stable nitrification producing an acceptable effluent quality should
exist at sludge ages of about 20 days when the wastewater temperature is
less than 30 C (86 F) , however, instability may exist at temperatures above
that level with the curve becoming very steep (indicating instability)
at temperatures in excess of 33 C (91.4 F).
Experimental Results on the Influence of Organics and Cyanides on
Nitrification Biokinetics
The chemicals selected for this phase of research are those which have
been shown to pass through first stage bioreactors or those chemicals found
in coal conversion wastewaters which are significantly resistant to
microbial degradation. The substances considered are 2,3,6 and 2,4,6 tri-
raethylphenol (TMP), 2-ethylpyridine, phenol (comparison), free cyanide,
complexed cyanide, thiocyanate, and a mixture of coal tar
acids derived from a local coke making facility.
As an example of some of the results to date, Figures 5 and 6
are plots of Vmax versus concentration of 2,3,6-trimethylphenol and
K versus concentration of 2,3,6-trimethylphenol. Figure 5 shows that
the critical concentration of 2,3,6-trimethylphenol is about 4.9 mg/L, with
levels above that severely depressing the biokinetic parameter of the
V term in accordance with the equation:
max
Log V = 0.853-1.117 Log (2,3,6-TMP) (8)
013.X
Figure 6 indicates that the critical concentration for
2,3,6-trimethylphenol is about 17.2 mg/L with severe depression of the
parameter K as trimethylphenol increases in accordance with the equation:
6
-------
Log (Km) = 1.839 - 1.069 Log (2,3,6-TMP) (9)
These relationships are combined with yield coefficient information
to provide plots of effluent ammonia as a function of the independent
variable sludge age in continuous culture systems.
A plot of the effect of 2,3,6-trimethylphenol level on sludge age for
nitrification is outlined on Figure 7. As can be seen, even small
levels of 2,3,6-trimethylphenol will dramatically effect steady state
effluent ammonia levels at a given sludge age.
Similar results with "shoulders"occuring at different concentrations
were found for the compounds above. A summary of shoulder values is listed
on table 2.
Table 2
Values Below Which Toxicant Has No Effect On Nitrification Kinetic
Parameter
Shoulder Values at pH = 8.0
Compound Vmax (gNH gVSS-Day) Km (mg NH /L)
-j -^
* Free Cyanide Toxic at all Values No effect on K
•* m
* Phenol Toxic at all Values No effect on K
m
Coal tar acid mixture 1.2 mg/L as phenol 5 mg/L as phenol
2,3,6 TMP 4.9 mg/L 17.2 mg/L
2-Ethyl-pyridine 10 mg/L 42 mg/L
2,4,6 TMP 30 mg/L 50 mg/L
Fe(CN) ~3 80 mg/L No effect on Km
SCN 236 mg/L No effect on K
m
-3 05
* Note: Nitrification Inhibition is proportional to (CN) and (Phenol) ' ,
free cyanide is far more toxic than phenol.
Figure 8 is a summary plot of required sludge age to meet a
10 mg/L NH effluent versus concentration of organic inhibitors, cyanides,
and thiocyanate. It should be noted that the concentration axis for
thiocyanate should be multiplied by 10. Examination of Figure 8 is
revealing in that it demonstrates that coal tar acids are more inhibitory
-------
than reagent phenol, low levels of 2,3,6-trimethylphenol are far more
inhibitory to nitrification then levels of 2,4,6-trimethylphenol, and free
cyanide is the most toxic to nitrifiers of all the chemical species studied.
Complexed cyanides and thiocyanates exhibited toxicities to nitrification
only at levels far greater than that for the other species studied and may
be considered "non toxic" to nitrifiers at usual levels found in coke plant
effluents.
For practical purposes, by drawing a horizontal line on Figure 8
at a sludge age of 10 days, the set of intersections of this line with the
curves for individual compounds provides insight into allowable levels of
these substance (assuming no synergism) for nitrification. Table 3
is a listing of these levels. They should be considered as allowable
levels of such trace substances in effluents so that biological nitrification
may take place in a stable manner.
Table 3
Maximum Levels of Contaminents to Permit Nitrification at a 10 Day
Sludge Age with Effluent NH0 = 10 mg/L at pH = 8.0
Compound
Free Cyanide
Coal tar acids (as 4-AA phenol)
Phenol
2,3,6 TMP
2-Ethylpyridine
2,4,6 TMP
Fe(CN)
-3
SCN
SUMMARY OF FINDINGS
Concentration mg/L
0.11 mg/L
1.7 mg/L
5.5 mg/L
7.8 mg/L
22 mg/L
39 mg/L
190 mg/L
660 mg/L
Based upon laboratory data obtained, the following conclusion may be
drawn concerning the effect of inhibition to biological nitrification.
1.0 Toxic inhibition to biological nitrification decreased in the
order of free cyanide, tar acid phenolics, phenol, 2,3,6-trimethylphenol,
2-ethylpyridine, 2,4,6-trimethylphenol, complexed cyanide and thiocyanate,
8
-------
2.0 All substances, except free cyanide and phenol, appeared to follow
a "shoulder effect" whereby low levels of inhibitor had no influence on the
rates of biological nitrification while higher levels had profound effects.
The absence of the shoulder effect is an indication that low levels of free
cyanide should be avoided for stable operation of biological nitrification
processes.
3.0 Calculations based on laboratory data are provided to show
maximum levels of contaminents that will still permit stable nitrification
at a 10 day sludge age at pH = 8 and effluent ammonia levels of 10 mg/L.
30°C.
4.0 Upper temperature limits for stable optimum nitrification is
5.0 Free (or unionized) ammonia is inhibitory at levels of 10 mg/L
or more. Free ammonia is mainly a function of wastewater total ammonia,
pH, and temperature.
References
(1) Neufeld, R. D. , Hill, A. J., Adekoya, D.O. "Phenol and Free Ammonia
Inhibition to Nitrosomonas Activity" Water Research, Vol 14, #12
ppl695 - 1703 (Dec. 1980)
(2) Neufeld, R. D., Rieder, C. B., Greenfield, J. H. "Influence of Elevated
Temperatures on Nitrification Biokinetics as Applied to Steel Industry
Wastewaters" report to AISI, project 78 - 395 (August 1, 1980)
(3) Neufeld, R. D., Greenfield, J. H. "Influence of Free and Complexed
Cyanides an Nitrification as Applied to Steel Industry Wastewaters"
report to AISI under project 78 - 395 (August 1, 1981.)
(4) Neufeld, R. D., Greenfield, J. H. "Factors Influencing Nitrification
Biokinetics as Applied to meeting BAT Requirements for the Iron and
Steel Industry" Final report to AISI under project 78 - 395
(August 1, 1982)
Acknowledgement: This project was supported by the Research Committee
of the American Iron and Steel Institute, and by the Industrial Environ-
mental Research Laboratory of the U.S. Environmental Protection Agency.
IN ADDITION: A computer simulation program of coke plant wastewater treatment
operations is being developed based on laboratory nitrification kinetic
behavior coupled with equations for multicompartment/single sludge systems.
Our objective is to simulate, and verify in bench scale, tbe optimal
operation of single sludge biological systems to accomplish phenolic,
SCN. and ammonia removals in a stable manner.
9
-------
Krn vs. Temperature
I
1.2
1.0
« Q8
0.6
<3 0.4
0.2
Vmax vs. Temperature
1 1 1 1 r
20
u " 3.78-.084 (T'C)
for Ti 3O-C
32 36
TEMP (-C)
FIGURE 1
40 44
100
80
60
40
30
20
o 10
- 3
I 6
4
3
log Km= -1.8829 +.O8228 (T)
for T i 30° C \ „ ©
log Km= L53- 0.0315 (T)
for 22«C i T i 30"C
10 20 30
TEMP CO
FIGURE 2
40
50
Effluent NHj Concentration
vs.
Sludge Age
0 02030405060708090 00 HO EO
SLUDGE AGE (doys)
Calculated Required SJudge Age
to meet Indicated NH3 Effluents
vs.
Temperoture
20
4O
FIGURE 3
FIGURE 4
10
-------
v» CONCENTRATION Of 2,3,6 - TRIMETMYLPHENOL
10.0
1.0
•e
i
0.1
to) Vm • .J53-LII7 to? (TUP)
IO OO IOOO
2,3,6 -Trinwthylphenol (m<)/!}
I
E
0000 °\
Yl CONCENTRATION Of 2 ,3,S - TRIMETHYLP-eNOL
1.0 100 COO
2,3,6 - Tnrrnmytarxnol (mq/l)
FIGURE 5
FIGURE 6
EFfECT OF 24,6-TBMETHrLPv£NCL LEVEL ON SLUDGE A3E
AS A D€S» A«3 OPERaTKJML WRafceTES FOfl
100
80
60
40
20 40 60 80 CO 120
>9« (doyt)
REQUIRED SLUDGE AGE TO MEET IOmq/1 Nl-L EFFLUENT
vs.
CONCENTRATION OF INHIBITOR
IOO
-i 1 1 \ 1 1 1 r
20 40 SO 80
Concentration (r
100 120
FIGURE 8
FIGURE 7
11
-------
Physical/Chemical Treatment
of Coke Plant Wastewaters
By
James R. Zwikl
Director of Environmental Control
Shenango, Inc.
and
Nicholas S. Buchko
Environmental Engineer
Shenango, Inc.
and
David R. Junkins
Senior Project Engineer
WESTON Designers-Consultants
For Presentation at
EPA Symposium on
Iron and Steel Pollution
Abatement Technology
for 1982
Pittsburgh, Pennsylvania
16-18 November 1982
DESIGNERS \*^/ CONSULTANTS
-------
Disclaimer
The work described in this paper was not funded by the
U.S. Environmental Protection Agency. The contents do
not necessarily reflect the views of the Agency and no
official endorsement should be inferred.
-------
PHYSICAL/CHEMICAL TREATMENT OF COKE PLANT WASTEWATERS
By: James R. Zwikl, Shenango Incorporated
Nicholas S. Buchko, Shenango Incorporated
David R. Junkins, Roy F. Weston, Inc.
BACKGROUND
Shenango Incorporated is a privately owned corporation head-
quartered in Pittsburgh, Pennsylvania. The Coke & Iron Division
of Shenango is located on Neville Island, Pennsylvania, and is
engaged in the production of iron, coke, and coke by-products.
The Coke & Iron Division consists of three major depart-
ments: Coke Plant, Blast Furnace, and Steam and Power. The Coke
Plant presently consists of two operating coke batteries (35
ovens/battery), a new 56 oven battery presently being heated up
and a by-products plant. The Blast Furnace department consists
of two furnaces (1 operating and 1 stand-by), while the Steam
and Power department consists of a boiler house which contains
four boilers, and power generating facilities. (Figure 1 is a
schematic of water distribution through the plant.)
In August 1971, in an effort to achieve compliance with the
applicable state effluent limitations, Shenango commissioned The
Chester Engineers to conduct a feasibility study on wastewater
treatment. The results of the study indicated that treatment of
a combination of coke plant waste streams constituting about 2.5
percent of the total plant wastewater of 24 mgd (90,840 m^/d)
would be required to meet the projected limitations for cyanide
and phenol.
The streams which would require treatment all originated
within the Coke Plant and included: waste ammonia liquor; excess
final cooler wastewater; light oil separator wastes; hot oil de-
canter wastes; and ammonia liquor cooling tower blowdown in the
event that cooler leakage should contaminate the blowdown (see
Figure 2).
Additional studies were conducted by The Chester Engineers
to determine the most cost-effective method of treatment. The
treatment concepts investigated included biological and physi-
cal-chemical (ozonation, activated carbon, alkaline chlorina-
tion, etc . ) -
-------
Process
Non-Contact Water
City^mgd
Water
0.04 mod
0.07 mgd
Coke Plant
2.67
tO.90
Evaporation
Blast Furnace
8.95 mgd
f 0.52 mgd
Evaporation
Steam & Power
mgd ^
2.24 mgd
24.1 mgd be
Outfal
0.79 mgd
Steam
0.22 mgd
Evaporation
Figure 1. Plant Water Flow Schematic
Shenango Incorporated
Pittsburgh, Pennsylvania
-------
COKE
PLANT
Hot Oil Decanter
0.096 mgd
Phenolate Storage Area
0.232 mgd
Light Oil Separator
0.014 mgd
Final Cooler
0.152 mgd
Cooling Tower Slowdown
0.306 mgd
Total Flow
To Wastewater
Treatment Plant
0.494 mgd Design Average
0.800 mgd Design Maximum
I
Figure 2. Coke Plant Wastewater Streams
-------
Factors considered which influenced the final decision were:
capital cost, projected operations and maintenance costs, land
requirements, future effluent limitations, sensitivity to plant
upsets, and necessary level of operating expertise required.
Activated carbon followed by alkaline chlorination was chosen
as the appropriate treatment process. A Water Quality Management
Permit to construct such a physical-chemical treatment facility
was issued by the Pennsylvania Department of Environmental Re-
sources (DER) on 25 April 1976.
