PB 198 348
SYSTEMS ANALYSIS OF EMISSIONS AND EMISSIONS CONTROL IN THE IRON
FOUNDRY INDUSTRY. VOLUME I. TEXT
A. T. Kearney and Company
Chicago, Illinois
February 1971
NATIONAL TECHNICAL INFORMATION SERVICE
Distributed , , .'to foster, serve and promote the
nation's economic development
and technological advancement.'
U.S. DEPARTMENT OF COMMERCE
This document has been approved for public release and sale.
-------
PB 198 348
-V'-y^^~;',,^S£. »">»S%W^f^^^^^^t^^?PSf^5^|^W'^
••/.•".-.- ^^^Kir.^Sfct ••--:• ^^ .-v^' i^: ^ /•^^••^^^vSS^ilJ.vJiJ^iS^S^^
NATIONAL TECHNICAL
INFORMATION SERVICE
V«. 23151
SYSTEMS ANALYSIS OF EMISSIONS
AND EMISSIONS CONTROL IN THE
IRON FOUNDRY INDUSTRY
VOLUME I TEXT
FEBRUARY, 1971
For
Division of Process Control Engineering
Air Pollution Control Office
Environmental Protection Agency
Contract No. CPA 22-69-106
Prepared by
A. T. Kearney & Company, Inc.
Chicago, Illinois
-------
AIR POLLUTION CONTROL OFFICE
SYSTEMS ANALYSIS OF EMISSIONS AND EMISSIONS
CONTROL IN THE IRON FOUNDRY INDUSTRY
VOLUME I TEXT
FEBRUARY. 1971
II
TABLE OF CONTENTS
INTRODUCTION
Acknowledgements
Purpose of the Study
Scope of the Study
Literature Search and Bibliography
Glossary of Terms
SUMMARY AND RECOMMENDATIONS
Scope and Objectives
Description of the Iron Foundry
Industry
Emissions Produced and Control
Capability
Quantification and Evaluation of
Emissions
Projection of Trends
Definition of Iron Foundry Air
Pollution Problems
Recommended Research & Development
PAGE
I - 1
I - 1
1-2
II - 1
II - 1
II - 3
II - 4
Technical Analysis of Control Technology II - 6
Economic Analysis of Control Equipment II - 6
Potential Modifications to Foundry
Practices and Emissions Producing
and Control Equipment
II - 8
II - 10
II - 13
II - 14
2 -
PAGE
IV
AIR POLLUTION AND THE IRON FOUNDRY
INDUSTRY
Description of the Iron Foundry
Industry
Iron Castings Production
Iron Foundry Population
Geographical Location of Iron
Foundries
Iron Foundry Melting Equipment
Other Equipment in Iron Foundries
Basic Characteristics of the Iron
Foundry Industry
Compilation of Foundry Data
In-Plant Foundry Visits
The Air Pollution Problem
DESCRIPTION OF BASIC FOUNDRY PROCESSES
Introduction
Process Flow for Gray, Ductile and
Malleable Iron
Process Flow and Descriptions of Basic
Foundry Operations
Raw Material Storage and Furnace
Charge Preparation
Raw Material Receiving and
Storage
Scrap Preparation
Furnace Charge Preparation
III - 1
III - 1
III - 2
III - 3
III - 4
III - 9
III - 11
III - 11
III - 12
III - 13
IV - 1
IV - 2
IV - 3
IV - 3
IV - 3
IV - 4
IV - 5
-------
- 3 -
- 4 -
SECTION
Iron Melcing
Cupola Furnaces
Electric Arc Furnace
Induction Furnaces
Reverberatory Furnace
Inoculation
Ductile Iron Production
Improvement of Mechanical
Properties
Duplexing
Molding, Pouring, Cooling and
Shakeout
Green Sand Molding
Centrifugal Casting
Dry Sand Molding
Shell and Hot Box Molding
Full Mold Process
Rheinstahl Process
Pouring and Cooling
Shakeout
Sand Conditioning
Receiving and Storage
Sand Preparation
Sand Reclamation
PAGE
IV - 6
IV - 7
IV - 15
IV - 19
IV - 23
IV - 23
IV - 24
IV - 24
IV - 25
IV - 25
IV - 26
IV - 27
IV - 27
IV - 28
IV - 29
IV - 29
IV - 29
IV - 30
IV - 30
IV - 31
IV - 31
IV - 33
SECTION PAGE
Cleaning, Heat Treating and
Finishing IV - 33
Cooling and Sorting IV - 34
Castings Cleaning IV - 34
Heat Treating IV - 36
Casting Finishing IV - 36
Casting Coating IV - 37
Coremaking IV - 37
Core Processes IV - 37
Core Department Process Flow IV - 40
Pattern Making IV - 42
V EMISSIONS PRODUCED AND EMISSION CONTROL
CAPABILITY
Emissions Produced in the Basic Processes V - 1
Raw Material Storage and Charge
Makeup V - 1
Melting V - 3
Magnesium Treatment for Producing
Ductile Iron V - 5
Molding, Pouring and Shakeout V - 6
Sand Conditioning V - 8
Cleaning and Finishing V - 9
Coremaking V - 10
Pattern Making V - 11
-------
- 5 -
SECTION PACE
Emission Control Capability V - 11
Emissions Controllable with Existing
Equipment Design V - 12
Farticulate Emissions Uneconomical
to Control V - 12
Cupola V - 14
Electric Arc Furnace V - 14
Magnesium Treatment for Producing
Ductile Iron V - 15
Mold Pouring V - 15
Coremaking V - 15
VI QUANTIFICATION AND EVALUATION OF EMISSIONS
Description of Sampling and Analytical
Techniques VI - 1
General Atmosphere Sampling VI - 1
Stack Sampling VI - 5
Necessity for Stack Sampling Program VI - 8
Test Reports Gathered and Analyzed VI - 8
Need for Additional Data VI - 9
Evaluation of Cupola Emissions VI - 10
Nature of Data Gathered VI - 10
Classification of Melting Processes
by Type and Design Parameter VI - 13
Variables Affecting Emissions VI - 14
Analysis of Emissions Data VI - 19
Discussion of Results VI - 26
Introduction VI - 26
Variables with Proven Effect on
Emissions VI - 27
- 6 -
SECTION PAGE
Cupola Design Variables VI - 29
Variables with Probable Effect
on Emissions VI - 30
Variables with Expected Effect
on Emissions VI - 32
Conclusion VI - 33
Evaluation of Emissions from Other
Sources VI - 37
Electric Arc Melting and Holding VI - 39
Magnesium Treatment for Producing
Ductile Iron VI - 43
Mold Pouring and Cooling VI - 46
Coremaking vi - 48
VII TECHNICAL ANALYSIS OF EMISSION CONTROL
TECHNOLOGY
Analysis of Emission Control Techniques VII - 1
Dry Centrifugal Collectors VII - 1
Historical Background VII - 1
General Characteristics of
Equipment VII - 2
Description of Specific Types VII - 2
Advantages of Dry Centrifugal
Collectors VII - 7
Disadvantages of Dry Centrifugal
Collectors VII - 7
Limitations of Dry Centrifugal
Collectors VII - 8
Wet Collectors VII - 9
Historical Background VII - 9
General Characteristics of Equipment VII - 9
A T KEARNEV 6k COMPANY
-------
- 7 -
- 8 -
SECTION
Description of Specific Types
Advantages of Wet Collectors
Disadvantages of Wet Collectors
Limitations of Wet Collectors
Fabric Filters
Historical Background
General Characteristics of
Equipment
Description of Specific Types
Advantages of Fabric Filters
Disadvantages of Fabric Filters
Limitations of Fabric Filters
Electrostatic Frecipitators
Historical Background
General Characteristics of Equipment
Description of Specific Types
Advantages of Electrostatic
Precipitators
Disadvantages of Electrostatic
Precipitators
Limitations of Electrostatic
Precipitators
Collection Efficiency of Emission
Control Equipment Systems
Weight
Particle Size Count
Opacity
Ground Level Concentration
PACE
VII - 11
VII - 18
VII - 19
VII - 20
VII - 20
VII - 20
VII - 22
VII —24
VII - 27
VII - 28
VII - 28
VII - 30
VII - 30
VII - 30
VII - 32
VII - 34
VII - 35
VII - 36
VII - 36,
VII - 37
VII - 39
VII - 40
VII - 41
SECTION PAGE
Factors Affecting Control Equipment
Collection Efficiency VII - 41
Combustion Devices VII - 52
Afterburners VII - 52
Catalytic Afterburners VII - 54
Application of Emission Control Equip-
ment Systems to Foundry Processes VII - 55
Raw Material Handling, Preparation
and Charge Makeup VII - 56
Cupola Melting VII - 57
Wet Caps VII - 58
Dry Centrifugal Collectors VII - 59
Wet Collectors VII - 60
Fabric Filters VII - 60
Electrostatic Precipitators VII - 61
Other Techniques VII - 61
Electric Arc Melting VII - 62
Emission Collection Equipment VII - 62
Furnace Hoods VII - 63
Other Methods VII - 64
Electric Induction Melting VII - 64
Other Types of Melting VII - 65
Inoculation VII - 66
Ductile Iron Production vll - 66
Mechanical Property Improvement
by Inoculation VII - 66
A.T KEARNEY & COMPANY. l»c
-------
SECTION
VIII
Mold Pouring, Cooling and Shakeout
Sand Preparation and Handling
Cleaning and Finishing
Coremaking
Miscellaneous Areas
Case Histories of Attempts to Control
Emissions in Iron Foundries
Melting Processes
Non-Melting Processes
ECONOMIC ANALYSIS OF EMISSION CONTROL
TECHNOLOGY
General
Capital Costs of Control Systems
Cupola Melting
Electric Arc Melting
Relation of Capital Costs of Control
Systems to Design Variables
Cupola Melting
Wet Scrubber Efficiency -
Pressure Drop Relationship
Electric Arc Melting
Other Melting Methods
Operating Costs of Control Systems
Relation of Operating Costs of Control
Systems to Design Variables
- 9 -
PAGE
VII - 66
VII - 68
VII - 69
VII - 70
VII - 71
VII - 71
VII - 72
VII - 87
VIII - 1
VIII - 2
VIII - 8
VIII - 9
VIII - 10
VIII - 10
VIII - 14
VIII - 16
VIII - 18
VIII - 18
VIII - 24
SECTION
IX
Effect of Emissions Control on Capital
and Operating Casts in Model Foundries
Model Foundry Development
Melting Systems Studied
Sizes of Foundries Studied
Levels of Operation Studied
Qualifications Regarding Model Foundry
Design Data
Equipment and Building Specifi-
cations
Melting Equipment
Emission Control Equipment
Buildings
Capital Costs
Operating Costs
Direct Material Costs
Conversion Costs
Emission Control Costs
Summary of Capital and Operating
Costs
Benefits of Emission Control
Installations
TECHNICAL AND ECONOMIC ANALYSIS OF
POTENTIAL MODIFICATIONS TO FOUNDRY
PROCESSES AND EQUIPMENT
Improvement of Emissions Control
Capability
Potential Modifications to Foundry
Practices and Emissions Producing
Equipment
Potential Modifications to Cupolas
and Cupola Operations
10 -
PAGE
VIII - 26
VIII - 26
VIII - 26
VIII - 27
VIII - 27
VIII - 28
VIII - 28
VIII - 29
VIII - 29
VIII - 31
VIII - 31
VIII - 32
VIII - 33
VIII - 34
VIII - 35
VIII - 37
VIII - 37
IX - 1
IX - 2
IX - 3
-------
- 11 -
SECTION
XI
XII
PAGE
Decrease Stack Gas Volume IX - 3
Decrease Coke Charge IX - 8
Preparation of Charge Materials IX - 11
Cupola Design Changes IX - 14
Change Melting Method IX - 15
Potential Modifications to Other
Melting Furnaces IX - 16
Modifications to Magnesium Treatment for
Producing Ductile Iron IX - 17
Modifications to Sand Hand]lag Systems IX - 18
Potential Modifications to Emissions
Control Equipment IX - 20
Impact of Potential Modifications to
Foundry Equipment and Practices IX - 22
PROJECTION OF TRENDS
Growth of the Iron Foundry Industry X - 1
Equipment Trends X - 4
Effects on Emissions X - 7
DEFINITION OF THE IRON FOUNDRY AIR
POLLUTION PROBLEM
General XI - 1
Nature of Foundry Emissions XI - 1
Inventory of Foundry Emissions XI - 4
Industry Problems XI - 8
OPPORTUNITIES FOR RESEARCH AND
DEVELOPMENT
Needs for Research and Development XII - 1
- 12 -
SECTION PAGE
Fundamental Knowledge XII - 1
Economic Factors XII - 2
Materials XII - 3
Processes XI1 - 3
Ongoing Research and Development
Projects XII - 3
Melting of Iron XII - 4
Emissions Collection XII - 5
Proposed Research and Development
Projects XII - 6
Fundamental Research Projects XII - 7
No. 1. Stack Sampling Program XII - 7
No. 2. Iron Oxide Formation
Program XII - 8
No. 3. Fine Oxide - Opacity
Relationship Program XII - 9
Economic Research Projects XII - 10
No. 5. Waste Heat Utilization XII - 10
No. 7. Centralized Scrap Pre-
paration Development XII - 11
No. 10. Waste Product Utilization XII - 12
Materials Development Projects XII - 13
No. 4. High Temperature Fabric XII - 13
No. 8. Low Emission Core Binder
Materials XII - 14
Development of New Equipment XII - 15
No. 6. Continuous Melting XII - 15
Furnace
No. 9. Agglomeration of Fine XII - 16
Emissions
A T KEARNEY & COMPANY INC
-------
VOLUME II - LIST OF EXHIBITS
III-l
III-2
III-3
I1I-4
III-5
III-6
III-7
IV-1
IV-2
IV-3
IV-4
IV-5
IV-6
IV-7
IV-8
IV-9
IV-10
Title
Iron Foundry Production Trends
Population Trends in the Foundry
Industry
Distribution of Iron Foundries, 1969
Geographical Distribution of Iron
Foundries
Iron Foundry Cupola Trends
Iron Foundry Electric Furnace Trends
Characteristics and Sources of Emissions
in Various Foundry Departments
Iron Foundry Process Flow
Process Flow Diagram - Gray, Ductile and
Malleable Iron
Summary of Gray Iron Specifications
Summary of Ductile Iron Specifications
Summary of Malleable Iron Specifications
Iron Foundry Scrap Specifications
Pig Iron and Ferroalloy Specifications
Process Flow Diagram - Raw Material
Storage and Charge Makeup
Process Flow Diagram - Melting Department
Electric Arc Furnace - Heat and Material
Balance
- 2 -
IV-12
IV-13
IV-14
IV-15
IV-16
IV-17
IV-18
IV-19
IV-20
IV-21
IV-22
IV-23
IV-24
IV-25
IV-26
IV-27
Title
Coreless Induction Furnace - Heat and
Material Balance
Process Flow Diagram - Molding, Pouring
and Shakeout
Process Flow Diagram - Cleaning and
Finishing
Process Flow Diagram - Sand Conditioning
Process Flow Diagram - Coremaking
Illustration of Conventional Lined
Cupola
Illustration of Water-Cooled Cupola
Illustration of Cupola Reaction Area
Illustration of Electric Arc Furnace
Illustration of Channel Induction
Furnace
Illustration of Coreless Induction
Furnace
Illustration of Reverberatory Furnace
Illustration of Magnesium Treatment
Methods for Producing Ductile Iron
Illustration of Pouring Station with
Horizontal Draft, Cantilevered Hood
Illustration of Shakeout Station
Illustration of Sand Muller
Illustration of Blast Cleaning Unit
-------
- 3 -
Number
VI-1
VI-2
VI-3
VI-4
VI-5
VI-6
Vl-7
VI-8
VI-9
VI-10
VI-11
VI-12
VI-13
VI-14
VI-15
VI-16
Title
Ringelmann Scale for Grading Density of
Smoke
Pertinent ASME Items Which Must Be
Conformed to by Parties Conducting a
Stack Sampling Test
Design Features of the Cupola
Design Features of the Electric Arc
Design Features of the Induction
Furnace
Classification of Lined and Unlined
Cupola Furnaces Found in Practice
Chemical Composition of Cupola
Particulate Emissions
Particle Size Distribution - Cupola
Emissions
Parameters of Cupola Furnaces -
Linear Regression Analyses of Emissions
Affected by Furnace Design Factors
Multiple Linear Regression Correlation
Matrices
Linear Regression Analyses Observations
Particulate Emissions vs Specific Blast
Rate for Acid Lined Cupolas
Effect of Specific Blast Rate and Coke Rate on
Particulate Emissions from Unlined Cupolas
Effect of Type of Scrap on Amount of
Iron Oxide Present
Results of Size Distribution and
Chemical Analysis for Three Electric
Arc Installations
Emissions Data from Electric Arc
Melting Furnaces
Number
VI-17
VI-18
VI-19
VI-20
VI-21
VI-22
VI-23
VI-24
VII-1
VII-2
VI I-3
VII-A
VII-5
VII-6
VII-7
VII-8
VII-9
VII-10
Title
Relationship between Rate of Emissions
and Heat Cycle for Electric Arc Melting
Treatment Agents for Producing Ductile
Iron
Magnesium Treatment Systems Emissions
Report for Ductile Iron Production and
Gray Iron Desulfurization
Molding Sand Gas Analyses
Molding Sand Gas Evolution and Hot
Permeability
Gas Volume Evolved as a Function of
Volatiles Contained in Molding Sand
Effect of Baking Time on Gas Generated dur-
ing Pouring for Various Baking Temperatures
Effect of Sand to Oil Ratio on Amount
of Core Gas Generated during Pouring
Cyclone Collector
High Efficiency Centrifugal Collector
Dry Dynamic Precipitator Collector
Wet Cap Collector
Wet Dynamic Precipitator Collector
Vane-Type Centrifugal Wet Collector
Multiple Tube-Type Centrifugal Wet
Collector
Orifice-Type Wet Collector
Centrifugal Spray Wet Collector
Marble Bed-Type Wet Collector
-------
- 6 -
VII-12
VII-13
VII-14
VII-15
VII-16
VII-17
VII-18
VII-19
VII-20
VII-21
VII-22
VII-23
VII-24
VII-25
VII-26
Title
Impingement Baffle Grid-Type Wet
Collector
Venturi Collector
Wet Collector Particle Collection
Limitations and Design Capacities
Cutaway View Showing Fabric Filter,
Flat or Screen-Type Bag
Cutaway View Showing Fabric Filter
Tubular-Type Bag
Intermittent Fabric Filter Collector
Continuous Automatic Fabric Filter
Collector
Reverse Jet Continuous Fabric Filter
Collector
Wet-Type Electrostatic Precipitator
Effluent Cleaning System
Dry-Type Electrostatic Precipitator
Effluent Cleaning System
Collection Efficiency of Emission
Control Equipment Systems
Grading of Test Dust
Overall Collection Efficiency on Test
Dust
Chemical Composition of Cupola Oust
by Weight
Grade Efficiency Curve, Dry Electro-
static Precipitator, High Efficiency
Cyclone
Calculation of Collector Efficiency
VII-28
VII-29
VII-30
VIII-1
VIII-2
VIII-3
VIII-4
VIII-5
VIII-6
VIH-7
VIII-8
VIII-9
Title
Grade Efficiency Curve for Fabric
Filter, Effect of Particle Size and
Length of Bag in Service on Fabric
Filter Efficiency
Relationship between Collection
Efficiency, Particle Size and Pres-
sure Drop for Venturi Scrubbers
Cupola Afterburner, Catalytic After-
burner Applied to Core Bake Oven
Process
Application of Emission Control Equip-
ment Systems to Foundry Processes
Conditions Affecting Installation Cost
of Control Devices
Investment Cost Equations for Equipment
Installed on Cupolas
Total Investment Cost vs Gas Volume
for Wet Scrubber on Cupolas
Total Investment Cost vs Gas Volume for
Fabric Filters on Cupolas
Total Investment Cost vs Gas Volume
for Mechanical Collectors on Cupolas
Approximate Melting Rates and Gas
Volumes for Lined Cupolas
Approximate Melting Rates and Gas
Volumes for Unlined Cupolas
Comparison of Gas Take-Off above
Charge Door and Below Charge Door.
Lined Cupola, Coke Ratio 8/1
High Energy Wet Scrubber Total Invest-
ment Cost vs Melt Sate for Unlined Cupola
8/1 Coke Ratio
-------
- 7 -
- 8 -
Number
VIII-10
VIII-11
VIII-1Z
VIII-13
VIII.-14
VIII-15
VIII-16
VIII-17
VIII-18
VIH-19
VIII-20
VIII-21
VIII-22
VIII-23
Title
High Energy Wet Scrubber Total Invest-
ment Cost vs Melt Rate for Lined Cupola,
8/1 Coke Ratio
Low Energy Wet Scrubber Total Investment
Cost vs Melt Rate for Unlined Cupola,
8/1 Coke Ratio
Low Energy Wet Scrubber Total Investment
Cost vs Melt Rate for Lined Cupola, 8/1
Coke Ratio
Fabric Filter Total Investment Cost vs
Melt Rate for Unlined Cupola, 8/1 Coke
Ratio
Fabric Filter Total Investment Cost vs
Melt Rate for Lined Cupola, 8/1 Coke
Ratio
Total Investment Costs for Wet Caps
Calculation of Wet Scrubber Efficiency
for Various Pressure Drops
Comparison of Cupola Outlet Dust Load-
ing and Pressure Drop for Wet Scrubbers
Approximate Exhaust Volumes for Electric
Arc
Installed Cost of Fabric Filter on
Electric Arc
Total Annual Costs for High Energy Wet
Scrubbers on Cupolas
Total Annual Costs for Low Energy Wet
Scrubbers on Cupolas
Relative Changes in Total Annual Costs
vs Pressure Drop for Wet Scrubbers
Total Annual Cost for Fabric Filters on
Cupolas
VIII-25
VIII-26
VIII-27
VIII-28
VIII-29
VIII-30
VIII-31
VIII-32
VIII-33
Title
Total Annual Cost for Fabric Filters on
Electric Arc
Comparison of Cost per Ton of Melt for
High Energy Wet Scrubber on Unlined
Cupola at Different Levels of Operation,
5/1 Coke Ratio
Comparison of Cost per Ton of Melt for
High Energy Wet Scrubber on Lined Cupola
at Different Levels of Operation, 8/1
Coke Ratio
Comparison of Cost per Ton of Melt for
Low Energy Wet Scrubber on Unlined
Cupola at Different Levels of Operation,
5/1 Coke Ratio
Comparison of Cost per Ton of Melt for
Low Energy Wet Scrubber on Lined Cupola
at Different Levels of Operation, 8/1
Coke Ratio
Comparison of Cost per Ton of Melt for
Fabric Filter on Unlined Cupola at Dif-
ferent Levels of Operation, 5/1 Coke
Ratio
Comparison of Cost per Ton of Melt for
Fabric Filter on Lined Cupola at Different
Levels of Operation, 8/1 Coke Ratio
Summary of Capital Costs to Produce
Iron under Various Production and
Operating Conditions
Summary of Operating Costs for Produc-
ing Iron under Various Production and
Operating Conditions
Capital and Operating Costs per Ton
versus Operating Hours per Year for
Cold Blast Cupola with Fabric Filter
(Alternate No.l)
-------
- 9 -
Number
VIII-34
VI1I-35
VIII-36
Title
Capital and Operating Costs per Ton vs
Operating Hours per Year for Hot Blast
Cupola with Wet Scrubber (Alternate
No. 2)
Capital and Operating Costs per Ton vs
Operating Hours per Year for Electric
Arc Furnace with Fabric Filter (Alternate
No. 3)
Capital and Operating Costs per Ton vs
Operating Hours per Year for Cureless
Induction Furnace with Afterburner on
Preheater (Alternate No. 4)
VOLUME III - LIST OF APPENDICES
APPENDIX TITLE
A Bibliography
B Data Bank
C Material and Heat Balance
D Detail Economic Cost Curves
E Emission Test Procedures
F Glossary of Terms
IX-1
Modifications to Cupola Melting Prac-
tices to Reduce Emissions
XI-1
XI-2
XI-3
XII-1
Inventory of Iron Foundry Emissions from
Melting Operations, 1969
Inventory of Iron Foundry Emissions from
Non-Melting Operations, 1969
Priority Rating Chart
Proposed Research and Development Projects
-------
NATIONAL AIR POLLUTION CONTROL OFFICE
CONTRACT HO. CPA 22-69-106
SYSTEMS ANALYSIS OF EMISSIONS AND EMISSIONS
CONTROL IN THE IRON FOUNDRY INDUSTRY
VOLUME I TEXT
FEBRUARY. 1971
I - INTRODUCTION
ACKNOWLEDGEMENTS
A. T. Kearney & Company, Inc. gracefully acknowledges Che
many people and organizations who cooperated and assisted in this
study.
We acknowledge the cooperation and assistance provided by
the'American Foundrytnen's Society and the Gray and Ductile Iron
Founders' Society in furnishing information and in obtaining the
cooperation of their member foundries. In addition, A.F.S. pro-
vided the use of its headquarters for review meetings with the
Industry Liaison Committee.
We acknowledge the assistance provided by the members of
the Industry Liaison Committee who were helpful in reviewing the
progress of the study and providing a critique of the report.
We acknowledge the assistance and cooperation of the many
iron foundries who provided A. T. Kearney & Company the courtesy
of plant visits and furnished data for this study and for the
data bank.
We acknowledge the assistance and cooperation of the various
manufacturers of foundry equipment and the manufacturers of
1-2
emission control equipment for providing valuable technical
and economic data.
Finally, we thank the staff of APCO for their assistance and
guidance.
This report was prepared by a project team of A. T. Kearney
staff and associated consultants. The overall direction was un-
der Mr. E. Stuart Files, Vice President of A. T. Kearney &
Company, Inc. The Project Director was Mr. Joseph H. Greenberg,
Principal, and the other members of the team were Mr. Robert E.
Conover, Senior Associate; Mr. Bernard Gutow, Associate; and
other members of the Kearney Organization. The associated con-
sultants who provided valuable assistance in carrying out of the
work were Mr. John Kane, Mr. George Tubich, and Mr. James Ewens,
all independent consultants in air pollution control and foundry
operations.
PURPOSE OF
THE STUDY
The Air Pollution Control Office has the task of developing
technology for a national program for control of air pollution
and, as a part of this program, is conducting a series of systems
analysis studies of various industries. These studies are being
conducted by the Division of Process Control Engineering. This
study is directed at the iron foundry industry, with particular
emphasis on the melting area. The primary goal of the study is
to define the air pollution problems of the iron foundry indus-
try, and to set priorities for research and development work that
-------
1-3
will lead Co improved emission control capabilities at reduced
cost.
SCOPE OF THE
STUDY
For the purposes of chis study, the iron foundry industry
was defined as those shops that melt iron (including iron and
steel scrap) in furnaces, pour the molten iron into molds, and
alloy and/or treat the iron in either the molten or cast state,
with processes limited to making gray, malleable, and ductile
cast iron. This definition excludes blast furnace processes,
including direct casting of blast furnace metal into molds, and
processes for converting iron to steel.
The following iron foundry facility and operations areas
were included in the study:
1. Raw materials receiving, storage, handling, prepara-
tion, and charging.
2. Iron melting, duplexing, holding, desulphurization,
inoculation, and magnesium treatment for producing ductile iron.
Equipment includes cupolas, direct and indirect electric arc fur-
naces, coreless and channel induction furnaces, stationary and
rotary fuel fired furnaces, inoculation and magnesium treatment
units.
3. Sand molding, pouring, cooling, shakeout, sand
handling, storage, preparation, and distribution.
4. Casting cooling, cleaning, grinding, heat treat-
ing, and painting.
1-4
5. Coremaking of oil-bonded, chemically bonded, and
other types of cores.
6. Other auxiliary areas in the iron foundry Chat
produce emissions.
LITERATURE SEARCH
AND BIBLIOGRAPHY
An extensive search of the literature was conducted to Identi-
fy and list in a single bibliography published material'pertinent
to the subject of this study. During the course of this litera-
ture search, approximately 191 information sources, including
serial publications, have been identified.
A total of 735 references of related material were cataloged,
including 511 articles in technical journals, 213 books, and
miscellaneous publications by industry, technical societies, and
governmental bodies, plus 11 patents.
Appendix A is a copy of the bibliography which is arranged
by calendar year and subdivided into categories of books, reports,
and articles.
GLOSSARY OF
TERMS
Standard iron foundry industry and air pollution terminology
are used throughout this report. Definitions of these terms are
provided in Appendix F, Glossary of Terms, found in Volume III
of this report.
-------
II - SUMMARY AMD RECOMMENDATIONS
SCOPE AND
OBJECTIVES
the Air Pollution Control Office has Che cask of developing
technology for a national program foe control of air pollution
and, as a part of this program, sponsored a series of systems
analysis studies of various industries which are primary sources
of air pollution. This study was directed at the iron foundry
industry, with particular emphasis on the melting area, and was
conducted by A. T. Kearney & Company, Inc.
The primary goal of the study was to define the air pollu-
tion problems of the iron foundry industry, and to set priorities
for research and development work that will lead to improved
emission control capabilities at reduced cost.
DESCRIPTION OF
THE IRON FOUNDRY
INDUSTRY
The number of gray, malleable and ductile iron foundries
in the United States hos decreased from about 3,200 in 1947 to
1,670 in 1969, with a trend of reducing by about 70 installations
per year. These foundries are located in 48 states, although
80% are located in only 13 states. Of all iron foundries, only
about 525, or 297., can be classified as medium and large
foundries, employing more than 100 people, and only 91 can be
called very large.
II - 2
The cupola is still by far the moat common method of melt-
ing iron, with about 1,970 cupolas in service In 1970. This Is
a decline of 2,500 in the past 25 years. This decline can be
expected to continue, as more foundries arc closed, and others
convert from cupola to electric melting. Electric arc installa-
tions are increasing rapidly as are electric induction
installations.
The information available regarding the other productive
operations in iron foundries which are sources of emissions is
not nearly as well documented as the melting area. However, we
know that all foundries employ some kind of molding practice,
with most iron foundries using sand molding. The modern trend
is toward automated or semiautomated molding, employing con-
veyors of some type for moving molds past each of the stations.
This can be expected to increase as more foundries automate
or mechanize their operations.
Coremaking is another area that is undergoing rapid
changes, with the trend being away from oil-bonded, baked sand
cores, toward chemically bonded, thermally cured cores.
The widespread distribution of iron foundries, and the
prominence of the cupola stack in nost communities in which an
iron foundry is located, have combined to label the iron foundry
as a major source of air pollution. It has been estimated that
iron foundries produce 243,000 tons of particulates in the
United States each year. Although some of these are now captured
-------
II - 3
by che presently used emission control equipment, a large per-
centage is being discharged Co che atmosphere, equal Co abouC
0.9% of all participate matter emitted by all sources.
The pollutants discharged by Che iron foundry indusCry can
be classified as:
1. Emissions from melting furnace operations.
2. Emissions from other dust-producing operations
within the plant.
3. Odors and gaseous compounds from both sources.
EMISSIONS PRODUCED
AND CONTROL CAPABILITY
Processes in the iron foundry can be divided into six
production areas:
1. Raw material storage, preparation and charging.
2. Metal melting.
3. Molding, pouring and shakeout.
A. Sand conditioning and reclamation.
5. Cleaning, heat treating and finishing.
6. Coremaking.
Each operation contains equipment and processes capable
of producing air pollutants in Che form of noxious fume, and/or
particulate matter. The latter ranges from metal dust from
grinding operations, that is relatively easy and inexpensive
to control, to extremely fine ferrous and nonferrous oxides and
smoke from melting furnaces that are very expensive and diffi-
cult to control.
II -
Over the years, increasing attention to foundry and other
industry sources of pollution has prompted development of
numerous methods and techniques for control of emissions from
these sources.
Of all techniques available to control foundry pollutants,
emission collection equipment systems are the most significant.
These systems, which include dry centrifugals, wet collectors,
fabric filters and electrostatic precipitatora, vary widely
in design, capabilities, cose and application.
QUANTIFICATION AND
EVALUATION OF
EMISSIONS
Although the majority of gray iron foundries currently
do not have collection equipment systems for control of emissions
from melting operations, the number of installations of these
systems has increased rapidly the last few years.
By far the majority of the foundries use cupolas, with
only about 15% of production made using other forms of melting.
The major components of particulate emissions from Iron melting
cupolas can be grouped into three major categories: (1) metallic
oxides (2) silicon and calcium oxides, and (3) combustible
materials.
The amounts of metallic oxides occurring in cupola emissions
are related to the presence of Che respeccive metals in Che
scrap charge and their partial vapor pressures at the temperature
of Che cupola melting zone.
-------
II - 5
II - 6
The oxides of silicon and calcium, representing the second
category, result from lining erosion, embedded molding or core
sand on foundry returns, dirt from the scrapyard adhering to
scrap, or from the limestone flux.
The third category of emissions, combustible material,
includes coke particles, vaporized or partially burned oil and
grease and other contaminants swept up Che stack by the top
gases.
The effects of 15 variables on the type and quantity of
cupola emissions have been evaluated by the use of a variety
-of analytical techniques. The analyses demonstrate that some
of the variables have strong and quantifiable effects and
others have insignificant effects on emission levels.
Although the cupola has been considered to be the principal
source of iron foundry emissions which are difficult or costly to
control, there are other areas in the foundry which also produce
these types of emissions. Some of these have been adequately
and economically controlled with existing designs of equipment,
while others have proven to present problems which have not
yet been completely solved.
The ease and cost of controllability can generally be re-
lated to two broad aspects: the problems which relate to cap-
turing the emissions with a minimum of infiltrated air and the
problems of adequately removing the offending fine particles
and gases from the captured effluents before releasing them to
the atmosphere. In general, those operations which in many
foundries take place in open, nonconcentrated, nonstationary
locations are difficult to control although in many cases,
the actual costs of control may not be high once the effluents
are captured. Such operations occur in the scrapyard, mold
pouring area, and the shakeout area.
The other areas which involve difficulty of capture and
high cost of control equipment include melting, inoculation and
coremaking.
TECHNICAL ANALYSIS
OF CONTROL TECHNOLOGY
A detailed analysis was presented of each of the types of
emission control equipment in use in each of the areas of the
iron foundry. This included a description of the equipment,
its application, historical background, and its relative posi-
tion as compared with other equipment. Case histories of past
attempts to abate emissions were cited. In general, it was
demonstrated that equipment and methods exist for control of
almost all foundry emissions, although in many cases the
economics are unfavorable, particularly for small foundries.
ECONOMIC ANALYSIS
OF CONTROL TECHNOLOGY
Estimates have been made of the capital costs and annual
costs of operation of control systems based upon analysis of
existing and proposed installations. The cost data obtained
were all basically for foundries with cupolas controlled with
-------
II - 7
wet caps, mechanical collectors, fabric filters, and high and
low energy wet scrubbers, and for electric arc furnaces con-
trolled with fabric filters.
The major factor that influences capital cost is the type
of melting equipment. The complexity of control systems is
greatest for the cupola, moderate for the electric arc, and
least for induction melting.
A number of design and operating variables are related
to the total cost of emission control systems. These include
melt rate of the furnace, pressure drop of the system, blast
volume in standard CFM and gas volume throughput in actual
CFM. A computerized multiple regression analysis of the
collected data has confirmed that total costs varied most
directly with actual gas volume throughput.
A comparison of the impact of annual costs was made for
wet scrubbers and fabric filters. These were the more popular
and commonly used types of control systems observed on foundry
melting equipment. These operating costs were given in terms of
cost per ton of metal melted for wet and dry collectors on
cupolas and arc furnaces.
A series of model foundries developed as a means of relating
the economics of emission control installation and operation to
the total costs of installing and operating iron melting depart-
ments. These models covered the complete melting department
of foundries with melting capacities of 5, IS, 30 and 50 cons
II - 8
per hour, and operating at varying levels, including 500, 1,000,
2,000 and 4,000 hours per year. Melting methods included
cupola, arc furnace and induction furnace. Costs of emission
control installation and operation were identified as a percentage
of total costs for each of these conditions.
POTENTIAL MODIFICATIONS
TO FOUNDRY PRACTICES
AND EMISSIONS PRODUCING
AND CONTROL EQUIPMENT
Of all foundry operations causing emissions to be released
to the atmosphere in some degree, the melting operation has re-
ceived the greatest attention. The effort to improve collection
capability of a lower cost has been concentrated on the problem
of the cupola as the source of the greatest amount of particulate
matter. The results of these investigations in recent years
have been a number of modifications offering at least a partial
solution to the problem.
Other work has been done in the areas of preheating,
magnesium treatment for producing ductile iron, pouring and
cooling, sand handling and coremaking.
Cupolas and cupola operations can be modified in an attempt
to achieve two goals: to decrease the cost of emissions collec-
tion and to decrease the quantity of emissions requiring
collection. One method of decreasing the volume of cupola
stack gas is to limit air infiltration by reduction of the
charging door area. Locating the gas take-off below the top of
the charge burden is a second method of reducing air infiltration
-------
II - 9
and thereby decreasing the volume of gas to be cleaned.
Reduction of coke in cupola charges offers some beneficial
results. The evaluation of cupola emissions indicates that
particulate emissions tend to decrease as the amount of coke in
the cupola charge decreases. Heating of the cupola blast air
has been recognized for many /ears as one method of increasing
the melting rate with the same coke charge or of maintaining
the melting rate with a smaller coke charge. Oxygen enrichment
of the blast air also results in a decrease in coke requirements
by reducing the 60% nitrogen in air which would otherwise
require heating to the iron meltiitg temperature in the cupola.
Natural gas injection, like oxygen enrichment, provides a means
by vhich cupola coke requirements can be reduced, melting rate
increased, melting costs lowered, and a modest decrease in par-
ticulate emissions can be realized.
Screening of limestone and coke before loading into the
charging bucket is an elementary and inexpensive step to pre-
vent breeze and dust from being charged into the cupola. Scrap
preparation is more often performed by scrap dealers than
foundry personnel due to the equipment cost and labor require-
ments. An increasing amount of today's scrap is automotive,
consisting of whole or broken engines or fragmented bodies.
The increase in the number of electric and reverberatory
melting furnaces in recent years, at a time when the number of
iron foundries have decreased dramatically, is evidence that
A T KEARVEY 61 COM PAN Y. Inc.
II - 10
one acceptable solution to the problem of the high cost of
emissions control is to replace the cupola by a furnace that
produces less emissions. Installation and melting costs of al-
ternative furnaces are not always lower, but the total of the ad-
vantages sometimes tips the scale in favor of cupola replacement.
Improvements in emissions control equipment rn the fore-
seeable future, with few exceptions, promise to be minor in
nature. These will be results of efforts by the manufacturers
of this equipment to improve quality of manufacturing, use of
better materials and redesign to raise efficiencies and lower
costs of their products.
PROJECTION
OF TRENDS
Total iron castings production, excluding ingot molds made
from direct blast furnace iron, has been projected to be approxi-
mately 17 million tons per year by 1980, or an average 2%
increase per year.
The number of foundries which will be in operation by 1980
has been projected to be approximately 1,100. Since the number
of medium- and large-sized foundries has been projected to in-
crease slightly, the entire drop from the 1969 total of 1,630
foundries to 1,100 foundries is expected to take place among
the small-sized companies. The combination of increasing output
and reduced number of producers is expected to continue to
increase the average output per foundry to an estimated 16,500
tons per year by 1980.
A T KEARS'EY fl. COMPANY. INC
-------
II - 11
The most significant changes in the types and uses of
equipment which are expected to take place in Che iron foundry
are in the melting department. The cupola, which once was
almost the only source of molten iron, has been rapidly declining
in number of installations, with the number in 1969 being only
about 45% of tKose In existence immediately after the war. This
decline has been projected to continue at an undiminished rate,
with the number of cupolas in 1980 estimated to be about 1,300.
In 1969, an estimated 85% of all iron melted in iron
foundries uas performed in cupolas. As the replacement of
cupolas by other forms of melting continues to accelerate, the
amount of iron melted in cupolas will decrease, with the amount
by 1980 projected to be as low as 50£ of the total. The electric
methods of melting, both direct arc and induction, will account
for almost 50%, with the small remainder being melted in fuel
fired, reverberatory furnaces.
The tonnage of malleable iron which is produced is expected
to remain relatively constant. On the other hand, the tonnage
of ductile iron has been increasing every year and is expected
to be about double the 1969 tonnage by 1960. This will result
in a corresponding increase in the amount of magnesium treatment
of iron which will be performed.
The trend toward mechanization of molding, pouring and
shakeout lines has been continuing at an increasing rate.
There is every reason to believe that within the next 10 years,
A T KEARNEY 8> COMPANY [NC
II - 12
the majority of iron foundries will use some form of mechanized
molding, pouring, shakeout and sand preparation.
Coremaking has also been undergoing a technological change,
with the trend being away from oil-bonded, baked sand cores,
toward chemically bonded, thermally cured and air cured cores.
The entire area of handling and preparation of materials
used in the foundry is continually being reviewed with respect
to the effect on emissions production and economy of collection.
Iron foundries are being more selective ir purchasing of scrap,
with more attention being given to elimination of combustible
materials from the scrap before purchase. Larger foundries
are installing scrap preparation facilities involving pre-
burning or cleaning to remove materials which will produce
emissions during melting.
Many of the trends toward electric or fuel fired melting
furnaces, mechanized molding, and continuous sand preparation
systems will have important effects in reducing the quantity
of emissions produced, or in making the emissions easier or
less costly to collect. The decisions made in selecting new
equipment are, therefore, being made to an increasing degree
with reference to the effect on reduction of emissions, or on
economy of emissions collection.
The use of a greater amount of pre-cleaned or burned scrap
will reduce the quantity of emissions from combustibles now
charged into furnaces along with the scrap. The trend away
A T KEARNEY & COMPANY, rue
-------
II - 13
from cupola melting and toward electric or fuel fired melting
furnaces will have a significant effect in reducing emissions.
The mechanization of molding, pouring, shakeout and sand
preparation facilities has not reduced the quantities of
emissions which are produced, but has resulted in confining
their production to fixed locations. Thus, it is relatively
simple In many cases to construct enclosures which can be used
as a means of collection of emissions with a minimum of in-
filtrated air.
DEFINITION OF
IRON FOUNDRY
AIR POLLUTION
PROBLEMS
The principal areas in which emissions which are difficult
or uneconomical to collect in the iron foundry have been
identified as iron melting, particularly in the cupola, and
in coremaking. Particulate emissions from all iron melting
operations have been estimated to total 243,000 tons per year,
of which approximately 182,000 have been estimated to be dis-
charged to the atmosphere. Emissions from each foundry source
were evaluated with regard to variables of quantity, particle
size, difficulty of collection, availability and cost of
equipment, economics, and reliability of equipment. A rating
system was developed to identify the areas of highest priority
with regard to additional development work required.
A.T KEARNEY ft COMPANY. INC
II - 14
RECOMMENDED RESEARCH
AND DEVELOPMENT
The gaps between existing knowledge, equipment and
techniques and the necessary levels required to efficiently and
economically control emissions in the iron foundry were
identified in the areas of fundamental knowledge, economics,
materials, and processes. Ongoing research and development
in these areas was identified, as far as information was avail-
able. A group of 10 research and development projects was
recommended to bridge the gap between existing and required
technology. For each of these projects the goals, procedure,
estimated time and costs, and priorities were developed. The
recommended projects are summarized as follows:
No. 1. Controlled stack sampling program of emissions
from cupolas to determine effects of design and operating
variables on the quantity and type of emissions.
No. 2. An extension of Project No. 1 to provide
quantitative data on effect of variables on iron oxide formation
in the cupola.
No. 3. An extension of Project No. 1 to provide data
on relationship between quantity of fine metallic oxides in
cupola gases and opacity of stack plume.
No. 4. Development of fabric materials for fabric
filters, which will be resistant to high temperatures and
corrosive gases.
No. 5. Development of Improved means for efficient
utilization of sensible heat and heat of combustion of cupola
A T KEARNEY & COMPANY INC
-------
II - 15
stack gases.
Ho. 6. Development of continuous Iron melting fur-
nace as a replacement for the cupola, using fuel firing or
electric power for economical, low emission melting.
Ho. 7. Development of economical, centralized scrap
preparation facilities to serve iron foundries.
Ho. 8. Development of core binder materials with low
level of acrid or noxious gas evolution during curing.
Ho. 9. Development of sonic or other method of
agglomeration of fine particles to make them easier to collect.
No. 10. Development of uses for waste products from
iron foundry melting operations.
A T KEARNEY & COMPANY. INC
III - AIR POLLUTION AND THE IRON
FOUNDRY INDUSTRY
DESCRIPTION OF THE
IRON FOUNDRY INDUSTRY
Iron Castings
Production
Annual castings production in Che United States has varied
widely, depending on the economy, with the ranges during the
postwar period as follows:
1 2
Iron Foundry Production '
Production Tons per Year
Type of Metal
All Metals
All Reported Cast
Iron
Cast Iron from
Iron Foundries
Gray Iron
Malleable Iron
Nodular Iron
Minimum
13,200,000
11,032,000
10,000,000
9,340,000
661,000
-
Maximum
20,800,000
17,084,000
14,486,000
11,936,000
1,227,000
1,570,000
Last 5-Year
Average
20,000,000
16,329,000
13,817,000
11,650,000
1,075,000
1,092,000
The complete castings production picture has been shown
graphically in Exhibit III-l. The data, as reported by the
Department of Commerce, included the production of ingot molds.
However, only about 30% of ingot molds are produced from gray
iron melted in cupolas, with the rest being produced from
direct blast furnace metal. Since castings production from
direct metal was excluded from this study by definition, and
additionally has already been covered by the iron and steel
industry study, the estimated portion of ingot mold production
A.T.KEARNEY a> COMPANY. INC
-------
Ill - 2
produced from this source was deducted from total iron castings
production.
Iron castings produced in foundries have accounted for
approximately 691 of total weight of castings production from
all metals. Nonferrous metals have accounted for approximately
8%, steel castings were approximately 10%, and ingot molds
from blast furnace iron accounted for the remaining 13%. From
these figures, it is evident that iron castings are by far the
most important classification of metal cast on a tonnage basis.
The rate of increase of iron castings production has been
found to be closely related to three industries—steel, con-
struction and automotive. Ingot mold castings production
amounts to approximately 8% of iron foundry production and
varies with steel industry production rates. Pressure pipe,
soil pipe and fittings production amount to approximately 201
of total iron foundry production and varies with domestic,
industrial and public construction levels. Miscellaneous
castings production covers the balance of 72% of iron foundry
output and varies most closely with automotive, truck, bus,
and agricultural vehicle production. This last item includes
gray iron, ductile iron and malleable iron castings.
Iron Foundry
Population
The population trends in the foundry industry have been
developed in Exhibit III-2. The total number of foundries of
all types has remained relatively constant during the postwar
III - 3
period, ranging from 5,000 to 5,800 and averaging approximately
5,400. However, the iron foundry population has shown a steady
decline, from 3,200 in 1947, to 1,670 in 1969. If this decline
continues, the number of iron foundries is projected to be approxi-
mately 1,100 by 1980. However, the average size of iron foundries
has been increasing steadily, with average annual production per
foundry going from 3,800 tons in 1947, to 5,300 tons in 1959, and
to 8,700 tons in 1969. By 1980, the average production per iron
foundry is projected to be approximately 16,500 tons per year.
An analysis of the population of iron foundries with
respect to size of foundry has shown that almost the entire
decline in foundry population has been among the small foundries,
with employment of under 100 per foundry. The number of medium-
sized foundries has remained almost constant, while the number
of large foundries has increased slightly. The number of
small foundries had declined by one-third from 1959 to 1969,
and is expected to further decline to only about half of the
1969 population by 1980.
Geographical Location
of Iron Foundries
The distribution of iron foundries by states and by major
metropolitan areas is shown in Exhibit III-3. The highest
concentration is in the states which border on the Great Lakes,
namely, Pennsylvania, Ohio, Michigan, Illinois, Wisconsin,
New York and Indiana. This group of seven states contains
-------
Ill - 4
almost half of all of the iron foundries in the United States
and more than half of the iron castings capacity. The State
of California contains the greatest concentration of iron
foundries in the western half of the country, with one-third
of the iron foundries in that 17-state area being in California.
Other areas of high iron foundry concentration are in the
southeastern states and in the northern states bordering on
the west bank of the Mississippi River.
The location of the principal concentrations of iron
foundries has been pinpointed into major metropolitan areas.
A total of 50 such areas accounts for more than two-thirds of
the iron foundries. Here again, the principal concentration
is in the industrial cities in the seven Great Lakes states,
with two-thirds of these centers being in, or bordering on,
those states. It is of further interest to note that the
decline in number of iron foundries in these areas was a con-
siderably lower rate than for the balance of the country,
emphasizing the fact that the greatest mortality in iron
foundries has been among the smaller companies located in the
smaller metropolitan areas.
The variations which have occurred in the distribution
of iron foundries by states during the period of 1963-1969 are
given in Exhibit III -It.
Iron Foundry
Melting Equipment
Trends in use of iron foundry melting equipment are given
for the postwar period on Exhibit III-5. The cupola has shown
III - 5
the most dramatic change, with the number of cupolas installed
in the United States declining steadily from 4,470 in 1947,
to 1,930 in 1969.3 This decline can be attributed to a com-
bination of several factors:
1. The decline in the number of iron foundries,
particularly the small foundries. For the most part the
foundries which ceased to exist were operators of cupolas.
2. The replacement of cupolas by electric melting
furnaces.
3. The replacement, in many foundries, of two or
more small cupolas by one 'larger unit.
The trend toward decline in the number of cupolas is
expected to continue for the foreseeable future, with the pro-
jected number being reduced to approximately 1,000 by 1980.
Although the reasons given above are expected to continue to
be the principal factors in this decline, other major factors
are most certainly the relatively high cost of installing
effective emission control equipment on cupolas and the 20-year
history of rapidly increasing cost of coke which is the source
of fuel as well as carbon in the cupola.
There have also been changes in the types of cupolas in
operation with an increasing percentage of the new installations
being hot blast units, many of them of the unllned, water-cooled
type. The percentage of new installations which are of the hot
blast type has increased from 24% during Che period of 1947-1956,
to 26% from 1957 to 1962, and to 47% from 1963 to 1968. At the
same time the number of unlined, water-cooled, hot blast cupolas
-------
Ill - 6
has been increasing from its beginnings in the mid-1950's
Co approximately 15% of all cupolas in service in 1969.
Trends in cupola installations have varied with the sizes
of the foundries. The small foundries have invariably installed
hot or cold blase, lined cupolas, but few water-cooled, unlined
cupolas. Medium-sized foundries have installed both lined and
unlined cupolas. The very large foundries have not installed
cold blast cupolas in recent years, although some lined, hot
blast cupolas have been erected. For the most part, the in-
stallations in the large foundries have been water-cooled, un-
lined, hot blast cupolas. These trends also have been greatly
influenced by the method of operation of the foundry. Contin-
uous tap melting for one or two shifts per day is well adapted
to hot blast cupola operation, either lined or unlined. On
the other hand, intermittent, batch melting is more suitably
served by a cold blast cupola.
The trend toward electric melting in iron foundries has
been accelerating rapidly, as shown on Exhibit 1II-6. Although
some scattered electric melting installations existed in iron
foundries prior to the mid-1950's, the great majority of the
installations were made during the period of 1960-1970. The
most recent census of foundries, taken in 1969, has revealed
that there were some 374 electric arc furnaces installed in
176 iron foundries in the United States. However, an analysis
of these installations has shown that more than half of these
foundries also produce steel castings, utilizing many arc
furnaces for this purpose. Since many foundries which produce
III - 7
both iron and steel castings use the same melting furnace for
both purposes, the actual number of arc furnaces used for iron
melting has been estimated to be approximately 200, located in
some 100 foundries. Most of these are used for melting, but
there are about 35 installations in which the arc furnaces are
used for holding, duplexing and superheating of iron.
