MIDWEST RESEARCH INSTITUTE
MRI
SUMMARY OF FACTORS AFFECTING COMPLIANCE BY
FERROUS FOUNDRIES
VOLUME II - APPENDICES A-E
FINAL REPORT
Date Prepared: April 30, 1981
EPA Contract No. 68-01-4139, Task No. 15
MRI Project Nos. 4310-1(15)
For
Division of Stationary Source Enforcement
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, D.C. 20460
Attn: Robert L. King
-------
DISCLAIMER
This report has been reviewed by the Division of Stationary Source
Enforcement, U.S. Environmental Protection Agency, and approved for pub-
lication. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
MRI-NORTH STAR LABORATORIES 10701 Red Circle Drive, Minnetonka, Minnesota 55343 • 612 933-7880
MRI WASHINGTON, D.C. 20006-Suite 250,1750 K Street, N.W. • 202 293-3800
-------
SUMMARY OF FACTORS AFFECTING COMPLIANCE BY
FERROUS FOUNDRIES
VOLUME II - APPENDICES A-E
FINAL REPORT
Date Prepared: April 30, 1981
EPA Contract No. 68-01-4139, Task No. 15
MRI Project Nos. 4310-1(15)
Division of Stationary Source Enforcement
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, D.C. 20460
MIDWEST RESEARCH INSTITUTE 425 VOLKER BOULEVARD, KANSAS CITY, MISSOURI 64110 • 816753-7600
-------
PREFACE
Midwest Research Institute has carried out a study for the Division of
Stationary Source Enforcement, Environmental Protection Agency, to review
the various technical and regulatory factors that affect the compliance of
ferrous foundries.
These appendices present the results of the study including characteris-
tics of the ferrous foundry industry, emissions from ferrous foundries, the
design, and operation and maintenance of emissions controls.
Mr. D. Wallace, Associate Environmental Scientist, Environmental Control
Systems Section, served as project leader, and Mr. J. Hennon, Senior Chemist,
and Mr. B. Boomer, Assistant Environmental Engineer of MRI contributed sig-
nificantly to the task. The assistance provided by Mr. A. Trenholm, Head
Environmental Control Systems Section and the guidance provided by Task Manager,
Mr. Robert L. King, throughout the project are gratefully acknowledged.
Approved for:
MIDWEST RESEARCH INSTITUTE
M. P. Schrag, Director!
Environmental Systems Department
April 30, 1981
111
-------
CONTENTS
Preface iii
Tables vi
Figures viii
Appendix A - Description of Ferrou Foundry Processes 1
A.I General process description 2
A. 2 Mold and core preparation 9
A.3 Furnace charge preparation 13
A.4 Melting and casting 16
A.5 Cleaning and finishing 27
A.6 Sand handling 29
References 32
Appendix B - Quantification of Particulate Emissions for Major
Foundry Emissions Sources 35
B.I Melting emissions 36
B.2 Nonmelting emissions 55
References 65
Appendix C - Description of Available Control Systems 67
C.I Cupola emission controls 68
C.2 Electric arc furnace controls 85
C.3 Shakeout and sand handling emission controls . . 101
C.4 Cleaning room controls Ill
References 119
Appendix D - Operation and Maintenance of Control Equipment . . 123
D.I Operation and maintenance of Venturi scrubbers . 124
D.2 Operation and Maintenance of fabric filters. . . 129
References 147
Appendix E - Procedures for Troubleshooting and Correction of
Baghouse Malfunctions 149
-------
TABLES
Number Page
A-l Chemical Composition of Ferrous Castings 9
A-2 Principal Organic Core Binders in Use in the United
States 12
A-3 Products of Thermal Decomposition of Sand Binders. ... 14
A-4 Functional Groups Observed in Infrared Absorption
Spectra of Condensed Liquid Phases 15
B-l Parameters of Cupola Furnaces-Linear Regression Analysis
of Emissions Affected by Furnace Design Factors. ... 38
B-2 Linear Regression Analysis Observation 39
B-3 Chemical Composition of Cupola Particulate Emissions . . 41
B-4 Particle Size Distribution-Cupola Emissions 42
B-5 Summary of Particulate Emission 44
B-6 Influence of Charging Practice on Dust Production at
Foundry A 45
B-7 Results of Cupola Testing in the Federal Republic of
Germany '. . . 49
B-8 Additional Cupola Emissions Data 51
B-9 Emissions Data from Electric Arc Melting Furnaces. ... 53
B-10 Electric Furnace Emissions Data 54
B-ll Size Distribution for Three Electric Arc Installations . 54
B-12 Pouring and Cooling Emissions 56
B-13 Shakeout Emissions 59
B-14 Sand Handling Emissions 61
VI
-------
TABLES (concluded)
Number Page
B-15 Cleaning Room Emissions 62
B-16 Emissions from Roof Exhausts at an Iron Foundry 64
C-l Comparative Properties of Most Popular Filter Fibers . . 84
C-2 Comparison of Emissions from Green Sand and Permanent
Mold Processes for Producing a 13-lb Uncored Casting
Under Ventilated Conditions 97
C-3 Typical Operations Found in Foundry Spent Sand Systems . 108
C-4 Downdraft Bench Exhausts Compared to Recommended Flows . 114
C-5 Description of HVLV Controlled Tools, Case E 116
D-l Maintenance for Plugging and Scaling Venturi Scrubber. . 127
D-2 Scrubber Maintenance . . . 128
D-3 Spare Parts Inventory for Venturi Scrubber 130
D-4 Type of Maintenance Required - Venturi Scrubber Systems. 131
D-5 Checklist for Routine Inspection of Baghouse 135
D-6 Baghouse Collector Maintenance 136
D-7 Approximate Cost of Replacement Bags 143
D-8 List of Replacement Parts for a Baghouse Filter 144
VII
-------
FIGURES
Number
A-l
A-2
A-3
A- 4
A-5
A-6
A-7
A-8
A-9
A- 10
A- 11
A-12
A-13
B-l
B-2
C-l
C-2
C-3
04
C-5
General Foundry Flow Diagram
Process Flow Diagram - Raw Material Storage and Furnace
Charge Make-up
Core and Mold Preparation
Melting and Casting
Sand Handling
Illustration of a Foundry Cupola
Illustration of an Electric Arc Furnace
Illustration of a Coreless Induction Furnace
Illustration of a Channel Induction Furnace
Illustration of a Reverberatory Furnace
Methods of Iron Inoculation
Process Flow Diagram - Cleaning and Finishing
Line Drawing of Canton Malleable 's Sand System Showing
Plowoff Points arid Resultant Sand Temperatures ....
Particle Size of Cupola Emissions at 5 Canadian Found-
ries
Average Particle Size Distribution for 7 U.S. Foundries.
Typical Cupola Wet Cap
Typical Cupola Scrubber System
Examples of Venture Scrubbers
Flooded Disk Scrubber
Typical Fabric Filter
Page
3
4
5
6
7
17
18
19
20
21
26
28
31
46
48
69
71
72
74
76
VI11
-------
FIGURES (continued)
Number
C-6
C-7
C-8
C-9
C-10
C-ll
C-12
C-13
C-14
C-15
C-16
C-17
C-18
C-19
C-20
C-21
C-22
C-23
C-24
' Reverse Air or Shaker Type Baghouse
Baghouse Showing Two Methods of Cleaning by Reverse Air
Flow
Pulse Jet Type Baghouse
Compartmentalized Reverse Air Cleaning
Capture Mechanism for EAF Melting Emissions
Canopy Hood
Building Evacuation
Sketch of Furnace Enclosure Design at Lone Star Steel
Company. .
Hawley Close Capture Hoods
Close Capture Hooding System for Electric Arc Furnaces .
ARMCO Incorporated Design for Tapping Pit Enclosure. . .
Iron Pouring Hood
Mold Cooling Conveyor Tunnel
Moveable Pouring Hood
Typical Shakeout Enclosures
Side Draft Hood
Double Side Draft Hood
Mixer and Muller Hood
Mixer and Muller Ventilation
Page
79
80
81
83
87
88
89
91
92
93
94
98
100
102
104
105
106
109
113
IX
-------
FIGURES (concluded)
Number Page
C-25 Swing Grinder Booth 113
C-26 Torch Cutoff Booth 118
D-l Poppette Valve 139
D-2 Typical Trough Hopper and Screw Conveyor Arrangement . . 141
D-3 Bag-Cell Plate Attachments 142
D-4 Typical Shaker Arrangements 145
-------
APPENDIX A
DESCRIPTION OF FERROUS FOUNDRY PROCESSES
-------
Simply, a ferrous foundry is a manufacturing operation which utilizes
scrap iron and steel to produce cast iron and steel products. However, as
explained in Section 3, the production of these castings can be accomplished
in a variety of ways. This appendix presents a more detailed description
of foundry processes than that described in Section 3. In addition, the
types of emissions that can be expected from each of the processes are de-
fined.
The appendix is divided into six sections. The first section describes
the basic foundry process and explains the differences in the processes used
to produce gray, ductile, and malleable iron and steel castings. The remain-
ing five sections describe the processes and emissions associated with the
following areas of foundry operations:
1. Core and mold preparation;
2. Furnace charge preparation;
3. Melting and casting;
4. Cleaning and finishing; and
5. Sandhandling system.
While all of these areas may not be present in every foundry and certainly
processes within each of these areas vary greatly between foundries, these
areas are typical of most ferrous foundries. The one other area that can
be found in many foundries is the pattern shop. However, all estimates of
uncontrolled emissions from pattern making were so low, that'it was not in-
cluded as a part of the study.
A.I. GENERAL PROCESS DESCRIPTION
The typical ferrous foundry (gray iron, ductile iron, malleable iron,
and steel) processes various grades of iron and steel scrap to form cast
products. The basic operations present in almost all foundries are: raw
materials handling and storage, core and mold preparation, melting, pouring
of metal into molds, and removal of castings from the molds. Other opera-
tions present in many but not all foundries include: (a) mold cooling; (b)
shakeout; (c) casting cleaning, and finishing; (d) sand handling and prepara-
tion; and (e) hot metal inoculation.
A general flow diagram for a ferrous foundry is presented in Figure
A-l. Block diagrams for raw materials handling, core and mold preparation
melting and casting, and sand handling are shown in Figures A-2 through A-5.
It should be noted that while almost all foundries will have operations fall-
ing in the basic areas of operations shown in Figure A-l, specific processes
vary from plant to plant and not all foundries will have all the operations
shown in the diagrams. These differences in processes are described in the
later sections of the appendix. The paragraphs below describe the general
process used in most foundries.
-------
__TJ:
L- Patterns
Core and
Mold Preparation
Figure A-l. General foundry flow diagram.
-------
From Foundry
From Machine—
Shop or
Scrap Dealer
Charging
Mechanism
Charging
Mechanism
i 1 A
Carbonates
Fluorides
1
Carbides
Figure A-2. Process flow diagram - raw material storage and furnace charge make-up.-*-
-------
Sand and
Binders or
Prepared
Sand
Return Sand
New Sand
and Binders
Green Sand
Molder
Shell or Hot
Box Mold
Machine
Dry Sand
Mixer
Cold Set Box
Machine
Core & Mold
Assembly
to Pouring
Area
Dry Sand
Mixer
Air &
Catalyst
Shell or
Hot Box
Core
No Bake
Core Box
Sand and
Binders or
Prepared
Sand
Sand and
Binders
Chemical
Catalyst
Figure A-3. Core and Mold Preparation.
5
-------
Ladle Me tallies
Additions Charge
*
Electric
Induction
j
|
Holding
Furnace
1
Flux
Charge
i
Furnace
, ,
Fuel
Charge
r
1 i -
Electric Cupola
Arc Furnace Furnace
1 i
1
Duplexing
Furnace
\
to. 1 nr
*
r
Jle
Inoculant
Figure A-4. Melting and Casting.-*
to Shakeout
-------
— ^^
Screen
1
Solid Reclninnlioii
Syslem
Kow Mnleiiul
Storage
Figure A-5. Sand Handling.
-------
As can be seen from Figure A-l, raw materials enter the foundry in one
of two areas, the furnace charge preparation area or the core and mold making
area. At the furnace charge preparation area the primary raw materials are:
iron scrap, borings and turnings, limited quantities of pig iron, and foundry
returns used for metallic content; coke for cupolas; and fluxing material
such as limestone, dolomite, fluorspar, and calcium carbonate. The primary
materials received at the core and mold area are: cleaned and dried sand;
cereal filler material; organic binders; and precoated sands for some types
of core or mold processes.
After arriving at the foundry, metallics are prepared for charging to
the furnace. The amount of treatment required is dependent upon the fur-
nace. The cupola can receive almost any type of scrap. However, the scrap
must be relatively free of oil and completely dry before it can be charged
to an induction furnace.
After preparation is completed, the metallics are charged to the fur-
nace. The major furnaces used in foundries are cupolas, electric arc fur-
naces, and electric induction furnaces. These three furnaces account for
75, 17, and 7 percent of iron foundry production, respectively. Reverbera-
tory or air furnaces account for most of the remaining production, but their
use is generally decreasing across the industry.
After the meltdown is complete, the iron is tapped into a ladle. In
some operations, particularly the production of ductile iron, inoculants
are then added to the ladle. Upon completion of the ladle addition, the
ladle is transported, generally by overhead rail, to the pouring area for
casting of the iron into molds.
Upon reaching the casting areas, the hot metal is poured into a mold
to produce an iron casting. The four types of molding processes which have
received most attention are green sand molds, shell sand molds, cold set
molds, and permanent molds or centrifugal casting. Of these, green sand
molding is by far the most prevalent. Details of these processes are dis-
cussed in the second section. If a sand mold is used, the mold and casting
are cooled and then transferred to a shakeout area where the casting is re-
moved from the sand.
After the casting is removed from the mold, the casting goes to the
cleaning room. Here the gates and risers are removed by chipping or grind-
ing. The casting is then cleaned by shotblasting or sandblasting. Further
grinding may then be necessary to finish the casting. The waste sand from
the shakeout is processed and reused for molding. About 2% of the sand is
continuously replaced to maintain sand quality.
Cast iron and steel are a family of materials that differ widely in
their properties. They are basically alloys of carbon and steel that also
include such elements as silicon, manganese, sulfur, and phosphorus. Table
A-l shows the basic ranges of composition for the various types of ferrous
castings.
-------
TABLE A-l. CHEMICAL COMPOSITION OF FERROUS CASTINGS4
Malleable iron
Gray iron (as white iron) Ductile iron Steel
Element
Carbon
Silicon
Manganese
Sulfur
Phosphorus
2.5-4.0
1.0-3.0
0.40-1.0 }
0.05-0.25
0.05-1.0
1.8-3.6
0.5-1.9
0.25-0.80
0.06-0.20
0.06-0.18
3.0-4.0
1.4-2.0
0.5-0.8
< 0.12
< 0.15
< 2.0b
0.2-0.8
0.5-1.0
< 0.06
< 0.05
Necessary chemistry also includes 0.01 to 1.0% Mg.
Steels are further classified by carbon content as follows: low carbon
< 0.20%, medium carbon - 0.20-0.50%, high carbon - > 0.50%.
Figure A-l depicts most of the basic differences in the process flow
for the production of each type of ferrous metal. The general flow in Figure
A-l was for a gray iron foundry. The major differences for the other
metals are the requirement of an inoculation step for ductile iron produc-
tion and heat treatment of all malleable iron and many steel castings. These
differences as well as differences in process types for the different metals
are described in greater detail in the sections below.
A.2. MOLD AND CORE PREPARATION
One of the preliminary steps in the production of ferrous castings is
the production of molds and cores. The mold gives the casting its basic
exterior shape while the cores are used to form indentations or the internal
shape of the casting, e.g., the cylinders in an engine block. The two sec-
tions below describe the various processes which can be used to produce molds
and cores respectively. In cases where the same process is used to produce
both molds and cores, a process description is included only in the section
which utilizes the process most extensively.
A.2.1. Molding Processes /
The foundry operator has many molding techniques from which to choose.
These include: green sand molding, dry sand molding, pit mold molding,
various types of chemically bonded sand molding, permanent mold casting,
die casting, investment casting, centrifugal ca-stings, plaster molding, ce-
ramic molding, and others. These processes are described in References 5
and 6. This section will include a discussion of green sand molds, dry sand
molds, and pit molds as these methods account for the vast majority of cast-
ings and have the greatest emissions potential. Centrifugal casting and per-
manent molds will also be discussed briefly. Since chemically bonded sand
-------
molding processes and emissions are similar to core making processes they
are described in Section A.2.2.
By far the greatest tonnage of castings is poured in green sand molds.
Green sand molds are made with a moist sand, and the moisture is retained
in the sand through the time when the metal is poured. The four steps in
green sand molding are: (a) preparation of the pattern; (b) preparation of
the sand (mulling); (c) making the mold; and (d) core setting. The two steps
in the process which have the potential for emissions are mulling and molding.
Water, sand and binding materials such as bentonite clay, sea coal, and cereal
additives are mixed in the muller. Since the materials are quickly wetted
during mulling the greatest potential for emissions occurs during the charg-
ing of materials, particularly binders, to the muller. However, these are
well controlled in most mechanized operations. Mechanized molding machines
of various types are then used to form the mold in two halves, the cope or
upper half and the drag or lower half. After both halves are formed, cores
are placed in the mold and the cope and drag are fastened together. Since
the sand is moist, the emissions from molding are generally quite low.
Dry sand molds are used most often for thick walled steel castings.
The sand mixes used for dry molding include such additives as pitch, sodium
silicate, gilsonite, cereal, molasses, dextrine, gluten, and resins.7 The
additives are mixed in a muller. The oils in the binder coat the sand grains
and leave the mixture in a green sand condition. This sand is then formed
into molds in machines like those used in green sand molding. The formed
molds are then baked in an oven at 300°F to 600°F which polymerizes the bind-
ers to form a hard rigid mold. The muller and, to a lesser degree, the mold
machines are sources of fugitive particulate emissions. The baking oven
may be a source of gaseous hydrocarbon emissions.
Pit molds, which are used to produce castings too large for a flask,
may be made in a pit by a bedding-in method. The pattern is set in a pit
in the position in which the casting is to be poured, and sand is rammed or
tucked under and around the sides of the pattern. The cope for the complete
mold may rest on the drag at or above floor level, and may be bolted down
to prevent run-out at the parting plane. Many foundries have a concrete-
lined pit equivalent to the size of the mold they customarily produce. The
mold may be rammed up, striking off the surface to produce the desired shape.
At times, when the design of the casting is such that a pattern cannot be
drawn out of the mold the entire mold cavity may be constructed with cores.8
These large pit molds are always dried.9
Several molding methods which do not utilize sand are used in ferrous
foundries. The two most predominate are permanent molds and centrifugal
casting. Permanent mold casting employs reusable molds fabricated from iron,
steel, or graphite. These molds, or dies, are clamped together in a machine,
usually by means of a hydraulic cylinder. The mold is prepared by coating
it with an insulating material such as acetylene soot. This process is called
blacking. Cores are then set into place, if required, and the mold closed.
Molten metal is introduced by gravity feeding. Solidification time can be
calculated and the mold preprogrammed to open shortly after the casting has
solidified to allow ejection of the red-hot casting. The only sand and bind-
ers employed in the process are those for castings that require cores. The
10
-------
dust, free silica, and binder decomposition products are eliminated in the
case of an uncored casting, or substantially reduced in quantity. The black-
ing operation may introduce carbon soot into the atmosphere as well as un-
burned acetylene. The duration of this blacking is generally only a few
seconds.10
Cast iron pipe is generally produced with permanent molds using the
centrifugal casting process. The molds are rotated about the longitudinal
horizontal axis during the pouring and cooling cycle. A dry sand core is
used to produce the bell end of the pipe. Cast iron pipe tonnage represents
about 20% of total cast iron production.6
A.2.2. Coremaking Processes
Cores are prepared by mixing clean sand with one of several types of
organic binders followed by a chemical or thermal setting process to form a
hard, rigid core. Coremaking processes can be identified as one of five
types: oven bake, shell, hot box, cold box or gassed core, and no-bake.
The level of usage of each of these types of corema.king and organic binders
associated with each are shown in Table A-2. Emissions from this area are
primarily organic vapors from the binders. The paragraphs below briefly
describe each of the processes and identify the organic emissions that can
be expected from each.
Oven-baked cores are formed in much the same manner as green sand and
dry sand molds. After the oven-bake core is molded, it is placed on a flat
core plate or formed core dryer and transferred to a gas- or oil-fired oven.
In the oven the light oil fractions and moisture are driven off, and the
core oil is polymerized. The primary emissions from core ovens are organic
acids, aldehydes, and photochemically active hydrocarbons.11 All ovens are
vented to the atmosphere, and some use afterburners or chemical scrubbers
to minimize organic emissions.
Shell coremaking or shell-molding is a process whereby cores or molds
having a thickness of 1/8 to 3/8 in. are produced. These are used for the
most part in applications requiring a great amount of precision. Sand and
approximately 5% thermosetting resin (usually having a phenol-formaldehyde
base) may be dry-mixed in a muller.12'13 The sands may also be prepared by
cold, warm, or hot coating. This mix is then blown into a metal box hous-
ing the pattern plate, which has been heated to a temperature of 350 to
700°F.1X The binder within 1/8 to 3/8 in. of the pattern is melted and the
material turned into a dough-like substance. Excess sand is dumped off,
and the shell is then hardened. The primary emissions from the process are
CO, formaldehydes, amines, ammonia, and phenols.
Hot-box binders are those resins that rapidly polymerize in the pres-
ence of acidic chemicals and heat to form a mold or core. The original hot-
box resins were developed by modifying urea-formaldehyde resins with the
addition of 20 to 45% of furfuryl alcohol. This type of hot-box resin is
commonly referred to as furan resin. The furan resins were then modified
with the addition of phenol to produce urea-phenol-formaldehyde hot-box
resins, which are referred to as phenolic resins or UPF resins. The UPF
11
-------
TABLE A-2. PRINCIPAL ORGANIC CORE BINDERS IN USE IN
THE UNITED STATES13
Binders
Approximate annual
current consumption
Organic
Oven bake
90 x 106
Heated core box
Shell
Hot box
Gassed core
No-bake
85 x 106
45 x 106
3 x 106
20 x 106
1. Oleoresinous
2. Urea-formaldehyde resins
3. Phenol-formaldehyde resins
4. Cereal binders
1. Phenol-formaldehyde novolaks
2. Furan resins (UFFA)
3. Phenol resins (UPF)
4. Phenol-modified resins
1. Cold box (isocyanate)
1. Air set (oil-oxygen)
2. Furan no-bake
3. Oil no-bake
4. Urethane (phenolic-isocyanate)
12
-------
resins have a pungent odor, and adequate ventilation at the coremaking machines
is required. More recently, urea-free phenol-formaldehyde-furfuryl alcohol
binders have been developed. These have a much lower volatile content and
odor compared with other hot-box resins as a consequence of eliminating urea
from the formulation.13
A two-part polyurethane cold-box binder system was developed about
1967 that required gassing rather than baking or heating to achieve a cure.
Part I of the system is a phenolic resin, and Part II is a polyisocyanate.
Both are dissolved in solvents. In the presence of a catalyst, triethyl-
amine (TEA) or dimethyl ethylamine (DMEA), the hydroxy groups of the liquid
phenolic resin combine with the isocyanate groups of the liquid polyisocya-
nate to form a solid urethane resin which serves as the sand binder. Fol-
lowing introduction of the catalyst into the cold box, air is used to sweep
any remaining vapors through the core, after which the core is removed from
the core box. The amine catalysts are volatile, flammable, organic liquids;
and excessive vapors present safety hazards.13
The so-called no-bake binders represent modifications of the oleoresin-
ous, urea-formaldehyde, phenol-formaldehyde, and polyurethane binder systems
previously described, in which various chemicals are incorporated to produce
polymerization in an unheated core box.12
Decomposition products of the various binders are presented in Tables
A-3 and A-4. It should be noted that these values were obtained by direct
venting of prepared cores and are not representative of in-plant ambient
levels.
A.3. FURNACE CHARGE PREPARATION
Materials required by the foundry melt department are metallics, flux-
ing material, and coke. In addition, refractory materials are generally
needed for furnace linings. The composition of the charge and required
charge preparation depend upon the type of furnace being used.
In addition to the metal charge cupolas require coke for fuel and flux-
ing agents to maintain the coke ash and metallic oxides in fluid form in
the slag. The only preparation associated with cupolas is a prescreening
of the charge, particularly coke and metallics, to limit the quantity of
fines charged to the furnace. This is not a general practice, but, as in-
dicated in Appendix B, screening may be used to reduce particulate emissions
from the cupola. Fugitive particulate emissions may be associated with the
screening process.
Since the heat in electric arc and electric induction furnaces is sup-
plied by electrical energy, only a small amount of coke used to control metal
quality is charged to these furnaces. Thus, the only material requiring
preparation is the metallic charge.
13
-------
TABLE A-3. PRODUCTS OF THERMA.L DECOMPOSITION OF SAND BINDERS
12
H1
-P-
Product
Carbon monoxide
Hydrogen cyanide
Methane
Ethylene
Acetylene
Carbon dioxide
Ammonia
Aldehydes (as
formaldehyde)
Phenol
Threshold limit
value (ppm)
50
10
-
-
-
5,000
25
2
5
Concentration in
Polyurethane
40,000
16
2,000
1,500
1,500
7,000
> 1,500
200
d
17.5 mg
effluent (ppm by volume)
Oil base
40,000
400
40,000
7,000
1,500
11,000
500
> 400
d
0.6 mg
Urea-
formaldehyde
40,000
320
2,000
1,500
1,500
7,000
1,500
400
d
1.5 mg
Phenolic
40,000
60
2,000
1,500
1,500
1,000
-
> 400
d
0.4 mg
All products except phenol were determined in the gas phase. The approximate volumes of the gas phases
collected from each binder material were as follows: polyurethane, 200 ml; oil base, 300 ml; urea-
formaldehyde, 1,000 ml; phenolic, 200 ml.
Threshold limit values (TLV) established by the American Conference of Governmental and Industrial
Hygienists.
Phenol was determined in the condensed liquid phase.
The values given are the total weights of phenol found in the condensed liquid phase.
-------
TABLE A-4. FUNCTIONAL GROUPS OBSERVED IN INFRARED ABSORPTION
SPECTRA OF CONDENSED LIQUID PHASES3>12
Binder material
Urea-
Functional group Polyurethane Oil base formaldehyde Phenolic
Aliphatic CH
Aromatic CH
Ester C=0
COOH
Aldehyde C=0
Amide
Secondary amide
Acidic OH
Phenyl
Substituted phenyl
The total weights of the condensed liquid phase collected from each
binder materail were as follows: polyurethane, 120 mg; oil base,
500 mg; urea-formaldehyde, 200 mg; phenolic 80 mg.
