GCA-TR-83-105G
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Ambient Standards Branch
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
Contract No. 68-02-3804
Work Assignment No. 15
EPA Project Officer
John Haines
COST ASSESSMENT OF REGULATORY
ALTERNATIVES FOR LEAD NATIONAL
AMBIENT AIR QUALITY STANDARDS
Revised Draft Report
by
William Battye
Rebecca Battye
Nancy Browne
Michael Clowers
Jeanne Eichinger
George Viconovic
July 1985
GCA CORPORATION
GCA/TECHNOLOGY DIVISION
Chapel Hill, North Carolina 27514

-------
DISCLAIMER
This Draft Final Report was furnished to the Environmental Protection
Agency by GCA Corporation, GCA/Technology Division, 213 Burlington Road,
Bedford, Massachusetts 01730, in fulfillment of Contract No. 68-02-3804,
Task No. 15. The opinions, findings, and conclusions expressed are those
of the authors and not necessarily those of the Environmental Protection
Agency. Mention of company or product name is not to be considered as an
endorsement by the Environmental Protection Agency. J
i
ii :

-------
ACKNOWLEDGEMENTS
This report reflects the effort of many investigators over an
extended period during which the data bases, methodological approaches,
analytical assumptions and the standards to be evaluated have undergone
considerable change. The authors wish to thank the following investiga-
tors for their participation in earlier phases of the study: Timothy Curtin,
James Lent, and Buddy Newman. We also wish to acknowledge the guidance and
contributions of the U.S. Environmental Protection Agency (EPA) project
officers on this study, John Haines and Robert Kellam.
iii

-------
TABLE OF CONTENTS
Acknowledgements		lit
Figures		vi
Tables		vii
Executive Summary		x
1.	Introduction		1
1.1. Purpose		1
1.2 Study Methodology			2
2.	Lead Emissions and Ambient Concentrations			6
2.1	Lead Emissions		6
2.2	Typical Ambient Lead Concentrations Near Emission Sources.. 8
2.3	Lead Concentration Distribution and Recent Trends		11
2.4	Effects of Averaging Time and Average Lead Concentration..	15
3.	Mobile Source Ambient Impacts and Costs		26
3.1	Current Ambient Impacts from Mobile Sources		27
3.2	Projected Ambient Lead Impacts from Mobile Sources		31
3.3	Interaction of Mobile Source and Stationary Source
Ambient Lead Impact		39
3.4	Lead Phasedown Costs		41
4.	Generalized Study Methodologies		45
4.1	Identification of Affected Industry Categories		46
4.2	Selection of Baseline Control Strategies		48
4.3	Estimation of Baseline Ambient Impacts		49
4.4	Source Interactions		52
4.5	Identification of Individual Affected Sources		54
4.6	Control Strategy Identification and Costing		54
4.7	Estimation of Lowest Cost of Attainment for Existing
Sources		57
4.8	New Source Costs		58
iv

-------
Table of Contents (continued)
5.	Industry-Specific Study Methods and Study Results		62
5.1	Primary Lead Smelting and Refining		62
5.2	Secondary Lead Smelters		68
5.3	Lead Ore Processing		75
5.4	Lead-Acid Battery Manufacture		80
5.5	Lead Alkyl Manufacture				88
5.6	Gray Iron Foundries		91
5.7	Alloy Steel Electric Arc Furnaces		97
5.8	Steel Electric Arc Furnaces		102
5.9	Iron and Bronze Industry - Sintering		108
5.10	Brass and Bronze Ingot Production		109
5.11	Primary Copper Smelting and Refining		117
5.12	Total Nationwide Control Costs		130
5.13	Uncertainty Analysis		133
6.	New Source Controls		140
7.	Costs of Waste Oil Combustion Regulations		142
Appendices
A.	Lead Emissions Inventory				146
B.	Raw Data Used in Predicting Mobile Source Impacts		153
C.	Sample Cost Calculation		158
D.	Costs of Industry-Specific Control Options and Techniques....	167
E.	Identification of Battery Plants Affected Under Alternative
NAAQS Levels		189
v
l-
/

-------
LIST OF FIGURES
Number	Page
2-1 Predicted maximum quarterly average ambient lead
concentrations around a large secondary lead smelter		12
2-2 Average 24-hour concentration of lead at various
elevations and set back distances		13
2-3 Trends in mean maximum quarterly lead concentration of
various monitor groups		16
2-4 Trends in overall annual average lead concentration for
various monitor groups		16
2-5 Frequency distribution of monthly-to-quarterly lead
concentration ratio for all lead monitors and for point
source monitors		20
2-6 Frequency distriubtion of monthly-to-quarterly lead
concentration ratio for microscale and neighborhood and
middle scale roadside monitors		21
2-7	Monthly average lead concentration plotted against
quarterly average		22
3-1	Quarterly and annual average ambient lead concentrations
at NAMS 03060013G01		32
3-2 Relative total lead burned in gasoline (1974 = 1.0).........	32
3-3 Maximum quarterly lead concentration at NAMS sites as
a function of totoal lead burned in gasoline		36
5-1 Conventional copper smelting process	 120
5-2 Sensitivity of NAAQS costs to Standard Level	 135
vi
i

-------
LIST OF TABLES
Number	Page
1	Nationwide NAAQS Control Costs For The Upper Limit Mobile
Source Background		xi
2	Nationwide NAAQS Control Costs For The Lower Limit Mobile
Source Background	 xii
2-1 Estimated Atmospheric Lead Emissions For The United States,
1981		7
2-2 Typical Maximum Quarterly Lead Concentration Near Lead
Emissions Sources		9
2-3 Frequency Distributions of Maximum Quarterly Lead
Concentrations By Site-Type		14
2-4 Properties of The Distribution of Maximum Quarterly Lead
Concentration For Various Site Types		17
2-5 Lead Concentration Trends For 1980 to 1983		18
2-6 Monthly-To-Quarterly Average Concentration Ratios For
Various Lead Monitors Categories		23
2-7 Monthly Concentration Limits Equivalent To The Quarterly
NAAQS Levels, In Terms of Allowed Average Concentration		23
2-8 Daily-To-Quarterly And Daily-To-Monthly Concentration For
Different Monitor Types		25
2-9	Monthly Concentration Limits Equivalent To The Quarterly
NAAQS Levels, In Terms of Expected Daily Concentration		25
3-1	Site Parameters And Ambient Lead Data For Selected
Miscroscale NAMS		28
3-2 Site Parameters and Ambient Lead Data For Selected
Middlescale NAMS			29
3-3 Site Parameters and Ambient Lead Data for Selected
Neighborhood Scale NAMS	.		30
3-4 Lead Burned In Gasoline		35
vii

-------
List of Tables (continued)
3-5	Projected Lead In Gasoline and Ambient Lead Impacts From
Mobile Sources		40
4-1	Current Capacity Utilization and Growth Rates For
Industries Affected by the NAAQS Alternatives		59
5-1	Nationwide NAAQS Costs For Primary Lead Smelters and
Refineries		69
5-2 Controlled Emission Factors and Source-Receptor
Coefficients For Secondary Lead Smelting		73
5-3 Nationwide NAAQS Costs For Secondary Lead Smelters
(October 1984 Dollars)		76
5-4 Baseline Emissions and Ambient Impacts From Lead Ore
Processing		79
5-5 NAAQS Controls Costs For The Lead Ore Processing Industry
(October 1984 Dollars)		81
5-6 Estimated Breakdown of Battery Plants By Size Range and
Control Status		84
5-7 Emission Factors For Lead-Acid Battery Manufacture		85
5-8 Nationwide NAAQS Costs For The Lead Acid Battery
Manufacturing (October 1984 Dollars)		89
5-9 Nationwide NAAQS Controls Costs For The Lead-Alkyl
Production (October 1984 Dollars)		92
5-10 Baseline Emissions and Ambient Impacts For Gray Iron
Foundry Cupola Furnaces		96
5-11 NAAQS Controls Costs For The Gray Iron Foundry Industry
(October 1984 Dollars)		98
5-12 Baseline Emissions and Ambient Impacts For Alloy Steel
Electric Arc Furnaces	 101
5-13 Nationwide NAAQS Costs For Alloy Steel EAFs (October 1984
Dollars)	 103
viii

-------
List of Tables (continued)
5-14 Baseline Emissions and Ambient Impacts For Steel Electric
Arc Furnaces		106
5-15 Nationwide NAAQS Costs For Carbon Steel EAFs	 107
5-16 Baseline Emissions and Ambient Impacts of Iron and Steel
Sintering	 110
5-17 NAAQS Control Costs For Iron Sintering	 Ill
5-18 Baseline Emissions and Ambient Impacts for Brass and
Bronze Production	 114
5-19 NAAQS Control Costs For Iron Sintering	 116
5-20 Domestic Primary Copper Smelters			.. 121
5-21 Fugitive Emission Control Techniques At Primary Copper	
Smelters	 126
5-22 Baseline Emissions and Predicted Ambient Lead Concentrations
After Control		128
5-23 NAAQS Control Costs For The Primary Copper Smelting
Industry	 129
5-24 Nationwide NAAQS Control Costs For The Upper Limit Mobile
Source Background	 131
5-25	Nationwide NAAQS Control Costs For The Lower Limit Mobile
Source Background	 132
6-1	NSPS Costs For The Lead-Acid Battery Industry (October
1984 Dollars)		 141
6-2	NSPS Costs For Alloy Steel Electric Arc Furnace (October
1984 Dollars	 141
7-1	Nationwide Costs Of Hazardous Waste and Used Oil
Combustion Regulations (1984 Dollars)		144
I
ix !

-------
EXECUTIVE SUMMARY
This report presents a cost analysis of the current national ambient air
quality standard (NAAQS) for lead, and a range of lead standard alternatives.
The current lead standard is 1.5 micrograms per cubic meter (ug/m ) averaged
over a calendar quarter. Other standards considered in this analysis were
0.25 ug/m^, 0.5 ug/m"^, 0.8 ug/m^, 1.0 ug/m^, 2.0 ug/m^, and 2.5 ug/m^. The
nationwide costs of these standards by 1990 or 1995 were calculated.
Stationary sources that would not attain the various alternative standards
were identified using atmospheric dispersion modeling and ambient lead
monitoring data. In projecting the attainment status of stationary sources in
urban areas with substantial mobile source emissions, a background mobile
source ambient impact was added to the predicted ambient impact for the
stationary source. Costs of controls needed to attain each alternative were
then calculated for the affected plants and summed to obtain nationwide costs.
Tables 1 and 2 present estimated nationwide capital and annual costs of
attaining the various alternatives for stationary sources under two possible
scenarios for mobile source lead emissions. At present, EPA lead-in-gasoline
phasedown regulations set a standard of 0.10 grams of lead per gallon of leaded
gasoline, effective January 1, 1986. EPA has also solicited public comments on
a possible total ban of lead in gasoline. The final form of lead-in-gasiline
phasedown regulations will determine the average mobile source background lead
concentration around sources in urban areas. Table 1 gives stationary source
costs calculated based on the current lead-in-gasoline phasedown program, while
Table 2 gives stationary source costs under the proposed total ban of lead in
gasoline.
Either lead-in-gasoline phasedown scenario would result in substantial
reductions of mobile source lead emissions by 1988, to the point where these
emissions will neither cause nor contribute to exceedences of any of the lead
NAAQS alternatives in 1990 or 1995. The current phasedown rule is expected to
result in a 620 million dollar per year (1984 dollars) increase in refinery
operating costs in 1986. By 1990, these costs will have dropped to 480 million
x

-------
TABLE 1. NATIONWIDE NAAQS CONTROL COSTS FOR THE UPPER LIMIT MOBILE "SOURCE BACKGROUND
Costs for noted NAAQS alternative (millions)

2.5
2.0
1.5
1.0
0.8
0.5
0. 25
U'lTAI. COSTS







Primary lead smelting
17.8
27.3
44.5
46.6
53.3
53.3
53.3
Secondary lead smelting
54.2
72.4
96.1
118.7
124.4
133.3
144.0
Lead ore processing
--
--
18.3
18.3
18.3
18.3
18.3
Lead-acid battery manufacture
1.6
2.1
2.1
3.5
4.1
5.2
6.5
Lead alkyl production
--

--
1.0
1.0
2.1
2. 1
Cray iron foundries
2.2
2.2
6.6
11.0
11.0
16.2
21.3
Alloy steel EAFs
—
--
—
--
—
6.8
8. 1
Steel EAFs
—
--
—
--
0.7
1.5
2.5
Tron and steel sintering
6.7
6.7
13.4
13.4
13.4
13.4
23.6
Urass and bronze production
0.4
0.7
0.7
1.4
2.3
3.6
5.6
Primary copper smelting
26.0
28.9
44.2
49.7
50.6
61.6
75.2
TOTAL
108.9
140.3
225.9
263.7
279.1
315.3
360. 3
ANNUAL COSTS
Primary lead smelting
4.5
6.7
11.9
12.6
14.5
14.5
14.5
Secondary lead smelting
16.4
22.0
28.9
35.5
36.8
38.5
40.8
Lead ore processing
--
--
7.3
7.3
7.3
7.3
7.3
Lead-acid battery manufacture
0.4
0.6
0.6
1.0
1.1
1.4
1.9
Lead alkyl production
--
—
—
0.4
0.4
0. 7
0. 7
Cray iron foundries
0.5
0.5
1.5
2.5
2.5
3.7
4.8
Alloy steel EAFa
—
--
—
—
—
1.8
2.5
Steel EAFs

--
>-
--
0.3
0.7
1.2
Iron and steel sintering
1.8
1.8
3.7
3.7
3.7
3.7
6.6
Urass and bronze production
0.1
0.2
0.2
0.4
0.6
1.0
1.6
Primary copper smelting
6.8
7.3
11.6
13. 3
13.5
16.5
20.0
TOTAL
30.6
39.1
65.7
76.6
•
O
00
89.8
101.8

-------
TABLE 2. NATIONWIDE NAAQS CONTROL COSTS FOR THE LOWER LIMIT MOBILE SOURCE BACKGROUND

Costs for noted NAAQS alternative (millions)

2.5
2.0
1.5
1.0
0.8
0.5
0.25
CAPITAL COSTS







Primary lead smelting
17.8
27.3
44.5
46.6
53.3
53.3
53.3
Secondary lead smelting
54.2
70.3
95.7
118.6
123.4
132.7
140.3
Lead ore processing

--
18.3
18.3
18.3
18.3
18.3
Lead-acid battery manufacture
1.6
2.1
2.1
3.5
3.8
4.8
6.3
Lead alkyl production

—
—
1.0
1.0
2.1
2. 1
Cray iron foundries
2.2
2.2
4.4
8.8
11.0
16.2
20.9
Alloy steel EAFs
—
—
--
--
—
--
8. 1
Steel EAFs

—
—
—
0.7
1.5
1.9
Iron and steel sintering
6.7
6.7
13.4
13.4
13.4
13.4
23.6
Brass and bronze production
0.4
0.4
0.7
1.4
1.9
3.6
5.6
Primary copper smelting
26.0
28.9
44.2
49.7
50.6
61.6
75.2
TOTAL
108.9
137.9
223.3
261.4
277.4
307.4
355.5
ANNUAL COSTS







Primary lead smelting
4.5
6.7
11.9
12. 6
14.5
14.5
14.5
Secondary lead 6meltlng
16.4
21.2
28. 7
35.4
36.5
38.3
40. 1
Lead ore processing

--
7.3
7.3
7.3
7.3
7.3
Lead-acid battery manufacture
0.4
0.6
0.6
1.0
1.1
1.3
1.8
Lead alkyl production
--
—
—
0.4
0.4
0. 7
0. 7
Cray iron foundries
0.5
0.5
1.0
2.0
2.5
3.7
4.7
Alloy steel EAFs
--
--
—
--
--
--
2.5
Steel EAFs
—
—
—
—
0. 3
0.7
1.0
Iron and steel sintering
1.8
1.8
3.7
3. 7
3. 7
3.7
6.6
Brass and bronze production
0.1
0.1
0.2
0.4
0.5
0.9
1.6
Primary copper smelting
6.8
7.3
11.6
13.3
13.5
16.5
20.0
TOTAL
23.8
30.9
53.4
62.7
66. 7
71.0
80. 7

-------
dollars per year, and additional costs for 1995 are estimated at 440 million
dollars per year. The lead-in-gasoline phasedown regulations, and thus these
costs, are independent of which NAAQS alternative is selected. Because of the
low projected mobile source ambient impacts after 1986, mobile source lead
emission controls over and above the lead phasedown requirements were not
considered in this study.
Another EPA activity which would affect ambient lead levels is a proposed
extension of hazardous waste management regulations to apply to marketing of
and burning for energy recovery of hazardous wastes and used oils. The
proposed rule would establish maximum levels of lead and other contaminants for
used oil burned in nonindustrial boilers. The maximum contaminant levels could
be achieved by dilution of the used oil with virgin oil prior to burning. The
proposed limit for lead is 100 parts per million (ppm), however a 10 ppm limit
is also being considered. Total capital costs of the proposed marketing rules
are estimated at 6.1 million dollars, and annual costs are estimated at 21
million dollars per year. These costs are also independent of which lead NAAQS
is selected.
All of the costs presented in Tables 1 and 2 represent costs of controls
in addition to those already in place. Especially in the case of primary and
secondary lead smelters, substantial costs already have been incurred in the
implementation of the current standard. Also, controls installed to meet OSHA
lead standards and EPA particulate emission limits generally act to reduce lead
ambient concentrations from lead-emitting industries. Due to the implementa-
tion process for emission regulations, there may be an overlap in the installa-
tion of control equipment that is considered to be "in place" and that required
to attain the various lead NAAQS alternatives.
The costs presented in Tables 1 and 2 and throughout this analysis are
incremental costs for attainment of the various NAAQS alternatives by 1990.
The incremental costs for the two alternatives in which the lead NAAQS is
relaxed do not include any cost savings for discontinuing the use of controls
that were installed to meet the current standard; thus costs for these
alternatives are biased upward. Conversely, costs for the lower alternatives
are biased downward, because controls could not be identified that would allow
some plants to attain these alternatives. Some of the plants for thich costs
were developed are currently uncontrolled. Controls for these plants may be

-------
required to meet the particulate matter NAAQS. Thus, some of the costs for all
of the NAAQS alternatives may be partially attributable to the particulate
NAAQS.
The attainment costs for 1995 are expected to be the sane as those for
1990 (in constant dollars) for two reasons. First, new stationary source costs
associated with attainment of the lead NAAQS after 1990 will be attributable to
new source performance standards. Second, because the background lead
concentration around stationary lead sources by 1990 is expected to be
negligible in most cases, the attainment status of a given plant will not be
affected by changes external to the plant. As noted above, the lead-in-
gasoline phasedown program will reduce mobile source contributions around
stationary sources to negligible levels. Also, because of the relatively small
number of stationary lead sources, most plants are not subject to background
concentrations from neighboring plants.
The results of this analysis are subfect to a number of limitations, due
to the uncertainty in the input data, and due to simplifying assumptions made
as part of the general methodology used. The majority of lead emissions from
major industries producing and using lead are fugitive emissions, because other
emissions are largely controlled. For other industries, lead emissions are due
to the presence of lead as a contaminant in feed materials, often at trace
levels. Thus, lead concentrations in emissions depend on the quality of feed
materials used, and are not easily quantified. In addition, dispersion
modeling and the costing of lead emission control systems are subject to
uncertianties and estimation limitations.
xiv

-------
1. INTRODUCTION
1.1 PURPOSE
Section 109(d)(1) of the Clean Air Act, as amended, states that the EPA
Administrator shall "complete a thorough review of the criteria published
under section 108 and the national ambient air quality standards promulgated
under this section and shall make such revisions in such criteria and
standards and promulgate such new standards as may be appropriate..." In
accordance with this requirement, the National Ambient Air Quality Standard
(NAAQS) for the lead is being reviewed. As part of this review procedure,
this report analyzes alternative standard levels and potential control costs
for meeting these levels. The report gives costs for the lowest net
annualized cost methods of achieving alternative NAAQS levels. The control
costs presented in these report will be used in an economic analysis as part
of the Regulatory Impact Analysis (RIA). In addition, ambient impacts
predicted in this study will be input to an exposure assessment for the RIA.
The current national primary and secondary ambient air quality standards
for lead are 1.5 micrograms/cubic meter (ug/m3) averaged over a calendar
quarter (40 CFR 50.12). Lead control costs required to meet this standard are
analyzed in this report, along with costs for the following alternative
standards: 0.25, 0.5, 0.8, 1.0, 2.0, and 0.5 ug/m^ quarterly averages. In
addition, the relationship between average lead concentration and averaging
time was studied to idenify monthly and daily lead concentrations that are
equivalent to various quarterly averages, in terms of the degree of protection
offered.
1
T

-------
1.2 STUDY METHODOLOGY
The first step in estimating nationwide control costs to meet an
alternative standard level is to project how many areas will not attain that
standard with baseline emission controls. For this analysis, attainment
status is projected for 1990 and 1995. Several causes of exceedances of the
alternative standards were studied: lead emissions from individual plants,
interaction of emissions from two or more plants, lead emissions from mobile
sources, and interation of mobile source and point source emissions. For
existing sources, baseline emissions controls are taken generally as those
controls currently in place. For new sources, the baseline is the emission
controls required by new source performance standards (NSPS). Ambient
concentration of lead around point sources were estimated using atmospheric
dispersion modeling of projected emissions. Where available, data on ambient
lead concentrations in the vicinity of point sources also were used.
For mobile sources, ambient monitoring data for lead near roadways were
used to predict 1990 and 1995 mobile source lead impacts by comparing time
variations of roadside lead concentrations with the time variation in the
nationwide average amount of lead used in gasoline. Mobile source ambient
impacts have been reduced dramatically since the institution of EPA
regulations requiring the phasing down of lead use in gasoline. The impacts
are expected to be reduced even further by 1990. Regresions developed based
on past data were then used with projections of future lead in gasoline to
estimate future mobile source lead impacts.
These projections indicate that mobile source ambient impacts will be
minor in comparison with stationary source impacts. Therefore, mobile source
controls over and above the lead phasedown program were not considered in
developing costs for the lead NAAQS alternatives. In general, most point
sources that must be controlled to meet an alternative NAAQS would require
control even in the absence of mobile source emissions.
For stationary sources, a least-cost algorithm was used to estimate the
costs of controls to meet each alternative lead standard. The total cost to
meet each alternative standard was then estimated by summing costs to
individual sources. In the least-cost algorithm, each plant is assumed to
install control equipment necessary to attain a given lead standard in a
manner that minimizes total net annualized costs for the plant.