Based upon the feasibility studies and concept report by The
Chester Engineers, final detailed engineering and preparation of
plans and specifications for the facility was undertaken by Roy
F. Weston, Inc., Designers and Consultants. Engineering and con-
struction was completed, and the plant subsequently began oper-
ations in September 1979.
WASTEWATER TREATMENT PLANT DESIGN BASIS
DESIGN PARAMETERS
As illustrated in Figure 2, the major sources of wastewater
discharged to the treatment plant include:
Sourc e Des ign Flow
Hot Oil Decanter 0.096 mgd
Phenolate Storage Area 0.232 mgd
Light Oil Separator 0.014 mgd
Final Cooler 0.152 mgd
Cooling Tower Slowdown 0.306 mgd
Total 0.800 mgd
The design characteristics of the combined flow to the
treatment plant and NPDES discharge limitations are listed in
Table 1.
DESIGN BASIS VS CURRENT CHARACTERISTICS
The basis for the hydraulic and parameter design of the
treatment plant are compared with current operating character-
istics in Table 2. It is important to note that the plant in-
fluent ammonia concentration is greater than three times the de-
sign basis. In addition, the influent pH is highly variable. The
significance of these variances from the design basis is dis-
cussed in subsequent sections of this paper.
-------
TABLE 1. DESIGN WASTEWATER CHARACTERISTICS
COMBINED FLOW TO WWTP
Parame ter
NPDES Discharge
Liraitat ions
Maximum
Flow
SS
NH3
Phenol
CN-A
O&G
0
100
(667
100
(667
300
(2,000
70
(467
135
(900
.80 mgd
mg/L
ppd)
mg/L
ppd)
mg/L
ppd)
mg/L
ppd)
mg/L
ppd)
--
(1 ,800
(560
0
(7
0
(5
(410
-
ppd)
ppd)
.1 mg/L
ppd)
.025 mg/L
ppd)
ppd)
--
(3,760
30
(1 ,120
0
(14
0
(10
10
(620
-
ppd)
mg/L
ppd)
.2 mg/L
ppd)
. 05 mg/L
(ppd)
mg/L
ppd)
8.0
6 to 9
-------
TABLE 2. COMPARISON OF DESIGN VS OPERATING
WASTEWATER CHARACTERISTICS
Parame t er
Flow, mgd
SS, mg/L
NH3, mg/L
Pheno 1 , rag/L
CN-A, mg/L
O&G , mg/L
pH
SCN
Desig_n Basis
0.80
100.
100.
300.
70.
135.
8. 0
Previous
Cond i t ions
(1980-1981)
0.35
69
199
69
45
8.0 to 9.0
255
Current
Cond it ions
(1981-1982)
0.38
135
332
88
32
8.0 to 9.0
322
-------
WASTEWATER TREATMENT PLANT DESCRIPTION
PROCESS DESCRIPTION
A block flow diagram illustrating the treatment process
utilized by Shenango for the removal of oil and grease, ammonia,
phenol, suspended solids, dissolved organics, and cyanides is
shown in Figure 3.
All the wastewater streams, except the hot oil decanter
overflow, initially are combined in an equalization basin. This
basin has a capacity of 260,000 gallons (984 m^) and a deten-
tion time of eight hours at the design flow of 0.80 mgd (3,028
m^/d) (16.4 hours detention time at the current flow rate of
0.38 mgd [1,438 m^/d]). The basin is divided into two sec-
tions to allow isolation of one portion of the basin for main-
tenance purposes. Each section of the equalization basin is
equipped with a 3 hp, single-speed mixer -
Underflow from the hot oil decanters is pumped to two dis-
solved air flotation (DAF) units for additional oil and grease
removal. Any oil skimmings collected are recycled to the by-
products plant. Each DAF unit has a diameter of 9 ft (3 m), and
detention time of 19 minutes at a design flow of 67 gpm (0.25
m^/d). The surface loading rate for the DAF units is 1,500
gpd/sq ft (61 m-Vm^-d). The underflow from the DAF units is
pumped to the equalization basin, and combined with the other
four waste streams at that point.
The combined flow from the equalization basin is pumped to
two 35-ft (11.6 m) diameter clarifiers for suspended solids re-
moval. The equalization basin discharge pumps may be operated
manually, or operated automatically according to level controls
in the basin. The detention time in each clarifier is six hours,
and the overflow rate is 420 gpd/sq ft (17 m^/m^-d) at the
design flow rate of 280 gpm (1.1 m^/d). The underflow from
the clarifiers is dewatered via a 4.5 ft (1.5 m) diameter (85 sq
ft [7.9 m^ ] of surface filtration) cloth vacuum filter. The
sludge cake generated is hauled to a landfill for ultimate dis-
posal .
The overflow from the clarifiers flows by gravity into a
holding tank, and is subsequently pumped to two dual media pres-
sure filters for additional suspended solids removal. The 10 ft
(3.3 m) diameter filters are downflow units with a design flow-
through rate of 7 gpm/sq ft (4.75 L/m^-s) of filter surface
area. The filters are utilized to prevent solids-clogging prob-
lems and decreased efficiencies in the downstream activated car-
bon columns. The effluent from the filters flows into an 18,000
-------
Hot Oil ^ Pump Sta.
Decanter No. 1
Phenolate
Storage—.- P^P Sta. __
Area No 2
Liaht Oil . ^ Pump Sta. ^
Separator " NO. 3
Cooling Tower
Slowdown
(Optional)
Final Cooler
Wastewater
Polymer
I
Flnrrulatinn Dissolved Oj| 55-gal.
Moccuiation .. y ^tnranp »- Hcu3c/Diaoonal
Tank Flotation Sk.mm.ngs Drum
Equal
Underflow
Dirty Backwash
1
Clean
Clarifiers Dual Backwash A .. . .
Tntinn _fc_ r*larifi*-'atir\n h~ FfflllRnt •- hJtaHi^ »^ \AMfrti- — — ACTIVateQ
Tank Filters Storage Carbon
Tank
Filtrate Vacuum
Filtration CbJ |NaOH
{ _^ Cyanide Final f \
Sludge Cake " Removal Discharge //
Figure 3. Block Flow Diagram Wastewater Treatment System
Shenango Incorporated
Pittsburgh, Pennsylvania
-------
gallon (68 m^) carbon feed/backwash water storage tank. This
tank is the water supply source for filter backwash water. Dirty
backwash water is routed back to the plant equalization basin.
Filtered wastewater is pumped from the backwash water stor-
age tank to the activated carbon system. Activated carbon util-
ized in the system is furnished through a service agreement with
the Calgon Corporation. The system consists of two activated
carbon columns operated in series to remove phenol and other
dissolved organics that might be present. The carbon columns are
located ahead of the cyanide removal system to minimize the for-
mation of chlorinated hydrocarbons.
The effluent from the carbon columns flows to the cyanide
removal unit operation. The cyanide system is a single-stage al-
kaline chlorination process that oxidizes cyanide to cyanate.
The effluent from the carbon columns initially discharges into
two 12-ft (4 m) diameter chlorine oxidation tanks where caustic
(NaOH) is added to raise and maintain the wastewater pH at 9.0,
and chlorine (Cl2) is simultaneously added to oxidize cyanide
to cyanate. The detention time in each tank at a design flow of
0.4 mgd (1,514 m^/d) is one hour. Caustic and chlorine addi-
tion is automatically controlled based on pH and ORP (oxidation
reduction potential) levels respectively in the oxidation tanks.
Chlorine is supplied to the wastewater treatment plant via 55
ton (49,900 kg) or 90 ton (81,720 kg) railroad cars although ton
cylinders are utilized during railroad car transfer periods. The
liquid chlorine is transferred from the tank cars, passed
through evaporators, and then fed via chlorinators to the oxida-
tion tanks. The effluent from the oxidation tanks is then dis-
charged as final effluent.
OPERATING EXPERIENCES
ACTIVATED CARBON SYSTEM
The Calgon activated carbon system consists of two 40,000
pound (18,160 kg) carbon vessels which are operated in series.
The system was first placed on line in September 1979 and has
operated efficiently with only relatively minor difficulties.
These difficulties included premature phenol breakthrough and
carbon fines breakthrough immediately following a carbon trans-
fer.
Based on the validation study conducted during the carbon
system design, it was estimated that the carbon usage rate would
be approximately 20 pounds (9 kg) of carbon per 1,000 gallons
(3.8 m-*) of wastewater for phenol levels averaging 159 mg/L
at a pH of 7.0 to 8.0. At a current average flow rate of ap-
proximately 0.38 mgd (1,438 m^/d), carbon breakthrough would
-------
be expected approximately every three days. Presently, the ac-
tual wastewater has a phenol concentration well below the design
limits and, consequently, the phenol breakthrough only occurs
approximately every seven days.
After several months of operation, an unusual trend began to
emerge. In several instances immediately following a carbon
transfer, the polished effluent would indicate phenol break-
through. A detailed survey of the situation revealed a two-fold
problem. First, an accumulation of solidified carbon had devel-
oped in the bottom (three to six feet) of the column. Secondly,
the high wastewater pH led to a very low phenol adsorption by
the carbon.
Although the reason for the solidification of the carbon was
never discovered, it has been projected that mechanical failure
in the carbon column underdrain system led to channeling of the
wastewater. This channeling in turn permitted a portion of the
carbon bed to solidify. A further study of the pH fluctuations
revealed the following: although the carbon system is designed
to operate at a pH of 7.0 to 8.0, the wastewater feed into the
carbon system is not pH controlled. Depending on occurrences
upstream of the facility, the pH can vary from 6.0 to 12.0, al-
though such extreme variations are unusual. Since the alkaline
chlorination system following the carbon system was operating
at a pH of 10.0 to 11.0 at the time, attempts were made to keep
the wastewater pH greater than 10.0 so that less caustic would
be required. Studies performed by Calgon*, however, indicated a
decrease in carbon efficiency as the pH increased above 9.0,
and severe upsets when the pH exceeded 11.0. According to Cal-
gon's report, "this phenomenon is expected due to the ionization
of the phenol molecule at a high pH. The ionic form of the phe-
nol molecule is very poorly adsorbed by the activated carbon."
Thus, as we can see from Table 3, as the pH increases from 9.0
to 11.0, the percentage of molecular phenol decreases from 90
to 10. At a pH of 11.0, only 10 percent of the phenol will be
in a molecular form. This results in approximately only 50 per-
cent phenol adsorption as the wastewater reaches the discharge
point of the first column. As further stated by Calgon, "when
the molecular form represents only a small percentage of the
total phenol content, the chemical equilibrium kinetics add to
the adsorption kinetics to extend the length of the mass trans-
fer zone and increase the probability of phenol leakage to the
effluent."
*Letter from A.J. Roy and W.M. Rogers, Calgon Corporation to
James R. Zwikl, Shenango, Inc., 12 May 1980-
10
-------
TABLE 3. RELATIONSHIP BETWEEN pH AND PERCENTAGE OF
MOLECULAR PHENOL PRESENT*
Approximate Percent of
Molecu 1 a r Pheno 1 PTt;
9 90
10 50
11 10
*Letter from A.J. Roy and W.M. Rogers, Calgon Corporation to
James R. Zwikl, Shenango, Inc., 12 May 1980.
11
-------
Excursions in pH above 11.0 can cause severe operational
problems if a carbon bed has been well loaded with phenol previ-
ously. Re-equilibrium to a lower phenol concentration can cause
desorption of previously adsorbed phenol, and produce effluent
levels in excess of those found in the influent. Figure 4 graph-
ically illustrates the relationship between phenol adsorption
and pH which was observed over a consecutive three week span*.
As can be observed, an increase in pH above 10.0 causes a cor-
responding increase in effluent phenol levels, while a decrease
in pH to less than 10.0 causes a similar decrease in effluent
concentrations.
Decreasing the pH levels to less than 8.5 at this installa-
tion, however, does not increase the operational efficiency of
the carbon system. As a result, the system pH is controlled be-
tween 8.5 and 9.5 to obtain the most efficient and economical
utilization of activated carbon.
CHLORINATION SYSTEM
The alkaline chlorination system is a single-stage process
designed to oxidize cyanide to cyanate.
The system consists of two completely mixed 15,000 gallon
(56 tn-*) cyanide oxidation tanks operated in parallel, two
8,000-pound (3,632 kg) per day evaporators (1 operating and 1
spare), two ejectors, and a 35,000 gallon (132 m^) caustic
storage tank with associated proportional feed pumps. The
chlorine is supplied as a liquid from either a 55 ton (49,900
kg) or 90 ton (81,720 kg) railroad tank car which is located on
a track adjacent to the chlorination building. Stand-by one ton
(908 kg) cylinders of chlorine are available during those peri-
ods when railroad cars are switched. Air padding of the chlorine
car is not required.