15
-------
Preparation may include:
1. Cutting to size by flame torch, breaking, or fragmentation;
2. Cleaning by degreasing or shotblasting;
3. Burning off coatings or oils; and
4. Drying or preheating.
If an electric arc furnace or cupola is used for melting, no prepara-
tion of the metallics (except possibly sizing) is needed. But, since the
presence of water or oil in the scrap cause an explosion hazard in induction
furnaces, scrap is frequently preheated before being charge to these furnaces.
Preheating is generally done in the charging bucket. Preheating mechanisms
include top or bottom fired radiant or flame heat, hot gases forced through
the charge bucket, the use of a double walled bucket with combustion between
the walls, rotary dryers, heated conveyors, or preheater furnaces.
Each of, the preparation processes may generate significant particulate
emissions at a particular foundry. In addition, preheati-ng of oily scrap
may result in gaseous organic emissions. However, since each of these pro-
cesses has limited application, they are not considered to be a serious par-
ticulate problem on a nationwide basis.
After the metallics are prepared, they are transferred to the furnace
for charging. The handling mechanisms for metallics, coke, and fluxes vary
considerably. Methods range from highly mechanized conveyor systems to manual
movement in wheelbarrows. Some transfer operations may release small amounts
of particulate, but again the source is not major.
A.4. MELTING AND CASTING
Operations which may be found in the melting and casting department of
a ferrous foundry include melting, superheating or duplexing, inoculation,
and pouring and cooling of the ferrous castings. The types of operation
and specific equipment utilized in these operations vary from foundry to
foundry dependent upon factors such as foundry size, type of metal cast,
type and size of casting, number of castings produced, energy availability,
and local environmental regulations. The sections below describe each of
these four classes of operation and the specific equipment that can be used
for each operation.
A.4.1. Melting
Ferrous foundries generally melt in one or more of four types of fur-
naces, cupola, electric arc furnace, electric induction furnace, and reverb-
atory (or air) furnace. These four types of furnaces are shown in Figures
A-6 through A-10.
16
-------
Skip-Hoisf Rail
(Iof2)
Brick Lining
Cast Iron Lining
Charging Door
Stack
Wind Box
Skip -Hoist Rail
Brick Lining
Cast Iron Lining
Charging Door
J- Refractory Lining Wafer
Charging
Deck
Steel Outer Shell
Blast Duct Stee' lnner Sne"
Water Inlet
Iron Trough
Taphole for Iron
(Slag Hole is 180°
Opposite)
Sand Bed
Door ( 1 of 2 )
Prop
Stack
Charging
Deck
Water Flow Between
Inner and Outer Shell
Skip-Hoist Rail
(Iof2)
Brick Lining
Cast Iron Lining
Charging Door
Carbon
Block
Slag and
v Iron Trough
\ Sand Bed
Door (1 of 2 )
Solid Steel
Shell
Water
Curtain
Water
Trough
Stack
Charging
Deck
Prop
Blast Duct
Wind Box
Water-Cooled
Tuyere
Carbon
Block
Slag
Slag and Dam
Iron Trough
Sand Bed
Door (1 of 2)
Conventional Cupola
Water-Cooled Cupola (Water-Wall)
Water-Cooled Cupola (Flood Cooled)
Figure A-6. Illustration of a Foundry Cupola. ^
-------
Carbon Electrodes
Spout
Slag
Furnace Tilted to Pour Rammed
Hearth
Metal
Door
Ladle
Figure A-7. Illustration of an Electric Arc Furnace.
16
18
-------
A. HYDRAULIC TILT CYLINDERS
B. SHUNTS
C. STANCHION
D. COVER
E. COIL
F. LEADS
G. WORKING REFRACTORY
H. OPERATORS PLATFORM
I. STEEL SHELL
J. TIE RODS
K. CLAMPING BOLTS
L. COIL SUPPORT
M. SPOUT
N. REFRACTORY BRICK
O. ACCESS PORT
P. LID HOIST MECHANISM
Figure A-8. Illustration of a Coreless Induction Furnace.
17
19
-------
LINING
{INSULATING
BACK-UP
HOT FACE
THROAT
CHANNEL
INDUCTOR ASSEMBLY
BUSHING
CORE
LADLE RETURN SPOUT
POUR SPOUTS
'COIL
Figure A-9. Illustration of a Channel Induction Furnace.
18
20
-------
Stack
Hearth
«y^-I-n-g.-...-rt--,s-rt——-n-~rr
—Af-.-jL u u.._u../::
Floor Level
Combination
Gas-Oil Burner
Figure A-10. Illustration of a Reverberatory Furnace.
19
21
-------
Hot metal production in ferrous foundries is divided among the foundries
approximately as follows:
Cupola - 75%
Electric Arc - 17%
Electric Induction - 7%
Reverberatory - <1%
The following sections briefly describe each of the furnaces and their as-
sociated emissions problems.
A.4.1.1 Cupola Furnaces--
The cupola furnace is the principal melting unit found in gray iron
and ductile iron foundries. The cupola is an upright cylindrically shaped
vessel which uses the heat from the charged coke to melt iron. The cupola
operation is continuous, with metallics, coke, and fluxing agents being
charged in layers near the top of the furnace and the molten iron tapped
from the bottom. The cupola bottom consists of two hinged doors which are
blocked closed during the operation but can be opened after melting is com-
pleted to dump the remaining charge. Before melting is started, the doors
are closed and the floor packed with 8 to 10 in. of sand to seal the cupola.
Combustion air for the melt is injected into tuyeres just above the
level of the sand. The taphole is also located at this level. For con-
tinuously operating cupolas, the slag and iron are tapped together and the
slag skimmed off in the runner or in a forehearth. For intermittent oper-
ations, the slag hole is located at the top of the level of the iron.
The charging door is located 15 to 25 ft above the bottom of the cupola.
The stack is extended above the charging door to sufficient height to clear
the roof of the foundry. Many times a burner or series of burners is in-
stalled immediately above or below the charging door to combust CO in the
stack gas.
Two factors of cupola design which are important from an air pollution
perspective are the type of lining used and blast air temperature. Cupolas
can generally be classified as one of three types with respect to cupola
lining: acid lined, basic lined, or unlined (also termed water cooled).
The conventional cupola has a refractory lining inside the shell which may
comprise either an acid or a basic material. Acid linings generally are /-
composed of silica brick. Basic linings are composed of dolomite or magne-
site brick. Many of the newer cupolas are unlined, using instead water
cooling on the exterior of the shell to prevent heat damage. This has the
advantage of decreasing downtime necessary for relining the cupola. The
effect of lining on cupola emissions is discussed in Appendix B.
The other major design factor which influences emissions is the type
of blast air used. Cupolas are generally classified as either cold blast
or hot blast. A cold blast cupola blows air at ambient temperature through
22
-------
the tuyeres. Most newer cupolas are of the hot blast variety. These cupo-
las generally use a regenerative system that utilizes heat from the cupola
exhaust gases to preheat the combustion air. The effect of blast tempera-
ture on cupola emissions also is discussed in Appendix B.
One industry trend which has improved cupola performance is the use of
the divided blast cupola. By the use of two rows of tuyeres that have sep-
arate air systems with blast equally divided between them, the operating
performance of a lined cupola has been substantially improved. When operat-
ing a cupola at a constant blast rate, the tapping temperature of the metal
is significantly higher than that obtained with one row of tuyeres at a simi-
lar charge coke consumption. Also, the charge coke consumption is reduced
and the melting rate increased for the same tapping temperature.20
One other type of cupola which has been used on a limited basis in
England is the "cokeless" cupola. The cokeless cupola is one which has
been converted to use oil or gas burners rather than a coke bed to provide
the heat for melting. The utilization of oil or gas appears to drastically
reduce emissions from the cupola. There are no known domestic cokeless cupola
installations.
The cupola is the largest single emissions source in the foundry in-
dustry. An uncontrolled cupola emits large quantities of both particulate
and CO. Small amounts of both SO and hydrocarbons may also be emitted from
the cupola. Emissions from the cupola are discharged through a stack at or
above the roof level of the foundry.
A.4.1.2 Electric Arc Furnaces--
The electric arc furnace is found in both iron and steel foundries and
is the principal melting unit in steel foundries. The electric arc furnace
is a refractory-lined, cup-shaped vessel with a refractory-lined roof. As
with the cupola the lining may be either acidic or basic. Three graphite
electrodes are placed through holes in the roof to provide the electrical
energy for melting iron.
Unlike the cupola, the electric arc furnace is a batch type operation.
The basic melt cycle consists of charging, melting or refining, and tapping.
Two additional steps that are often found are backcharging and, in steel
foundries, oxygen lancing. The paragraphs below describe these five opera-
tions and the emissions associated with each operation.
An electric arc furnace can be charged through a side door; or the roof
can be removed and the furnace charged through the top. The top charging
method predominates in ferrous foundries. The charge is introduced to the
furnace through the use of a charge bucket or, in smaller, less mechanized
foundries, by hand. Top charging produces emissions which are not controlled
in most of the existing plants. Emissions result from: (a) vaporization
and partial combustion of the oil introduced with any scrap, borings, turn-
ings, and chips which are contained in the charge; (b) oxidation of organic
matter which may adhere to the scrap; and (c) liberation of sand particles
which are introduced into the furnace on the surface of casting returns.
High oil content is characteristic of the least expensive scrap, e.g., swarf
23
-------
(turnings, chips, and borings) from machine operations. Charging emissions
are essentially made up of particulate matter, carbon monoxide, hydrocarbon
vapors, and soot.
When the furnace is ready for the melting cycle, the electrodes are
lowered through the roof to a position first above the metallic charge and
then energized. Melting is accomplished from the heat supplied by radia-
tion from the arc formed between the electrodes and the metallic charge,
radiation from the furnace lining and resistance of the metal between the
arc paths. During melting operations (meltdown, slagging, and refining),
emissions consist of: (a) particulates as metallic and mineral oxides gen-
erated from vaporization of iron and transformation of mineral additives;
(b) carbon monoxide from combustion losses of the graphite electrodes, car-
bon raisers and carbon in the metal; and (c) hydrocarbons from vaporization
and partial combustion of oil remaining in the charge. During melting,
emissions escape from the furnace through electrode annuli (holes), the slag
doors, the roof ring (the joint between the furnace shell and roof), and
sometimes the tap spout.21
Steel furnaces are sometimes backcharged, i.e., a second charge is added
to the furnace, as soon as sufficient volume is available in the furnace.
(Iron furnaces are generally charged only one time.) Backcharging produces
a violent eruption of emissions with a strong thermal driving force. The
amount of pollutants generated during this phase of the operation is probably
higher than during either the first charge or during treatment of the molten
bath in the transfer ladle.22
Oxygen lancing in steel furnaces is used mainly for adjustment of the
chemistry of the steel, for speeding up of the melting process, and for su-
perheating of the bath. Oxygen lancing results in increased temperature,
gas evolution, and generation of particulates (particularly iron oxide) and
carbon monoxide. Oxygen lancing can be carried out with moderate rates of
oxygen addition, thereby avoiding excessive generation of high temperatures,
gas evolution, and particulate emissions. However, extended periods of oxy-
gen lancing can increase energy consumption, refractory wear, oxidation of
the bath, and change the production rate.
When the melting and refining cycle is completed, the electrodes are
raised and the roof is removed. The furnace is then tilted by as much as
45°, and the refined metal is tapped into a ladle. Emissions during tap-
ping consist primarily of fine metallic fume. Because of higher metal tem-
peratures, tapping emissions from steel furnaces tend to be greater than
those from iron furnaces.
A.4.1.3 Electric Induction and Reverberatory Furnaces--
The two types of electric induction furnaces used in foundries are the
channel induction and coreless induction furnaces. The coreless induction
furnace is most frequently used for iron melting. The coreless induction
furnace is a cup-shaped vessel which uses electrical energy to induce eddy
currents in the metallic charge to produce molten iron. Since wet or oily
scrap can lead to explosions in the furnace, the scrap is generally cleaned
and is often preheated before charging. Very clean scrap generally leads
24
-------
to much less particulate emissions than for the cupola or the electric arc
furnace and no CO and hydrocarbon emissions. Hence, these furnaces are often
uncontrolled. In that case, the total furnace operation becomes a fugitive
particulate emission problem. As a result of the low pollutant levels, in-
duction furnaces are finding increased use in ferrous foundries.
Reverberatory fuel-fired furnaces are used in foundries for both melt-
ing and duplexing or refining of malleable iron. These furnaces are rectan-
gular or cylindrical structures which are horizontally fired with powdered
coal, oil, or gas. Generally, the furnaces are fired from one end, with
waste gases removed from the opposite end. These furnaces generally have
low-level particulate emissions and are often uncontrolled.
A.4.2 Superheating or Duplexing
Superheating or duplexing is used in malleable iron foundries to raise
the temperature of the white iron in a slag free atmosphere to complete the
refining process. A separate duplexing furnace is essential if a cupola is
used as the primary melting unit. This separate furnace is generally an
electric arc furnace or a reverberatory furnace.
Since most of the impurities in the scrap are released during the melt-
ing cycle, emissions from the duplexing furnace are minimal. Any emissions
that are released from the furnace will be fine metallic fumes.
A.4.3. Inoculation
Iron inoculation is an operation used primarily in the production of
ductile iron. During inoculation a nodularizing agent, most frequently
magnesium, is added to the molten gray iron. This agent causes the flake
carbon found in gray iron to become graphite spheroids. This chemical trans-
formation produces a material which is less brittle than gray iron.
The magnesium (or other nodularizing agent) is generally added to the
molten metal after it has been tapped into the ladle. Several of the methods
used to inoculate ductile iron are shown in Figure A-ll. Modi23 describes
these processes in more detail and discusses their advantages.
The addition of the nodularizing agent to the ladle produces a violent
reaction accompanied by a highly visible emissions stream. The primary con-
stituents of this plume are magnesium oxide and metallic fumes. These emis-
sions are generally fine particulate. The quantity of emissions generated
is dependent upon the magnesium recovery rate (the rate at which magnesium
is retained in the metal).
Matter24 indicates that 75 to 80% of the ductile iron produced in the
United States is inoculated with the pour over and sandwich methods. In
the pour over method the nodulizing alloy is placed in the bottom of the
ladle and the hot metal is poured on top. This method results in 20 to 30%
inoculant recovery. With the sandwich method, the alloy is covered with 1
to 2% steel punching or plate or ferrosilicon. This allows a greater amount
of hot metal to be poured before the reaction starts and results in magnesium
recovery of 40 to 50%.24
25
-------
"SANDWICH
'TRIGGER"
POUR-OVER
THROW-IN
PLUNGING
Figure A-ll. Methods of Iron Inoculation.
26
-------
Industry personnel indicate that newer methods of inoculation results
in magnesium recovery of 50 to 90%.2S One of these methods which shows par-
ticular promise inoculates the metal in the mold rather than in the ladle.
Matter indicates that magnesium recoveries of 80 to 90% have been obtained
with in mold inoculation.
A.4.4. Pouring and Cooling
The final operation in the metal casting area is the pouring of the
molten iron into the mold and subsequent cooling of the casting. The types
of pouring operations found in ferrous foundries vary widely depending upon
the type of mold used and the degree of mechanization in a particular foundry.
This discussion will focus on operations involving sand molds as these are
the type most frequently found in foundries. Pouring of metal into sand
molds also has a greater potential for emissions than pouring into permanent
molds. The paragraphs below describe two major classes of pouring operations,
mechanized pouring lines and floor pouring.
Mechanized pouring lines are generally found in medium to large found-
ries which produce small to medium sized castings. The pouring line has
one or more pouring ladles located along a conveyor. These ladles may be
stationary or may be capable of moving parallel to the conveyor. The models
are placed on a conveyor and moved to the pouring station. After the pour-
ing operation is complete the mold and casting are carried by the conveyor
through a cooling area, often an enclosed "tunnel" made of sheet metal.
"Floor" pouring is found in small to medium sized foundries which gen-
erally do not have sufficient capital to finance mechanization and in larger
foundries which produce castings that are too large to be transported by
conveyor. In these foundries the mold is placed on an open floor or in a
pit and the ladle is transported to the mold, generally by overhead pulley.
When the ladle reaches the mold, the molten iron is poured into the mold
and the casting is then cooled in place.
Emissions problems are comparable for both processes. The emissions
are contained in a relatively high-temperature/high moisture, buoyant stream.
The constituents of the stream are fine metallics from the hot metal and or-
ganics produced by thermal decomposition of the binders. Some CO may also
be emitted during pouring and cooling. The moist, buoyant stream, the or-
ganic emissions, and the disperse nature of the source make control of these
sources difficult.
A. 5. CLEANING AND FINISHING
After the casting has been cooled it must be removed from the mold,
cleaned and finished into a final product. The specific cleaning and finish-
ing operations will vary depending upon the type of metal cast, the type of
mold used to produce the casting, the size of the casting, and the degree
of mechanization. A general flow diagram of the cleaning and finishing area
is presented in Figure A-12. The discussion below will focus on those areas
having the greatest potential for emissions.
27
-------
from Pouring
_L
Shakeout
Remove Gates
& Risers
Figure A-12.
Process Flow Diagram - Cleaning and Finishing.
28
26
-------
After the casting has cooled it must be removed from the mold. If a
sand mold is used, this process is generally called shakeout. Shakeout me-
thods probably vary more from plant to plant than any other operation with
the possible exception of mold and core making. In foundries where large
pit molds are used, the sand is often removed from the mold with front end
loaders and shovels. In small, nonmechanized foundries methods consist of
dumping the molds on the floor, use of pneumatic tools to break sand loose,
and manual removal with shovels. However, the most typical manner of re-
moval is to place the flask on a vibrating screen. The sand is knocked
loose from the casting and falls through the screen and the castings are
carried on the vibrating screen to a conveyor and then move on to other
cleaning steps. One newer method of shakeout substitutes a rotating screen
for the traditional vibrating screen. .In any case the emissions consist of
dust from the dried sand, organic residue from binders, and water vapor.
Some type of exhaust ventilation and particulate control device is usually
found. However, based on foundry visits, capture is sometimes of limited
effectiveness.
The next step is removal of the sprues, gates, and risers if these have
not been knocked off during shakeout. These appendages can be knocked off
manually with hammers, cut off with abrasive, band, or friction cutting, or
removed with an oxygen torch. During knockoff and cutting operations, silica
dust, from burned-in sand, and metallic particles are liberated to the foundry
environment and potentially to the atmosphere. However, these are primarily
larger size particles and should be contained. A major emissions problem
is torch cutoff of large castings. Torch cutoff appears to emit large quan-
tities of iron oxide.
After the sprues, gates and risers are removed, chipping hammers and
various types of grinders are used to remove other irregularities from the
castings. The castings are then subjected to abrasive blast cleaning and/
or tumbling to remove any remaining scale or burned-on sand from the sur-
face. Each of these operations is a source of particulate emissions which
contain silica dust and metallic particles.
The cleaning room is one of the major concerns of the foundry and gov-
ernmental safety and health personnel with respect to industrial hygiene.
As such the development of adequate internal controls for the cleaning room
is a focus for both groups. Future activities in such control development
will impact on air pollution problems related to the cleaning room.
A.6. SAND HANDLING
In foundries which practice sand molding, the sand is reused many times.
In general, the sand is taken from the shakeout, reconditioned, and returned
to the muller for reuse. About 2 to 5% of the sand used is replaced on a
daily basis to maintain sand quality.27
Again the specific sand handling steps will vary depending upon the
degree of mechanization in the foundry. At unmechanized foundries the sand
may be dumped on the floor during shakeout, transferred manually by front
end loader to a screening operation and transferred manually again to a stor-
age pile near the muller from where it is manually charged to the muller.
29
-------
Each of these steps will generate particulate emissions containing silica
dust and binding materials.
A diagram of a modern high volume sand system is presented in Figure
A-13. This system has been modified to use the Schumacher system described
in Section 4. However, it shows the basic steps that are found in a mech-
anized process. If the damp sand were not introduced at the shakeout, each
of the reconditioning steps would generate particulate emissions.
30
-------
r\
Floor
Sand
Hopper
Prepared
Squeezer
Line
Shakeout
Automatic
Line
Shakeout
Floor
Shake-
out
160° F
8..
O.
o
160°F
t !
O
90° f at
Machines
o
1/1
11
6
c
o
t
01
200 Ton Sand
Storage Bin
Squeezer Line
105° F
100 Ton
Capacity
Drum
_. Muller
Plow
f
-^ Sand -^
Dilution Plow
Prepared
111 ( 4 1
III III
^- Shakeout Sand Belt
1
200 Ton Sand
Storage Bin
Automatic
Machine Line
105° F
100 Ton
Capacity
Drum
Muller
-•
1
Bucket
Elevator
Drum Sand
Cooler
( Not Used )
New
150 Ton
Muller
Planned
^
Sand Storage Bin Temperature
Varies with Ambient:
85-90°F AMB. '= 115° F Bin (9-11-691
t
Aerator
Magnetic
Separator
90° F at
Machines
73° F AMB. = 105° F Bin (12-9-69)
Figure A-13.
Line Drawing of Canton Malleable's Sand System Showing Plowoff Points and Resultant
Sand Temperatures
-------
APPENDIX A REFERENCES
1. A. T. Kearney Company. Systems Analysis of Emissions and Emission Control
in the Iron Foundry Industry, Volume II: Exhibits. PB198 349. U.S.
Environmental Protection Agency, February 1971, Exhibit IV-8.
2. Reference 1, Exhibit IV-9.
3. Reference 1, Exhibit IV-14.
4. Heine, R. W. , C. R. Loper, Jr., and P. C. Rosenthal. Principles of
Metal Casting. McGraw Hill, 1967, p. 467, 493.
5. Reference 1, pp. 24-51.
6. Sylvia, J. G. Cast Metals Technology. Addison Wesley. Reading, MA.
1972, pp. 32-61.
7. A. T. Kearney Co. Systems Analysis of Emissions and Emissions Control
in the Foundry Industry, Volume 1, Text. PB198-348. U.S. Environmental
Protection Agency. February 1971, p. IV-27.
8. Reference 6, p. 98.
9. Reference 4, p. 31.
10. Bates, C. E. and W. D. Scott. Better Foundry Hygiene Through Permanent
Mold Casting. Southern Research Institute. January 1976, p. 18.
11. Danielson, J. A. Air Pollution Engineering Manual - Los Angeles County
Air Pollution Control District. National Center for Air Pollution Control,
1967, p. 314.
12. American Foundrymen's Society. Molding, Coremaking, and Patternmaking.
1972.
13. Bates, C. E. and L. D. Scheel. Processing Emissions and Occupational
Health in the Ferrous Foundry Industry. American Industrial Hygiene
Association Journal. August 1974, pp. 452-462.
14. Reference 9, p. IV-28.
s
15. Metals Handbook,,..Volume 5. American Society for Metals. 1970, p. 337.
16. Reference 6, p. 255.
17. Reference 1, Exhibit VI-21.
18. Reference 1, Exhibit JVI-20.
32
-------
19. Reference 6, p. 254.
20. American Foundrymen's Society, Cupola Handbook, 4th Edition Revised,
American Foundrymen's Society, Des Plaines, IL, 1975.
21. Fennelly, P. F. and P. D. Spawn. Air Pollutant Control Techniques for
Electric Arc Furnaces in the Iron and Steel Foundry Industry. U.S. En-
vironmental Protection Agency. Research Triangle Park, NC. Publication
No. EPA-450/2-78-024. June 1978, 221 p.
22. Georgieff, N. T. Addendum to Standards Support and Environmental Impact
for Electric Arc Furnaces in the Gray Iron Foundry Industry. U.S. Environ-
mental Protection Agency. Research Triangle Park, NC. December 1976.
Unpublished.
23. Modi, E. K. Comparing Processes for Making Ductile Iron. Foundry.
July 1970, pp. 42-49.
24. Matter, D. Nodularizing Methods. Quality Ductile Iron - Today and
Tomorrow. Proceedings of a Joint AFS/DIS Conference. Octover 14-16,
1975.
25. Wallace, D. W. and C. C. Cowherd. Fugitive Emissions from Iron Foundries.
EPA-600/7-79-195. U.S. Environmental Protection Agency. Research Triangle
'Park, NC. August 1979, p. 46.
26. Reference 1, Exhibit IV-13.
27. Reference 25, p. 31.
28. Modern Casting, August
33
-------
APPENDIX B
QUANTIFICATION OF PARTICULATE EMISSIONS FOR MAJOR
FOUNDRY EMISSIONS SOURCES
35
-------
As indicated in the description of foundry processes (Section 3), the
typical ferrous foundry has numerous operations which may be sources of both
particulate and gaseous emissions. The scope of this study did not permit
detailed quantification of all operations and pollutants. Since the state
and local agencies contacted were primarily concerned about particulate emis-
sions from foundries, this study was focused on the most significant sources
of particulate emissions.
An initial review of foundry particulate emissions identified six opera-
tions or areas of operation as potentially the most significant sources of
foundry emissions. These operations are: (a) cupolas; (b) electric arc
furnaces; (c) pouring of castings into sand molds and subsequent cooling of
the castings; (d) casting shakeout from sand molds; (e) the total sand han-
dling system; and (f) the cleaning room.
The remainder of the discussion is divided into two sections. The first
is an evaluation of melting emissions data. The second describes the emis-
sions data for nonmelting operations.
B.I MELTING EMISSIONS
As indicated in Section 3, the four types of furnaces currently used
in ferrous foundries are cupolas, electric arc furnaces, electric induction
furnaces, and reverberatory furnaces. However the initial review of the
data indicated that both reverberatory and induction furnaces have relatively
low particulate emissions. AP-42 indicates that emissions from induction
furnaces are only 1.5 Ib of particulate per ton of metal charged for gray
iron foundries1 and 0.1 Ib particulate per ton of metal charged for steel
foundries.2 Data from several furnaces in the Philadelphia area indicate
that emissions from reverberatory furnaces are only about 2 Ib of particu-
late per ton of metal charged.3 Since emissions from these furnaces were
low and production was limited, only emissions data for cupolas and electric
arc furnaces were examined in detail. The results of the examination are
presented below.