-------
An alternate assumption might be that all lead sources around which the
standard is exceeded would install reasonably available control technologies
(RACT). There is a significant difference in estimated costs predicted using
these two assumptions. A least-cost algorithm was chosen for use here because
(1) it models more realistically the behavior of firms and control agencies in
achieving a "negotiated" control strategy for each firm (as part of the State
Implementation Plan [SIP] process), and (2) the bubble policy as well as
recently proposed prevention of significant air quality deterioration (PSD)
and new source review rules define a pollutant "source" as an entire plant.
The significance of these rules is that each plant is considered as an entity
and can apply controls to the sources it chooses as long as the plant emission
limit is achieved. Thus, RACT probably will not be applied on a
source-by-source basis. Given that a plant has some discretion on which
sources to control, it seems reasonable that their decisions will be made to
minimize costs.'- Due to the need to compare cost estimates provided in this
report with the benefits of an ambient standard, a least cost method is the
preferred approach because at allows the estimation of incremental costs.
Incremental costs must be compared with incremental benefits for the various
NAAQS alternatives.^
In some industries affected by the lead NAAQS, plant closures may occur
due to economic conditions. For instance, in the case of alkyl lead
additives, a decrease in domestic demand of about 70 percent is expected by
1990, causing some plant closures. For gray iron foundries, numerous plant
closures have occurred in the recent past, and this trend is expected to
continue. The cost effects of plant closure will be addressed in the Economic
Impact Analysis for the RIA.	1
In this cost assessment, several source categories and individual sources
were identified as needing additional emission controls at the level of the
current lead NAAQS (1.5 ug/m^). Some of these categories and sources may not
in fact, require control under State Implementation Plans (SIPs) for meeting
the current standard. Conversly, some source categories may require control
under the current SIPs that are not identified in this study as needing
additional control. Also, this report presents projected air lead
concentrations and emissions, identifies levels of control needed, and
3
t

-------
predicts that certain sources may not be able to comply with the current
standard. These scenarios may be quite different from information presented
in the current lead SIPs. While this situation may on the surface appear to
be an inconsistency, it is in fact a natural consequence of the two processes
employed -- cost assessment and SIP development. This report necessarily
relies on a more "broad-brush" national approach in which time and resource
limitations preclude the kind of detailed area-by-area analyses required for
actual SIPs. Furthermore, while the authors consulted the lead SIP files
during the preparation of this report, several lead SIPs have not yet been
completed as of this writing.
.1
4

-------
REFERENCES FOR SECTION 1
1. Wilson, J.H. Jr., D.L. Keyes, and J.W. Lent. Cost and Economic Analysis
of Regulatory Alternatives for NO2 NAAQS. June 1982.
5

-------
2. LEAD EMISSIONS AND AMBIENT CONCENTRATIONS
Although some lead in the atmosphere results from natural process such as
the wind erosion of lead-bearing dust, lead emissions from human activity
greatly outweigh these natural processes, both in the magnitude of lead
emissions and in their contribution to ambient lead levels. Lead emissions
result not only from the processing and use of lead and its compounds, but
also from the processing and use of materials contaminated by lead. This
section presents an overview of lead emissions and ambient concentraions
resulting from them.
2.1 LEAD EMISSIONS
An inventory of lead emissions in the U.S. is shown in Table 2-1;
assumptions and data sources used to derive the values given in the table are
detailed in Appendix A. As Table 2-1 shows, gasoline combustion in
automobiles currently accounts for the largest share of total lead emissions.
These emissions result from the use of alkyl lead compounds as antiknock
additives in gasoline. Although the use of leaded gasoline has been phased
down by about 75 percent since 1974, emissions from its use are still
substantial.
Stationary sources are divided into three categories in
Table 2-1 -- combustion sources, metallurgical processes, and other industrial
processes. Lead emissions from crankcase oil combustion result from
contamination of automotive oil with lead due to leaded gasoline use. Some
solid wastes also are contaminated with lead. Coal and residual oil contain
lead in trace quantities. Emissions from combustion of these fuels are
substantial because of the large amounts of these fuels that are burned.
i
6

-------
TABLE 2-1. ESTIMATED ATMOSPHERIC LEAD EMISSIONS
FOR THE UNITED STATES, 1981
Annual emissions
Source category	(Mg/yr)
Mobile Sources
Gasoline combustion	35,000
Stationary Sources
Waste oil combustion	830
Solid waste incineration	319
Coal combustion	950
Oil combustion	243
Ore crushing and grinding	325
Primary lead smelting	921
Secondary lead smelting	631
Iron and steel production	533
Gray iron production	295
Primary copper smelting	30
Other metallurgical	54
Lead alkyl manufacture	245
Lead acid battery manufacture	85
Portland cement production	71
Miscellaneous	223
TOTAL	40,700
7

-------
Lead emissions result from the processing of lead ores, and from primary
and secondary lead smelting. Because lead is present in other metal ores, it
is also emitted from iron and steel production and from the smelting of copper
and other metals.
Lead emissions from other industrial processes generally result from the
use of lead to produce products such as lead-acid batteries, leaded glass, and
alkyl lead gasoline additives. Lead emissions from Portland cement
manufacture result from the presence of trace quantities of lead in the cement
raw materials.
The nature and chemical form of lead emissions depend on their origin and
the mechanism of formation. Combustion processes and smelting processes,
because of their high temperature, emit submicron particulate fumes. Lead
emissions from handling and mechanical processes, such as ore processing,
comprise larger particles. Chemical forms of lead emissions include elemental
lead (Pb), its oxides (PbO, PbC>2, Pb03, etc.), lead sulfates and sulfides
(PbSo^, PbS, etc.), lead chlorides and bromides (PbCl2» PbClBr, PbBr2» etc.),
and alkyl lead (Pb(CH3)4 and Pb(C2H5)4).
2.2 TYPICAL AMBIENT LEAD CONCENTRATIONS NEAR EMISSION SOURCES
Although leaded gasoline combustion is the primary source of lead
emissions in the U.S., the majority of exceedences of the current NAAQS
results mainly from stationary source emissions. The average quarterly lead
concentration detected by roadside ambient lead monitors (5-15 meters from the
road) was 0.77 micrograms/cubic meter (ug/m^) for 1983, and none of the
roadside monitors then in place have showed exceedences of the current
standard. This absence of exceedences for such a large emission source
results from the wide spatial dispersion of automotive admissions.
For similar reasons, many of the major stationary sources of lead
emissions listed in the previous section do not cause correspondingly numerous
or significant exceedences of the current standard. Table 2-2 summarizes
typical maximum quarterly lead concentrations around lead emission sources.
Ambient monitoring results are presented where data are available. Where
monitoring data are unavailable, the table presents the results of dispersion
modeling studies described in Sections 3 and 5.
8
T

-------
TABLE 2-2. TYPICAL MAXIMUM QUARTERLY LEAD CONCENTRATIONS
NEAR LEAD EMISSIONS SOURCES
Source category
Ambient lead
concentration
(ug/m^)
Data source
Mobile Sources
Gasoline combustion
0.36-1
Monitoring a'^
Stationary Sources
Waste oil combustion
Commercial	0.15
Utility	<0.1
Solid waste incineration	<0.01
Coal combustion	<0.01
Oil combustion (residual)	<0.01
Ore crushing and grinding	1.7
Primary lead smelting	6.6
Secondary lead smelting	3.0
Iron and steel production	<0.5
Gray iron production	<0.5
Primary copper smelting	0.8
Leaded brass production	0.3
Leak alkyl manufacture	1.6
Lead-acid battery manufacture 0.5
Portland cement production	<0.01
Leaded glass production	<0.1
Modeling
Modeling
Modeling
Modeling
Modeling
Modeling
Monitoring
Monitoring
Modeling
Modeling
Monitoring
Modeling
Modeling
Modeling
Modeling
Modeling
aFrom the SAROAD computer file for ambient lead data.
^Concentrations are for monitors located 5-15 meters from well-traveled
road (>50,000 vehicles/day).
9

-------
The table shows that the combustion sources of lead emissions have low
ambient lead impacts. This is not surprising in the case of coal and residual
oil combustion and solid waste incineration. The magnitude of nationwide lead
emissions from these categories (Table 2-1) is due more to the volume of
material burned and the numbers of sources more than to high concentrations of
lead in fuel or solid waste. Also, these sources have high stacks which
promote dispersion of emissions.
The low ambient impacts predicted for waste oil combustion are somewhat
surprising, as waste oil contains about 1200 ppm lead. The predicted impacts
are low because pure waste oil is expected to be burned only in small
commerical-sized heating units. Where waste oil is burned in large units, it
is added to other fuels. Also, in large units burining waste oil, particulate
emission controls are often used, and these would reduce lead emissions in the
form of particulate matter. The low predicted impacts for small waste oil
combustion units result partially from limitations in atmospheric dispersion
modeling. These small combustion sources typically have short stacks. The
short stack height, coupled with building wake effects and stack downwash,
would cause the plume to impact very close to the source. However, dispersion
models generally cannot estimate concentrations closer than 100 meters from a
source. Thus, dispersion modeling may not give the maximum ambient impact for
these sources.
The high ambient impacts around primary and secondary lead smelters result
mainly from low level fugitive emissions from these plants. Process emission
sources are typically well controlled for these categories. The fugitive
emission sources include materials handling, furnace upsets, and furnace
charging and tapping operations. The high ambient impacts around smelters are
localized; concentrations typically fall below 1.5 ug/m^ within 1 km, and
below 0.1 ug/m^ within 5 km.
For iron and steel production, gray iron production, lead-acid battery
manufacture, brass and bronze production, and primary copper-smelting the
typical lead impacts as shown in Table 2-2 are well below the 1.5 ug/m^
standard. However, for uncontrolled or poorly controlled facilities in these
source categories, ambient impacts are expected to be substantially higher
than those presented, resulting in exceedences of the current standard.
10
T

-------
Spatial variations in ambient lead concentration are illustrated for
stationary and mobile sources in Figures 2-1 and 2-2. Figure 2-1 shows ambient
concentrations predicted by dispersion modeling around a secondary lead
smelter. Sources of lead emissions from the smelter include both fugitive and
stack sources. As the figure shows, the ambient concentration drops rapidly
with distance from the smelter. Figure 2-2 graphs the lead concentration near
a roadway as a function of height and distance from the road. The concentra-
tions shown are averages of monitored lead levels for April and May 1980 near a
road with average deily traffic of about 60,000 vehicles per day. The lead
concentration near the road is approximately constant between 5 and 15 meters
from the road. The results of other EPA modeling studies indicate that the
concentration declines rapidly for distances greater than 15 meters.
2.3 LEAD CONCENTRATION DISTRIBUTIONS AND RECENT TRENDS
The distributions of maximum quarterly lead concentrations at ambient lead
monitoring sites were studied using the computerized SAROAD ambient monitoring .
data base (December 1984). Distributions were determined for the entire set of
monitors, for sites designed to monitor concentrations near stationary sources,
and for sites designed to monitor roadside lead concentrations due to mobile
source emissions. Roadside lead monitors are divided into three subsets based
on distance from the roadway. Microscale roadside sites are closest to the
monitored roadways, middle scale roadside sites are at an intermediate
distance, and neighborhood sites are at a greater distance from the monitored
roadway. A detailed definition of the three types of roadside monitor site-
types is given in Section 3.
Monitoring data were studied both for the period 1980 to 1983 and for the
most recent year of data, 1983. Table 2-3 shows the frequency with which the
O
maximum quarter fell into five concentration ranges -- 0 to 0.5 ug/m , 0.5 to
1.0 ug/m1.0 to 1.5 ug/m"', 1.5 to 2.0 ug/m"', and over 2.0 ug/m^. Table 2-4
gives other properties of the distribution of maximum quarterly concentration,
including: the minimum and maximum observations, the frequency distribution
peak (or mode), the median and mean, and the 80th, 90th, 95th, and 99th
percentile levels.
11

-------
0.5
1.0 km
Figure 2-1. Predicted maximum quarterly average ambient
lead concentrations around a large secondary
lead smelter (vig/m3)
12
n

-------
1.40-
.1.30-
1.20-
1.10
«E 1.00-
J 0.90-
0
2 0,80 -
5c
UJ
u
1	°-70"
o
UJ
mrni 0.60"
0.50-
0.40-
0.30-
0.20-
0.10 -
1	r
i I r
E^-ElEliliVAT 1 on
TOWER
HO. 2
TWER
MO. 1
TOWER
NO. 3
I
6 8 10 12 14 16 18
MONITOR SETBACK DISTANCE, meters
20 22
Figure 2-2. Average 24-hour concentration of lead
at various elevations and setback distances.
13

-------
TABLE 2-3. FREQUENCY DISTRIBUTIONS OF MAXIMUM QUARTERLY LEAD CONCENTRATIONS BY SITE-TYPEa
ssses:s:csBSBsessss:aESSBei:estE6stsBtscesass:sssec:sBc&£:s&BBBBs:s3EaBSs:siss&ssa:
Percent of site-years in concentration ranges1
Site type/time frame <0.5 0.5-1.0 1.0-1.5 1.5-2.0 >2.0
3SSSSS
.b
Total Total
site- number
years of sites
1980 through 1983
All monitors	69
Stationary source	62
Microscale roadside	17
Middle scale	39
Neighborhood scale	61
1983 only
All monitors	71
Stationary source	53
Microscale roadside	16
Middle scale	53
Neighborhood scale	72
22
16
40
42
32
19
24
47
33
28
5
8
27
18
5
5
9
32
13
0
2
3
13
0
0
1
2
5
0
0
2
11
3
0
2
3
11
0
0
0
1189
159
30
33
62
360
45
19
15
29
414
56
18
15
29
360
45
19
15
29
aData are the numbers of site-years for which the maximum quarterly concentrations fall
within the designated concentration ranges. To be included, a site-year must have four
valid quarters of data.
b	1
Concentration ranges are In units of ug/m .

-------
As illustrated in Tables 2-3 and 2-4, the total number of lead monitors is
much larger than the sum of the numbers of monitors designed to cover station-
ary and mobile lead emission sources. This discrepancy results from recent
changes in lead monitor siting guidelines, which have caused a major reworking
of the national monitoring network. The overall monitor network includes many
monitors which were positioned prior to the promulgation of the current siting
guidelines. These monitors were sited to determine urban lead levels, and did
not generally conform to guidelines for roadside monitors or point source
monitors.
Trends in lead concentrations over the period 1980 to 1983 were analyzed
both for maximum quarterly concentration and annual average concentration. In
order to avoid biases relating to changes in the monitoring network, the trend
study included only sites where valid data were available for three of the four
years studied. (A valid site-year must have valid data for all four quarters.)
The trend study was done for stationary source monitors; microscale, middle
scale, and neighborhood scale roadside monitors; and the entire set of lead
monitors. Figures 2-3 and 2-4 illustrate trends in maximum quarterly lead
concentration and annual average lead concentration, respectively, for the five
types of monitors. Table 2-5 presents the data illustrated in the figures, and
shows the numbers of sites meeting the study criteria for each of the monitor
types. From the difference between monitor counts given in Tables 2-3 and 2-4
and those given in Table 2-5, it can be seen that only a small percentage of
the roadside monitors in service in 1983 were in service at the beginning of
the study period.
2.4 EFFECTS OF AVERAGING TIME ON AVERAGE LEAD CONCENTRATION
When the ambient lead standard was originally proposed in 1977, the
averaging time for the primary standard was specified as a calendar quarter.
Steady-state lead-in-blood levels for adults adjust to changes in air lead
concentration in approximately 60 days. The one month averaging time was
originally proposed because of the risk of exposure to young children, whose
faster metabolisms would presumably result in faster adjustments to air lead
concentrations.^" After reviewing public comments on the proposed standard, EPA
promulgated the NAAQS with an averaging time of a calendar quarter. The

-------
MICftC SCALE
iTATIOMARV SOURCES
... NEiGHBwB-H00D SCALE
lUjl'L- iU-Lt .Y~,.
ALL WONITW,^
1.S7S
1.581
Tear
r
!.$>€
Figure 2-3. Trends in mean maximum quarterly lead
concentration for various monitor groups.
1 ,5 •
1 .A
1.3
1 .2
1 ,1
1.0
0.3
0.3
0.7
0.6
0.3
O.A
0.3
C.2
0.1
c.o
K.
STATIONARY SOURCES
Xx,
MICROSOALE
NEIGHSOfiHOOQ 5CAL:
$10CL£ SCALE
ALL MOMlfOS?i!£r-
1 ,~i 73
1 .T'3.1
Year
1 C.® •
Figure 2-4. Trends in overall annual average lead
concentration for various monitor groups.
16


-------
TAHLE 2-4. l'ROl'ERTTES OF T1IE DISTRIHUTION OF MAXIMUM QUARTERLY LEAD CONCENTRATION FOR VARIOUS SITE TYPES
:ssssecssssBB = =suss=ssssBBessBBBB=sasBSSBasssBS8aBBsesa£BBSSBe3srcaB=sassss=sseaesssBssa|)S3ss = 3=s = s = ssscsBeB^ssaast
Other distribution benchmarks	Total Total
Minimum		:	 Maximum site- number
Site type/time frame value Mode Median Mean	80% 90% 95% 98% 99% value years of sites
1980 through 1983
All monitors	0	0.3	0.4	0.53	0.6
Stationary source	0	0.2	0.4	0.95	1.1
Mlcroscale roadside	0.3	0.5	0.9	0.99	1.4
Middle scale	0.2	0.4	0.6	0.68	1.0
Neighborhood scale	0.1	0.4	0.5	0.51	0.7
1983 only
All monitors	0	0.3	0.3	0.51	0.6
Stationary source	0	0.2	0.5	0.88	1.1
Mlcroscale roadside	0.3	0.6	0.8	0.83	1.2
Middle scale	0.3	0.4	0.5	0.60	1.0
Neighborhood scale	0.1	0.3	0.4	0.40	0.5
1.0	1.4	2.2	3.4	13.7	1189	414
2.2	3.6	(6)	(10)	13.7	159	56
1.7	1.8	I I	2.6	30	18
1.1	1.2	I I	1.5	33	15
0.9	1.1	(1.4) I	2.1	62	29
1.0	1.4	2.9	(4.1)	6.7	359	359
3.0	(3.7)	I I	5.5	45	45
II	I I	1.4	18	18
I	I	I I	1.1	15	15
I	I	I I	1.0	29	29
a
Data are lead concentrations In ug/m3. To be Included a site year must have four valid quarters of data.
b
I Indicates that Insufficient observations are available to calculate the given percentile levels.

-------
TABLE 2-5. LEAD CONCENTRATION TRENDS FOR 1980 TO 1983
Mean concentration (ug/m3)	Number
	-		of
Site type3	1980 1981 1982 1983 sites
Maximum quarter
All monitors
0.79
0.61
0.52
0.39
181
Stationary source
1.75
1.47
0.61
0.48
27
Microscale roadside
1.87
1.28
1.32
0.99
3
Middle scale
0.88
0.63
0.75
0.60
4
Neighborhood scale
0.86
0.64
0.52
0.50
9
aual average





All monitors
0.55
0.44
0.36
0.30
181
Stationary source
1.24
0.98
0.44
0.41
27
Microscale roadside
1.25
0.92
0.89
0.80
3
Middle scale
0.56
0.47
0.43
0.44
4
Neighborhood scale
0.60
0.46
0.38
0.37
9
aTo be included, a site must have four valid quarters of data for
each of three out of the four years studied.

-------
possibility of elevated air lead levels within the quarterly period was not
expected to have a substantial impact on public health. However, data from
recent studies have prompted EPA to consider shortening the averaging time for
the standard. Therefore, a review of ambient lead data was conducted to
identify maximum monthly and daily average levels which would be equivalent to
various quarterly levels, in terms of the degree of protection afforded to the
public.
Ratios between maximum monthly concentrations and quarterly average
concentrations were calculated for the entire set of SAROAD lead monitors, and
for three lead subgroups: point source monitors, microscale roadside monitors,
and the combination of middle and neighborhood scale roadside monitors. In
addition, frequency distributions for the monthly-to-quarterly concentration
ratio were determined. These distributions are shown in Figures 2-5 and 2-6.
As these figures show, the monthly-to-quarterly concentration ratio is
generally between 1.1 and 1.3, and seldom exceeds 2. If identical numbers of
samples are collected in all months of a quarter, the theoretical maximum
monthly-to-quarterly ratio is 3. Figure 2-3 shows a small peak at 3.0 for the
overall monitor population, and some ratios over 3. Values in excess of the
theoretical maximum of 3 can and do result when far fewer samples are collected
during the highest concentration month than during the other two months.
The largest discrepancy between monthly average concentration and
quarterly average concentration is in the case of point source monitors, where
the average monthly-to-quarterly concentration ratio is about 1.4. The
monthly-to-quarterly lead concentration ratios or microscale and neighborhood
scale monitors are somewhat lower than the average ratio for the monitor
population as a whole.
Table 2-6 presents mean ratios, medians, distribution peaks (modes) and
various upper-limit levels for the distributions. The table also gives the
correlation coefficient of the ratio with the measured average concentration.
The low correlation coefficients given in Table 2-6 suggest that ratios studied
are largely independent of average lead concentration. Figure 2-7 illustrates
the dependence of maximum monthly concentration on quarterly concentration.
This figure presents data for 500 monitor-quarters selected randomly from the
10,711 monitor-quarters in the SAROAD system. As Figure 2-7 shows, the maximum
19

-------
2,4 -T
K>
O
CO
13
e
w
in
3
O
.e
4J
<11
o
c
0)
N
P
o
u
o
o
G
V
3
o*
(U
Pm
1 .6 -
1 .+ -
1 .2
1 -
0,8 -
0.6 -
0.4 -
0,2 -
0 ••
r^1
V

q
Point
sources
All
\ monitors
b
\
\

_,w--
w
N..
~1~
*1——I*-™*,!	^3"
—I	~—J—|fl -^--^-4^. •—ui—=^W»-yj—(jj-—ID¦* ¦ 1 jI — Hi -
_r-
I .4
1.8	2,2	/:.£	J
Ratio of monthly to quarterly lead concentration
3.-1-
Figure 2-5. Frequency distribution of monthly-to-quarterly lead concentration
ratio for all lead monitors and for point source monitors.

-------
!•' 0 -r—
bO
f
j-.
- !