The number of arc furnace installations for iron melting
has been increasing at a rate of about 15 furnaces per year.
If this rate were to continue, the number of such furnaces
could be expected to reach approximately 350 by 1980. However,
as previously noted, the combined effects of high costs of
installing emission controls on existing cupolas, and the
rapidly rising costs of coke, will in all probability accelerate
the rate of replacement of cupolas by electric melting furnaces.
The number of arc furnaces melting iron will, therefore, be
higher than the straight line projection, possibly in the range
of 400 to 450 furnaces by 1980. For the most part, the electric
arc furnace installations have been replacing cupolas in
existing foundries, although there have been several new -
foundries built in recent years in which arc furnaces were
installed.
In 1969, there also were 495 coreless Induction furnaces
installed in 191 Iron foundries. Approximately half of these
furnaces were in foundries which also produced steel castings.
Since many of these Iron and steel foundries use the same
furnaces for melting both metals, the actual number of these
-------
Ill - 8
coreless induction furnaces which are used for melting iron is
estimated to be approximately 300, located in some 125 foundries.
Most of the coreless induction furnaces are used for melting,
but about 50 of them have been used for holding, superheating
and pouring of iron. The number of coreless induction furnace
installations has been increasing at a rate of approximately
50 per year. As was the case for the arc furnaces, this trend
will probably accelerate, resulting in an estimated 700-800
furnaces in iron foundries by 1980.
In 1969, there were reported to be 489 channel induction
furnaces installed in 134 iron foundries. Although a few of
these are used as melters, most of them are used for holding,
duplexing, superheating and pouring of iron. These have been
increasing at a rate of about 25 installations per year. If
this rate continues, there will be an estimated 600 units in
service by 1980.
Fuel fired melting and duplexing furnaces which are used
in iron foundries generally fall into two classifications:
the reverberatory-type, tilting furnaces used for melting and
the stationary, reverberatory, air furnace used for duplexing
of malleable iron. There has been an increase in the number
of reverberatory melters in recent years, with the number in
service now estimated to be approximately 100, in about 80
foundries. These are increasing at a rate of about SO per year
and may be expected to reach a total of about 250 installations by
1980, if this rate were to continue.
Ill - 9
The reverberatory air furnace was once found in every
malleable iron foundry. However, no new furnaces of this type
are now being built, and the existing installations have de-
clined to about 75, located in about 55 foundries. These can
be expected to continue to decline as they are replaced by
electric furnaces. By I960, the number in service is estimated
to have declined to about 40 units.
Preheaters, a type of fuel fired equipment associated with
electric arc and electric induction melting furnaces, have
found increasing application in the iron foundry. These units
are used to both preheat scrap and to burn off combustible
materials that nay be embedded in or coated on the scrap,
before charging into the melting furnace. The number of such
units Is still relatively few--no more than 50—but their use
is increasing rapidly, and many of the induction melters will
be equipped with preheaters in the future.
Other Equipment
in Iron Foundries
Most of the iron foundries employ a form of sand or
ceramic molding, utilizing a variety of molding equipment
ranging from manual to fully automated and mechanized. The
principal exceptions to this are those foundries producing
cast iron pipe by pouring metal into rotating metal molds and
those using rotating or stationary permanent molds. Although
a large tonnage of iron is produced in the pipe foundries,
there are relatively few such foundries by comparison with the
-------
Ill - 10
total number of iron foundries. There are approximately 35
foundries in this category who produce about 20% of total iron
foundry production.
The remaining iron foundries, with few exceptions, use
sand as a molding material, and are therefore equipped with
sand handling storage and conditioning facilities; mold pour-
ing, cooling and shakeout facilities; casting cooling, cleaning
and finishing facilities; and, in many cases, coremaking
facilities. Because of the wide variety of types, arrangements,
sizes and utilization of the equipment in these areas of the
iron foundry, it is impractical to attempt to classify or
tabulate facilities in all of the foundries, and this has not
been done in this report.
In general, the increasing costs of labor, and the prob-
lems of recruiting and training foundry labor, have resulted
in a trend toward mechanization as a means of reducing the
labor content in iron casting costs. An increasing number
of iron foundries are installing continuous or mechanized
molding lines, sand preparation, casting cleaning, coremaking,
and other facilities. This trend, which increases the facilities
cost in the foundry, has also been a factor in the continuing
reduction in the number of small foundries.
Ill - 11
BASIC CHARACTERISTICS
OF THE IRON FOUNDRY
INDUSTRY
Compilation of
Foundry Data
The large number of iron foundries in existence, reported
to be 1,670 in 1969, and the general lack of detailed informa-
tion on each of these foundries made impractical a complete
detailed tabulation of data on every foundry.
Detailed information on approximately 170 foundries, 10%
of the total number, is compiled in a data bank included in
Appendix B. The nature of the data compiled includes:
1. Coded identification of foundry.
2. Type and number of furnaces.
3. Melt rate of each furnace.
4. Type of air pollution control equipment on the
furnaces.
5. Collection efficiency of the air pollution con-
trol equipment.
6. Types, capacities, and number of molding, sand
conditioning, coremaking, shakeout, cleaning, painting, and
other emissions producing equipment.
7. Emission test data resulting from stack testing
done at the foundry where available.
Because of the confidential nature of some of the informa-
tion which was provided by many companies, all of the foundries
are identified by a code number so that the information is
presented on a basis of non-identification with any individual
company.
-------
Ill - 12
In-Plant Foundry
Visits
Pace of the information for the data bank was developed
from in-plant foundry visits made in 1970 by A. T. Kearney &
Company, Inc. and also in 1967 as a part of a survey made
jointly by the Department of Commerce and NAPCA. A total of
approximately 100 foundry visits was made in these two surveys.
The purpose of the visits was twofold:
1. To obtain emissions data to be used in a mathe-
matical analysis that will identify the effect of certain
design and operating variables on the type and quantity of
particulate emissions.
2. To obtain firsthand reports from users of control
equipment as to the effectiveness, and the operating or main-
tenance problems experienced with different types and manufacture
of equipment.
A careful analysis of iron foundries in the United States
was made to insure that the total emissions data collected
reflected, to the degree possible, the effects of all identified
variables in the melting systems design and operation. The
steps preliminary to the selection of foundries to be visited
were based on an analysis and classification of melting
facilities, which is described in detail in Section VI.
THE AIR POLLUTION
PROBLEM
The pollutants discharged by the iron foundry industry
can be classified as:
1. Emissions from melting furnace operations.
Ill - 13
2. Emissions from other dust-producing operations
within the plant.
3. Odors and gaseous compounds from both sources.
Exhibit III-7 has been prepared to identify the principal
sources of emissions in iron foundries and to indicate the
types of emissions which are produced in each. A more detailed
description of these processes, the emissions produced, and the
methods of collection and control will be found in subsequent
sections of this report.
The physical difficulties of satisfactory collection of the
emissions are not easily solved and, in most cases, costs of
satisfactory collection are quite high. Gases from foundry fur-
naces are hot and must be cooled before collection. If recircu-
lated water is used for cooling and dust collection, corrosion
problems may be introduced. Cost of fresh water is high, requir-
ing recirculation in most cases. Most metallic oxides from melt-
ing operations are extremely small in size and require very effi-
cient equipment for collection. In some cases, condensed metals
present explosive hazards which must be considered.
Particulate emissions have been a point of focus for concen-
trated efforts in the control of air pollution. However, gaseous
emissions and odors from the foundries have not been given much
attention, and the foundry industry now has to take steps to sup-
press these discharges into the atmosphere. Most of the odors xn
foundries result from core baking and shell molding operations,
-------
Ill - U
but the common gaseous emissions also include vapors from melting
oily metal scrap, painting operations, inoculation of metal, and
from metal pouring into molds.
Air pollution control equipment has been generally considered
a nonproductive expense, involving a major investment on the part
of the industry without benefit of payback. The operation and the
maintenance costs can be high and the equipment is generally con-
sidered to have a relatively short life. All these factors have
an effect on the selection of the type of equipment to be used in
a foundry installation and usually require a high level of man-
agement decisions.
These factors, which represent the nature of the air pollu-
tion problem for the iron foundry industry, indicate the need for
a detailed study of the problems facing the iron foundry industry
and development of technology for control of air pollution from
these sources at reduced cost.
Ill - 15
REFERENCES
1. Foundry Magazine. "Foundry Census," July, 1970,
pp. 94-5.
2. U.S. Dept. of Commerce. "Iron and Steel Foundries
and Steel Ingot Producers, Summary for 1968," July, 1970.
3. Foundry Magazine. "Inventory of Foundry Equipment,"
1968, 1960, 1954, 1957.
-------
IV - DESCRIPTION OF BASIC
FOUNDRY PROCESSES
INTRODUCTION
The Iron foundry consists of a number of distinct but
strongly interconnected operations. All foundries utilize
certain basic operations consisting of:
1. Raw material storage and handling.
2. Melting.
3. Pouring into molds.
Other processes present in most, but not all, foundries
include:
1. Molding.
2. Sand preparation and handling.
3. Mold cooling and shakeout.
4. Casting cleaning, heat treating, and finishing.
5. Coremaking.
6. Pattern making.
A simplified, schematic flow diagram encompassing most of
these processes is presented in Exhibit IV-1.
Each operation contains equipment and processes capable
of producing emissions which may include gas, fume, smoke, vapor
and particulate matter. The latter can range from metal dust
from grinding operations, that is relatively easy and in-
expensive to collect, to extremely fine ferrous and nonferrous
oxides from melting furnaces that are very expensive and
IV - 2
difficult to collect. The sources of these emissions are
schematically indicated in Exhibit IV-1, and the operations are
described in the following paragraphs of this report.
PROCESS FLOW FOR
CRAY, DUCTILE AND
MALLEABLE IRON
The basic process flows for gray, ductile, and malleable
iron are shown in Exhibit IV-2. In general the flow is similar
for each of the three types of iron, except for the variations
shown in the exhibit.
The most common flow pattern in the iron foundry industry
is the one for gray iron production. The process flow for
ductile iron differs from gray iron principally by the addition
of two operations—the magnesium treatment to nodularize the
graphite in the iron and the press straightening which is
sometimes required.
Malleable iron process flow is also similar to gray iron
flow with the exception of an annealing operation required to
convert the as cast "white" iron into malleable iron, and the
press straightening sometimes required to correct the warping
that results from the annealing process. In other regards the
process flow for malleable iron is the saute.
Specifications for various classes of gray, ductile, and
malleable iron are tabulated in Exhibits IV-3, IV-4, and IV-5.
-------
IV - 3
PROCESS FLOW AND DESCRIPTIONS
OF BASIC FOUNDRY OPERATIONS
Detailed process flow diagrams for each of the basic iron
foundry production operations are depicted in referenced exhibits,
except for cupola melting. Heat and material balances for cupolas
are included in Appendix C with a description of the mathematical
model developed for calculating the heat and material balances.
Brief descriptions of the operations are given on the following
pages. More detailed information on the basic operations can be
obtained from referenced sources.
Raw Material Storage
and Furnace Charge
Preparation
(a) Raw Material
Receiving and Storage
The raw materials used for iron production fall into Che
following classifications:
1. Metallics:
(a) Pig iron.
(b) Iron and steel scrap.
(c) Turnings and borings (loose or briquettes).
(d) Ferroalloys.
(e) Foundry returns.
(f) Alloys such as nickel, copper, tin and
magnesium.
2. Fluxes:
(a) Carbonate type (limestone, dolomite, soda
ash).
(b) Fluoride type (fluorspar).
(c) Carbide type (calcium carbide).
IV - 4
3. Fuels-Coke (for cupolas).
4. Refractories.
These materials, except for foundry returns, are received
by railcar or truck, and are unloaded and stored in the foundry
scrapyard. Although open scrapyards are still common, the use
of covered storage areas is becoming more widespread to reduce
che rusting of ferrous scrap and the weathering and degradation
of coke and limestone. Covered storage is also desirable for
induction melting scrap which must be dry when charged into the
furnace. Scrap metals, pig iron, and foundry returns are usually
stored in piles which may be unconfined, but are often separated
by walls to form open bins. Coke and stone may also be stored
in piles, but are commonly stored in confined bins to facilitate
charge makeup.
There are no industry standards for foundry coke. Size
classifications up to 6" x 8" are available depending on the
requirements of each foundry. The range of proximate analysis
values for foundry coke are as follows:
Volatiles 0.77. - 1.257.
Ash 4% - 12%
Fixed Carbon 867. - 95%
Sulfur 0.5% - 1.0%
Specifications for ferrous scrap metals are tabulated in
Exhibit IV-6. Pig iron and commonly used ferroalloy specifica-
tions are shown in Exhibit IV-7.
-------
IV - 5
(b) Scrap Preparation
For Che most part, scrap materials, including foundry
returns, are used In the as-received form. Many scrap dealers
prepare scrap to foundry specifications, thus eliminating any
need for additional preparation by the foundry. Where scrap
preparation la required, the operations may involve any com-
bination of the following:
1. Cutting to size by flame torch, shear, breaking,
or fragmentizing.
2. Cleaning by degreasing, by steam or by shot-
blasting.
3. Burning of surface coatings or oils.
A. Drying or preheating.
With the exception of the cutting operations, scrap
preparation is not widely performed for cupola melting, or for
top charged electric arc furnaces. However, for electric in-
duction furnaces, a greater degree of preparation is necessary
to obtain dry, clean scrap of Che proper size.
(c) Furnace Charge
Preparation
The methods of makeup and handling of melting furnace
charges vary widely from completely manual systems where all
materials are hand shoveled or carried, to highly mechanized
systems where one man can control the handling, weighing, and
loading of all rav materials.
Charges are normally loaded directly into the furnace
charging bucket, skip, or similar container. The prescribed
IV - 6
combination of metallics, flux, and coke (for cupolas) is weighed
either before loading or while loading.
Where preheating or drying operations are performed, they
are commonly done directly in the charge containers. This can
be accomplished by heating from above with a flame or radiant
burner, or by use of a double-walled bucket in which combustion
is accomplished within the wall cavity. Other methods include
the use of rotary dryers, heated conveyors, and preheat furnaces.
Process flow for scrap preparation and change makeup is shown
in Exhibit 1V-8.
Iron Melting
The process flow in the foundry melting department is
depicted in Exhibit IV-9. Four types of melting furnaces rep-
resenting over 98% of the installed melting systems are
considered in the flow diagram. A 1968 study, covering about
75% of all foundries, showed the following distribution of
melting furnaces in the iron foundry industry.
Iron Foundry Melting Furnaces - 1968
Furnace Type
Cupola
Electric Induction
Electric Arc
Other Types
Total
Number Installed
1,232
73
42
29
1 .376
Percent of Total
89.5%
5.3%
3.1%
2.1%
100.0%
-------
IV - 7
The furnace census does not show the number of reverbera-
tory furnaces In use, but it would be expected that they
constitute the majority of the 2.U indicated as "Other Types."
Despite the low incidence of use, this method of melting is of
interest because of its reported low emission of particulate
matter, and its increasing use in small foundries.
The process flow diagram shows the metallic and flux
materials charged into each type of melting furnace. The
reverberatory furnace is heated by coal, natural gas or oil,
while the induction and arc furnaces obtain their heat from an
electric induction coil or an electric arc. In the cupola,
coke is a portion of the furnace charge and the heat required
to melt the iron is derived from the combustion of the coke
in contact with the metallic and fluxing charge materials.
The molten metal from the furnaces is tapped directly into
a ladle for direct mold pouring, into a transfer ladle for
conveying to the molding area, or into a heated or unheated
holding device.
From a cupola, the hot metal can be transferred to a
forehearth, or an electric induction or arc furnace for holding,
or into a reverberatory duplexing furnace if the final product
is to be malleable iron. With the exception of the forehearth,
a similar selection of holding or duplexing furnaces is avail-
able for electric or fuel fired melting furnaces. As the metal
is available and required, it is removed from the holding or
duplexing furnaces for transfer to the molding area.
IV - 8
(a) Cupola Furnaces
The cupola is a vertical furnace with a normally circular
cross section which is charged alternately with metal to be
melted, fuel in the form of coke, and a fluxing material, to
produce molten iron of a specified analysis and temperature.
Many fundamental cupola designs have evolved through the years,
two of which are widely in use at this time--the conventional
refractory lined cupola and the more recent development, the
unlined, water-cooled cupola.
1. Conventional Lined Cupola. For all cupola designs,
the shell is made of rolled steel plate. In the conventional
design, an inside lining of refractory material is provided to
protect the shell from the operating temperature. The cupola
bottom consists of two semicircular, hinged steel doors,
supported in the closed position by props during operation, but
able to be opened at the end of a melting cycle to dump the
remaining charge materials. To prepare for melting, a sand bed
6 to 10 inches deep is rammed in place on the closed doors to
seal the bottom of the cupola.
Combustion air is blown into the wind box, an annular
duct surrounding the shell near the lower end, from which it
is piped to tuyeres or nozzles projecting through the shell
above the top of the rammed sand. The tap hole through which
the molten iron flows to the spout is located at the level
of the rammed sand bed. For nearly all continuous tap operations,
the slag also is discharged through the tap hole and separated
from the iron by a skimmer in the runner. For intermittent
-------
IV - 9
capping, molten Iron collects in the well with the slag float-
ing on its surface, and a slag hole is located at the level
representing the height of the maximum amount of iron desired
to collect in the well. An opening is provided in the cupola
shell 15-25 feet above the bottom plate for charging the cupola.
The charging door opening varies in size according to the in-
tended method of charging and the diameter of the cupola. The
upper stack is extended sufficiently to pass through the building
roof and provide the required, natural draft. On open top cupolas,
a spark arrester is fitted to the top of the stack to reduce
the hazard of fire.
The shell of a conventional lined cupola is protected
with a refractory lining. The lining is usually built up using
high-duty fireclay shapes for acid melting and magnesite shapes
for basic melting. An Illustration of a typical conventional
lined cupola is depicted in Exhibit IV-16.
2. Water-Cooled Cupola. The unlined, water-cooled
cupola utilizes a steady flow of cooling water on the outside
of the unlined shell generally from below the charging door to
the tuyeres, and an inside lining of carbon blocks below the
tuyeres to the sand bed, to protect the shell from the interior
temperature. Conventional lining is used at the charging
door level and in the upper stack. In other regards, the water-
cooled cupola is essentially the same as the lined cupola.
Exhibit IV-17 shows two types of water-cooled cupolas.
3. Blast Air. Combustion air is supplied to the
cupola in appropriate amounts depending upon the desired melting
rate and the size of the coke charge, at a pressure related to
IV - 10
the volume and height of the burden. The air is piped from the
blower, which can be a centrifugal or positive displacement
type, to the wind box, and thence to the tuyeres, spaced equally
around the cupola shell. In a normal design, a single row of
tuyeres is used. In the balanced blast cupola, the blast air
is divided between three rows of tuyeres with the provision for
adjusting the flow of air through each tuyere. Varying the flow
in a regular sequence prevents bridging of the charge at the
tuyere and results in greater combustion efficiency and higher
melting rate.
The temperature of the blast air has a strong effect
on coke requirements, melting rate, and cost of melting. The
heat supplied by warm or hot blast decreases the BTU's required
from the coke and therefore permits the coke charge to be de-
creased, or the melting rate to be increased. Where the cost
favors natural gas or oil over coke, the heating of the blast
air with these fuels can lower the cost of melting. Early
efforts at heating the blast air resulted in temperatures of
300°-400° F. Later, temperatures of 1000°-1200° F were used
successfully and, recently, a successful installation with
1800° F blast air has been reported.
Most air preheater installations in the United States
are externally fired, although many attempts have been made to
successfully recover the latent and sensible heat of the cupola
with varying results. Some of the problems in developing
successful units were maintaining the heat exchanger, keeping
the heat transfer surfaces clean while passing dirty gas through
-------
IV - 11
them, and designing a system low enough in cost to make the
installation economically feasible.
4. Fuel Injection. Numerous attempts have been made
to inject coal, oil, or gas into cupolas to replace a portion
of the coke required to melt the iron. Successful results have
been reported for natural gas injection resulting in lower
melting costs. No tests using coal or oil have been reported
in recent years.
5. Oxygen Enrichment. A number of foundries are
injecting oxygen into the cupola with resulting increased melt-
ing rates. One potential advantage of this technique is that
the stack gas volume is decreased, since the oxygen used for
combustion does not carry with it the inert volume of nitrogen
found in air.
6. Closed Top Cupolas. In either recuperative hot
blast cupolas, or where there is a requirement that cupola
stack gases be cleaned before releasing them into the atmosphere,
the normal practice is to close the top of the cupola and draw
the gases off either just below the cap, or below the charging
door. If collection is above the charging door and the door
is always open, air is infiltrated through the opening sub-
stantially increasing the volume to be handled, and therefore
increasing the cost of the control system. If the door is
normally closed, the infiltrated air volume will be lower but
oxygen present in the gases may be insufficient to bum the CO
to CC^. A compromise is sometimes reached where doors are
provided that permit infiltration of enough air to aid in
complete combustion of the CO.
IV - 12
7. Afterburners. The cupola gas rising from the
top of the burden is rich in carbon monoxide. When mixed with
air infiltrated through the charge door opening, this gas
becomes a combustible mixture and it is desirable that it be
ignited at this point. Combustion normally occurs here because
of the gas temperature, but the flame is sometimes extinguished
when the charge is discharged from the charging bucket. To
assure reignition, a gas torch, or afterburner, is often
installed in the cupola just above or below the charging door.
The afterburner performs the additional function of incinerating
oil vapors, coke particles and other combustibles.
8. Cupola Operation. At the beginning of the melt-
ing cycle, a coke bed is placed on the rammed sand bottom and
ignited, preferably with a gas torch or electric igniter.
Additional coke is added to a height of four or five feet above
the tuyeres after which regular layered charges of metal,
limestone and coke are placed on the coke bed up to the normal
operating height. Typical cupola charges for gray, ductile,
and malleable iron are included in Appendix C. The air blast
is turned on and the melting process begins. As the coke is
consumed and the metal charge is melted, the furnace contents
move downward in the cupola and are replaced by additional
charges entering the cupola through the charging door.
As the metallic charge moves downward, it is preheated
by the hot gases resulting from combustion. These gases con-
sist of carbon monoxide, carbon dioxide, nitrogen, hydrogen,
and sulfur dioxide, the latter principally from the sulfur
-------
IV - 13
contained in che coke. Aa the metal enters the melting zone, the
atmosphere is of a highly reducing nature with no free oxygen.
The molten iron, trickling down over the incandescent coke and
increasing in temperature, reaches the combustion zone where the
atmosphere becomes oxidizing in nature, due to the presence of
oxygen from the blast air. In this part of che cupola, the
iron reaches the desired tapping temperature.
Blast air entering the cupola through the tuyeres
contains 21% oxygen which combines quickly with the carbon in
the coke as follows:
C + 02-» C02 + 175,900 BTU/pound mole
The oxidizing zone in which this reaction occurs is designated
the combustion zone. It is the zone of highest temperature and
extends from the tuyeres to a level where the following reaction
occurs:
C +• 2 -» 2 CO - 69,700 BTU/pound mole
The reduction of C02 to CO starts before all oxygen in the
blast air is consumed. The maximum COj concentration is be-
lieved to be approximately 147.-187. at the boundary of the
oxidation and reduction zones at a maximum temperature of 2,800°
-3400° F. Both reactions noted are reversible and proceed in
both directions depending upon conditions at different levels.
The reactions almost cease in che preheat zone as energy is
used to preheat the charge materials and the gas temperature
ia lowered to the reaction temperature below which further
reduction of carbon dioxide to carbon monoxide will not occur.
A pictorial description of a cupola reaction area is shown in
Exhibit IV-IB.
IV - W
'• Cupola Furnace Operations Model. Approximately
300 iron foundry melting systems in the United States are
currently equipped with some type of air pollution control
equipment, ranging from wet caps to fabric filters, wet
scrubbers, and electrostatic precipitators. A detailed investiga-
tion of about 170 of the cupola systems included in this total
showed that they could be characterized by only 32 classifica-
tions, each one representing a specific combination of cupola
design factors.
In Appendix C, the heat and material balances for each
class are tabulated and the classifications are depicted
pictorially. The heat and material balances were derived by
the use of a computerized mathematical model designed to cal-
culate these balances for a given set of cupola operating
conditions. Input data were obtained from the records of about
50 foundries visited in 1970 and from test data In the files
of various city and state air pollution control boards. Details
of the computer program developed to make the calculations are
included in Appendix C.
In addition to the description of the heat and
material balance model, Appendix C includes the following data:
~ Specific Inputs Required for the Model.
Input data required tor the calculations
are tabulated.
- Chemical Reactions Considered in the
Model.Chemical reactions considered to
have a qualitative and quantitative effect
on emissions produced in a cupola used
to produce iron are listed.
- Material and Heat Balance Relationships
Considered in the Model'. Theoretical and
empirical material balance relationships
upon which the model is based are listed.
Similar relationships for the heat balance
calculations are shown.
-------
IV - 15
Sample Outputs for Material and Heat
Balances. Sample outputs for material
and heat balances are included in the
appendix. The appendix indicates the
format of the computer output. Material
balance inputs are acceptable either in
pounds of charge materials or a percentage
of total charge weight. The model cal-
culates the balance and prints the outputs
in pounds per ton of molten iron.
Heat balance outputs are printed out in
thousands of BTU s per hour. The pro-
gram also computes the volume of the top
gases, and a theoretical analysis of the
stack gases including air infiltrated
through the charging door if the input
indicates that the cupola is operated
with the door open.
Composition of Materials. To simplify
the computation, the model makes use of
standardized material compositions for
metallic charge materials and products,
fluxes, coke ash, and cupola refractory
linings.
(b) Electric Arc
Furnace
1. Description. The direct arc electric furnace
consists of a refractory lined, cup shaped, steel shell with
a refractory lined roof through which three graphite or carbon
electrodes are inserted. The shell is arranged for tilting to
discharge the molten metal. Charging of the metal is accom-
plished by chuting through a door opening in the side of the
shell, for fixed roof furnaces, or by raising the roof and swing-
ing it aside to permit the use of a bottom dump charge bucket
for removable roof furnaces. Foundry size furnaces generally
range in diameter from about 3'0" to 12'9" with holding capa-
cities of 500 pounds to 25 tons, and melting rates from 250
IV - 16
pounds to 12 tons per hour. Larger furnaces have recently been
installed for melting iron, one large production foundry having
recently put into service a 65-ton capacity arc furnace with
a melting rate of about 22 tons per hour. For iron melting
with light scrap, the inside depth of the furnace is approximately
757. of the inside diameter. Furnace shells are of welded con-
struction, often with a water-cooled section at the top edge.
The furnace roof is provided with a circumferential water-cooled
ring, and water-cooled electrode glands. Tilting mechanisms
are mechanical or hydraulic, the latter being of later and more
satisfactory design from the standpoint of precise controllability.
The furnace is provided with a tapping spout and a working door,
with a second door for large furnaces.
The electrodes, supported by water-cooled holders,
are provided with a hydraulically or mechanically operated
raising and lowering mechanism to maintain the arc, especially
during meltdown. The electrodes and roof on a top charge
furnace are supported Independently from the shell and are
arranged to permit raising to clear the shell and rotating to
one side for charging.
A transformer vault adjacent to the furnace is re-
quired to house the current and potential transformers, circuit
breakers, high and low voltage wiring, low voltage bus bars,
secondary furnace cable connections, and instrument panel. A
control panel and operator's console are required near the
furnace. Multi-voltage, tap charger transformers are used to
-------
IV - 17
make voltage changes as necessary during the melting cycle.
The furnace bottom refractory lining can be either
acid with a rammed ganister mix over silica brick, or basic
with a rammed magneaite mix over silica brick, although acid
lined furnaces are generally used for melting iron. Sidewall
lining is generally a high alumina fire brick. The bottom
lining is subjected to abrasion and crushing during charging,
but commonly remains serviceable for months. The sidewall is
subjected to both abrasion and the effect of the molten iron
and slag and requires patching between heats. The roof lining
made of high alumina brick is subjected Co high temperature and
effects of metal and slag splatter from the arc. A typical
electric arc furnace with supporting equipment is illustrated in
Exhibit IV-19.
2. Electric Arc Melting. For top charge furnaces,
the charge materials are usually weighed into bottom dump buckets.
When charge preheating is used, the heating may be accomplished
in a specially designed, refractory lined bucket, a double
walled bucket, or the charge transferred to a preheating chamber,
and then moved to the furnace. A cost-saving potential exists
in preheating the charge in the charge bucket, by substituting
a lower cost fuel to raise the charge temperature to 1200°-
1400° F.
In charging a side charge furnace, the metal charge
is often weighed Into a trough-shaped bucket, which is trans-
ferred to the furnace and tilted to permit the charge materials
to slide into the furnace door opening. When using the arc
IV - 18
furnace as a hot metal holder, the hot metal from the primary
melter can be transferred to the arc furnace by ladle, or by
refractory lined launder, and discharged into the charging
opening.
Exhibit IV-10 shows the composition of a typical
charge, heat, and material balance. The carbon required in
addition to the carbon content of the scrap is derived from
the petroleum coke added to the charge and from the electrode
material consumed during the melting cycle. Since there is no
coke ash requiring fluxing as in the cupola, only small
quantities of flux materials are needed. Chemical reactions
occurring within the furnace also are few compared to the
cupola, since there is little combustion. The fume generated,
particularly at the start of meltdown, is the result of locally
intense heat caused by arcing of the electrodes before a molten
pool of metal is formed. Oxides of iron and other elements
contained in the charge materials are formed at this time,
4
by vaporizing some of the charge.
3. Indirect Arc Furnace. The indirect arc or rocking
arc furnace is more commonly used for nonferrous and high alloy
melting than for iron melting, but is found in many smaller
iron foundries. It consists of a refractory lined horizontal
drum which rotates about its longitudinal axis for charging
and tapping. One electrode is located in each end, with the
resultant arc being struck above the charge, resulting in
melting from an indirect arc as contrasted with the direct arc
-------
IV - 19
in the more commonly used arc furnaces. These furnaces are
builc only in small sizes, from a few hundred pounds Co four-Con
capacities.
(c) Induction Furnaces
1. Description. The induction furnace is a cup- or
drum-shaped vessel that converts electrical energy into heat to
melt the charge. Unlike the electric arc furnace, no electrodes
are required. The induction furnace converts electrical energy
into heaC by utilizing Che transformer principle in which a
magnetic field is set up when the primary coil of the transformer
is energized. The magnetic field at a high flux density induces
eddy currents in the charge which are converted to heat by the
electrical resistance of the charge itself.
The heat is developed mainly in the outer rim of the
metal in the charge, but is carried to the center by conduction
until the metal is molten. The electrical energy is converted
into heat by induction in two ways. In the channel induction
furnace, the metal charge surrounds the transformer core,
thereby forming a loop or channel. In the coreless induction
furnace, the metal heated is both the core and secondary coil.
Furnace coils are water-cooled to prevent heat damage.
The amount of heat developed in the furnace is a
function of the electrical current frequency and the intensity
of the magnetic field. Theoretically, there is no limit to the
temperature attainable; however, a practical limit exists due
to material limitations such as refractory lining of the furnace.
IV - 20
At present, commercial linings are available for temperatures
somewhat over 3000° F.
2. Operating Characteristics. The induction furnace
lends itself to either semicontinuous or batch-type operations
and can be used as a melting furnace or for holding or duplex-
ing operations. Generally, the coreless furnace is better
adapted to melting and superheating, whereas the channel furnace
is better suited to superheating, holding, and duplexing,
although it is also sometimes used for melting. Exhibits IV-20
and IV-21 illustrate channel and coreless induction furnaces.
Coreless furnaces are available in high, medium and
low frequency types. Installations for iron melting and hold-
ing are generally limited to the low or line frequency of 60
cycles, and medium frequencies of 180 cycles.
The coreless design permits melting of a cold charge for
each cycle, allowing frequent analysis change. The energy is
constantly induced into the entire charge so that temperature
differentials within the melt are minimal. This is not the
case in a channel induction furnace where energy is induced
only in the molten metal actually in the channel. Despite
the ability of the coreless furnace to melt cold charges, most
production foundries maintain a heel of molten metal in the
furnace to increase efficiency and lower thermal shock to the
refractory lining. The scrap charge must be dry for this
method of operation to prevent explosions that would result
from addition of moisture to a molten metallic bath.
-------
IV - 21
3. Charging. Melting losses are low and the recovery
of alloying additions Is high In induction melting of gray iron.
The charge may consist of a single lump of metal, a number of
small pieces of selected scrap, or even turnings. Charges are
usually made up of steel scrap, cast iron scrap, foundry returns,
and ferrosilicon and carbon in suitable quantities to adjust
composition. Pig iron is seldom used. Turnings and borings
are readily melted in a coreless furnace, although care must
be taken to avoid oxidation before melting has occurred.
When available, turnings or borings can comprise up
to 20% of the charge in a coreless furnace. Carbon adjustment is
accomplished by the addition of pelletized coke. Preheating of
the charge is considered beneficial from the standpoint of re-
moving moisture and oil as well as reducing melting costs and
shortening the melt cycle.
Charge materials are commonly weighed into a bottom
dump charging bucket, although small furnaces can be charged
manually. In some cases, the charge is collected in pans placed
on the working platform and is dropped into the furnace either
by tipping the pan or by raking the pieces out of the pan and
into the furnace. Charging into the furnace is from the top.
Additions of carbon and ferrosilicon are made directly into
it
the furnace.
4. Melting. The induction furnace is basically
a batch melter. Continuous melting is possible with a recently
developed coreless furnace and approximated in the conventional
design by partial tapping of a melt followed by a makeup charge
of cold scrap.
IV - 22
A strong stirring action from electromagnetic forces
is exerted on the molten bath in a coreless furnace. The
stirring action helps to obtain a homogeneous melt when ladle
additions for carburizing and analysis correction are made.
The mixing action is less in a channel-type furnace.
The analysis of the iron from induction melting is
highly reproducible due to the general lack of unpredictable
or undesirable reactions in the furnace. Induction melting
does not alter the chemical composition of the molten metal
from the original charge. However, since it is not a refining
process, care must be exercised in the makeup of the charge.
The process flow for coreless induction melting is shown in
Exhibit IV-11.
Induction furnaces are more efficient superheaters
than cupolas or arc furnaces. For this reason, high tapping
temperatures are more practical with coreless and channel
induction furnaces, and these units are often used for super-
heating, and where hot metal must be held for an extended period
of time, induction furnaces have been used for feeding automated
casting lines with the metal being tapped as required, and
often held at pouring temperature when molding is delayed for
some reason.
5. Furnace Sizes. Coreless induction furnaces are
found in an extremely wide range of sizes, from under 100 pounds
capacity, to recently designed installations of 70 tons and
larger. Similarly, channel induction furnaces vary widely in
size, from a few hundred pounds to 260 tons in holding capacity.
-------
IV - 23
Power inputs range from under 100 kw, Co 17,000 lew for coreless
furnaces, and up to 4,000 kw for channel furnaces.
(d) Reverberatory
Furnace
The reverberator? type of fuel fired furnace is found in
two types of applications in the iron foundry. The large,
stationary reverberatory or air furnace is associated with
malleable iron foundries, where it has long been used as a
duplexing unit in conjunction with a cupola. These furnaces
are generally powdered coal fired, although oil and gas are
also used. Host are not melters, but are used to receive
molten iron from the cupola, and to refine and superheat it
for pouring. These furnaces are long, rectangular units with
arched or suspended roofs, generally fired from one end, and
with waste gases exhausting into a stack from the opposite
end. Temperatures of 2900° F and higher are reached in these
furnaces. Holding capacities range up to 40 tons. * A
typical furnace of this type is shown in Exhibit TV-22.
the second type of reverberatory fuel fired furnace is
used for melting. It is generally small in size, up to two
tons capacity, and tilts for pouring. Furnaces of this type
are found in small foundries, where economical installations
and low emission melting are desired.
Inoculation
Inoculation is a process applied either in the production
of -ductile iron, where it is known as magnesium treatment, or
IV - 24
to improve the mechanical properties of castings.
(a) Ductile Iron
Production
In the production of ductile iron, agglomeration of
spheroidal graphite is accomplished by treatment of the base
iron while in the ladle. A number of inoculating agents in-
cluding magnesium, calcium, cerium, and yttrium are suitable,
although magnesium is most widely used. The relative economics
and availability of these materials have resulted in the common
use of magnesium for this purpose. Many techniques have been
developed to place the inoculant in the molten iron using
magnesium in the pure form, in a variety of alloys, and im-
pregnated in coke. No single form has been universally accepted,
although nickel-magnesium, copper-magnesium, magnesium-silicon-
iron, nickel-silicon-magnesium alloys, and magnesium impregnated
coke are frequently employed. The speed and violence of the
reaction is due in part to the concentration of magnesium in the
alloy or coke. Some of the treatment techniques used to produce
g
ductile iron are shown in Exhibit IV-23.
(b) Improvement of
Mechanical Properties
The inoculation process is applied to molten gray iron
to obtain a desired shape, size, and distribution of graphite
in the finished casting. The effect of an inoculant depends
on the type and amount used, the temperature and condition of
the molten iron at the time of addition, and the amount of time
that elapses between inoculation and pouring.
-------
IV - 25
Inoculants contain various amounts of carbon, silicon,
chromium, manganese, calcium, titanium and other elements.
Inoculant effects include decreasing chill depth and increasing
tensile strength.
An inoculant can be introduced Co the molten metal in the
cupola spout, forebearth, transfer ladle, pouring ladle or mold.4
Duplexing
Where large tonnages are needed, or where a continuous
supply of molten iron Is required to pour conveyorized molding
lines, duplexing melting systems are often employed. In duplex-
ing, iron is melted in one furnace, usually a cupola, and refining
and temperature control are achieved in a second furnace. The
melter is continuously tapped into the second furnace (an air
or electric unit) which is then tapped intermittently into
transfer ladles for pouring into molds. The second furnace
accomplishes carbon adjustment and serves to decrease analysis
variations of metal coming from the cupola. Duplexing is also
used In the production of malleable iron.
Molding, Pouring,
Cooling and Shakeout
Many molding materials and types of equipment have been
developed for the production of iron castings. Techniques found
in current practice include green sand, dry sand, shell, hot box,
the full mold process, permanent metal molds, and the Rheinstahl
process where stationary molds are used with both lined and un-
lined rotating molds for centrifugal casting.
IV - 26
Exhibit IV-12 illustrates the process flow for molding,
pouring, and shakeout.
(a) Green Sand
Molding,
Of all the types of molding in use today, by far the
greatest tonnage of castings is produced from green sand molds.
They are commonly made of silica sand, clay, water, and organic
binders such as cereal binders mulled together to form a moist
mixture suitable for use with manual or automated molding
machines and sand slingers. Green sand molds are the least
costly of all molds to produce.
A typical molding operation is done with the aid of jolt-
squeeze draw machines. The complete operation is often performed
in two separate machines. A jolt-squeeze-rollover-draw machine
for drag molding and a jolt-squeeze-draw machine for cope mold-
ing are used. Holding machines capable of molding cope and
drag simultaneously in one molding cycle are also utilized.
Mold production by a single machine can also be performed
by match plate molding machines or on an automatic molding line.
In automatic molding lines, the entire green sand molding
sequence is automatic with preset time cycles.
Very large green sand molds, beyond the capacity of con-
ventional molding machines, are often made with a sand slinger.
This machine propels small amounts of sand Into the mold with
enough force to eliminate the need for machine jolting.
Green sand molds are occasionally skin dried or air dried
to produce better finish. Drying removes moisture or other gas
-------
IV - 27
forming materials In che sand at the mold-metal interface.
Molds which have to be akin dried are either faced with sand
mixtures containing thennosetting additives or they are washed
or sprayed with a refractory mold coating to prevent metal
penetration.
(b) Centrifugal
Casting
Cast iron pressure and soil pipe is normally made in
permanent molds. The molds are rotated rapidly about their
longitudinal, horizontal axis during che pouring and cooling
cycles producing a dense, high quality casting. A dry sand or
shell sand core is inserted in the mold to form the bell end of
the pipe only. Cast iron pipe tannage represents approximately
20% of the total iron casting production. Centrifugal casting
is also used for production of other types of castings, generally
of a symmetrical design.
(c) Dry Sand
Molding
Dry sand molding is usually applied to thick section
castings of fairly large size or weight such as machine beds,
cylinders, and heavy gears. The sand mixes for this type of work
contain certain additives such as pitch, sodium silicate,
gilsonite, cereal, molasses, dextrine, gluten, and resins.
These additives set at the drying temperature employed, or by
chemical reaction with curing agents, and produce a high dry
strength and rigid mold walls.
IV - 26
In general, dry sand molding Involves assembling the mold
through the use of a number of parts of the mold which have been
prepared using a dry sand mix. These smaller mold parts are
prepared in the same type of equipment as used in green sand
molding where the dry sand mix is poured into the permanent metal
mold and compressed to form a section of the larger mold. These
sand molds are then baked in an oven where drying or baking takes
place at temperatures of 300°-600° F. After cooling, the baked
dry sand molds are assembled to form the complete mold for the
casting.
(d) Shell and Hot
Box Molding
The shell molding process for making castings is used
where dimensional accuracy and surface finish of the part is
important. This molding technique has found wide acceptance
in the foundry industry. The sand used consists of a mixture
of dry sand grains and synthetic resin binder. The resin muse
be a thermosetting type since the strength developed by Che mold
depends entirely on the strength of the binder after the mold
has been heated.
The equipment used for molding consists of a metal box
which houses the pattern plate and is heated to a temperature
of 350° to 700° F. The sand mix is then blown into this hot
box where the resin partially thermosets and builds up a coherent
sand shell next to pattern. A mold release agent or a parting
agent is necessary so that the ejector pins can push the shell
-------
IV - 29
off the patterns. The cleaned shell halves, thus produced, may
be assembled and poured. Shell molds may require cores and the
9
core setting is done prior to the assembly of shell halves.
(e) Full Mold
Process
The full mold process is a patented recent development which
Involves the use of materials such as polyurethane or polystyrene
foam to fabricate patterns where small runs of complicated
designed castings are to be made. In this process, the pattern
is either machined from a single block, or indivual pattern pieces
are cut and then glued or fastened together. These patterns, with
attached gating system, are then buried in sand. Various cold
setting sand mixes are commonly used. Some work has been done
with unbonded sand. Little success has been achieved with green
sand since the ramming pressures deform the patterns. Metal is
poured into the mold, vaporizing the pattern and replacing it
with metal. An advantage of full molding is that pattern design
does not have to consider pattern removal from the sand since
each pattern is destroyed in the pouring process.
(f) Rheinstahl
Proceaa
Another recent development, the Rheinstahl process, con-
sists of using cement as a binder and a fluidizing agent that
promotes a high density casting mix that results in a cement-
bonded, fluid molding sand. The fluidizer is added to the sand
as it is being prepared in a muller or paddle-type mixer. The
IV - 30
reported advantage of the Rheinstahl process is that ramming
time is practically eliminated since pouring can occur almost
10
immediately after sand is placed in the mold.
(g) Pouring and
Cooling
In a high volume production foundry, finished molds are
set out on continuous car-type or indexing mold conveyors, or
on pallet conveyor lines. In smaller production and jobbing
foundries, completed molds are set out on the foundry floor
or on gravity roller conveyors.
In mechanized production foundries the molds are usually
poured within minutes of completion. Most small foundries,
however, may pour only a few hours a day, or only two or three
days a week. In these cases the molds are usually set out on
the foundry floor until hot metal is available. A pouring
station in a mechanized foundry is illustrated In Exhibit IV-24.
(hi Shakeout
In many foundries the casting is separated from the mold-
ing sand manually. Even in partially mechanized foundries where
completed molds are set out on gravity roller conveyors for
pouring and cooling, the molds are dumped manually and the
castings picked or hooked out of the sand and placed in a tote-
box for further cooling and moving to the cleaning room.
Larger and more mechanized foundries tend to use a heavy-
duty vibrating screen for shaking-out. The sand flows through
-------
IV - 31
the screen openings Co Che return, or shakeout sand system, for
transfer Co Che sand condicloning system. The cascings are
removed from the shakeout grid manually, by a hoisC, or by action
of Che vibracing screen, for cooling and sorting. Exhibit IV-2S
shows a shakeout work sCat ion.
Sand Conditioning
Sand is a basic foundry material used in substantial volume
by almost all foundries as a molding medium. The complete sand
system consists of receiving and storage facilities, molding
sand preparation, and the handling equipment required to transfer
the sand to its points of use, preparation, and discard. In
addition, some high production foundries using very large amounts
of sand have reclamation systems for returning used sand to its
original condition, or as near to it as is practical and
economical.
The sand conditioning process flow diagram is given in
Exhibit IV-14.
(a) Receiving and
Storage
Raw materials requiring handling and storage after re-
ceiving include the silica sand and various additives such as
clay, carbonaceous materials, synthetic resins, and cereal
binders. In addition, some foundries doing shell molding pur-
chase premixed shell sand. Silica sand is received by car or
truck, unloaded and conveyed manually, mechanically, or
pneumatically, and stored in covered areas or in enclosed bins
or silos. Shell premix may be shipped in bulk or in multiwall
IV - 32
bags depending on usage. Solid additives are received in bags
or fiber drums, and liquid additives in metal barrels.
(b) Sand Preparation
In past times, naturally bonded molding sand was univer-
sally used for green sand molding. Preparation of the sand for
reuse consisted of adding water and some new sand to replace
the amount sticking to, or embedded in, the surface of the
castings. This sand has been almost completely replaced by
clean silica sand to which is added water and the various
binders required to produce the desired molding characteristics.
Sand preparation in the majority of today's foundries is
accomplished on the foundry floor by adding moisture, binders,
and new sand to the used sand. A number of types of mobile and
portable equipment are used to cut, screen, mix, and aerate the
sand to prepare it for reuse.
Mechanized foundries are generally equipped with fixed
sand preparation equipment to which the shakeout and spill sand
is transported by conveyors and elevators. After preparation,
including magnetic separation of tramp iron, screening, addition
of binders and moisture, mulling, and possibly aeration, the
prepared sand is conveyed to the molders. Exhibit IV-26
illustrates a typical sand mulling unit.
High production foundries normally reuse the molding sand
a number of times each shift. If the sand is not effectively
cooled after each use, the temperature builds up to the point
-------
IV - 33
where evaporative cooling of the prepared sand occurs. A
significant amount of moisture added in the miller can be lost
before the sand can be formed into a mold resulting in a change
in che characteristics of the sand. This condition plus the
heat remaining gives rise to casting defects. Conventional
cooling is accomplished in the muller by directing a flow of
cool air through the mixer. A new concept, absorption cooling,
has recently been patented and is in use in several foundries
with reported success. This process is described in Section
LI
IX.
(c) Sand Reclamation
Three basic types of sand reclamation units are available--
dry, wet, and thermal. In the dry system the sand is first
screened and crushed to reduce the lumps and is Chen discharged
into a dry scrubber where the grains are subjected to the
blowing action of low pressure, high velocity air. The result-
ing impaction and scrubbing action between Individual grains
cleans the sand. For wet reclamation the sand, after screening
and crushing, is slurried with water end pumped into a. wet
scrubber equipped with a rapidly revolving propeller. The sand
is agitated violently, causing multiple collisions of the grains.
The slurry is next pumped into a classifier where the fines are
removed. Finally, after draining, the sand is dried and cooled.
The thermal reclaimer consists of a multiple hearth furnace
into which the screened and crushed sand is fed. The sand is
heated to about 1300° F and plowed from shelf to shelf within
the gas fired chamber. Cooling and screening to remove the
9, 13
sand fines complete the reclamation.
IV - 34
Cleaning, Heat Treating
and Finishing
Cleaning and finishing of castings are the final opera-
tions performed in the foundry. Cleaning of castings generally
refers to the operations involved in the removal of sand and
scale; sprues, gates, and risers; and fins, wires, chaplets or
other metal not & part of the casting. The castings, after
they have been separated from sand at the shakeout screen, are
cooled in boxes or on a conveyor which moves them to the cleaning
and finishing area. For gray iron castings, the gating system
may be broken off by impact in the shakeout, or they may be re-
moved on the casting delivery conveyor. Heads and gates on
ductile iron castings may require burning or sledging.
These operations ace shown on Exhibit IV-13.
(a) Cooling and
Sorting
Following shakeout and sprue removal, castings are cooled
before being sent to the cleaning and finishing area. Cooling
can be accomplished in a variety of ways, depending on the size
and shape of the castings, the rate at which castings are pro-
duced, and the degree of mechanization in the foundry. Several
types of conveyors—pallet, flat chain, belt, and overhead
chain—are used in high production shops, while other shops cool
in boxes or simply by piling the castings on the floor. Sorting
is generally done after a sufficient amount of cooling and in-
volves separating the different products into handling containers.
-------
IV - 35
(b) Castings Cleaning
The removal of sprues, gates and risers Is usually the
first operation in the cleaning room. This is done by impact,
shearing, abrasive cutoff, band or friction sawing, or torch
cutting depending upon the size of the casting and the type of
iron. Abrasive cutoff wheels can be used for removing gates
and risers on castings made of hard- or difficult-to-saw alloys.
Gates and risers on large ductile castings are most conveniently
removed by gas cutting. Band or friction sawing may be used
where it is desirable to follow the contour of the casting.
Surface cleaning operations ordinarily follow removal of
the gating system for all foundry castings except malleable iron
castings which are heat treated before surface cleaning operations.
Abrasive operations such as shotblasting, sand blasting, and
tumbling are normally used. Wire brushing, buffing, pickling,
and polishing operations may also be performed in some cases.
Shotblasting the casting surface is a rapid means of re-
moving sand and scale. The abrasive action removes the mold
sand, scale, and burrs. There are several methods for propelling
the abrasives used in blast cleaning and a great variety of
abrasive mediums may be used. The common abrasive mediums used
in foundry application are sand, grit and shot. Grit consists of
angular metallic particles and the round metallic particles form
the shot.
IV - 36
Two basic methods employed for blasting are:
1. Mechanical blasting, in which the abrasive is pro-
pelled by means of a power-driven, rapidly rolling bladed wheel.
2. Air blasting, in which the abrasive is propelled
through a nozzle by compressed air.
Of the two methods, mechanical blasting is the more widely
used in foundries.
Various shot blast machines are available including rotary
tables, tumbling units, rotary drums, continuous cabinet types,
portable hand-operated units, and cleaning rooms for very large
castings. An illustration of a commonly used cabinet type shot-
blast unit is shown in Exhibit 1V-27.
(c) Heat Treating
Iron castings are heat treated to improve physical
characteristics and to improve machineability. Typical heat
treating temperature ranges for gray iron are:
Stress Relief 1000°-1250° F
Anneal 1250°-1725° F
Normalize 1725° F
Quench and Temper 1550°-1600 F
Annealing of white iron castings to produce malleable iron is
done at 1600°-1750° F. Batch-type or continuous furnaces may
be used depending upon the type of heat treating and production
requirements.
-------
IV - 37
(<0 Casting Finishing
Following surface cleaning and heat treating, castings
are finished to remove gate and riser pads, chaplets, wires,
parting line flash not a part of its final dimensions. Chipping
hammers and grinders are used for these operations.
Castings that can be handled manually are trimmed or
ground on bench, floorstand, or portable grinders. Swing-frame
or portable grinders are used for trimming castings that are too
heavy to be carried or held by hand.
