A COST ANALYSIS OF AIR-POLLUTION
CONTROLS IN THE INTEGRATED
IRON AND STEEL INDUSTRY
Contract No. PH 22-68-65
to
DIVISION OF ECONOMIC EFFECTS RESEARCH
NATIONAL AIR POLLUTION CONTROL
ADMINISTRATION
DEPARTMENT OF HEALTH, EDUCATION,
AND WELFARE
May 15, 1969
(Complementary Final Technological
Report is also dated May 15, 1969.)
BATTELLE MEMORIAL INSTITUTE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
-------
A COST ANALYSIS OF AIR-POLLUTION CONTROLS
IN THE INTEGRATED IRON AND
STEEL INDUSTRY
(Contract No. PH 22 -68-65)
to
DIVISION OF ECONOMIC EFFECTS RESEARCH
NATIONAL AIR POLLUTION CONTROL ADMINISTRATION
DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE
May 15, 1 969
by
Thomas M. Barnes; Principal Investigator
H. W. Lownie, Jr.; Project Director
BATTELLE MEMORIAL INSTITUTE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
-------
Battelle Memorial Institlrte . COLUMBUS LABORATORIES
505 KING AVENUE COLUMBUS, OHIO 43201. AREA CODE 614, TELEPHONE 299-3151 . CABLE ADDRESS: BATMIN
May 29, 1969
Mr. Norman Plaks
Division of Process Control Engineering
National Air Pollution Control
Administration
57 lOW 00 s te r Pike
Cincinnati, Ohio 45527
Dr. Paul Kenline
Division of Economic Effects Research
National Air Pollution Control
Administration
1055 Laidlaw Avenue
Cincinnati, Ohio 45237
Mr. N. G. Edmisten
Division of Economic Effects Research
National Air Pollution Control
Administration
411 W. Chapel Hill Street
Durham, North Carolina 27701
Gentlemen:
A Cost Analysis of Air-Pollution Controls
in the Integrated Iron and
Steel Industry
(Contract No. PH 22-68-65)
Two copies of the subject report are being sent to both Mr. Plaks and
Mr. Edmis ten, and 96 cop ie s to Dr. Kenline.
For the companion Final Technological Report, two copies are being sent to
both Mr. Edmisten and Dr. Kenline, and 96 copies to Mr. Plaks.
Very truly yours,
~~\.
H. W. Lownie, Jr.
Project Director
HWL:jls
-------
-.----
TABLE OF CONTENTS
SECTION I
INTRODUCTION.
SCOPE AND OBJECTIVES OF THE PHASE I STUDIES
ORGANIZA TION AND CONTENTS OF THIS REPORT.
SECTION II
SUMMARY.
Section III: A Profile of the American Steel Industry.
Industry Resources for Growth and Change
Distribution of Costs in a Hypothetical Steelworks.
Section IV: The Technical Determinants of Air -Pollution
Control Cost.
Priorities, Problems, and Opportunities in Pollution-Control
Technology
Section V: Specific Cost/Effectiveness Investigations
Methodology
Results and Analysis
Section VI: Findings and Recommendations.
SECTION III
A PROFILE OF THE AMERICAN STEEL INDUSTRY.
INDUSTRY RESOURCES FOR GROWTH AND CHANGE
Operating Margins
Cash Generation.
Return on Equity.
Indebtedness
New Capital Spending
Overview
DISTRIBUTION OF COSTS IN A HYPOTHETICAL STEELWORKS.
Introduction
Costs: Raw Materials.
Costs Above Raw Materials
Total Annual Costs and Revenues
R.J:!,FERENCES.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
Page
I-I
I-I
1-3
II-I
II-2
II-2
II-3
II-4
II-5
II-6
II-6
II-6
II-7
III - 1
III - 6
III-7
III - 7
. III-IO
. III - 1 0
. III-IO
. III - 13
. III-14
. III-14
. III-14
. III - 1 7
. III - 1 7
. III- 24
-------
TABLE OF CONTENTS
(Continued)
SECTION IV
THE TECHNICAL DETERMINANTS OF AIR-POLLUTION CONTROL COST
Receipt, Storage, and Handling of Raw Materials.
Coking
Charging and Coking.
Pushing.
Coke Quenching
Coke Handling.
By-Product Processing.
Preprocessing Raw Materials
Sintering
Pelletizing.
Ironmaking .
Char ging
Smelting.
Casting and Flushing
Steelmaking
Open-Hearth Furnaces.
Basic Oxygen Furnaces.
Electric-Arc Furnaces.
Vacuum Degassing of Steel
Continuous Casting of Steel
Steel Shaping
Primary Breakdown .
Conditioning, Reheating, and Hot Rolling.
Acid Pickling.
Cold Rolling and Cold Forming
Steel Finishing
Miscellaneous Operations.
Gas Distribution.
Powerhouse
Plant Waste Incineration
Priorities, Problems, and Opportunities in Pollution-Control
Technology
Priorities
Problems
Opportunities.
SECTION V
SPECIFIC COST/EFFECTIVENESS INVESTIGATIONS
METHODOLOGY.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
Pal7p.
IV -1
IV -1
IV-Z
IV-Z
IV -5
IV-5
IV -5
IV -5
IV -6
IV -6
IV-6
IV-8
IV-8
. IV -1 0
. IV - 1 0
. IV - 11
. IV - 11
. IV - 14
. IV -14
. IV - 1 9
. IV-19
. IV-19
IV - 1 9
. IV - 2 3
. IV - 2 3
. IV - 2 3
. IV -24
IV -24
. IV -24
. IV - Z 5
. IV - 2 5
. IV - 2 6
. IV - 2 6
. IV - 2 7
. IV - 2 8
V-I
V-3
-------
TABLE OF CONTENTS
(Continued)
Page
Engineering Estimates.
Definition of the Form of the Model. .
Modeling, Pretesting, and Interpretation.
V-3
V-4
V -10
RESULTS AND ANALYSIS.
V-14
Coke -Oven Charging.
Coke-Oven End-Door Seals
Coke -Oven Pushing.
Quenching of Coke
By-Product Processing.
Coke -Oven Gas Systems
Ore and Flux Handling.
Sintering
Pelletizing.
Blast-Furnace Charging
Ironmaking Slag Disposal.
Steelmaking
Electric-Arc Steelmaking.
Basic Oxygen Steelmaking.
Open-Hearth Steelmaking.
Ste e Imaking: Ove rvie w .
Teeming of Molten Steel
Slab and Billet Conditioning
Hot-Rolling and Hot- Working of Steel.
V-14
V-14
V -15
V -15
V -15
V-15
V -16
V -16
V -18
V -18
V -19
V -19
V -19
V -21
V-23
V-23
V-25
V-25
v-26
SECTION VI
FINDINGS AND RECOMMENDATIONS.
VI-l
Criteria for Planning Future Phases
Organization and Objectives of Phases II, III, and IV .
Phase II: Study of Process Segments Now Largely Controlled.
Assembly of Data.
Data Processing.
Application of Cost Models
Phases III and IV: Study of Other Process Segments.
VI-2
VI-3
VI-4
VI-4
VI-6
VI-6
VI-7
APPENDIXES
A.
LIST OF STEELWORKS BY TYPES
A-I
B.
EXAMPLE SUMMARY SHEETS.
B-1
C.
COSTS AND PERFORMANCE OF CONTROL SYSTEMS AND
CONTROL EQUIPMENT
C-I
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
Figure III-l.
Figure III-2.
Figure 1II-3.
Figure 1II-4.
Figure 1II-5.
Figure 1II-6.
Figure III-7.
Figure III-8.
Figure IV - 1 .
Figure IV -2.
Figure IV -3.
Figure IV -4.
Figure IV -5.
Figure IV-6.
Figure IV -7.
Figure IV -8.
Figure IV - 9.
Figure IV -10.
LIST OF FIGURES
Geographic Distribution of Raw Steel Production in the
United States in 1967 .
Distribution of Steel Sales Among U. S. Companies, 1967.
Raw Steel Production and Finished Steel Shipments, 1958-67.
Operating Margins in the U. S. Steel Industry, 1958-67
Cash Flows Generated by the U. S. Steel Industry, 1958-67 .
Returns on Equity Investment in the U. S. Steel Industry.
Ratio of Long-Term Debt to Total Assets in the
U. S. Steel Industry, 1958-67 .
Flow Chart for 1,693,000 Net Tons of Products Per Year
for Hypothetical Steel Corporation's Plant in Ohio
Typical Flow Sheet for a By-Product Coke Plant.
By-Product Plant Flowsheet
Typical Flow Sheet for a Sintering Plant
Typical Blast-Furnace Operation With a Burden Consisting
Mainly of Sinter, and With Injection of Natural Gas.
Open-Hearth Furnace Operating With Hot-Metal Practice
Consisting of 60 Percent Hot Metal and 40 Percent Steel
Scrap (Ore Practice) .
Illustrative Flow Diagrams of Open-Hearth Dust-
Collecting Systems.
Basic Oxygen Furnace Operating With 70 Percent Hot
Metal and 30 Percent Steel Scrap.
Examples of Gas -Cleaning Systems for BOF
Steelmaking Furnaces.
Example of Electric-Furnace Steelmaking Practice.
Examples of Electric-Furnace Direct-Extraction
Emission-Control Systems With Baghouses
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
Page
III - 2
III-4
III - 5
III - 8
III - 9
III - 11
. III - 12
III - 1 5
IV-3
IV-4
IV-7
IV-9
IV -12
IV - 13
IV - 15
IV - 1 6
IV - 1 8
IV - 2 0
-------
Figure IV -11.
.t' igure IV -12.
Figure V-I.
Table I-I.
Table 1-2.
1 able III-I.
Table III-2.
Table III-3.
lable 1II-4.
Table III-5.
Table 1II-6.
Table 1II-7.
Table 1II-8.
Table III-9.
LIST OF FIGURES
(Continued)
Examples of Electric -Furnace Dust-Collecting Systems
Using Wet Scrubbers.
Examples of Electric-Furnace Shop-Roof Emission-
Control Systems
General View of a Curved Surface in Cartesian Space
LIST OF TABLES
The Principal Operations of the Integrated Iron and
Steel Industry
Economic Studies and Researches Included in Phase I Studies
of the Integrated Iron and Steel Industry.
Annual Capital Spending as a Percentage of Fixed Assets.
Prices and Amounts of Major Items Consumed in the
Hypothetical Steel Corporation I s Ohio Plant
Estimated Operating Cost for Making Hot Metal in
Hypothetical Steel Corporation's Ohio Blast Furnace at
a Rate of 1,200,000 Net Tons Per Year.
Estimated Operating Costs for Steelmaking in a BOF Shop
With an Annual Capacity of 1,500,000 Net Tons of
Molten Steel.
Estimated Operating Costs for Steelmaking in an Electric
Furnace Shop With an Annual Capacity of 500,000 Net
Tons of Molten Steel
Estimated Operating Cost for Continuous Casting of
1,440,000 Net Tons of Slabs and 480,000 Net Tons of
Billets.
Estimated Operating Cost for Making 167,400 Net Tons of
Wire Rod and 270,000 Net Tons of Merchant Bar From
Continuously Cast Billets
Estimated Operating Cost for Hot-Rolled Coils and Sheets.
Estimated Operating Cost for Cold-Rolled Coils and Sheet.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
Page
. IV -21
. IV - 2 2
V-s
1-2
1-4
. III-13
. III-16
. III-18
. III - 1 9
. III - 1 9
. 1II-20
. III-21
. III - 2 1
. III-22
-------
Table Ill-lO.
Table V-I.
Table V -2.
Table V -3.
Table V -4.
Table V -5.
Table V -6.
Table V -7.
Table VI-I.
LIST OF TABLES
(Continued)
Summary of Estimated Costs and Revenues for Hypothetical
Steel Corporation's Ohio Plant.
Air-Pollution Control Problems Considered in Phase I
Studie s .
Data, Modeling Results, and Pretests for Sintering Plants
Data and Modeling Results for Pelletizing Plants.
Data, Modeling Results, and Pretests for Electric-Arc
Furnaces.
Data, Modeling Results, and Pretests for Basix Oxygen
Furnaces.
Data, Modeling Results, and Pretests for Open-Hearth
Furnaces.
Data and Modeling Results for Scarfing Machines.
Process Segments of the Integrated Iron and Steel Industry,
Categorized According to the State of Progress in Air-
Pollution Control
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
Page
. III-23
V-II
V -17
V -18
V-20
V-22
V-24
V-26
VI-2
-------
z
o
~
u
::)
c
o
0::
~
Z
-
.
-------
I-I
SECTION I
INTRODUCTION
The National Air Pollution Control Administration (NAPCA) is an agency of the
U. S. Department of Health, Education and Welfare. As such, it is charged with cer-
tain research responsibilities under the terms of the Air Quality Act of 1967 (Public Law
90 -148). In fulfilling these responsibilities, NAPCA is sponsoring analytical studies of
air-pollution control in selected industries.
When completed, the industrial studies will describe
(1) The present status of air -pollution control technology
(2) The cost of controlling air-polluting emissions
(3) The needs for research and development toward better controls.
An important feature of the industrial studies is that the companies within each
studied industry have been requested by NAPCA to participate directly. It is recognized
that joint efforts, with data and guidance supplied by the industries themselves, will best
assure that the studies are responsive and authoritative.
NAPCA plans to use the results of the industry studies, togethe r with information
about other air -pollution sources, to project both regional and national summaries of the
status and cost of air -pollution control. The projections will be used to asses s the cos.t
effects of proposed or changing air-quality standards. At the same time, the technical
problems and priorities identified during the studies will be used to help plan and direct
appropriate research and development programs.
In accord with the overall objectives, this analytical study of the integrated iron
and steel industry was jointly sponsored by two divisions of NAPCA. This is an economic
(cost) report prepared for the Division of Economic Effects Research which is responsible
for determining the economic impact of air-quality standards upon industries and other
sources of air pollution. The other sponsoring group, which receives a complementary
technical report, is the Division of Process Control Engineering, which is responsible
for technical aspects of Federally sponsored research and development on the control of
air pollution from stationary sources.
SCOPE AND OBJECTIVES OF THE PHASE I STUDIES
The integrated iron and steel industry, defined for present purposes by the general-
ized list of operations in Table I-I, was selected for analysis because a number of its key
processes can contribute to air pollution. Within the steel industry, there are many chal-
lenging, often unique, technical problems in controlling or preventing emis sions. Metal-
lurgical fumes are characteristically finer in particle size, higher in opacity, and other-
wise more complex than other process emissions such as those from power generation.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
1-2
The economic cost analysis and planning of controls is also quite complex, mainly be-
cause of delicate competitive balances between process alternatives. Because the steel
industry is made up of physically large components, capital decisions inevitably commit
large sums of money; considerations of money already invested tend to retard change in
some instances.
TABLE I-I.
THE PRINCIPAL OPERATIONS OF THE INTEGRATED
IRON AND STEEL INDUSTRY
I.
Receipt, storage, and handling of raw materials
II.
Coking of coals and recovery of coal chemicals
III.
Pre-treatment of iron-bearing materials by agglomeration
IV.
Smelting of iron ores with coke to produce iron
V.
Conversion of iron into steel, and remelting of scrap
VI.
Processing of raw steel into saleable shapes
VII.
Finishing of steels; especially surface treatment
VITI.
Ancillary operations, e. g., steam and power generation
Note: Some Definitions Pertinent to Ste:l-Industry StUdies
Integrated Steelworks: A plant or group of adjacent plants which conducts a number of
the above operations including at least IV, V, and VI. Tonnagewise, most steel
is produced in these.
Secondary Steelworks: A plant which conducts Operations V and VI (at least), but which
does not conduct Operations II, III, and IV. Numerically, most steelworks are of
this type.
Steel Processing Plant: A plant which obtains raw or semifinished steel from other plants
and emphasizes Operations VI and/or VII.
Ironworks: A plant based on Operations I to IV; producer of pig iron, ferromanganese, or
other ferrous alloys.
Coke Plant: A plant based on Operation II, and which produces no iron or steel.
This study is identified as Phase I of a series.
It has the specific purpose of
(I) Analyzing present costs for air -pollution control, and determining
procedures for estimating total costs for controlling to such
standards of air quality as may become established, and
(2) Defining problem areas and needs, thereby providing one basis for
establishing priorities for fede rally sponsored research and de-
velopment projects.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
1-3
Purpose (1) corresponds to the goals of this project for the Division of Economic Effects
Research, and has been emphasized in this cost report. Table 1-2 lists the work of the
Phase I cost studies in general terms.
ORGANIZA TION AND CONTENTS OF THIS REPORT
In preparing a separate final report for each of the two sponsoring divisions of
NAPCA, it has been intended that each shall be a self-sufficient document oriented to
divisional purposes and needs. Accordingly, background technical information has been
included in this economic report, and background economic information has been
similarly incorporated into the technical report.
Section I is
to orient readers
expressed in this
the research that
this introduction to the project and to the economic report; it is prepared
inside and outside of NAPCA to NAPCA's overall purposes as they are
project. The tables in Section I define the area of study and describe
was performed.
Section II is a summary of significant results obtained from the Phase I cost/
effectiveness studies. It contains commentary on the evolution of steelmaking macro-
economics as affe cting the industry's ability to support new air -pollution control activities.
It reviews the areas of interaction between control costs and control technology and un-
derscores problems and marginal situations where an adequate response to the Air Quality
Act of 1967 must depend upon technical progress. The summary section also presents
the overall findings from the development of models and methodology for comprehensive
analysis. Throughout Section II, the emphasis is upon implications for the continuation
of the research.
Section III is a three-part presentation of the business economics of steelmaking,
and includes geographic and organizational profiles of the industry, a lO-year business
profile with emphasis on cash generated and returns obtained, and a presentation of the
approximate distribution of costs within a hypothetical Ohio steelworks. The purpose of
Section III is to orient economic analysis to the nature of the steel industry, its economic
size and strength, and its internal cost structures. Together, the presentations of
Section III form a body of necessary background to the more comprehensive analysis
planned for the future.
Section IV is a compilation of technical background information derived from
interim reports and the complementary final technological report. Section IV has two
purposes: one is to introduce the elements of steelmaking which are considered to be
sources of air pollution (and therefore objects of economic study), and the other is to
point out subjects of technical difficulty or inordinate control cost where permanent
solutions are still to be attained. The section concludes with an overall statement of
the steel industry's air-pollution priorities and problems, categorized according to
their relative basis in technology or economics. This analytical statement includes
some preliminary opinions about the likelihood of attaining solutions through applied
research.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
1--
TABLE 1-2.
1-4
ECONOMIC STUDIES AND RESEARCHES INCLUDED IN PHASE I STUDIES
OF THE INTEGRATED IRON AND STEEL INDUSTRY
I.
The integrated iron and steel industry was segmented and sub-
segmented into processes and operations suited to the selective
analysis of air-pollution problems and controls. This segmentation
is presented in Table V-I.
II.
Data were compiled from the literature, from industry sources both
by direct visit and by special solicitation, from the suppliers of
control apparatus, and from state governments. These data were
augmented by engineering estimate.s of control costs prepared by
the Swindell-Dressler Company(a) from information in their own
files or procured by them.
III.
Statistics characterizing the steel industry, as prepared by the
American Iron and Steel Institute and others, were studied and
analyzed to obtain suitable economic profiles and projections as
a background for future studies and as a basis for the interpretation
of Phase I data.
IV.
An expression for use in the modeling of control costs was devised
as an analytical tool in compiling and adjusting estimates of reg-
ional or national control costs on a process-by-process basis.
V.
By reference to all data available from all sources,
model expression was pre-tested and evaluated as
fifteen representative steelworks operations.
the proposed
applied to
VI.
The relative usefulness of the various data-gathering, data-generation,
and data-reduction techniques was critically reviewed and evaluated
as a prelude to subsequent study.
(a) Swindell-Dressler Company was a subcontractor to Battelle for some aspects of this Phase I study.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
1-5 and 1-6
Section V is the heart of this report. It presents the results of the specific
analyses performed in this project. The first part of Section V describes the purposes
and methods of the cost studies, and the second part gives the results for 15 specific
studies of representative steelmaking operations. For each operation studied, data
drawn from non-industry sources or estimated by the study subcontractor, Swindell-
Dressler, were modeled and compared with data obtained directly from within the steel
industry. The results of mathematical modeling and pretesting of the models are given
in all instances where data were sufficient to permit a meaningful test. Most results of
the special cost studies performed by Swindell-Dressler were incorporated into the
specific case studies of Section V.
Section VI presents Battelle's findings and recommendations and focuses upon the
experiences gained in obtaining, reducing, and analyzing data that characterize the cost
of air-pollution controls in steel plants. Section VI assesses the future application of
methods and models developed in Phase I, and describes the main limitations to be
observed as the studies are continued into characterization of other process segments
and broader representations of the steel industry.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
>-
~
~
~
~
::>
en
-
-------
I--
I
II-I
SECTION II
SUMMARY
The overall objectives of this study of the steel industry are to determine:
(1) The present status of air-pollution control technology
(2)
The cost of applying present technology to the lessening
of air pollution by steelworks
(3)
The subjects on which research and development are needed
to improve the technology and economics of air-pollution
control.
The study was conducted in 1968 and early 1969 by Battelle Memorial Institute,
Columbus Laboratories, with substantial subcontracted assistance from Swindell-
Dressler Company of Pittsburgh and voluntary assistance from a number of steel com-
panies that contributed information and guidance throughout the project.
This Final Economic Report is complete in this volume. It includes the report
proper, plus Appendixes A and B, and Appendix C is the final report of the Swindell-
Dressler Company. A complementary volume is entitled "Final Technological Report
on a Systems Analysis Study of the Integrated Iron and Steel Industry".
The partition of the final Battelle reports reflects the design of this study to serve
both the Division of Economic Effects Research and the Division of Process Control
Engineering of the National Air Pollution Control Administration.
In the economic (cost) aspects of the study, Battelle and Swindell-Dressler had the
following general objectives:
(I) To deve lop background depicting the
American stee 1 industry, as a basis
controlling air pollution
economic position of the
for study of the costs of
(2) To develop technical background indicative of the general prob-
lems and priorities of steelworks air-quality endeavors
(3)
To prepare for specific and comprehensive cost analyses by
developing methods for obtaining and analyzing data represen-
tative of the present and future costs of controlling steelworks
emissions to any given standard of effectiveness.
The third objective was emphasized. As Phase I of an expected series, this cost-
analysis study was to prepare for and lead into more detailed phases.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
II-2
This report contains four se ctions other than the Introduction and this Summary.
Three of them (Sections III, IV, and V) deal topically with the above general objectives.
The last (Section VI) offers Battelle's recommendations to NAPCA for the continuation
of economic studies into Phase II and beyond. The following paragraphs summarize
the findings and recommendations from each of the main sections of this report.
Section III: A Profile of the American Steel Industry
The United States continues as the world's largest steelmaker, despite recent in-
roads made by imported steel from Japan and Western Europe. Steelmaking is practiced
in many states (even Hawaii), but is heavily concentrated in the northeast central belt of
states from Illinois east to Pennsylvania. There are two principal types of steel plants:
integrated plants and secondary plants. Integrated plants manufacture iron from its
oxide ore, then convert the iron to steel in a separate operation. Secondary steelworks
smelt no ore, but produce steel by remelting scrap steel and repeating the refining
steps. In addition to the 49 primary (integrated) and 100 or more secondary steelworks,
the United States has a number of plants that perform smaller segments of the steel-
making job. These include coke plants, ironworks, and steel rolling and finishing
plants. At the present time, high capital costs for construction of all-new integrated
steel plants, together with very favorable trends in the cost of scrap, have encouraged
construction of new secondary steelworks. However, the major share of postwar
growth in overall steelmaking capacity has occurred through modernization and expan-
sion of existing integrated plants.
The economy of scale inherent in integrated steelmaking, and the advantages of
vertically integrating operations from mines to warehouses, have led to the formation
of some very large enterprises comparable to the big automakers and the larger oil
companies. But economies of distribution (especially for low-priced products such as
concrete-reinforcing bar, or for very specialized products such as tool steels) have also
led to the formation of some small enterprises on a profitable basis. The larger mills
are vulnerable to the ills of mass production. To fill modest orders, they must main-
tain large inventories which incur costs that erode the natural advantages of high-volume
production. Little steelmakers operate close to their markets in every sense of the
word, and compete vigorously and successfully with large integrated steel companies.
Industry Resources for Growth and Change
Financially, the steel industry (as a whole) has some problems. In a labor market
of constantly increasing cost per man-hour, steelmakers have spent large amounts of
capital to mechanize and automate operations, or to make them more productive. The
costs of purchased materials and services have gone up steadily, and have not been off-
set in full by rising prices. The paid-in revenue per ton of finished steel shipped has in
fact remained almost constant from 1958 to 1967. Operating margins have narrowed
generally, and have become extremely erratic since 1963. This, combined with rising
fixed charges for the new equipment, has squeezed profits to a level well below the level
typical for durables manufacturing. More and more of the cash flows generated by
steelmakers have been in the form of depreciation, depletion, and amortization
allowances.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
II-3
Low return on investors I equity, together with the erratic behavior of steel
earnings (intensified by the growing load of fixed charges), has made it difficult for the
steelmakers to issue securities on favorable terms. The indebtedness of steelmakers
has grown substantially, and there have been a number of mergers, absorptions, and
other upheavals in recent years. In 1969, an investor can still buy some steel stocks
well below their book values.
Since 1960, capital spending by steelmakers, especially for modernization, has
taken a sharp upturn. It now exceeds 20 percent of fixed assets annually. If these
capital expenditures yield the expected results, steelmakers may attain a turnaround in
their fiscal performance and regain some equilibrium. Otherwise, the general squeezes
and the erratic effects of leverage might continue or intensify. The biggest concern of
all is that imports will prevent full utilization of new and modernized facilities.
Distribution of Costs in a Hypothetical Steelworks
Hypothetical Steel Corporation, founded by Battelle (for the purposes of this study)
on the shores of Lake Erie, in Ohio, produces steel for the sole purpose of illustrating
where the steelmaking dollar is spent. The following tabulation summarizes this
information:
Cost per Year
Cost per Shipped Net Ton
Item
Iron ore, coal, limestone, and
finishing additions
Purchased steel scrap
$40,835,536
20,000,466
11,090,250
$24.10
11. 80
Other materials and fuels
6. 60
Labor, utilities, refractories,
purchased services, and other
conversion costs
Total
55,490,124 32.80
11,057,400 6.20
72,000,000 42.50
$210,473,776 $124.00
Sales and administration expense
Fixed charges on capital
The profitability of Hypothetical Steel Corporation is only 4 percent on sales (before
taxes), and 1. 8 percent (pretax) on investment. This is a result of high fixed charges
in the new plant. An older plant of the same type might incur only about $60 million in
fixed charges, and profits would be more than double those cited.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
II-4
Section IV: The Technical Determinants of
Air -Pollution Control Cost
Both particulate matter and undesirable gases are generated in the operations of
steel manufacture. Nearly all plant activities (except the actual generation of electricity
from steam) give rise to some form of air pollution.
Depending on the weather and the dryness of raw materials, the release of partic-
ulate dirt at hundreds of transfer points may be one of the biggest problems in a steel
plant. Immense amounts of ore, coal, limestone, and intermediates made from these
raw materials are unloaded, stocked, reclaimed, batched, blended, crushed, screened,
charged to processes, and otherwise handled on a round-the-clock basis. At each point
where the materials are in motion or exposed to drafts, a wisp of dust may arise -
trivial in itself, but contributing to a severe overall effect. Control over this system is
exercised by wetting, shrouding, exhausting through dust collectors, and numerous other
means; but existing materials-handling arrangements are often difficult to reshape for
modern dust-control standards. A particularly difficult problem arises in the charging
of coke ovens because the coal releases both dust and fumes.
Particulates also are released in the internal operations of smelting iron ore and
converting the re suIting iron to steel. Ore s abrade as they de scend in the stack of the
blast furnace, and the resultant dusts are carefully collected for recycling. But the
sintering operation required to re-agglomerate the dusts (and to utilize the finer ores)
is a center of dust raising. Sintering requires forced drafts, both for formation and for
cooling of the sinter, and the task of cleaning sintering exhausts is a major one. As
molten iron emerges from the blast furnace, it begins almost immediately to release
airborne graphite called kish. This kish forms and escapes continually whenever the
blast-furnace iron is handled in a molten state. In steelmaking, dust and dirt from
scrap is compounded with fumes from the oxidizing reactions that convert blast-furnace
iron to steel. Modern practices based on accelerating these reactions with injected
oxygen can generate a plume of red iron-oxide dust visible for many miles. The steel
companies have taken particular pains to contain, collect, and either process or dispose
of the dusts generated in blast furnaces, sinter plants, and steelmaking furnaces.
Other releases of particulates occur in the operations of quenching incandescent
coke, in the conditioning of slabs and billets with jets of oxygen, and in the high-speed
operations of the final stages of hot rolling. For all such instances, controls have been
established wherever the configuration of existing equipment facilitates a reasonably
practical approach. The oldest air-pollution controls are applied to cleaning blast-
furnace gas, which has value as cleaned. Newer control devices, at ever-higher cost
per unit of control achieved, have been installed either with new equipment or in antici-
pation of or in re sponse to abatement action .on the part of authoritie s. It is safe to say
that new steelworks apparatus is designed for and equipped with dust-collection
equipment as a matter of course. Coke plants are a thorny exception.
Gaseous emissions arise principally from the conversion of coal to coke, from the
reactions of ironmaking slag with the atmosphere, and from the combustion of sulfur-
bearing fuels. Remarkably, the biggest fuel-using device in any plant (the blast fur-
nace) cleans sulfur from its own waste gases and releases it through the action of water
and air upon the slag. The problems of sulfur dioxide from steaming boilers, from
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
II-5
heating furnaces, and from open flames are familiar, and steelmakers tend to seek clean
fuels to minimize these emissions. Not so easily controlled are the gases released by
the essential operation of coking coal. The gases and vapors around a coke plant include
coal smoke and tar fume s, both considered hazardous to humans. The fuel gas produced
as a by-product of the coking proce s s contains appreciable hydrogen sulfide which may
be stripped only at considerable expense. Steelmakers have made little progress in
coping with gaseous air pollutants despite ambitious, often aggressive research conducted
both here and abroad for 20 years or more. A great deal of additional technical progress
will be needed before present public expectations can be satisfied.
Priorities, Problems, and Opportunities in
Pollution-Control Technology
Steelmakers have a major overall problem of satisfying the public 1 s expectations
for air quality, and this problem fragments itself into dozens (perhaps hundreds) of sub-
problems within each plant. For this reason, air-pollution control is more difficult to
attain in steelmaking than in most other industrial operations. Priority is generally
given to the specific emission-control problems in the following process segments:
. Coking, especially the charging and pushing of coke ovens
. Steelmaking, especially during use of oxygen
. Sintering, especially highly fluxed practices
. Raw-materials transfers of all kinds
. Casting and handling of hot metal from the blast furnace
. Oxygen scarfing of semifinished steel
. Flushing and disposing of blast-furnace slag
. Utilization or disposal of coke -oven gas.
The above listing is not in any order of priority - each of these problems must be solved.
Contributors to the problems listed above include (l) difficulties in shrouding,
hooding, or otherwise accomplishing containment of the emissions; (2) maintenance of
the numerous seals and closures required in any containment system; (3) collection of
certain materials which are at once abrasive, corrosive, and not amenable to electro-
static precipitation; and (4) disposal of collected materials not suited to some form of
recycling. For a few kinds of emissions, notably sulfurous gases from slag, the
mechanisms causing the release of fume are not known. This lack of fundamental
knowledge also extends in part to generation of iron oxide fumes during steelmaking. It
is now generally recognized that metallurgical fumes are different and much harder to
collect than pollutants from many other industrial processes. For this reason, public
standards based on normal types of problems are frequently inappropriate in one or
more ways.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
II-6
The opportunities for technical gain include improvement of devices for the detec-
tion and measurement of air pollutants, but only if these devices are integrated with
public standards of adequate compliance. Other opportunities center on the revision or
even replacement of problem processes such as coking and sintering, to make them
more suited to air -pollution control. Some have suggested that coking as we know it
must be made obsolete, but they have not suggested how. Finally, there is an oppor-
tunity to unify collection systems, to improve the engineering design of the simpler (and
presently less effective) dust-collection devices, and thereby to improve the overall
cost and effectiveness of controls.
Section V: Specific Cost/Effectiveness Investigations
Battelle and Swindell-Dressler set out originally to collect and analyze large
amounts of cost and effectiveness data representing existing air-pollution controls in
steel plants. However, this procedure was retarded by a general shortage of published
and other public information on costs and effectiveness, and also by blocks to the trans-
fer of such information from steel companies to outsiders. In view of these blocks, a
second approach was taken, based on preparation of engineering estimates of control
costs by Swindell-Dressler Company. The estimates were used as the basis of the
Phase I work to develop and pretest econonllc models of control costs as a function of
process throughput and effectiveness attained.
Methodology
The Swindell-Dressler estimates were used together with inputs from the steel
industry. First, a generalized mathematical model was defined to represent the
probable variation of control costs with varying tonnage throughput and control effec-
tiveness. This model may take the form
Cost = A x (Throughput)b x (Effectiveness)g
where A is a constant of proportionality, and band g are exponents denoting the sensi-
tivity of costs to changes in throughput and effectiveness, respectively. Other factors
may be added as required to depict specific cost situations. It was shown that the above
model (as generalized) can represent any orderly, normal system in which either ris ing
throughput or rising effectiveness occasions higher costs. The general model was
solved by regression analysis of the Swindell-Dressler data for each of six steelmaking
situations, as scaled and expanded by Battelle.
Results and Analysis
The cost-effectiveness data obtained from the steel industry were applied to pre-
testing of the analysis model, but in most cases there were insufficient examples to
establish coherency. The pretest work established that general engineering estimates
and models derived from them can represent grouped industry data much better than they
can represent individual items within the group. Whereas the average absolute error of
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
II-7 and II-8
all pretests was about 23 percent, the aggregate overall error for all estimates summed
was only slightly over 9 percent when compared to cost and effectiveness data supplied
by steel companies for use in checking the models. In two cases, it was possible to per-
form analyses of limited data from industry sources and prove that they could yield
coherent cost models. In general, the generic model was verified as applied to real
data, and the efficacy of engineering estimation was verified as applied to a grouping of
individual cases.
Section VI: Findings and Recommendations
As a result of the experiences gained in collecting, processing, and analyzing
cost-effectiveness data, Battelle recommends that NAPCA's Division of Economic
.t..ffects Research proceed with a Phase II which should be based on a complete analysis
of costs for air -pollution control in steelworks process segments where controls of one
kind or another have been widely applied. This Phase II effort should be based mainly
upon industry data, and the use of engineering estimates should be minimized. To
facilitate the free exchange of information, a method is suggested whereby the steel
industry may participate without revealing actual data to NAPCA or to NAPCA's con-
tractor. This may be accomplished by processing data to generate the required cost
models under conditions of demonstrably adequate security. Once the data have been
processed, and the results examined for coherency, the original data may be discarded
or returned to their sources. The guiding principle is that once a coherent model has
been formed by regression analysis, the data entering the analysis have no value of
themselves and could not really be used in the rest of the cost-projection procedure.
Phases III and IV are outlined in brief. For Phase III, estimates of sensitivity to
process throughput and control effectiveness would be used along with both industrial
data and engineering cost estimates to prepare cost models for a group of steelworks
process segments not generally under good control at this time. In Phase IV, engineer-
ing estimation alone would be used with accrued experience to estimate order-of-
magnitude costs for the control of air pollution from process segments for which prac-
tical control methods do not yet exist. The information from Phase III would have a
practical value in national and regional projections; information from Phase IV would
measure mainly the degree of present impracticality, because all of the costs would
probably be inordinately high.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
-.
W
.J
LL
o
a::::
a..
>-
a::::
l-
II')
:::>
c
z
-
-------
III - 1
SECTION III
A PROFILE OF THE AMERICAN STEEL INDUSTRY
The United States produced about 127 million net tons of raw steel in 1967, more
than any other nation. The Soviet Union ranked second with 113 million tons, and Japan
was third with 68.5 million tons. These three nations, together with the countries of
Western Europe, account for over 83 percent of world steel production. (1)':'
On a per-capita basis, the U. S. average production of about 1300 pounds of raw
steel per person is consistently exceeded by West Germany, Luxembourg, Belgium,
Sweden, and Czechoslovakia. Japan exceeded U. S. per capita production for the first
time in 1967. (2) The United States imports steel heavily from these countries (except
Czechoslovakia), and also from Canada, France, Italy, and the United Kingdom. Total
imports of iron and steel exceeded 11. 4 million net tons in 1967 and 17.9 million tons in
1968. (3) Imports are becoming a sore subject to U. S. steelmakers, who have been
urging Congress to impose limits.
Inasmuch as manufacturing industries and the steel industry are interdependent,
steel production is heaviest in the main industrial states, and especially in the belt
from Western Illinois to Eastern Pennsylvania. Figure III-l illustrates the distribution
of raw steel production according to geographic divisions established by the American
Iron and Steel Institute. (4) Although steel production tends to concentrate near supplies
of coal and ore, sparse markets have retarded growth in the mining areas of Wisconsin,
Minnesota, and Michigan. Cheap lake transportation has permitted extensive steel-
making developments along the southern shores of Lake Michigan and Lake Erie. The
heaviest concentrations of steelmaking extend in strips from Bethlehem to Baltimore,
from Pittsburgh to Cleveland, and along the south shore of Lake Michigan. Other active
steelmaking centers include Detroit, Buffalo, and Birmingham. Appendix A, following
Section VI of this report, lists all steel plants in the United States (as of 1967) by owner-
ship, location, and general type. The number of principal furnaces is also shown as a
guide to plant magnitudes, but this number is not definitive because steelmaking furnaces
are built in a wide range of sizes. Plants performing only cold operations and heat-
treating have been excluded from the Appendix; their air -pollution problems contribute
less than the other plants to the problems of the industry.
Most steel in the United States is produced in 49 integrated steel plants';";' of the
type listed in Part I of Appendix A. Integrated plants are outnumbered, however,
by smaller secondary steelworks characterized by melting furnaces and small rolling
mills. At one time, the remelting of scrap steel was marginally economic except for
high-priced products or in the more remote areas of the nation. This is no longer true,
because the technology of secondary steelmaking has advanced, and the relative cost of
prime remelting stock has declined substantially. The transition may be illustrated by
a comparison of the critical prices:
"References for this section are given on page III-24.
""See definitions of plant types, footnote to Table 1-1.
8ATTELLE MEMORIAL INSTITUTE - COLUM8US LA80RATORIES
-------
ID
»
-i
-i
IT!
r
r
IT!
~
IT!
~
o
~
»
r
-
Z
(II
-i
-i
C
-i
IT!
I
o
o
r
c
~
m
c
(II
r
»
m
o
:u
»
-i
o
:u
IT!
(II
8 (Includes
Hawaii)
8
H
H
H
I
N
FIGURE III-I.
GEOGRAPHIC DISTRIBUTION OF RAW STEEL PRODUCTION IN THE UNITED STATES IN 1967
Diameters of circles correspond to relative output for states and regions shown.
Source: AISI Annual Statistical Report, 1967.
-------
Heavy-melting steel
composite (No.1)
(a) 2240 pounds.
scrap,
III - 3
Price per Long Ton(a)(5) Percent
1952 1967 Change
$ 8.45 $ 10. 70 +26
53.08 63.00 +19
64.20 96.40 +50
80.60 122.20 +51
41. 89 27. 63 -34
Mate rial
Iron ore, Mesabi Bessemer,
Lake Erie docks
Basic pig iron, Mahoning Valley
Steel billets, Pittsburgh
Hot-rolled strip, Pittsburgh
Some forecasters, including Battelle, have indicated that high-tonnage production
of steel remelted from scrap will grow sharply during the decade 1971-1980. (6,7) Rel-
ative costs of controlling air pollution for primary versus secondary steelmaking could
have an important influence on the se developments.
The plants not categorized as either integrated or secondary steelworks include a
few isolated coke plants (one is the world1s largest), some blast-furnace plants producing
pig iron and ferromanganese, and a number of plants which process raw or semifinished
steel to salable products. Most large plants of the latter type are captive to and coor-
dinated with large integrated steelworks. Captive steel-processing plants are favored as
a way of penetrating new market areas because freight rates for steel-in-process are
moderate compared with rates for finished steel. Independent hot-strip mills have all
but disappeared since 1960.
Present capital markets do not favor growth of small non-integrated steelmakers
to become integrated operations. The last all-new integrated steel company was Kaiser
Steel Corporation, formed in the 19401 s. There are, however, routes other than direct
growth which require less capital. For example, Acme Steel Company at Riverdale,
Illinois, experimented with the use of cupolas to provide hot metal for steelmaking, then
merged in 1964 with Interlake Iron to obtain hot metal from South Chicago - Interlake
had blast-furnace capacity to spare, and a rail connection between the plants was estab-
lished at modest cost. It remains to be seen whether infusions of capital from con-
glomerates can help the little steelmakers to grow; Northwestern Industries has ac-
quired Lone Star Steel, and Colt Industries now owns Crucible Steel Corporation of
America. On the West Coast, Oregon Steel Mills is undertaking to integrate its
electric-arc steelworks by addition of facilities for solid-state reduction of Peruvian
iron ores. This venture is technically unique in the U. S., and will be watched closely
by the secondary steelmaking industry. (8)
Figure III-2 characterizes the American steel industry by size of enterprise as
measured by 1967 sales and operating revenues. (9) In 1950 and later years, this charac-
terization has become blurred by extension of steelmakers into the manufacture of pre-
fabricated buildings and other steel-consuming businesses. (Figure III-2 excludes the
captive steelmaking operations of Ford Motor Company and International Harvester's
Wisconsin Steel Division; both would probably rank among the top twelve';'. )
"Both plants ha ve three blast furnaces.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
I7j
H
0
c::
:;c
M
H
H
H
I
N
m .
» en I\)
-i tJ
-i 0 ~
III ~
r >i 1-3 ~
n
r (1) :;c
III .. H
tP
3:: ~ c:: Z en
III 1-3 c
3:: H 3
0 0 0"
~ Z (1) CD
...
» O
r 0
I7j - 0
- en
z ("to en
1/1 (1) 1-3 b
j (1)
...... M N
-i ...
5' M IQ
C (1)
-i g. t" CII
III -
m en ~
I ("to >
0 >i t" (")
'< 0
0 M 3 ffi
r I7j en '0
C .... 0
3:: ::s > J
m ~ ~ (1) (ij
c CII
1/1 n 0
....
r III Z (")
...... 0 c ~
» 3
m ~ c::
0 c
. C
;II III I\)
» ...... en - I\)
'< (1)
-i m . 0.
0 .... ()
~ m ~
~ 0
III
1/1 .... ~
~ "0
'" I\)
-J ~ en
H
M
en
~
....
~
'"
-J
Cumulative Percent af 1967 Sales and Operating Revenue
00
I\)
o
<0
o
o
o
~
o
en
o
~
o
.."
o
CD
o
U1
o
o
--- I I I .I I I
--- Two companies have sales over $ 2600 million annually
-....~--- - -I -
Six companies have sales from $700 million to ij5-
~ $1800 million annually 2J
(1) -
'" CII
-
-"
3
~~t:-:-ompanie:- 0-
-
(1)
0.
5-
\ have sales fr:om $ 0.1 -
a
to $ 0.4 bi II ion 3"-
-
\ annually a.
c
CII
~ -
~ '<
CII
o
CD -
\ CJ)
..
~-
\ P5
<.JJ-
\ 0"
-
o' -
\- J
---
.\
H
H
H
I
~
-------
130
120
110
I/) 100
c
t2
-
Q)
Z
-
o
I/)
c
.Q
90
.-
~
80
70
60
III - 5
f;:ob
,...q;
.~
"!\.~
,>CJ
o~
~
o~
~
50
.65
66
67
63
64
62
Year
1958
60
61
59
FIGURE III-3. RAW STEEL PRODUCTION AND FINISHED STEEL
SHIPMENTS, 1958-67
Source: AISI Annual Statistical Report, 1967.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
III - 6
The overall growth of steelmaking in the United States is illustrated by Figure III-3;
the trend lines are a least-squares fit to the data for 10 years. (10) Production of raw
steel exceeds shipments by the amount of scrap generated in rolling and finishing steps.
Battelle has projected that production of raw steel will reach 157 million net tons annu-
ally by 1975, and that the three principal steelmaking processes will share the business
as follows(ll):
Steelmaking Process
Basic open-hearth (integrated)
Percent of Raw
Steel Production
1965 1975
Basic oxygen process (integrated)
Other pneumatic (integrated)
72.0 37.0
17. 6" 46.0
0.4 0.0
10.0 17.0
100. 0 100. 0
Electric arc furnace (mainly remelting)
Total
The trend in proces s selection is a clear one as of 1969.
is sensitive to changes in the proportion of imports.
The trend in tonnage produced
INDUSTRY RESOURCES FOR GROWTH AND CHANGE
The growth or revitalization of any independent enterprise depends upon its ability
to obtain funds, both from operations and in the capital markets. The resources avail-
able to the steel industry may be measured, and the measurement is appropriate be-
cause these resources govern the rate and extent to which the industry may respond to
demands for better air-pollution controls. The focus in this part of Section III is upon
sales, costs, taxes, earnings, dividends and growth of equity, and depreciation allow-
ances. The appropriate questions are as follows:
(l) How well has the steel industry maintained its operating margins
in the face of cost pressures?
(2) How much cash does the industry generate?
(3) How have equity investors fared in terms of capital gains and
distributions?
(4) What has been the trend in long-term indebtedness?
(5) How do present rates of capital expenditure compare with other
capital-intensive industries?
The major source for this part of the study is the American Iron and Steel Institute's
Annual Statistical Report for 1967, which presents financial data covering a 10-year pe-
riod and embracing about 94 percent of crude steel production. (12) (Much of the other
6 percent is produced by Ford Motor Co. and International Harvester. )
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
III - 7
Operating Margins
For this analysis, operating margin is defined as the net billing value of products
shipped - less labor, materials, freight, purchased services, and other direct costs.
T lobor applied to non-operating purposes, and income other than from operations, have
been omitted. Figures III-4a and 1II-4b portray total operating margins and margins
per net ton of finished steel shipped for each of the 10 years, 1958-1967. The trend
line in Figure III-4a is plotted on a least-squares basis.
From the trend of operating margins, it is clear that the price changes in the
steel market have not covered the increase in basic costs. In fact, operating revenues
per ton stee 1 shipped declined from a 1958 -60 average of about $209.30 to a 1965 -67
average of $208.00. ':' Perhaps equally revealing is the wide annual fluctuation of mar-
gins since 1960. Increasingly, steel-company financial performance has been influenced
toward erratic behavior by strike threats in both the steel and automotive industries.
When a steelworkers I contract is up for renewal, steel users acquire large hedge in-
ventories and then liquidate them (or buy steel abroad). When an auto strike looms,
automakers hedge by overproducing to keep the showrooms full, then slack off after the
settlement is reached.
Interestingly enough, the 1958-1960 3-year average of direct labor costs per ton
of finished steel was about $79.80 as compared with a 1965-1967 3-year average of
$78.55. The principal rise has been in costs for materials, supplies, and purchased
services, which rose by nearly $2 per ton of steel between the 1958-1960 and 1965-1967
averaged periods. (13)
The main conclusion from examination of operating margins is that they have be-
come both smaller and more erratic in the past decade.
Cash Generation
Net generation of cash is defined for this study as net after-tax income, plus al-
lowances for depreciation, depletion, and amortization. (It is assumed in the case of
non-linear depreciations that the tax allowances subtracted from pretax income include
a reserve for taxes effectively deferred by depreciation practices.) Figures III-5a and
1II-5b illustrate the total and unit generations of cash by the steel industry between 1958
and 1967. (14)
The cash picture is a healthy one. Cash flows in recent years seem to have been
more stable than sales and costs, and the trend is clearly upward on a per-ton basis. A
major contributor to favorable cash flows is the cut in taxes resulting in part from the
7 percent investment credit enacted in 1962. The total of Federal and state taxes aver-
aged for 1958-1960 was $16.20 per net ton of steel shipped, as compared with
$11.15 on average for 1965-1967. (15) Investment credits were worth about $1. 55 per ton
of shipments in 1967. (16)
"In part, this reflects shifts in the product mix, order quantities, and other factors unrelated to pricing.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
3200
3000
VI
~
.2
<5
CI 2800
'0
VI
.~ 2600
~
2400
2200
c:
t2 36
+-
Q)
Z
~
cf 34
VI
~
o
o 32
CI
a.
42
40
38
30
28
26
III- 8
1958
60
67
59
61
62
Year
63
64
65
66
Total Operating Margin = Sales Less Direct Labor and Materials Costs
1958
60
62
Year
67
63
64
65
66
59
61
b.
Operating Margin Per Net Ton Finished Steel Shipped
FIGURE III-4. OPERATING MARGINS IN THE U. S. STEEL INDUSTRY, 1958-67
Source: AISI Annual Statistical Report, 1967.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
2400
2200
en
...
o
8 2000
-
o
~ 1800
o
:2:
c:: 26
~
...
C1)
a.. 24
en
...
o
o
o
1600
1400
28
22
20
1958
59
III - 9
60
62 63
Year
66
67
64
65
61
a.
Total Cash Generated = Net Income Plus Allowances for Depreciation,
Depletion, and Amortization
1958
59
60
61
63
65
67
66
62
Year
64
b.
Cash Generated Per Net Ton of Finished Steel Shipped
FIGURE III-5. CASH FLOWS GENERATED BY THE U. S. STEEL
INDUSTRY, 1958-67
Source: AISI Annual Statistical Report, 1967.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
III - 1 0
Return on Equity
Annual gain or return on investors I equity may be defined as the sum of dividends
paid in cash plus net increases in reinvested earnings. This gain or return, studied
as a proportion of total stockholders I equity, indicates the standing of the steel industry
as an investment opportunity. Figure 111-6 illustrates how steel-industry investors
fared from 1958 to 1967. (17) As in the case of operating margins, the graph is char-
acterized by a downtrend and by erratic behavior in recent years. Accordingly, the
private investor sees steels as a "depressed" area of the industrial economy and as
a doubtful equity investment. In the 3 "wild" years, 1965-1967, average returns on
equity amounted to only 6. 15 percent, lower than is usually obtained on high-grade
bonds. Forbes ranked the steel industry 22nd of 23 industry groups for profitability as
of January, 1969. (18)
To counter this bleak picture, steel companies have made an effort to improve
their appearance to investors, and stocks of the eight largest companies closed in
January, 1969, about 30 percent above their 1968 low points. (19) The major change has
been to use straight-line depreciation for accounting purposes, although accelerated
methods are still used for tax calculations. Situations of risk have arisen because the
capital stock of a number of sound companies (such as Armco) have been selling below
book value. Less conservative accounting practices, stable output at healthy levels,
appropriate prices, continuation of the investment credit, and a successful repulsion of
imports all appear to be necessary if the industry is to gain and hold the investor con-
fidence required for equity financing.
In de bte dne s s
The ratio of long-term (funded) debt to total assets is a measure of the use of debt
by the steel industry in recent years. This ratio is shown in Figure 111-7 for the years
1958-1967. (20) The percent relative increase, which has been one of the strong trends,
merely confirms the above conunents about the industry's problems in attracting equity
financing. However, judging from the steadiness of cash flows, the steel industry does
not seem to be overextended in debt financing. It is noteworthy that most preferred
stocks outstanding at the beginning of the decade have now been retired. (21)
New Capital Spending
In the 3 -year period 1958-1960, average new capital spending by steelmakers was
about $18.50 per net ton of finished steel shipped, and exceeded the depreciation, deple-
tion, and amortization charges by about $8 per ton or 76 percent. In the 3 -year period
1965-1967, capital spending had risen to an average of $23.40 per net ton of shipments,
but depreciation charges also rose, and the margin of new spending over depreciation
and other writeoffs was down to 70 percent. Accordingly, apparent growth rates through
capital spending have not changed much. Some of the increase in total spending is trace-
able to a rise in the cost index for capital goods over the period. Most of the expended
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
12
>-
== "
::J
CT
Q)
.E
$2
'0
-
c: 9
Q)
~
Q)
a.
en 8
0'
.5
c:
...
0 7
w
-0
Q)
-
en
Q) 6
f:
Q)
a:::
c: 5
en
Q)
0' 4
c:
o
~
U
-0 3
c:
o
en
-0 2
c:
Q)
-0
>
0
0
III - 11
/
/
.. ~ l
~ ~ /
ITJe
\ 7 ~ ~~
\ / ................... 1000
I
I
I
...
y
10
1958
59
60
62
Year
63
65
67
66
64
61
FIGURE III-6. RETURNS ON EQUITY INVESTMENT IN THE U. S.
STEEL INDUSTRY
Source: AISI Annual Statistical Report, 1967.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
II)
-
Q)
II)
:r. .18
o
-
o
r-
.......