B.I.I CUPOLA EMISSIONS
The cupola furnace is the greatest single emissions source in ferrous
foundries. As such, it has been subject to more study than all other foun-
dry emissions sources combined. A literature search identified three major
studies of cupola emissions which examine the relationship between cupola
design and operating parameters and particulate emissions. Each of these
studies is discussed in the following three subsections. These discussions
are followed by a brief description of other available data.
B 1.1.1 U.S. EPA Systems Evaluation of Cupola Emissions4'5
As a part of an overall evaluation of emissions and emissions control
in iron foundries, the A.T. Kearney Co. performed a detailed statistical
analysis of the relationship between cupola emissions and the following de-
sign and operating variables:
36
-------
Cupola Design Variables
Lining: acid, basic, or unlined.
Blast Temperature: cold, warm, or hot.
Blast Heating: external or recuperative.
Charging: side or top charge.
Gas Take-Off: below charge, above charge, top of stack.
Afterburner: with afterburner or without afterburner.
Charge Door: open or closed.
Fuel Injection: with or without fuel injection.
Oxygen Enrichment: with or without oxygen enrichment.
Cupola Operating Variables
Specific Melting Rate
Specific Blast Rate
Metal to Coke Rate
Blast Air Temperature
No actual testing was done as a part of this study, but emissions data
were obtained from operators during plant visits, testing laboratories, equip-
ment manufacturers, state control agencies and published literature. These
emissions data were combined with a large foundry data bank compiled by Kearney
to allow analysis of the emissions.
The data which were used in the analysis are shown in Tables B-l and
B-2. The test methods associated with the data in these tables are not known.
Reference 4 suggests that different methods were used to obtain the data
but that ASME procedures described in PTC 21-1941 and PTC 27-1957 were fol-
lowed by most foundries visited. While the differences in test methods and
insufficient data to completely define the effect all variables on emissions
limited the study, the following results were obtained.
1. The analyses show that cupola emissions rates are not significantly
affected by design factors of the furnaces within the parameters established
by current United States design practices. These factors include the method
of blast heating, top or side charging, charging door size and whether or
not the opening is closed or open, the location of the gas take-off above
or below the door, or an open stack permitting the gases to escape out the
top. In addition, no significant effect on emissions rates was found for
specific melting rates.
2. Eight of the 12 unlined cupolas have emission rates greater than
the median rate of 20.8 Ib/ton while all but one of the acid lined cupolas
is below the median.
3. Acid lined cupolas show a significant correlation between emissions
and blast rate expressed by the equation:
E = 0.05 + 0.07B, where
E = Particulate emissions (Ib/ton melt)
B = Specific Blast Rate (SCFM/ft2 furnace area)
37
-------
TABLE B-l. PARAMETERS OF CUPOLA FURNACES-LINEAR REGRESSION ANALYSIS OF EMISSIONS
AFFECTED BY FURNACE DESIGN FACTORS
U)
00
Foundry
Number
151
12
5
146
12
50
37
26
152
7
45
-69
134
150
9
9
35
125
160
-71
84
29
18
67
69
67
Furnace
Classi f i-
cation
10
27
14
17
32
16
14
14
16
18
10
29
6
24
23
14
4
9
2
11
9
4
30
9
13
9
Lining
Type
4
1
1
1
2
1
1
1
1
1
4
1
4
1
1
1
4
4
4
4
4
4
2
4
1
4
Blast
Design
1
1
3
3
1
3
3
3
3
3
1
1
1
2
2
3
2
1
3
1
1
2
3
1
3
1
Blast
Heating
3
3
1
1
3
3
3
3
2
3
1
3
3
1
3
3
3
1
3
1
3
Top Open
or Closed
1
2
1
1
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
1
2
1
Charging
Top or Side
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
Gas
Takeoff
1
8
1 .
1
8
1
1
1
1
2
1
2
1
2
1
1
1
1
1
2
1
1
8
1
8
1
After-
burner
0
0
2
0
0
-1
2
2
-1
2
0
0
-1
1
2
2
2
2
0
0
2
6
-1
1
0
1
Charging
Door Open
or Closed
1
1
1
2
1
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
1
1
1
1
Fuel
Injection
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Oxygen
Enrich-
ment
0
0
0
1
0
0
0
0
0
0
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
Emissions
Ib./ton Melt
7.5
9.6
11.4
12.1
12.4
15.1
17.4
18.3
19.5
19.9
20.4
20.6
20.8
22.9
36.0
37.6
40.4
40.4
40.5
44.7
45.7
46.6
48.5
50.0
53.4
66.3
Note: See Reference 7 for description of cupola furnace parameter codes.
-------
TABLE B-2. LINEAR REGRESSION ANALYSIS OBSERVATION
Foundry
Number
Cupola
Classifi-
cation
Particulate
Emissions
Ib./ton
Specific
Melt Rate
T/Hr./S.F.
Specific
Blast Rate
SCFM/S.F.
Metal to
Coke Ratio
Temperature
°F
OJ
Acid Lined Cupolas
12
5
37
26
7
150
9
9
27
14
14
14
18
24
23
14
Basic Lined Cupola
18
Unlined
151
45
35
125
160
84
29
67
67
30
Cupolas
10
10
4
9
2
9
4
9
9
9.5
11.4
17.4
18.3
19.9
22.9
36.0
37.0
48.5
7.5
20.4
40.4
40.
40.
45,
46.
66,
50.0
0.56
.73
.64
.63
.71
.78
.57
.57
0.48
0.50
.52
.76
.55
.36
.60
.31
.63
.70
269
364
317
274
194
231
462
462
357
248
238
324
244
317
238
252
352
352
11.
8
6
8
9
10.
10
10
6
9
9
8
10
8
7
6
6
7
1,100
70
70
70
700
750
750
70
1,000
1,400
600
1,000
1,000
1,000
750
1,200
1,400
-------
3. Unlined cupolas have a significant correlation between emissions
and both coke rate and specific blast rate expressed by the equation:
E = 57-6.6C + 0.1B, where
E = Particulate Emissions (Ib/ton melt)
C = Metal to coke ratio
B = Specific blast rate (SCFM/ft2 furnace area)
4. Data are inconclusive with respect to the usage of oxygen enrich-
ment. Data from one plant indicate that the grain loading increased by a
factor of about 2.5 with 4% oxygen enrichment of the blast air. However
other industry sources indicate that even though the emissions concentra-
tions increase,the resulting decrease in melting time results in lower total
emissions with oxygen enrichment.
5. In general, if all other factors are equal, the use of bricquettes
increases emissions.
6. Surprisingly neither the coke rate nor the blast temperature cor-
related highly with emissions for all data (both had an index of less than
0.3).
7. Specific test data are not available to relate emissions to coke,
limestone, and scrap quality. However, visual observation of the plume in-
dicates that the degradation of coke and limestone do effect emissions. In
addition, the collection of Si02 and metallic oxides in the particulate point
to the effect of charge quality on emissions.
8. Limited data were also compiled on the chemical composition and
particle size of cupola emissions. These data are presented in Table B-3
and B-4. The wide variation of both composition and size may result from
differences in test method. However, it is possible that these variations
are an indication of large differences in emissions from cupola to cupola.
B.I.1.2 Canadian Department of Energy, Mines, and Resources Study11'12
In an attempt to provide insight into two problems confronting the
foundryman, the lack of knowledge on the nature and extent of cupola emis-
sions and questions concerning the reliability and practicability of emis-
sions sampling methodology, the Physical Metallurgy Division of the Canadian
Department of Energy, Mines, and Resources conducted a study of cupola emis-
sions. The study consisted of the development of a sampling method for cu-
polas, emissions sampling at five foundries for two weeks each and a sixth
foundry for one week, and detailed analyses of the particulate samples for
particle size and composition. Although the study examined both particulate
and gaseous emissions, this discussion will describe only the portion of
the study related to particulates.
The sampling method is described in detail in Reference 11. The tests
consisted of multipoint samples using two perpendicular probes to simultan-
eously traverse the stack at a height ranging from 2.5 to 6 stack diameters
above the charging door. Depending upon the height of the ports above the
40
-------
TABLE B-3. CHEMICAL COMPOSITION OF CUPOLA PARTICIPATE EMISSIONS9
Percent by Weight in Cupola Effluent
Foundry
Number
66
85
90
113
116
146
150
Iron
Oxide
11-1%
14.7%
-
8.6%
10.0%
33 . 0%
11.6%
Magnesium
Oxide
-
1-3%
-
-
5.0%
-
1.0%
Manganese
Oxide
-
-
-
3.7%
10.0%
1-0%
5.5%
Lead Aluminum Zinc Silicon
Oxide Oxide Oxide Dioxide
12.3%
1.4% - - 28.7%
56.3%
.05% - 31.8%
5.0 % 1.0% 10.0%
5.0% - 38.0% 20.0%
20.0% 1.4 % 14.7% 30.1%
Calcium
Oxide
-
-
42 . 0%
3.1%
3.0%
1.0%
1.1%
Combustibles
-
24.0%
0.9%
27.0%
5-0%
-
-
Note: Quantities as reported. They do not add up to 100%.
-------
TABLE B-4. PARTICLE SIZE DISTRIBUTION-CUPOLA EMISSIONS10
Cumulative Percent by Weight
Diameter in Microns
Foundry
9
14
18
26
32
67
67
146
151
A1
B1
C1
I2
22
32
42
A2
B2
Sources: 1. The
-1 -2
30%
64%
13%
0.6%
0 7%
0 7%
Cupola and Its
-5
50%
82%
2%
28%
54%
14%
2%
4%
11%
8%
18%
17%
24%
26%
25%
24%
-10
65%
98%
12%
45%
86%
15%
3%
5.5%
13%
12%
25%
26%
28%
30%
32%
41%
-20
82%
99%
34%
55%
98%
15%
19%
8%
7%
32%
17%
38%
36%
23%
32%
34%
47%
-50
90%
92%
60%
99%
21%
25%
99%
99%
13.7%
53%
28%
62%
53%
42%
44%
41%
32%
-100
99%
99%
99%
99%
99%
99%
99%
75%
75%
69%
56%
69%
-200
99%
99%
80%
94%
89%
61%
81%
Operation,
Third Edition, 1965,
American Foundrymen's Society,
p. 82.
2. Air Pollution Engineering Manual,
Public Health Service Publication,
No. 999-AP-40, 1967
Department of Health, Education, and Welface.
42
-------
charging door, either 6 to 10 points were used per diameter. All samples
were isokinetic.
The mass emissions data from the particulate tests are shown in Table
B-5. Two observations regarding the data in Table B-5 are of particular
interest. The first is the wide variation in emissions between cupolas.
The average emissions for an average of 4 tests (results which should be
comparable to EPA Method 5) range from 4.2 to 64.0 Ib/ton. An observation
of even greater signifiance is the wide variation in emissions from a par-
ticular cupola. This variation is reflected in the last 3 columns of the
table. Columns 6-8 show the number of tests needed to assure with 95% con-
fidence that the real mean is within ±50%, ±25%, and ±10% respectively of
the sample mean. Note that a minimum of 6 and a maximum of 12 tests are
needed to state (with 95% confidence) that the true mean is within ±25% of
the sample mean and that at least 4 tests are needed to state (with 95% con-
fidence) that the real mean is within ±50% of the sample mean.
As a part of the study, experiments were conducted at Foundry A to de-
termine the effect of charging practices on cupola emissions. The data from
these experiments are presented in Table B-6. The data indicate that at
least a 40% and perhaps as much as a 60% reduction was obtained from the
use of screens and other precautions to limit the amount of loose sand, rust,
and coke fines charged to the furnace.
The particle size was measured by screening the captured particles with
screens as fine as 400-mesh and by optical and electron microscopic analyses
of the minus 400-mesh fraction. The results are shown in Figure B-l. The
chemical composition of the various size fractions is described below.
Comparison of Si02 and total emission rates shows that for all size
ranges at all foundries, Si02 is the primary constituent of the dust. Com-
parison of the combustible and total dust emission rates indicates that a
significant and relatively constant fraction is combustible in the coarse
fractions of the dust but this fraction becomes insignificant in the sub-
micron dust. Similar behavior is exhibited by Fe, except at foundry A where
Fe comprises a significant fraction of the sub-micron dust. Pb and Zn dis-
play trends opposite to that of iron and combustible material; with the ex-
ception of foundries A and B, the lead content of the dust increases with
decreasing dust size, attaining significant levels of emission and concen-
tration in the sub-micron fraction.12
B.I.1.3 U.S. Department of Energy Study of the Effect of Operating
Parameters on Cupola Emissionsia
As a part of a study for the U.S. Department of Energy, the Pennsylvania
State University tested cupolas at seven operating foundries for particle
size distribution. An analysis of variance (ANOVA) routine was then used
to analyze the effects of blast rate, iron coke ratio, blast temperature,
and cupola size on particulate emissions.
The particulate sampling train is a modification of the one used by
Warda and Buhr11 and is described in detail in Reference 13.
43
-------
TABLE B-5. SUMMARY OF PARTICULATE EMISSION
11
No.
of
Days
1
2
4
8
1
2
4
12
22
2
4
12
22
2
6
12
21
2
6
12
26
2
4
8
14
2
4
8
X,
Ib/ton
75.5
64.0
47.4
44.9
10.0
10.75
14.31
15.48
11.1
19.4
19.8
17.9
11.75
10.2
9.7
8.2
53.0
43.0
37.8
34.1
3.8
4.2
4.8
No. of Tests Requireda for
95% Confidence Limits of
±50% ±25% ±10%
38
26
12
0.19
0.64
1.16
66
42
32
26
a No. of tests required was calculated using the confidence interval relationship.
S
p = X ± tQ -, ^/n where
p = true mean
X = sample mean
t. -„ = t - statistic for a 95% confidence interval
s = sample standard deviation
n = number of tests
44
-------
TABLE B-6. INFLUENCE OF CHARGING PRACTICE ON DUST PRODUCTION AT FOUNDRY A11
Dust Production Rate
x,
Period (Ib/ton)
Days 1, 2, and 3 51.8
Days 4 and 5, Runs 1 and 2 43.6
Day 5, Run 3 25.0
Day 6 43.5
Day 8 18.5
s>
(Ib/ton)
17.6
5.5
5.0
3.5
1.5
Remarks
Standard charging practice, scrap small and
rusty.
As during days 1, 2, and 3 except that scrap
consisted of large plates with less rust.
As during days 4 and 5 except that loose sand
was removed from foundry returns by a screen.
As during days 4 and 5 (Runs 1 and 2) except
that plate scrap was completely free of rust.
As during day 5 (Run 3) except that extra pre-
cautions were taken to prevent charging of
loose sand, rust, and coke fines.
-------
50
£
"o
40
0)
O
OH
-------
Particle size was measured with an in-stack impactor with a 10 pro presepara-
tor. For both total particulate and particle size the sampling rate was main-
tained within ±10% of isokinetic. Reference 13 does not indicate whether
single point or multiple point sampling was used.
The emissions data from the seven tests are not presented in Reference
13. However, the following conclusions are drawn based on the ANOVA routine.
1. In seven of eight cases the blast rate was shown to have signifi-
cant relationship to emissions.
2. In no case was any statistical significance shown between the iron-
coke ratio and emissions.
3. The generation of particles in the ^ 2 [Jm range appears to be strongly
related to blast temperature and melt rate.
4. Generation of particles in the respirable region appears highest
at 315°C (600°F) with higher or lower temperatures giving reduced particle
loadings.
5. Melting metal at a higher rate produces a higher loading of fine
particles.
Data are presented on the overall particle size at each of the found-
ries. These are shown in Figure B-2.
B.I.1.4 Analysis of Cupola Emissions in the Federal Republic of Germany
14
In order to develop an improved data base to aid cupola operators in
the selection of appropriate control equipment, dust emission measurements
were carried out on 35 different types of cupola installations. Data were
compiled on cupola design characteristics, operating parameters, exhaust
gas characteristics and dust emissions. These data are presented in Table
B-7.
As indicated in Table B-7, dust emissions samples were obtained at dif-
ferent points in cupola exhaust systems. Samples were collected at veloci-
ties slightly higher than the average gas velocity at multiple points in
the exhaust stream. (Reference 14 indicates that the experimental error
from using this nonisokimetic sampling method is small.) Each test lasted
at least 3.5 hr, and no measurements were made during start-up, interuptions,
blower shutdown or blowdown of the charge. Particle size was also,measured
for the collected particulate using sieving and, for particles <,.&3 |Jm, by
Bahco Classification. Additional information on the sampling method is pre-
sented in Reference 14.
47
-------
100
50
10
0)
o
u
il
Q.
Q
0.5
0.1
0.01
I I
I T
A Foundry A
A Foundry B
O Foundry C
• Foundry D
o Foundry E
• Foundry F
v Foundry G
I I
I I
10 20 50 80 90
% Particles < Dp
99
99.99
Figure B-2. Average Particle Size Distribution for 7 U.S. Foundries.
13
48
-------
TABLE B-7. RESULTS OF CUPOLA TESTING IN THE FEDERAL REPUBLIC OF GERMANY
Test
No.
1/64
2/64
3/64
4/64
5/64
6/64
7/64
8/64
9/64
10/64
11/64
12/64
13/64
14/64
15/64
1/65
2/65
3/65
4/65
5/65
6/65
7/65
8/65
9/65
10/65
11/65
12/65
13/65
14/65
15/65
1/66
2/66
3/66
4/66
5/66
Blast
Type3
II
C
11
II
II
II
II
C
C
C
Bll
II
II
II
C
II
C
C
H
II
Bll
H
H
H
II
C
C
II
11
C
C
c
C
C
H
Cupola
Type
D
A
A
D
C
C
D
A
A
A
D
D
D
D
A
C
A
A
D
D
0
D
D
D
D
D
A
D
I)
A
A
A
A
A
A
Melting
Rate
(T/lir)
1,320
1,100
1,045
880
990
724
825
968
880
880
1,793
1,650
1,595
781
990
715
902
925
990
935
2,090
715
935
1,210
1,045
913
1,100
770
935
583
880
770
935
880
990
Coke Ratio
(T coke/ 100 T)
10.6
14.15
9.33
12.0
10.0
14.4
' 13.25
15.0
12.7
13.2
21.0
10.7
12.3
11.56
11.2
9.5
11.7
13.36
11.0
11.0
15.5
12.5
13.0
15.0
14.6
11.74
13.0
11.24
10.7
12.0
10.2
12.32
12.0
12.0
12.4
Test
Pointc
D
AtlV
A
A+F
1)
B
D
A
A
B
D
D
E+II
D
A+IV
D
A
A
U
D
D
D
E+ll
D
D
0
A
1)
D
A
A
A
A
A
A
Top Gas Total
(gr/scf)
4.02
4.67
2.23
3.06
3.13
4.82
1.73
2.95
3.02
3.23
(24.5)
3.53
3.16
5.48
4.08
(7.52)
3.36
4.52
3.86
5.20
(7.31)
2.93
9.55
4.49
5.48
4.98
4.74
2.80
4.87
3.50
4.60
4.91
4.27
7.45
6.93
Total
Emissions
(Ib/ton Fe)
10.6
18.3
6. 1
10.8
8.7
20.0
6.2
12.2
11. 1
12.1
136.2
9.6
10.7
17.7
13.1
21.5
11.6
17.4
12.3
16.6
29.7
10.6
34.5
18.0
21.8
16.7
17.7
9.1
14.5
12.1
14.1
18.1
15.0
26.3
25.6
Fraction
< 63 l-iin
0.41
0.47
0.49
0.51
0.52
0.57
0.53
0.22
0.43
0.48
0.96
0.69
0.53
0.45
-
0.19
0.46
0.32
0.37
0.43
0.35
0.38
0.45
0.43
0.49
0.19
0.32
0.53
0.48
0.45
-
-
0.43
-
"
Emissions
< 63 M1"
(Ib/ton Fe)
4.35
8.60
2.99
5.51
4.52
11.40
3.29
2.68
4.77
5.93
131.8
6.62
5.67
7.97
-
4.09
5.34
5.57
4.55
7.14
10.40
4.03
15.52
7.74
10.68
3.17
5.66
4.82
6.96
5.44
-
-
6.45
-
"
11 = Hot Blast C = Cold Blast Bll = Basic Hot Blast
Cupola types. A = Cupolas in which the top gas is not used; i.e., cold blast cupolas and hot blast cupolas with independent
blast heating. B = Hot blast cupolas fitted with radiation recuperators. C = Hot blast cupolas in which the cupola gases
are drawn off above the charging door into a nearly radiation or convection recuperator. D = Cupolas (primarily hot blast)
in which all or part of the gas is drawn off through a ring-top duct or draw-off channel just below the charging door.
Sample Points. A = Waste gas (gas mixture which results from the addition of secondary air before the top gas escapes to the
the atmosphere) in cupola stack.
-------
The emission factors for the cupolas shown in Table B-7 as a function
of blast type (excluding the two basic hot blasts) are:
Total Particulate (Ib/T iron) Particulate < 63|Jm (Ib/T iron)
Range Average Range Average
Cold Blast 11.1-26.3 15.4 2.68-8.6 5.36
Hot Blast 6.1-34.5 15.0 2.99-15.52 6.55
The particulate emissions with diameter of less than 63 |Jm are probably
biased low in that they are based on sieve analysis of the sample. It is
quite likely that some agglomeration occurred in the sample train and hence
sieving would identify less particulate in the sub 63 |Jm range than actually
leave the stack.
It is significant that the mean total particulate emissions from cold
blast and hot blast cupolas are almost the same. However, the range in emis-
sions is much greater for hot blast cupolas. Patterson et. al.,14 attribute
this to a wider range in design and operating characteristics for the hot
blast cupolas rather than effect of blast temperature.
An appreciable difference can be seen between cold blast and hot blast
cupolas regarding particulate emissions in the sub 63 [Jm range with hot blast
cupolas having 20% greater emissions. Reference 14 indicated that this in-
crease may have resulted from the larger quantities of small steel scrap
charged to the hot blast cupolas. These small pieces of steel scrap may
result in greater formation of fine iron oxide particles.
Patterson et.al.,14 also analyzed the relationship of particulate emis-
sions to cupola operating parameters. The conclusions that can be drawn
from these analyses are:
1. The density and quantity of particulate with diameter § 63 |Jm de-
finitely increase with increasing blast rate. However there is no signifi-
cant relationship of emissions with diameter < 63 (Jra to blast rate.
2. Increased coke ratios lead to higher emissions of particles with
diameter < 63 pm.
3. No relationship was found between the amount of steel scrap in the
charge and total particulate emissions. However, emissions of particles
with diameter < 63 [Jm definitely increased with an increase in steel content
in the charge.
B.I.1.5 Other Emissions Data
The emissions tests described above comprise a small portion of the
tests that have been conducted on cupolas both by control agencies and foun-
dries. Since the other data compiled during the study are not as amenable
to analysis as those presented above, they are not described in detail.
However, for the sake of completeness, these data are presented in Table
B-8.
50
-------
TABLE B-8. ADDITIONAL CUPOLA EMISSIONS DATA
Reference
Crabaugh (15)
NEDS (16)
AFS (17)
Drake (18)
BCIRA (19)a
No. of Cupolas
Tested
10
45
Unknown
Unkown
1
Emissions
Range
5.5-27
1.3-244
5-29
30-45
(Ib/ton charge)
Average
15.2
35.3
-
-
1.25-1.85 1.55
This is a test of a cokeless cupola which has been converted to use oil
and natural gas. The emission rate is in Ib/ton of metal charged.
The one test of particular significance in Table B-8 is the BCIRA test
of the cokeless cupola. The cupola which is installed at Hayes Shell Cast,
Ltd. is a 5 ton/hr cupola which has been converted completely to gas/oil
firing. No form of particulate control is used, and yet the cupola has an
average emission rate of 1.55 Ib/ton of metal charged.
B.I.1.6 Summary of Cupola Particulate Emissions Data
In summarizing the results of the cupola emissions tests presented above,
several observations are worthy of note. First, particulate emissions vary
over a rather wide range (greater than one order of magnitude). The analyses
indicate that these variations are, at least in part, a result of different
design and operating parameters. In fact, the effect of these parameters
can sometimes be quantified.
These factors certainly have an impact upon enforcement practices. Given
the wide range of emission factors, the utilization of an average emission
factor to enforce process weight regulations has some drawbacks. On the
other hand, the data can provide an enforcement tool. If compliance test
data and associated operating characteristics are available for a particular
cupola, the analyses presented above can be used to estimate the impact of
process or operational changes on the compliance status of that cupola.
Another factor, which may be particularly useful for small foundries,
is the result of testing at Foundry A11 shown in Table B-7. At this foun-
dry screening of the scrap and careful handling to prevent charging of loose
sand, rust, and coke fines resulted in a 50% reduction in emissions. This
practice may be an economically feasible way of reducing emissions in smal-
ler foundries where the cost of fabric filter systems is prohibitive.
These are just two ways in which the data can be used in developing
enforcement strategies. Other possibilities include the use of regression
51
-------
analyses to develop better estimates of uncontrolled emissions for a parti-
cular cupola or comparison of the relative impact of an existing and new
installation.
B.I.2 ELECTRIC ARC FURNACE EMISSIONS
The second major source of particulate emissions from the melting de-
partment is the electric arc furnace (EAF). Emissions from the electric
are furnace occur at five stages of the operation: charging, melting and
refining, backcharging, oxygen lancing (generally used only in steel found-
ries), and tapping. Since each of these different emissions points presents
a unique control problem, emissions data will be presented for each of these
operations when available. However, emissions data are not available in
nearly the detail for EAF's as for cupolas. Hence detailed analysis of emis-
sions with respect to EAF operating characteristics is seldom possible.
An earlier EPA study found that emissions range from 4 to 40 Ibs of
particulate per ton of metal charged with an average emission factor of 13.8
lb/ton.20 The data (presented in Table B-9) show no correlation between
furnace size and emissions. However, they do indicate some relationship
between melt cycle and emissions.
Data are also available which indicate a relationship between the me-
tallic content and cleanliness of scrap charged to the furnace. Coulter
performed several tests under identical conditions but varied the clean-
liness and quality of scrap charged. The results (see Table B-10) showed a
100% increase in emissions (in lb/ton) when dirty scrap with large amounts
of metallic impurities were used in the charge.21 Data from Kane (also in
Table B-10) substantiate this.22 Emissions from Test 2 in which highly oxi-
dized scrap was used are about 30% higher than the average emissions for
the other three tests.
Oxygen lancing is often practiced in steel foundries to achieve proper
chemical composition of the melt. Although data are insufficient to quantify
emissions from oxygen lancing, it is known that lancing increases both the
volume of gas generated and the particulate concentration in the exhaust
stream. It has been found that the average concentration of particulate in
the exhaust stream is 1.1 to 3.7 gr/dscf. However, during lancing peak values
of 5.2 to 6.5 gr/dscf are found.