-------
AVC-C4
8 H
~ cP

T
o-o
0-5
1.0 1-5 2-0 2-5 3-0
QUARTERLY AVERAGE LEAO CONC (UG/M-»31
LEGEND : SOURCE. O ~ ~ NQNPQ1NT 4 a a POINT
3-5
4.0
Figure 2-7. Monthly average lead concentration
plotted against quarterly average.
22

-------
TABLE 2-6. MONTHLY-
VARIOUS
•TO-QUARTERLY AVERAGE CONCENTRATION RATIOS FOR
LEAD MONITORS CATEGORIES
Distribution
property
All
monitors
Point
source
monitors
Microscale
roads ide
monitors
Neighborhood
scale
roadside
monitors
Mean ratio
1.35
1.40
1.27
1.27
Median
1.28
1.31
1.23
1.25
Distribution peak
1.20
1.25
1.20
1.20
Upper limits (one tailed)



80 percent
90 percent
95 percent
98	percent
99	percent
1.50
1.68
1.89
2.24
2.62
1.60
1.82
2.09
2.43
2.86
1.30
1.56
1.79
1.80
2". 00
1.39
1.50
1.67
1.89
2.09
Correlat ion
coefficient3
0.028
-0.021
-0.058
0.068
Correlation coefficient of ration with quarterly average concen
tration.
TABLE 2-7. MONTHLY CONCENTRATION LIMITS EQUIVALENT TO THE QUARTERLY
NAAQS LEVELS, IN TERMS OF ALLOWED AVERAGE CONCENTRATION

Monthly
concentration
equivalents
(ug/m3)
NAAQS

Point
Microscale
Middle and
level
All
source
roads ide
neighborhood
(ug/m3)
monitors
monitors
monitors
scale sites
0.25
0.34
0.35
0.32
0.32
0.50
0.68
0.70
0. 64
0.64
0.80
1.08
1.12
1.02
1.02
1.00
1.35
1.41
1.27
1.27
1.50
2.03
2.11
1.91
1.91
2.00
2.70
2.81
2.55
2.55
2.50
3.38
3.51
3.18
3.19
23

-------
monthly concentration is roughly a straight line function of quarterly average
over the concentration range of greatest importance, 0 to 1.5 ug/m3. This
confirms that the ratio of maximum quarterly concentration to quarterly average
concentration is roughly constant over this range.
The monthly concentration limit that would offer the same degree of
protection, in terms of the measured average concentration, as a given
quarterly limit can be calculated for a given type of monitor by simply
multiplying the quarterly limit by the appropriate mean ratio from Table 2-6.
Table 2-7 lists the monthly concentration limits that would be equivalent to
the quarterly NAAQS levels currently under study.
Monthly limits determined by the above method offer the same degree of
protection as the corresponding quarterly limit in terms of the long term
average (monthly or quarterly) concentration that would be permitted. An
alternate method of defining the equivalence of monthly and quarterly limits is
in terms of the maximum short term concentration that would be expected, for
instance the expected maximum daily concentration. The monthly limit which
would give the same expected daily maximum as a given quarterly limit can be
calculated by multiplying the quarterly limit by the the mean ratio of maximum
daily value to quarterly average, and dividing by the mean ratio of maximum
daily value to monthly average. Table 2-8 gives ratios of maximum daily value
in a quarter to quarterly average, and maximum daily value in a given month to
monthly average for the various monitor types. Table 2-9 lists the monthly
concentration limits that would be equivalent to the quarterly NAAQS levels
currently under study in terms of the expected maximum daily concentration.
24

-------
TABLE 2-8. DAILY-TO-QUARTERLY AND DAILY-TO-MONTHLY CONCENTRATION
FOR DIFFERENT MONITOR TYPES
Site type
Average daily
to monthly
concentration
ratio
Average daily
to quarterly
concentration
ratio
All monitors
Point sources
Microscale roadside monitors
Neighborhood/middle scale roadside
1.72
Maximum daily concentration in the quarter.
Maximum daily concentration in the month.
TABLE 2-9. MONTHLY CONCENTRATION LIMITS EQUIVALENT TO THE QUARTERLY
NAAQS LEVELS, IN TERMS OF EXPECTED DAILY CONCENTRATION

Monthly
concentration equivalents
(ug/m3)
NAAQS

Point
Microscale
Middle and
level
All
source
roadside
ne ighborhood
(ug/m3)
monitors
monitors
monitors
scale sites
0.25
0.35
0.38
0.33
0.33
0.50
0.71
0.75
0.66
0.67
0.80
1.13
1.20
1.05
1.07
1.00
1.41
1.50
1.32
1.34
1.50
2.12
2.26
1.98
2.01
2.00
2.83
3.01
2.64
2.68
2.50
3.54
3.76
3.30
3.35
25
i

-------
3. MOBILE SOURCE AMBIENT IMPACTS AND COSTS
In 1973, EPA promulgated regulations requiring the phasing down of the use
of alkyl lead antiknock additives in gasoline. These regulations were revised
in 1982 and in 1983; the most recent rule, issued in March 1985, sets a
standard of 0.10 grams of lead per gallon of leaded gasoline (gplg)> effective
January 1, 1986.^ As a result of these phasedown regulations, the use of lead
gasoline additives in the U.S. was reduced from about 180 gigagrams (Gg) in
1974 to about 55 Gg in 1982,^>3 an(i is expected to be reduced to about 3 Gg
(2% of the 1974 amount) in 1986.^ Correspondingly, lead concentrations near
roadways have been reduced dramatically since 1974 and are expected to be
reduced further in the future, especially if a total ban on leaded gasoline
being considered for 1988 is put into effect.
The rest of this chapter details the data and techniques used to predict
the ambient impacts of mobile source lead emissions, the methods used to
consider the interaction of mobile and stationary source lead emissions, and
the costs of the lead phasedown program. To determine whether mobile source
emissions alone would cause exceedences of any alternative NAAQS levels, the
upper limits of mobile source ambient impacts were projected for 1990 and
1995. In addition, mean mobile source ambient impacts were predicted for 1990
and 1995, to determine the approximate contribution of mobile sources to
concentrations in the vicinity of point sources of lead. These predictions
were made by first quantifying the relationship between past trends in
roadside concentrations and lead additive use, and then combining this
information with projections of future lead additive use. The upper limit of
ambient lead impacts from mobile sources, assuming that leaded gasoline is not
banned and contains 0.10 gplg, is projected to be 0.15 ug/m^ in 1990, and
0.13 ug/rn-^ in 1995. Thus, because these values are less than the lowest NAAQS
alternative being considered (0.25 ug/m^) and would be even lower if leaded
gasoline were banned, mobile source emissions by themselves are not expected
to exceed whatever new standard is set. However, mobile source emissions may

26

-------
contribute to exceedences in the vicinities of urban point sources. The
average contribution of mobile source emissions to the lead concentration near
a point source in 1990 was estimated to be between 0.05 ug/m^ and 0.02 ug/m^
for the 0.10 gplg standard, or between 0.02 ug/m^ and 0.01 ug/m^ if a total
ban on leaded gasoline were imposed.
3.1 CURRENT AMBIENT IMPACTS FROM MOBILE SOURCES
Regulations promulgated by EPA in 1980 require that, beginning in 1981,
ambient lead impacts from mobile sources be monitored in any urbanized area
with a population greater than 500,000. The primary objective for National
Air Monitoring Stations (NAMS) in such areas is to monitor in the locations
where the population exposures are expected to be the highest, consistent with
the averaging time of the NAAQS. For monitoring ambient lead concentrations,
NAMS siting criteria require (at a minimum) in each urban area (1) a
microscale or middle scale station located in the area of expected maximum
concentrations, and (2) a neighborhood scale station located in an area that
combines poor air quality with high population density, but that is not
necessarily an area of expected maximum concentrations. Both of these
stations are to be located adjacent to a roadway with an average daily traffic
volume (ADT) greater than 30,000 vehicles per day.
Each of the three types of stations mentioned above represents a different
spatial scale:
•	miscroscale--defines the concentrations in air volumes associated
with area dimensions ranging from several meters up to about
100 meters;
•	middle scale--defines the concentration typical of areas up to
several city blocks in size with dimensions ranging from about
100 meters to 0.5 kilometer;
•	neighborhood scale--defines concentrations within some extended area
of the city that has relatively uniform land use with dimensions in
the 0.5 to 4.0 kilometers range.
Currently, 115 NAMS have been approved for lead, including 32 microscale,
28 middle scale, and 55 neighborhood scale sites. Tables 3-1, 3-2, and 3-3
27

-------
TABI.E 3-1. SITE PARAMETERS AND AMBIENT LEAD DATA FOR SELECTED
MICROSCALE NAMS *
oaiaaaHi
I a BOOBBC
• nannaaaixnBBBnaHnaaanBnanpB3BaaDnBnBe
IDOSBB C3 S3 S3
Site number
Probe
hei ght
(meters)
Avg. roadway
volume (1000
vehi clea
per day
Distance
to road
(meters)
Annual
average
concentratlon,
1903 (ug/mS)
Max i mum
quarterly
concentration,
1983 (ug/m3)
01O30OO20GO1	2	30	9	0.B6	1.04
053900001101	5	25	6	0.70	1.10
100S60024G01	3	100	15	1.15	1.39
101960084H01	3	104	14	1.02	1.15
1517B0009H01	3	75	12	0.40	0.43
210120041F01	2	64	13	0.44	0.60
220240002F01	5	35	7	0.51	0.59
243O0OOO4FO1	2	69	7	0.63	0.70
310960002F01	2	97	14	0.73	0.87
320040022H02	3	35	7	0.74	0.95
330660017F01	2	78	.12	0.64	0.69
334680052F01	6	150	10	0.73	0.90
372200018F02	3	15	11	0.44	0.50
381460082F01	4	132	14	1.22	1.37
397140047H02	4	36	9	0.36	0.41
402140003F01	3	64	15	1.22	1.30
451700027F01	3	16	5	1.30	1.50
Mean value	0.77	0.91
Sample std. deviation	0.31	0.35
¦ BBHnianBBanBnBanaHaBBiinBHiiBnBBBiinnnHnnannBHaMnBBHnaBBnBnHiiiMittBBnaniiiaiiaBBaanBttBannBa
*0nly eiteo with -four quarters o-f data -for 1983 were included.

-------
TABLE 3-2. SITE PARAMETERS AND AMBIENT LEAD DATA FOR SELECTED
MIDDLE SCALE NAMS*
inenaaoiHBnaeanaaBaaaiBnaaaHnniaHBBnHai
Avg. roadway	Annual	Manimum
Probe volume (1000	Distance	average	quarterly
height vehicles	to road	concentration,	concentration,
Site number (meters) per day	(meters)	1983 (ug/m3)	19B3 (ug/m3)
030600017G01
4
19
23
0.70
1.08
056980004101
4
10
14
0.34
0.44
057160001101
4
13
19
0.28
0.35
090020027101
10
34
33
0.23
0. 32
101840001G01
4
38
37
0.31
0.41
141220026H01
B
177
13
0.45
0. 52
222160007F01
4
81
51
0.76
0.99
261040002G01
13
53
43
0.35
0. 40
2623S0002H01
6
14
56
0.21
0.33
402160004F01
4
12
16
0. 47
0.36
410300007F01
5
90
23
0.69
0.92
451310029H01
7
79
24
0.29
0.37
451BB0003F01
7
13
21
0.60
1. 00
451SB0023F01
6
70
50
0. 70
0. 90
512200099F01
13
16
28
0.29
0.32
Mean value	0.44	0.139
Sample etd. deviation	0.19	0.29
BBBcanBnaBnaaB»naiBBinnaaiinnaBnnRanniaMiHninaiaainnnBMnMaBnnnain«aiaaBBaRaE)Bnnnp3ia«BnBnannBaaiinBatg
*0nly sites with four quarters of data for 19B3 were included.

-------
TABLE 3-3. SITE PARAMETERS AND AMBIENT I.EAD DATA FOR SELECTED
NEIGHBORHOOD SCALE NAMS*
BnnaiianaBaaBnniannnHBiipaH«BnBnBiniBnMiiniHiinuniaMMBaaBD>MnBBUiannnaiBtBnBiHiiaaiBiiantaiBttnnRBOiaanBnniBi
Avg. roadway	Annual	Maximum
Proba volume (1000	Distance	average	quarterly
height vehiclBo	to road	concentration, concentration,
Site number (meters) per day	(meters)	1983 (ug/m3)	1983 (ug/m3)
030600013Q01
13
10
23
0.36
0. 34
052700001101
6
24
130
0.21
0.30
0341B0103101
13
9
87
0.60
0. 90
056535001101
7
18
38
0.30
0. 50
056860004I01
9
5
15
0.26
0.29
141220052H01
4
140
120
0.49
0.57
151520014H01
10
25
300
0.30
0.39
152040038H01
15
70
83
0. 38
0. 47
172340001F01
5
11
80
0. 14
0. 15
220380002F01
12
75
130
0.31
0. 40
222160011F01
14
46
106
0.42
0.55
260030001GO1
4
33
100
0.24
0.29
281880019G01
11
64
81
0.09
0.10
312320003F01
11
33
92
0.39
0.42
320040019H01
4
1
12
0.35
0.49
3346B0079F01
14
30
1 17
0.36
0. 47
372200033F01
13
9
400
0. 13
0.14
3814600801- 01
4
19
150
0.21
0.28
390400002001
3 .
27
95
0.22
0. 23
397140049H01
B
75
130
0.46
0. 57
402160003FO1
8
49
100
0. 19
0. 23
410300017F01
14
123
132
0.42
0. 60
451700010G01
•3
8
32
0.21
0. 30
434370034F01
IS
50
100
0.40
0. 60
483240005F01
9
43
103
0. 17
0. 24
Mean value



O
• -
«
Q
0. 40
Sample std. deviation


0. 13
0. 18
HBOiacinaniinMHHBnH
HHBBaBHBnnMMIlia

I
B
1
e
i
i
H
B
1
I
I
B
I
BanmannnanaunMBaatBaauMMiaa
nnunnnnnn
*0nly sites with
•four quarters
of data
•for 1983 were
included.


-------
summarize ambient data for monitors that had complete data (4 quarters) for
1983. For each site, annual average and maximum quarterly concentrations are
given, as well as physical site characteristics. Concentration data are
summarized with means and standard deviations.
3.2 PROJECTED AMBIENT LEAD IMPACTS FROM MOBILE SOURCES
The first step in predicting future ambient lead concentrations from the
projected amount of lead additives burned in gasoline is to model the past
relationship between these two variables. Section 3.2.1 discusses the use of
literature information and available monitoring data in defining the
variables, and introduces two techniques used to quantify the relationships
between them. Sections 3.2.2 and 3.2.3 detail the application of these
two techniques to predict ambient lead concentrations for 1990 and 1995.
3.2.1 Definition of Variables and General Methodologies
To calculate the total annual amount of lead burned in gasoline (the
independent variable) in the calendar years 1974 through 1983, information on
average gasoline lead contents reported to EPA by refiners was combined with
Department of Energy figures on total U.S. gasoline consumption. Table 3-4
presents this information as well as references and details on calculations.
Two separate indicators of the trend in ambient lead concentrations were
chosen as dependent variables: the highest quarterly average concentration
for a given calendar year, and also the annual average concentration.
Figure 3-1 shows quarterly average and annual average concentrations at one
neighborhood scale site since 1974 (near the start of the lead phasedown
program). The concentration fluctuates seasonally, but there is a general
downward trend, illustrated by the line highlighting the maximum quarterly
averages and by the line connecting the annual averages. For comparison,
Figure 3-2 shows the downward trend in total lead burned in gasoline per year,
relative to 1974 = 1.0 (see Table 3-4).
31

-------
MAXIMUM CALENDAR QUARTERS
CALENDAR QUARTERS
u
2.0
ANNUAL AVERAGES
u
©•©
1974
1973
1976-»
1977 -»¦ 1978
1979
1980
1981
1982
1983 —»
YEAR
Figure 3-1. Quarterly and annual average ambient lead concentrations
at NAMS 030600013G01
l.o
i »
Ui ^
-J 9
o
0 —
1974
1983
1982
1981
1980
1973 -> 19 79
YEAR
1976 -»• 1977
Figure 3-2. Relative total lead burned in gasoline (1974 - 1.0)
32

-------
Maximum quarterly average (MQA) and annual average (AA) lead
concentrations for the calendar years between 1974 and 1983 were compiled for
all current NAMS. For any given monitor, however, only those years during
which all data met SAROAD summarization criteria (OAQPS Guideline 1.2-040,
Vol. 3, Sec. 2.3.0) were included in the study; this eliminated many
site-years of data.
Two techniques were used to describe the past relationship between the
variables and then to project future ambient lead concentrations from
projected lead-in-gasoline: regression analysis and ratio analysis. Applying
ordinary least squares regression analysis to data from past years produces a
model of the form:
AA (or MQA) = PQ + Pi(GASPB)
where PQ and p^ are the regression coefficients (intercept and slope,
respectively) of a straight line, and GASPB represents the annual lead burned
in gasoline. Thus, once the coefficients are known, annual average or maximum
quarterly concentrations can be predicted by substituting into the equation
the projected amount of lead burned in a given future year.
If, however, the intercept is equal to zero, then a second method, ratio
analysis, can be applied. In this case, the following equation describes the
relationship between the variables:
AA (or MQA) = C • (GASPB)
or:
AA (or MQA) = C	.
GASPB
where C is a direct-proportionality constant, representing the value of the
ratio of the independent and dependent variables. A zero intercept, implying
negligible background concentrations in the absence of lead in gasoline, makes
sense for this analysis for two reasons. Background concentrations can be
caused by point source contributions or by re-entrainment of lead from
contaminated roadside dust. However, the NAMS chosen for this analysis are

33

-------
monitors for mobile sources of lead, and are not affected by point sources of
lead. In addition, by 1990, the level of lead in gasoline will have been very
low (0.10 gplg maximum) for four years, implying that contamination of
roadside dust (and therefore lead re-entrainment) will be minimal. Thus,
intuitively, a zero intercept and the use of ratio analysis appear to be a
better choice than regression analysis.
3.2.2 Application of Regression Analysis
The best monitoring sites for use in regression analysis were those with
data extending back into the mid-1970's, because these show clearly the
downward trend in ambient lead concentrations over time. However, only six
NAMS-approved sites have such data, providing a total of two sets of
32 observations each to be used in conjunction with the data on annual
gasoline lead burned (GASPB) given in Table 3-4. The relationships between
total lead burned in gasoline and the maximum quarterly concentrations for
each of the six sites from 1974 to 1982 are shown in Figure 3-3; each site's
data, and all site data combined as well, show an apparent correlation between
the two variables. If annual average concentrations were graphed in place of
maximum quarterly concentrations, a similar plot would result.
Using the two 32-observation datasets, two simple linear regressions of
the form given in Section 3.2.1 were derived. The evaluative model statistics
(R^ and the standard error of the estimate, Sy/X) indicated fairly good
models, but the residuals plots showed pronounced heteroskedasticity
(nonhomogeneity of variance, increasing as the value of GASPB increased); this
is a violation of an OLS regression assumption, and implies that calculated
confidence limits and significance tests do not apply. To correct the
heteroskedasticity, all values of AA (or MQA) and GASPB were divided by the
GASPB values (a common correction technique), a new regression was run using
these transformed values (e.g., [AA/GASPB] = P0 + Pi[1/GASPB]), the equation
then multiplied through by GASPB to return the variables to their original
metrics (i.e., AA = P0[GASPB] + P^ — note that the coefficients reverse
places), and the evaluative statistics calculated by hand. This yielded the
following two equations:

-------
TABLE 3-4. LEAD BURNED IN GASOLINE
Gasoline Usage
(1000 bbl/day)a
Year
Leaded
Unleaded Total
Average
lead content
of leaded
fuel(g/gal)b
Total
lead
burned
(Gg)c
Total
lead burned
relative
to 1974
1974
6537
0
6537
1.79
179
1.0
1975
NA
NA
6675
1.82
174d
0.97
1976
5424e
1554
6978
2.02
168
0.94
1977
5201
1976
7177
2.03
162
0.91
1978
4891
2521
7412
1.94
146
0.82
1979
4236
2798
7034
1.85
121
0.68
1980
3512
3067
6579
1.38
75
0.42
1981
3324
3264
6588
1.15
59
0.33
1982
3130
3409
6539
1.24
60
0.34
1983
2975
3647
6622
1.14
53
0.30
NA - not available
aReference 3; 1 bbl = 42 gallons.
^Reference 2.
cIn calculating the total amount of lead compounds burned, the average lead
content in unleaded gasoline was taken as 0.014 grams/gallon (Reference 2).
^Because data on leaded and unleaded fuel use in 1975 are unavailable, the
total amount of lead burned in gasoline in 1975 was estimated by taking the
average of the 1974 and 1976 values.
eLeaded and unleaded fuel use rates in 1976 were estimated based on
nine months of data (Reference 4).
35

-------
3.0
uyati
! 19821	Cl980>
\j\ /	l*1
(1979)	(19771(1973)
(19781 )l97S 11974)
J'OI	i'3'CU U3'
-0 SITE ID 0306000I3G0I
•ASITE ID 056400003
-~ SITE ID 0565350011
-X SITE ID 056800004
-+ SITE ID 056980004
-^7SITE ID 101760001 G
2.0
.0
0
±
J	L
0.5	1-0
RELATI VE TOTAL LEAD BURNED IN GASOLINE (1974 = I ]
Figure 3-3. Maximum quarterly lead concentration at NAMS sites as a
function of total lead burned in gasoline.
36

-------
AA = 0.0671 +	0.00706 (GASPB)	R2 = 0.81
(sfi= 0.068)	(s~= 0.001)	sv/x = 0.21
P	P	n - 32
MQA = 0.119 +	0.0101 (GASPB)	R2 = 0.70
(s~= 0.124)	(sft= 0.002)	sv/x = 0.38
P	P	n - 32
The residuals plots appeared randomly scattered and heteroskedasticity was
eliminated.
Using these regression equations to predict future ambient lead
concentrations appeared questionable, though, because at the low levels of
GASPB projected for 1990 and 1995, the intercept would contribute much more to
the predicted AA or MQA value than the (p^[GASPB]) term; that is, the
lead-in-gasoline contribution to roadside lead levels would be lower than the
unexplained contribution. This is counter-intuitive, as discussed in
Section 3.2.1. Using ratio analysis instead of regression analysis to make
predictions, however, hinges on establishing that the intercept is not
significantly different from zero, as discussed below.
The standard errors (sp) for both intercepts were very high (relative to
the magnitudes of the intercepts themselves), indicating that the intercepts
might be equal to zero. Using the null hypothesis H0:P0 = 0 and the
alternative hypothesis H^: PQ > 0, prob-values were calculated and looked up
in a Student's-t table. For both models, the prob-value lay between 0.15 and
0.20, implying that H0 could be rejected with only 807,-857. confidence (the
usual level for rejection is 95% confidence). Thus, as expected, the
hypothesis that the intercept is different from zero is not accepted. An
intercept equal to zero is substantiated both statistically and intuitively.
3.2.3 Application of Ratio Analysis
As discussed in Section 3.2.1, ratio analysis is based on these two
equations:
AA = CX
GASPB
37

-------
MQA = C2
GASPB
In order to base calculation of and C2 on as much past data as possible,
the 32-observation/6-site dataset used in regression analysis was expanded to
include all site-years that had four complete quarters of data. Each of the
final MQA and AA datasets consisted of 122 observations representing the years
1975 through 1983 (primarily 1981-1983), and the following numbers of each
site type: 17 microscale sites (28 observations), 18 middle scale sites
(32 observations), and 28 neighborhood scale sites (62 observations). These
data are given in Appendix B.
Calculation of and C2 and of projected lead-in-gasoline use allows
estimation of mean future mobile source impacts, confidence levels for these
means, and upper limits (99th percentiles) for future maximum quarterly and
annual average concentrations. The first step in calculating and C2 was to
create two new datasets (one for AA, one for MQA) by dividing each value of
MQA or AA by the corresponding GASPB level for the year of measurement,
producing two sets of ratios. Creating histograms from these two 122-point
sets produced log-normal distributions. To calculate distribution properties
and confidence intervals, the data were transformed to produce normal
distributions by taking the natural log of each ratio. Next, the geometric
mean and standard deviation of each ln-transformed distribution of ratios was
calculated, along with an upper limit on each mean ratio (using: x + [t^Q11 *
[s/Vn]), and the 99th percentile of each distribution of ratios (using: x +
^.01J * Cs ]) • Then, to put the means, upper limits on means, and percentiles
back into their original metric, the inverse logs of each of these were
taken. Finally, to arrive at values for AA or MQA (and upper limits and
percentiles) for 1990 and 1995, each ratio was then multiplied by the
predicted values of GASPB for those years. Predicted GASPB was calculated as
in Table 3-4, using values given in the Federal Register^- on projected
gasoline usage, and assuming that leaded gasoline would not be phased out,
would contain 0.10 gplg, that no fuel switching would occur, and that unleaded
gasoline would contain 0.014 gpg lead. Values of GASPB and predicted future
38
T