Tumble finishing may be used in providing final finish on
some castings. This process involves the tumbling or rolling of
parts in rotating barrels or the agitation of parts in shaker
containers. Vibratory tubes or other comparable equipment for
cleaning, shine rolling and burnishing are also used.
(e) Casting Coating
The large majority of iron castings are ready for ship-
ment after finishing operations are completed. Some casting
specifications, however, require that a surface coating be
applied for rust proofing or other reason. Rust preventive
paints are usually applied by spraying, fogging, or dipping,
depending upon the type of coating, casting size and shape, and
production requirements.
The painting facility may include conveyors, spray booth,
dip tank, a conveyor system, and a drying oven.
IV - 38
Coremaking
Cores are normally made of silica sand and organic or
inorganic binders. The selection of the core formulation and
process best suited to a particular application requires consid-
eration of many factors Including green strength, dry strength,
porosity, core complexity, quantity of cores required, and raw
material, equipment and production costs.
(a) Core Processes
The major coremaking processes in current use for castings
are;
- Oil Sand Cores
- Shell Cores
- Silicate Bonded Cores
- Furan Cores
Hot Box Cores
- Alkyd - Isocyanate Cores
Phenol - Isocyanate Cores
Oil sand cores are widely used although silicate and resin
bonded cores are being used in greater numbers each year.
Vegetable or mineral oils are commonly used as binders. Cereal
binders and clay are often used in conjunction with core oils.
The cereal binders, mostly derived from corn, are added to im-
prove green and dry bond, decrease the oil required, and improve
collapsibility of the core. Clay is often added in small amounts
to increase the green strength.
Shell cores for iron casting are usually made with phenol-
formaldehyde resin and round grain silica sand. Large users of
-------
IV - 39
shell core sand find It economical to buy bulk resin and coat
Che silica sand themselves. Premixed sand and resin is avail-
able commercially for the smaller user for whom this is not
economical.
Hollow shell cores are made by the investment process,
and small, solid cores can be made in a hot box machine. No
oven is required for shell cores since they cure quickly in the
coremaking machine.
Silicate bonded cores are made in a molding or core blow-
ing machine, end set by the application of carbon dioxide in a
manner that permits the gas to completely permeate the core.
Since the storage life of silicate bonded sand is short when
exposed to the air, due to absorption of C02, the mixed sand
must be stored in covered containers.
Furan air-set cores employ resins made from furfuryl
alcohol, ureas, and formaldehydes. The resins are mixed with
core sand and a phosphoric acid activator in conventional mix-
ing equipment.
Hot box core resins include furfuryl alcohol, urea-
formaldehyde and phenol urea-formaldehyde. The liquid resin
is mixed with the core sand and activated in conventional
mixing equipment. An exothermic reaction between the resin
and the activator progresses quickly when the mixture enters
the heated core box in a core blower.
The alkyd-isocyanate process involves the use of a synthetic
oil binder which, when mixed with sand and activated chemically,
IV - 40
produces cores that cure at room temperature. The key to the
process is the polymerization reaction that provides advantages
in terms of time and thoroughness of set.
Phenol-isocyanate Is a new no-bake process that was in-
troduced in 1970 and Is based on a urethane resin system. One
type of process consists of a binder and catalyst used In equal
portions. An advantage of this process is that strength and
hardness properties are obtained at binder levels'lower than
with other systems.
(b) Core Department
Process Flow
Exhibit IV-15 illustrates the process flow Cor the making
of sand cores. Sand is purchased in bulk quantities and ship-
ments are stored In silos or bins. Small quantities of core
premixes, resins, binders and other additives are received in
bags or drums and stored indoors.
The sand storage bin is often located above the mixer to
permit discharge to a weigh hopper directly above the mixer
inlet. Other dry and liquid materials are measured by weight
or volume and added to the mixer at the appropriate point in
the mixing cycle. The core sand mix is discharged from the
mixer and transferred to the core machines by conveyor or tote-
box.
The selection of the proper machine for forming the core
will be determined, in part, by the type of sand mix employed,
such as shell or hot box mixes which require specially designed
equipment for core box investment and heating. Other processes
-------
IV - 41
permit the use of a variety of machines. Simple cores required
in small quantities are made manually at a core bench using a
wooden or metal core box. Cores with a constant cross section
can be made in the desired length in a povered or hand-operated
extrusion machine. Molding machines of all types including
rollover, jolt and squeezer machines, and sand slingers are
Still in use in many foundries, but the use of core blowing and
shooting machines has become more widespread in recent years
except for large cores.
After forming, those cores that achieve a primary or com-
plete set while in the core machine require no special handling
since they have sufficient structural strength to permit placing
on storage racks or in tote pans for storage or further process-
ing. Cores requiring an oven bake or gasing are placed on a
flat core plate or a formed core dryer providing rigid support.
Oil sand cores requiring baking are transferred to gas or
oil-fired ovens. Light oil fractions and moisture in the sand
are evaporated, followed by oxidation and polymerization of the
core oil.
Core finishing operations consist of cleaning, sizing,
and assembling. The cleaning operations include trimming,
brushing, coating, mudding, and venting. The trimming and
brushing removes fins and excess sand. Core coatings, or
washes, improve the surface of the casting and insulate the cores
from the molten metal.
IV - 42
Holes or depressions in a core are filled and, if addi-
tional vent holes in the core are required, they are made at
this point.
Sizing of cores includes gaging of the core and filing or
grinding to the required dimensions.
Complex cores sometimes require that they be made in
separate parts and assembled into final form with core pastes.
Occasionally other methods of fastening cores together are used
13
such as bolting.
Pattern Making
Foundry patterns are normally made of wood or metal.
Patterns for small production runs tend to be the former and for
large production runs, the latter. Wood patterns generally
have a shorter useful life, although they can be repaired more
easily than the metal patterns which usually have a higher first
cost. A large production requirement, however, often results
in a lower pattern cost per mold with metal molds.
Typical pattern making steps include cutting the wood or
metal pieces to size, fastening the pieces together, and paint-
ing, varnishing and mounting where required. Match plates are
often cast from the original patterns.
Wood pattern shop equipment includes different types of
saws, planers, joiners, lathes, edgers, routers and drill
presses. Metal pattern making equipment includes typical machine
shop tools.
-------
IV - 43
REFERENCES
1- U.S. Department of Commerce Neve. Business and Defense
Services Administration, BD 69-14, March 13, 1969.
2. A Systems Analysis Study of the Integrated Iron and
Steel Industry. Division of Process Control Engineering, National
Air Pollution Control Administration, Department of Health,
Education and Welfare, Battelle Memorial Institute, May 15, 1969.
3. The Cupola and Its Operation, published by the American
Foundrymen's Society, 3rd ed., 1965.
4. Metals Handbook; Volume 5 Forging and Casting, pub-
lished by the American Society for Metals, 8th ed., 1970.
5. "Continuous Iron Melting," R. S. Amala and J. B.
Walker, Modern Casting. May, 1970, p. 45.
6. "Some New Factors Affecting Cupola Operation," H. W.
Lownle, Modem Casting. 1966.
7. Principles of Metal Casting. R. Heine, McGraw-Hill
Book Co., 1955, pp. 550-553.
8. "Comparing Processes foi Making Ductile Iron," Dr. E.
Modi, Foundry. July, 1970, pp. 42-48.
9. Molding Methods and Material, published by the American
Foundrymen's Society, 1st ed., 1962.
10. Casting Materials Company.
11. "Beat Hot Sand Problems with Absorption Cooling,"
A. J. Wagner, Modern Casting. August, 1970, p. 46.
12. "Progress in Coremaking," W. 0. Ferguson, Foundry.
August, 1970, pp. 53-59.
IV - 44
13• Foundry Core Practice. H. Dietere, published by the
American Foundrytnen's Society, 3rd ed. , 1966.
-------
V - EMISSIONS PRODUCED AND
EMISSION CONTROL CAPABILITY
EMISSIONS PRODUCED IN
THE BASIC PROCESSES
Emissions consisting of particulace matter, fume, smoke,
or gas are a by-product of every foundry process and almost
every foundry operation. It is necessary that the source of
these emissions and the factors affecting their type and quantity
be identified so that the foundry emissions problem can be
evaluated.
The type, concentration, and size of typical foundry emissions
are tabulated in Exhibit III-7 by department and operation. These
emissions are identified in the following paragraphs and evalu-
ated in a later section of the report.
Raw Material Storage
and Charge Makeup
The handling, preparation, and charge makeup of the basic
foundry raw materials — scrap metal, coke, and limes tone--may
produce emissions in a variety of ways. The storage of coke and
limestone over extended periods may result in the disintegration
of these materials from the action of the sun, rain, and repeated
freezing and thawing. Ferrous scrap rusts rapidly. Subsequent
handling during the makeup of furnace charges may cause lime-
stone dust, coke breeze and small quantities of rust to be re-
leased into the environment. At conveyor transfers, as well as
storage bins, weigh hoppers and the location where these
V - 2
materials are placed in charge buckets, emissions tend to
increase due to tumbling of the material. Rehandling of coke
results in additional disintegration and, to a lesser degree,
this Is also true of limestone.
The preparation of metallic charge materials including
the breaking and cutting of large scrap, removing cutting oil
residue from machine shop turnings and borings in preparation
for briquetting, and cleaning of foundry returns represent an
additional potential source of emissions. The industry trend,
at least for the larger foundries, is Co purchase scrap to
definite specifications to eliminate Che need for scrap pre-
paration by the foundry, except for foundry returns. A few
foundries make briquettes from internally generated turnings
and borings, but most companies in this position sell the
turnings and borings and purchase briquettes according to their
needs. For the foundry that does convert its machine shop
scrap into briquettes, the removal of oil on the turnings
represents a source of potential emissions if the scrap is in-
cinerated, and to a much lesser degree if the scrap is centrifuged.
Shot blasting of foundry returns may produce emissions in the
form of dust from the embedded sand and from broken shot. Shot
blast machines are normally furnished with dry mechanical col-
lectors or fabric filters.
-------
V - 3
V - 4
Melting
The melting department is responsible for the greatest
amount and heaviest concentration of emissions in the foundry,
producing the need for control equipment on cupolas, electric
arc furnaces, preheaters and dryers. Emissions from coreless
induction furnaces are insignificant due to the quality of
scrap and the fact that no combustion takes place in the units.
Channel induction furnaces also produce minimum amounts of
emissions and are seldom provided with emission control equip-
ment •
The cupola is the largest single source of emissions,
producing fume, smoke, particulate matter and gas. Concentra-
tions are affected by the quality and quantity of charge materials,
Che use of techniques such as oxygen enrichment and fuel injection,
the volume and rate of combustion air, and the melting zone tem-
3
perature.
The electric arc furnace produces somewhat less emissions
than the cupola because no fuel or combustion air is required.
Combustion in an arc furnace is limited and results from the in-
clusion of combustible elements in the charge and from burning of
the electrodes. Concentrations may be great enough, when used as
a prime melter, to require emission control equipment. Emissions
Include smoke, fume, dust, and possibly some oil vapors in con-
centrantions that are affected by the quality and composition
of charge materials and the temperature of the bath. Emis-
sions are greatest during the melt-down phase of the cycle,
and less after melting is completed, when the molten metal is
covered with slag.
Induction melting produces light concentrations of emis-
sions consisting of fume, smoke, and oil vapor. Control
devices are usually not provided or required. The smoke and
oil vapor usually derives from small amounts of cutting oil
adhering to the steel or iron scrap, and can be eliminated by
preheating prior to charging into the induction furnace.
The reverberatory or air furnace for melting or duplexing
produces light to moderate concentrations of emissions. Com-
bustion occurs within the furnace but the gas or oil fuel is
burned in highly efficient burners above the metal bath.
Smoke and fume are produced in this type of furnace. The
smoke results from combustion of oil on the scrap and other
combustible materials in the charge. Metallic oxides appear
in the emissions as in any melting furnace, and are the result
of nonferrous contaminants in the charge material, vaporized in
the molten bath with a portion of the iron scrap. The concen-
trations are related to the partial pressures of the oxides at
the melting temperature.
Preheaters or dryers for furnace charges are another
source of emissions in the melting department. They are
rarely used in conjunction with cupolas or reverberatory
-------
V - 5
furnaces since these are highly efficient preheaters in them-
selves. The use of a preheater as a means of increasing
melting capacities of electric arc and induction furnaces,
and to insure dry furnace charges essential to the induction
furnace, transfers the majority of the emissions production
from the furnace to itself by burning combustibles which
would otherwise burn in the furnace. The resulting emissions
are therefore the same type and appear in the same quantity
as would be found In these furnaces without preheating.4
Coreless induction furnaces used as holding or super-
heating furnaces charged with molten iron emit minor amounts
of metallic oxide fume and are rarely provided with emission
control equipment.
Magnesium Treatment for
Producing Ductile Iron
Treatment of molten iron with magnesium compounds to
produce ductile iron is a significant source of foundry emis-
sions. Emissions from this operation consist primarily of
magnesium oxide representing up to 65% of the metal injected.
The high loss occurs since the boiling point of magnesium is
far below the temperature of the molten iron. The reaction
is often violent, depending on the form of magnesium used,
and shielding or enclosing of the ladle at this time is re-
commended for this reason as well as to insure maximum col-
lection with a minimum dilution with Infiltrated air.
V - 6
Molten metal in a furnace or ladle produces a shimmering
heat wave above its surface which is often mistaken for emis-
sions rising from the bath. The phenomenon is an optical
illusion resulting from the varying index of refraction of
the unequally heated air strata, and is present where molten
metal is used.
Molding, Pouring
and Shakeout
The molding operation is not a major contributor to
foundry emissions. In green sand molding, moisture in the
sand acts as a dust suppressant. Small quantities of dry
parting compound are emitted when the mold halves are dusted
with this material. Liquid partings used to prevent molding
sand from sticking to metal patterns or match plates have a
kerosene base. When sprayed on the patterns, a portion of
the vehicle vaporizes, and the solids such as stearic acid
are sprayed into Che air in the immediate environment. Sea
coal is also used as a mold spray and is released into the
atmosphere. Concentrations are very light.
Emissions from the pouring operation are much more severe
than from molding and are usually more difficult to capture.
The hot metal, when poured into the mold, first ignites and, as
oxygen in the mold is exhausted, thermally degradates such
materials in the sand as sea coal, cereal and synthetic binders,
and core binders. Steam is formed in green sand molds from the
moist sand.
-------
V - 7
In che full mold process, the complete pattern is consumed.
Concentrations are usually relatively low for a single mold, but
the multiplicity of molds produced increases the severity of the
problem. Most of the emissions are steam, vapor, and smoke.
In the full mold process and for many of the synthetic binders,
emissions may be toxic, and only the low concentration per mold,
coupled with general foundry ventilating systems, prevents phys-
iological reactions in molders, pouring crews and shakeout men.
Examples of potentially toxic emissions are styrene, low mole-
cular weight polystyrene, ethylbenzene, methyl chloride, chlorine,
hydrogen chloride and decomposition of evaporative products in
addition to CO, CO. and HjO.
The pouring of ductile iron results in further emissions
of magnesium oxide from the magnesium treatment process. The
concentration of smoke, fume, gas, and vapor is related to the
hot metal temperature, length of time between pouring and shake-
out, and the quantities of binders, moisture and parting com-
pounds in the molding sand.
At the shakeout, the action of separating castings from
molds brings hot castings into contact with moist molding sand
not adjacent to the sand-metal interface at the time of pour-
ing. This results in additional smoke, steam and vapor of
the same type emitted during the pouring operation. Concen-
tration of emissions is momentarily high, but the casting is
cooler than the molten metal while pouring, the sand is quickly
separated from the casting, and the emissions are often able
to be contained and removed through the use of ventilated hoods.
V - 8
Sand Conditioning
New molding and core sand are ordered to a desired screen
test for specific use and always include some fines. The es-
cape of fines into the atmosphere varies with the method of
handling. Closed systems such as pneumatic conveyors release
only small amounts and are provided with exhaust connections
at the inlet and the receiver. Systems using belt conveyors
and bucket elevators for dry sand may release fines and dust at
transfer points between conveying units. Many smaller foundries
unload and transfer sand to floor level bins manually, or with
front-end loaders. Load and unload points are generally un-
controlled. The handling of conditioned molding or core sand
presents fewer problems than new sand because of the moisture
content and binder additives and therefore control equipment
and hoods are generally not required. Shakeout or return sand
produces more emissions because it has been partially dried
from contact with the hot metal. Introduction of fresh spill
sand from the molding floor and excess prepared sand helps
considerably in cooling, moistening and decreasing dust and
fines from being released to the atmosphere. It is considered
good practice to enclose the transfer points of conveyor systems
handling dry sand, and to provide exhaust connections at these
locations and at the vibrating or rotating screen and the return
sand storage bin.
Moderate concentrations of fines, dust and binder materials
are emitted at the sand mixer. Concentrations are increased if
-------
V - 9
the muller is equipped for sand cooling. This is accomplished
by directing a blast of cooling air either over or through the
sand while it Is being mixed. The air blast entrains small
particles and must be exhausted to a control device to separate
Che dust from the air blast
Cleaning and
Finishing
Cleaning and finishing operations produce emissions less
troublesome than other foundry processes. Particulate matter
is generally larger and easier to capture and separate from
the air stream even though concentrations can occasionally be
moderate to heavy. The metallic dust and particles removed
from grinding operations have a high density permitting the
use of less expensive dry mechanical collectors but require
higher entralnment velocities for removal from the working
surface. Particulate concentration is dependent upon type
and surface speed of the grinding wheel and the amount of
pressure exerted by the grinder. Chipping operations produce
such large particles that control equipment is not required
except when chipping castings with burnt-in sand. In this
event, good foundry practice dictates the wearing of a face
mask. Abrasive shotblastine produces high concentrations of
metal particles, sand dust, and broken shot but modern blast
machines are provided with high efficiency fabric filters
designed for the purpose. The concentration of these emis-
sions is a function of the quantity of embedded sand on the
castings, fracture strength of the shot, and length of time
V - 10
in the blast cabinet or room. Sand blasting, now seldom used,
produces high concentrations of sand dust with concentrations
related to air pressure, blast sand characteristics and length
of time required for cleaning. Sand blast operators wear ven-
tilated hoods to prevent inhalation of sand dust.
Emissions from annealing and heat treating furnaces are
minimal except when the castings have previously been oil
quenched. Concentration of the resulting smoke is a function
of the amount of oil residue on the casting surfaces. Paint-
ing is infrequently done by foundry departments. Emissions
from this operation consist primarily of vapors from thinners
and concentrations depend on the type and quantity of the
volatiles.
Coremaking
Emissions problems in the core department are similar to
those encountered in the handling of molding sand. No major
difference exists for the handling and storing of new sand.
The coremaking operation using synthetic binders and heated
core boxes often results in vapor and gases; some of them
may be toxic or acrid in nature. A similar situation exists in
some of the synthetic binders that set at room temprratures, and
to a lesser degree with oil sand coremaking. Oil sand cores
require oven baking to dry and polymerize the oil binders.
Core oven emissions can be highly odorous. Incineration is the
most satisfactory way to control these emissions, using after-
burners or catalytic combustion chambers. Concentrations of
-------
V - 11
V - 12
Che emissions vary with Che Cype of binder used. Wlch che
wide variety of binders now available and others in Che develop-
ment stage, conservacive and flexible design criteria for col-
lection ductwork and concrol systems would be recommended.
Patcern Malting
Wood pattern shop emissions consist primarily of wood
dust and chips. Concentration can be high for each machine
tool but normal utilization factors of the machines reduce
the system capacity. The wood dust, having a low density
and-laree surface area, is easily entrained.
EMISSION CONTROL
CAPABILITY
The current state of the art of foundry emissions control
does not fully satisfy the needs of the industry. Although,
on a purely technical basis, virtually all particulate and
most gas, smoke and fume emissions can be controlled, the cost
of such control for several basic foundry processes is often
beyond the financial ability of the small- and medium-sized
foundries, which comprise the large majority of the industry,
to support. The current foundry need is for the development
of equipment to control emissions from cupolas, electric arc
furnaces, ductile iron production by magnesium treatment,
mold pouring stations, and coremaking operations at costs that
do not threaten the profitability, and even the existence, of
the small foundry.
Emissions Controllable
with Existing Equipment
Design
Relative controllability of foundry emissions is indicated
in Exhibit III-7 on a comparative basis. Emissions more dif-
ficult to collect are by and large Chose with large concentrations
of very fine particles, five microns and smaller. Conversely,
emissions easier to collect are those consisting entirely of
large particles.
Particulate Emissions
Uneconomical to
Control
Generally, all foundry emissions are uneconomical to
control, since the collected material has little or no value,
and its collection adds no value to the foundry's product.
The installation and operating costs of control systems vary
over a wide range, and it is necessary to identify those
operations requiring highly uneconomical systems beyond the
financial capability of the majority of foundries.
The cost of an emissions control system depends on the
following variables:
Properties of emissions, including size distri-
bution, density, chemical composition, corrosiveness, solu-
bility, combustibility, and concentration.
Difficulty of capturing the emissions in an air,
gas, or water stream of moderate temperature and volume.
Difficulty of separating emission particles from
the captor medium.
-------
V - 13
Properties of Che raaCCer to be collected are generally
fixed by the process and raw materials, although modification
of the equipment could possibly alter the properties. Assum-
ing them to be fixed for a given operation, the first consid-
eration Is Che cost of capture. If the operation occurs In an
enclosed and fixed location such as a melting furnace or oven,
capture may be relatively simple and may be accomplished at
low cost although emission collection and separation costs
could be high. If the location of the operation Is not fixed
and occurs in Che open, such as pouring of molds set out on
the foundry floor, then capture Is difficult and more expensive.
In the latter case, with pouring emissions dispersed through-
out the plant, much of the air in the building must be pro-
ceased trough the control system to collect the emissions.
A system of this capacity would be expensive to Install and
operate.
The third factor of system cost is the difficulty of
particle separation from the captor medium. Large, dense
particles, such as metallic fragments from grinding operations,
can usually be separated by the use of relatively low cost dry
centrifugal collectors. Submicron-slzed metallic oxide particles
from a melting furnace, however, require more costly collection
equipment such as high energy wet scrubbers or fabric filters.
An analysis of foundry operations identifies processes
combining difficulty of emission capture and collection with
V - 14
high cost of the control equipment. These are briefly dis-
cussed below while an evaluation of these emissions are pre-
sented In Section VI and relative costs are tabulated In
Exhibit III-7.
(a) Cupola
Iron melting in a cupola produces heavy concentrations
of particulate emissions in a gas stream up to 2000° F. In
most cupolas, large amounts of air are infiltrated through
the open charging door greatly increasing the cost of the
control system, since the size and cost of a collector is
directly related to the volume of gas to be handled.
(b) Electric Arc
Furnace
The electric arc furnace when melting iron produces
moderately heavy concentrations of particulaces with a large
percentage of the total particles below five microns in size.
The gas stream is well over 2000° F, requiring cooling by
infiltrating air or by water sprays.
(c) Magnesium Treatment
for Producing Ductile
Iron
This process results in heavy concentrations of extremely
fine metallic oxides, principally magnesium oxide.
(d) Mold Pouring
Mold pouring produces only moderate concentrations of
emissions, but in many foundries capture is very difficult.
In addition, the emission may be toxic or acrid.
-------
V - 15
(e) Coremaktne
Coremaking and baking or heating for setting of the
binders result In moderate concentrations of sometimes toxic
and acrid emissions. Smoke and vapors from floor or core-
making operations are difficult to capture with only moderate
dilution and are therefore expensive to control. Core baking
oven emissions can be controlled with afterburners or cata-
lytic combustion chambers.
These five processes represent the sources of foundry
emissions not readily controllable, or highly uneconomical to
control.
V - 16
REFERENCES
1. Conference on Foundry Ventilation and Dust Control,
Harrogate, England, 1965. British Cast Iron Research
Association, Alvechurch, Birmingham.
2. Foundry Core Practice. H. W. Dletert, American
Foundrymen's Society, Des Plaines, Illinois, 3rd ed., 1966.
3. The Cupola and Its Operation. American Foundrymen's
Society, Des Plaines, Illinois, 3rd ed., 1965.
4. Air Pollution Engineering Manual. U. S. Department
of Health, Education and Welfare, Publication 999-AP-40, 1967.
-------
VI - QUANTIFICATION AND
EVALUATION OF EMISSIONS
DESCRIPTION OF SAMPLING AND
ANALYTICAL TECHNIQUES
Measurement or sampling of emissions from a foundry
source can be conducted in the general atmosphere near Che
suspected source or in a specific stack. The sampling alter-
native used depends on the objectives of the measurement which
Include:
1. Determining whether stack emissions are in vio-
lation of existing ordinances.
2. Providing a guide in selection of proper control
equipment.
3. Evaluating the effectiveness of control equip-
ment after installation.
Various techniques and methods are available for sampling
emissions in the atmosphere or stack, and Che adequacy and
effectiveness of these methods vary widely depending on which
one is employed. *•
General Atmosphere
Sampling
Many different area or atmosphere sampling methods are
used in various sections of the country. Normally, the con-
centration of dust collected by area sampling methods Is re-
ported in micrograms per cubic meter and coarse dust fallout
A T KF-KRVFY ft rOMPAVY Ivr
VI - 2
in tons per square mile per month, pounds per 1,000 square
feet per 30 days, or as a soiling Index. Such samplings are
conducted off the premises of a particular plant and a local
agency procedure is followed rather than one related to a
specific Industry.
Various factors such as frequency of sampling, wind,
humidity, topography and analytical procedures are important,
no matter which method is used to sample contaminants in the
general atmosphere.
Usually, particle fall is best determined by frequent
short time cumulative samplings rather than one continuous
cumulative sampling over a long period of time. The summa-
tion of a series of samples can produce data approximately as
accurate as, and more economically than, a continuous sample.
Meteorology factors such as wind and humidity, and also
topography observations including the relation of differences
in surface elevation of the surrounding area, hills, valleys,
terrain and buildings, must be recorded when atmosphere sam-
pling of emissions is made.
How the sample data will be analyzed, and the type of
equipment used, must be determined prior to sampling so that
the proper sampling procedure is pursued. The adequacy of any
quantitative general atmosphere sampling method is limited
-------
VI - 3
VI -
because of the effects which diffusion, turbulence, wind veloc-
ity, wind direction and various other factors have on measure-
ment results.
In addition to the quantitative methods discussed, dis-
charge quality can be estimated by observing the opacity or
discharge appearance from a stack. Opacity scales of light
obscuring degree, an adaptation of the Ringelmann Scale, are
often referred to in air pollution codes.
The Ringelmann Scale shown in Exhibit VI-1 is a chart for
grading the black density of s-nc'.cc- from exhaust stacks and has
been used in the United States since the early 1900's. The scale
was originally designed for the purpose of indicating the degree
of combustion control in coal burning equipment. It was not
designed to cover metal melting emissions.
With Che Ringelmann Scale, an observer compares smoke
emission to the chart and estimates black density of the
smoke. This method relies entirely on the personal evaluation
of an observer when viewing stack discharge. Because smoke
emission colors other Chan black also occur, variations of
the Ringelmann Scale have been developed which refer to an
equivalent opaciCy rule. This rule scates that if an emis-
sion, regardless of color, obscures an observer's view to
such an extent Chat smoke of a given Ringelmann number would,
the emission is equivalent to a smoke plume of that number.
Metallurgical processes such as iron melting often emit
particulate matter down to one micron or less in diameter.
These small particles often exhibit light scattering effects,
and cause the appearance of a high level of opacity when emitted
from an exhaust stack in a gaseous stream. When the Ringel-
mann Scale is used to determine concentration or weight of parti-
culates in metallurgical process exhausts, the presence of these
submicron particles tends to lead to erroneous conclusions.
Some air pollution control regulations specifically limit
the discharge of visible emissions in addition to limiting par-
ticulate concentrations of those emissions. For these regulations
Che legal criterion is the Ringelmann equivalent opacity test,
even though the application of a visual test for compliance is
subjective and its reproducibility subject to atmospheric con-
ditions. Some other state air pollution boards have adopted
statements such as the following regarding the use of a Ringel-
mann Scale: "The Ringelmann chare shall be used for grading the
light obscuring power of smoke. It shall not be used for
determining metallurgical fume emissions or measuring the opac-
ity of non-combustion process emission." However, it is a
simple and rapidly applied test that will probably continue to be
used for judging compliance to visible emissions standards until
such time as test devices are developed that provide reliable,
inexpensive and reproducible results.
-------
VI - 5
Stack Sampling
The collection of representative samples of gaseous and
partlculate matter flowing through a stack requires specialized
equipment, skills and knowledge and therefore is a more compli-
cated procedure than general atmosphere sampling. More than
one sampling method is usually used when sampling mixtures of
particulate matter and gaseous compounds, especially if sepa-
ration of the sample is required.
Of the number of methods presently in use for stack sam-
pling, those developed by the American Society of Mechanical
Engineers are the most widely employed. The ASME has published
two Power Test Codes--(1) PTC 21-1941 "Dust Separating Apparatus"
and (2) PTC 27-1957 "Determining Dust Concentration in a Gas
Stream"—which are followed when the effectiveness of an emis-
sion control equipment system is evaluated and/or to determine
if an ordinance has been violated. An adaptation of the ASME
Codes with special emphasis on the specific problems of testing
foundry cupolas has recently been adopted by the American
Foundrymen's Society and the Gray and Ductile Iron Founders'
Society, Inc. The recommended practice is discussed later
in this section, and a copy is included in Appendix E. The
data presented in this report are from cupola tests made
according to the provisions included in the ASME Codes and
were run before the adoption of the recommended practice.
The ASME lists four basic steps to follow in evaluating
the effectiveness of an emission control equipment system:
1. Secure a representative emission sample.
VI - 6
2. Filter and measure the dust contained in the
sampled gases.
3. Measure the sampled gas volume.
4. Measure dust conditions such as temperature,
pressure, composition and various other factors.
In following this four-step procedure, agreement must be
reached by the parties to the test on certain pertinent items.
A listing of these items appears in Exhibit VI-2.
The ASME provides a number of tables to report and
analyze test results, major headings of which appear below.
General Information.
- Description, Dimensions, etc.
- Test Data and Results.
- Pitot Tube and Dust Samples Data.
- Dust Caught by Separator.
- Size Analysis of Dust Samples.
- Efficiencies.
- Fuel and Gas Analyses.
- General Data
There are 69 lines of data called for in the above nine
tables. The code manual, PTC 21-1941, can be referred to
for a complete description of all test procedures including
1 2
tables for reporting results.'
The ASME procedures described in PTC 21-1941 and
PTC 27-1957 were followed by the foundries visited where
emission test results were obtained.
-------
VI - 7
The ASME test procedures on dust concentration and dust
separating equipment are adequate and will produce accurate
results If followed by a trained team of specialists.
Of particular importance in any stack sampling method
is determination of emission particle size since these data
will largely determine overall collection efficiency of a
control equipment system. Considerable care must be exercised
in collection of samples so that agglomeration and shattering
of particles will not occur. Particles 44 microns and larger
in size are normally separated by standard screens. Particles
passing through the 325-mesh screen (44 microns or smaller)
can be estimated or quantified by one of several methods.
1. Microscope Counting - A small portion of the
dust sampled is suspended in a suitable liquid medium and
placed in a dust counting cell where the particles in each
size range are counted. The ranges vary but separations of
up to 5, 6-10, 11-20 and 21-44 microns are usually made.
2. Settling Method - A sample of the collected dust
is suspended in alcohol in a beaker. Separation is then made
by filling the beaker to a mark 8 or 10 centimeters from the
bottom, stirring until completely mixed, letting stand for a
specified time and quickly siphoning off the liquid down approxi-
mately 6 centimeters below the starting level. The settling
time is computed to give the size range desired and the opera-
tion is repeated approximately three times. The particle size
range is then checked by microscope in a dust counting cell and
the dust particles are separated either by filtration or by
centrifuging for drying and weighing.
VI - 8
3. Centrifuge Separation - This technique essen-
tially consists of separating one-gram samples of dust into
sizes. The specialized test equipment required is manufactured
in Sweden.
4. Air Elutriation - This method consists of separat-
ing microscopic particles into size ranges by air elutriation.1
NECESSITY FOR
STACK SAMPLING
PROGRAM
Test Reports Gathered
and Analyzed
Emissions data from 45 stack sampling tests tabulated in
the foundry data bank were available for the evaluation of
cupola emissions as functions of furnace design and operating
variables. The data were obtained from foundry operators
during plant visits, testing laboratories, equipment manufac-
turers, state pollution control boards, and published litera-
ture. Stack sampling tests were not conducted specifically
for this study. The collected data were not adequate to
completely define and quantify the effect of all melting
variables on emissions levels. The analyses did, however,
identify certain design variables that have no significant
effect on emission levels and a number of variables having
either a quantifiable effect or a definite, if not completely
defined, effect on emissions levels.
-------
VI - 9
Need for Additional
Data
Additional testing of cupolas would be required to defi-
nitely establish mathematical relationships between all opera-
ting variables and emissions rates. The program to achieve
this purpose would require the use of experimental cupolas of
several different sizes operating under laboratory conditions.
It has been determined that the results of such a program,
while of considerable future interest, are not essential for
the formulation of the recommended research and development
projects developed in this study, and for the completion of
this report.
EVALUATION OF
CUPOLA EMISSIONS
Nature of Data
Gathered
The evaluation of cupola emissions and the development
of emission levels as a function of cupola design factors and
operating practices required analysis of three types of Input
data: the design criteria for each cupola considered, the
actual operation at the time the stack tescs were made, and
the emission data at the cupola outlet, or control system
inlet.
Detailed information on a total of 481 Iron melting
systems in 1968 foundries is tabulated in the foundry data
bank. Cupola furnace installations represent 329 of this
total. The cupolas are classified according to type of lining,
blast air, charging, gas take-off, and use of afterburners,
VI - 10
oxygen enrichment, and fuel injection. In addition, the data
bank lists cupola dimensions, capacities, blast air volumes
and temperatures, afterburner sizes and locations, and other
appropriate data required to specifically describe the instal-
lation. Operating practices, including charge makeup, type
and condition of scrap, and coke, are also shown.
Partlculate emission levels, while commonly reported in
various ways, are tabulated in the data bank in terms of
grains per standard cubic foot of gas, and for the analysis,
are converted to pounds per ton of metal melted. This form
removes the effect of varying amounts of infiltrated air on
concentrations of partlculates. Emissions listed in the data
bank show control equipment inlet and outlet concentrations
as follows:
Test Location Number Reported
Inlet concentrations only 31
Outlet concentrations only 70
Inlet and Outlet concentrations 22
The amount of emission data available from the foundries
included in the data bank is limited even though the foundries
were selected on the basis of either having control systems
already installed or in the late planning stage.
A number of factors result in the limited amount of data
available.
1. Foundries are often reluctant to disclose the
results of stack testing, even on a non-identification basis,
when the results Indicate that the control equipment is either
not meeting the applicable code, or is just barely meeting it.
-------
VI - 11
VI - 12
2. Because of the high cost, few foundries care Co pay
for stack testing co determine particulate concentrations from
uncontrolled cupolas already known to violate the code, preferring
instead to invest those funds in the control equipment.
3. Equipment manufacturers do noc, as a rule, re-
quire cupola emission levels when quoting control systems.
Equipment designs are generally based on years of experience
and rules of thumb, with design contingencies usually large
enough to cover variations in type and quantity of emissions
from one installation to the next. Testing of only the final
emission is performed following equipment installation to
prove that contract guarantees have been met.
4. Stack sampling tests are expensive, and few
tests are made without a compelling need.
5. Stack testing required by the various emissions
control boards to prove compliance with local or state codes
are at the expense of the foundry. Since the control board
is concerned only with emissions from the control system,
this is by and large the only location in the entire melting
and control system that is tested.
For certain systems, such as fabric filters and high
energy wet scrubbers known to be operating properly in the
9W. efficiency range, the inlet concentration can be esti-
mated very closely from the collector outlet emission level,
or the amount of material collected. This is not feasible
with dry centrifugal, low and medium wet scrubbers, or wet
caps where the total efficiency can vary as much as 20%
depending on the type and particle size distribution of the
particulate matter.
Classification of Melting
Processes by Type and
Design Parameter
The features of melting furnaces are organized in an
orderly format to provide a basis for evaluating the foundry
data gathered from various sources. The furnaces are classi-
fied initially by type—cupola, electric arc, electric induc-
tion, and reverberatory or air furnace. The first three types
of melters are further divided by design parameters to produce
the "family tree" patterns indicated in Exhibits VI-3, VI-4,
and VI-5. Each path from the top to the bottom of the exhibits
represents a theoretical classification. An eleven-digit code
describing each possible classification is used in order to
facilitate identification of the system, and for ease in re-
trieval of information from a computerized data bank. A
description of the code is included in Appendix B.
Most theoretical combinations of furnace design parameters
are either impractical or mutually exclusive, and do not exist
in practice. Exhibit VI-6 shows the 32 cupola classifications
identified as practical designs in current use. The tally
marks indicate the distribution of existing cupola designs
among these 32 classes, for the furnaces in the data bank for
which emissions data were obtained. The tally marks are not
necessarily additive.
-------
VI - 13
Variables Affecting
Emissions
Exhibit VI-7 portrays the major components of particulate
emissions from iron melting cupolas and the percentage by
weight of the various materials determined by chemical analysis
of emissions from seven cupolas. The nine components can be
grouped into three major categories: (1) metallic oxides,
(2) silicon and calcium oxides, and (3) combustible materials.
It has been reported that the amount of metallic oxides
occuring in cupola emissions is related to Che presence of the
respective metals in the scrap charge and their partial vapor
pressures at the temperature of the cupola melting zone. All
metallic oxides except those of iron indicate the presence of
nonferrous contaminants or alloying additions in the metallic
scrap. For example, zinc oxide could result from the presence
of galvanized scrap or from die castings, lead oxide from
terne plate, aluminum oxide from aluminum scrap, and chromium,
copper and cadmium oxides from plated materials. Iron oxides
are always to be found in cupola emissions, the concentration
dependent on such factors as scrap thickness, degree of sur-
face corrosion, and temperature in the melting zone.
The oxides of silicon and calcium, representing the second
category, result from lining erosion, embedded molding or core
sand on foundry returns, dirt from the scrapyard adhering to
scrap, or from the limestone flux.
VI -
The third category of emissions, combustible material,
includes coke particles, vaporized or patially burned oil and
grease, particles from fabric coatings from automotive scrap,
and other similar contaminants swept up the stack by the top
gases.
The range of concentrations of the emission components
reported by the seven foundries listed in Exhibit VI-7 is in
4
general agreement with other reported data as shown below.
Chemical Composition of Cupola Dust
Component
S102
CaO
MgO
FeO(Fe203,Fe)
MnO
Mean Range
20%-40%
3- 6
2- 4
1- 3
12-16
1- 2
Scatter Values
2 -18
0.5 -25
0.5 - 5
5 -26
0.5 - 9
In addition to the presence of these materials in the
furnace charge and the temperature of the oxidizing atmosphere
in the melting zone, other variables would be expected to in-
fluence the amount of cupola emissions. Specific blast rate,
when increased, would be expected to result in greater emis-
sions by entrairanent of metallic oxides and mechanical dusts,
such as coke and limestone. A portion of the entrained par-
ticles is filtered out of the gas stream by the burden, with
a higher burden offering greater opportunity of particle
capture. It is well recognized that emission rates are greater
during burn-down, due in part to increased temperatures re-
sulting in larger gas volumes, higher gas velocity, lower
-------
VI - 15
collecting ability of the smaller burden height, and possibly
greater formation of metallic oxide vapors in the melting
zone. Furthermore, the height of the reducing zone is shorter,
with less potential for reduction of the already formed oxides.
It would be further expected that cupola emiss'ions would
vary directly as the percent of coke in the charge, and some
researchers have reported such a trend. This is a reasonable
theory since degradation of the coke while weighing, charging,
and moving downward in the cupola shaft will result in an in-
crease of coke dust in the furnace. Therefore, any change in
operating practice resulting in a decrease in the coke charge,
including heating of the blast air, or injection of an auxil-
iary fuel, should have a beneficial effect on the amount of
particulates emitted.
The use of an afterburner, properly designed and installed,
will decrease the quantity of combustible particles released to
the atmosphere or control system. Sufficient oxygen must be
provided through the charging door to permit complete combus-
tion, and the upper cupola stack must extend far enough to per-
mit time for comubstion before the particles are exhausted to
the atmosphere or to the emissions control equipment. Defi-
ciency in either factor will tend to negate the potential ad-
vantage of the afterburner.
VI - 16
Operating practices have noticeable effects on emissions
t.v.1.. The use of wood or paper produces for igniting the coke
hed results in smoke during this part of the operating cycle.
FluctUating burden height can result in higher mission rates^
Coke and limestone should be handled carefully to li.it degrad
tion, and should be screened prior to weighing in order to limit
che addition of dust to the charge. Shotblasting of foundry
returns and cleaning of oily scrap 111 result in lower missions.
It has been theorized that design of the cupola can have
an effect on the type and quantity of paniculate emissions.
An objective of the study was to determine whether test data
analyzed on this basis substantiate the theory. The effort
expended to investigate possible relationships between e^s-
Si0ns levels and cupola design includes the identification of
32 design classifications representing a large majority
of cupola installations in this country, the visits to found-
rles representative of these classifications, and collection
of detailed design, operating and emissions data. "~™
daca for this analysis are shown in the table on the
page.
T KEARNEY & COMPANY.
-------
VI - 17
Sources of Foundry Data for Determination of Effect
of Cupola Design Parameters on Emissions
Melting Systems
Electric
Cupola Induction
Electric
Arc Total
(D
33
8
Source
Foundry Visits 29
Interview Guide 8
Only A/P Control
Boards 9 _z_
Total 46 (D*
Note: * This foundry has both electric induction
and arc furnaces as primary melters, and
is counted only once.
Particle size distribution of cupola emissions indicated
by both respondent foundries and by installations reported in
the literature are tabulated in Exhibit VI-8. A definitive
relationship between size distribution and chemical composition
of emissions has not been discovered in the literature search,
nor is Investigation of the possible relationship within the
scope of the study. It would be expected, however, on the
basis of other information, that a high percentage of particles
less than five microns in size would coincide with a finding
of a substantial percentage of metallic oxides. Similarly, a
high percentage of particles greater than 44 microns in size
would correspond to large amounts of Si02 from foundry returns
and dirty scrap, and combustibles such as coke breeze.
VI - 18
Analysis of
Emissions Data
Cupola design parameters and particulate emission rates
for 17 furnace classifications are tabulated in Exhibit VI-9.
This is a condensed summary of the melting systems, reported
in sufficient detail to permit complete classification of the
furnaces. Data from foundries reporting identical, or nearly
identical, emission rates from two or more cupolas of the same
classification are averaged and the data are reported only once
in the exhibit. This averaging prevents undue emphasis being
placed on any particular foundry and testing technique. Lining
characteristics, blast air temperature and method of heating,
charge opening, location of gas take-off, and use of after-
burners are considered in the multiple regression analysis to
determine the relative effect of design parameters on cupola
emission rates. In certain cases where data are limited and
furnaces differ only in one or two parameters, forced compari-
sons are made.
The results of the linear regression analyses show no
clear relationships between furnace emissions and any of the
cupola design parameters. Two trends, possibly warranting
further investigation in the future, are apparent:
1. Eight of the 12 unlined cupolas have emissions
above the median rate of 20.8 pounds/ton of melt, while the
emissions of the 13 acid lined cupolas are below the median
-------
VI - 19
rate. The two reported emissions rates Cor basic lined cupolas
permit no conclusions to be drawn since one was above and the
other below the median.
2. Those cupolas reported as using briquettes in
the metallic charges all have emissions rates greater than
average of all foundries for which emissions rates are avail-
able.
Exhibit VI-10 portrays two correlation matrices from a
second series of multiple linear regression analyses. In this
setice, specific melt rate, specific blast rate, metal to coke
ratio, and blast temperatures are compared to particulate emis-
sions to determine the possible existence of correlations. The
following units are employed in the analysis:
1. Specific melt rate - Melting rate in tons per
hour per square foot area of the melting zone.
2. Specific blast rate - Blast rate in standard
cubic feet per minute per square foot area of the melting zone.
3. Metal to coke ratio - Pounds of metallic charge
per pound of coke charge.
4. Blast temperature - Degrees Fahrenheit. Exhibit
VI-11 tabulates the data for the variables used in the analyses.
The first matrix is the result of data from acid lined
cupolas and the second from unlined cupolas. Insufficient
data are reported from basic lined furnaces to permit a similar
analysis for this classification.
A T KEARNEY & COMPANY. INC
VI - 20
The acid lined cupola matrix indicates a significant.
correlation between emissions and blast rate expressed by
the formula
E = .05 + .07 B
where E = particulate emissions in pounds per ton of melt,
and B = specific blast rate in SCFM per square foot furnace
area. The line represented by this equation is shown on
Exhibit VI-12.
The matrix indicates that specific blast rate provides
the best correlation with particulate emissions of all varia-
bles considered, followed by specific melt rate, blast air
temperature, and metal to coke ratio in descending order.
The regression analysis program, after calculating the
correlation index of the individual variables, calculates the
combined index of the two variables with the highest individual
indices. The process is continued by adding variables one at a
time in descending order of their individual indices until all
have been included, or until the index Improvement by the ad-
dition of one fails to increase the combined effect by a sig-
nificant amount.
In this case, the correlation index of 0.6530 for specific
blast rate is not improved significantly by the inclusion of
specific melt rate in the analysis. This implies that specific
melt rate is not an important factor in the generation of par-
ticulate emissions.
-------
VI - 21
The second correlation matrix listed on Exhibit VI-10 Is
for unllned cupola data and Indicates a significant correlation
between emissions and the two variables, coke rate and specific
blast rate, expressed by the formula
E •» 57 - 6.6 C + 0.1 B
where E = participate emissions In pounds per ton of melt,
C - metal to coke ratio, and B - specific blast rate in SCFM
per square foot furnace area.
The regression analyses indicate that specific blast rate
has a significant effect on emission rate for both acid lined and
unlined cupolas. Coke rate, however, is of greater importance
than blast rate for unlined cupolas and insignificant for acid
lined cupolas. No reasonable explanation can be offered for this
difference between acid lined and unlined cupola emissions. Ad-
ditional observations could possibly show a stronger correlation
between coke rate and emissions for acid lined cupolas.
Oxygen enrichment and natural gas fuel injection have been
presented in recent years as techniques to reduce coke require-
ments, or to increase melting rates when using the same metal
to coke ratio. These techniques have been partially accepted
by the industry because of their substantial advantages but
little research and development work has been done to date
that establishes their effect on cupola emissions. It is
possible at this time to report on only one research effort
for oxygen enrichment and one for gas injection. No broad
conclusions can be drawn from these limited data, but trends
demonstrated are believed to be valid.
VI - 22
Over 20 cupolas are identified In the general foundry
data as making use of oxygen enrichment to increase the total
Oj content in the blast air from 21% to 257., commonly reported
as 4% enrichment. Only one foundry, however, has reported
emissions levels with and without the use of additional oxygen
as follows:
Emissions with 4% 02 enrichment - 1.84 gr/SCF
Emissions without 02 enrichment - .71 gr/SCF
These emission levels are reported for an 88-inch lined
diameter cold blast, acid lined cupola melting 52 tons per
hour with an exhaust gas volume of 40,000 SCFM and a coke
rate of 9.3 to 1. The data are inconclusive since they are
not complete, but they do show an increase in emissions re-
sulting from oxygen enrichment. Information from other sources
indicate that although total emissions are increased, the
improvement in the melting rate with oxygen enrichment results
in a slightly lower emission rate per ton of metal melted.
Additional testing is required to definitely establish the
effect of oxygen enrichment on emission levels.
Several research programs are currently in progress to
determine the effects of natural gas injection as a replace-
ment for part of the coke charge. The results of one such
program are shown below.
Coke
Burner Height Replaced with Production Emissions
Inches Cas Percent Tons/Hour Pounds/Ton
_
50
50
0%
30
40
14.8
20.1
20.3
67.8
57.1
58.5
-------
VI - 23
The tests were run on a 90-inch diameter, cold blast
cupola lined to 72-inch diameter, using a 4,000-pound metallic
charge, 600-pound coke charge and 130 pounds of limestone.
During the tests, the blast air was kept constant at 8,000
CFM, even though coke was replaced by natural gas on equivalent
BTU content basis. The combustion air for the gas was provided
stoichiometrically at a 10 to 1 ratio. The data show decreases
in stack emissions of 15.75% for a 30% reduction in coke
charged, and 13.7% for a 40% reduction in coke charged.
The emission rate of 67.8 pounds per ton reported for
the control condition with no coke replacement is several
times higher than shown in Exhibit VI-12 for a specific blast
rate of 272 CFM/SF. Two special conditions, one inherent in
the test program and the other a factor of weather conditions,
could account for the discrepancy.
First, in the cest program, the blast air was held constant
at 8,000 CFM although the coke charge was decreased by as much
as 60%. This quantity of air was further increased 36.5% by
the natural gas and combustion air for replacement at 30% of
the coke charge and 54.8% by the gas and air for the 40%
replacement. It is reasonable to expect this increase in
gas volume to augment entrainment of sand and small coke
particles thereby increasing emission rates. Second, in answer
to a query regarding the unusually high reported emissions,
VI - 24
the author reported that the tests were performed at a time
when the scrapyard was wet and muddy resulting in the proba-
bility of significant amounts of soil adhering to the scrap
charge, thereby adding to the stack emissions.
Although the emissions appear to be excessive, the data
can be accepted as demonstrating the relative effect of the
injection of natural gas on emissions levels.
The injection of other hydrocarbon fuels including coal
and fuel oil has been reported in the literature. Less im-
portance is attached to these efforts than the injection of
natural gas, and no data pertaining to the effect of these
fuels on emissions have been reported.
Discussion of Results
(a) Introduction
The effect of a number of variables on the type and quantity
of cupola emissions has been evaluated by the use of a variety of
analytical techniques. These variables include the following:
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 after-
burner
-------
VI - 25
Charge Door - open or closed.
Fuel Injection - wlch 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
The analyses demonstrate that some of these variables have strong
and quantifiable effects and others have statistically insig-
nificant effects on emission devices. Six of the variables
evaluated show evidence of the existence of relationships but
lack of data prevents determination of the degree of correlation.
Finally, for two variables, no proof of correlation can be found,
although logic demands that a causal relationship exist.
(b) Variables with
Proven Effect
on Emissions
The established correlation between emission quantities
and coke rate suggests that any operating procedure resulting
in changing the coke rate will affect the quantity of partic-
ulate emissions. This includes heating of the blast air and
oxygen enrichment or fuel injection with an attendant decrease
in the coke requirement. The analysis for unlined cupolas
calculated the effect of coke ratio and specific blast rate
on emission quantities as E = 57 - 6.6 C + 0.1 B. If only
A T KEARNEY & COMPANY Ivr
VI - 26
the coke rate is considered, the relationship is E = 106.4 - 8.6 C.
The index for the latter is lower than when both variables are
considered jointly, but the equation can be used to approximate
the single effect of coke rate on the emissions level. This
effect for unlined cupolas is to decrease emissions by 8.6 pounds
per ton of melt for each unit increase in the metal to coke ratio.
It is not possible to state the exact effect of oxygen
enrichment in conjunction with changes in the coke rate be-
cause of limited data. It appears probable that the potential
beneficial effect of the increase in coke rate permissible with
four percent enrichment will be partially offset by an increase
of very small, difficult to collect metallic oxides, resulting
from increased oxidation in the melting zone.