Limited data are also available on the particle size of electric furnace
emissions. The data from three foundries (see Table B-ll) indicate that
essentially all EAF emissions are less than 50 [Jm in diameter and that about
90% (disregarding Foundry A) are respirable (i.e. < 10 |Jm in diameter).23
However, during the four tests conducted by Kane22 (see Table B-10)
analysis of periodic grab samples by electron microscope indicated that about
95% of all particles were smaller than 0.5 |Jm and almost no particles were
larger than 2 Mm in diameter. This would indicate that EAF emissions are
both difficult to control and have a maximum impact on human health.
52
-------
TABLE B-9. EMISSIONS DATA FROM ELECTRIC ARC MELTING FURNACES26
No.
1
2
3
4
5
6
1
8
9
10
11
12
13
14
15
16
17
18
19
Furnace
Shell
Diameter
Feet
11.0
12.0
8.0
12.0
7.0
12.0
8.0
7.0
7.0
7.0
7.0
9.0
9.0
11.0
9.0
9.0
8.0
11.0
12.0
Furnace
Charge
Tons
15
20
5
20
3
25
5
3
2
2
3
6
6
18
6
6
4
14
19
Furnace
Cycle
Hours
1.15
1.5
1.0
2.5
- 1.75
4.0
1.0
1.75
2.0
1.3
2.0
2.3
2.0
3.0
1.2
1.75
2.0
1.75
1.7
Emissions
Produced
Ib/ton Charge
12.0(Est.)
6.0
20.0
18.3
10.0
4.0
40.0
12.7
10.7
13.4
5.3
15.3
12.8
6.1
29.4
12.7
11.0
7.5
15.0
Sources: 1- 4 Foundry Visits
5- 9 AFS Foundry Air Pollution Manual
10-19 Los Angeles Air Pollution Manual
53
-------
TABLE B-10. ELECTRIC FURNACE EMISSIONS DATA
Source Test No.
Kane 1
2
3
4
AP-40 1
2
a These data are for a
TABLE B-ll.
Particle Size
Distribution, Microns
Less than 1
Less than 2
Less than 5
Less than 10
Less than 15
Less than 20
Less than 50
Scrap Quality
Average
Poor (Highly Oxidized)
Average
Average
Normal
Dirty, Subquality
Emissions
Ib/ton
4.5
7.5
5.8
5.7
9.3a
18. 6a
50 and 75 ton steel furnace.
SIZE DISTRIBUTION FOR THREE
ARC INSTALLATIONS
o
Foundry A Foundry
5% 8%
15% 54%
28% 80%
41% 89%
55% 93%
68% 96%
98% 99%
ELECTRIC23
B Foundry C
18%
61%
84%
91%
94%
96%
99%
a Foundry A provided an agglomerated sample and is, therefore,
less representative.
54
-------
The above data were developed for iron foundries. It has been esti-
mated that the emissions from the melting and refining operations for steel
furnaces average about 16.0 Ib/ton with charging and tapping emissions esti-
mated to be 1.6 Ib/ton.24 No information is available on the method of de-
velopment of these numbers, or on the range of emissions from steel furnaces.
B.2 NONMELTING EMISSIONS
The nonmelting foundry operations which are sources of particulate emis-
sions can be classified as fugitive emissions sources. These sources are
so defined because, in the absence of auxiliary ventilation systems, the
emissions from these sources enter the foundry environment and are exhausted
to the atmosphere through doors, windows, roof monitors, and exhaust vents
rather than through a confined stack.
Emissions data from all fugitive emissions sources; including those in
foundries, are scarce. There are two reasons for the lack of data. First,
fugitive emissions are hard to measure, and problems are associated with
most test methods as described in Reference 25. Second, until recently fu-
gitive emissions sources were not a major concern for either control agencies
or industry personnel charged with air pollution control. As a result there
was little effort to develop reliable emissions estimates.
Due to the lack of data it is often necessary to develop engineering
estimates of emissions from fugitive sources. The remainder of this section
presents both test data and engineering estimates of emissions for pouring
and cooling operations, shakeout, sand handling, and the cleaning room.
It is important that the reader remember that the data in these sections
have a much lower reliability than the emissions data presented earlier for
melting emissions. The emission factors are often best estimates and are
never based on more than a limited number of emission tests of uncertain
accuracy. As such the reader should carefully evaluate the data before using
them for planning or enforcement purposes.
B.2.1 POURING AND COOLING EMISSIONS
If sand molds are used, pouring and cooling operations appear to be
one of the significant sources of emissions in the foundry. As the hot
metal is poured into the mold, metallic fumes and products of combustion
and decomposition of the binder systems are released to the foundry environ-
ment. The quantity of these emissions is probably related to such factors
as mold composition, mold size, sand to metal ratio, metal temperature and
pouring rate. However, data are insufficient to quantify the effect of these
factors. Available emissions estimates for pouring and cooling operations
are presented in Table B-12. The sources of the data are described below.
As a part of an EPA study to develop test methods for fugitive emis-
sions, Kalika conducted a series of quasi-stack tests on actual pouring emis-
sions in an iron foundry. A quasi-stack test is one in which a hooding or
enclosure system is placed over the operation and the emissions are exhausted
55
-------
TABLE B-12. POURING AND COOLING EMISSIONS
o
Data Source
Kalika (26)
Bates et.al. (27)
Gutow (28)
Gutow (28)
NEDS (16)
NEDS (16)
Method of , No. of
Determination Tests
TF
TB 1
U
U
TF 4
E 8
Emission Factor (Ib/ton
of metal poured)
Range Average
0.55-4.5
8.3
5.10C
10.3d
0.9-25 11.3
0.09-19 8.8
Particle
Size Data
95%
60%
90%
-
< 5(Jm
> 50|Jm
> 50|Jm
-
-
a Number in parentheses is Reference number.
b TF = Test on a full scale operation(s).
TB = Test on a bench scale or pilot scale operation.
E = Engineering estimate.
U = Unknown.
c Pouring only.
d Cooling only.
56
-------
through the hood into a duct. A Method 5 train is then used to sample emis-
sions from in the duct. As shown in the table, emissions from these tests
ranged from 0.55 to 4.5 Ib/ton of metal poured.
In another study conducted for NIOSH, quasi-stack tests were run on
both bench-scale and pilot-scale pouring and cooling operations. The pour-
ing and cooling of 30-lb cube casting resulted in a total of 54.61 g of
particulate,22 which gives an emission rate of 8.3 Ib/ton. Based upon con-
centration profile data, this has been separated into 4.0 Ib/ton for pour-
ing and 4.3 Ib/ton for cooling.27 Data on concentrations of organic gases
evolved during pouring and cooling are also presented. However, data are
insufficient to determine emission factors.
Gutow28 has also developed emission factors for iron pouring and cool-
ing. The emission factor given for pouring is 5.10 Ib/ton of melt with 60%
of the particles greater than 50 (Jm. If it is assumed that particles greater
than 50 |Jm settle in the foundry, the factor for emissions escaping to the
atmosphere is 2.0 Ib/ton of melt. The cooling emission factor is 10.30 Ib/ton
of melt with 90% of the particles being greater than 50 (jm. Under the same
assumption as above, the emissions which escape to the atmosphere are 1.0
Ib/ton.
As a part of this study the National Emissions Data System (NEDS) was
surveyed to obtain all available emissions data. Foundry data identified
in the system had been obtained by one of these four methods:
Emissions testing.
National Air Data Branch approved emission factor.
Material balance.
Guess.
The four sources for which data were obtained from emissions testing had
emissions ranging from 0.9 to 25 Ib/ton with an average of 11.3 Ib/ton. The
other three methods were combined into a single category based on engineer-
ing estimates. The emissions in this category ranged from 0.09 to 19 Ib/ton
with an average emission factor of 8.8 Ib/ton.
The emissions from pouring and cooling (or casting) are particularly
important because most of these sources are not controlled. As can be seen
from Table B-12 emissions factors developed through testing have a wide range
of 0.55 to 25 Ib/ton of metal cast. Given this wide range of emissions and
average emissions in the range of 6 to 10 Ib/ton, this source warrants fur-
ther investigation.
B.2.2 SHAKEOUT EMISSIONS
After the castings have been solidified, they are taken to the shakeout.
Here the metal casting is removed from the sand mold by one of the methods
described in Appendix A. The removal of the casting from the mold releases
moisture which has been trapped in the mold, dust from the cereal and sea-
coal binders which have dried during casting, and products from the thermal
decomposition of organic binders that occurs as they are exposed to air.
57
-------
Available emissions data for the shakeout operations are presented in Table
B-13.
As a part of the NIOSH testing described in Section B.2.1, Bates and
Scott sampled emissions from shakeout of the same 30 Ib casting. Emissions
were found to be 3.15 Ib/ton of casting.26
As a part of an EPA study to collect data on organic emissions from
ferrous foundries, SASS train samples were gathered upstream and downstream
from a wet scrubber on a shakeout. The individual castings in the molds
weighed about 9 Ib with total iron per mold of about 140 to 160 Ib. The
sand to metal ratio was about 5 to 1. The resultant uncontrolled emissions
were 14 Ib particulate per ton of metal cast while the controlled emissions
were 0.086 Ib/ton of metal cast.27
Gutow estimated an emission factor of 32.20 Ib/ton of metal poured with
90% of the emissions > 50 |jm. Again assuming that those particles > 50 (Jm
settled in the foundry, total emissions to the atmosphere from an uncontrolled
source are 3.2 Ib/ton of metal poured.28 It should be noted that if the
particulate is collected in a hooding system, malfunction of a control device
could result in total emissions of greater than. 3.2 Ib/ton.
If a singular high emission factor of 88 Ib/ton is discarded, the 11
tests in NEDS indicate that emissions range from 0.17 to 18 Ib/ton with an
average of 3.2 Ib/ton. The average emission factor of 15.7 Ib/ton from the
NEDS engineering estimates appears quite high in relation to the test data
presented above.
Kane has estimated that a typical foundry pouring 5 ton/hr of hot metal
will collect emissions of 100 to 700' Ib/hr and discharge an additional 10
to 25 Ib/hr to the atmosphere. This results in a total uncontrolled emis-
sion factor of 22 to 145 Ib/ton of metal poured.29 This factor appears quite
high in relation to other emission factors. In examining Kane's data, it
appears that the calculated emissions are based on an average exhaust of
15,000 to 30,000 scfm and grain loadings ranging from 1 to 3 gr/scf. This
exhaust range appears reasonable in comparison to an exemplary system iden-
tified in a NIOSH study which exhausted 26,200 cfm (at near atmospheric con-
ditions)30 for a shakeout within the design parameters described by Kane.
The dust concentrations are also less than the 3 to 5 gr/scf suggested by
A.T. Kearney31 for shakeout operations with 50% of these emissions in the 2
to 15 |Jm range. Given this corroberation, the data do not appear that un-
reasonable. It should be noted, that Kane estimates the controlled emissions
to be in the range of 2 to 5 Ib/ton of metal poured.
Again given the wide range of both estimates and test data (0.17 to 18
Ib/ton), additional effort is needed before a reliable estimate of shakeout
emissions can be developed.
B.2.3 SAND HANDLING EMISSIONS
In those foundries using green sand molds, the sand which is removed
from the castings must be conditioned and returned to the molding area for
58
-------
TABLE B-13. SHAKEOUT EMISSIONS
o
Data Source
Bates (27)
Gutow (28)
Kane (29)
Method of , No. of
Determination Tests
TB 1
U
U
Emission Factor (Ib/ton
of metal poured)
Range
-
20-140
Average
3.15
32.20
-
Particle
Size Data
98% <15Mm
90% >50|Jra
Mean Size
0.5-1.5|Jm
NEDS (16)
NEDS (16)
Baldwin (27)
TF 12 0.17-87(18)° 10.2(3.2)C
E 60 0.23-88 15.7
TF 14
a Number in parentheses in Reference number.
b TF = Test on a full scale operation(s).
TB = Test on a bench scale or pilot scale operation.
.E = Engineering estimate.
U = Unknown.
c Values in parentheses represent upper end of range and overage if
the single high value of 88 Ib/ton is removed.
59
-------
reuse. During this period of time the sand is relatively dry and suscepti-
ble to extensive dust emissions. As described in Appendix A, the process
steps may vary considerably between foundries. These process differences
will result in different emission rates. Sand handling emissions data are
not sufficient to estimate emissions for unit operations within the handling
process. Thus the emissions data presented in Table B-14 and discussed be-
low are for the total sand handling process starting from the point where
the sand leaves the shakeout until it is ready to enter the muller. The
reliability of these estimates is low, and the data should be used with care.
As can be seen from Table B-14, the only emissions test data available
are those obtained from the NEDS survey. These data indicate that sand hand-
ling emissions range from 0.6 to 50 Ib/ton of sand handled.
Kane has estimated that the emissions collected from the sand system
range from 150 to 500 Ib/hr with an additional 5 to 15 Ib/hr emitted to the
atmosphere for a plant melting 5 ton/hr.29 Assuming a sand to metal ratio
of 6:1, a sand handling rate of 30 ton/hr is estimated. Thus emissions range
from 3 to 10 Ib/ton of sand handled.
Gutow has estimated emissions from the sand system and mulling to be
21.8 and 20.6 Ib/ton melt respectively with the sand system including emis-
sions from dry sand handling, screening, and sand drying and reclamation.
Since 90% of these emissions are estimated to be > 50 pm, the amount of emis-
sions escaping to the atmosphere are 2.2 Ib/ton of melt for the sand system
and 2.1 Ib/ton of melt for the muller if the sources are uncontrolled.28
Note that these emission factors are based on the quantity of metal melted.
Since typical sand to metal ratios range from 4:1 to 10:1, the emission fac-
tors would both be well below 1 Ib/ton of sand handled.
B.2.4 CLEANING ROOM EMISSIONS
The cleaning room is probably the most diverse area with respect to
emissions sources in the iron foundry. Emissions can come from such varied
sources as abrasive cleaning (shot, sand, or tumble blast), torch cutoff,
air-carbon arc cleaning, chipping, core knockout, and grinding. Depending
on the operations used, emissions from this area of the foundry can contain
course metal dust, metallic oxide fume, sand dust (from either the abrasive
or sand which is "burned in" on the metal surface), and grinding wheel bond
material. It is known that emissions from the cleaning room are dependent
on such factors as type of casting, types of molds and cores used, quality
control in the pouring operation, types of operations used in cleaning, and
work practices applied during cleaning. However, sufficient data are not
available to quantify emissions from most unit operations, and data are cer-
tainly not sufficient to determine the effect of the operating parameters
described above on emissions.
The only operations for which test data are available are abrasive clean-
ing and grinding. These data are presented in Table B-15. The only test
data shown in the table are from the NEDS survey for shot and sand blasting.
The emissions estimates for the tests range from 27 to 500 Ibs/ton of metal
cleaned. It should be noted that the 500 Ibs/ton is not an isolated high
60
-------
TABLE B-14. SAND HANDLING EMISSIONS
Data Source
Gutow (28)
Kane (29)
NEDS (16)
NEDS (16)
Gutow (28) e
NEDS (I6)e
Method of , No. of
Determination Tests
U
U
TF 6
E 53
U
E 30
Emission Factor (Ib/ton
of metal poured)
Range
5-10d
0.6-50
0.08-93
-
0.11-24
Average
21.80°
-
3.2d
I4d
20. 6C
4.2d
Particle
Size Data
90% >50(Jm
90% >50|Jm
a Number in parentheses is Reference number.
b TF = Test on a full scale operation(s).
TB = Test on a bench scale or pilot scale operation.
E = Engineering estimate.
U = Unknown.
c Ib/ton melt.
d Ib/ton sand.
e Mulling only.
61
-------
TABLE B-15. CLEANING ROOM EMISSIONS
Q
Data Source
Gutow (28)
Kane (29)
NEDS (16)
NEDS (16)
Operation
Shot Blast
Grinding
Grit Blast
Room
Airless Blast
Grinding
Shot or Sand
Blast
Shot Blast
Grinding
Method of No. of
Determination Tests
U
U
U
U
U
TF 7
E 100
E 41
Emission Factor (Ib/ton) Particle
Range Average
15.5
1.6
40-100
40-60
2-3
27-500 257
0.065-241 59
0.4-125 40
Size Data
80% >50Mm
80% >50|Jm
a Number in parentheses is a reference number.
b TF = Emissions testing on a full scale operation(s).
TB = Emission testing on a bench or pilot scale operation(s).
E = Engineering estimate.
U = Unknown.
62
-------
value as several other tests estimated emissions to be in the 250 to 4501b/
ton range. Since these numbers appear excessively high, the NEDS printout
was reviewed to identify reasons for the high values. It was found that in
most cases the test data were for controlled sources with efficiencies rang-
ing from 98 to 99.9%. Slight changes in the efficiency (as little as 1 or
2%) can make a large difference in the uncontrolled emission factor. However,
one source, with a control efficiency of 70%, still had an uncontrolled emis-
sion factor of 330 Ib/ton of metal cleaned. Thus, it is possible that these
high factors may be correct.
Gutow28 has estimated total emissions for shot blasting and grinding
to be 15.5 and 1.6 Ibs/ton respectively. However, 80% of the particulate
are estimated to have a diameter > 50 (Jm. Again, assuming those particles
larger than 50 pm settle in the foundry, the emissions to the atmosphere
from an uncontrolled source are estimated to be 3.1 Ib/ton for shot blast
and 0.3 Ib/ton for grinding. However, for controlled operations, the po-
tential emissions should the control device malfunction are the total emis-
sions values.
Kane29 has estimated that for a typical foundry melting 5 ton/hr, 200
to 500 Ib/hr of particulate are collected and 2 to 5 Ib/hr of particulate
are discharged to the atmosphere from the grit blast room. An additional
200 to 300 Ib/hr of particulate are collected and 2 to 3 Ib/hr of particu-
late are discharged to the atmosphere from an airless blast. Finally 10 to
15 Ib/hr of particulate are collected and 0.5 to 1 Ib/ton of particulate
are discharged from grinding wheels. These data were used to calculate the
emission factors shown in Table B-15.
The impact of emissions from the cleaning room vary depending on the
operation. Abrasive cleaning is accomplished in a closed structure and,
with proper operating procedures, the emissions are confined in an exhaust
system which is amenable to control. These emissions are commonly controlled
in iron foundries. Emissions from the other cleaning room sources are fugi-
tive and are not controlled to the same degree as those from blasting. Even
though the emission factor for blasting is higher than the other cleaning
room sources, it may have a lesser impact on air quality than other sources
due to the high degree of control.
B.2.5 SUMMARY OF FUGITIVE SOURCE EMISSIONS DATA
In reviewing the emissions data presented above, it is apparent that
the various fugitive emissions sources are a significant contributor to par-
ticulate emissions from ferrous foundries. Emissions test data indicate
fugitive emissions from nonmelting operations emit the following quantities
of particulate:
Units Range "Best" Estimate
Pouring and Cooling (Ib/ton metal poured) 0.55-25 6-10
Shakeout (Ib/ton metal cast) 0.17-18 3-4
Sand Handling (Ib/ton sand) 0.6-50 3
63
-------
Test data are insufficient to estimate emissions from other sources. Note
that the emissions for sand should be multiplied by a factor ranging from 4
to 10 (depending on the sand to metal ratio) to compared total emissions
from the other sources.
The impact of these emissions is substantiated by the data from Cookman
and Johari shown in Table B-16.32 These data were compiled by using a high
volume air sampler to measure particulate concentrations immediately below
roof exhaust fans at an operating foundry. Fan exhaust rates were than used
to calculate the particulate emission rate. These data indicate that both
the sand system and the pouring area are significant contributors to atmo-
spheric emissions.
TABLE B-16.
EMISSIONS FROM ROOF EXHAUSTS AT
AN IRON FOUNDRY32
Location
Penthouse
Shakeout Roof
Molding Roof
Furnace Roof
Pouring Roof
Emission Rate
Ib/hr
9.6
0.4
0.3
2.8
4.2
Particulate Characteristics
Composition
Almost all Sand
>95% Sand
>5% Iron
Sand/Iron
99% Iron
75% Iron
25% Sand
Size
1 to 50pm
1 to SOpra
0.1 to 30|Jm
0.1 to 20|jm
0.1 to 20|Jm
Mostly Fine
0.1 to 20(Jm
64
-------
REFERENCES FOR APPENDIX B
1. U.S. Environmental Protection Agency. Compilation of Air Pollution
Emission Factors. Second Edition. April 1973. p. 7, 10-1.
2. Reference 1, p. 7, 13-2.
3. Miller, W. C. "Reduction of Emissions from the Gray Iron Foundry In-
dustry." Paper 71-134. Presented at the 64th Annual Meeting of the
Air Pollution Control Association, June 27 - July 2, 1971.
4. A.T. Kearney Co. "Systems Analysis of Emissions and Emissions Control
in the Iron Foundry Industry, Volume I, Text." PB 198 348, U:S. En-
vironmental Protection Agency. February 1971.
5. A.T. Kearney Co. "Systems Analysis of Emissions and Emissions Control
in the Iron Foundry Industry, Volume II, Exhibits." PB 198 349 U.S.
Environmental Protection Agency. February 1971.
6. Reference 5. Exhibit VI-9.
7. A.T. Kearney Co. "Systems Analyses of Emissions and Emissions Control
in the Iron foundry Industry, Volume III, Appendices." U.S. Environmental
Protection Agency. February 1971. Appendix B. Exhibit II.
8. Reference 5. Exhibit VI-11.
9. Reference 5. Exhibit VI-7.
10. Reference 5. Exhibit VI-8.
11. Warda, R.D. and Buhr, R.K. "A Method for Sampling Cupola Emissions."
AFS Transactions. Volume 81. p. 24-31. 1973.
12. Warda, R.D. and Buhr, R.K. "A Detailed Study of Cupola Emissions."
AFS Transactions. Volume 81. p. 32-37. 1973.
13. Davis, J.W. and Draper, A.B. "Effect of Operating Parameters in Cupola
Furnaces on Particulate Emissions." AFS Transactions, p. 287-296.
1973.
14. Patterson, W., Weber, E., and Engles, G. "Dust Content of Cupolas for
Cupolas of Different Designs and Modes of Operation.". The British
Foundryman. p. 106-117. March 1972.
15. Crabaugh, H.R., Rose, A.M., and Chass, R.L. "Dust Fumes from Gray
Iron Cupolas - How They are Controlled in Los Angeles County." Air
Repair, 4(3). p. 125-130. November 1954.
65
-------
I I
I '
16. National Emissions Data System listing of ferrous foundry emissions
sources, July 1979.
17. Foundry Air Pollution Control Manual. 2nd Edition. American Foundry-
men's Society. Des Plaines, Illinois. 1967.
18. Drake, J.F. and Kennard, T.G. "Closed Top System in Cupola Stack Emis-
sions Control." American Foundrymen, 17(2). p. 55-57. February 1950.
19. "Recent Tests on the Cokeless Cupola." Foundry Trade Journal, p. 234-
235. February 19, 1976.
20. Reference 5. Exhibit VI-16.
21. Kane, J.M. and Sloan, R.V. "Fume Control - Electric Melting Furnaces."
American Foundryman, 18(5). p. 33-35. November 1950.
22. Reference 5. Exhibit VI-15.
23. Fennelly, P.F. and Spawn, P.O. "Air Pollutant Control Techniques
for Electric Furnaces in the Iron and Steel Foundry Industry." U.S.
EPA Publication No. EPA-45012-78-024. June 1978.
24. Wallace, D.W. and Cowherd, C. "Fugitive Emissions from Iron Foundries."
U.S. Environmental Protection Agency. Publication No. 60017-79-195.
August 1979. p. 32-37.
25. Kalika, P.W. "Development of Procedures for Measurement of Fugitive
Emissions." U.S. Environmental Protection Agency. Contract No. 68-
02-1815. July 1975.
26. Bates, C.E. and Scott, W.D. "Better Foundry Hygiene Through Permanent
Mold Casting." Contract No. 1 R01 OH 000456-01. NIOSH. January 1976.
p. 64-66.
27. Baldwin, V.H. Jr., "Environmental Assessment of Iron Casting," U.S.
Environmental Protection Agency, EPA-60012-80-021, January 1980. p.
67-71.
28. Gutow, B.S. "An Inventory of Iron Foundry Emissions." Modern Casting.
p. 46-48. January 1972.
29. Kane, J.M. "Air Pollution Ordinances." Foundry, p. 104-107. October
1952.
30. "An Evaluation of Occupational Health Hazard Control Technology for
the Foundry Industry." U.S. Department of Health Education and Welfare.
DHEW (NIOSH) Publication No. 79-114. October 1978. p. 218-222.
31. Reference 5. Exhibit III-7.
32. Cookman, M.A. and Johari, 0. "Measurement of Iron Foundry Particulate
Emissions." Foundry Management and Technology. October 1974. p. 78-79.
66
-------
APPENDIX C
DESCRIPTION OF AVAILABLE CONTROL SYSTEMS
67
-------
This appendix describes, in some detail, emissions control systems that
are available for five operations or areas of operation: (a) cupolas; (b)
electric arc furnaces; (c) pouring and cooling; (d) shakeout and sand hand-
ling; and (e) the cleaning room. The appendix is divided into five sections
that describe systems for these areas. The first section on cupola emissions
provides detailed descriptions of both wet scrubbers and fabric filters as
well as descriptions of gas conditioning equipment. The remaining sections
are devoted primarily to discussions of the capture systems for these fugi-
tive sources. The primary removal devices used with these capture systems
are also fabric filters and wet scrubbers. The features of these devices
relevant to the source are described in Sections C.2 through C.5. However,
the reader should refer to section C.I for more detailed descriptions.
C.I CUPOLA EMISSION CONTROLS
More effort has been focused on the control of emissions from the cu-
pola than from any other ferrous foundry emissions source. Early control
attempts (prior to 1940) used baffle plates, spark arrester screens and dry
mechanical collectors to eliminate sparking and dust buildup on foundry roofs.
Since that time, various equipment including wet caps, low and high energy
scrubbers, electrostatic precipitators and fabric filters have been installed
on foundry cupolas. Some cupolas in smaller foundries still use wet caps
for control. However, most foundries now have installed either fabric fil-
ters or high energy scrubbers, primarily venturi scrubbers, on the cupola.