-------
impacts for all sites types combined (a sort of national average) are given in
the first two columns of Table 3-5.
To allow site-type-specific predictions, the two 122-observation datasets
were each split into three subsets—one for microscale sites (28 points), one
for middle scale sites (32 points), and one for neighborhood scale sites
(62 points)--and the same manipulations as described above were performed on
each subset. The resulting predictions are shown in the first two columns of
Table 3-5.
Finally, all calculations were repeated, using the levels of GASPB that
would result if the use of leaded gasoline were banned, and all gasoline used
contained 0.014 gpg lead.^ Results are given in the last two columns of
Table 3-5.
To determine whether mobile source emissions alone would cause exceedences
of any NAAQS, only the predictions for maximum quarterly concentrations were
considered, as these represent more of a worst-case situation that the annual
average values; specifically, the 99th percentile predictions for microscale
sites were used: 0.15 ug/m^ for 1990, 0.13 ug/m^ for 1995. Thus, because
these values are less than the lowest NAAQS alternative being considered
(0.25 ug/m^), mobile source emissions by themselves are not expected to exceed
whatever new standard is set.
3.3 INTERACTION OF MOBILE SOURCE AND STATIONARY SOURCE AMBIENT LEAD IMPACTS
In projecting the attainment status of areas with both mobile and
stationary sources of lead emissions, a background mobile source ambient
impact was added to the projected impact for stationary sources. As shown in
Table 3-5, the average mobile source impact in 1990 could range from
0.01 ug/nr* (the value calculated for neighborhood scale sites, with a ban on
leaded gasoline) to 0.05 ug/m^ (calculated for microscale sites, with leaded
gasoline in use and containing 0.10 gplg). In analyzing NAAQS costs for
stationary sources, costs were studied for both the highest mean mobile source
impact (0.05 ug/m^) and the lowest mean impact (0.01 ug/m-*). It should be
noted that, in both of these cases, mean mobile source impacts were used, not
upper or lower limits. Clearly, some urban sources are located in areas where

-------
TABLE 3-5. PROJECTED LEAD IN GASOLINE AND AMBIENT LEAD IMPACTS FROM
MOBILE SOURCES

Projections
(leaded oasoline
at 0.1& gplg)
Projections
(total ban on
leaded gasoline)

1990
1995
1990
1995
lead burned in gasoline (Sg/yr)
2.B
2.*
1.4
1.4
Mobile source aebient iapacts (ug/a3l




Annual average values




All site types coabined




K;an iapact
Upper luit an eean*
9?th percentile of distribution
0.021
0.027
0,076
0.01B
0.023
0.066
0.011
0.013
0.038
0.011
0.013
0.03B
Micrcscale sites only




Hean iapact
Uooer liait on aean»
99th percentile of distribution
0.039
0.047
0.11
0.033
0.041
0.092
0.019
0.024
0.053
0.019
0.024
0.053
Riddle scale sites only




Kean iapact
Uoper halt on aean*
99th percentile of distribution
0.021
0.025
0.055
0.018
0.021
0.047
0.010
0.012
0.028
0.010
0.012
0.028
Neighborhood scale sites only




flean iapact
Uooer liait on «ean*
99th percentile of distribution
0.016
0.018
0.046
0.014
0.016
0.039
o.ooei
0.0092
0.023
0.0081
0.0092
0.023
Haxiaua quarterly values




All site types coabined




(lean iapact
Uoper liait on lean*
99th percentile of distribution
0.029
0.035
0.11
0.024
0.030
0.091
0.014
0.018
0.053
0.014
0.018
0.053
Hicroscale sites only




flean iapact
Uooer liait on lean*
99th percentile of distribution
0.048
0.059
0.15
0.041
0.051
0.13
0.024
0.030
0.073
0.024
0.030
0.073
Kiddle scale sites only




Mean iioact
Uooer liait on aean*
99th percentile of distribution
0.029
0.035
0.385
0.024
0.030
0.073
0.014
0.017
0.043
0.014
0.017
0.043
Neighborhood scale sites only




Hean iapact
Uooer liait on Bean*
99th percentile of distribution
0.022
0.015
0.070
0.018
0.021
0.060
0.911
0.012
0.035
0.011
0.012
0.035
*991 confidence
40




-------
the background will be higher than the 0.05 ug/m^ average, while others are
located in areas where the background may be lower than 0.01 ug/m^. Thus, the
use of a constant background concentration results in some inaccuracy in the
cost estimates for individual sources. However, it is not expected to bias
the overall nationwide cost results.
3.4 LEAD PHASEDOWN COSTS
In 1982, the original 1973 lead phasedown rules were replaced with
regulations requiring that the concentration of lead in leaded gasoline be
less than 1.1 grams per gallon.^ The annual operating costs attributable to
these regulations were estimated at 110 million dollars (1984 dollars), or
about 0.2 cents per gallon of leaded gasoline.^ in January 1986, the 1.1 gplg
rule will be replaced by a 0.10 gplg standard, which will create additional
refining costs. Initially, the additional costs in 1986 will be 620 million
dollars per year (1984 dollars), or 0.6 cents per leaded gallon; by 1990, they
will drop to 480 million dollars per year, and the estimated additional costs
for 1995 are 440 million dollars (0.4 cents per gallon). For comparison of
mobile source costs with stationary source costs in this study, the costs
attributable to the 1986 phasedown rules were used.
The best method over and above the lead phasedown program for reducing
mobile source lead emissions in areas with high lead concentrations would
probably be local restrictions on leaded gasoline sales. Another potential
method would be to lower the speed limits in these areas. At lower speeds,
most lead in gasoline is deposited in the engine and exhaust system instead of
being emitted to the atmosphere. The fraction of lead emitted increases with
vehicle speed.®
For a number of reasons, mobile source lead emission controls over and
above the lead phasedown program were not considered in developing costs of
compliance for the lead NAAQS alternatives. Because the average mobile source
background around point sources should not exceed about 0.05 ug/m^, the mobile
source contribution to an exceedence of the alternative levels would be minor
in comparison with the point source contribution. For instance, for the
0.25 ug/m^ alternative, the contribution of mobile source background to any

-------
violation would be at most about 20 percent. In general, most point sources
that must be controlled to meet an alternative NAAQS would require control
even in the absence of mobile source background. Also, where stationary
source controls are applied, they generally are efficient enough so that they
reduce the ambient impact from the source to a level at which there is not an
exceedence even when the background is added.
Finally, if mobile source controls were to be used, the scale at which the
controls would be required would necessarily be much larger than the scale of
the exceedence to be eliminated. As shown in Section 2, the scale of an
exceedence in the vicinity of a stationary source is typically less than
1 kilometer. However, automobiles being operated in the vicinity of the
source will typically have come from locations tens of kilometers from the
source, and may have come from locations hundreds of kilometers away. Thus,
for example, to effect the elimination of an exceedence in an area one square
kilometer in size, control of mobile sources or leaded gasoline sales may be
required for tens or hundreds of square kilometers.
42

-------
REFERENCES FOR SECTION 3
1.	Regulation of Fuels and Fuel Additives: Federal Register. 50:9386-9399.
March 7, 1985.
2.	Supplementary Guidelines for Lead Implementation Plans—Updated
Projections for Motor Vehicle Lead Emissions (EPA-450/2-83-002). U.S.
Environmental Protection Agency, Research Triangle Park, N.C. March 1983.
3.	Energy Information Administration. Petroleum Supply Monthly.
DOE/EIA-0109 (83/03), U.S. Department of Energy, Washington, D.C.
December 1984.
4.	Energy Information Administration. Petroleum Supply Monthly.
DOE/EIA-0109 (78/05), U.S. Department of Energy, Washington, D.C.
May 1978.
5.	Regulation of Fuels and Fuel Additives. Federal Register.
47:49322-49334. October 29, 1982.
6.	Supplementary Guidelines for Lead Implementation Plans
(EPA-450/2-78-038). U.S. Environmental Protection Agency, Research
Triangle Park, N.C. August 1978. p. 153.
43


-------
4. GENERALIZED STUDY METHODOLOGIES
Costs of attaining the alternative lead NAAQS levels were developed for
each of the industries that would be affected by any of the NAAQS
alternatives. These were then summed to obtain nationwide costs. For each
affected industry, one of two basic methodologies was used to obtain control
costs, depending on the nature of the industry. For the primary lead smelting
and alkyl lead manufacturing industries, which comprise a small number of
plants, each plant was considered separately. For larger industries, a "model
plant" approach was used. In this approach, plants were grouped into classes
based on size, process type, control status, and other characteristics. Costs
or other impacts for a class of plants were then estimated based on the
operating parameters of a "model" or hypothetical plant considered to be
representative of the class.
A number of steps were involved in developing costs both for industries
comprising small numbers of plants and for larger industries. First, the
affected industries were categorized and control strategies to be used as
baselines in developing the attainment costs were identified. Ambient impacts
were then determined for the baseline emissions, and emission reductions
necessary to attain the various NAAQS alternatives were quantified. Control
options were identified for each emission source and costs were developed for
each option. Controls to achieve each NAAQS alternative at the lowest net
annualized cost were then selected from among the available options, and the
costs of these selected control systems were summed to obtain total plant and
industry costs of attainment. Thus, as noted in Section 1, the costs
developed in this study are the least net annualized costs of attaining the
various NAAQS alternatives.
It is important to note that this study has been conducted over a period
of several years beginning in late 1982. The original study plan has been
extended to include the additional low level NAAQS alternatives of 0.25 and
0.5 ug/nP, and to include a study of the effect of averaging time on measured

45

-------
lead concentration. Thus, the study is based on data of a number of ages.
For instance, the data used to identify affected industries are from 1982.
The analysis of major lead-emitting source categories is based on data from
late 1983. Finally, the analysis of smaller emitters is based on data from
late 1984. The following discussion deals with basic methodologies and
assumptions used in this study. Specific methodologies and assumptions used
in the analysis of individual industrial categories are presented in
Section 5.
4.1 IDENTIFICATION OF AFFECTED INDUSTRY CATEGORIES
Industry categories which would be affected under the lead NAAQS
alternatives were identified by two methods. First, industries associated
with exceedences of the current NAAQS of 1.5 micrograms/cubic meter (ug/nP)
would be affected by that alternative and by the lower alternative levels of
0.25 ug/nP, 0.5 ug/m?, 0.8 ug/m? and 1.0 ug/nP. Second, as part of the
exposure assessment for the lead Regulatory Impact Analysis (RIA), dispersion -
modeling was conducted for all significant stationary lead sources in ten
urbanized areas. Significant sources of lead emissions were identified using,
the December 1982 updates of the National Emissions Data System (NEDS) and the :
Hazardous and Trace Emissions System (HATREMS) computerized data files. The
ten urbanized areas studied are:
•	Butte, Montana;
•	Chicago, Illinois;
•	Dallas, Texas;
•	Minneapolis-St. Paul, Minnesota;
•	Nashville, Tennessee;
•	Philadelphia, Pennsylvania;
•	Phoenix, Arizona;
•	Tampa, Florida;
•	Washington, D.C.; and
•	Waterbury, Connecticut.
46

-------
The results of these studies were used extensively in identifying industries
which would be affected under the NAAQS alternatives.
Several industries which were found to be major source categories in the
lead emission inventory (Section 2) currently are not associated with
exceedences of the 1.5 ug/m^ NAAQS. For coal and residual oil combustion,
overall emissions of lead are substantial; however projections based on lead
emission factors and past EPA modeling studies for other pollutants indicated
that ambient levels caused by individual members of these categories or
dispersed groups are very low. Therefore, these industries were eliminated as
significant contributors to ambient lead levels prior to the modeling studies
for urbanized areas.
Other industries identified as major sources of total lead emissions were
not included in the modeling for urbanized areas because sources were not
located in the ten areas studied. For these industries, dispersion modeling
was conducted, or, where available, ambient data from nearby monitors were
reviewed to determine whether sources could be affected by any of the NAAQS
alternatives.
The following industry categories were identified as potentially affected
under one or more of the lead NAAQS alternatives:
•	primary lead smelting;
•	secondary lead smelting;
•	lead ore processing;
•	lead-acid battery manufacture;
•	alkyl lead manufacture;
•	gray iron cupolas;
•	alloy steel electric arc furnaces;
•	carbon steel electric arc furnaces;
•	iron ore sintering;
•	brass and bronze production; and
•	primary copper smelting.
47


-------
Source categories that were studied by one or more of the above methods, and
that are not expected to be affected under any of the lead NAAQS alternatives
include:
•	copper ore processing;
•	Portland cement production;
•	waste oil combustion;
•	municipal waste incineration;
•	sewage sludge incineration;
•	leaded glass manufacture;
•	residual oil combustion; and
•	coal combustion.
4.2 SELECTION OF BASELINE CONTROL STRATEGIES
The baseline control strategy for a particular industry is the set of
controls used as a benchmark for the estimation of the impacts of a regulatory
action on the industry. For this study, the baseline controls selected were
those that are currently in use. These controls include those installed to
attain the current NAAQSs for lead, particulate matter, and sulfur dioxide,
and those installed to meet new source performance standards (NSPS). Because
the State implementation process for the current lead NAAQS is not yet
complete, baseline controls in many cases do not achieve the current standard.
Baseline controls are not necessarily the same for all members of an
affected industry, but may represent a distribution of controls across the
industry. Controls currently in use depend on plant size and plant location.
For instance, controls vary depending on whether a plant is in an attainment
or a non-attainment area for particulate matter, and on the degree of
completion of the implementation process for the lead NAAQS in the State where
the plant is sited. For new plants, baseline control was taken as the degree
of control required by NSPS.
48

-------
4.3 ESTIMATION OF BASELINE AMBIENT IMPACTS
Ambient lead impacts were estimated for sources with baseline controls in
order to determine the degree of emission control that would be required under
each of the lead NAAQS alternatives. The baseline ambient impacts were taken
as the maximum average lead concentration resulting from plant emissions for a
calendar quarter.
4.3.1 Dispersion Modeling
Atmospheric dispersion modeling results were the major input in the
estimation of baseline ambient impacts. As noted in Section 4.1, atmospheric
dispersion modeling was conducted for all significant sources of lead in ten
urbanized areas. For each of these urbanized areas, the area covered included
all of the counties associated with the urbanized area in the 1970 census.*
Modeling was conducted for each plant in the primary lead smelting and
lead alkyl manufacture source categories. In addition, dispersion modeling
was performed for a hypothetical lead ore processing plant, two hypothetical
leaded glass plants, a commercial heating unit fired by waste crankcase oil, a
sewage sludge incinerator, and a municipal waste incinerator.
The Industrial Source Complex - Short Term (ISCST) model was used in most
cases. The model was chosen over the ISC-Long Term (ISCLT) model to allow the
assessment of short term exposures as well as long term exposures. Where
ISCST was used, terrain effects were neglected. For five primary lead
smelters, terrain effects were considered important and the COMPLEX I model
was used to take them into account. These models were used to predict, for
each source, the daily average ambient lead impact at an array of receptors
for a full year. The receptor network was radial, with 160 receptors in 16
directions in 22.5° sectors, at distances of 0.2, 0.3, 0.5, 0.7, 1, 2, 5, 10,
and 20 kilometers from the source. For the modeling in the urbanized areas,
receptors were also located at the centroids of all of the census tracts. The
daily average concentrations were averaged to obtain impacts for four calendar
quarters.
49

-------
For the urbanized area exposure assessment modeling, the meteorological
data used were from aerometric stations in those cities. For modeling of
individual sources, the data used were from the closest station for which data
have been preprocessed for short term EPA dispersion models, and were the most
recent data available. Intermediate plume rise was assumed, and rural or
urban settings for the models were selected as appropriate based on plant
location. Stack downvash effects were taken into account, using NEDS stack
parameters. Building wake effects were included for emissions released at
roof level. Generally, building dimensions were assumed to be 50 meters by
50 meters by 10 meters high. Particle deposition was taken into account where
large particles were known to be emitted. Generally, however, particle size
data were not available, so particle deposition effects were not included.
4.3.2	Emissions Data used in Modeling Studies
The primary source of emissions data for dispersion modeling was the
computerized NEDS file. Stack parameters were taken directly from NEDS.
Where possible, lead emission rates were calculated using the operating rates
given in NEDS and lead emission factors published by EPA.2>3
For lead-acid battery manufacture and alkyl lead additive manufacture,
the NEDS files were incomplete. For primary and secondary lead smelters,
fugitive emissions generally are not addressed in NEDS. For these four source
categories emissions data were developed from published sources and contacts
with EPA Regional Office personnel and State and local agency personnel. The
derivation of model inputs for these categories is described in further detail
in Section 5.
4.3.3	Baseline Ambient Impacts for Industries with Small Numbers of Plants
For the lead alkyl manufacture and primary lead smelting industries,
which comprise a small number of plants, costs were developed on a
plant-by-plant basis. Therefore, baseline ambient impacts were estimated for
each individual plant. In the case of lead alkyl manufacture, the baseline
impacts were estimated based entirely on dispersion modeling results. The
50
¦}

-------
baseline impact for each plant was taken simply as the maximum quarterly
average concentration predicted by dispersion modeling for four calendar
quarters.
For four of the five primary lead smelters, ambient lead concentration
data are available from nearby monitors. For two of these smelters, source
apportionment studies have been conducted as part of implementation programs
for the current lead NAAQS. Data from ambient monitors and source
apportionment studies were used in place of or in conjunction with dispersion
modeling results to predict baseline ambient impacts for primary lead
smelters. The development of baseline ambient impacts for primary lead
smelters is discussed in more detail in Section 5.
4.3.4 Baseline Ambient Impacts for Industries with Large Numbers of Plants
For most of the industries considered in this study, dispersion modeling
was conducted for only a small set of sources. Thus, it was necessary to
extrapolate these modeling results to estimate ambient impacts for the entire ;
population of plants. This was done using source-receptor coefficients. A
source-receptor coefficient is defined as the ambient impact of a source : r
divided by its emission rate. In this study, the source-receptor coefficient -
for a given source was calculated by dividing the maximum quarterly
concentration by the emission rate for the source. Source-receptor
coefficients were calculated for sources modeled in the ten urbanized areas
listed above.
It was assumed that the maximum ambient impact occurs outside the plant
boundary. For lead-emitting industries comprising large numbers of plants,
most plants are located in urban areas where large tracts of property are
expensive and difficult to obtain. These plants were assumed to own property
to at most 150 meters away from the actual emission sources. Because the
maximum lead impacts predicted by dispersion modeling occur at distances of
200-300 meters, even for low lying sources, these maxima generally will be
outside plant property.
Source-receptor coefficients generally vary with emission height,
emission temperature, and other stack parameters; and with atmospheric
1
51 :
T

-------
conditions. The coefficient is independent of emission rate. It is also
independent of plant size if stack velocity and temperature are independent of
plant size. For a given type of operation in a given industry, stack
parameters generally are similar across the industry. Therefore, much of the
variation in the source-receptor coefficient for a given type of operation
results from variations in meteorological parameters from site to site.
Nevertheless, differences of up to an order of magnitude were found in
source-receptor coefficients for various operations.
For each type of lead emitting operation, the mean source-receptor
coefficient was calculated for all operations of that type modeled in the ten
cities assessed. Ambient impacts for sources outside the ten cities were then
calculated by multiplying their emission rates by the appropriate
source-receptor coefficients. Source-receptor coefficients used for this
analysis are given in Section 5.
4.4 SOURCE INTERACTIONS
For other criteria pollutants, the interaction of emissions from large
numbers of plants is a major cause of exceedences of ambient standards.
Because of the distances that separate major lead emitting sources, however,
interaction of emissions from separate plants were not considered in this
study.
Figure 4-1 shows the lead concentrations predicted by dispersion modeling
for census tract centroids in central Dallas, Texas. The figure does not
include the entire Dallas metropolitan area, but shows only the largest
concentration of emission sources. Plants G and H are secondary lead smelters
and Plant A is a Portland cement plant. Plants G and H are large sources of
lead emissions, each of which causes significant exceedences of the current
standard of 1.5 ug/m^. Although the sources are less than 10 kilometers
apart, there is no source interaction. If the emissions from plants G and H
were lowered sufficiently to result in attainment of the current standard, the
plants would have to be moved to within about 0.5 kilometer for the
interaction of their emissions to cause a violation.
52

-------
Based on the studies conducted in the ten cities for the exposure
assessment, sources generally are not located in such close proximity. In
fact, in all of the modeling done for the ten cities, 110 instances of
interactions between sources were found that would not be eliminated when both
sources are brought into attainment at their plant boundaries.
A study was also made of all sources of lead emissions nationwide, using
the NEDS files. The source receptor coefficients discussed in the previous
subsection were used in conjunction with NEDS data on particulate emission
rates and published data on particulate lead concentrations to estimate the
maximum ambient lead impact from each lead-emission source. For each pair of
plants located within ten kilometers of one another, the ambient impact of
each plant was estimated at the site of maximum ambient impact for the other.
This was done by assuming that the ambient impact for each plant declines as
the reciprocal of the distance from the plant and using plant location data
given in NEDS. Source interaction was considered to occur where the ambient
impact from one plant makes more than a ten percent difference in the emission
reduction necessary for another plant to achieve an NAAQS. By this method,
source interaction was found to occur for a total of only five plants
nationwide for the 0.5 ug/m^ alternative and none for the 2.5 ug/m^
alternative.
The neglect of source interactions in this study may have resulted in
some underestimation of total costs. Because of the absence of source
interaction in the exposure assessment cities and the small number of
interacting pairs detected in the cursory NEDS analysis, this bias is not
expected to be substantial in comparison with the total control costs.
Interactions of stationary source emissions with mobile source emissions
were taken into account for urban sources by superimposing the stationary
source impacts on a calculated average mobile source background level. Two
mobile source background concentration levels were studied, 0.01 ug/m^ and
0.05 ug/m^. These levels cover the range of possible average mobile source
backgrounds which might exist in 1985. The derivation of these average
background levels was based on two possible final forms for lead-in-gasoline
phasedown rules. Details of the calculations are given in Section 3.
53