-
.c
Q)
CI .17
E
...
~
I
0'
c:
o
......J
; .16
-
o
a:
.20
.19
.15
.14
III - 12
1958
60
62 63
Year
67
64
65
66
59
61
FIGURE III-7. RATIO OF LONG-TERM DEBT TO TOTAL ASSETS IN THE
U. S. STEEL INDUSTRY, 1958-67
Source: AISI Annual Statistical Report, 1967.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
III - 1 3
money has been spent for renewal and modernization. Forbes ranked the steel industry
17th of 23 industries in rate of growth as of January, 1969. (22)
There has been considerable publicity about revitalization and modernization of
steelmaking facilities, and remarkable strides have been taken at some plants. In 1969,
capital spending by the steel industry will exceed $2300 million for the third year in
succession, or roughly 20 percent of total fixed assets in each year. Table III-l com-
pares capital spending for the iron and steel industry with capital spending in other
American industries, and shows that spending by steelmakers has consistently outpaced
both the other durables industries and also the comparable oil and chemicals
industries. (23,24)
..
TABLE III-I.
ANNUAL CAPITAL SPENDING AS A PERCENT AGE
OF FIXED ASSETS
Source:
Chemical Economics Handbook, Stanford Research Institute,
current Section 219.2430, April, 1968, and Quarterly
Financial Reports, Federal Trade Commission, Securities
and Exchange Commission, 4th Quarters 1959, 1960, 1966,
and 1 967.
Industry Group (U. S. A. )
Annual New Capital Spending,
percent of fixed assets
Average, Average,
1958-1960 1965-1967
Chemicals and Allied Industries
12.4 15.7(a)
9.7 10.0
11. 2 13.5
11. 7 14.8
11. 1 13.3
Iron and Steel Manufacturing
Petroleum and Allied Industries
Durable Goods Industries
All Manufacturing
(a) 1968-1969 proportions are even higher.
Overview
The American steel industry has been concerned about its own future for years,
with cause. In the 10 years 1958-1967, labor costs rose by 26 percent per man-hour,
and materials costs per ton of finished steel rose by almost 6-1/2 percent. Yet, total
revenues per ton of steel shipped did not change markedly. Note that this constant
revenue per net ton of steel shipped despite price increases, may be due to a change in
the product mix. This change, which resulted in a lesser proportion of high-priced
products being shipped may have been caused by the growing influx of steel imports.
During this 1 a-year period, the squeezes were met by outstanding technical performance
coupled with substantial new investments in mechanization, expansion of capacity, and
automation. The results are apparent in a steady growth in cash flows generated. Tax
relief related to investments has been instrumental in fostering the changes required.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
III - 14
. Nonetheless, earnings and return on equity have sagged badly, and have become
quite erratic as a result of uneven market conditions. Imports of steel, once trivial,
have become the principal competition to both coastal and inland producers. The in-
dustry has been responding with even heavier spending on new facilities and better equip-
ment. Private investors have been disinterested in steel's troubles, at least until
recently - a great share of steelmaking growth and change has been based upon long-
term debt. This can tend to further depress earnings on equity and to increase the
erratic effects of leverage.
Because high new capital costs for additional air-pollution controls are expected
to confront steelmakers within the next decade, some additional money must be found.
It must be expected that steel prices will adjust as required to meet costs of operation
and to pay for new hardware, on a basis that will leave the industry attractive to sources
of equity financing.
DISTRIBUTION OF COSTS IN A HYPOTHETICAL STEELWORKS
Introduction
Studies of costs for air-pollution control gain perspective if they may be referred
to other elements of the cost of producing finished steel from iron ore. To facilitate
NAPCA'S study of the operating costs for making steel in the United States, Hypothetical
Steel Corporation was "establishedll in Ohio, near Lake Erie, with an annual production
capacity of just under 2 million tons of raw steel. The capital investment for this all-
new plant was estimated at $600 million, or $300 per ton of annual capacity for raw
steel.
The Hypothetical Steel Corporation's Ohio plant produces 1. 5 million tons of
molten steel in a basic oxygen furnace (BOF) and 0.5 million tons of molten steel in the
electric arc steelmaking furnace (EF). A flowsheet showing the assumed product mix
is given in Figure III-8. The molten steel from the BOF is continuously cast into slabs
for the production of flat products. The molten steel from the electric furnace is con-
tinuously cast into billets for the production of rod and bar. The total annual production
of finished products was taken to be l, 693,000 net tons. Recycled steel scrap from the
casting and finishing operations was estimated to be 267,400 net tons.
Costs: Raw Materials
The costs of producing iron and steel are broken down into four major items:
(1) major raw materials costs, (2) costs above raw materials, (3) fixed charges, and
(4) sales and administration costs.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
BOF Steelmaking
m
>
-i
-i
111
r
r
111
~
111
~
o
~
>
r
Hot metal from blast furnace
Steel scrap
Finishing additions
Total
Molten steel produced
Products from continuous
casting machine
(Slabs)
z
I/)
-i
-i
c
-i
111
I
o
o
r
c
~
m
c
I/)
r
>
m
o
;U
>
-i
o
~
111
I/)
Steelmaking Metallics Flowchart
1,200, 000
494, 040
10,500
1,704,540
!
1,500,000
~
1,440, 000
I
1,368,000 HR(a) sheet
f
t
368,000 HR strip
+
349,600 HR sheet
for sale
t
322,000 CR coils
306,000 CR sheets
for sale
t
1,000,000 HR coil
I
970,000 HR pickled coil
t
950,600 CR(b) coil
t
950,500 annealed coil
t
922, 000 CR coil annealed
and tempered
,
t
600,000 annealed
coils for sale
EF Steelmaking
Steel scrap
Finishing additions
Total
Molten steel
produced
Products from continuous
casting machine
(Billets)
Rod for sale
Bar for sale
Total annual production of finished products for sale = 1,693,000 net tons
(a) HR = hot-rolled.
(b) CR = cold - rolled,
Recycled Scrap Scale and Loss
534, 134
~
537, 634
1
500,000
~
480,000 80,000
~
167, 400} 37,800 4,800
270, 000
43,200 28,800
......
......
......
I
.-
U1
18,400
24,000 6, 000 loss
19,400
100
28,500
16,000
26'i,400
39,600
FIGL:I'E III-S,
FLOW CHART FOR 1, 693, (JOO NET TONS OF PRODUCTS PER YEAR FOR HYPOTHETICAL STEEL CORPORATll)!\:'$ PLANT IN OHIO
-------
III - 1 6
Principal raw materials for ironmaking for the Hypothetical plant include taconite
pellets, coal, limestone, and oil for injection into the blast furnace. For steelmaking,
the principal raw materials are metallics (steel scrap, hot metal, and finishing alloy
additions). For casting, rolling, and finishing, the principal raw material is molten or
semi-finished steel generated within the plant.
The prices and amounts of some major items consumed in the Hypothetical Steel
Corporations's Ohio plant are given in Table III-2. These costs were not summed be-
cause the total would include duplications and would not be representative of the actual
cost. For example, all of the coal is converted to coke before consumption in the blast
furnace. Over 50 percent of the limestone is converted to burnt lime and then is con-
sumed in the steelmaking furnaces. The value of recycled scrap is established for
internal bookkeeping, and does not represent an expenditure for steelmaking materials.
TABLE IlI-2. PRICES AND AMOUNTS OF MAJOR ITEMS CONSUMED IN THE HYPOTHETICAL
STEEL CORPORA TION'S OHIO PLANT
Nominal Total Materials
Amount, net Unit Cost, Cost per Year,
Item tons per year dollars per net ton dollars
Taconite pellets 1,920,000 $ 13.64 $26,188,800
Coal(a) 1,036,800 10.02 10,388,736
Coke 720,000 15. 72(b) 11,318,400
LimestOne(c, d) 650,000 2.20 1,430,000
Finishing additions 14,000 202.00 avg. 2,828,000
Oil 48,000 17.10 8,208,000
Oxygen 125,000 12.00 1,500,000
Electrodes 2,375 582.00 1,382,250
Steel scrap 1,028,174
Purchased 760,474 $ 26.30(e) 20,000,466
Recycled 267,400 25. oo(f) 6,685,000
(a) Coal is used to produce coke for blast-furnace production of pig iron.
(b) Materials and operating cost only. Fixed charges on coke plant and auxiliaries add at least
$2 per ton of coke.
(c) limestOne is used as flux in the blast furnace. Limestone is also used to produce burnt lime
for the steelmaking furnaces.
(d) Burnt lime made from this limestone is valued at $16 per net ton.
(e) The price for purchased scrap is based on a mixture of grades that is considered suitable for
either BOF or EF steelmaking.
(f) The valuation of recycled scrap is for internal bookkeeping and may be assigned at any
reasonable level.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
III - 1 7
Costs Above Raw Materials
The cost above raw materials (sometimes called conversion cost) includes all
items other than raw materials, fixed charges, and sales and administration costs.
nature of most Ilcost-above'l items tabulated below is self-evident:
The
(1) Electric power and oxygen for steelmaking furnaces
(2) Graphite electrodes for electric furnace steelmaking
(3) Burnt lime for steelmaking furnaces
(4) Refractories for furnaces and ladles
(5) Repairs and maintenance
(6) Direct and indirect labor
(7) Utilities, including water and natural gas, plus electric
power used for purposes other than melting
(8) Miscellaneous supplies including tools, lubricant, and
chemicals for laboratory analyses.
(9) Yard switching and slag disposal
(10) General production overhead.
The items of cost-above for iron and steel production were determined item by
item, and are illustrated in Tables 1II-3 through III-5. For casting, rolling, and fin-
ishing, the cost-above has been shown as a lump sum, as in Tables 1II-6 through III-9.
Total Annual Costs and Revenues
The summary of estimated total annual costs and revenues for the Hypothetical
Steel Corporationls Ohio plant is given in Table III-lO. The total annual operating cost
as detailed in Tables III-3 through III-9 is $127.4 million, or about 58 percent of sales.
Fixed charges (based on 12 percent of the capital investment of $600 million':') total $72
million per year, or about 33 percent of sales. Sales and administration costs (based on
5 percent of the selling price of finished products) are $11 million. The total annual cost
for manufacturing and marketing finished steel is $210. 5 million. Income from sales is
$221.1 million; the profit prior to income taxes is about 4 percent of sales, and the return
on capital investment before income taxes is 1. 8 percent. These returns are low, pri-
marily because of high fixed charges in the all-new plant. Existing U. S. plants were
built for typically $250 or less per ton of annual capacity, whereas a new plant in 1969
typically could cost up to $350 or more per ton of annual capacity.
*Fixed charges include amortization, interest, local taxes, and insurance. Fixed charges on an annual basis were taken as a
minimum 12 percent of the total capital cost. A generally descriptive but flexible breakdown could be 5 percent for deprecia-
tion' 5.5 percent for interest, and 1. 5 percent for local taxes and insurance.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
III - 1 8
TABLE III-3. ESTIMATED OPERATING COST FOR MAKING HOT METAL IN
HYPOTHETICAL STEEL CORPORATION'S OHIO BLAST
FURNACE AT A RATE OF 1,200,000 NET TONS PER YEAR
Item
Unit Cost
Net Tons per Net
Ton Hot Metal
Operating Cost per
Net Ton Hot Metal
Materials
Taconite pellets
Coke (see below)
Limestone
Oil
Excess gas credit
Dust and sludge credit
Subtotal for materials
$13. 64/NT
15.72/NT
2.20/NT
17.10/NT
1. 60
0.60
0.26
0.04
$21.82
9.43
0.57
0.68
(0.71)
(0.06)
$31. 73
Cost Above
Labor
Utilities
Refractories
Distributed reline cost
Maintenance and repairs
Miscellaneous supplies
General overhead
Subtotal for cost above
Total operating cost (excluding fixed charges)
$5.00/man-hour
0.50
1.00
0.15
0.60
0.75
0.78
0.45
$ 4.23
$35.96
Estimated Cost of Producing One Net Ton of Coke
Item
Unit Cost
Net Tons per Net
Ton Hot Metal
Operating Cost per
Net Ton Hot Metal
Coal $10.02/NT 1. 44/NT
Credits for coke breeze, coke-oven gas, and coal chemicals
$14.43
(4.71)
9.72
6.00
Cost above
Total operating cost for coke
(excluding fixed charges)
$15.72
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
III - 1 9
TABLE IlI-4. ESTIMATED OPERATING COSTS FOR STEELMAKING IN A BOF SHOP WITH AN
ANNUAL CAPACITY OF I, 500, 000 NET TONS OF MOLTEN STEEL
Item
Operating Cost per Net
Ton of Molten Steel
Unit Cost
Requirements per
Net Ton Steel
Burnt lime
Refractories
Oxygen
Labor
Repairs and maintenance
Utilities
Yard switching and slag disposal
Miscellaneous supplies and service
General overhead
Subtotal for cost above
Total operating cost for molten BOF steel
(excluding fixed charges)
Metallics
Hot metal
Steel scr a p
Finishing additions
Subtotal for metallics
Cost Above
$ 35. 96/NT
26.30/NT
202. OO/NT average
0.791 NT
0.339
0.007
$28.41
8.92
1.41
$38.74
1. 04
1. 34
0.89
3.00
2.00
0.40
0.45
0.40
0.50
$10.02
$48.76
$ 16.00/NT
130 pound
12.00/NT
$ 5.00/man-hour
148 pound
0.60 man-hour
TABLE III-5. ESTIMATED OPERATING COSTS FOR STEELMAKING IN AN ELECTRIC FURNACE
SHOP WITH AN ANNUAL CAPACITY OF 500,000 NET TONS OF MOLTEN STEEL
Item
Operating Cost per Net
Ton of Steel
Unit Cost
Requirements per
Net Ton Steel
Electric power
Electrodes
Burnt lime
Refractories
Oxygen
Repairs and maintenance
Labor
Utilities
Yard switching and slag disposal
Miscellaneous
General overhead
Subtotal for cost above
Total operating cost for molten electric furnace steel
(excluding fixed charges)
Metallics
Steel sera p
Finishing addition
Subtotal for metallics
Cost Above
$ 26.30/NT
202. OO/NT average
1. 068 NT
O. 007 NT
$28.09
1.41
$29.50
3.84
2.76
0.64
1. 60
0.07
1. 80
4.00
0.45
0.40
0.35
0.50
$16.41
$45. 91
$ O. 008/kilowatt-hour
0.291/pound
16.00/NT
480.0 kilowatt-hour
9. 5 pound
O. 04/NT
12.00/NT -
11. 0 pound
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
5.00/man-hour
0.8 man-hour
-------
III - 20
TABLE III-6. ESTIMATED OPERATING COST FOR CONTINUOUS CASTING OF 1,440,000
NET TONS OF SLABS AND 480,000 NET TONS OF BILLETS
Item
Amount, net tons
per year
Unit Cost
Total Operating Cost
per Year'
Total Operating
Cost per Net Ton
Billets and Slabs
Major Materials
Molten steel 500, 000
Scrap (credit) (20, 000)
Subtotal major materials
Cost Above
Total operating cost per net ton of billets
(excluding fixed charges)
Major Materials
Molten steel 1,500,000
Scrap (credit) (60,000)
Subtotal major materials
Cost Above
Total operating cost per net ton of slabs
(excluding fixed charges)
Continuously Cast Billets
$ 45. 91/NT
(25.00)/NT
$22,955,000
(500,000)
$47.82
(1. 00)
$50.79
(1.00)
$46.82
4.50
$ 51. 32
$49.79
4.00
$ 53.79
Continuously Cast Slabs
$ 48. 76/NT
(25.00)/NT
$73,140,000
(1,500,000)
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
rn
III - 2 1
TABLE III-7. ESTIMATED OPERATING COST FOR MAKING 167,400 NET TONS OF WIRE ROD AND
270,000 NET TONS OF MERCHANT BAR FROM CONTINUOUSLY CAST BILLETS
Item
Amount, net tons
per year
Unit Cost
Total Operating Cost
per Year
Total Operating
Cost per Net Ton
of Wire Rod or Bar
Major Materials
Continuously
cast billets 180, 000
Scrap (credit) (10,800)
Subtotal major materials
Cost Above
Total operating cost per net ton of wire rod
(excluding fixed charges)
Major Materials
Continuously
cast billets 300,000
Sera p (credit) (27, 000)
Subtotal major materials
Cost Above
Total operating cost per net ton of merchant bar
(excluding fixed charges)
Wire Rod
$51. 32/NT
(25.00)/NT
$ 9,237,600
(270,000)
$55.18
(1. 61)
$53.57
14.75
$68.32
Merchant Bar
$51.32/NT
(25.00)
$57.04
(2. 50)
$15,396,000
(675,000)
$ 54. 54
12.25
$66.79
TABLE III-8. ESTIMATED OPERATING COST FOR HOT-ROLLED COILS AND SHEETS
Item
Amount, net tons
per year
Major Materials
Slabs
Scrap (credit)
Subtotal major materials
Cost Above
Total operating cost per net ton of hot-rolled coils
(excluding fixed charges)
1,440,000
(43,200)
Major Materials
368,000
(18,400)
Hot -rolled coils
Scrap (credit)
Subtotal major materials
Cost Above (for shearing)
Total operating cost per net ton of hot -rolled sheet
(excluding fixed charges)
Unit Cost
Total Operating Cost
per Year
Total Operating COSt per
Net Ton of Hot-Rolled
Coil and Sheet
Hot-Rolled Coils
$53.79
(25.00)
$77,457,600
(1,080,000)
$56.62
(0.79)
$55.83
7.20
$63. 03
Hot-Rolled Sheet
$63.03
(25.00)
$23,195,040
(460,000)
$66.35
(1.32)
$65.03
3.50
$68. 53
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
L
III - 22
T ABLE III -9. ESTlMA TED OPERATING COST FOR COLD -ROLLED COILS AND SHEET
- --------.
Item
Amount, net tons
per year
Unit Cost
Pickled Hot-Rolled Coils
Major Materials
Hot -rolled coils
Scrap (credit)
Subtotal major materials
Cost Above
Total operating cost per net ton of pickled coils
$63.03
(25.00)
1,000,000
(24,000)
Cold -Rolled Coils
Major Materials
Pickled coils
Scrap (credit)
Subtotal major materials
Cost Above (cold rolling)
Total operating cost per net ton of cold-rolled coils
970,000
(19,400)
$67.85
(25. 00)
Annealed Cold -Rolled Coils
Major Material
Cold -rolled coils
Cost Above (for annealing)
Total operating cost per net ton of annealed cold -rolled coils
950,600
$72.47
Total Operating Cost
per Year
$63,030,000
(600,000)
$65,814,500
(485,000)
$68,889,982
Major Materials
Annealed and Tempered Cold -Rolled Coils
Annealed cold -rolled
coils
Scrap (credit)
Subtotal major materials
Cost Above (tempering)
Total operating cost per net ton of annealed
and tempered coils
950,600
(28,600)
$74.97
(25.00 )
Cold -Rolled Sheet
Major Mater~als
Tempered cold -rolled
coils
Scrap (credit)
Subtotal major ma terials
Cost Above (shearing)
Total operating cost per net ton of cold-rolled sheet
322,000
(16,000)
$79.52
(25.00)
$71,266,482
(715,000)
$25,605,440
(400,000)
Total Operating
Cost per Net
Ton of Product
$64.97
(0.62)
$69.23
(0.51)
$64.35
3.50
$67. 85(a)
$72.47
2.50
$68.72
3.75
$72. 47(a)
$74.97(a)
$77.30
(0.78)
$83.68
(1. 31)
$76.52
3.00
$79.52(a)
$82.37
3.50
$85.87(a)
(a) Product cost does not include fixed charges.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
III-23
TABLE III-10. SUMMARY OF ESTIMA TED COSTS AND REVENUES FOR HYPOTHETICAL
STEEL CORPORATION'S OHIO PLANT
Operating Cost Sales Price, Revenue,
Products Quantity Per Net Ton Per Year dollars per net ton dollars per year
Wire rod 167,400 $68.32 $ 11,436,768 $130.00 21,762,000
Merchant bar 270,000 66.79 18,033,300 126.00 34,020,000
Hot -rolled sheets 349, 000 68.53 23,958,088 117.00 40,903,200
Cold -rolled sheets 306,000 85.87 26,276,220 144.00 44,064,000
Cold -rolled coils 600,000 79.52 47,712,000 134.00 80,400,000
1,693,000 $127,416,376 221,149,200
Overall Manufacturing and Sales Costs
Total operating costs
Fixed charges, 12 percent of
capital investment ($600 million)
$127,416,376
72,000,000
Sales and administration, 5 percent of
selling price of finished products
Estimated overall costs (excluding
Federal income taxes)
11,057,400
$210,473,776
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
III - 24
REFERENCES
(1) Stahl und Eisen, Heft 9, Jahrgang 88, May 2, 1968, seite 465.
( 2 ) Ibi d.
(3) American Metal Market, 76, No. 23, February 5, 1969, p 1.
(4)
American Iron and Steel Institute, Annual Statistical Report, 1967, New York, 1968,
page 68.
(5)
American Metal Market Company, Metal Statistics 1968, 61st ed., New York, 1968,
pp 175, 185, 285, 288, and 306.
(6) Sims, C. E., "What is Ahead in the Next 25 Years for Electric Furnace Steel-
making", Journal of Metals, 20, No.2, February, 1968, p 44.
(7)
Miller, J. R., "Prereduction and Steelmaking", Address presented to AIME
Pittsburgh Chapter, Process Metallurgy Section, February 6, 1969.
(8)
Ibi d.
(9)
IISteel Industry Financial Analysis", Iron Age, April, 1968.
(10) AISI, op. cit., P 8.
(11 )
Miller, op. cit.
(12) AISI, op. cit., pp 9, 10, 11.
(13) Ibid., P 10.
(14) Ibid.
(15) Ibid.
(16)
Iron Age, op. cit.
(17) AISI, op. cit., pp 9, 10 (some data from 1966 edition also).
(18) Forbes, Company Group Reports, January 1, 1969, P 164.
(19) McManus, G. J., "Steel's Profit Image", Iron Age, 204, No.6, February 6,
1 969, p 44.
(20) AISI, op. cit., P 9.
(21) Ibid.
(22) Forbes, op. cit., p 164.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LASORATORIES
-------
III-25 and III-26
(23)
Chemical Economics Handbook, Stanford Research Institute, current section
219.2430, April, 1968.
(24) Quarterly Financial Reports, FTC-SEC, U. S. Dept. of Commerce, 4th Quarter
Issues for 1959, 1960, 1966, and 1967.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
>-
(.')
o
-J
o
Z
:J:
U
w
I-
z
o
I-
::;)
-J
-J
o
a..
>
.
-------
IV-l
SECTION IV
THE TECHNICAL DETERMINANTS OF
AIR-POLLUTION CONTROL COST
This section presents an abbreviated general outline of air-pollution control
technology as background for the cost studies. More detail on all of these discussions
may be found in the companion Final Technological Report.
Receipt, Storage, and Handling of Raw Materials
Raw materials (principally coal, ore, and limestone) usually are crushed at their
points of origin and sized prior to shipment. Conventionally, ore is crushed to a 4-inch
top size, limestone is crushed and sized to a typical range of 1-1/2 by 2-1/2 inches, and
coal is crushed to a top size of 1 inch or finer. There is a trend to perform additional
crushing and sizing of ores at the mines so that the ore as shipped is ready for charging
to blast furnaces. If crushing and screening are performed at the steelworks, shroud-
ing is required to hold dust losses to a minimum. Raw materials are received at the
steelworks via lake or ocean freighter, by barge, or by truck or railroad car.
Open stockpiles of coal (1/2 to 1 inch is a common top size) may reach a height
of 100 feet and cover up to 10 acres. They can be serious sources of dust during windy
weather. In comparison to coal, stockpiles of ore and limestone are considered to
be less serious sources of emission because the materials are denser and typically
coarser in size. No truly effective control exists for dust blow-off from stockpiles of
coal. Spraying with pitch, oil, or plastics, sometimes used on the dustier coals, is
of limited effectivenes s because the piles are constantly being broken by addition or
removal of material. The large scale of most stockpiles makes it impractical to con-
sider shrouding the entire area. Fine ores intended for sintering contain appreciable
dust and are sometimes wetted to maintain some control over dusting.
Transfers from stockpiles into steel-plant operations are usually by means of
overhead clam-bucket gantry cranes to bottom-dump cars operating on elevated rail-
ways, or by endless rubber belts for upward movement and gravity chutes for down-
ward movement. Dust is created at each transfer point. Outdoor belts are usually
covered but not enclosed, and dust blow-offs can occur during transport of the mate-
rials. Emission control at indoor transfer points may be achieved with cyclone dust
collectors. In Japan, free-fall dusting in the transfer of coal is minimized by under-
ground reclaiming from stockpiles; this practice is not in general use in the U. S.
Massive scrap metal must sometimes be cut up to size it for charging into steel-
making furnaces. Cutting is done outdoors with oxygen torches. Often the preheat is
lost when the lance is moved, whereupon the oxygen jet contacts cold metal. This re-
sults in the evolution of iron oxide fumes. The total amount of fume from this opera-
tion and the amount of iron oxide dust generated from handling scrap is relatively
small, but locally can be quite dense at times. No emission control ordinarily is used
in these operations.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV-2
Fugitive dust from handling of raw materials during receiving, from wind-swept
stockpiles, and from free fall at transfer points is under generally poor control. The
usual motivation to prevent or contain dusting is to improve recovery for economic
reasons, to improve housekeeping (coal dust is a fire hazard), and to reduce fall-out
in areas adjoining the mill.
Coking
Coke is the principal fuel of the blast furnace, and the most important source of
energy in integrated steelmaking. Unfortunately, the charging and pushing of coke
ovens is one of the major uncontrolled sources of air pollution in an integrated steel-
works. Practical methods for effective control of coke-plant emissions are not yet in
sight.
Nearly all metallurgical coke is made in by-product ovens. In the use of by-
product ovens, volatile matter released by heating coal in the absence of air is pro-
cessed to recover tar, ammonia, and solvents. After the extraction of these, a
medium-Btu fuel gas remains. This gas may be used for underfiring the coke ovens,
as well as for heating elsewhere in the steel plant. Coke-oven gas is free of partic-
ulates, but usually contains hydrogen sulfide.
A modern coke oven receives a charge of 10 to 30 tons of coal through ports at
the top. Each oven may be 10 to 18 feet high, 30 to 60 feet deep, and 14 to 20 inches
wide, and as many as 100 ovens may be set together in a battery. Flues between the
ovens carry hot gases which heat the oven walls to an incandescent temperature. The
charged coal usually has a top size of 1/8 inch. After a coking period of 16 to 20 hours,
doors at both ends of the oven are opened and a ram pushes the coke, now typically
4-inch pieces, into special railway cars. At a quenching tower, the hot coke is deluged
with water to cool it. The coke is then screened to size and sent to the blast furnaces.
A flow sheet for a typical by-product coke plant is presented in Figure IV -1.
The gases and volatiles released from coal during coking are drawn into ascen-
sion pipes at the ends of the coke ovens and thence into collector mains. The gases
include carbon monoxide, hydrogen, methane, hydrogen sulfide, ammonia, and nitro-
gen. The volatiles include anthracene and other tarry compounds, benzene, toluene,
xylene, naphthalene, phenols, and pitch. Recovery of these by-products starts as the
cooled raw gases leave the mains. A flow sheet for a typical by-product plant is given
in Figure IV-2.
Charging and Coking
Steam, flame, smoke, and fine coal rush out of the oven ports during charging.
During coking, leakage of aromatic vapors may occur at the charging ports and at the
end doors. Leakage can be minimized by adequate maintenance of the seals, but a satis-
factory solution to the problem of emissions during charging has not been found. Some
improvement has resulted from the use of steam jets in the raw-gas ascension pipes to
increase the draft on the open oven during charging. An alternate method tried in
Europe is the use of a special larry car that drops the coal through sleeves lowered
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
Air
IV-3
Cleaned Coal
3100 pounds
.
Coal Dust
15 pounds
Products of Combustion
(Heating)
.
Coke Oven Gas
560 pounds
Hot Coke
2220 pounds
.
QUENCH TOWER
Water +
Spent Liquor
( . Denotes emission sites)
FIGURE IV-l. TYPICAL FLOW SHEET FOR A BY-PRODUCT COKE PLANT
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
1-
IV-4
COKE OVENS
DOWNCOMER
REHEA TER
Tar
Ammonia and
Phenols
SCRUBBER
Steel Plant Use
Weak Ammonia Liquor
Further Processing
FIGURE IV-2. BY-PRODUCT PLANT FLOW SHEET
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV-5
into the oven and collects the emissions in shroud pipes. An attempt is made to burn
the vapors with the shroud air, then to wash the waste gases. Structural limitations of
present American ovens may not permit direct adoption of this system; in any event it
is imperfect. An alternative pneumatic charging method is being worked on in the
United States.
Pushing
When the incandescent coke is pushed out of the oven onto an open quench car, a
thermal draft is induced in the immediate surroundings. Fine particulate s are blown
high into the atmosphere. Most of this dust comes from abrasion of the coke during
pushing, but there may be smoke from incompletely coked coal. Good coking practice
results in reduced emission of smoke during pushing, but does not eliminate dusting.
The near-future prospects for development of a pushing system having good control of
emissions are poor.
Coke Quenching
The open quench car contains the newly pushed coke on a sloping hopper bottom
with side gates made of grating. The car is moved to the quench tower: a large
chimney that fits over the car. Water sprays in the chimney deluge and cool the coke,
and the rising cloud of steam in the chimney lifts coke dust into the atmosphere. Most
of this dust appears to fall out in the vicinity of the tower. Baffles installed in a quench
tower will reduce the emission of particulates into the atmosphere by 75percent or more.
Coke Handling
The quenched coke is dumped onto a sloping brick wharf and conveyed to sizing
and screening operations. Steel and rubber shrouding at the transfer points, at the
crusher, and at the screens minimize dust loss. Indoor transfer points may utilize a
cyclone and fan to collect the dust. Fines are removed before the sized coke is trans-
ferred to the blast furnace. The fines are usually transferred to the sinter plant. Com-
pared to charging of the ovens with coal and pushing of the hot coke, dusting during coke
handling is not a serious problem, except in windy weather.
By-Product Proces sing
Usually only minor emissions occur at the primary (hot) end of the by-product
system because it is under negative pressure. Some odor of free vapor may be present
at the tar collectors and decanters, and at locations where the liquor runs in lines that
are not fully closed. Ammonia and organic fumes may be particularly strong at sumps
where decanted liquor and other flush liquor is recycled to the sprays for cooling the
collector mains. In older plants, the addition of sulfuric acid to the ammonium sulfate
precipitator tank can cause fuming.
Tars may contain polycyclic aromatic hydrocarbons which may be hazardous to
health. Emissions from vents on tar processing and storage tanks have been led through
scrubbers to absorb or destroy the fumes. However, the tars tend to condense and foul
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
-
IV-6
the scrubbers. Vapors of the light oils are both toxic and flammable. Condensers and
some process tanks are vented, and the sweet aromatic vapors often pervade a large
area on calm days. Some leakage is inevitable, but the high fire-hazard level encour-
ages preventive measures irrespective of needs to control air quality.
Preprocessing Raw Materials
Sintering
Sintering plants convert iron-ore fines and metallurgical dusts into an agglomer-
ated product that is coarse enough for charging into the blast furnace. The iron-bearing
materials are moistened and mixed with fine coke to form a bed on a slow-moving
grate. The bed is kindled under an ignition hood; then a forced draft of air keeps the
bed burning until a clinker is formed. Air is blown onto the sinter to cool it after dis-
charge from the main grate. Then the cooled sinter is screened and transported to
the blast furnace. Sinter fines are recycled.
Sintering machines process a wide variety of feed materials and produce a con-
siderable amount of emissions. Points of emission in a sinter plant are designated by
the circles in the flow sheet in Figure IV -3.
Minor amounts of dust are created in the handling and grinding of raw materials.
Other emissions include dust sucked through the grate bars into the windbox, combus-
tion gases from ignition and firing, and dusts generated in the screening and cooling
operations. Complete combustion during sintering makes it unlikely that the exhaust
gas contains unburned hydrocarbons. However, the coke -oven gas used for ignition
and the sulfur in the sinter mix contribute sulfur dioxide to the combustion products.
Sinter dust may contain particles of iron oxides, fluxes, and silicates.
Multicyclones, electrostatic precipitators, venturi scrubbers, mechanical col-
lectors, and baghouses have been used in various combinations at the various points of
emission.
Control of sintering emissions is considered to be at a low level, especially in
manufacture of fluxed sinters. Operation and maintenance costs are high. Control of
emissions to a level under 0.05 grain/ sd is possible, but clearly expensive.
Pelletizing
Pellets are made by rolling fine iron ores mixed with a binder to form damp balls
about 1/2 inch in diameter. The balls are dried and fired to harden them. Pellets are
usually made at or near the mine sites, rather than in the steel plants. In this respect,
pelletizing differs from sintering, which is usually done at the steel plant.
Concentrates received at the pelletizing plant are moist, and dust generation
during receiving is not a problem. Bentonite is received in covered hopper cars and is
unloaded in special bins that meter the material into the pelletizing operation. Partic-
ulate emissions are magnetite, hematite, and bentonite. The minor amounts of dust
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
Coarse ore
r-------'
Open- hearth I
~__A.ulr.- -_J
(. Denotes emission sites)
IV -7
Limestone
.
.
Coke
.
.
FIGURE IV-3. TYPICAL FLOW SHEET FOR A SINTERING PLANT
MIXING
SINTER MACHINE
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
----------. -
IV-8
generated in the plant are handled usually by simple cyclones or baghouses. The indura-
tion operation (heat hardening) is conducted at relatively low air-flow rates, and forma-
tion of particulate emissions usually is not substantial. Although finished pellets are
strong and abrasion loss is low, during loading for shipment a considerable amount of
dust is released.
Ironmaking
The blast furnace is a massive refractory-lined structure about 100 feet high and
up to 30 feet or more in diameter at the hearth. Iron ores, fluxes, and coke (jointly
called the burden) are charged at the top of the furnace through a series of seals (called
bells), Air preheated in regenerative stoves is forced through ports (called tuyeres)
arranged around and just above the hearth. The air (sometimes augmented with oil, gas,
oxygen, or steam) reacts immediately with ignited coke to generate hot reducing gases
which, in turn, react with the oxygen in the ore to reduce it to iron. As the burden
moves downward into the fusion zone, the iron becomes molten and collects in the
hearth. Fluxes in the burden react with impurities in the ore and coke and form a layer
of molten slag that floats on top of the molten iron. Periodically, the iron is cast into
ladle cars that deliver it to the steelmaking furnaces. Slag is tapped from the furnace
and transported to a dump area, or is granulated with water. The gas ascending in the
blast furnace is removed at the top, stripped of dust, and then used as fuel to fire the
regenerative stoves of the blast furnace. It is also used as fuel in the powerhouse and
for other operations. Figure IV -4 shows a flow diagram of blast-furnace operations;
the circles indicate emiss ion sites.
Charging
Modern blast-furnace burdens consist of sinter, pellets, screened ore, or a mix-
ture of these, plus coke and fluxes. Raw materials are moved from the stockpiles to
surge hoppers (called pockets) at the blast furnace. Coke is usually transferred to the
pockets by conveyor belt; ore and flux are transferred by bottom-dump cars. Materials
drawn from the pockets are weighed and transferred to the top of the blast furnace by a
skip hoist. Usually a scale-mounted car is used for the transfer of ore and flux from
the pockets to the skips, and a conveyor is used to transfer coke.
The raw materials are dumped from the skips into a receiving hopper at the top
of the furnace. In step-wise fashion, the materials are dropped through a series of
two or more seals with intermediate hoppers. This seal system minimizes the escape
of furnace gases as charges pass successively from the outer hopper to the furnace.
Operations in the stockhouse are dusty, and shrouding at transfer points aids in
confining the dust. Conveyor belts raise less dust than transfer cars. The transfer
at the top of the furnace is highly exposed, but partial shrouding is possible. Leakage
of gases, which also contain particulates, can develop in the bell system as a result of
wear or distortion of seals. In high-pressure blast furnaces, a steam sy"stem main-
tains back pressure between closed seals during burden transfer.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV -9
.
.
.
.
Natural gas,
45 pounds
Heated air,
4, 300 pounds
BLAST FURNACE
Slag,
460 pounds
.
.
Top gas,
6,450 pounds
Pig iron,
2, 000 pounds
.
Other plant use,
4, 850 pounds
Atmosphere
Combustion air,
1, 090 pounds
Heated air,
4,300 pounds
Blast air,
4, 300 pounds
Combustion
prod uets,
2,460 pounds
.
<8
Denotes emission sites)
FIGURE IV-4. TYPICAL BLAST-FURNACE OPERATION WITH A BURDEN CONSISTING MAINLY OF
SINTER, AND WITH INJECTION OF NATURAL GAS
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV - lO
Smelting
Slips during smelting are sudden movements of the burden in the furnace. In
some practices, small, periodic, controlled slips are caused deliberately. Large,
uncontrolled slips may generate high pressures that are relieved by bleeders (safety
valves) in the uptakes (gas-collection mains) at the top of the furnace. In such cases,
relief of the pressure may be accompanied by a discharge of dust and gas into the
atmosphere. Improved raw materials have reduced the dust loading in the furnace and
have also minimized abnormal operating conditions that cause slips. Control on some
blast furnaces has advanced to the stage where bleeders open rarely and only for short
intervals.
The raw blast-furnace gas consists chiefly of steam, nitrogen, hydrogen, carbon
monoxide, and carbon dioxide. The carbon monoxide content is about 25 to 30 percent.
The hydrogen content ranges from 1 to 6 percent and varies with blast moisture and
fuel injection. At least some of the gas discharges continuously through leaks at the
bells and in areas where the furnace shell is pierced for instruments and coolers.
Dust entrained in the blast-furnace gas results from abrasion of the burden during
charging and during the early stages of passage down the blast furnace.
The flue gas leaving the blast furnace is dusty, and is cleaned to a concentration
of les s than 0.01 grain/ sd to as sure that clogging or slagging reactions do not occur
when the gas is used as a fuel to fire the regenerative blast stoves. About 30 percent
of the gas usually is used to heat the stoves; the rest is used as fuel for other in-plant
heating purposes, or flared. Commonly a gas-cleaning system will include a dust
catcher and a primary washer. Additional equipment is used only as dictated by the in-
plant use of the gas.
Casting and Flushing
Blast-furnace iron is saturated with carbon. As soon as it emerges from the
furnace, graphite flakes are rejected and rise to the surface where air currents sweep
them into the atmosphere. Manganese vapor is also given off and oxidizes to form a
fine dust, but the amount is quite small. The graphite "kish" is a nuisance because it
is readily windblown and is difficult to remove after it has settled. Kish control during
casting consists of running the iron short distances to minimize the amount that can
form and escape before the iron flows into closed ladles.
The volume of slag from the blast furnace is similar to the volume of hot metal.
Slag is flushed out of the furnace prior to and during each cast of iron. The sulfur load
in the slag is 6.5 to 9.5 pounds per ton of hot metal. During flushing, some of the
sulfur may react with oxygen and moisture in the air to form sulfurous gases near the
slag runners. Long runners at older furnaces give greater surface exposure and in-
creased fouling of the air. Most slags are transported in open ladles to dump areas
where hydrogen sulfide is evolved as the slag cools and weathers. In some newer
practices, the slag runs a short distance and is granulated with high-pressure water.
However, the formation of hydrogen sulfide may continue in the granulation pit at low
temperature.
A small portion of the molten iron from American blast furnaces is solidified into
solid "pigs" for distribution. The kish problem discussed above is severe during
pigging.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV - 11
Steelmaking
Open-Hearth Furnaces
With the growth of other steelmaking processes, the percentage of steel made in
open-hearth furnaces has decreased, but 55 percent of U. S. raw steel was made by
this process in 1967. Open-hearth furnaces range in holding capacity from 25 to
600 tons. Time to produce a heat ranges normally from 8 to 12 hours, but this time
can be shortened by lancing the bath with oxygen.
Special charging machines charge iron ore, limestone, scrap iron, and scrap
steel to the open hearth at the start of a heat. An intense flame from combustion of
oil, tar, coke-oven gas, or natural gas travels the length of the furnace to heat the
solid charge. The hot waste gases are led through regenerative chambers called
checkers, and the flow of the flame is reversed each 15 to 20 minutes to let the newly
heated set of checkers preheat the combustion air. Hot metal from the blast furnace
is added by pouring from a large ladle into a spout set temporarily in a door of the
open-hearth furnace. Carbon and impurities are oxidized from the bath to convert the
charge into steel. After the heat is refined to the desired composition, the molten
steel is ready for deoxidation and tapping. The tap hole is at the base of one wall and
discharges into a ladle. To shorten the time of open-hearth heats, oxygen may be
injected into the bath by lances extending through the roof of the furnace. Oxygen con-
sumption may range from 600 to 1000 cubic feet per ton of steel. A flow sheet for a
typical hot-metal operating practice is shown in Figure IV -5. The circles indicate
points of emission.
Emis sions. Minor emissions of iron oxide occur during the charging and tapping
of open-hearth furnaces, but the main emissions are in the combustion gases. Partic-
ulates in the combustion product consist of red iron oxide and magnetic iron oxide. The
gas also contains sulfur compounds and fly ash from the use of fuel s such as oil, tar,
and coke-oven gas. Firing with oil is reported to create a lower average dust loading
than firing with tar.
The amount of dust generated varies at different stages of the steelmaking proc-
ess and with the particular practice. Typical dust loadings for oxygen-lanced open-
hearth furnaces are estimated to be 20 to 30 pounds per net ton of raw steel. Without
oxygen lancing, about 8 to 10 pounds of dust per net ton of steel is emitted.
The slag pockets, checker chambers, and flues to waste-heat boilers provide
opportunities for settling-out of dust from the combustion gases, and thus served as
fairly efficient dust collectors until the advent of oxygen lancing. The use of oxygen
lancing in open hearths increases the dust loading and generates large volumes of
metallurgical fume. Effective control of the emissions is obtained with electrostatic
precipitators. Venturi scrubbers and baghouses are also being used. Illustrative flow
diagrams of open-hearth dust-collecting systems are presented in Figure IV -6.
Teeming. Teeming is the pouring of liquid steel into cast iron molds where the
steel solidifies into ingots. In the teeming area, ingot molds on a string of cars are
filled with steel from the ladle. In 1967, about 94 percent of raw steel produced in the
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV -12
Heated fuel oil
and steam,
258 ounds
Steel scrap,
876 pounds
Scrap yard,
60 pounds
8
(8
Denotes emission sites)
FIGURE IV-5. OPEN-HEARTH FURNACE OPERATING WITH HOT-METAL PRACTICE CONSISTING OF
60 PERCENT HOT METAL AND 40 PERCENT STEEL SCRAP (ORE PRACTICE)
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
Two parallel,
electrostatic
precipitators
U. S. Steel Corp.
Fairless, Pa. - 1953
IV - 13
Republic Steel Corp.
Buffalo District- 1964
Bethlehem Steel Corp.
Sparrows Pt., Md. - 1963
FIGURE IV-6. ILLUSTRATIVE FLOW DIAGRAMS OF OPEN-HEARTH DUST-COLLECTING SYSTEMS
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
~- .-- ------ -.- _.- ~- - -
IV -14
United States was poured into ingot molds. The use of tar and bitumens as mold coat-
ings has been curtailed in the past decade, with a resulting decrease in the amount of
visible emission during teeming. However, with some present teeming practices,
visibility can become highly restricted at the teeming station. Little information is
available on methods of controlling the teeming emissions.
Basic Oxygen Furnaces
It is estimated that by 1970 nearly half of the steel in the United States will be
made by the basic-oxygen-furnace (BOF) process. The furnace is a pear-shaped shell
lined with refractory brick. It is charged through the top, and tilted down for tapping.
Typically, the charge consists of 70 percent blast-furnace hot metal and 30 percent
scrap, plus fluxes. A water-cooled lance impinges oxygen at high velocity on the sur-
face of the charge to promote agitation mixing of the oxygen with the molten bath.
Rapid oxidation of carbon, silicon, manganese, and some of the iron occurs. These
exothermic reactions supply heat to melt the scrap and reach the tapping temperature.
Some impurities from the charge enter the slag. A typical ISO-ton BOF can produce
a heat in about 33 minutes.
A flow sheet for a typical BOF process is presented in Figure IV-7.
major emissions are indicated by the circles.
Points of
Emissions. Kish is formed during the charging of hot metal. It consists of
flakes of graphite, and may include fragments of iron oxide and traces of quartz and
calcite.
During the blow with oxygen, the predominant particulate emission is fine iron
oxide fume. If galvanized scrap is part of the charge, zinc oxide in the collected dust
makes the dust unsuitable for sintering for feed to the blast furnace. For this reason,
the trend is to divert galvanized scrap to open hearths.
The amount of dust per net ton of raw BOF steel range s from 20 to 50 pounds.
1968, an average of 40 pounds per ton was reported by one plant.
In
Flow diagrams for typical gas-cleaning systems for the BOF are shown in Fig-
ure IV-8. Serious air pollution is avoided by the use of these cleaning systems.
Teeming of BOF heats is the same as was described for open-hearth heats.
Electric-Arc Furnaces
In 1967, electric-arc furnaces produced about 11 percent of the total raw carbon
steel made in the United States and 36 percent of the alloy and stainless steels. About
59 percent of all electric-furnace heats were carbon steels. Electric-arc furnaces are
refractory-lined cylindrical basins having a capacity of up to 200 tons or more. In
1968, 40 percent held less than 50 tons, 36 percent held from 50 to 90 tons, and 24 per-
cent were over 90 tons in holding capacity.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV - 15
Oxygen,
138 pounds
BASIC OXYGEN FURNACE
.
Raw steel,
2,000 pounds
Slag,
263 pounds
Scrap,
77 pounds
Dust,
45 pounds
Off gas,
168 pounds
Rolling
operation,
2,000 pounds
Scrap yard,
77 pounds
Dust collector
(8
Denotes
emission sites)
(a) Charge scrap plus cooling scrap = 678 pounds.
FIGURE IV-7.
BASIC OXYGEN FURNACE OPERATING WITH 70 PERCENT
HOT METAL AND 30 PERCENT STEEL SCRAP
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
1-
Inland Steel Company
Indiana Harbor, Ind. - 1965
-
IV - 16
Wisconsin Steel
South Chicago Ill. - 1964
FIGURE IV-S. EXAMPLES OF GAS-CLEANING SYSTEMS FOR BOF STEELMAKING FURNACES
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV - 1 7
High-amperage electricity at moderate voltage is passed through large graphite
electrodes extending down through the roof of the furnace. The charge consists of
selected steel scrap plus fluxes and alloying elements to achieve the desired composi-
tion. Melting is accomplished by the heat of the arc formed between the electrodes and
the charge.
Emis sions. Preheating of the scrap is not yet a common practice for shortening
the heat time. Where preheating is practiced, it is done most commonly with air-fuel
burners. Oxy-fuel burners are used to only a limited extent. The scrap is rarely
heated to above 1800 F; thus preheating creates no significant emission problems unless
dirty fuels are used or combustible dirt (such as oil or rubber) is in the scrap.
Electric-induction furnaces melt special alloys on a small scale. If the charge is
free of tramp combustibles, emission from these furnaces is minor and is collected
readily in simple equipment.
Emissions in the electric-arc furnace can originate from light scrap that oxidizes
readily, from dirty scrap (a major source), and from oxygen lancing. A flow sheet for
a typical electric-arc furnace process is shown in Figure IV-9. The main emissions
are fumes from scrap preheating, iron oxide dust from the melting operations, and
furnace off-gases.
The amount of dust released per net ton of electric-furnace steel depends on the
condition of the scrap and whether or not oxygen lancing is used. Dirty scrap can raise
the dust emissions from a normal level of 8 to 15 pounds to as high as 40 pounds per
net ton of steel. It has been estimated that oxygen lancing produce s 20 percent of the
total emissions. The composition of the off-gas from the electric furnace varies with
the practice. The chief constituents are carbon monoxide, carbon dioxide, nitrogen,
and oxygen.
Emis sions leave the furnace through the electrode ports in the furnace roof, out
of the tapping spout and slagging door and, in the case of top-charged furnaces, through
the open furnace top during charging. Three main types of systems are used to collect
the emissions: (1) hoods over and around the furnace, (2) direct extraction under draft
from inside the furnace, and (3) shop-roof hoods.
In the first type of system, hoods are fitted at the points of emission, and ducts
pass the emissions to the dust collector. Hoods of this type must be movable when
used with top-charging furnaces. They tend to obscure visibility from the crane
operator's cab. Warpage of the hoods is a problem, and they seldom last 1 year.
Direct extraction by drafted duct from inside the furnace causes shop air to flow
into the furnace and thus minimizes the discharge of emissions through the doors and
ports. This system increases roof life and may decrease electrode consumption. It
affects recovery of alloying elements in the steel bath, and may create some difficulties
with carbide slags used in refining of special steels. Heat exchangers must be used
unless the duct length is adequate to cool the gases by radiation to a safe entry temper-
ature into the dust collector.
In the shop-roof extraction system, the shop building serves as the collector hood.
Ducts in the roof of the building exhaust the emissions to the dust-collecting system.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV - 18
ELECTRIC FURNACE
.
.
o
Scrap yard,
60 pounds
<8
Denotes emission sites)
FIGURE IV-g. EXAMPLE OF ELECTRIC-FURNACE STEELMAKING PRACTICE
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV - 19
Teeming practice is the same as for open hearth and BOF heats. Some
electric-furnace shops use roof-extraction exhaust systems in the teeming building to
collect the fumes.
Flow diagrams of dust-collecting systems that use baghouses or wet scrubbers
for direct extraction and for hood extraction are presented in Figures IV-I0 and IV-II.
Figure IV-12 is the flow diagram for emission control by a shop-roof extraction
system.
Vacuum Degassing of Steel
For certain critical applications, steel
methods are used: (1) stream degassing, (2)
degassing.
is degassed by vacuum treatment. Three
circulation degassing, and (3) ladle
The source of emissions in vacuum degassing is the molten steel in the vacuum
chamber. The principal gases emitted are carbon dioxide, carbon monoxide, and
hydrogen. The violent agitation of the molten steel and the vaporization of some
metallics generates dust.
The steam ejectors used to create the degassing vacuum also serve as scrubbers.
Thus, dusts are normally not released to the atmosphere. One source of information
states that about 10 pounds of dust are collected in degassing 100 tons of steel.
Continuous Casting of Steel
Continuous casting is an alternative to conventional teeming into ingot molds. In
continuous casting, molten steel is poured into a water-cooled, bottomless copper mold
and cools quickly to the shape of the mold, from which it is withdrawn continuously.
The estimated continuous casting capacity in the United States in 1968 was about 7 mil-
lion net tons per year, and is expected to double in 1969.
Emissions during continuous casting are markedly lower than during teeming into
ingot molds, because the rapeseed oil used as a mold coating creates only a small
amount of fume. Continuous casting is conducted in one location rather than over a
large area. This permits the use of a localized fume-collecting system. The pouring
tundish may be blanketed with a reducing gas to minimize oxidation of the steel, and
for vacuum-degassed heats the stream from the tundish to the mold is shrouded with
argon.
Steel Shaping
Primary Breakdown
Soaking pits are reheat furnaces in which steel ingots are brought to a controlled
temperature for rolling. They are fired with blast-furnace gas, coke-oven gas, or a
mixture. The amounts of particulate emissions are small, but gaseous emissions may
include sulfur dioxide if the coke-oven gas has not been freed of hydrogen sulfide. No
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
Lukens Steel Co.
Coatesville, Pa. - 1964
IV - 20
I
I
I
r .t --,
L Same I
1- -
I
I
I
I
I
I
I
r J-l
L~ar~
I
I
t
Lukens Steel Co.
Coatesville, Pa. - 1966
U. S. Steel Corp.
Chicago, Ill. - 1961
FIGURE IV-10. EXAMPLES OF ELECTRIC-FURNACE DIRECT-EXTRACTION EMISSION-CONTROL SYSTEMS
WITH BAGHOUSES
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV - 21
Two water-cooled
evaporation chambers
Armco Steel Corp.
Butler, Pa. - 1959
Armco Steel Corp.
Houston, Texas - 1966
FIGURE IV-II. EXAMPLES OF ELECTRIC-FURNACE DUST-COLLECTING SYSTEMS USING WET SCRUBBERS
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
Top-charge
furnaces,
five 60 ton
one 20 ton
Ingot
pouring
area
Six Baffle Control Hoods,
one per furnace
Six roof exhaust hoods,
9-1/2 x 16-1/2 feet
Four Distribution ducts
Jones & Laughlin Steel Corp.
Warren, Mich. - 1966
FIG URE IV -12.
IV -22
Top-charge
furnaces,
one 70 ton
two 100 ton
Bethlehem Steel Corp.