In addition most cupolas use gas or oil fired afterburners prior to
the control device to control carbon monoxide and organic materials. The
two reasons for controlling these gases are: (a) to comply with emissions
regulations, (b) to prevent explosions in the gas cleaning system, and (c)
to reduce problems caused by condensation of organic materials in the pri-
mary control device. These afterburners generally consist of three direct
flame burners located at 120° positions around the cupola stacks. The burners
are located below the charge door upstream from the exhaust takeoff for the
control device.
The remainder of this section briefly describes the primary particulate
control devices used on cupolas. The first section describes wet caps, the
second describes verturi scrubbers, and the final section discusses fabric
filters and the associated gas stream conditioning mechanisms. Greater de-
tail can be found in References 1 and 2.
C.I.I Wet Caps
The wet cap was one of the earliest controls applied to the cupola.
Some of the earliest systems were installed in the 1940's. Wet caps are
still found on cupolas in smaller foundries located outside metropolitan
areas.
Shown schematically in Figure C-l, the wet cap is essentially a low
pressure wet scrubber. The wet cap is an expanded shell that extends above
the top of the cupola stack and requires no auxiliary fan to draw gases
68
-------
\—Entrained
*— Moisture
Eliminator
CO Gas Vent
Drain
Figure C-l. Typical Cupola Wet Cap.3
69
-------
through the device. The wet cap consists of one or more inverted cones sur-
rounded by a collection trough. A water stream is passed over the cones,
forming a water curtain through which the emission stream must pass before
exiting to the atmosphere. Most large particles impinge on the curtain and
are carried to the collection trough.
Wet caps are not effective in controlling the fine particulate in cupola
emissions. Data indicate that wet caps will not collect particles less than
10 |Jm in diameter4. Limited data indicate that the best wet caps can only
limit emissions to about 0.1 gr/scf with other installations having much
higher emissions5.
C.I.2 High Energy Wet Scrubbers
Because the wet caps described above do not effectively remove the fine
particle fraction of cupola emissions, improved control equipment has been
required at many foundries. High energy venturi scrubbers have been applied
to most large cupolas and to some medium sized cupolas.6 Generally, two
types of scrubbers have been used in foundries, the conventional venturi,
which has a fixed throat orifice, and the flooded disk scrubber which has a
variable throat orifice. The paragraphs below discuss the basic operating
principles of the venturi scrubber, describe the designs of the two types
identified above, identify factors which effect scrubber performance, and
present available data on the performance of wet scrubbers installed on
cupolas.
The primary collection mechanisms involved in a venturi scrubber are
the impingement of particles on droplets and the condensation of liquid on
the particles. Impingement is attained by accelerating the gas stream to
high velocities (200 to 600 ft/sec) in the venturi throat. When water is
introduced into the high velocity stream it is atomized into tiny droplets.
Since these droplets are at a relatively low velocity with respect to the
gas stream, the particles are collected on these droplets through impaction.
Particle conditioning through condensation also occurs when saturated streams
are cooled in the venturi.
A typical cupola scrubber system is shown in Figure C-2. After the
gas stream leaves the venturi throat it goes through a mist eliminator (a
cyclonic or impaction collector) where the particle laden droplets are re-
moved from the gas stream. The captured liquid is generally cleaned in a
settling tank or pond and reused in the scrubber. A quench spray or wet
cap upstream from the scrubber is used to protect the scrubber material
from heat damage and modulate temperature variations which might affect
scrubber performance.
Some typical examples of the conventional venturi scrubber are shown
in Figure C-3. The water introduced at right angles to the gas stream by
means of spray nozzles (Figure C-3a) or by means of a weir box (Figure O3b, c)
The latter method is used most frequently as it allows a much greater degree
of water recirculation.8'9 High energy scrubbers for cupolas use approxi-
mately 8 gpm per 1000 cfm of stack gas with 1/2 to 1 gpm per 1000 cfm re-
moved from the system with the collected particulate.10
70
-------
Failsafe
Cupola Cap
Combustion
Zone
Air Indraft
Through
Charge Door
Burner
Cyclonic Type
Demister
Settling/ Coarse
Recycle Sludge
Tank
/7
From Water Cooling
Tower
To Water Cooling
Tower
Sound
pO|
(1(1 IT Attenuation
/ \
Cupola
Thicking
Tank Concentrated
Fines to Recycle
Disposal Pump
FOUNDRY CUPOLA
Figure C-2. Typical Cupola Scrubber System.
71
-------
Figure C-3. Examples of Venturi Scrubbers
11
72
-------
Another type of venturi scrubber frequently used on cupolas is the
flooded disk scrubber shown in Figure C-4.
In operation, water flows from a large central support pipe and floods
over a nonrotating disc. The shearing action of the gas at the edge of the
disc atomizes the water into fine droplets which collide with and capture
the fume particles.
To maintain optimum pressure drop at different gas flows, the disc is
raised or lowered. This increases or decreases the annular area through
which the gases must pass. In this way, the scrubber is able to provide
the same efficiency performance over a wide variation of gas volume ranges.
Because of the temperature and volume increases during burn down, the dust
collector for either above or below charge level takeoff must be oversized
to handle the extra volume. Spray cooling permits the collector to be sized
for normal operation with a small allowance for burn down. The scrubber
can be adjusted manually or automatically, to maintain optimum cleaning
conditions over a range of gas rates.
For both types of scrubbers, factors which affect the choice of a con-
trol system are: (a) particle size distribution; (b) pressure drop; (c)
liquid to gas ratio; and (d) effect of the gas stream on scrubber materials.
The efficiency of wet scrubber is dependent upon the size distribution
of the particulate emissions and the pressure drop across the scrubber.
Data in Appendix B indicate that as much as 50% of the cupola emissions may
be less than 5 |Jm in diameter with a significant quantity in the submicron
range. It is estimated that pressure drops of 60 to 80 in. H20 (and in some
cases up to 100 in. H^O) will be needed to control these emissions.
The liquid-to-gas ratio is basically a function of the inlet solids
content, inlet gas temperature and method of water introduction. As indi-
cated earlier, about 8 gpm/1000 cfm is required for cupola applications.
Because of the sulfur oxides generated from the coke and the silica
content of cupola emissions, this emissions stream is both corrosive and
errosive. Data collected during the study indicate that it is often neces-
sary to construct the venturi throat and separator from stainless steel.
It is suggested that in some cases that the fan housing be epoxy coated and
the fan housing be made of stainless steel.8 Some care should also be taken
to prevent wear in the water recirculation system.
As indicated earlier the efficiency of a wet scrubber is dependent upon
the pressure drop across the scrubber. Davis, et al., have indicated that
high energy scrubbers are capable of reducing outlet particulate loadings
to about 0.05 gr/scf.13 This is substantiated by AFS, which indicates that,
depending upon quality of scrap, scrubbers with pressure drops in the range
of 50 to 70 in. HgO will reduce emissions to 0.05 gr/scf with reductions to
0.03 gr/scf for scrubbers with 100 in. H^O pressure drop.8
73
-------
Clean Gas Out
Dirty Gas In
Water In
Slurry Out
Figure C-4. Flooded Disk Scrubber.12
74
-------
C.I.3 Fabric Filter
A second control device which effectively collects particulate emis-
sions from the cupola is the fabric filter or baghouse. In a fabric filter
the particulate laden gas stream passes unidirectionally through a woven
fabric, deposits the particulate on the fabric and emerges clean on the op-
posite side. Currently, fabric filters are applied to most small cupolas
and to some medium sized cupolas.6 The paragraphs below describe the phys-
ical design, collection mechanisms, auxiliary equipment, operating charac-
teristics, and control effectiveness of fabric filters.
The fabric filter or baghouse is a relatively simple mechanism. A
typical baghouse (shown in Figure C-5) consists of an inlet air plenum, a
series of tubular or envelope filters, an exhaust air plenum, a dust col-
lection hopper, and an exhaust fan. The particulate laden gas stream enters
the inlet air plenum and passes to the "dirty side" of the fabric filters.
As the gas stream passes through the filter a "cake" is formed on the fil-
ter. This cake is the primary collection surface and is responsible for
the high efficiency removal of particulates by the filter. The gas stream
then passes through to the "clean side" or outlet plenum and out to the
atmosphere. As the dust layer continually builds on the bags they must be
cleaned periodically to avoid excessive pressure drops. Part (or all) of
the bags are cleaned together at regular intervals by a shaker or reverse
air mechanism. The dust falls into the hopper and then is removed from the
baghouse.
In addition to the baghouse, two additional pieces of equipment are
necessary if fabric filtration is to be used, an afterburner and a gas cool-
ing mechanism. The paragraphs below discuss the need for these devices and
briefly describe typical systems.
As indicated earlier, afterburners are often installed in cupolas with
control equipment to oxidize the CO and prevent explosion damage in the con-
trol system. If fabric filtration is used, afterburners are an essential
part of the control system. In addition to controlling CO emissions, these
afterburners also decompose any oils and tars that are emitted from the
charging of dirty, oily scrap. If these oils and tars reach the baghouse
they stick in the pores and may cause bag blinding.
If these pollutants are to be fully controlled, the stack temperature
at the afterburner must be maintained at 1300° to 1500°F, depending on the
type of material charged.15 Also, to ensure complete combustion, three burn-
ers located at 120° intervals around the stack should be used rather than
single burner.
The maximum temperature at which any commercially available filtering
media can continuously operate is 550°F'. Since the gas stream leaves the
afterburner at 1300° to 1500°F, it must be cooled before entering the bag-
house. Three cooling mechanisms are available: dilution, evaporation, and
radiation.
75
-------
Figure C-5. Typical Fabric Filter.
14
76
-------
Dilution cooling is simply adding ambient air or exhaust streams at or
near ambient temperature to the cupola exhaust stream in sufficient quantity
to reduce the temperature to the desired level. If ambient air is used,
dilution cooling has the disadvantage of increasing the volume of gas pro-
cessed and hence the size and cost of control equipment. In addition, since
the outlet concentration of a fabric filter is relatively insensitive to
changes in inlet concentration, an increase in the volume of gas processed
effectively increases the total mass emissions.
If it is possible to combine the cupola exhaust stream with other ex-
haust streams at or near ambient temperatures (e.g., the exhausts from sand
handling or grinding) dilution cooling may be economically feasible. During
plant visits, foundry personnel indicated that more information is needed
on the types of exhaust streams that can be combined for cleaning. Two con-
cerns raised by plant personnel were possible synergistic effects of dif-
ferent streams on bag life and the types of media that should be used with
combined streams.
Evaporative cooling involves the introduction of water into the gas
stream in the evaporative cooler (or quencher). The hot gas stream sup-
plies the heat necessary to vaporize the water resulting in a cooling of
the gas stream. It is important that the evaporative cooler be designed to
ensure that no droplets are carried over to the filter. This is best accom-
plished if the water is introduced as a fine spray or as a mist. Droplet
size for hydraulic nozzles operating with relatively high pressure is 100
to 200 [Jm. For the temperature difference of the cupola quencher a reten-
tion time of about 0.5 sec is needed to ensure droplet evaporation.16
Two additional factors should be considered when using evaporative cool-
ing. First a backup dilution cooling and by-pass system should be included
in the system to avoid bag burnout in the event of a quencher failure. Second,
since the gas is at or near saturation leaving the quencher, the baghouse
should be sufficiencly insulated to prevent the temperature from dropping
below the dewpoint. Condensation in the baghouse can result in both bag
blinding and corrosion of bags and the housing.
The most frequent application of radiant cooling in a cupola is in the
use of an indirect air to air heat exchanger. This mode of cooling is found
in recuperative systems, where the exhaust air is passed through a chamber
with indirect heating coils. Air is transported countercurrently on the
inside of the coils for use as cupola blast air. The use of hot blast has
the effect of reducing the coke consumption and increasing the cupola melt
rate.
Other baghouse design features that are important to both the foundry
operator and the air pollution control agency charged with regulating the
foundry include:
77
-------
Location of the fan with respect to the baghouse;
• Cleaning mechanism;
• Filter media; and
Air-to-cloth ratio.
These factors which affect baghouse performance, ease of operation, and pos-
sible regulatory actions are described below.
In cupola baghouses the fan may be located either upstream of the filter
resulting in a "positive pressure baghouse" or downstream from the filter
resulting in a "negative pressure baghouse." In most cases positive pres-
sure baghouses are used on cupola systems primarily because of easier moni-
toring and maintenance of the bags.17 In addition, the baghouse acts as a
muffler to reduce fan noise. From a regulatory standpoint, the use of a
positive pressure baghouse may result in greater difficulty in enforcing
regulations. The gas stream from a positive pressure system is generally
exhausted through a roof monitor rather than through a well defined duct
making both emissions and opacity measurements more difficult. One advan-
tage of the positive pressure system is that it does allow easy inspection
of the clean side of the bags during operation.
Various cleaning mechanisms are used to remove the collected particu-
late from the filter media. The choice of cleaning methods depends pri-
marily on the filter material and particulate characteristics. The methods
used on fabric filters associated with most metallurgical furnaces are shaker,
reverse air, and pulse jet with shaker and reverse air being the predominate
methods. These mechanisms, shown in Figures C-6, C-7, and C-8 are described
below.
Mechanical shaking of the bags can be accomplished in many different
ways. The most common method is to use a motor attached to a common frame-
work at the upper end of the bag as shown in Figure C-6. The motor then
moves the framework in a variety of different ways causing the dust to fall
into the hopper.
During shaking, air flow should be stopped, or some of the particulate
will penetrate the bag and be emitted to the atmosphere. Generally the air
flow is stopped by dividing .the baghouse into compartments and closing off
individual compartments for cleaning while continuing to operate the remainder
of the baghouse. However, on some small cupolas which operate for only a
short time it may be possible to shake only at the end of a melt.
Shaking may be accomplished on either a preset timed sequence or manu-
ally as needed to maintain proper pressure drop. Several foundries con-
tacted during the study indicated that bag life and equipment performance
had been improved by using manual rather than automatic shaking.
78
-------
Shaker Motor
D A • i -T-> i -^^^^••••^^•a^j^^ai^mjmjj
Reverse Air and Clean Air Plenum
Dust Conveying Rotary
System Discharge
Dirty Air
Clean Air
Figure C-6. Reverse Air or Shaker Type Baghouse.
18
79
-------
Exhaust Inlet
Blower
Blow Ring Makes Contact
-with Cloth Tube
Dust Hopper
Inlet
A. Filter Cylinder
B. Wire Retainer
Collars
Tube Sheet
Nozzle or Orifice
Timer
Collector Housing
J. Air Manifold
K. Upper Plenum
Inlet
Hopper
Solenoid Valve
Exhaust Outlet
Figure C-7.
Saghouse Showing Two Methods of Cleaning by Reverse
Air Flow.1-9
80
-------
Dirty Air
Clean Air
Compressed Air
Dust Conveying Rotary
System Discharge
Figure C-8. Pulse Jet Type Baghouse
18
81
-------
If the dust releases fairly easily from the fabric, a low-pressure re-
versal of the flow may be enough to loosen the cake without mechanical agi-
tation. To minimize flexural attrition of the fabric, it is supported by a
metal grid, mesh, or rings and is usually kept under some tension. The sup-
port is usually on the clean side of the tube or bag, although dirty-side
support can help to keep the sides of the bag or the panels sufficiently
apart to allow the cake to fall to the hopper.26
As illustrated in Figures C-7 and C-9, there are several ways of accom-
plishing flow reversal. In addition to the standard dampers on each compart-
ment, each one can have its own reversing fan. A few models have a traveling
apparatus that goes from bag to bag or from panel to panel, blocking off
the primary flow and introducing some air in the reverse direction with a
secondary blower. Perhaps a simpler method is to take advantage of a suc-
tion on the dirty side or a relative pressure on the clean side, without
using a blower as shown in Figure C-9.
Any flow volume reversed through the filter must be refiltered. Thus,
in addition to taking cloth out of the system for cleaning, this cleaning
method increases the total air flow in the remainder of the system. The
net increase in air/cloth ratio is normally 10% or less.20
Two types of pulse cleaning which are frequently used are the plenum
pulse system and the pulse jet system. The plenum pulse method attempts to
overcome some of the difficulties associated with other methods of cleaning.
In this kind of equipment a sharp pulse of compressed air is released in
the plenum chamber giving rise to some combination of shock, fabric defor-
mation and flow reversal. The result is the removal of the dust deposit
without more than a brief interruption of the filtering flow. The fabric
receives a minimum of flexural wear, and the filter installation is smaller
because the fabric is in use practically all the time.
The main distinction of pulsed equipment is the brief cleaning time,
typically around 1/10 of a second. The very low ratio of cleaning time to
filtering time makes pulsed equipment useful at high dust loadings.
The pulse jet cleaning method illustrated in Figure C-8 is essentially
similar to plenum pulse cleaning. The difference is that in pulse jet clean-
ing each bag is individually pulsed whereas in plenum pulse cleaning the
whole compartment of bags is pulsed via introduction of pulsing air in the
plenum chamber.
Fabric filters are constructed from several different media. Table C-l
lists various media that are used on ferrous foundry processes and shows
some of the properties of these media.
Because of the high temperature of the cupola emissions, fiberglass
Nomex, and Teflon are the media most frequently used. Reference 17 suggests
that if fluorspar is used in the charge material, fiberglass bags should
not be used as florides rapidly dissolve glass fibers. In cases where
fluorspar is used in the charge and high temperature filtration is desired,
Teflon and Teflon-coated bags can be used.
82
-------
OPTIONAL, TO AVOID
TEMPERATURE CHANGES
—I
PRESSURE
F: COMPARTMENTS FILTERING
R: COMPARTMENT BEING CLEANED BY DAMPERED
CONTROL FROM SUCTION SIDE OF SYSTEM
Figure C-9. Compartmentalized Reverse Air Cleaning.
20
83
-------
TABLE C-l. COMPARATIVE PROPERTIES OF MOST POPULAR
FILTER FIBERS.21
Fabric
Cotton
Wool
Nylon
Dynel®
Polypropylene
Orion®
Dacron®
Nomex®
Teflon®
Fiberglass
Type
Yarn
Staple
Staple
Filament
Spun
Filament
Spun
Filament
Spun
Spun
Filament
Spun
Filament
Filament
Spun
Filament
Spun
Bulked
Maxi-
mum
Temper-
ature
Fahren-
heit
180
215
225
160
200
250-
275
275+
400
400-
450
550
Acid
Resis-
tance
Poor
Very good
Poor
Good
Excellent
Good to
Excellent
Good
Poor to
Fair
Excellent
Fair to
Good
Fluoride
Resis-
tance
Poor
Poor to
Fair
Poor
Good
Poor
Poor to
Fair
Poor to
Fair
Good
Poor to
Fair
Poor
Alkali
Resis-
tance
Good
Poor
Excellent
Good
Excellent
Fair to
Good
Good
Excellent
Excellent
Fair to
Good
Flex
Abrasion
Resis-
tance
Very good
Fair
Excellent
Fair to
Good
Excellent
Good
Very
Good
Excellent
Fair
Fair
84
-------
The major design factor which affects the efficiency of the filter is
the air-to-cloth ratio, that is the ratio of gas entering the filter (cfm)
to the total surface area of the filtering fabric (ft2). The ratio chosen
is generally dependent on the particle size of the emissions with lower air-
to-cloth ratios used for emissions streams with fine particulate. Selection
of the ratio is based on past experience rather than on performance equa-
tions. For the cupola the air-to-cloth ratio is generally about 2:1 with
air flow based on the maximum volume reached during the burndown period.17
In reviewing permit applications, agency personnel should determine that
baghouses are designed for the maximum flow generated during burndown rather
than for average operating conditions.
In examining the effectiveness of fabric filters in controlling cupola
emissions, it is helpful to review fabric filter collection mechanisms.
Particulate are collected by fabric filters primarily through two mechanisms,
impaction and Brownian diffusion. Those particles greater than 1.4 pm in
diameter are collected at nearly 100% efficiency and particles smaller than
0.1 to 0.2 Mm in diameter are collected by diffusion mechanism. The filter
loses efficiency in the 0.2 to 1.4 range with a minimum efficiency of about
10% at 0.9 Mm22
Data in Appendix B indicate that most cupola emissions fall in the size
range for which the fabric filter is highly efficient. Thus, it is expected
that fabric filters should perform well on cupolas. Davis et al., indicate
that fabric filters will reduce effluent in the cupola exhaust stream to
0.01 gr/scf and that manufacturers will guarantee outlet concentrations of
0.02 gr/scf.13 Thus, fabric filters are an effective means for controlling
cupola emissions.
C.2 ELECTRIC ARC FURNACE CONTROLS
Unlike the cupola described earlier, the exhaust stream from the elec-
tric arc furnace (EAF) is not emitted through a well defined stack. Hence,
the emissions control system for an EAF must include a hood or exhaust system
to capture the gas stream. This "capture" mechanism is then connected by
ductwork to a particulate collection device, generally a fabric filter.
The sections below describe the capture systems and removal devices
used on electric arc furnaces. These sections briefly summarize the infor-
mation that has recently been compiled as background for a new source perfor-
mance standard for electric arc furnaces.23 The reader can find greater
details on these systems in that document.
C.2.1 Capture Systems
A control system for EAF emissions must capture emissions from three
distinct stages of the operation: melting and refining (including oxygen
lancing), charging, and tapping. The paragraphs below describe the capture
mechanisms that can be used for each of these stages.
85
-------
The capture mechanisms for melting emissions are roof hoods, side draft
hoods, and direct evacuation. These systems are illustrated in Figure C-10.
A roof hood is mounted on the furnace roof and draws gases through the
annuli between the electrodes and the openings in the roof. Extensions of
the hood may collect emissions from the pouring spout and slag door. Cap-
ture efficiency ranges from 95 to 100% for melting and refining emissions.
The side draft hood is also mounted on the furnace roof but one side
is open to permit movement of the electrodes. The hood collects furnace
gases from the electrode holes. A side draft hood is simpler than a roof
hood and places less weight directly on the furnace but it requires a larger
exhaust volume. The emission capture efficiency ranges from 90 to 100%.
Direct furnace evacuation requires an extra hole in the furnace roof.
Exhaust gases are withdrawn from beneath the furnace roof. Often this sys-
tem is not considered for use in small furnaces due to a lack of space for
the extra hole and since pressure variations in the furnace may be too rapid
for automatic control of dampers in the exhaust duct. In addition, direct
evacuation may act to weaken the roof refractory in small diameter furnaces.
The capture efficiency is similar to the side draft hood - 90 to 100%.
Most often electric arc furnaces are charged by removing the furnace
roof and dropping scrap from buckets above the furnace. Emissions from the
charging process often require emission control. Methods used for capture
of the charging emissions include canopy hoods, building evacuation, bay
evacuation, furnace enclosures, and close capture hoods.
Canopy hoods illustrated in Figure C-ll are most commonly used for the
collection of charging emissions at foundries. The hood is placed as close
above the furnace as possible but must allow clearance for overhead cranes
and access to furnace electrodes. The hood may only run during charging
and tapping stages or may run through the complete cycle, and must be phys-
ically large enough and draw through a large volume of air to insure effec-
tive capture of emissions. Impingement on overhead equipment and crossdrafts
in the shop can lower the collection efficiency. Devices such as curtain
walls and air curtains have been used to overcome crossdraft problems. The
capture efficiency of a canopy hood can be 80 to 90% with the lower figure
considered a more typical value when considering potential crosswinds.
Building evacuation (see Figure C-12) is a method of collecting various
emissions from the foundry. A very large volume of air is withdrawn to ob-
tain an emission capture range of 95 to 100%. It is possible to produce a
net increase in emissions to the atmosphere when considering the power re-
quirements of a building evacuation system.
Bay evacuation systems can produce an emission capture efficiency similar
to building evacuation. Each shop bay is separated from other bays by air
locks and/or soundproof doors, and each bay is .evacuated separately. In
the bay system, the problems with cross drafts found in the canopy hood and
building evacuation methods would be reduced.
86
-------
(a) Roof Hood
II
(b) Side Draft Hood
(c) Direct Shell Evacuation
Figure C-10. Capture Mechanisms for EAF Melting Emissions.
23
87
-------
Figure C-ll. Canopy Hood.23
88
-------
///////// ROorTop
/y///77//Ar+
PARTITION TO
SHIELD AREA
FROM CROSS-
DRAFTS
TO DUST
COLLECTOR
AIR INTAKE
OPENINGS
Figure C-12. Building Evacuation.2-^
89
-------
A metal shell surrounding the furnace and tapping area connected to a
take-off duct can provide effective collection of charging, melting, and
tapping emissions. A furnace enclosure system may require only 30 to 40%
of the air volume needed for an efficient canopy hood. Crossdrafts are not
a problem in the furnace enclosure. Capture efficiencies for the method
are not well documented. An example of a furnace enclosure is shown in
Figure C-13.
The close capture collection method shown in Figures C-14 and C-15
uses small capture devices close to the source. A rectangular hood that
completely surrounds the electrodes is used to evacuate melting and re-
fining emissions using minimum exhaust volumes. An annular ring hood swings
over the furnace top during charging to capture emissions and is rotated
back to the furnace side for storage during melting. Emissions are evac-
uated radially through slots in the ring. Separate hoods can also be placed
at the slag door and the tap spout. The close capture method reduces the
exhaust flow rate and requires less energy than other methods. However,
close capture may not be as efficient as other methods since charge and tap
hoods do not completely enclose emission sources.
Many of the previously mentioned methods for collecting charging emis-
sions also evacuate tapping emissions. In addition, a tapping pit enclosure
was designed in which metal is drained into a ladle in an enclosed pit.
Gas from the pit is exhausted to a control device. This system is shown in
Figure C-16. Hoods located over a tapping area is another method of emis-
sion capture.
C.2.2 Particulate Removal for EAF Systems
Virtually all EAF's in the United States use fabric filters as a par-
ticulate removal device. The basic design features and collection mechanisms
are identical to those described earlier for cupola systems. The paragraph
below describe the operating variables for EAF fabric filters.
As with cupolas, the EAF can operate with either a positive or negative
pressure unit. New baghouses installed on EAF's tend to be of the positive
pressure type because of lower capital costs and simple inspection procedures
for detecting damaged bags as well as lower fan noise. However, some nega-
tive pressure units are still used since these generally require less fan
maintenance and less operating horsepower than the positive pressure type.25
No data were obtained on the air-to-cloth ratio required for EAF fil-
ters. However, Szabo and Gerstle indicate that EAF's in the steel industry
have air-to-cloth ratios of 2.5:1 to 3:I.26 It is likely that emission char-
acteristics of foundry and steel EAF's are similar indicating that an air-
to-cloth ratio of 2.5 - 3:1 is reasonable for foundry EAF's.