-------
4.5 IDENTIFICATION OF INDIVIDUAL AFFECTED SOURCES
Using baseline ambient impacts and background ambient lead
concentrations, individual plants were identified that would be affected under
the various lead NAAQS alternatives. For most of the affected industries,
individual plants were studied to determine their status with respect to the
various NAAQS alternatives. In general, a plant was determine to be affected
under a given alternative if the sum of the baseline ambient lead impact and
the background concentration from mobile sources around the plant would exceed
the given alternative.
4.6 CONTROL STRATEGY IDENTIFICATION AND COSTING
For the major lead-emitting sources in the industry categories identified
as potentially affected under the various NAAQS alternatives, emission control
options were developed and assessed. These were developed using a number of
sources. Major sources used in control option development include:
•	State implementation plans for the current lead NAAQS;
•	background information documents (BIDs) for NSFS for particulate
matter and lead emissions;
•	control techniques guideline (CTG) documents for particulate matter
emissions; and
•	other EPA publications.
Control efficiencies for the various options also were obtained from these
sources.
For each control technique, capital and annualized costs were developed.
The approach taken in the development of these costs depended on the number of
plants in the affected industry category. For the primary lead smelting and
alkyl lead gasoline additive manufacture categories, which comprise a small
number of plants, control costs were estimated for each individual plant. For
larger industries, control costs were developed for model plants, representing
various size categories.
54

-------
4.6.1 Capital Costs
The two components of total capital costs are direct capital costs and
indirect capital costs. Direct capital costs include such items as equipment
purchase costs, costs of instrumentation and process controls, ductwork and
piping, electrical work, structural and foundation costs, and direct labor
costs associated with construction and installation. Indirect capital costs
include such items as engineering and design costs, contractor fees,
supervisory expenses, and startup and performance test expenses. Capital cost
estimates for the control options included in each control technique were
developed from cost data obtained from manufacturers of emission control
equipment, industry trade associations, government publications, and standard
cost references.
The standard cost references used in developing the cost estimates
included:
•	"Estimating Costs of Air Pollution Control Systems." (Parts I
through XI). Chemical Engineering: and
•	Gard, Inc. Capital and Operating Costs of Selected Air Pollution
Control Systems (EPA-450/5-80-002). U.S. Environmental Protection
Agency, Research Triangle Park, N.C., December 1978.
BIDs and CTGs for particulate matter emissions were also major sources of cost
information.
An additional component of capital cost associated with the installation
of emission control techniques at an existing plant is the retrofit cost
penalty. This cost, which can range from 10 to 70 percent of total installed
capital cost, is due to space limitations and additional structural and
ductwork requirements for a retrofitted control system.
4.6.2 Annualized Costs
The two components of total annualized costs are annualized capital
charges and annual operating and maintenance costs for the control options.
Annualized capital charges include capital recovery of the total capital costs
55
i

-------
and the miscellaneous capital charge of taxes, insurance, and administration.
The capital recovery factor used is a function of the interest rate selected
and the general life expectancy of the piece of equipment or capital
investment. The interest rate that was used in developing the capital
recovery factors is 10 percent and the average life of the control options
included in each control technique ranges from 5 to 25 years. In addition, a
standard factor of 4 percent of the installed capital cost was used for the
miscellaneous annual charge of taxes, insurance, and administration. The
annual operating and maintenance costs include such items as operating and
maintenance labor, utilities, fuel usage, repair materials, and overhead. As
with the capital costs, the total annualized cost estimates for each control
option were developed using cost data obtained from the industry trade
associations, government reports, standard cost references, equipment
manufacturers, and architectural and engineering firms.
In addition to total annualized costs for the control options in each
control technique considered, annual savings and costs associated with product
recovery or disposal were estimated. Product recovery is achieved when an
emission control device captures a pollutant that can be recycled into the
production process, sold as a byproduct, or used as fuel or a fuel substitute,
These annualized savings are presented as a separate component of the cost
analysis, but are deducted from the total annualized cost of each respective
control technique to give a final total net annualized cost or savings. It
should be noted that annualized savings due to product recovery, like other
capital and annualized cost components, do not necessarily apply to all
control techniques or options.
4.6.3 Cost Updating
All costs or savings (capital and annualized) were updated to October
1984 dollars. Capital costs were updated using the appropriate factor from
the Chemical Engineering Plant Index. Annual operating and maintenance costs
were calculated from these updated capital costs and the October 1984 costs
for such items as utilities and labor. Annualized costs were updated in two
steps where possible. The annualized capital charges were recalculated using
1
56

-------
the updated capital costs. Annual operation and maintenance costs were then
updated using the appropriate cost indices. In some of the sources used, the
annualized costs are not broken down such that this methodology could be
employed. In these cases, the Chemical Engineering Plant Index was used to
escalate the entire annualized costs.
4.7 ESTIMATION OF LOWEST COST OF ATTAINMENT FOR EXISTING SOURCES
Although interactions of emissions from two or more separate plants are
not a major cause of NAAQS exceedences, exceedences often are caused by the
interaction of different emission sources at a single plant. Thus, for plants
with a number of lead-emission operations, various control strategies were
compared to determine which would achieve each NAAQS alternative at the least
cost.
4.7.1 Least Cost Control Strategies
The development of a least cost control strategy for a particular plant
began with a review of the ambient impact near the plant. , First, the point
with the highest ambient impact around the plant was found. This impact was
then apportioned between the various sources at the plant based on dispersion
modeling results or other source apportionment studies. The result was a set
of ambient impacts for individual sources within the plant. When a control
strategy is applied to a source, the percentage reduction in the ambient
impact from the source is the same as the percentage reduction in the emission
rate. For each NAAQS alternative, a target level was defined by subtracting
the mobile source background from the NAAQS. This target represents the level
below which the total plant impact would have to be reduced in order for the
NAAQS alternative to be attained. Mission reductions were then applied to
the individual source ambient impacts in order to identify the set of control
options that would reduce the total plant impact below the target level at the
lowest total cost. Costs used in this exercise were the net annualized
control costs. After the identification of the lowest-cost controls to attain
a particular alternative at the point of maximum baseline impact, controlled
57

-------
impacts were estimated for other receptors to ensure that the control strategy
would result in attainment at all points around the plant.
This general methodology was used both for individual plants in the
primary smelting and alkyl lead manufacturing industries, and for model plants
representing larger industries. After the identification of the least-cost
set of controls for a particular plant, control costs were summed for the
plant to obtain total capital and annualized costs of attainment.
4.7.2 Control Cost Aggregation
For the primary smelting and alkyl lead manufacture source categories,
total industry-wide costs were developed simply by summing costs for the
individual plants. For larger industries, costs of control were developed on
a model plant basis, and aggregate costs also were developed using a model
plant methodology.
The first step in developing aggregate costs for the larger industries
was to group the affected plants into classes by size. The total costs of
attainment for each size range were then calculated by multiplying the number
of affected plants in the range by the cost of attainment for the model plant
representing the range. The total costs for the various size ranges were then
summed to obtain an industry-vide cost.
4.8 NEW SOURCE COSTS
4.8.1 Industry Growth
For each of the industries identified as potentially affected under the
NAAQS alternatives, industry growth was estimated in order to determine
whether new sources will be built by 1995. Table 4-1 gives current capacity
utilization and growth rate estimates for these industries. The references
used to develop these estimates and the projection periods for the growth
rates are also listed.
It was assumed that when a source category reaches a capacity utilization
of 90 percent, new capacity will be added. Thus, new capacity is expected to
58


-------
TABLE 4-1. CURRENT CAPACITY UTILIZATION AND GROWTH RATES FOR INDUSTRIES
AFFECTED BY THE NAAQS ALTERNATIVES
S=SS=SS6SSSSCSBBBBBBBaSBBBSesCBSSSBBeBee8BBBBBaBesBaBaBBBSBSBSBBEBBBaeBSBBae=SBBeB3SSSSBSBSBS:
Current Estimated
capacity growth rate	References

utilization
(percent)
(percent/
year)a
Capacity
Growth
Projection
period
Primary lead smelting
83
1.0-2.5b
4
4,5
1978-1990
Secondary lead smeltins
50

4
4,5
1978-1990
Lead ore processing
68
<1.0
4,5
6
1979-1985
Lead-acid battery manufacture
75
3.0
7
7
1983-1988
Alkyl lead manufacture
NA
c
J

8
1983-1990
Gray iron production
68d
2-3
9
10
1985-1989
Alloy steel electric arc furnaces
NA
•
4.0®
—
11,12*
1982-1988
Steel foundry electric arc furnaces
68
2-3
9
10
1985-1989
Iron and steel sintering
68d
2-3d
9
10
1985-1989
Brass and bronze production
82
f
9
13
1985-1989
Primary copper 6melting
82
-1.0
9
10
1985-1989
aExcept for electric arc furnaces, growth rates are for product deliveries.
^The estimated increase in total lead consumption ranges from 1.0 to 2.5 percent per year. The
lower estimate (Reference 2) is the average increase in total lead consumption since 1970.
The higher estimate is the Bureau of Mines projection of the increase in lead consumption
through 1990 from a 1978 base.
Consumption of alkyl lead in the U.S. is expected to decrease by over 90 percent by 1990
(Section 3)j with the exact rate depending on the final lead phasedown rules. Foreign exports
of alkyl lead are also expected to decline.
^Average capacity utilization and growth for the iron and steel industries.
eGrowth in production capacity.
^A decline in brass and bronze production is expected over the next several years.

-------
be added to the lead-acid battery manufacturing industry beginning in 1990.
New electric arc furnaces for the alloy steel industry currently are being
built and industry projections indicate this growth will continue through the
1990s. In this study, it was assumed that the growth will continue through
1995 .
For the lead smelting industry, based on the low growth projections for
lead mining and the low capacity utilization for secondary lead production, it
was assumed that most of the projected increase in lead demand will be filled
by secondary lead smelters rather than by primary smelters. No new capacity
was projected for either industry based on current growth estimates.
4.8.2 Cost Estimates
For both lead-acid battery manufacture and alloy steel electric arc
furnaces, capacity is expected by industry sources to be added by construction
of new large facilities rather than by expansion of existing facilities.7
This is also assumed to be the case for gray iron production and brass and
bronze production. In all cases, controls required by NSPS will be sufficient
for attainment of all of the NAAQS alternatives. Thus, the costs of
attainment for new sources are expected to be attributable entirely to NSPS.
60

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
REFERENCES FOR SECTION 4
1980 Census of Population, Volume 1—Characteristics of the Population,
Chapter A—Number of Inhabitants, Parts 1 to 52 (PC70-1-A). U.S.
Department of Commerce, Washington, D.C. 1982.
Compilation of Air Pollutant Emission Factors, Third Edition, and
Supplements 1 through 14 (AP-42). U.S. Environmental Protection Agency,
Research Triangle Park, N.C. 1977 to 1983.
Control Techniques for Lead Emissions (EPA-450/2-77-012). U.S.
Environmental Protection Agency, Research Triangle Park, N.C. 1977.
Mineral Commodity Summaries - 1983. U.S. Bureau of Mines. January 1983.
The International Competitiveness of the U.S. Non-Ferrous Smelting
Industry and the Clean Air Act. American Mining Congress. April 1982.
Engineering & Mining Journal. March 1982 and January 1983.
Telecon. W. Battye, GCA/Techology Division, with R. Burkhard, Battery
Council International Marketing Committee. April 16, 1983. Growth in
the lead-acid battery industry.
Regulation of Fuels and Fuel Additives. Federal Register.
47:49322-49334. October 29, 1982.
Federal Reserve Statistical Release: Capacity Utilization:
Manufacturing, Mining, Utilities, and Industrial Materials. G.3(402) The
Federal Reserve System, Washington, D.C. June 17, 1985.
1985 Industrial Outlook. U.S. Department of Commerce. January 1985.
Electric Arc Furnaces and Argon—Oxygen Decarburization for Vessels in
Steel Industry - Background Information for Proposed Revision to
Standards (EPA-450/3-82-020a). U.S. Environmental Protection Agency,
Research Triangle Park, N.C. 1983.
EPRI Journal. Electrical Power Research Institute. June 1982.
Telecon. Michael Clowers, GCA/Technology Division, with the Brass and
Bronze Ingot Institute, April 10, 19&5.
61
¦>

-------
5. INDUSTRY-SPECIFIC STUDY METHODS AND STUDY RESULTS
The general study methodology was described in the previous section.
Because of differences in available data, however, the study methods differed
between the various lead-emitting industries. This section details the
industry-specific methods used and presents the study results for each
affected industry. For each industry, a description is given of processing
techniques used and the competitive structure. The assumed baseline controls
and the methods of determining baseline ambient impacts are detailed. The
controls required to meet each alternative NAAQS (or to approach it as closely
as possible) are then described and costs are presented for the NAAQS
controls. Finally, the total NAAQS costs for each industry are aggregated to
obtain nationwide NAAQS costs.
5.1 PRIMARY LEAD SMELTING AND REFINING	. .. .
5.1.1 Process Description^
The primary smelting of lead ore concentrates (SIC 3332) produces lead
metal and other metal byproducts. The four basic production processes of
primary smelting are: sintering; reduction; drossing; and refining.
Sintering - Lead ore concentrates must be agglomerated and desulfurized
prior to being used as the primary feed material in the reduction process.
Agglomeration of the lead ore concentrates occurs in bedding or blending
plants where the concentrate is mixed with sulfur-free fluxes, such as lime,
silica, and iron-bearing materials, and pelletized. This pelletized material
or sinter charge is loaded onto pallets that are attached to a continuous
conveyor belt system in a sinter machine. Heated draft air is forced through
the moving pallets, which are perforated and slotted to accommodate this
process. The sinter charge is heated to approximately 1000°C (1800°F), which
is just below its melting point. Oxidation of the sinter charge occurs at
62

-------
this temperature and sulfur dioxide is released. Properly sized sinter is
then sent to the blast furnace for reduction; undersized sinter, which usually
results from insufficient desulfurization, is recycled into the sinter process.
Reduction - The blast furnace used in the reduction process is a water
jacketed shaft furnace supported by a refractory base. Pressurized combustion
air is injected into the furnace through evenly spaced tuyeres that are
located near the furnace base.
The material mix that is charged into the blast furnace consists of sinter
(80 to 90 percent of total charge) metallurgical coke (8 to 14 percent), and
other charge and recycled material. The combustion of the metallurgical coke
supplies the necessary carbon monoxide and heat to reduce the sinter to
bullion lead and byproducts. The byproducts includes speiss (arsenic and
antimony compounds); matte (copper sulfide and other metal sulfides); and slag
(silicates). These materials are tapped off from the bullion lead for further
processing and metal recovery.
Drossing - The bullion lead that is tapped from the blast furnace is
cooled to approximately 370° to 430°C (700° to 800°F) in dross kettles. In
this process, impurities such as copper, zinc, arsenic, and antimony are
precipitated according to their solubility limits and collected on the surface
of bullion lead as a dross. Sulfur bearing materials, zinc, or aluminum may
be added to the drossed bullion to further reduce its copper content.
Refining - In the refining process, bullion lead is cast in iron kettles
and the following refining operations occur:
•	zinc combines with gold and silver to form an insoluble intermetallic
substance at operating temperatures (Parke's Process);
•	zinc is removed under vacuum;
•	calcium and magnesium are added to form an insoluble compound with
bismuth (Betterson Process), which is skimmed off; and
•	sodium hydroxide and sodium nitrate are added for the removal of
remaining traces of metal impurities.
63

-------
The refined lead produced is of 99.990 to 99.999 percent purity.
5.1.2	Industry Structure^
Currently, there are five domestic primary lead smelting/refining
complexes and one lead refinery. The annual primary lead capacity of these
complexes is 595,000 Mg (656,000 tons), with individual plant capacities
ranging from 76,000 to 209,000 Mg (84,000 to 230,000 tons). Primary lead
production accounted for slightly less than 50 percent of total domestic lead
production for 1982 and this percentage should decline to around 40 percent by
the year 2000 as secondary lead smelters increase their market share (U.S.
Bureau of Mines projections). These projections would change if the Bunker
Hill smelter in Kellogg, Idaho, were to be restarted. This smelting/refining
complex was closed in 1982 and would need extensive retrofit before resuming
any primary lead production.^ For this industry, it is estimated that no new
capacity will be required through 1990. This estimate is based on the low
growth projection (less than one percent per year) for the lead ore processing
industry, which provides the basic feed material for primary lead smelters,
and the low capacity utilization rate of 50 percent for secondary lead
smelters, which leaves a large quantity of unused capacity to meet any
increase in total lead demand. The five domestic primary lead smelters are
presently operating at 83 percent of capacity and may exceed 90 percent of
capacity during future peak demand periods.^>3
5.1.3	Emission Sources and Control
Lead-bearing particulate is emitted from process, fugitive process, and
fugitive dust sources at primary lead smelters and refineries. Process
emissions, such as those generated in furnace smelting operations, are well
controlled through the use of extensive hooding and ducting systems, high
efficiency baghouses, and dispersion of emissions from high stacks. The
fugitive emission control techniques that are utilized in the primary lead
smelting and related industries includes

64

-------
•	hooding and ducting of fugitive process emissions to baghouses or
scrubbers;
•	paving, traffic control, wet suppression and soil stabilization for
unpaved haul roads and general open work areas; and
•	wet suppression, surfactants, and enclosures for stockpiles and
conveyor systems.
Due to the well-controlled nature of process emission sources, fugitive
process and fugitive dust emission sources were the primary sources evaluated
for the installation of various control techniques to achieve the alternative
lead NAAQS levels. The baseline emissions and control techniques used in this
evaluation were those that currently exist at each plant.
5.1.4 Industry Specific Methods
5.1.4.1 Estimation of Baseline Ambient Impacts
For the primary lead smelting and refining industry, the baseline ambient
impact was estimated for each lead source at each individual plant. Lead
emission parameters used in the analysis of this industry were based on data
obtained through contacts in late 1983 with State and federal environmental
agencies and plant personnel.4~H Source apportionment studies have been made
for two of the primary lead smelters in the development of State
implementation plans for the current lead NAAQS.9,10 These studies used
chemical and particle size analyses of ambient monitor samples and emission
samples to apportion ambient lead impacts between the individual sources at
the smelters. The results of these source apportionments were used in the
present study as baseline ambient impacts for individual sources.
For the other three smelters and the primary refinery, baseline ambient
impacts were estimated using atmospheric dispersion modeling. The COMPLEX I
model was used for point sources, and the ISCST model was used for area
sources. Terrain effects were taken into account for point sources, but no
dispersion models were identified which could include terrain effects for area
sources. In each case, the meteorological data input to the models were the
most recent available preprocessed data from the closest meteorological
65

-------
station. The maximum distance from any smelter to the nearest meteorological
station was about 100 miles. Plant site meteorological data would have been
preferable to the nearest-station data because of possible impacts of terrain
on meteorological parameters. Plant site data have been gathered in many
cases, but these data either are incomplete, or have only been processed for
long term dispersion models. Short term models were used in this study in
order to allow the assessment of short term exposures as well as long term
exposures. Additional details on the general modeling protocols used in this
study are give in Section 4.
Ambient monitoring data are available for two of the three modeled
smelters and for the primary refinery. For one of the smelters and the
refinery, these data were in agreement with the dispersion modeling results.
In the case of the other smelter, the concentration predicted by dispersion
modeling at the plant boundary was about a factor of 2.5 higher than the
concentration measured at a monitoring site at the same location. For this
plant, the monitored concentration was taken as the total plantwide baseline
ambient impact. The monitored concentration was apportioned between sources
within the plant based on the modeling results to obtain baseline impacts for
the individual sources at the smelter. (This was done for each source by
multiplying the monitored plant boundary concentration by the ratio of the
predicted ambient impact for the source at the boundary to the total ambient
impact for the plant at the boundary.)
5.1.4.2 Mobile Source Background Concentration
In the two smelter source apportionment studies discussed above,
background lead concentrations from mobile sources were measured at 0.36 ug/m^
and 0.20 ug/m^. Both of these smelters are located in small cities. The lead
refinery is also located in a small city, and is expected to be subject to a
similar mobile source background. When the lead phasedown program is taken
into account (using the methods given in Chapter 3), predicted mobile source
backgrounds in 1990 and 1995 are less than 0.03 ug/m^ for these three
sources. The concentration is negligible in comparison with the ambient
impact of even a well-controlled lead smelter or refinery. Mobile source
backgrounds around the remaining three smelters in 1990 and 1995 are also
T
66

-------
assumed to be negligible because of their rural locations. Thus, in the
identification of necessary emission reductions around primary smelters, the
mobile source background was taken to be zero in all cases.
5.1.5 NAAQS Control Strategies and Costs
Using the results from the calculation of the baseline ambient impacts and
background ambient lead concentrations, individual plants were identified
which would be affected under the various lead NAAQS alternatives. A plant
was determined to be affected under a given alternative if the sum of the
baseline ambient impact and background ambient lead concentration from mobile
sources at any point outside the plant boundary exceeded the NAAQS level.
Control techniques that were evaluated for emissions from fugitive process
and fugitive dust sources included:
•	additional hooding, ducting, and ventilation systems for product
handling, sinter machines, blast furnaces, dross kettles, and
refining operations;
•	replacement of existing control systems with more efficient systems;
and
•	fugitive dust control systems, such as paving and wet suppression.
In addition, the construction of new stacks or the extension of existing
stacks were considered in several cases. The height increases considered were
within the definition of good engineering practice. The higher stack allowed
for better dispersion of primary lead plant emissions.* Acquisition of land
around the smelters was also considered as a possible means of attaining NAAQS
alternatives. Because a policy on it is under review within EPA, and because
of the location of public roadways or urban areas near plant property, land
acquisition was judged not to be a feasible means of compliance.
*EPA is currently revising its stack height regulations in response to a court
ruling (Sierra Club vs. EPA, 719.2d 436 (D.C. Cir. 1983), cert. denied, 52
U.S.L.W. 3929 (U.S. July 2, 1984)). Credit in dispersion modeling for stack
height increases, even up to good engineering practice may be affected by the
forthcoming EPA rulemaking on this matter. Estimated costs of attainment for
primary smelters should be interpreted in light of this possibility.
67

-------
Based on results of dispersion modeling, control strategies were developed
for each of the six domestic primary lead plants. The control strategies
include the control techniques and related costs that would be necessary for
each smelter to attain each of the lead NAAQS alternatives under
consideration. These controls are in addition to existing emission control
techniques. The costs associated with the control techniques evaluated are
presented in Appendix D. In addition, the total industry capital and
annualized control costs associated with achieving each lead NAAQS alternative
are presented in Table 5-1.
No control strategies were identified which showed attainment of the
lowest NAAQS levels all six of the primary lead plants. Given the
uncertainties in emissions estimates and dispersion modeling, it is possible
that the plants would attain the lower NAAQS levels using the identified
controls. In addition, other controls are available which were not assessed
in this study because of difficulties in ascertaining their effectiveness.
5.2 SECONDARY LEAD SMELTERS
5.2.1 Process Description
Secondary lead smelters (SIC 3341) process a variety of lead-bearing
material and scrap to produce lead and lead alloy ingots, lead oxide, and lead
pigments. The basic production processes at a secondary lead smelter are:
scrap treatment; smelting; and refining/casting.
Scrap Treatment - In this process, metal and nonmetal contaminants are
removed from the lead-bearing material and scrap. This may include such
operations as battery breaking, crushing, and sweating. Sweating separates
lead from high-melting metals by heating the scrap in direct-fired rotary or
reverberatory furnaces. The partially purified lead is periodically tapped
from these furnaces for further processing smelting or pot furnaces.
Smelting - Smelting results in the production of purified lead by melting
and separating lead from metal and nonmetal impurities and by reducing oxides
68