The alkaline chlorination system was designed to operate as
follows: partially treated wastewater is fed in equal volumes
into the bottom of each oxidation tank. Continuous readout ORP
and pH probes located near the top of each tank measure and
record both values on the main control panel. The pH recording
controllers modulate the caustic pumps to maintain the proper pH
levels within the oxidation tanks. Presently, a pH of 9.0 to 9.5
is maintained- The ORP recording controller regulates the amount
*Letter from A.J. Roy and W.M. Rogers, Calgon Corporation to
James R. Zwikl, Shenango, Inc., 12 May 1980.
12
-------
1. Letter (ram A.J. Roy and W.M. Rogers, Calgon Corporation to James R. Zwftl,
Shenango. Inc.. 12 May 1980.
Figure 4. Relationship Between pH and Phenol Concentration
-------
of chlorine gas injected into the system through the ejectors.
Laboratory analyses performed on the actual wastewater indicate
that maintaining an ORP of approximately 120 to 180 mv should
provide sufficient oxidation of cyanide to cyanate.
To date, efforts to place the system in the fully automatic
mode of operation have proven unsuccessful. Although the pH con-
trol system has operated efficiently in the automatic mode, the
ORP control system has not. The ORP system operation becomes
very erratic within 4 to 48 hours after being placed in the
automatic mode. It appears as though some interfering compounds
inhibit the automatic operation of the ORP sensing and, conse-
quently, the chlorine feed controller instrumentation.
The cyanide oxidation system is currently being operated in
the manual mode, with resultant effluent cyanide "A" levels
equalling or approaching the design limits while attempts are
being made to resolve the problem. Several sets of bench-scale
and full-scale tests have been performed in order to identify
an ORP/cyanide oxidation relationship. A curve similar to that
indicated in Figure 5 would be expected to characterize the
cyanide, alkaline chlorination process. Bench-scale tests of
composite carbon column effluent samples were conducted at sys-
tem pH's of 9.5 and 10.5. The test results are presented in
Tables 4 and 5 and indicate that chlorine dosage rates of 2,000
mg/L or less should be sufficient to oxidize the cyanide to per-
mitted levels. At the plant's current influent flow rate of 0.38
mgd (1,438 m3/d), this is equivalent to 6,300 ppd (2,858 kg/
d) of chlorine, which is well within the treatment plant's chlo-
rination system capacity. Unfortunately, graphical analysis of
ORP (i.e., cyanide levels), and chlorine feed rate data collect-
ed during full-scale field tests, have yielded a horizontal line
with no distinct breakpoint, as shown in Figure 6, indicating no
consistent relationship exists between these parameters. There-
fore, it appeared that ORP was not a viable parameter for moni-
toring and controlling the cyanide oxidation system in our proc-
ess.
Since attempts to correlate ORP levels with cyanide "A" lev-
els that are consistently below 2.5 mg/L have not proven totally
successful, additional means to permit quick, accurate and fre-
quent analyses of cyanide "A" in the influent and effluent chlo-
rination system streams are required. A continuous readout, on-
line cyanide "A" monitor is currently under development, and the
first full-scale instrument is scheduled to be tested at She-
nango's wastewater treatment facility in hopes that it will
stabilize operation of the cyanide oxidation unit. Additional
process studies are also being investigated.
14
-------
at
"Z
n
o>
o>
oc
Q.
Cyanide Concentration -
Decreasing
Figure 5. Typical ORP Versus Cyanide Concentration Relationship
15
-------
Total Cyanide, mg/L CN 104
Amenable Cyanide, mg/L CN 97
Ammonia, mg/L N 54
Thiocyanate, mg/L CNS 240
ORP, mv 40
j> a m p 1 _e _N
-------
jParame ter *>£ mj>^_e_N£^_^ Sampl e No . 3
Total Cyanide, mg/L N 83 96
Amenable Cyanide, mg/L CN 77 88
Ammonia, mg/LN 35 33
Thiocyanate, mg/L CNS 310 290
ORP, mv 20
§a.inj>.Le_N°._i_.2 Analyses at 90 minutes and pH 9.5
Chlorine Addition, mg/L Cl
° 2000 50° 00
ORP, mv 100 120 140 320 410
Total Cyanide, mg/L CN 8.6 6.0 6.2 6.2 6.2
Amenable Cyanide, mg/L CN 1.8 0.2 0.4 --- ---
Ammonia, mg/L N 34 35 14 2 2
Thiocyanate, mg/L CNS 190 72 <5 <5 <5
Sanrgle No. _ 3_AnjaJ.^j5 £j5 _ at 90 minutes and pH 9.5
ORP, mv --- 130 220 400 460
Total Cyanide, mg/L CN --- 7.6 --- --- ---
Amenable Cyanide, mg/L CN --- 0.0 --- --- ---
Ammonia, mg/L N --- 33 23 42
Thiocyanate, mg/L CNS --- 57 5 5 ---
17
-------
00
Oxidation
_ Tank No. 1
Oxidation
Tank No. 2
Figure 6. Cyanide Oxidation Field Test Results
-------
SCALING
In early 1982 operating problems began to surface with the
chlorine injector pumps. When the pumps were disassembled for
maintenance it was discovered that all the wetted parts of the
pump were coated with a grayish-white substance. In addition,
the chlorine oxidation tanks were also found to be coated with
a similar type of substance which was approximately l/4~in.
(0.635 cm) thick. Laboratory analyses revealed that the deposit-
ed material was calcium carbonate. Shortly thereafter it was
also discovered that the chlorine-ejector water supply lines
were reduced in diameter by approximately 1 to 2 in. (2.5-5 cm)
due to scale buildup.
Further analyses were conducted to determine the scaling po-
tential of the wastewater in the oxidation tools through the use
of the Langelier Saturation Index (LSI). An LSI of +1.7 was cal-
culated, indicating that the wastewater has a tendency to form
scale.
Analyses also confirmed a decrease in total hardness across
the oxidation tanks. Calcium precipitates as calcium carbonate,
thus explaining the increase in suspended solids in the final
plant e ffluent .
Several options are available as possible solutions to the
scaling problem. These include: determining possible process
changes upstream of the treatment facility; precipitating the
column in the clarifiers; chemically neutralizing the wastewater
through the use of a deposit control agent; or reducing the sys-
tem pH in an attempt to retard the formation of scale.
The initial alternative chosen consisted of lowering the pH
from 10.5 to 9.0. This is the simplest and most cost effective
means of attempting to regulate the formation of scale, and is
currently being evaluated.
CHEMICAL AVAILABILITY
The basic reactions which occur in the alkaline chlorination
process require the use of large quantities of caustic and chlo-
rine. Unfortunately, the market for these two chemicals has
fluctuated dramatically over the past several years. Since both
caustic and chlorine are produced simultaneously, the demand for
either chemical influences the cost and availability of the
other -
19
-------
The wastewater treatment facility consumes approximately
three to four tons (2,724 to 3,632 kg) of chlorine and 1,400
gallons (5.3 m^) of caustic daily. Therefore, both the cost
and availability of either chemical will dramatically affect
both the efficiency and operating costs of the facility.
Since the plant start-up in September 1979, the cost of
chlorine has actually decreased by 30 percent, while the price
of caustic has increased by 106 percent. The overall effect is
a net chemical cost increase of $1.28/1,000 gallons of waste-
water treated .
WASTEWATER TREATMENT PLANT PERFORMANCE
CURRENT PERFORMANCE
The current operating performance of the treatment plant is
indicated in Table 6. Excellent phenol removal in excess of 99
percent is consistently obtained from the system.
As detailed in this paper, however, problems still remain to
be overcome in the efficient operation of the alkaline chlorina-
tion portion of the plant. It is anticipated that the efficient
control of the chlorine feed which may be obtained through the
use of the continuous cyanide "A" analyzer will increase the cy-
anide "A" removal efficiency from 87 to 96 percent.
Influent ammonia levels, as stated earlier, are approximate-
ly 200 percent higher than originally anticipated. As a result,
an insufficient amount of excess chlorine is available to en-
tirely oxidize the ammonia and, thus, a lower than anticipated
removal efficiency is obtained.
The suspended solids removal data of 77 percent is very mis-
leading. A suspended solids removal of approximately 100 percent
is generally achieved between the clarifiers, dual-media fil-
ters, and activated carbon columns. The precipitation of calcium
carbonate (as previously discussed) in the alkaline chlorination
system is responsible for the apparent decreased efficiency in
solids removal.
Thiocyanate removals of 72 percent have also been observed
as the thiocyanate is oxidized to cyanate.
BPT PERFORMANCE
The plant effluent characteristics are compared with the
current NPDES and Best Practicable Control Technology Currently
Available (BPT) discharge limits for by-product coke plants in
Table 7. It can be observed that except for cyanide "A" the
20
-------
TABLE 6. TREATMENT PLANT PERFORMANCE
Influent
Paramet er
Flow, mgd
SS, mg/L
NH3> mg/L
Phenol, mg/L
CN-A, mg/L
PH
SCN, mg/L
Average
0.38
135
332
88
32
8.8
322
Range
15-
74-1
35-
4-
2-
106-
472
,233
238
102
12
484
Effluent Percent Removal
Average
0.38
21
175
0.40
4.0
9.0
102
Range Average Range
N/A
3- 53 77 19- 98
21-941 52 2- 83
0.004- 1.4 >99.5 98.3- 99.9
0-12 87 58-100
N/A
11-338 72 49-96
21
-------
TABLE 7. TREATMENT PLANT PERFORMANCE VS NPDES PERMIT AND PROMULGATED BAT/BCT LIMITS
Past Discharge
Parameter Levels (1980-1981)
mg/L
(ppd)
SS
O&G
NH3-N
CN-A
Phenol
PH
Benzene
Avg . Max .
15.6 30
(44.1) (101)
144 569
(427) (1,604)
5.9 17.2
(17.9) (79)
0.7 3.8
(2.0) (10.9)
9
0.0046
(0.013)
Present Discharge
Levels (1981-1982) NPDES Limits BPT Limits*
mg/L mg/L (ppd)
(ppd) (ppd)
Avg . Max . Avg . Max . Avg . Max .
21.3 53
(64.1) (173) (1,800) (3,760) 393 759
10
(410) (620) 33 98
175 941 -— 30
(545) (2,606) (560) (1,120) 274 822
4.0 12.2 0.025 0.05
(9.7) (35) (5) (10) 66 197
0.4 1.4 0.1 0.2
(1.3) (4.2) (7) (14) 4.5 13.5
6 to 9 6 to 9 6 to 9
Naphthalene<0.010
<( 0.030)
BAP
<0.010
<(0.030)
— — — -v — — — — — _
BAT/BCT Limits*
(ppd)
Avg . Max .
(393) (759)
(33) (98)
(97) (194)
(0.13) (0.26)
6 to 9
(0.06)
(0.06)
(0.06)
*At production level = 1,500 T/Day - Based on regulations promulgated on 5/27/82 (40 CFR Part 420).
-------
plant is meeting current NPDES discharge limits. In addition,
the plant discharge is achieving BPT limits with the exception
of ammonia level exceedances.
BAT/BCT PERFORMANCE
Table 7 also compares the current discharge levels with the
recently promulgated Best Available Technology Economically
Achievable/Best Conventional Technology* (BAT/BCT) limits for
by-product coke plants. As can be readily observed, the plant
does not meet BAT or BCT limits for ammonia or phenol. The
limits for benzene, benzo(a)pyrene, and napthalene are currently
being met based on limited data.
It does not appear economically feasible to sufficiently
reduce the discharge of phenol and ammonia at this facility-
The BAT limits
performance on a Ib
for phenol are
phenol/1,000 Ib
BAT Limit
Maximum 0.0000859 Ib phenol/
1,000 Ib coke
Average 0.0000430 Ib phenol/
1,000 Ib coke
compared to treatment plant
coke basis below:
WWTP Effluent
0.00660 Ib phenol/
1,000 Ib coke
0.00068 Ib phenol/
1,000 Ib coke
The treatment plant effluent values reported are based on
long term data collected by Shenango and represent discharge
levels following treatment by 80,000 pounds (36,320 kg) of acti-
vated carbon. Although a minimum of 0.000007 Ib phenol/1,000 Ib
coke has occasionally been achieved at the physical-chemical
plant, the present system is not capable of achieving such a
discharge limit over prolonged periods of operation. In fact,
the average discharge level is 16 times higher than the BAT lim-
its.
Currently higher than anticipated influent ammonia levels
are being observed at the treatment plant even though a free and
fixed ammonia still is presently operating efficiently in the
by-products plant. This has resulted in the inability to achieve
discharge limits. The proposed BAT ammonia level of 0.0322 Ib
ammonia-N/1 ,000 Ib coke is presently being exceeded by a factor
of 9 (0.2868 Ib NH3-N/1,000 Ib coke). Considerable further
investigation is required before this problem can be resolved.
*As published in 40 CFR Part 420, 27 May 1982.
23
-------
PRIORITY POLLUTANTS
As previously shown in Figure 3, Shenango's wastewater
treatment scheme includes activated carbon followed by alkaline
chlorination. This was done to minimize the formation of chlori-
nated hydrocarbons in the treatment plant effluent.