Los Angeles, Cal. - 1966
EXAMPLES OF ELECTRIC-FURNACE SHOP-ROOF EMISSION-CONTROL SYSTEMS
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV - 23
economical commercial method is available for removal of the sulfur dioxide in the
exhaust gases.
During breakdown of the ingots by rolling into billets, blooms, or slabs, the majoI
emission is steam; emission of dust is minor.
Conditioning, Reheating, and Hot Rolling
After ingots are rolled to billets, blooms, or slabs, they may be cooled, in-
spected, and conditioned by grinding, chipping, or scarfing surface defects with hand
torches. More complete surface conditioning is sometimes obtained via all-over
scarfing with oxygen in special scarfing machines. Then the steel slabs and billets are
reheated before they enter the hot-strip mill.
The fine particulates generated during grinding are airborne only a short distance
in the area. Sometimes they are collected at the grinding station. Billet scarfing
causes a metal loss ranging from 3 to 6 percent, but most of the loss is spatter - not
fume. No emission controls are used in hand scarfing, except that hoods are some-
times used if the shop practice calls for extensive hand scarfing. The loss in machine
scarfing of slabs is up to 2.5 percent, and up to 7 percent for blooms. The emissions
are chiefly iron oxides and are collected with electrostatic precipitators or high-energy
scrubbers. Machine scarfing would result in serious air pollution if control systems
were not used.
Emissions during reheating are mixtures of carbon monoxide, carbon dioxide, and
nitrogen from the combustion of natural gas. Emissions at the mills consist mostly of
steam, which is confined to the area of the mills. The emission of fine iron oxides at
strip finishing stands is cons idered to be significant enough that some mills collect these
emissions with high-energy scrubbers.
Except for scarfing operations, no serious
tions of conditioning, reheating, and hot rolling.
forging and hot-forming operations.
emission problem exists in the opera-
This conclusion also applies to hot-
Acid Pickling
Acid treatment, called pickling, is used to clean the oxidized surface of hot-
rolled steel in preparation for cold rolling. Either hydrochloric or sulfuric acid is
used. Acid fumes are the chief emissions during pickling. Hoods exhaust these fumes
to a wet scrubber and thence to a packed tower. One source states the collection rate
is about 100 grains of hydrochloric acid per ton of steel. For pickling of low tonnages
of steel, the fume system may be too costly, and the fumes are then ejected to the
atmosphere through a roof exhaust.
Pickling can contribute moderately to air pollution if control systems are not
used.
Cold Rolling and Cold Forming
In high- speed cold
cant applied to the rolls.
rolling, a water-oil mist is generated from the emulsion lubri-
Collection of this emission is by mechanical mist eliminators
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
~ --
IV -24
or wet scrubbers. Neither cold-rolling nor cold-forming operations contribute signif-
icantly to air pollution.
Steel Finishing
Continuous strip lines lend themselves to convenient control of emissions during
coating operations. Coatings include zinc, tin, terne (lead alloy), aluminum, chro-
mium, nickel, copper, phosphate, and a variety of paints. Prior to coating, the strip
is heated to carbonize grease or oil films, and the lightly oxidized surface is cleaned
in an acid tank or by the action of a reducing atmosphere in a furnace. A hot alkaline
solution may be used instead of heat to remove the grease film before pickling.
In dip coating (zinc, aluminum, and terne coat), the preheated strip passes
through a flux layer on top of the metal bath. In electrolytic processes (zinc, chro-
mium, nickel, and copper), the cleaned strip enters directly into the plating bath.
Recent proces ses for coating with chromium and nickel involve coating with chemicals
which are then reduced to metal in a hydrogen atmosphere. Phosphate coatings are
produced by dipping steel in a dilute acid phosphate solution saturated with a metal such
as zinc, cadmium, aluminum, or lead. The metal surface is converted to an insoluble
crystalline phosphate coating. Paints are applied to cleaned strip by an electro-
phoretic process, by dipping, or by roller coating. Painting operations emit solvent
vapors which can be collected by an exhaust system that disperses the vapors to the
atmosphere outside of building. During baking, the evolved vapors are combusted
or exhausted.
In these finishing operations, emission of particulates is negligible. Heating of
the steel results in stack gases having the normal carbon monoxide-carbon dioxide mix-
ture from combustion of natural gas. Acid mists from pickling and vapors from
electroplating baths are readily collected by hoods and removed in wet scrubbers and
packed towers.
Miscellaneous Operations
Gas Distribution
Blast-furnace gas and coke-oven gas are used as in-plant fuels. Blast-furnace
gas has a heat value of about 80 Btu/ sd. It is cleaned of particulates to a high degree
to make it suitable for heating the blast stoves. Exces s blast-furnace gas is used in
the powerhouse, for underfiring of coke ovens, and in soaking pits of reheating fur-
naces. Because its caloric value is low, it is often enriched with natural gas or coke-
oven gas.
Coke-oven gas has a heat value of 500 to 550 Btu/ sd. It is often used for under-
firing coke ovens. The remainder is used for (1) ignition of the bed in sintering plants,
(2) injection into the blast furnaces, (3) open-hearth fuel, (4) enrichment of blast-
furnace gas, and (5) powerhouse fuel. If the coke-oven gas is not stripped of its hydro-
gen sulfide, its use as a general fuel results in the release of objectionable sulfur
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
"_-
IV -25
oxides in the products of combustion. The one exception is its use as an auxiliary
fuel by injection in the blast furnace. In this case the sulfur goes into the slag. The
technology exists for desulfurizing coke-oven gas, and some coke plants are equipped
to do it. However, the removal and recovery of sulfur as elemental sulfur, sulfuric
acid, or ammonium sulfate is not economically practical.
In the various processes where coke -oven gas is used as a fuel (other than in the
blast furnace), the products of combustion will contain sulfur as oxides. The amount
of sulfur will depend on the proportion of coke -oven gas making up the fuel. Coking of
a low-sulfur coal results in about 3.8 grains of hydrogen sulfide per standard cubic
foot of coke -oven gas. If used undiluted, this gas would yield an appreciable amount
of sulfur dioxide upon combustion. Coke -oven gas containing up to 0.9 grain of hydrogen
sulfide per standard cubic foot is considered to be metallurgically acceptable for
general mill use. Usually no control is used on the sulfur oxide emis s ions when
hydrogen sulfide -bearing coke -oven gas is used as a fuel.
Powerhouse
The powerhouse raises steam used for compressing air for the blast furnace,
generating electricity, distilling coke by-products, pumping coke-oven and blast-
furnace gases, powering of forges and presses, warming residual oil and tar lines,
and for comfort heating and general utilities.
Almost any available fuel can be burned in the boiler station. The principal fuel
is blast-furnace gas, and the need for additional fuel is balanced out chiefly with non-
coking coals having typically 2 percent sulfur. Coke-oven gas and coke breeze are
used if there is a surplus.
Blast-furnace gas is a clean fuel; but coal, coke-oven gas, tar, oil, and coke
breeze generate sulfur oxides and perhaps some fly ash upon combustion. Methods
for economical treatment of flue gases for the removal of sulfur are still under devel-
opment. If treatment of the flue gases becomes mandatory, it may become economical
for mills to diminish steam-raising to the minimum.
With automatic combustion control of gaseous fuels in the boilers, generation of
particulates is insignificant and the stack gas is not cleaned. Electrostatic precipi-
tators are used to collect particulate emissions (smoke and fly ash) from firing with
other fuels. Except for the exhausting of sulfur oxides to the atmo sphere when firing
with sulfur-bearing fuels, emission-control systems are effective in preventing pollu-
tion of the air.
Plant Waste Incineration
Open-dump burning of steelworks' waste and refuse can create considerable
smoke. One mill reports the recent installation of a double-fired incinerator designed
to burn all combustible materials cleanly. A fan draws the off-gases through a wet
scrubber designed to reduce the stack emissions.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
_.
IV -26
Priorities, Problems, and Opportunities in
Pollution-Control Technology
The foregoing review has shown that there are numerous occasions and sites of
possible air-polluting emissions within an integrated steelworks. Because of these
multiple sources, air-pollution control for steelmakers is a complex and substantial
problem. Whereas the public (including the steelmakers) has a single overall objective
(cleaner air), this objective is fragmented into dozens of separate pollution-control
goals within the plants. To obtain suitable overall progress at feasible rates of invest-
ment and of technical development, it is important to assign priorities, to anticipate
major problems well in advance, and to identify all possible opportunities for technical
gain.
Priorities
In discus sing the overall problem of steelworks air pollution, highest priority
should be given to control of coke-plant emissions. In particular, primary concern
should be directed to emissions of coal smoke and aromatic gases during charging of
coal, early stages of coking, and pushing of the finished coke. Imperfect controls at
these critical points allow the release of dirty and noxious substances not unlike those
released by unregulated burning of soft coal in our cities decades ago. Considerable
research and development effort has been expended on control of coking emissions, but
unfortunately without satisfactory results to date.
Assignment of lower priorities (by both steelmakers and the public) tends to be
based upon the apparent amounts of material emitted from various operations. Because
they are visible, particulates have received the most attention until recently. Visible
airborne particulate emissions are potentially greatest for the following operations
(not necessarily in the order shown):
.
Steelmaking (especially when oxygen is used for refining)
.
Sintering (especially for modern fluxed sinters)
.
Transfer, crushing, screening, and blending of raw materials
.
Casting and handling of hot metal from the blast furnace
.
Scarfing and grinding of semifinished steel.
Most steelmakers have made at least some progress toward control of fumes and dusts
from the above source s, but in each plant there are special difficulties to be over-
come. Although the technology for capture of dust by baghouses, scrubbers, or electro-
static precipitators may be well-developed, economic and technical problems may arise
where the emission occurs over large areas (as in stocking, storage, and reclaiming
operations) or from numerous individual sites (as in sinter processing or coke handling).
Many stocking/reclaiming systems require several dumps from overhead clamshell
buckets as the ore or coal moves into or out of stock. These are difficult to shroud
over an area of several acres.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV - 2 7
Of the unwanted gaseous emissions from steel plants, two must be assignea !a5"
priority because they release sulfurous gases to the atmosphere. Two steelworks sub-
stances, the coke -oven gas and the blast-furnace slag, collect and carry about 75 per-
cent of all sulfur that enters the plant. If the gas is used as a fuel, the hydrogen sulfide
in it oxidizes and is exhausted as sulfur dioxide. When the slag is disposed of, it may
give off a fraction of its hydrogen sulfide upon weathering, whenever it is wetted by
rain or plant water. Economically practical controls for these gaseous emissions are
not yet available.
Lower priorities are assigned to control of the emissions from conveyance of
raw materials (mostly by rubber belt), from blow-off at stockpiles, from steelworking
operations such as rolling, and from the multitudinous dips, baths, and sprays used in
steel finishing. Where blow-off from stored solids has intruded into the community, it
has usually been controlled by wetting the materials. Dust collectors are sometimes
used on high-speed rolling mills, as much for shop cleanliness as for control of air-
pollution. The finishing processes are generally amenable to some form of hooding as
appropriate to control emis sions.
In this steelmaking study, lowest priority is assigned to emissions from the
blast-furnace process gases and from powerhouse operations. Because the economic
incentive to clean blast-furnace gas for use as an in-plant fuel is strong, it is normally
cleaned to 0.01 grain (particulates) per cubic foot, or better. Blast-furnace gas con-
tains little or no sulfur; thus its use for firing stoves or for steam-raising creates no
air-pollution problems. Where powerhouses are partly fired with tar or fos sil fuels,
the emissions are properly classed with those of the power industry. Most steelmakers
would be able to convert boilers to use only blast-furnace gas and natural gas, if this
became necessary.
Problems
Air-pollution control technology embraces the activities of prevention, contain-
ment, and collection of airborne emissions, and these activities may be considered
together or separately as the occasion requires. Containment and collection obligate
steelmakers to arrange for re-use or disposal of the materials collected.
In the case of coking, the big problem is containment of smoke so that it may be
collected or destroyed. A conventional coke-oven battery may have a mile or more of
seals associated with its end doors; the operations of charging and pushing require
containment of emissions from 400 to 500 additional locations per battery of 100 ovens.
Although the ducts and hoods applied to this task might be made movable, the overall
problem is such that one control authority has proposed that entire batteries might be
"encapsulated" within some kind of building. If this were done, very careful engineering
of building ventilation and exhaust systems would be required to assure that overheated,
toxic, and/ or explosive atmospheres did not form within. Many students of the coke-
oven problem feel that a wholly new approach to the coking of coal is preferable, but
nothing practical has yet been developed. Steelmakers are concerned that present steps
to install shrouds and hoods on conventional coke ovens may prove to be both inadequate
and very expensive.
For particulates arising from ironmaking and steelmaking operations, the main
'- problems are tho se related to containment and disposal, because collection devices
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV -28
are reasonably efficient. An exception occurs in the case of fluxed sinter, where the
substances are abrasive toward fabric filters, corrosive toward wet scrubbers, and
bipolar (attracted to both wires and plates) in electrostatic precipitators. The cost of
providing hoods, ducts, and fans to carry metallurgical fume into collection systems is
very high, and technical problems arise in fitting and maintaining hoods near hot opera-
tions. Although blast-furnace and sinter-plant dusts may be recycled directly into the
sintering operation (coke fines are also usable), the iron oxide fumes from electric-arc
steelmaking may require careful burial if there is no adjacent sintering operation.
Similarly, certain fumes collected from steelmaking with partly galvanized scrap must
be buried because the zinc ferrites they contain are harmful to blast-furnace refrac-
tories if recycled in the sinter. The containment of elusive kish particles during cast-
ing and handling of blast-furnace iron is a knotty problem because the required drafts
tend to cool the molten iron too much.
Although the desulfurization of coke-oven gas may be accomplished by use of soda
solutions in closed-cycle equipment, present apparatus does not do a very thorough job.
It appears that the cost of cleaning this gas to anticipated standards may far exceed the
value of the gas as a fuel. It is not possible to avoid the problem by deciding to flare
or vent the gas, because the sulfur compounds would still be released into the atmo-
sphere. As in the case of coking generally, a wholly new approach may be needed.
Even if the coal-sulfur could be entirely retained in the coke, almost all of it would be
carried into the blast-furnace slag. Slag fumes are another complex problem: the
chemical mechanisms which cause slag to emit some of its sulfur are complex and
imperfectly understood.
Opportunities
Research and engineering activities conducted by steel companies to solve steel-
works emission problems sometimes appear to be directed toward the meeting of public
standards which are in turn based upon incomplete understanding of the control task.
Investigations appear to have proceeded to the point where there appears to be general
re cognition that "metallurgical fume" has characteristics different from "smoke" from
some other types of processes. This leads to a general conclusion that there might be
good reason for applying different standards of measurement or control to "metallurgical
fume" than to other types of "smoke". However, what this difference reasonably should
be or could be remains undefined. Indeed, even definition of what constitutes
"metallurgical fume II is hampered by the sparcity and lack of dependability of published
and available data on the amounts, compositions, size consist, and other characteristics
of emissions from steelworks processes. Because a large steelworks can hardly be
ignored by an air -quality organization, there has been a tendency to apply to steelworks I
emissions some measurements (such as opacity and process weight) adapted from in-
formation obtained in studies of other industries, and quite possibly not ideal for control
of steelworks I emissions. Steelmakers dislike this situation, and quite clearly they
should have the opportunity to improve future air-quality standards by determining and
disseminating more information pertaining to the actual characteristics of their
emis s ions.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
IV-29 and IV-3D
Metallurgical coke made in present-style ovens is not absolutely necessary for
the reduction of iron ore (but conventional ironmakers are doubtful about some of the
alternatives that have been suggested), By-products from coke ovens no longer have
the value that they once did. Present designs of coke ovens seem to lend themselves
more to expedients than to true control of emissions. There seems to be a substantial
opportunity awaiting the development of an entirely fresh approach to the whole problem
of providing dense, stable carbon fuel for the blast furnace. The dearth of good ideas
in this problem area would indicate an opportunity for creative engineering of alternate
ways to stabilize coal carbon. Note that the blast furnace is an effective desulfurizing
reactor. There is no inherent reason why the coaly aromatic compounds (now removed
from coal in coking) could not someday be charged (in stabilized form) into the blast
furnace and utilized directly. Nonrecovery ovens have been proposed as one way of
doing this.
Emission-control costs for particulates are raised by the costs of providing draft
from dozens of different fans, and also by the costs of collecting emissions adjacent to
many sources. There is opportunity to lower these costs and improve effectiveness in
some cases by manifolding both the drafts and their emissions into larger collectors
and bigger, more efficient fans of the type used in turbocompressors. ':' The exhaust
might be usable as prewarmed combustion air for stove burners and powerhouse boilers,
or as feed air to the blast-furnace turbocompressors. If turbine "superfans" were
used, sensitive monitoring of the dust-collection devices would be needed to guarantee
against admission of dusty air into the turbines.
The simplest du~t-collection devices, plenums and cyclones, have fallen into
disuse because they are not efficient enough. There seems to be an opportunity to re-
design and improve these types of collectors, if only to reduce the load on more
sophisticated equipment.
The above short list of potential opportunities contains no inexpensive or short-
term projects, because most opportunities in that category were followed up long ago
by steelmakers and equipment suppliers.
.Emission-control standards based on process weight per stack tend to raise legal blocks to manifolding.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
VI
W
C
:J
I-
VI
VI
VI
W
Z
W
>
I-
U
w
u.
u.
w
.........
I-
VI
o
u
>
-------
V-I
SECTION V
SPECIFIC COST/EFFECTIVENESS INVESTIGATIONS
Section 305 of the Air Quality Act of 1967 instructs the Department of Health,
'/!'ducation, and Welfare to prepare ". . . a comprehensive study of the economic impact
of air quality standards on the Nation's industries, . . . including an analysis of. . . the
cost of controlling emissions to attain such standards of air quality as may be estab-
lished. . . II The law also requires annual re -evaluation of these studies and estimates.
The economic research activities within the present Battelle project have been an early
step in compliance with this part of the law. Specifically, Battelle's task in Phase I
has been to develop and pretest methods to be used in the comprehensive study, and to
obtain preliminary estimates of control cost by application of the methods.
Before the detail and results of this Phase I cost study are presented in numerical
form, the progress of the study will be described in narrative for the orientation of the
reader, and the elements of the study method will be discussed.
At the outset of the investigation, the plan was for project teams from Battelle
and from the Swindell-Dressler Company to collect a large mass of information relating
to both technical and cost aspects of air-pollution control in the steel industry. Sources
of this information were to include the published literature, the files of both investigating
organizations, and data to be obtained from steel companies, manufacturers of emission-
control equipment, governmental bodies, and any other appropriate places.
To facilitate communication with the steel industry, NAPCA appointed an Industry
Liaison Committee of 11 men, each representing a major steel company. Early in the
project, a meeting was held with this Committee to explain the objectives of the work
and to solicit cooperation by. the steel industry. During the course of the project, two
interim reports (issued by Battelle and Swindell-Dressler to NAPCA) were distributed
to each member of the Committee and were discussed at two joint meetings of the
Committee with NAPCA representatives and members of the project team. Drafts of
the final technical and economic reports were also distributed to members of the
Committee and discussed in a fourth meeting.
Although the Industry Liaison Committee was advisory, the interest of its mem-
bers and the importance of the companies represented affected the data-gathering effort.
Field visits and other contacts between the project team and steelmakers were mostly
to plants represented on the Liaison Committee and were at times and places arranged
by Committee members.
After the first few months of the study, it became apparent that extensive cost
information in complete and usable form was not available either from the literature
or from contact with the steel companies. ':' Information for these cost studies must
be complete with respect to (1) general description of the nature of the specific steel-
making process, including approximate throughput in tons or other units per unit of
time; (2) general description of the air-pollution control system, including flow ratings
as appropriate; (3) capital cost, together with the year of installation; (4) operating
. The remainder of this discussion is concerned solely with the cost-versus~effectiveness part of the study. Technical studies are
covered in a complementary report.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
v-z
cost; and (5) effectiveness of the system, for example in terms of weight of particulate
matter per unit of exhaust from the control apparatus. Although much information
bearing on these aspects was collected from the literature and from cooperating steel
companies, in only 16 cases was the cost information regarding a single process seg-
ment complete enough to permit direct analysis and interpretation of the cost of that
ins talla tion.
The original intention had been to collect sufficient inputs of cost information to
permit direct statistical analysis and the development of specific cost models that could
be used to predict costs for other situations. The sparsity of complete data from the
expected sources required a modification of procedure. After it became apparent that
data would not be available in sufficient quantity to serve directly as a base for prepara-
tion of cost models, the Swindell-Dressler staff undertook to develop a body of cost
estimates using a number of sources, The estimates covered capital and operating
costs for air -pollution control equipment. They were developed by applying the tech-
niques of engineering-project estimation to the available data for a number of steel-
works process segments. Because of the impossibility of allowing for all eventualities
that might affect costs in specific situations, the Swindell-Dressler estimates were
prepared for typical situations and for installations of new process facilities (except in
the case of open-hearth steelmaking). Because most emission-control equipment is
engineered to clean effluent to about 0.05 grain of particulates per standard cubic foot,
the estimates were based on this reference point. Swindell-Dressler also developed
both theoretical and practical estimates of the changes in cost that might be expected
if other levels of effectiveness of control were applied.
The results of the Swindell-Dressler studies and estimates are presented in
Appendix C.
The Swindell-Dressler estimates were then subjected by Battelle to an analysis
that led to the development of several primary cost models for each process segment
for which Swindell-Dressler had prepared cost estimates. Each such model is a mathe-
matical expression that can be solved for the annual cost of controlling emis sions from
a particular steelworks process segment. To use the primary model for a situation
where cost is to be estimated, two parameters are needed: (1) annual tonnage through-
put for the process segment (or holding capacity for batch processes) and (2) effective-
ness of the control equipment expressed as level of emissions or other quantifying
criteria. If these two parameters can be stated, then to the degree that the model is
correct, solution of the mathematical expression yields an estimated value for the
annual cost of the control applied or to be applied. Annual cost calculated in this
manner from the model combines (1) fixed charges, the annualized portion of capital
and capital-related cost, >:' and (2) annual operating costs.
With completion of the modeling of the Swindell-Dressler estimates, the study
entered a pretesting phase in which the object was to compare costs as reported by steel
companies for actual installations with calculated cost estimates derived by applying the
model to the throughput and effectiveness data for the installations. The concept was
that to the extent that reported costs agreed with costs calculated from a model, the
primary model derived from Swindell-Dressler's estimates could be demonstrated to
have reasonable validity for use in predicting costs in other situations.
"Taken at 20 percent of 1968-based capital costs.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-3
Battelle encountered two important problems in this type of pretesting and predic-
tion. First, although the models were based on a large background of what is thought
to be valid and reliable engineering and cost information, they were developed for
typical stylized systems that almost certainly would differ in design and environment
from any particular and specific steelworks case. Second, although cost and effective-
ness data may be reported as they apply to a specific set of conditions, experience dur-
ing this study has shown that there are several reasons why such reported figures may
deviate substantially from "typical". Experience has also shown that such deviations
usually are not identified when the data are submitted. Because of the interaction of
these two types of problems, it is reasonable to expect that when anyone specific instal-
lation is considered, the annual cost computed by use of the primary model might differ
substantially from the reported cost. However, if the model is reasonably correct for
the typical situation used in deriving it, then testing of the model against a large number
of accurate reported cost values should show reasonably good statistical agreement.
METHODOLOGY
The following elements of method were applied to the cost/effectiveness
Phase I:
(1)
studies of
Swindell-Dressler prepared engineering estimates of costs for air-
pollution control.
(2)
The form of a representative model was defined in mathematical terms,
and a version of this model was applied to the Swindell-Dressler estimates.
(3) Data were gathered from several sources for use in pretesting.
(4) The findings from modeling and pretesting were interpreted.
This part of Section V describes the procedures and results obtained as experi-
enced for specific inquiry into 15 selected steel-industry process segments. Section VI
suggests recommended future procedures and methods as determined from the
experiences of Phase I.
Engineering Estimates
The final report of Swindell-Dressler Company (Appendix C) is a basic reference
for the modeling and pretesting studies of Phase I. It contains engineering estimates
of capital and operating costs for several types of air-pollution controls as applied to a
number of steelworks processes. It also contains information about the factors that
may cause actual control costs to vary from case to case, especially in the event of
allowable emissions higher or lower than 0.05 grain per cubic foot.
The Swindell-Dressler estimates are representative of the type that might be pre-
pared by NAPCA or contractors wherever control-cost data are required but are not
available from industry sources. Note, however, that Swindell-Dressler was unable to
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-4
prepare similarly accurate estimates of control cost for situations where control
technology has not been worked out, e. g., for the charging of coke ovens. It is also
to be noted that the estimating procedure, which is based on experience, does not neces-
sarily extrapolate well to situations where experience is meager. (For example, the
Swindell-Dressler engineers express concern about the accuracy of estimating costs for
the unusual application of fabric filters to open-hearth furnaces.) Caution should be
exercised in preparing and using estimates in situations where there is meager expe-
rience and where twofold to threefold errors may occur.
The Swindell-Dressler estimates produced for the Phase I studies are a useful
body of information for the following three purposes:
(1) They des cribe the probable costs of controlling certain forms of
steelworks air pollution, for the typical case
(2) They permit the calculation of primary mathematical models which
summarize control costs as a function of process throughput and
effectiveness, and which facilitate interpolations
(3) They can be compared, directly or in modeled form, with costs
reported for actual situations, as a pretest of the estimating and
modeling procedures.
The degree of agreement between the modeling forms and expressions and the
estimates prepared by Swindell-Dressler helps to evaluate the modeling method as
applied to highly coherent, orderly sets of data. The degree of agreement found between
the Swindell-Dressler estimates (or models made from them, if the models are accurate)
and the costs reported by the steel industry help to determine generally the usefulness
of engineering estimates as data supplements. But these pretests do not determine the
applicability of the modeling methods to analysis and characterization (and eventually
projection) of steel industry data as such. Applicability could be determined accurately
only by statistical assessment with real cases.
Definition of the Form of the Model
Battelle proposed in November, 1968, that a simple equation of the general
form
Cost = (AZ) x (BY) x (CX) . . . x (Nm)
could be used to correlate and evaluate economic cost-effectiveness data pertaining to
either (1) capital costs or (2) annualized capital costs combined with operating costs,
for air-pollution controls. A specific adaptation of the general equation was made for
primary application and pretesting of the idea in Phase I. The adaptation proposed that
annual overall cost (operating costs plus fixed charges of 20 percent of 1968 capital
costs) may be described by
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V -5
Cost = Al x (Throughput)b x (Effectiveness)g
where cost (C) is expressed in dollars per year, A is the IIstandard" overall cost per
year at reference throughput and effectiveness, throughput (T) is expressed in millions
of tons per year with a reference point of 1 million tons, and effectiveness (E) is
expres sed in arbitrarily chosen units, with a refe rence point of E= 1 corresponding to
the national or regional norms for air -pollution control. According to these definitions,
the trivial exponent (1) on term A may be dropped to give the short algebraic form:
C = A x Tb x Eg
For a specific process segment operated at a throughput of 1 million tons per year
at normal or reference control effectiveness, C=A irrespective of the numerical values
of the exponents band g. These exponents des cribe the sensitivity of cost to variations
from reference throughput and reference effectiveness. This is an abbreviated or
minimum version of the model, adjusted to the specific purpose of modeling Swindell-
Dressler's estimates. The proposed general model is not limited to only throughput
and effectiveness factors, and it is expected that additional factors would be required
for accurate modeling of industrial data. For example, higher control costs are
expected for sintering plants operated at high basicity ratio, and an additional factor
and sensitivity exponent would be required to express this effect. The criterion for
adding factors to the model is that they must take the value 1.0 under some arbitrary
reference condition. The use of extra factors is justified wherever they can improve
the fit of the equation to the data.
The suggested primary modeling expression depends for its usefulness upon a
selective and judicious definition of E, the effectiveness coefficient, for each process
segment studied.~' Where emis sions are in the form of particulates suspended in air or
other gases, and where measurements are feasible, E may be objectively defined in
terms of the dust loading in the outlet gases emanating from the control device. For
example, many gas -cleaning systems are built to the specification that the outlet air or
gas shall co"ntain less than 0.05 grain of total particulates per standard cubic foot. This
amount corresponds roughly to the threshold of dust visibility in many cases, and
satisfies some air-quality codes. ,;<>;, Because this specification was used uniformly by
Swindell-Dressler, it was found to be convenient to define
E = 0.05
R
where R is the outlet loading of particulate matter in grains per standard cubic foot of
gas. This definition is essentially simple. It takes the value
"Industry representatives point out that the present definition of E as given here may be unsatisfactory in statistical modeling of
industrial data, because the construction or upgrading of industrial air-pollution controls is based on specific performance
criteria for specific locations and situations.
"'The origin of the 0.05 grain standard is to be found in early work in Allegheny County, Pa., which included studies by United
States Steel and Jones & Laughlin Steel on emissions from open-hearth stacks. This work was done prior to use of oxygen in
steelmaking. Much metallurgical fume, and especially oxygen-refining fume, is quite visible at a concentration of 0.05 grain
per cubic foot, thus the reference point may no l_onger be satisfactory. California's code is based upon a lower grain loading.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-6
E=.05=1
.05
for the standard or reference condition where R = .05 grain per cubic foot outlet loading,
and it varies in the appropriate sense, rising as the outlet dust loading decreases, and
vice versa. In other kinds of situations, such as for gaseous emissions or in non-
ducted systems, E may become more complex or even a matter of judgment.
A typical primary cost model, developed in this study from Swindell-Dressler
estimates, takes the form
C = $623,000 x TO. 67 x EO. 19
(This particular example is for wet scrubbers applied to control of dust in a sinter
plant.) At reference conditions of 1 million tons throughput (as sinter produced) and
0.05 grain of particulates per cubic foot in the scrubber exhausts, T (throughput) and
E (effectivenes s) both take the value 1. 0. Thus C takes the value $623,000 per year
(which equals A) for the total or overall cost, which is operating cost plus the annualized
portion of capital cost.
For any other condition of tonnage and effectiveness, appropriate values of T and
E must be inserted into the model expression and raised to the powers shown. For
example, if the rate of sinter production is 2 million tons per year, then T becomes 2,
and if the desired effectiveness is 0.03 grain per cubic foot of effluent, E becomes
.05 = 1 67
.03 . .
The exponent 0.67 applied to the new value of T indicates that the cost, C, will
increase by about 59 percent when the throuput of sinter is doubled, and the exponent
0. 19 applied to the new value of E will increase the cost by an additional 10 percent.
The new overall cost is obtained by the expression
C = $623,000 x 2°.67 x 1.67°.29 = $623,000 x 1.59 x 1.10 = $1,090,000
which combines the effects of changes in throughput and effectiveness.
The application of the general model to NAPCA's task of analyzing costs for air-
pollution control at various levels of effectiveness is best made through statistical
analysis of large amounts of industrial information. For example, data on the control
costs, throughputs, and attained effectiveness of controls in 30 sinter plants might be
reduced by regression analysis to derive best-fit values for A, b, g, and exponents for
other factors as may be required. The resulting model would then be used to project
costs for control of sinter plants on a regional or national basis. Directory information
would give (or permit estimation of) throughputs for each plant, and the values of E
would be assigned by NAPCA. The overall regional or national cost would be the sum
of individual plant costs as projected by the model. To assess costs for the whole steel
industry, a separate model could be prepared for each process segment contributing to
pollution-control costs. ,;,
"In the example of the Swindell-Dressler estimates, separate models were prepared for each type of control system too. This
would not be necessary in the analysis 'of industrial data because it could be assumed that the least costly system for a given
level of control would be installed in each case.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-7
Two defects appear in this ideal approach. The first is that some steelworks
emissions are uncontrolled or poorly controlled, and for these processes no amount of
industry assistance can produce data for analysis to form models. ':' Also, the gaps in
the steel industry's internal information systems may limit the number of inputs avail-
able for even the best-controlled process segments, In light of these situations, the
best route to analysis, although a limited one in terms of accuracy, may be to supple-
ment or synthesize data by careful study and engineering estimation procedures, In
some instances, however, these procedures can yield only order-of-magnitude esti-
mates, Section VI develops the possible methodologies in greater detail.
The mathematics of the proposed primary cost model are simple, and a discussion
of this topic aids in appreciation of the model's validity for this study. Consider three-
dimensional space, with three axes of this space intersecting at right angles. This is
so-called Cartesian space, which also corresponds to the real world. Name the axes
C, T, and E, such that various distances along each axis correspond to various values
of cost, throughput, and effectiveness from the primary model expression. Figure V-l
illustrates the axes as intersecting heavy lines.
If either T (throughput) or E (effectiveness) is zero, C (cost) must be zero, be-
cause no control cost should be incurred if no control is achieved or if the process is
inoperative. Thus the value of C, which is the height of points above the basal plane
T-E, must be zero along the T and E axes. For nonzero (and implicitly positive) values
of T and E, C has some positive value represented by a height above the T-E plane.
The following facts are known about values for C:
(l) At the location whe re both T and E equal 1, the value of C is A;
A is the height above the plane
(2) For any constant value of E, increases in T produce increases in C
(3) For any constant value of T, increases in E produce increases in C.
Logically, then, the C -values form a surface that trends upward as the values of T and
E increase. The rate and curvature uptrend depend on the exponents band g. In
Figure V -1, the surface has been depicted to show how two different curved tendencies
(one concave, one convex) can combine to form the surface of C -values.
The application of the primary model may be made entirely with conventional
"canned" computer programs requiring no detailed appreciation of the mathematics
involved. The point of reviewing the mathematics is that it may be seen that the equa-
tion proposed describes a general surface that may be curved or moved by adjusting the
values of b, g, and A, This surface has the property that it is monotonic, i. e., always
upward-trending as either T or E increases. Finally, it may be complexly curved as
shown in the figure. Clearly, this kind of model surface may be fitted to the real world
of air -pollution control costs as well as any other. Battelle proposes that this is the
simplest surface having the required properties.
.Emissions in the uncontrolled category include a number of less-significant contributors to the overall problem.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
- ------
V-8
u
.ll1cr.
eoSil1g
Ihr~
'PUt
I I
.
. et'\ess,
.. \\ec\ \"
reas\t'\Q
lnC
o
C = A. Tb . E'; ( b I)
FIGURE V-I. GENERAL VIEW OF A CURVED SURFACE IN
CAR TESIAN SPACE
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-9
Data Gathering
Two different approaches to the gathering of cost data from steel companies were
tried by Battelle. The first took the form of direct all-day interviews conducted between
September, 1968, and February, 1969. Twelve steel companies were visited; ten of
these were represented on the Liaison Committee. These visits were intended to be
comprehensive in two senses - they were to cover all process segments in each plant,
and they were to obtain all relevant technical and economic data.
For purposes of the study of cost versus effectiveness for air-pollution controls,
the all-day visits were not particularly productive. Three kinds of data "blocks" were
encountered:
(1) Technological blocks were encountered wherever air -pollution problems
have not been solved in practical ways. (A large number of steelworks
emis sions fall into this category. )
(2) Information blocks appeared in the case of newly-installed control
equipment having no cost history, and also where plant accounting
systems were not amenable to read-out or analysis of operating
cos ts for air -pollution controls.
(3) Policy blocks were encountered in varying degrees from company to
company. (Many steel companies characteristically do not reveal data
descriptive of capacities, rates of operation, or costs. The restrictive
policies are of long standing, and apply especially to inquiries regarding
costs. )
Most larger steel companies do exchange current technical and operating statistics
(excluding costs) on an intramural basis through the American Iron and Steel Institute,
but the usual public release of this information is in the form of industry totals and
averages published from time to time by the AISI. In April, 1969, the Board of Directors
of AISI approved Battelle's use of certain quarterly reports for the last two quarters of
1968. The approval was granted as a one-time, nonrecurring use for the purpose of
these studies, and certain confidentiality precautions would be observed. Battelle was
unable to make use of the AISI quarterly operating reports in this present study because
access to the reports was granted too late in the contract period.
A second data-gathering approach was made in which both the scope and the depth
of inquiry were considerably abbreviated. Nine steel companies, each represented on
the Liaison Committee, were invited to prepare summary sheets describing their cost/
effectiveness experiences in controlling air-pollutant emissions from 15 selected
process segments within their plants. The blank summary sheets are bound into this
report as Appendix B. They request the following information:
(I) Identification (company and person responding':')
(II) Type and capacity of the operation for which information is offered (for
evaluation of the throughput, T)
(III) Brief non-technical description of air-pollution controls in use
(IV) Estimated capital costs and escalation factors (to determine the capital
portion of overall costs)
"Identification was included solely to permit the Battelle project staff to call back to resolve questions.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-IO
(V) Estimated operating costs (qualified as to credits included and
man-hours applied)
(VI) Estimated overall effectiveness of control (objective or subjective
estimate s according to measurability).
The selection of 15 specific proces s segments for inquiry in the second approach
was based on several criteria, including severity of the problems, repre sentation of a
number of process types, and focus upon those situations for which the steel industry
should have the best information. A worksheet decision matrix showing this selection
was bound into the Second Interim Report dated January 31, 1969, and the list of process
segments with selections marked for emphasis is given in Table V-I.
The second (summary sheet) approach to information-gathering did not overcome
the technological blocks. Battelle's willingness to accept estimates and subjective
judgments may have cleared away some informational blocks, but the summary sheets
added only a little to the information already obtained by interviewing. Some companies
indicated that the summary-sheet approach might have been more acceptable if the
sheets were submitted to and processed by AISI or by the Industry Liaison Committee.
These trusteeship possibilities were not explored because time in Phase I was too short,
but they do offer potential for future use.
The use of summary sheets was adjudged to be a good information-gathering
technique because:
(1) They constitute a specific inquiry, with uniform definitions and with
certainty that the same information is asked of every source
(2) They are convenient for the steel companies to handle because they
minimize the physical work of preparation and because they may be
mailed to other locations if the information sought is decentralized
in the company or in a given plant
(3) They are amenable to trusteeship processing if necessary.
A mass-mailing approach to information-seeking is inappropriate. The proper
use of the summary sheets is as a format for steel-company responses after personal
contact has been made. In future inquiries, the summary sheets could be left at the
end of the interview for later preparation and return.
Modeling, Pretesting, and Interpretation
The combined efforts yielded usable information pertinent to eight of the process
segments selected for intensive study. Of these, four had been studied extensively by
Swindell-Dressler. In all, 17 mathematical models were derived by analysis of the
Swindell-Dressler estimates. The specific steps taken were as follows:
(1) The Swindell-Dressler estimates (prepared for a reference outlet grain
loading (R) of 0.05 grain) were expanded to reflect two additional levels
of effectiveness. Costs were altered according to factors derived by
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V -11
TABLE V-I. AIR-POLLUTION CONTROL PROBLEMS CONSIDERED IN
PHASE I STUDIES
Coking Operations
Coal receiving and storage
Coal handling
. Coke-oven charging
. Coke-oven seals
. Coke-oven pushing
. Coke quenching
Coke handling
. By-products recovery
. Coke-oven gas systems
Ironmaking Operations
Ore and flux receiving
and storage
. Ore and flux handling
. Sintering
. Pelletizing
. Blast-furnace charging
Blast-furnace smelting
Casting and flushing
Pigging (of iron)
. Slag disposal
Power Generation
Fluid-fuel boilers
Solid -fuel boilers
Steelmaking
Scrap handling
Flux handling
Hot-metal handling
. Steelmaking
. Steel teeming
Steel working
Soaking. reheating, etc.
Primary breakdown
. Conditioning of slabs
. Hot rolling/forging
Acid pickling
Cold rolling/forging
Finishing Operations
Cleaning and degreasing
Hot-dip coating
Painting
Cold-dip coating
Electrocoating
General
Plant waste incineration
(unloading. stOcking, and reclaiming)
(blending and pulverization)
(end-door seals in particular)
(wharfing, transfer, breaking. screening)
(particularl y vents and drains)
(specifically desulfurization of the gas)
(unloading, stOcking, and reclaiming)
(transfer, comminution, sizing)
(as practiced at mining sites)
(stockhouse, skip, and bell operations)
(wind and tOp-gas systems)
(either by granulation or by dumping)
(unloading, cutting-up, baling, and charging)
(unloading, storage, reclaiming, and charging)
(melting, refining, and tapping - all methods)
(ingot practice or continuous casting)
(includes annealing and heat treatment)
(scarfing and grinding)
(includes all hot-forming operations)
(sol vent or detergent bath and spray steps)
(hot galvanizing, tinning, temeing)
(conversion coating, oiling, etc.)
(buming of garbage. tar. etc.)
. Denotes problems receiving specific economic emphasis.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-12
Swindell-Dressler from theoretical studies and other sources (also
presented in full in Appendix C). Where effectiveness-scaling factors
were not available, they were estimated by Battelle.
As an example, Swindell-Dressler reported an overall operating cost
of $277,000 for an electrostatic precipitator applied to an electric-arc
furnace of ISO-ton nominal capacity. Swindell-Dressler also reported
a variation of 10 percent in operating cost for electrostatic precipitators
applied to arc furnaces in cases where the outlet grain loading varied to
0.02 grain (costs up 10 percent) or 0.125 grain (costs down 10 percent)
per standard cubi;£oot. The following scalings resu1t~
R, grains per
cubic foot
0.05
0.02
O. 125
E = .05/R
1
2.5
0.4
Cost Estimated
$277,000 (base)
$277,000 xLI
$277, 000/1. 1
= $305,000
= $251, 000
In the case of electric-arc furnaces equipped with other kinds of gas-
cleaning equipment, Swindell-Dressler did not report effectiveness-
scaling factors. Working from other Swindell-Dressler estimates for
venturi scrubbers, and from commentary on the variance for baghouses,
Battelle estimated a factor of plus and minus 15 percent for scrubbers
and plus and minus 10 percent for baghouses.
(2) The Swindell-Dressler estimates were also scaled to obtain values for
the throughput, T. In the case of sintering, the calculation was based
on a 330-day operating year. In the case of pelletizing, 8000 operating
hours per year was used. For the three steelmaking processes, scaling
was based on an 8000-hour year with tap-to-tap times of 2.5 hours for the
arc furnace, 1 hour for the basic oxygen furnace, and 8 hours for the
open hearth.
(3) Finally, the Swindell-Dressler data were scaled according to factors
(given in the text of Appendix C) to represent steelmaking shops as
follows:
(a) An electric-are-furnace shop with two furnaces operating simultan-
eously and ducted to a single collector
(b) A basic-oxygen furnace shop with alternate use of one of two furnaces
ducted into a single collector
(c) An open-hearth-furnace shop with six furnaces, each having its
own dust-collection system.
(4) The data resulting from these scaling operations (data are given in full
in the next part of Section V) were analyzed by multiple linear regression
against the linearized form of the primary modeling equation:
log (C) = log (A) + b x log (T) + g x log (E)
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-13
The regression analyses yielded specific models, including values for
A, b, and g in each case. The closeness of fit was also reported as a
part of the regression statistics. Where it appeared that data for different
apparatus applied to the same process segment might agree well, com-
bined regressions were performed. This technique, where successful,
Ie s sens the number of individual models required.
The purpose of modeling, aside from characterization of Swindell-Dressler1s esti-
mates of control costs in a form easily interpolated, was to determine how well the
modeling expression could represent the values and variances estimated by Swindell-
Dressler and by Battelle.
Prete sting followed. The purpose of prete sting was to determine how well the
Swindell-Dressler data, as represented by the models, agreed with cost-effectiveness
information obtained from steel companies. This form of pretesting did not ascertain
how well the primary model might fit reported data from the steel companies. A mini-
mum of ten, perhaps more, inputs would normally be required to generate a model of
anyone process segment using steel-company data only. The steps of pretesting were
as follows:
(1) The data reported by steel companies were scaled to obtain values of
C, T, and E on a basis comparable with the scaling used for the Swindell-
Dressler data. E values were computed according to the definition E =
.05/ R, where R is grains of particulates per cubic foot of effluent gases.
T values were scaled from reported capacities by applying the same tap-
to-tap times (for batch steelmaking processes) as had been used to scale
the Swindell-Dressler figures.
(2) The coefficients and exponents (A, b, and g) obtained in the modeling of
the Swindell-Dressler estimates were applied to the throughput and
effectiveness values for each of the pretest cases reported by industry,
and estimated overall operating costs were computed. These computed
costs were then compared to costs derived by adding reported operating
costs to reported capital costs (annualized at 20 percent per annum) as
given by the industry sources. The discrepancy of comparison was cal-
culated in each case by the following formula:
(Cost Reported) - (Cost Estimated From Model)
Discrepancy =
.01 x (Cost Reported)
(Units of discrepancy were percent of reported cost. )
The specific results of all modeling and pretesting operations are presented in the
next (and concluding) portion of Section V. Battelle has not yet analyzed sufficient data
for any of the sources of steelworks air pollution. It would be premature to apply the
results of modeling the Swindell-Dressler data or any other results from Phase I to
direct characterization of regional or national control costs. From 10 to perhaps
30 reliable inputs of data representing actual installations would be required to establish
a single model with reliability suited to the purpose of projection.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-14
RESULTS AND ANALYSIS
This part of Secti(:>n V presents the specific results obtained from the application
of the proposed methodology to 15 steelworks process segments where air-pollution con-
trols are needed. The selection of these segments was illustrated in Table V-I. The
data from the Swindell-Dressler estimates were extensively scaled, as has been
described, and the data from actual installations in steel plants were also scaled in the
case of batch processes such as steelmaking. Data submitted by steelmakers were not
identified as to source, and information offered in incomplete form (generally lacking
some costs) were not used at all. A shortage of industry data in some instances limited
the amount of pretesting that could be accomplished. Seven of the 15 topics were not
represented by industry cost data.
The following discussions and tables are arranged topic by topic.
of the results of economic modeling and pretesting from Phase 1.
They present all
Coke-Oven Charging
Swindell-Dressler sought to prepare estimates for air-pollution controls applied to
coke-oven charging, but was unable to do so for lack of cases to study. There are a
few partially successful control systems in Canada and Europe, but most are not very
effective, and few or no cost or effectiveness data have been released. Similarly, steel-
industry sources in the United States responded that no true controls exist for air pollu-
tion from this process segment.
Coke-Oven End-Door Seals
Sources within the steel industry reported the existence of maintenance programs
for coke -oven door seals. There may be a mile or more of such seals on a battery of
100 coke ovens. Two specific reported figures were as follows:
Annual Cost
Coke Production
Reported Effectiveness
$243, 700
1 . 4 million tons
"Best available technology yields
inadequate control"
$68, 700
O. 84 million tons
"Meets public standards fully"
The above costs work out to 17.4 cents per ton of coke in the first case, and 8.2 cents
per ton in the second case. The degree to which these reports were not coherent with
one another illustrates the problem of collecting data. Although most companies have
seal-maintenance programs, the results are quite variable. Additional inputs to this
study were blocked for lack of information.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
. "
I
I
V-15
Coke -Oven Pushing
Swindell-Dressler could not locate, nor did industry sources provide, cost-
effectiveness information on the subject of pushing of coke. Two industry sources stated
that their efforts center on coking the coal thoroughly and evenly so that no uncoked coal
(the principal source of smoke) is pushed. The ability of a plant to do this depends upon
factors such as the required operating rate, the regularity of demand for coke, the rank
and quality of coals used, and the age, condition, and design of the ovens. Quantitative
economic study of cost versus effectiveness was blocked by technological deficiencies.
.lnere are apparently no specific controls applied in the United States.
Quenching of Coke
One source in the steel industry reported information on the use of louvered baffles
within the quenching towers in one plant. The baffles mechanically restrain the coke
particles driven upward by thermal draft during quenching. Costs at the cited plant
(with capital costs annualized at 20 percent) amount to $12,600 annually. Operating costs
were reported to involve only minor maintenance. The plant produces 1. 4 million tons
of coke per year; thus, costs are about 1 cent per ton of coke. There are other plants
with similar installations, but further study was blocked by lack of information on the
part of the steel companies; most installations of quench-tower baffles are quite recent.
By-Product Processing
Steel-industry sources furnished information about two installations covering dif-
ferent sectors of the by-product plant. One such installation (which ignites and flares
s"urplus coke-oven gas) costs about 39 cents per ton of coke to own and operate. A dif-
ferent installation (for which no costs were reported) collects, scrubs, and burns fumes
from the processing of tar. Neither installation is a comprehensive control. Quantita-
tive study of this topic was blocked primarily because overall control of by-products
fumes is not yet practiced.
Coke -Oven Gas Systems
There is some equipment installed in this country for the purpose of desulfurizing
coke -oven gas, but some of it is idle. ':' Industry sources gave costs for two installations
(both old) which annualize to $96 and $23, respectively, per million cubic feet of gas
desulfurized. The effectiveness of these installations, which are based on two different
ways of absorbing hydrogen sulfide in a soda solution, is only about 65 to 70 percent
(as removal of inbound hydrogen sulfide). It was concluded that even partial desulfuriza-
tion carries a cost of from 4 to 18 cents per million Btu in the gas.
The advanced age of the less-expensive installation (dating originally from 1928,
with modifications and rebuilds in 1940 and 1948) makes the annualization of its capital
costs unreliable. The higher figure (18 cents) is probably more representative, but the
'Stream desulfurization of coke-oven gas concentrates but does not alleviate the air-pollution problem unless the sulfur is
recovered from the hydrogen sulfide.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-16
fuel value of coke-oven gas is only about 35 cents per million Btu (more or less depend-
ing on inte rnal accounting practices). More complete desulfurization could double costs;
thus Battelle concluded that the cost of completely desulfurizing coke -oven gas would
probably exceed its nominal value as a fuel. Note that the existence of the coking pro-
cess obligates the steelmaker to burn or dump this gas in one way or another. Further
quantitative study of costs was blocked by the sparsity of control systems and by the
lack of cost information on hand within the steel companies.
Ore and Flux Handling
The literature and the steel companies which were contacted reported no compre-
hensive controls. One respondent reported the occasional use of fog nozzles to dampen
dusty fluxes. Coal is considered to be more of a handling problem than ore or flux
because coal particles average smaller and lighter, hence become airborne more read-
ily. Analysis of the economics for control of this process segment was blocked by lack
of comprehensive technology and by inadequate cost information on present practices.
Sintering
Swindell-Dres sler reported on the estimated cost of controls for both the draft sys-
terns and the materials -handling areas of sinter plants. Two levels of throughput (1000
and 6000 tons per day) were considered. These figures were annualized on a 330-day
basis to 0.33 and 2.0 million tons per year, respectively. Factors of cost variation
with varying control effectiveness were estimated by Swindell-Dressler for electro-
static precipitators, and these same factors were applied by Battelle to the use of wet
scrubbers and fabric filters as well. ':<
The results of modeling and pretesting the Swindell-Dressler estimates for sinter
plants are given in Table V-2. Data were furnished by two steel companies for pre-
testing purposes. One of the reported cases was based on reworking of an old dust-
collection system; the other was based on high-efficiency collection in the materials-
handling system, but low effectiveness of collection in the draft system.
The models represented the estimated data well, and explained all of the varia-
tion in log (C). This "accuracy" was forced because there are only two values of
throughput in the Swindell-Dressler data, such that a perfect line may be drawn between
them on a plot of C versus T. When the data for wet scrubbers, electrostatics, and
fabric filters were combined for analysis, the modeling process represented log C to
within 7 percent. This corresponds to an error of up to about 25 percent in annual cost
if the combined model were used. Data received from industrial sources were inade-
quate to test the direct statistical application of the modeling method.
'The potential error introduced by this procedure is moderate. If the value of the g exponent is in fact 0.1 or 0.3 instead of the
0.2 derived by assumption, the maximum error (in the range of E stUdied) is 11 to 15 percent for C.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V -17
The estimated standard cost of controls (roughly 40 to 60 cents per ton of sinter)
is almost 10 percent of the value of the sinter to ironmakers.
TABLE V -2.
DATA, MODELING RESULTS, AND PRETESTS FOR
SINTERING PLANTS
Control Used
Cost per Year
Throughput,
million tons
Grains Dust
at Outlets
Effectiveness
Coefficient
Low-energy
wet scrubbers
E1e ctros ta tic
precipitators
Fabric
filters
Data: As Expanded With Effectiveness Factors
$
251 000 = C(a)
,
298,000
353,000
827,000
989,000
1, 184,000
0.33 = T
0.33
0.33
2.0
2.0
2.0
O. 125 = R
0.05
0.02
0.125
0.05
0.02
0.4 = E
1.0
2.5
0.4
1.0
2.5
0.4
1.0
2.5
0.4
1.0
2.5
0.4
1.0
2.5
0.4
1.0
2.5
Models Obtained by Regression Analysis of Above Data
182,000 0.33
217,000 0.33
259,000 0.33
590,000 2.0
708,000 2.0
851,000 2.0
162 OOO(a) 0.33
,
193,000 0.33
231,000 0.33
560,000 2.0
673,000 2.0
810,000 2.0
O. 125
0.05
0.02
O. 125
0.05
0.02
O. 125
0.05
0.02
O. 125
0.05
0.02
C = $623,000 x T' 67 x E' 19
C =$449,000xT.66xE.20
C = $417,000 x T' 69 x E' 20
C = $489,000 x T' 67 x E' 19
For low-energy wet scrubbers:
For electrostatic precipitators:
For fabric filters:
For combined data:
Results of Pretesting
Reported Industry Data Calculated
Cost Throughput Effectiveness Cost
$ 347,000(C) 1.053 0.5 $ 406,000
$1,494,000 5.2(3units) 0.5approx.(d) $1,701,000
Percent
Discrepancy(b)
-17
-14
(a) Variation of cost with effectiveness was estimated.