-------
TABLE 5-1. NATIONWIDE NAAQS COSTS FOR PRIMARY LEAD SMELTERS AND REFINERIES3
Lead NAAQS
Alternatives (ug/m3)
Costs (millions)
Installed Total
capital annual
Number
of plants
affected
Number
unable to
comply^
0.25
53.3
14.5
6
6
0.5
53.3
14.5
6
5
0.8
53.3
14.5
6
4
1.0
46.6
12.6
6
4
1.5
44.5
11.9
6
2
2.0
27.3
6.7
6
0
2.5
17.8
4.5
5
0
aBackground ambient lead level for all
six plants are assumed
to be negligible.
^Controls could not be
with some of the NAAQS
identified which
alternatives.
would allow
some smelters to comply
69

-------
to elemental lead. The two types of furnaces used for smelting are
reverberatory and blast. The reverberatory smelting furnace produces a
semi-soft lead product containing 3 to 4 percent antimony. Blast furnaces
produce hard or antimonial lead, which contains 10 percent antimony. As with
other smelting and melting production processes, these furnaces are charged
with a variety of material, including lead scrap, recycled lead dust, fluxes
(such as limestone), and coke. The charge material is heated by combustion
air or gases and the slag is tapped off. The molten lead tapped from the
furnace is put into a holding pot and then cast into large ingots.
Refining/Casting - The refining/casting production process consists of the
following operations:
•	Remelting - Materials charged for remelting are usually lead alloys
which require no further processing before casting;
•	Alloying - Alloying furnaces simply melt and mix lead ingot with
alloying materials. Common lead alloying materials include antimony,
arsenic, copper, nickel, and tin;
•	Refining - Refining furnaces remove antimony and copper from lead
ingots to produce soft lead or remove arsenic, copper, and nickel to
produce hard lead. Various compounds may be added or injected during
the refining operation to form drosses, which are skimmed off. In
addition, sawdust may be added to separate lead from the dross and to
reduce some of the lead oxide to elemental lead; and
•	Oxidizing - Oxidizing furnaces are either kettle or reverberatory
furnaces which oxidize lead and entrain the product in the combustion
air stream. The product is recovered in high efficiency baghouses.
5.2.2 Industry Description^-
There are approximately 49 domestic secondary lead smelters which account
for slightly over half of total domestic lead production.12 These plants
are located primarily near large urbanized areas due to the supply of scrap
lead (mostly used automobile and truck batteries) and the proximity of their
main customers (lead-acid battery manufacturers). The basic industry trend is
toward larger, more efficient secondary lead smelters, and many smaller, less
efficient operations have been phased out during the last decade. Based on
70

-------
U.S. Bureau of Mines estimates, secondary lead smelters, which are operating
at a capacity utilization rate of approximately 50 percent, have adequate
existing unused capacity to meet the projected increase in total lead
demand.This should allow secondary lead smelters to increase their market
share of lead supply from just over 50 percent presently to nearly 60 percent
by 2000.
5.2.3	Emission Sources and Controls^
Lead bearing particulate matter is emitted from process, fugitive process,
and fugitive dust emission sources at secondary lead smelters.13 Process
emissions, which are generated during the smelting/refining processes
described above, are well-controlled through the use of extensive hooding and
ducting systems, high efficiency baghouses and scrubbers, and dispersion of
emissions through high stacks. Fugitive emissions result from incomplete
capture of process lead emissions, storage and handling of lead-bearing
materials, and reentrainment of dust from plant property due to wind or
vehicle traffic. Most secondary smelters are equipped with capture systems
and control devices for fugitive emissions from furnaces. Also, in order to
meet OSHA regulations, housekeeping measures are usually taken to minimize
lead dust levels.13,14
5.2.4	Industry Specific Methodologies
For this study, it was assumed that baghouses are used to control process
emissions. Also, it was assumed that most process fugitive emissions are
captured by hoods and ducted to baghouses. The lead emission sources
considered in this study include:
•	controlled process emissions,
•	process fugitive emissions escaping capture systems, and
•	battery breaking emissions.
71

-------
Controlled emission factors for these sources are given in Table 5-2.13,15
These emission factors were used instead of plant-specific emission data,
because plant specific data are not available for fugitive emissions.
Table 5-2 also presents average source-receptor coefficients for the above
emission sources. These were obtained using dispersion modeling results from
the urbanized area exposure assessment. Dispersion modeling was conducted for
secondary smelters in the following metropolitan area:
•	Chicago, Illinois;
•	Dallas, Texas;
•	Minneapolis- St. Paul, Minnesota; and
•	Tampa, Florida.
As noted in Section 4, the source-receptor coefficient is defined as the ratio
between the maximum ambient impact and the emission rate.
The baseline ambient impact was calculated for each smelter by summing the
baseline ambient impacts for the individual sources within the smelter. The
baseline impact for a given source was obtained by multiplying the production
rate by the product of the emission factor and the source-receptor coefficient
from Table 5-2. Smelter production rates were obtained (late 1983) from the
Bureau of Mines and from the EPA National Emissions Data System.^
Because most secondary lead smelters are located in urban areas, the urban
mobile source background concentrations derived in Section 3 were used. These
concentrations, 0.01 and 0.05 ug/m^, bracket the range of potential average
urban mobile source impacts for the two lead-in-gasoline phasedown programs
currently under study. Costs of compliance for secondary lead smelters were
estimated for each of these background levels.
It was assumed that the maximum ambient impact occurs outside the plant
boundary. For the secondary lead smelting industry, most of the plants are
located in urban areas where large tracts of property are extensive and
difficult to obtain. These plants were assumed to own property to a distance
no greater than at most 150 meters aways from the actual emission sources.
Because the maximum lead impacts predicted by dispersion modeling occur at
72

-------
TABLE 5-2.
CONTROLLED EMISSION FACTORS AND
SOURCE-RECEPTOR

COEFFICIENTS FOR SECONDARY LEAD
SMELTING

Baseline


emission
Average source-

factor
receptor coefficient
Source
(g/Mg)a
ug/mJ per Mg/yr
Controlled furnace
emissions 200
0.006
Fugitive emissions


Charging area
28.1
/ \
Tapping area
19.A
I )
General furnace^
18.8
\ °" /
Battery breaking
26
\ /
Reference 15.
^General furnace fugitive emissions were apportioned 50 percent to
charging and 50 percent to tapping.
73

-------
distances of 200-300 meters even for low lying sources, these maxima generally
will be outside plant property.
Using the baseline ambient impacts calculated and background ambient lead
concentrations, individual plants were identified which would be affected
under the various lead NAAQS alternatives. For the secondary lead smelting
industry, each plant was studied to determine its status with respect to the
various NAAQS alternatives. In general, a plant was determined to be affected
under a given alternative if the sum of the baseline ambient lead impact and
the background concentration from mobile sources around the plant exceeded the
alternative.
The major source of uncertainty in the analysis of secondary lead smelters
was the selection of baseline control levels and the identification of
baseline emission factors. Emission factors for fugitive sources were based
on tests conducted at a single plant. Fugitive emission controls at
individual smelters may be much more effective or less effective than the
controls used at the tested facility. In addition, fugitive emission factors
could not be obtained for some sources, and these sources were not included in
this study. Alternatively, some of the sources considered in this study, such
as battery breaking, probably are not present at some smelters. Because
fugitive emissions are not included in the National Emissions Data System, and
because of the size of the secondary smelting source category, it was not
feasible to make an in-depth study of the fugitive sources at each plant.
Another source of uncertainty for the secondary smelting category is the
selection of 150 meters as the typical distance to the plant boundary. In
some cases, secondary smelters may own more property around the plant. As
noted above, the point of maximum ambient impact from secondary lead smelters
is 200 to 300 meters from the emission sources. Thus, for smelters owning
property more than 200 to 300 meters from the plant, attainment of the NAAQS
alternatives would be less costly than this study has indicated.
5.2.5 NAAQS Controls and Costs
NAAQS controls assessed for the secondary lead smelting industry include:
-j
74

-------
•	building enclosure and process modifications for battery breaking;
•	enclosure and process modification for furnace charging; and
•	additional hood and process modifications for metal and slag tapping.
Table 5-3 gives industry-wide costs required to achieve the various lead NAAQS
alternatives. Appendix D presents data on emissions and costs used to develop
the values given in Table 5-3. Control strategies could not be identified
which would show attainment of the lower alternatives for up to 49 smelters.
Given the uncertainties of the study methodology, however, it is possible that
some of these smelters could attain the lower levels with the identified
controls. Also, controls are available which are not assessed because of
difficulties in ascertaining their effectiveness. These include work
practices for furnace charging and tapping, and other work practices.
5.3 LEAD ORE PROCESSING17>18>19
5.3.1 Process Description
Ore processing at lead mines (SIC 1031) comprises a two stage process; the
first stage being the dry process operations of crushing and grinding of
stockpiled ore, and the second stage of the wet process operation of
flotation. In the dry process operations, stockpiled ore (such as galena and
other lead-bearing ores) is size reduced by jaw and gyratory crushers in
combination with grizzly bars, screens, and conveyors which feed into fine ore
bins. Some mining operations with centrally located vertical shafts employ
underground crushers for the initial size reduction step. Grinding of the
size reduced ore is accomplished through the use of rod mills, which have less
tendency than other grinding methods to overgrind the soft and friable
galena. Ball mills, in closed circuit with hydrocyclones, are used for fine
grinding, which prepares the ore for the flotation process. Floatation is the
primary method used for the concentration of lead-bearing ores. Depending on
the mineralogy of the ore, two or more floatation cells may be required to
75

-------
TABLE 5-3. NATIONWIDE NAAQS COSTS FOR SECONDARY LEAD SMELTERS
(OCTOBER 1984 DOLLARS)	\
Background and
NAAQS
level (ug/m3)
Costs (millions)
Installed
capital
Total
annual
Number
of plants
affected
Number
unable to
comply^
Background = 0.05
0.25	144.0	40.8	49	40
0.5	133.3	38.5	44	38
0.8	124.4	36.8	41	11
1.0	118.7	35.5	39	9
1.5	96.1	28.9	31	3
2.0	72.4	22.0	28	0
2.5	54.2	16.4	18	0
Background = 0.01
0.25	140.3	40.1	49	38
0.5	132.7	38.3	41	33
0.8	123.4	36.5	40	11
1.0	118.6	35.4	38	8
1.5	95.7	28.7	29	2
2.0	70.3	21.2	23	0
2.5	54.2	16.4	18	0
aControls could not be identified that would allow some smelters to comply
with some NAAQS alternatives.
76
i

-------
achieve the designed percent of lead in the concentrate. The concentrate is
then shipped to one of the five domestic primary lead smelters for further
processing.
5.3.2	Industry Structure^' 18,19,20,21
Currently, there are thirteen domestic lead ore processing plants, located
primarily in the New Lead Belt of Missouri. Based on a capacity utilization
rate of 68 percent for this industry and a projected increase in total lead
consumption of less than one percent per year, no new lead ore processing
plants are predicted through 1995. Selected expansions of existing plants may
occur during this period; however, two recent expansion projects have been
abandoned due to technical and economic factors.20
5.3.3	Emission Sources and Controls
Particulate lead emissions are generated from both fugitive dust sources
and fugitive process sources at lead ore processing operations. The fugitive
dust sources include reentrainment from haul roads and lead ore stockpiles,
and wind blown dust from open work areas and conveyors. The fugitive process
sources consist of the crushing and grinding operations involved in size
reduction of the lead-bearing ores, including conveyor transfer points and dry
product loading. To control these particulate emissions, lead ore processing
plants currently use wet dust systems and medium energy wet scrubbers for
fugitive process sources.Due to the wet nature of the process operations
involved, little if any particulate lead emissions are generated during
flotation.
5.3.4	Industry-Specific Methodologies
For the lead ore processing industry, it was assumed that process sources
at these plants (crushers, grinders, screens, and storage hoppers) were
controlled through the use of hooding and ducting systems and medium energy
77

-------
(1.5 kPa) wet scrubbers (about 85 percent efficiency). The fugitive emission
sources at ore processing plants (i.e. lead ore storage piles) were assumed to
be uncontrolled. Emission factors for process and fugitive emission sources
are presented in Table 5-4.18,22 These represent the level of control
described above. These emission factors, along with published stack
parameters for two model sizes (270 and 540 megagrams per hour throughputs),
were used in the dispersion modeling analysis.
Table 5-4 also presents the baseline ambient impacts for the above
emission sources and model plant sizes. These were obtained from dispersion
modeling results for the two model plant sizes using St. Louis, Missouri,
meteorological data. The St. Louis data were used for this industry because
it was assumed to be the most representative available meteorological data for
the New Lead Belt area of Missouri, where the majority of domestic lead ore
processing plants are located. The concentrations presented in Table 5-4 are
for the plant boundary, which is assumed to be one kilometer from the
processing plant. Because of the rural location of lead ore processing
plants, the background lead concentration from mobile sources was assumed to
be negligible.
Using the baseline ambient impacts calculated for the model plants, each
of the 13 domestic lead ore processing plants were grouped into one of the
two sets, represented by the model plant sizes, to determine individual plant
impacts. Five of the existing ore processing facilities correspond roughly to
the 270 Mg/day model plant, while the remaining eight correspond to the
540 Mg/day model plant. Control strategies were developed to bring the two
groups into compliance with the lead NAAQS alternatives.
5.3.5 NAAQS Control and Costs
Based on emission factors for uncontrolled area sources (lead ore
stockpiles) and plant process sources controlled by medium energy (1.5 kPa)
wet scrubbers, dispersion modeling indicates that the lead ore processing
model plants would exceed the various ambient lead concentrations under
consideration. The addition of fugitive dust controls at these operations
78

-------
TABLE 5-4. BASELINE EMISSIONS AND AMBIENT IMPACTS
FROM LEAD ORE PROCESSING
Source
Baseline
emissions
(g/sec)a
Baseline
ambient
impact,
(pg/m3)
270 Mg/hr plant


Process sources0
2.32
1.6
Storage piles
0.0005
—
540 Mg/hr plant


Q
Process sources
2.74
1.7
Storage piles
0.0007
—
Based on particulate emission factors from Reference 22 and lead
ore compositions from Reference 18.
Based on dispersion modeling for the two model plants with St.
Louis, Missouri meteorological data. Concentrations presented
are for the plant boundary, assumed to be 1 km away from the
processing plant.
c
Baseline control for all process sources was assumed to be
medium energy scrubbers.
¦j
79

-------
(i.e. wet dust suppression systems) would have an insignificant impact on
ambient lead concentrations. Thus, the required reductions must be
accomplished through the increased control of plant process emission sources.
Control costs were developed for the retrofit of high energy (3.75 kPa)
venturi scrubbers at existing lead ore processing plants. These control
devices will allow the plants to achieve all levels of the ambient lead
concentrations under consideration.
Appendix D gives the costs of attaining the lead NAAQS alternative for
individual plants. All of the plants are expected to be affected by NAAQS
less than or equal to the current standard. Table 5-5 gives industry-wide
costs for lead ore processing.
Control strategies could not be identified that would show attainment of
the lower NAAQS alternatives for lead ore processing plants. Given the
uncertainties of the study methodology, however, it is possible that some of
the plants could attain the lower levels.
5.4 LEAD-ACID BATTERY MANUFACTURE
5.4.1 Production Processes.!!
The manufacture of lead-acid batteries (SIC 3691) begins with the casting
of lead grids and the production of lead oxide powder. The lead oxide is
produced by oxidizing either molten lead or lead ingots in air. The oxide is
mixed with sulfuric acid and water to produce a paste. Pasting machines are
used to force the paste into the interstices of lead grids. The pasted grids,
called plates, are cured and then sent to a three process operation comprising
plate stacking, burning, and element assembly. Plates are stacked in an
alternating positive and negative block formation, with insulators between
them, to produce elements. Then, in the burning operation, leads are welded
to tabs on each positive or negative plate. An alternative to this operation
is the cast-on-strap process, in which molten lead is poured around the plate
tabs to form the connection, and positive and negative terminals are then
welded to each such connected element. The completed elements are assembled
in battery cases either before of after the formation step. In formation, the
80

-------
TABLE
5-
5.
NAAQS CONTROLS COSTS FOR THE LEAD ORE PROCESSING
INDUSTRY (OCTOBER 1984 DOLLARS)3
Lead NAAQS	Costs (millions)	Number	Number
alternative	Installed	Total	of plants	unable to
(ug/m^	capital	annual	impacted	comply*5
0.25	18.3 7.3	13	13
0.5	18.3 7.3	13	13
0.8	18.3 7.3	13	8
1.0	18.3 7.3	13	0
1.5	18.3 7.3	13	0
2.0	0	0
2.5	0	0
aThe background ambient lead level for lead ore processing facilities is
assumed to be negligible.
^Controls could not be indentified that would allow the lead ore processing
facilities to comply with some NAAQS alternatives.
81

-------
plates are immersed in dilute sulfuric acid and connected to a direct current
source. Formation can be done either in the battery cases (wet formation) or
in an open tank (dry formation).
Defective parts are either reclaimed at the battery plant or are sent to a
secondary lead smelter. Lead reclamation facilities at battery plants
generally are small pot furnaces. Approximately 1 percent of the lead
processed at a typical lead acid battery plant is recycled through the
reclamation operation.
5.4.2	Industry Structure
Lead acid storage batteries are produced in many sizes, but the majority
is produced for use in automobiles and fall into a standard size range. A
standard battery contains about 11.8 kg (26 lb) of lead, of which about half
is present in the lead grids and half in the lead oxide paste.
Lead acid storage battery plants range in production capacity from less
than 500 batteries per day to about 10,000 batteries per day (bpd). In 1979,
it was estimated that there were 190 lead-acid battery plants in the U.S. of
which fewer than 100 had the capacity to produce over 500 bpd.23 Since that
time, because of a decline in battery demand, many plants have been closed.
The average capacity utilization is only about 75 percent for the companies
supplying most of the batteries sold in North America (eight companies account
for about 95 percent of sales). Optimum capacity utilization is about
90 percent.^
Battery demand is expected to increase by less than 3 percent per year.^
Thus, new capacity is not expected to be required until 1990. Industry
sources expect new capacity to be added through construction of new plants in
the 10,000 bpd range rather than through expansion of existing plants.^
5.4.3	Emission Sources and Controls^
Lead oxide emissions result from the discharge of air used in the lead
oxide production process. In addition, lead-bearing particulate matter is
generated in the grid casting, paste mixing, lead reclamation, and
82

-------
three process operations. These particulates are usually collected by
ventilation systems to reduce employee exposure to airborne lead.
Fabric filters are incorporated into all lead oxide production operations
for product recovery. Fabric filtration is also often used to control
emissions from the three-process operation. The paste mixing operation
consists of two phases. The first, in which dry ingredients are charged to
the mixer, results in emissions of lead oxide and usually is vented to a
baghouse. For the second phase of the cycle, when moisture is present in the
exhaust stream, the paste mixer is generally vented to an impingment scrubber;
however fabric filtration can also be used for this phase. Grid casting
machines are sometimes vented to an impingement scrubber. Lead reclamation
facilities are generally also vented to impingmement scrubbers
5.4.4 Industry-Specific Methods
5.4.4.1 Assessment of Baseline Control Status and Emission Rates
The emission controls used for battery production operations vary widely
from plant to plant. This variation results mainly from the variation from
plant to plant in the design of the battery production operations themselves.
Because lead-acid battery plants generally are not major sources of any
criteria pollutants other than lead, only about one third of them are listed
in the National Emissions Data System (NEDS).l^ Thus, data were not available
to assess the control status of each lead-acid battery plant. However, from a
survey of the plants included in NEDS, it was possible to assess the
frequencies of control and typical control efficiencies for various
operations. Based on the sources included in NEDS, it was estimated that all
lead reclamation facilities were controlled with an average efficiency of
99 percent; 83 percent of paste mixing facilities were controlled with an
average efficiency (for controlled facilities) of 97.7 percent; 79 percent of
three-process operation facilities are controlled with an average efficiency
of 98.5 percent; and 55 percent of grid casting facilities were controlled
with a an average efficiency of about 95 percent. As noted above, lead oxide
83


-------
TABLE 5-6. ESTIMATED BREAKDOWN OF BATTERY PLANTS
BY SIZE RANGES AND CONTROL STATUS1
Facilities with
£
controls

Estimated number of plants

Smallb
MediumC
Large^
PbO; Rec

5
7
4
PbO; Rec; PM

1
2
1
PbO; Rec; PM;
3-P
7
10
6
PbO; Rec; PM;
3-P; GC
15
23
13
Total

28
42
24
"PbO" refers to lead oxide production; "Rec" to lead reclamation; "PM" to
paste mixing; "3-P" to the three-process operation; and "GC" to grid
casting.
400-1,000 batteries per day (bpd),
:1,000-3,500 bpd.
*0ver 3,500 bpd.
84

-------
TABLE 5-7. EMISSION FACTORS
FOR LEAD-ACID BATTERY MANUFACTURE

Emission factors
(g/battery)

uncontrolled
controlled
Grid casting
0.35
__a
Paste mixing
1.13
0.026
Lead oxide production^
—
0.05
Three process operation
4.79
0.072
Lead reclamationc
0.63
0.006
aControl of grid casting was not found to be required to meet any of the
NAAQS alternatives.
^Lead oxide production facilities are equipped with baghouses for product
recovery.
cIn a survey of NEDS, lead reclamation was found to be controlled in all
cases.
85

-------
production incorporates fabric filtration for product recovery. No
correlation was found in the survey of NEDS between control status and
facility size.
Table 5-6 gives the assumed baseline control status of the lead acid
battery manufacturing industry. Existing plants were divided into twelve
categories based on their size and the control status. It was assumed that
the frequency of control for the facilities listed in NEDS is characteristic
of the industry as a whole. Thus, lead reclamation and lead oxide production
were taken to be controlled in all cases. It was also assumed that, at a
given plant, paste mixing is controlled first, followed by the three-process
operation, and finally grid casting. This assumption was based on the results
of the NEDS survey and on the cost-effectiveness of controlling these
facilities
Table 5-7 gives the uncontrolled and controlled emission factors used in
this study to determine baseline emissions. The uncontrolled values are
published EPA emission factor.^5 The controlled factors are based on the
uncontrolled factors, and the average control efficiencies detected in the
NEDS survey.
5.4.4.2 Identification of Affected Sources
Dispersion modeling for the urbanized area exposure assessment indicated
source-receptor coefficients for lead-acid battery manufacture ranging from
0.089 ug/m^ to 0.51 ug/m^. In calculating the source-receptor coefficients,
it was assumed that the maximum ambient impact from a battery plant occurs
outside the plant boundary. The location of the maximum impact ranged from
200 to 300 meters from the emission sources. Because most lead-acid battery
plants are located in urban areas, where land is difficult and expensive to
obtain, it was assumed that plants generally own land to at most 150 meters
away from the plant buildings. Dispersion modeling was conducted for
lead-acid battery plants in the following urbanized areas:
•	Chicago, Illinois;
•	Dallas-Ft. Worth, Texas;
86