Analyses were conducted across the treatment system to de-
termine the ability of the facility to remove priority pollu-
tants. Five days of samples were collected and composited at
the locations shown in Table 8. As shown in the table, only 11
organic priority pollutants were detectable in the influent to
the activated carbon system. Adsorption by activated carbon re-
moved the majority of pollutants with the exception of acryloni-
trile, which was oxidized in the alkaline chlorination portion
of the plant. The treatment facility showed 100 percent removal
of the above listed organic priority pollutants with the excep-
tion of phenol, which was reduced in concentration by 99.3 per-
cent.
The use of alkaline chlorination treatment did, however,
cause the formation of chloroform. A concentration of 27.5 Mg/L
was measured in the chlorination system effluent. The mixture of
the treatment plant effluent with the remaining facility waste-
water, however, resulted in a chloroform concentration of ^5
|jg/L at the point of discharge to the Ohio River.
In 1979-1980, EPA conducted a coke plant wastewater treat-
ment study* at Shenango's Coke & Iron facility. The pilot plant
flow scheme utilized during the EPA study included alkaline
chlorination followed by activated carbon which is reverse to
the flow scheme utilized by Shenango. Priority pollutant sam-
pling results from EPA's study are compared in Table 9 to those
compiled by Shenango. The analytical results indicated that ex-
cept for phenol, the concentration of organic priority pollu-
tants in Shenango's effluent was less than or equal to the lev-
els reported in the EPA study. The levels of acrylonitrile,
chloroform, and 1 ,2-dichloroethene were particularly lower in
Shenango's effluent than in EPA's pilot plant discharge.
ECONOMICS
The capital cost required to construct Shenango's physical-
chemical wastewater treatment facility was $2,300,000 in 1977-
1979 dollars. Of this total, approximately fc660 ,000 (29 percent)
*E P A~Ff "f Tue"n"t ~Gu i d e lines Division, Report No. EPA-600/2-8 1-
053, Research and Development, PhysicaI/Chemical Treatment
of Coke Plant Wastewater, April 1981.
24
-------
TABLE 8. PRIORITY POLLUTANTS REMOVAL TRACKING DIAGRAM
Priority Pollutants Removal Tracking Diagram
Ohio River
E
•* F
/ (T) '
I r-
Intake) jj
H,0 p
1
I
last
irea
team &
ant
~nkp
'lant * Equallxatton -*.
© G
CI.riflc.Hon -* ff™n J
Activated
Carbon
Sample Point Concentration, M g/L
Parameter
Acrylonitrile
1 2
<100 6,000
1,2-Dichloroe thane O 60
Benzene
Toluene
Napthalene
Acenaphthylene
Phenanthrene
Phenol
Fluor ene
Anthracene
<5 5,835
<5 2,060
<10 3,610
<10 150
<10 19
<25 29,000
<10 9
<10 9
Methylene Chloride <5 ND
Chloroform
2,4
<5
,6-Trichlorophenol <25
3 4
5,000 <100
ND <10
80 4.6
ND <10
1.1 <10
ND <10
ND <10
417 209
ND <10
ND <10
9 <10
27.5
24
5
<100
<5
<5
<5
^10
^10
(10
<25
^10
^10
<5
<5
<25
Sample /->.
Point ^J
Number
P ? ?//
t Cyanide | i J,(^(
Removal \ - \
Net Percent
Removal
100
100
100
100
100
100
100
100
100
100
100
25
-------
TABLE 9. COMPARISON OF SHENANGO'S AND EPA1 S
PLANT PRIORITY SAMPLING RESULTS
METALS
Parameter
Shenango WWTP
EPA Pilot Plant*
Arsenic, mg/L As
Aluminum, mg/L Al
Barium, mg/L Ba
Calcium, mg/L Ca
Copper, mg/L Cu
Iron, mg/L Fe
Magnesium, mg/L Mg
Manganese, mg/L Mn
Nickel, mg/L Ni
Selenium, mg/L Se
Sodium, mg/L Na
Zinc, mg/L Zn
Mercury, mg/L Hg
0.068
0.
17
1.85
2.38
4.2
0.05
0.12
0.049
3,150
0.14
<0.002
VOLATILE COMPOUNDS
Acrolein, Mg/L
Acrylonitrile, Mg/L
Benzene, Mg/L
Bromomethane, Mg/L
Bromodi chlorome thane,
Bromoform, Mg/L
Carbon Tetrachloride,
Chlorobenzene, Mg/L
Chloroethane, Mg/L
2-Chloroethylvinyl ether, Mg/L
Chloroform, Mg/L
Chloromethane, Mg/L
Dibromochloromethane, Mg/L
<100
<100
4.6
Mg/L
Mg/L
27.5
Average
1.3
0.007
15
0.04
3.39
4.1
0.34
0.19
0.24
6,417
0.12
2,015
12
ND
ND
ND
630
ND
Range
0.072 0.03 - 0.22
0.8
0.04
11
0.01
0.0
3.5
0.19
0.2
0.13
1.9
0.12
20
0.08
7.7
4.9
0.50
0.05
0.34
3,900 -8,000
0.08 - 0.22
0.0023 0.0003-0.0061
ND** -8,500
ND
ND
ND
ND
ND
ND
-3,200
*EPA Effluent Guidelines Division, Report No. EPA-600/2-81-053,
Research and Development, PhysicaI/Chemical Treatment of Coke
Plant Wastewater, April 1981.
**ND - Not Detected.
26
-------
TABLE 9. (continued)
VOLATILE COMPOUNDS (CONTINUED)
Parameter
Shenango WWTP
EPA Pilot Plant*
1,1-Dichloroethane, pg/L
1,2-Dichloroethane, pg/L
1,1-Dichloroethene, pg/L
trans-1,2-Dichloroethene, pg/L
1, 2-Dich lor o pro pane, pg/L
cis-1,3-Dichloropropene, pg/L
trans-1,3-Dichloropropene, pg/L
Ethylbenzene, pg/L
Methylene chloride, pg/L
1,1,2,2-Tetrachloroethane, pg/L
Tetrachloroethene, pg/L
1,1,1-Trichloroethane, pg/L
1,1,2-Trichloroethane, pg/L
Trichloroethene, pg/L
Trichlorofluoromethane, pg/L
Toluene, pg/L
Vinyl chloride, pg/L
4-Chloro-3-methylphenol,
2-Chlorophenol, pg/L
2,4-Dichlorophenol, pg/L
2,4-DimethyIphenol, pg/L
2,4-Dinitrophenol, pg/L
2-Methyl-4,6-d initrophenol,
2-Nitrophenol, pg/L
4-Ni trophenol, pg/L
Pentachlorophenol, pg/L
Phenol, pg/L
2,4,6-Trichlorophenol, pg/L
ACID EXTRACTABLES
Pg/L <25
<25
<25
<25
<250
pg/L <250
<25
<25
<25
<209
<24
Average
25
ND
ND
ND
ND
BASE-NEUTRAL EXTRACTABLES
Acenaphthlene, pg/L
Acenaphthylene, pg/L
Anthracene, pg/L
Benzo(a)anthracene, pg/L
Benzo(b)fluoranthene, pg/L
Benzo(k)fluoranthene, pg/L
Benzo(a)pyrene, pg/L
19
ND
ND
ND
ND
ND
ND
ND
0.08
0.3
ND
ND
ND-64
ND
ND
ND
ND-60
ND
ND
ND
ND
ND
ND
ND
ND-1
ND-3
ND
ND
27
-------
TABLE 9. (continued)
BASE-NEUTRAL EXTRACTABLES (CONTINUED)
Parameter Shenango WWTP EPA Pilot Plant*
Average Range
Banzo(g,h,i)perylene, Mg/L <25 ND ND
Benzidine, Mg/L <10 ND ND
Bis(2-chloroethyDether, Mg/L
Bis(2-chloroethoxy)methane, Mg/L
Bis(2-ethylhexyl) phthalate, Mg/L
Bis(2-chloroisopropyl)ether, Mg/L
4-Bromophenyl phenyl ether, Mg/L
Butyl benzyl phthalate, Mg/L (10
2-Chloronaphthalene, Mg/L \ 10
4-Chlorophenyl phenyl ether, Mg/L <^10
Chrysene, Mg/L <(lO ND ND
DibenzoCa ,h)anthracene, Mg/L <25 ND ND
Di-n-butyphthalate, Mg/L
1,3-Dichlorobenzene, Mg/L
1,4-Dichlorobenzene, Mg/L
1,2-Dichlorobenzene, Mg/L
3,3'-Dichlorobenzidine, Mg/L
Diethylphthalate, Mg/L
Dimethylphthalate, Mg/L
2,4-Dinitrotoluene, Mg/L <10 ND ND
2,6-Dinitrotoluene, Mg/L
Dioctylphthalate, Mg/L
1,2-Diphenylhydrazine, Mg/L
Fluoranthane, Mg/L <10 ND ND
Fluorene, Mg/L <10 ND ND
Hexachlorobenzene, Mg/L OO ND ND
Hexachlorobutadiene, Mg/L
Hexachloroethane, Mg/L
Hexachlorocyclopentadiene, Mg/L
Indeno (l,2,3-cd) pyrene, Mg/L <25 ND ND
Isophorone, Mg/L <10 ND ND
Naphthalene, Mg/L <10 1.3 ND-9
Nitrobenzene, Mg/L
N-Nitrosodimethylamine, Mg/L
N-nitrosodi-n-propylamine, Mg/L
N-Nitrosodiphenylamine, Mg/L ^10
Pnenanthrene, Mg/L <10 0.3 ND-3
Pyrene, Mg/L <10 ND ND
2,3,7,8-Te trachlorodigenzo n-p-
dioxin, Mg/L ND
1,2,4-Trichlorobenzene, Mg/L
28
-------
was required for mechanical equipment, and the remaining
$1,640,000 (71 percent) was required for the actual plant con-
struction.
Annual operating and maintenance costs have increased con-
siderably since the plant start-up in September 1979. The major
factor contributing to this dramatic increase has been the ef-
fects of inflation on the price of caustic (106 percent since
September 1979) and activated carbon (45 percent since Septem-
ber 1979). In the first seven months of 1982, these two operat-
ing supplies accounted for approximately 62 percent of the
monthly operating cost. A breakdown of the 1982 operating costs
is presented in Table 10.
Based on 1982 cost figures through July, the treatment fa-
cility has an operating and maintenance cost of approximately
$181,000 per month. This is an equivalent cost of $6.65/ton of
coke at present reduced production rates, or $5.87/ton at 1981
production rates, which represent a more normal business con-
dition. An increase in production costs in this magnitude sig-
nificantly influences a company's ability to remain competitive
in the open market.
When analyzed from a water through-put basis, the operating
and maintenance costs are equivalent to $15.86/1,000 gallons
($4.19/m3) of influent wastewater. Unfortunately, even with
such a high operating cost, all of the BAT limits cannot be
achieved even though BAT, or "state of the art" technology is
currently being utilized by Shenango.
SUMMARY
In order to meet discharge limitations, and in anticipation
of rapidly approaching BAT requirements in September 1979, She-
nango Incorporated installed an activated carbon/alkaline chlo-
rination wastewater treatment facility at its Neville Island
Coke Plant. The optimization of the carbon system, along with
the on-going attempts to effectively automate the chlorination
system, has resulted in a treatment facility which can be con-
sidered BAT technology, but which cannot totally achieve BAT
1imi t a t ion s .
Although the facility does an excellent job of removing
organic priority pollutants, certain BAT limitations, such as
phenol, cannot be achieved even though the carbon system removal
efficiency consistently averages greater than 99.5 percent.
29
-------
TABLE_.H)_._ OPERATING & MAINTENANCE CO ST D IS TRIBUT10 N
Operat
Item
ing Labor
Maintenance Materials and Services
Operat
(Caust
ing Supplies
ic , Chlorine, Carbon)
Laboratory Services
Electr
Miscel
TOTAL
icity*
laneous
Cost/Month
$ 26
16
121
2
12
2
$180
,600
,200
,400
,000
,000
,600
,800
Perc
O&M
14
9
67
1
6
1
100
ent
Cost
.7
.0
.2
.1
.6
.4
. 0
*Based on a cost of fe0.108/KWHr (weighted average of purchased
and produced electricity).
30
-------
When the effects of inflation from 1979 to 1982 on caustic
and activated carbon are analyzed, the economic achievabi1ity of
certain BAT limitations through physical/chemica 1 treatment must
be questioned. A treatment cost of greater than $15 per 1,000
gallons (i>4/m3) of wastewater demands that cost/benefit con-
siderations should be made before implementation of more strin-
gent guidelines is imposed.
31
-------
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of
the Agency andvno official endorsement should be inferred.
TREATMENT OF COKE PLANT WASTEWATER WITH OR WITHOUT BLAST FURNACE
SLOWDOWN WATER IN A TWO-STAGE BIOLOGICAL FLUIDIZED BED SYSTEM
by
S.G. Nutt, Senior Project Manager, Canviro Consultants Ltd.,
178 Louisa Street, Kitchener, Ontario, Canada, N2H 5M5
H. Melcer, Head, Biological Processes Section, Wastewater
Technology Centre, Environment Canada, P.O. Box 5050,
Burlington, Ontario, Canada, L7R 4A6.