(Reported cost) - (Calculated cost) .. b b
(b) Percent discrepancy = m thLS and su sequent ta les.
.01 x (Reported cost)
(c) Estimated in advance. of installation.
(d) Materials handling system effectiveness = 10; windbox system effectiveness = 0.15. (Modeled separately for this pretest. )
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
l~ ---
I
I
V-18
Pelletizing
Swindell-Dressler estimated costs for wet-scrubber controls in the materials-
handling portion of a 1. 5 million ton-per-year pellet plant, and for cyclones on the dryer
exhausts in the same plant. Swindell-Dressler also furnished estimates for a cyclone
control system in a 60 ton-per-hour plant using a shaft furnace. These latter data were
scaled to 0.4 million tons annually on the basis of 8000 operating hours per year. A
variation cost with effectiveness amounting to 20 percent for a 2. 5-fold change in aver-
age outlet particulates loading was estimated by Battelle to facilitate modeling, as pre-
sented in Table V -3. This model was also forced; no true fitting was involved. At any
rate, the estimated costs did not include dust controls for the loading of boats or trains
at the pellet-shipping point, and this may be the most important problem.
Steel companies furnished no information on pelletizing-dust controls.
TABLE V-3.
DA TA AND MODELING RESULTS FOR
PELLETIZING PLANTS
Controls Used
Cost per Year
Throughput,
million tons
Grains Dust
at Outlets
Effectiveness
Coefficient
Data: As Expanded With Effectiveness Factors
Cyclones and $143, 700(a) 0.5 0.125 0.4
wet scrubber 172, 500 0.5 0.05 1.0
207,000 0.5 0.02 2.5
Cyclone and 176 OOO(a)
wet scrubber 1.5 O. 125 0.4
,
211,000 1.5 0.05 1.0
253,000 1.5 0.02 2.5
Model Obtained by Regression Analysis of Above Data
For combined data:
C = 196,000 x TO. 18 x EO. 20
Results of Pretesting: None - no cost data supplied for actual installations.
(a) Variation of cost with effectiveness was estimated.
Blast-Furnace Charging
Systems for the control of dusting in stockhouses and furnace charge hoppers exist,
but they differ considerably from place to place. The sources in the steel industry fur-
nished no data on capital or operating costs for this equipment, apparently because costs
for these controls are not separable from other charging and stockhouse costs. Most
controls are confined to the screening operations and transfer points within the
stockhouses.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-19
Ironrnaking Slag Disposal
Neither the literature nor the steel companies contacted by Battelle furnished cost
information on the prevention or containment of sulfurous fumes during the flushing and
disposal of blast-furnace slag. However, this topic is being researched by the American
Iron and Steel Institute at present.
Steelmaking
The general subject of air-pollution controls for steelmaking was studied carefully
by Swindell-Dressler, and the resulting estimates were the most complete of the entire
study. Similarly, steel companies furnished a relatively substantial amount of informa-
tion. Specific re sults of modeling and prete sting are pre sented separately for the differ-
ent processes.
Electric-Arc Steelmaking
Swindell-Dressler prepared estimates for the costs of controlling emissions in new
installations, for three sizes of arc furnaces, and by three different collection methods.
Specific estimates of the variability of cost with effectiveness were made for two of the
control systems, and Battelle assumed a variability factor for purposes of modeling the
third system. Data, modeling results, and pretest information are given in Table V-4.
An attempt was made to prepare a combined model for all three types of controls, but
the agreement was poor. In scaling both the estimates and the data obtained from steel
com.panies, a tap-to-tap time of 2.5 hours was used.
The cost data obtained from steel companies for use of baghouses to collect arc-
furnace fume included three sets of points. This permitted a separate solution to derive
parameters A, b, and g of the modeling expression. Although the resulting expression
is too narrowly based in data to be dependable for projections, it is interesting to con-
trast the parameters of the Swindell-Dressler baghouse model with those obtained in
the case of the data from steel companies:
Source of Data Modeled
Model Parameters
A b-L-
Swindell-Dressler; value
for g estimated by
Battelle
$285,000
$450,000
0.66
0.96
o. 11
1.5
Steel companies
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V -20
TABLE V-4. DATA, MODELING RESULTS, AND PRETESTS FOR
ELEC TRIC -ARC FURNACES
Throughput, Grains Dust Effectiveness
Control Used Cost per Year million tons at Outlet Coefficient
Data: As Expande d
High-energy $149,000 = da)
wet scrubbers 0.16 = T 0.125 = R 0.4 = E
164,000 O. 16 0.05 1.0
189,000 0.16 0.02 2.5
429,000 0.96 0.125 0.4
472,000 0.96 0.05 1.0
542,000 0.96 0.02 2.5
614,000 1.6 0.125 0.4
676,000 1.6 0.05 1.0
778,000 1.6 0.02 2.5
Electrostatic
precipitators 95,000 O. 16 0.125 0.4
105,000 O. 16 0.05 1.0
116,000 0.16 0.02 2.5
251,000 0.96 0.125 0.4
277,000 0.96 0.05 1.0
305,000 0.96 0.02 2.5
363,000 1.6 0.125 0.4
400,000 1.6 0.05 1.0
440,000 1.6 0.02 2.5
Fabric filters 77 OOO(a) o. 16 0.125 0.4
,
86,000 O. 16 0.05 1.0
95,000 0.16 0.02 2.5
247,000 0.96 0.125 0.4
274,000 0.96 0.05 1.0
304,000 0.96 0.02 2.5
354,000 1.6 0.125 0.4
393,000 1.6 0.05 1.0
436,000 1.6 0.02 2.5
Models Obtained by Regression Analysis of Above Data
For high-energy wet scrubbers: C = $505,000 x T' 61 x E' 13
For electrostatic precipitators: C = $296,000 x T' 57 x E' 11
For fabric filters: C = $285,000 x T. 66 x E. 11
Re suIts of Prete sting
Reported Industry Data Calculated
Cost Throughput(b) Effectiveness Cost
Percent
Discrepancy
$691,000
65,200
705,000
350,000
0.64
0.134
1.6
O. 171
1
1
1
2.5
$385,000
75, 600
492, 000
295,000
+44
-16
+30
+16
-
(a) Variation of cost with effectiveness was estimated.
(b) As scaled: 2.5 hours tap-to-tap.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V -21
The discrepancies are large. Although they do not detract from the estimating proce-
dures used by Swindell-Dressler, they do illustrate that the estimates may not agree
well with limited amounts of data obtained for actual installations. The indicated value
for g is especially distant from the g derived from Battelle 's assumed effectiveness
factor and is probably quite distorted by the pretest at E = 2.5, which is an unusual case.
According to the parameters, the industry data are reasonably coherent. Before
NAPCA could justify extensive use of estimating procedures, those procedures would
have to be verified by similar direct comparisons. The result is not conclusive, but it
serves as a warning against use of untested estimates such as the effectiveness -s caling
estimate applied by Battelle in this case.
From the Swindell-Dressler estimates, the cost of air-pollution controls for
electric-arc steelmaking (at standard throughput and effectiveness) appeared to be about
30 to 50 cents per ton of steel. The industrial data, however, indicated costs ranging
as high as $2 per ton of steel, even with throughput standardized to holding capacity.
Two of the pretests were poor, and indicate that the estimates may be low or that the
industry figures are not to comparable bases.
The normalization of throughput data to a tap-to-tap time of 2.5 hours (as a refer-
ence point) is based on the idea that air-pollution controls must be designed and operated
to accommodate emission peaks for batch processes. The duration of off-peak periods
is of importance secondary to the rate of emissions during the peaks.
Basic Oxygen Steelmaking
Swindell-Dressler prepared estimates of control costs for three sizes of new basic
oxygen furnaces and for three types of control on each size of furnace. These data were
normalized for operation of one of two side -by-side furnaces connected to a single sys-
tern, and throughput scaling was based on a tap-to-tap time of 1 hour and an operating
year of 8000 hours. Scaling for changes in effectiveness was based on Swindell-Dressler
estimates in the case of wet scrubbers and electrostatic precipitators, and was estimated
by Battelle in the case of baghouses. Table V -5 gives the data, the models, and all pre-
testing results.
The six available pretests showed moderate to large discrepancies, which fell to
both sides of the Swindell-Dressler estimates. Although the estimates indicated
reference-point costs of just under $1 per ton of steel, data reported by steel companies
were not coherent and showed a range of 60 cents to $1.90 per ton.
A combined model determined for wet scrubbers and electrostatic precipitators
appeared to represent the Swindell-Dressler estimates within :1:20 percent. Pretesting
of this combined model was attempted via regression analysis of the six sets of pretest
data (all refer to either wet scrubbers or electrostatics), but the resulting model param-
eters were not coherent. (This means the reported data do not fit an orderly surface.)
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
TABLE V-5.
V -22
DA TA, MODELING RESULTS, AND PRETESTS FOR BASIC
OXYGEN FURNACES
Control Used
Cost per Year
Throughput,
million tons
Effectiveness
Coefficient
Grains Dust
at Outlet
High-energy
wet scrubbers
Electrostatic
precipitators
Fabric filters
Data: As Expanded
$ 774,000 0.8 0.125 0.4
821,000 0.8 0.05 1.0
895,000 0.8 0.02 2.5
1,332,000 1.6 O. 125 0.4
1,413,000 1.6 0.05 1.0
1,540,000 1.6 0.02 2. 5
1,845,000 2.4 O. 125 0.4
1,958,000 2.4 0.05 1.0
2,135,000 2.4 0.02 2. 5
616,000 0.8 O. 125 0.4
678,000 0.8 0.05 1.0
746,000 0.8 0:02 2.5
1,107,000 1.6 0.125 0.4
1,218,000 1.6 0.05 1.0
1,340,000 1.6 0.02 2.5
1,553,000 2.4 O. 125 0.4
1,708,000 2.4 0.05 1.0
1,879,000 2.4 0.02 2.5
448, ooo(a) 0.8 o. 125 0.4
498,000 0.8 0.05 1.0
553,000 0.8 0.02 2. 5
848,000 1.6 O. 125 0.4
942,000 1.6 0.05 1.0
1, 046, 000 1.6 0.02 2.5
1,220,000 2.4 O. 125 0.4
1,354,000 2.4 0.05 1.0
1,503,000 2.4 0.02 2.5
Models Obtained by Regression Analysis of Above Data
For wet scrubbers:
C = $987,000 x T. 79 x E. 08
C = $819,000 x T' 84 x E' 10
C = $ 611, 000 x T' 91 x E' 11
For ele ctrostatic precipitators:
For fabric filters:
For combined data representing
scrubbers and electrostatics: C = $899,000 x T' 82 x E' 09
Results of Pretesting
Cost
I
$1,495,000
$2, 180,000
$1,200,000
$1,694,000
$1,270,000
$2,260,000
Reported Industry Data
Throughput{ b) Effectiveness
Calculated
Cost
$1,733,000
1,413,000
1,479,000
964,000
1,467,000
1,551,000
Percent
Discrepancy
-16
+35
-23
+48
-16
+31
2.04
1.60
2.20
0.88
2.00
1.92
1
1
0.5
2.5
1
2.5
(a) Variation of cost with effectiveness was estimated.
(b) As scaled: 1 hour tap-to-tap.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-23
The principle that a good model will represent a group of cases better than it will
represent the individuals within the group is borne out by the model for basic oxygen
furnaces. Whereas the average absolute discrepancy of pretests was 28 percent on an
individual basis, the discrepancy for the combined figures was 14 percent. Additional
pretests should continue to lower this combined discrepancy. If they do, the primary
model may be valid for projections despite poor fit to any given case.
Open-Hearth Steelmaking
The Swindell-Dressler estimates for air-pollution control costs for three sizes
of existing (not new) open-hearth furnaces and for three types of control systems were
scaled, modeled, and pretested as shown in Table V -6. Scaling overall was based on a
shop of six furnaces with separate collectors, and a tap-to-tap time of 8 hours was
assumed. Specific cost versus effectiveness factors were prepared by Swindell-Dressler
for wet scrubbers and for electrostatic precipitators, and Battelle estimated the scaling
factor for changed effectiveness of fabric filters. Cost-effectiveness data reported by
steel companies permitted four prete sts.
As in the case of electric -arc steelmaking, there was an opportunity for direct
analysis of some of the data furnished by sources within the steel industry. The last
three prete sts shown in Table V - 6 refer to use of electrostatic precipitators, and the
coherent discrepancies illustrate the internal consistency of these data. Solution of
three equations in three unknowns gave the following preliminary model:
C = $1 133 000 x T' 73 x E' 22
, ,
The corresponding m.odel from Swindell-Dressler estimates was:
C = $1 045 000 x T' 66 x E. 09
, ,
Comparison of these models indicates moderately good agreement for the factor A and
for the sensitivity exponent for throughput (T), but as in the case of the electric-arc
furnace there was a difference between the theoretical and reported exponents denoting
sensitivity to effectiveness. This was regarded as additional evidence that there is gen-
eral agreement between the mathematics of the model and the industrial facts, and also
as evidence that the theoretical determination of sensitivity to effectiveness may give
incorrect estimates.
Steelmaking:
Overview
The three principal steelmaking processes are all batch processes. It was found
that in dealing with these processes, values for throughput must be scaled arbitrarily.
A dust-collection system must be designed to trap the dust at the time of peak emission;
thus its cost depends upon the amount of emission per unit time over what may be a
short part of the steelmaking cycle. Because variations in practice and in the type of
steel being made can lead to a two-fold or greater disparity in tap-to-tap time in any
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-24
TABLE v-6.
DATA, MODELING RESULTS, AND PRETESTS FOR
OPEN-HEARTH FURNACES
Throughput, Grains Dust Effectiveness
Control Used Cost per Year million tons at Outlet Coefficient
Data: As Expanded
High-energy
wet scrubbers $ 732,000 0.36 0.05 0.25
776,000 0.36 O. 125 0.4
846,000 0.36 0.2 1.0
1,750,000 1.2 0.05 0.25
1,855,000 1.2 0.125 0.4
2,023,000 1.2 0.2 1.0
3,862,000 3.6 0.05 0.25
4,093,000 3.6 O. 125 0.4
4,463,000 3.6 0.2 1.0
Electrostatic
precipitators 490,000 0.36 O. 125 0.4
534,000 0.36 0.05 1.0
582,000 0.36 0.02 2.5
1,079,000 1.2 O. 125 0.4
1,176,000 1.2 0.05 1.0
1,282,000 1.2 0.02 2.5
2,230,000 3.6 O. 125 0.4
2,430,000 3.6 0.05 1.0
2,647,000 3.6 0.02 2.5
Fabric filters 353 OOO(a) 0.36 o. 125 0.4
,
406,000 0.36 0.05 1.0
467,000 0.36 0.02 2.5
829,000 1.2 O. 125 0.4
954,000 1.2 0.05 1.0
1,097,000 1.2 0.02 2.5
1,857,000 3.6 O. 125 0.4
2, 136,000 3.6 0.05 1.0
2,455,000 3.6 0.02 2.5
Models Obtained by Regression Analysis of Above Data
For wet scrubbers: C = $1,774,000 x T' 72 x E' 10
For electrostatic precipitators: C = $1,045,000 x T' 66 x E' 09
For fabric filters: C::: $844,000 x T. 72 x E- 15
Results of Pretesting
Cost
Reported Industry Data
Throughput Effectiveness
$2,242,000
1,706,000
1,776,000
2,340,000
1.8
1. 52
1. 98
2.7
1.6
1.6
0.8
1
Calculated
Cost
$2,840,000
1,437,000
1,608,000
2,013,000
Pe rcent
Discrepancy
-27
+16
+ 9
+14
(a) Variation of cost with effectiveness was estimated.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-25
type of steelmaking furnace, standardization is required. The throughput is properly
determined by multiplying tonnage capacity (per batch) by nominal operating hours per
year (usually 8000), and dividing by a characteristic tap-to-tap time. For a ISO-ton
electric furnace, the tap-to-tap time might be 90 minutes in a modern reinforcing-bar
plant, and 4 to 5 hours for the manufacture of aircraft-quality alloy steel with a two-
slag practice. Using the typical tap-to-tap time of 2. 5 hours for both cases gives the
same value for throughput.
150 x ~~~O = 480,000 tons per year
in each shop, although actual throughputs would be 800,000 tons per year for the rebar
plant and as low as 240,000 tons per year for the alloy-steel plant. The dust-collection
systems probably would not differ very much.
The average absolute discrepancy between pretest data and steelmaking cost
models was almost 24 percent, but the aggregate error for the 14 pretests was only
11 percent. Tpis result supports the general conclusion that modeling is more accurate
for a group of installations than for individual installations within the group.
Teeming of Molten Steel
Swindell-Dressler prepared no cost estimates for teeming-fume controls, and
steel-industry sources submitted no information. Lack of suitable applied technology
can account for this, because the only teeming-fume collectors known to Battelle are
either whole-shop exhaust systems or hoods applied to continuous-casting facilities.
Both types of equipment are rare among the information sources used to date.
Slab and Billet Conditioning
The Swindell-Dressler data and the primary models derived from them are pre-
sented for scarfing operations in Table V -7. Two control systems were studied at two
different capacities, thus the data "forced" the model and no relevant test of the model-
ing procedure was made. Volume-throughput ratings were converted to tonnage-
throughput ratings via an arbitrary scaling factor of 10 tons per year per dm gas
thr oughput.
The cost data submitted by steel companies showed an annualized cost of $95,000
per year for s carfer fume collectors applicable to a throughput of 1.6 million tons of
steel annually. For this installation, the outlet loading is controlled to 0.02 grain per
cubic foot; thus the effectivenes s ratio is about 2.5. Direct pretesting was not feasible
because the cfm-throughput rating on this equipment was not given. More information
is needed on the relationship between costs, air-volumes handled, and steel-scarfing
capacity to permit better scaling of data.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V-26
TABLE V-7.
DATA AND MODELING RESULTS FOR
SCARFING MACHINES
Control Used
Cost per Year
Throughput,
million tons
Grains Dust
at Outlet
Effectiveness
Coefficient
High-energy
wet scrubbers
Electrostatic
precipitators
Data: As Expanded
$ 86,000 0.5 0.125 0.4
101,000 0.5 0.05 1.0
122,000 0.5 0.02 2.5
144,000 1.0 0.125 0.4
168,000 1.0 0.05 1.0
203,000 1.0 0.02 2.5
70,000 0.5 0.125 0.4
83,000 0.5 0.05 1.0
98,000 0.5 0.02 2.5
104,000 1.0 0.125 0.4
123,000 1.0 0.05 1.0
145,000 1.0 0.02 2.5
Models Obtained by Regression Analysis of Above Data
For wet scrubbers: C = $170,000 x T' 74 x E. 19
For electrostatic precipitators: C = $123,000 x T. 57 x E' 18
Results of Pretesting
None
Hot-Rolling and Hot- Working of Steel
A number of hot-strip mills, especially the larger and faster ones, have dust col-
lectors at the finishing stands to pick up the fine particles of scale that are thrown into
the air during rolling. Other uses of air-pollution controls in hot-working may also
exist, but no data or cost information was furnished on any of these installations. It
was concluded that the release of suitable data was blocked because cost information on
pollution-control equipment was not separable from rolling costs.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
.
'"
z
o
I-
<
o
z
w
~
~
o
u
W
D::
-
>
-------
VI-l
SECTION VI
FINDINGS AND RECOMMENDATIONS
In the cost/effectiveness studies of Phase I, Battelle and Swindell-Dressler sought
to develop and pretest methods to be used by NAPCA in preparing national and regional
analyses of the costs of controlling air pollution from steelworks operations. This effort
led to a clarification of some of the specific problems now facing NAPCA's Division of
.J:!.,conomic Effects Research. In this Section, the problems are categorized, methods
for overcoming them are suggested, and a specific program of future work toward the
required analyses is proposed.
There are 37 different process segments as listed by Battelle for purposes of
economic study (Table V-I). To complete the research required by the Congress, costs
for controlling air pollution from each of these process segments must be determined at
several levels of effectiveness. The main determinant of future economic study methods,
and the primary determinant of the degree of accuracy to be obtained in those studies, is
the degree to which the individual air-pollution problems within each process segment
have been solved in the practical sense. Three technological categories may be
identified:
Category A.
Process segments for which practical air-pollution controls
exist and have in fact been applied in a number of steel
plants
Process segments for which technically reasonable methods of
air -pollution control may exist, but for which few or no spe-
cific applications have been made in the plants
Category B.
Category C.
Process segments for which practical air-pollution controls
have not yet been developed.
Table VI-l lists the 38 process segments according to Battelle's present opinion of their
approximate categorization (under the above criteria) for future studies.
The assignments of process segments to technological categories as given in
Table VI-l are intended as a guide to the content of future economic studies. The judg-
ments upon which these assignments are based are subject to a number of exceptions.
For example:
(1) Assignment to Category A is based on Battelle's
day standards of satisfactory emission control.
defined and subject to change.
understanding of present-
The standards are ill-
(2) Assignment to Category A does not apply for certain variants in the
processes; e. g., for the manufacture of highly-fluxed blast-furnace
sinter.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
VI-2
(3) Assignment to Category B implies technical and economic practicality:
a relative state which these studies may disprove for any given case.
(4)
Technical studies now in progress may (hopefully) alter the assignments
to Category C over a short span of time.
Accordingly, the assignments should be regarded as approximate and subject to both
exception and change.
TABLE VI-I. PROCESS SEGMENTS OF THE INTEGRATED IRON AND STEEL INDUSTRY, CATEGORIZED ACCORDING
TO THE STATE OF PROGRESS IN AIR-POLLUTION CONTROL
Category A: Practical air-pollution controls exist and are applied to many emission sites: segments amenable to statistical
an al ysis .
Coke-oven seals
Coke quenching
Coke-oven gas system
Coke handling
Flux handling
Hot metal handling
Steelmaking furnaces
Sintering
Pelletizing
Blast-furnace charging
Blast-furnace smelting
Hot-dip coating
Electrocoating
Conditioning of slabs,
Hot rolling
Acid pickling
Cold rolling
billets
Plant-waste incineration
Category B: Practical means of controlling air pollution may exist but are not widely applied; segments amenable to
engineering estimation.
Coal receiving and stOrage
Coal handling
By-products recovery
Ore and flux receiving and storage
Ore and flux handling
Pigging of iron
Soaking, reheating, etc.
Primary rolling of ingots
Scrap charging
Steel teeming
Cleaning and degreasing
Painting
Cold-dip coating
Category C: Practical means of controlling air pollution do not exist.
Coke-oven charging
Coke-oven pushing
Fluid - fuel boilers
Solid-fuel boilers
Casting and flushing
Slag disposal (from ironmaking)
Scrap preparation
Criteria for Planning Future Phases
Battelle suggests that planning of future cost and/ or economic studies, beginning
with those of Phase II, should be oriented to the technological categorization presented
in Table VI-I. The proposed criteria for planning of further studies are as follows.
( I)
They should deal with the facts and costs of actual existing control
installations fir st, and with speculative costs and uncertain tech-
nologies last.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
VI-3
This criterion permits NAPCA to concentrate first on the
actual expenditures by steelmakers, which Congress wants
to know and which the steelmakers consider to be an im-
portant input to policy formation. This criterion also em-
phasizes that NAPCA and NAPCA's contractors shall seek
to understand present conditions as fully as possible before
projecting costs into areas of little or no hard knowledge.
(2) The studies should be based as fully as possible upon the actual cost/
effectiveness experiences of the steel companies for present installations.
This criterion emphasizes the primary accuracy of measured
costs as compared to engineering estimates of cost. In deal-
ing with process segments from Categories Band C where
estimates are essential to cost studies, confidence will be
improved if scaling exponents (b and g in the model described
in Section V) may be estimated by analogy between the uncer-
tain situations and previously studied plant situations.
(3) The studies should be based on nationally dispersed input samples, and
regional estimates should be produced by projection.
(4) In no instance should projection be used to estimate control costs for a
specific process segment in a specific plant.
Whereas the use of samples drawn from the entire nation
should probably produce better overall accuracy than the use
of a smaller sample from within a region, reasoning from
the general cost model to any specific case is not valid.
Such argument would be equivalent to sewing a suit of
clothes for the average U. S. male and then expecting it to
fit most people.
(5) Continuing effort should be made to enlarge the active cooperation and
the direct assistance of steel companies in preparing the analyses.
The steelmakers have shown willingness (as a group) to
help keep these studies on firm ground technically and eco-
nomically. However, as individual companies they are
concerned that their competitive positions may suffer
through too much disclosure. Battelle is convinced that
much more disclosure is desirable, and that study plans
should be of such a nature as to encourage further disclosure.
Organization and Objectives of Phases II, III, and IV
In accord with the five criteria proposed above, Battelle recommends the continu-
ing cost-effectiveness studies for control of air pollution from steel plants should be
organized in three phases that need not be completely consecutive. Overlap is accept-
able, but the completion of Phase II should precede the final efforts on Phase III, and the
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
VI-4
completion of Phase III should precede the final estimates of Phase IV. The three pro-
posed phases, as suggested by Battelle, will deal with the process segments in their
technological categories:
Phase II would be a statistical modeling study of process segments in
Category A.
Phase III would be a mixed statistical and estimation study of process
segments in Category B.
Phase IV would be a cost study of the process segments in Category C,
conducted entirely by estimation.
In suggesting the above phasing, Battelle hopes to obtain the most-needed information as
early as possible, and then to use this information to improve the validity of the studies
that have weaker foundations. It is also intended that the passage of time (as the re-
search proceeds) will bring technical growth that will tend to promote process segments
from Category C to Category B, and from Category B to Category A. Overall, the pro-
posed phasing is a route chosen for minimum uncertainty of results.
Phase II: Study of Process Segments Now Largely Controlled
The suggested methodology for Phase II is to assemble all pos sible data de scribing
air-pollution control costs, process throughputs, and control effectiveness for the 18
process segments to be studied. Using the general mathematical model developed in
Phase I, researchers should reduce these data by regression analysis. The result of
data-processing will be 18 or more cost models, at least one for each process seg-
ment under study. Using the current AISI Iron and Steel Works Directory of the United
States and Canada and other sources, NAPCA may then compile an inventory of popula-
tion for each proces s segment, by regions. Application of the models will project costs
for the entire population of all 18 process segments, regionally and as a total for the
nation, at any selected level of effectiveness. The modeling process will also suggest
the mean effectiveness now attained where controls have in fact been installed.
Assembly of Data
Lack of availability of large amounts of steel industry cost information (the first
step of the methodology described above) was a defect in the conduct of Phase I. NAPCA,
for its part, acted from the point of view that it is advantageous to industry for these
cost studies to be made in advance of the setting of air-quality standards. NAPCA stated
that the steel industry has a stake and an obligation in helping to avoid arbitrary air-
quality decisions by giving these studies a foundation of economic understanding. Indeed,
a review of testimony will show that steelmakers and other industrialists requested this
approach when the Congress was considering air-quality legislation. .
The steel industry, for its part, did not foresee that intensive cost studies would
intrude into their cost and production records. Such records have long been considered
a part of steelmakers I private business knowledge, to be held secret for purely business
reasons from competitors, unions, suppliers, and government alike. Indeed, the
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
VI-5
AISI-sponsored exchanges of current technical operating data are considered by steel
companies to be an advanced form of cooperation. Additionally, the steel industry has
just been through a period of annoyance at the hands of the popular press; much of it
based on sensational interpretation of water-quality data released in confidence (without
control costs) to Federal and state agencies. From this viewpoint, as surances that
individual data-offerings will not be applied to individual abatement actions tend to be
received with skepticism.
Battelle suggests that these studies can be performed in a manner that is palatable
to all concerned. It is clearly possible to characterize the cost of air-pollution controls
with respect to process throughput and control effectiveness on a regional or national
scale. without perusal of private data. NAPCA or NAPCA's contractors can in fact do the
job without ever seeing the numbers. Battelle's methodology during Phase I has been
based all along on the reduction of data to descriptive mathematical expressions, and it
is the expressions, not the data, that are important. Regression analysis, properly
applied, is an automatic procedure for the digestion and precise description of moderate
to large masses of coherent data.
It follows that these studies, and particularly those of Phase II, are not in the least
affected if data are supplied and processed without examination by NAPCA or NAPCA's
contractors. The examination of individual data, like the examination of individual
beans, can serve no useful purpose except the culling-out of spoiled ones. And just as
beans are culled by electric-eye, the data can be culled while it is processed simply by
applying statistical tests of coherence during processing. This procedure may be a
trivial safeguard, for the most intelligent examination of the data is the examination
made by the accountants and engineers who prepare it.
Battelle recommends that the steel industry representatives and NAPCA repre-
sentatives should arrange mutually satisfactory ways to get data into analytical opera-
tions without disclosing the data as such. Mechanically, the best route is probably for
the companies to punch all of the data into conventional BO-column data-processing
cards. A single card can contain all of the data required for a single air-pollution
control installation:
Columns 1-5: Number of process segment for which data are
provided
Columns 6-10: Year of most recent expansion, rebuild, or
installation
Columns 11-20: Process throughput, tons of product or inter-
mediate produced per normal year
Columns 21-25: Number of process units, for batch processes
Columns 26-50: Batch capacities, in tons (separated by commas)':'
~Or other qualifying data, such as sinter basicity, if use of expanded mathematical models is contemplated.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
VI-6
Colunms 51-60:
Colunms 61-70:
Colunms 71-80:
Overall capital cost of air-pollution controls
Direct operating cost of air-pollution controls
Effectiveness of controls - form of data entry
varies with process segment.
Cards of this kind can be brought or mailed to a meeting, or submitted to a disinterested
trustee for accumulation.
Data Processing
Accumulated data cards could f0rm a large deck. This deck should then be sorted
to condense separate decks for each of the process segments or other subdivisions (there
are three subdivisions of steelmaking, for example). Decks for each process segment
may then be processed through individual programs which will perform the scaling oper-
ations and the estimation of the effectiveness coefficient, E. This intermediate step will
result in a new deck with only the process segment or sub-segment number and the input
values of C, T, and E on each card.
The secondary decks may then be fed directly into the regression-analysis pro-
gram. This program may be chosen from a number of alternative forms, or prepared
specially for the work of Phase II. Computer commands causing the taping or other out-
put of inbound data should be deleted, and tests of significance may be added to guard
against wild data entering and biasing the operations. Because regression analysis is
performed rapidly by computers, all conceivable models and options should be tried and
compared. Specifically, there are several options in the analysis of batch processes,
according to whether the sizes of individual process units are considered to be important
or not. Also, the effect of nominal sinter basicity should be tried as an expansion of
the basic model. Only a few days would be required to exhaust the possibilities so that
the original and secondary data decks may be de stroyed.
Security procedures for data processing may be arranged easily. One or more
observers representing the steel industry could be present for the entire data-processing
session, and could bring the data to the processing center and take them away or de-
stroy them at the end of the job. When the computer is set up for preprocessing and
analysis operations, the observers may catalog which data-storage units are connected
to the system. When the work is finished, the observers may supervise the clearing of
all data-storage areas, and verify the clearances by simple tests. Of course, the in-
dustry observers should be qualified systems engineers, preferably with experience on
computers similar to the one used.
Application of Cost Models
In cases where alternative modeling forms have been used to process input data
during the analytical step, the key to selection among the results is the reported lIindex
of determination", sometimes denoted as p2 (rho squared). This parameter denotes
mathematically how well each model fits the data it is supposed to represent. An in-
terim report for Phase II could be prepared at the completion of analyse s to pre sent the
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
VI-7
models obtained, to establish choices between alternative forms, and to examine the
overall validity of the procedure as given by the indices of determination.
It is conceivable that at this point some changes of direction may be required.
Process segments represented by marginal amounts of data may be cited for gathering
of additional data and reprocessing. Others may prove to be poorly fitted in spite of
generous data supply, and these process segments may be relegated to analysis by other
approache s in Phase III or IV.
Where coherent models are obtained, which should be in the large majority of
cases, they may be applied directly to the characterization of regional and national con-
trol costs. For this purpose, an inventory of process segment throughputs is required
for each area studied. An inventory of physical facilities such as sinter plants, blast
furnaces, coke ovens, and steelmaking furnaces and mills is available from the AISI in
the directory previously cited. Conversion of this inventory into throughputs may be
accomplished entirely by estimation, or (preferably) with the help of the steel companies
themselves. A convenient way to assign throughputs would be to utilize the pertinent
AISI operating reports, or appropriate figures from them. These reports have been
made available to Battelle. Effectiveness data are not required; NAPCA will stipulate
levels of effectiveness for some studies, and may use the average from the original
data in other cases such as the determination of present costs and effectiveness.
The elements of population, together with effectiveness values or stipulations,
could be entered on cards for computer processing. For each process segment, a com-
puter would be used to apply the model to each element of the population in turn, deter-
mining a cost and accumulating it into regional or national totals as required. As in-
dicated in the criteria presented early in this Section, the individual costs thus com-
puted would have no validity or usefulness of themselves, thus the computer output would
properly be confined to totals only.
Completion of the regional and national projections would end Phase II. In planning
this work, the essence of success is to arrange for data transfers satisfactory to both
the steel industry and NAPCA.
Phases III and IV: Study of Other Process Segments
The succes sful arrangement and accomplishment of data transfers in Phase II could
become the basis for continued cooperation between NAPCA and the steel industry. If
so, the work of Phases III and IV would be greatly facilitated.
In Phase III, the object would be to combine cost-effectiveness data for limited
numbers of actual installations, exponents of sensitivity determined from the experiences
of Phase II, and straightforward engineering estimation in a three -way approach to the
modeling of air -pollution controls in Category B. This work could also lead to the forma-
tion of mathematical cost models, somewhat similar to those formed in Phase I from
Swindell-Dressler's estimates.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
VI-8
Inasmuch as "blind" data proce s sing is inapplicable to Phase III, a certain amount
of reliance would be placed on companies who do have controls installed on Category B
process segments to step forth and describe their experiences to the contractor. There
would be no need to identify sources, but the actual data would be of interest in the steps
prior to modeling. The proce ss of engineering estimation in Phase III could be improved
if the estimators were invited to use specific plants as bases for estimation. Estimates
would turn out better, on average, if reviewed by plant engineers. At least six to eight
inputs to modeling of each process segment should be used. Where three actual installa-
tions may be studied, the number of estimates required would be from three to five.
Modeling, analysis, and application would be as in Phase II, with the recognition of
greater uncertainty in the answers.
In Phase IV, pure estimation is to be applied to characterization of possible costs
for controlling situations where truly practical controls do not exist. There can be no
attempt to determine what the costs will actually be. One can only hope to determine
the minimum attainable costs for existing technology. An example will illustrate the
procedure.
In the case of the charging and pushing of coke ovens, true controls do not exist
and the segments are assigned to Category C. To attain control, some apparent alter-
natives seem to be to (1) totally enclose the coke ovens or (2) make coke in some other
way. Although neither of these alternatives may appear to be truly practical, they may
be evaluated to the nearest order of magnitude - that is, we can estimate whether each
will cost $1, $10, or $100 per ton of coke. For some kinds of alternatives slightly
more precise determinations may be possible. The approximate costs may be estimated,
and after consideration of a number of alternatives, one or two may stand out as least
expensive. The cost for the least-expensive solution considered is the temporary cost
answer for that process segment.
It should be apparent that many heads should be brought to bear on this kind of
estimation. In ~hase IV, NAPCA might find it efficient to sponsor seminars for spe-
cialists in the particular processes being studied. At these seminars, the problems of
estimation would be outlined, and the group would generate and edit a list of ideas,
selecting those which are seemingly "least impractical" to be tested by estimation. The
regional and national costs re suIting from Phase IV will necessarily be inaccurate, and
may be useful mainly for assigning research priorities. However, there is always the
chance that one or more potentially practical ideas may come from the process of ex-
amining existing impractical possibilities in a serious manner.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
V\
~
D::
o
~
-I
W
W
I-
V\
LL
o
I-
V\
-I
<
.
-------
A-I
APPENDIX A
LIST OF STEELWORKS BY TYPES
PART 1. INTEGRATED STEELWORKS
(These combine the operations of smelting iron ore, conversion
of iron to steel, and shaping of raw steel into saleable forms. )
Company, Plant,
and Location
Number of
Blast Furnaces
Company, Plant,
and Location
Number of
Blast Furnaces
Alan Wood Steel Company
Swede Furnaces, Swedeland and
Ivy Rock, Pennsylvania
International Harvester Company
Wisconsin Steel, South Chicago, lllinois
3
2
Jones & Laughlin Steel Corporation
Armco Steel Corporation
Ashland, Kentucky
Houston, Texas
Hamilton and Middletown, Ohio
3
1
3
Aliquippa,
Pittsburgh,
Cleveland,
pennsyl vania
Pennsylvania
Ohio
5
5
2
Bethlehem Steel Corporation
Bethlehem, Pennsylvania
Sparrows Point, Maryland
Lackawanna, New York
Johnstown, Pennsylvania
Kaiser Steel Corporation
Fontana, California
4
5
10
7
5
Lone Star Steel Company
Lone Star, Texas
1
CF&I Steel Corporation
Pueblo, Colorado
4
McLouth Steel Corporation
Trenton, Michigan
2
Crucible Steel Company of America
Midland, Pennsylvania
National Steel Corporation
3
Weirton, West Virginia
Great Lakes, Ecorse, Michigan
4
4
nptroit Steel Corporation
Portsmouth, Ohio
Ford Motor Company
Dearborn, Michigan
3
Republic Steel Corporation
Youngstown, Ohio
Warren, Warren and Niles, Ohio
Massillon-Canton, Ohio
Cleveland, Ohio
Buffalo, New York
South Chicago, Illinois
Gulfsteel, Gadsden, Alabama
4
1
2
6
2
1
2
2
Granite City Steel Company
Granite City, Illinois
2
Inland Steel Company
Indiana Harbor, East Chicago, Illinois
8
Sharon Steel Corporation
Roemer, Farrell, Pennsylvania
2
Interlake Steel Corporation
Chicago and Riverdale, Illinois
2
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
A-2
PART 1. (Continued)
Company, Plant,
and Location
Number of
Blast Furnaces
Company, Plant,
and Location
Number of
Blast Furnaces
United States Steel Corporation
Duquesne, Pennsylvania
Edgar Thomson, Braddock,
pennsyl vania
Homestead, Rankin and Munhall,
Pennsylvania
Gary, Indiana
South Chicago, Illinois
Fairless, Fairless Hills,
Pennsyl vania
Fairfield District, Jefferson County,
Alabama
Geneva, Utah
National, McKeesport, Pennsylvania
Lorain, Ohio
Youngstown, Ohio
Duluth, Minnesota
5
6
5
12
11
3
8
3
4
5
4
2
Wheeling- Pittsburgh Steel Corporation
Monessen, Pennsyl vania
Steubenville, Ohio
3
5
Youngstown Sheet and Tube Company
Campbell, Campbell, Ohio
Brier Hill, Youngstown, Ohio
Indiana Harbor, East Chicago,
Indiana
4
2
3
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
A-3
PART II. SECONDARY STEELWORKS
(These combine the operations of steelmaking - usually by remelting
scrap - and shaping of raw steel into saleable forms.)
Company, Plant,
and Location
Number of
Melting Furnaces
Company, Plant,
and Location
Number of
Melting Fumaces
Alco Products, Incorporated
Latrobe, Pennsylvania
Allp~heny Ludlum Steel Corporation
Brackenridge, Pennsylvania
Dunkirk and Watervliet,
New York
Femdale, Michigan
Special Metals, New Hanford,
New York
Allison Steel Manufacturing Company
Tempe, Arizona
American Compressed Steel Corporation
Cincinnati, Ohio
Armco Steel Corporation
Baltimore, Maryland
Butler, Pennsylvania
Kansas City, Missouri
Sand Springs, Oklahoma
Torrance, Califomia
Atlantic Steel Company
Atlanta, Georgia
Babcock & Wilcox Company (The)
Bea ver Falls, Pennsylvania
Baldwin-Lima-Hamilton Corporation
Burnham, pennsyl vania
Bethlehem Steel Corporation
Steelton, Pennsylvania
Los Angeles, Vemon, Califomia
Seattle, Washington
Border Steel Rolling Mills, Incorporated
Vinton, Texas
2
Borg- Warner Corporation
New Castle, Indiana
Chicago Heights, illinois
4
2
3
2
1
19
8
2
3
3
3
2
5
8
10
15
Braebum Alloy Steel Division, Continental
Copper and Steel Industries, Incorporated
12
8
Braebum, Pennsylvania
9
Byers Company, A. M.
Am bridge, Pennsylvania
2
Cabot Corporation
Pampa, Texas
Cameron Iron Works, Incorporated
1
Houston, Texas
6
7
3
1
3
Carpenter Steel Company (The)
Reading, Pennsylvania
New England, Bridgepon, Connecticut
Ceco Corporation (The)
Lemont, Illinois
Milton, Pennsylvania
2
C F & I Steel Corporation
Roebling, Burlington County,
New Jersey
7
7
Columbia Tool Steel Company
Chicago Heights, Illinois
Continental Steel Corporation
Kokomo, Indiana
4
3
2
Copperweld Steel Company
Aristoloy, Warren, Ohio
2
Crucible Steel Company of America
Syracuse, New York
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
iH-
PART II. (Continued)
A-4
Company, Plant,
and Location
Number of
Melting Furnaces
Company, Plant,
and Location
Number of
Melting Furnaces
Cyclops Corporation
Mansfield, Ohio
Bridgeville, Pennsylvania
Titusville, Pennsylvania
Driver Company, Wilbur B.
Newark, New Jersey
Eastern Stainless Steel Corporation
Baltimore, Maryland
Edgewater Steel Company
Oakmont, Pennsylvania
Erie Forge & Steel Corporation
Erie, Pennsylvania
Etiwanda Steel Producers, Incorporated
Etiwanda, California
Fink! & Sons Company, A.
Chicago, Illinois
Firth Sterling, Incorporated
McKeesport, Pennsylvania
Florida Steel Corporation
Tampa, Florida
Croft, North Carolina
Harper Company, H. M.
Morton Grove, Illinois
Harsco Corporation
Harrisburg, Pennsylvania
Hawaiian Western Steel Limited
Ewa, Hawaii
Heppenstall Company
Pittsburgh, Pennsylvania
Philadelphia, Pennsylvania
Intercoastal Steel Corporation
Chesapeake, Virginia
8
10
3
5
7
1
2
2
2
7
3
2
2
3
1
2
6
1
Interlake Steel Corporation
Wilder, Kentucky
Jessop Steel Company
Washington, Pennsylvania
Owensboro, Kentucky
Jones & Laughlin Steel Corporation
Warren, Michigan
Jorgensen Company, Earle M.
Seattle, Washington
Joslyn Mfg. and Supply Company
Fort Wayne, Indiana
Judson Steel Corporation
Emeryville, California
Kankakee Electric Steel Company
Kankakee, Illinois
Kentucky Electric Steel Company
Coalton, Kentucky
Keystone Steel and Wire Company
Peoria, Illinois
Knoxville Iron Company
Knoxville, Tennessee
Laclede Steel Company
Al ton, Illinois
Latrobe Steel Company
Latrobe, Pennsylvania
Le Toumeau, Inc. (R. G.)
Longview, Texas
Lukens Steel Company
Coatesville, Pennsylvania
Mesta Machine Company
West Homestead, Pennsylvania
New Castle, pennsyl vania
3
6
2
6
2
3
3
2
1
5
3
2
5
3
9
5
4
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
A-5
PART II. (Continued)
Company, Plant,
and Location
Number of
Melting Furnaces
Company, Plant,
and Location
Number of
Melting Furnaces
Mississippi Steel Corporation
Flowood, Mississippi
2
Soule Steel Company
Long Beach, California
1
National Forge Company
Irvin, Pennsylvania
3
Southern Electric Steel Company
Birmingham, Alabama
2
North Star Steel Company
St. Paul, Minnesota
1
Southwest Steel Rolling Mills
Los Angeles, California
2
Northwest Steel Rolling Mills, Inc.
Seattle, Washington
Structural Metals, Inc.
2
Seguin, Texas
2
Northwestern Steel & Wire Company
Sterling, Illinois
5
Texas Steel Company
Fort Worth, Texas
4
Oregon Steel Mills
Portland, Oregon
3
Timken Roller Bearing Company (The)
Canton, Ohio
9
Owen Electric Steel Company of
South Carolina
Cayce, South Carolina
2
Union Electric Steel Corporation
Carnegie and Harmon Creek,
Pennsyl vania
1
Pacific States Steel Corporation
Union City, California
4
United States Steel Corporation
Johnstown, pennsyl vania
Torrance, California
3
4
Phoenix Steel Corporation
Phoenixville, Pennsylvania
Claymont, Delaware
5
7
Vasco Metals Corporation
Monroe, North Carolina
Latrobe, Pennsylvania
Monaca, Pennsylvania
7
11
1
Pollak Steel Company (The)
Marion, Ohio
1
Washburn Wire Company
Phillipsdale, Providence County,
Rhode Island
4
Porter Company, Inc., H. K.
Connors, Birmingham, Alabama
Huntington, West Virginia
3
2
Washington Steel Corporation
Roanoke Electric Steel Corporation
Roanoke, Virginia
Houston, Pennsylvania
2
3
Wickwire Brothers, Inc.
Roblin Steel Corporation
Dunkirk, New York
Cortland, New York
2
3
Simonds Steel Division, Wallace
Murray Corporation
Lockport, New York
5
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
A-6
PART III. STEEL PROCESSING PLANTS
(These obtain raw or semifinished steel from other plants
for processing to saleable shapes. List is edited by
omission of plants with no hot-fabricating operations.)
Company, Plant, Number of Company, Plant, Number of
and Location Reheat Fumaces and Location Reheat Fumaces
Allegheny Ludlum Steel Corporation Lockhart Iron & Steel Company
West Leechburg, Pennsylvania 1 Vulcan Iron, McKees Rocks,
Pennsyl vania 3
American Chain & Cable Company, Inc.
Page, Monessen, Pennsylvania 2 Michigan Seamless Tube Company
South Lyon, Michigan 1
Arkansas Steel Corporation Rosenberg, Texas 2
Magnolia, Arkansas Unknown
Missouri Rolling Mill Corporation
Armco Steel Corporation St. Louis, Missouri 2
Am bridge, Pennsylvania 4
Zanesville. Ohio 4 Northem Steel Incorporated
Medford, Massachusetts 1
Babcock & Wilcox Company (The)
Milwaukee, Wisconsin 5 Phoenix Manufacturing Company
Joliet, Illinois 3
Bethlehem Steel Corporation
Lebanon, Pennsylvania 1 Poor & Company. Inc.
Bums Harbor, Porter County. Troy, New York 2
Indiana 12
South San Francisco, California 3 Republic Steel Corporation
Borg-Warner Corporation Elyria, Ohio 2
Cleveland, Ohio 2
Chicago, Illinois 2
Franklin, Pennsylvania 1 Roblin Steel Corporation
Copperweld Steel Company North Tonawanda, New York 1
Shelby, Ohio 4 Tredegar Company
Glassport, Pennsylvania 2
Richmond, Virginia 2
Cyclops Corporation United States Steel Corporation
Pittsburgh. Pennsylvania 10
Irvin, Dravosburg, Pennsylvania 5
Hoster Investment Company Gary Sheet & Tin, Gary, Indiana 5
Pittsburgh, Califomia 1
Oklahoma City, Oklahoma 2 Ellwood, Pennsylvania 11
Fairless, Fairless Hills, Pennsylvania 1
Jersey Shore Steel Company Gary Tube, Gary, Indiana 3
South Avis, Pennsylvania 1 South, Worcester. Massachusetts 1
Cleveland. Ohio 3
Keystone Steel and Wire Company Joliet, Illinois 5
Chicago Heights, Illinois 1 Wheeling- Pittsburgh Steel Corporation
Lake Erie Rolling Mill, Inc. Benwood, West Virginia 1
Tonawanda. New York 1
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
A-7
PART IV. IRONWORKS
(These are based upon the operations of smelting iron from
iron ore, or ferrous alloys for use in steelmaking. )
Company, Plant, Number of Number of
and Location Coke Ovens Blast Furnaces
Interlake Steel Corporation
Erie, Pennsylvania 58 1
Toledo, Ohio 151 2
Jackson Iron and Steel Company (The)
Jackson, Ohio None 1
Lavino and Company, E. J.
Sheridan, pennsyl vania None 1
Lynchburg, Virginia None 2
National Steel Corporation
Hanna, Buffalo, New York None 4
Republic Steel Corporation
Troy, New York None 1
Thomas, Birmingham, Alabama 65 2
Shenango, Incorporated
Neville Island, Pennsylvania 105 2
Sharpsville, Pennsylvania None 2
Tonawanda Iron Division, American Radiator
and Standard Sanitary Corporation
North Tonawanda, New York None 1
United States Pipe and Foundry Company
Birmingham, Alabama 240 3
United States Steel Corporation
Cleveland, Ohio None 2
Woodward Iron Company
Woodward, Alabama 256 4
Rockwood, Tennessee 44 2
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
A-8
PART V. COKE PLANTS
(These are based upon the conversion of coal to coke and the
recovery of organic byproducts from the coking process.
Beehive plants are omitted as are plants producing
foundry coke. )
Company. Plant,
and Location
Number of
Coke Ovens
Alabama Byproduct Corporation
Tarrant, Alabama
203
Allied Chemical Corporation
Ashland, Kentucky
Buffalo, New York
Ironton, Ohio
196
120
168
Citizens Gas & Coke Utility
Indianapolis, Indiana
168
Connecticut Coke Company
New Haven, Connecticut
70
Donner-Hanna Coke Corporation(a)
Buffalo, New York
200
Eastern Gas & Fuel Associates
Philadelphia, Pennsylvania
74
Empire Coke Company
Holt, Alabama
60
Indiana Gas & Chemical Corporation
Terre Haute, Indiana
60
Koppers Company, Inc.
Kearny, New Jersey
St. Paul, Minnesota
125
65
Milwaukee Solvay Div. Pickands Mather
and Company
Milwaukee, Wisconsin
200
Sharon Steel Corporation
Fairmont, West Virginia
60
United States Steel Corporation(b)
Clairton, Pennsylvania
1375
(a) Owned jointly by Republic Steel and National Steel.
(b) Includes alloy blast furnace and rolling mills.
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
'"
t-
W
W
:I:
'"
>-
D:::
-------
B-1
SUMMARY A -- CHARGING OF COAL TO BYPRODUCT COKE OVENS
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
Tn:.
1.
Identification
Company:
Data entered by:
II.
Capacity of Coking Operations Represented by This Summary
Typical annual coke production, net tons, wet =
Number of batteries =
Number of ovens =
Description of Equipment and Activities Represented
On the reverse side of this sheet, briefly describe equipment and/or activ-
ities used to control smoke and dust during charging of coke ovens. Use
block diagrams to illustrate sequential equipment; omit numerical data.
TU.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land, and apparatus not specifically used to control pollution.
Include equipment, installation, engineering, research, start-up, mainten-
ance tools and facilities, inventories, and incremental costs for changes
in production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
Year of start-up:
; Annual escalation has averaged about
%
V.
Estimated Operating Cost of Air-Pollution Control System
Exclude depreciation, taxes, interest, insurance, and all other fixed costs.
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon production cost. Credit materials re-
claimed and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
17T
Estimated Overall Effectiveness of Controls Represented by This Summary
Check one:
----- Air pollution from charging is completely suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
For ducted systems, indicate range of outlet grain loadings observed:
----- grains to ----- grains particulates per cubic foot.
from
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-2
SUMMARY B -- PREVENTION OF LEAKAGE FROM END DOORS OF COKE OVENS
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
III.
I.
Identification
Company:
Data entered by:
II.
Capacity of Coking Operations Represented by This Summary
Typical annual coke production, net tons, wet =
Number of batteries =
Number of ovens =
Description of Equipment and Activities Represented
On the reverse side of this sheet, briefly describe equipment and activ-
ities used to minimize leakage from end doors of the coke ovens. Omit
numerical data.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land, and apparatus not specifically used to prevent leakage.
Include equipment and tools, installations, engineering, research, inven-
tories, and changes in production equipment to accommodate leakage-control
programs or devices.
Estimated installed capital cost = $
Year of installation: Annual escalation has averaged
%
V.