Although most EAF particulate emissions are less than 20 [Jm in diameter,
fabric filters have been shown to effectively control these emissions. Emis-
sions test data indicate that properly designed and operated fabric filters
can achieve effluent concentrations of 0.007 gr/dscf. 6 Fabric filter ven-
dors also indicate that the above effluent limitation can be met.
90
-------
Roof
R°°f
/v/\
i
4
Concrete
Floor
~;
t
5'
I
i'
y
\
£—%f
0'
?oof Trusses i^
)
c- Top Charge Door
Front Charge Doort
r, 0
1 1 1 1 1
1 'jj '
(*•" ~x ^'-Alloy Addition Chute
! Furnace -7 ^s*
\ ' O
1
"••« "^ -J"""*- — Tcpping Exnausf Due?
ar Enclosure Door 1 |
Ladle |
-J—_ __U
1 ___!
5laq PT^ . _.'i . _'.j. . , _JV
t ^- v
SiDE VIEW
Fronf Charge Doors —j
Top Charge Door—7 //
Air Curtain F
Alloy Addition
Chute
r~~~?
Furnace
z:
_J
,
To Control Device
— Main Exhaust Due*
''Damper
Concrete Floor
Ladl
Topping Exhaust Duct
Shop Floor
FRONT VIEW
Figure C-13. Sketch of Furnace Enclosure Design at Lone Star Steel Company.
23
91
-------
HOOD EXHAUSTING
SLAG DOOR
ELECTRODE AREA
ENCLOSED WITH
RECTANGULAR HOOD
HOOD ENCLOSING
TAP SPOUT
(STATIONARY)
SWIVEL JOINT
*TO
BAGHOUSE
ANNULAR RING HOOD
SWINGS OVER
FURNACE TOP
DURING CHARGING
ANNULAR RING HOOD
IN PLACE TO COLLECT
CHARGING EMISSIONS
HOOD ENCLOSING
TAP SPOUT
TO
BAGHOUSE
Figure C-14. Hawley Close Capture Hoods.
23
92
-------
Figure C-15. Close Capture Hooding System for Electric
Arc Furnaces.
24
93
-------
To
Control
Device
Furnace
Figure C-16. ARMCO Incorporated Design for Tapping Pit Enclosure.
94
-------
One aspect of furnace operation which must be considered in the design
of the EAF control system is the use of oxygen lancing in steel foundry EAF's
to control the carbon content of the steel. The use of oxygen lancing has
been found to increase the temperature, volume, and dust content of the gases.
In addition, oxygen lancing creates significant amounts of CO (one furnace
had a measured CO content of 80% above the metal in the furnace) which are
not combusted in the furnace because of a lack of oxygen and elevated tem-
peratures.28
The impact of oxygen lancing on emission characteristics influences
the choice of control system parameters. First, if direct evacuation is
used, the system must be equipped with a gas combustion chamber to reduce
the potential for explosion in other parts of the control system. Care
should be taken in the design of this chamber to inhibit particulate fall-
out in the chamber. One plant experienced difficulty when elevation of the
exhaust gas temperature during lancing by 200°C to a maximum temperature of
1500°C caused the dust to become fritted into a hard mass in the chamber.29
The system must also be designed to handle the elevated temperature
and gas volume during lancing. As indicated above, the gas temperature at
one installation increased by 200°C from 1300°C to 1500°C during oxygen
lancing with an associated increase in volume. The gas cooling system and
baghouse should be designed to handle this temperature increase. It has
been determined empirically that provision for an incremental gas volume
during lancing of 15 times the oxygen lancing rate, measured at normal tem-
perature and pressure, and then corrected to the temperature and saturation
level appropriate to conditions at the control device inlet provides a good
estimate for design purposes.28
C.2.3 Pouring and Cooling Controls
The mold pouring and cooling area has proven to be one of the most
difficult areas to control in the ferrous foundry. The degree of control
which can be attained is dependent upon the type of casting produced and
the type of pouring process used by the foundry. In fact, for large cast-
ings that are poured in pit or floor molds, no known control mechanisms are
available. Control systems for other types of operation are described below.
As with the EAF, the pouring and cooling emissions stream must first
be captured and then the particulate must be removed in a particulate col-
lection device. An alternative control is the replacement of the sand molds
with a less polluting method. Currently three methods are available which
might be used to reduce emissions or capture the emissions stream. For new
foundries which produce a large number of the same type of casting, a per-
manent mold made of graphite or metal rather than sand can be used to reduce
emissions. A second alternative is the use of a stationary hooded pouring
station in conjunction with an enclosed cooling conveyor. The final alterna-
tive, which has been used to control toxic fumes in some nonferrous foundries,
is the use of a moveable hooding system attached to the ladle. This system
which has been used in nonferrous foundries is probably limited to relatively
small ladles and castings. It is also of limited effectiveness since smoke
may be emitted from the mold for as much as 30 min after the pour is com-
pleted.
95
-------
It should be noted that there is some question as to whether the par-
ticulate can be adequately removed from the emissions stream even if captured.
Given the fine particle size (95% < 5 pm) and low grain loading (about 0.04 gr/
scf) obtained by Bates and Scott on a pilot scale unit,36 it is quite likely
that even fabric filters will not be effective in controlling these emissions.
No foundries having particulate collection devices on the pouring and cooling
line have been identified in this study.
The sections below describe the three particulate reduction or gas
stream capture systems identified above.
C.2.3.1 Permanent Mold Casting—
The standard process for production of gray iron castings has used green
sand molds with sand cores bound by organic binders. For many years "per-
manent" or reusable molds have been used to produce small cast iron parts.
However, recent developments have made it possible to extend the process to
other high volume castings.30
The permanent mold-casting technique uses reusable molds of iron, steel
or graphite which are held together by a machine. The mold is coated with
an insulating material and cores are set into place. After the hot metal
is poured into the mold and allowed to solidify, the mold is opened. The
maximum time from the beginning of a pour until the mold is released is
about 3 rain.
Tests of emissions from 13-lb cast iron blocks produced from a per-
manent mold and green sand mold were reported in Reference 30. The results
of these tests are giyen in Table C-2. The permanent mold technique re-
sulted in a 99% reduction in particulate emissions and a 99% reduction in
hydrocarbon emissions. If capture of the remaining pollutant is deemed
necessary, the stationary mold machines are relatively easy to hood. The
technique has the additional advantage of reducing emissions in the sand
handling and core-making areas and virtually eliminating shakeout problems.
It should be noted that this method can be economically applied only
in those foundries producing adequate volume of identical castings. Ref-
erence 31 suggests that a minimum of 2,000 castings per mold is required to
make this system competitive with green sand molding. In addition, this
control method is more appropriate for new or significantly modified found-
ries than for foundries currently in operation with adequate process equip-
ment. Finally, it has been suggested that there may be size limitation in
the use of the process.
C.2.3.2 Mold Pouring Hood/Conveyor System-
One system that has been used to remove pouring and cooling emissions
from the foundry environment is a hooded pouring station followed by an en-
closed cooling conveyor. In these systems the molds are moved by conveyor
or on tracks to a push-pull hood such as that shown in Figure C-17. The
ladle is then moved to the pouring station and the metal is poured into the
molds and the emissions are exhausted into the conveyor. In some cases floor
fans are used behind the workers to negate possible effects of crossdrafts
and building ventilation systems. After pouring is completed, the molds
96
-------
TABLE C-2. COMPARISON OF EMISSIONS FROM GREEN SAND AND PERMANENT
MOLD PROCESSES FOR PRODUCING A 13-LB UNCORED
CASTING UNDER VENTILATED CONDITIONS.30
Green sand
(S:M ratio = 7:1)
Permanent mold
Time of emissions
Dust loading
Calculated total weight of
particulate evolved on a
one-casting basis
Maximum hydrocarbon concentration
Average hydrocarbon concentration
Maximum carbon monoxide concentration
Average carbon monoxide concentration
1 hr
0.04052 gr/scf
5.5 g
1,800 ppm
460 ppm
1,350 ppm
250 ppm
3 min
0.01017 gr/scf
0.15 g
125 ppm
100 ppm
100 ppm
> 50 ppm
97
-------
Ladle
Figure C-17. Iron Pouring Hood.
32
98
-------
and castings are moved through a cooling conveyor such as the one shown in
Figure C-18. Commercially available models of these systems are described
in References 32 and 33.
Reference 34 suggests a flow rate of about 150 to 175 cfm/linear foot
of hood, with slot velocities of 1,500 fpm for the pouring hood. Exhaust
takeoffs every 8 to 10 ft are recommended. The enclosed smoke hood for the
cooling conveyor will require about 75 to 100 cfm/linear foot of hooding
with a minimum flow of 200 fpm through all openings. Exhaust takeoffs
should be located on approximately 60-ft centers.
One manufacturer indicates that the velocity through control areas for
the pouring hood will generally be in the range, of 150 to 200 ft/min.35
The system has an air supply system to properly distribute flows across open
areas. Hood lengths range from 10 to 200 ft. Most pouring hoods are 50 to
70 ft long. Exhaust connections to the plenum are on approximately 20-ft
centers and supply connections are usually midway between the exhaust con-
nections .32
The effectiveness of these type systems in capturing pouring emissions
was verified by Envirex in a study of occupational health hazard control
technology for NIOSH.36 In examining a number of foundries for various types
of internal foundry controls, Envirex found two systems similar to the one
described above which adequately captured emissions. These two systems are
described below.
In one foundry a large plenum extending over the pouring line was com-
posed of two separate air sections. The upper section conveyed fresh air
which was aimed downward and outward to help contain the emissions. The
lower section provided exhaust drawing the emissions from the molds in the
air stream. The resultant air pattern took the form of a loop. This con-
trol method was very effective in keeping emissions out of the workers'
breathing zone. The supply air feed for the 60 ft pouring line was 52,000 cfm
and the exhaust was 78,000 cfm. The net air exhausted was somewhat higher
than the ACGIH recommended flow range of 12,000-18,000 cfm for a 60 ft con-
veyor line controlled by exhaust air only.3
In another foundry the pouring hood enclosed the entire mold conveyor
leaving only a 4.3 ft high opening for pouring the molds. A slotted takeoff
at the upper back of the hood provided 500 cfm/linear ft of exhaust along
the entire length of the pouring line at a slot velocity of 1500-2000 fpm.
This amount of exhaust was substantially higher than the ACGIH recommended
range of 200-300 cfm/linear ft for partially enclosed pouring stations.
Even though the exhaust flow was higher than recommended, the pouring hood
required the push air to capture all of the smoke from the burning molds.
Smoke was generated just below the front edge of the hood, and the rising
thermal draft needed to be deflected toward the back of the hood for com-
plete capture.37
The total exhaust flow from the pouring line of 23,800 cfm was cleaned
by a fabric collector before being discharged to the outdoors. No data could
be obtained on the operating parameters or effectiveness of the filter. The
99
-------
Figure C-18. Mold Cooling Conveyor Tunnel.
32
100
-------
untempered fresh air supply was distributed at a rate of 775 cfm/linear ft
from grilles located 11.2 ft above the foundry floor. The fresh air pro-
vided an average of 160 fpra of downdraft velocity at the front lower edge
of the hood, which helped to contain the billowing smoke within the hood.
However, when the molds passed a certain short section of the pouring line
where the fresh air was not distributed properly, the downdraft velocity
dropped below 100 fpm and smoke escaped.
C.2.3.3 Portable Exhaust Hoods--
For those pouring operations where a stationary pouring area is not
feasible, the most efficient solution appears to be a portable exhaust hood
attached to the pouring ladle (see Figure C-19). Reference 39 suggests that
this exhaust system can be used with either monorail or crane and can capture
emissions with a flow of 1,500 cfm/ladle. No evidence was compiled during
the study to suggest that such a hood had been applied in ferrous foundries.
Since many of the emissions from pouring are a result of combustion of the
organic binders in the mold which occurs for as much as 30 minutes after
pouring, it is likely that the hood will be of limited effectiveness.
C.3 SHAKEOUT AND SAND HANDLING EMISSION CONTROLS
All ferrous foundries that cast in sand molds potentially generate
significant quantities of fine dusts having a high silica content during
the handling of dry sand or sand/binder mixtures. The greatest potential
for emissions is the spent sand cycle, i.e. the handling, transfer, and sand
conditioning operations starting with the shakeout (or other casting removal
process) and ending with the muller. Since control for several individual
operations is often integrated into a single system, this section will dis-
cuss available technology for all parts of the spent sand cycle.
Information gathered from a review of the literature and from control
agency and foundry personnel, indicates that control systems are generally
well developed for shakeout and sand handling and that these sources present
no real control problem. However, since specific sand handling operations
vary considerably from foundry to foundry, it is not possible to define a
"best system" of emissions control for foundry sand handling. The discussion
below will focus on the controls that are available and will identify process
conditions that might limit the effectiveness of the systems.
The discussion is divided into three sections. These three sections
discuss capture or dust inhibition techniques for shakeout, sand handling,
and transfer and mulling operations. The final part of each section des-
cribes the particulate removal devices that are used to control emissions
from the spent sand cycle.
C.3.1 Shakeout
As described in Appendix A, the removal of the casting from the mold,
generally termed shakeout, can be accomplished in a number of different ways
including manual removal of sand by front end loaders and shovels, pneumat-
ically shaking the casting from the mold, and, most commonly, vibrating
101
-------
Crane
Ladle
Flexible
Hose
Hood
Mold
Figure C-19. Moveable Pouring Hood.
102
-------
screens. The mechanism and degree of emissions captured is primarily de-
pendent on the type of process used and on the size of the casting. For
large castings and nonmechanized foundries where the sand is removed manu-
ally from the casting little can be done to capture the emissions. On the
other hand several acceptable methods are available for capture of emissions
from the typical vibrating shakeout. These methods are described below.
If the size of castings and foundry operating practices permit, the
preferred method of capture is enclosure of the vibrating shakeout. Examples
of typical enclosures are shown in Figure C-20. The AFS control manual sug-
gests that at a minimum air flow should be 200 cfm/ft2 of grate area and
provide a face velocity of not less than 200 ft/min through the hood open-
ings.40 The American Conference of Governmental Industrial Hygienists
(ACGIH) recommends that for cool castings the flow should be at least 150
cfm/ ft2 of grate area with a face velocity of 200 fpm.41 For hot castings
the ACGIH recommends the same flow as AFS.4'1 The ACGIH also recommends that
10% of the flow should be exhausted through the hopper.
Enclosed shakeout operations were observed during two plant visits.
Both of these systems were operating well with no visible emissions from
the hood openings. One system that appeared particularly successful em-
ployed a series of rubber strips about 4 inches wide across each of the
openings to reduce the amount of open area. Rexnord in their examination
of one foundry shakeout found that although the actual flow was less than
ACGIH recommendation (2200 cfm vs 2430 cfm) the hood performed adequately.
Some problems were encountered when the hood was exposed to a cross-draft
from a nearby door.44 These observations and information from agency per-
sonnel indicate that a properly designed enclosure will capture nearly all
fugitive emissions from shakeout.
The experience during one plant visit is worthy of note. The foundry
had an enclosed shakeout that allowed no visible emissions. However, the
shakeout was fed by a vibrating conveyor that had no capture system. As
the molds and castings moved down the conveyor, much of the sand was shaken
away from the casting resulting in significant quantities of visible emis-
sions. Thus even though the shakeout was well controlled, the choice of
auxilliary equipment and lack of control of that equipment resulted in exces-
sive emissions. This situation highlights the need to design operations
and controls for the total handling process rather than for single opera-
tions .
Should the size of the casting or the method of handling not allow the
use of shakeout enclosures, side draft or double side draft hoods are recom-
mended. These options are shown in Figure C-21 and C-22 .
For a side draft hood serving a shakeout less than 6 ft wide, a flow
rate of 500 cfm/ft2 of shakeout grate is recommended. It is also recommended
that sufficient air be exhausted from the shakeout hopper to provide a down-
draft of 40 ft/min through the grate. For shakeouts greater than 6 ft wide,
hoods should be used on "two adjacent sides again with a total air flow of
500 cfm/ft2 of grate area. Flow rates should be increased if (a) the cast-
ings are quite hot, (b) sand to metal ratio is low; or (c) cross drafts are
high.40
103
-------
-0 HOOD
SUCTION
Molds in
here—
Mold
conveyor—
D
V
I
FIG. 25
D
Working openings,
keep as small as
possible.—7\
/ \
Shake-out -
-Castings
out here.
Figure C-20. Typical Shakeout Enclosures. 2>
104
-------
CONTROL OPENING" FRONT
OPENING -<-2 END OPENINGS
Figure C-21. Side Draft Hood.
45
105
-------
-Blank wall in fhis position is
almost as good as double hood.
Minimum clearance-
Ridgidly braced
DOUBLE SIDE-DRAFT
Proportions same as single side -draft hood except for overhang.
ShakeouJ grate 30
Plenum chamber and slots full
length of shakeout - in tunnel.
Figure C-22. Double Side Draft Hood.47
106
-------
Kane suggest that a properly operated side draft hood is about 90% effec-
tive in capturing shakeout emissions.46
For the double side draft hood air flows of 300 cfm/ft2 of grate are
suggested for cool castings and 400 cfm/ft2 of grate area for hot castings.
Again it is recommended that 10% of the volume be exhausted from the shake-
out hopper and that higher flows be used if the castings are quite hot,
sand to metal ratio is low or cross drafts are high.41 It can be assumed
that the efficiency of a double side draft hood is at least as high as a
side draft hood (i.e. if the hood is properly designed it should be least
90% efficient).
The emissions removal systems most frequently used on shakeout emis-
sions are low energy wet scrubbers (AP of about 10 in 1^0) and fabric fil-
ters. The basic systems are similar to those described for cupolas. The
paragraphs below describe available data on the performance of these sys-
tems .
Kane and Kearney indicate that the inlet loading from a properly operat-
ing side draft hood is 0.5 to 1.0 gr/scf.46*48 Kearney also indicates that
a typical outlet loading for a low energy scrubber is 0.01 gr/scf for shake-
out emissions. Data in Appendix C indicated that shakeout mass emissions
were reduced from 14 to 0.08 Ib/ton by a wet scrubber. Thus the scrubber
is about 98 to 99% efficient in controlling shakeout emissions.
No design data are available on the fabric filters used with shakeout
systems. However, based on limited foundry visits it appears that shaker
cleaning mechanisms are preferred for shakeout filters. Some industry per-
sonnel had indicated that moisture in the gas stream may cause problems with
filter operation. However no such problems were identified during plant
visits. No data are available on the effectiveness of fabric filters in
controlling shakeout exhaust. However it is estimated that, given the na-
ture of the shakeout emissions particulate loadings at least as low as 0.01
gr/scf can be attained.
C.3.2 Handling Emissions
After the sand enters the shakeout hopper it may undergo any of a num-
ber of handling, transfer, and conditioning steps before reaching the muller.
Table C-3 provides a partial listing of the types of the types of handling
operations that might be found at a foundry.
Each of these operations is a potential emissions problem. However
both industry personnel and control agencies indicate that the control sys-
tems for these operations are well developed and widely applied. Thus lit-
tle effort was expended in identifying exemplary capture systems for the
spent sand system. The interested reader can find descriptions of generally
accepted systems in References 40 and 41. The particulate removal devices
are the same as those for shakeout.
107
-------
TABLE C-3. TYPICAL OPERATIONS FOUND IN FOUNDRY SPENT
SAND SYSTEMS
Belt conveyor handling sand
Vibrating conveyor handling castings and molds
Vibrating conveyor from shakeout hopper
Bucket elevator
Gravity feed transfer chute
Magnetic separation conveyor
Bucket elevator
Revolving screen
Vibrating screen
Sand crusher
Sand reclaimer
Front end loader dump
Sand storage bin
Aerator
A alternative concept (U.S. Patent No. 3,461,941) has been developed
which has the potential for control of fugitive dust emissions from most
sand handling operations other than shakeout by reducing rather than cap-
turing emissions. The process is called the Schumacher Sand Process Sys-
tem. The normal sand-to-metal ratio in a green sand foundry is between 5
and 7:1. The Schumacher process utilizes a sand processed to metal ratio
of 20:1. This is the quantity of sand put through the muller. However,
the extra sand is not utilized to produce molds. It is diverted to an
inundator. Here the hot dry sand taken off the shakeout is mixed with the
moist sand from the muller to produce a moist cool sand. This sand is then
taken through the normal sand handling processes. However, the now moist
sand presents no emissions problems. Tests near transfer stations indicate
that dust concentrations are reduced by as much as 99% by application of
the system.49
The system requires little additional equipment (the inundator and a
small amount of additional mulling equipment) and is estimated to cost sub-
stantially less than equivalent collection systems. The system is claimed
to have the additional advantages of saving binder loss and producing cooler
sand for the molding line.
C.3.3 Mulling
Information gathered during the study indicates that mulling is not a
major emissions problem and that it is generally well controlled in most
foundries. Figures C-23 and C-24 show the emission capture systems that
are suggested by the ACGIH. The particulate removal devices generally as-
sociated with the systems are low energy wet scrubbers. No data were col-
lected on either the capture or removal efficiency of these systems. How-
ever, based on operations during plant visits and conversations with industry
and control agency personnel, both are expected to be quite high.
108
-------
Hood behind skip between—y i i
rails. Q=250LWcfm X7
To prevent condensation, insulation or strip -
heaters may be necessary or use
dilution fitting
Skip
Opening for skip
loading
0- ISO cfm/sq ft through all openings but not less than.
Mixer diam, feet
4
6
7
8
10
Exhaust, cfm
75O
9OO
I05O
1200
1575
\-~-Boffle
Muller
Figure C-23. Mixer and Muller Hood.
50
109
-------
Loading
ho,
Mutter
To prevent condensation,
insulation or strip heaters may
be necessary or use
dilution fitting.
Tight enclosure
Side hood or
booth
/-Enclosing hood
Bond hopper
\/-Low-velocity duct
used with cooling
type muller.
Cooling fan
blow-through
arrangement
Minimum exhaust volume
Location
Batch hopper
Bond hopper
Muller:
4't diameter
6' diameter
7'diameter
8'diameter
lO'diameter
Muller type
No cooling
Notel
600
Note 2
750
900
1050
1200
1575
Blow-thru
cooling
600
6OO
Note3
a
a
a
a
a
Draw -thru
cooling
Notel
600
Note 3
H
1
Figure C-24. Mixer and Muller Ventilation.
51
110
-------
C.4 CLEANING ROOM CONTROLS
Of all foundry operations, the cleaning room may be the most difficult
to control. The cleaning room generally comprises a large number of individ-
ual operations, each of which may be a source of particulate emissions.
Most of the operations are small sources of fugitive emissions. However,
together, these operations may be a significant source of emissions.
The discussion of cleaning room controls will focus on four areas:
(a) abrasive blasting and tumbling; (b) abrasive cutoff; (c) torch cutoff;
and (d) cutting and grinding. Much of the information presented in these
sections was obtained from an extensive NIOSH study of occupational health
hazard controls in a number of foundries.36 Unfortunately the data in this
study were obtained under strict confidentiality agreements and the found-
ries could not be contacted to obtain further data on the control systems.
Hence the data on the capture systems are quite good, but little information
is available on the removal devices.
C.4.1 Abrasive Blasting and Tumbling
For worker safety, it is necessary that abrasive blasting and tumbling
be tightly enclosed operations. Thus, if the process equipment is properly
sealed and if the ventilation system operates properly this is a stack rather
than a fugitive emission problem. Typical exhaust systems for a blast room
and tumble mill are found in Figures VS-101 and VS-113 respectively of the
ACGIH control manual.41 Envirex does indicate that best control can be main-
tained with preventative maintenance on shot-sand separators, coarse debris
screens, and chamber seals.52 They did not indicate that any fugitive emis-
sions problems were identified from properly operating systems.
The particulate removal devices most frequently used on these opera-
tions are low energy wet scrubbers and fabric filters.46-47 No data are
available on the design characteristics of either system. However, Kearney48
does indicate that the typical outlet grain loadings are 0.01 to 0.05 gr/scf
for scrubbers and 0.01 gr/scf for fabric filters compared to an inlet load-
ing of 0.5 to 5 gr/scf.
C.4.2 Abrasive Cutoff
Abrasive cutoff is the removal of metal, primarily gates, risers, and
sprues, from the casting using saws or abrasive wheels. Emissions of both
silica from burned on sand and metallic particles generally result.
Envirex suggests that when possible the castings should be precleaned
by tumbling or abrasive blasting before abrasive cutoff is used. This will
result in most of the silica and some of the appendages being removed in a
much more easily controlled operation. In addition, the close capture ven-
tilation system or grinder booth described below are suggested.52
In one of the foundries visited, commercially available close-fitting
hoods were attached to a mobile ductwork system that permitted exhaust from
swing grinding in any position of the wheel. Ductwork mobility was provided
by two mechanisms:
111
-------
1. A series of freely rotating duct elbows.
2. A carriage moving along an overhead exhaust plenum.
Movement of the swing grinder back and forth by the investigator was accom-
plished with some difficulty, although it was hard to determine how much of
the resistance was caused by the ductwork system and how much by the over-
head crane.54
The hood functioned as a receiving hood, positioned to receive the swarf
as well as the fine dust carried with it. Exhaust from each swing grinder
was measured at 2400 cfm. This was low when compared to most exhausted booths
that require 100-150 cfm/ft2 of booth opening. To capture all of the swarf
without interferring with the process, an adjustable bottom chute on the
hood allowed repositioning for the various sizes of grinding wheels used.
However, the adjustable bottom chutes on both machines were wired in fixed
positions. This non-adjusted position was correct for capturing the entire
grinding swarf for the small wheel at Station B but it was set too high and
only captured half of the swarf at Station A where a larger wheel was used.54
Typical booths which should be used for an abrasive cutoff saw and a
swing grinder can be found in Figures VS-401 and VS-414 of the ACGIH manual.41
The booth for the swing grinder is shown in Figure C-25. The saw booth is
similar.
As with abrasive cleaners, the particulate removal devices most fre-
quently used with abrasive cutoff are fabric filters and wet scrubbers.47
No data are available on the effectiveness of either device. However it
seems reasonable to assume that it would be similar to abrasive cleaning.
C.4.3. Chipping and Portable Grinding
Emissions from chipping and grinding with portable tools are the most
difficult of all cleaning room emissions to control because of both the num-
ber and mobility of the operations. During their study Envirex identified
four means of capturing these emissions: (a) downdraft benches; (b) high-
velocity, low-volume (HVLV) hoods; (c) retractable ventilated booths; and
(d) mobile extration hoods. However, each of these measures is limited both
in applicability and effectiveness. Only the first two systems were examined
in detail, and they are described below.