I.J. Marvan, Manager, Dearborn Environmental Consulting Services
P.O. Box 3060, Station A, Mississauga, Ontario, Canada, L5A 3T5
P.M. Sutton, Senior Process Engineer, Dorr-Oliver Inc.,
77 Havemeyer Lane, Stamford, Conn. 06904
ABSTRACT
Pilot-scale treatability studies conducted at Environment Canada's
Wastewater Technology Centre have demonstrated that the two-stage
biological fluidized bed is an effective high-rate system for treatment of
coke plant Wastewater alone and in combination with blast furnace blowdown
water. Greater than 90 percent removal of total nitrogen from undiluted
coke plant wastewater was achieved at a total system HRT of 16 hours. A
similar degree of treatment of a combined wastewater containing coke plant
wastewater and blast furnace blowdown water was achieved at a total system
HRT of 4.5 hours. In both cases, removal of conventional contaminants
including FOG, phenolic compounds and CNS approached or exceeded 90 percent
consistently.
Capital costs associated with treatment of 4950 rn^d""1 (1.31 MUSGPD)
of combined wastewater were essentially identical to capital costs
associated with treatment of 1300 nP-d"1 (0.34 MUSGPD) of coke plant
wastewater. Although annual direct operating costs for treatment of the
combined wastewater exceeded those related to coke plant wastewater
treatment by approximately $100,000, on a unit cost basis ($ per m3
treated) operating costs for treatment of the combined wastewater were less
than half those estimated for treatment of coke plant wastewater alone.
-------
INTRODUCTION
Byproduct coking operations and blast furnace operations are major
sources of wastewater in an integrated steel mill. Effluents from bypro-
duct coke ovens contain high concentrations of ammonia, thiocyanate, pheno-
lic compounds, cyanide and a variety of complex hydrocarbons. Slowdowns
from blast furnace gas cleaning recycle' systems contain lesser, albeit
significant, levels of ammonia and cyanide, but are relatively small con-
tributors of organic contaminants.
Biological treatment processes are commonly applied to coke oven
wastewaters to oxidize phenolic compounds and conventional organic contami-
nants. Recent research has shown that biological nitrification of coke
plant wastewaters can be achieved under conditions of stringent SRT control
(1,2,3)^ Treatment of blast furnace gas cleaning wastewater, beyond that
necessary to effect suspended solids removal and to control scaling and
corrosion in the gas scrubber system, has typically involved the applica-
tion of physical-chemical methods to oxidize ammonia and cyanide. The most
commonly applied process is alkaline breakpoint chlorination ^'. Biolo-
gical treatment methods are seldom applied to blast furnace blowdown
streams due to the low concentrations of organic contaminants and the
possible inhibitory effect of heavy metals present in the blowdown stream.
Several researchers have discussed the possibility of biologically
treating a combined stream of coke plant wastewater and blast furnace blow-
down water ^' . This treatment approach may eliminate the use of fresh-
water dilution commonly practised by steel mills during treatment of coke
plant wastewater. Furthermore, the utilization of a single treatment
facility for the combined wastewater stream may provide significant econo-
mic advantage. The technical feasibility of combined treatment of coke
plant wastewater and blast furnace blowdown water was recently demonstrated
at laboratory-scale '-*'.
Pilot-scale treatability studies conducted at Environment Canada's
Wastewater Technology Centre in Burlington, Ontario over the past two years
have demonstrated that the biological fluidized bed process, operated in
the pre-denitrification—nitrification flow mode, is an effective high-rate
system for treatment of coke plant wastewater alone and in combination with
blast furnace blowdown. The fluidized bed system was shown to offer signi-
ficant advantages in terms of reduced reactor volumes due to the high bio-
mass concentrations maintained on the fluidized support media. The pre-
denitrif ication-nitrification flow mode was shown to reduce the oxygen
requirements of the biological processes due to the anoxic removal of a
large fraction of the wastewater organics as well as to minimize the pro-
cess requirements for supplemental carbon and alkalinity.
Detailed results of the coupled fluidized bed biological treatment of
coke plant wastewater alone (6,7,8) an(j ^n combination with blast furnace
blowdown vy.iUJ have been presented elsewhere. In this paper, these per-
formance data are briefly reviewed and process design and cost data are
presented for coupled biological fluidized bed systems designed to treat
coke plant wastewater with and without blast furnace blowdown water.
-------
EXPERIMENTAL PROGRAM
PILOT PLANT DESCRIPTION
The process flowsheet of the coupled fluidized bed pilot plant is
shown in Figure 1. The plant configuration used for treatment of coke
plant wastewater alone and in combination with blast furnace blowdown water
was essentially the same with the exception that the facilities for supple-
mental methanol addition to the feed were not required during treatment of
coke plant wastewater alone. The system was designed to operate as a two-
stage separate sludge pre-denitrification-nitrification system to effect
phenol oxidation, thiocyanate oxidation, cyanide removal and complete
nitrogen control in separate anoxic and oxygenic biological reactors.
DO/pH CONTROL
SAND
WASTE
BIOMASS-*
BLAST
FURNACE
BLOWDOWN
SETTLED
SLUDGE
FIGURE 1. PROCESS FLOWSHEET OF TWO-STAGE FLUIDIZED BED
BIOLOGICAL TREATMENT SYSTEM
The anoxic denitrification reactor was cylindrical, 150 mm in diameter
and 4.3 m in height. Fluidized bed height was controlled at approximately
2.2 m, equivalent to an empty bed reactor volume of approximately 31
litres, although some variation in reactor height was allowed, depending on
operating conditions. The reactor had a conical inlet design to improve
influent flow distribution.
The oxygenic nitrification reactor was also cylindrical, 290 mm in
diameter and 4.3 m in height. Fluidized bed height was controlled at
approximately 3.9 m, equivalent to an empty bed reactor volume of approxi-
mately 210 litres.
- 3 -
-------
The nitrification reactor had a downflow inlet design to reduce turbu-
lence and shear in the reactor. Pure oxygen was supplied to the reactor
through a downflow oxygen contactor provided by Dorr—Oliver Inc. Oxygen
supply was automatically controlled by an analog PID feedback controller in
response to effluent dissolved oxygen concentration.
Tankage and pumps were provided to recycle treated effluent to the
anoxic reactor and the oxygenic reactor as required to maintain fluidiza-
tion of the reactors, to return nitrified effluent for denitrification and
ensure adequate oxygen transfer for the biological processes. A primary
clarifier was included in the process flow sheet but was not found to be
necessary for efficient pilot plant operation.
The pilot plant included feed systems to supply phosphoric acid as a
biological nutrient, sodium bicarbonate for pH control and as a source of
supplemental alkalinity for the nitrification reactions, and, in the case
of treatment of the combined wastewater, methanol as a supplemental carbon
source for the denitrification reactions. Temperature control was provided
to maintain the reactors in the range from 25°C to 30°C.
The support media in both reactors was quartzite sand with an effec-
tive size (d!0) of 0.48 mm and a uniformity coefficient (d60/d!0) of 1.23.
SOURCES OF PILOT PLANT FEED
Coke plant wastewater was obtained from the byproduct coke plant at
Dofasco Inc. in Hamilton, Ontario. The wastewater consisted of ammonia
still effluent obtained from the settling sump of a free and fixed-leg
still treating excess flushing liquor from the coke oven operation plus a
small flow of fractionator bottoms from the Phosam plant.
Blast furnace blowdown was obtained from a second integrated steel
mill which included four blast furnaces served by two separate water recir-
culation systems. Recirculation system A, serving three of the four fur-
naces, was operated initially at a blowdown of approximately 50 percent
but, during the course of the pilot plant study, blowdown was reduced to
approximately 15 percent. Recirculation system B, serving the remaining
blast furnace, operated at a blowdown of approximately 15 to 20 percent
throughout the study. Sinter plant gas scrubber water discharged to recir-
culation system A.
EXPERIMENTAL PROCEDURES
The sampling schedules and analytical procedures followed during the
pilot plant treatment of coke plant wastewater with and without blast fur-
nace blowdown were similar and have been described in detail elsewhere
(6,8,10).
The fluidized bed pilot plant was operated on coke plant wastewater
alone for a period of approximately 18 months between July 1980 and
December 1981. The experimental program consisted of an extended start-up
and acclimation period followed by approximately 12 months of non-steady
state operation. Process operating conditions during the non-steady state
- 4 -
-------
period were based on the results of the pseudo - steady state studies.
Process design data were developed based on 48 consecutive days of non-
steady state operation during which five batches of coke plant wastewater
were treated in the fluidized bed pilot plant. The results of the non-
steady state operating period are reviewed in this paper. Previous papers
have presented the results of the start-up and pseudo-steady state opera-
ting periods (o>'/.
Subsequent to the completion of the coke plant wastewater treatability
studies, the fluidized bed pilot plant was operated for a period of approx-
imately 6 months from January 1982 to July 1982 on a combined wastewater
containing coke plant wastewater and blast furnace blowdown water. Ratios
of blast furnace flowdown water to coke plant wastewater of 2 to 1 and 3 to
1 (by volume) "were used to cover the range of conditions anticipated at
full-scale. A ratio of 2:1 was used during the first four months and a
ratio of 3:1 was used during the final two months of plant operation.
Blowdown water from recirculation system B was used in the first mixed
batch of feed covering the first 3 1/2 months of plant operation as system
B contained a significantly higher nitrogen concentration than system A.
All subsequent feed batches utilized blowdown water from recirculation sys-
tem A. During the investigation, the process loading on the biological
reactors was progressively increased to determine the minimum reactor
hydraulic retention times required to maintain nitrification of the com-
bined wastewater and to achieve a treated effluent quality comparable to
that attained during treatment of coke plant wastewater. Due to the short-
term nature of the experimental program, biological steady state conditions
in terms of SRT were not attained in the process after step changes in the
hydraulic loading.
RESULTS
COKE PLANT WASTEWATER TREATMENT
The mean process operating conditions for the anoxic and oxygenic
reactors comprising the fluidized bed pilot plant during the non-steady
state operating period are summarized in Table 1. The average system
hydraulic retention time (HRT) was 17.7 hours. The HRT in the anoxic deni-
trification reactor was 2.4 hours. The HRT in the oxygenic nitrification
reactor was 15.3 hours. Mean system SRT was 42.0 days.
The process performance, based on 48 consecutive days of plant opera-
tion, is summarized in Table 2. As evident from these performance data,
removal of conventional carbonaceous contaminants (FOG) was consistently
high. Phenolic compounds were removed to a median level of 0.16 rng'L" ,
equivalent to a removal efficiency in excess of 99.9 percent.
As indicated by the data presented in Figure 2, efficient nitrifica-
tion was consistently maintained despite a two-fold variation in the oxi-
dizable nitrogen loading on the process. Median effluent NF^-N concentra-
tion was 1.3 mg'L"1, equivalent to a removal efficiency exceeding 99 per-
cent. The total cyanide content of the coke plant wastewater was reduced
- 5 -
-------
to 6.1 rng-L * from a median value of 8.0 mg TCN.L *- Based on a limited
amount of data collected at the end of the pilot plant study, the concen-
tration of cyanide compounds amenable to chlorination in the treated
effluent was reduced to the detection limit (0.01 mg-L~l).
TABLE 1: MEAN PROCESS OPERATING CONDITIONS DURING
TREATMENT OF COKE PLANT WASTEWATER
PARAMETER
HRT (hours)
SRT (days)
Reactor VS (g'L"1)
Recycle Ratio
Effluent D.O. (mg-L"1)
Temp. (°C)
Ammonia Loading*
(gNH3-N-g VS'i-d"1)
DENITRIFICATION
REACTOR
2.4
13.3
16.3
47.5
-
25.2 - 28.0
-
NITRIFICATION
REACTOR
15.3
62.4
14.5
270.0
2.9 - 3.7
25.5 - 28.0
0.021
SYSTEM
17.7
42.0
-
-
-
-
-
* Includes contribution of CNS to oxidizable nitrogen concentration,
TABLE 2: PROCESS PERFORMANCE DURING TREATMENT OF
COKE PLANT WASTEWATER
PARAMETER
FOC
CODF
Phenolics
TKNF
NH3-N
CNS
TCN
N02-N
N03-N
TN
SS
FEED*
700 + 51.
2750 + 297-
413 + 33.
247 + 41.
113 + 57.
346 + 61.
8.0 + 1.4
246 + 41.
— —
EFFLUENT*
56.0 + 4.
478.0 + 90.
0.16 + 0.03
11.9 + 1.6
1.3 + 0.6
2.4 + 0.5
6.1 + 1.4
3.0 + 0.9
1.6 + 0.8
16.5 + -
232 + 76.0
REMOVAL
(%)
92.0
82.6
>99.9
95.2
99.4**
99.4
23.8
93.3
—
* Concentration in mg-L *• based on 48 consecutive days of operation.