The injection of natural gas with combustion air provided
at a 10 to 1 ratio does not alter Che oxidizing atmosphere in
the melting zone. It would therefore be expected that emission
rates with gas injection would approximate the results of the
regression analysis as the coke charge is decreased, for a
constant blast air to coke ratio.
The decrease of coke in the charge, while beneficial from
the standpoint of emissions rates and overall melting cost,
presents some metallurgical problems that must not be over-
looked. The coke in addition to providing heat for melting
also serves as a source of carbon often required to meet the
desired analysis. This need is minor when the metallic charge
consists primarily of foundry returns, cast iron scrap, and
-------
VI - 27
pig iron. In recent years, Che increasing cost of pig iron
has resulted in a shift away from this material toward steel
scrap with drastically lower carbon content. This kind of
charge requires a compensating carbon source such as graphite
or the substitution of Cabot coke or Carbo-coke for the stand-
ard coke. No data are available to show the effect of these
materials on emission levels.
The second major factor bearing a strong relation to
emission levels is specific blast rate. Exhibit VI-11 shows
that the specific blast rate varies over a range from 194 to
462 cubic feet per minute, per square foot of furnace area.
Maximum blast rate is determined by the capacity of the blower
and Its motor. Rates below the maximum can be adjusted to
suit the hot metal requirements since a decrease in airflow
slows Che coke combustion rate and the melting rate. Converse-
ly, an increase in blast rate increases the temperature in the
melting zone and the melting rate. The higher melting zone
temperature will tend to increase vaporization resulting in
greater amounts of metallic oxides, and the increase in air-
flow rate will increase its entrainment capability.
Specific blast rate gives no indication of the maximum
rate of gas flow in the cupola, which is strongly affected by
channeling within the burden. However, it is reasonable to
expect that as the blast rate increases, the maximum potential
gas flow rate will be directly affected, increasing the cap-
ability for entrainment of particulate matter.
A T KEARVFY & fOVPAVY
VI - 28
(c) Cupola Design
Variables
The analyses show that cupola emissions rates are not
significantly affected by design factors of the furnaces with-
in 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 Che
gases Co escape out the top. In addition, no significant
effect on emissions rates was found for specific melting rates.
(d) Variables with
Probable Effect
on Emissions
Quantitative determination of correlation has not been
possible from the available data for six factors showing prob-
able effect on emissions levels. Results of test programs
have been presented for oxygen enrichment and fuel injection,
but additional data are required for quantifying the effects
of these practices. Comparison of emission data from cupolas
with charges including briquettes, or not including briquettes,
shows that wich ocher factors being equal, the use of bri-
quettes results in larger emissions rates. Examination of
emissions from acid brick lined cupolas shows lower average
levels than do unlined cupolas but it is not established
whether this condition results from the acid lining icself or
Che fact Chat coke races are generally higher for lined cupolas.
A condicion difficult Co assign quantitative values to
A T KF^RVFY * r o M P» W
-------
VI - 29
is the degree of degradation of coke and limestone when these
materials are charged. Care in handling, degree of weathering,
and efficacy of screening, when done at all, are all factors
determining the quantity of breeze and dust charged into the
cupola. There is general agreement that a direct relationship
exists between the quantity of fine particles of these materials
in the charge and cupola emissions. Only visual observation of
stack plumes bears out Che supposition, since no test data are
available for positive verification.
A final variable with a strong indication of a causal
relationship perhaps requires no proof of existence--that of
the presence in cupola emissions of nonferrous oxides result-
ing from charge contamination. The quantities of oxides of
zinc, lead, tin, copper, manganese, and the like can only
derive from contaminants or alloying materials in the metallic
scrap charge.
Even though scrap is purchased to specification, and
Che analysis is quoted by the dealer, it is possible that
significant quantities of undesirable alloys or contaminants
can be included in a shipment. In many cases, the existence
and quantity of these materials can be determined only by an
analysis of the cupola emissions, slag, and iron.
It would pose an almost impossible probe 1m to obtain a
completely representative sample of scrap charge to determine
a relationship of composition of input to that of the col-
lected particulate matter. There are strong indications that
A T KT^PS'FY ft fOMP^VY
VI - 30
inclusion of burned-m sand on foundry returns and dirt from
the scrapyard results in increased S102 concentrations in the
emission. Here again the collection of data is extremely
difficult—both the determination of the quanciCy of sand in
the charge and Che discribuCion of the sand in Che particulate
emissions and the slag. A similar indication of probable
correlation, without sufficient data Co establish an index
of correlation, exists for the scrap metal thickness and
degree of surface corrosion with iron oxide concentration in
cupola emissions. The problem again relates to the difficulty
of measuring corrosion and determining a reasonably accurace
figure for surface area per unic weighc of Che metallic charge.
Exhibit VI-14 shows Che approximate surface area per ton and
pounds of rust per ton for various ferrous scrap materials.
An additional uncertainty exists in the amount of iron oxide,
if any, converted to elemental iron in the reducing zone and
melted, or being trapped in the slag.
(e) Variables with
Expected Effect
on Emissions
Two variables expected to show large indices of correlation
with emissions levels surprisingly did not. Blast temperature
shows in Exhibit VI-10 an expectedly high correlation of 0.874
with coke rate, indicating that the percentage of coke in the
charge is usually decreased when the blast temperature is in-
creased. Unpredictably, however, neither the coke rate with
an index of 0.223 nor the blast temperature with an index of
0.294 correlated strongly with emissions levels. For unlined
A T KF«iR"-'FV ft
-------
VI - 31
cupolas, practically no correlation exists between blast rate
and any of the independent variables or the emissions level.
No justification for this situation can be established.
No decrease in emissions clearly attributable to the use
of afterburners was recognized by the analyses. A properly
designed and installed system of afterburners will, in addition
to burning carbon monoxide to carbon dioxide, reduce combus-
tibles to ash, provided sufficient oxygen is available and
the residence time is long enough for complete combustion.
The unexpected lack of correlation between the use of after-
burners and emissions level can only be explained by one or
more of the following:
1. Afterburner not large enough or improperly
installed.
2. Afterburner not operating at time of reported
tests.
3. Insufficient retention time.
4. Insufficient oxygen.
5. No combustibles in stack.
The last two items are highly improbable, but any or all of
the other alternatives are possible.
(f) Conclusion
The lack of correlation between standard furnace design
factors and emissions levels requires that the explanation
for the wide variance in type and quantity of emissions lies
with cupola operating factors. This is borne out by the fact
T i, r \ o v r v q, rn v p \
VI - 32
that all variables proven to affect emissions levels, or in-
dicating a probability of affecting emissions levels, relate
more to the operation of the cupola than to its design. These
operating factors can be easily divided into two quite distinct
groupings with some cross effects from one group to the other
The first group consists of variables related directly to
cupola operation, including specific blast rate, blast tem-
perature, type of lining, and operating variables of the after-
burner. The afterburner itself is an emission control device
but adjustment of gas and combustion air is considered here as
a variable for the melting system. These variables are rela-
tively inflexible and are determined by required, or desired,
operating characteristics.
The second group of variables concerns the quantity and
quality of charge materials. These include coke rate, oxygen,
natural gas, coke and limestone dust, briquettes often contain-
ing oil or cementitious materials, and contaminants or alloying
materials in the metallic charge. These factors are highly
variable, often from minute to minute, and are more controllable.
Insufficient data prohibit the quantitive evaluation of
the total effect of all variables in the first group compared
to all variables in the second group. The data suggest, how-
ever, that the type and quantity of cupola emissions are
affected more by the quantity and quality of charge materials.
Certainly little or no limestone dust, coke particles, or oil
vapor and other combustibles will appear in the emission unless
A T K F. * R V F Y A r O M I • \ VV I v r
-------
VI - 33
these materials are charged Into the cupola. Similar state-
ments can be made for zinc, lead, aluminum, chromium, cadmium,
copper, silicon and other oxides, particularly when their for-
mation is abetted by the injection of oxygen, or high blast
rates.
Certain relationships expected to be identified by the
analyses were not discovered. Blast air temperature, with a
demonstrable effect on coke rate, was expected to show a
secondary effect on the emissions level. The use of after-
burners in the cupola stack has been shown to aid in the in-
cineration of combustibles, the result of which would be to
lower the emission levels to at least a noticeable degree.
The fact that these relationships were not identified might be
attributed to two factors possioly affecting all the analyses:
the quantity and quality of the test data.
The limited availability of data has been documented, as
well as the reasons for the short supply. The variable quality
of Che information is due first to the lack of understanding
by foundry operators of the accuracy required of data for
analytical purposes, and second to the several stack testing
techniques in use by different testing laboratories. Where
possible, questionable data were confirmed, and if verification
was not possible, the data were discarded. The greater problem
is in the results of stack sampling tests.
Stack testing is not an exact science at this time and no
single technique has been accepted by the industry. Methods
and equipment used to obtain the data listed in the foundry
» T i r > p M F v (v r- n v P * w i v <•
VI - 34
data bank have been discussed earlier. Repeatability of re-
sults is difficult with any given technique by a single testing
firm, even for & stable emissions producing system. With
relatively unstable conditions in cupola furnaces, and the
generally poor working conditions existing at the top of
cupola stacks, variation in results would be expected. With
this situation further compounded by the use of different tech-
niques, equipment, and testing companies to obtain data for
comparison and analysis, the confidence level of the data must
suffer, despite the high degree of professionalism of the
laboratories performing the tests. As a result of this con-
dition, all data used in the analyses have undergone critical
evaluation before acceptance. Even with this method of re-
viewing data, it is believed that identification of less strong
relationships might be missed in the evaluative processes, and
that only strong relationships will be recognized.
The problem of comparing stack test results obtained with
different test procedures is expected to end with industry
acceptance of the recently published "Recommended Practice for
Testing Particulate Emissions from Iron Foundry Cupolas." The
procedure was developed by a joint industry committee sponsored
by the American Foundrymen'a Society and the Gray and Ductile
Iron Founders' Society, Inc. The committee investigated the
theorectical and practical aspects of stack testing and held
discussions with representatives of the various testing lab-
oratories to gain additional insight before writing the pro-
cedure.
A T KE-iRVFY * COMPANY
-------
VI - 35
The recommended practice, included in Appendix E, is
based upon the American Society of Mechanical Engineers
Performance Test Code 27-1957, Determining Dust Concentration
in a Gas Stream with certain modifications to satisfy the
special conditions existing in the iron melting cupola instal-
lation. Three special conditions are identified as follows
and recommended test procedures are established for each:
1. Sampling raw particulate emissions in the cupola
stack where gas flow fluctuates, flow rate is low, gas tem-
perature is high, and dust loading is extremely uneven.
2. Sampling raw particulates emissions in the inlet
duct to the dust collector where velocities and loadings are
more uniform, flow rate is greater, temperatures are more
moderate, and humidity is high.
3. Sampling cleaned gases at the dust collector
outlet where velocities and flow rates are nearly uniform,
and dust loadings are low.
EVALUATION OF EMISSIONS
FROM OTHER SOURCES
Although the cupola has been considered to be the princi-
pal source of iron foundry emissions which are difficult or
costly to control, there are other areas in the foundry which
also produce these types of emissions. The earlier discussions
in this report covered the equipment and operations in the iron
foundry which produce emissions. Some of these have been ad-
equately and economically controlled with existing designs of
A T KEARNEY & COMPANY. IMC
VT - 36
equipment, while others have proven to present problems which
have not yet been completely solved. In Exhibit V-l, chcse
emissions sources are classified according to the case and
economy with which they can be controlled.
The ease and cost of controllability can generally be
related to two broad aspects: the problems which relate to
capturing the emissions with a minimum of infiltrated air and
the problems of adequately removing the fine particles and
gases from the capture medium before their release to the atmo-
sphere. In general, those operations which in many foundries take
place in open, nonconcentrated, nonstationary locations are
difficult to control although, in many cases, the actual coses
of control may not be high once the emissions are captured.
These operations take place in the scrapyard, mold pouring area,
and the shakeout area.
The other areas w'.iich involve difficulty of capture and
high cost of control equipment include arc melting, magnesium
treatment to produce ductile iron and coremaking, The dis-
cussion in this section concentrates on these areas.
A general problem encountered in analyzing these areas,
other than the cupola, is that relatively little quantitative
information is available regarding the quantity and nature of
emissions generated. As a result, much of the data in the
following paragraphs were developed by analytical methods
A T KEARNEY & COMPANY I x e-
-------
VI - 37
and by adaptation from similar operations in other Indus Cries,
racher Chan from accual field information.
Eleccric Arc
Me1Cing and
Holding
The number of electric arc melting installations in iron
foundries is relatively small, with less Chan 50 known to
exisC in 1959, and approximacely 200 in 1969. This is a small
faccor when compared with che cupola. However, Che Crend co-
ward elecCric melting, both arc and induccion, is increasing
rapidly, and Che elecCric arc furnace will most cerCainly be
a growing faccor in che iron foundry.
Because of ics greaCer importance in Che melcing of sCeel,
much accenCion has been given Co che emissions problems from
arc melcing in Che sCeel indusCry and in che sCeel foundry.
Practically all of Che emissions daCa which have been gathered
have been in chese induscries as discinguished from che iron
foundry indusCry. SCeel producclon and melcing in electric
arc furnaces can be classified inCo Cwo broad groups. The
first involves basic melting practice in which refining as
well as melcing is performed and in which one or Cwo slags are
formed Co eliminate impurities. The second involves acid
melcing practice, in which almost no refining is done and in
which only a minimal slag is produced.
Acid sceel melcing is very similar Co iron melcing in Che
arc furnace, with che principal difference being che common use
VI - 38
of oxygen injecced inCo Che molcen sCeel as a means of rapid
reducclon of carbon content, while iron is melced without che
use of oxygen and, in face, usually must have the carbon con-
Cent raised by carbonaceous additions. With che excepcion of
Che gases evolved during oxygen lancing, che emissions evolved
in both Cypes of acid melcing are similar.
The emissions from iron melcing in Che arc furnace come
from Cwo principal sources--the burning or vaporization of
combustible materials which may be in the charged raw materials,
and the burning of the electrodes and some of Che charge metal-
lies during meltdown. In both cases, the greatest evolution
of gases occurs during the early part of the cycle, when melt-
down takes place and when the electric power consumption is
highest. Although the type and quantity of emissions from
combustion of impurities in the charged maCerials is highly
variable depending on Che nature and cleanliness of Chese
materials, Che gases produced from combustion of Che electrodes
are a known and comparatively conscanC and calculable source
of emissions. Approximacely 9-11 pounds of electrodes are con-
sumed per con of iron melted, producing approximacely 30 pounds
of CO and C02 gases, plus 150 pounds of N'2 which comes from the
air induced inCo che furnace. Additionally, a small quanCiCy
of Che mecallics, principally iron, is oxidized and emitted as
oxide fumes.
A T KF«iRVEY * COMI'\VY
-------
VI - 39
Four of Che iron foundries visiced during the course of
this study were electric arc melting installations, all having
fabric filter collectors for emission control. Samples from
the collectors of three of these were anlayzed for particle size
distribution and for chemical analysis. The results of these
tests, together with reported emission rates per ton of iron
melted, are given tn Exhibits VI-15 and VI-16. Data on the
furnace sizes and dust collector system size are also given.
Although the collected data are insufficient for Che
type of analysis which was conducted for Che cupola, they can
be combined with other uata for acid steel melting as well as
iron melting Co form cerCain general conclusions. Exhibit
VI-16 gives these additional results for a variety of installa-
tions ranging from 2- to 25-ton capacities, all melting in aci v
-------
VI -
representing about 10% of Iron Connage cast. Despite this,
there has been relatively little information gathered regarding
the nature and quantity of emissions produced during this
practice. The treatment agent is commonly a form of magnesium
which can be introduced into the molten iron to produce the
desired effect. Exhibit VI-18 illustrates the various methods
. by which this can be accomplished.
The reaction produced during the magnesium treatement
process is a violent one, but of short duration, although
the degree of the violence varies with the form and method
of introduction of the magnesium. Because of this, only a
relatively low percentage of the magnesium which is introduced
is actually involved in the reactions which produce ductile
iron.'Vh-L'.i .-lie remainder being vaporized and expelled as a
fume. The actual yields vary from as low as 15% to as high as
75%, depending on the treatment agent used and the rapidity
with which it is added to the iron bath. The yield factor which
is most generally accepted is about 30% to 357..
Although several ductile iron foundries have installed
emissions control systems on the treatment stations, none has
attempted to measure the quantity of emissions per ton of iron
treated, or to analyze the emissions to determine the particle
size distribution or the chemical composition. The analysis,
therefore, is based on an analysis of the known reactions
which occur, and the normal yields which are expected during
the process.
A T KF \RS-FY S. TOMPIXY 1 - r-
VI - 42
Magnesium is the principal agent resulting in emissions,
since the alloying materials which are used as carriers of
the magnesium either dissolve in the iron or oxidize to form
slag. A major exception to this is the use of magnesium im-
pregnated coke which evolves CO and C02 gas as well as MgO
fume. The boiling point of magnesium is about 2,025° F, which
is well below the temperature of molten iron and accounts for
the violence of the reaction which takes place. The magnesium
in the inoculant is used up in three ways:
1. Some magnesium will react with any sulfur pre-
sent to form MgS, and become part of the slag. Although iron
which is to be used for ductile iron production is generally
pretreated with a basic material such as Na2C03 or CaC03 to
remove sulfur, there is usually about .02% to .03% of sulfur
remaining. This will be effectively eliminated by the mag-
nesium, using about 0.5 pounds of magnesium per ton of iron.
2. A small quantity of magnesium will dissolve in
the iron, to the extent of about .04%. This amounts to about
0.8 pounds of magnesium per ton of iron.
3. The remaining magnesium will boil off, forming
MgO upon contact with the air. The amount of magnesium added
will vary from 0.12% to 0.30% of the iron treated or from 2.4
to 6.0 pounds of magnesium per ton of iron. Deducting the
1.3 pounds of magnesium which was consumed by sulfur reaction
or dissolved in the iron leaves from 1.1 to 4.7 pounds of
magnesium per ton of iron treated to form MgO fume. This
will result in about 2 to 8 pounds of MgO fume generated per
ton of iron treated.
-------
VI - 43
The fume from the magnesium treatment process will be
largely MgO, with this material accounting for from 60% to
80% of the total, depending on the form In which the magnesium
was Introduced and the violence of the reaction. The more
violent reactions, particularly when silicon-magnesium alloys
are used, will also produce SlOj particles In the emissions.
Iron oxide, as Fe203, will also be found in the emissions and
will constitute the second most important material present,
after MgO.
Particle size of the emissions will be fine for the MgO
and Fe203 portions, with the silica and alumina particles
generally of larger size. These particles are under or>°
micron in size and are difficult to collect, requiring the
use of fabric filters or high energy wet scrubbers.
Exhibit VI-19 gives the reported results from the magne-
sium treatment station of a large gray and ductile iron foundry.
This station was used for ductile Iron treatment, desulfuri-
zation and ferrosilicon inoculation, which explains the pre-
sence of such elements as sulfur and calcium in the catch.
The amount of magnesium in the inoculant was 2.25 pounds per
ton of iron treated. At a yield of 35%, this resulted in 1.45
pounds vaporized, giving 2.4 pounds per ton of MgO. This
amounts to 73% of the emissions actually captured.
Mold Pouring
and Cooling
Molding sands consists of silica, zircon, olivine, chamotte,
and occasionally other mineral grains bonded with clay, ben-
tonlte, Portland cement, plaster of paris, petroleum residues
A T KEARNEY & COMPANY Ivr
VI - 44
wood
r:~=~
Wood flour
Sea coal
Cereal binder
Silica flour
0.5%-
2.5%- 8.0%
0.5%- 1.0*
0.0%-15.0%
:r:::— rr
into the surrounding atmosphere. The natur
-20. While these data refer to
:;::::;:,;;«
-
A r K ¥
-------
VI - 45
from decomposition of water vapor, while the CO comes from
combustion of organic materials.
The volume of gas formed is illustrated in Exhibits VI-21
and VI-22, for various mold materials. Gas evolved ranges from
74.4 to as high as 824 cubic feet per cubic foot of sand at
1,800° F. Only a small portion of the sand adjacent to the
sand-metal interface approaches this temperature and gas forma-
tion drops off rapidly as the distance from the interface in-
creases. Although relatively small amounts of participates are
involved, the potential toxicity of the unburned combustibles makes
collection an important factor. The high temperatures associated
with pouring often result in burning of gas as it leaves the
molds. This afterburning is desirable to completely convert the
combustibles to COj and water vapor and to eliminate explosion
and toxicity hazards, particularly if the mold contains oil sand
cores.
Coremaking
Emissions resulting from coremaking operations are general-
ly in the form of vapor, dust and gas, the type and amount depend-
ing on the nature of the core mix and the coremaking process.
The core mix is typically comprised of silica sand, binder and
moisture. The binders used in coremaking include linseed oil,
core oil, wheat flour, sulphite, pitch, oiless binders, resins,
silica flour, fire clay, wood flour, iron oxide, bentonite,
and silica sand.
A T KFM«N-EY et COMPAN-Y
VI - 46
Core binders that generate a considerable volume of gas
on pouring of the mold are undesirable. A typical core mix
for malleable iron castings might be as follows:
Sand Cereal Moisture Oil Binder
92%-98 .75%-!. 257. 0%-57. 07.-17. 0%-,5%
Core mixes for gray iron castings vary greatly according
to Che general size of the casting and the specific application
for the part. The rate of gas volume generated in a core
during the pouring process is largely a function of baking
time and temperature. Exhibit VI-23 illustrates the effect of
baking time on the volume of gas generated at various baking
temperatures. A review of the curves quickly points out that
the gas content is reduced by baking at higher temperatures.
Resin binders of the type normally used in shell cores
present varying degrees of hazards due to the toxicity of the
decomposition products. Dermatitis is the principal effect
caused by an excess of free phenol, formaldehyde, hexamethylene-
tetramine or alcohol. The extent of the hazard depends upon
the specific agent and the tolerance level for that agent.
Phenol, for example, can cause dermatitis and do organic damage
to the body at levels exceeding five parts per million. Form-
aldehyde is a nuisance at levels exceeding five parts per
million. Hexamethylenetetramine can cause skin irritations
with direct contact.
Other toxic and irritating materials include furfuryl
alcohol, ethyl alcohol, methyl alcohol, urea, carbon monoxide
A T KFARVEY ft COMPANY Ivc
-------
VI - 47
and silica dust. These can be released during shell operations.
Each has varying minimum levels of concentration before Its
toxicity or Irritation are critical or a nuisance. Ventila-
tion becomes the important factor in minimizing these hazards.
The sand to oil ratio has a bearing on the volume of gas
generated in a core during pouring. The effect of sand to
oil ratio on the amount of core gas given off during pouring
is Illustrated in Exhibit VI-24. The relative amounts of
gas produced by various core binders is given in the following
table.
Core Binder
Linseed
Petroleum
Urea resins
Cereal
Cubic Centimeter
Gas per Gram
380 - 450
350 - 410
300 - 600
550 - 660
A T KEARNEY 8= COMPANY, lie
VI - 48
REFERENCES
1. Foundry Air Pollution Control Manual. American
Foundrymen's Society, 1967.
2. jr,n-
Anoaratus. American Society of
Mechanical Engineers, PTC 21-1941.
3 "Influence of Melting Method and Charge Composition
on Cupola Effluent," Mark M. S. Chi and F. EKman, paper pre-
sented before 1970 AFS Casting Congress and Exposition,
Cleveland, Ohio.
4. -,.T-i~*— tstaubunK. G. Engels and E. Weber,
Giesserie-verlag G.m.b.H. , Dusseldorf, translated by P. S.
Cowen, Gray and Ductile Iron Founders' Society, Cleveland,
Ohio.
5 -Gas injection Lovers Cupola Melting Costs, Increases
H.ltin. Rates," J. A. Davis, e^al, J-ndrv, June, 1969,
pp. 66-73.
6 "Readers Comment. Letter from Institut fur
CiessereitechniU," Dusseldorf, Germany, Foundrv, December,
1969, p. «.
A T KEARNEY & COMPANY. I
-------
VII - TECHNICAL ANALYSIS OF
EMISSION CONTROL TECHNOLOGY
ANALYSIS OF EMISSION
CONTROL TECHNIQUES
Over the years, increasing attention to foundry and other
industry sources o£ emissions has prompted development of
many types of equipment for control of the emissions from
these sources. As the significance, complexity and scope
of the problem grew, new designs were introduced or existing
ones modified.
The problems arising from each type of foundry emissions
and the equipment best suited for control vary with the
nature of the specific problem. These systems, which include
dry centrifugals, wet collectors, fabric filters and electro-
static precipitators, vary widely in design, capacities,
capabilities, cost and application.
Dry Centrifugal
Collectors
(a) Historical
Background
Low pressure drop centrifugal collectors were first developed
In about 1880 to remove dust from the secondary circuit of dust
louvres. Higher pressure loss centrifugal collectors were in-
troduced in the 1920's in Europe. At about this time, small
diameter cyclones connected in parallel were being used in the
United States. Although first applied to boiler fly ash pro-
cessing, they were to become the usual application for grinding
T hi « T» V F V ft I- O V V •,
VII - 2
and chipping operations. Design improvements increasing the
collection efficiency were made in Che 1940's and early 1950's
making it possible to apply centrifugal collectors to cupolas,
(b) General Characteristics
of Equipment
Dry centrifugal collectors are essentially low energy
units operating on the principle of mass force action on
individual dust particles. Gravity, inertia, and centrifugal
force act to separate the dust particles from an airstream. It
is iiroortant that this type of collector operate within design
limitations to maintain proper efficiencies. If the gas flow
drops below normal, inertia or centrifugal forces are reduced
and collection efficiences are lowered. Excess moisture in the
gas stream also results in reduced effectiveness by clogging
the system.
Physical characteristic of the dust is an important factor
in the application of centrifugal collectors. Particle size,
shape and density determine, in large part, the ability of
these collectors to function properly. The collection efficiency
required for a particular application is also a major factor
in selection of a centrifugal collector. High efficiency
requirements may preclude the use of this type of collector
with its limited effectiveness.1
(c) Description of
Specific Types
Four basic types of dry centrifugal collectors currently
in use are: cyclone, high efficiency cyclone, high efficiency
A T KEARNEY & COMPANY 1 v <-
-------
VII - 3
centrifugal and dynamic precipitator.
1. Cyclone - This unit type is essentially a conical
chamber with a dirty air inlet on the periphery, a clean air
exhaust at the top, and a dust removal opening at the bottom of
the cone. In a common cyclone design, contaminated gases are
tangentially led into the unit so that a rotational gas flow
is produced. The dust particles travel towards the outside and
bottom of the chamber under the influence of centrifugal force.
The particulate matter enters a collection bin or hopper and
settles, while the cleaned gases, which also traveled to the
bottom of the cone during the centrifugal action, leave we
cyclone through a pipe or tube located in the middle of the
system.
The dust collection efficiency of cyclones depends on
the pressure drop which increases with the inlet velocity
of the dust gas mixture in the collector. The larger the pressure
drop, the smaller the Limiting dust grain size, and the higher
the collection efficiency. As gas velocity increases, collection
efficiency will increase rapidly. However, at a-certain point,
a further increase in gas velocity will not raise efficiency by
any noticeable amount. Collection efficiency of cyclones also
depends on other factors such as the quantity and nature of the
dust, their surface properties, electrical charge and moisture
content.
Electrical charge is important with regard to submicron
particles. If emissions are charged similarly, they tend to re-
pel each other with the result that collection efficiency drops.
A T.KEARVEY & COMP".VY 1 «. f
VII - 4
Conversely, agglomeration takes place and efficiency increases
when positively and negatively charged particles are present.
In addition to the dirty gases being tangentially led
into the cyclone collector, there are other input methods used.
Inlet spiral or axial and radial guide vanes, on the inlet side
of the dust collector, can also be utilized. An example of a
cyclone is shown in Exhibit VII-1.
2. High Efficiency Cyclone - This type of unit is
similar to regular cyclones in design and operation. Differences
exist in Che diameter and shape of the two units. High effi-
ciency cyclones usually have a smaller diameter chamber and a
longer cone than regular cyclones.
Pertaining to the collection operation of these units,
higher efficiency can be obtained in comparison to regular cy-
clones. This is achieved by the increased velocity that exists
with high efficiency cyclone units and greater centrifugal
force resulting from the smaller diameter. Energy requirements
of high efficiency cyclones are somewhat greater than regular
cyclones. Pressure drop is in the 3- to 6-inch water gauge range
with high efficiency units whereas regular cyclones obtain
approximately a 1.5- to 3-inch level. With the greater energy
requirements of a high efficiency unit, a more powerful fan and
drive are needed.
With these units, dust discharge must be continuous to
a dust storage bin that is often equipped with an air lock or
flap gate discharge device. Collected dust must not be stored
in the cone bottom since reentrainment and loss of efficiency
-------
VII - 5
would resulc. High efficiency units generally operate on suction
but can also operate on the pressure side of the fan.
Maintenance of these units requires periodic inspection
of interior surfaces for wear. The discharge device should also
It
be examined for proper operation.
3. High Efficiency Centrifugal - High efficiency
centrifugal collectors consist of a series of small cyclone-type
units in a single housing. Different designs of this unit type
are available. One type utilizes swirl rings that help produce
a cyclone motion in the airstream. The cleaned air is exhausted
out the center section of the cyclone and the dust settles into
a storage hopper. Another type utilizes multiple tubes and a
secondary air circuit to achieve a higher collection of small
particulate matter. In the operation of this system, dirty air
enters through an inlet that has a slot for maintaining a uniform
distribution of the air. The incoming air moves tangentially to
provide the cyclonic action to hurl the particles to the outer
periphery of the tube. The dust and a portion of the primary air
are bled off. The clean air passes out through the end of the
tube into a main duct. Larger dust particles then fall into a
hopper and the remaining dust and secondary air are drawn through
another set of tubes. The air also enters tangentially in these
tubes and the remaining dust particles move outward and exit into
a hopper. The clean air passes back through the center of the
tubes and reenters on the inlet side of the unit. If this type of
unit is used on cupolas, other equipment such as cupola caps,
pressure relief vents, gas burners, cooling towers and high tem-
perature ductwork is required.
A T KK\HM:Y & COMPASS i^c
VII - 6
Due to the high internal velocity of the air, various
component parts may tend to wear. This requires inspection of
all tubes, wall surfaces and hoppers to assure proper function-
4
ing. Pressure drops are in the range of four to six inches
water column, and collection efficiencies somewhat higher than
for high efficiency cyclones. The small dust outlet size can
be plugged more readily and poor distribution within che
collector can cause short-circuiting within the dust hopper
which will reduce collection efficiency.
An example of a high efficiency centrifugal unit is
illustrated in Exhibit VII-2.
4. Dynamic Precipitator - This dry centrifugal col-
lector consists of three main components: an impeller, housing,
and a dust chamber. The impeller's purpose is to move the
air, guide the airflow and force the dust particles to the pe-
riphery by centrifugal force.
The housing has two compartments--one for cleaned
air which comes from the impeller and is projected out the dis-
charge, and the other for dirty air which is discharged from the
impeller. This dirty air goes into a hopper after which a
decrease in velocity takes place that allows the particles to
settle. The cleaned air is then returned into the outer compart-
ment. Collected material is usually emptied from hoppers and
disposed of in the dry state.
There are modifications which can be made to the sys-
tem described in the previous paragraph. If larger particles
or heavier concentrations exist, a skimmer may be used. The
A T KEARNEY & COMPANY I Mc
-------
VII - 7
operation of a skimmer Is similar to that of a low energy
mechanical collector. In addition, aftercleaners are sometimes
used if the dust Is nontoxic. Aftercleaners assist in removing
small participate matter and also allow filtered air to be
returned to the plant. Maintenance of the system is required
4
for such areas as internal surfaces and bearings.
An example of a dynamic precipitator is shown in
Exhibit VII-3.
(d) Advantages of Dry
Centrifugal Collectors
Certain dry centrifugal designs such as cyclones and high
efficiency cyclones are relatively simple in operation and do
2
not require a large investment cost.
Common cyclones are simple to maintain although inspection
of portions of the system are required. Furthermore, dynamic
precipitator installations can be modified in the future with
the addition of skimmers or aftercleaners. This permits a
4
more flexible collection operation.
Finally, protective devices can be installed on some designs
6
of these systems to enable gas bypass in case of failures.
(e) Disadvantages of
Dry Centrifugal
Collectors
These systems are subject to abrasion and therefore require
periodic inspection to operate properly. Common cyclones require
inspection usually to a lesser degree than other dry mechanical
systems because of their larger scale application to coarse
particle problems.
VII - 8
If dry centrifugal collector systems are not operated
within their airflow design specifications, collection efficien-
cies achieved will be lower than what the system is capable of
capturing. Furthermore, compared to the other three basic
emission control systems, dry centrigual collectors achieve the
lowest collection efficiency on fine and medium-sized particle
dust.
With dry centrifugal collectors, a dry dust disposal
problem exists depending on the type of refuse handling equip-
ment utilized by the particular foundry. Dry dust is difficult
to wet, and disposal often results in secondary pollution.
Certain systems such as high efficiency centrifugal units,
when installed on cupolas, require additional equipment for
dust handling and disposal which adds to investment requirements.
(f) Limitations of Dry
Centrifugal Collectors
Dry centrifugal collectors have the greatest limitation
of the four emission control systems under consideration regard-
ing particle size that can be captured. The smallest particle
dynamic precipitators and high efficiency centrigual units can
effectively collect is approximately 10 microns in size. Common
cyclones have a greater limitation in that the smallest particle
4
size they effectively collect is in the 20-40 micron range.
These limitations prevent their effective application to
a number of foundry processes. Dry centrifugal collectors are
A T KEIRN'EV & COMPAVY Ivr
-------
VII - 9
not applied to electric melting processes, mold cooling, pour-
ing, core and paint ovens or oil burn-off furnaces. Finally,
these collection systems are normally not capable of operating
at temperatures above 750 F unless special heat resisting alloys
are employed.
Wet Collectors
(a) Historical
Background
Wet collectors have been applied in foundries to control
emissions since the late 1930's. Low energy units were first
introduced, including the \^~ cap which became popular during
World War II.
As residential areas developed closer to foundry operations,
nuisance complaints from dust settlements increased. Conven-
tional medium energy wet scrubbers such as centrifugal units
were introduced in the 1950's to help solve this problem.
Later, the public demand for removal of finer particles
from cupolas encouraged the application of high energy scrubbers
such as Venturis. With the advent of the venturi scrubber, a
substantial increase in wet cleaning efficiency was introduced.
(b) General Characteristics
of Equipment
Wet collectors all utilize water, or a water solution, as
a means of capturing and removing dust particles from an air-
stream. In general, three basic principles are involved:
1. Saturation of the Airstream - A wet collector
must first saturate the airstream to retain dust particles
A T KEIRN'EY 8t COMPANY INC
VII - 10
captured. If the water use 1 for collection is allowed to cv.ipn-
ratc, the dust patticles w. ..1 be released.
2. Wetting - Wetting of dust particles is necessary,
since the particles cannot be captured unless their surface is
wet.
3. Separation - After thorough wetting jnd i.ipi.in
of the dust, the water droplets containing the dust particles
are separated from the airstream.
Wet collectors can safely handle high temperatures, iijjli
moisture cgntent and fine clay bond particles without obstruc-
tion to air moving or particle collective functions. Further-
more, this type is most practical for foundries having an exist-
ing supply of low cost water and the ability to dispose of col-
lected waste in slurry or sludge form. In fact, collector de-
signs will be determined to a considerable extent by the foundr<<
water supoly and water clarification system being employed. Urn
systems will incorporate self-contained water circulating systems
with either integral or auxiliary settling tanks for dewatering
and removing collected sludge.
For large production foundries, central sluicing systems
are sometimes employed to eliminate multiple disposal points
and in-plant haulage of the collected sludge. Collectors drain
continuously to the sluicing system. Collected material is
settled out in large settling tanks or tailing ponds with water
usually recirculated after clarification.
A T hEARNEY & COMPANY IMC
-------
VII - 11
(c) Description of
Specific Types
Seven specific types of wet collectors used are: static
washer, dynamic precipitator, centrifugal, orifice, centrifugal
spray, flooded bed and venturi.
1. Static Washer - Static washers, the siiiplest form
of wet collectors, are used where moderate to heavy concentration
of dust particles and lower efficiencies are expected. This
design usually consists of a shell or housing containing banks
of water sprays, simple scrubber plates, additional sprays, and
moisture separators. Spray towers or stacks equipped with water
sprays fall in this classification. The lower section of the
collector is cone-shaped to collect and drain the collected
liquid. Additional equipment required for these collectors in-
cludes pumps and/or piping for circulation of water, a means of
disposing of the collected sludge and fans in all cases except
the wet cap which relies on natural draft. The basic principle
for cleaning with a static washer is that the dirty gases come
in direct contact with sheets or sprays of water which flow in
the opposite direction of the gases.
The most common type of cupola static washer is a wet
cap which is shown in Exhibit VII-4. Designed relatively simply,
wet caps usually consist of one or more inverted cones surrounded
by a collection trough. By directing a stream of water over one
or more of the cones, placed above the top of the cupola, a water
curtain is formed through which cupola gases must pass and by
which heavier particles of dust are removed from the gas stream.
Water in a wet cap operation can be applied in numerous
ways: an open waterline can be centered above the apex of the
A T KF«iRVFY ft rOMPIN'V INT
VII - 12
cone flow by gravity; a spray nozzle can be mounted above the
apex of the cone; sometimes clusters of the cone arc arranged
horizontally and vertically; a ring of nozzles may be used;
water can be discharged from a hole in the cone apex and flow
downward by gravity; a continuous plate is also used whereby
water enters in sheets; with the cone hollow, water can be
discharged from nozzles mounted at the periphery; and finally,
a small receivable water distributing head is sometimes mounted
above the main cone and water enters from an adjustable slot at
the periphery of the small cone.
2. Dynamic Precipitator - This type is used to collect
light to heavy concentrations of fine particles such as found
in shakeouts and other sand-handling applications. Essentially,
the dynamic precipitator is a combined exhaust fan and collector.
This design consists of a prepackaged unit made up of an inlet
duct with a spray nozzle, a multi-bladed impeller and a special
housing that separates the cleaned air from the dirty slurry.
The basic cleaning process consists of dirty gases
entering the inlet duct where they are mixed with water.
Specially designed blades then remove particles by capturing the
dust on a moving film of water on the blade surfaces. The
force of the moving wetted blades provides the energy to wet the
particles, and the centrifugal force discharges and separates
the collected dust in the form of a slurry while exhausting the
cleaned air. An example of a wet dynamic precipitator is shown
in Exhibit VII-5.
Dynamic precipitators must be periodically inspected
to assure that the fan rotor and spray nozzles are not clogged.
A T KEARNEY & COMPANY l»c
-------
VTT - 13
The rotor must also be checked for wear and buildup on blades.
This could cause an unbalanced operation and excessive loads on
fan bearings.
3. Centrifugal - Centrifugal wet collectors are used
to remove dust from sand conditioning equipment, casting lines
and cupolas. The two basic designs available in this caccgorv
are the vane type and multiple tube type. The vane-type collec-
tor consists of a vertical cylindrical housing and ono or marc
stationary vanes and impingement plates. The tube type utilizes
a rectangular housing surrounding inlet collector tubes, a water
eliminator section and a clean air section. Both types utilize
simple water piping to produce a film of water on the surface of
the vanes. Additional equipment includes pumps and/or piping
for vater circulation, a method of disposing of sludge and an
exhaust fan.
With a vane-type centrifugal collector, dirty gases
enter the bottom of the main housing at a tangent which provides
the centrifugal force to impinge the larger particles against
the lower section of a shell. The gases and smaller particles
flow upward through vanes which produce a turbulent action.
Dust is removed by coming in contact with the wetted impingement
plates and water droplets. The gas then passes through a. separa-
tor where the free moisture is removed. An example of a vane-
type centrifugal collector is shown in Exhibit VII-6.
In the operation of the tube-type centrifugal, dust-
laden gases are led in tangentially at the upper end of the
collector and then move through multiple tube-type canals.
Centrifugal force hurls the dust particles against the wetted
peripheral surface of the unit. The slurry is then separated
FY & COMP^VV l-ic
VII - 14
from the clean air by being forced down and out the bottom of
the collector. The clean gases move parallel to Che water
stream towards the outlet at the lower end of the dust collector.
Exhibit V1I-7 illustrates a tube-type wet centrifugal collector.
4. Orifice - This collector type is used in sand
handling and cleaning processes. Orifice units are rectangular
and employ one or more impeller orifices that are submerged in
water. The clean air side of the main housing is equipped with
baffle plates to separate entrained moisture. Orifice collectors
utilize the principle of centrifugal force and the mixing
action of dust-laden gases and water to achieve cleaning.
The first step of the cleaning cycle consists of con-
taminated air entering the collector horizontally after which
ic is violently pulled down through water. Fan suction con-
tinuously draws water and air up where it impinges against a
plate and sprays off. This action causes great turbulence, which
in turn wets the dust particles. The gases are then exhausted
out the collector after which a drag conveyor continuously
removes the settled sludge. An orifice-type unit is shown in
Exhibit VII-8.
Water is continually reused and since a water curtain
is produced by airflow, no pumps and nozzles are required.
Furthermore, orifice collectors normally do not require extensive
maintenance because the only moving device is an exhaust fan.
In addition to fan maintenance, occasional inspection and clean-
4
ing of the impeller and moisture separator are also required.
An important operating requirement of baffle-type
4
A T KFMIVFY A COMPANY
-------
VII - 15
orifice collectors is that the water level must be synchronized
with the amount of air flowing through the unit if constant col-
lection and reasonable pressure drop are to be maintained.
Recent orifice designs automatically adjust operating water
Q
levels when the exhaust volume varies.'
5. Centrifugal Spray - This unit is used on a wide
range of foundry applications from shakeouts to cupolas. The
centrifugal spray collector consists of a cylindrical shell in
which are mounted a pump, spray generator, and fan, all on a
common shaft. A separator is irounted on the upper section of
the shell.
Collection is achieved by combining high kinetic
energy water droplets with the contaminated gas. Dirty gas
enters the unit and is directed under a stationary impinger
where heavy particles are impacted and carried off by liquid
water. After this, the air is drawn up through a high velocity
water spray, where impingement of the smaller dust particles
takes place. The cleaned gases continue upward through a moisture
separator and finally are exhausted from the unit. Exhibit VII-9
illustrates a centrifugal spray wet collector.
A variation of the design described above is a scrub-
bing unit consisting of a cone bottom housing and a pump spray
generator device. This unit is usually coupled with other units
such as separate exhausters, hot gas quenchers, moisture elimina-
tors, etc. to make a cupola cleaning system. The cone bottom
unit operates on the same principle as the cylindrical shell
VII - 16
for the removal of airborne water which is handled by the mois-
ture eliminators.
These units utilize much of their input horsepower to
create a high energy water spray. This is different than other
wet-type units which overcome static pressure drop across the
collector.
Centrifugal units require periodic inspection and ser-
vicing of pumps, spray generator cage, bearings, the main shaft
4,
and drive motor.
6. Flooded Bed - The two basic types of flooded bed
collectors are marble bed and impingement baffle grid. Both
designs have common characteristics and also some differences.
Each employs a cylindrical shell with a cone-shaped bottom
section, fixed blade moisture separators at the top of the shell
and spray nozzles below the first stage of scrubbing. Further-
more, both types operate on the principle of direct contact of
dirty gases with water spray in conjunction with venturi-designed
induced turbulence. In a typical marble bed collector, gases
enter the collector near the bottom and then turn upward into a
water spray. These water sprays are used to wet the lower
portion of the bed. This causes the larger dust particles to
separate and drop down. The gases and water are then drawn
up by an exhaust fan through a bed of glass or other type balls
where smaller particles are captured by the restricted flow.
The cleaned gases ascend further and move across moisture
separators and then out the collector. The dust-laden slurry
is drained down and into a sludge hopper. An example of a
A T KF".RVFY
-------
VII - 17
marble bed collector is shown in Exhibit VII-10.
Efficiency of this unit can be increased by additional
layers of balls. Spheres packed in contact with each other
constitute a multiplicity of Venturis to achieve this result.
The impingement baffle grid collector uses one to
three grids which consist of a perforated plate and impingement
plate above each perforation. An essential difference of the
impingement baffle grid from the marble bed unit is in the water
supply. Each scrubber section of the grid type has an individual
water supply in addition to the water spray located in the cone
section. Exhibit VII-U shows a grid-type flooded bed
4
collector.
7. Venturi - This unit is usually applied where heavy
concentration of both coarse and extremely fine particles exists
and high efficiencies on subraicron particles are desired. In
a typical cupola system, hot gases move from the furnace to a
quencher where they are cooled by a water spray. Here, the
large particulate matter is removed. From the quencher, gases
are directed to the venturi where they are accelerated to a high
velocity and brought into contact with multiple jet streams of
vater. The water and gases are drawn through the constricted
throat of the venturi at such a high velocity that the water
atomizes into minute droplets and dust particles collide with
the droplets and become wetted. This wetting occurs in the dif-
fusion area downstream from the venturi throat. The dirty water
is carried out the bottom and the gases pass through a moisture
VII - 18
separator, are cooled and Chen exit through n saturated gas stack.
A high pressure exhaust fan, piping and electrical controls an-
part of the system. Exhibit VI1-12 illustratcs n venturi
collector.
Venturi scrubbers achieve a high efficiency by virtue
of a high pressure drop occurring across the venturi throat.
Some Venturis are designed with variable throats so that a more
uniform pressure drop can be maintained with variable gas flow.
A variation of the venturi described above combines a
marble flooded bed collector with a venturi section mounted un-
derneath. Nearly all of this system is encompassed in one large
chamber. In the operation of a flooded bed venturi, dirty gases
enter the collector where they are met by a cooling water spray.
The gases then turn upward into the venturi and the large par-
ticles drop out of the airstream into a cone section. Gases
and fine particles pass across the venturi and the particles
become wetted. The marble bed and moisture eliminator act to
4
further clean the gases and remove water drops from the effluent.
(d) Advantages of
Wet Collectors
Wet collectors can be applied to a variety of foundry appli-
cations. Certain designs, such as wet caps, require a relatively
low iiitial investment cost, limited space needs, no induced
draft fan and are rather simple to maintain.
Venturi scrubbers have good collection efficiencies down
to particles of 0.1 or 0.2 microns in size but also have high
£,
energy requirements to achieve this cleaning capability.
A T KEARNEY & COMPANY Isc
-------
VII - 19
In addition, wet collectors are capable of handling hot
gases, sticky dust and high moisture content. In some cases,
disposal of sludge can be achieved if foundries have slurry
pumps, settling ponds or tanks, drag chain sludge removal
1
conveyors or other similar equipment.
Finally, these system types are safer from fire hazards
than fabric filter collector designs.
(e) Disadvantages of
Wee Collectors
One major disadvantage is that operating costs can be high
for wet collectors in relation to other collectors due to water
usage, corrosion, and the chemical treatment needed for water.
Also, the high energy units, such as Venturis, require a high
power usage.
Another disadvantage is that since the cleaned gas exiting
from the foundry is saturated with moisture, a plume of condensed
steam develops which is visible until the vapor trail evaporates
This condition can be mitigated through the use of additional
de-misters. Furthermore, submicron dust sizes are difficult to
extract in dewatering equipment. Acid pickup from burning coke
and oil from scrap also occur.
If complete control of all solid emissions is required at
a later date, it is often impractical to make certain equipment
modifications to wet collectors such as use of aftercleaners.
However, there have been instances where medium energy systems
have been designed with the intent of making future equipment
A T *r»n^rv ft i- o M P s vs I •. r-
VII - 20
modifications to increase cleaning capabilities. Finally, duo
to Che need for conservation of water, additional equipment for
2
circulation and clarification is required.
(f) Limitations of
Met Collectors
Exhibit VII-13 indicates the smallest particle that can
be collected by the seven types of wet collectors. As shown,
static washers are not capable of collecting dust particles
under 10 microns in size and thus have the greatest limitations
in this regard. Conversely, venturi scrubbers, being able to
collect particles as small as O.S microns in size, have the
4
least Limitation.
Wet collectors are not applied for collection in certain
foundry areas such as boiler fly ash, pouring, mold cooling
and pulverizing. Also they are rarely used on woodworking and
oil burn-off furnaces.
Capacity ranges available for the different types of
4
vet collectors are also shown in Exhibit VII-13.
Wet collectors have no practical limitation regarding
temperature or moisture conditions that may exist during opera-
tion.6
Fabric Filters
(a) Historical
Background
Until the mid-1930's, foundry fabric collectors consisted
A T KEMI^FY & COMP"lVY
-------
VII - 21
of large U-shaped frames, originally of wood, Co which a cotton
cloth or canvas was cacked or otherwise attached to both sides
of the structure. Panels were installed vertically with the
top sealed to an adjacent header so dust-laden air was forced
through the fabric with the cleaned air gathered in the inside
of the support structure and exhausted through the open top.
Periodically, the collector was stopped and the accumulated dust
load dislodged by vibration.
During the early 1930's, design improvements were intro-
duced tc provide more effective vibration of the filter cloth,
more positive seals between fabric and support, and faster media
replacement. It was during this period that Che current concepts
of the envelope and cloth tube types were developed. Sectional-
ized, continuously operated collectors were a logical refinement
for larger foundry applications where the collector could not
be stopped with the frequency needed to maintain the desired ex-
haust flow.
It was the pressure on the industry by the Los Angeles Air
Pollution District that generated the use of glass fabric for
cupola gas cleaning during the late 1940's and early 1950's.
Typical early installations consisted of relatively crude de-
signs where long tubes were suspended from a mechanism that was
manually operated to "rock" the cubes during shutdown. Produc-
tion of glass cloth with a silicone lubricant permitted this
mild flexing action without rupture of the glass fibers. The
A T h r » n v r v
<-OM p \ \ v
VII - 22
sensitivity of glass fabric to flexing encouraged the develop-
ment of collectors of the reverse flow concept to provide more
effective collected dust removal and continuous operation of
uie collector.
Fabric filters for electric arc furnaces in the gray and
malleable iron foundry use exhaust control methods and generally
orlon or dacron fabric based on practices developed in steel
foundries. Local exhaust hood concepts date back to development
work by the farm implement industry foundries in the mid-1940's.
While medium pressure loss wet scrubbers were used on early in-
stallations, they gave way to the higher efficiency of the fabric
c
filter by the mid-1950's for electric arc furnace fume collection.*
(b) General Characteristics
of Equipment
The basic principle of fabric collection consists of dust-
laden air passing through a filter media. The media acts as a
base to build a mat of the collected material on the dirty air
side. As the dust builds up on the fabric, resistance to airflow
increases. Periodically, the airflow must be stopped and the
accumulated cake removed. This can be done by vibrating the bag
and letting the dislodged material fall into a hopper. It can
also be accomplished by using reverse or secondary airflows.
Fabric filters are quite versatile in their contaminant
removal capabilities. Since particles as small as 0.2 microns
in size can be captured, excellent collection efficiencies can
be achieved. Furthermore, dust concentrations can be heavy
A T KFMJVEY A roMP^VV T-
-------
VII - 23
or light and capacity ranges can also vary with minimum effect
on efficiency.
A significant factor regarding operation of fabric filters
is the ratio of cubic feet per minute of air handled in the sys-
tem to the square feet of cloth area of the collector. This is
usually referred to as the air to cloth ratio. By lowering the
ratio, less air velocity through the fabric is achieved, and
bag wear decreases. However, economic factors also need to be
4
considered in determining the ratio.