Estimated Operating Cost of Air-Pollution Control Program.
Exclude depreciation, taxes, interest, insurance, and all other fixed costs.
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon production cost. Credit any noticed
increase in yield of gas and byproducts.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overall Effectiveness of Controls Represented by This Summary
Check one:
----- Leakage from oven doors is completely eliminated
Control program described is thorough and meets public standards fully
Control program is based on best available technology but is inadequate
Other:
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-3
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
SUMMARY C -- PUSHING OF COKE FROM BYPRODUCT COKE OVENS
I.
Identification
Company:
Data entered by:
II.
Capacity of Coking Operations Represented by This Summary
Typical annual coke production. net tons. wet =
Number of batteries =
Number of ovens =
III.
Description of Equipment and Activities Represented
On the reverse side of this sheet. briefly describe equipment and/or activ-
ities used to control smoke and dust during pushing of coke ovens. Use
block diagrams to illustrate sequential equipment; omit numerical data.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land. and equipment not specifically used to control pollution
Include equipment. installation. engineering. research. start-up. mainten-
ance tools and facilities. inventories. and incremental costs for changes
in production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
Year of start-up:
; Annual escalation has averaged about
%
V.
Estimated Operating Cost of Air-Pollution Control System
Exclude depreciation. taxes. interest. insurance. and all other fixed costs
Include labor. supervision. maintenance. utilities. materials. major repair
allowances. and direct effects upon production cost. Credit materials re-
claimed and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year = Credits total $( )
VI. Estimated Overall Effectiveness of Controls Represented by This Summary
Check one:
----- Air pollution from pushing is completely suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
For ducted systems. indicate range of outlet grain loadings observed:
----- grains to ----- grains particulates per cubic foot.
from
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
SUMMARY D
B-4
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
-- QUENCHING OF COKE MADE IN BYPRODUCT COKE OVENS
III.
I.
Identification
Company:
Data entered by:
II.
Capacity of Coking Operations Represented by This Summary
Typical annual coke production, net tons, wet =
Number of batteries =
Number of ovens =
Description of Equipment and Activities Represented
On the reverse side of this sheet, briefly describe equipment and/or activ-
ities used to control escape of dust during quenching of byproduct coke.
Use simple diagrams to illustrate methods; omit numerical data.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land, and apparatus not specifically used to control pollution.
Include equipment, installation, engineering, research, start-up, mainten-
ance tools and facilities, inventories, and incremental costs for changes
in production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
Year of start-up:
Annual escalation has averaged about
%
V.
Estimated Operating Cost of Air-Pollution Control System
Exclude depreciation, taxes, interest, insurance, and all other fixed costs.
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon production cost. Credit materials re-
claimed or conserved and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overall Effectiveness of Controls Represented by This Summary
Check one:
----- Air pollution from quenching is completely suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
If measurements have been made, indicate pounds coke dust lost per ton of
coke quenched (other than to sump)-----
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-5
SUMMARY E -- CONTAINMENT OF FUMES AND VAPORS FROM COKE-OVEN BYPRODUCTS
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
III.
I.
Identification
Company:
Data entered by:
II.
Capacity of Coking Operations Represented by This Summary
Typical annual coke production, net tons, wet =
Number of batteries =
Number of ovens =
Description of Equipment and Activities Represented
On the reverse side of this sheet, briefly describe equipment and/or activ-
ities used to control escape of fumes and vapors from the byproduct sys-
tems of the coke plant. Use simple diagrams to illustrate methods; omit
numerical data.
TU
Estimated Installed Cost of Pollution-Control Equipment
Exclude land, and apparatus not specifically used to control pollution
Include equipment, installation engineering research, start-up, maintenance tools and
facilities, inventories, and incremental costs for changes in production
equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
Year if start-up;
Annual escalation has averaged about ----- percent
V.
Estimated Operating Cost of Air-Pollution Control System
Exclude depreciation, taxes, interest, insurance, and all other fixed costs.
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon production cost. Credit materials re-
claimed or conserved and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overall Effectiveness of Controls Represented by This Summary
Check one:
Fumes and vapors from the byproduct system are completely contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-6
SUMMARY F -- DESULFURIZATION OF COKE-OVEN GAS PRIOR TO DISTRIBUTION
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
III.
1.
Identification
Company:
Data entered by:
II.
Capacity of Coke-Oven Gas System Described in This Summary
Typical annual production of coke-oven gas for all uses, millions of
cubic feet at I atmosphere pressure and 60 F =
Percent used to underfire ovens =
Percent sold =
Percent used in mixture with other fuel gases =
Description of Desulfurization Equipment and Process
On the reverse side of this sheet, briefly outline the steps taken to remove
hydrogen sulfide from raw coke-oven gas prior to distribution. Use block
diagrams to illustrate sequential equipment; omit numerical data.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land, and equipment not specifically used to desulfurize gas.
Include equipment, installation, engineering, research, start-up, mainten-
ance tools and facilities, inventories, and incremental costs for changes
in production equipment (e.g. booster pumps) to accommodate desulfurizing
apparatus.
Estimated installed capital cost = $
Year of start-up:
Annual escalation has averaged about ----- percent
V.
Estimated Operating Cost of Gas Desulfurization System.
Exclude depreciation, taxes, interest, insurance, and all other fixed costs
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon gas distribution cost. Credit sulfur
values reclaimed if any and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Overall Effectiveness of Desulfurization Represented by This Summary
State total sulfur or hydrogen sulfide present in unmixed coke-oven gas
after desulfurization, in units convenient to you:
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-7
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
SUMMARY G -- UNLOADINGt STORAGEt AND RECLAIMING OF IRON ORES AND FLUXES
III.
I.
Identification
Company:
Data entered by:
II.
Capacity of Bulk Handling Operations Represented by This Summary
Typical annual receipts/consumption of natural and processed iron
and limestone or dolomite for blast-furnace and sinter plant uset
tonst natural basis =
Percentage of receipts arriving during lake navigation season only =
Briefly describe unloading/stocking/reclaiming system if not based on ore
ores
net
bridges:
Description of Equipment and Activities Represented
On the reverse side of this sheet, briefly describe measures taken to sup-
press and/or collect handling dusts from iron ores and fluxes. Omit numbers.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land, and equipment not specifically applied to pollution control.
Include equipmentt installationt engineeringt researcht start-up, maintenance
tools and facilitiest inventories, and incremental costs for changes in
production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
Year of start-up:
Annual escalation has averaged about ----- percent
V.
Estimated Operating Cost of Air-Pollution Control Systems
Exclude depreciationt taxest interest, insurance, and all other fixed costs.
Include labort supervisiont maintenancet utilities, materials, major repair
allowancest and direct effects upon production cost. Credit materials
reclaimed or conserved and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overall Effectiveness of Controls Represented by This Summary
Check one:
----- Air pollution from bulk handling is completely suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-8
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
SUMMARY H
-- SINTERING OF FINE ORES, FLUXES, AND RECYCLED DUSTS
III.
I.
Identification
Company:
Data entered by:
II.
Capacity of Sintering Operations Represented by This Summary
Typical annual sinter production, net tons =
Basicity:-----% at B/A =
Number of sintering strands =
-----% at B/A =
Description of Equipment and Activities Represented
On the reverse side of this sheet, briefly describe equipment used to
control dust in all sectors of the sintering operation. Use simple
block diagrams to illustrate sequential equipment; omit numerical data.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land, and apparatus not specifically used to control pollution
Include equipment, installation, engineering, research, start-up, maint-
enance tools and facilities, inventories, and incremental costs for changes
in production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
Year of start-up:
Annual escalation has averaged about
%
V.
Estimated Operating Cost of Air-Pollution Control System
Exclude depreciation, taxes, interest, insurance, and all other fixed costs
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon production cost. Credit materials re-
claimed and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overal Effectiveness of Controls Represented by This Summary
Range of outlet grain loadings observed for main exhaust stacks, -----
to grains particulates per cubic
Range of outlet grain loadings observed
to grains particulates per
Check one:
foot.
for cooler/handling exhausts,
cubic foot
----- Air pollution from sintering is completely suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-9
SUMMARY I -- PELLETIZING OF IRON-ORE CONCENTRATES
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
III.
I.
(do not disclose pellet company name)
Identification
Reporting Company:
Data entered by:
II.
Capacity of Pelletizing Operations Represented by This Summary
Typical annual pellet shipments, net tons =
Induration furnaces:
Number =
Type:
Description of Equipment and Activities Used to Control Air Pollution
On the reverse side of this sheet, briefly describe equipment used to
control dust in pelletizing, induration, and handling operations. Use
simple block diagrams to illustrate sequential equipment. Omit numerical
data and controls applied during mining and concentration of ores.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land, and apparatus not specifically used to control pollution
Include equipment, installation, engineering, research, start-up, maint-
enance tools and facilities, inventories, and incremental costs for changes
in production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
Year of start-up;
Annual escalation has averaged about ----- percent
u
Estimated Operating Cost of Air-Pollution Control Systems
Exclude depreciation, taxes, interest, insurance, and all other fixed costs
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon production or shipping cost. Credit
materials reclaimed and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overall Effectiveness of Controls Represented by This Summary
Range of outlet grain loadings observed for induration furnaces,
to
grains particulates per cubic foot.
Check one:
Air pollution from pelletizing and pellet handling or loading is
completely suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-lO
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
SUMMARY J
-- CHARGE PREPARATION AND CHARGING OF BLAST FURNACES
1.
Identification
Company:
Data entered by:
II.
Capacity of Ironmaking Operations Represented by This Summary
Typical annual hot metal production, net tons =
Number of blast furnaces =
Overall burden is roughly
Typical coke/NTHM =
percent iron ore
pounds
percent pellets
percent sinter
percent other
III.
Description of Equipment and Activities Represented
On the reverse side of this sheet, briefly describe equipment and/or activ-
ities used to control dust in the stockhouses and furnace tops.* Exclude
bleeders as a source; omit numerical data.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment.
Exclude land, and apparatus not specifically used to control pollution.
Include equipment, installation, engineering, research, start-up, mainten-
ance tools and facilities, inventories, and incremental costs for changes
in production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
Year of start-up:
Annual escalation has averaged about
%
V.
Estimated Operating Cost of Air-Pollution Control System
Exclude depreciation, taxes, interest, insurance, and all other fixed costs.
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon production cost. Credit materials re-
claimed and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year = Credits total $( )
VI. Estimated Overall Effectiveness of Controls Represented by This Summary
Check one:
----- Air pollution from charging is completely suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
*The exact scope of this Summary is: discharge of raw materials to stockhouse
bins or pockets; stockhouse operations including screening, weighing, and
disposal of undersize; materials hoisting to furnace top; discharge of skips
and operation of distributor; opening of small and large bells. Cost of
normal sealing bells and hoppers is not considered a control cost.
BATTELL.E MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-ll
SUMMARY K -- PREVENTION OR CONTAINMENT OF SULFUROUS FUMES FROM IRONMAKING SLAG
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
III.
I.
Identification
Company:
Data entered by:
II.
Capacity of Ironmaking Operations Represented by This Summary
Typical annual hot metal production. net tons =
Number of furnaces =
Typical slag/NTHM =
pounds
Description of Equipment and Activities Represented
On the reverse side of this sheet. briefly describe methods of slag dis-
posal and means employed to control emission of hydrogen sulfide. sulfur
dioxide. and other sulfurous gases during flushing. handling. and disposal
or accumulation of slag.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land. and apparatus not specifically used to control pollution.
Include equipment. installation. engineering. research. start-up.
ance tools and facilities. inventories. and incremental costs for
in production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
main ten-
changes
Year of start-up:
Annual escalation has averaged about
percent
V.
Estimated Operating Cost of Air-Pollution Control Systems
Exclude depreciation. taxes. interest. insurance. and all other fixed costs.
Include labor. supervision. maintenance. materials. utilities. major repair
allowances. and direct effects upon production cost. Credit materials re-
claimed (sulfur values?) and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overall Effectiveness of Controls Represented by This Summary
Check one:
Sulfurous gases from ironmaking slag are wholly suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-12
SUMMARY L --MELTING, REFINING, FINISHING, AND TAPPING OF STEEL
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
III.
I.
Identification
Company:
Data entered by:
II.
Capacity of Steelmaking Operations Represented by This Summary
Typical annual crude steel production, net tons =
Number of furnaces =
Type of furnaces =
Description of Equipment and Activities Represented
On the reverse side of this sheet, briefly describe equipment and/or activ-
ities used to control air-pollution from steelmaking operations. Use
block diagrams to illustrate sequential equipment; omit numerical data
but for BOF shops state how many furnaces operate at once.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land, and apparatus (such as waste-heat boilers) not specifically
used to control air pollution.
Include equipment, installation, engineering, research, start-up, mainten-
ance tools and facilities, inventories, and incremental costs for changes
in production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
Year of start-up:
Annual escalation has averaged about ----- percent
V.
Estimated Operating Cost of Air-Pollution Control System
Exclude depreciation, taxes, interest, insurance, and all other fixed costs.
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon production cost. Credit materials re-
claimed and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overall Effectiveness of Controls Represented by This Summary
Check one:
----- Air pollution from steelmaking operations is completely suppressed
or contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
Range of outlet grain loadings observed for stacks =
grains particulates per cubic foot.
grains to
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-13
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
SUMMARY M -- TEEMING OF MOLTEN STEEL
III.
I.
Identification
Company:
Data Entered by:
II.
Capacity of Teeming Operation Represented by This Summary
Typical annual tonnage of steel poured, net tons =
----- Continuous casting;
Ingot practice, type:
Description of Equipment and Activities Represented
On the reverse side of this sheet, briefly describe equipment and/or activ-
ities used to control smoke, dust, and fume during teeming of steel. Use
block diagrams to illustrate sequential equipment; omit numerical data.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land, and apparatus not specifically used to control pollution.
Include equipment, installation, engineering, research, start-up, mainten-
ance tools and facilities, inventories, and incremental costs for changes
in production equipment to accommodate pollution-control devices.
Estimated installed .capital cost = $
Year of start-up:
Annual escalation has averaged about ----- percent
V.
Estimated Operating Cost of Air-Pollution Control Systems
Exclude depreciation, taxes, interest, insurance, and all other fixed costs.
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon production cost. Credit materials re-
claimed ahd increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overall Effectiveness of Controls Represented by This Summary
Check one:
----- Air pollution during teeming is completely suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based on best available technology but are not adequate
Other:
If a ducted exhaust system is used to control teeming fumes, give range of
outlet grain loadings observed:
per cubic foot.
grains to
grains particulates
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-l4
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
SUMMARY N -- CONDITIONING OF SLAB. BLOOM. OR BILLET SURFACES
III.
I.
Identification
Company:
Data entered by:
II.
Capacity of Conditioning Operations Represented by This Summary
Typical annual throughput of semifinished steel. net tons =
Type of conditioning: ----- grinding
scarfing by----- hand;
machine
Description of Equipment and Activities Represented
On the reverse side of this sheet. briefly describe equipment and/or activ-
ities used to control airborne dusts during conditioning of steel. Use
block diagrams to illustrate sequential equipment; omit numerical data.
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land. and apparatus not specifically used to control pollution.
Include equipment. installation. engineering, start-up. research. mainten-
ance tools and facilities. inventories, and incremental costs for changes
in production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
Year of start-up;
Annual escalation has averaged about ----- percent
V.
Estimated Operating Cost of Air-Pollution Control Systems.
Exclude depreciation, taxes. interest. insurance. and all other fixed costs.
Include labor. supervision. maintenance. utilities. materials. major repair
allowances, and direct effects upon production cost. Credit materials re-
claimed and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overall Effectiveness of Controls Represented by This Summary
For ducted systems. indicate range of outlet grain loadings observed:
from
grains to
Check one:
grains particulates per cubic foot.
----- Air pollution from conditioning is completely suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based upon best available technology but are not adequate
Other:
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
B-15 and B-16
SUMMARY 0 -- HOT-WORKING OF STEEL BY FORGING OR ROLLING
COST/EFFECTIVENESS STUDY OF STEELWORKS AIR-POLLUTION CONTROLS
III.
I.
Identification
Company:
Data entered by:
II.
Capacity of Steelworking Operations Represented by This Summary
Typical annual throughput of semifinished steel, net tons =
Type of operation: Typical piece weight
tons
Description of Equipment and Activities Represented
On the reverse side of this sheet, briefly describe equipment and/or activ-
ities used to control airborne dust and fume during hot-working of steel.
Use block diagrams for illustration of sequential equipment; omit numerical
data,
IV.
Estimated Installed Capital Cost of Pollution-Control Equipment
Exclude land, and apparatus not specifically used to control pollution.
Include equipment, installation, engineering, research, start-up,
ance tools and facilities, inventories, and incremental costs for
in production equipment to accommodate pollution-control devices.
Estimated installed capital cost = $
main ten-
changes
Year of start-up:
Annual escalation has averaged about ----- percent
V.
Estimated Operating Cost of Air-Pollution Control System
Exclude depreciation, taxes, interest, insurance, and all other fixed costs.
Include labor, supervision, maintenance, utilities, materials, major repair
allowances, and direct effects upon production cost. Credit any materials
reclaimed and increases in life of equipment.
Estimated annual operating cost = $
Man-hours direct labor/year =
Credits total $(
)
VI.
Estimated Overall Effectiveness of Controls Represented by This Summary
For ducted systems, indicate range of outlet grain loadings observed: from
grains to ----- grains particulates per cubic foot.
Check one:
----- Air pollution from hot-working is completely suppressed or contained
Controls described do a thorough job and meet public standards fully
Controls are based upon best available technology but are not adequate
Other:
BATTELLE MEMORIAL INSTITUTE - COLUMBUS LABORATORIES
-------
1___-
.
.-
0:::
o
Q.
W
0:::
0:::
W
..J
'"
'"
W
0:::
C
I
..J
..J
W
C
Z
~
'"
u
-------
APPENDIX C FOR
FINAL ECONOMIC REPORT ON COST ANALYSES
for
A SYSTEMS ANALYSIS STUDY OF THE
INTEGRATED IRON AND STEEL INDUSTRY
(Contract No. PH 22-68-65)
to
DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE
NATIONAL AIR POLLUTION CONTROL ADMINISTRATION
DIVISION OF ECONOMIC EFFECTS RESEARCH
May 15, 1969
COSTS
AND PERFORMANCE OF CONTROL
AND CONTROL EQUIPMENT
(Sub-Contract Tasks)
SYSTEMS
by
P. R. K1auss, Principal Investigator, Economic
P. L. Sieffert, Principal Investigator, Technical
J. F. Skelly, Sub-Contract Task Force Director
Swindell-Dressler Company
II OM';". of Pullman Incorporated
441 Smithfield Street
PittwurSh. Pa. 15222
-------
TABLE OF CONTENTS
APPENDIX C
Summary and Conclusions
Primary Categorization of Costs
The General Structure of Capital Costs
Categorization Schemes
The General Problem of Capital Cost Estimating
The General Structure of Operating Costs
The General Problems of Estimating Operating Costs
Costs of Control Systems
Effect on Cost of Multiple Furnace Installations
Special Problems Encountered When Installing New
Equipment in Existing Plants
Data Sources, Procedure, Precision of Results
Alternate Systems
Sinter Plants
Pelletizing Plants
Coke Oven
Blast Furnace
Basic Oxygen Furnace
Open Hearth Furnace
Electric Arc Furnaces
Scarfing
HCL Pickling Line
Cold Rolling Mill
Power Plant Boilers
Sample Calculation - Operating Cost
Method of Determining Exhaust Gas Volume
Capital Cost Breakdown
Gaseous Pollutants
Area Ventilation and Emission Control
Effect of Efficiency Specifications Greater than
Current Legal Requirements
Performance Equations
C- 1
C- 2
C- 2
C- 4
C- 6
C- 8
C- 9
C- 10
C- 11
Control
C- 11
C- 13
C- 15
C- 17
C- 24
C- 29
C- 30
C- 32
C- 36
C- 40
C- 45
C- 48
C- 50
C- 52
C- 55
C- 56
C- 58
C- 65
C- 65
C- 68
C- 69
-------
Theoretical Performance Variables
Control System Cost Changes
Conclusions
C- 71
C- 73
C- 93
Condition of Particulate
Temperature and Humidity
C- 96
C-109
C-110
C-110
C-lll
C-1l7
C-l22
Factors Affecting Gas Cleaner Performance
Gas Volume Changes
Pressure Drop
Dust Loading
Adaptability to Removal of Gaseous Pollutants
-------
C-I
SUMMARY AND CONCLUSIONS
In this appendix, the cost of air pollution control equipment is estimated
for various processes of the integrated iron and steel industry. Factors
affecting the performance of control equipment and the effect of performance
level on cost are discussed.
The purpose of the cost estimating presented here is to provide average
data on the emission control cost per production unit, as an initial step in
the formulation of a model for calculating the national cost of air pollution
control in the industry. A technique of estimating costs for this purpose is
described and used. The more detailed technique for estimating the cost of a
specific installation to suit the particular conditions and specifications is
not called for here. Without that detail, as outlined below, these estimates
cannot and should not be used to determine the cost of control for a specific
installation. An engineering analysis for that purpose, using classical esti-
mating techniques, is available from a number of engineering firms, and should
be used if the cost of control is desired for, e.g., company A's furnace B in
city C.
The costs presented in the tabulations here are based on process parameters
representative of average modern practice. Costs will vary from plant to plant
depending on such specifics as raw material properties, details of the process
as applied, product properties desired, unusual materials of construction or
unusual combinations of equipment occasioned by special corrosion or abrasion
conditions, details of integration of equipment into plant lay-out, etc. Typical
methods of control which are, or could be, applied to the average process are
cost estimated. Typical options for gas cooling are incorporated in each system;
this choice, together with the average process effluent level (both gas quantity
and particle concentration), determines the capacity and power ratings of
equipment.
The effect on cost of unusual arrangements for combined control systems
and for area ventilation is discussed. The effects on cost of unusual adapta-
tions of equipment to existing plants and facilities are discussed.
Also presented here is an indication of the theoretically determined
difference in cost for a more effective control system of the types typically
used. This is not the cost of altering an existing installation, with its
specific needs. These cost differences also are intended as input to the
model for national costs. The development of the model with cost data from
systems as used today, or as designed, will yield the ultimate tool for
determining national cost, and provide an average comparison to the initial
input data presented here.
Certain nominal unit costs have been established as bases for calculating
operating costs. These costs will vary with monetary fluctuations over the
life-span of equipment. In adjusting field data as input to the model devel-
opment calculations, these costs would be normalized. Electrical energy is
standardized at $50/installed HP per year (based on a standardized 330 oper-
ating days per year x 24 hours/operating day = 7,920 operating hours/year).
This corresponds generally to 3/4~/KWH for large motors. Labor cost is set
at $5.00 per manhour including all welfare and fringe costs. Ratios of real,
local costs to these standardized values may be used as factors for adjusting
reported operating costs.
-------
C-2
PRIMARY CATEGORIZATION OF COSTS
Many costs arise during the life span of an industrial project, from
the earliest planning to the final demolition of the obsolete plant.
These costs are commonly assembled into three categories:
1.
Capital Costs: Cash outlays associated with
planning, engineering, purchasing, construction
and startup of the installation. Such costs occur
only once during the life of the installation.
2.
Operating Costs: Charges associated with the
operation, maintenance and financing of the plant
during its period of productivity. These costs
are repetitive in nature, constituting a continual
flow of cash away from the operating organization.
3.
Demolition and Salvage Costs: Cash transactions
arising while the facility is being dismantled and
sold off. Some of the cash flows are expenses and
some are income. The algebraic sum constitutes
either the Demolition Cost or the Salvage Value
depending upon the direction of the net cash flow.
These items occur only once during the life of the
plant, in which respect they are related to the
Capital Costs in (1) above.
THE GENERAL STRUCTURE OF CAPITAL COSTS
The capital cost of any plant or facility is the sum of many separate
cash payments made to suppliers of component parts, to workers of many kinds
for their labor, to consultants and contractors for their services, to
shippers for transportation of components to the construction site, etc.
This can be expressed as
CT = cl + C2 + c3 + ...... + cn
where CT is the total capital cost and cl' c2' c3'
cash payments as described above.
..... c
n
are individual
In most industrial projects, the total number of these payments, n, is
very large. To simplify accounting and cost analysis, they are usually
grouped into a relatively small number of categories. Each category contains
the amounts of all payments related to some recognizable sub-division or
functional aspect of the total project. If these categories are called
Cl' C2' C3' CM'
Cl = cl + c2 + c3
C2 = c4 + Cs + c6 + c7
C = C + C + C
M e m n
-------
1-
C-3
The sum of
individual cash
the categories.
the categories equals the total cost, provided
payment appears once and only once, in one and
That is,
CT = Cl + C2 + C3 + ......+ CM
that each
only one of
In a typical situation, the total number of payments, n,
the range of thousands, but M, the number of cost categories,
10 to 20. Various categorizing schemes may be, and have been
categories may be functional in nature:
might be in
might be only
used. Some
Cl =
materials cost
C -
2 -
field labor cost
C =
3
engineering labor cost
C =
4
freight cost
etc., etc.
In such a scheme, Cl is the sum of all materials costs on the project
while C2 is the total cost of all labor required for field erection of all
materials.
Another system of categorization can be based on major
of the installation. Each category would include all costs
labor, freight, engineering, etc.) associated with one part
component parts
(materials,
of the plant.
c =
1
cost of furnaces
C =
2
cost of gas cleaning equipment
C =
3
cost of water supply system
C =
4
cost of buildings
etc., etc.
The choice of categorization scheme is usually a matter of tradition
or convenience, as long as each cost item appears once and only once in the
array.
As was already noted, these cost items occur only once in the life of
the project. Moreover, these costs are clearly defined cash transactions,
determined by market action. Unlike certain components of operating cost
(discussed elsewhere), they are generally unaffected by accounting practices,
tax procedures and other constraints arising from policy de~isions.
The planning, engineering and construction of an industrial plant
often encompasses a time span of two to four years. During these years,
currepcy inflation may be great enough to have a significant effect on
costs arising in the later phases of the overall program.
-------
C-4
CATEGORIZATION SCHEMES
A typical categorization scheme for industrial projects divides
capital costs into the following ten categories:
1.
Material: This includes the purchase cost of all materials,
machines, component parts, cement, structural steel, etc.
which are required in the field to make up the complete
operating installation. Net purchase costs are often F.O.B.
point of origin.
2.
Erection Labor and Supervision: This item includes wages
and salaries, payroll taxes, welfare benefits, etc. for all
persons employed at the construction site in the installation
of the material items listed above. Some estimating pro-
cedures involve the preparation of separate estimates for
the labor force and for a supervisory group of engineers.
This procedure may be needed when supervision is supplied by
an organization that is not responsible for employing the
general construction personnel.
3.
Freight: This covers the cost of transporting all of the
materials from their respective points of origin to the
construction site.
4.
Special Tools: This cost category includes rental and
transportation charges for special tools or equipment that
may be needed at the construction site. Excavating equipment
and large hoisting machinery are often rented for brief periods
of time during a construction project because the amount of
work to be done by them does not warrant their outright
purchase for one job.
5.
Taxes and Insurance: This covers the payment of necessary
sales taxes, permits and charges for insurance protection
as required during the course of the project.
6.
Engineering: In this category are collected all of the costs
associated with the central engineering and design aspects of
the project. This includes the services of engineers and
other personnel involved in the design, purchasing, and
general management of the project. The costs of these
services are usually calculated at a standard rate which
provides for salaries, fringe benefits, occupancy, and
departmental overhead. In addition, this category can
logically include general overhead and fee to the central
engineering organization.
7.
Client Engineering and Coordination: During the design and
construction of a new plant installation, the company or other
organization which will operate the finished plant usually
participates in the engineering and general management of
design and construction. This may involve the preparation of
specifications, review of design drawings, review of purchase
orders, participation in development of field schedules, etc.
Charges in this category should include those directly
associated with the personnel involved together with an
appropriate share of overhead.
-------
C-5
8.
Startup:
operation
it into a
This covers costs associated with preliminary
and test of the new equipment in order to bring
condition of reasonable operating efficiency.
9.
Inventory: This working capital item
which must be tied up in inventory of
process, maintenance supplies, etc.
covers those moneys
raw materials, goods in
10.
Land: This covers the fair value for the land area assigned
to the installation. In general, it includes not only the
land occupied by processing or manufacturing equipment but
also land utilized for the storage of materials and products
immediately preceding and following the processing unit. In
building a new installation, there is often a cost arising
from the preparation of the land site. This may involve
grading, filling, removal of old structures, etc., and must
not be omitted from the calculation of total investment.
The total investment is the sum of all of the foregoing items. This
total will not be the total payment to contractors, suppliers, and construction
labor because of the presence of items 7 through 10 above. Nevertheless, all
of these items (1 - 10) make up the total investment required to erect the
new industrial installation and bring it into normal working condition. It
is this total investment cost which enters into the cost computations else-
where in this project.
The detailed arrangement and tabulation of individual cost items in
the total as defined above can be handled in a variety of ways. The
particular method selected is largely a matter of individual preferences.
This may be based upon accounting practices well established in earlier
projects or upon cost classifications needed for tax or operating control
purposes.
In one method of tabulating costs, the costs are arranged according to
the piece of equipment which is concerned; the cost of a single item of
equipment would include material, erection labor and supervision, freight,
engineering, etc. This leads to an estimate composed of a group of cost
figures which are the total installed costs of individual pieces of equip-
ment or of operating subsections.
Another approach to the problem is based upon functional lines. Costs
are arranged in accordance with the ten categories given above. This
method does not display the total installed cost of a single piece of
equipment but does reveal the total cost of each function. In particular,
it discloses the total cost of field labor and the total cost of engineering
services. Control of these two functions is often considered to be an
important matter by the managers of engineering contracts. In general, this
latter scheme will be used in the present investigation, with some reduction
in the number of categories, because of the generalized nature of the
estimates involved.
-------
C-6
THE GENERAL PROBLEM OF CAPITAL COST ESTIMATING
The estimator of capital costs wishes to predict the total cost, ~,
of a new installation to be designed and constructed at some later time.
His principal working technique is that of extrapolation into the future,
using design data on the new plant and past cost experience with similar
facilities. Various estimating procedures are available, differing in
the amount of work they entail, and in the accuracy of the resulting
estimate. In general, the more laborious methods are needed to achieve
the more accurate results. As a result, a choice must be made in any given
case between estimating precision and estimating cost.
The most economical estimating procedures generally involve a direct
estimate of ~, total capital cost of a new plant, derived from a historical
record of the cost of similar plants. In practice, however, such methods
give very rough, imprecise estimates because the design of the new plant
usually differs in major respects from its predecessors. Since many com-
ponent parts of plants undergo only small changes with time, it is usually
possible to obtain more precise results by making separate estimates of
component costs.
Component cost estimates may be taken from historical records, or from
recent market quotations. Both of these sources unavoidably contain poten-
tial errors which pass along into the total cost, CT' The probable error
in the total is a weighted average of the probable errors in all of the
components. The weighting factors are based on the relative costs of the
individual components, and the average is calculated as a root-mean square.
As might be expected, this procedure assigns greatest importance to the
most expensive components, with proportionately less emphasis on the cheaper
items. In practical estimating, therefore, high-priced components must
receive a great deal of attention if the final estimate is to be accurate.
Smaller, inexpensive parts of the plant may be treated in a more approximate
manner without seriously distrubing the total cost.
One consequence of this situation is a trend toward increasing the
number of components to be estimated separately. It is argued that this
will eliminate high-priced components and improve estimating accuracy
because no single error will have a large weighting factor associated
with it. Experienced estimators know this to be true, to some extent, but
that a limit exists which cannot be passed. Even though the number of
components separately estimated becomes very large, each component price
estimate (however small it may be) still contains its error. The percent
error in the total is always a weighted average of the percentage errors
in all the component prices.
The cost of preparing an estimate increases as the number of estimated
components increases. This acts to restrain the tendency toward enlargement
of the number of estimated components, which is reinforced by the unavoidable
total error even with a large number of components, as described in the
previous paragraph.
-------
C-7
It is important to realize that some design work must be done before
the actual estimating can begin. This design work, which is expensive,
must generate enough data about each component in the estimate to permit
the setting of a component price. For a fully detailed estimate, the
design cost may be almost as great as that needed for actual construction.
For example, a detailed estimate of the cost of foundations in an
industrial plant might cover the following items:
Earthwork:
Machine Excavation
Trench Excavation
Hand Excavation
Trucking and Hauling
Backfill
Deep Foundations:
Bearing piles
Sheet piles
Walers
Buildings, mats and piers
Spread footers
Grade beams
Footings
Walls - below grade
Walls - above grade
Heavy equipment
Heavy mats
Elevated slabs
Shored slabs and columns
Earth slab paving
Formwork:
Reinforcing:
Bar
Mesh
Miscellaneous
Steel:
Anchor bolts
Embedded steel
Embedded railroad tracks
Miscellaneous steel
Checker plate and grating
Buildings, mats and piers
Spread footers
Grade beams
Continuous footings
Walls - above grade
Walls - below grade
Heavy equipment
Heavy mats
Concrete:
-------
C-8
Concrete: (continued)
Elevated slabs
Shored slabs and columns
Earth slab paving
Fine grading for paving
Batch plant
Waterproof walls and piers
Vapor barrier and waterstop
Joint materials
Color, sealer and grout
Finishes:
Steel Trowel
Screed finish
Brick
Floor hardener
Separate estimates are to be made of labor and materials for each
of the above items. It is evident that this array requires the making of
a very detailed design together with an equally detailed compilation of
historical data.
The design and estimating of a complete plan on this basis is very
expensive, and will be done only under extremely competitive conditions.
Moreover, this kind of analysis is not possible until a specific plant
site has been selected and its characteristics have been determined.
The objectives of the present study can better be served by using a
small number of components. In fact, the elaborate detail set forth above
would be inappropriate because this work is not directed toward any single
plant location.
THE GENERAL STRUCTURE OF OPERATING COSTS
Two classes of transactions enter into the operating cost of a
manufacturing plant. One is composed of direct cash expenditures for
labor, raw materials, fuel, electric power, maintenance supplies, etc.
The other is a group of costs whose magnitudes are determined in part
by managerial policy decisions of various kinds. Depreciation charges and
general overhead burden are of this latter type.
Like capital costs, the multitude of individual operating cost items
is usually arranged into a small number of categories. Some of these have
been mentioned in the preceding paragraph. Since operating costs, unlike
capital cost, arise continually over the many operating years of the plant's
life span, they are usually collected and reported for comparatively short
periods of time. Most often, the year is the time interval chosen for
steady-state analysis, in order to eliminate season effects and daily
fluctuations caused by minor events in plant operation. The total
operating cost over this period can then be related to the total production
during the same time to arrive at a useful value for the unit production
cost.
-------
C-9
Thus, if 0* is the unit production cost, ° is the
cost, and W is the total number of units of proauction,
given year,
total operating
all during a
°T
0* =-
W
0T = 01 + 02 + 03 + 04 + 05 + ...... + ON
where 01 =
operating labor cost
raw materials cost
°2 =
°3 =
maintenance labor and materials
° =
4
electric power cost
° =
5
depreciation charges
° =
6
working capital charges
etc., etc.
During the operating life-span of a typical plant, there will be
substantial changes in technology, administrative techniques, social
practices, markets, state regulations, interest rates, currency values, etc.
These evolutionary changes may lead to substantial modifications in the
unit production cost.
THE GENERAL PROBLEMS OF ESTIMATING OPERATING COSTS
The estimation of the cost of the two types of transactions described
on the previous page brings two different problems to the estimator. The
first of these, the direct cash outlays for labor, maintenance ,power , etc.,
can best be handled with the help of historical records of actual plant
experience. Cost records are generally kept by operating companies for
organizational cost centers based on considerations of product management,
administration structure, etc. When these cost centers cover the equipment
of interest to the estimator, plant records can supply directly the historical
basis needed for close estimating. When the cost centers do not correspond
to the operating area being examined by the estimator, the direct con-
struction of an historical base is not possible. In that case the required
cost components must be arrived at by theoretical calculations, personal
recollections, intuition, etc.
Unfortunately for the present study, steel companies do not generally
maintain cost centers around their pollution control activities. It is
the general practice to use cost centers which include both production units
and control facilities within a single perimeter. At this time there are
only scattered cost data available on operating labor and maintenance for
air pollution control installations in the steel industry.
-------
C-IO
The second type of component in operating cost is that whose
magnitude is established by policy decisions relating to depreciation,
overhead, capital charges, etc. In principle, depreciation should be
based in a simple, non-controversial way on the actual life of the
equipment and its ultimate salvage value. In practice, the prediction
of the life and salvage value of pollution control equipment is uncertain.
Operating conditions are usually severe, maintenance practices vary, and
the danger of obsolescence is great. Depreciation rates therefore are
strongly influenced by policy.
The allocation of corporate overhead involves even more difficult
policy questions. Charges for the use of capital in control equipment
require predictions of interest rates, profitability of alternative
investments, and future credit rating. The estimator is clearly working
in a very imprecise area when he considers these problems.
As a result, operating cost estimating is inherently uncertain and
must not be expected to lead to results of high precision. The selection
of optimum or preferred pollution control equipment or processes should
not be based upon small differences between the operating cost of
alternative designs.
COSTS OF CONTROL SYSTEMS
A control system is considered to be made up of all the items of
equipment and their auxiliaries which are used solely for the general
abatement of atmospheric pollution in the neighborhood of the steel
works. Typically this will include a collecting hood or gas collecting
pipe at the furnace, ductwork, spray cooler, dust collector, fan and
motor, and stack. Included also will be structural steel, foundations,
control instruments, insulation, piping, water treatment, and electric
power supply facilities for the entire gas cleaning system. (Water treat-
ment includes all those items required for gas cleaning water uses and
sufficient for avoiding a water pollution problem.) Excluded are those
equipment items which, while they may contribute to the functioning of
pollution abatement equipment, would be used for process or economic
reasons even if there were no pollution abatement requirements.
The cost of land occupied by pollution abatement equipment has not
been included. It is recognized that such land has a real value but a
satisfactory method for estimating it has not been established. Costs
associated with preparation of the site, start-up operations and working
capital are also not included. Certain portions of a control system
occupy or utilize parts of steel plant buildings and, therefore, might be
charged with a share of general building costs. This item has not been
estimated here. In calculating operating costs no attempt has been made to
allocate a portion of general overhead to control systems.
Capital and operating costs in the following tabulations are based upon
collectors whose efficiency can be relied upon to produce an outlet dust
loading of 0.05 grains/SCF of gas. A later section of the report contains a
discussion of the relationship between cost and collection efficiency. In
general, higher efficiency is more costly.
-------
C-ll
EFFECT ON COST OF MULTIPLE FURNACE INSTALLATIONS
Pollution abatement equipment becomes increasingly cheaper as the
number of furnaces that can be connected to one common control system
increases. The largest saving in multiple furnace situations can be
realized from the alternate scheduling sequence of furnace operation.
Furnace shops using arc furnaces and BOF vessels which have high and
low gas emission periods seldom reach peak conditions at the same time.
This is generally due to material handling which limits the furnaces in
a common shop to being charged and teemed in succession. These furnaces
can be lanced or blown alternately which will permit designing the
multiple furnace control system for a reduced thermal capacity and reduced
air volume. A single collecting system sized on this basis for two furnaces
will handle approx. 1.7 times more fume than for a single furnace appli-
cation and a three furnace system could be designed to handle 2.5 times
more. This results in a large reduction in capital cost, although there
is a loss in operating flexibility. Where flexibility is important, a
design compromise is possible by using a single collector whose size will
permit peak operations on all furnaces at the same time. When more than
three units are to be served by a common control system, it is best to
assume that several furnaces will have to be at peak load together.
The following tabulation
the effect of combinations on
of control equipment.
is suggested as a rough guide for estimating
capital cost; for all processes and types
1.
Separate collectors
on each furnace
1 2 3
100% 200% 300%
100% 170% 250%
Number of Furnaces
2.
One collector to handle
peak loads on all
furnaces at once.
3.
One collector to handle
only one peak load at a
time.
100%
140%
200%
For example, the capital cost for control equipment serving
furnaces together, with both able to run at peak loads, would be
of the cost for a single furnace installation.
two
170%
SPECIAL PROBLEMS ENCOUNTERED WHEN INSTALLING
NEW CONTROL EQUIPMENT IN EXISTING PLANTS
Existing plants can and have been revamped to accommodate modern dust
collecting equipment. A grass roots plant affords the flexibility of selecting
and installing cleaning equipment and duct work for maximum efficiency and
minimum capital cost. Providing and adapting fume abatement equipment for an
existing plant can in many cases be very expensive, especially if satisfactory
land is not available for locating this new equipment. In many cases the fume
collecting equipment must be located on top of the building roof which requires
-------
C-12
strenghtening all the supporting columns and trusses. A more serious situation
would require placing the fans and motors at roof level. For wet scrubbers in
such a case the weight due to the much higher scrubber horsepower requirements
can become so costly that it might be necessary to use an inferior type of
collector. The second-best collector may ultimately cost the customer more in
capital expenditure, maintenance, operating cost and efficiency.
The most costly aspects of designing a new fume collection system for
an operating facility often involves unusual and unorthodox arrangements of
fume pickup at the furnace. This is in part due to lack of space for supports
and interference with existing structures or obstructing personnel, vehicle
and crane approaches. To avoid these conditions may require an alternate type
of pickup at the furnaces, which, in turn, could dictate the type of apparatus
used for separation of the fume and particulate matter from the gases. In one
existing arc furnace shop, for example, it was most desirable to employ a direct
shell tap extraction from the arc furnaces, but in this shop very little free
area would then have been left for water cooled ducts, spark boxes, or cooling
chambers. Consideration was given to running ducts under the teeming building
but this alternative would have been extremely expensive and would have caused
considerable shut-down time. Moreover, the added furnace roof loading would
have required expensive revamping to support the extra weight of the water-
cooled elbow. The only practical solution was to install roof-truss hoods over
the furnaces. This system had the inherent disadvantage of moving 4 to 5 times
the air volumes required by the direct shell tap. The additional air volume
resulted in much more capital investment on fans and cleaning equipment. This
fume collection system was a compromise design forced by the limitations of
existing facilities. A newly designed plant could include direct shell taps on
the furnaces, resulting in much less expensive equipment and in a considerable
reduction in operating expense.
Limited space around an existing plant may require locating new control
equipment on the roof, as previously discussed, or possibly in a location so
remote as to make the duct runs much longer than would otherwise be needed. The
capital cost increases because of the extra ducting and the added air friction
losses. The increased static pressure requires a larger fan of greater horse-
power. The result will be increased investment and higher operating costs for
the life of the system.
Older control equipment at existing plants has often influenced the
selection of a new type of collector system. A plant already operating with
wet scrubbers will, if at all possible, try to adapt the new system to similar
cleaning equipment especially if there is enough reserve built into the plant
slurry system to handle the additional loading. Maintaining the same pattern
of equipment will not necessarily be the best selection for the process involved
and may increase the capital outlay as well as operating costs of the new
installation. A grass-roots plant is not generally as much influenced by such
continuity factors.
The amount of shut-down time required to install the fume collectors
will add to the cost of the installation by the amount of lost production. The
amount of down time that can be tolerated will influence the type, location and
duct routeingof the system. Any deviation from the most direct design such as
would be practical for a grass roots plant will therefore add cost to the install-
ation.
-------
1--.
C-13
Detailed design and estimate studies will usually be needed if
estimates are desired for new control equipment in existing plants.
general cost data given in this report will not be reliable in such
cost
The
cases.
DATA SOURCES. PROCEDURE. PRECISION OF RESULTS
The primary data used in the preparation of cost estimates were
taken from a number of sources. These were:
1.
Estimate files of the Swindell-Dressler Company.
These files ranged in age from the immediately
current back to 1962. All costs were adjusted for
price escalation to bring them to present levels.
2.
Cost information supplied by certain steel companies
relating to their own plant facilities. Such data were
generally in the form of total costs for complete
installations. These were adjusted for inflation
by means of the same factors used on the estimate
data from Swindell-Dressler files.
3.
Cost information supplied by certain manufacturers
of pollution control equipment. This information
was presented in response to specific requests by
Swindell-Dressler and came in the form of budget
figures.
The Swindell-Dressler file data are detailed estimates prepared to meet
the requirements of particular competitive situations. Each estimate
assembled the costs for a specific location at a specific moment in time,
and this was done in much detail based upon a combination of firm price
quotations from suppliers and historical data about labor productivity at
the location in question. Each estimate, therefore, contains cost elements
which are influenced by local conditions not necessarily applicable to
other plants and geographical locations. A simple compilation of these
estimates would not have been adequate for the needs of the present investi-
gation.
The data from these many particular cases have been rearranged into
a more generalized form. The primary tactic used here has been that generally
followed by other government agencies and students of cost estimating.
This is to segregate the capital costs of the principal items of equipment
in the installation and to prepare smoothed, adjusted values for these
equipment items over a wide range of operating capacities. As is
well known, such smoothed values usually form a straight line graph on log-
log paper, with the slope of the line being related to certain characteristics
of the type of equipment involved. In this study smoothed material cost
data did give satisfactory linear graphs with slopes that were reasonably
related to the type of equipment.
-------
C-14
Following the practice of others in this field, the minor and bulk
materials were estimated on the basis of ratios applied to the costs of
the principal items. Labor costs were estimated for each principal item
and bulk category by the use of standard Swindell-Dressler factors
relating labor to material costs. The resulting labor figures apply to
the Pittsburgh area but may be adjusted for other locations through the
use of regional labor indices.
The capital costs included only facilities for loading the collected
dust or sludge into trucks for transportation elsewhere. No other disposal
costs or by-product values have been assigned. Central engineering costs,
overheads and fees were based upon a standard sliding scale generally used
by contract engineers.
It is believed that the general precision of the capital cost estimates
is such that most specific plant situations will fall within i15% of the
tabulation values. In more statistical terminology it might be suggested
that the standard deviation is about ilO - 12%. It is to be expected that
any specific plant location which presents unusual cost problems associated
with layout, structure, power supply, etc. might fall outside these limits.
In such cases a detailed plant design and estimate should be prepared if
accurate capital cost data are required. As previously noted, the accuracy
of operating cost values is influenced by many factors which may vary
considerably from one company to another. The selection of control equip-
ment should not be based upon small differences in operating cost estimates.
Operating cost estimates present tabulated costs for the following
items:
1.
Electric Power
2.
Maintenance
3.
Operating Labor
Direct Operating Cost - The sum of above three items.
4.
Depreciation
5.
Capital Charges
The items included in Direct Operating Cost are those cost elements
which are the direct cash outlays discussed on page C-8. They are, to some
extent, under the control of the plant operating management. The costs
assigned for Depreciation and Capital are, as noted on page C-10 based upon
policy decisions not generally under the control of management at the plant
operating level.
Electric energy is calculated at a standardized rate of 3/4 C per kilowatt
hour. The cost of make-up water is not included as such. Operating labor
cost was calculated at the nominal value of $5.00/man hour including all
welfare and fringe costs.
-------
C-15
Maintenance is taken at a nominal cost of 4% of the total investment
shown in the capital cost tables. This figure has been the subject of
considerable discussion and has been retained in this analysis because it
is believed to represent a reasonable value over the total life of the
equipment. Most steel plant maintenance cost records do not make a
separate accounting for each pollution control installation. It is,
therefore, difficult to arrive at an exact, numerical evaluation of total
maintenance during the life of a piece of control equipment. In actual
practice, maintenance expenditures are not uniform from year to year.
In many cases major maintenance outlays occur only after the passage of
several years of operation. Moreover, as might be expected, maintenance
costs usually increase during the service life of an item of control
equipment. There have been some cases reported where major costs were
experienced early in the life of a control installation. It is believed
that these cases should properly be attributed to inadequate engineering
rather than standard maintenance. These incidents were more common some
years ago when knowledge of pollution control engineering was not as
extensive as it is today. There have also been cases reported in which
major modifications were made to control equipment after it had been in
service for some years. Some times these episodes were caused by changes
in the operating practice of the process segment which placed greater
burdens on the control equipment. This type of cost is not considered to
be a part of maintenance. The 4% figure is retained in this study be-
cause it is believed to represent a reasonable value for good maintenance
in well designed equipment when calculated over the entire life of the
installation.
In those cases where filter bag replacement represents a major
maintenance cost item, the system, less bags, is given the 4% maintenance
charge; and the cost in material and labor for bag replacement at a reason-
able average rate (18 months for Sinter plants, 2 years for steelmaking
shops) is added for the net maintenance cost listed.
Depreciation is calculated on a straight line method using total
investment with an expected life of ten years. Other studies of depreciation
have suggested longer service life times, but these are considered to be
greater than average plant experience will confirm. Advancing technology
and rising standards give importance to the factor of technical obsolescence.
Capital charges are taken at 10% per annum. It is believed that this
will be reasonable in the light of rising interest rates and local taxes.
Annual calculations are based on 330 operating days per year, 24 hours
per day. This gives a total of 7,920 operating hours per year.
ALTERNATE SYSTEMS
In the tabulations which follow, several alternate control systems are
included, for most processes. Aside from cost, other factors enter into the
selection of a system. The nature of the process to some extent dictates or
precludes the use of a particular type of control. For example, some collected
dusts can be reused directly in the process or used as burden in the plant
-------
C-16
following agglomeration; the wet or dry state of the collected dust may afford
a convenience to disposition of the dust according to the current practice of
a particular plant.
The gas may be reusable as a process material in the plant, where changes
in its temperature and humidity by the cooling system would have to be consid-
ered in the total plant energy economy. The local cost of treated water, space
requirements for retreatment facilities, and possible difficulties in using
water near the process vessel may affect the choice of wet or dry systems.
The particle size and concentration of the effluent determine whether
high efficiency gas cleaning equipment (high-energy wet scrubber, electrostatic
precipitator, or baghouse) is needed, based on particle size vs. efficiency
experience data for different types of collectors. This data is largely in
the form of proprietary design curves in the files of equipment manufacturers.
For the purposes of this study, the dividing line between low-and high-energy
wet scrubbers is 12 inches of water pressure drop across the collector, with
high-energy applications generally using several times this pressure drop.
The nature of the dust collector equipment to be used largely determines
the extent and method of cooling the process gas. The process effluent may
vary in temperature from 100°F for material transfer point ventilation to
3000°F or higher for furnace exhaust. This gas may be quenched by air dilution
or water sprays to a lower temperature, or undergo a heat exchange to cool
without adding material to the effluent stream. If the gas is combustible,
it may initially be burned, with an excess of air to insure completness of
combustion, where a fire or explosion hazard would exist. A wet-type cleaner
may treat water-quenched or hot gases directly. The electrostatic precipi-
tator used on highly (electrically) resistive particles requires a degree of
cooling and humidification control to be effective and of economical construc-
tion. On the other hand, over-cooling or over-quenching can result in conden-
sation with resultant corrosion, collector surface fouling, and dust handling
problems in dry precipitators, and baghouses as well. Thus, in general
- excess air is added to the effluent stream where combustion occurs,
in a water-cooled or other heat-exchange vessel;
- water addition completes the cooling for a wet system;
- indirect cooling by heat exchange will provide the most economical
cooling to about 500°F in dry systems;
- added humidification by water sprays usually completes the treatment
of gases prior to electrostatic precipitation;
- air dilution for bag temperature control usually completes pre-
baghouse cooling.
Finally, to achieve economical fan power levels, the gas volume is kept
low by gas cooling especially with high-energy wet scrubbers. This ultimate
effluent gas, if sulfur oxides persist in significant degree to this point in
the system, must have sufficient lift in the form of thermal or mechanical
energy, or stack height to disperse in the atmosphere.
-------
C-17
SINTER PLANTS
The following tables contain capital and operating cost data for two
sizes of sinter plants. Sinter plant control systems are usually designed so
that one control unit handles gases coming from the windbox while a separate
control unit receives dust collected at several points in the material
handling system.
The attached estimate gives separate figures for the windbox and
materials handling operation. Various combinations of types of collection
equipment are used on sinter plants and it is therefore necessary to offer
separate values for these two zones of collection. The total cost for a
given sinter plant will be the sum of the cost for the windbox and the cost
for the materials handling.
The tables do not include system components through recovery cyclones
or windbox fans. Booster fans are included. Modifications only are in-
cluded in the windbox stack item cost.
The capacity of the sintering machine for the tabulated gas volume
and dust collection cost is based on the average nominal capacity, making
normal sinter. Variations will occur with differences in the burden. For
example, self-fluxing sinter capacity may be as much as 35 percent higher
than a machine's normal capacity. (Symposium on Sinter Plants, Discussion,
Iron and Steel Engineer, June, 1959.) In other reports no such change is
noted. With self-fluxing sinter, more particulate matter passes through
the cleaner.