** Includes contribution of CNS to oxidizable nitrogen concentration.
As shown in Figure 3, the oxidized nitrogen (NOT-N) concentration in
the denitrif icat ion reactor effluent was reduced to less than 1 mg-L"-'-
consistently without supplementary carbon addition. The raw feed FOC/TKN
ratio averaged approximately 2.8 during the non-steady period, significan-
tly lower than the ratio of 3.5 found by Bridle et_ _al_ (D to be required
in suspended growth pre-denitrification systems treating coke plant waste-
water.
- 6 -
-------
330
329
30O
273
•T 230
J? "
" 2OO
1 '"
F tso
O «91-
z
o
u 100 -
1 2 S 1O 2O 3O 4O SO SO 7O 8O 9O 93 98 99
PERCENT OF OBSERVATIONS LESS THAN OR EQUAL TO STATED VALUE
FIGURE 2. REMOVAL OF OXIDIZABLE NITROGEN AND THIOCYANATE FROM
COKE PLANT WASTEWATER
-I
9
o
CC
t-
UJ
U
z
o
u
o
z
12 3 1O 2O 3O 4O 3O 6O 7O BO 9O 93 99 99
PERCENT OF OBSERVATIONS LESS THAN OR EQUAL TO STATED VALUE
FIGURE 3. REMOVAL OF OXIDIZED NITROGEN FROM COKE PLANT
WASTEWATER
- 7 -
-------
Total nitrogen removal efficiency over 48 consecutive days of non-
steady state operation was approximately 93.3 percent.
COMBINED COKE PLANT WASTEWATER AND BLAST FURNACE SLOWDOWN TREATMENT
The characteristics of the individual components of the blended feed
and the average concentrations of contaminants in the combined wastewater
used as feed to the fluidized bed pilot plant during the six month investi-
gation are shown in Table 3. The coke plant wastewater used in the com-
bined feed was similar to that treated previously in terms of the concen-
tration of nitrogen compounds and thiocyanate but contained lower concen-
trations of organic contaminants. The characteristics of the combined
wastewater were more strongly affected by the characteristics of the coke
plant wastewater than by the ratio of blast furnace blowdown to coke plant
wastewater utilized.
TABLE 3: CHARACTERISTICS OF INDIVIDUAL WASTE COMPONENTS AND
COMBINED WASTEWATER
PARAMETER
(mg-L-1)
FOG
Phenolics
NH3-N
CNS
TCN
CNA
TKN
Total Zinc
COKE PLANT
WASTEWATER
550.
241.
183.
290.
5.0
2.4
307.
0.4
BLAST FURNACE BLOWDOWN
SYSTEM
A
11.
5.1
22.1
3.3
5.3
5.2
28.2
20.3
SYSTEM
B
8.
0.1
33.4
1.7
1.6
1.5
40.0
6.1
COMBINED
WASTEWATER
215.
119.
66.4
98.1
1.8
0.5
111.
2.0
The range of operating conditions applied during the six—month treata-
bility study are summarized in Table 4. As discussed, the experimental
program involved sequential changes in the process loading to determine the
minimum reactor HRT's required to maintain efficient nitrification and
denitrification in the system. Total system HRT ranged from a minumum of 4
hours to a maximum of approximately 20 hours.
The data presented in Table 5 indicate that removal of organic conta-
minants (measured as FOC and phenolic compounds), thiocyanate and cyanide
was consistently high throughout the investigation despite step changes in
process loading and variations in feed quality. The removal efficiency of
phenolic compounds and thiocyanate exceeded 99 percent throughout the
period of pilot plant operation. Removal of nitrogen compounds (NH3~N, TKN
and NO^-N) depended on the process loading conditions as well as the feed
character is tics.
-------
TABLE 4: RANGE OF PROCESS OPERATING CONDITIONS DURING
TREATMENT OF COMBINED WASTEWATER
PARAMETER
HRT (hours)
SRT (days)
Reactor VS (g-L"1)
Recycle Ratio
Effluent D.O. (mg-L"1)
Temp. (°C)
Ammonia Loading*
(g NH3-N-g VS-i-d"1)
DENITRIFICATION
REACTOR
0.5 - 2.6
13. - 47.
8.9 - 26.8
7-50
-
24 - 31
-
NITRIFICATION
REACTOR
3.5 - 17.5
72. - 650.
14.9 - 26.2
50 - 270
1.8 - 5.6
24 - 31
0.005 - 0.037
* Includes contribution of CNS to oxidizable nitrogen concentration.
TABLE 5: REMOVAL OF FOG, PHENOLIC COMPOUNDS, THIOCYANATE AND
CYANIDE FROM COMBINED WASTEWATER
PARAMETER
FOG
Phenolic Compounds
CNS
TCN
CNA
FEED*
215. + 66.
119. + 36.
98.1 + 32.
1.8 + 1.3
0.5 + 0.8
EFFLUENT*
23. + 6.
0.09 + 0.04
0.7 + 0.5
0.8 + 0.9
0.1 + 0.3
REMOVAL
(%)
89.3
>99.9
99.3
57.9
77.8
* Concentration in mg-L 1 based on 196 consecutive days of operation.
Maximum specific denitrification rates for the combined wastewater
were not significantly different than the rates attained during treatment
of coke plant wastewater alone, averaging approximately 0.15 g NOT-N-g
VS~l-d~l. Complete denitrification of the combined wastewater was
attained at significantly lower FOC/TKN ratios than those required for
similar treatment of coke plant wastewater alone. Mass balances around the
denitrification reactor indicated the anoxic removal of b-etween 0.84 and
1.26 mg FOG per mg NO-p-N removed compared to the anoxic removal of 2.8 to
3.0 mg FOG per mg NO^-N during treatment of coke plant wastewater.
The most significant impact of variations in process Loading on system
performance was related to nitrification efficiency. As shown in Figure 4,
operation at oxidizable nitrogen loadings of up to 0-031 g Nt^-N-g
VS"l'd~~l resulted in treated effluent Nl^-N concentrations consistently
less than 2.0 mg-L"1. At higher loadings, instability LIT the efficiency
of nitrification was evident as indicated by the variability in treated
effluent NH3~N concentrations at these loading conditions.
- 9 -
-------
'_, 12
o>
E
- 10
o
z
O o
O 8
i-
z
UJ
D
UJ
2 -
A
O
A
RUN
1
2
3
4
5
6
7
LOADING
(g-g"1^'1)
0.005
0.008
0.013
0.031
0.037
0.033
0.030
RUN 5
RUN 6
RUNS
1,2,3,4,7
I I
1 5 1O 20 40 60 8O 9O 95 9899
PERCENT OF VALUES LESS THAN OR EQUAL TO STATED VALUE
FIGURE 4. EFFLUENT NH3~N CONCENTRATION DURING TREATMENT OF
COMBINED WASTEWATER
Figure 5 indicates the effect of process loading on effluent quality
in terms of Nt^-N concentration. At loadings up to 0.031 g NE^-N-g
VS~l-d~l, average effluent ammonia nitrogen concentrations of less than 1
mg-L~l were attained. At higher loadings, a deterioration in effluent
quality was evident as the loading exceeded the maximum specific nitrifica-
tion rate of the system. For the purpose of comparison, specific nitrifi-
cation rates achieved during treatment of coke plant wastewater alone were
approximately 0.022 g NH3-N-g VS^-d"1.
Nitrification efficiency was impaired during periods when supplemental
carbon as methanol was unnecessarily added to the denitrification system.
Zinc concentrations of up to approximately 3.4 mg-L~l did not appear to
affect nitrification in the fluidized bed system. At concentrations of 7
mg Zn-L"^, there appeared to be some inhibition to the nitrification sys-
tem but further data are necessary to confirm this observation. Prelimi-
nary results indicated that zinc could be precipitated from the combined
wastewater by adjustment of pH to the range of 9 to 10.
The suspended solids content of the treated effluent from the flui-
dized bed system was significantly lower during treatment of the combined
wastewater compared to that achieved during treatment of coke plant waste-
water alone. The average effluent suspended solids concentration during
treatment of the combined wastewater was approximately 20 mg-L~l compared
to an average level in excess of 200 mg-L"1 during treatment of coke
plant wastewater (Table 2).
- 10 -
-------
I6
o
z 5
o
UJ
.
u.
tr
tu
0.00
I
O.01 0.02 0.03
PROCESS LOADING (g-g
0.04
FIGURE 5. EFFECT OF PROCESS LOADING ON NITRIFICATION EFFICIENCY
PROCESS DESIGN AND COST ESTIMATES
DESIGN BASIS
Process design details for a coupled biological fluidized bed system
to treat coke plant wastewater and a combined stream of coke plant waste-
water and blast furnace blowdown water were developed based on the pilot-
scale treatability studies. Design flows, wastewater characteristics and
treated effluent quality for each case are summarized in Table 6. Raw
wastewater qualities were based on the actual characteristics of coke plant
wastewater and blast furnace blowdown measured during the pilot plant
studies. Design flow of coke plant wastewater was 1300 m^-d"! in both
cases. Design flow of blast furace blowdown water was 3650 rn^'d"^-, equi-
valent to 2.8 times the flow of coke plant wastewater.
Process design details for the coke plant treatment system and the
combined wastewater treatment system are compared in Table 7. The process
design flowsheet for the coke plant wastewater treatment system is shown
schematically in Figure 6. A comparative schematic for the treatment system
- 11 -
-------
designed for the combined wastewater is shown in Figure 7. As indicated in
Table 7, the coke plant wastewater treatment system was designed to provide
-a total system HRT of 16 hours compared to a system HRT of approximately
4.2 hours for equivalent treatment of the combined wastewater. A floccu-
lator-clarifier was included in the coke plant wastewater treatment flow-
sheet to achieve the required effluent suspended solids concentration.
Effluent suspended solids removal was not necessary during treatment of the
combined wastewater.
TABLE 6: PROCESS DESIGN BASIS FOR FLUIDIZED BED SYSTEMS
PARAMETER
Flo* (m3^-1)
FOC (mg-L-1)
NH3-N (mg -L"1)
CMS (mg-L-1)
TKN (mg-L-1)
NOT-N (ng-L-1)
SS (ng-L-1)
RAW WASTEWATERS
COKE PLANT
WASTEWATER
1300.
700.
100. ^ 50.
300.
ZOO.
-
BLAST FURNACE
SLOWDOWN
3650.
10.
30. +_ 10.
1.
40.
COMB INED
WASTEWATER
4950.
190.
48. +. 20.
80.
82.
-
TREATED EFFLUENTS
COKE PLANT
WASTEWATER
1300.
50.
1.
2.
-
15.
30.
COMB INED
WASTEWATER
4950.
20.
1.
1.
15.
30.
COKE PLANT
WASTEWATER
1300 m'-d"1
25O kg-cT1
85% H3PO4
6OO kg-d~'
90 X CaO
WASTE BIOMASS
195 kg VS- d~1
ANOXIC
DENTRIFICATKDN
REACTORS
WASTE BIOMASS
9O kg VS d"1
OXYGEN
195O kg-d
-t
OXYGENIC
NITRIFICATION
REACTORS
FLOCCULATOR
CLAR1FIER/
10.400 m'-d"1
-X- OXYGEN REQUIREMENTS ARE STATED
AS A MAXIMUM VALUE.
FIGURE 6. PROCESS DESIGN FLOWSHEET FOR TREATMENT OF COKE PLANT WASTEWATER
- 12 -
-------
TABLE 7. DESIGN SPECIFICATIONS FOR FLUIDIZED BED TREATMENT SYSTEMS
PARAMETER
HRT (hours)
Fluid Bed Ht . (m)
Reactor Ht . (m)
Reactor Area (m^)
Hydraulic Flux (m-min~^)
Recycle Ratio
Sand Inventory (tonnes)
COKE PLANT
WASTEWATER SYSTEM
DN REACTOR
2.0
5.5
7.6
19.9
0.40
8.0
87-
NIT REACTOR
14.0
6.4
7.6
119.
0.48
62.7
609.
COMBINED
WASTEWATER SYSTEM
DN REACTOR
0.53
2.6
4.6
42.2
0.40
4.0
87.
NIT REACTOR
3.7
6.4
7.6
119.
0.48
15.8
609.
COKE PLANT
WASTEWATER
130Om3-d~'
BLAST FURNACE
BLOWDOWN
.-I
315 kg-d
85% H3P04
-t
1660kg-d
SOX CaO
SUPPLEMENTAL
CARBON —
(AS REQUIRED)
WASTE BIOMASS
2OO kg VS-d
-I
WASTE BIOMASS
6O kg VS • d '
COMBINED
WASTEWATER
4950m3 d~1
ANOXIC
DENITRIFICATION
REACTORS
OXYGEN
I4OO kg-d
-1
OXYGENIC
NITRIFICATION
REACTORS
FINAL
EFFLUENT
19.800 m3-d~(
OXYGEN REQUIREMENTS ARE STATED
AS A MAXIMUM VALUE.
FIGURE 7. PROCESS DESIGN FLOWSHEET FOR TREATMENT OF COMBINED COKE PLANT
WASTEWATER/BLAST FURNACE BLOWDOWN
CAPITAL AND OPERATING COST ESTIMATES
On the basis of the process designs developed for the biological flui-
dized bed systems, preliminary cost estimates were developed for major
capital items and for direct operating costs (chemicals and power) related
to the treatment systems.