The filter cloth is usually a specially woven cotton. How-
ever, dacron or orlon is used for temperatures reaching 225° F,
and glass for 550° F temperatures. The units must be operated
above the dew point temperature of the airstream so that moisture
does not form to produce a mud cake to hamper air passage through
the media.1' *' 5
Two basic types of cloth filters are available—the flat-
or envelope-type bag and the tubular-type bag. The flat- or
envelope-type bag generally provides a more compact filter and
better cleaning because the flat bag is kept taut. Dust is
collected on the outside surface of flat bags. They also require
a rigid framework so that collapse of the bag can be prevented.
Exhibit VII-14 illustrates a flat-type bag used on a fabric
filter.
The tube- or stocking-type bag is suspended from the shaker
mechanism at the top and fastened on the bottom. Dust is
A T KEARNEY & COMPANY
VII - 24
captured on the inside of the tube.'
tube unit.
Exhibit VII-15 shows a
The key to successful operation and continued efficiency
of fabric collectors is planned maintenance. Bags need to be
inspected for leaks and cleaned. Shaker mechanisms and baffle
plates have to be checked for proper cleaning action and wear.
The valves and fan also require inspection to determine if they
are functioning properly.
The results of numerous investigations performed on dust
separation processes with fabric filters can be summarized as
follows:
- With increasing temperature, the collection
efficiency drops and the filter resistance
increases.
- The separation of particles larger than 0.5
microns takes place through an impingement
and a sieving effect. Particles smaller than
0.5 microns are separated mainly by diffusion
and electrostatic forces.
- The separation of particles takes place mainly
on the outer surface of the filter fabric.
The filter resiseance naturally increases with
thicker filter fabric, but the collection
efficiency does not improve in the same propor-
tion.
- The smaller the diameter of the fibers, the
smaller the particles which can be separated.
Stronger fibers have to be incorporated into
technical filter fabrics for strength.
- The filter resistance of fabrics is directly
proportional to the air or gas throughput.2
(c) Description of
Specific Types
Three types of fabric filters used are: intermittent, con-
tinuous automatic, and reverse jet continuous. The basic
A T K J- "iRVE Y & COMP^S•Y l^r
-------
VII - 25
principle of fabric collection previously described applies,
in general, to each of the three systems except for bag clean-
ing.
1. Intermittent - This unit type lends itself to
collection of widely scattered dust of an intermittent nature
where the entire system can be shut down periodically to be
cleaned. This type of collector can be used on any system that
can be interrupted or bypassed during the cleaning cycle.
Either the fabric filter, vibrating mechanism and exhaust fan
arc packaged in one unit or the exhaust is separate.
The intermittent-type collectors in the smaller sizes
often are factory assembled, eliminating the need for field
erection. Larger fabric collectors are available that contain
up to 12,000 square feet of cloth area. Field erection of these
«i 5
units is required in many cases.
Exhibit VII-16 shows an example of an intermittent
collector.
2. Continuous Automatic - With a continuous automatic-
type collector, one section of the collection system can be shut
down and remaining sections will continue to operate. The
system consists of multiple collection sections, usually not
less than four or six, each of which can be programmed to clean
for a desired time period, progressing through the entire collec-
tor automatically. High concentrations of dust can normally be
handled and a fairly uniform pressure drop is achieved with this
type of unit. '
r n M i* \ v ^
VII - 26
Some reverse flow designs rely on flexing during
collapse and reinflation to dislodge collected dust without
mechanical vibration. In these designs, the reverse flow of
air through the bag has to be carefully controlled or it could
damage the bags. A sudden full reinflation of Che collapsed
bag could result in bag failures at the points where severe
fiber flexing occurs such as along the edge of a seam. One manu-
facturer handles this problem by gradually introducing the flow
of reverse air. Then, the reverse air valve is kept open tempo-
rarily when the filtering process begins, to allow a gradual
buildup of pressure. Once the proper pressure is restored, the
5, 10
reverse air valve is closed.
Continuous automatic-type systems require additional
equipment such as spare sections, damper assemblies, multiple
shaker units and field assemblies.
Also, the damper and shaker units require additional
maintenance. An example of a continuous automatic system is
shown in Exhibit VII-17.
3. Reverse Jet Continuous - Reverse jet continuous
units are used where a high concentration of fine particles
is being collected or where the material must be continually
collected. This unit utilizes a. secondary source of air to dis-
lodge the dust cake. As the dust particles accumulate, airflow
resistance through the filter increases. At a predetermined
resistance level, a high velocity air jet penetrates the filter
tube through a blow ring moving along the tube. This air jet,
combined with the flexing action of the filter tube, dislodges
-------
VII - 27
1, «
the accumulated dust which is then carried down into a hopper.
Secondary air is provided through use of an attached
blower. One manufacturer's design utilizes an externally located
pressure blower that is connected to the main collecting unit
by a flexible heavy-duty, industrial-type hose. With the pres-
sure blower, various types of air intake filters are used to
control incoming air. Environmental dust conditions and the
specific maintenance requirements of the system are two factors
determining the type of air intake filter to use.
The advantage of this system is that cleaning is
usually more efficient than shaking. Also, less cloth area per
cubic foot per minute is required compared to other fabric-type
collectors. Operating costs and maintenance expenses are
1» 4
reportedly higher than shaker-type fabric collectors.
Exhibit VII-18 illustrates a reverse jet continuous
fabric collector.
(d) Advantates of
Fabric Filters
Well-designed fabric filter installations provide excellent
collection efficiency on dust particles down to 0.2 microns in
size and can be designed to handle large gas volumes. In
addition, dispersion of exit gases into the atmosphere does not
present a steam plume problem.
Fabric filters normally operate with a low pressure drop,
three-inch to five-inch, and, therefore, have low power require-
ments. Fabric units are not normally subject to corrosion,
A T KEARVEY & COMPANY lie
VIT - 28
L, 4
and furthermore, no disposal of waste water is required.
(e) Disadvantages of
Fabric Filters
Initial equipment and erection cost can be high and large-
space requirements are usually necessary. Filter bags need
periodic replacement, adding to the maintenance costs. Added
safety precautions are sometimes necessary for cleaning in-
flammable gases. Furthermore, there is the danger of tempera-
tures sinking below the dew point as a result of evaporative
cooling which would cause moisture to form and clog filters.
Carry-over of glowing particulate matter represents a danger
to natural or synthetic fiber bags.
Dry material disposal can be a problem as particle sizes
of collected material decrease. Refuse handling and secondary
dust dispersion are factors that must be considered in this
. 1. 2, 4
regard.
(f) Limitations of
Fabric Filters
In a normal installation, the smallest particle size that
can be effectively captured by a fabric filter collector is
approximately 0.2 microns in size. Special equipment modifi-
4
cations could result in increased efficiencies.
There are temperature limitations that must be maintained
with respect to different fabrics as shown on the following
49
page.
-------
VII - 29
Fabric
Cotton
Wool
Nylon
Dacron
Nomex
Orion
Creslan
Dynel
Poly-
propylene
Teflon
Fiber glass
Filcron
Melting Temperature
Decomposes at 302° F
Chars at 572° F
480° F
482° F
Decomposes at 700 F
o
Softens at 482 F
o
Softens at 475 F
Softens at 325° F
333° F
o
Decomposes at 750 F
Emits toxic gas at
4 SO0 F
o
1,470
Softens at 505° F
Recommended
Maximum Continuous
Operating Temperature
180° F
200° F
200° F
275° F
350° F
260° F
250° F
160° F
200° F
400° F
550° F
270° F
Fabric filters are not applied for collection in certain
foundry areas such as boiler fly ash, pulverizing and oil burn-
off furnaces, nor are they suitable for handling emissions from
oily scrap and phenols from shell molds and cores.
Pertaining to capacity ranges, small intermittent fabric
collectors are designed to handle from 300 to 2,000 cubic feet
per minute. Capacity rates can be varied quite widely for large
intermittent, continuous automatic and reverse jet continuous-
type fabric collectors.
A T kFM»V|-Y & rOMPAVY
VII - 30
Fabric materials have certain limitations that appear note-
worthy. Wool is not suitable when dirty gases contain tar or
have an acidic effect. Glass fabrics are less flexible than wool
and cannot withstand strong mechanical stresses. Finally, glass
is not suitable for handling fluoride compounds since the lubri-
cants and fibers are subject to chemical attack.
Electrostatic
Precipitators
(a) Historical
_ Background
The first effective electrostatic precipitator was designed
in 1883 for cleaning top gases from a blast furnace. Since
production of direct voltage through friction was unsatisfactory,
further application was not attempted at that time. In 1906,
an electrostatic precipitator was introduced that used a
mechanical high voltage rectifier. In 1916, a dry-type
itator was designed followed by wet systems shortly thereafter.
The electrostatic precipitator for cupola gas cleaning has
been more widely accepted in Europe than in the United States
where the trend in recent years has been strongly in favor of
high energy wet scrubbers. Only two precipitator installations
in foundries have been reported in this country in the last two
years-one in Texas and the most recent in Indiana.
(b) General Characteristics
_ of Equipment __
This type of collector utilizes high voltage, low amperage
direct current for operation. A complete system usually consists
A T KEARSFY & COMPANY \-.<~
-------
VII - 31
of the following: precipitator housing, discharge electrode,
collecting electrodes, gas distribution system, cleansing de-
vices, dust and sludge removal and electric power supply facili-
ties. Wires or rods are suspended from an overhead high voltage-
charged grid between grounded plates. These wires are hung on
approximately 2- to 3-foot centers and carry 50,000 to 100,000
volts. When the voltage reaches the corona stage, the gases
passing through are ionized, forming negative and positive ions.
With the wires negative and the plates positive, the negative
ions travel across to the positive collecting plates and in the
process negatively charge the dust particles which also then are
attracted to the positive plates. Removal of dust particles
varies depending on whether the unit is a wet- or dry-type pre-
cipitator.
Performance of electrostatic precipitators essentially de-
pends on the characteristics and relationship of the suspended
particles, velocity of the gas passing through the collector
and electrical resistivity of the material deposited on the elec-
trodes.
Pertaining to the suspended particles, it is important to
know their size, amount of particles in suspension and chemical
analysis. Particle size will have a bearing on whether primary
collection preceding the precipitator is required. The amount
of particles in suspension has an effect on design efficiency
required to meet emission codes. Chemical composition relates
to particle size since some nonferrous metals product fine par-
ticles.
* T K F M» V F Y (* f n w «" S VY !>»•
VII - 32
The velocity of the gas passing through the collector is
determined by the length of time the particulate matter must
remain in an electrical field. This period determines the
number and size of collection tubes needed to treat the total
volume of gas at this velocity.
Electrical resistivity pertains to the measurement of the
ability of a particle to resist an electrical unipolar charge.
This measurement indicates the amount of conditioning necessary
and the length of time a particle needs to remain under the
influence of the electrical field. Resistivity varies over a
wide range depending on the temperature and moisture content
of the gases.1' 2' *
(c) Description of
Specific Types
Electrostatic precipitators can be either wet- or dry-
type collectors:
1. Wet Type - These units usually consist of a cir-
cular shell or casing that is divided into two gas compartments
by a header plate. Suspended from the header plate are a number
of tubes that form gas ducts. Discharge electrodes are suspended
in the tubes from an insulated framework. Dirty gases pass from
the lower compartment through the tubes for cleaning and then
into the upper compartment. Dust particles are then removed from
the surface of the collecting pipes by water flushing.
The discharge electrodes are energized by electrical
equipment consisting of a manual or automatic control unit,
high voltage transformer and rectifying equipment. Control
A T KEARVEY & COMPtVY l^<-
-------
VTI - 33
equipment is usually all electronic. Power level or sparking
rates act as electrical feedback signals that activate the
control equipment. With this type of equipment system, electri-
cal input under varying process conditions such as temperatures,
flow rates, dust particle size and resistivity is optimized.
Rectifying equipment can be either mechanical, motor driven,
electronic or solid state.
Wet-type units are frequently built vertically.
In this arrangement, gas flows through the filter in a vertical
direction and normal1;, from bottom to top. This results in a
longer period of stay for dust particles in the collector be-
cause of the gravitational force and produces better separation
1 2 A
of coarse particles. ' '
Exhibit VII-19 indicates a wet-type electrostatic
precipitator gas cleaning system.
2. Dry Type - The basic structure of these units con-
sists primarily of a rectangular shell or casing from which are
suspended a number of collecting electrode plates. These plates
are parallel to each other and equally spaced to form channels
through which gases pass. High tension discharge electrodes are
suspended vertically in the center of the channels. Collected
material is removed from the surface of the collecting electrodes
by a continuous rapping device.
Dry process electrostatic precipitators are usually
built on a horizontal basis. A horizontal arrangement allows the
gas to move in this direction which facilitates the cleaning o£
A T h r MI ** E Y & f O M P \ V ^
VII - 34
collecting electrodes by means of rapping or shaking. Dust docs
not fall against the gas stream but only at right angles to it.
The rapping mechanism usually consists of a number of
hammers that are driven pneumatically or by means of impact
shafts striking against the electrodes in regular intervals.
The electrodes can also be deflected horizontally out of their
normal position over a cam shaft and then made to strike against
an anvil. The discharge electrodes are rapped with hammers
that are lifted by tie rods and then released. Regular mainte-
nance is necessary for the electrode rapping equipment.
The collecting electrodes should represent a freely
oscillating system with a minimum damping capacity. This allows
a uniform distribution of the oscillations over the plate sur-
face, so that particles of all sizes can be removed from the
124
electrodes during rapping. ' '
A dry-type electrostatic precipitator gas cleaning
system is illustrated in Exhibit VII-20.
(d) Advantages of
Electrostatic
Precipitators
These unit types provide excellent removal of solid
particles when the collector is properly designed to enable such
performance. Power requirements are substantially less for
electrostatic precipitator units than for other induced draft
collection systems.
Furthermore, these collector types can be designed to
handle moisture, oil and similar conditions that are unsuitable
A T KEARVEY 8. COMPANY Ivr
-------
VII - 35
for certain other collector designs. Finally, these units are
flexible in that additional cells can be added to existing
equipment to meet changing process requirements or more
124
demanding codes. ' '
(e) Disadvantages of
Electrostatic
Precipitators
For small gas volumes—50,000 CFM or less—costs are dis-
proportionately high on a per CFM basis because of power equip-
ment and engineering cost requirements if the unit is not an
"off che shelf" design. Maintenance expense Is greater than units
having a lower order of collection.
Gas temperatures and water vapor content need to be closely
controlled due to the resistivity characteristic of particles.
Fluctuations of cupola stack temperature can make it difficult
to maintain proper temperature of gases going to the collector.
If the effluent emitting from the cupola contains iron and
silicon oxide and the collector is operating at temperatures of
300° F or above, problem' could develop. At these temperatures,
the resistivity oE silica will increase to the point where it
is necessary to condition the gases, by adding moisture. This
moist condition will accommodate silica, but the iron oxides
will tend to coagulate on the collector plate and be difficult
Co remove.
Other disadvantages are chat a precleaner is sometimes
necessary for heavy loads and a dry dust disposal problem may
VII - 3f>
develop. Finally, there is a potential risk involved with
using high voltage equipment. ' * ' '
(f) Limitations of
Electrostatic
Preeipitators
The smallest particle that can be collected with a normal
electrostatic precipitator installation is 0.1 microns in size.
Increased efficiencies are possible through equipment modifi-
cations .
These collectors are limited in their application to
different' foundry processes and consequently have been very
rarely used. Furthermore, these units should not normally be
o 1, 2, 4
used when temperatures exceed approximately 700 F.
Collection Efficiency
of Emission Control
Equipment Systems
The success of a control equipment system will depend upon
its ability to remove sufficient emissions to provide an
acceptable order of cleanliness. Criteria for evaluating
collection efficiency will be determined by the objective of
the system and often by regulations defining acceptable emis-
sion levels.
Accurately defining equipment efficiency is difficult
since inlet loadings vary widely in concentration, particle size
ranges and dust composition for most foundry operations. This
occurs not only between similar operations in different plants,
but also with any given application, almost minute by minute
A T KF \nvrv ft
-------
VII - 37
during a normal working day. Consequently, designers of control
equipment find it difficult to define exit or cleaned gas con-
ditions other than as a function of incoming particulates. The
lower the order of collector performance, the more difficult
it will be to predict exit conditions except in terms of over-
all percentage of removal or percentage removed by particle
size fraction.
Yardsticks used to express collection efficiency of
emission control equipment include- (a) weight, (b) particle
size count, (c) opacity and (d) ground level concentration.
(a) Height
The most acceptable yardstick for expressing collection
performance is on the basis of the weight of emissions collected.
Different scales are used in this regard including pounds per
hour, pounds per unit of process weight (tons melted, and
sand conditioned), pounds per 1,000 pounds of gas and grains
per standard cubic foot of gas.
Regarding the grains per standard cubic foot of gas scale,
performance Is calculated on the basis of the ratio of weight
of particulate per unit of gas entering the collector to the
weight of particulate per unit of gas trapped in the collector.
This ratio is then expressed as a percentage. An example of
how efficiency is calculated in this manner is given in the
tabulation shown on the following page.
A T Kr»BVrV 5.
VII - 38
Dust entering collector -- 1.0 grain/cubic foot
Dust leaving collector -- O.I grains/cubic foot
Dust trapped in collector-- 0.9 grains/cubic Coot
Thus, 0.9 grains/cubic foot divided by 1.0 grain/cubic foot =
4
90% dust collector efficiency.
Exhibit VII-21 indicates relative efficiencies of collector
systems based on outlet loadings in grains per cubic foot of
gas for various foundry applications. These are typical per-
formances that can vary based on differences in melt charge,
dust characteristics, particle size distribution, collector
design and numerous other factors. By referring to the figures
on Exhibit VII-21, collectors can be ranked according to effi-
ciency as shown on the following tabulation.
Type Collector Collection Efficiency
Fabric Filter Highest
30"-70" wet scrubber and
electrostatic precipitator
6"-30" wet scrubber
Low efficiency cyclone
Wet cap Lowest
These efficiency rankings are general, overall indications
that are not related to costs, collector design, specific appli-
cations or other factors. They are simply an indication of the
relative performance that has been experienced with these sys-
tems.
A T KFARWEY St COMPANY Ivr
-------
VII - 39
(b) Particle
Size Count
A count of particle sizes collected is another method wide-
ly used to express efficiency of collection equipment. This
requires knowing the fractional efficiency of a collector at
various micron size ranges and also the particle size distribu-
tion.
Particle size efficiency curves are available for most
commercial collector designs, although any such base curve
must be modified for the influence of particulate concentration,
particle shape, specific gravity, chemical characteristics and
various other factors to fit a specific situation. Where inlet
conditions cannot be accurately specified because of variable
composition and rate of contaminant release, new processes not
yet in operation, or high cost of data development, specifica-
tion of collector performance on a fractional basis is usual.
A convenient way of comparing relative performance of
various collectors by this method is to determine overall col-
lection efficiency based on a standard test dust. This was
accomplished by 1956 and the test results are presented in the
following paragraphs.
The test was conducted on a silica powder which has a grad-
ing about the same as typical fly ash from a pulverized fuel
boiler. The grading of this dust is shown in Exhibit VII-22.
As shown in the exhibit, the largest percentages of dust were
in the 20- to 75-micron size ranges and also under 2.5 microns.
VII - 40
Based on this standard dust, collection efficiency was then
calculated for various collectors, the results of which appear
in Exhibit VII-23.
The efficiencies shown at five microns were taken from
grade efficiency curves developed for each collector. A
ranking of overall collector efficiency based on these results
is shown in the following tabulation.
Type Collector Collection Efficiency
Fabric filter Highest
Venturi scrubber
Wet impingement scrubber
Electrostatic Precipitator
Spray tower
Self-induced spray deduster
High efficiency cyclone
Medium efficiency cyclone Lowest
As discussed previously, these figures serve to place the
collectors in order of merit in terms of collection efficiency
only. A more meaningful assessment includes consideration of
costs and other relevant factors.
(c) Opacity
Opacity, or discharge appearance from a stack, is another
yardstick often used to measure collection efficiency of control
equipment. The Ringelmann Scale shown in Exhibit VI-1 is one
of the most commonly used measures of this type.
A T KFM»VFV
-------
VII -
A major limitation of opacity measures, which were discussed
in Section VI, is that discharge appearance from a stack will
seldom indicate the degree of collection. For example, where
only coarse particles are released, such as snag and portable
grinding, there would be no visible escapement if exhausted dust
were discharged to the atmosphere without collection equipment.
On the other hand, removal of 90% of the metallic oxides
generated by the electric arc furnace could make little or no
reduction in the visible plume.
Because acceptable discharge appearance for most foundry
processes occurs where solid concentrations do not exceed 0.05
grains per cubic foot of gas, this weight measure is often sub-
stituted for performance in terms of equivalent opacity.
(d) Ground Level
Concentration
The objective of air pollution and public nuisance
legislation is, in part, one of controlling the air quality at
ground level. However, measured in terms of tons per square
mile, micrograrss per cubic meter, or as a soiling index, it can
only indirectly be related to the condition of a discharged
quantity of process air because of diffusion, turbulence, wind
velocity and direction, and contributions from other sources.
Factors Affecting
Control Equipment
Collection Efficiency
A number of operating conditions and other factors affect
gas cleaning performance. Certain of these have more effect
T i\r » n vr v
VII - 42
on efficiency than others while some conditions are difficult
to evaluate because of limited factual knowledge on their
influence. Furthermore, some conditions appear to affect all
control systems while others affect only one or a few systems.
1. Melting Operations - The cupola melting operation
results in exhausting large volumes of high temperature gas
laden with dust particles and fume. It is difficult to arrive
at valid generalizations on how melting operations affect
collection efficiency as evidenced from the following statements-
It would be helpful to the average
foundryman if an average emission
load from the cupola melting opera-
tion could be established. This
is impossible, however, due to the
varied operations from one foundry
to another, and due to the complex
melting cycle, metal composition
and operations of various foundries....
The volumes of gas exhausted and the
quantities of coarse and fine materials
making up the total emission vary in
cupola operations depending on melting
rate, charging methods and cleanliness
of materials. "
This conclusion was also reached in an analysis of
cupola furnace metallurgy. It was found that values for coke
rate, blast volume, and blast temperature generally applied in
practice result in such widely varying melting conditions, that
a universally valid statement on the effect of chemical con-
stituents in charge materials on the characteristics of the
reaction products, namely gas dust, cannot be made.
This problem is further exemplied from the following
findings:
The raw gas contents ascertained during
the...survey did not show any relationship
A T KFM»VFY ft rOMPVVS
-------
VII - 43
between dust concent on Che one hand and
coke race and specific melting race on the
other.... The dust content of the waste
gas from basic hot blast cupolas was es-
pecially high on account of the large
proportion of small-sized steel scrap
material in the burden. If one calculates
the corresponding dust emission per ton
of iron from the top gas dust contents,
there appears to be an increase with in-
creasing coke rate.... Because of uncer-
tainties in the calculation of top gas
and waste gas quantities per ton of iron
in each case, many of these values are not
very reliable.2
It can be established chat one of the factors affect-
ing collector efficiency is the raw gas dust content. Further-
more, it has been shown that data are incomplete on the effect
of melting operations on dust content, except for the general-
ization that dust content is influenced. Therefore, ic can be
concluded that melting operations affect collector efficiency,
but the exact degree of influence is unknown at this time.
2. Gas Flow Rate - Gas flow rate variations result
in velocity changes and thereby influence control equipment
efficiency and pressure drop. The influence on efficiency
varies with the control system being considered.
With dry centrifugal cyclones, the degree of particle
collection is proportional to the amount of centrifugal force
exerted on dust particles in a waste gas stream. In conical
portions of a cyclone, centrifugal force, and thereby collection
efficiency, is greatest at high gas velocity and with a small
cyclone diameter. However, there are limitations to this appli-
cation. In a cyclone operation, there is an inlet velocity
beyond which turbulence increases more rapidly than particle
migration to the wall. When such conditions are reached, with
A T KFMJVFY « ^OVP^S-Y l^r
VTT -
further increases in flow rate, collection efficiency actually
16
decreases.
Pertaining to wet collectors, the following statement
indicates the effect of gas flow on efficiency.
The faster the particle is traveling,
the less likely it is to avoid colli-
sion. Therefore, the efficiency in
scrubbers is a function of the velocity
in the collection area, which in turn
is related to the overall pressure drop.
So, high energy scrubbers in general are
more efficient than low energy scrubbers.1-
• *
In the operation of fabric filters, Che velocicy at
which the gas passes through the fabric must be low. "This is
because higher velocities lead to compaction of the dust cake,
which results in high pressure drop, or worse, in breakdown of
the filter bed, allowing coarse parcicles to pass. Therefore,
to enable normal bag filters to operate at high efficiency, it
is necessary to provide large areas for filtration and avoid
frequent cleaning. The reverse jet filter overcomes this
limitation to a great extent by removing deposited dust by a
reverse current of air. Thus, the buildup of dust is avoided
14
which eases the limitation on operating velocity.
Fabric filters are adaptable to gas rate variations
because a barrier for particulate removal is developed. How-
ever, these systems are subject to pressure drop variations
and generally speaking, the cleaning system will not deliver at
a constant rate when the pressure drop increases.
Gas flow rates affect electrostatic precipitator
performance in that the higher the migration velocity of the gas,
T >• r < n v r v
r- n M p v vv ' • r
-------
VII - 45
the larger the Increase in collection efficiency. Further
factors must be considered in this regard, though, including:
Surface area.
Particle composition and chemical analysis
of dust.
- Characteristic curve of current and voltage.
Flashover frequency.
- Velocity and pressure in precipitator.
Gas composition, temperature and moisture.
Dust resistance.
Finally, optimum gas flow velocities have not generally
been established with certainty for any of the control processes
because they are greatly influenced by properties of dust and
carrier gas as well as by the design of the collection equipment.
3. Carrier Gas Temperature - Gas temperature has a
principal influence on the volume of carrier gas. This, in turn,
influences the size and cost of the collector and the concentra-
tion of the contaminant per unit of volume. Furthermore:
In electrostatic precipitation, both dust
resistivity and the dielectric strength
of the gas are temperature dependent.
Wet process cannot be used at temperatures
where the liquid would either freeze, boil,
or evaporate too rapidly. Filter media can
be used only in the temperature range within
which they are stable. The structure must be
of materials that retain their integrity at
their operating temperatures.
Last, low temperature gases from a stack
following control equipment disperse in the
atmosphere less effectively than high tempera-
ture gases. Consequently, benefits derived
from partial cleaning accompanied by cooling
17
A T KEARNEY & COMPANY I
VTI - 46
may be offset if the cool stack gas cannot be
well dispersed. This is a factor of impor-
tance in wet cleaning processes for hot gases,
where the advantage gained by cleaning is
sometimes offset near the plant by downwash
from the stack because the stack gas is
cooled. 17
4. Carrier Gas Pressure - Pressure of the carrier
gas is not of prime importance in particuLate collection ex-
cept for its effect on gas density, viscosity, and electrical
properties. It could be a factor, however, in deciding whether
a high energy wet scrubber or another type of control system is
best suited for a particular application. An available source
pressure can be used to overcome the high pressure drop across
the scrubber, reducing the high power requirements that could
limit the use of scrubbers.
5. Carrier Gas Viscosity - This gas property has an
effect on efficiency as shown below:
Viscosity is of importance to collection
techniques in two respects. First, it
is important to the removal mechanisms
in many situations (inercial collection,
gravity collection, and electrostatic
precipitation). Particulate removal
techniques often involve migration of the
particles through the gas stream under
the influence of some removal force. Ease
of migration decreases with increase in viscosity
of the gas stream. Second, viscosity in-
fluences the pressure drop across the col-
lector and thereby becomes a power con-
sideration.17
6. Carrier Gas Humidity - Gas humidity affects the
performance of control systems as evidenced in these statements:
High humidity may contribute to accumula-
tions of solids and lead to the caking and
blocking of inertial collectors as well as
-------
VII - 47
caking on a filter medium. It can also re-
sult in cold spoc condensation and aggrava-
tion of corrosion problems. In addition,
the presence of water vapor may influence
the basic mechanism of removal in electro-
static precipitation and greatly influence
resistivity. In catalytic combustion, it
may be an important consideration in the
heat balance that must be maintained. Even
in filtration, ic may influence agglomeration
and produce subtle effects.!'
7. Electrical Properties of Carrier Gas - This is a
factor regarding electrostatic precipicators because the rate
or ease of ionization influences particle removal. The in-
tensity of Brovnian motion, which states that particles in motion
in the neighborhood of the collecting droplets diffuse to them
and are intercepted, and gas viscosity, both increase with gas
temperature. These factors are important characteristics that
relate to the sonic properties of the gas stream. Increase in
either property will tend to increase the effectiveness with
which sonic energy can be used to produce particle agglomeration.
8. Dust Composition - An analysis of cupola dust
composition from different foundries is tabulated in Exhibit
VII-2A. As indicated by these results, dust composition varies
quite widely. This is important with respect to collector
efficiency since dust constituents with higher specific weights
will sink down easily in air as a result of mass forces. Also,
the carbon content of dust makes wetting more difficult and thus
2
puts higher requirements on wet scrubbers.
9. Dust Loading - Dust loading will influence differ-
ent types of collectors in different ways. For example, cyclone
efficiency will increase at higher dust loadings, but extremely
17
VII - 48
high loading may overtax the rapping and shaking mechanism of
fabric filters ?nd electrostatic precipitators. Furthermore,
processes such as sonic agglomeration are quite sensitive to
changes in dust loading.^
10. Dust Solubility - This factor is important re-
garding absorption and scrubbing. In absorption, which relates
to uet collectors, the degree of solubility is an indication
of the ease of removal of the contaminant. In scrubbing, solu-
bility will provide a secondary removal mechanism to assist
the basic separating forces.
11. Dust Sorbabilitv - Sorbability, or the ease with
which a contaminant can be removed in wet collectors by absorp-
tion, is a function of temperature, pressure of the system,
chemical composition of the dust and sorbent and also solubility.
The effect on efficiency varies depending on the interaction of
these factors.
12. Electrical Properties of Dust - Electrical and
sonic properties are an influence in collector performance as
the following evidence indicates:
Electrical properties are considered to be
a contributing factor influencing the build-
up of solids in inertial collectors. In
electrostatic precipitators, the electrical
properties of the contaminant are of paramount
importance in determining collection efficiency
and influence the ease with which it is removed
by periodic cleaning. In fabric filtration,
electrostatic phenomena may have direct and ob-
servable influence upon the process of cake
formation and the subsequent ease of cake removal.
In spray towers or other forms of scrubbers in
A T KEARVEV & COMPANY
-------
VII - 49
which liquid droplets ace formed and contact be-
tween these droplets and contaminant particles is
required for particle collection, the electrical
charge on both particles and droplets is an im-
portant process variable. The process is most
efficient when the charges on the droplet attract
rather than repel those on the particle. Sonic
properties are significant where sonic agglomera-
tion is employed."
13. Dust Toxicity - The degree of dust toxicity
will influence collector efficiency requirements and may neces-
sitate the use of equipment that provides exceptionally high
efficiency. However, the removal mechanism of the collector
is not affected hy toxicity.
14. Particle Size. Size Distribution. Shape and
Density - The size, shape and density of particles are the pro-
perties that determine the magnitude of forces resisting move-
ment of a particle through a fluid. These forces are a major
factor in determining the effectiveness of removal by collection
systems. Furthermore, these forces are balanced against some
removal force applied in the control device and the magnitude
of the net force for removing the particle will determine the
effectiveness of the collection equipment.
Size, shape and density of particles are related to
terminal settling velocity which is significant in selecting
control equipment. This relationship is expressed by Stokes
law which equates the velocity at which particles fall at a
constant speed because of a balance of the frictional drag
force and the downward force of gravity to the properties of
the particle and the viscosity of the fluid through which it
A T KFM»"EY 8. rOMI'*VY Isr
VII - 50
settling. The equation for Stokes law is:
Vt - Ppgdp2/18uf
where: Vt = terminal settling velocity in centimeters/
seconds
Pp - particle density in grams/cubic centimeter
g = acceleration due to gravity in centimeters/
seconds 2
dp » particle diameter in centimeters
uf = fluid viscosity, poise in grams/centimeter
second
This equation is for particle sizes less than 100
microns.
Particle size and size distribution are especially
significant with regard to collection efficiency as shown below;
Any dust collector, operating under specific
flow conditions and with a given gas, will
have a collection efficiency corresponding
with each particle size or particle size- ,g
distribution passing through the collector.
Particle size distribution of dusts found in air pollution work
satisfies the normal standard distribution curve.
Additional conclusions on the effects of particle size
on efficiency appear relevant.
The effect of particle size on efficiency
of a collector can be demonstrated by col-
lecting dust samples on the inlet and out-
let of the operating collector. Knowing
the Inlet and outlet dust concentration
and the inlet and outlet size distributions,
collection efficiency can be plotted as a
function of particle size on arithmetic
coordinates. The resulting plot, known as
a. "size-efficiency" curve, describes the
performance capabilities of the collector
in question when operating at stated gas
flow conditions (velocity, viscosity, etc.).
A T K F * P V I? •>
-------
VII - 51
The collector would have a different "size-
efficiency" curve for each new gas flow
condition, but the curve would remain un-
changed for different dust concentrations
and different size distributions. Thus,
for any particle size passing through a
given collector, a definite and predictable
- collection efficiency will be observed, flow
conditions remaining constant. "Size-
efficiency" curves are especially useful in
predicting overall collection efficiency for
a specific dust having known size-distribution
characteristics.17
Typical examples of how particular gas cleaning
equipment efficiency is affected by particle size and particle
size distribution is presented in the following discussion.
Exhibit VII-25 gives the typical grade efficiency
curves for a high efficiency cyclone and electrostatic precipi-
tator. The density of dust that the curves are based on is 2.7
grams per cubic centimeter. Fractional efficiencies for various
micron size ranges were then read from these curves and applied
to a standard test dust shown in Exhibit VII-22. These results
are shown in Exhibit VII-26.
As shown in Exhibit VII-26, the percent of dust dis-
tribution in each micron size range times the percent efficiency
for that range gives an overall collection efficiency per range.
The sum of these individual range efficiencies equals the over-
all efficiency for the collector. In this particular test, the
cyclone collector was followed by an electrostatic precipitator
in one total installation rather than each system tested in-
dividually.
Particle size has an interesting effect on efficiency
of fabric filters, as Exhibit VII-27 indicates. Theoretically,
A T KF^RVFY ft r n M P \ VY 1 * <•
VII - 52
efficiency is approximately 100% at zero micron size but dips
down to 10% for 0.9-micron particles and then increases back to
1007. again at 1.6 microns. The reason for the dip is that
impingement and diffusion effectiveness is lowest at approxi-
14
mately the 0.9-micron size.
Another interesting effect that particle size (and the
length of time a fabric is in use) has on fabric filter per-
18
formance is also shown in Exhibit VII-27.
When new fabric has been placed in service, efficiency
will be low until dust cake builds up. After cleaning, once
the fabric is loaded, there is still sufficient dust adhering
to the fabric to maintain the dust mat needed for efficient
collection and the collector will perform in the normal operat-
ing range.
The relationship of particle size, pressure drop and
efficiency of venturi scrubbers is shown in Exhibit VII-2B.
As shown, pressure drop has a significant effect on
efficiency with respect to all particle sizes
Combustion
Devices
In addition to basic emission control equipment systems,
combustion devices can be used to control foundry pollutants.
(a) Afterburners
This equipment, used on cupolas, can consist of a combustion
chamber, gas burner, burner controls and temperature indicator.
In an afterburner operation, contaminated gases are mixed thor-
oughly with the flames and the burner combustion gases in the
-------
VII - 53
afterburner throat. Next, the gases pass into the main section
of the afterburner where velocity is reduced. Combustion re-
action is Chen completed and the incinerated air contaminants
and combustion gases pass out of the cupola.
It is important to position afterburners properly so that
complete mixing of gases will be achieved. In addition, suf-
ficient oxygen must be available to burn all combustible con-
taminants completely and ignition time has to be provided in the
cupola stack. A distance from the top of the charging door to
the top of the stack of at least 25 feet is usually required for
proper combustion.
Afterburners are used to complete oxidation of combustible
contaminants in cupola melting operations. In this regard, the
burning of carbon monoxide into carbon dioxide reduces the ex-
plosion hazard. Afterburners also reignite gases that may be
extinguished by a cupola charge or charging bucket. The ad-
vantage from a public relations standpoint of a reduced discharge
appearance by using atterburners usually will outweigh the for-
mation of some additional dust particles through the reduction
of coke cinders to ash.
Efficiency of afterburners essentially depends on the de-
gree of mixing of contaminated gases with the flames and burner
combustion gases within the afterburner unit. Also important
in this regard are operating temperature, retention time of
gases within the 'afterburner and concentration and types of
contaminants to be burned.
VII - 54
Afterburners are often used in conjunction with basic
emission control equipment systems. They are especially useful
when the cupola charge contains oils, as the greater portion of
visible smoke comes from unburned oils.
Afterburners have limitations in that cupola exhaust gases
with carbon monoxide in certain concentrations need to be in a
temperature range of approximately 850° to 900° F. This is
important since exhaust gases must be heated to approximately
1200° to 1400° F to initiate oxidation which means that fuel
costs for operating afterburners can be excessive if too much
heating is required.1' 6> l9 Exhibit VII-29 illustrates a typi-
cal afterburner unit.
(b) Catalytic
Afterburners
This process, used on bake ovens, consists of an after-
burner housing containing a preheating section, if required, and
a catalyst section. Frequently, contaminated gases are delivered
to the catalytic afterburner by the fan exhausting the oven.
The gases then pass into the preheat zone where they are heated
to the temperature required to support catalytic combustion.
Here, some burning of contaminants usually occurs. The preheated
gases then pass through the catalyst bed where the remaining
combustible contaminants are burned.
The purpose of the afterburner is to complete the oxi-
dation of combustible gases to an odor- and color-free condition
before exhausting them to the atmosphere. These gases will
n, T I r • P V 1- V
-------
VII - 55
burn at a temperature of 650 -1000° F, in Che presence of a
suitable oxidizing catalyst, such as platinum. A significant
savings in fuel cost can be realized in the preheating of com-
bustible gases to this lower temperature instead of the
1500°-2000° F required without a catalyst.
Efficiency of catalytic combustion units depends on several
factors. Contact of gases with the catalyst, uniform flow of
gases through the catalyst bed, temperature, catalyst surface
area, nature of materials being burned and oxygen concentration
are important in this regard.
A limitation of catalytic units is that, because of some
materials of construction, these devices cannot be operated at
1 19
temperatures necessary to produce complete combustion. '
Exhibit VII-29 depicts a catalytic combustion device venting a
core baking oven.
APPLICATION OF EMISSION
CONTROL EQUIPMENT SYSTEMS
TO FOUNDRY PROCESSES
Although the majority of gray iron foundries currently do
not have collection equipment systems for control of emissions
from melting operations, the number of installations of these
systems has increased rapidly the last few years. The 1967
survey conducted by the Business and Defense Services Adminis-
tration and the National Air Pollution Control Administration
indicated that 204 of 1,376 foundries had some type of emission
control system for melting processes. Current estimates appear
VII - 56
20
to indicate this figure has approximately doubled. A
tabulation of different emission collection equipment designs
and their particular application to foundry processes is shown
in Exhibit VII-30.
In addition to the many dust collection equipment systems
in use, a variety of hoods, ventilating and exhaust systems
and various other techniques are employed to capture or exhaust
foundry emissions.
For purposes of discussion, the iron foundry has been
divided into areas of activity in which characteristic emission
control problems occur. These are as follows:
- Raw material handling, preparation and charge makeup.
- Cupola melting.
- Electric arc melting.
- Electric induction melting.
- Other types of melting.
- Inoculation.
- Mold pouring, cooling and shakeout.
- Sand preparation and handling.
- Cleaning and finishing.
- Coremaking.
- Miscellaneous areas.
Raw Material Handling,
Preparation and Charge
Makeup
The charge yard and charge makeup areas in most iron
-------
VII - 57
foundries have not normally been considered important sources
of air pollution and therefore, with rare exceptions, have not
had air pollution control systems installed. This has occurred
for two reasons. Most of the emissions which come from these
areas are dusts of relatively large particle sizes which settle
readily. Second, few "ixed emission points exist in typical
yards where control can easily be applied. In a few cases,
ventilation systems and dry centrifugal collectors have been
installed in the charge makeup area, where it is located inside
an enclosed building.
The one area which has been receiving attention in recent
years Involves those foundries in which metallic charge materials
are either burned to remove nontnentallic coatings or accompany-
ing nonmetallic debris, or are preheated to remove moisture or
oily coatings. Since these operations are almost always per-
formed in a fixed combustion unit of some type, emission
control systems are relatively easy to apply. Medium energy wet
collectors have been used where oil fume and vapor were present,
and dry centrifugal collectors were applied where dry dusts
existed.
Cupola Melting
Because of the widespread use of cupolas for melting, the
severity and complexity of the cupola emissions problem, the
difficulty of determining the nature and amount of emissions,
and the generally high collection equipment costs, more attention
has been devoted to emission control of this foundry process
VTI - 58
than any other. The presence of these factors as well as other
variables such as the wide range of cupola operating character-
istics has resulted in more control methods and techniques being
employed on cupolas than on any other foundry process. In fact
every known method, from simple spark screens to complicated
systems such as electrostatic precipitators , has been tried
with varying degrees of success. Although selection of clean-
ing equipment varies with the purpose of the installation,
recent attention has centered on those techniques which have
been most successful in high efficiency control of emissions,
such as high energy wet scrubbers and fabric filter baghouses.
The problem of selecting gas cleaning equipment for cupolas
depends essentially on the degree of efficiency required, need
to meet existing pollution codes, and the economic factors of
capital and operating costs.
Vet caps, dry centrifugal collectors, wet collectors,
fabric filters and electrostatic precipitators are the different
collection systems which have been used for cupola emission
control.
In the following paragraphs, the principal types of equip-
ment in use on cupolas are briefly described.
(a) Wet Caps
These collectors, used more frequently on cupolas than any
other type, are placed directly on top of cupola stacks and thus
do not require any gas-conducting pipes and pressure-increasing
-------
VII - 59
blowers. Wet caps are most practical in plants having an
existing supply of low cost water and the ability to dispose
4
of collected dust in sludge form. Furthermore, some type of
wet cap system is often employed in conjunction with high energy
wet collector installations on cupolas.
Approximately 95 gray iron foundries were reported to have
20
cupola wet caps in 1967.
Once the most popular type of cupola emission control, the
low efficiency of the wet cap has caused it to decline in use
in recent years. Progress is now being made in developing
higher efficiency wet caps with multiple spray sections.
(b) Dry Centrifugal
Collectors
These low energy units are designed for light to moderate
particulate concentrations. In a cupola installation, ductwork
and an exhaust fan to draw gases to the collector are required.
These systems also necessitate capping the cupola and install-
ing a cooling spray to reduce the temperature of exhaust gases
flowing to the collector. Often, dry centrifugal units are used
as precleaners of hot blast cupola top gases prior to feeding
21
into a recuperator. Furthermore, this type of collector is
an integral part of most high efficiency emission collection
systems. In 1967, approximately 15 gray iron foundries re-
portedly had dry centrifugal installations which were not part
of a larger cupola emission collection system. The low effi-
ciency of the dry centrifugal collector has resulted in almost
VII - 60
no new installations on cupolas in recent years, unless they
were part of a larger system.
(c) Wet Collectors
Several different medium and high energy designs have
been applied on cupolas. A wide range of capacities and
collection efficiences is available. These systems are usually
used where moisture and/or high temperature are present in the
emission. A complete installation requires ductwork, an exhaust
fan and capping of the cupola. As with wet caps, these systems
are most practical where low cost water and sludge disposal
4, 21
equipment are available. Although only 30 gray iron found-
ries were reported to have cupola wet collectors in 1967, recent
trends indicate that installations of this type of system are
increasing more rapidly than any other.
(d) Fabric Filters
When cupola collection efficiencies of 99% or higher are
required, fabric filters are often selected. Although various
fabric materials are available, glass fabric is typically chosen
because of its resistivity to high temperatures. Complete
installations can include numerous components such as a baffle,
raised cupola stack and lid, ductwork, exhaust fan, spray
coolers and other items in addition to the fabric filter unit.
Another type of installation involves using heat exchangers
instead of spray coolers. Fabric filter units can be installed
21
to handle more than one cupola if desired. Approximately 39
-------
VII - 61
gray iron foundries were reported to be equipped with fabric
20
filters on cupolas in 1967.
(e) Electrostatic
Precipitators
Rare applications of these systems have been made on
cupolas. Excessive co ts, and operating and maintenance problems
have limited the use of these systems. Only one gray iron
foundry was reported to have a cupola electrostatic precipitator
2O
installation in 1967. Additional installations have been made
in the past few years.
(f) Other Techniques
In addition to application of the major classes of collec-
tion equipment, afterburners and preheaters can be used to
control emissions from cupolas.
1. Afterburners - In cupola installations, after-
burners or gas igniters can be employed for burning the com-
bustible top gases, thereby reducing the opacity of particles
discharged from the stack and for eliminating explosion hazards
from cupola gases. Afterburners are located just below or
21
opposite the charging door.
2. Preheaters - Burning of unburned products of
combustion can also be accomplished at times with a type of
blast air preheater which burns exhaust gases from the cupola.
Not only is thermal efficiency of the cupola capable of improve-
ment, but the preheater acts as a settling chamber for collect-
ing coarse dust.
21
A T K r v P «c r v * r n M p» v > ? ^ -
VII - 62
Electric Arc
Melting
Compared with the cupola, electric arc melting has in the
past received a great deal less attention regarding air pollution
control for several reasons. Many of the emission problems for
this process had already been substantially resolved with
existing designs of equipment. Furthermore, electric arc melt-
ing accounts for a relatively small portion of total foundry
melting in the United States, although this percentage is
increasing rapidly. Finally the emissions from electric furnaces
are substantially less than from the cupola.
A number of differences exist between the electric- arc
and cupola air pollution problem which are significant regard-
ing control techniques. First, the electric arc melting process*
and emissions problem are less complex. Second, since the
average particle size of electric arc emissions is considerably
smaller than that of the cupola, different collection objectives
exist. Finally, more uniform electric arc operating condifiioio
and lower total emissions tend to simplify the design of control
equipment for this process.
(a) Emission Collection
_ Equipment _
In 1967, approximately 24 gray iron foundries were reported
to have some type of air pollution control equipment for
electric arc melting processes, but the number of installations
has increased substantially in the last few years.
-------
vir - 63
1. Fabric Filters - Of the major categories of
emission collection equipment, fabric filters are best suited for
electric arc furnaces and have been most frequently applied. This
is due to the extremely fine particle size of dust and fume
emitted from electric arc furnaces. Complete installation of a
fabric filter unit to w*ie furnace includes ductwork, an exhaust
fan to draw gases to the collector and a means of collecting
the gases from the furnace. l Approximately 20 gray iron foundries
reportedly had fabric filter installations on electric arc melt-
ing in 1967.
2. Wet Scrubbers and Electrostatic Precipitators -
These collection systems are rarely used on electric arc furnaces.
Wet scrubber limitations include the existence of too much fine
dust and high energy requirements. Electrostatic precipitators
can encounter exhaust volumes too low for their design require-
ments. Four foundries were reported to have wet scrubber instal-
20
lations in 1967 on electric arc melting processes.
(b) Furnace Hoods
Electric arc furnaces are also equipped with various types
19
of hoods to capture emissions. Arrangement of electrodes and
gear above the furnace top as well as the charging and operating
method largely determines the hood type applied. Often, some
type of hood is used in conjunction with an emissions collection
system.
1. Full Roof Hood - This type of hood is attached to
the top ring of the furnace. It requires stiffening to prevent
sagging at high temperatures and protection of electrodes to
VII - 64
prevent short-circuiting.
2. Side Draft Hood - Another design often employed is
the side hood. This unit is located on the side of the roof
close to the electrodes to produce a lateral type of control.
An overhead hood at the charging door is also often used with
the side hood.
3. Canopy Hood - Canopy hoods are located above the
craneway, and do not interfere with furnace operating procedure.
Effectiveness of these units is limited though, due to equip-
ment requirements needed to handle the large volumes of infiltra-
ted air.21
(c) Other Methods
1. Fourth Hole Ventilation - In this system, a water-
cooled probe is directly connected to the furnace roof. The
probe maintains a carefully controlled draft in the furnace
body.
2. Snorkel - This technique is similar to the fourth
hole ventilation method except that the extra hole serves as a
natural pressure relief opening for the furnace.
Electric Induction
Melting
Very little attention has been directed toward control of
emissions from electric induction melting furnaces. Since no
combustion and only limited metal oxidation occur in this type
of furnace and because relatively clean scrap is used for charge
-------
VII - 65
material, no serious emissions problem exists for induction
melting of iron. Induction furnaces normally operate with a heel
of liquid metal, and for this reason, metal scrap must be dry and
free of oil. Gas fueled dryers and preheaters have been used to
achieve this condition, often with operating cost benefits
resulting from the use of the lower cost fuel to preheat the
charge. The burning of oil residue on the scrap produces ob-
jectionable emissions requiring the use of emission control equip-
ment. Afterburners and wet scrubbers, either separately or in
combination, are often used to reduce these emissions to accept-
able levels. With oil and other combustibles removed from the
furnace charge, emission control equipment is not required on
induction melting furnaces to satisfy current codes.
Other Types
of Melting
The other principal type of furnace used for melting of
duplexing of iron is the stationary or rotating, fuel fired,
reverberatory furnace. The stationary furnace, often referred
to as the *ir furnace, was once universally used for duplexing
when producing malleable iron. Its use has declined with the
increase in production of ductile iron, but it is still found in
malleable iron foundries. The emissions from this type of furnace
come principally from the combustion of coal oil or gas fuel,
plus some slag and iron oxide which is carried up the stack with
the products of combustion. The older installations all exhausted
into the atmosphere through a stack or chimney. Medium energy
wet scrubbers and fabric filter bag collectors have been applied
in a few cases.
VII - 66
The rotary reverberatory furnace has been only recently
utilized in small installations in iron foundries. Since a
small quantity of emission in the form of waste products of com-
bustion and slag particles is given off, these installations are
not equipped with collectors.
Inoculation
(a) Ductile Iron
Production
Original installations for producing ductile iron by mag-
nesium treatment either exhausted directly into the foundry build-
ing, or were equipped with a ventilation hood which then exhausted
into the atmosphere. In recent years, these stations have been
equipped with collecting hoods, or have been installed in
enclosed rooms, and the resultant gas is drawn off by means of
an exhaust fan, into a dust collection unit. Medium energy wet
scrubbers and fabric filter baghouses have been used for emission
collection in this area.
(b) Mechanical Property
Improvement by
Inoculation
Since inoculation in this area produces no major emissions
problem, emission control techniques have not been applied.
Mold Pouring, Cooling
and Shakeout
Control of emissions resulting from pouring and cooling of
molds has been common for several decades for those high volume
production foundry installations where finished molds are set out
-------
VII - 67
on continuous car-type mold conveyors, providing fixed locations
for pouring, cooling, and shakeout operations. With this type of
equipment, side draft hoods are often provided for the pouring
area and side or bottom draft hoods at the shakeout, with the mold
cooling conveyor between these two points fully hooded with sheet
metal. Ducting is commi nly provided from each area to a single
control device, usually a wet scrubber or dry centrifugal
collector.