The addition of
in the windbox gas.
wet scrubber system.
oily turnings and borings to the burden generates oil mist
One solution to this is the use of a very high energy
The temperature of the windbox gases is determined
process used on the machine. The gases are moist. Too
can result in corrosion -causing condensation and tacky
problems in dry collectors.
by the sintering
low a temperature
dust handling
-------
C-18
SINTER PLANT (WINDBOX) - WET SCRUBBER (LOW ENERGY)
Gas Volume - ACFM @ 325°F
Plant Capacity - TPD
1. Material*
2. Labor
3. Central Engineering
4. Cl ien t Engineering
TOTAL
105tOOO
ltOOO
CAPITAL COST
$193tOOO
100tOOO
72 t 000
l8tOOO
$383tOOO
OPERATING COST ($/Yr.)
1.
Electric Power
2.
Maintenance
3.
Operating Labor
Direct Operating Cost
4.
Depreciation
5.
Capital Charges
TOTAL
*
$ 20tOOO
l5tOOO
30tOOO
$ 65tOOO
38tOOO
38tOOO
$14ltOOO
Note:
1)
Prices: 1969 base.
For items included in materialst see pages C-10 and C-17.
630tOOO
6tOOO
$880tOOO
440tOOO
245tOOO
6ltOOO
$lt626tOOO
$llOtOOO
66tOOO
80tOOO
$256tOOO
l63tOOO
l63tOOO
$582tOOO
-------
C-19
SINTER PLANT (WINDBOX) - ELECTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 3250F
Plant Capacity - TPD
1. Materia1*
2. Labor
3. Central Engineering
4. Client Engineering
TOTAL
1.
Electric Power
2.
Maintenance
3.
Operating Labor
Direct Operating Cost
36,000
4.
Depreciation
5.
Capital Charges
TOTAL
105,000
1,000
CAPITAL COST
$180,000
88,000
72 , 000
18,000
$358,000
OPERATING COST ($!Yr.)
$ 12,500
14,500
20,000
$ 47,000
36,000
* For items included in materials, see pages C-IO and C-17.
$119,000
Note: 1) Prices: 1969 base.
630,000
6,000
$800,000
385,000
225,000
56,000
$1,466,000
$ 77 , 000
59,000
30,000
$ 166,000
147,000
147,000
$ 460,000
-------
C-20
SINTER PLANT (WINDBOX) - FABRIC FILTER
Gas Volume - ACFM @ 3250F
105,000
Plant Capacity - TPD
1,000
CAPITAL COST
1. Materia1*
2. Labor
3. Central Engineering
4. Client Engineering
TOTAL
$154,000
72,000
58,000
14,000
$298,000
OPERATING COST ($/Yr.)
1.
$
8,500
Electric Power
2.
18,000
Maintenance
3.
Operating Labor
20,000
Direct Operating Cost
46,500
4.
Depreciation
30,000
5.
Capital Charges
30,000
$ 106,500
* For items included in materials, see pages C-IO and C-17.
Note :l)This system is rarely used.
2) Prices: 1969 base.
630.000
6,000
$800,000
340,000
222,000
55.000
$1,417,000
$ 45,000
87,000
30,000
162,000
142,000
142,000
$ 446,000
-------
C-2l
SINTER PLANT (MATERIAL HANDLING) - WET SCRUBBER (LOW ENERGY)
Gas Volume - ACFM @ l350F
48,000
Plant Capacity - Tons/Day
1,000
250,000
6,000
CAPITAL COST
1.
Materia1*
$194,000
2.
Labor
146,000
$420,000
480,000
3.
Central Engineering
81,000
184,000
TOTAL
20,000
$441,000
46,000
$1,130,000
4.
Client Engineering
1.
Electric Power
OPERATING COST ($/Yr.)
$ 36,000
$
95,000
2.
Maintenance
17,600
45,500
Direct Operating Cost
15,000
$ 68,600
$
40,000
180,500
3.
Operating Labor
4.
Depreciation
44,000
113,000
TOTAL
44,000
$156,600
$
113 , 000
406,500
5.
Capital Charges
* For items included in materials, see pages C-10 and C-17.
Note:
1)
Prices: 1969 base.
-------
C-22
SINTER PLANT (MATERIAL HANDLING)-ELECTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 1350F
48,000
Plant Capacity - Tons/Day
1,000
250,000
6,000
CAPITAL COST
1. Materia1*
2. Labor
3. Central Engineering
4. Client Engineering
TOTAL
$149,000
$420,000
91,000
195,000
60,000
133,000
15,000
$315,000
33,000
$781,000
1.
Electric Power
OPERATING COST ($/Yr.)
$ 7,000
$ 41,000
2.
Maintenance
12,500
31,000
3. Operating Labor 15,000
Direct Operating Cost $34,500
4. Depreciation 31,500
5. Capital Charge 31,500
TOTAL $97,500
*For items included in materials, see pages C-IO and C-17.
Note: 1) Prices: 1969 base.
20,000
$92,000
78,000
78,000
$248,000
-------
C-23
SINTER PLANT (MATERIAL HANDLING) - FABRIC FILTER
Gas Volume - ACFM @ 1350F
48,000
1,000
Plant Capacity - Tons/Day
CAPITAL COST
L. Materia1*
2. Labor
3. Central Engineering
4. Client Engineering
TOTAL
$120,000
68,000
49,500
12,500
$250,000
OPERATING COST ($/Yr.)
1.
$
9,000
Electric Power
2.
12,800
Maintenance
3. Operating Labor 15,000
Direct Operating Cost 36,800
4. Depreciation 25,000
5. Capital Charges 25,000
TOTAL $ 86,800
*For items included in material, see pages C-IO and C-17.
Note:
1)
Prices: 1969 base.
250,000
6,000
$350,000
166,000
116,000
29,000
$661,000
$ 38,000
36,500
20,000
94,500
66,000
66,000
$ 226,500
-------
C-24
PELLETIZING PLANTS
The following tables contain cost data for a pelletizing plant
of 1,500,000 tons per year. This is a commonly used plant
capacity with larger output being achieved through use of parallel
units. It is not likely that many pelletizing plants will be
built whose capacity is less than that shown here. It is believed
that the costs presented are reasonably typical of the several
types of moving grate equipment now in use.
Like the sinter plant, several control systems are used at different
points on the unit. The total cost is the sum of the cost at the
dryer exhaust and the materials handling dust points.
Many pelletizing plants hold to the shaft furnace design, using
multiples of the 60 ton/hr. furnace. A system of cyclones is
included in this section for the cleaning of the process gas
leaving the furnace. And also, air from the cooling unit and material
handling points at the discharge station is cleaned ~eparately.
-------
C-25
PELLETIZING PLANT (MOVING GRATE - DRYER EXHAUST) - CYCLONE
Gas Volume - ACFM @ 2500F
320,000
1,500,000
Plant Capcity - Tons/Year
CAPITAL COST
1.
$ 180,000
Material*
2.
Labor
95,000
3.
Central Engineering
69,000
4.
Client Engineering
17,000
$ 361,000
TOTAL
OPERATING COST ($/Yr.)
1.
Electric Power
$ 20,000
14,000
30,000
$ 64,000
36,000
36,000
$ 136,000
2.
Maintenance
3.
Operating Labor
Direct Operating Cost
4.
Depreciation
5.
Capital Charges
TOTAL
*
For items included in materials, see pages C-10 and C-24.
Note:
1)
Prices: 1969 base.
-------
.- -
C-26
PELLETIZING PLANT (MOVING GRATE - MATERIAL HANDLING)
- WET SCRUBBER (LOW ENERGY)
Gas Volume - ACFM @ 700F
55,000
Plant Capacity - Tons/Year
1,500,000
CAPITAL COST
1.
Material*
$
80,000
2.
Labor
54,000
3.
Central Engineering
36,000
4.
Client Engineering
9,000
TOTAL
$ 179,000
OPERATING COST ($/Yr.)
1.
Electric Power
$ 22,000
7,000
10,000
$ 39,000
18,000
18,000
$ 75,000
2.
Maintenance
3.
Operating Labor
Direct Operating Cost
4.
Depreciation
5.
Capital Charges
TOTAL
*
For items included in materials, see pages C-10 and C-24.
Note:
1)
Prices: 1969 base.
-------
C-27
PELLETIZING PLANT (SHAFT FURNACE - PROCESS EXHAUST)-CYCLONES
Gas Volume - ACFM @ 4600F
Plant Capacity - Tons/Hr.
125,000
60
CAPITAL COST
1.
Material*
$135,000
2.
Labor
82,000
3.
Central Engineering
55,500
4.
Client Engineering
TOTAL
13,500
$286,000
OPERATING COST ($/YR.)
1.
Electric Power
$ 23,000
2.
Maintenance
11,300
3.
Operating Labor
Direct Operating Cost
15,000
$ 49,300
4.
Depreciation
28,600
5.
Capital Charges
TOTAL
28,600
$106,500
*
For items included in materials, see pages C-IO and C-24.
Note:
1)
Prices: 1969 base.
-------
C-28
PELLETIZING PLANT (SHAFT FURNACE - MATERIAL HANDLING)
CYCLONES AND WET SCRUBBER
(LOW ENERGY)
Gas Volume - ACFM @ 700F 30,000 19,000
Plant Capacity - Tons/Hr. 60
CAPITAL COST
1. Material'\' $ 45,000 $35,000
2. Labor 32,000 22,000
3. Central Engineering 23,000 19,000
4. Client Engineering 6,000 5,000
$106,000 $81,000
TOTAL $187,000
OPERATING COST ($/Yr.)
1. Electric Power $ 4,000 $ 7,500
2. Maintenance 4,000 3,500
3. Operating Labor 5,000 5,000
$ 13,000 $16,000
Direct Operating Cost $ 29,000
4. Depreciation 10,500 8,000
5. Capital Charges 10,500 8,000
$ 34,000 $32,000
$ 66,000
* For items included in material, see pages C-IO and C-24.
Note: 1)
l5000ACFM capacity for once a week cleaning routine.
2)
Prices: 1969 base.
-------
C-29
COKE OVEN
Cost data are not presented for control of emissions from coke ovens.
The engineering problems involved are still being investigated, both in
the U.S.A. and abroad. Satisfactory control equipment, with proven
industrial performance, has not yet been developed.
-------
C-30
BlAST FURNACE
The attached table presents cost information for a typical modern large
blast furnace. It is anticipated that most future blast furnaces in the
United States will be of this size or greater. They will probably have
the type of wet scrubbing system shown here. Older units with combinations
of several types of control equipment are not likely to be copied in the
future.
Blast furnace gas cleaning costs should be divided between emission control
and normal plant operation. In the absence of an industry consensus, it is
suggested that an equal share be allocated to each of these accounts. The
portion of the top gas which is fine-cleaned for the blast furnace stoves
is shown for a number of plants on pages C-3l to C-37 of the technical
counterpart of this report, entitled, "Final Technological Report on a
Systems Analysis Study of the Integrated Iron and Steel Industry," May 15, 1969.
-------
.-
C-31
BLAST FURNACE - WET SCRUBBER (TWO STAGE, HIGH ENERGY)
Wind Rate - SCFM
Gas Volume - SCFM
150,000
210,000
CAPITAL COST
1. Materia1*
2. Labor
3. Central Engineering
4. Client Engineering
TOTAL
$1,427,000
636,000
360,000
90,000
$2,513 , 000
OPERATING COST ($/Yr.)
TOTAL
$ 20,000
100,000
40,000
$ 160,000
251,000
251,000
$ 662,000
1.
Electric Power
2.
Maintenance
3.
Operating Labor
Direct Operating Cost
4.
Depreciation
5.
Capital Charges
,'(
For items included in materials, see pages
C-IO and C-30.
Note:1)These are total costs of cleaning the furnace top gas
of particulate matter. Since this operation serves
the ends of both emission control and plant operational
requirements (material recovery and fuel conditioning
for re-use), a share of the cost should be apportioned
to each account. It is suggested, in the absence of an
industry consensus, that the shares be equal.
2)Prices: 1969 base.
-------
C-32
BASIC OXYGEN FURNACE
The following pages contain cost data on several sizes of basic oxygen furnaces.
They assume that a new plant is being designed and that the pollution control
equipment is included in the original design. The figures cover a single furnace
only. It is recognized that various combinations of multiple units are used
in actual practice. The influence of this is discussed on page C-ll.
Heat extracting hoods are included. These are total combustion systems
for typical oxygen-blow rates, with excess air used for a portion of the
cooling. Water additions for saturation in wet scrubber systems and for
humidification (considerably less than for saturation) in electrostatic pre-
cipitator systems completes the cooling typically. For baghouse systems, the
gas would be kept dry, using air dilution at the hood and before the baghouse
with heat exchange means between.
-------
C-33
BASIC OXYGEN FURNACE - WET SCRUBBER (HIGH ENERGY)
Gas Volume - ACFM @ l800F
220,000
100
Furnace Size - Tons
CAPITAL COST
1. Material~'" $910,000
2. Labor 490,000
3. Central Engineering 250,000
4. Client Engineering 60,000
TOTAL $1,710,000
OPERATING COST ($/Yr.)
1. Electric Power $ 207,000
2. Maintenance 68,000
3. Operating Labor 40,000
Direct Operating Cost $ 315,000
4. Depreciation 171,000
5. Capital Charges 171,000
TOTAL $ 657,000
440,000
200
660,000
300
$1,460,000 $1,960,000
790,000 1,060,000
390,000 470,000
100,000 120,000
$2,740,000 $3,610,000
$ 432,000 $ 664,000
110,000 145,000
60,000 80,000
602,000 889,000
274,000 361,000
274,000 361,000
$1,150,000 $1,611,000
Note:
1)
* For items included in material, see pages C-IO and C-32.
One Furnace System. For effect on cost of combined cleaning
systems on multiple furnace shops, see page C-ll.
2)
These estimates cover full combustion systems in which all of
the gas leaving the converter is mixed with an excess quantity
of air. All of the carbon monoxide is therefore burned to carbon
dioxide. This is the common industry practice in this country.
Systems have been designed which collect this gas in a substantially
unburned state. Such non-combustion systems may offer certain
economies. The exact extent of these economies has not yet been
generally recognized in the industry.
3)
Prices: 1969 base.
-------
C-34
BASIC OXYGEN FURNACE - ELECTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 5000F
375,000
100
Furnace Size - Tons
CAPITAL COST
1. Materia1'>'( $ 900,000
2. Labor 450,000
3. Central Engineering 250,000
4. Client Engineering 60,000
TOTAL $1,660,000
OPERATING COST ($/Yr.)
1. Electric Power $ 90,000
2. Maintenance 66,000
3. Operating Labor 20,000
Direct Operating Cost $ 176,000
4. Depreciation 166,000
5. Capital Charges 166,000
TOTAL $ 508,000
785,000
200
1,200,000
300
$1,600,000 $2,250,000
800,000 1,100,000
410,000 550,000
100,000 140,000
$2,910,000 $4,040,000
$ 210,000 $ 310,000
116,000 162,000
30,000 40,000
$ 356,000 $ 512,000
291,000 404,000
291,000 404,000
$ 938,000 $1,320,000
Note:
* For items included in material, see pages C-10 and C-32.
1) One Furnace System. For effect on cost of combined cleaning systems
on multiple furnace shops, see page C-11.
2) Prices: 1969 base.
-------
C-35
BASIC OXYGEN FURNACE - FABRIC FILTER
Gas Volume - ACFM @ 2750F 288,000 600,000 892,000
Furnace Size - Tons 100 200 300
CAPITAL COST
1. Materia1''c' $ 660,000 $1,280,000 $1,840,000
2. Labor 360,000 690,000 990,000
3. Central Engineering 200,000 340,000 470,000
4. Client Engineering 50,000 90,000 120,000
TOTAL $1,270,000 $2,400,000 $3,420,000
OPERATING COST ($/Yr.)
1. Electric Power $ 43,000 $ 89,000 $ 130,000
2. Maintenance 59,000 112,000 160,000
3. Operating Labor 20,000 30,000 40,000
Direct Operating Cost $ 122,000 $ 231,000 $ 330,000
4. Depreciation 127,000 240,000 342,000
5. Capital Charges 127,000 240,000 342,000
TOTAL $ 376,000 $ 711 , 000 $ 1,014,000
* For items included in material, see pages C-lO and C-32.
Note:
1)
One Furnace System.
on multiple furnace
For effect on cost of combined cleaning systems
shops, see page C-11.
2)
This system is used in Europe, but so far has not had an American
application.
3)
Prices: 1969 base.
-------
C-36
OPEN HEARTH FURNACE
The following tables contain cost data for open hearth furnaces. They are
based upon the addition of gas cleaning equipment to an existing furnace
shop. It is not likely that many new open hearths will be built in the
future. The figures shown are for a single furnace. The effect upon cost
of multiple furnace combinations are discussed on pageC-llof this report.
It is assumed that waste heat boilers and boiler fans are existing at the
furnaces, and needed stack modifications are included in the estimates,
along with booster fans. Waste heat boilers, while they contribute to
pollution control by cooling the gases (without adding additional material
to the gas stream which would increase size and cost of subsequent equipment),
also serve the plant energy economy, and have been in general use on open
hearth furnaces having no abatement equipment. Thus, they are not included
in the cost of air pollution control and no credit is assigned for steam
produced.
The estimates are based on averaged data for current oxygen-blown furnaces
of different sizes, charged typically with 50% hot metal, 50% cold scrap.
The typical gas cleaning equipment begins with boiler exhaust gas at 5000F
and 18% moistur~.(Steam augmentation is assumed during the dry gas period
after hot metal addition when fuel and atomizing steam rates are low, and
during low-rate initial oxygen lancing when gas temperature is low and the
checker water cooling sprays are not used.) Thus, temperature and humidity
control are minimized for dry gas cleaning systems. This gas volume per ton
furnace capacity diminishes on the average with increasing furnace capacity,
and is cleaned directly in an electrostatic precipitator system. The gas
is cooled by air dilution before a baghouse collector. The wet scrubber
saturates the gas.
-------
C-37
OPEN HEARTH FURNACE - WET SCRUBBER
(HIGH ENERGY)
Gas Volume - ACFM @ 180°F
30,000
90,000
240,000
Furnace Size - Tons
60
200
600
CAPITAL COST
1. Material~'( $160,000 $430,000 $1,000,000
2. Labor 85,000 230,000 540,000
3. Central Engineering 60,000 140,000 280,000
4. Client Engineering 15,000 35,000 70,000
TOTAL $320,000 $835,000 $1,890,000
OPERATING COST ($/Yr.)
1. Electric Power $ 24,000 $ 77 , 000 $ 210,000
2. Maintenance 13 , 000 33,000 76,000
3. Operating Labor 40,000 60,000 80,000
Direct Operating Cost $ 77 , 000 $ 170,000 $ 366,000
4. Depreciation 32,000 83,500 189,000
5. Capital Charges 32,000 83,500 189,000
TOTAL $ 141,000 $ 337,000 $ 744,000
~'( For items included in materials, see pages C-10 and C-36.
Note: 1) One Furnace System. For effect on cost of combined gas cleaning
systems on multiple furnace shops, see page C-ll.
2)
A variance from this design and cost is noted for a tar-fired
furnace with no waste heat boiler. Gas volume at higher temp-
eratures before and after saturation, and other factors lead
to a 60% higher cost.
3)
For a discussion of unusual problems encountered when installing
new collecting equipment at existing furnace shops, and an in-
dication of cost variances, see page C-1l.
4)
Prices: 1969 base.
-------
'------- --
i
C-38
OPEN HEARTH FURNACE - ELECTROSTATIC PRECIPITATOR
o
Gas Volume - ACFM @ 500 F
29,000
60
85,000
225,000
Furnace Size - Tons
200
600
CAPITAL COST
1. Materia1*
2. Labor
3. Central Engineering
4. Client Engineering
TOTAL
$130,000
70,000
$320,000 $700,000
170,000 380,000
110,000 200,000
30,000 50,000
$630,000 $1,330,000
52,000
13,000
$265,000
1.
Electric Power
OPERATING COST ($/Yr.)
$ 5,000
$ 15,000 $ 45,000
25,000 54,000
30,000 40,000
$ 70,000 $ 139,000
63,000 133 , 000
63,000 133 , 000
$196,000 $ 405,000
2. Maintenance 11 , 000
3. Operating Labor 20,000
Direct Operating Cost $ 36,000
4. Depreciation 26,500
5. Capital Charges 26,500
TOTAL $ 89, 000
* For items included in material, see pages C-lO and C-36.
Note:
1)
One Furnace System. For effect of combined gas cleaning systems
on a multiple furnace shop, see page C-ll.
2)
For a discussion of unusual problems encountered when installing
new collectors at existing furnace shops, and an indication of
cost variances, see page C-ll.
3)
Prices: 1969 base.
-------
C-39
OPEN HEARTH FURNACE - FABRIC FILTER
Gas Volume - ACFM @ 2750F
45,000
60
135, 000
200
350,000
600
Furnace Size - Tons
CAPITAL COST
1. Materiab\' $75,000
2. Labor 40,000
3. Central Engineering 36,000
4. Client Engineering 9,000
TOTAL $160,000
$210,000 $530,000
120,000 300,000
80,000 180,000
20,000 45,000
$430,000 $1,055,000
1.
Electric Power
OPERATING COST ($/Yr.)
$ 8,000
2.
Maintenance
7,700
$ 22,000 $ 54,000
21,000 51,000
30,000 40,000
$ 73,000 $ 145,000
43,000 105,500
43,000 105,5.00
$159,000 $ 356,000
3. Operating Labor 20,000
Direct Operating Cost $ 35,700
4. Depreciation 16,000
5. Capital Charges 16,000
TOTAL $ 67,700
* For items included in material, see pages C-IO and C-36.
Note:
1)
One Furnace System. For effect on cost of combined gas cleaning
systems on a multiple furnace shop, see page C-ll.
2)
This system is not currently in general use, but it has been
successfully applied in the U.S.
3)
For a discussion of unusual problems encountered in installing new
collecting equipment at existing furnace shops, and an indication
of cost variances, see page C-ll.
4)
Prices: 1969 base.
-------
L-
C-40
ELECTRIC ARC FURNACES
The following pages contain cost data relating to electric arc furnaces de-
signed for production of carbon steel. The figures are for completely new
installations. The special problems encountered when installing new control
equipment in existing plants were discussed on page C-ll. Each cost value
applies to a system of two furnaces with a common gas cleaner capable of
handling only one furnace at peak loads at any given time. For effect on
cost of a different system of multiple furnace control see page C-ll.
The volumes listed are based on typical oxygen blowing rates used in furnaces
making carbon steel from cold scrap. Oxygen and exhaust rates may be con-
siderably higher when making stainless heats. An excess of air would typ-
ically be added to the furnace gases for complete combustion of carbon
monoxide and hydrocarbons (the latter, during the melt-down of oily scrap),
and for cooling. These mixed gases would then be water quenched in wet
scrubbing, and also in pre-conditioning of the particles before electrostatic
precipitation, though less water would be used in the latter system to avoid
condensation and to optimize the temperature and humidity conditions for
effective precipitation. Although, for baghouse collection, these gases also
could be water quenched to some extent, effecting an economy in collector
size,it is more typical to cool by use of a radiating exchanger, and to
finish the cooling with controlled air dilution just before the baghouse.
-------
C-41
ELECTRIC ARC FURNACE - WET SCRUBBER (HIGH ENERGY)
o
Gas Volume - ACFM @ 180 F
36,000
25
13 7 , 000
150
210,000
250
Furnace Size - Tons (each)
CAPITAL COST
1.
Material*
$173,000
$511,000
2.
Labor
93,000
277 ,000
$723,000
388,000
3.
Central Engineering
67,000
162,000
215,000
TOTAL
17,000
$350,000
40,000
$990,000
54,000
4.
Client Engineering
$1,380,000
OPERATING COST ($!Yr.)
1. Electric Power $ 40,000 $174,000 $265,000
2. Maintenance 14,000 40,000 55,000
3. Operating Labor 40,000 60,000 80,000
Direct Operating Cost $ 94,000 $274,000 $400,000
4. Depreciation 35,000 99,000 138 , 000
5. Capital Charges 35,000 99,000 138,000
TOTAL $164,000 $472,000 $676,000
* For items included in materials, see pages C-10 and C-40.
Note:
1)
Two Furnace System, alternating peak loads. For effect on cost of dir-
ferent combinations of furnaces per cleaning system see page C-ll.
2)
See also the section on Special Problems Encountered When Installing
New Control Equipment in Existing Plants, page C-ll.
3) A variation from these costs is noted in a case of a single furnace
cleaning system where, after correction for the savings in a 2-
furnace system, the cost would be 40% higher than indicated here.
Remote placement of the scrubber is one factor in this variation.
4)
Prices:
1969 base.
-------
C-42
ELECTRIC ARC FURNACE - ELECTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 5000F
48,000
25
185,000
150
280,000
250
Furnace Size - Tons (each)
CAPITAL COST
2.
Labor
85,000
$465,000 $652,000
251,000 352,000
151,000 197,000
38,000 49,000
$905,000 $1,250,000
1.
Materia1~'(
$159,000
3.
Central Engineering
61,000
TOTAL
15,000
$320,000
4.
Client Engineering
OPERATING COST ($/Yr.)
1. Electric Power $ 8,000 $ 30,000 $ 60,000
2. Maintenance 13 , 000 36,000 50,000
3. Operating Labor 20,000 30,000 40,000
Direct Operating Cost $ 41,000 $ 96,000 $ 150,000
4. Depreciation 32,000 90,500 125,000
5. Capital Charges 32,000 90,500 125,000
TOTAL $105,000 $277 ,000 $ 400,000
* For items included in materials, see pages C-lO and C-40.
Note:
1)
Two Furnace System, alternating peak loads. For effect on cost of
different combinations of furnaces per cleaning system see page C-ll.
2)
See also section on Special Problems Encountered When Installing New
Control Equipment in Existing Plants, page C-ll.
3)
Prices: 1969 base.
-------
C-43
ELECTRIC ARC FURNACE - FABRIC FILTER
Gas Volume - ACFM @ 2750F
60,000
25
230,000
150
350,000
250
Furnace Size - Tons (each)
CAPITAL COST
1.
Materia1~'(
$120,000
2.
Labor
60,000
$441,000 $654,000
209,000 321,000
140,000 196,000
35,000 49,000
$825,000 $1,220,000
3.
Central Engineering
44,000
TOTAL
11 , 000
$235,000
4.
Client Engineering
OPERATING COST ($/Yr.)
1.
Electric Power
$
8,000
$ 40,000 $ 52,000
39,000 57,000
30,000 40,000
$109,000 $ 149,000
82,500 122,000
82,500 122,000
$274,000 $ 393,000
2.
Maintenance
11 , 000
Direct Operating Cost
20,000
$ 39,000
3.
Operating Labor
4.
Depreciation
23,500
TOTAL
23,500
$ 86,000
5.
Capital Charges
* For items included in materials, see pages C-IO and C-40.
Note:
1)
Two Furnace System, alternating peak loads. For effect on cost of
different furnace combinations per cleaning system see page C-ll.
2)
See also section on Special Problems Encountered When Installing
New Control Equipment in Existing Plants, page C-ll.
3)
Prices: 1969 base.
-------
C-44
ELECTRIC ARC FURNACE
Combination Direct Evacuation Control and Furnace Canopy-
Type Area Ventilation System - Fabric Filter
Gas Volume - ACFM @ l400F
Shop Size - 2 Furnaces @ Tons (each)
125,000
20
750,000
120
CAPITAL COST
1.
Material*
$240,000
$1,200,000
2.
Labor
102,000
480,000
3.
4.
Central Engineering
96,000
353,000
TOTAL
24,000
$462,000
88,000
Client Engineering
$2,121,000
1.
Electric Power
OPERATING COST ($/Yr.)
$ 18,500
$ 100,000
2. .Maintenance
21,000
98,000
3.
Direct Operating Cost
30,000
$ 69,500
40,000
$ 238,000
Operating Labor
4.
Depreciation
46,000
212,000
5.
Capital Charges
TOTAL
46,000
$161,500
212,000
$ 662,000
* For items included in material see pages C-IO and C-40.
Note:
1)
Two Furnace System, alternating peak loads. For effect on cost
of different furnace combinations per cleaning system see page C-ll.
2)
See also the section on Special Problems Encountered When Installing
New Control Equipment in Existing Plants, page C-ll.
3)
Prices: 1969 base.
-------
C-45
SCARFING
The following table presents cost data on scarfing units. These
units are of two different sizes. The smaller size is usually
employed when the billets to be handled are never larger than about
50 inches. Larger billets will require the larger gas cleaning
equipment. The material cost excludes the cost of the Smoke
Tunnel. In wet cleaning systems on a scarfer, the water circuit
is normally coupled to an existing slab mill water treatment system,
so that slurry treatment is excluded in this case.
-------
C-46
SCARFING - WET SCRUBBER (HIGH ENERGY)
Gas Volume - ACFM @ 100°F
50,000
100,000
CAPITAL COST
1. Materia1~\-
2. Labor
3. Central Engineering
4. Client Engineering
TOTAL
$114,000 $176,000
66,000 96,000
48,000 68,000
12,000 17,000
$240,000 $357,000
OPERATING COST ($/Yr.)
1. Electric Power $ 38,000 $ 75,000
2. Maintenance 10,000 14,000
3. Operating Labor 5,000 7,000
Direct Operating Cost $ 53,000 $ 96,000
4. Depreciation 24,000 36,000
5. Capital Charges 24,000 36,000
TOTAL $101,000 $168,000
,\-
For items included in materials, see pages C-10 and C-45.
Note:
1)
Prices: 1969 base.
-------
C-47
SCARFING - ELECTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 100°F
50,000
100,000
CAPITAL COST
1. Materia1*
2. Labor
3. Central Engineering
4. Client Engineering
TOTAL
$135,000 $204,000
85,000 112,000
57,000 76,000
14,000 19,000
$291,000 $411,000
OPERATING COST ($/Yr.)
1.
Electric Power
$
8,000
$ 18,000
2.
Maintenance
12,000
16,000
3.
Operating Labor
Direct Operating Cost
5,000
$ 25,000
7,000
$ 41,000
4.
Depreciation
29,000
41,000
TOTAL
29,000
$ 83,000
41,000
$123,000
5.
Capital Charges
*
For items included in materials, see pages C-IO and C-45.
Note:
1)
Prices: 1969 base.
-------
C-48
HCL PICKLING LINE - WET WASHER
The following table presents cost data on a spray washing system for an
HCL Pickling Line acid fume removal system. Most modern lines are now
sized for 80 inch strip. Fiberglass material is used for all duct and
stack work. Fume is scrubbed by successive spray and eliminator units.
For optional acid brick lined tunnel (6 ft. sq.) to outside fume collectors,
add $184 per foot of length to capital cost total, and $44 per foot of
length to annual operating cost total.
-------
C-49
HCL PICKLING LINE - WET WASHER
Gas Volume - ACFM @ 100°F
130,000
Line Capacity
80 inch at 1,000 FPM
CAPITAL COST
1. Materia1~'<
2. Labor
3. Central Engineering
4. Client Engineering
TOTAL
$ 81,000
30,000
23,000
6,000
$140,000
OPERATING COST ($/Yr.)
1.
Electric Power
$ 10,000
2.
Maintenance
6,000
3.
Operating Labor
Direct Operating Cost
5,000
$ 21,000
4.
Depreciation
14,000
5.
TOTAL
14,000
$ 49,000
Capital Charges
~'<
For items included in materials, see pages C-10 and C-48.
Note:
1)
Prices:
1969 base.
-------
C-so
COLD ROLLING MILL - MIST ELIMINATOR
The following table presents cost data for an eliminator system to remove
the palm oil and water mist emission at roll stands of a typical, large,
five stand tandem cold rolling mill. The suction of the system picks up
mist from closure plate enclosed areas at each stand, carries it through
a tunnel to two mist eliminators and fans. The ventilation air thus
cleaned is discharged up a stack. The treatment of collected oil for re-
use or disposal is not included.
-------
C-51
COLD ROLLING MILL - OIL MIST ELIMINATION
Gas Volume - ACFM @ l100F
200,000
Mill Size
80 inch, 5 stand tandem
CAPITAL COST
1. Materia1~'(
$ 85,000
2.
Labor
62,000
3.
Central Engineering
29,000
4.
Client Engineering
7,000
TOTAL
$183,000
OPERATING COST ($/Yr.)
1.
Electric Power
$ 18,000
2.
Maintenance
7,000
3.
Operating Labor
7,000
Direct Operating Cost
$ 32,000
4.
Depreciation
18,000
5.
Capital Charges
TOTAL
18,000
$ 68,000
*
For items included in materials, see pages C-IO and C-50.
Note:
1)
Prices: 1969 base.
-------
C-52
POWER PLANT BOILERS
The following pages contain cost data on several sizes of in-plant boiler
houses. They assume that smoke and fly ash control equipment is being in-
stalled on an existing coal-fired boiler. The figures cover a single boiler
only. Various combinations of multiple boiler-collector units are used in
actual practice, with savings in larger sizes dissipated in additonal duct,
dampers and complicated setup. The stack is considered to be already existing.
Multiple cyclones, when used for primary collecting, are included as they
yield no process advantage to the boiler. Booster fans are included.
Mechanically fed coal-fired boilers may achieve acceptable fly ash control
with multi-cyclones alone. However, large modern boiler houses in integrated
steel plants would usually use pulverized coal firing for efficiency and quick
regulation of firing rate as well as ease of combined or auxiliary firing with
blast furnace or coke oven gas. Pulverized coal's higher percentage of fly ash
a finer size grading requires the use of high efficiency control equipment,
of which the electrostatic precipitator is almost solely used (often in con-
junction with a mechanical primary collector~ as it is more economical than
wet scrubbing. The exhaust gas usually contains a significant amount of sulfur
dioxide, which promotes effective cleaning with a smaller precipitator than
would be required without it. The hot, buoyant gases leaving the precipitator
disperse more readily than if cooled by scrubbing or for baghouse cleaning.
with
Sulfur dioxide emission suppression, using limestone injection with baghouse
collection or absorbtive solution scrubbing, currently undergoing tests for
public utility application, may eventually displace electrostatic precipitation
of fly ash. But the trend in steel plant boilers is toward relatively pollution-
free fuels, particularly gas and oil. Combustion devices to prevent carbon
monoxide emissions are considered 100% process beneficial, and not funded as
pollution control equipment. The formation mechanism and control techniques
for nitrogen oxides emissions are currently under study; a preventive method will
likely be sought for their limitation. The development of acceptable soot
build-up removal means remains a problem.
-------
C-53
POWER PLANT BOILER
Mechanically Fed, Coal Fired BOi1er-Mu1ticyc1one Collector
o
Volume, ACFM @ 600 F
32,000
96,000
Boiler Size, pounds steam/hr.
50,000
150,000
CAPITAL COST
1.
Materia1*
TOTAL
$20,000 $ 60,000
10,000 30,000
10,000 24,000
2,500 6,000
$42,500 $120,000
2.
Labor
3.
Central Engineering
4.
Client Engineering
OPERATING COST ($/Yr.)
1. Electric Power $ 2,300
2. Maintenance 1,700
3. Operating Labor 7,000
Direct Operating Cost $11,000
4. Depreciation 4,300
5. Capital Charges 4,300
TOTAL $19,600
$
7,000
5,000
15,000
$ 27,000
12,000
12,000
$ 51,000
,,<
For items included in materials, see pages C-IO and C-52.
Note:
1)
2)
One Boiler System.
Prices: 1969 base.
-------
C-54
POWER PLANT BOILER
Pulverized Coal Fired Boiler - Electrostatic Precipitator
Volume, ACFM @ 3000F
100,000
200,000
CAPITAL COST
1.
Material*
$260,000
$440,000
2.
Labor
140,000
230,000
3.
Central Engineering
100,000
170,000
TOTAL
25,000
$525,000
45,000
$885,000
4.
Client Engineering
1.
Electric Power
OPERATING COST ($/Yr.)
$ 28,000
$ 55,000
2. Maintenance 21,000
3. Operating Labor 30,000
Direct Operating Cost $ 79,000
4. Depreciation 52,500
5. Capital Charges 52,500
TOTAL $184,000
36,000
40,000
$131,000
88,500
88,500
$308,000
*
For items included in materials, see pages
C-IO and C-52.
Note:
1)
2)
One Boiler, Two Precipitator System.
Prices: 1969 base.
-------
1__-
[
i
C-55
SAMPLE CALCULATION - OPERATING COST ($/Yr.)
The sample illustrates the calculations performed in arriving at the
operating cost for a fabric filter installation on a 150 ton Electric Arc
Furnace. Electric power costs (@ 3/4C per kwh) is obtained by calculating
the total horsepower of all motors (plus power to lights and instruments)
and multiplying by a cost per horsepower factor. Power to a fan motor is
calculated by applying an efficiency to the power required for reversible
adiabatic compression. This latter quantity is called "Air H.P." Power
to a water pump motor is similarly calculated, the reversible pumping power
requirement being called "Water H.P."
1.
Electric Power (*)
Basis: $50/HP/Yr 800 HP Motor
$50/HP/Yr x 800 HP =
$ 40,000/Yr
2.
Maintenance
Basis:
4% of Capital Cost
Capital Cost $825,000
0.04 x $825,000 =
$ 33,000/Yr
6,000/Yr**
3.
Depreciation
Basis:
10% of Capital Cost
Capital Cost $825,000
0.10 x $825,000 =
$ 82,500/Yr
4.
Capital Charges
Basis: 10% of Capital Cost
Capital Cost $825,000
0.10 x $825,000 =
$ 82,500/Yr
5.
Operating Labor
Basis: 3/4 Man/Shift or 18 Manhours/Day
$5.00/Manhour
18 MH/Day x $5.00/MH x
330 Opr.Day/Yr =
$ 30,000/Yr
TOTAL
$274,000/Yr
*
See following page for notes.
** The difference from the 4% standard maintenance
cost with bag replacement cost figured as
described on page C-15.
-------
C-56
*1.
Electric Power Cost @ $0.0075/KWH
Operating Days = 330 Days/Yr
1 HP = 0.746 KW
$0.0075/KWH x 330 Days/Yr x 24 Hr/Day x 0.746 KW/HP
1
x .89 Motor Eff. $50/HP per Yr
*2.
Air HP = 0.0001575 PQ
P = Static Pressure, in. water
Q = Volume, CFM
Motor HP = Air HP
Eff.
Eff. = Efficiency - Range 60 to 70%
GPM x H
Water HP = 3,960
*3.
GPM = Gallons per Minute
H = Head, in Ft.
Water HP
Motor HP = Eff.
Eff. = Efficiency - Range 75 to 85%
METHOD OF DETERMINING EXHAUST GAS VOLUMES IN SIZING
COLLECTING SYSTEMS FOR PRICING
The following sample illustrates the method of calculating the capacity
of the collector in each estimated system. In general, the exhaust gases are
cooled in transit through the system, so that successive items of equipment
in the system will have different volumetric capacities due to gas volume
changes with temperature and with the material additions (dilution air or water
vapor) added to effect cooling.
The starting point is to determine a typical process exhaust gas composition
and volume rate per unit of process throughput (as SCFM/ingot ton). In some
cases this is determined solely by the oxygen lancing rate which generates the
maximum exhaust volume during a steelmaking heat. In the open hearth case,
since fuel and air are customarily added to the furnace during lancing, and
waste heat boilers are generally used for cooling the exhaust gases, typical
volumes of gas at the boiler outlet condition were selected as a starting volume
for the gas cleaning system.
-------
C-57
In the blast furnace, scarfing, sinter plant windbox, pelletizing (dryer or
process), and power plant cases typical modern practice was used as a basis
for determining the process exhaust volume. In materials handling and mist
pick-up cases, where in-drawn ventilation air entrains particles, mist and
vapors to be controlled, typical modern systems were studied to determine
ventilation rates for adequate emission containment and to ensure the inclusion
of sufficient pick-up points to contain a plant's effluent according to the extent
that current technology can meet current standards.
Sample of volume determination method:
1.)
2.)
Peak oxygen rate to process = 1500 SCFM at 32°F
Combustion with air of carbon monoxide
Carbon monoxide (maximum) = 3000 SCFM
2ca + 02 + ~ N2 = 2C02 + 79 N2
21 21
produced.
+ excess air
Combustion products
CO 3000 SCFM
2
N2
+ Excess air
Excess air
1 x 79 x 3000 SCFM = 5640 SCFM
2' 21
= 500% in a typical case
= 5 x 100 x 5640 SCFM = 35,700 SCFM
79
Total 44,300 SCFM
44,300 SCFM x 1.7 (=Factor for two furnaces with alternating peak loads.)
= 75,500 SCFM
3.)
Cooling the gases
The combustion occurs in a water-cooled, double-wall duct where
cooling occurs by radiation and convection of heat to the walls. The
gases leaving this section will typically be at about 12000F. The
size of such indirect heat exchanger will be determined by combining
heat transfer and heat balance equations in an iterative calculation,
based on certain reasonable assumptions of water temperatures, gas
velocity, and water circuit capacity. Optimizing the total cooling
and gas cleaning system is an extensive design task, so that typical
equipment for each system has been selected for this study's estimates.
The cooling by air or water additions to the gases at 12000F involves
a heat balance for calculating resultant volume.
~Mml\n + MwHw = ~~Hn
M = pound moles of each component.
H = enthalpy of each component at conditions.
m = each component of uncoo1ed gas.
n = each component of cooled gas.
w = water at spray water temperature.
-------
C-58
For a final temperature of 500°F, suitable for an electrostatic precip-
itator, about 20% moisture is required by this analysis.
75,500 SCFM + .8 = 94,500 SCFM
94,500 SCFM x (500 + 460)OR -
492°R -
185,000 ACFM @ 500°F
CAPITAL COST BREAKDOWN
The following tables illustrate the relative importance of various
components in total material costs.
This is a very rough breakdown, and variations occur due to capacity
and type of system. However, the relative orders of magnitude are well
maintained. Certain conclusions can be drawn from this tabulation con-
cerning the sensitivity of the total to local conditions. Foundations and
structure may change considerably without having a marked effect on the total.
Very often, a local requirement which tends to increase structure will simul-
taneously reduce foundations. The figures used for these two components are
based upon simple structures supporting the collector near grade, and a soil
bearing value of 4,000 lbs. per square foot.
The stack and fan components are rather closely related to gas volume
and collector type. They are therefore relatively well defined. Electrical,
while an important component, is predicted with comparative certainty from
horsepower.
The key cost element is the collector itself, and it is to this item
that the estimator gives the greatest attention. Generally this will involve
obtaining a price quotation from a reliable manufacturer, although the published
literature also contains useful information.
The second category, labor et aI, is estimated on the basis of
anticipated labor costs for each of the components in Total Material.
Typical rules for this calculation are:
(a)
(b)
Collector:
Labor is about 35% of Material
Fan, motor and starter:
Material
Labor is about 15% of
(c)
(d)
Stack:
Labor is about 100% of Material
Ductwork:
Labor is about 100% of Material
(e)
(f)
Steel:
Labor is about 30% of Material
Foundations:
Labor is about 130% of Material
(g)
Electrical:
Labor is about 150% of Material
-------
C-59
MATERIAL BREAKDOWN
Sinter Plant - Wind box Gas Cleaning
Wet Electrostatic Fabric
Scrubber Precipitator Filter
1. Foundations 4 3 4
2. Ductwork and Stack 2 5 4
Modifications
3. Collector 30 68 71
4. F an and Motor 7 7 10
5. Structural 2 3 3
6. Electrical 7 9 5
7. Water Treatment & Piping 46 1 1
8. Controls 2 4 2
Total 100% 100% 100%
Sinter Plant - Material Handling -
Dust Collection
1. Foundations 4 3 4
2. Ductwork and Stack 12 18 21
3. Co llec tor 28 47 46
4. Fan and Motor 7 5 7
5. Structural 4 7 10
6. Electrical 8 17 9
7. Water Treatment & Piping 35 1 1
8. Controls 2 -1. -1.
Total 100% 100% 100%
-------
C-60
MATERIAL BREAKDOWN
Pelletizing Plant (Moving Grate) - Dust Collection
Wet
Cyclone Scrubber
1. Foundations 5 2
2. Ductwork and Stack 15 5
3. Collector 45 40
4. Fan and Motor 14 20
5. Structural 7 5
6. Electrical 11 8
7. Water Treatment and Piping 1 18
8. Control 2 2
Total 100% 100%
Pelletizing Plant (Shaft Furnace)
Process Exhaust Material Handling
Cyclones Cyclones Wet Scrubber
1. Foundation 2 4 2
2. Ductwork and Stack 10 20 12
3. Collector 41 34 39
4. Fan and Motor 18 13 18
5. Structural 11 12 7
6. Electrical 12 13 8
7. Water Treatment & Piping 2 1 12
8. Controls 4 --.l 2
Total 100% 100% 100%
-------
C-6l
MATERIAL BREAKDOWN
Blast Furnace
1. Foundations
2. Ductwork and Stack
3. Collector
4. Fan and Motor
5. Structural
6. Electrical
7. Water Treatment and Piping
8. Control
Total
Two Stage Venturi
Scrubber System
3
10
46
7
6
25
3
100%
-------
C-62
MATERIAL BREAKDOWN
Basic Oxygen Furnace
Wet Scrubber Electrostatic Fabric
(High Energy) Precipitator Filter
1. Foundations 4 3 2
2. Ductwork and Stack 30 36 37
3. Collector 10 31 32
4. Fan and Motor 9 5 6
5. Structural 6 6 5
6. Electrical 7 8 7
7. Water Treatment and Piping 31 7 7
8. Controls 3 4 4
Total 100% 100% 100%
Open Hearth .Furnace
Wet Scrubber Electrostatic Fabric
(High Energy) Precipitator Filter
1. Foundations 3 2 2
2. Ductwork and Stack 21 25 25
Modifications
3. Collector 15 40 42
4. Fan and Motor 6 2 4
5. Structural 9 9 9
6. Electrical 11 10 8
7. Water Treatment and Piping 33 9 7
8. Controls 2 3 3
Total 100% 100% 100%
-------
C-63
MATERIAL BREAKDOWN
Electric Arc Furnace (Direct
Extraction Fume System)
Wet Scrubber Electrostatic Fabric
(High Energy) Precipitator Filter
1. Foundations 4 3 2
2. Ductwork and Stack 22 31 35
3. Collector 15 35 34
4. Fan and Motor 9 6 7
5. Structural 7 6 7
6. Electrical 10 10 9
7. Water Treatment and Piping 30 5 2
8. Controls .2 4 4
Total 100% 100% 100%
Electric Arc Furnace (Combination Direct
Evacuation Control and Furnace Canopy- Type
Area Ventilation System)
1. Foundations
2. Ductwork and Stack
3. Collector
4. Fan and Motor
5. Structural
6. Electrical
7. Water Treatment and Piping
8. Controls
Total
Fabric Filter
3
37
27
10
10
7
1
5
100%
-------
C-64
MATERIAL BREAKDOWN
Scarfing
Wet Scrubber Electrostatic
(High Energy) Precipitator
1. Foundations 2 2
2. Ductwork and Stack 12 20
3. Collector 30 55
4. Fan and Motor 24 7
5. Structural 5 3
6. Electrical 16 7
7. Water Circuit 7 2
8. Controls 4 4
Total 100% 100%
Power Plant Boiler
Electrostatic
Cyclone Precipitator
1. Foundations 6 3
2. Ductwork 13 33
3. Collector 40 20
4. Fan and Motor 19 2
5. Structural 8 19
6. Electrical 12 17
7. Water Treatment and Piping 0 1
8. Controls 2 5
Total 100% 100%
-------
C-65
GASEOUS POLLUTANTS
The present report does not present cost data on equipment for the
control of gaseous pollutants. Methods for the chemical treatment of gases
for the removal of sulfur and nitrogen oxides are still under development.
Reliable plant cost data will not be available for some time.
Volatiles emitted during the processing of coke oven by-products can
generally be controlled by careful operating control of leaks, drips, drains,
and vents. Any waste gases from flare stacks will probably contain sulfur
oxides, for which treatment methods are not commercially available.
AREA VENTILATION AND EMISSION CONTROL
While the technology for cleaning of effluent material contained in
exhaust ducts from enclosed processes has reached a state of development
where clearly defined practices and equipment can be specified, the means
to clean areas where process materials enter or leave the process enclosure
and to clean the ventilated air from shop structures and outside handling
areas is only now developing. Until the sizing and alternate methods have
been tested by sufficient application, and competitive pricing has evolved,
a definitive estimate of the cost and performance of truly adequate control
means is premature.
The ventilation air volumes may be many times the volume of the gases
cleaned in ducted exhaust circuits from the process; and explosion hazards
at times occur with the influx of air. An example estimated here at the
current level of development is the electric arc furnace melt shop with a
combination of direct evacuation control at the furnace and a canopy above
the furnace. This system provides containment and control during all phases
of the heat cycle when the furnace roof is in place, and, in addition, good
control in preventing fume from escaping from the building during those
operations when there is no local containment at the furnace, such as
charging, teeming,and slagging. The added volume to the collector is I to
2.5 times greater than with furnace flue gas treatment alone, or 2 to 3.5
times greater for the total system. In a case where the canopies are
installed higher, at the roof truss, with no direct furnace evacuation,
the volume is 4 to 5 times greater than it would be if shell evacuation
alone were to be used. The basic oxygen furnace fume system, sized for
peak volumes during oxygen lancing, could be fitted with auxiliary hoods
and dampers to accommodate the hot metal charging and teeming area at low
level, utilizing this peak evacuation capacity for area ventilation during
off-blow periods. The same external operations at the open hearth would
require added exhaust capacity, used in turn on each furnace of the shop.
Drafts in the shop seriously effect the "catch" of open hoods, especially
high-lofted canopies as applied to arc furnaces.
Alternate means are being applied for exhausting and fume removal.
These include various pickup devices:
-------
C-66
1.
Close fitting hoods (with relatively low volume
required) applied to pickling tanks and roll stands
for mist pickup.
2.
Low auxiliary hoods and partial enclosures applied
to pouring operations of hot iron or steel, or the
crushing, screening, loading and discharging of dry
materials (sinter, ore, coal, coke, fluxes and other
chemicals).
3.
Tunnels as applied to scarfing units and conveying
lines.
4.
High canopies with isolation dampers for selective
ventilation of high concentration dust areas, and
total building air-change systems are currently
being evaluated at a few melt shops. Buildings
to enclose extensive areas of material handling
and open processing with many dust generation
points or discharges that are difficult to control
at the source, are used to some extent now (at
crushing and screening stations, for example).
In principle, the enclosure of such an area with cleaning and possibly
recycling of the ventilation air therefrom could effect a reduction in volume
and system complexity compared to that for many high pickup canopies. In
practice, however, while emissions to the atmosphere could be significantly
reduced, hazards would in many cases accompany returning air from the collector
discharge to the workspace, limiting application of this principle. The
magnitude of the task suggests the need for less costly, more effective,
close-to-source control means. The volume required for adequate entrainment
of emissions varies greatly, becoming much larger and less effective when
pickup devices are farther removed from the source.
And while concentrations of pollutant material can be measured at points,
the open-air distribution of concentration cannot be adequately profiled. The
concentration of an air borne material beyond the plant area is subject to the
weather and fall-out variables. Therefore, research is needed to quantitatively
evaluate an area atmosphere by means that could be used for design criteria by
equipment manufacturers and would give correlated information in performance
guarantee tests and the abatement inspector's spot check.
With enough application of engineering design, less expensive means of
controlling presently uncontained volumes will evolve. Plant design can
accomplish some grouping of high dust areas to reduce ventilation requirements.
Process change and new equipment design will increasingly consider pollution
problems as a factor. Building design, currently based on natural ventilation
means, could undergo changes to reduce the extent and facilitate the means of
ventilation. And with optimization of means, a more realistic cost level will
in time evolve.
-------
C-67
Some prior cost tables give estimates of costs for ventilating dust
and mist areas and cleaning the captured air around several processes. The
estimates represent the most adequate systems currently being applied or
quoted for process ventilation needs to supplement ducted process gas
exhausting and cleaning:
Sinter plant material handling,
Arc furnace canopy-type area ventilation,
Scarfing tunnel evacuation,
Pickling line mist removal,
Rolling mill mist pickup.
-------
C-68
EFFECT OF EFFICIENCY SPECIFICATIONS GREATER
THAN CURRENT LEGAL REQUIREMENTS
In many localities, current legal codes specify a permissible particulate
emission at the stack of not more than 0.05 grains per dry standard cubic foot
of gas (or equivalent) exhausted to the atmosphere. Some facilities have met
this requirement or even exceeded it with even fine, sub-micron sized steel-
making dust, using high efficiency filters, scrubbers, and precipitators.
Manufacturers have been able to guarantee this performance with their equipment
in a variety of applications. Also it is noted that blast furnace gas has been
cleaned as finely as 0.005 gr./DSCF when necessary for reuse of the gas in high
energy burners and fine checkerwork of the blast stoves (although this is a
coarser dust than from steelmaking).