Capital cost estimates for the coupled fluidized bed system are sum-
marized in Table 8. In both cases, it was assumed that pipeline oxygen
would be available from the cryogenic generator associated with the BOF.
It was also assumed that the lime requirements could be satisfied from the
lime feeding system associated with the fixed—leg ammonia still and costs
- 13 -
-------
for a dedicated lime slaking facility and lime slurry feed system were not
included in the cost estimate. Reactor costs were based on parallel anoxic
and oxygenic fluidized beds.
TABLE 8: COMPARISON OF CAPITAL COSTS RELATED TO COKE PLANT TREATMENT
WITH AND WITHOUT BLAST FURNACE SLOWDOWN
ITEM
1. Equipment, instrumentation and
engineering for reactors and
clarif ier
2. Concrete reactors, oxygenators,
pump sumps and clarifier
3 . Pumps and pump rooms
4. Piping, valving (erected basis)
5. Installation of equipment, motor
control center, wiring, site work
TOTAL
COST (CANADIAN $ x 1000)
WITHOUT BLAST
FURNACE SLOWDOWN
1,010.
460.
292.
283.
312.
2,357.
WITH BLAST
FURNACE SLOWDOWN
1,050.*
413.*
*
300.
276.
300.
2,339.
* No clarifier required for treatment of combined wastewater
As shown in Table 8, the capital costs associated with treatment of
1300 m-^-d"^ of coke plant wastewater were essentially the same as those
associated with treatment of 4950 m^-d"! of combined coke plant waste-
water and blast furnace blowdown.
Annual direct operating costs for chemicals and power for both treat-
ment systems are summarized in Table 9. The chemical requirements for each
system were shown in the process design flowsheets (Figures 6 and 7).
TABLE 9: COMPARISON OF ANNUAL DIRECT OPERATING COSTS RELATED TO
COKE PLANT TREATMENT WITH AND WITHOUT BLAST FURNACE BLOWDOWN
ITEM
1. Phosphoric Acid
85% H3P04 @ $0.90 per kg
2. Lime — bulk quicklime
(90% CaO) @ $0.061 per kg
3. Oxygen @ $0.06 per kg*
4. Polymer @ $6.30 per kg
5. Power @ $0.03 per kwh
TOTAL ANNUAL COST
UNIT COST ($ per m3 treated)
COST (CANADIAN $)
WITHOUT BLAST
FURNACE BLOWDOWN
82,100.
13,400.
28,500.
2,300.
24,600.
150,900.
0.32
WITH BLAST
FURNACE BLOWDOWN
169,300.
37,000.
21,900.
0.
24,250.
252,450.
0.14
* Based on pipeline oxygen
- 14 -
-------
Based on the design wastewater quality, supplemental carbon addition would
not be necessary in either system to maintain complete denitrification.
Operating costs for treatment of the combined wastewater were approximately
$100,000 per year higher than similar costs for treatment of coke plant
wastewater. The cost difference is primarily the result of increased costs
associated with higher phosphoric acid utilization. On a unit cost basis
($ per m^ treated), operating costs associated with treatment of the com-
bined wastewater were less than half the costs associated with treatment of
coke plant wastewater alone.
CONCLUSIONS
Based on pilot-scale investigations of the coupled fluidized bed pro-
cess for nitrification and denitrif ication of coke plant wastewater with
and without blast furnace blowdown water, the following conclusions can be
drawn:
• To achieve greater than 90 percent removal of total nitrogen
from undiluted coke plant wastewater required a total system HRT
of approximately 16 hours. A similar degree of treatment of
combined wastewater containing coke plant wastewater and blast
furnace blowdown water was achieved at a total system HRT of
approximately 4.5 hours.
• In both cases, removal of conventional contaminants including
FOG, phenolic compounds and CNS approached or exceeded 90 per-
cent consistently. Effluent quality in terms of FOC, phenolic
compounds, CNS, TCN and suspended solids was superior during
treatment of the combined wastewater compared to that achieved
during treatment of coke plant wastewater.
• Estimated capital costs associated with treatment of 4950
m^'d"^- of combined wastewater were essentially identical to
capital costs associated with treatment of 1300 m->-d~l of coke
plant wastewater.
• Annual direct operating costs related to treatment of coke
plant wastewater with blast furnace blowdown water were approxi-
mately $100,000 per year higher than those estimated for treat-
ment of coke plant wastewater alone. On a unit cost basis ($
per m^ treated), operating costs associated with treatment of
the combined wastewater were less than half those associated
with treatment of coke plant wastewater alone.
ACKNOWLE DGEMENT S
The authors wish to express their appreciation to Dofasco Inc. for
their co-operation throughout the study and their financial support of the
combined wastewater treatability investigation. In addition, the contribu-
tions of Mr. R.R. Evans of Dorr-Oliver Inc. to the development of the cost
estimates is gratefully acknowledged.
- 15 -
-------
REFERENCES
1. Bridle, T.R., Bedford, W.K. and B.E. Jank, "Biological Nitrogen
Control of Coke Plant Wastewaters", Prog. Wat. Tech., 12, (1980).
2. Medwith, B.W. and J.F. Lefelhocz, "Single-Stage Biological Treatment
of Coke Plant Wastewaters with a Hybrid Suspended Growth - Fixed Film
Reactor", Proc. of the 36th Purdue Industrial Waste Conf., 68, (1981).
3. Wong-Chong, G. and J.D. Hall, "Single-Stage Nitrification of Coke
Plant Wastewater", Proc. of the Symposium on Iron and Steel Pollution
Abatement Technology for 1980, EPA-600/9-81-017, 395, (1981).
4. Hofstein, H. and H.J. Kohlmann, "An Investigation of Foreign By-
Product Coke Plant and Blast Furnace Wastewater Control Technology",
Proc. of the Symposium on Iron and Steel Pollution Abatement Techno-
logy for 1980, EPA-600/9-81-017, (1981).
5. Gauthier, J.J., Jones, D.D., Wilson, L.W. and C.R. Majors, "Combined
Biological Treatment of Coke Plant Wastewater and Blast-Furnace
Recycle - Water System Slowdown", Proc. of the 36t^ Purdue Industrial
Waste Conference, 77, (1981).
6. Nutt, S.G., Melcer, H. and J.H. Pries, "Two-Stage Biological Fluidized
Bed Treatment of Coke Plant Wastewater for Nitrogen Control", pre-
sented at the 54th Water Pollution Control Federation Conference,
Detroit, Michigan, (1981).
7. Nutt, S.G., Melcer, H. , Marvan, I.J. and P.M. Sutton, "Treatment of
Coke Plant Wastewater in the Coupled Pre-Denitrification Fluidized Bed
Process", presented at the 37th Purdue Industrial Waste Conderence,
Lafayette, Ind., (1982).
8. Nutt, S.G. and I.J. Marvan, "Biological Fluidized Bed Treatment of
Complex Industrial Wastes for Nitrogen and Contaminant Control", Final
Report to Environmental Protection Service, Environment Canada (in
press) .
9. Melcer, H. , Nutt, S.G., Marvan, I.J. and P.M. Sutton, "Combined Treat-
ment of Coke Plant Wastewater and Blast Furnace Slowdown Water in a
Coupled Biological Fluidized Bed System", presented at the 55th Water
Pollution Control Federation Conference, St. Louis, Missouri, (1982).
10. Nutt, S.G. and I.J. Marvan, "Biological Fluidized Bed Treatment of
Combined Coke Plant Wastewater and Blast Furnace Slowdown", Final
Report to Environmental Protection Service, Environment Canada, (in
press) .
- 16 -
-------
PREREGISTRANTS
SYMPOSIUM ON IRON AND STEEL POLLUTION ABATEMENT TECHNOLOGY FOR 1982
The William Penn Hotel
Pittsburgh, PA
November 16-18, 1982
Michael R. Alford
U.S. EPA
401 M Street, S.W. (PM-219)
Washington, DC 20460
202/382-2698
Gary A. Amendola
U.S. EPA
25089 Center Ridge Road
Westlake, OH 44145
216/835-5200
Gopal Annamraju
PEDCo Environmental, Inc.
11499 Chester Road
Cincinnati, OH 45246
513/782-4700
Jacques Antoine
L.E.C.E.S.
B.P. 36
Maizieres-Les-Metz,
(8) 780-21-11
FRANCE 57210
Cornelis A. Aronds
Estel Hoogovens BV
Postbus 10000
1970 CA Umuiden, NETHERLANDS
Oil 31251 095232
Charles W. Askins
Babcock & Wilcox
P. 0. Box 401
Beaver Falls, PA 15010
412/846-0100, Ext. 2694
Franklin A. Ayer
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, NC 27709
919/541-6260
K. Shankar Banninthaya
Allegheny County Health Department
301 39th Street, APC
Pittsburgh, PA 15201
412/578-8117
Leslie L. Beck
U.S. EPA
OAQPS, MD-13
Research Triangle Park, NC 27711
919/541-5601
Peter N. Bibko
U.S. EPA, Region III
Curtis Bldg., 6th and Walnut Streets
Philadelphia, PA 19106
212/264-2525
Thomas R. Blair
Radian Corporation
P. 0. Box 9948
Austin, TX 78766
512/454-4797
David G. Boltz
Bethlehem Steel Corporation
Room B-252 Martin Tower
Bethlehem, PA 18016
215/694-2721
Peter Brand
Hatch Associates, Ltd.
21 St. Glair Avenue East
Toronto, Ontario, CANADA M4T 1L9
416/962-6350
Nick Browne
Bischoff Environmental Systems
135 Cumberland Road
Pittsburgh, PA 15237
412/364-8860
Nicholas S. Buchko
Shenango Incorporated
200 Neville Road
Pittsburgh, PA 15225
412/777-6655
D. F. Cairns
National Engineers and Associates
7777 Bonhomme Avenue
St. Louis, MO 63105
314/725-7210
S. Charles Caruso
Mellon Institute
4400 Fifth Avenue
Pittsburgh, PA 15213
412/578-3320
Rodney A. Cerny
The H. K. Ferguson Company
One Erieview Plaza
Cleveland, OH 44114
216/523-5690
-------
Keith L. Cherryholmes
Envirodyne Engineers, Inc.
12161 Lackland Road
St. Louis, MO 63141
314/434-6960
Ronald J. Chleboski
Allegheny County Health Department
301 39th Street, APC
Pittsburgh, PA 15201
412/578-8101
Jesse R. Conner
SolidTek, Inc.
5371 Cook Rd., P. 0. Box 888"
Morrow, GA 30260
404/361-6181
Alan Conners
Pennsylvania Engineering Corporation
32nd Street and AVRR
Pittsburgh, PA 15201
412/288-6800
Joseph G. Crist
U.S. Steel Corporation
125 Jamison Lane
Monroeville, PA 15146
412/372-1212, Ext. 2906
Thomas Cuscino, Jr.
Midwest Research Institute
425 Volker Boulevard
Kansas City, MO 64110
816/753-7600
Edwin R. Daly
Krupp Wilputte Corporation
152 Floral Avenue
Murray Hill, NJ 07974
201/464-5900
Richard A. D'Amico
Busch Co.
904 Mt. Royal Boulevard
Pittsburgh, PA 15223
412/487-7100
Richard P. de Filippi
Critical Fluid Systems, Inc.
25 Acorn Park
Cambridge, MA 02140
617/492-1631
Lex de Jonge
Prov. Dept. of Public Work and Water
Zijlweg 245
2015 CL Haarlem, NETHERLANDS
023-319350
Paul J. DeCoursey, III
Michael Baker, Jr., Inc.
4301 Dutch Ridge Road, Box 280
Beaver, PA 15009
412/495-7711, Ext. 365
Anthony T. DeQuittner
ATD Consultants
314 Pleasant Avenue
McMurray, PA 15317
412/941-5653
James H. Dougherty
Roy F. Weston, Inc.
Weston Way
West Chester, PA 19380
215/692-3030, Ext. 240
Lynn R. Emerson
Calgon Corporation
P. 0. Box 1346
Pittsburgh, PA 15230
412/777-8445
Terry R. Fabian
PA Dept. of Environmental Resources
Rm 850 Kossman Bldg., 100 Forbes Ave
Pittsburgh, PA 15222
412/565-2336
Patrick C. Falvey
NUS Corporation
Park West II, Cliff Mine Road
Pittsburgh, PA 15275
412/788-1080
Simon Feigenbaum
Allegheny County Health Department
301 39th Street, APC
Pittsburgh, PA 15201
412/578-8103
Robert L. Felt
Allegheny County Health Department
301 39th Street, APC, Building 7
Pittsburgh, PA 15201
412/578-8115
George P- Fenton
Busch Co.
904 Mt. Royal Boulevard
Pittsburgh, PA 15223
412/487-7100
Marjorie J. Fitzpatrick
JACA Corp.