Envirex examined four foundries using downdraft benches (Cases A through D)
The design features of each of the systems as found in the NIOSH report is
presented below.56
Case A—A small bench was specially designed by foundry engineers
to be very rugged and compact. The workbench surface was con-
structed of replaceable wood slats, spaced to permit air flow.
Wood was used because it tended to deaden the noise. Easily re-
moveable dust trays permitted fast removal of grinding debris and
access for recovery of tools. Benches were baffled as far as
practical on the back and on the side facing the hopper which
112
-------
-Branch take -off at top or back. Central location
or multiple branches if several booths art used.
Additional adjoining
booths if needed.
45°slope
Booth encloses grinder
frame and suspension. —'
Grinder to operate in or
close to face opening.—
4' -6'- large opening - face
velocity -100 to ISO fpm -
never below IOO fpm
2'-O" - 2'-6* -small opening -
grinder in front - face velocity:
ZOO fpm
Minimum duct velocity =30OOfpm
Entry loss =0.5 VP
NOTE: Small local exhaust hoods mounted behind
grinder wheel may trap the stream of sparks,
out are usually not effective in control of
air-borne dust.
Figure C-25. Swing Grinder Booth.
55
113
-------
helped to direct air movement past the process and provide a shield
between adjacent workers. Exhaust flows through three of the four
benches evaluated were less than the ACGIH recommended flow range
(Table C-4).
TABLE C-4. DOWNDRAFT BENCH EXHAUSTS COMPARED TO RECOMMENDED FLOWS.
Bench area, ft2
Exhaust flow, cfm
A
4
430-680*
B
12
1730-2670*
C
24
50001
D
9.2
1150-1700*
ACGIH recommended
exhausts§, cfm 600-1000 1800-3000 3600-6000 1380-2300
§Reference 41, print no. VS-412.
* Flow measured in 4 benches.
t Flow measured in 1 bench.
Case B--A downdraft bench somewhat larger than in Case A also utilized
wood slats as a work surface. This bench was also custom designed
with the same features as Case A. An additional feature was a
distribution duct located under the bench top providing uniform
velocity distribution over the surface of the bench. Much of the
chipping and grinding was performed in the front section of the
bench nearest the operator, and thus it was important that adequate
downdraft be available in this section. A partial booth arrange-
ment was created by two vertical sides above each bench which re-
duced short-circuiting of air and shielded the grinding swarf be-
tween workers. In three of the four benches evaluated, flows were
within the ACGIH recommended flow range (Table C-4). The fourth
was slightly below the minimum recommended flow rate.
The exhaust air in Case A and B was cleaned by fabric filters
before being discharged outdoors.
Case C--This was a commercially available downdraft bench with a
built-in air cleaner section and fan. The metal grating used as
a work surface was large in area and the work height was low,
specifically to handle large castings. Workers could stand on
the bench and move around the casting as required. The bench was
too low, however, for grinding on smaller castings, causing the
operator to work in sitting and kneeling positions. A large rub-
ber mat on which to kneel was located in front of each bench.
The downdraft velocity through this bench was much higher in back
than in front of the bench where much of the work was performed.
The bench was baffled on three sides: in the back by the air
cleaner, and on the two sides by removable wooden baffles.
114
-------
The exhaust air from each bench was cleaned by a reusable viscous
metallic prefilter, followed by a disposable fiberglass after-
filter. Rated efficiency of the filtering system was 99.7% at
0.5 microns and larger. The air cleaning system was located in
the back of the bench and the exhaust discharge stack for each
bench recirculated the cleaned exhaust air into the plant air
space.
Because clogging of the air cleaning system on each bench could
cause reduction in flow and, therefore, reduction in capture ef-
ficiency of the bench, a manometer was used to monitor the dif-
ferential pressure across the filters, with a point on the scale
marked for disposable filter replacement. Since the equipment
had just been installed, required filter replacement frequency
had not yet been established. The downdraft air flow was well
within the ACGIH recommended flow at the bench measured (Table
C-4).
Case D--Like Case C, this was also a commercially available downdraft
bench with a built-in air cleaner section and fan, but this bench
was smaller, higher, and circular, permitting good access for work-
ing around a casting, as well as rotation without using a hoist.
The benches were located in semi-enclosing booths.
The installation had a history of operation and it had been found
that filters in the air cleaner section of the bench needed to be
replaced every two weeks. Filter changing had taken place four
days before the field survey, but downdraft velocity measurements
at the benches showed that some benches were drawing considerably
less air than others. In two of four benches measured, the flows
were at or below the minimum ACGIH recommended flow range (Table
C-4).
In general Envirex found that the benches worked well when the tool
was used on the outside of the casting, close to the bench with the swarf
directed toward the bench. However the effectiveness was limited when the
swarf was directed away from the bench or when the grinding was on an internal
part of the casting. The use of pneumatic tools also reduced capture effec-
tiveness. No quantitative estimate of the capture efficiency is available
for the operation.
Another foundry visited by Envirex used HVLV hoods to capture emissions
from chipping and grinding. Close capture hoods were retrofitted to some
of the portable grinding tools to capture fines generated by the grinding
process using the high velocity, low volume method. The exhaust system
functioned by creating an indraft velocity sufficiently in excess of the
dust generation velocity to capture dust as it was produced. High suction
pressure at hood slots could produce air velocities ranging from 6,000 to
39,000 fpm. A description of the hoods is presented in Table C-5.
The hoods were exhausted through long flexible hoses which were attached
to vacuum inlets located at each of the casting cleaning stations. The vacuum
manifold pipes joined and the entire flow was cleaned by a fabric collector
before being discharged outdoors.
115
-------
TABLE 05. DESCRIPTION OF HVLV CONTROLLED TOOLS, CASE E.
Tool name
Cup type surface
grinder
Abrasive cut-off
saw
Cone wheel
grinder
Diameter,
in.
6
9
1
Speed,
rpm
6,000
6,000
6,000
Hood description
Fitted hood wrapping 3/4 of the way
around the wheel.
Fitted hood enclosing all but work
area of saw.
Slot located on tool shaft adjacent
to wheel.
Radial grinder
6,000 or Adjustable extractor head located
12,000 adjacent to periphery of wheel.
* Reference 41, print no. VS-802.
Reference 41, print no. VS-801.
c Reference 41, print no. VS-804.
The HVLV system was designed for 10 vacuum points, but only five were being
used. Exhaust flow measurements for the individual tools were not made dur-
ing the survey but rather, system operating conditions were measured and
compared against design values.
Total system exhaust - actual
- design
Static pressure
actual
design
1770 cfm
1950 cfm
14 in. Hg
12 in. Hg
Inspector observations indicated that use of the hoods noticeably re-
duced dust emissions especially when working on flat surfaces. However,
because of the added difficulty in using the tools, workers are often op-
posed to their use and, in some cases, a particular job will not allow the
of hooded tools.
Based on the experiences of Envirex, conversations with foundry per-
sonnel, and comments from OSHA personnel, it appears that capture systems
for emissions from portable tools are not well developed. Available equip-
ment may provide control in some cases but generally acceptable methods have
not been developed. No data were found on the effectiveness of the partic-
ulate removal devices associated with these systems.
C.4.4 Torch Cutoff
The final emissions source that is examined is the use of an oxy-acetylene
torch to cut risers off large castings. 'This generally results in the emis-
sion of large quantities of fine iron oxide fumes. In general capture of
116
-------
these emissions is accomplished through the use of a large exhausted and
air supplied booth. Envirex examined the booth described below.
At one foundry a large booth (18 ft wide x 17 ft deep x 14 ft high)
was designed to evacuate the torching fumes (see Figure C-26). A design
exhaust flow of 20,000 cfm and a supply air flow of 16,000 cfm resulted in
a face velocity of 125 fpm at the booth entrance. The canvas curtain shown
in Figure C-26 effectively increased the velocity over the rest of the open-
ing as well as preventing the escape of fumes from the booth.
The inspector found that the booth worked quite well as long as the
torch was pointed directly toward the back of the booth. However if the
torch was positioned at an angle, the exhaust was not sufficient to overcome
the air nozzle blast. A roll back effect was observed, and some escaped
the booth.
No data are available on the particulate removal device used with this
system. However, given the amount of fine iron oxide in the emissions stream,
it is assumed that only a fabric filter will effectively collect these emis-
sions .
117
-------
EXHAUST TAKE-OFF
PLAN VIEW
(Cei1 ing not shown)
FRONT VIEW
SIDE VIEW
Figure C-26. Torch Cutoff Booth.
56
118
-------
APPENDIX C REFERENCES
1. Cupola Handbook. 4th Edition. American Foundrymen's Society. Des
Plaines, Illinois. 1976. pp. 102-117.
2. A. T. Kearney Company. Systems Analysis of Emissions and Emissions
Control in the Iron Foundry Industry. Volume I. Text, PB 198 348.
U.S. Environmental Protection Agency. February 1971. pp. VII - 9 -
VII - 62.
3. A. T. Kearney Company. Systems Analysis of Emissions and Emissions
Control in the Iron Foundry Industry. Volume II. Exhibits. PB 198
349. U.S. Environmental Protection Agency. February 1971. Exhibit
VII - 4.
4. Reference 3. Exhibit VII - 13.
5. Reference 2. pp. VII - 75.
6. Davis, J. A., E. E. Fletcher, R. L. Wenk, and A. R. Elsea. Final Re-
port on Screening Study on Cupolas and Electric Arc Furnaces in Gray
Iron Foundries. EPA Contract No. 68-01-0611. Task No. 8. (1975).
7. Reference 1. pp. 105.
8. Reference 1. pp. 106.
9. Control Techniques for Particulate Air Pollutants. National Air Pol-
lution Control Association. U.S. Department of Health Education and
Welfare. Washington, D.C. 1969. pp. 4-46.
10. Reference 1. pp. 108.
11. Reference 9. pp. 4-47.
12. Reference 1. pp. 109, 110.
13. Reference 6. pp. IV - 23.
14. Reference 3. Exhibit VII - 15.
15. Reference 1. pp. 113.
16. Reigel, S. A. and L. Rheinfrank Jr. "Cooling Hot Gases." Pollution
Engineering. November/December 1970. pp. 32-34.
119
-------
17. Reference 1. pp. 114.,
18. Szabo, M. F. and R. W. Gerstle. Operation and Maintenance of Partic-
ulate Control Devices on Selected Steel and Ferroalloy Processes.
EPA-600/2-78-037. U.S. Environmental Protection Agency. Washington,
B.C. March 1978. pp. 3-45 and 3-46.
19. Reference 6. pp. IV-6
20. Reference 18. pp. 2-101 thorugh 2-104.
21. Reference 1. pp. 111.
22. Reference 6. pp. IV-7.
23. Electric Arc Furnaces in Ferrous Foundries - Background Information
for Proposed Standards - Draft EIS. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. April 1980. pp. 4-1 through
4-52.
24. Wallace, D. and C. Cowherd, Jr. Fugitive Emissions from Iron Foundries.
EPA 600/7-79-195. U.S. Environmental Protection Agency. Research
Triangle Park, N.C. August 1979. pp. 62.
25. Reference 23. pp. 4-44 and 4-45.
26. Reference 18. pp. 2-95.
27. Reference 23. pp. 4-47.
28. Davies, E. and W. T. Crosly, The Control of Fume from Arc Furnaces,
Special Report No. 85, the Iron and Steel Institute, London, 1964, pp.
133-143.
29. Scriven, D. Oxygen-Assisted Electric Air Furnace Operating Ironmaking
and Steelmaking (Quarterly) (4), 1974, pp. 193-200.
30. Bates, C. E., and W. D. Scott. Better Foundry Hygiene through Permanent
Mold Casting. Southern Research Institute. January 30, 1976.
31. Bates, C. E. Profit Potential in Permanent Mold Iron Castings. Foundry,
November 1972.
32. Dust, Fume and Smoke Hoods for Shakeouts, Pouring Stations, Mold Con-
veyors. Bulletin 574, Schneible Company. Holly, Michigan.
33. Dust and Fume Control Systems. Catalog 12745-WG. Kirk and Blum Manu-
facturing Company.
34. Design of Sand Handling and Ventilation Systems. American Foundrymen's
Society. 1972.
120
-------
35. Personal communication with Mr. A. S. Lundy, Schneille Company.
*-
36. Enivrex, A Rexnord Company. An Evaluation of Occupational Health Hazard
Control Technology for the Foundry Industry. National Institute for
Occupational Safety and Health. Cincinnati, Ohio. October 1978.
37. Reference 36. pp. 332.
38. Reference 36. pp. 348.
39. Melting and Pouring Operations. American Foundrymen's Society. 1972.
40. AFS Foundry Environmental Control. No. 5. Design of Sand Handling
Ventilation Systems. American Foundrymen's Society. Des Plaines,
Illinois. 1972. pp. 5-5.
41. Committee on Industrial Ventilation. "Industrial Ventilation, A Manual
of Recommended Practice." 16th Edition. American Conference of Govern-
mental Industrial Hygienists. Lansing, Michigan. 1980. pp. 5-15.
42. Reference 40. pp. 5-12.
43. Reference 41. pp. 5-14.
44. Reference 36. pp. 305, 308.
45. Reference 40. pp. 5-11.
46. Kane, J. M. "Air Pollution Ordinances." Foundry. October 1952.
47. Reference 41, pp. 5-15.
48. Reference 3, Exhibit VII-21.
49. Reference 24. pp. 56.
50. Reference 41. pp. 5-10.
51. Reference 41. pp. 5-11.
52. Reference 36. pp. 104.
53. Reference 36. pp. 98-99.
54. Reference 36. pp. 187.
55. Reference 41. pp. 5-48.
56. Reference 36. pp. 158-165.
121
-------
APPENDIX D
OPERATION AND MAINTENANCE OF CONTROL EQUIPMENT
123
-------
One of the major contributors to foundry noncompliance is the malfunc-
tion of air pollution control equipment. The results of this study indicate
that these incidences of malfunction can be significantly reduced by the
proper operation and maintenance of control equipment, particularly those
control devices on the melting furnaces. This appendix outlines general
operation and maintenance procedures for venturi scrubbers and fabric fil-
ters. The discussion relies heavily on data compiled for an earlier EPA
study of operation and maintenance procedures for iron and steel control
devices.1 While these procedures were not developed specifically for foundry
cupolas and electric arc furnaces, the similarity of the emissions charac-
teristics for foundry and iron and steel processes makes the analysis quite
applicable.
D.1 Operation and Maintenance of Venturi Scrubbers
The typical scrubber system associated with ferrous foundry cupolas
consists of a gas prequencher to reduce the temperature of the cupola ex-
haust, a flooded disc or fixed throat venturi scrubber, a mist eliminator
with sump, recirculation pumps, and an induced draft fan. Each of these
components can be a source of malfunctions; however, the main problems iden-
tified during the study were fan bearing and wiring failure, feedwater nozzle
plugging, and corrosion and erosion of the venturi throat and mist eliminator.
It appears that proper operation and maintenance of the scrubber can reduce
the occurrence of these problems. The sections below describe typical oper-
ational procedures that can be used during start-up, normal operation and
shut down and some routine maintenance procedures that can be used to improve
equipment performance.
D.I.I Operating Procedures--
D.1.1.1 Preoperation--Before start-up, all major items of equipment,
connecting pipes and auxiliaries must be inspected, cleaned, and tested.
In newly installed systems, the first step should be an air test of fans
and duct, and a hydraulic test of pipes and valves to check for leaks and
instabilities. A water test of the system should also be carried out to
ensure that equipment, instruments, and control/safety systems are working
properly. The items which should be checked during preoperational tests on
a flooded disc venturi scrubber are summarized below:
0 FD/ID Fan
Electrical controls
Fan bearing coolant system
Alignment
Lubrication
Vibration sensors
Bearing temperature sensors
Belt tension, pump rotation, pump alignment, lubrication, seal water,
packing, pressure gauge, suction and discharge valves, motor bearing
temperature, hydraulic system (for flooded disc control pump).
124
-------
° Control Systems
Flue gas bypass
Flooded disc pressure drop
Makeup water rate
Recirculation sump level
Slurry density
Slurry purge rate
0 Safety System (interlocks and alarms)
High flue gas pressure
Low level in sump
High and low density
0 Utilities
Electric power
Instrumentation air
Process water
Process return water
D.I.1.2 Start-up—To start a system for operation or for a water test,
the procedure described in the designers' operating manual is generally used.
Typical steps for the start-up of a new venturi scrubber installation are
outline below:
1. Close all drain valves.
2. Turn on the circuit breakers for all instruments and electric valves.
3. Set all monitoring instruments to zero reading.
4. Start the service water system and raise the water level in the
sump to the defined level.
5. Turn on the recycle pump circuit breaker, and start the operating
and standby pumps.
6. Turn on the circuit breaker for the disc control pump, start the
disc control pump, and adjust the high and low limits of the pres-
sure drop indicator.
7. Close the flue gas bypass dampers and start the fan.
8. Check the scrubber pressure controller and the system monitoring
instruments.
D.I.1.3 Shutdown--A general procedure for a planned shutdown of a
flooded disc scrubber system is outline below.
125
-------
1. Turn the flue gas damper to the bypass position and stop the fan.
2. Close the makeup water and slurry couple valves.
3. Stop the recycle pumps (both operating and standby).
4. Open the drain valves at the slurry pumping lines and flush the
lines, gauges, and pumps with water.
5. Stop the disc control pump and leave the disc in the fully raised
position.
6. Open the drain line on the pressure gauge to the throat and disc
and allow .the line to drain.
D.I. 1.4 Normal Operation—Under normal operating conditions, all con-
trol variables should be operated in the defined ranges. These control var-
iables include the scrubber pressure drop, recycle pump rate, makeup water
rate, slurry density, slurry purge rate, and recirculation sump level.
An abnormal condition is indicated by an alarm. If the problem cannot
be corrected by the operator, under certain circumstances an interlock will
open the flue gas bypass damper and shut down the scrubber system.
Alarm conditions involved in the system are outlined below:
1. Scrubber pressure drop —
Alarm condition may be a result of a malfunctioning pressure drop
controller, failure of the disc control pump, a jammed disc, or a
rapid change of the boiler load.
2. Slurry density —
An alarm condition may occur because of a malfunctioning control,
a defect in the density control valve, a malfunction in the sump
level control, or a makeup water rate change.
3. Recirculation sump level --
An alarm condition may be due to a malfunctioning control or exces-
sive or insufficient slurry in the sump.
4. Others —
An alarm condition may be caused by plugged lines, closed valves,
pump trouble, or fan trouble.
D.I.2 Inspection and Maintenance During Normal Operation--
Many items checked before operation should be inspected during routine
maintenance, which generally includes unplugging lines, nozzles, pumps, etc.;
replacement of worn equipment parts, erosion/corrosion prevention liners,
126
-------
and instruments (level indicators, density indicators, etc.); and repairing
damaged components (when practical from the standpoint of labor and mate-
rials) .
Table D-l indicates the manpower requirements for maintenance due to
scaling and plugging for both the wet approach and liquid injection venturi
scrubbers.
TABLE D-l. MAINTENANCE FOR PLUGGING AND SCALING VENTURI SCRUBBER2
(From interview with P. Wechselblatt, Chemico)
Type of
venturi
scrubber
Wet
approach
Liquid
injection
Type of
Plugging
Mechanical
cleaners
1 man/shift/
mo
1 man/shift/
mo
Cylinder
cleaners
1 man/ shift/
mo
1 man/shift/
mo
problem
Scaling
Chemical
cleaning
3 men/shift/
wk
3 men/shift/
wk
Hand
cleaning
1 man/shift/
wk
1 man/shift/
wk
Table D-2 lists maintenance requirements for two ranges of pressures,
and various lining materials and gas characteristics. This table should be
useful in the selection of scrubber liners or venturi units for the various
iron and steel applications, including iron foundry cupolas and sand system
scrubbers.
The following check list is based on problems encountered in scrubber
operation. These should be checked routinely and corrected according to
the manufacturers' recommended procedures.
Check the scrubber disc to ensure even water distribution across its
surface.
Check erosion and corrosion of all scrubber internal surfaces, especially
corrosion underneath scale buildup. Repair as necessary.
Clean and descale all scrubber internal surfaces. While descaling,
exercise care to prevent damage to the linings.
Check the disc operation and perform maintenance on the hydraulic pack-
ing.
Check the nozzles for buildup and/or damage. Repair or replacement
may be necessary.
Check for solids buildup in blowdown lines. Cleaning may be effected
without system shutdown, and a flush connection may be installed to
prevent this condition in the future.
127
-------
TABLE D-2. SCRUBBER MAINTENANCE
(From Interview with P. Wechselblatt, Chemico)
N5
oo
Pressure droj^
> 30" AP
Type
of liner
Ceramic
Silicon
carbide
Cement
Rubber
Rubber
Plastic
Steels
Carbon
316
304
Inconel
625
Hastoloy
Life
cycle,
years
3-4
1
1
5
2-6
2-6
2-6
2-6
2-6
Repair
time
2 men/wk
2 men/wk
2 men/2 wk
2 men/ 2 wk
Patchable
Patchable
Patchable
Patchable
Patchable
< 30'
Life
cycle,
years
10
4
5
10
Indefin-
ite
6
6
6
6
6
" AP
Repair
time
2 men/wk
2 men/wk
2 men/2 wk
2 men/2 wk
1 day
Patchable
Patchable
Patchable
Patchable
Patchable
Gas
Corrosive
Poor
Poor
Excellent
Excellent
Poor
Excellent
(arid)
Good
(arid)
Good
Excellent
characteristics
Corrosive
Abrasive and abrasive
Excellent Good (mildly
corrosive)
Poor Good (mildly
corrosive)
Good Good
Good Good
Fair Fair
Fair Fair-good
Fair Good
Good Good
Good Good
Comments
For cutting type
particles for ero-
sive but not sharp
particles. ' •
Patchable lining.
Good for chlorides.
Not good on
chlorides.
Except for S03
and Cl".
-------
Check for corrosion, erosion, and leaks in lines where protective liners
may have deteriorated. Replace liners as required.
Check operation of mist eliminator. Formation of droplets can be caused
by excessive gas flow rate, plugged drains from the moisture eliminator,
or condensation in the outlet duct.
Check pumps for wear, seal water, packing, and smooth operation.
Check dampers and damper linkages for proper positioning and wear.
Fan check should include lubrication, fan bearing coolant, belt wear,
and belt tension, and impeller erosion/corrosion.
Inspect all interior surfaces and condition of mist eliminator and sump
during major outages.
Exterior inspection should include a check for leaks in all process
and control lines, ductwork, and expansion joints.
Note the condition of all instruments, e.g., level probes and density
probes with regard to solids buildup. It is impractical and usually
impossible to remove solids buildup on the probes. In many cases the
probes must be replaced.
Perform a final check for proper operation of density sensors, pres-
sure drop control, and level elements.
Spare Parts - The minimum inventory is one of each part for each venturi
scrubber. The inventory for a venturi system is given in Table D-3.2
Manpower Requirements - The preceding discussion has given an indica-
tion of maintenance items, maintenance times, and spare parts inventory for
a venturi scrubber system. Table D-4 completes this picture by presenting
the types of personnel generally required to perform maintenance on various
parts of the venturi scrubber system.2
D.2 OPERATION AND MAINTENANCE OF FABRIC FILTERS
The typical fabric filter control system applied to a ferrous foundry
cupola consists of a cooling mechanism, usually a prequencher, to cool the
gas stream to about 450°F, the fabric filter (or baghouse) including its
cleaning mechanism, a fan which may be either upstream or downstream from
the baghouse, and a dust removal system to handle the captured dust. As
with the scrubber system, each of the components of the system is subject
to breakdowns which can lead to a malfunction of the entire system and ex-
cessive emissions from the cupola. Proper operation and maintenance of the
system will reduce the frequency of the malfunctions to low levels (in some
cases 1 to 2% of the operating schedule).
129
-------
TABLE D-3. SPARE PARTS INVENTORY FOR VENTURI SCRUBBER2
(From interview with P. Wechselblatt, Chemico)
OJ
o
Type of parts
Section of
the
system Motors
Scrubber
Separator
Fan X
Pump(s) X
Mist eliminator
' Mist
elimi-
nator
modules Seals Bearings
X X
X X
X
Reamers
(50% of Packing
Impeller total) material
X X
X
X
Adjust-
able
throat-
damper None
X
-------
TABLE D-4. TYPE OF MAINTENANCE REQUIRED - VENTURI SCRUBBER SYSTEMS2
(From interview with P. Wechselbatt, Chemico)
Section of
the system
Scrubber
Separator
Laborer
X
X
Type of worker
Wastewater
treatment
Electrical Plumber operator
Mechanical
Fan X
Pump XX X
Piping, valves X
Water treatment XXX X
equipment
131
-------
This section will consider all aspects of the system but the cooling
mechanism. The section is divided into four parts: (a) preoperational
checks; (b) startup; (c) shutdown; and (d) maintenance during normal opera-
tion. Again, much of the material in the following sections is excerpted
from Szabo and Gerstle1 with slight modification to make it more applicable
to foundries.
D.2.1 Preoperational Checks—
The following checks are recommended prior to start-up:
0 Test control air lines (hydrostatically).
0 Check air dryers that supply control air to the bag filters.
0 Check dust removal system.
0 Inspect collapse air fans for alignment and rotation.
0 Check seals at gas inlet, collapse air, and gas outlet damper.
0 Check baghouse compartments, remove debris.
0 Check filter bags for proper installation and tension, including
a check of proper bag connection.
0 Check and sweep thimble floors clean. Dust buildup on floor dur-
ing operation is a positive indication of a broken bag.
0 Calibrate pressure drop recorders and transmitters.
0 Check pressure taps for leakage.
D.2.2 Start-up—
The operation of a fabric filter system is automatic. However, start-up
and shutdown are extremely critical.
When the new equipment is started for the first time, the fan should
be checked for correct direction of rotation and speed. The ducting, col-
lector housing, etc., should be checked for leaks. Gas flows and pressures
should be checked against the design specifications. Instruments should
then be checked for correct reading and calibration adjustments made as
necessary. Control mechanisms, and especially all fail-safe devices, should
be checked for operability.3
At the first start-up of the system, and also whenever new bags have
been installed by the maintenance crew, the bags should be checked after a
few hours of operation for correct tension, leaks, and expected pressure
differential. Initial temperature changes or stress induced during the
cleaning cycle can pull loose or burst a bag. It is wise to record at
least the basic instrument readings during this start-up period on new bags,
for ready reference and comparison during later start-ups.3
132
-------
During any start-up, transients in the dust generating process and surges
to the filter house are probable and ought to be anticipated. Unexpected
temperature, pressure, or moisture has often badly damaged a new installa-
tion. In particular, running almost any indoor air or combustion gases into
a cold filter can cause condensation on the walls of the baghouse and on
the bags leading to blinding and corrosion. Condensation in the filterhouse
may void the manufacturer's guarantee. It can be avoided by preheating the
filter or the gas.3 Since most cupolas and EAF's operate on an intermittent
basis, it may be necessary to heat the gases above dewpoint in a bypass mode
during each start-up. The filter can then be brought on line.