-------
•	Nashville, Tennessee;
•	Philadelphia, Pennsylvania; and
•	Tampa, Florida.
The stack parameters used in the dispersion modeling (emission height,
velocity, and temperature) were obtained from the NEDS file^ and published
sources.^»23 These were not dependent on source type of size. Thus, the
wide range in source-receptor coefficients for lead-acid battery plants
results mainly from differences between meteorological conditions in the above
urbanized areas. The source-receptor coefficients were not distributed
normally or log-normally, but were instead uniformly distributed between the
observed minimum and maximum values.
Each of the 12 plant groups shown in Table 5-6 were studied to determine
the number of plants which will be affected in the group under the NAAQS
alternatives. This was done by calculating the probability that a model plant
representative of this group will be affected, and multiplying by the number
of plants in the group. A size of 500 batteries per day (bpd) was chosen to
represent plants in the 0-1000 bpd size range, and sizes f 2000 and 6500 bpd,
respectively, were selected to represent the 1000-3500 bpd and 3500-and-over
ranges. Each of the four control status groupings were studied separately for
each size range. The methods used to estimate the numbers of affected plants
in each size and control-status group are detailed in Appendix E.
Host lead-acid battery plants are located in urban areas. Therefore, the
upper and lower limit urban mobile source backgrounds derived in Section 3
were used. These concentrations, 0.01 and 0.05 mg/rn^, bracket the range of
possible average urban mobile source impacts for the two lead-in-gasoline
phasedown alternatives. Costs of attainment for lead-acid battery plants were
estimated for both possible background levels..
87

-------
5,4.5 NAAQS Controls and Costs
Control strategies identified in this study to meet the NAAQS alternatives
were retrofit of fabric filters to uncontrolled paste mixing and three-process
operations. Control of grid casting was not found to be required for any of
the NAAQS alternatives. Also, small plants were not found to be affected
under any of the alternatives. Table 5-8 gives the total costs and the
numbers of affected plants for each of the NAAQS alternatives. Appendix D
presents cost data used to develop the figures presented in this table.
5-5 LEAD ALKYL MANUFACTURE
5.5.1 Production Processes^
Commercially, lead alkyl compounds are produced by the batch alkylation of
sodium-lead alloy (SIC 2869). The basic step in this process is the reaction
of sodium-lead alloy with an excess of ethyl chloride in autoclaves at 70°C to
75°C as described by the following equations
4 NaPb + 4C2H5C1—«-(C2H5)4—Pb + 4NaCl + 3Pb
This reaction also can be carried out continuously in a process developed by
duPont. Details of the continuous process are not available.
In the batch process, sodium-lead alloy is produced by combining molten
lead, virgin or recycled, with sodium in a 9-to-l ratio in an airtight alloy
pot. The molten alloy is solidified and flaked in an oil-cooled flaker. The
flaked alloy is discharged into the autoclaves. Ethyl chloride then is added,
and the liquids and solids are intimately mixed.
Initially the batch is heated to start the reaction; the exothermic
reaction is maintained at 70°C to 75°C by external cooling and ethyl chloride
reflux. When the reaction is complete, the autoclave is vented through a
condenser for excess ethyl chloride recovery, and the reaction mass is
discharged into stream stills for separation.
88
i

-------
TABLE 5-8. NATIONWIDE NAAQS COSTS FOR THE LEAD ACID BATTERY
MANUFACTURING (OCTOBER 1984 DOLLARS)2
Background
and NAAQS
level (ug/m^)
Costs (thousands)
Installed
capital
Total
annual
Number
of plants
affected
Background = 0.05
0.25
0.5
0.8
1.0
1.5
2.0
2.5
6,477
5171
4,082
3,502
2,128
2,085
1,553
1,862
1421
1,129
971
594
578
431
21
13
11
9
5
4
3
Background = 0.01
0.25
0.
0.
1.
1.
2.
2.
6260
4,881
3792
3502
2128
2085
1553
1781
1,342
1050
971
594
578
431
20
14
10
9
5
4
3
89

-------
From the reaction mixture containing NaCl, tetraethyl lead, unreacted
lead, and dissolved ethyl chloride, ethyl chloride is removed and collected in
a brine condenser system. A water-cooled condenser then is used to collect
water and the tetraethyl lead product. Non-condensible vapors are vented to
the atmosphere. The separation operation lasts for approximately two hours,
and the residue in the still then is sluiced to a sludge pit for lead recovery.
Collected tetraethyl lead and water is decanted, and then tetraethyl lead
is purified by air blowing or washing with dilute aqueous solutions of
oxidizing agents to oxidize bismuth compounds present in the raw lead. The
clean tetraethyl lead is filtered, and stabilizing agents are added. The
sludge, consisting of fine lead particles, water, dissolved salt, and traces
of tetraethyl lead, is leached with water in the sludge pit to separate lead
and salt. The sludge pit bottoms, consisting mostly of lead, are dried in an
indirect steam dryer. The dried sludge is fed to a gas-fired reverberatory
furnace to recover lead.
5.5.2	Industry Structure
Tetraethyl lead and tetramethyl lead are produced as primary anti-knock
gasoline additives in two manufacturing facilities operated by duPont and
Ethyl Corporation. As a result of decreased demand for gasoline additives,
the capacity for lead alkyl manufacture has decreased substantially in the
recent past. The number of lead alkyl plants in the U.S. has decreased from
six, in 1975, to two plants operating at present. The Ethyl plant is also
expected to remain in operation, at least in the near future, serving the
small U.S. market and also exporting some lead additives.
5.5.3	Emission Sources and Controls
The reverberatory furnace is the major lead emissions source. As apparent
from the reaction equation, less than one-fourth of the lead introduced into
the autoclaves actually reacts to form lead alkyl. The reverberatory furnace,
therefore, processes approximately four times the amount of lead in the lead
alkyl product. The emissions consist mainly of particulate lead oxide and are
similar to emissions from secondary lead smelters.
90

-------
5.5.4 Industry-Specific Methods
For the alkyl lead industry, the baseline ambient impact was estimated for
each lead source at each individual plant, using the ISCST atmospheric
dispersion model. Lead emission parameters used in the analysis were obtained
through contacts with State and federal environmental agencies and published
sources.?
Very little data are available for duPont's Deepwater, New Jersey, plant.
Claims of confidentiality necessitated making assumptions concerning the
plant's operation. Essentially, a model lead alkyl plant was developed to
represent the duPont plant based on the processes used at the recently closed
Ethyl and Nalco plants. The plant was assumed to have two major lead emission
sources resulting in a maximum average ambient concentration of 1.193 ug/m^ on
a quarterly basis. Because of the suburban location of the Deepwater plant,
the mobile source background concentration around this plant in 1990 and 1995
was assumed to be neglible.
5.5.5 NAAQS Control Strategies and Costs
One fabirc filter would be required for the model plant to achieve the 1.0
and 0.8 ug/m^ alternatives; and two fabric filter systems would be required to
meet the 0.5 and 0.25 ug/rn-^ alternatives. Costs for fabirc filters for the
model plant are given in Appendix D. Costs of attainment for the plant are
tabulated in Table 5-9.
5.6 GRAY IRON FOUNDRIES
5.6.1 Process Description^
Gray iron, which is defined as an iron with a carbon content of 2 to 4
percent, is produced from the melting and casting of scrap metal, foundry
returns, and pig iron at foundry operations (SIC 3321). The four basic
production processes at a gray iron foundry are: raw material handling; metal
91


-------
TABLE 5-9. NATIONWIDE NAAQS CONTROLS COSTS FOR THE LEAD-ALKYL PRODUCTION
(OCTOBER 1984 DOLLARS)3
Lead NAAQS	Costs (thousands)	Number
alternative	Installed	Total	of plants
ug/m^)	capital	annual	affected
0.25
2,080
749
1
0.5
2,080
749
1
0.8
1,040
375
1
1.0
1,040
375
1
1.5
—
—
0
2.0
—
—
0
2.5
—
—
0
aThe background ambient lead level for the lead alkyl plant was assumed
to be negligible.
92

-------
melting; mold and core production; and casting and finishing. Because lead is
present as a contaminant in the raw materials, the production of gray iron
results in lead emissions.
Raw Material Handling - The raw material handling operations include the
receiving, transfer, and storage of raw materials. These materials include
pig iron, metal scrap, foundry returns, fluxes (limestone, soda ash,
fluorspar, calcium carbide), coke, and binders. After receipt, these
materials are transferred to storage areas until required for the production
of gray iron.
Metal Melting - The three major types of furnaces used in the metal
melting process at gray iron foundries are; cupolas, which use heated
combustion air to melt the furnace charge; electric arc, which uses three
graphite electrodes; and electric induction, which uses the magnetic field
created from alternating current that energizes the metal coils surrounding
the furnace. The basic steps in the metal melting process include:
•	furnace charging, in which metal, scrap, alloying material, coke, and
flux are charged into the furnace;
•	melting;
•	backcharging, which involves the addition of metal and alloying
material;
•	refining and treating;
•	slag tapping; and
•	metal tapping into ladles or molds.
Mold and Core Production - Holds are forms used to shape the exterior of
castings and cores are shapes used to make the interior voids of castings.
Both are made from sand and various binding materials and then heated or dried
for form retention.
93

-------
Casting and Finishing - After tapping from the furnace, molten metal may
be treated through the addition of magnesium, lime, soda ash, or graphite.
The treated metal is then poured into molds and allowed to cool. The
partially cooled castings are placed on a vibrating grid where mold and core
sand is shaken away, collected, and recycled. Burrs, risers, and gates are
broken or ground off in the finishing process to match the contours of the
castings. In the final step, castings are shot blasted to remove remaining
mold sand and scale.
5.6.2 Industry Description
The casting produced at gray iron foundries are used primarily in the
production of industrial equipment, such as automobiles and trucks, farm
machinery, and machinery for heavy industries. The gray iron foundy industry
has undergone a dramatic change during the last 40 years, as new, larger
capacity operations have forced smaller, less efficient foundries from the
market. It is estimated that the number of gray iron foundries has decreased
from approximately 1,500 in the early 1970's to around 700 to 800 at present.
Total industry production has remained relatively constant during that period
and capacity utilization for this industry is currently around 68
percent.28>29 Although overall growth for the non-ferrous castings industry
is predicted to be about 4 percent per year,28 much of this growth is expected
to occur in electric furnaces.29 in fact, the Electric Power Research
Institute predicts that electric furnaces will replace fuel-fired furnaces
such as cupolas. Thus, growth in gray iron by fuel-fired furnaces is not
expected to exceed 2 percent per year.
5.6.4 Industry Specific Methodologies
As stated in the process description section, the gray iron foundry
industry uses three major types of furnaces in the metal melting process:
cupolas, electric arc, and electric induction. Based on emission calculations
and dispersion modeling results, the only emission source at gray iron
94

-------
foundries that contributed significantly to exceedences of the lead NAAQS
alternatives were uncontrolled cupola furnaces. The NEDS files indicated that
there were eleven of these furnaces of varying capacities. The dispersion -
modeling results for this industry were based on sources located in the
following urbanized areas:
•	Chicago, Illinois;
•	Dallas, Texas;
•	Minneapolis/St. Paul, Minnesota; and
•	Philadelphia, Pennsylvania.
To determine the baseline ambient impacts of these furnaces, each was
classified into one of three-model plant categories based on furnace charging
capacity; small (23 megagrams per heat); medium (91 Mg/heat); and large (272
Mg/heat).30 Using particulate emission factors developed for each of these
model plants sizes2?>30 an(j an average lead content for the particulate of
three percental, annual baseline lead emissions were calculated. These
emissions were then multiplied by a source-receptor coefficient for this
industry of 0.15 ug/m per Mg/yr, baseline ambient impacts were developed. The
inputs used in developing these impacts are presented in Table 5-10.
Host gray iron foundries are located in urban areas. Therefore, the upper
and lower limit urban mobile source background levels derived in Section 3
were used. These concentrations, 0.01 and 0.05 ug/m-*, bracket the range of
possible average urban mobile source impacts for the two lead-in-gasoline
phasedown alternatives. Costs of attainment for gray iron foundries were
estimated for both possible background levels. As with other stationary
sources located in urbanized areas, it was assumed that gray iron foundries
owned property up to 150 meters from the emission source.
5.6.5 NAAQS Control Strategies and Costs
Industrial source categories that would not attain one or more of the lead
NAAQS alternatives were determined from dispersion modeling of sources in ten

95

-------
TABLE 5-10. BASELINE EMISSIONS AND AMBIENT IMPACTS FOR GRAY IRON
FOUNDRY CUPOLA FURNACES



Baseline
Baseline
Plant
Furnace
Baseline
lead
ambient
identification
Capacity
percent
emissions
impact
number
(Mg/hr)
control3
(Mg/yr )b
(ug/m3)c
1
13.6
62
6.4
0.96
2
6.9
95
10.5
1.57
3
36.4
0
12.5
1.87
4
43.6
0
10.0
1.49
5
5.9
98
8.0
1.20
6
40.9
95
1.9
0.29
7
12.0
0
4.7
0.70
8
16.9
71
3.8
0.58
9
104.6
95
3.9
0.58
10
13.8
75
1.9
0.28
11
5.9
98
2.1
0.32
12
14.5
25
1.7
0.25
13
45.5
97
7.3
1.10
14
36.4
97
4.2
0.63
15
32.5
78
8.2
1.23
16
25.5
95
10.9
1.63
17
22.3
99
2.5
0.38
18
21.8
92
2.1
0.31
19
4.7
48
1.6
0.23
20
36.4
0
43.5
6.52
21
7.1
0
3.8
0.58
22
5.5
0
2.6
0.38
23
6.9
53
1.7
0.26
24
8.3
0
3.7
0.55
25
14.3
95
1.8
0.26
*From the December 1984 updata of NEDS.
bBaseline lead emissions were calculated based on controlled emission
rates given in NEDS and a lead concentration in particulate of
3 percent (reference 30).
°Based on a source-receptor coefficient of 0.15 ug/m3 per Mg/yr.
96

-------
selected urbanized areas. Based on this modeling, it was determined that
uncontrolled cupola furnaces at gray iron foundries would exceed at least one
of the lead NAAQS alternatives. The control strategy developed for this
industry consists of the installation of a side draft hood ducted to a
baghouse control system. Costs of attainment for this industry are presented
in Table 5-11. Appendix D gives background data on the development of control
techniques and costs. Table 5-11 projects that some plants will not be able
to attain the current NAAQS and lower alternatives. Given uncertainties
involved in the use of an average lead-in-particulate concentration and an
average source-recptor coefficient, it is possible that the plants would
attain the lower NAAQS using the identified controls. In addition, due to
plant closures associated with industry trends toward larger plants, some of
the plants on which this analysis was based may not be in operation by 1990.
5.7 ALLOY STEEL ELECTRIC ARC FURNACES
5.7.1 Process Description^
Direct electric arc furnaces (EAFs) (SIC 3313) are refractory-lined,
cylindrical vessels that have a bowl-shaped hearth and domed roof. Three
graphite or carbon electrodes are used to heat charged material in the furnace
and are raised and lowered through holes in the furnace roof as needed.
Thebasic processes used in the production of alloy steel by EAF's are:
charging; meltdown and refining; and tapping. Because lead is present as a
contaminant in the raw materials, the production of alloy steel results in
lead emissions.
Charging - A drop-bottom (clam shell type) bucket is used to charge EAF's
with iron and steel scrap; alloying materials such as nickel, copper, and
molybdenum; lime, which is used as a flux material to reduce the sulfur and
phosphorus content of the molten steel; and coke. Charging is accomplished
through opening of the furnace roof.
97
T

-------
TABLE 5-11. NAAQS CONTROLS COSTS FOR THE GRAY IRON FOUNDRY
INDUSTRY (OCTOBER 1984 DOLLARS)
Lead NAAQS and Costs (thousands)
mobile source Installed Total
background (ug/m^) capital annual
Number
of plants
impacted
Number
unable to
comply3
Background
= 0.05



0.25
21,330
4,828
14
14
0.5
16,200
3,665
9
7
0.8
10,980
2,494
5
3
1.0
10,980
2,494
5
3
1.5
6,588
1,496
3
2
2.0
2,196
499
1
0
2.5
2,196
499
1
0
Background
= 0.01



0.25
20,871
4,720
13
13
0.5
16,200
3,665
9
7
0.8
10,980
2,494
5
3
1.0
8,784
1,995
4
3
1.5
4,302
998
2
2
2.0
2,196
499
1
0
aControls could not be identified
various NAAQA alternatives.
that would allow
some plants
to attain

98

-------
Melting and Refining - After charging, the furnace roof is closed and the
electrodes are lowered to the charge material. Electrical currents are passed
through the electrodes to cause arcing. This arcing heats the charge and a
molten pool of melt is formed in the center of the hearth. The furnace many
be backcharged one ore more times during the melting process. The slag formed
during melting is tapped off prior to tapping of the molten metal.
Tapping - Tapping of the molten metal is accomplished through tilting of
the furnace vessel and pouring the product into ladles. Ferromanganese alloy,
ferrosilicon alloy, and aluminum may be added to the ladel to adjust the
oxygen content of the alloy steel. In addition, alloying materials such as
aluminum, zirconium, titanium, vanadium, and boron may be added to achieve the
desired product. Bottom tapping of EAFs is a fairly new advance that is being
used at some new facilities.
5.7.2	Industry Description-^
EAFs are utilized typically in semi-integrated and nonintegrated steel
mills and speciality steel shops. The semi-integrated steel mills use direct
reduced iron along with iron and steel scrap as a soure of ferrous
material.The nonintegrated mills operate melting units, such as EAFs, casting
units, and fabrication mills to produce limited product lines for regional
markets. Those with annual production of less than 544,200 megagrams (600,000
tons) are classified as mini-mills. Speciality steel shops are similar to
nonintegrated mills except for an additional secondary refining of the molten
metal tapped from EAFs. The use of EAFs in the steel production industry is
expected to continue an upward growth trend, as this process accounts for an
increasing share of total steel output. It is estimated that 30 new EAFs will
come on-line by 1990 and an additional 11 by 1995.31•32
5.7.3	Emission Sources and Controls^!
The charging, melting/refining, and tupping operations at EAFS contribute
to the emission of fugitive process particulate lead. Currently, the typical

-------
control system used for process emissions is direct-shell evacuation and
fabric filter control system. For fugitive process emissions, the control
techniques that are typically used include fabric filters, single or segmented
canopy roofs, scavenger duct, cross-draft partitions, and enclosed dust
handling equipment.
5.7.4	Industry Specific Methodologies
To determine the baseline ambient impacts from alloy steel EAF's, each of
these sources that appeared in the NEDS file was classified into one of three
model plant sizes based on furnace capacity. The three sizes developed were:
23 megagrams per heat; 91 Mg/heat; and 272 Mg/heat.^l The furnaces appearing
in NEDS were thus classified and emission factors^ and operating rates were
used to esimate total particulate emissions. For the alloy steel EAF
industry, dispersion modeling results were based on sources located in the two
urbanized areas of Chicago, Illinois, and Philadelphia, Pennsylvania.
It was then assumed that the particulate emissions had an average lead
content of two percent. From this, lead emission estimates were developed for
each source. Using a source-receptor coefficient of 0.12 ug/m^ per Mg/yr for
alloy steel EAFs, baseline ambient impacts on a source-by-source basis were
calculated. These estimates are presented in Table 5-12. From these
estimations, it was determined that only uncontrolled alloy steel EAFs would
exceed the various lead NAAQS alternatives.
As with other source categories located primarily in urban areas, the
upper and lower limit mobile source background levels derived in Section 3
were used. Costs of attainment for alloy steel EAFs were estimated for both
possible background levels.
5.7.5	NAAQS Control Strategies and Costs
Based on the dispersion modeling and emission calculations as described
above, it was determined that alloy steel EAFs would cause exceedences of the
various NAAQS alternatives being studied. Control strategies designed to
¦)
100

-------
TABLE 5-12. BASELINE EMISSIONS AND AMBIENT IMPACTS FOR ALLOY STEEL
ELECTRIC ARC FURNACES
Plant
identification
number
Furnace
capacity
(Mg/hr)
Baseline
percent
control3
Baseline
lead
emissions
(Mg/yr )b
Baseline
ambient
impact
(ug/m3)c
1
36.4
99.9
4.8
0.58
2
52.7
90.5
2.3
0.28
3
77.3
99.5
4.9
0.59
4
9.1
0
3.8
0.46
aFrom the December 1984 update of NEDS.
Baseline lead emissions were calculated based on controlled emission
rates given in NEDS and a lead concentration in particulate of
3 percent (reference 30).
CBased on a source-receptor coefficient of 0.12 ug/m3 per Mg/yr.
101
i

-------
bring these sources into attainment with the lead NAAQS alternatives were then
developed. The three strategies developed contain the same process emission
control system (direct-shell evacuation and fabric filter control system), but
vary in the degree of fugitive process emission control.
Costs of attainment for the various lead NAAQS alternatives on an
industry-wide basis are given in Table 5-13. Appendix D presents data used in
the development of control techniques and costs. Table 5-11 projects that
some plants will not be able to attain the lower NAAQS alternatives. Given
uncertainties involved in the use of an average lead-in-particulate
concentration and an average source-receptor coefficient, it is possible that
the plants would attain the lower NAAQS using the identified controls.
5.8 STEEL ELECTRIC ARC FURNACES
5.8.1 Process Description^
Direct electric are furnaces (EAFs) (SIC 3313) are refractory-lined,
cylindrical vessels that have a bowl-shaped hearth and domed roof. Three
graphite or carbon electrodes are used to heat charged material in the furnace
and are raised and lowered through holes in the furnace roof as needed. The
basic processes used in the production of steel by EAFs are: charging;
meltdown and refining; and tapping. Because lead is present as a contaminant
in the raw materials, the production of steel results in lead emissions.
Charging - A drop-bottom (clam shell type) bucket is used to charge EAF's
with iron and steel scrap; alloying materials such as carbon raiser,
manganese, and silicon; lime, which is used as a flux material to reduce to
sulfur and phosphorus content of the molten steel; and coke. Charging is
accomplished through opening of the furnace roof.
Melting and Refining - After charging, the furnace roof is closed and the
electrodes are lowered to the charge material. Electrical currents are passed
through the electrodes to cause arcing. This arcing heats the charge and a
molten pool of melt is formed in the center of the hearth. The furnace may be
¦)
102