Collection systems have been, and still are, uncommon for
those smaller production and jobbing foundries where completed
molds are set out on the foundry floor or on gravity roller
conveyors, and where the pouring and cooling utilize a sub-
stantial percentage of the molding floor. The problem for this
type of operation is related more to the cost of capture of the
emissions with a minimum amount of infiltrated air, than to
separation of the emission. With pouring and cooling in non-
fixed locations, without hoods to capture emissions, a large part
of the air in the foundry would require handling. The cost
would be prohibitive due to the volume to be processed. A
partial solution to the problem in the pouring area of non-
ferrous foundries has been provided by a traveling vent attached
to the pouring ladle bail, and ducted by means of flexible tubing
and specially designed connecting ducts to a suitable emission
control unit. This technique permits capture of emissions result-
ing from mold pouring with a minimum of infiltrated air. It is
possible that similar methods could be used in iron foundries.
VII - 68
Additional venting is required during subsequent cooling, how-
ever, and this is not practical when the ladle is moved on to
pour the next mold, with the result that significant emissions
are still not collected.
Large casting?. such as automotive dies and machine beds,
can be cast by the full mold process. Generally, no central
pouring station is provided and the smoke generated is released
directly into the foundry building, creating an in-plant
problem.
The current method for controlling the smoke is through
the use of ventilating fans. A properly designed arrangement
of fans and makeup air systems may produce a relatively clear
shop environment, but as the emissions are exhausted from the
foundry, an air pollution problem is created. The problem is
further complicated by the fact that ventilating fans exhaust
large volumes of air and are not designed to be connected to a
duct and collector arrangement.
Sand Preparation
and Handling
Processes such as mechanical sand handling systems and sand
mixing or reconditioning produce emissions that have received
attention over the years. Since most gray iron foundries make
green sand molds that produce moisture when the mold is poured,
medium energy wet collectors are best suited for effluent
control. Occasionally, cotton or wool fabric filters are
employed only when dry sand conditions exist. Also, medium
-------
VII - 69
energy dry centrifugal collectors are sometimes used. Often,
some type of hood is used to capture emissions in sand con-
veyor systems especially at transfer points. As with many
other processes, ductwork and an exhaust fan are required in a
complete collection system.
Cleaning and
Finishing
The various cleaning and finishing operations produce a
significant emissions problem that has received considerable
attention regarding control.
Dusts from gate and riser removal are generally controlled
with local exhaust systems connected to dry centrifugal collec-
tors, medium energy uet collectors, or possibly fabric filters.
Sometimes exhaust hoods are provided above the work station.
Other cleaning processes such as abrasive shotblasting and
tumbling are commonly controlled with fabric filters or medium
energy wet collectors. Applications of dry centrifugal
collectors are also made for abrasive cleaning processes.
Most of the trimming and finishing operations generate
emissions and require control. Chipping and grinding operations
are normally provided with local exhaust hoods connected to
either high efficiency centrifugals or fabric filters. Wet
collectors are used if central sluicing systems are employed
or where grinding exhaust is combined with other cleaning room
operations.
VII - 70
Surface painting requires ventilation to reduce the hazard
due to volatile materials being atomized in the air. Exhaust
systems are generally used where dip painting is performed.
Open tank installations are also provided with local ventilation
hoods.
Heat treating furnaces for malleable!zing or for other
treatments of iron castings present the usual problem of
emissions from combustion of liquid or solid fuels except where
electric or gas fired radiant tube heating is concerned. In
most foundries, these are exhausted into the foundry building,
or through a stack to the atmosphere. Medium energy wet
scrubbers are an effective means of cleaning coal burning
exhaust gases, but have not been applied in many cases.
Coremaking
Efforts have been made to control certain coremaking
emissions, but the gases emitted from bake ovens and shell
core machines are a serious problem difficult to control.
Usually these gases are permitted to exhaust to the atmosphere
through a ventilation system. Sometimes, catalytic combustion
devices are used on core ovens to burn gases to noncotnbustible
analysis.
Other coremaking processes present a less serious air
pollution problem capable of control. In coreblowing or core-
shooting, fabric filters are usually selected if control equip-
ment is desired. In rare instances, medium energy wet collectors
-------
VII - 71
are used. For core grinding, fabric filters and medium energy
dry mechanical collectors and wet scrubbers are frequently
selected.
Miscellaneous
Areas
Finally, emission control equipment is applied in certain
non-manufacturing areas in foundries including pattern shops and
crating or boxing for shipping, where woodworking operations
occur. Dry centrifugal collectors are commonly used to collect
wood dust and chips from these operations. Machine shops and
metal pattern shops usually present problems of collecting
dust from machining or grinding of cast iron. Dry centrifugal
collectors are commonly used for this purpose.
CASE HISTORIES OF
ATTEMPTS TO CONTROL
EMISSIONS IN IRON
' FOUNDRIES _
Although there were individual attempts to eliminate or
control foundry emissions before World War II, the serious
efforts at emission control started about 1945. Initial efforts
concentrated on removal of undesirable dust and fume from
Inside the foundry buildings and from work a teas where large
of emissions were -generated. Tbe principal means of
accomplishing this was to simply ventilate the buildings and to
eXhaust the emissions to £he atmosphere. As Che problem
in severity, many foundries begarv -to -appiy existing
of dust colLecttDH-«^ai-pment to individual emissions
.producers _in the foundry. Gradually, the designs of dust
VJI -_Z2_
collectors began to be directed toward the individual foundry
requirements, and the present types of equipment began to
evolve.
The history of the development of .foundry emission control
has been traced fr -m actual case histories which were obtained
from .reports "f p.iM'if ig°n"1°r, -\»t? -from equipment manu-
facturers, published articles, and private sources.
Because of the differences in the historical development
of emission control equipment for various foundry processes
over the years, .tne-dlsciission is diuided into t-we sections:
(a) melting processes and (b) non-melting processes.
Melting Processes
Prior to the 1940's, emission control techniques for the
cupola, which was the principal melting practice employed at
the time, essentially consisted of screen cage spark and dust
arresters or chain barriers. The screen cages, placed on top
of cupolas, were simple and inexpensive designs, primarily
used for reducing the emission of cinders that caused damage to
fraiynliy rojofs, g"«•«•»«•«= .aaJ rifTMnTr""""" -asd uhioh were also a
nuisance to the immediate .area. Some cinders were removed,
but since screen cages were coarse to prevent excessive resis-
tance to gas flow, this technique did little in catching the
-also emitted. r)**i*r,c suspended from weather caps
located near the top.of the cupola were even, less -effective.
21, 23
-------
VII - 73
One of Che first emission control equipment systems
employed on cupolas and the most widely used in the 1940's
was a wet cap. These collectors came into prominence because
they were simple, economical systems and a major improvement
over screen cages and chains in reducing cupola emissions that
led to roof damage and neighborhood nuisance problems.
Success in controlling cupola emissions with early wet
cap designs varied widely. Some foundries fabricated their
own wet caps while others bought commercially available units.
Some foundries ex^tilanced increased cleaning efficiency with
larger amounts of water. Others were less successful and had
excessive corrosion and prohibitive maintenance costs when water
usage was increased in efforts to obtain higher cleaning
efficiencies. Furthermore, a successful wet cap design for one
cupola installation did not always produce similar results on
another cupola having the same outward size and characteristics,
due to Che many variables in cupola operating condicions and the
21
wide range of solids emitted.
Although the first reported installation in the United
24
States occurred in 1936 on a midwestem automotive foundry,
wide use of wet caps actually appear to be an outgrowth of black-
out efforts in England during the early part of World War II.
Another reported installation occurred in 1941, but shutdown of
the foundry because of the war prevented testing the wet cap
system. Many of the initial wet cap installations which
occurred in the early 1940's consisted of a shell and one to
three spray nozzles located near the top of the shell.
VII - 74
The first wet cap installation where effectiveness of the
gas cleaning system was reported occurred in L945. Six cupolas
at a midwestem automotive foundry were each equipped with a
double-cone wet cap system consisting of 70 water spray nozzles.
Although no quantitative test data were revealed, the instal-
lation was considered successful from the standpoint of reduced
emissions deposited on adjacent roofs and grounds.^*
Because of the success of the automotive foundry wet cap
installation in 1945, the exact same wet cap design was employed
on six new cupolas at a gray iron foundry located in the Mid-
west in 1947. On the first day of operation, the blast on one
cupola was turned on before the pump supplying water to the
collector could be started. This caused the cone of the scrubber
to become thoroughly heated, and when water entered between the
cones, the pressure pulled one of the stays out of the lower cone,
forcing a stream of water down into the cupola stack. Water to
the collector was not cut off for fear of damage, so the bottom
on the cupola was dropped and the operation discontinued.
Due to this unfortunate experience, the hollow cone multi-
spray wet cap was discarded. The difficulties with the multi-
sprays were attributed mainly to dirt, pipe scale and boiler
scale in the mill water. Maintenance cost was also considered
prohibitive.
Later, the original wet cap design was remodeled by
removing the upper cone and multiple nozzle arrangement and
-------
VII - 75
installing a single large nozzle over the apex of the remaining
lower cone. Tests made on the cupolas with the remodeled wet
caps showed that .09 to .13 grains per cubic foot of particulates
were collected. The resulting emission levels were lower than
that experienced by many other foundries with wet cap installa-
tions.
An analysis of particle sizes was made as a further check
on the performance of these wet caps, and it was found that
about 1% by weight of the escaping particulate matter exceeded
40 microns in diameter. Of the remaining 99% by weight, over
99% of the particles, by count, were less than 5 microns.
24
Wet cap designs introduced in the mid-1940's, such as those
consisting of a steel cone, nozzles discharging water in a
curtain and a deflector ring, were an improvement over the shell
design and are basically the same as current concepts.
In the late 1930's, attempts were made to design other
wet-type cupola gas cleaning systems to prevent disgruntled
neighbors from bringing suits for property damage. One such
attempt occurred at a New England gray iron foundry in 1937.
Gases from two capped cupolas were drawn downward through duct-
work by a fan to the collector. The heavy particles sank
downward by gravity and some of the lighter particles were
removed when passed through a water spray. It was reported that
125 pounds of particulates per hour, consisting of coke particles,
iron oxide particles, and silicious material, were removed from
26
the gas stream.
VII - 76
During the late 1930's and 1940's, various dry centrifugal
collector designs were also used on cupolas but less frequently
than wet caps. Some attention was directed in this area, how-
ever, because of the advantage of gravity and centrifugal forces
in these systems. Particles too heavy to be supported by the
upward velocity of the waste gas settled back into the stack,
or into expansion-type dry collection chambers. Fine dust
collection was poor, and overall collection efficiencies of
these dry centrifugal systems were in the range experienced with
wet caps.
Dry centrifugal systems used during this period included
cyclones, impact chambers, zigzag baffle arrangements and down-
up expansion chambers. Problems with high temperature, corrosion
and handling the collected dust prevented more widespread use
of dry mechanical or centrifugal collectors on cupolas at that
time.
Dry centrifugal cleaning principles proved to be more
effective when combined with a wet cap system. A combination dry
centrifugal, wet cap system installed on eight cupolas in a
midwestern automotive foundry in 1948 was reported to be approxi-
mately 90% efficient. This system type consisted of a flame
suppressor that introduced an atomized mixture of water and air
into the rising gases. Then a combination cyclone-type blower
unit, conical cap and water sprays further wetted dust particles
27
and washed them down into discharge pipes.
-------
VII - 77
VII - 78
During the early 1940's, systems to control electric arc
furnace emissions made use of building roof monitors, roof fans
and exhaust hoods. These techniques helped reduce fumes in the
plant by exhausting them to Che atmosphere but no dust collection
was achieved.
In the mid-1940"s, foundries began to apply emission control
equipment to electric arc melting processes as the following
case describes. A roof-type ventilating system was installed
above an electric arc furnace in a midwestem steel foundry in
the 1940's. During warm weather, the system worked adequately,
but gases and dust collected in the building during cold weather
and made working conditions unpleasant. In an attempt to over-
come this problem, foundry management Installed a wet centrifugal
unit with a cyclone precleaner in 1945. In 1950, oxygen began
to be used in the melting practice and this increased contamina-
tion during injection. When the local smoke code could not be
met due to this contamination, a battery of spray nozzles was
added to the precleaner. This modification reduced dust in the
airstream, but wind carried the contaminated water over nearby
homes, causing a nuisance. A water eliminator was then added
28
to the system and it was reported that complaints stopped.
During the middle and late 1940's, further improvements in
cupola emission control involved use of afterburners in the
cupola stack to reduce the visible black plume from unburaed
volatiles, largely oil from scrap. Also, gas or electric ignit-
ers were frequently substituted for traditional wood fire for bed
coke ignition. The burning wood often produced an objection-
able pungent odor and smoke from partial combustion of the
residuals.
Until the late 1940's, very few foundries had any effective
type of dust collection equipment system for control of melting
emissions. The systems that had been installed, mostly wet
caps, were largely due to aroused public opinion of residents
located near foundries. In the late 1940's though, the dramatic
air pollution difficulties at Donora, Pennsylvania, the adoption
of the Los Angeles Air Pollution Control District Code, increased
neighborhood complaints, publication of the American Society of
Mechanical Engineers smoke regulation and ordinance model, and
other factors prompted development of additional emission control
equipment and refinement of existing techniques. This was the
first major concerted effort to develop high efficiency emissions
control equipment.
Efforts to meet the strict Los Angeles County code with
wet caps and other designs previously employed during the 1940's
proved unsuccessful. The foundry industry and control equip-
ment manufacturers were not aware that cupola emissions contained
a large percentage of particles too small to be effectively
collected by the equipment in current use. After many tests, it
was determined that the fabric filter was best suited to meet
the code provisions for foundry emissions. Fabric filters had
been used successfully for many years in other industries, but
this was the first application to foundry melting processes.
-------
VII - 79
The dramatic increase in collection efficiency that could
be experienced with fabric filters is demonstrated by the report
on a cupola installation in Los Angeles in 1952. Close to 1007
of all dust was collected with a fabric filter consisting of
crlon bags.
However, only rare applications were made subseqently of
fabric filters with orIon bags due to temperature limitations.
Since orlon usually could not be used as a filter material when
temperatures exceeded 275 F and because cupola -gas temperatures
are much higher, attention focused on other filter material that
could withstand higher temperatures, such as fiber glass fabric.
Pressure on the foundry industry by the Los Angeles Air
Pollution District resulted in the development and use of glass
fabric for cupola gas cleaning In the late 1940's and early 1950's.
Typical early installations consisted of relatively crude designs
where 15-foot-long, 11-1/2-inch-diameter tubes were suspended to
a mechanism that was manually operated to "rock" the tubes during
shutdown. Production of glass cloth with a silicons lubricant
permitted this mild flexing action without rupture of the glass
fibers. However, problems with bag wear and the sensitivity
of glass fabric to flexing encouraged development of fabric
collectors of the reverse flow concept in the early 1950's. This
provided more effective dust removal and continuous operation of
the collector. However, visible plumes emitting from the
foundry stack still presented a problem.
VII - SO
Experimental tests on electrostatic precipltators in the
late 1940'8 were successful in the Los Angeles area and the
system was in use in other industries, but attempts to apply
precipitators to cupolas were unsuccessful at the time due to
maintenance and other problems.
Meanwhile, intensified Interest in melting emissions con-
trol also had effects on electric arc furnaces which,were being
installed in increasing numbers during the 1950's. The full
roof hood concept that was used in the raid-1940"s was modified
to that of a local side hood which improved control of electric
arc emissions. Fabric filters also began to be used for higher
collection efficiencies.
Nationally, there waa a reluctance to follow the glass
fabric practice of Los Angeles for cupola emission control,
especially in the colder northern climates. The multiple dry
centrifugal was introduced for such systems in the mid-1950's
and tests conducted in 1956 at a Wisconsin foundry equipped with
this type collector resulted in efficiencies ranging from 59%
to 84%.29
Some attention was directed toward other dry centrifugal
designs also, such as high efficiency cyclones. Tests conducted
on this system type in the early 1950's produced cleaning
efficiencies of 65% to 83%. Generally speaking, less success
was achieved with high efficiency cyclones than multiple dry
centrifugals and this latter type was used on cupolas more
-------
VII - 81
frequently when mechanical systems were selected. Dry centrifugal
collector design concepts in use during the early 1950 's for
cupolas are basically the same as current concepts.
In the late 1940's and early 1950's, considerable attention
continued to focus on wet cleaning systems for emission control.
Even though a number of these systems could not meet the strict
Los Angeles code, continued refinements, such as capping the
cupola to confine the gases and prevent combustion in the stack,
and development of more complex control techniques increased
cleaning capabilities. For example, in 1949, a Los Angeles
foundry was able to meet existing codes with a cleaning system
involving a charging bell, dry centrifugal and wet scrubber.
In other sections of the country, many foundries were con-
tinuing to use wet caps, but performance was generally below
that obtained by foundries meeting the strict Los Angeles code.
A northern foundry, in 1951, installed a wet cap and the major
benefit was the reduction in deterioration of the foundry build-
ing roof. Complaints from housewives stopped also, but some
problems were experienced with the pump because of the acidic
condition in the water system.
In the mid- and late 1950's, continued public demand for
better cleaning efficiencies prompted development of medium
energy wet scrubbers. These systems were more successful than
the lower energy wet caps, but not as effective in dust removal
as fabric filters. An installation of this type on a 96-inch
VII - 82
cupola in the early 1960's produced dust outlet loadings of
.035 grains per standard cubic foot which was well below exist-
ing code levels.
Development of medium energy wet scrubbers also led to
applications on electric arc furnaces where efficiencies of
approximately 75%-857. for solid particles were achieved.
Problems experienced with these systems included Little altering
of discharge appearance because of the large percentage of sub-
micron dust and corrosion from oil scrap.
Even though dust collection efficiency was increasing with
the development of new systems, complaints still persisted for
removal of the visible plumes from cupolas. This encouraged
application of the high energy wet scrubber. Although used in
other industries since the early 1950's, it appears that the
first cupola installations were put in service by an automotive
foundry in the mid-19601 s, using below charge door gas take-off.
At least one installation of that time period condensed the
water vapor ahead of the exhaust fan to provide for a smaller
fan and drive motor and, at the same time, eliminated the dense
steam plume.^
Actual results that could be achieved with this system
type were demonstrated by a high energy venturi scrubber instal-
lation in an Indiana foundry in 1967. Emissions were controlled
to about .1 pounds/1,000 pounds of gas or .05 grains per standard
cubic foot.33
A T k F \ n v r v
-------
VII - 83
In another example, a medium energy flooded bed wet scrubber
tested early in 1969 showed that New Jersey air pollution codes
probably could not be met. The results of 5 tests averaged
274 pounds/hour of emissions whereas code standards allowed 205
pounds/hour. The problem cited was that the cupola was not
y
functioning according to design operating conditions. In
an effort to correct the problem, the collector was upgraded to
a high energy unit in August, 1969. Collector modifications
made were:
- Addition of a high energy fume agglomerating section
in the original scrubber housing.
- Provision for the use of 15 to 20 GIW liquor per
1,000 CFM of gas.
- Installation of a new fan impeller.
Tests made after the modifications Indicated emission
levels of approximately 125 pounds/hour--we11 below the exist-
ing code.
35
In the 1960's, foundries continued installing wet caps be-
cause of their simple, inexpensive design and collection capabili-
ties regarding larger dust particles. Replies to a survey con-
ducted by the American Foundrymen's Society, which was documented
in 1967, indicated that some foundries were satisfied with per-
formance of their wet caps while others were not. Most of these
wet caps were installed to prevent nuisance or damage to the
immediate plant property.1
Some foundries that previously Installed wet caps modified
VII - 84
their cleaning systems when local codes became more strict. For
example, management of a large midwestern production foundry was
faced with a problem of reducing emissions from its 2 60-inch
hot blast cupolas that were already equipped with wet caps.
In 1968, a variable annulus scrubber was added to the system
which, when tested, resulted in an emission level of .12 grains/
standard cubic feet (.23 pounds dust/1,000 pounds gas). The wet
caps were provided with high efficiency non-clogging multiple
flight nozzles and corrosion resistant deflectors to augment
the oeripheral water curtain. The modified collection system
increased previous efficiency levels by 50% in terms of weight
36
of solids collected.
Fabric filters were also often applied on cupolas during"the
1960's. Many of these were on the West Coast where codes were
strict. This system was an improvement over initial designs, and
glass bag material, because of its higher temperature resistiv-
ity, was the usual selection. Reports from two California foun-
dries and one in the Midwest, that installed fabric filters in
the 1960's, indicated collection efficiencies of over 98%. One
of these cases cited high maintenance costs due to bag replace-
ment, but another reported maintenance costs to be low.1'" 37» 3B
Replies to questionnaires distributed by the American
Foundrymen's Society, and documented in 1967, indicated that
foundry management felt fabric filter performance was good. One
case was reported where bags could not be kept in repair because
of seam splitting and the foundry intended to replace the fabric
collector with a wet scrubber.
-------
VII - 85
Renewed interest developed In the 1960's regarding electro-
static precipitacors for cupola emission control. One report
on a California foundry, documented in 1967, indicated collection
efficiencies of 96.6%. Another installation though, on a
large automotive foundry in the Midwest, proved to be unsatis-
factory and was eventually discontinued. High maintenance costs
22
were cited in this latter case study. Until recently, because
of high costs, maintenance problems, and operating problems,
electrostatic precipitators were rarely used to control cupola
emissions. Two precipitators have been Installed in the last
18 months. The most recent installation is reportedly operating
satisfactorily. No maintenance cost data are available because
of their short operating history.
Some attention was directed toward use of a dry multiple
cyclone collector on cupolas in the 1960's. Replies to the
American Foundrymen's Society questionnaire indicated that found-
ries were obtaining efficiencies from 61%-75% with such equip-
ment. These were results from five tests on different installa-
tions .
Current trends appear to indicate that high energy wet
scrubbers and fabric filters are the types of emission control
equipment systems usually selected for cupolas because of their
high efficiency and success in meeting existing codes.
Even though present wet cap designs do not appear to satisfy
most air pollution codes, these systems are still being purchased
by foundries.
VII - 86
In the 1960's, increasing fabric filter installations were
also made on electric arc furnaces. An example of such an in-
stallation pertained to an eastern railway signal foundry in
the early 1960's. The electric furnace in this foundry melted
about 3 tons per hour and was equipped with an exhaust hood
that had a 24-inch-diameter telescoping and swivel connection
to provide constant ventilation throughout melting and tapping.
The dust collector was equipped with approximately 4,800 square
feet of orIon filter cloth which provides a 1.7 to 1 air to
cloth ratio. To safeguard the orIon bags, a bypass damper was
installed in the exhaust piping at the collector inlet. Al-
though no exact collection figures were made available, dust
and fumes were no longer considered a problem.
39
The American Foundrymen's Society survey disclosed that
fabric filters did a very good job of emission control for elec-
tric arc furnaces. In one foundry, dust collected per ton of
charge was 8.8 pounds when scrap was handled manually. After a
magnet crane was installed to handle scrap, the dust collected
per ton of charge was 13.4 pounds. When scrap was handled man-
ually, much of the dirt would fall to the ground and accumulate,
while with the magnet, the dirt would enter the furnace and exit
as dust. Complaints concerning these fabric filter installations
included:
- Dry dust disposal.
- Bag failure.
- Dust buildup on the clean air side of bag.
-------
VII - 87
- Dust sticking to the side of collection hoppers.
Finally, the survey disclosed that a roof-mounted hood with
baghouse was preferred to a roof tap and baghouse system.1
Because of the problems with controlling cupola emissions
and other factors, some foundries turned to other melting
processes, such as electric induction. This trend began in the
early 1960's and it was the general belief that induction fur-
naces offered relief from air pollution legislation pressures.
At this writing, no control equipment installations have been
reported on €?"»»dry induction furnaces. A series of tests
conducted at an East Coast foundry in 1968 indicated that
coreless induction melting and a charge preheater could be
operated in most installations without pollution control devices
40
and yet comply with existing or pending emission limits.
Non-Melting
Processes
Contrary to the many developments and refinements in melt-
ing emission control equipment systems during the past 25 to 30
years, emission control of non-melting processes such as shake-
out, sand handling, cleaning, finishing, coremaking and pattern
shopwork is substantially the same now from a technology stand-
point as in the mid-1940's. Practically all of the major
emission control problems in these areas had been resolved by
World War II except for a few improvements which were made in
later years. This appears to have occurred for a number of
reasons.
First, emission control of non-melting processes has proven
VII - 88
to be more economical and less complicated. Also, these process-
es, when compared to melting, do not contribute as much on an
individual basis to the foundry emissions problem. Consequently,
less attention has recently been directed toward non-melting
processes regarding emission abatement. Reports regarding past
attempts to control emissions in these areas are limited in
comparison to the widespread interest in solving melting air
pollution problems.
Foundry cleaning room operations were among the earliest
non-melting processes where attempts were made to control or
capture emissions. Prior to the 1930's, sand blast rooms and
tumbling mills were in extensive use and local exhaust systems
removed the dust, quantities of burned sand, scale and spent
abrasive. This produced a number of advantages. First, visi-
bility for the blast room operator was improved and also, since
the fractured dust fines were removed, better cleaning efficiency
of the abrasive was possible. In addition, a shorter tumbling
mill cleaning cycle was obtained by preventing accumulations of
dislodged dust from cushioning the impact forces of the tumbling
barrel stars. However, it was recognized by some foundries that
discharge of such quantities of dust to the atmosphere without
collection equipment would cause a neighborhood nuisance.
The hollow trunion design of early tumbling mills made dust
collection feasible and dry mechanical, low energy cyclones as
well as fabric filters were used for this purpose. Fabric
filters were probably more frequently applied and early designs
-------
VII - 89
were crude, usually made of wood. Dust collection was fairly
good but there were problems with cloth vibration, dust escap-
ing between the fabric and the support and also filter media
replacement.
Prior to the 1930's, local exhaust systems were used for
various grinding processes especially where sand grinding was
employed. Dust particles torn from the cast metal surfaces
by the grinding wheels were relatively large and low energy
cyclones or fabric filters were used during this early period.
Fabric filters were often employed when a single system was
used to exhaust grinders and abrasive cleaning or tumbling mill
operations. Fabric filter emission cleaning efficiency was
fairly good but the use of cyclones was marginal. Although
there were no visible particulates in the cleaned air from the
collector, oxidation of the settled material discolored buildings
and walkways in the area. When higher efficiency cyclones were
introduced, this became less of a problem.
The introduction of mechanical sand handling systems and
central shakeout equipment in the larger foundries in the early
1930's made dust control through local exhaust ventilation feasi-
ble. Dust was released in great quantities in specific locations
as compared to the more general release in smaller foundries
where floor molding was usually employed. As long as the sand
was dry, dust could be collected with fabric filters. However,
dry sand required baked sand molds or the storage of the poured
flask in an outdoor storage yard for cooling. Most gray iron
castings were, and still are, made in a green sand mold and shaken
A T *F»"VrY* ' 'i "I F' \ V Y F « r
VII - 90
out of the flask as the hot casting has solidified. Under these
conditions, steam in quantities was released by the hot, moist
sand. This moisture, picked up by the exhaust air heated by
the hot castings and sand, made application of fabric filters
impractical for most gray iron foundry shakeout and sand systems.
Moisture would condense on the filter surface and combine with
the sand fines to prevent an effective flow of exhaust air.
This condition was dealt with in latter years when wet scrubbers
were introduced.
Early foundry wood pattern shops used exhaust systems and
later low energy dry mechanical cyclones for dust collection.
Since particle sizes from sawdust, wood chips and slivers are
large, and also because low velocities are required, collection
was good and no significant problems existed. Through the years,
cyclones continued to be used and are the usual selection today
for woodworking processes.
Fabric filters have also been employed over the years in
woodworking areas, often as a final cleaner when cleaned air is
returned to the work space during the heating season.
The introduction of wet scrubbers in the mid-1930's improved
dust collection capabilities in most non-melting areas including
cleaning, grinding, shakeout and sand handling. This was espe-
cially significant in shakeout and sand handling processes where
wet sand created dust collection problems. A midwestern malle-
able iron foundry that installed a wet dynamic precipitator re-
ported a collection efficiency of 99% in 1937. This test occurred
A T K F \ R V r
•I I" > N N F ^ r
-------
41
VII - 91
on a shakeout operation handling four large castings per minute.
Another foundry reported a 98% collection efficiency on a sand
mixing process with a wet dynamic precipitator. This test
occurred in the late 1930's on the East Coast with a collector
41
rated at 5,400 cubic feet per minute. In a cleaning room
application, a wet dynamic precipitator was reported to have
collected 99% of the dust emitted from a tumbling mill operation.
41
This occurred in 1939 at a foundry located in Tennessee.
Successful application of wet scrubbers during the mid- and
late 1930's stimulated substantial product development although
by the mid-1940's, designs had stabilized along the lines of
current four-inch to seven-inch pressure loss collectors.
Early cleaning room practices underwent changes that had
a favorable impact on dust collection capabilities in this area.
Tumbling mills were rapidly replaced by airless blasting, in the
1930's. Also, sand as an abrasive was replaced by grit or shot,
originally chilled cast iron but currently cast steel pellets.
Metal abrasive created less dust from shattered abrasive fines
and had a higher cleaning efficiency than sand. This became
increasingly significant when airless blast equipment was
introduced. The abrasive development was a major stimulant
toward automatic cleaning devices-which by the late 1940's had
largely replaced the manually directed hose in a sand blast
room, except for large or occasional parts, and the tumbling
mills. Most abrasive cleaning units, such as tumbling mills,
were originally discharged to fabric filters, but when four-inch
A 7 K K * II M"
VII - 92
to seven-inch pressure loss wet collectors were developed, these
systems were frequently used. The wet collectors had advantages
especially in the larger production foundries where central
sluicing systems were in use.
By the late 1940's and early 1950's, advances in wet collector
designs made collection efficiencies of 97% or higher typical
for sand handling systems. These systems consisted of conveyors,
elevators, screens, lines and mixers. From 150 to 500 pounds of
dust per hour were collected.
Over the years, many non-melting processes were equipped
with various types of hoods to prevent emissions from discharging
into the plant. Often, they were used in conjunction with a
dust collector. Enclosing hoods and side hoods are two types
used in shakeout processes and these have demonstrated a 97%
and 90% efficiency respectively. This was reported in the early
1950's.42
Reports in the early 1950's on the use of dry centrifugal
collectors on metal grinding processes indicated that particu-
lates were reduced to .002 grains per cubic foot. These processes
were usually two-wheel grinding stands and collectors captured
approximately 15 pounds of cast iron dust per hour
42
In the early 1950*s, medium energy wet scrubbers and fabric
filters were the usual installation on grit blast and airless
blast abrasive cleaning processes. Particulates were reduced to
approximately .02 grains per cubic foot which resulted in approxi-
42
mately 800 pounds of dust being captured per hour.
-------
VII - 93
VII -
As pressures on industry to increase air pollution efforts
persisted in the 1950's and early 1960's, foundries stepped up
application of emission control techniques to non-melting process-
es. More attention began to be directed to such areas as mold
lines, coreraaking, scrap preparation and magnesium treatment for
ductile iron production.
In the mid-1960*s, mechanically energized wet collectors
were installed in a production gray iron foundry in an attempt
to improve mold line ventilation. Outlet samples obtained
yielded dust loading results that were considered lower than
usually achieved in a relatively uncontaminated factory atmos-
phere.*3
One large foundry had suffered extensive damage when the
core ovens exploded due to exhaust stacks becoming plugged with
organic residues. A centrifugal spray collector was installed
with facilities for adding chemicals Co react with the organic
airborne residues. It was reported tn 1965 that the stack
hazard was eliminated and the combustible and polymerized resins
44
removed.
Another example of coremaking fume control pertains to the
experience of a midwestern foundry. In this case, a catalytic
combustion system was installed to complete oxidation of contam-
inants that had been giving off objectionable odors. Complaints
that had been raised by nearby residents ceased after the system
45
was installed.
In preparing scrap for melting, centrifuges were often used
to reduce excess oil. Driers or kilns were applied also by a
number of foundries to drive off moisture and burn off oil from
scrap before processing. Furthermore, abrasive cleaning equip-
ment was used to remove burned sand from foundry returns. This
reduced dust loss in transferring these returns between different
points in the foundry.'
A West Coast foundry reported in the late 1960's that plant
conditions improved and emissions were reduced as a result of a
fabric filter installation on a ductile Iron magnesium treatment
station. Less emissions were captured when ductile Iron was not
being produced. This system was also controlling desulphurization.5
Efforts continued to be directed in the 1960's toward con-
trol of heavy dust producing areas such as cleaning, finishing,
grinding and shakeout.
A centrifugal spray collector controlled outlet loadings in
a cleaning department to .003 grains/standard cubic feet. This
result was obtained by an outside testing laboratory in the mid-
1960 's in a high production gray Iron foundry.*
An orifice wet scrubbing collector was installed in the
grinding and cleaning area of a midwestern foundry in the mid-
1960 's. The dirty air was drawn via suction pipes from points
near each piece of equipment. The breathing area around each
grinder was reported co be substantially dust free and workers
did not require masks. Much of the exhaust piping formerly
-------
VII - 95
VII - 96
employed at each grinding station continued to be used, but
required alteration.
A large midwestern automotive foundry recently reduced
emissions in their finishing department to .05 grains per stan-
dard cubic foot. Wee collectors, centrally located on the plant
roof, achieved this performance. The central location was re-
ported to permit ease of maintenance and repair.
A midwestern foundry recently equipped its shakeout area
with an orifice wet scrubber. The collector was mounted on an
overhead steel structure Chat allowed headroom in working areas.
Dust was drawn into the collector by two fans and after scrub-
bing, the sludge deposited onto a conveyor and then dropped Into
a drum for removal. The collector, rated at 20,000 CFM was re-
ported to have Improved operations and working conditions in the
plant.46
In another shakeout example, tests performed by an indepen-
dent laboratory showed outlet loadings from a centrifugal spray
collector to be .0006 to .0007 grains per standard cubic foot.
This test was documented in 1965.**
Dust collection equipment also continued to be used in
pattern shops. An East Coast foundry, attempting to control dust
from woodworking equipment operations, installed a roof-mounted
bag collector in the 1960's. Although no quantitative collec-
tion figures were available, it was reported that the desired
level of air cleanliness was being achieved. Each woodworking
machine was equipped with a hood which the dust traveled through.
before being captured by the collector.
REFERENCES
L. Foundry Air Pollution Control Manual. American
Foundrymen's Society, 1967.
2. Cupola Emission Control, translated by P.S. Cowen
published by the Gray and Ductile Iron Founders' Society, Inc.
3. "Dust, Fume and Smoke Suppression," A. Grlndle,
Iron and Steel Engineer. July, 1951, p. 92.
4. Dust Collectors. American Foundrymen's Society, 1967.
5. Personal Notes of John Kane.
6. "Available Control Equipment for the Foundry Cupola
and Their Performance, Operations, and System Characteristics,"
John Kane, presented at American Foundrymen's Society Annual
Meeting, May 7, 1956.
7. "Dust Control," R. Whit lock., Ceramic Age. October, 1965.
8. "Foundry Cupola Dust Collection," W. Witheridge,
Heating and Ventilating. December, 1949, pp. 70-84.
9. American Air Filter, Rolo-Clone-Type N. Bulletin No.
277E, February, 1967.
10. Pangborn Division, The Carborundum Company, Air
Pollution Control by Pangborn.
II. American Air Filter, Reverse Jet Fabric Dust Collector,
Bulletin No. 279E.
12. "A Discussion of Some Cupola Dust Collection Systems-
Part 3," C. R. Wledemann, Modern Casting. January, 1970,
pp. 72-74.
-------
VII - 97
VII - 98
13. "Choosing Your Electrostatic Precipitator," E.P.
Stansny, Power. January, 1960, p. 62.
14. "The Design and Performance of Modern Gas Cleaning
Equipment," C. J. Stairmand, Journal of the Institute of Fuel.
February, 1956.
15. The Cupola and Its Operation, published by the
American Foundrymen's Society, 1965.
16. Introduction to Air Quality Management. Training Course
Manual in Air Pollution, U.S. Department of Health, Education
and Welfare, pp. 8-10.
17. Air Pollution Manual, published by American Industrial
Hygiene Association, 1968.
18. Torit, Dust Collectors, Series 130, January, 1966.
19. Air Pollution Engineering Manual. U.S. Department of
Health, Education and Welfare, Public Health Service, Publication
No. 999-AP-40, 1967.
20. Private communication.
21. Control of Emissions from Metal Melting Operations.
American Foundrymen's Society.
22. Personal notes of James Ewens.
23. "Cupola Dust Collection," W. Witheridge, The Foundry.
March, 1950, p. 88.
24. Cupola Gas Scrubbers. 0. J. Brechtelsbauer, American
Foundrymen. February, 1955, pp. 34-37.
25. "Collectors on Cupolas Clean Waste Gases," 0. Allen,
The Foundry. November, 1945, p. 89.
26. "Control of Cupola Stack Emissions," J. Drake and T.
nard, The Iron Age. April, 1949, p. 88.
27. "Around Detroit," 0. H. Allen, The Foundry. March, 1949,
124.
28. Control Emissions from the Electric Furnace. L. Krueger.
29. American Air Filter, Case History Report, March, 1959.
30. "A Study of Cupola Design and Operating Factors That
luence the Emission Rates from Foundry Cupolas," R. C.
gies, reprinted from American Foundrymen's Society Transactions,
5.
31. Cupola Fly Ash Suppression. R. M. Overstried, pp. 550-
32, The Study of the CVX Wet Gas Scrubber in Its Applications
Power Stations. Foundries, and Iron Ore Mills. John Tailor.
33. "Computerized Hot Blast Cupola Serves Induction Holding
nace in Production of Milton Iron," Industrial Heating.
ober, 1967.
34. National Dust Collector Corp., Report No. 201, May 19,
9.
35. National Dust Collector Corp., Report No. 216,
tember 16, 1969.
36. Foundry Facts, published by the American Coke & Coal
micals Institute, No. 5, October, 1968.
37. "Costs, Efficiences and Unsolved Problems of Air Pol-
ion Control Equipment," B. D. Bloomfield, APCA Journal.
. 17, No. 1, January, 1967, pp. 28-32.
-------
VII - 99
38. "Capturing Cupola Emissions," Modern Castings. October,
1969, pp. 54-56.
39. "Controlling Dust and Fumes from an Electric Funace,"
Foundry. September, 1960.
40. "Induction Furnaces, Preheaters and Air Pollution,"
T. Steffora, Foundry. August, 1968, pp. 82-86.
41. American Air Filter, Roto-Clone Efficiency Studies on
Foundry Dust Control, August, 1945.
42. "Foundry Dust Problems and Air Pollution Control,"
John Kane, Foundry. October, 1952.
43. "Performance Testing Data on Mechanically Energized
Spray Wet Type Dust Collectors," R. Jamison, Air Engineering.
June, 1965.
44. Performance of Centri-Spray Wet Type Dust and Chemical
Fume Collectors, Orlan Arnold, Ajem Labs, June 11, 1965.
45. "Air Pollution Problem Solved," Foundry, no date.
46. "Controlling Foundry Dust," G. Medley, Foundry. June,
1966, pp. 203-204.
47. "A Total Approach," Metal Progress. May, 1969.
48. "Foundry Dust Gets The Air," Iron Age. January 21,
1965.
49. "Cooling Hot Gases," S. A. Keigel and L. Rheinfrank,
Jr., Pollution Engineering. November/December, 1970.
VIII - ECONOMIC ANALYSIS
OF EMISSION CONTROL TECHNOLOGY
GENERAL
This section deals with the costs of air pollution control
systems in the iron foundry industry. Estimates have been made
of the capital costs and operating costs of emission control
systems, based upon analysis of existing and proposed installa-
tions. Additionally, the capital and operating costs of melting
departments have been developed for a variety of types and sizes
of installations, and the effects of costs of emission control
on the overall costs have been determined.
The basic data used in the development of the cost esti-
mates were obtained from the following sources:
1. The in-plant surveys which were conducted as part
of this study.
2. Information supplied by selected manufacturers of
pollution control and melting equipment. This information was
generally in the form of bid figures.
3. Cost information and operating data obtained in
the process of developing the foundry data bank.
4. Cost information and operating data from the
engineering files of A. T. Kearney & Company, Inc. based on
assignments in development of iron foundries.
The cost data obtained are basically for foundries with
cupolas controlled with wet caps, mechanical collectors, fabric
A T KEARNEY «t c OMP1NV l-.r
-------
VIII - 2
filters, and high and low energy wet scruubers, for electric arc
furnaces controlled with fabric filters, and for induction furnaces
with preheat systems.
CAPITAL COSTS OF
CONTROL SYSTEMS
An emission control system was considered to include all
items of equipment and their auxiliaries whose sole purpose is
to abate atmospheric pollution from a cupola or electric furnace
in an iron foundry. The capital costs associated with the
estimates included the following cost categories:
- Basic Equipment
- Auxiliary Equipment
- Engineering Cost
- Installation Cost
The summation of these cost categories for a control system
constituted the total investment cost. The cost of land upon
which the control system is located has not been included.
The specific items included within each cost category are
as follows:
1. Basic Equipment - This cost includes all taxes
and shipping charges in addition to the "flange-to-flange"
price of the basic equipment.
2. Auxiliary Equipment - The cost for auxiliary equip-
ment includes the cost of all items essential to the successful
operation of a control system. This category has been further
VIII - 3
divided into the following subdivisions-
(a) Air movement equipment costs including:
- Fans and blowers.
- Motors, starters, wire, conduit,
switches, and other electrical
components.
- Hoods, ductwork, gaskets, and dampers.
(b) Liquid movement equipment costs including:
- Pumps.
- Motors, starters, wire, conduit, switches,
and other electrical components.
- Piping and valves.
- Settling tanks.
(c) Storage and disposal equipment costs
including.
- Dust storage hoppers.
- Sludge pits.
- Draglines, trackway, and roadway.
(d) Support construction costs including-
- Structural steelwork.
- Foundation, and piers.
- Insulation.
- Vibration and antiwear materials.
- Protective cover.
(e) Instrumentation costs including the
measurement and control of:
- Air and liquid flow.
- Temperature and pressure.
- Operation and capacity.
A T KF^RVFY Sc
A T KEAHNTY 8c COMPANY
-------
VIII - 4
- Power.
- Opacity of the flue gas (smoke meters, etc.).
3. Engineering Cost - This cose allocates the research
and engineering expenditures required for the selection of the
specific control system, including such items as:
(a) Material specification.
(b) Gas stream measurements.
(c) Pilot operations.
A. Installation Cost - This cost includes the follow-
ing items when applicable:
(a) Labor to install.
(b) Site preparation.
(c) Building modification.
(d) Design contingency.
(e) Field office charges including supervision
and engineering.
(f) Inspection.
(g) Protection of existing facilities.
(h) System start-up.
(i) Profit reduction attributable to plant
shutdown for installation.
Exclusive of the value of the land the equipment is located
on, the summation of the foregoing items represents the total
investment required to purchase and erect the pollution control
equipment, connect it to the furnace, and bring it into operation.
This total investment cost is the figure used in analyzing capital
coses of Installations.
VIII - 5
fne different installations for which cnpit.il cost dnta
were obtained represent systems whose ages range from the
present to 20 or more years. To compensate for differences in
installation dates, investment costs were converted to a common
base of 1969 dollars. This adjustment was made by using the
material, construction, and labor indices published by Engineer-
ing News-Record, March, 1970.
The basic and auxiliary equipment costs are the main com-
ponents of the total capital cost. These equipment costs varied
on the average from 42% to 66% of the total capital cost.
The collection equipment costs for wet scrubbers (limited
to the collector, exhauster and motor) varied more widely as a
function of pressure drop. These component prices are affected
by the energy required. Increases in pressure drop for example,
result in larger horsepower needs which are reflected in greater
investment costs in the motor, starter, and all other electrical
components. In the following chart, the relative variation is
shown. The reference index is unity for 20-inch pressure drop.
3-
10 20 30 40 50 60
PRESSURE DROP
70
80
' •» x- p % c f r« *
-------
VIII - 6
The curve illustrates that the cost of the basic control
equipment is more than twice as great for a 40-inch pressure drop
system than for a 20-inch pressure drop. The relative change from
40-inch to 80-inch is about 20%. In addition to the basic control
equipment cost increase, there will be increases in the necessdrj
appurtenances such as electrical starters, wiring, controls, etc.
Not included in the above curve are the costs for water clari-
fication, quencher, engineering, cupola caps, and erection.
The total equipment costs for scrubbers represent on the
rverage about 65% of the total investment costs. Erection costs
are about 35% of the total cost and approximately 54% of the
total equipment price.
The overall impact of the basic equipment costs (collector,
exhauster, and motor) on the total investment cost is less
significant and well within the shaded range of the investment
cost curves discussed later. The impact of the range in pressure
drops does, however, affect operating costs, due in part to
higher electrical energy requirements. This will be further
covered in the discussion on operating costs.
The following information illustrates the average ratios
and ranges observed for cupola installations.
A T K F \ R V F Y
VIII - 7
Equipment Costs as a Percentage of
Total Investment Cost
Control System
Wet Caps
Mechanical Collectors
Low Energy Wet Scrubbers
High Energy Wet Scrubber
Fabric Filter
Equipment Cost/Total Investment
AverageRange
42%
55%
65%
66%
65%
367. - 677.
367, - 797,.
48% - 80%
487. - 85%
417, - 827.
The ratio of equipment to total investment varies consider-
ably on an individual foundry basis. The wide variance in the
range of equipment cost as a percent of total investment is
caused by several factors. The data represent installations at
many foundries which have many different requirements. Some
foundries had available space for the control equipment while
others required some plant modifications to install the equipment.
The age of.the foundry affects the cost. In some new foundries,
the pollution control equipment was designed as an integral part
of the facility, while in old foundries, additional costs must
be incurred for adaptation of facilities.
These factors and others, as reviewed in Exhibit VIII-1,
represent unique differences between one foundry and another.
The differences affect the total investment cost, as represented
by the ranges given in the previous table.
The major factor that influences capital cost is the type
of melting equipment. The complexity of control systems is
A T KE«iRVFV & ro.MP*VV
-------
VIII - 8
greatest for the cupola, moderate for the electric arc, and
least for induction melting.
Cupola Melting
Air pollution capital costs are most varied for this method
of melting, and control solutions encompass a wide variety of
collector designs. Often, the collector type will determine
the required system components. Separate studies were made of
each of the four major control methods. The factors that have
the greatest influence on the investment costs include those
listed as follows:
1. Number of Cupolas Served on a Single System.
2. Materials of Construction - Corrosion is a major
problem in cleaning cupola gases by wet methods. Temperature
is another major factor. Service factors influence the selection
of materials of construction. Investment costs will vary signifi-
cantly when the unit is constructed of corrosion-resistant metals
such as stainless steel or when the ductwork is refractory lined.
3. Water Supply and Solids Disposal - Investment in-
creases for wet systems when drain water is clarified in a separate
settling tank, chemically treated, and recirculated by pumps back
to the collector. These costs are obviously less for systems
using dry collection methods.
4. Erection Costs - Costs vary greatly with prevailing
labor rates. In addition, the location of the control system,
whether on a new or existing site, plant arrangement, and special
rigging required, will affect costs.
A T KEARNEY & COMP\XY
VIII - 9
Electric
Arc Melting
A typical air pollution control system for thi« method of
melting consists of a local exhaust hood on, or at, the furnace
roof, duct connections from the hood to the collector, and n
fabric filter collector. The factors which influence the major
components of an electric arc control installation include the
following-
1. Exhaust Hood - The exhaust hood will vary with the
diameter of the furnace. Also influencing costs are the problems
'-~f .attaching or supporting the hood to the furnace structure and
the ease of providing a suitable fixed exhaust point that will
permit the furnace roof and hood to swing for charging and tilt
for slagging and pouring.
2. Duct Connections - The cost of ducts will vary
with the size of the ducts, which are a function of exhaust volumes,
the length of duct run and the amount of interference encountered.
3. Dust Collector - These costs are essentially a
function of exhaust volume. The cost of power supply for col-
lector motors depends upon the collector location and the plant
power system.
4. Installation - Erection costs vary with prevailing
labor rates, and problems of adapting to existing conditions.
Factors which contribute to cost variations are summarized
in Exhibit VIII-1.
* P V r V A- r- o V
-------
VIII - 10
RELATION OF CAPITAL COSTS
OF CONTROL SYSTEMS TO
DESIGN VARIABLES
A number of design and operating variables are related to
the total cost of emission control systems. These include melt
rate of the furnace, pressure drop of the system, blast volume
in standard CFM and gas volume throughput in actual CFM. A
computerized multiple regression analysis of the collected data
has confirmed that total costs varied most directly with actual
gas volume throughput.
The greatest differentiation in capital costs due to outlet
dust loading existed with wet scrubbers, for low energy systems,
under 30-inch static pressure drop, and high energy systems, over
30-inch static pressure drop. The costs of these two categories
have been considered separately.
Cupola
Melting
The total investment costs of high and low energy wet
scrubbers, mechanical collectors, and fabric filters vary directly
with gas volume throughput in actual cubic feet per minute. Invest-
ment cost equations were obtained by multiple regression analysis,
in which the equations obtained were the best approximation of
total investment costs from the available data. The investment
cost equations for pollution control equipment installed on
cupolas are presented in Exhibit VIII-2. The column in Exhibit
VIII-2 entitled "Limits of Observation" gives the ranges of gas
volumes for which each equation.is validfr -
A T KEARNEY * COMPtNY Ivr
VIII - 11
As shown in this exhibit, excellent correlations were ob-
tained with high energy wet scrubbers and fabric filters. A
perfect correlation coefficient is 1.0. Although the correlation
coefficients for low energy wet scrubbers and mechanical collectors
were lower, they were considered to be extremely good. Graphs
of each equation are plotted in Exhibits VIII-3 through VIII-5.
The range shown on each graph represents the standard deviation
of the statistical analysis.
The total investment curves all relate total investment?
costs in 1969 dollars to gas volume throughput. Gas volume
throughput is related to cupola operating condition* and—£unaee-
design. The significant factors which influence gas volume in a
cupola are the following:
- Blast volume.
- Size of charging door.
- Method of cooling exhausted ga-ses. •
- Location of exhaust gas take-off.
- Gas temperature at the collector and fts~exhatzstez*.-
Blast volume is a function of the oxygen needed to burn the
coke. Metal to coke ratios can be-increased- to produce-higher.
melt rates by preheating the blast air, by oxygen enrichment -of•
the blast-air, and by the -us* of-& snpplemeanT fuel. Houupgt,~
in present-day practice with an open charging -door and wtefr^gas
take-off above the door, the variation in- blast volume -caaecd by -
these factors will become a less significant factor in the total
gas throughput handled by the control system than the amount of
infiltrated air drawn in through the charging door.
A T KEARNEY & COMPANY Ivc
-------
VIII - 12
Charging door openings require an inward flow of air to
confine cupola gases and contaminants. Many designers of air
pollution abatement equipment specify an indraft based upon 300
FPM through the door opening, which results in significant addi-
tional exhaust volumes that must be handled by the exhauster and
collector. Even such volumes are subject to wide variations where
charge door openings are exposed to aspirating crosswinds.
Charge door openings are determined by the method of charging and
the size of charging equipment, and consequently may have little
relation to melting rate. The effect of air infiltration may
increase gas volumes to as much as 2007. where gas take-off is
above the charge door.
Hot gas in the cupola stack must be cooled to safe tempera-
tures for the control equipment and exhauster. The degree of
cooling will differ with the type of system but the exhaust gas
is normally cooled down to about 500° F for mechanical or glass
fabric collectors and below 200° F for wet scrubbers and synthetic
fabric collectors. The usual methods of cooling are by water
sprays or air cooled ducts. However, cooling of gases by use of
water from the high temperatures in the cupola stack introduces
another significant gas volume, that of the water vapor generated.
Water vapor can increase the total gas volume handled by the con-
trol equipment by 20% to 357..