550 Pinetown R.oad
Fort Washington, PA 19034
215/643-5466
-------
Joseph R. Ford
Babcock & Wilcox
p. 0. Box 401
Beaver Falls, PA 15010
412/846-0100, Ext. 2449
James R. Forrelli
Betz Converse Murdoch, Inc.
5777 Baum Boulevard
Pittsburgh, PA 15206
412/361-6000
Steve Francis
Annco Inc.
1808 Crawford Street
Middletown, OH 45043
513/425-3476
R. W. Fullerton
U.S. Steel Corporation
125 Jamison Lane
Monroeville, PA 15146
412/372-1212
Walter I. Goldburg
Group Against Smog and Pollution
1237 Murdoch Road
Pittsburgh, PA 15217
412/624-4317
Jeffrey H. Greenfield
University of Pittsburgh
School of Engineering
Pittsburgh, PA 15261
412/624-5380
Murray S. Greenfield
Dofasco Inc.
P. 0. Box 460
Hamilton, Ontario, CANADA L8N 3J5
416/544-3761, Ext. 3681
Kenneth 0. Groves
Dow Chemical U.S.A.
2020 Dow Center
Midland, MI 48640
517/636-3246
James A. Hagarman
Calocerinos & Spina
1020 Seventh North Street
Liverpool, NY 13088
315/457-6711
Andrew M. Halatin
Busch Co.
904 Mt. Royal Boulevard
Pittsburgh, PA 15223
412/487-7100
Nona L. Hancock
National Engineers and Associates
7777 Bonhomme Avenue
St. Louis, MO 63105
314/725-7210
Penelope Hansen
U.S. EPA
401 M Street, S.W. (WH-565A)
Washington, DC 20460
202/382-4756
Robert Harshman
Green International, Inc.
504 Beaver Street
Sewickley, PA 15143
412/761-2770
Fred C. Hart
Fred C. Hart Associates, Inc.
530 Fifth Avenue
New York, NY 10036
212/840-3990
Jim Haskill
Environment Canada
Water Pollution Control Directorate
Ottawa, Ontario, CANADA K1A 1C8
819/997-2270
J. 0. Hawthorne
U.S. Steel Corporation
125 Jamison Lane
Monroeville, PA 15146
412/372-1212
Jorgen G. Hedenhag
Flakt, Inc.
1500 E. Putnam Avenue
Old Greenwich, CT 06870
203/637-5401
C. P- Heijwegen
Estel Hoogovens BV
Postbus 10000
1970 CA Umuiden, NETHERLANDS
02510-92451
Hans-Friedrich Hinrichs
Saarbergwerke AG
Triererstrasse 1
6600 Saarbrucken, W GERMANY
0681-405-3787
C. F. Hoffman
U.S. Steel Corporation
125 Jamison Lane
Monroeville, PA 15146
412/372-1212
-------
Thomas W. Hoffman
Midrex Corporation
One NCNB Plaza, 38th Floor
Charlotte, NC 28280
704/373-1600
Gary Hudiburgh
U.S. EPA
401 M Street, S.W. (EN-336)
Washington, DC 20460
202/755-0750
Michael Jasinski
GCA Corporation/Technology Division
213 Burlington Road
Bedford, MA 01730
617/275-5444
John D. Jeffery
GCA Corporation/Technology Division
213 Burlington Road
Bedford, MA 01730
617/275-5444
Gate Jenkins
U.S. EPA
401 M Street, S.W.
Washington, DC 20460
202/382-4801
Ch. R. Josis
CRM (Centre de Recherches Metallurgiques)
Rue Ernest Solvey, 11
4000 Liege, BELGIUM
41/52.70.50
Randy Junkins
Roy F. Weston, Inc.
Weston Way
West Chester, PA 19380
215/692-3030, Ext. 313
Woody Kawaters
TRC Environmental Consultants, Inc.
800 Connecticut Blvd.
East Hartford, CT 06108
203/289-8631
Maria Kelleher
MacLaren Engineers
33 Yonge Street
Toronto, Ontario, CANADA M5E 1E7
416/365-7337
William Kemner
PEDCo Environmental, Inc.
11499 Chester Road
Cincinnati, OH 45246
513/782-4700
Richard W. Klippel
Cajjcerinos & Spina
102;, Seventh North Street
Liverpool, NY 13088
315/i57-6711
Michael J. Kozy
Dravo Corp.
One Oliver Plaza
Pittsburgh, PA 15222
412/566-5115
E. Radha Krishnan
PEDCo Environmental, Inc.
11499 Chester Road
Cincinnati, OH 45246
513/782-4700
Kenneth Krupinski
U.S. Steel Corporation
125 Jamison Lane, MS-57
Monroeville, PA 15146
412/372-1212,. Ext. 2472
Cezary Krzymowski
Illinois EPA
1701 First Avenue, DAPC
Maywood, IL 60153
312/345-9780
Robert H. Lace, Sr.
The Munters Corporation
E. 0. Box 6428
Ft. Myers, FL 33911
813/936-1555
Joseph F. Lennon
Ford Motor Company
One Parklane Blvd.,
Dearborn, MI 48126
313/322-1227
628 Parklane W.
Eddy Lin
Pacific Environmental Services, Inc.
465 Fullerton
Elmhurst, IL 60126
312/530-7272
Joseph H. Lucas
Mellon Institute
4400 Fifth Avenue
Pittsburgh, PA 15213
412/578-3409
Richard W. Masters
U.S. Steel Corporation
600 Grant Street
Pittsburgh, PA 15230
412/433-6553
-------
Robert C. McCrillis
U.S. EPA
IERL, MD-63
Research Triangle Park, NC 27711
919/541-2733
Robert M. McMullen
Bethlehem Steel Corporation
Room B-252, Martin Tower
Bethlehem, PA 18016
215/694-7116
Connally E. Hears
U.S. EPA, Region VIII
1860 Lincoln Street (8AW-AP)
Denver, CO 80295
303/837-3711
Daniel J. Metzger
National Engineers and Associates
7777 Bonhomme Avenue
St. Louis, MO 63105
314/725-7210
Andrew C. Middleton
Koppers Company, Inc.
440 College Park Drive
Monroeville, PA 15146
412/327-3000, Ext. 5132
E. L. Mihelic
U.S. Steel Corporation
125 Jamison Lane
Monroeville, PA 15146
412/372-1212
Jaffer Mohiuddin
Pacific Environmental Services, Inc.
8230 Boone Boulevard, Suite 450
Vienna, VA 22180
703/790-0272
Samuel M. Murphy
KY Dept. of Environmental Protection
Ft. Boone Plaza, 18 Reilly Road
Frankfort, KY 40601
502/564-3382, Ext. 308
Daniel J. Murray
U.S. EPA
25089 Center Ridge Road
Westlake, OH 44145
216/835-5200
Theodore W. Nelson, Jr.
Peter F. Loftus Corporation (Illinois)
223 W. Jackson Boulevard
Chicago, IL 60606
312/939-0581
Richard L. Neraeth
Republic Steel Corporation
3100 East 45th Street
Cleveland, OH 44127
216/622-6376
Ronald D. Neufeld
University of Pittsburgh
Department of Civil Engineering
Pittsburgh, PA 15261
412/624-5362
Pauline Nixon
University of Pittsburgh
209 South Braddock Avenue #2
Pittsburgh, PA 15221
412/963-1999
Stephen G. Nutt
Canviro Consultants, Ltd.
178 Louisa Street
Kitchener, Ontario, CANADA N2H 5M5
519/579-3500
Bernie O'Barsky
National Steel Corporation
2800 Grant Building
Pittsburgh, PA 15219
412/263-4746
Hidehiro Obata
Retired Civil Engineer
Ave. Alvares Cabral, n. 917, apt. 203
3000 Bela Horizonte, BRAZIL
Telephone Not Furnished
Andrew P. Pajak
Michael Baker, Jr., Inc.
P. 0. Box 280
Beaver, PA 15009
412/495-7711
George Pashel
Bethlehem Steel Corporation
Room B-252, Martin Tower
Bethlehem, PA 18016
215/694-6108
James W. Patterson
Illinois Institute of Technology
IIT Center
Chicago, IL 60616
312/567-3535
Stephen F. Pedersen
PA Dept. of Environmental Resources
600 Kossman Bldg., 100 Forbes Ave.
Pittsburgh, PA 15222
412/565-5103
-------
Earl R. Pennell
Wheatland Tube Co.
Council Avenue
Wheatland, PA 16161
412/342-6851
Michael S. Peters
Structural Metals, Inc.
P. 0. Box 911
Seguin, TX 78155
512/379-7520
Jacques Pivont
Groupement de la Siderurgie
Rue Montoyer 47
1040 Brussels, BELGIUM
02/513.38.20
Richard T. Price
PEDCo Environmental, Inc.
11499 Chester Road
Cincinnati, OH 45246
513/782-4621
B. A. Procyk
U.S. Steel Corporation
125 Jamison Lane
Monroeville, PA 15146
412/372-1212
Joe Quigley
CAMETCO, Inc.
P. 0. Box 266
Duquesne, PA 15110
412/466-6897
John Reggi
WV Air Pollution Control Commission
1911 Warwood Avenue
Wheeling, WV 26003
304/277-2662
Hooite Rugge
Kaiser Steel Corporation
P. 0. Box 217
Fontana, CA 92335
714/350-5448
John S. Ruppersberger
U.S. EPA
IERL, MD-63
Research Triangle Park, NC 27711
919/541-2733
William R. Samples
Wheeling-Pittsburgh Steel
Duvall Center
Wheeling, WV 26003
304/234-2936
Peter W. Sawchuck
Koppers Company, Inc.
440 College Park Drive
Monroeville, PA 15146
412/327-3000, Ext. 5151
Robert Schlosser
JACA Corp.
550 Pinetown Road
Fort Washington, PA 19034
215/643-5466
Stephen M. Schwartz
American Iron & Steel Institute
1000 16th Street, N.W.
Washington, DC 20036
202/452-7275
John P. Shaughnessy
MikroPul Corporation
10 Chatham Road
Summit, NJ 07901
201/273-6360
Donald W. Simmons
National Steel Corporation
2800 Grant Building
Pittsburgh, PA 15219
412/263-4395
Larry L. Simmons
Energy & Environmental Management, I
P. 0. Box 71
Murrysville, PA 15668
412/247-5124
Anthony M. Skicki
PA Dept. of Environmental Resources
P. 0. Box 2063, Water Quality Mgmt.
Harrisburg, PA 17120
717/787-8184
John R. Smith
Koppers Company, Inc.
440 College Park Drive
Monroeville, PA 15146
412/327-3000, Ext. 5309
Peter D. Spawn
GCA Corporation/Technology Division
213 Burlington Road
Bedford, MA 01730
617/275-5444
D. E. Splitstone
U.S. Steel Corporation
125 Jamison Lane
Monroeville, PA 15146
412/372-1212
-------
Robert M. Stein
AWARE, Inc.
P.-O. Box 40284
Nashville, TN 37204
615/794-0110
Richard D. Stern
U.S. EPA
IERL, MD-63
Research Triangle Park, NC 27711
919/541-2547
Anton Telford
Illinois EPA
2200 Churchill Road
Springfield, IL 62706
217/782-2113
C. Joseph Touhill
Michael Baker, Jr., Inc.
4301 Dutch Ridge Road, Box 280
Beaver, PA 15009
412/495-7711, Ext. 306
Martin G. Trembly
U.S. EPA
25089 Center Ridge Road
Westlake, OH 44145
216/835-5200
Stephen Vajda
Jones & Laughlin Steel Corporation
900 Agnew Road
Pittsburgh, PA 15227
412/884-1000
Ralph W. Waechter
PA Dept. of Environmental Resources
600 Kossman Bldg., 100 Forbes Ave.
Pittsburgh, PA 15222
412/565-7957
Anna W. Wallace
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, NC 27709
919/541-6967
William L. West
Republic Steel Corporation
P. 0. Box 6778, Room 520-R
Cleveland, OH 44101
216/622-5911
Roger C. Westman
Allegheny County Health Department
301 39th Street, APC
Pittsburgh, PA 15201
412/578-8103
Robert H. Wills, Jr.
Crucible, Inc.
P. 0. Box 977
Syracuse, NY 13201
315/487-4111
Leon W. Wilson, Jr.
U.S. Steel Corporation
125 Jamison Lane
Monroeville, PA 15146
412/372-1212, Ext. 2206
L. D. Wisniewski
Republic Steel Corporation
P. 0. Box 6778, Env. Control Dept.
Cleveland, OH 44101
216/622-5910
W. S. Workman
U.S. Steel Corporation
125 Jamison Lane
Monroeville, PA 15146
412/372-1212
John E. Yocom
TRC Environmental Consultants, Inc.
800 Connecticut Blvd.
East Hartford, CT 06108
203/289-8631
Earle F. Young, Jr.
American Iron & Steel Institute
1000 16th Street, N.W.
Washington, DC 20036
202/452-7271
James R. Zwikl
Shenango, Inc.
200 Neville Road
Pittsburgh, PA 15225
412/777-6654
------- |