This quantity "0.05" is not necessarily an ultimate measure of the effluent
quality that can be obtained. It came into use in the early 1960's, on the.
basis that an open hearth furnace stack plume containing fume at such a concen-
tration had an "acceptable" appearance in many steelmaking areas. The value
"0.05" correlated approximately with the maximum efficiency of electrostatic
precipitators normally offered by manufacturers at that time for collecting
this fume. However, the rapid growth in the use of oxygen lancing of steel-
making furnaces had led to larger quantities of finer fume in their waste
gases today.
A stack plume cleaned to this level is not an invisible plume. The very
fine steelmaking fume escaping at the stack effects a much larger degree of
scatter of transmitted light than the larger particles previously encounteredl,
and thus may be visible even in low concentrations. And yet, visibility of
an exhaust plume persists as a means of checking collector performance, since
it is a very simple comparison of "equivalent opacity" of the plume against
the Ringelmann Smoke Chart.
Local code limitations based on Ringelmann opacity judgments may find a
concentration of 0.05 grains/DSCF of steelmaking fume unsatisfactory. Where
local codes are based on a schedule of allowable fume emission weight per ton
throughput of processed material, the permissible fume rate customarily
decreases for larger production equipment, so that above 30 - 40 tons per
hour, the 0.05 level of control will not often be adequate.
Thus, the widespread use of 0.05 grains/DSCF as a general limiting level
for emissions led to its choice as a basis for calculating the size and cost
of collectors for each process in the tabulations in the appendix. But, in
recognition of the use of more restrictive enforcement methods in some steel-
making areas, and because of the trend in promulgating air quality criteria
which may suppress the emission sources in an area to an increasing degree,
the following indications are drawn of the difference in cost for fume
collecting systems capable of an efficiency beyond the currently practiced
or currently attainable level.
E. R. Watkins & K. Darby, The Application of Electrostatic Precipitation
to the Control of Fume in the Steel Industry, Fume Arrestment, Special
Report 83, of the Proceedings of the Autumn General Meeting of the Iron
and Steel Institute (Brit.) 1964, Wm. Lea and Co., Ltd., London, p. 24.
-------
C-69
Performance Equations
The performance equations of gas cleaners, as currently understood and
applied, in selecting the size and operating parameters for a particular
cleaning application have this in common - they are of the form:
-F(x)
T) = 1 - e
where T) = collection efficiency
or l-T)
= penetration, dust loss, or outlet
concentration as a fraction of the
inlet concentration to the gas
cleaner. It corresponds to some
figure like 0.05, for example:
0.05 (grains/DSCF)
l-T) = inlet conc.(grains/DSCF)
In
(I-f) = -F(x)
F(x) is a function of the size and
operating parameters of the collector.
For a bag filter, Stairmand2 has indicated
-S ~
T) = l-e 0 D ,
~' = function of (~) as shown2
lOO;q=p:. +H'-Ttt-ITf r-'-
J ,T-1- r--- -- - --1 . -..
I , .
L.. ,.,D'_H__-- '....' J= -+-J
,/ ; \
~. 7Spherical Obstacles
D f' '-.~- ~j'-"--+- -.'---r-l D' /D(Spheres)= zero
_1_.--. .. J--1~~ I at Dg/Vf == 24d1
' "": i ,D /D(Cylinders)=zero
. . ,- ~. ,- ! at Dg/Vf == 16
Cylindrical ~ I I
fobstac1es - ,..--:' .c--t--t--j"7-r-t-t -
L.L_..L I - ._-i-~~.-
o 1 2 3 4 5 16
~
Vf
Fig. 1, Relation between Dg/Vf and
target efficiency~ D'/D
2
C. J. Stairmand, Dust Collection by Impingement and Diffusion, Trans.
I. Chem. E., Vol. 28 (1950)
-------
C-70
where D = fiber diameter
g = gravitational constant
v = velocity of gas at filter face
- Q - flow rate of gas = Filter Ratio
- A - area normal to flow
f = settling velocity of particle, as
from Stoke's Law
S
o =
total projected area of all fibers in the filter,
cross section of filter bed
both normal to the gas flow
For an electrostatic precipitator3,4 the Deutsch equation is
T1 = l-e-A'f' /Q
where A' = collecting surface area
, [ 2(k-l)] rE2 d 'f
f = 1 + (k+2) 61TjJ = n t
k = dielectric constant of particles
velocity
E = electric field strength
r = particle radius
jJ = gas viscosity at temperature
And for wet scrubbers, Semrau's5 correlation yields;
- 1 -a(PG + PL)Y - 1 -a(Pr)Y
T1 - -e - -e
where PG = contacting power of gas stream
= 0.157 FS
FS = pressure loss across scrubber, in. water,
exclusive of loss due only to velocity
changes or friction losses across dry
portions of the equipment.
PL = contacting power of liquid stream
= O.583PF qL
Q
PF = liquid feed pressure, psig.
qL = liquid feed rate, g.p.m.
Pr = PG + PL = HP/lOOO CFM, based on Q
Q = actual gas flow at the scrubber, CFM
a,y = constants for a particular dust, related
to particle size and size distribution
-------
C-71
Theoretical Performance Variables
To increase the efficiency of a collector, whose performance is
describable by this logarithmic decay type function, it is necessary to
increase F(X). The variable flow and equipment parameters comprising F(X)
for a particular dust are respectively:
Bag Filter So' ~ or
A'E2 -
Precipitator Q~ -
S DA
0' Q
2L W n E2
A V ~
2L W n E2 - 2LE2
n b W V ~ - bV~
where W = collector surface span normal to flow
L = collector surface length in direction of flow
n = number of collecting ducts
b = separation of collecting surfaces
Scrubber FS and PF(qL/Q')
where qL/Q' is the liquid/gas ratio (gal/1000 CF)
a.
Particle property effects:
1.
Increasing f increases nbag filter
f = 4g r2 p
18~
where p is the density of particle.
50 larger, denser particles are collected more easily.
2.
Increasing r increases n .. Again, larger
precl.pl.tator.
particles are more easily attracted to the collector.
Increasing the dielectric constant of the particulate,
and decreasing resistivity by pre-conditioning via
temperature and humidity (or S02 addition) increases
n precipitator.
3.
Increasing a or y increases n bb' as can be
scru er
estab1ishedS. Both increase with particle size.
3
J. S. Lagarias, Predicting Performance of Electrostatic Precipitators,
Journal APCA (1963) Vol. 13, No. 12
M. Robinson, A Modified Deutsch Efficiency Equation for Electrostatic
Precipitation, Atmospheric Environment, Permagon Press 1967, Vol. 1,
pgs. 193 - 204.
4
5
K. T. Semrau, Correlation of Dust Scrubber Efficiency, Jour. APCA,
Vol. 10, No.3, June 1960, pp. 200 - 207.
-------
b.
c.
d.
C-72
Dust collector geometry effects:
1.
Increasing filter thickness or mat density, decreasing
air/cloth ratio by using larger bag surfaces - increases
nbag filter.
Increasing precipitator length in the flow direction, or
decreasing plate spacing or tube diameter (within limits
of electrical stability) - increases n "" t Since
prec1p1ta or.
the dust loading decreases in the flow direction, it is
possible to achieve an economy by successive stages of
precipitation, each optimized electrically for maximum
efficiency at the respective loading it will see, rather
than simply extending the first stage field.
2.
3.
Decreasing the throat area of a scrubber increases its
pressure drop and increases n bb This can be done
scru er.
by variable geometric arrangement or increasing water rate.
Utility parameter effects:
1.
A partially blinded filter will be more efficient but at
the cost of higher pressure drop and fan horsepower.
2.
Increasing electric field strength increases n ""t
prec1p1 ator
within the limits imposed by the geometry of the collector
and dust properties with respect to sparking. This limit
can be approached more closely with safety if automatic
controls are used to regulate the discharge. Power use
rises.
3.
Venturi Scrubber. Increasing water usage or delivery
pressure in a scrubber increases n bb Increased
scru er.
gas pressure drop gives improved efficiency at the cost
of fan horsepower.
Flow effects:
1.
Q
Even though an increase in face velocity (A) gives a
higher theoretical efficiency in the inertial effect
range, the effect is reversed in dealing with small
particles «l~). And for a filter with a fixed pressure
drop and fixed cleaning routine, the dust buildup will
dominate, so that if increased loading blinds the filter,
causing spillage and less net cleaning, then the following
holds. Decreasing the quantity of gas treated, or using
a larger filter for lower face velocity increases nb
filter. ag
2.
Decreasing the amount of gas treated by lowering
velocity and increasing residence time increases
if distribution of the gas is maintained uniform
plates.
precipitator
nprecipitator
between the
3.
Increasing the quantity of gas treated or increasing throat
velocity increases n bb' by increasing pressure loss
scru er
across the constriction, with an increase in PG or fan
horsepower.
-------
C-73
e.
Temperature effect on viscosity:
Increasing temperature increases ~ gas.
1.
Decreases nbag filter
2.
Decreases nprecipitator
Increasing temperature increases the quantity of gas handled,
again lowering these efficiencies.
Besides altering flow and settling or drift velocity,
temperature also endangers the bags, structures and
mechanisms of the collectors. But filters and dry
precipitators must have an inlet temperature above
the water vapor (and sulfuric acid) dew point to avoid
corrosion and dust caking on the collector, and causing
dust handling problems in disposal conveyors.
3.
Temperature effects the scrubber mainly in increasing the
gas flow, and increasing the saturation water requirement.
Control System Cost Changes
It is a property of decay functions of the aforementioned type
at high efficiency, an increasingly large change in the exponent is
for a small increment in efficiency.
that
required
ELECTROSTATIC PRECIPITATION
For example, an electrostatic precipitator vendor6 reports that the
following increases in precipitator unit size are attendant to the respective
efficiency changes:
Overall Efficiency6
for a Particular Dust
Outlet Loading with 5.0
grains/DSCF Input Loading
Size of Precipitator
Box and Unit Cost6
90%
0.5
x
99%
0.05
2X
99.9%
0.005
3X
This tabulation excludes ductwork, water sprays, hood with its cooling
auxiliaries, stack; but includes the precipitator and its electrical compo-
nents. The fan and motor size and cost, for a precipitator increase (IX),
would be affected by an increment corresponding to an increased static
pressure of about 1-1/2 inches of water (the loss through box X), with the
volume remaining unchanged,
6
private communication,
Pangborn Corporation
-------
C-74
for a precipitator increment, X
S.P. + 1.5
Horsepower increment = S.P. x H.P.
S.P. + 1.5
Fan pressure increment = x S.P.
S.P.
for the total system fan.
Fan volume unchanged.
The following field data7 are illustrative of this:
~
RELAiIVr: 51Z E OF' UN 1-(
100 ~'!i'; jj i:~-- ~'-I-._-_._o.~-~'l-~'-"'---~ y~.- -~= - - - .,.- ~'-'..!:~- _~6f>
j;'-~~=-l---f==r---+---f~-l-i I :1 R.cl(1tio,1\5hiP::~~izc and
90 I I ,- .l--'.rt----j-'--r-j-----I Col!ecllon Ej'j''',ICl1CY
80 __L_:___- 1__:---Ilr--_.L--!-___L_:.- . It m?)' Iw of,,!I\~('i'i"t to 1I,',Ie Ihat th,':
~ I If, I I I (ledll,'IIlI.~ of .)!;).noo ,'ublc feel 1'''1'
10 ---- -. I .-----.Jr-'--l'--1-L \ minute of 01\\'1\ hearth wa,II' g:H,"-; re-
-- -' II =R' L: I quircd 5::>,:Wn ,ql,w.re fed. of _c()I!I'L'lil:.~
eo 11 -'~]--'-, ,--- ",.. surfal:c, for al\ efhCI(~I\(,Y "f ooTIi~>OO
o 10000 30000 60000 70000 90000
20000 400<;10 60000 80000 100000
;.
CO!..LCC.i'lNG !WR FA'E ~A'
!
.
Z
.
>-
U
""I
W
U
h.
L\.
..,
IN
5 a.. F T.
Fig.2
R~lotion~hip-precipitQtor $izc on coll<:ction efficiency.
7
Collection of Metallurgical
Furnace, Jour. APCA, Vol. 14,
A. C. Elliott and A. J. LaFreniere, The
Fumes from an Oxygen Lanced Open Hearth
No. 10 (1964), p. 401.
-------
C-75
The above variation in size corresponds to Deutsch's Law
-A'f'
(l-n) = e
Q
for a particulate of homogeneous size, shape, density,
and composition.
n - inlet loading -R - l-~ where R = outlet loading.
- inlet loading - I.L.'
-constant2 x length
R = constantl x e
log R = constant3 + constant4 x length
(cost)
for a given process and precipitator.
However, real particulate varies in size, density, and susceptibility
to charging (depending on surface and compositional variables) - so that
the least collectable particles remain after each treatment, lowering the
efficiency of subsequent treatments.8 Case 2 below illustrates this with
arbitrary efficiencies:
(Case 1), Deutsch's Law variation,
5 grains -+ Box .5 -+ Box .05 -+ Box .005
R: -+ -+ -+
SCFD X X X
T): 90% 90% 90%
net n: 90% 99% 99.9%
net size: X 2X 3X
(Case 2),
R:
5 -+ BOxx -+ .5 -+ BOXX -+ .1 -+ BOXX -+ .03 -+ BOXX -+ .012 -+ BOXX -+ .006 -+ Box 005
5/l2X -+ .
n: 90%
net n: 90%
net size: X
80% 70% 60% 50% 40%
98% 99.4% 99.76% 99.88% 99.9%
2X 3X 4X 5X 5-5/l2X
A body of blast furnace data 9 for a number of operating furnaces at
various precipitator loadings yields the following progression, which shows
this trend,
net T): 90% 95% 98% 99% 99.5% 99.9% 99.99%
net size: .55X .86x 1.4X 2X 2.65X 4.5X 7.5X
8 G. Penney, Carnegie-Mellon University, Symposium on Gas-
Solids Separation, January 14, 1969
9
B. R. Berg, Development of a New, Horizontal-Flow, Plate-
Type Precipitator for Blast Furnace Gas Cleaning, Iron &
Steel Engineer Year Book, p. 786
-------
C-76
Cost data from precipitator manufacturers indicate a close correspon-
dence to Case 1 in variation of cost (= constant x length) with efficiency.
Guarantees are made on efficiency rather than outlet loading because the
precipitator is not adequately adjustable for cleaning a higher inlet dust
concentration to the same outlet level (say 0.05). In fact, the higher
loading may reach a point where spark-over occurs; so automatic electrical
controls are used to maintain the highest collection efficiency just short
of spark-over. (Large loading differences require design selection of
plate spacing and voltage optimized for the loading and dust properties of
the individual process effluent). The maximum guarantee is presently about
99.5%, although higher efficiencies (around 99.7%) can be reached.
The successive lowering of efficiency found with addition of identical
precipitation units can be compensated; since each successive unit sees a
lower dust loading, plates can be spaced more closely, and voltage optimized
in each succeeding section, while avoiding spark-over. Still, each type of
dust must be tested to determine its collectability as a function of precip-
itator length.
The following tables show some estimated cost differences from the system
cost tabulations for 0.05 grains/DSCF for processes cleaned by electrostatic
precipitation to various outlet dust concentrations. The variation is based
on the Deutsch Law. Capital cost differences include:
Materials:
precipitator plus a fraction of electrical.
Labor:
corresponding to each of above at standard factors.
Engineering:
scaled fraction materials plus labor.
Annual operating cost differences include 24% of capital cost difference
(for capital charges, depreciation and maintenance), plus a fraction of the
electric power for precipitator and fan horsepower increments. Only small
variations were noted for capacity of the cleaner, so that only one size of
cleaners are included for processes previously estimated at 0.05 grains per
SCFD, in several sizes. This study gives cost differences for new equipment,
not alteration costs.
-------
C-77
EFFICIENCY (% DIFFERENC~ AT)
SINTER PLANT - WINDBOX - ELECTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 325°F
630,000
Plant Capacity - TPD
6,000
Outlet Loading (R)
for .8 grains. *
SCFD lnput
Capital Cost Difference (~KR)'
%K
. 0 S
.125
-29
.05
o
.02
+29
Annual Operating Cost Difference (~CR)'
%C
. as
.125
-23
.05
o
.02
+23
Annual Direct Operating Cost
Difference (~DR)' %D.05
.125
-17.5
.05
o
.02
+17.5
*4 grains %
SCFD effluent precleaned by 80~ efficient recovery cyclones.
R
10g10(~)
.05
10g10(2.5)
-~K -~C -~D
R R R
= 29% = 23% = 17.5%
-------
C-78
EFFICIENCY (% DIFFERENCE AT)
SINTER PLANT - MATERIAL HANDLING - ELECTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 135°F
250,000
Plant Capacity - TPD
6,000
Outlet Loading (R)
grain.
for 1 SCFD 1nput
Capital Cost Difference (6KR)'
%K.05
.125
-18
.05
o
.02
+18
Annual Operating Cost Difference (6CR)'
%C.05
.125
-15
.05
o
.02
+15
Annual Direct Operating Cost
Difference (6DR)' %D.05
.125
-10
.05
o
.02
+10
(~)
loglO R.05
-6K -6C -6D
- R- R- R
- 18% - 15% - 10%
loglO(2.5)
-------
C-79
EFFICIENCY (% DIFFERENCE AT)
BASIC OXYGEN FURNACE - ELECTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 500°F
785,000
Furnace Size - Tons
200
Outlet Loading (R)
for 4 grains.
SCFD 1nput
Capital Cost Difference (~KR)'
%K.05
.125
-9
.05
o
.02
+9
Annual Operating Cost Difference (~CR)'
%C.05
.125
-10
.05
o
.02
+10
Annual Direct Operating Cost
Difference (~DR)' %D.05
.125
-11
.05
o
.02
+11
(~)
log10 R.05
log10(2.5)
-~K- -~C -~D
-~---A-~
- 9% - 10% - 11%
-------
C-80
EFFICIENCY (% DIFFERENCE AT)
OPEN HEARTH - ELECTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 500°F
85,000
Furnace Size - Tons
200
Outlet Loading (R)
for 5 grains. t
SCFD l.llpU
Capital Cost Difference (6KR)'
%K.05
.125
-10
.05
o
.02
+10
Annual Operating Cost Difference (6CR)'
%C.05
.125
-9
.05
o
.02
+9
Annual Direct Operating Cost
Difference (6DR)' %D.05
.125
-7
.05
o
.02
+7
(--B-)
loglO R.os
loglO(2.5)
-6K- -6C -6D
- ~-~-~
- 10% - 9% - 7%
-------
'C-81
EFFICIENCY (% PIFFERENCE AT)
ELECTRIC ARC FURNACE - ELECTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 500°F 1
185,000
Furnace Size - Tons 2
150
Outlet Loading (R)
f 3 grains
or SCFD input
Capital Cost Difference (6~),
%K.05
.125
-10
.05
o
.02
+10
Annual Operating Cost Difference (6CR)'
%C.05
.125
-10
.05
o
.02
+10
Annual Direct Operating Cost
Difference (6DR)' %D.05
.125
-9
.05
o
.02
+9
(~ )
log10 R.os
log10(2.5)
-6KR -6CR -6DR
10% = 10% - 9%
Note:
1 Assumes humidification of process fume is capable
of maintaining particle resistivity in satisfactory
collection range.
2 Two furnace system.
-------
C-82
EFFICIENCY (% DIFFERENCE AT)
SCARFING - ELEqTROSTATIC PRECIPITATOR
Gas Volume - ACFM @ 100°F
100,000
Outlet Loading (R)
grain
for 1 SCFD input
Capital Cost Difference (~~),
%K.os
.125
-19
.05
o
.02
+19
Annual Operating Cost Difference (~CR)'
%C.os
.125
-18
.05
o
.02
+18
Annual Direct Operating Cost
Difference (~DR)' %D.os
.125
-17
.05
o
.02
+17
(~ )
logi0 R.os
loglO(2.5)
- -~~ - -~CR -~DR
- 19% - 18% = 17%
-------
C-83
WET SCRUBBING
In the case of the venturi scrubber, a vendor10 reports the following:
Inlet Loading, Grains/DSCF
1.0 3.8 5.0 10
90% 97.4% 98% 99%
Efficiency
96.2% 99% 99.24% 99.62%
Outlet LoadinglO
Grains/DSCF
Capital
Cost 10
.10
x
.038
1.43X
The operating expenses vary similarly for a venturi scrubber as effi-
ciency is increased. This is shown in Figure 3 and 411 for an open hearth
application where a decrease in outlet loading from 0.1 to 0.01 grains/SCFD
results in more than doubling the annual operating cost of the fan. For a
given size adjustable venturi, the increased efficiency requires only an
increase in available horsepower to the fan and selection of a higher
pressure fan, and operating power consumption increases directly with the
pressure drop.
PRESSURE DROP VS
SCRUBBER PERFORMANCEll
lCLEANED - GAS
V
0.1
0.08
0.06
DUST LOADING, grains / scfd
0.02
ORE AND LIME BOIL
rAND WORKING PERIOD
0.04
CHARGING, MELT DOWN
AND HOT METAL
0.01
0.008
26
28
30
32
34
36 .
38
40
Fig. 3
35
Fig. 4
PERMANENT PRESSURE DROP, inches of water
45 55 65 75
FAN OPERATING COST, thousands of dollars/yr
10
private communication, Pangborn Corporation
11
Bishop, C. A., et a1, "Successful Cleaning of Open-Hearth Exhaust Gas
with a High-Energy Scrubber," Jour. APCA, 11 (2), 83-87, (February 1961)
-------
C-84
The above venturi-cleaned open hearth application involves oxygen lancing
during the periods noted on the upper curve. Dust loading was low (.82 to
.87 grains/SCFD during oxygen periods, and .35 to .45 grains/SCFD during the
charging, melting and hot metal addition periods). When this data is corrected
to a typical peak 5 grains/SCFD loading for today~ oxygen lanced furnaces it
yields the following correlation:
(grains)
R, outlet loading SCFD
6P, venturi pressure drop (in.w.)
.125
34.7
.05
41
.02
48.2
6PR R -0.178
~o = (~)
. 5
However, Semrau5 has applied his scrubber correlation to data from Basse12
for the regression line of a plot of non-lanced open hearth gas cleaning
efficiency vs. pressure drop at various operation conditions. This gives,
for a peak 5 grains/SCFD inlet loading
R (grains/SCFD)
6P (in.w.)
.125
44
.05
78
.02
136
6PR
t:,P.05=
62
(....!L) - .
.05
The higher numerical exponent seems more in line with results from other
steelmaking fume.
Basse'J2b1ast furnace data gives:
(grains)
R, outlet loading SCFD
6P venturi pressure drop (in.w.)
.125
15.8
.05
23
.02
33.2
6PR
6P.05
=
R - 1t03
(--) .
.05
-------
C-85
Basse's data for an electric furnace making 20% ferro-silicon show:
tiP 1 S 3
R (-L)-'
tiP . 0 S = . 0 S
For the typical scrap-charged electric arc furnace. wet scrubbing applications
are sparse and data are not available for a scrubbing power-efficiency correlation.
Venturi gas cleaning data o~ the basic oxygen furnace have been developed13
(grains)
R. outlet loading SCFD
tiP. venturi pressure drop (in.w.)
.125
27
.05
41
.02
60
tiP
giving 2
tlP.05
R -.417
(:os)
Data from a pilot size conventional venturi scrubber applied to clean
scarfing machine effluent have been published14
(grains)
R. outlet loading SCFD
tiP. venturi pressure drop (in.w.)
.125
34
.05
60
.02
108
tlPR
-=
tlP.os
R -.631
(-:os)
12
B. Basse. Gases Cleaned by the Use of Scrubbers. Blast Furnace &
Steel Plant. November. 1956. 44. p. 1307:
13
H. P. Willet. D. E. Pike. The Venturi Scrubber for Cleaning Oxygen
Steel Process Gases. Iron & Steel Engineer. 38. July 1961. p. 126.
14
American Air Filter Co. bulletin 2~4-l0M-3-65-CP
-------
C-86
The following tables give some estimated cost differences for processes
cleaned by wet scrubbers of the high energy types. The variation is based on
the preceding scrubber application data. Capital cost differences include:
Materials:
Fan and motor plus fraction of electrical.
The venturi itself is assumed adjustable
and of sufficient strength for the higher
pressure difference across its walls. Water
rates are unchanged.
Labor:
Corresponding to each of above at standard
factors.
Engineering:
Scaled fraction of materials plus labor.
Annual operating cost differences include 24% of (capital differences)
plus electric power for horsepower increments.
-------
C-87
EFFICIENCY (% DIFFERENCE
AT)
BASIC OXYGEN FURNACE - WET SCRUBBER
Gas Volume - ACFM @ 180bF
440,000
Furnace Size - Tons
200
Outlet Loading (R)
for 4 grains I
SCFD nput
Venturi Pressure
Capital Cost Difference
(lIKR' %K.os)
Drop (liP, in.w.)
.125
27.5
-4
.05
41
o
.02
60
+5.5
Annual Operating Cost
Difference (lICR' %C.os)
.125 27.5
.05 41
.02 60
-6
o
+9
Annual Direct Operating
Cos t Dif ference (WR' %D. 0 S )
.125
27.5
41
-8
o
.05
.02
60
+12
An empirical relationship is indicated
(liP R ) 41 R - 417
- -1 =- [(-). -1]
liP. os 19 .05
lI~ lICR lIDR - (lIPR -1) R - 417
11.8% = 19.4%= 25.9% - liP. os [(.05). -1]
Doubling the venturi pressure drop would cause a 25.9% increase in direct
operating costs. Venturi loss of 41 in.w. is 85% of system loss, which accounts
for 40% of total horsepower (including an unchanged water pumping and treatment
system) in this case. Power cost is about 72% of direct operating costs.
(40 x .85 x electric power fraction) + (11.8 x maintenance fraction) = 25.9%
Note: One furnace System
lI~ lICR -
5.5% = 9% -
lIDR -
12% -
lIPR-lIP.OS - lIP.os
60-41 - 19
-------
C-88
EFFICIENCY (% DIFFERENC~ AT)
OPEN HEARTH - WET SCRUBBER
Gas Volume - ACFM @ 180°F
90,000
Furnace Size - Tons
200
Drop (t.P, in.w.)
Capital Cost Difference
(t.~, %K. 05)
Outlet Loading (R)
grains
for 5 SCFD Input
Venturi Pressure
.20
32.6
-11
.10
50.2
- 6.5
.05
78
o
Annual Operating Cost
Difference (t.CR' %C.os)
.20 32.6
.10 50.2
.05 78
-15.5
- 9
o
Annual Direct Operating
Cost Difference (t.DR' %D.OS)
.20 32.6
.10 50.2
.05 78
-19.5
-12
o
t.~ t.CR - t.DR
-6.5% = -9% - -12%
t.PR-t.P.OS - t.p.os
50.2-78 - -27.8
(R ) 78 R - 62
t::.os-1 = -27.8[(.05)' -1]
t.KR t.CR - t.DR
18.2%= 25.1% - 33.5% =
(t.P R )
t.p.os-l
=
R - 62
[(-)' -1]
.05
No te:
One Furnace System
-------
C-89
EFFICIENCY (% DIFFERENCE. AT)
SCARFING - WET SCRUBBER
Gas Volume - ACFM @ 100°F
100,000
Outlet Loading (R)
1 grain
for SCFD input
.0833
Venturi Pressure
Drop (tiP, in.w.)
Capital Cost Difference
(tlKR' %K.os)
44.3
-9
.05
61
o
.03
81.S
+11
Annual Operating Cost
Difference (tlCR' %C.os)
.0833 44.3
.OS 61
.03 81.S
-17
o
+21
Annual Direct Operating
Cost Difference (tlD , %D 05)
R .
.0833 44.3
.OS 61
.03 81.S
-23
o
+28
tI~ tlCR tlDR tlPR-tlP.OS - tiP. 05 (tlPR ) 61 R - 631
11% = 21% = 28% = 81.S-61 - ~ tlP.os-1 = 20.S [(.Os)' -1)
tI~ tlCR tlDR-
32.6% = 62.S% = 83.4%-
(tiP R )
tiP. 05 -1
R - 631
[(-)' -1)
.OS
-------
C-90
FABRIC FILTRATION
For an acceptable, constant dust penetration through a fabric filter,
the face velocity or air volume to cloth area ratio must decrease with
decreasing particle size and density or increased inlet loading. For a
lower allowable penetration, the face velocity would similarly decrease.
Thus, a more difficult or more thorough cleaning job would involve in-
creased cost to provide more filter surface area. This is exemplified
in the extreme case of a reverse jet-cleaned filter where face velocities
are the highest encountered.
A reverse jet cleaned fabric filter calculated from the charts below15
for 70% by weight of the dust loading less than 10 microns and S.G. above
2.0 yields the following information on sizing:
For Inlet Dust Loading Efficiency Required for Filter Ratio Q/A Filter
Grains/DSCF 0.05 Grains/DSCF Outlet (CFM/Sq.Ft.) Area
5 99 % 16.4 X
10 99.5 % 13.7 1.2X
20 99.75% 9.6 1. 7X
25 99.8 % 7.6 2.lX
The effective filtering body is the dust cake layer on the bags.
This does not at this time seem amenable to treatment which will improve
efficiency. However, the bag, when new, and to a lesser extent when
cleaned, holds little dust cake so that the fabric, with its small,
dust laden fibers, is the basic filter until the filter cake layer
reforms. As the small fibers break in service, the bag loses filtration
capability. Additionally, the lower flow resistance of a cleaned bag
passes a greater volume of air at reduced cleaning efficiency than when
dust-coated; but at a higher velocity which betters the collectability
of larger particles and worsens the diffusional efficiency dominating
small particle collection.
An adequately designed baghouse will have a bag-cleaning cycle suited
to the inlet dust loading from the process to which it is applied. This
cycle is often automatically adjustable, so that the filter maintains the
same average (time-wise) efficiency with variations in inlet dust loading
and gas volume. The bag-cleaning period will begin when the collected dust
causes the pressure drop through the filter to reach a set-point pressure.
In addition, the fabric weave and material is chosen with the special
character of the process effluent in mind (such as particle size distribution).
Economic factors (bag life and initial cost differences) also enter this choice,
but increased efficiency can only be achieved by choosing from a group of
fabrics which will give cleaning to the required level. Present practice
usually gives efficiencies of 99%+, and bag filters frequently give the highest
efficiencies of the applicable cleaning devices considered for a process; so
this selective optimization does not offer much potential except as currently
ongoing research reveals new materials and weaves.
113uffalo Forge Company, Bulletin AP650
-------
~ .50
o
.~ 60
~
A 0 70
C-91
>-
...
The following non",.". "I'" ... ,',., ,.,
as a convenient means of selecting
Filter Ratio for pl'eliminal'y detel"
mination of the size Ael'olurn Dust
Colleclor that will best satisfy the
needs of YOlll' installation.
In many inslanees the nomograph will
pl'ovide determination of the optimum
Filter Ratio. Because of the great vad-
ety of possible sel'vjee condi tions and
the effecl of the chamcteristics of
specific dusts. final det el'mina t ions of
Filter Ratio will he made by Buffalo
Forge Company. This proceduI'e pro-
vides the greatest assurance of correct
and economic select ion of equipment
for YOlll' installation.
IIOW TO (:SE
In o)"(ler to selecl Fillel' Ratio, three
condit ions pertaining to your specific
dust collection job ar~ needed. They m'e:
a. The approxima te percen tage, by
weight, of dust pm'tides 10 microns
or smaller.
b. Dust content of the air entedn" the
Aeroturn Col1e,~tol' expressed in t~rrns
of grains (7000' pel' lb.) pel' cubic fool.
Use average 01' normal values for both
dust and air quantities,
e, Specific gravity of the material to
be collected.
TO ('SI';:
1. From appropdate point on vel,tical
stale A draw hori7.0ntal line intel'sect-
ing sloping line 8.
2, From appropriate point on vertical
scale C draw horizontal line intel's('(:t-
ing the sloping line which r(,pl'esents
the proper specific gravity range for
the material to be collccted.
3, Now, draw a straight line between
points selected in stel;s 1 and 2 above,
The intel'section of this line with hOl'i-
zontal scale n gh'es the Fillel" Ratio,
This value may now be used in the Size
Selection Chat"\' on the next page. to
detel'mine the Ael'oturn Dust Collectors
applicable to your I'equiremenl.
As a case in point, a process having a generally large particulate
may be adequately cleaned by a certain bag to .05 grains/SCFD. If lower
outlet loading is required, a suitable bag, which gives similar results
on a process with finer effluent, may be substituted. The overall cost may
or may not be larger. The choices are presently limited by limited test
results on filtration properties of fabrics, and state-of-the-art in fabric
technology with regard to dust abrasion, flexural durability, and chemical
and temperature resistance. Electrostatic interactions of various fabrics
with dust particles may prove to be significant.
~o
~ 80
Q;
.D
~ 90
a"
100
30
22 20
25
20
15
C
10
..:
oj
....
'"
!?
C) 5
Z
0
<
9
..
on
::>
0
Fig. 5
-------
COLLEc.rOR SIZE 15
C-92
36.000
~.OOO
32.000
~.OOO
28.000
26.000
2'.000
22.000
i e
o'
~ 20,000 ~
~ .f
u ~
0( 18.000 ~
..
0( "'
V ~
.. ~
.. 16.000
... '"
;; ~
".000
12.000
10.000
'.000
6.000
'.000
2.000
Fig. 6
1,000
1.200
',400
t 1.600
o
~
~ 1.800
..
0(
..
..
~ 2.000
2,200
2.400
2.600
2.800
3.000
3.200
3,400
3,600
FILTER SIZE
200 16.8
400 16.10
16.12
J2-8
600 16-14
16-16
16.18
800 32-10
16.20
48.8
J2-12
32-16
48.10
]2-14
48.12
32-18
32.20
48.14
64-12
48-16
48.1,8 64-14
48.20
----.,
HOW TO tJSE
64.8'
This chart provides a conven-
ient and accurate means for
selecting the applicable size or
sizes of Aeroturn Dust Col-
lectors when Filter Ratio and
required Air Cleaning Capacity
are known.
64.10
1) Draw a line from the re-
quired Capacity through the
applicable Filter Ratio to in-
tersect the Filter Area scale.
2) From this point of inter-
section, draw a horizontal line
through blocks designatin~
Filter Size selections for de-
sired Capacity.
64.16
3) If horizontal line passes
through more than one Filter
Size, first size intersected will
be most economical. Subse-
quent selections will be less
economical.
64-18
4) For capacities larger than
shown: Use 'h the l'equired ca-
pacity in the above procedure.
Filter Size thus selected must
be doubled for full capacity.
64.20
No clear correlation has been advanced relating efficiency to operating
parameters:
A higher pressure drop may be expected to increase filtration
action at the cost of additional power, but the trend of such
variation is not known.
A higher filter face velocity (higher air/cloth ratio) theoret-
ically yields a higher efficiency of collection for particles
large enough to be governed by inertial laws, but these are
ordinarily cleaned to near 100% efficiency anyway, so the filter
size is governed by loading. The small particles which escape
collection migrate under diffusional impulses, and higher effi-
ciency here would increase with residence time (lower face
velocity, lower air/cloth ratio, thicker filter media). The
relative effects of these coacting collection mechanisms is not
sufficiently understood at present for use in practical design.
These foregoing factors are insufficiently defined to use at present for
an economic study of the effect on costs of changed efficiency requirement.
-------
C-93
CONCLUSIONS
A situation of diminishing returns is indicated by performance
of the exponential type and in many cases a 0.05 grains/SCFD outlet
tration becomes a practical maximum level for improving efficiency,
though it is by no means an absolute limit.
equations
concen-
even
The state of the art, then, allows the gas cleaner manufacturer to predict
performance, design a collector, and guarantee it with some confidence to
about 0.05 grains/SCFD outlet loading for particles greater than 2 microns
in size. At lower outlet levels, his experience is limited. And the large
change in size or operating parameters required for further small efficiency
increases would magnify the uncertainties known to exist 16,17,4,18 in these
simplified exponential relations and experience-based empirical constants used
in them by the engineer.
Measurement techniques used to determine dust loading in the ducted stream
before and after the collector leave much to be desired, especially where
small concentrations and even smaller changes in concentration are to be used
as evidence of guaranteed performance or violation. Lack of homogeneity of
most dusts from iron and steel processes make the use of monitored data (light
scattering or transmission, for example) difficult to interpret, or the equip-
ment difficult to calibrate, for all the variations in dust composition, size,
gas flow rate, etc. caused by process changes during a heat cycle, or from heat
to heat. Isokinetic sampling (sampling at stream velocity) with traversing
probes involve much averaging (in time and space) with calculation and readjust-
ment continuing during the traverse - this costly and of questionable accuracy.
Null probes, too, operate with a significant degree of error in trying to bal-
ance small pressure differences. Neither approach to isokineticity can give
a time history of emission rate during the course of a rapidly changing heat
cycle as only two or three traverses can be run at best in an hour. Gas density
(composition and state) and moisture content data should be monitored continuously
and used as input to sampling rate determinations during the course of a sampling
test; for deviations here can seriously effect the loading measured as grains
of dust per dry standard cubic foot of carrier gas.
16 Electrostatic Precipitation, Weakness in Theory, G. W. Penney,
Mechanical Engineering, October 1968, p. 32.
17 Turbulent Gas Flow and Electrostatic Precipitation, M. Robinson,
Jour. APCA, April 1968.
18 M. W. First, L. Silverman,
Industrial Fabric Filters,
p. 581.
Predicting the Performance of Cleanable
Jour. APCA, Vol. 13, No. 12, Dec. 1963,
-------
C-94
From such quantitative data as can be obtained, control equipment is
designed, often with a costly excess performance factor built-in, and guar-
anteed somewhat conservatively. The guarantee is proven (or indicated) by
standard sampling tests, and no assurance is given that any particular level of
Ringleman chart greyness will not be exceeded. Research is needed to find
a method to inexpensively quantify dust concentrations; and agreement is
needed to correlate design and enforcement bases of measurement.
Very fine particulate matter, because of greatly extended surface area,
causes a much greater scattering of light, even in small concentration. A
Ringleman comparison must thus in some way account for the nature of the
emission being sampled to indicate relative concentration. If this correlation
can be made, then this economical method of testing might be used to obtain
both adequate design data and unquestioned legal evidence.
In the case of very fine steelmaking dusts from open hearth, electric arc,
and basic oxygen furnaces, the collector performance is difficult to predict
because,
a.
The particle size distribution determination is difficult
to quantify with present methods for sampled dust, and
the correlation of this data to "in situ" dust in the
furnace effluent gas is in doubt. (Large discrepancies
in reported BOF dust sizing is a case in point). The
smaller the size the greater the difficulty.
b.
Agglomerative properties of the dust are not well
established and the effect of this on sampled dust
sizing and on collection mechanisms in the gas cleaners
is not well understood.
c.
The mechanism of collection upon which the performance
equations are based (inertial and electrostatic forces)
tend toward zero efficiency in the size range of the
bulk of steelmaking dusts «2 microns), where molecular
interactions dominate the motion of particles.
Actually, any attractive interactions or agglomerative tendency would
be beneficial to particle collection on a clean collecting element, but
joining particles into larger, inter-adhesive masses would tend to blind
a filter matrix (lessening gas handling capacity), or interrupt electrostatic
precipitator field propagation about the wires and plates, and make the
collector surface hard to clean off and the dust hard to handle. This in
some cases necessitates close control of temperature and humidity.
-------
C-95
For low velocity collectors (inherently large and thus economically
inefficient for larger particle collection), a diffusional mechanism can
give significantly large collecting efficiencies. (The effect is greatest,
in theory, near zero microns size, and decreases with increasing particle
size.) A middle ground"exists around 0.9 microns in a bag filter where
minimum efficiency can be as low as 10% - exactly in the center of concen-
tration of some 70% of steelmaking dust. This is shown in an efficiency -
particle size relation drawn by Stairmand19 for a new, unused bag, for
which the effect is most pronounced. See also figure 8, and for electro-
static precipitators, figure 11, where this effect seems to be indicated.
I
a
I
~ =='
10
'I.
10 - i":-
10 = V
= - -'.-', = -
I,
1
1
o
.
,.."C\,. I'll WIC.OM'
Fig. 7
New Fabric Filter19
~
.
"
.
r
t
~
Further research effort is indicated to:
a.
Develop techniques for confident particle size distribution
data representing the dust as it exists in the effluent gas.
b.
Determine the extent of agglomerative effects and their
agency in gas cleaning and effluent sampling mechanisms.
c.
Utilize the diffusion mechanism for small particles in
an optimum way while retaining economical and efficient
inertial mechanisms for large particle collection. If
the valley of low efficiency between the size ranges where
diffusion and inertia are effective cannot be narrowed by
this development, then another tack at development must
investigate other gas-solid interaction phenomena for
possible use in gas cleanirig. Particle interaction effects
may be important here.
Develop economical methods to measure
adequate for design purposes and well
used for obtaining enforcement data.
In view of the foregoing difficulties it maybe' concluded that changes
in legally required efficiency levels (to outlet loadings below about 0.05
grains/DSCF) would at this time be based on much questionable design measure-
ment, and theory (whose extensiqn into this range is also questionable). The
cost of such changes, as indicated by present understanding of the mechanisms
of collection with proven equipment, would become increasingly great for
collection efficiency changes of very small magnitude-changes which can only
be measured with an error of the same order as the change sought.
d.
dust concentrations -
correlated to methods
19 C.J. Stairmand, Design and Performance of Modern Gas-Cleaning Equipment,
Jour. Institute of Fuel (Brit.), Feb. 1956, p. 58.
-------
C-96
FACTORS AFFECTING GAS CLEANER PERFORMANCE
The processes in the iron and steel industry can and do depart from design
capacity and operating conditions for a number of reasons:
Economic pressures dictate the continued improvement
in productivity of an installed furnace.
Technological improvements make possible significant
increases in productivity (such as the introduction
of oxygen blowing to open hearth and electric furnace
steelmaking) of a new or existing facility.
Batch handling of especially specified heats or runs
of varying sizes and treatments.
Slack market conditions may require output cutbacks.
And with changes in productivity effluent quantities increase or diminish
both in gas volume and loading. Operating conditions in the gas cleaning sys-
tem can vary with these conditions as we11aswith the weather, gas utilization
program, raw material charge, etc. And non-continuous or batch type metallur-
gical processes vary during the course of a heat in both quantity and condition
to the effluent.
To maintain satisfactory gas cleaning performance under these conditions
it is necessary to have anticipated these factors in designing the pollution
abatement system, rather than specifying for average conditions. Maximum capa-
city should be installed or adaptation to additional capacity provided. Adjust-
able equipment can often be used to optimize performance over a range of oper-
ations.
Provision should be made also in the initial installation to meet, or to
add and adapt equipment to meet, expected future requirements of the pollution
control codes both as to dust content of effluent and treatment of objectional
gas and solid chemicals in the effluent.
Assuming proper design and selection of equipment, which would usually
give superior performance over an extended life span with timely maintenance,
and would offer the economies of optimization--any variation or variability in
the process, control equipment or performance would generally require an added
cost. And any unique feature of a particular gas cleaning application (part-
icle size, dust loading, corrosion, etc.) would generally require a departure
from a more general system design (and cost).
The following exerpt from G. Punch20 summarizes the performance factors
required for effective particulate removal:
20
Gas Cleaning in the Iron & Steel Industry, Part II: Applications, G.
Punch; Fume Arrestment, Special Report (83) of the proceedings of the
Autumn General Meeting of the Iron & Steel Institute (Brit.), 26 Nov-
ember 1963. (1964), Williams Lea & Co., Ltd., London, p. 10.
-------
C-97
Punch, Gas Cleaning
The Clean Air Act and the increasingly wide use of oxygen in both the
classical and the recently developed top-blown converter processes
have combined to create an urgent need for highly efficient cleaning
of high-temperature effluent gases containing submicron iron oxide
fume to the visibility threshold of 0.05 grains/CF. In order to sat-
isfy this need, manufacturers of gas cleaning equipment had first to
find how collectors which had already been well proved in other fields
could be adapted to applications of which they had had no previous
experience. This entailed not only the establishment of the empir-
ical design parameters concerned with efficiency, but also a very
close consideration of the ability of each type of collector to cope
with unavoidable variations in gas volume, temperature, humidity,
solids concentration, etc.
The flexibility of any given type of collector (i.e. its ability
to operate efficiently without breakdown over a wide range of con-
ditions) is much more important in practice than its theoretical
efficiency at constant flowrate and temperature, etc., and the best
unit for any given application will often not be the one which a
comparison of efficiency and cost based on idealized operating con-
ditions would indicate.
Every manufacturer who can offer a complete range of equipment
must weigh very many factors before finally offering one particular
type of collector. He may be handicapped, particularly in the case
of a completely new installation, by a shortage of basic process
data, but he can usually arrive at a fairly accurate assessment of
the relative strengths and weaknesses of the possible units.
Although the size distribution and shape of dust or fume parti-
cles are of course the factors which determine the fundamental suit-
ability or otherwise of any given design of collector for a partic-
ular application, other characteristics of the solids, the carrier
gas, and the process itself must ... also be carefully considered and
their effect on the collection device evaluated before a final sel-
ection is made.
The agglomerating propensities of the solid particles are im-
portant because they determine the size distribution of the particles
presented to the collector. The extent to which agglomeration into
clusters or chains of particles will have proceeded, and hence what
the effective particle size will be immediately before the process
of final collection is begun, cannot be accurately predicted, and in
practice allowance is made for it in the empirical design constants
used by equipment manufacturers. Agglomeration after collection
affects the caking properties of dry material, making it more easily
released from filter fabrics, less liable to re-entrainment during
precipitator rapping, and more easily settled from liquid effluent.
The electrical resistivity of the material to be collected is
of the utmost importance if a dry precipitator is to be used.
-------
':-----
C-98
Punch, Gas Cleaning
If the collected material is not free-flowing when dry it may
create dust handling problems. Hygroscopic dust will give rise to
similar difficulties in 'dry' collectors, unless humidity and the
temperature of solids and gas can be maintained at safe levels
by control of the process, lagging, external heating, warm air
purging, or by a combination of these.
For the collection of dusts which are corrosive when wet the
obvious choice is a dry type of unit, unless there is a risk of
condensation. If the wasregases contain water vapour which comes
from the process itself, or has been added for cooling or con-
ditioning them, and sudden temperature surges are likely, elab-
orate precautions against condensation may be needed, anB a more
compact wet unit constructed from corrosion-resistant materials
may be more economical as well as more reliable.
The physical and chemical characteristics of the carrier gas
must also be carefully considered when a collector is being chosen.
The effect of variations in gas temperature and humidity, in par-
ticular, must be carefully investigated especially if, as is al-
most always the case, they accompany or cause changes in gas volume
and dust characteristics during and after collection. These fac-
tors are affected by the method of hooding, cooling, and volume
and temperature control, but no matter how carefully these are
engineered the characteristics of the process may still cause the
collector to be subjected to cond i tions which are far from ideal
and impair its operation either directly by affecting the collec-
tion process, or indirectly by hindering dust discharge or caus-
ing structural damage. Collectors of different types are more
or less susceptible to different non-ideal conditions, as shown
in Table I. The table is only intended to indicate some of the
fundamental strengths and weaknesses of hig~efnciency dedusters
in relation to fluctuating operating conditions of one sort or
another, and is not intended to be a comprehensive summary; it
does, however, demonstrate the importance of factors which have
nothing to do with the properties of particles.
Additionally, Table I from Punch20 indicates operating conditions
which affect the dust collector efficiency at a peak level of equipment
maintenance and factors which require regular attention (cleaning the collec-
tor surface adequately to match dust loading, temperature control in dry
collectors to minimize moisture and heat deterioration and maximize dust
removal and handling properties) to insure peak efficiency throughout the
life of the equipment.
The effect of operating conditions on the effective performance of the
gas cleaning function, the effect of those conditions which cause maintenance
difficulties and shorten service life, and the effect of those conditions
peculiar to a particular furnace type or process on the design and selection
of gas cleaning equipment--are best judged in the light of operating and
design experience. The literature contains scattered discussions of this sort.
-------
"-
Punch, Gas Cleaning
C-99
TABLE I
EfTc;:t on collector performance of fluctuating operating conditions
Dry pbr.e precipitator Fabric filter Scrubber
Irrigated precipitator
Ten,pcraturc
Humidity
Flowrate
Corrosive
solids or gas
Ir.'ct
\"(I:!\.'~'tltr.,tio!1
Corrosion can be
avoided by accurate
temperature control,
insulation, auxiliary
heating, bypassing, or
corrosion-resistant
materials of construction.
Filter fabric may be
damaged.
J :-.itial design must be Efficiency not affected
based on peak loading. by increased loadings:
effect on pressure drop
depends on duration of
surges, but can be reduced
by tern porary increase
in cleaning intensity.
Normally up to 650' F
with star:dard construc-
tion but momentary
peaks of I OOO'F can
be tolerated. Tempera-
rure must be selected
to suit electrical
characteristics of dust.
Insufficient moisture
may lower efficiency
by increasing dust
resistivity. High
humidity with low.
temperature may ca Jse
condensation, possible
corrosion, insulator
and plate cleaning and
dust disposal diffi-
culties. Accurate
control of spray
cooling essential.
Efficiency increased
if flowrate reduced,
although gas distribu-
tion may deteriorate.
Normal maximum
temperature c!cpends on
fibrc used. Up to say
275'F with o-ganic
synthetics, 600 'F with
fibreglass. Hi!;her peaks
tolerable but reduce
bag-life dispropor-
tionately.
Operation below dew-
point leads to bag-
cleaning troubles.
Chemical and physical
damage to fabric likely.
Dust disposal difficulties.
EtJ1ciency little affected
by tlowrate. Pressure
drop reduced as vol~me
falls.
Normally below 200'F
with pn:saturatio:1.
Surges can be prevented
if maximum water rate'
always used in saturator.
Efficiency unaffected by
changes in humidity,
providing gas remains
near saturation.
~'ater-rate and.'or throat
area must be adjusted to
compensate for chat'ges
in inlet volume.
Alternativdy volume
may be kept constant
by air-addition.
Special materials of
construction will prevent
corrosion. High-pressure
(high top speed) stainless
steel fan impellers can
give trouble.
Efficiency increased by
opaation at lower
flowrates.
Special materials of
construction eliminate
corrosion but price may
rule out.
Initial design must be
based on peak loading.
!1"!!tiai dt:sig!,l ;nusc be
based on p,'ak loading.
--..----
The following sections are discussions of emission cleaning for three iron
and steel industry processes which present difficult problems of equipment
selection, performance, and maintainability. These discussions were chosen
for their concise and comprehensive consideration from an application point
of view of the critical factors of equipment use. While they center on
British practice (where raw materials, processes, codes, etc. have some var-
iance from general American practice), the discussion of each factor remains
pertinent, with perhaps some difference in degree, to a consideration of a
corresponding American plant. 50 that, after considering all the conditions
existing on a particular job of equipment application, an engineer may find
somewhat more or less difficulty in his case.
-------
--
C-lOO
SINTER PLANT20
MAIN STRAND GASES
The gases withdrawn from the main strand of a sinter machine present
a fairly difficult gas cleaning problem, not, as in most other iron-
and steelmaking applications, because high efficiencies must be
achieved on very fine particles, but because of~her characteristics
of the dust and the gases themselves.
Volumes are great and the use of medium and high pressure drop
collectors would involve large non-productive power consumption.
The waste gases contain large quantities of both sulphur oxides
and water vapour. Consequently they have a high (acid) dewpoint
so that condensation and corrosion are a constant danger, aggra-
vated by the wide fluctuations of temperature which occur from
time to time.
The coarser fractions of the dust burden are exceedingly
abrasive.
Hence the ideal dust collector will have the following character-
istics:
A pressure drop as low as possible.
Ability to operate efficiently over a wide range of tempera-
tures without ill effect from occasional dampness of dust and
collector internal surfaces.
A construction which minimizes condensation, lends itself to
reasonably economical corrosion prevention, and is not susceptible
to plugging during the occasional but inevitable periods of oper-
ation below dewpoint.
Freedom from abrasion troubles, preferably by complete avoid-
ance of high velocities, otherwise by pre-collection of the coarse
abrasive dust fractions prior to passing the gases through any
collector in which high velocities are used.
Dust Characteristics
The particle size analysis of the dust content of sinter strand gases
can vary between quite wide limits.... The type of dust to be dealt
with depends on the mix fed to the strand,i.e. proportions of home
and foreign ores and return fines, and also on whether or not the
burden is conditioned in a pelletizing drum. It must be remembered
that changes in dust composition occur as the rate of sintering alters
and the relationship between temperature, flame-front penetration,
and position on the strand varies.