The location of the gas take-off also contributes to the
overall cost of the control system. The placement of the take-
off ,b»low the charge door reduces the total gas volume handled
A T KFM^r^ & COMPANY l^r
VIII - 13
since indraft air is at a minimum. The cost for control equipment
for any given type of control system is also subject to a number
of other factors that have less correlation with melting rate than
the gas volume.
Exhibits VIII-6 and VIII-7 give the approximate melt rates
and gas volumes for lined and unlined cupolas. The increase in
gas volume due to a gas take-off above the charge door is quite
apparent when compared to a below the charge door take-off.
Exhibit VIII-8 presents a single comparison of gas volumes for
gas take-offs above the charge door and below the charge door for
a lined cupola.
Operating a cupola with an open charge door with an above
the door gas take-off results in a greater total gas volume.
The cost of the control equipment is then based upon this higher
gas volume. Efforts to reduce the total gas volume have been
successful in several ways, as described in the following list:
1. Providing doors to partially seal the charge open-
ing except during the charging period.
2. Reducing the charge opening through the use of a
vibrating feeder instead of a bottom drop or skip bucket charger.
3. Enclosing the skip hoist to eliminate the wind
effect and reduce the net open area.
It. Exhausting the cupola gases from below the charging
door level.
5. Using a top charge double door design.
A T KFARNTY gt roMI'\» \
-------
VIII - 14
The amount of water vapor in the exhaust gas can be reduced
by use of indirect heat exchangers which use the heat in the exhaust
gas for heating the blast air. With a below the door take-off,
some wet collector systems quench the gases without burning off
the CO. This results in exhaust gas temperatures of 400° F-700° F
instead of the more usual 1500° F found with afterburning, and
also results in a corresponding reduction in the volume of water
vapor produced.
The total investment costs can be approximated by first
determining the total gas volume generated from Exhibits VIII-6
and VIII-7 and applying this volume to Exhibits VIII-3 through
VIII-5 for the cost of the specific system type. However, for
ease of interpretation, Exhibits VIII-9 through VIII-14 have
been prepared to relate installation costs for high and low energy
wet scrubbers and for fabric filters directly to melting rate
of cupolas in tons per hour. The melting rates were based on an
average metal-coke ratio of 8:1. Wet cap installations are
directly related to cupola diameter with relatively little
relationship to melting rates or gss flow. Total investment
costs for wet caps as a function of cupola size are given in
Exhibit VIII-15.
Wet Scrubber Efficiency-
Pressure Drop Relationship
The wet scrubber curves have been separated into high
energy--over 30-inch pressure drop--and low energy—under 30-
inch pressure drop. The relative effectiveness of each is shown
in Exhibit VII-21. High energy scrubbers on cupolas are up to
A T KTMISFY «. COMPANY i^~
VIII - 15
six times more effective than low energy systems, with n tvpicnl
outlet loading of 0.05 gr/SCF for 30-inch to 70-inch pressure
drop systems compared to 0.3 gr/SCF for systems under 30-inch
pressure drop.
The degree of gas cleaning obtained at various pressure
drops is illustrated in Exhibit VII-28. Wet scrubber efficiencies
in capturing particles exceeding 1 micron in size are 90% or more
for pressure drops of 10 inches and above. At 30-inch pressure
drop the efficiency exceeds 99% for particles of 1 micron and
larger. However, for fine particle sizes, in the range of 0.2
micron and below, a 30-inch pressure drop system is only 957.
efficient.
The overall collection efficiency is determined by the par-
ticle size distribution in the effluent gas and the ability to
capture these particles. For a constant pressure drop, a scrub-
ber has a different collection efficiency for each particle size.
The overall efficiency is the weighted sum of efficiencies for
each particle size.
Exhibit VIII-16 has been derived as a theoretical presenta-
tion of overall collection efficiencies of wet scrubbers at various
pressure drops. This was done for a typical particle size distri-
bution and is based upon data found in the available literature.
As shown in the exhibit, the efficiencies in percent of weight
collected range from 95.4% at 5-inch pressure drop to 99.8% at 60-
inch pressure drop. The results obtained are theoretical and not
A T KEARNEY Sc COMPNVY 1-.
-------
VIII - 16
representative of actual efficiencies which may be obtained at
these pressure drops. However, the exhibit does illustrate the
trends in increased collection efficiency at increased pressure
drops.
Increased collection efficiencies due to higher pressure
drops result in a decreased outlet loading. The amount of the
outlet loading or discharge is influenced by local air pollution
codes. Many codes provide for a maximum allowable discharge.
A plot of the outlet loading versus pressure drop for found-
ries for which data were available is given in Exhibit VIII-17.
The outlet loading is expressed in pounds/ton which factors the
melt rate into the total allowable discharge of pounds/hour
expressed in process weight tables of many air pollution codes.
This exhibit illustrates the expected increase in pressure
drop required to reduce outlet loading. An approximation can
easily be made to determine the energy necessary to satisfy the
maximum allowable discharge for a foundry for those codes which
utilize process weight tables. The Los Angeles County Code, for
example, limits the maximum discharge to 10 pounds/hour for a
process weight of 10,000 pounds/hour (5 tons/hour). The result-
ing outlet loading is 2 pounds/ton, and from Exhibit VIII-17
the pressure drop necessary to satisfy this condition is 30
inches.
Electric Arc
Melting
Although the electric arc furnace is used in the iron foundry
A.T KEARNEY & COMPANY. IMC
VIII - 17
for both melting and for holding, the economic analysis has been
directed only at the melting application. The reason for this is
that only a few holding furnaces have been equipped with emissions
control systems because the level of emissions for holding only
is relatively low when compared with melting. The economic
analysis was also limited to fabric filter collectors on arc
furnaces, again because only a feu installations have been made
using other methods.
Fabric filter installations on electric arc furnaces have
been related to the diameter of the furnace, rather than the tons
per hour melted. Exhaust volume and furnace diameter relation-
ships approximate the figures in the table below and in Exhibit
VIII-18.
Exhaust Volumes for Electric Arc Furnaces
ROOf
Diameter
6 feet
8 feet
10 feet
12 feet
14 feet
CFM
Local Hood
20 ,000
26,000
30,000
33.000
36,000
CFM
Canopy Hood
55,000
70,000
90,000
115,000
150,000
Fabric filters can be of the intermittently cleaned,
single section design, or the compartmented, continuous duty
design. When each melting furnace has its own fabric filter, the
collector can be stopped and serviced at the end of each melting
cycle. For multiple furnaces controlled by a single collector
A T KEM*VFY fie rriMfivv i.
-------
VIII - 18
and for large exhaust volumes, the continuous duty desLgn is
usual. On this basis, cost estimates could be approximated as
shown in the tabulation below.
Approximate Installed Cost,
S/ACFM for Electric Arc
Roof
Diameter
6 feet
8 feet
10 feet
12 feet
U feet
Local
Intermittent
Operation
$2.10 -
1.90
1.85
1.85
l.BO
Hood
Continuous
Service
$2.50
2.50
2.35
2.30
2.25
Remote Canopy Hood
Intermittent
Operation
$1.25
1.25
1.25
1.25
1.25
Continuous
Service
$1.85
1.75
1.70
1.60
1.60
Exhibit VIII-19 gives the approximate total installed cost of
fabric filter collectors on electric arc furnaces. This cost has
been related to furnace size and has also been converted to melt-
ing rates usually associated with various furnace diameters.
Other Melting
Methods
Induction melting furnjces, as well as induction holding
units, have only rarely been equipped with emission control sys-
tems. This is also true for reverberatory methods of melting.
For this reason, information has been lacking regarding costs of
installations on this type of equipment. Therefore, data on
costs have not been prepared for this study for these types of
installations. However, the cost of induction melting will be
discussed in a later section dealing with melting department
costs.
A T KEARNEY & COM HANY
VIII - 19
OPERATING COSTS OF
CONTROL SYSTEMS
A comparison of the impact of operating costs has been made
for wet scrubbers and fabric filters. These were the more popu-
lar and commonly used types of control systems observed on found-
ry melting equipment.
The annual operating costs associated with air pollution
abatement equipment are comprised of several components, includ-
ing the following:
- Fixed charges
- Maintenance
- Operating costs
- Water treatment
- Depreciation
Fixed charges are calculated on the basis of total invest-
ment cost. A rate of 13% per year has been used to account for
interest, insurance and taxes.
Maintenance includes labor costs and routine replacement
costs, such as bags in a fabric filter. Maintenance expenditures
are not necessarily uniform from year to year and will increase
after the passage of time. New equipment may require little or
no maintenance but maintenance requirements will generally in-
crease as the system grows older.
Data collected have Indicated that an annual expenditure of
four to five percent of the original equipment cost for wet
scrubbers provides an adequate amount for good maintenance. In
A T KEARNEY & COMPANY IKC
-------
VIII - 20
VIII - 21
the case of fabric filters, the maintenance figure was estimated
to be about five percent per annum of the original equipment cost.
These percentages were reduced to about two percent for smaller
foundries which melted less than 1,000 hours per year, with
lower pollution control system utilization.
In addition, a separate amount was added for bag replace-
ment. This vas based on a 100X bag replacement each year for
foundries with continuous melting and reduced to a 100% bag
replacement each third year for foundries with intermittent
melting. The replacement cost was $10 per bag. Manufacturers
of fabric bags have indicated that In a large number of cases,
the replacement of filter bags has been less frequent than once
a year, with bags sometimes lasting five years or more.
The number of filter bags required for cupolas and electric
arc furnaces with local hoods was calculated for an air to cloth
ratio of two to one. For electric arc furnaces with a canopy
hood, the air to cloth ratio used was four to one.
The maintenance costs used in the analysis do not include
costs associated with equipment modifications after installation.
A few foundries visited reported a large expenditure for changes
early In the life of the control equipment, and some reported a
cost for modifications after several years.
A disproportionate Increase in maintenance costs may possibly
occur in systems with large pressure drops. The increased horse-
power and system requirements due to greater pressure drops may
create conditions that accelerate the maintenance costs. This
has not been determined for this study.
Operating costs include electrical power and makeup water.
For wet scrubbers, these are the major costs. Electrical power
is required for the fans and pumps. Power requirements vary with
the equipment size and pressure drop. The total power require-
ments for high and low energy scrubbers are 15 HP/1,000 ACFM and
S HP/1,000 ACFM respectively. These formed the basis for cal-
culating operating costs for scrubbers. Power requirements for
fabric filters were based upon 2 HP/1,000 ACFM.
Requirements for makeup water in gas cooling and scrubbing
are related to equipment size. The data collected have Indicated
a correlation of makeup water to equipment size for wet scrubbers
at about 2 gallons/1,000 ACF.
Electrical energy and demand costs were calculated at
rates based upon the prevailing power rates in the Midwest for
process industries. Water costs were also based upon rates in
Large midwestern cities. The water cost used in the analysis is
$.275/1,000 gallons.
Many foundries utilize afterburners in cupolas to insure
:he ignition of the top gases. The combustion of the top gases
burns off any carbonaceous matter present in the exhaust gases.
Dther foundries rely on afterburners to specifically burn off
:arbonaceous material. A review of the capacities of the burners
jtllized indicates that the selection is somewhat independent of
A. T. KEARNEY ft COMPANY INC
A T KEARNEY » COMPANY. INC
-------
VIII - 22
cupola size or melt rate and is rather arbitrary. Several
foundries had afterburners of such size and capacity to provide
substantially more energy than required for complete combustion.
The use of afterburners has not been considered in the total
annual costs. However, an approximate cost per operating hour for
each million BTU's of gas burned Is $.60, based on midwestern
gas costs.
Water treatment includes the chemical treatment to the
scrubber water in the sludge tank. This includes flocculation
and neutralization. Foundries requiring water treatment indi-
cated a range of $.03 to $.085 per ton of iron for chemical
costs. For the purpose of this analysis, $.05 per ton was used.
Depreciation was calculated on a straight line method based
upon total investment. The useful life of control equipment
was taken as 10 years. In several foundries, however, the equip-
ment was found to have a longer service life, but rapid advances
in technology and more stringent local codes may tend to accelerate
technical obsolescence of these older units.
The annual calculations of cost for wet scrubbers are based
upon four levels of foundry operation: 500, 1,000, 2,000 and
4,000 hours/year. It is reported that in large foundries operat-
ing 4,000 hours or more per year, rather than shutting down
cupolas during "off" shifts and over a weekend, they may be run
at a reduced level. This results in additional operating hours
for the control equipment, which may cause a slight increase in
the cost estimates.
VIII - 23
Exhibits VIII-20 and VIII-21 present approximate total
annual costs for high and low energy scrubbers on cupolas and are
presented for operating levels of 500, 1,000, 2,000 and 4,000
hours/year. The curves relate total dollar costs versus gas
volume for a single system pressure drop. As the pressure drop
of a control system increases, the power requirements will in-
crease. A doubling of the pressure drop, for example, results
in a doubling of the required horsepower.
Exhibit VIII-22 presents the relative change in total .
annual costs as pressure drop increases. These were calculated
for each operating level. It becomes readily apparent from this
exhibit how a change in pressure drop affects wet scrubber annual
costs.
If pressure drop increases from 30 to 40 inches for a 5-ton-
per-hour melt rate and a 500-hour/year operating level, the in-
crease in annual costs would be from an index of 1 to about 1.04
or a 4* rise. The corresponding increase in theoretical removal
efficiency, from Exhibit VIII-16, would be a rise of 0.1%. A
larger foundry operating at 4,000 hours per year and at 30-ton-
per-hour melt rate would experience an annual cost increase of
approximately 6% for a pressure drop increase from 40 to 50 inches.
The increase removal efficiency for this example would be about
0.1%.
The detailed information to make up Exhibit VIII-22 has
been plotted as total annual costs in dollars, versus gas volume
for a family of pressure drop curves. This was done for each
A T KEARNEY & COMPANY INC
A T KEARNEY «> <
INC
-------
VIII - 24
operating l&vel. These curves are presented in the detail
economic cost curves in Appendix D.
Tn the case of fabric filters, it was assumed that gas
cooling is accomplished by air and not by a water quench. This
removes the need for makeup water and water treatment. The total
annual operating costs for baghouse installations on cupolas are
illustrated in Exhibit VIII-23. These costs are presented for
4,000 hrvirs/year, 2,000 hours/year and 1,000 hours/year. The
comparable costs for fabric filter installations on electric arc
furnaces are given in Exhibit VIII-24 for canopy hood and local
hood exhausts. Operating levels of 4,000 hours/year, 2,000
hours/year and 1,000 hours/year are also presented.
RELATION OF OPERATING
COSTS JF CONTROL SYSTEMS
TO DESIGN VARIABLES
The annual costs of wet scrubbers and fabric filter
collector;; on cupolas, as presented in Exhibits VIII-20 through
VIII-23. £.re all based on gas volume throughput expressed in
ACFM. acwever, there is a relationship between gas volume and
melt raci for different metal to coke ratios. Exhibits VIII-6
and VIi:-~ gave this relationship for lined and unlined cupolas.
By 'j-ambining Exhibits VIII-6 and VIII-7 with Exhibits
VIII-20 rrroueh VIII-23, the annual cost of wet scrubbers and
fabric f.trers on cupolas are expressed as a function of melt
rate for i variety of metal to coke ratios. Also included are
A T KF.ARNI V A € <>P>
VIII - 25
comparative costs for gas take-offs above and below the charge
door.
The curves of annual cost versus melt rate are presented in
Appendix D. Sets of curves are given for operations of 4,000
hours/year, 2,000 hours/year and 1,000 hours/year for low and
high energy wet scrubbers on cupolas and for fabric filters on
cupolas. Both lined and unlined cupolas are considered.
As expected, installations with gas take-offs 'above the
charge door have higher annual costs than similar control equip-
ment installations but with gas take-offs below the charge door.
The total annual costs for scrubbers and fabric filters on
cupolas, as illustrated in Appendix D, have also been modified
to present the costs on a basis of cost per ton of melt. Esti-
mates of cost per ton of melt are given in Appendix D for the
various operating conditions and operating hours discussed
earlier.
A summary of the cost-per-ton curves, comparing 4,000-hour/
year, 2,000-hour/year and 1,000-hour/year levels of operation,
are given in Exhibits VIII-25 through VIII-30. These summary
curves are given for low energy and high energy wet scrubbers
and fabric filters on cupolas for a constant metal to coke ratio.
From these exhibits, it becomes apparent that the cost per
ton of melt rises rapidly as the size of the foundry operation
decreases. The cost increases, for example, for a high energy
scrubber from S3/ton at a 4,000-hour year to over $8/ton at a
A T KhARMEY * COMP*NY I-
-------
VIII - 26
VIII - 27
at a 1,000-hour year. Lower levels of operation would exceed
$12/con or more.
EFFECT OF EMISSIONS
CONTROL ON CAPITAL AND
OPERATING COSTS IN MODEL
FOUNDRIES
Model Foundry
Development
The effect of adding emission controls, on the capital and
the operating costs of the melting department in the iron foundry,
was developed for a variety of different melting methods,
different sizes of foundry, and different Levels of operation.
The various alternates which were evaluated are described below.
(a) Melting Systems
Studied
1. Cupola melting, cold blast, lined, below door gas
take-off, with unheated forehearth, with fabric filter emission
control.
2. Cupola melting, hot blast, unlined, water-cooled,
below door gas take-off, using channel-type induction holding
furnace, with a high energy wet scrubber emission control system
on the cupola.
3. Arc furnace melting, using channel-type induction
holding furnaces, and provided with a fabric filter emission
control system on the arc furnace.
4. Induction furnace melting, crucible type, no
holding furnace, using scrap preheaters, and with afterburners
provided in the preheater stack for emission control.
(b) Sizes of Foundries
Studied
Sizes of foundries in terms of tons per hour melting capacity
were selected at levels of 5, 15, 30, and 50. The range included
in these melting rates covers most of the foundries operating in
this country with the exception of a few very large captive plants
and some very small operations.
(c) Levels of Operation
Studied
The operating levels included in the study, in terms of hours
per year of melting operation, were selected as 500, 1,000, 2,000,
and 4,000. These correspond respectively to melting 2 hours per
day, 4 hours per day, 8 hours per day, and 16 hours per day.
Because of practical operating limitations on some types of
melting equipment, not all melting methods were analyzed at all
melt rates and operating levels.
A T KFARS'EY * COMPANY lie
A T KEARNEY & COMPANY I •« o
-------
VIII - 28
Qualifications Regarding
Model Foundry Design Data
The intent of the model foundry economic development in
this study was to provide comparative figures relating the capital
and operating costs of melting departments of foundries of varying
sizes and rates of operation, and utilizing different methods of
melting, to the requirements and costs of emission control. The
actual design and operating data were based on commonly utilized
equipment and raw materials, for the foundry sizes and rates of
operation which were selected, and thus represent average con-
ditions rather than actual operating foundry cases. It was recog-
nized th»»t further refinement of these design and operating data
vould probably result in cost figures which would more closely
approximate actual costs in many foundries. However, the figures
which were developed are believed to be reasonably representative
and are useful for comparative purposes.
Equipment and
Building Specifications
The model foundry melting departments which were developed
for this economic analysis were intended to represent typical
conditions for the methods of melting and the sizes of foundries
which were evaluated. For purposes of comparison, similar facil-
ities for auxiliary operations and services were used in each
case. The only exception was that the smallest size melting
department did not contain a full complement of auxiliary items,
in accordance with usual design practice in these small instal-
lations.
it. T KEARNEY St COM PAN-. I »,
VIII •• 29
The melting department was considered to encompass the
facilities and operations from raw materials receiving through
molten iron holding. This included raw material unloading,
scrapyard, charge makeup, charging, melting, scrap preparation,
emissions collection, slag handling, water system, holding
furnace or forehearth, buildings and services.
(a) Melting
Equipment
The melting systems covered were those in principal use:
cold blast, lined cupola; hot blast, unlined water-cooled cupola;
electric arc furnace; and coreless induction furnace. Channel
induction holding furnaces were used for the water-cooled cupola
and the electric arc installations, unheated forehearth was
used for the lined cupola, and no holding furnace was used for
the induction furnaces. High energy wet scrubbers were applied
to unlined cupolas, fabric filters to the arc furnaces, and a
minimum afterburning system on the scrap preheater for the induc-
tion furnaces. An alternate of fabric filter collectors was also
considered for the lined cupolas.
(b) Emission Control
Equipment
Emission control equipment selected was based on those
items normally used to control emissions for a particular melting
method and includes basic as well as auxiliary equipment.
Melting alternate Ho. 2 is equipped with a wet
scrubber emission control system. A wet scrubber's capability
to capture particulate matter is a function of the amount of
A T KEARNEY & COMPANY. IMC
-------
VIII - 30
energy expended in the system which is measured by pressure
drop. The higher the pressure drop, the greater the amount of
particulate matter collected and the higher the capital and
operating cost for the system.
Wet scrubber designs were selected that would satisfy the
more stringent air pollution codes. For the purpose of this
study, the air pollution code for Los Angeles County was selected
as a basis for determining wet scrubber efficiency cost require-
ments.
The Los Angeles Code establishes a maximum allowable dis-
charge per hour based upon process rates. Exhibit VIII-17
indicates the minimum pressure drop required to capture various
emission discharge levels. From this information, it is possible
to determine the pressure drops required to satisfy the Los
Angeles Code according to melting rates established for this
study. These figures are shown below:
Melt Rate, Ton/Hour 5 15 30 50
L.A. Code Maximum Discharge
(pounds/hour)
10
22.22 40
40
Outlet Loading (pounds/ton)
1.48
1.33
.8
Wet Scrubber Pressure Drop
Required to Meet L.A. Code 30 in. 40 in. 45 in. 60 in.
Melting alternate Nos. 1 and 3 are equipped with fabric
filters. Since fabric filters collect nearly 100% of all par-
ticulates and therefore meet existing air pollution codes in
analysis of their relative capabilities and efficiencies for
VIII - 31
this investigation uas not required. The only point that needs
clarification is that the fabric filters selected for the arc
furnaces are equipped with a local hood rather than a canopy type.
(c) Buildings
A determination was made of the appropriate amount of
building space which would be required to melt iron based on the
five alternates selected. This included provision for all aux-
iliary building requirements such as transformers and similar
equipment. No provision was made for cost of land.
Capital Costs
This consists of all melting, holding, auxiliary equipment
and buildings as well as emission control systems required for
a typical melting department.
Melting and holding equipment capital costs are current
figures obtained from equipment suppliers and data from foundry
installations. Costs of equipment installations including
foundations and services, such as engineering, have been based on
percentages of original new equipment cost. Building costs per
square foot are based on midwestern construction figures and
vary according to the type of melting method employed. Emission
control equipment capital costs were obtained from information
previously developed which is shown in Exhibits VIII-3, VIII-4,
VIII-18 and VIII-19.
Exhibit VIII-31 summarizes capital costs required to install
melting departments for the various alternate melting methods
A T KEARNEY & COMPANY. INC
A T KF^PS-FY* <- o V r> vv
-------
VIII - 32
VIII - 33
and foundry sizes which were considered. The details of these
capital costs are given in estimates in Appendix D. That portion
of total capital costs accounted for by emission control ranges
from 0.5% to 19.41 depending on the production and operating
conditions involved. More specifically, these ratios vary by
the conditions defined as shown in the table.
Percent of Emission Control Tost
Hi 1C Inn <'o«t« - EVr Tno i'f I rnii
Alternate Number Lowest
1 11. 61
2 12.11
3 5.91
* O.S1
Operating
Costs
Melt Rate Operating
Tons/Hour Hours Highest
5 2.000 19.47.
IS 2.000 16. 3'/.
4.000
50 2,000 12. 11.
4,000
30 2,000 0.6S
50 4,000
Melt Rate
Tons /Hour
SO
50
5
S
IS
Operating
Hours
2.000
A, 000
2,000
4,000
SOO
1,000
2.000
WO
1.000
2.000
4.10(1
These costs were developed per ton of hot metal at the
spout of the melting or holding furnace and include direct
materials, all conversion costs and emission control equipment
operating costs related to the various methods established.
Exhibit VIII-32 gives a summary of total operating costs
per ton required to produce iron based on the alternates under
investigation. These figures are briefly summarized in the
table on the following page.
'itfn'/II.'iur
KiniK'r lu r W.ir
1
2
J
*
,,
son
5IJ7
.MI
182
1 .000 J.OOO 4.000
•IlOJ S TO
81 S7<1
lij 106
128 101
l-i
i .onii
$ 88
125
93
J.ono 'i
S!"
Jh
lJh
82
.{Hill
s;-.
UK
no
74
10
J.OMO 4.IIOl>
•US f(."
71 • >
0" 7t,
80 71
MI
1 «lOi» 'Mil
,;i .,-
i-
-.
i« i •>
(a) Direct Material
Costs
The metal to be produced was selected as gray cast iron hav-
ing 35,000 PSI tensile strength which is a grade commonly produced.
To provide a meaningful comparison, all alternatives were studied
on the same basis. The direct materials used for the various
melting furnaces were selected as representative charge propor-
tions which would provide the desired grade of iron. Prices of
direct materials are based on published quoted prices for the
second quarter of 1970, Chicago area. For both the cold blast and
hot blast cupola melting methods, a metal to coke ratio of 8 to
1 has been used in the study. The direct material costs per
ton for the five melting methods are shown in the table below.
Melting Method
Lined Cupola - Cold Blast
Unlined Cupola - Hot Blast
Electric Arc
Induction Coreless Type
Direct Materials S/Ton
$51.09
47.14
44.69
47.06
These direct material costs have been applied to all melting
rates and all production levels studied. No attempt was made
A.T.KEARNEY & COMPANY INC
A T KEARNEY 8> COMPANY. Ixc
-------
VIII - 34
VIII - 35
ion Costs - IVr Toil of L run
to evaluate the effect on scrap prices of volume buying by the
larger foundries which would operate 2,000 or 4,000 hours per
year. Also, the type of casting being produced affects the
yield and the available percentage of casting returns. Approx-
imately one-third of the metallics were assumed to be r.-turns.
The details of the direct material proportions and per ton
costs are shown in Appendix D.
(b) Conversion
Costs
Conversion costs consist of all direct and indirect labor,
supervision, supplies, maintenance, depreciation, capital charges,
utilities and allocated expense required for charge makeup,
->_lic,..£i-ng, scrap handling and preparation if required, melting
and holding. Allocated expense consists of that portion of
other foundry functions such as general management, quality
control and administration which are necessary to operate a
foundry melting department.
Conversion costs, calculated for each of the alternatives,
were based on operating data obtained from foundries. The data
were modified as required to fit the particular operating con-
dition.
Conversion costs per ton for the various conditions in-
vestigated are shown in the table on the following page.
AllrrnntL Opi-rnrinR.
1
2
3
4
S 66
_
144
144
1 ,000
$45
-
85
•b
2.000
535
38
55
S6
4.000
.
$29
-
-
1 .11(10
$33
-
7S
57
2 .0(1(1
S2ft
27
'.8
>8
4 .000
$20
1'P
11
IK
> .11110 '. .0(1(1
$J2 Sid
22 II-
4 1 i«
V. "•
2,0(10 '. ,(l(l(>
JI9 -1^
!•> 1 .
17 •'
10
Conversion cost details are shown in Appendix D.
(c) Emission Control
Costs
This consists of all costs associated with operating an
emission control system and includes capital charges, maintenance,
utilities and depreciation.
Emission control operating costs were derived from data
previously developed which are depicted in Exhibits VIII-20,
VIII-23 and VIII-24.
1. Alternative costs. Emission control equipment
operating costs per ton for the different conditions investigated
are shown in the table.
ltolt Uace ]
"
-
4
,10 00 S 6 00 S3 50
3 50 S2 50
22 40 12 10 6 30
46 27 17
15 .
1 ,000 2 .000
S4 00 S2 33
2 00
5 '.7 3 27
20 1'
4.000
51 33
1 25
2 13
30 3
?2 33
1 83
2 45
II
4.000
SI 33
1 17
1 60
10
2 ,000 - .010
51 ;n _-' no
1 65 °»
1 <* 1 :=
11 "
A T KEARNEY 6t COMPANY l-it-
A T KFM5VFY «c rOMP*VV
-------
VIII - 36
VIII - 37
That portion of total operating costs accounted for
by emission control ranges from 0.1% to 10.6% depending on the
alternative condition. More specifically, these ratios vary by
the conditions established as shown in the table below.
Percent of Emission Control Cost
to Total Operating Costs
Alternate
Number
I
2
3
4
Melt Rate
Lowest Tons/Hour
I
1
I
0
2
.si so
.61 SO
.81 50
.11 50
. Wet scrubber
Operating
Hours
4,000
4,000
4,000
2,000
4.000
costs and
Highest
7.97.
3.9%
10.6%
0.37.
Melt Rate Operating
Tons/Hour Hours
5
5
5
5
efficiencies.
soo
2,000
500
500
To properly
evaluate wet scrubber performance for this investigation, It is
necessary to determine the relationship between costs, efficien-
cies and pressure drops. As mentioned previously, the Los
Angles County Air Pollution Code formed the basis for wet scrubber
requirements.
Exhibit VIII-16 indicates theoretical wet scrubber particu-
late removal efficiencies and corresponding pressure drops for a
typical particle size distribution. Based on these figures, the
efficiencies developed are presented below:
Melt Rate Tons/Hour 5 15 30
Pressure Drop
Particulate Removal
Efficiency
30
96.5%
40
98.9%
45
99.5%
60
99.8%
An increase in the pressure drop for each melt rate
would result in an increase in removal efficiency. This con-
dition would also increase wet scrubber operating cost as shown
in Exhibit VIII-22. In developing Exhibit VIII-22, a pressure
drop of 30 inches was selected as an index. Curves are given
for each of the four operating levels; i.e., 500, 1,000, 2,000
and 4,000 hours per year.
As shown in Exhibit VIII-22, a pressure drop increase
from 30 to 40 inches would result in an increase in the wet
scrubber annual costs index from 1.0 to 1.05 for the 2,000-hour-
per-year operating condition. This represents a cost increase
of 5%. Therefore, the 5-ton, 2,000-hour wet scrubber operating
cost previously shown would increase from $3.50 to $3.71 if a
40-inch, instead of a 30-inch, pressure drop scrubber was
selected.
Summary of Capital
and Operating Costs
A T.KEARNEV & COMPANY. INC
Exhibits VIII-33 through VIII-36 depict the relationships
of capital costs to operating costs per ton for the various
production and operating conditions investigated. These figures
were obtained from Exhibits VIII-31 and VIII-32.
BENEFITS OF EMISSION
CONTROL INSTALLATIONS
The major emphasis of this chapter has been directed to-
ward estimating the costs associated with purchasing, erecting,
and operating emission control equipment. In this section, the
direct and indirect benefits to the foundry for installing
emission control equipment are discussed.
A T KEARNEY & COMPANY Ivc
-------
VIII - 38
During che field interviews for obtaining data, foundries
were asked to identify any benefits they had experienced from
controlling air pollution. In a vast majority of the cases
the installations were new, and it was too early for the foundry
to ascertain if any benefits would accrue. Several of these
foundries expressed an anticipated savings in roof maintenance
due to less frequent cleaning.
Roof raaintenanace and gutter cleaning was the mosC fre-
quent benefit listed. The reduction in particulars emissions
also resulted in less frequent cleanup in other areas of the
plant. Maintenance and building deterioration were reduced,
resulting in less cleaning and painting. One foundry reduced
the cleanup required at the air intakes, heater coils, and
ductwork.
Many foundries expressed a direct savings in plant main-
tenance costs. The reported savings ranged from $600 to
$5,000 per year. A large foundry, producing 30,000 tons per
month, indicated a savings in equipment and building maintenance
as great as $200,000 per year. The manpower requirements for
maintenance were reduced. Where one man was required constantly
at one'tounary without control equipment, the installation of
pollution control equipment reduced the manpower needs for
pollution-related maintenance to one man every four weeks.
Other foundries indicated a general improvement in the in-
plant environment due to the elimination of cupola gases escap-
ing at the charge door and dispersing through the plant. The
VIII - 39
atmosphere on the charging floor has become clearer and the
generally cleaner foundry atmosphere resulted in better light
conditions. As a result of the generally healthier working con-
ditions, employee relations were improved. Several foundries
indicated that as a by-product of the cleaner foundry environ-
ment, employee complaints and personnel turnover have been
reduced. One large foundry reportedly experienced a savings
of as much as $200,000 per year due to fewer personnel problems.
Most foundries located close to residential and commercial
areas stated that with control equipment, atmospheric conditions
in the neighborhood improved to degrees that neighbors' com-
plaints virtually ceased.
Other benefits attributed to the installation of control
equipment were that potential hazardous areas of toxic gases
were eliminated and refractory replacement in cupola stacks were
reduced.
Although no sound economic use of the particulate matter
has been stated, many foundries use the by-product for land fill".
A T KEARNEY fie COMPANY INC
-------
VIII - 40
REFERENCES
1. County of Los Angeles, California. Air Pollution Con-
trol District, Rules and Regulations, Regulation IV, Prohibitions.
A T KEARNEY & COMPANY. INC
IX - TECHNICAL AND ECONOMIC ANALYSIS
OF POTENTIAL MODIFICATIONS TO
FOUNDRY PROCESSES AND EQUIPMENT
IMPROVEMENT OF
EMISSIONS CONTROL
CAPABILITY
Three Approaches are possible in the effort to improve
foundry emissions capability at a lower cost: the modification
of emissions producing equipment to reduce emissions, changes in
operating practice, and improvement of emissions control equip-
ment. Modification of emissions producing equipment often
requires a change in operating practice. Because of this
condition, potential changes in these two areas are discussed
together, and improvements of collecting equipment are discussed
separately.
A systems concept must be applied to the analysis, par-
ticularly with regard to costs. An increase in scrap metal cost,
for example, resulting from a change in quality specifications,
or special pretreatment, may have an overall cost benefit if
a less costly control device can be used to collect the emissions.
An overview of industry efforts Coward this basic problem
shows no major technological breakthrough, but rather an ongoing
effort to better existing equipment through improved design and
the use of new materials.
Some results of innovative thinking are apparent such as
the reverse draft, "upside-down cupola," and a cupola emissions
-------
IX - 2
cooling system using a counterflow of a granular material such
as sand. The successful operation of these particular systems
has not yet been proven, although a full-size test of the new
cupola design is now under way with published results expected
in the near future.
Due to the proprietary and often confidential nature of
research and development effort by equipment manufacturers,
detailed information on various projects under way at this time
cannot be discussed unless public disclosure has already been
made. The material in this section is therefore limited to
recent developments not yet fully accepted by the industry,
or to concepts advanced by industry representatives outside the
equipment manufacturing companies.
POTENTIAL MODIFICATIONS
TO FOUNDRY PRACTICES
AND EMISSIONS PRODUCING
EQUIPMENT
Of all foundry operations causing emissions to be released
to the atmosphere in some degree, the melting operation has re-
ceived the greatest attention from emissions control boards,
the general public, foundry operators, and equipment manufac-
turers.
The effort to improve collection capability at a lower
cost has been concentrated on the problem of the cupola as the
source of the greatest amount of particulate matter. The result
of these investigations in recent years has been a number of
modifications offering at least a partial solution to the prob-
lem.
r i. r « l> NT"
IX - 3
Other work has been done tn che areas of preheating,
magnesium treatment for producing ductile iron, pouring and
cooling, sand handling and coremaking.
Potential Modifications
to Cupolas and Cupola
Operations
Cupolas and cupola operations can be modified in an
attempt to achieve two goals-to decrease the cost of emissions
collection and to decrease the quantity of emissions requiring
collection. With collection equipment and operating costs
directly related to the amount of gas to be processed, a direct
approach to the cost problem is to decrease the gas volumes.
A regression analysis, described in Section VI, showed no
significant relationship between cupola design factors and
emission quantities. The analysis did, however, indicate that
emission would be lessened by modifications to operating
practices resulting in decreasing the coke charge and eliminating
contaminants from the charge material.
(a) Decrease Stack
Gas Volume
One method of decreasing the volume of cupola stack gas
is to limit air infiltration by reduction of the charging door
area. This requires a change in the charging equipment since a
large door opening is required for the entrance of drop bottom,
core bottom, and end or side dump charging buckets. In place
of buckets, the charge materials can be discharged into the CUpola
by chute or oscillating conveyor. Some concern over possible
-------
IX - 4
charge segregation was expressed when this modification was
first proposed, but evidence from several systems now in opera-
tion shows that the problem is not serious with the lighter
density charges in general use at this time. Cupolas up to 48
inches in diameter have been charged successfully by means of
a chute, and cupolas up to 84 inches in diameter by means of an
oscillating conveyor, using a standard skip hoist to raise the
charge materials from the yard level.
One foundry has recently installed a new water-cooled
cupola with the charging opening reduced from a normal 96
square feet in area to 16 square feet. Charging equipment
consists of a conventional skip hoist, and a 42 inch wide
vibrating feeder. The feeder is arranged so that it can be
moved forward into the charge door opening for charging and
retracted when the material is completely discharged into the
cupola. Reduction of the door opening has reduced the in-
spirated air 83% and total gas volume 59%.
A cost comparison of systems using conventional and con-
veyor charging methods was prepared to aid in final selection
of equipment. The analysis included equipment and operating
cost estimates for the two systems furnished with emission
control equipment to limit emissions to 0.15 and 0.05 grains per
standard cubic foot of exhaust gas. The comparison shows equip-
ment cost savings of approximately $80,000 using the small
charging door and conveyor charging. Annual operating cost
savings are about $7,500 for the 0.15 gr/scf system, and $18,500
A T KEARNEY & COMPANY I»c.
IX - 5
for the .05 gr/scf system using the small charging door opening.
The system is reported as working satisfactorily with no adverse
effects due to charge distribution or segregation.
Locating the gas take-off below the top of the charge
burden is a second method of reducing air infiltration and
thereby decreasing the volume of gas to be cleaned. This method
requires the use of an induced draft fan to substitute for the
stack effect and is used on cupolas with emission, control equip-
ment. Placing the take-off approximately 3 feet below the top
of the charge effectively limits air pulled down through the
burden to about 10% of the blast air volume, independent of the
charging door area.
There are two alternative methods of processing the gas
after removal from the cupola. In one, the gas passes through
a combustion chamber where secondary combustion occurs with the
assistance of a gas burner and with only sufficient air added
to support the combustion. Combustibles including carbon
monoxide are completely burned, after which the gas is cooled,
cleaned, and discharged to the atmosphere. For the other
alternative, the gas is immediately cooled, cleaned, and dis-
charged back into the cupola stack below the level of the
charging door, where it can be ignited by afterburners to remove
the carbon monoxide content, if desired. A portion of the
cleaned gas, which contains essentially no oxygen, is drawn into
the gas take-off and recycled through the flooded disc scrubber.
A certain hazard exists in cooling the gas after take-off prior
A T KFM3NFY & COMP'kVV 1 v«-
-------
IX - 6
to immediate ignition in a combustion chamber. The negative
pressure in the take-off enhances the possibility of inward air
leakage through any opening, and also of drawing air downward
through the burden, to form an explosive mixture of CO and 02.
Discharge of the cleaned and cooled gas into the cupola below
the charging door insures against inspiration of air through
the charging door, but extra care must be taken to insure that the
take-off ductwork is airtight when installed. A good preventive
maintenance program is desirable to insure that it remain so.
Reduction of charging door size with conveyor or chute
charging and below charge door gas take-off systems are effective
in reducing the quantity of infiltrated air. However, they lend
themselves more readily to new installations than to modifica-
tions of existing cupola melting systems. The expense of
relocating an existing skip loader or charging machine and
possibly altering or installing a charging floor for the feeder
usually offsets the operating cost savings. Only if the exist-
ing cupola is charged by crane or monorail is it probably
economically feasible. Installation of a below charge door
take-off on an existing cupola can also be too costly to result
in an early payback. Lack of space adjacent to the existing
cupola may not permit installation of a combustion chamber or
cooling device to quickly lower the gas temperature to a safe
level.
One manufacturer of emissions control equipment has re-
cently reported on a system with the gas take-off located opposite
« r « P v~ v •• r
IX - 7
the charging door opening. The total gas volume, including
inspirated air, is reportedly reduced about 507. by the special
design of the take-off. The infiltrated air sweeping across
the cupola stack from the door opening to the gas tak«--off acts
as an air curtain preventing escape of the off-gases out the
open charging door or vertically into the upper stack. This
system, if technically successful, has the advantage of re-
quiring relatively simple and inexpensive alterations to an
existing cupola. The gas take-off should be somewhat less
expensive at the charging door level than near the top of the
cupola since it will be smaller and require less supporting
structure than at the higher elevation. Furthermore, the re-
sulting gas volume requiring cleaning is independent of charge
door area and whether the charge door opening and cupola top
are open or closed.
A fourth technique for reducing air infiltration is pro-
viding doors for the charging inlet. This method has been in
use for some years with indifferent results, due to mechanical
and electrical problems with door operating controls, and with
warping and lining problems with the doors themselves. Severe
operating conditions have inhibited the development of reliable
mechanisms for opening and closing the doors at the desired
times. Furthermore, sudden increases in gas volumes when the
doors are opened either overload the collection system, or
require that it be designed for the full open door capacity.
The result of the first is puffs of incompletely cleaned gases
A T KEARNEY Sc COMPANY
-------
IX - 8
and of the second, no savings In Che system cost.
The problems related to charging doors have resulted in
the discontinuance of many installations after repeated damage
to the doors by charging buckets. The most successful and long-
lived have been those installations with the simplest of
mechanical controls. In order to permit sufficient air in-
filtration for secondary combustion, doors are either provided
with openings in the bottom half, or are made shorter than the
cupola door opening. Installation costs ranging from $1,500 to
$3,500 per cupola have been reported depending upon the cupola
size and type of operating mechanism and control.
(b) Decrease Coke
Charge
Reduction of coke in cupola charges offers some beneficial
results. The evaluation of cupola emissions, described in
Section VI, indicates that particulate emissions tend to decrease
as the amount of coke in the cupola charge decreases. In addi-
tion, the replacement of the coke with lover cost fuels results
in lower melting costs under certain conditions.
Foundry operators contacted in the course of the study
expressed considerable dissatisfaction with the increasing cost
of coke coupled with a reported decrease in quality. In several
cases a distinct credibility gap exists with regard to proximate
analyses submitted by vendors. An experienced cupola operator
can quickly judge the coke quality by the furnace operation and,
in many cases, his judgment varies substantially from the
A T KFM»S'FY ft r•0^1P^VV 1^*-
IX - 9
analysis submitted with each shipment.
Heating of the cupola blast air has been recognized for
many years as one method of increasing the melting rate with
the same coke charged or of maintaining the melting rate with a
smaller coke charge. Use of hot blast for only the latter
reason, however, is rarely practiced. The economic benefit
of increasing the melting rate with hot blast is too much
greater than the more moderate savings obtained by using less
coke at the same melting rate. Neither have installations of
air heating units been made for the specific purpose of reducing
emissions level. Only two hot blast cupolas listed in Exhibit
VI-11 are reported as having emissions below the median of
36.5 pounds per ton of melt. All unlined cupolas in the exhibit
use warm or hot blast to overcome the greater heat loss from the
unlined shell and to increase the melting rate.
Hot blast is universally achieved by means of a fuel fired
heat exchanger. Natural gas is readily available in many parts
of the country for about $0.06/therm (100,000 BTU) for the
medium-sized user. The coke replaced by the gas costs approxi-
mately $45.00/ton delivered, representing a cost of $0.155 per
100,000 BTU. The operating cost savings amounts to approximately
one dollar per million BTU of heat in the blast air.
Oxygen enrichment of the blast air also results in a
decrease in coke requirements by eliminating the 80% nitrogen
in air which would otherwise require heating to the iron melting
A T KFARS'FY & COMPANY I - r
-------
IX - 10
temperature in the cupola. Like blase heating, however, oxygen
enrichment is invariably used to increase melting rate by in-
creasing the amount of oxygen available to the coke. An
accompanying benefit of the 80% reduction in volume of the
oxygen compared to the air it replaces is that specific blast rate
as related to melting rate is reduced. The earlier emissions
analyses showed the correlation between specific blast rate and
emissions rate.
Enrichment of blast air with 4% oxygen has been reported
to re."."1'- in savings up to $1.50 per ton of mct^l melted by
permitting the use of less costly metallies in the charge and
by increasing the melting rate up to 25%. Test data quantifying
the reduction of particulate emissions resulting from oxygen
enrichment are not available. The facts suggest, however, that
the installation of such a system only for the purpose of re-
ducing emissions would be uneconomical, if the cupola were
operated at the normal melting rate and full advantage were taken
of decreased coke charge.
Natural gas injection, like oxygen enrichment, provides
a means by which cupola coke requirements can be reduced, melt-
ing rate increased, melting costs lowered, and a modest decrease
in particulate emissions realized. The major research effort
in this field has been briefly described in Section VI. Results
of tests 12 and 11 in the research program showed melting rate
increases of 37%, melting cost savings of 51.22 and $2.01 per
ton, and emissions rates down about 15% when 30% and 60% of the
A T hFARVFY
IX - 11
coke was replaced by natural gas. The reported cost savings
include only the following unit operating costs:
Coke $40.00/ton
Natural Gas 0.06/therm (100,000 BTU)
Carbon 0.04/pound
Silicon 0.159/pound
Amortization of the estimated $20,000-$35,000 equipment invest-
ment is not included. The discussion in Section VI emphasizes
the fact that emissions rates would probably be lowered more if
blast air had been reduced approximately in proportion to the
decrease in coke charge, but no test data are available for this
condition.
(c) Prepartion of
Charge Materials
The evaluation of cupola emissions in Section IV points
out that many of the emissions can derive from the presence of
undesirable materials in the charge. These materials include
coke and limestone dust, sand, nonferrous metals, and a variety
of combustibles such as oil, grease, and synthetics. These
materials can be more or less completely removed from the charge
by more effective preparation of the coke, limestone, and
metallic scrap, with a resultant decrease in cupola emissions.
Screening of limestone and coke before loading into the
charging bucket is an elementary and inexpensive step to prevent
breeze and dust from being charged into the cupola. A simple
grizzly bar screen for the mechanized foundry, or the use of a
fork rather than a shovel in the manual foundry, will suffice.
This precaution is commonly taken, but an occasional foundry,
-------
IX - 12
IX - 13
Large or small, is found where Che absence of screening is
obvious from Che puff of black dust from the top of the cupola
as each charge is dumped. The investment is extremely nominal
and there is almost no operating cost.
Removal of embedded molding and core sand from foundry
returns can be effected with a shotblast machine at an estimated
cost of $2.00 to $2.50 per ton. This is infrequently done,
however, due to its cost, or the lack of available equipment. A
five-minute blast cycle is sufficient to clean the returns.
Other scrap preparation is more often performed by scrap
dealers than foundry personnel due to the equipment cost and
labor requirements. More and more of today's scrap is auto-
motive, consisting of whole or broken engines or fragmented
bodies. Charging of whole engines into cupolas, as a significant
number of foundries do, results in emissions problems from the
engine oil, grease, nonferrous, and alloying materials from the
engine components. Whole engines are available from scrap dealers
at a reported cost of $24-530 per ton. Engines that have been
fragmented and stripped of the large majority of nonferrous
components reportedly sell for about $50 per ton. Some scrap
dealers are equipped co break, wash, and strip engines on con-
tinuous lines consisting of a breaker, detergent washing tank
and a supporting conveyor system including a picking or sorting
conveyor where nonferrous materials are removed. This operation.
which removes all oil and grease from the scrap, adds $3.50-$4.00
per ton to the cost of the material.
Cupola scrap as purchased from dealers contains a minimum
of nonferrous materials, in large part because these materials
command a higher price, and it is to the dealers' interest to
separate them out for sale for the larger amount. Nevertheless,
a certain amount of copper alloys and even more of the lower
value nonferrous metals find their way to the foundry scrapyard.
The large mechanized foundry is not generally able to sort out
undesirable scrap due to the volume and method of handling
employed. On the other hand, the operator of a smaller foundry,
particularly one with manual or semi-mechanized charge makeup,
can instruct scrapyard personnel to set aside questionable scrap
for little additional labor cost.
Another problem, gaining importance with the increased
use of automotive scrap, is the inclusion of synthetic fibrous
material in shredded steel scrap derived from auto bodies. This
material, if not consumed in the furnace or by the afterburners,
tends to clog the collection equipment. Except for the possi-
bility of this type of contaminant and its relatively low
density, shredded automotive scrap offers an advantage for
cupola use in that the analysis is nearly constant, with little
nonferrous metallic inclusions. Specifications for this
material should limit nonmetallic inclusions, by requiring
incineration or other suitable process.
A T
A T KF1RVEY
-------
IX - 14
IX - 15
(d) Cupola Design
Changes
While minor design changes of conventional cupolas have
been shown to have little effect on cupola emissions rates, one
newly developed cupola design is so revolutionary as to warrant
special mention.
The cupola is top charged through two charging bells
roughly similar to a blast furnace. Blast air, preheated to
approximately 1100° F, enters the cupola through tuyeres located
somewhat below the bottom bell, raising the temperature of the
coke to the ignition point. The metal charge is melted in this
oxidizing zone and the molten metal along with the products
of combustion flow downward through the coke bed in a reducing
atmosphere. A portion of the particulate matter in the gas is
absorbed by the coke and the molten metal and slag. At the
bottom of the cupola, the molten metal, slag and gas flow into
an enclosed forehearth where outside air is mixed with the gas
and secondary combustion is achieved. At this point, particulate
and gaseous combustibles are burned. At the same time, the
velocity of the gas stream is sharply decreased as it flows into
the forehearth, and additional particulate matter settles out on
the slag layer where it is absorbed. The hot gases pass through
a recuperative blast air heater and are exhausted to the atmos-
phere.
The design is currently being tested on a full scale in-
stallation. Preliminary test results show emissions levels
rell below 0.05 grains per standard cubic foot, although some
•perating problems are still being worked on. No installation
>r operating cost data have yet been developed.
(e) Change Melting
Method
The increase in the number of electric and reverberatory
citing furnaces in recent years, at a time when the number of
ron foundries have decreased dramatically, is evidence that
ne acceptable solution to the problem of the high cost of
missions control is to replace the cupola by a furnace that
reduces less emissions. Installation and melting costs as
hown in Section VIII are not always lower, but the total of
he advantages sometimes tips the scale in favor of cupola re-
lacement. Exhibits III-5 and III-6 show the population changes
n types of melting systems in recent years and projections for
he next ten years.
Growth patterns in the population of electric arc and in-
action furnaces have been well documented. In addition,
eported sales of reverberatory furnaces for iron melting are
eported to be increasing as the result of the recent clean
ir laws. Stack sampling test results for a reverberatory
jrnace without any emissions control equipment melting one ton
ar hour of gray iron show emissions of 0.82 pounds. This
nissions rate is well below current minimums and explains the
nterest of the smaller foundries in the use of reverberatory
jrnaces despite problems of superheating the molten iron. This
ype of furnace in capacities of 500, 1,000, 2,000 and 4,000 pounds
A T KFARVFY 8e rOMPXNY I •. f
-------
IX - 16
per hour are sold completely wired, piped and lined ready for use.
Installed costs including shipping charges, hood, stack, skip
loader, and reasonable connection costs are estimated to be as
follows:
500 pounds/hour - $15,000
1,000 pounds/hour - 18,000
2,000 pounds/hour - 22,500
4,000 pounds/hour - 28,000
Capacities given are for melting and superheating gray iron to
2750° F. Capacities for melting malleable iron to 2650° F are
double those shown. Manufacturers of reverberatory furnaces
estimate that sales of these units will total 50-75 in 1970,
and that at least this many will also be sold in 1971.