It is particularly important to bring the temperature of cupola gases
up before the filter is on line as the sulfates created by the coke condense
in the range of 225 to 325°F and are extremely corrosive on the bags and
baghouse structure. Agency persons contacted during the study indicated
that this practice often results in excessive visible emissions during the
firing of the coke bed. No solutions to this problem were identified dur-
ing the study.
A typical sequenced start-up procedure for a large continuous automatic
multicompartment fabric filter using either reverse air, shake, or combina-
tion cleaning is:
1. Check to see that all system monitoring instruments are reading
zero; especially fan motor ammeters and compartment pressure
manometers.
2. Close all system dampers except tempering air damper (if used).
This includes main compartment isolation dampers, reverse air
dampers (if used), and fan modulation dampers.
3. Start material handling system including any motorized airlock
devices and screw conveyors. Hoppers should be empty on start-up.
4. Sequentially start main fans allowing each to come to speed before
starting next fan.
5. Start separate reverse air fan if used and allow to come to speed.
6. Engage fan modulating damper circuit(s).
7. Engage tempering air damper circuit (if used).
8. Slowly open main compartment isolation dampers. If dampers are
opened too quickly bags will pop open, ultimately resulting in
failure.
9. Engage compartment cleaning circuit.
10. Check normalcy of readings on system monitoring instruments;
especially fan motor ammeters and compartment pressure manometers.
133
-------
D.2.3 Shutdown--
The main precaution in shutting down the filter system is prevention
of moisture in the filterhouse. Condensation can occur due to cooling of
gases containing moisture, particularly combustion gases, if they are not
completely purged from the filter system and replaced with drier air before
the filter cools down. This can also happen with air at ambient moisture
levels if the filter is in a colder location. To prevent condensation, the
systems should be purged carefully on shutdown and then sealed off completely.
Alternatively, a flow of warm air can be continued through the filter during
the shutdown, which also helps prevent condensation when it is started up
again. A shutdown procedure is summarized below:
1. After the process has been stopped and emissions have ceased, allow
baghouse to track through one complete cleaning cycle; this will
purge system of process gas and collected dust.
2. Stop main fans.
3. Stop separate reverse air fan if used.
4. Allow material removal system to operate for 1 hour or until sys-
tem is purged of collected material. This is imperative for a
fabric filter on a shakeout as the combination of moisture and
binders may result in bag blinding or hopper plugging if the sys-
tem is not cleaned prior to shutdown.
D.2.4 Maintenance During Normal Operation--
Maintenance of fabric filters in iron foundries centers around the bags
and the moving mechanical parts in the hostile interior of the baghouse (i.e,
dampers, screw conveyors, and shaker linkages). The same maintenance proce-
dures can be applied to baghouses operating on electric arc furnaces or
cupolas in ferrous foundries. Table D-5 presents a checklist of items that
require regular inspection.
Plant personnel must learn to recognize the symptoms that indicate po-
tential problems in their fabric filter, determine the cause of the problem
and remedy it, either by in-plant action or by contact with the manufacturer
or other outside resource.
For example, high pressure drop across the system is one symptom for
which there could be many causes, e.g., difficulties with the bag cleaning
mechanism, low compressed-air pressure, weak shaking action, or loose bag-
tension. Many other factors can cause excessive pressure drop, and several
options are usually available for corrective action appropriate to each cause.
Thus the ability to locate and correct malfunctioning baghouse components
is important and requires a thorough understanding of the system. A detailed
list of trouble-shooting and corrective measures is given in Appendix E.
Table D-6 presents the frequency of failure of basic fabric filter parts,
including the frequency of inspection and inspection time, as well as the
time required for repairs.
134
-------
TABLE D-5. CHECKLIST FOR ROUTINE INSPECTION OF BAGHOUSE
Component
Check for:
Shaker mechanism (S)
Bags
Magnehelic gauge or
manometer
Dust removal system
Baghouse structure
(housing, hopper)
Ductwork
Solenoids, pulsing valves
(RP)
Compressed air system
(RP, PP)
Fans
Damper valves (S, PP, RF)
Doors
Baffle plate
Proper operation without binding;
loose or worn bearings, mountings,
drive components; proper lubrica-
tion.
Worn, abraded, damaged bags; con-
densation on bags; improper bag
tension (S) (RF; loose, damaged
or improper bag connections.
Steadiness of pressure drop
(should be read at least daily).
Worn bearings, loose mountings,
deformed parts, worn or loose
drive mechanism, proper lubrica-
tion.
Loose bolts, cracks in welds;
cracked, chipped, or worn paint;
corrosion.
Corrosion, holes, external damage,
loose bolts, cracked welds, dust
buildup.
Proper operation (audible com-
pressed air blast).
See above; proper lubrication
of compressor; leaks in headers,
piping.
Proper mounting, proper lubrica-
tion of compressor; leaks in
headers, piping, balance.
Proper operation and synchroniza-
tion; leaking cylinders, bad air
connections, proper lubrication,
damaged seals.
Worn, loose, damaged, or missing
seals; proper tight closing.
Abrasion, excessive wear.
RP-reverse pulse; PP-plenum pulse; S-shaker; RF-reverse flow.
135
-------
TABLE D-6. BAGHOUSE COLLECTOR MAINTENANCE2
Item
Frequency of
breakdown
Frequency of
inspection
Time required
to perform inspection Time to repair
Type of
person
to repair Comments
INSIDE BAG COLLECTION
Bags
5 in. <)> 14 ft Monthly Monthly
8 in. ij> 22 ft Monthly Monthly
12 in. 30 ft Monthly Monthly
Door seals 2-4 yr Monthly
Cleaning mechanism
Shaker 6 mo Monthly
Reverse air 2 yr Monthly
£J Dust removal system
°^ Screw conveyors 1-2 yr 6 me
Air locks 1-2 yr 6 mo
Pneumatic 2-3 yr 6 mo
Baffle plate 4 yr 1 yr
Damper valves 2-3 yr Monthly
OUTSIDE BAG COLLECTION
All bags Monthly Monthly
Cleaning mechanism
Pulse jet 2 yr Monthly
Pulsing plenum 2 yr Monthly
Door seals 2-4 yr Monthly
Dust removal system
Screw conveyors 1-2 yr 6 mo
1.5-3 man-h/100 bags3 10-30 min/bagb
2-4 man-h/100 bags3 15-45 min/bag^
2.5-5 man-h/100 bags3 20-60 min/bag
5 min/door
5 rain/row
15 min
1 hour
30 min
1 hour
30 min
15 min/valve
1 hour/door
30 min/row
2 hours
2-4 hours
1-2 hours
8 hours
8 hours
1-24 hours
2 min/row
2 min/row
5 min/door
1 hour
(concluded)
30 min
1 hours
1 hour/door
2-4 hours
Laborer
Laborer
Laborer
Laborer
Maint. man
Maint. man
Maint. man
Maint. man
Maint. man
Maint. man
Maint. man
0.6 man-hour/100 bags 6-10 min/bag Laborer
Maint. man
Maint. man
Laborer
Maint. man
Complete
replacement
2 years same
Same
Assume top
bag removal
-------
TABLE D-6. (concluded)
Item
Air locks
Pneumatic
Baffle plates
Damper valves
Frequency of
breakdown
1-2 yr
2-3 yr
4 yr
2-3 yr
Type of
Frequency of Time required person
inspection to perform inspection Time to repair to repair Comments
6 mo
6 mo
1 yr
Monthly
30 rain
1 hour
30 min
15 min/valve
1-2 hours Maint. man
8 hours Maint. man
8 hours Maint. man
1-24 hours Maint. man
a .
, of fluorescent particles and black light are used).
Low value is total changeout/bag and high value is individual bag change.
Three-man crew minimum.
-------
Following is a discussion of major fabric filter components requiring
routine maintenance.
D.2.4.1 Inlet Ducting—Common problems such as abrasion, corrosion,
sticking or plugging of dust, and settling must be dealt with on a routine
basis. Abrasion can be reduced by using special materials at bends in duct-
ing. Corrosion can be minimized by supplying insulation, especially in the
long duct runs, which are most susceptible to moisture condensation. Regu-
lar inspection will help control plugging and settling problems in ducts.
D.2.4.2 Blast Gate and Flow Control—Problems with flow control equip-
ment are reported frequently.3The blast gate valve is especially vulnerable
and should be checked periodically and adjusted. Filter compartment inlet
dampers are a high-maintenance item, and spare parts should be stocked.3 A
bad damper seal can shorten the life of bags in a shake-type system, and
caking bags, if not replaced, can foul valves on the clean side of the bag-
house and cause them to malfunction. The most popular dampers for compart-
ment isolation are air cylinder-operated poppets acting vertically (see
Figure D-l). Several users mentioned problems with push rod guides when
dampers were made to act horizontally. Maintenance on these dampers con-
sists of periodic inspection and replacement of packing and solenoids. The
wafer and seat were not indicated as presenting severe maintenance problems.
Damper failures can sometimes be detected by observation of- a differential
pressure chart. As the dampers open and close, the differential pressure
swings. If a damper fails, the absence of this pressure swing leaves a
"gap" on the differential pressure chart. This is one reason for ensuring
proper operation of the pressure monitors. If a high differential pressure
is signaled, the dampers are routinely checked for proper operation. If
not, the operator must observe damper operation through the complete cycle
directly at the baghouse.
D.2.4.3 Fans—Fans and blowers are reported to be a large problem area,
particularly those located on the dirty side of the baghouse where material
can accumulate on the vanes and cause them to get out of balance.3 Corro-
sion and abrasion can also cause problems. Condensation and corrosion in
the fan may be alleviated with duct and fan insulation.3 Most fan housings
can be drained, and the drains should be checked on a regular basis.
Air flow and fan speed should be measured periodically and belt condi-
tion and tension determined; the fan should also be checked for direction
of rotation. These checks can be combined with routine lubrication proce-
dures .
D.2.4.4 Entrance Baffles—Baffles may be added to improve distribution
of the gas to each compartment and bag. They should be adjustable, however.
They may cause problems by accumulating dust or abrading too rapidly.
D.2.4.5 Hopper—Hoppers are a common problem in any fabric filter sys-
tem. Dust flow can be facilitated by the use of vibrators and/or heaters
(if they work properly); by lining the hoppers with anti-friction material;
by the use of air-pulsed rubber-lined hoppers; by placing poke holes in the
side of the hoppers; or by insulation if condensation is a problem.
138
-------
WAFER
SEAT
PUSH ROD
AIR CYLINDER
OPERATOR
Figure D-l. Poppette Valve.'
.139
-------
Trough-type hoppers with integral screw conveyors are by far the most
common material handling systems in the ferrous metallurgical industry.
Dust storage in baghouse hoppers is a common industry practice, although
this frequently results in dust bridging and subsequent sledgehammering of
hoppers to break the dust bridge. Hopper vibrators are not generally used
because of expense and the tendency of vibrators to pack the dust and ag-
gravate the problem if vibration amplitude and frequency are not correctly
selected. Regular inspection (once per shift) of the hopper is mandatory
to alleviate suction-removal system or bridging problems before they become
serious.
The screw conveyor flighting inside the hoppers is supported every 10
to 15 ft by nonlubricated sleeve-type hanger bearings (see Figure D-2).
Wear on these sleeves and on outboard packed bearings is the major screw
conveyor maintenance problem. The most common sleeve material is cast iron,
although Babbitt, wood, and various other materials have been used.
D.2.4.6 Bag Replacement—The most expensive maintenance operation for
fabric filter systems is the complete change of a set of bags. This is ac-
complished by having a crew of two to six men enter the baghouse and dis-
connect each bag at the cell plate and top suspension level and install a
new bag in its place. Two bag attachment techniques are illustrated in
Figure D-3. The purchase price of replacement bags is given in Table D-7.
D.2.4.7 Tension--The amount of bag tension required for best overall
performance varies according to the make of the equipment. Correct tension
is a function of filter dimensions and cleaning mechanism. A bag that is
too slack can fold over at the lower cuff, bridge'across, and wear rapidly.3
Too much tension can damage the cloth and the fastenings. Shaket cleaning
in particular seems to require a unique combination of tension, shake fre-
quency, and bag properties for best results.3 In any event, the manufac-
turer's recommendations should be followed and the tension checked period-
ically, especially a few hours after installing a new bag.
D.2.4.8 Spare Stock—It is advisable to have a complete set of filter
elements in stock in case of an emergency. The spare filter elements should
be clearly labeled and kept well-separated from used filter elements.3 Ta-
ble D-8 presents a typcial list of items that should be stocked, the approxi-
mate quantities, and if the parts are not stocked, the approximate delivery
time and cost (1977 dollars).
D.2.4.9 Inspection Frequency—External maintenance inspection of the
filter house is usually performed daily, whereas the filter elements them-
selves are typically inspected once a week to once a month.3
D.2.4.10 Shake Cleaning—Shaker mechanisms are generally simply sup-
ported from each end by knife-edge bearings set in grooved blocks. A frac-
tional horsepower motor is used with a yoke linkage to oscillate the shaker
bars (see Figure D-4). Shaker mechanism maintenance is centered around the
drive arrangement. Periodic lubrication of bearings and checking of align-
ment are required. The shaking machinery should also be checked periodi-
cally for wear.
140
-------
0 = 60° MIN
70° BETTER
SCREW CONVEYOR
FLIGHTING
BAGHOUSE
HOPPER
SIDE WALL
BOLTED FLANGE
U" - TROUGH
FLANGED DISCHARGE SPOUT TO
GATHER UP SCREW CONVEYOR
OR AIR LOCK DEVICE
Figure D-2. Typical Trough Hopper and Screw Conveyor Arrangement.
141
-------
BAG.
STAINLESS
STEEL CLAMP
N>
Y////A
CELL PLATE
GAS FLOW
BAG CUFF
CELL PLATE
CUFF WITH
•SPRING STEEL-
BAND
>GAS FLOW
/
tt
THIMBLE CONNECTION
SNAP BAND CONNECTION
Figure D-3. Bag-Cell Plate Attachments.'
-------
TABLE D-7. APPROXIMATE COST OF REPLACEMENT BAGS
Material
1977 cost
(dollars)
Nylon (5.3 oz/yd2)
Sewn in ring
Polyester (7 oz/yd2)
Sewn in ring
Fiberglass (silicon/graphite finish; 9 oz/yd2)
Sewn in ring
Fiberglass (10% PTFE finish; 9 oz/yd2)
Sewn in ring
Top caps (mild steel, 12 in. dia)
Stainless steel clamps
0.64/ft2
2.00 each
0.31/ft2
1.30 each
0.26/ft2
1.25 each
0.42/ft2
1.50 each
2.80 each
1.75 each
143
-------
TABLE D-8. LIST OF REPLACEMENT PARTS FOR A BAGHOUSE FILTER2
-C-
-C-
Type of part
Bags
Door seals
Mechanism
Shaker
Reverse air
Pulse jet
Pulsing plenum
Screw conveyor
Air locks
Pneumatic
Baffle plates
Damper valves
% of total parts
in baghouse that
should be stocked
15
20
20
100
20
20
20
100
20
Delivery time
if not stocked,
weeks
4-8
2-4
2-4
6-10
2-4
2-4
8-10
8-10
8-10
4-6
4-6
Estimated cost
%
See Table D-7
10/seal
10/item
25/item
3/item
5/item
10/item
10/item
100/plate
10/item
Comments
Typical 18" x 48" door
Bearings, knife blades,
belts
Belts
Valve rebuild kit
Solenoid valves, seals,
cylinders
Bearings
Seals
Variable
Solenoid valves, seals,
cylinders
-------
ROCKING MOTION
SHAKER BAR
TENSION NUTS
BAG CAP
CLAMP
Figure D-4. Typical Shaker Arrangements,
145
-------
If the bags are not being cleaned properly, sometimes a minor adjustment of
the shake amplitude or frequency can markedly improve cleaning. If a safe
amount of shaking still does not properly clean the clots, it may be neces-
sary to reduce the filtration velocity for a few hours.3
D.2.4.11 Reverse-Flow Cleaning—With this type of cleaning, the only
maintenance requirement is to check the rate of flow (back pressure) and
the timing periodically to keep the residual drag at an economical level.
D.2.4.12 Shake and Reverse-Flow Cleaning - As in the case of shake
cleaning, wherever the bag is flexed the rate of wear is apt to be high.
Maintenance procedures outlined for the shake and reverse-flow methods also
apply here.
D.2.4.13 Pulse Jet Cleansing--Since there are almost no moving parts
in the pulse type apparatus, hardware maintenance is reduced in comparison
with other cleaning methods. However, excessive use of air cleaning pres-
sure can damage bags by overstretching them. Corrective measures include
reduction of the frequency of cleaning, the use of another type of bag fab-
ric, or reduction of the abrasiveness of the dust. If the baghouse is in a
cold climate, the compressed air lines should be checked periodically in
winter to ensure that they are not frozen. This can easily be accomplished
by listening for the air pulse during the cleaning cycle.
D.2.4.14 Instrumentation--Proper operation of fail-safe mechanisms
and automatic control instrumentation is very important to the safety of
the filter cloth.3 The location of all sensing instruments should be checked
to see that the proper temperature, air flow, etc. are being measured. All
instruments should be calibrated after installation and rechecked monthly
for sensor location, leaks (manometer), sticking, and legibility.3 Instru-
ment readings covering one complete operating cycle should be recorded for
future use in routine checks and troubleshooting. This record should be
posted beside each instrument.
146
-------
APPENDIX D REFERENCES
1. Szabo, M. and R. W. Gerstle. Operation and Maintenance of Particulate
Control Devices on Selected Steel and Ferroalloy Processes. EPA-60012-
78-037. U.S. Environmental Protection Agency. Research Triangle Park,
North Carolina. March 1978.
2. Industrial Air Pollution Guide. PEDCo Environmental Inc. Chapter 7.0.
EPA Contract No. 69-01-4147. (Draft report).
3. Billings, C. E. and J. Wilder. Handbook of Fabric Filtration Technology.
Volume I. Prepared by GCA Corporation for National Air Pollution Control
Administration. Contract No. CPA-22-69-38. December 1970.
147
-------
APPENDIX E
PROCEDURES FOR TROUBLESHOOTING AND CORRECTION OF
BAGHOUSE MALFUNCTIONS
149
-------
APPENDIX E
PROCEDURES FOR TROUBLESHOOTING AND CORRECTION OF
BAGHOUSE MALFUNCTIONS
(RP-reverse pulse; PP-plenum pulse: S-shaker; RF-reverse flow)
Symptom
Cause
Remedy
High baghouse pressure
drop
Baghouse undersized
Bag cleaning mechanism
not adjusted properly
Compressed air pressure
too low (RP, PP)
Repressuring pressure
too low (RF)
Shaking not strong
enough (S)
Isolation damper valves
not closing (S, RF, PP)
Consult manufacturers.
Install double bags.
Add more compartments
or modules.
Increase cleaning fre-
quency. Clean for
longer duration. Clean
more vigorously.
Increase pressure.
Decrease duration and/
or frequency. Check
dryer and clean if
necessary. Check for
obstruction in piping.
Speed up repressuring
fan. Check for leaks.
Check damper valve
seals.
Increase shaker speed.
Check linkage. Check
seals. Check air sup-
ply on pneumatic opera-
tors .
Bag tension too loose
(S)
Pulsing valves failed
(RP)
Cleaning timer failure
Tighten bags.
Check diaphragm.
Check pilot valves.
Check to see if timer
is indexing to all con-
tacts. Check output on
all terminals.
150
-------
APPENDIX E (continued)
' PROCEDURES FOR TROUBLESHOOTING AND CORRECTION OF
BAGHOUSE MALFUNCTIONS
Symptom
Cause
Remedy
Not capable of remov-
ing dust from bags
Low fan motor amperage/
low air volume
Excessive reentrain-
ment of dust
Incorrect pressure
reading
High baghouse pressure
drop
Fan and motor sheaves
reverse
Ducts plugged with
dust
Fan damper closed
System static pres-
sure too high
Fan not operating
per design
Belts slipping
Condensation on bags
(see below).
Send sample of dust
to manufacturer. Send
bag to lab for analysis
for blinding. Dry clean
or replace bags. Reduce
air flow.
Continously empty hopper.
Clean rows of bags ran-
domly, instead of sequen-
tially (PP, RP).
Clean out pressure taps.
Check hoses for leaks.
Check for proper fluid
in manometer. Check
diaphragm in gage.
See above.
Check drawings and
reverse sheaves.
Clean out ducts and
check duct velocities.
Open damper and lock
in position.
Measure static on both
sides of fan and review
with design. Duct
velocity too high. Duct
design not proper.
Check fan inlet config-
uration and be sure
flow is even.
Check tension and adjust.
151
-------
APPENDIX E (continued)
PROCEDURES FOR TROUBLESHOOTING AND CORRECTION OF
BAGHOUSE MALFUNCTIONS
Symptom
Cause
Remedy
Dust escaping at source Low air volume
Ducts leaking
Improper duct balanc-
ing
Improper hood design
Dirty discharge at
stack
Bags leaking
Bag clamps not seal-
ing
Failure of seals in
joints at clean/dirty
air connection
Insufficient filter
cake
Bags too porous
See above.
Patch leaks so air
does not bypass source.
Adjust blast gates in
branch ducts.
Close open areas around
dust source. Check for
cross drafts that over-
come suction. Check for
dust being thrown away
from hood by belt, etc.
Replace bags. Tie off
bags and replace at
later date. Isolate
leaking compartment
if allowable without
upsetting system.
Check and tighten
clamps. Smooth out
cloth under clamp
and re-clamp.
Caulk or weld seams.
Allow more dust to
build up on bags by
cleaning less frequently.
Use a precoating of
dust on bags (S, RF).
Send bags in for per-
meability test and re-
view with manufacturer.
Excessive fan wear
Fan handling too much
dust
See above.
152
-------
APPENDIX E (continued)
PROCEDURES FOR TROUBLESHOOTING AND CORRECTION OF
BAGHOUSE MALFUNCTIONS
Symptom
Cause
Remedy
Excessive fan vibration
High compressed air
consumption
Reduced compressed air
pressure (RP, PP)
Improper fan
Fan speed too high
Buildup of dust on
blades
Wrong fan wheel for
application
Sheaves not balanced
Bearings worn
Cleaning cycle too
frequent
Pulse too long
Pressure too high
Damper valves not
sealing (PP)
Diaphragm valve
failure
Compressed air con-
sumption too high
Dryer plugged
Check with fan manu-
facturer to see if fan
is correct for applica-
tion.
Check with manufacturer.
Clean off and check to
see if fan is handling
too much dust (see above)
Do not allow any water
in fan (check cap, look
for condensation, etc.).
Check with manufacturer.
Have sheaves dynamically
balanced.
Replace bearings.
Reduce cleaning cycle
if possible.
Reduce duration (after
initial shock all other
compressed air is wasted)
Reduce supply pressure
if possible.
Check linkage. Check
seals.
Check diaphragms and
springs. Check pilot
valve.
See above.
Replace dessicant or
bypass dryer if allowed.
153
-------
APPENDIX E (continued)
PROCEDURES FOR TROUBLESHOOTING AND CORRECTION OF
BAGHOUSE MALFUNCTIONS
Symptom
Cause
Remedy
Premature bag failure
decomposition
Moisture in baghouse
High screw conveyor
wear
Supply line too small
Compressor worn
Bag material improper
for chemical composi-
tion of gas or dust
Operating below acid
dew point
Insufficient preheat-
ing
System not purged
after shut-down
Wall temperature below
dew point
Cold spots through
insulation
Compressed air intro-
ducing water (RP, PP)
Repressuring air caus-
ing condensation (RF,
PP)
Screw conveyor under-
sized
Conveyor speed too
high
Consult design.
Replace rings.
Analyze gas and dust and
check with manufacturer.
Treat with neutralizer
before baghouse.
Increase gas temperature.
Bypass at start-up.
Run system with hot air
only before starting
process gas flow.
Keep fan running for at
least 5-10 minutes after
process is shut down.
Raise gas temperature.
Insulate unit. Lower
dew point by keeping
moisture out of system.
Eliminate direct metal
line through insulation.
Check automatic drains.
Install aftercooling.
Install dryer.
Preheat repressuring air.
Use process gas as source
of repressuring air.
Measure hourly collection
of dust and consult manu-
facturer.
Reduce speed.
154
-------
APPENDIX E (continued)
PROCEDURES FOR TROUBLESHOOTING AND CORRECTION OF
BAGHOUSE MALFUNCTIONS
Symptom
Cause
Remedy
High air lock wear
Material bridging in
hopper
Frequent screw conveyor/
air lock failure
High pneumatic con-
veyor wear
Air lock undersized
Thermal expansion
Speed too high
Moisture in baghouse
Dust being stored in
hopper
Hopper slope insuffi-
cient
Conveyor opening too
small
Equipment undersized
Screw conveyor mis-
aligned
Overloading components
Pneumatic blower too
fast
Piping undersized
Elbows too short
radius
Measure hourly collec-
tion of dust and con-
sult manufacturer.
Consult manufacturer to
see if design allows
for thermal expansion.
Reduce speed.
See above.
Remove dust continuously.
Rework or replace hop-
pers.
Use a wide flared trough.
Consult manufacturer.
Align conveyor.
Check sizing to see that
each component is capa-
ble of handling a 100%
delivery from screw
conveyor.
Reduce blower speed.
Review design and re-
duce speed of blower
or increase pipe size.
Replace with long
radius elbows.
155
-------
APPENDIX E (concluded)
PROCEDURES FOR TROUBLESHOOTING AND CORRECTION OF
BAGHOUSE MALFUNCTIONS
Symptom Cause Remedy
Pneumatic conveyor pipes Overloading pneumatic Review design.
plugging conveyor
Reference: Szabo, M. F. and R. W. Gerstle, Operation and Maintenance of
Particulate Control Devices on Selected Steel and Ferroalloy
Processes, EPA 600/2-78-037, U.S. Environmental Protection
Agency, March 1978, Appendix C-4.
156
------- |