Dust Loadings
The general level of dust concentration is affected greatly by
the nature of the material fed to the machine and can vary from
plant to plant between 0.1 and 1.0 grains/NCF and may occasionally
reach 1.2 grains/NCF. The rate of solids emission is very sensitive
to variations in the progress of the sintering process along the
length of the strand. Dust is mainly generated early in the sinter-
ing process and again when the flame-front reaches the bottom of the
bed. It has been suggested that in the intermediate zone the in-
-------
C-lOl
Punch, Gas Cleaning
creased moistness of the lower part of the bed causes it to act as
a crude filter and hence to pass less dust. It has been found that
as complete sintering approaches the discharge end of the strand,
i.e. as the mean hottest windbox number increases, the dust load-
ing rises noticeably.
Gas Temperature
Gas temperatures usually fluctuate between
but 100-1500C is the most common range....
(U.S. practice is in the range 300-4000F.)
600C and a maximum of 2000C
Gas Composition
The only constituents of the waste gases which are important from
the gas cleaning point of view are water vapour and sulphur oxides,
both of which affect the frequency and severity of condensation in
the cleaning system. There will usually be about 10% water vapour
by volume in the gases and the sulphur oxide content, expressed
as S02, may be as high as 1.5 grains/NCF. Unfortunately, no acid
dewpoint figures are availiable, but water dewpoints as high as 500C
are encountered and acid dewpoints considerably higher than this
must therefore occur. So far as condensation and corrosion are con-
cerned, the relatively high proportions of water vapour and oxides
of sulphur in the gases complicate the design and selection of gas
cleaning equipment....
(But they tend to faciiitate electrostatic precipitation.)
ChQice of Dust Collector
In the authors' opinion the sulphur oxide content of the gases rules
out wet methods of collection, since these would result in difficult
liquid effluent problems, and saturated gases having hardly any
thermal lift and still containing some sulphur oxides would consti-
tute an air pollution problem worse in some respects than the orig-
inal one.
The choice of a dry collector will be dictated by the quantity
and size range of the dust in the case under consideration, the
space available, the pressure drop which can be tolerated and the
outlet loading required. Generally speaking, particularly for
dusts at the coarser end of the normal range.. . and if an. outlet con-
centration of 0.15 grains/NCF can be tolerated, a settling chamber
will be adequate. If, say, 0.10 grains/NCF is the highest accept-
able outlet loading, or the dust is finer, or a settling chamber
cannot be accommodated within the space available, cyclones may be
used, but their pressure drop (up to 6 inwg) is a disadvantage and
they must be specially constructed to withstand erosion by abrasive
dust particles. For a stack loading of less than 0.10 grains/NCF a
more efficient type of collector must be used.
If outlet loadings down to 0.05 grains/NCF are required, the
only suitable device is the electrostatic precipitator. Were it
not for the constant danger of condensation, the fabric filter
would be a possibility, but a filter fabric would 'blind' when oper-
ated under moist conditions. It is true that the silicone-treated
-------
C-l02
Punch, Gas Cleaning
fibreglass fabric (which would have to be used in any case to with-
stand the high maximum temperature) is much less susceptible to plug-
ging than are the natural and organic synthetic cloths, and has
been found to regain its porosity on drying out, but there would
always be a risk of the cloth becoming 'starched' with soluble salts
and failing prematurely through what can only be described as crack-
ing. Fibreglass, which has poor flex resistance in the first place,
is exceptionally vulnerable to this sort of trouble. This type of
collector also compares unfavourably with the precipitator from the
points of view of pressure drop, spa~e requirement, and maintenance
cost, and would not be recommended for main strand gas cleaning.
Although, in common with all other collectors, a precipitator
for this application has to contend with occasional condensation, its
operation is not unduly affected by moist conditions, providing pre-
cautions are taken against corrosion, and providing it has efficient
rapping gear which will clear any...build-up and prevent progressive
deterioration in its performance. The water vapour and sulphur oxides
in the waste gases 'condition' the dust and together with the rela-
tive coarseness of the dust....
(...assist the precipitation process.)
The Head Wrightson sinter machine installed at the works of the
Skinningrove Iron Co. Ltd, Saltburn-by-the-Sea, is provided with a
Head Wrightson/Research Cottrell dry plate precipitator. The mach-
ine was designed to process a wide variety of home and foreign ore
mixes, and experience indicated that the dust burden in the waste
gases could be reduced by a simple settling chamber from 1.0 to 0.3
grains/NCF.* The gas volume from the 16 x 6ft square windbox machine
is 180,000 CFM. The precipitator has two treatment zones, energized
by a 15 kVA 230 mA transformer-rectifier set and operates at a
treatment velocity of 6.8 ft/s. In view of the expected intermit-
tent operation, it was thought advisable to fabricate the collector
plates in copper-bearing 'Corten' steel (O.l%C max., 0.1 - 0.3%Si,
0.5 - 1.0'~n, 0.3 - 0.5%Cu, 0.5 - 1.5%Cr, 0.1 - 0.2%P) and these
have withstood the adverse conditions very well without noticeable
deterioration. The interior of the precipitator shell is protected
with gunned aluminous cement and the whole unit is thermally in-
sulated to minimize condensation. The precipitator...was designed
to operate at an average temperature of 300oF, and at an efficiency
of 86.7%, corresponding to an outlet loading of 0.04 grains/CF. The
design performance has been...achieved and, although the sinter
plant has worked on a one or two shift per day basis and the precip-
itator has undergone an abnormal number of start-ups, there has
been no deterioration of its internals. The sinter fan was inspec-
ted in August 1963, 20 months after commissioning, and showed no
sign of wear other than a general smoothness over the faces of the
blades; it is estimated that it will operate for at least another
3 - 4 years without requiring maintenance. The machine had produced
250,000 tons up to the time of the inspection. Reduced fan mainten-
ance and plant downtime are two useful indirect benefits of efficient
main strand gas cleaning.
The dust discharged from the precipitator hoppers is conditioned
in a pelletizing drum and the pellets produced are returned to the
process via the return fines conveyer.
*(This is lower than typical loading in American practice.)
-------
C-I03
Punch, Gas Cleaning
DISCHARGE ~ND EXHAUST SYSTEM
The whole of the discharge end of the sinter machine is usually
completely enclosed; 100 tons or more of dust per day may be re-
leased by the equipment in this area (i.e. the end of the strand
itself, the breaker, hot screen, and discharge to cooler). Air
volumes vary with the size of sinter machine and the completeness
of hooding, and are between 30,000 and l50,000CFM. Gas temperatures
are usually between 400 and l500C. Both the loading and the size
range of the entrained dust are affected by the designs of hoods
employed, and the exhaust volumes allocated to them, but dust
burdens are typically in the range 4 -6 grains/NCF of which 80%
might be <100 ~m and 10% <10 ~m.
Careful hood design, combined with adjustment of individual
exhaust rates during commissioning, can reduce both grain loadings
and the proportion of coarse abrasive particles carried in the
gases. It is relatively easy to obtain collection efficiencies of
90-95% by means of simple high-efficiency cyclones, and the stack
discharge in such cases will contain about 0.5 grains/NCF of dust,
90% of which is <10 ~m. At this sort of grain loading the stack
plume does not appear offensive; all the same it represents a very
high rate of solids emission (up to 700lb/h on a large plant), and
more and more interest is being shown in alternative higher-
efficiency collection methods.
For cleaning the tip end emission the fabric filter and dry
plate electrostatic precipitator are two obvious possibilities. Wet
methods can be employed (self-induced spray units are fairly often
used in the USA), but are not to be recommended because they intro-
duce a secondary (liquid) effluent problem, and are liable to suffer
from wet-dry interface troubles and sometimes from sludge discharge
problems. At first sight, the fabric filter would appear to be
ideally suited to this application, providing it is designed so as
to avoid excessive scouring of the bags by abrasive dust, and prop-
erly maintained so that a small leak in one bag cannot 'grit-blast'
a hole into an adjacent one and start a rapid and messy chain re-
action. The first requirement is quite easily satisfied, but the
second is not so straightforward and a short period of neglect
could have expensive and inconvenient consequences in the form
of extensive bag replacements and operation at reduced capacity.
From the point of view of efficiency and capital cost, the fabric
filter is a 'good buy', but running costs are a most important
factor, and cannot be accurately forecast.
While the operating characteristics of the dry plate precip-
itator are quite predictable for this application, discharge end
precipitation is difficult because of the high resistivity of the
dust at the gas temperatures normally encountered, when the moisture
content is less than about 1.5% by volume. In cold, dry weather
the water vapour content may be as low as 0.5% by volume, and under
these conditions unstable precipitator conditions are liable to
occur at temperatures around 600C.
The addition of relatively small quantities of water vapour,
sufficient to raise the volume percentage to 2.0, leads to a marked
-------
C-I04
Punch, Gas Cleaning
improvement in precipitator performance, as does the addition of 100
ppm of S02. If a guaranteed efficiency is to be maintained under
every circumstance, and at all times, and if water vapour or S02 can-
not be added, the precipitator will be perhaps three times as large
as a unit which will operate satisfactorily under all but the driest
conditions. It is therefore well worthwhile either to mix in gases
from some other part of the sinter system or to add steam. If the
problem of conditioning can be overcome this is a very straight-
forward precipitator application....
(Application of either dry system at the discharge end when water
cooling of the sinter is employed could involve re-introduction of
moisture control problems.)
QUENCH GASES
The quenching of hot fines in pug mill or drum gives rise to large
quantities of fine dust, particularly during periods of erratic
plant operation.
In a typical installation the volume of gas vented from the
drum was 7600 NCFM at 40-l20oC, containing between 5% and 24% water
vapour by volume. It was found that the dust loading was greatly
affected, not only by the quantity and distribution of spray water,
but also by the quality of sinter being made. During normal oper-
ation of the machine the loading was found to vary between 1.3 grains/
NCF when sintering was complete, and 4.7 grains/NCF when incompletely
sintered material was being discharged from the strand. Shortly after
commissioning, before the sprays had been adjusted and while the
operation of the machine was abnormally erratic, the mean dust con-
centration had been 4.8 grains/NCF (corresponding to a rate of dis-
charge of nearly 400 lb/h) and the peak loading 33.2 grains/NCF in gas
volumes of 8500-ll,000NCFM. This illustrates the effect of plant
operation on stack emissions. The final emission rate averaged 140
lb/h compared to 280 lb/h from the main stack and 135 lb/h from the
tip end cyclone stack.
The quench stack dust is rather fine (99% <100 ~m, 30% <10 ~m,
10% <3 ~m). To date, to the best of the authors' knowledge, no
attempt has been made to clean the gases, but if cleaning were re-
quired in an existing plant a self-induced spray washer or an orifice
scrubber would be the best solution, unless the gases could be hand-
led by an existing discharge end cleaning system. In a new plant
the authors would recommend mixing the quench gases with the dis-
charge end exhaust air to give a conditioned mixed gas stream capable
of being cleaned in a precipitator of....
(conservative size to 0.05 grains/NCF.)
The problem can be entirely avoided if the hot returns are conveyed
direct to the mixer (preheating the mix often has much to commend it).
THE OPEN-HEARTH FURNACE20
The open-hearth furnace emits waste gases equivalent to between 74,000
and 134,000 NCF/ton of crude steel. Volume rates of flow are usually
-------
C-I05
Punch, Gas Cleaning
within the range 170 - 280 NCFM/ton of furnace capacity. Fume load-
ings vary from one period of the melting cycle to another. During
charging and melting down concentrations of less than 0.5 grains/CF are
usual, and during fett1ing they are even lower. The highest fuming
rates occur during refining and lancing, and loadings of 6 grains/NCF
are common during oxygen injection. The concentration of fume is of
course affected by the volume of excess air which is allowed to
enter the furnace as well as by the fuming rate. .
The composition of the furnace waste gases depends on the fuel
used. The most important constituents from the gas cleaning point
of view are water vapour and sulphur oxides. The percentage of
water vapour may be as low as 2% or as high as 25% and a concentration
of sulphur oxides calculated as S02 of 3.52 grains/CF has been re-
ported for a producer gas fired furnace. A sulphur oxide concentra-
tion of 0.36 grains/CF has been reported for furnaces using 70%
coke-oven gas and 30% pitch-creosote. In general high acid dewpoints
are to be expected and if dry collection is to be used condensation
must be guarded against.
Providing suitable precautions are taken against condensation,
a dry cOllector may be used, and numerous dry plate precipitators
have been installed in OH melting shops in recent years. Purely
from the precipitation point of view, OH fume collection is fairly
straightforward,
(with automatic controls)
due largely to the conditioning effect of the water vapour and sul-
phur oxides in the gases, but the fume tends to be 'sticky' and an
efficient rapping system is essential. The collector casing must be
well insulated to minimize condensation, and if the unit is to operate
under pressure the top insulator housings must be pressurized with
warm air to keep the insulators dry. Dust should preferably only
be stored in the hoppers in an emergency because it tends to bridge,
and it may be advisable to heat the hopper sides. Some condensation is
bound to occur at start-up, and it is advisable to clear as much
collected dust as possible from the interior of the precipitator
while it is shut down. If this is not done conveyers and dust dis-
charge valves may become clogged with moist dust. If possible the
precipitator should only be energized when it has reached its normal
operating temperature, so that little dust is collected in it when it
is sweating. If these...precautions are observed the dry plate pre-
cipitator will operate continuous1y...if not, severe build-up, elec-
trical and operating difficulties, and corrosion will be experienced.
A dry plate precipitator installation on a 250 ton tilting OH
furnace......fo110ws a waste heat boiler and ID fan, and is designed
to clean 78,900 CFM of furnace gases at a maximum temperature of
2800C. The design inlet loading is 5 grains/NCF during oxygen lanc-
ing and the outlet cleanness 0.04 grains/NCF. The precipitator is
insulated, the hoppers are steam - heated, and the insulator compart-
ments on top of the unit are pressurized with 600 CFM of air at 2000F
to prevent outward leakage of dirty gas and to keep the insulators
both dry and clean. The precipitator has three treatment zones each
of which is energized by a 21 kVA, 250 mA transformer rectifier set.
This is quite a good example of a precipitator fitted into a very
restricted site, utilizing turning vanes to reduce inlet and outlet
duct sizes without detriment to gas distribution.
-------
C-I06
Punch, Gas Cleaning
The fabric filter may be used for OH gas cleaning but is more
susceptible than the precipitator to condensation troubles, has a
much higher power consumption, and requires more space. A filter
serving one of the Ajax furnaces was reported to operate at a pres-
sure drop of 8 inwg and to have a bag-life of only 20 weeks. There
seems to be no reason why filters of modern design using improved
high-temperature fabrics should not operate satisfactorily at a pres-
sure drop of 4 -5 inwg with a bag-life of a year or more, but prolonged
pilot-plant testing would be needed to prove the durability of the
filter fabric.
Both the irrigated electrostatic precipitator and the high-
energy scrubber are capable of cleaning OH fume to 0.05 grains/CF
or better, but they would have to be constructed from espensive
corrosion-resistant materials and would create secondary problems
of liquid effluent treatment and loss of stack gas buoyancy.
ARC FURNACES 2 0
Furnace Pressure Control
For consistently good fume control at minimum rates of extraction,
automatic control of furnace pressure is essential. The indicated
pressure which it is necessary to hold within the furnace depends on
the position of the pressure pick-up. The accuracy of control re-
quired is of the order of ~ 1.0 inwg for furnaces melting OH grades
of steel but may be as fine as t 0.03 inwg for a furnace producing
alloy steels. The control system used must have a high speed of
response if it is to cope with sudden fluctuations within the furnace.
Gas Cooling or Conditioning
Temperature at the outlet of the combustion chamber may be upwards
o
of 1000 C and the gases must be cooled before they can be cleaned.
The methods available are air dilution, indirect cooling by heat ex-
changer, and evaporative cooling. It is considered that the latter
is often the best compromise on the grounds of simplicity, final gas
volume, space requirements, and initial cost.
However, the type of collection device used will often dictate
the Inanner in which cooling is carried out. With wet methods of
collection, a comparatively small spray tower may be used (without
fine control of the cooling sprays) and air dilution or indirect
cooling would be pointless. If dry precipitation is preferred, the
gases must be conditioned (most simply with water) and if a spray
conditioning tower is required for this reason the gas will be spray
cooled to the desired precipitator operating temperature. The fabric
filter does not require pre-humidification of the gases for efficient
operation and, is, moreover, exceptionally vulnerable to condensation.
The preferred method of cooling in this case will depend upon whether
the filter fabric is organic-synthetic (e.g. OrIon or Terylene) and
therefore not suitable for operation at over l30oC, or fibreglass,
which will withstand up to 250oC. In the former case air dilution
-------
C-l07
Punch, Gas Cleaning
or indirect cooling may be used, but in the latter spray cooling
should present not difficulties providing a good control system
is fitted.
System Capacity and Safety
....The details of safety require that...very conservative assumptions
are made. The problem of explosion hazards has been considered in
recent papers.
Air may enter the system at the air break between elbow and
fixed fume pipe and at the combustion chamber, as well as through the
furnace openings. The volume of air entering by each of these routes
is unimportant providing (a) that control of fume is obtained and
(b) that the final waste gas volume is such that even if ,combustion
has been incomplete an explosive mixture cannot be formed.
The combined effects of combustion and dilution have been cal-
culated for the lancing period, and are shown in Table II. However,
the rate of evolution of combustion following the addition of oily
scrap cannot be predicted, and it must be remembered that in practice
the operation of a fume cleaning system must take second place to the
production of steel; allowance must also be made for occasional de-
ficiencies in the standard of both operation and maintenance of
cleaning systems. Hence, although under ideal conditions an 02 to
waste gas ratio of 10:1 would no doubt be adequate, it is recommended
that a ratio of not ~ess than 15:1 be used.
Current understanding of the explosion problem is incomplete:
explosions have been reported even in conservatively designed systems
following errors in operation, and it is considered more prudent to
use theory to predict the magnitude of apparent safety margins rather
than to reduce these to the point where (due to the intrusion of
incalculable factors) they do not exist, and a variation in the
process or a mistake by an operator can cause an explosion.
'J',\BI.E II
",r"cts of com hll';timt and dilution
-----.
Ratio of waste
gas:oxygen
injection nowratc:
I~ U Carbon
monoxide if no
cOI1Jh,Jstion occurs
Approx, ~o
combustion for safe
op<:ration (based on
IOO'~" oxygen
utilization)
--------.-- - ----- ... .-.-------.---- ---------.
22: )
16: )
15: 1
)oJ: )
10: 1
6: 1
5: 1
9.)
12.5
!3'?,
14.3
20.0
33,3
40.0
Nil
Nil
5~'~1
10~.~
32(~,
50%
55%
'----'---'--
GAS CLEANING
The furnace gases may be cleaned to 0.05
(wet or dry), fabric filter, high-energy
scrubber-precipitator.
grains/CF by precipitator
scrubber, or combination
-------
--
C-108
Punch, Gas Cleaning
Dry plate precipitation is relatively straightforward providing
the gases are properly conditioned. It is therefore ideally suited
to direct extraction systems but much less so for hood or conventional
hood vent installations. (A) 75 ton furnace...has been fitted with
direct extraction fume control equipment and fume is to be collected
by a dry-plate electrostatic precipitator (... unit referred to below).
The lancing rate of this furnace is 1200 CFM and the volume during
lancing, after combustion and cooling 49,200 CFM. The precipitator
is designed to clean a total of 83,200 CFM from the existing furnace
and another which is to be added in the future, from 6.5 to 0.05
grains/NCF.
Furnace gases will pass through a water-cooled elbow and refract-
ory-lined fixed duct connected by a power-operated movable sliding
sleeve, into a gas burner followed by a combustion chamber. They will
be cooled and conditioned in the rectangular spray tower and will en-
ter the precipitator at a temperature of 5000F.
The fabric filter is theoretically ideal, having a uniformly
high efficiency irrespective of throughput but it must be carefully
designed and protected against condensation. Filtering velocities
may also be as low as 2 ft/min so that space limitations will often
exclude this type of cleaner.
The high-energy scrubber operating at a pressure drop of 30 inwg
or more will do a satisfactory fume-cleaning job
(U.S. codes would require about 45 inwg)
and its compactness is a great advantage, particularly when the
available space is limited. Power may be saved by regulating the
fan in an efficient manner to suit the rate of exhaust required for
fume control and the pressure drop needed at different periods of
the melt to give the statutory final gas cleanliness, but this is
only practicable if the pressure drop of the scrubber can be adjusted
to the desired level over a wide range of f10wrates.
It must be stressed, that in the long run, regular maintenance and
attention to operating conditions affect the cost and effectiveness of
any gas cleaning unit. The incorporation of automatic controls, operator-
proof operating controls, scheduled preventive maintenance, anticipation of
adverse process conditions and raw material possibilities are important to
the continued performance of gas cleaning equipment after the guarantee
period.
We will concern ourselves more specifically with the following parameters
which affect gas cleaner performance:
1.
Effect of gas volume changes on collection efficiency of a dust
collector.
2.
Effect of pressure drop within the gas cleaner on efficiency
and capacity of collector.
3.
Effect of dust loading: effect of collector surface renewal on
pressure drop, volume and collecting efficiency.
-------
C-109
4.
Effect of particulate as generated in each matallurgical pro-
cess (particle density, particle size, size distribution) on
efficiency of each applicable dust removal device.
5.
Effect of temperature on efficiency of and gas volume to col-
lector, and required gas conditioning for cooling and humid-
ification before dust removal.
a.
Gas analysis as it affects conditioning required prior to
cleaning and exhausting. (Refer back to Punch's examples
with respect to combustibles, 802' water vapor.)
Corrosion and the use of water.
b.
c.
Abrasion and chemical effects of dust.
6.
Adaptability of the particulate removal system to removal of
gaseous pollutants.
1.
Effect of gas volume changes on dust removal efficiency---
As previously indicated, the volume of effluent gas emitted by a metal-
lurgical process may vary greatly during a heat, or according to changing
production level of the process. And since the efficiency of dust removal
changes when volume changes, it becomes necessary to
---operate at constant volume with air substituted for effluent gas
deficiency,
---or, use a gas cleaning device which adjusts itself to volume
changes, or is adjustable to satisfactory efficiency over a range
of volume.
Self-induced or orifice washers (of the Rotoclone type) and certain
fluidized bed scrubbers can adjust themselves, essentially at constant
efficiency. Adjustable throat venturis, orifice-wedge and flooded disk
scrubbers can be adjusted to suit a range of gas flow. These and other wet
scrubbers can also be uneconomically flooded to achieve the same effect.
Multiple units (nested cyclones; parallel scrubbers, precipitator tubes
or ducts; multiple venturis, baghouse filter tubes) can be partially blocked
off to maintain high (design) efficiency at reduced volume, with economy of
water and power use.
---or, design for maximum possible effluent volume, and "over-clean" at
reduced volumes.
-------
~ --
C-110
GAS DISTRIBUTION21
It will be appreciated that the efficiency of a precipitator is
greatest when the velocity of the gas through the cross-section of the
electrode system is uniform, and no gas is bypassing the electrode sys-
tem. This is ensured by the construction of...models...of the precipi-
tator and inlet flue system. The flow conditions in the model are ad-
justed to give the same Reynolds number as the full-scale plant, allowance
being made for sca£ factors, gas viscosity, and density. The flow pattern
in the model is corrected using splitters and baffles, their position
being determined by experiment. Such model tests permit requirements to
be worked out in advance, and avoid the difficulties in carrying out such
work on site on the finished plant.
2.
Effect of pressure drop within the gas cleaner on efficiency and
capacity of collector---
Electrostatic precipitators will experience negligible change
resistance to flow in operation because of large cross section and
build-up conditions (temperature and humidity) and regular rapping
removal.
in
control of
for dust
Bag filters, when new, have very low resistance and efficiency. Sometimes
a pre-coat of dust is applied to make the initial cleaning of process fume
more effective, for the buildup of dust increases both efficiency and pressure
drop, until the cleaning (by shaking or reverse flow of air) cycle is initiated
(often by a pressure signal). Then efficiency will be at a lower (but still
effective) level until the dust layer reforms on the fabric.22 Wet scrubbers
increase in efficiency with increased resistance due to mechanical constriction
of the throat area or added water input. The proportionality of change as
attributed to Semrau's correlation was described earlier.
Cyclone's efficiency also depends upon pressure drop. These are only
used with coarser, easily collected dusts, however, and usually with a view
to product recovery as much as to gas cleaning. As such, they may usually be
regarded as process equipment. The rules relating pressure drop, capacity and
efficiency are available in the Air Pollution Engineering Manua1.23
3.
Effect of dust 10ading---
An electrostatic gas cleaner is in principle a constant efficiency de-
vice, so that any change in inlet loading should be reflected proportionally
in the outlet stream loading. However, in actuality changes in dust build-up
21
E. R. Watkins & K. Darby, The Application of Electrostatic Precipitation
to the Control of Flume in the Steel Industry. Fume Arrestment, ibid.
22
M. W. First, L, Silverman, Predicting the Performance of Cleanable Indus-
trial Fabric Filters, Jour. APCA, 11, 12, Dec. 1963, p. 581.
23
Public Health Service Publication No. 999-AP-40
-------
C-lll
occur, adversly affecting the propagation of a uniform electric field. Plate
spacing must be designed to accommodate the condition of heaviest expected dust
loading. Automatic controls are often required to maintain an optimum elec-
tric field without spark-over.
A well designed bag filter will be unaffected by a change in inlet
loading except that automatic cycling of the bag cleaning system will adjust
t.o the change within its limits of variabi li ty.
A wet scrubber will yield constant efficiency for a given pressure drop.
Therefore, a change in inlet loading will be reflected proportionately in the
outlet loading. However, wet scrubbing systems can be very adaptable to
changing conditions, provided sufficient power is applied. A venturi
throat can easily be closed to maintain a given effluent level with increased
dust generation in the process. A process whose fume output varies widely
with time could be matched by cleaner adjustments to maintain a constant
acceptable output of fume.
4.
Relationship of particulate as generated by different processes to
collecting efficiency.
In Appendix C, "Characteristics of Emissions," of the technical counter-
part of this report, entitled ,~ Systems Analysis Study of the Integrated
Iron and Steel Industry" (May 15, 1969), some data are presented on the nature
of particulate material as generated by various processes and conveyed by
gases emitted from the process vicinity. This dust is generally non-uniform
from one particle to another and from process to process. The differences
may be categorized as particle size, shape, density, and composition.
The mechanisms of particle collection on which gas cleaning equipment
are based vary in collecting efficiency generally with particie physical pro-
perties. The chemical nature of the dust may affect its susceptibility to
electric charging. (This would primarily affect electrostatic precipitation,
but could be a second order effect in wet scrubbing and fabric filtration)
and interaction with water droplets. (Solubility and chemical activity would
affect the water cycling, dust handling, and collector surface maintenance in
wet collectors.)
Some general variations in the efficiency of collectors with these par-
ticle properties can be drawn. Stairmand19 has presented grade efficiency
curves (typical of industrial collectors in the mid-1950's) for various
types of dust collecting equipment. These show the efficiency of collecting
particles of a given size. The curves were based on test results using a
standard dust (TableTII) with a 2.7 specific gravity. These curves can be
used to indicate the relative applicability of each type of equipment to
different process fumes. As shown in figures 14 to 25, the efficiencies
generally are lower (often dropping abruptly) for finer grades of dust.
Some devices are more economical to operate but generally do not clean fine
particles from gases as well as others.
By making a density correction, the curves can be applied to dusts for
which particle size distribution data (as in TableITI) are known. Some dis-
tribution data is given in the aforementioned Technological Report, Appendix C.
No quantitative data are available upon which to base corrections for particle
-------
C-1l2
shape, composition, and surface differences, so such an application of the
curves will not be quantitatively precise.
Particle size distribution data available for this kind of analysis are
inadequate in some measure. The size ranges reported are usually too large
and require excessive averaging in the region of greatest variation in
efficiency on the grade efficiency curve--the fine particle size region.
Steelmaking dust is largely concentrated in this region. Whether or not
averaging according to the log-probability distribution would be applicable
to distributions having given data ranges such as 0 - 1 micron or 0 - 5
microns is not known.
The shape of the grade efficiency curve may be shaped somewhat by
process variables which alter the properties of the dust, by conditioning of
the dust by humidity and temperature control in the case of electrostatic
precipitation, by collector geometry (affecting treatment time) and energy
input. See Table IV. The development of a family of curves should be under-
taken, showing grade-efficiency variations at different levels of pertinent
operating variables (such as scrubbing energy level or electrostatic precipi-
tation treatment duration). Note that steelmaking fume can generally be
adequately removed with a scrubber pressure drop of 40+ inwg., or by electro-
static precipitators whose geometry, control, and energization are specifically
selected from that application. Using the given curves, however, to analyze
the effect of each type of treatment on actual process dusts will give a
comparison of dust property effects on performance of each collector--at least,
relative indications may be drawn. Efficiencies measured in the field, of
equipment collecting dust from the actual processes, dusts whose properties
would also be tested under the collecting conditions, would allow precise com-
parisons, more precise (and probably more economical) designing, and knowledge-
able predictions of performance. Such data is generally not available, not
very good (due to difficulties of measuring particle size and dust concentra-
tion with accuracy under conditions of collection, or difficulties of correl-
ating standard test results--which,too,are costly and limited--to the conditions
of temperature, humidification, agglomeration, dispersion, etc., under which
a dust would be collected), or undisclosed (being the proprietary tools of a
competitive collector industry). So Stairmand's contribution, while limited,
is available and useful for relative comparisons.
Stairmand's fabric filter curve (see Figure 7) is not reproduced here,
as it is based on a theoretical calcu~~tion for a new filter. A more typical
grade efficiency is given by Figure 82 , based on test results. However, both
indicate that the efficiency would n02 go to' zero as particle size approaches
zero. This is discussed by Stairmand in terms of a diffusional collecting
mechanism which comes into play at an increasing rate as particle size dimin-
ishes. The zero drop-off in the other grade efficiency curves represents the
failure of the inertial impaction mechanism to collect small particles.
Stairmand's discussion includes diffusion of particles to water droplet targets
as well as filter media, so this effect should also apply to wet scrubbing.
As indicated by the U-shape of curves in Figure 11, the mechanism of diffusion
seems also to apply to electrostatic precipitation. This mechanism, then,
would suggest an upward alteration of the grade-efficiency curves. (Variations
24
Whitby, K. T., and Lundgren, D. A., Technical Report Aug. 1961,
Fractional Efficiency Characteristics of a Torit Unit Type Cloth
Collector, Torit Mfg. Co., St. Paul, Minn.
-------
C-1l3
in this effect will occur with temperature and particle concentration.)
By designing low flowrate collectors, and optimizing inlet conditions, one
could take advantage of this mechanism with fine dusts.
The investigation of this diffusional mechanism (what it does in present
collectors, and its potential for new equipment development), the testing and
sampling techniques which contribute to the dirth of data, and filling the
data gaps should be undertaken to assist proper design of collectors and pre-
diction of costs.
The grade efficiency data shown is applied in Table V to various
process dusts with known particle size distributions. By multiplying each
of the size .range limits of these distributions by the square root of the
ratio
specific gravity of process dust
standard dust specific gravity
the curves can be used to obtain the average efficiency of collection for the
weight fraction of dust within those corrected size range limits. By mult-
iplying each such fractional efficiency by the corresponding weight fraction,
and summing, the net efficiency for the collector and dust combination is
obtained. But, note that it is only a relative indication of efficiency, and
is not to be taken absolutely and compared to a measured efficiency from the
field. We have found just such comparisons to show calculated precipitator
efficiencies low, and calculated wet scrubber efficiencies (using extrapolations
to Stairmand's 22 inwg. pressure drop situation) high for open hearth dust and
low for BOF dust (using the actual size distributions in each case for cal-
culating efficiencies).
Table III Grading of W.C.3 Test Dust19 and Sample Efficiency Calculation for
Self-Induced Spray Collector (lIP=6.l inwg.)
Size Range Median Size Wt. Fraction % Efficiency of Weighted
of Grade,~ in Grade,~ in Grade Median Size (Fig.23) Efficiency
104 - 150 127 .03 x 100 3.0
75 - 104 89.5 .07 x 100 7.0
60 - 75 67.5 .10 x 100 10.0
40 - 60 50.0 .15 x 100 15.0
30 - 40 35.0 .10 x 100 10.0
20 - 30 25.0 .10 x 99.5 9.95
15 - 20 17.5 .07 x 98.5 6.90
10 - 15 12.5 .08 x 97.5 7.76
7.5- 10 8.75 .04 x 96.5 3.86
5 - 7.5 6.25 .06 x 95.0 5.7
2.5- 5 3.75 .08 x 90.0 7.2
o - 2.5 1. 25 .12 x 63.0 7.56
1.00 93.93 Total
Efficiency
-------
C-114
Table IV
VARIOUS DEDUSTING SYSTEMS TREATING 60,000 CU.FT./MIN. OF DUSTY GASES AT 68°F
(The grade efficiency curves for each type of collector are shown
in figures 14 to 25. 19)
TYpe of
EquIpment
Efficiency
on
at.mdnrd
dUlt, %
(\)
Prenure
drop
in. W.8.
Medium-efficiency 6S.3 3.7
cyclone.
Hiirh-efficien~y 84.2 4.9
cyclones
Tuhular cyclones 93.8 4.3
-.-..-------
Irri~:\ted cyclone" 91.0 3.9
------
Low prC:BlIre.. 74.2 104
~;~rc.n~~tluIQr
Electrouatic pre- 94.1 0.6
cipitDlors
Irrii.ted elrc!ro- 99.0 0.6
Itatic precipit:l-
tors
Frnme-t)-pc fabric 99.9 4.0
niter
Revene-jet fabric 99.9 5.0
filter
- --
Spray tower 96.3 1.4
\Vet impinacment 97.9 6.1
Icrubber
Self-induced 93.S 6.1
Ipray. collectnr
Venturi scrubber 99.7 22.0
Di,intcjfrator 98.S
(1) See Table III
19 C. J.
Jour.
W.ter
Ula8e.
1101./
1,000
cu. fro
4.0
2.S
18.0
3.0
0.6
7.0
S.O
Stairmand, Design and Performance of Modern Gas-Cleaning Equipment,
Institute of Fuel (Brit.), Feb. 1956, p. 58
-------
o
...,...--
./
/
J
I
I
I
I
I
!IOO
..
i 10
t
='0
~
- 00
e
..20
::t
~
10 ..0 .0 10
...._TICIA ''''. WICIIIO...,.
.... I'
Medium dlidcncy, high.d'.rou.lhpul cyclone.
£lficimcy .II j micrOftJ .. 21'1.
~ - - -- - -
I '
I /
I.
I I
"
. ----j ~rl0l-
f--
I I I I
I I I I
~IOO
=
..
r 10
.
..
='0
~
-00
.
~
~ 10
.. ..
10 10 JO 40
'''''''CLI I:'II,,-..IC,.OMI.
Fic. It
ut" diamclU dry md iUi'Jled c)'dont.
Bd8ctcncy It , mtcrons .. 87%
C-1l5
'0
T
I
I
I
~IOO
~
r 10
.
..
~.o
~
-'0
2
~
.. 20
::t
g
100
20 40 60 10
M,TlCLI ,'U. ..,caONI.
.... IS
Hip efficiency (Sort, cone) cyclone.
EIficWw:Y .II , ,!,icrona .. 7) ''',
o
v
/
I I
I
I
... 100
~
r '0
~.
~ 60
U
~
- '0
.
~
.. 20
::t
g
so
10 20 )0 .0
P"RTICI.I s,n. W'CRONI.
f1a. .8
Low pre:uurc drop ctlhal..r cyclone.
Etlicicncy .I' , microns" ur.
t:.:: =~~!------~~~=-~~--I:-/~~~~]
tU --. -j-- -------------- ~~=---~1
'::.=q=~~#~=~=~====
„'- ::(iJ~;-_R~I~; l-6~ :;;::r~:.~-:--/:-~--
I l ~}'-----a-- i I
! ;~~= ~~~if!:~":_.-~~,,~~F--=-=:..-~
j 50F--- ----'------------'--' - ..-.----
! :I~~~ ~=~~~ ----~-T=-=--==-
~ ~:[~J==~-~--, ,---j-==---=..
O.O~ OJ 0.5 1.0 5.0
PARTiCLE SIL£ - MiCRONS
Fig.
8
Fit- It
Fabric filler (tee Fi,.. 7 Ind S)
Efficiency 11.5 miCTOfU" 99.9~,
o
~- '7
I
/
~.oo
i .0
..
~.O
~
-00
~
~ 10
::t
d
. '0
'''.''CLI
It 30
IIZI. "'C-'O"'.
'-.. 22
Sdl.~ \prav (1111(("10'.
!lfi..ltlK)' II .5 miL"",'" 9}~.
..100
~
~ '0
.
..
~.o
~
:40
.
~
ro
o
) . , . ., .
'A"TlCL' Iln, wrCAOIIII.
. Pit,. :u
Ventv,; KrUbber
Eflicimcy II j miLTOIUI - ~.6%
'00,
o
~
I
I
,.
~'~O
..
: '0
.
..
~ . 0
~
-'0
.
~
u 20
::t
~
10 10 JO
,...rICLI Iiti. MICA-OlliS.
'-.. ..
Small dilmttCt, Nbular cyclonCi.
ElfIdcncy II.' ",i.::roN - 891.
'0
.0
o
/' - - - -i-- - - -
-
,
,
,
I
- UUUGAU:O - -
-n- ... I
I I
I I I
e'OO
r 10
.
..
~ 10
~
-'0
~
u 20
::t
g
so
I 10 IS 10
---TICLI SUI!. UICR?t~S.
FII- It
ElCCV'OIUtic prccipitiuOI'. Irri.ued-cffi~i'.:nI;Y II S mi~rQns .. 9A~/.
Dr1-cfficienl;)' II ~ micror., ... 9; ~'.
,.
,.. 100
~
r 10
t
~IO
~
-'0
7
i
.. 20
~
o
I 10 II 10
....TICLI 1111. IoI'C.ONI.
.... 21
Spl'lytwCl'.
Efficiency II , microns ... 94Y.
II
I
f-- '--
I
...100
~
r 10
.
..
~ 10
~
~.o
.
~
roO
14'. .
'""jUICI.I SIll. NICROHS.
FII- Z2
Wet inpl"..:mcnt scrubber.
emc.iaK'y .u , microns'" 91:',
'0
~IOO
i .0
'0
7
-/
-,
I
I
T
.
.. .
= '0
~
-'0
.
~
.. 20
~
g
.~ . , . '7 .
flA.TlCL.1 Sltl. "'CAONS.
FII- IS
Dilinrccr1ror au ...aha.
Eftictenq" II .5 mi.:ron. ... 98 Yo
..
Grade Efficiency Curves for Various Co11ectors19
Typical of Performance Required in the mid-1950's.
-------
Table V
C-116
RELATIVE EFFICIENCY OF COLLECTING EQUIPMENT
FOR VARIOUS PROCESS DUSTS. AND COLLECTORS
'0
t:
(1j
1-0
~
tI)
1-0
Q)
~
t:
OM
tI)
Particle Specific Gravity
4.0
I 1 grains
n et Loading, SCFD
Required Efficiency, %
to 0.05 grains/SCFD
Predicted Efficiency, %, for:
Cyclones
High Throughput
High Efficiency
Multicyc10ne
65
91
98.5
97
Wet
Wet Scrubber
Low Energy
Spray
97
99.75
98.25
Wet Impingement
Self-Induced
High Energy
Disintegrator
99.32
99.98
Venturi
Electrostatic Precipitator
Dry
Wet
99.00
99.86
99.99
Fabric Fil ter
1-0
t: Q)
o I ~
"M ~ t:
~~"M
c.JQ)tI)
(1j tI) .
1-0 lID,....
~t:t:1ID
"M"M t:
~ ~"M
::ItI)::I,.\I\
~ (1j~ (1j
~~~13
2.7
4
98.73
59
90
98
96
97
99.52
98
98.95
99.95
95.5
98.95
99.99
t:
Q)
lID
;:...
8 Q)
c.J
CJ (1j
"M t:
tI) 1-0
(1j ::I
~~
5.0
4 - 8
98.9-
99.38
N.A.
"
"
"
N.A.
"
"
72
85*
64**
97.8
.c
~
1-0
(1j
Q)
::r:
t:
Q)
8'
5.2
2 - 7
97.5-
99.28
N.A.
"
"
"
N.A.
"
"
88
94.5*
83**
99.5
CJ
1-0
<
c.J
"M Q)
1-0 c.J
~ (1j
CJ t:
Q) 1-0
~ ::I
~~
3.93
3 - 6
98.33-
99.17
N.A.
"
"
"
N.A.
"
"
86
94*
81***
86**
99.5
P-
o
1-0,....
Q 1-0
Q)
Q) ~
1-0 (1j
::I ~
tI)
tI)
Q) t:
1-0 "M
p., ~
3.7
4.9
4.3
3.9
1.4
6.1
6.1
22*
.6
.6
4.0
A pressure drop of 40+ inwg. normally is required to clean steelmaking
fume to the specified 0.05 grains/DSCF with a venturi scrubber.
The codes in the mid-1950's when the grade efficiency data were typical
in Britain were less restrictive. Precipitator designs today can be
guaranteed to 99.5% efficiency if required. But the figures indicate
relative co11ectabi1ity.
*** With proper humidification,**
*
**
-------
5.
Effect of temperature---
C-117
From the gas laws, volumes vary directly with absolute
temperature. This has several implications for the gas cleaning system:
1000
900
800
~ 700
100
200
100
a)
The fan power requirement is proportional to volume.
So, many systems cool the gases to a minimum practical
temperature before entry into the fan unless thermal lift
must be maintained to get rid of noxious gases in the
effluent. Since positive pressure ordinarily makes for
simpler structure in a precipitator, and best efficiency
requires several hundred degrees of temperature, this
benefit is ordinarily not available for a forced draft
precipitator fan. However, in the wet scrubber, where
efficiency depends on a high gas stream energy, at a
high power consumption level, the cooling of gases is
particularly beneficial.
Precipitators and especially bag filters are limited in
the temperature at which they operate. Thus cooling as
a pre-conditioning step is required before entry to the
gas cleaner on many processes. The method of cooling
also effects the volume of gas to be handled.
b)
7i
~/f7
~' .C
~,~
,-
~84
O~
~Q.
\.?a:
~~
I.) U,J
~~
a:
r~
:3 - 2
9';;2
o
o
Required
Weter Ipmy,nq (b)
I
450°F
bOO"F
-=
I~ !
,
'Indlre,t cooilnq (,j
(wolte neat boder)
J
1300 1400 1500 IKD i700 I&..!O frO LCZY)
INLET GAS TEMPER.ATURE. OF
Fig. 1025 Waste gas volume variation for different gas condiriotled tell/-
peratures by air infiltration and water spraying
W.P.C. Ungoed & W.F. Needham, Waste Heat Boilers on Open Hearth
Furnaces. . Fume Arrestment. Special Report 83, British Iron &
Steel Institute~ Ibid.
21
E.R. Watkins and
Precipitation to
Fume Arrestment,
K. Darby, The Application of Electrostatic
the Control of Fume in the Steel Industry.
ibid.
-------
-- -------------------.-
C-118
Effect of Humidity.
Water additions to the gas stream or gas cleaner system is a frequently
required pre-conditioning step. The humidity of the gas entering a baghouse
must be sufficiently below the dew point to preclude corrosion of the structure
and clogging of the bags with moist cake. On the other hand, quenching is
an economical way to cool, avoiding expensive radiation ducting or excessively
large components to handle dilution air.
Humidity is sometimes critical in the electrostatic precipitation of
certain dusts of high resistivity. Again, it can be coupled with cooling
quench in the pre-conditioning zone, but care must be taken to stay above
the dew point temperature. No liquid effluent results from these cases.
It should be noted that in both of the above systems, the acid dew point
is also critical from a corrosion point of view in those cases where sulfur
dioxide is a significant process effluent component, such as the open hearth
and some coal burning processes. While it has been noted that sulfur dioxide
can be beneficial in the precipitation of some process dusts, this benefit
is likely to be lost as sulfur dioxide regulations take effect.
Water flushing of elbows, fan blades, and other parts of the systems
has been effectively used to inhibit impingement abrasion and to prevent
dust buildup.
Corrosion, abrasion, dust buildup and excessive temperature are the
most frequent maintenance problems on a gas cleaning system. Where water
is used as a remedy, careful pH control is important, as is solids build
up in a recirculating water system.
Except for this preventive maintenance application, water effluents
usually are the result of wet scrubbing or wetted surface precipitation.
(This later finds application to electric arc furnace fume cleaning where
satisfactory particle resistivity for collection is difficult to achieve, and
in sinter plant and blast furnace applications where coarse, abrasive dust
must be removed).
EFFECT OF ELECTRICAL RESISTIVITY OF DUST FUME21
Much has been written on the subject of the effect of the
resistivity of the dust on precipitator efficiency and it
is not proposed to go deeply into this aspect of the subject...
It can be shown, however, that for dust of a very high resistivity,
precipitation efficiency can be seriously reduced (see Fig. 11
and section on particle sizing), and, for resistivities higher
than lOllohm-cm, difficulties are likely to be encountered. There
is some divergence of opinion between different investigators on the
value of resistivity at which difficulty is likely to be en-
countered and it is thought that this is due to a number of
factors difficult to control, such as the degree of packing of the
dust, so that in practice different forms of apparatus can disagree
to a considerable extent. At the same time resistivities
measured by anyone form of apparatus, when used by an experi-
enced operator, can be relat~d to precipitator performance.
-------
C-1l9
10"
9 .
.,e
~
\0/
'"
II;
2
o
>=
~
0::
Q
ir
80
E
'Ii
I A'c t"lfloce
I LD O:I)"~~'''' .
~ Oc.f't'. ",.,,0;" , ,,'''OCt,
"u"'o1oo
Watkins, Electrostatic Precipitation
4 Open .heorth furnace
~ Oeut.CO"'ZOf Ion
lodl~ procest
6 Open -heart'" furnoce
100
~-
I HoqIIly '0""'''' ~.., (id]""",.c")
qlY1nq senous reverse lONZotlOl'\
.
~
~
~
:;; .
~ 10
a:
2 ~I' 01 I after use 0/ canddooninq
oqe"' 10 ,..;..co 'olill.."y
Dry Iype prec1poi01Or
J Normal well candltloned dUll
~
~
4 Hlqhly ,olill... dull u'''''I ..et
type lftCipilOIot
70 - 2
Fig. 11
,
4
,
b
t ~. ----L,,---.J
8 ~ ~ ~ ~ ~ w n ~
PARTICLE SIZE . I'm
10'
100
~oo
600
Variarion of precipiration efficiency wirh parricle size
Fig. 12
E/ecr"",/ ruistiviry of red oxide fume from
o"Ygtn-b/own srll/malli", proctsses
various
The electrical resistivity of most dusts and fume depends
on the nature and condition of the surface of the dust part-
icles, rather than on the material of which the dust is
composed; the resistivity is in practice often determined by
adsorbed layers of vapor, such as water, sulphuric acid, or
ammonia. These usually arise from reactions taking place in
the furnace or vessel to which the precipitator is attached;
for instance, high-sulphur fuel oil used in firing OR furnaces
can produce sulphuric acid, and this in turn is adsorbed by
the dust. Where the dust resistivity is high, suitable layers
to reduce the resistivity of the dust can be provided by the
injection of one of the conditioning agents listed above into the
flue before the precipitators. In practice, however, in this
country, it has not so far been found necessary to supply any
artificial conditioning agent to red oxide dust ...,' although
difficulties have been reported from abroad. Figure 12 shows the
resistivity plotted against temperature for fume originating from
LD converters,OR furnaces, arc furnaces, and ladle desiliconization
processes.
-------
C-120
Watkins, Electrostatic Precipitation
It will be seen that the resistivity is below 10" ohm-cm
in all cases except for fume from the arc furnace. In the
case of OR furnaces the dust is normally 'conditioned' by
the water vapor and sulphur trioxide resulting from the
combustion of the fuel used to fire the furnace.
In the case of the arc furnace, there is normally no
such supply of conditioning agent in the gases leaving
the furnace as curve 1, which is the resistivity of a
dust sample taken immediately at the furnace outlet and
is typical of a highly resistive dust without the con-
ditioning surface layer. When a precipitator is attached
to an arc furnace, it is necessary, in view of the high
temperatures involved, to cool the gases, usually by means
of a water spray tower, to an economical level for the pre-
cipitator. This has the effect of reducing the gas volume
to be treated; and at the same time, the dust is 'conditioned'
by water vapor and the resistivity curve assumes the shape
shown for the other fume with the peak value below the limit
for efficient precipitation.
EFFECT OF PARTICLE SIZE ON PRECIPITATION EFFICIENCy21
It can be shown from calculations on the forces acting on
charged particles that the efficiency of an electrostatic pre-
cipitator should decrease with decreasing particle size; this,
however, is not normally borne out in practice and many
commercial applications of precipitation are on processes
in which much of the fume is submicron, as for instance, blast~
furnace gas cleaning and the red oxide fume evolved from oxygen
blowing processes.
Figure 11 shows the relationship with precipitation efficiency
and particle size for a number of dusts, the first three re-
lating to' dry precipitation, and number 4 to a wet precipitator.
Curve No.1 was obtained on a dust whose electrical resistivity
was of the order ofl0130hm-cm, and the precipitator was ex-
hibiting signs of severe reverse ionization; it is interesting
to note that the fall-off in efficiency becomes increasingly
serious for particle sizings below 20 ~m. Curve No.2 was ob-
tained upon the same dust when the resistivity had been de-
creased by the use of a conditioning agent, while curve No.3 was
obtained on a dust whose resistivity was well below the limit
of 1011ohm-cm quoted in the section on resistivity. Curve No.4,
obtained on the wet electrofilter, has a fall-off comparable with
the lowest resistivity dry dust (curve No.3); in this case, although
the dust was initially highly resistive, the resistivity of the
dust in the precipitator was reduced to a safe level by the cooling and
natural conditioning effect of the spray tower preceding the pre-
cipitator; in addition, since the particles were deposited on a
moving flow of water, the effect of high resistivity would be of no
consequence in any case.
-------
C-121
Watkins, Electrostatic Precipitation
Since this fume consisted of non-magnetic particles of
spherical form it was possible, using the electron micro-
scope, to continue the grading beyond the limit of most
forms of grading apparatus; these gradings indicated that
there was no serious fall-off of precipitator efficiency
for particle sizings down to the order of .Ol~m.
While the theory of the motion of the dust particles
under the effect of the electric field assumes that the
dust is deposited as individual particles, there is in
practice a strong tendency for very fine fume to agglomerate
into masses consisting of hundreds of fine particles, such
agglomerates behaving as single, much larger particles in
the electric field, with the result that efficiency is higher
than would be theoretically calculated for such a fume. This
is normally considered to be one of the explanations why the
precipitator fails to obey the basic theory. It is also one
of the difficulties of carrying out dust gradings and limits
their value, since clearly what is required of a dust grading
apparatus is the grading including the effect of agglomeration,
and one of the debatable points in dust grading methods is the
energy which should be used to disperse the agglomerates formed
in the precipitators in a dust grading apparatus. An interesting
feature is the action of the conditioning agent, as illustrated
by curves I and 2, since it would appear that, in addition to
reducing the resistivity of the dust, the agglomerating pro-
perties are also materially improved. The authors consider that
for efficient precipitation it is necessary, particularly in a
dry precipitator, for the dust to have the correct agglomerating
properties in addition to a suitable electrical resistivity value.
DUST CONCENTRATION AND ELECTRODE DESIGN21
In recent years a considerable amount has been written on the
subject of high-corona electrodes and their importance in the
precipitation of red oxide fume. It is the experience of the
authors of this paper that high-emission electrodes can only use-
fully be applied where suppression of the corona discharge current
takes place with the normal type electrode. Its effect in this
instance is to bring the current discharge back to what has been
found to be a reasonably normal value for efficient precipitation.
Corona suppression is caused when the dust concentration is very
high and the particle sizing extremely small, the effect being to
blanket the corona electrode and to inhibit corona discharge in
much the same way that the control grid around the cathode of an
electronic valve can be used to limit the current flowing across
the valve from anode to cathode. High-corona current electrodes
should be confined to the stages of a precipitator where corona
suppression is likely to take place add dust concentrations are
heavy.
-------
C-l22
6.
Adaptability of particulate removal systems to removal of
gaseous pollutants.
Studies on the injection of dry, powdered limestone, dolomite,
manganese dioxide, alumina and other metal oxides to process gases
containing sulfur dioxide indicate that some 30 - 60% of the S02
can be absorbed by the additive and removed in the particulate re-
moval system, as an added inlet loading.
A bag filter could do
gives the added benefit of a
good test results. Alkaline
injection.
this effectively. A wet scrubber system
liquid absorbtion stage and has yielded
solutions may be used without the powder
A catalytic oxidation process unit could be inserted in series
following a precipitator so the high temperature at which the pre-
cipitator operates could be used in the oxidation. The process scrubbing
would preclude following with a baghouse or precipitator.
These systems are in the development phase and their use is
contingent on economics and competitive process developments.
------- |