The reverberatory furnaces described above are all batch-
type melters. A continuous type of fuel fired furnace has been
designed and installed on an experimental basis in an operating
foundry. Test data were not available at the date of issue of
this report, but the designers were hopeful of achieving the
dual benefits of lower cost melting than is possible in the
cupola and reduced emissions generation.
Potential Modifications
to Other Melting Furnaces
It has been demonstrated that melting emissions from
induction and reverberatory furnaces are generally below
minimums established by the various control bodies. No modifica-
tions to these types of furnaces, for the purpose of reducing
emissions, have been reported. Electric melting is a batch
IX - 17
process requiring a forehearth or holding furnace to provide
the continuous supply of molten metal possible from a cupola.
One major manufacturer has reported the development of a con-
tinuous melting coreless induction furnace permitting a better
basis for comparison with cupola melting. The furnace is "U"-
shaped with a horizontal coil located in the center. Preheated
charge materials are placed in the charging leg and the molten
metal is discharged from the top of the other leg. No decrease
in melting emissions results from the new design.
Emissions -f-om electric arc furnaces tabulated in the
general foundry data range from 6 to 28 pounds per ton of metal
melted. The variation in rate is related to operating practice
rather than to details of furnace design. For this reason, no
investigations into possible design modifications to reduce
emissions are being carried out as far as is known. Clean,
uncontamlnated scrap will, however, tend to decrease emissions
as it does in the cupola. The use of an afterburner to burn
combustible materials is particularly beneficial, since fabric
filters are almost universally used with arc furnaces.
Modifications to
Magnesium Treatment
for Producing Ductile
Iron
Inoculation of molten iron with some form of magnesium
for producing ductile iron is inherently a violent process
that is enhanced by the need to provide a strong stirring action
for insuring complete mixing of the inoculant with the iron.
While the process is basically the same today as when it was
-------
IX - 18
first introduced to the industry, much progress has been made
in the development of different carriers for the magnesium and
methods of introduction of the inoculant into the metal bath
to slow the reaction, promote greater retention, and decrease
the resulting emissions. Despite these advances the problem
is severe, particularly when the process is performed in the
open foundry area. Attempts have been made to contain the
emissions in order to decrease environmental hazards as well as
to lower the cost of capture of the emissions. These have
included the use of pressure ladles, pressure chambers, ladle
covers with or without connections to an exhaust and collecting
system, and enclosed inoculation stations. The pressure
systems are relatively expensive and time-consuming. Compared
to less costly systems, they can usually be justified only when
the pressure is used to improve the effectiveness of the
nodularization. Ladle covers at specific inoculation stations
have a nominal cost compared to pressure systems. Inoculation
station enclosures can be installed for $2,500 or less, com-
plete with ductwork to an existing collector and prevent the
magnesium oxide and other fume from blanketing the whole
melting, pouring and mold making area.
Modifications to
Sand Handling
Systems
Foundry molding sand when moist does not produce emissions
while being conveyed from the mixer to the molding machines.
After pouring, however, the heat absorbed from the molten metal
evaporates a large part of the moisture, and sand fines become
A T KEM*SEY & COMP*XV I - «-
IX - 19
airborne at the conveyor transfer points. Conventional control
for this condition is to enclose these points with hoods and
connect all transfers to a dust collector. A new concept of
dust control, recently patented, blankets the return sand with
moist prepared sand, effectively limiting the dusting normally
encountered. The dust eliminating feature of the system is
actually an added benefit, since the concept was developed to
solve the problem of hot molding sand.
In operation, excess molding sand is mixed and circulated
through the system with the overflow from molding requirements
discharged onto the return sand conveyor immediately following
the shakeout. The ratio of sand prepared to metal poured is
in the range of 20 or 30 to 1. Several benefits result from
this excess of sand. The heat of the shakeout sand is absorbed
by the prepared sand that blankets it, resulting in a very
nominal total temperature rise in the total mass. Convection
cooling on the conveyor system dissipates most of the heat.
Because the proportion of shakeout sand to the total mass is
small, only small amounts of additives, such as bonding
materials and water, are required, and mulling time can be
sharply decreased. Additives are metered to the shakeout sand
immediately after the shakeout, and before the overflow prepared
sand is discharged on the conveyor. With moist sand forming the
top layer on the belt, no dusting occurs, and no dust collection
is required except at the shakeout. The decrease in dust
collection requirements not only results in lower investment
needs, but also lowers operating and maintenance costs, and
A T KFMJVFY Sc rOMP\VV I N «-
-------
IX - 20
air makeup and heating costs. Sand storage facilities can be
reduced by as much as 80% since the majority of the sand is
constantly recirculating and being cooled. Installations have
been cited where savings of 50% in investment for the sand
storage bin, sand coolers and aerators, and dust collectors
have been realized for this process compared to conventional
sand systems. Some additional capacity for millers and
conveying equipment may, however, be required which will tend
to offset some of the savings. A highly variable cost not
included in the above must also be considered. Due to the patents
outstanding for this process, royalty payments are required for
its installation and use. Molding sand dust control systems for
conventional installations located in existing foundries are often
extremely costly due to the necessity of fitting ductwork into
crowded areas and around obstructions. A cost estimate for an
"average" installation of this kind is not possible due to the
possible variations for each specific case.
POTENTIAL MODIFICATIONS
TO EMISSIONS CONTROL
EQUIPMENT
Improvements in emissions control equipment in the fore-
seeable future, with few exceptions, promise to be minor in
nature. These will be results of efforts by the manufacturers
of this equipment to improve quality of manufacturing, use of
better materials and redesign to raise efficiencies and lower
costs of their products. Three modifications have been
A..T KEARNEY & COMPANY. Ivc
IX - 21
reported that could be of significant importance: the develop-
ment of a high temperature metallic fabric for fabric filters,
the use of a granular filter bed, and a system of cooling
cupola emissions in a rain of sand particles followed by a
fabric filter. No economic analysis is possible for these de-
velopments since cost data are not yet available.
Patent applications have been reportedly filed for a
metallic fabric suitable for use as a dry filter system with a
maximum gas temperature approaching 2000° F. Few additional
details are available at this time. The reported expanded
temperature range would permit installation adjacent to the fur-
nace with little or no gas cooling by water sprays or a convec-
tion cooler. The system drastically reduces the danger of des-
truction by the heat or glowing particles carried over into the
filter. In addition, the metallic fabric would be relatively
immune from action of hydrofluoric acid resulting from fluorspar
additions to the charge. The system is presently in the experi-
mental and pilot model stage and has not been tested in a full
scale installation.
A foreign manufacturer of control equipment has developed a
control device using a granular material as a filter bed for
cleaning furnace gas. The filter bed material is reported to
be a nearly spherical natural material with a close diameter
tolerance that will clean furnace gas to 0.05 gr/scf. Some
systems are reportedly working satisfactorily in Europe. Sales
representation in the United States is now supposedly being
A T KEARVEV & COMPANY INC
-------
IX - 22
negotiated. Full disclosure of details is expected within a
feu months.
Another emissions control device utilizing granular material
like sand for cooling cupola gas has recently been announced in
this country. The use of sand size material as a heat transfer
medium is effective because of the extremely large surface area
per unit weight. This patented system uses a rainfall of sand
in an enclosed tower through which the cupola emissions rise.
The cooled gas is cleaned in a conventional fabric filter, and
the heated sand is discharged into a second tower in which
fresh air rising in counterflow cools the sand to its original
temperature. The cooling air also passes through a fabric
filter before being exhausted to the atmosphere. The cooled
sand is conveyed to the upper level for reuse. All equipment
utilized in the system is standard and has been proven in
operation for many years. While sand is abrasive and equipment
used for its transport requires regular maintenance, the prob-
lems are familiar to foundry personnel and require no special
maintenance skills.
IMPACT OF POTENTIAL
MODIFICATIONS TO
FOUNDRY EQUIPMENT
AND PRACTICES
Exhibit IX-1 tabulates ten potential modifications to
cupola melting practices or equipment resulting in either de-
creased cost of emissions collection or lower emissions rate.
Operating cost savings are significant for those items that reduce
the volumes of stack gas required to be cleaned. The first
A T hEARVEY & COMPANY I •>. c
IX - 23
modification, small charging door and conveyor or chute charging,
is reported by cupola manufacturers as gaining wide acceptance,
and the majority of recent orders specify this feature. It is
often, however, not an economical modification to an existing
installation. The second item in the exhibit also is effective
but usually it is too costly to adapt an existing cupola to be-
low charge door gas take-off. The "at charge door" take-off,
listed third, has not yet been proven from a technical stand-
point. Since it can easily be installed on an existing cupola,
its success means that any foundry using a cupola for melting
can take advantage of one of three economical methods to sub-
stantially decrease the cost of collecting cupola emissions.
The second and third categories of modifications listed in
Exhibit 1X-1 include methods to reduce cupola emissions up
to 25%. The modifications to reduce coke requirements offer
operating cost savings up to $2.00 per ton of melt but result
in only moderate decreases in emissions. Preparation of
charge materials to reduce or eliminate certain emissions com-
ponents results in additional operating costs up to $4.00 per
ton of melt. None of these modifications, however successful
in reducing specific emissions, will result in the ability to
use lower cost collection equipment, since they do not affect
the release of all fine metallic oxide particles. Thus,
despite the additional cost of up to $6 or $7 per ton, the
same high energy systems will be required.
It can therefore be said that it is less costly to provide
A T KEARNEY & COMPANY INC
-------
IX - 24
equipment to collect whatever emissions are produced in a rupol,-:
than to control the input of materials causing the emissions,
particularly when steps are taken to keep gas volumes to a
practicable minimum.
The other practical approach to the problem of the cost
and difficulty of collecting cupola emissions is to replace the
cupola with a different kind of melting furnace. Foundries
of medium size and larger, depending upon local cost and
availability of raw materials and fuel, often find electric
induction or arc furnaces economically advantageous. The small
foundry, and even some medium-sized foundries, for whom electric
melting investment requirements are too great, may well find
reverberatory furnaces a low cost answer to the problem. Cer-
tainly the entrepreneur with limited capital, wanting to enter
the industry, will find the reverberatory furnace with its low
first cost, and low emissions rate, the most practical approach
to the choice of iron melting equipment.
The excess prepared molding sand system is expected to have
relatively widespread effects on sand preparation in iron found-
ries since it seems to solve the ubiquitous problem of hot moid
ing sand. Other potential modifications in non-melting applica-
tions are not expected to result in rapid changes in normal oper?t.-
ing methods or emissions collection capability. Similarly no r.aji.L
technological breakthroughs in emissions control systems, other
than those already discussed, are known at this time.
A T KFARKEY
X - PROJECTION OF TRENDS
GROWTH OF THE
IRON FOUNDRY
INDUSTRY
The previous discussion in Section III, as well as
Exhibits III-l through III-6, presented the historical
development of the iron foundry industry, particularly during
the postwar period. The situation can be described as one
in which the small, individually owned entrepreneurships
which once characterized the industry, are rapidly ceasing
to exist, and are being replaced by larger, corporately owned
foundries, with an ever-increasing percentage of the latter
being captive foundries. The growth of iron castings output,
combined with the reduction in number of foundries by 50%
in a period of less than 25 years, has resulted in more than
doubling the average output per foundry.
It is particularly significant that the reduction in
number of iron foundries has been almost entirely among the
small-sized companies. While the medium- and large-sized
foundries have remained relatively constant in number, the
small foundries have declined to under half of their number
during the postwar period. The reasons for this are not
difficult to trace. The increasing costs of labor and the
greater difficulty in recruiting and holding skilled foundry
A T KEARNEY & COMPANY INC
-------
X - 2
labor have resulted in a. continuing increase in the degree of
mechanization in foundries. It was once possible to operate a
foundry on a very small investment in equipment, and to use
the same crew to perform all of the functions of melting,
molding, pouring, cleaning, etc. The modern foundry requires
a high investment cost, with labor being fixed in single area
jobs, necessitating relatively continuous operation of the
shop.
To all of these problems, the additional consideration of
environmental control is now being added. This generally
requires the installation of relati-ely high cost air, water,
noise, and other control systems, which further add to the
capital and operating costs without, by themselves, contri-
buting to production or earnings. This has caused many
marginal operators to go out of business, and can be expected
to continue to have this effect in the next decade. An
analysis of the iron foundry population in the state of
California during the postwar period has shown that the periods
when strict enforcement of air pollution codes were introduced
were also the periods of most rapid decline in the number of
iron foundries.
The growth of production in the iron foundry industry
has been projected by estimating the growth rates for each of
the principal segments identified previously, based on the
A T KEARVEY & COMPANY INC
X - 3
growth of the user industry identified with that segment. The
three were then combined to provide the input for the total
estimated growth. Total iron castings production, excluding
ingot molds made from direct blast furnace iron, has been
projected to be approximately 17 million tons per year by
1980, or an average of 27. increase per year.
Iron castings production growth, as measured by tons per
year of output, has been somewhat behind the trends of some
of the other Industries, and has not kept up with the trends
shown in the early 1950's. This has probably been caused,
in part, by competition from other products and materials, such
as steel weldments, nonferrous castings and nonmetallics. The
relatively moderate Increase in tonnage of output has also
been caused by the improvement in technology which has enabled
thinner wall iron castings to be produced, resulting in an
increase in the number of castings per ton of weight. On a
basis of quantity of castings, therefore, the growth of output
of the iron foundry industry has more than kept up with the
national gross product growth.
The number of iron foundries which will be in operation by
1980 has been projected to be approximately 1,100. Since the num-
ber of medium- and large-sized foundries has been projected to
increase slightly, the entire drop from the 1969 total of 1,630
foundries to 1,100 foundries is expected to take place among the
A T KEARWEY & COMP\VY I ..-r-
-------
X - 4
small-sized companies. The combination of increasing output and
reduced number of producers is expected to continue to increase
the average output per foundry to an estimated 16,500 tons per
year by 1980. This is almost four times the average production
immediately after the war.
EQUIPMENT TRENDS
The most significant changes in the types and uses of
equipment which are expected to take place in the iron
foundry are in the melting department. The cupola, which
once was almost the only source of molten iron, has been
rapidly declining in number of installations, with the number
in 1969 being only about 45% of those in existence immediately
after the war. This decline has been projected to continue
at an undiminished rate, with the number of cupolas in 1980
estimated to be about 1,300. Although part of this decline
has been traced directly to the reduction in the number of
small-sized foundries, it has also been related to several
other factors. The principal one has been the replacement of
the cupola by other methods of melting, electric arc,
electric induction, and fuel fired reverberatory furnaces.
An additional factor has been the replacement, in many foundries,
of two or more cupolas by a single, larger-sized unit.
In 1969, an estimated 85% of all iron melted in iron
foundries was performed in cupolas. As the replacement of
A.T KEARNEY & COM PAN Y. INC.
X - 5
cupolas by other forms of melting continues to accelerate,
the amount of iron melted in cupolas will decrease, with
the amount by 1980 projected to be as low as 50% of the
total. The electric methods of melting, both direct arc
and induction, will account for almost 50%, with the small
remainder being melted in fuel fired, reverberatory furnaces.
The electric arc furnace population has been projected
to be about 350 units by 1980. Electric induction furnaces
have been projected to total about 1,350 units both coreless
and channel types, by 1980. Reverberatory furnaces have been
estimated to increase to about 250 units by 1980. All of these
types of furnaces can be used for holding, duplexing and
superheating, as well as for melting of iron. Therefore, the
total of 1,950 furnace units other than cupolas is not to be
considered entirely as replacements for cupola melting. In
fact, probably as much as one-third of them will not be used
for melting, but will perform the other functions of holding,
duplexing and superheating.
Although the tonnage of malleable iron which is produced
has remained relatively constant, the number of reverberatory
air furnaces used for duplexing has been declining steadily,
as these types of fuel fired furnaces have been replaced by
electric furnaces, principally arc furnaces. On the other
hand, the tonnage of ductile iron has been increasing every
A T KEARNEY 8. COMPANY
-------
X - 6
year, and is expected Co be about double the 1969 tonnage by
1980. This will result in a corresponding increase in the
amount of magnesium treatment of iron which will be performed.
The trend toward mechanization of molding, pouring and
shakeout lines has been continuing at an increasing rate.
Most of the large- and medium-sized foundries employ some
degree of mechanized molding. With the development of depend-
able and economical automated molding systems for small-sized
molds, many of the small foundries have installed mechanized
lines. Similarly, there are an increasing number of foundries
in which continuous and automated sand handling and preparation
systems have been installed. There is every reason to believe
that within the next 10 years, the majority of iron foundries
will use some form of mechanized molding, pouring, shakeout
and sand preparation.
Coremaking has also been undergoing a technological
change, with the trend being away from oil-bonded, baked sand
cores, toward chemically bonded, thermally cured and air
cured cores. Core processing is becoming mechanized and
automated. Fuel fired core ovens are being replaced with
dielectric ovens, or with processes in which cores are cured
directly in the core boxes.
The entire area of handling and preparation of materials
used in the foundry is continually being reviewed with respect
A T hEARMEY & COM PAW Ivc
X - 7
to the effect on emissions production and economy of collection.
Iron foundries are being more selective in purchasing of scrap,
with more attention being given to elimination of combustible
materials from the scrap before purchase. Larger foundries
are installing scrap preparation facilities involving pre-
burning or cleaning to remove materials which will produce
emissions during melting. Screening of coke and stone, and
use of higher quality briquettes are recognized as a means of
eliminating fines which will tend to blow out of furnaces
during melting.
It is believed that the next generation will see an
increasing trend toward centralized scrap preparation yards
which will service many iron foundries in a given area. This
will help to provide cleaner scrap at lower costs.
EFFECTS ON
EMISSIONS
Many of the trends toward electric or fuel fired melting
furnaces, mechanized molding, and continuous sand preparation
systems will have important effects in reducing the quantity
of emissions produced, or in making the emissions easier or
less costly to collect. The decisions made in selecting new
equipment are, therefore, being made to an increasing degree
with reference to the effect on reduction of emissions, or
on economy of emissions collection.
A T KEARNEY 81 COMPANY INC
-------
X - 8
The use of a greater amount of pre-cleaned or burned
scrap will reduce the quantity of emissions from combustibles
now charged into furnaces along with the scrap. Screening
to remove fines will also help to accomplish this same goal.
The trend away from cupola melting and toward electric
or fuel fired melting furnaces will have a significant effect
in reducing emissions. The electric induction furnaces and
the fuel fired reverberatory furnaces produce a very small
quantity of emissions. The electric arc furnace emissions
are substantial in quantity, but are much tasier and more
economical to collect than are cupola emissions.
The mechanization of molding, pouring, shakeout and
sand preparation facilities has not reduced the quantities
of emissions which are produced, but has resulted in confining
their production to fixed locations. Thus, it is relatively
simple in many cases to construct enclosures which can be used
as a means of collection of emissions with a minimum of
infiltrated air. This not only reduces the size and cost of
the collection system, but makes it more effective, reducing
the amount of unconfined emissions which are released in the
foundry building.
A T KEARNEY & COMPANY INC
XI - DEFINITION OF THE IRON FOUNDRY
AIR POLLUTION PROBLEM
GENERAL
The widespread distribution of iron foundries and the
prominence of the cupola stack in most communities in which
an iron foundry is located have combined to label Che iron
foundry as a major source of air pollution. This opinion is
often strongly held by the downwind neighbors of a foundry
using an uncontrolled cupola for iron melting, and air pollu-
tion control bodies receiving complaints of foundry emissions.
It is estimated that 243,000 tons per year of particulate
matter are generated from iron foundry melting operations, of
which about 182,000 tons are emitted to the atmosphere, the re-
mainder being collected. In addition, 1,504,000 tons per year
of particulate matter are estimated from non-melting operations
of which about 77,000 tons are estimated to be emitted to the
atmosphere. Total particulate emissions released to the atmos-
phere are therefore estimated at 259,000 tons per year. Based
upon a 1968 inventory by NAPCA, of emissions from all sources, it
is estimated that the iron foundry contributes about 3.57. of all
particulate matter emitted by industrial sources, and about 0.9%
of all particulate matter emitted by all sources.
NATURE OF
FOUNDRY EMISSIONS
Particulate and gaseous emissions result from most iron
foundry operations, with the largest concentrations being
A T KEARVEY & COMPANY Isc
-------
XI - 2
emitted from cupola and direct electric arc furnace melting.
Cupola emissions, as described earlier, include particulate
matter such as coke particles and ash, metallic fume, smoke oil
vapor, and dust from sand and limestone. Gaseous emissions from
the cupola include carbon monoxide from incompletely burned coke,
carbon dioxide, water vapor, nitrogen, and small amounts of sul-
fur dioxide and fluorine compounds. The last two emissions
quickly hydrolize to form corrosive acids.
Electric arc furnace melting results in similar types of
missions except for those deriving from the combustion of coke,
and include metallic fume, oil vapor dust, and smoke and gase-
ous compounds from the combustible materials in the furnace
charge. Emissions from reverberatory and electric induction
furnaces are negligible, and control equipment is rarely
employed on this equipment.
Drying and preheating cause emissions similar to the
cupola except that the fuel most commonly used is natural
gas instead of coke.
Many operations, such as sand handling, raw material
storage and handling, charge makeup, molding and sand condition-
ing produce only mechanical dust emissions. This particulate
matter is neither corrosive nor toxic, as some components of
cupola emissions, nor an asphyxiant like carbon dioxide. A
degree of hazard does exist however.from breathing the dust-
laden air over extended periods of time. Similar mechanical
dusts result from several cleaning operations including
A T KEARN-EY & COMPANY Isc
XI - 3
abrasive cleaning and grinding. These metallic dusts present
quite the same kind of health hazards if completely uncontrolled.
The increased use in recent years of chemically bonded
core and molding sands has given rise to an additional type of
emission. Both the setting reactions and the combustion of the
bonding material when the mold is poured result in smoke and
gaseous emissions that can be unpleasant and mildly toxic.
A small percent of iron castings are painted before ship-
ment. The process produces gaseous emissions consisting of
thinners and, depending on the equipment used, paint spray.
Many foundries continue to use oil bonded cores requiring
oven baking for setting. Gaseous emissions are produced
consisting of the more volatile fractions of the core oil and
some particulate matter such as smoke.
The adverse nature of these various particulate and
gaseous emissions produced by foundry processes has long been
recognized. Furthermore, to the distinct credit of the indus-
try, much has been accomplished through the years to control
emissions and to improve both the foundry and atmospheric
environment. It is also important to note that efforts in
this direction were begun many years before air pollution
concerns achieved national prominence.
It has been stated earlier that control devices for cupolas
were in use as early as 1938, and efforts to control non-
melting processes date back at least to the 1920's. While
A T KEARS'EY & COMPANY INC
-------
XI - 4
early equipment was crude and only partly effective, industry
pressure has resulted in steady improvement in efficiency and
more widespread use of control equipment.
INVENTORY OF
FOUNDRY EMISSIONS
The analysis of cupola and electric furnace emissions and
• the factors affecting the rates of emissions showed that an
average of 20.8 pounds of particulate emissions are produced
per ton of metal melted in an iron foundry cupola, and that
an average of 13.8 pounds of parciculate emissions per ton of
netal melted result from direct electric arc furnace iron pro-
duction.
Exhibit XI-1 shows the total estimated particulate
emissions generated by melting operations in foundries using
cupolas and direct electric arc furnaces in 1969. The
exhibit shows total quantities for each of nine geographical
regions and the nationwide totals based on the molten iron
production for the year and the above emission rates. Based
on a survey of iron foundries, considering the number and
capacity of furnaces equipped with control systems, the
effectiveness of the control systems, and the number of un-
controlled furnaces, it is estimated that 75% of the particu-
late emissions generated are presently being released to the
atmosphere.
A T KEARNEY 8. COMPANY. Ivr
XI - 5
Exhibit XI-1 also shows estimated quantities of carbon
monoxide generated and emitted. The first estimate is based
on an average cupola operating with a 7/1 coke ratio, using
coke with a carbon content of 91%, and with 11.6% carbon
monoxide in the top gas. Under these conditions, 276 pounds
of carbon monoxide is generated per ton of metal melted.
The amount of carbon monoxide emitted to the atmosphere
is dependent on a number of factors including the temperature
of the top gas, the availability of infiltrated air to provide
oxygen for combustion, the completeness of combustion, and
the percent of the total time that burning of the carbon
monoxide occurs. With sufficient oxygen from the infiltrated
air and with constant combustion, the carbon monoxide content
should be completely burned. Several factors tend to work
against this ideal condition, including the flame being
extinguished by each charge addition, lack of immediate re-
ignition either without an afterburner, or with an improperly
directed flame from an afterburner, varying carbon monoxide
content precluding constant combustion, and variable air supply.
A conservative estimate of 50% combustion efficiency has been
applied to the quantities of total carbon monoxide generated
to obtain the estimated weight of this gas emitted into the
atmosphere.
A T KEARVEY & COMPAVY 1 *c
-------
XI - 6
The results of the calculations can be summarized as
follows for 1969 nationwide production levels:
Total castings produced
Total molten iron produced
Total particulate emissions
generated
Total carbon monoxide
generated
Total particulate emissions
emitted
Total carbon monoxide
emitted
16,614,000 Tons
24,367,000 Tons
243,000 Tons
2,924,000 Tons
182,000 Tons
1,462,000 Tons
Particulates small enough to remain suspended in air over
an extended period are defined as aerosols. The maximum diameter
of such particles h«« h^en variously identified as from 20 to 100
microns. Using 50 microns as a limiting diameter, the aerosols
resulting from iron melting operations amount to approximately
56% by weight of the total emissions generated. On this basis,
the suspendible particulate matter generated by melting opera-
tions in 1969 amounted to 136,000 tons, of which approximately
102,000 tons was emitted to the atmosphere. Particulates over
50 microns diameter, totaling 80,000 tons for 1969, being too
large to remain suspended, settled out in a short time depending
on meteorological conditions.
The above data are derived only from cupola and electric
arc furnace operation. Emissions from other melting equip-
ment including induction furnaces and reverberatory furnaces
are negligible, not only because of conditions inherent to
these types of furnaces but also because generally cleaner
scrap metal is used for furnace charges and a relatively
A T KEARNEY & COMPANY Ivc
XI - 7
small percentage of the total iron is melted in these furnaces.
Preheating of less clean scrap for charging into induction
furnaces will add significantly to the emissions inventory
only when the process is substantially more widely used than
it is now. At its present level of application, preheater
emissions are also negligible.
Emissions previously described from non-melting foundry
processes, with a single important exception, are often
controlled as a standard practice, generally affect only the
foundry environment, and are released to the atmosphere only
in minor quantities compared to cupola and electric arc
furnace emissions. The concentration of these emissions at
their source can be substantial as in the case of the shake-
out, abrasive cleaning, and griding, but the particles emitted
are often large with a relatively high settling rate. The
portion of the particulate matter escaping the normal collec-
tion ductwork tends to settle out within the foundry building.
An analysis of non-melting operations indicates that 114.91
pounds of emissions are generated for each ton of metal melted,
but that only 5.83 pounds of this total are released to the at-
mosphere. Normal collection practices and settling out within
the foundry building account for the difference between these
two quantities. Exhibit XI-2 shows the total estimated quantities
generated and emitted to the atmosphere for the same nine geo-
graphical regions tabulated in Exhibit XI-1. The total estimated
weight of non-melting suspendible emissions from non-melting
operations in 1969 is shown in the exhibit to be 76,600 tons.
A T KEAR^CY & COMPANY. Ivc
-------
XI - 8
The non-melting emissions posing the greatest current
problem are those resulting from coremaking. A minor problem
existed in the past when practically all cores were made from
oil sand. This type of core, however, is thermally cured in
a core oven, and the emissions are relatively easy co capture
from the core oven stack for afterburning. The use of organic
chemical bonding agents, that are becoming more and more widely
used, intensifies the problem since these produce emissions
extremely difficult to capture due to their method of appli-
cation.
Molds or cores made from air set sand are often set out
on foundry floor or on racks while the sand sets. The local
environment in this situation is extremely poor. Not only is
it Difficult to capture the emissions over a large floor area,
but the dilution of the gaseous emissions by the air makes
the resulting mixture difficult and expensive to burn in any
type of afterburner.
The situation in many foundries for thermally cured
chemical binders when making shell or hot box cores causes
similar problems for the local environment as well as after-
burning. The resulting odors can be detected beyond the
foundry property in many cases.
INDUSTRY
PROBLEMS
The air pollution laws established in recent years have
created many serious problems for the iron foundry industry
A T KEARNEY & COMPANY. IMC
XI - 9
beyond the determination of the proper emission control equip-
ment to insure compliance with local codes.
Code limitations on permissible emissions are anything
but static. More than a few foundries have invested substantial
amounts of capital in equipment to meet the first generation
codes, only to discover, a short time later, that restrictions
have been made more severe, requiring costly reworking of the
control equipment. Current codes do not limit gaseous emissions
but all foundry owners are fearful that st'fingent" limitations
on carbon monoxide, sulfur dioxide and other gases are forth-
coming in the near future. These conditions mean that the
equipment purchased now must not only insure compliance with
present laws but also with future laws not yet drafted, or
at the very least, must be flexible enough to permit upgrading
in the future.
The emission control system has been characterized as a
profit-sharing silent partner that adds nothing to the value
of the foundry product but continually demands its share of
the profit. This is an unwelcome situation at best for the
large financially secure production foundry. It can be
extremely serious for the typically small, privately owned
foundry with low profitability and with cash-flow problems.
For these reasons many foundries have been slow to react to
the need for emission controls. In some states, a minor form
of relief is granted by waiving sales taxes on purchases of
equipment for air and water pollution control. However, other
assistance and incentive programs in terms of tax relief,
A T KEARNEY & COMPANY INC
-------
XI - 10
XI - 11
faster write-offs, low Interest loans, assistance grants,
and similar approaches are necessary to promote rapid and
complete compliance without forcing the smaller firms out of
business. A national program applying uniform standards in
all areas that will aid all foundries In installing emission
control systems is required. The development of a suitable
program requires additional study.
State air pollution control laws vary widely in the methods
of determining permissible emissions from foundries. One of
the methods becoming more widely used is the process weighc
basis in which permissible emissions are related to Che total
weight of all materials except air Introduced into the furnace
for a given time span. Some codes differentiate between
existing furnaces and new furnaces, or between jobbing foundries
and production foundries, and specify different permissible
emission rates for each. Few of the codes, however, differen-
tiate between foundries that nay have the same process weight
per hour but operate widely different total hours a week. Thus
the small foundry melting only 10 hours a week requires the same
size and type of control system as another foundry with the same
size cupola, but is melting 80 hours a week. The financial im-
pact of such a relationship can be seen in the cost per ton curves
and summary of operating costs given in the exhibits for Section
VIII. The economic burden of a fabric filter for cupolas, with
identical process weights of IS ton per hour, for example, but
operating at 1,000 and 4,000 hours per year can be seen on Exhibit
VIII-32 to be $4 per ton and $1.33 per ton respectively.
It is evident that application of the same process weight
standard is not an equitable method and thnt consideration should
be given to the use of a longer time span than one hour. Addi-
tional study would be required to develop a more reasonable and
equitable basis for applying process weight standards.
Despite the rapid technological advances in emissions con-
trol techniques and equipment in the past few years, a gap still
exists between foundry needs and equipment to meet those needs
efficiently and at a cost the foundry industry can afford to pay.
With improvements in emission control capability desired
for each operation in all foundry departments, a basis for
establishing a degree of need and a priority ranking for
research and development projects is obvious. Page one of
Exhibit XI-3 presents a matrix listing the major emissions-
producing operations versus the important characteristics of
the emissions and the control equipment installation required
for suitable control. Page two of the exhibit shows the rating
code appropriate for each characteristic, and the priority
assigned to each of three ranges of the total rating for each
operation. The operations indicating the greatest need, and
therefore the highest priority, are cupola melting and core-
making. Research and development projects for improving
emission control capabilities in these areas are included in
recommendations developed in the following section.
A T HEABVEY 8t COMPANY IK c
A T KEA8SEY & COMPAVY INC
-------
XI - 12
REFERENCES
1. Nationwide Inventory of Air Pollutant Emissions. 1968.
U.S. Department of Health, Education, and Welfare, Public Health
Service, Environmental Health Service, Publication No. AP-73.
A T KEARNEY & COMPANY. INC
XII - OPPORTUNITIES FOR RESEARCH
AND DEVELOPMENT
NEEDS FOR RESEARCH
AND DEVELOPMENT
Although iron founding has been carried on for thousands
of years, the application of emission control techniques have
only been made in iron foundries for some 30 years. Neverthe-
less, in those few years, the techniques and equipment have
been improved to the point where it can now be reasonably
stated that means exist for emission control for practically
all iron foundry operations. However, in many cases, these
emission control methods or equipment are not ecoi'or.uc-1 tr
install and operate, making their use economically o.-C'.. jiL- ^.
particularly in small foundries. The needs for research
development for emission control in the iron found'-.' •••uustr>
are therefore largely economically oriented, ained • pt.-vi.di.ng
equipment and methods which can be used by all SL -.> .-:" found-
ries without forcing them out of business.
Fundamental Knowledge
Much of the presently used control equipsPC was adapted
from other industrial uses, or was developed with little or no
knowledge of the fundamental's of emission evoli c.ion and the
nature of the emissions. I-t is believed that more basic under-
standing of these phenomena will provide necessary input for
development of more efficient and economical means for control-
ling emissions.
A T KEARNEY & COMPANY. Ixc
-------
XII - 2
The results of the regression analysis which was conducted
as a part of this investigation clearly demonstrated the lack
of quantitative data which related emissions to cupola design
and operating features. Additionally, the data which were col-
lected from stack tests on cupolas also demonstrated the lack
of correlation which results when the testing is not done under
controlled conditions, in which all facts relating emissions
produced with operating factors and charge material variations
are not clearly recorded. The research programs which are dis-
cussed are therefore proposed to be performed by qualified
laboratories under controlled conditions. Among the questions
which must be answered are the following:
1. What are the basic relationships between melting
furnace design and operating features, and the production and
evolution of emissions?
2. What are the mechanisms by which very fine metal-
lic oxides are formed, and what variables affect their forma-
tion?
3. What causes iron oxide formation, and what varia-
bles affect rate and quantity of production'
Economic Factors
The economic factors involving emission control have been
shown to be the most important area for further research and
development, particularly if methods are to be developed which
can be applied to small foundries and at the same time permit
A T KEARNEY & COMPANY
XII - 3
them to survive. Research and development efforts are needed
to answer the following questions.
1. What can be done to reduce costs of collection
of fine particles?
2. What means can be developed to reduce costs of
installation and operation of dust collection systems?
3. What means can be developed to economically
utilize waste heat in cupola stack gases7
4. Can uses be found for collected waste materials
which will aid in improvement of economy of collection?
Materials
1. Can fabric materials be developed for fabric
filter collectors which will resist high temperatures and
corrosive gases, and at the same time not be excessively
costly?
2. Can core binder materials be developed which
will not evolve acrid and noxious fumes during curing?
Processes
What alternates can be developed for cupola melting
which will combine continuous operation with economy of
melting and low emissions production?
ONGOING RESEARCH AND
DEVELOPMENT PROJECTS
Various research and development projects, covering a
number of aspects of iron foundry facilities and operation,
have been publicized or are known to be in progress. Since
A, T KEARNEY & COMPANY INC
-------
XII - A
many companies do not publicize their research and development
activities, there are undoubtedly other projects which are now
in progress but which have not been made known to the industry.
Among the known projects, some of which were discussed in ear-
lier sections of this study, are the following.
Melting of Iron
1. A reverse draft, or what has been referred to as
an "upside-down cupola," has been developed by Mechanite Metals
Co. and is now in experimental operation at Combustion Engi-
neering Company's Mbnogahela foundry. The intent of this cu-
pola is to reduce emissions evolution by forcing the blast air
downward through the cupola burden.
2. A continuous, gas fired shaft furnace for melting
of iron has been developed by Battelle Memorial Institute and
is in experimental operation at the Cooper-Bessemer Company
foundry. The purpose of the furnace Is to provide economical,
low emission melting of Iron.
3. A continuous, induction heated tneitiag furnace
for iron has been developed by Ajax Magnethermic Corp. and
General Motors Corp., for the purpose of providing low emis-
sion, economical melting of iron.
4. A gas injection program on a cupola is being
conducted jointly by Battelle Memorial Institute and Campbell,
Wyant and Cannon at their foundry in Muskegon. The purpose is
to provide greater economy in melting, reduced coke consumption,
A T KEARNEY & COMPANY. INC
XII - 5
and possibly reduced emissions. Similar programs arc also
being conducted in other iron foundries in the United States
and in Europe.
Emissions Collection
1. A new high temperature fabric with a new fabric
filter collector design is being developed by a commercial
foundry.
2. A new type of sand bed dust collector is being
developed by the Lurgi Company in Germany.
3. A new type of dust collector utilizing a rain
of sand to collect dust and heat, and also to transfer the
heat to combustion blast air, has been developed by Butler
and Kutny.
4. A collection system utilizing the heat of com-
bustion of the cupola gases to preheat blast air has been
developed by Holley, Kenney & Schott and is in operation at
Majestic Iron Works foundry.
5. A new low cost method of wet collection of fine
particles is reported to be in development by National Dust
Collector Co.
6. A program of waste product utilization has been
conducted by Zoller Castings Company, in which collected cupo-
la dust has been used as a fertilizer.
A T KEARNEV Sc COMPANY. INC
-------
XII - 6
PROPOSED RESEARCH AND
DEVELOPMENT PROJECTS
Based on Che needs for research and development as pre-
viously described, the ongoing projects, and the gap which
exists between available technology and needs, a series of
potential research and development projects has been assembled
and is described in the following paragraphs. These projects
nave also been summarized in Exhibit XII-1. The projects have
been grouped into associated programs involving fundamental
research, materials and equipment development, economic im-
provements, and standards development. A pri.un.cy rating has
been applied to each project. However, because of the previ-
ously noted fact that control technology and equipment actually
are available for almost every iron foundry source of emissions,
although not always at economical conditions, none of the pro-
grams has been rated as highest or urgent priority.
Estimated costs have also been applied to each project.
In general, they were based on the work being performed by a
research institute, university, or industrial research depart-
ment. The actual costs of research projects are difficult to
determine with any accuracy. The costs given are, therefore,
intended only as order-of-magnitude figures to enable an assess-
ment to be made of the relative value of benefit of each pro-
ject to the projected cost.
A T KEARNEY 8. COMPANY INC
XII - 7
Fundamental Research
Projects _
Project Nos. 1, 2 and 3 are a series of related programs,
involving research into the relationships which exist between
the quantity and type of emissions which are produced in the
cupola process, and the variables of design, operation and raw
materials used. These programs are proposed to be carried out
on an experimental cupola by a qualified research institute or
a university with an experimental foundry. They can also be
carried out in a commercial foundry if a cupola can be set
aside '-~r the work.
No. 1. Stack Sampling
The first project involves a controlled stack sampling
program on a cupola to determine the effect of cupola design,
operating and raw material variables on the type and quantity
of emissions which ore produced. The goal of the program will
be to provide quantitative information, which is not now avail-
able to enable designers of cupolas and emission controls as
well as foundry operators, to optimize design and operating
variables, to result in production of a minimum amount of
emissions and to make them easier to collect. The stack samp-
ling must be accomplished under controlled conditions in which
the various design, operating and raw materials factors can be
individually varied, and the effects on emissions noted. A
program of from six months to one year in duration is believed
to be required to accomplish the desired results. The cost of
A T KEARNEY * COMPANY, lie
-------
XII - 8
XII - 9
such a program, including a team of three or four investigators,
cupola equipment, test apparatus, and raw materials, is esti-
mated to be in the range of $150,000 to $200,000. A top level
priority has been assigned to this program with relationship to
the other recommended iron foundry programs, since it is funda-
mental in nature and needed to carry out other recommended work.
There is no known research which is being carried out in this
area.
No. 2. Iron Oxide Formation
.. Program
The second project is an extension of the stack sampling
program and probably should be carried out by the same re-
searchers. The purpose of the work will be to conduct re-
search which will determine the mechanism of iron oxide for-
mation in the cupola, and the effect of operating and raw
materials variables on the quantity of iron oxide which was
formed. Exhibit VI-14 suggested a possible relationship be-
tween the type of scrap used and the potential amount of iron
oxide which is present. This should be confirmed in the pro-
posed research program. The work has been estimated to be
able to be carried out in a period of four to six months by a
team of three researchers. The estimated cost of such a pro-
gram, including cost of the team, the experimental cupola,
test apparatus and raw materials, is in the range of $50,000
to $75,000. A medium level priority has been assigned to this
work, which should be valuable in potentially reducing the
amount of iron oxide present in cupola emissions. There is no
known research which is being carried out in this area.
No. 3. Fine Oxide-Opacity
Relationship Program
The third program is an extension of both the stack samp-
ling program and the research into iron oxide formation, and
also can be carried out by the same researchers. The purpose
of the work will be to establish a relationship between the
presence of fine metallic oxides in cupola stack gases and
the opacity of the gases. The results of this program can be
used to develop new opacity test standards which can more
effectively be used as a means of measurement of emissions
levels than the presently used Ringelmann chart. The time
required for this work has been estimated to be three to four
months, using a staff of two. The estimated cost for the work,
including the research team, raw materials and test apparatus
is in the range of $50,000 to $75,000. There is no known
research which is being carried out in this area.
The total fundamental research programs which have been
proposed can be combined into a single program to be conducted
by the same organization. If this is done, the work could be
accomplished in about a year with a team of about six researchers,
at a cost in the range of $200,000 to $300,000.
A T KEARNEY & COMPANY. INC
A T KEARNEY & COMPANY. Ivc
-------
XII - 10
Economic Research
Projects
We have previously noted that the principal area in which
new development work remains to be done involves the economic
effect of emission control on foundry costs. Rigid enforce-
ment of present control standards, combined with necessity of
using existing designs of equipment, would have Che effect of
forcing many of the smaller sized foundries out of business,
and seriously affecting the earnings of medium and even large-
sized foundries. The group of programs included in the economic
area includes a variety of approaches to the problem, involving
improved raw materials, waste heat and waste product utilization,
and reduced costs of collection.
No. 5 Waste Heat Utilization
The purpose of a development project for greater utiliza-
tion of waste heat in cupola stack gases is to improve the
economics of cupola melting and emission control. Two potential
sources of heat energy are proposed to be investigated. The
first covers the sensible heat in the hot cupola stack gases,
which is in the range of 180 to 320 BTU per pound of gas, while
the second covers the heat of combustion of the CO in the stack
gas, which is in the range of 4,350 to 8,700 BTU per pound of
gas. The work is proposed to be accomplished by a research
institute, or by a builder of melting or emission control
equipment. A program of about six months' duration involving
a staff of three researchers is proposed to investigate past
attempts at waste heat utilization, evaluate them, and develop
A T KEARSEV & COMPAVV Ivc
XII - 11
a new method which will combine sensible heat and heat of
combustion recovery. The estimated cost of such a program is
in the range of $100,000 to $150,000. The benefit will be
achieved by reduced coke consumption during melting, which
will also reduce costs of melting and reduce emissions levels
from the cupola. Although the waste heat recuperator, utilizing
the Griffin system, was commonly used in cupola installations
more than 20 years ago, the high costs of maintenance and low
reliability have made it obsolete. Known development programs
include a proposed system for cleaning gases and recuperation
of heat by use of a sand stream through which the gases pass1
and a combination gas scrubber and combustion unit for pre-
o
heating air.
No. 7. Centralized Scrap
Preparation Development
The high costs of scrap preparation, with particular ref-
erence to removal of combustibles and nonferrous metals, has
resulted in poorly prepared scrap being used in many foundries.
The purpose of the proposed development project is to enable a
centralized scrap preparation facility to be installed in a
community, serving numerous iron foundries and providing pro-
perly prepared scrap of minimum cost. This will promote use
of this scrap and will result in reduced emissions from melting
operations. This program can be carried out by an engineering
firm, or by a large scrap metal organization. A three co four
month study using a team of two engineers is proposed, at an
estimated cost of $30,000 to $50,000. A medium level priority
A T KEARNEY «c COMPAVY Ive
-------
XII - 12
has been assigned to this project. No known work is now being
carried out in this area. However, a study was made by the
U. S. Bureau of Mines on dismantling of junk automobiles to
3
produce quality scrap.
No. 10. Waste Product
Utilization
At present, the collected emissions from iron foundry
operations are not utilized for any purpose of value other
than for landfill and, in fact, are costly to remove from
foundry sites for disposition. The purpose of this proposed
project-is to develop potential uses for these waste products
which may help pay for the costs of collection and may even
provide some revenue. The development project would consist
of an investigation into past and present attemps to utilize
waste products from iron foundries and similar Industries, an
investigation of the analysis of typical waste products, and
the development of potential uses for these products. This
work is proposed to be carried out by a research institute, a
university, or an engineering company. A program of three to
four months' duration utilizing a staff of one or two investi-
gators is envisioned, at an estimated cost of $20,000 to $30,000.
Because of the marginal possibility of a positive result, a
low level of priority has been assigned to this project.
Development work in this area has been carried out by the
Zoller Castings Company, which has used collected emissions as
fertilizer on farmland, and by Swindell Dressier Co., which has
^
conducted a study on the use of steel plant wastes.
A T KEARNEY & COMPANY IHC
XII - 13
Materials Development
Projects
In certain areas involving evolution or collection of
emissions in the iron foundry, equipment exists for the required
purpose, but optimum materials have not yet been developed to
make use of known equipment or techniques. The projects in
this group are directed toward development of new or improved
naterials.
No. 4. High Temperature
Fabric
Fabric filter collectors are known to be the most efficient
n-eans of emissions control. However, their use in melting
operations has been limited, and their cost of operation has
been increased by the lack of availability of fabric materials
which will resist high temperatures over 500° F and at the same
time resist the corrosive action of stack gases which result
fron use of fluorspar in the cupola charge. A program of de-
velopment of temperature and corrosion-resistant fabrics, re-
lated if necessary to changed design of the baghouse collector
system, is proposed. Such a new material will result in an
increased.utilization and reduced cost of operating fabric
filter collectors. This work can be carrie'd out by a research
institute, or a manufacturer of fabric materials. The program
has been estimated to involve at least six months, using a
staff of three researchers, with appropriate test equipment.
The cost has been estimated to be in the range of $75,000 to
$100.000. Because of existence of several research projects
A T XtARNEY & COMPANY. Isc
-------
XII - 14
in this area, a medium level priority has been assigned. Known
development work is being carried out by a research company and
a. manufacturer on a new fabric material combined with a new
collector design. Additional work has been conducted by NAPCA
on endurance tests of high temperature fabrics,5 and endurance
tests on fiber glass fabrics.6
No. 8. Low Emission Core
Binder Materials
The preparation of oil bonded sand cores, which were
cured in an oven, results in oil fumes which are relatively
easy to capture, and which can be incinerated by use of a
catalytic combustion unit. However, the increasing use of
chemically bonded core materials which are cured in the molds
or on racks without use of an oven, produces fumes which are
difficult to capture, are acrid and often toxic in nature,
and which are difficult to convert into harmless, odorless
gases. The proposed research program involves the development
of new core binder materials which produce the desired physical
properties in cores and which do not give off noxious fumes
during curing. The goal of the work will be to eliminate the
fumes which are generally associated with coremaking operations
and with subsequent pouring of molds. This work is proposed
to be performed by a research institute, or by a manufacturer
of chemicals used in cores. A six-month program Involving
three researchers at an estimated cost of §100,000 to $150,000
has been envisaged. A medium priority level has been assigned
A T KEARNEY & COMPANY INC
XII - 15
to this work. Although no research programs in this area have
been reported, it is believed that some of the manufacturers
of core additive materials are engaged in such work.
Development of
New Equipment
The principal area of new equipment development is in
melting of iron, with specific reference to replacement of
the cupola by other melting methods which will still provide
some of the desirable features of the cupola, without some
of the undesirable features. The important other area for
new equipment development is in new dust collection systems
of lower cost and higher efficiency. Two such areas have
been selected for proposed research and development projects.
No. 6. Continuous Melting
Furnace
The principal advantages of the cupola for iron melting
are that it is a continuous nielter and an economical melter,
and it is capable of using a wide variety of metallic charge
materials. The principal disadvantage, which has been well
documented in this study, is that it produces a large quantity
of emissions which are difficult and costly to collect. The
alternate methods of melting which are now in use—electric
arc, induction and reverberatory are all batch-type melters
and, with the exception of the electric arc, are not capable
of using a wide range of metal scrap in the charge. The
proposed development project involves the design and construc-
tion of a continuous, fuel fired or electrically heated,
A T KEARVEY & COSTP-xv>- i.
-------
XII - 16
melting furnace, which will be economical to operate, and
which will produce a low level of emissions. It should also
be capable of being built and installed for a relatively low
costs. The benefits of such a furnace will be largely felt
by the smaller foundries which are now faced with the com-
bination of high cost melting and emission control installa-
tions. The development program can be best carried out by
a research institute, a furnace builder, or a large foundry
with a development department. A program of up to one year,
utilizing a staff of four people, at an estimated cost of
$200,000 to $300,000 is believed to be necessary. In view of
known research now being carried out in this field, a medium
level of priority has been assigned to the project. Two pro-
jects of this type are known to exist. The first involves a
joint development by Battelie Memorial Institute and Cooper
Bessemer Co., covering a gas fired, vertical shaft furnace.
The second involves General Motors Corp. and Ajax Magnethermic
Corp. and covers a continuous induction heated melting furnance.
No. 9. Agglomeration of
Fine Emissions
The cost of emission control, particularly on melting
operations, increases rapidly as the particle size to be
captured decreases. Submicron sized particles are very
costly to collect, requiring high pressure-drop wet scrubbers,
baghouses or electrostatic precipitators. However, if it
were possible to agglomerate fine particles into coarse
particles, efficiency of collection could be improved, and
A T KEARNEY St COMPANY. Im.
XII - 17
collection costs could be reduced. The proposed projects is
directed toward the development of a means of agglomeration
of Cine particles into coarse particles, using sonic or
electrical means for accomplishing this. This work can be
accomplished by a research institute, or by the development
laboratory of an emission control equipment builder. A
project involving five to six months of time, utilizing a
staff of three, at an estimated cost of $75,000 to $100,000
has been contemplated. A low level of priority has been
assigned to this work. No known research is being carried
out in this field.
A T KEARNEY 8, COMPANY I»c
-------
XII - 18
REFERENCE
1. Butler, Kucny, "A New Approach for Cupola Emission
Control," Modem Casting. June, 1970.
2. "Cupola Off-Gas Scrubbing with Recuperation", "33"
Magazine. 1969.
3. Dean, "Dismantling a Typical Junk Automobile to
Produce Quality Scrap," Bureau of Mines, No. 2350, December,
1969.
4. M. A. Osman, "Preparation of Useful Products from
Steel Plant Dusts," Swindell-Dressier Company, 1969.
5. J. M. Yacher, "High Temperature Fabric Filter Study,"
NAPGA, 1969-70.
6. Spaite, Harrington, "Endurance of Fiberglass Filter
Fabrics," Mapea, 1967.
7. AraaLa, Walker, "Continuous Induction Iron Melting,"
A.F S. Transactions, 1970.
(IBLIOORAPHIC DAT* <• »'»<" «•• (*•
fHMT APTO-OGad 1
Systems Analysis of Emissions and Emissions Control
Iron Foundry Industry Volume I - Text
In the
?. Auhorh)
». f»«i«va| O>B»iuini «•«* ""I Ad 22. PtictL .
P^raCL»SSiriED ^I?lt70
A T KEARNEY S COMPANY I-s c
-------
This report was furnished to the
A1r Pollution Control Office by
the A. T. Kearney In fulfillment
of Contract No. CPA 22-69-106.
------- |