-------
TABLE 5-13. NATIONWIDE NAAQS COSTS FOR ALLOY STEEL EAFs
(OCTOBER 1984 DOLLARS)
Background and
NAAQS
level (ug/m3)
Costs (thousands)
Installed
capital
Total
annual
Number
of plants
affected
Number
unable to
comply3
Background = 0-05
0.25	8,069	2,457 2 2
0.5	6,818	1,826 1 2
0.8	—	--0 0
1.0	—	—00
1.5	—	—00
2.0	—	--0 0
2.5	—	—00
Background = 0.01
0.25	8,069	2,457 2 2
0.5	--	— 0 2
0.8	--	--0 0
1.0	—	—00
1.5	—	—00
2.0	--	—00
2.5	—	—00
aControls could not be identified that would allow some plants to comply with
the lower alternatives.
103

-------
backcharged one or more times during the melting process. The slag formed
during melting is tapped off prior to tapping of the molten meltal.
Tapping - Tapping of the molten metal is accomplished through tilting of
the furnace vessel and pouring the product into ladles. Ferromanganese alloy,
ferrosilicon alloy, and aluminum may be added to the ladle to adjust the
oxygen content of the alloy steel. Bottom tapping of EAFs is a fairly recent
advance that is being used at some new facilities. After the tapping process,
the molten melt is poured into casts for cooling.
5.8.2	Industry Description-^
EAFs are utilized typically in semi-integrated and nonintegrated steel
mills and speciality steel shops. The semi-integrated steel mills use direct
reduced iron along with iron and steel scrap as a source of ferrous material.
The nonintegrated mills operate melting units, such as EAFs casting units, and
facrication mills to produce limited product lines for regional markets.
Those with annual production of less than 544,200 megagrams (600,000 tons) are
classified as mini-mills. Speciality steel shops are similar to nonintegrated
mills except for an additional secondary refining of the molten metal tapped
from EAFs.
5.8.3	Emission Sources and Control-^
The charging, melting/refining, and tapping operations at EAFs contribute
to the emission of fugitive process particulate lead. Currently, the
typicalcontrol system used for process emission are side draft hoods and
fabric filters for small furnaces and direct-shell evacuation and fabric
filter control systems for medium and large furnaces. For fugitive process
emissions, the control techniques tht are typically used include fabric
filters, single or segmented canopy roofs, scavenger duct, cross-draft
partitions, and enclosed dust handling equipment.
The average lead concentration in particulate for steel EAFs was taken to
be 2 percent. It was assumed that the source-receptor coefficient for steel
104

-------
EAFs is the same as that for alloy steel EAFs, or 0.12 ug/m^ per Mg/yr
(Section 5.1.4). Table 5-14 gives baseline lead emissions and ambient impacts
were calculated for individual steel EAFs using these factors and particulate
emission rates from NEDS. As with other stationary sources located primarily
in urbanized areas, the costs of attainment were estimated the upper and lower
limit mobile source background levels derived in Section 3.
5.8.4 Industry Specific Methodologies
To determine the baseline ambient impacts from steel EAF's, each of these
sources that appeared in the NEDS file was classified into one of three model
plant sizes based on furnace capacity. The three sizes developed were: 3.6
megagrams per heat; 9.1 Mg/heat; and 22.T Mg/heat.^2 The furnaces appearing
in NEDS were thus classified and emission factors^ and operating rates were
used to estimate total particulate emissions.
5.8.5. NAAQS Control Strategies and Costs
Based on the dispersion modeling and emission calculations as described
above, it was determined that steel EAFs would cause exceedences of the
various NAAQS alternatives being studied. Control strategies designed to
bring these sources into attainment with the lead NAAQS alternatives were than
developed. The three strategies developed contain the same process emission
control system (side draft hood or direct-shell evacuation and fabric filter
control system), but vary in the degree of fugitive process emission control.
Costs of attainment for the various lead NAAQS alternatives on an
industry-wide basis are given in Table 5-15. Appendix D presents data used in
the development of control techniques and costs. Table 5-15 projects that
some plants will not be able to attain the current NAAQS and the lower
alternatives. Given uncertanties involved in the use of an average
lead-in-particulate concentration and an average source-receptor coefficient,
it is possible that the plants would attain the lower NAAQS using the
identified controls.
105


-------
TABLE 5-14. BASELINE EMISSIONS AND AMBIENT IMPACTS FOR STEEL
ELECTRIC ARC FURNACES
Plant
identification
number
Furnace
capacity
(Mg/hr)
Baseline
percent
¦.a
control
Baseline
lead
emissions
(Mg/yr)b
Baseline
ambient
impact
(ug/m3)'
1
1.5
0
2.9
0.35
2
7.9
97.2
3.1
0.37
3
48.2
95.0
2.0
0.23
4
9.1
0
4.5
0.54
5
9.1
0
6.1
0.73
6
33.2
96.5
4.4
0.53
7
3.9
0
3.6
0.43
8
17.9
0
7.2
0.86
9
9.1
0
2.4
0.29
10
9.1
99.9
16.0
0.48
11
8.5
99.0
6.1
0.73
aFrom the December 1984 update of NEDS.
bBaseline lead emissions were calculated based on controlled emission
rates given in NEDS and a lead concentration in particulate of
3 percent (reference 30).
cBased on a source-receptor coefficient of 0.12 ug/m3 per Mg/yr.
106

-------
TABLE 5-15. NATIONWIDE NAAQS COSTS FOR CARBON STEEL EAFs
(OCTOBER 1984 DOLLARS)
Background and
NAAQS
level (ug/m3)
Costs (thousands)
Installed
capital
Total
annual
Number
of plants
affected
Number
unable to
comply3
Background
0.25
0.5
0.8
1.0
1.5
2.0
2.5
= 0.05
2,459
1,451
654
1,189
670
287
9
5
2
0
0
0
0
4
2
1
1
1
0
0
Background = 0.01
0.25
0.5
0.8
1.0
1.5
2.0
2.5
1,949
1,451
654
998
670
287
8
5
2
0
0
0
0
4
2
1
1
1
0
0
aControls could not be identified that would allow some plants to comply with
the current NAAQS and the lower alternatives.
107

-------
5.9 IRON AND STEEL INDUSTRY - SINTERING
5.9.1	Process Description^
Sintering operations are used in iron and steel manufacture at large,
integrated mills. The sintering process converts fine materials into an
agglomerate that can be charged to a blast furnace. Feed materials to
sintering operations include fine iron ore, coke breeze, fluxstone, mill scale
and flue dust. These materials are mixed with water, and then placed on a
grate that moves through the sinter furnace. In the furnace, the coke in the
feed material is ignited, providing the necessary heat for surface melting and
agglomeration. Fused sinter is discharged at the end of the furnace, and is
then crushed and screened, with any undersized material recycled to the
furnace. Sinter product is cooled either by air or by a water spray, and is
then fed to blast furnace.
5.9.2	Emission Sources and Controls^
Combustion air from the sintering process contains entrained sinter
particulate, and is generally ducted to a particulate control device.
Fugitive particulate emissions emulate from materials handling, the sinter
furnace charge and discharge ports, and the cooling and crushing operation.
In this study, controls were identified for sinter fugitive emissions from
materials handling, and from the sinter furnace area. For materials handling,
the controls identified were covering of trucks and railcars, with ventilation
to a fabric filter. For the furnace area, the controls identified were
hooding and ventilation of the furnace ports, and replacement of the crushing
circuit with a closed system.
5.9.3	Industry Specific Methodologies
The average lead concentration in particulate matter for sintering
fugitives was taken to be 1.7 percent; and the average source-receptor
coefficient for the sintering operation was assumed to be the same as that for
108
"i

-------
gray iron production, or about 0.15 ug/m^ (Section 5.6.4). Table 5-16 gives
baseline lead emissions and ambient impacts were calculated for individual
sintering operations using these factors and particulate emission rates from
NEDS. As with other stationary sources located primarily in urbanized areas,
the costs of attainment were estimated using the upper and lower limit mobile
source background levels derived in Section 3.
5.9.4 NAAQS Control Strategies and Costs
Costs for the sinter materials handling and sinter furnace emission
control systems identified in this study are given in Appendix D. Table 5-17
gives total estimated attainment costs for sinter operations for the various
NAAQS alternatives. The table indicates that some plants will not be able to
attain the lower NAAQS alternatives. Given uncertainties involved in the use
of an average lead-in-particulate concentration and an average source-receptor
coefficient, it is possible that the plants would attain the lower NAAQS using
the identified controls.
5.10 BRASS AND BRONZE INGOT PRODUCTION
5.10.1 Process Description-^
The raw materials used in brass and bronze production (SIC 3341) are
almost entirely derived from scrap materials with virgin metals used only to
adjust the composition of the product as desired. These raw materials are
subjected to a series of sorting, classification, and preparation steps before
undergoing the actual ingot production process. The ingots produced are
generally further processed by remelting, shaping, rolling, extracting, etc.
in brass or bronze mills to produce final products or intermediate products
for delivery to brass and bronze manufacturing facilities.
The scrap raw materials are prepared for the production of ingots by
mechanical hydrometallurgical, or pyrometallurgical treatment. These
processes remove the undesireable scrap components leaving a mixture of metals
as close as possible to the composition of the desired product.
->
109

-------
TABLE 5-16. BASELINE EMISSIONS AND AMBIENT IMPACTS OF IRON AND STEEL SINTERING
Plant
identification
Operation
Baseline
percent
control3
Baseline
lead
emissions
(Mg/yr )b
Baseline
ambient
impact
(ug/m3)C
1
Material handling
0
2.7
0.33
2
Furnace
82.0
61.5
S. 21
3
Furnace
75.0
14.9
1.79
4
Material handling
99.9
1.8
0.22
5
Furnace
96.0
3.4
0.40
aFrom the December 1984 update of NEDS.
bBaseline lead emissions were calculated based on controlled emission
rates given in NEDS and a lead concentration in particulate of
3 percent (reference 30).
°Based on a source-receptor coefficient of 0.12 ug/m3 per Mg/yr.
110

-------
TABLE 5-17. NAAQS CONTROL COSTS FOR IRON SINTERING
(OCTOBER 1984 DOLLARS)
Lead NAAQS and
Costs
(thousands)
Number
Number
mobile source
Installed
Total
of plants
unable to
background (ug/m^)
capital
annual
affected
comply3
Background = 0.05




0.25
23,560
6,550
4
2
0.5
13,400
3,680
2
1
0.8
13,400
3,680
2
0
1.0
13,400
3,680
2
0
1.5
13,400
3,680
2
0
2.0
6,700
1,840
1
0
2.5
6,700
1,840
1
0
Background - 0.01




0.25
23,560
6,550
4
1
0.5
13,400
3,680
2
1
0.8
13,400
3,680
2
0
1.0
13,400
3,680
2
0
1.5
13,400
3,680
2
0
2.0
6,700
1,840
1
0
2.5
6,700
1,840
1
0
aControls could not be identified that would allow some plants to achieve the
lower alternatives.
Ill

-------
The production of brass and bronze ingots takes place in one of three
types of furnaces: direct-fired reverberatory, indirect-fired or electric.
The raw materials are charged to the furnace and when the charge attains the
proper heat and impurities have been drawn off into the slag, the molten metal
is tested for its alloy composition. Adjustments are made as needed and the
metal is brought to the ideal pouring temperature for the specific alloy and
the ingots are poured.
5.10.2	Industry Structure^
There has been a steady decline in the number of plants and production
since the peak year of 1966 despite the high wartime demand during the late
sixties and early seventies. There were 60 producers in 1969 and only 35 in
1978. This negative growth is expected to continued for two major reasons:
the first is a decline in market demand for certain brass and bronze products,
and the second is substitution of other materials or technologies for the
previously used brass or bronze. Fiber optics for telephone and data
transmission is expected to cause even further declines in brass and bronze
demand.
5.10.3	Emission Sources and Controls-^
Lead bearing particulate matter is emitted from process and fugitive
process sources at brass and bronze ingot production plants. The extent of
the lead emissions depends on the lead composition of the scrap and other
charged materials, temperature, type of furnace and various operating factors
such as methods of charging, slagging, alloying and pouring ingots The lead
fraction of particulate matter emissions ranges from 1-12 percent for most
operations.
The process emissions can be collected and ducted to a high-efficiency
baghouse for control of all particulate matter fractions. A few high pressure
drop venturi scrubbers are used in the industry but their overall control
efficiency (60 percent) is significantly lower than that of baghouses.
•)
112

-------
5.10.4 Industry Specific Methodologies
Particulate emissions from brass and bronze production plants were
obtained from the December 1984 update of NEDS. Particulate emissions from
brass and bronze production range from 2.5 to 56 percent lead; however, for
the majority of brass and bronze alloys, the lead content of particulate
emissions is about 15 percent. This average lead content was used to estimate
lead emissions from the individual plants; an average source-receptor
coefficient of 0.90 ug/m^ was used to estimate baseline ambient lead impacts.
This source-receptor coefficient was calculated using dispersion modeling
results for brass and bronze production plants in Chicago and Phoenix; gray
iron foundries in Chicago, Dallas, Minneapolis - St. Paul, and Philadelphia;
and alloy steel plants in Chicago and Philadelphia. Modeling results for the
iron and steel industry were used in the study of brass and bronze plants
because of the similarity in emission and stack parameters.
Baseline lead emissions and ambient impacts determined using the above
methods are given in Table 5-18. The costs of attainment for the brass and
bronze industry were estimated for the upper and lower limit mobile source
background levels (Section 3).
There were no data available concerning fugitive emissions and they were
therefore not considered in this analysis. The source impact for each plant
was determined and the emissions reductions necessary at the various lead
NAAQS alternatives were calculated. The control of choice for the industry,
high efficiency baghouse, was applied in the instances where reductions were
necessary.
5.10.5 NAAQS Controls and Costs
NAAQS control for the brass and bronze ingot production plants is limited
to the application of high-efficiency baghouses control of furnace emissions.
Table 5-19 gives industry-wide costs to achieve the various lead NAAQS
alternatives. The costs associated with the control techniques evaluated are
presented in Appendix D.
113

-------
TABLE 5-18.
BASELINE EMISSIONS
AND AMBIENT
IMPACTS FOR BRASS
AND

BRONZE PRODUCTION






Baseline
Baseline
Plant
Furnace
Baseline
"lead
ambient
identification capacity
percent
emissions
impact
number
(Mg/day)
control3
(Mg/yr)b
(ug/m3)c
1
34.2
0
7.4
0.60
2
45.2
0
3.5
0.28
3
0.2
0
8.4
0.69
4
6.9
0
2.7
0.22

• 10.3
0
2.9
0.23



TOTAL
0.45
5
3.3
0
7.2
0.59

6.4
0
2.4
0.20

6.5
0
1.1
0.09



TOTAL
0.87
6
2.0
0
1.8
0.15


0
1.2
0.10

3.3
0
0.6
0.05



TOTAL
0.29
7
29.5
0
11.6
0.94

29.5
0
12.5
1.02



TOTAL
1.96
8
109.1
95
3.2
0.26

76.4
85
0.6
0.05

109.1
95
0.8
0.06

218.2
95
4.8
0.39



TOTAL
0.76
9
17.5
0
8.1
0.66
10
18.2
80
5.3
0.43
11
54.5
0
9.3
0.76
--CONTINUED--
114

-------
TABLE 5-18 (CONTINUED)



Baseline
Baseline
Plant
Furnace
Baseline
lead
ambient
identification
capacity
percent
emissions
impact
number
(Mg/day)
control
(Mg/yr)
(ug/m3)
12
8.7
0
1.5
0.12

37.6
0
6.5
0.53

2.2
0
0.3
0.02

1.2
0
4.7
0.38

28.4
0
0.6
0.05



TOTAL
1.10
13
45.2
0
4.1
0.33


0
13.2
1.08



TOTAL
1.41
14
49.1
95
0.6
0.05

469.1
95
2.6
0.23



TOTAL
0.28
15
32.7
98
1.1
0.09

27.3
98
0.2
0.01

32.7
98
0.9
0.07

45.8
98
0.6
0.05

45.8
99
0.6
0.05

10.9
98
0.2
0.01

21.8
39
2.3
0.18



TOTAL
0.47
16
65.5
50
78.3
6.40
17
30.5
0
2.1
0.17

30.5
0
2.0
0.16

6.5
0
0.2
0.01

6.5
0
0.2
0.01



TOTAL
0.36
18
23.3
0
3.3
0.27

16.1
0
2.3
0.18



TOTAL
0.45
19
25.1
0
7.5
0.61

25.1
0
0.8
0.06



TOTAL
0.67
aFrom the December 1984 update of NEDS.


^Baseline lead
emissions were
calculated based
on controlled
emission
rates given in
NEDS and a lead concentration in particulate
of
3 percent (reference 30).
CBased on a source-receptor coefficient of 0.09 ug/m3 per Mg/yr.
115

-------
TABLE
5-
19. NAAQS CONTROL COSTS FOR IRON SINTERING
(OCTOBER 1984 DOLLARS)
Lead NAAQS and
mobile source
background (ug/m^)
Costs (thousands)
Installed
capital
Total
annual
Number
of plants
affected
Number
unable to
comply3
Background
0.25
0.5
0.8
1.0
1.5
2.0
2.5
= 0.05
5,575
3,613
2,300
1,440
745
745
404
1,568
980
624
390
202
202
109
19
12
8
5
3
3
2
4
2
0
0
0
0
0
Background
0.25
0.5
0.8
1.0
1.5
2.0
2.5
0.05
5,575
3,609
1,896
1,440
745
404
404
1,568
871
515
390
202
109
109
19
11
7
5
3
2
2
4
2
0
0
0
0
0
aControls could not be identified that would allow some plants to achieve the
lower alternatives.
116

-------
5.11 PRIMARY COOPER SMELTING AND REFINING
5.11.1 Process Description
The material processed at U.S. primary copper smelters (SIC 3331) are
predominately sulfide ore concentrates, of which the most common copper
bearing constituent is chalcopyrite (CuFeS2). The copper content of the ore
as minded is generally less than 1 weight percent, and flotation methods are
used to produce a copper concentrate that serves as the feed to the primary
smelting plant. A typical copper ore concentrate contains 15 to 30 weight
percent copper and large amounts of iron, sulfur, and silicon. Other
important constituents are lead, antimony, zinc, cadmium, bismuth, selenium,
and arsenic.^
The copper bearing minerals in the ore concentrate are separated from the
other materials and converted to 99 percent pure "blister copper" at the
primary copper smelter. There are three major processes involved at most
smelters: roasting, smelting, and converting.
5.11.1.1 Roasting
Copper ore concentrates require roasting if they have a low copper content
relative to sulfur and other impurities. Roasting involves heating the ore
concentrate to about 650°C (below the melting point of the charge) and has
three functions: (1) eliminating a portion of the sulfur as SO2;
(2) converting a portion of the iron sulfides to iron oxides; and (3) removing
volatile impurities such as antimony, arsenic, bismuth, and cadmium. The
roasted product is known as the calcine.
Currently, there are two types of roasters used—multiple hearth roasters
and the newer fluid bed roasters. For a given degree of sulfur removal, fluid
bed roasters require a shorter concentrate residence time and produce a higher
strength SO2 offgas stream than do multiple hearth roasters.
117

-------
5.11.1.2 Smelting
In the smelting process either hot calcine from the roaster or unroasted
ore concentrates are melted with suitable fluxes in a smelting furnace to
chemically reduce copper and higher iron oxides and to liberate sulfur as
SC>2. This process results in the production of two products: a copper-rich
matte, which generally contains 15 to 50 percent copper, together with iron,
most of the sulfur and any contained percious metals; and a copper-poor slag,
the waste flux high in silicon and iron and containing small amounts of copper
and other oxide impurities. These "phases" are mutually insoluable and
separate within the furnace, the slag floating on top of the matte. The matte
and slag are drawn off (tapped) separately, with the slag being discarded or
recycled and the matte being sent to the converter for further processing.
There are several types of smelting furnaces currently used in domestic
copper smelters: reverbatory furnaces, electric furnaces, flash furnaces, and
continuous (Noranda) furnaces. The reverbatory furnace is the oldest
technology and differs from the rest in that it produces an offgas stream with
a relatively low SC>2 content, generally between 0.5 and 1.5 percent SC>2» as
opposed to the other furnaces which produce offgas streams having greater than
3.5 percent SO2.
5.11.1.3 Converting
The molten matte produced in the smelting furnace is transferred to the
converter where flux materials are added to provide a medium in which any iron
contained in the matte can be removed as slag. There are two stages in the
converter operations. During the "slag blow," the iron sulfide in the matte
is preferentially oxidized to iron oxide. The iron oxide dissolves in the
flux and forms a ferrous slag layer that is removed from the converter and
recycled to the smelting furnace. Following the slag blow, the molten copper
sulfide is converted to metallic blister copper and sulfur dioxide during the
"copper blow." The SO2 content of the offgas is higher during the copper blow
than during the slag blow.
118

-------
The 99 percent pure blister copper from the converter is subsequently
refined to reduce the level of impurities. Refining consists of a series of
operations in which blister copper is first oxidized to CuC>2 and then reduced
back to metallic copper. Figure 5-1 shows a conventional copper smelting
process.
5.11.2	Industry Structure
Currently, there are seven companies operating twelve domestic primary
copper smelters with a combined production capacity of 1,421,000 Mg
(1,563,100 tons).38 This represents an 18 percent reduction in capacity from
1984 due to the closing of three smelters: the ASARC0 smelter in Tacoma,
Washington (90,000 Mg capacity), and the Phelps Dodge smelters in Ajo
(64,000 Mg capacity) and Morenci (160,000 Mg capacity), Arizona.^7 In 1984,
domestic primary copper smelter production was 989,720 Mg (1,088,692 tons),
approximately 57 percent of capacity.
The Copper Range/White Pine smelter has been closed since 1982, but
negotiations, with intent to reopen the mine and smelter are currently
underway.^9 No growth in U.S. capacity is expected and a contraction of the
industry is probable over the short term; an average annual growth rate of
1.7 percent per year from 1983 to 2000 is predicted.^® Table 5-20 shows the
company, location and estimated production capacity of the twelve operating
smelters.
5.11.3	Emission Sources and Controls
Lead-bearing particulates are emitted from process, fugitive process, and
fugitive dust sources at primary copper smelters.
5.11.3.1 Process Emissions
Process emissions can be well controlled through the use of extensive
hooding and ducting systems, baghouses, contact sulfuric acid plants,
119

-------
ROASTI.NG
•	multiple hearth
•	fluid bed
calcine
"grssn feed"

SMELTING

• reverbatory

• electric

• continuous

• flash




slag	matte
i
1

CONVERTING



blister copper
1 r
REFINING
Figure 5-1. Conventional copper smelting process.
120

-------
TABLE 5-20. DOMESTIC PRIMARY COPPER SMELTERS
Company
Smelter location
Copper Production
Capacity3 Mg/yr (tons/yr)
ASARCO
El Paso, Texas
91,000
(100,000)

Hayden, Arizona
182,000
(200,000)
Tennessee Chemical
Copperhill, Tennessee
13,600
(15,000)
Copper Range
White Pine, Michigan
52,000
(57,000)
Inspiration Consolidated
Miami, Arizona
136,000
(150,000)
Kennecott
Garfield, Utah
254,000
(280,000)

Hayden, Arizona
71,000
(78,000)

Hurley, New Mexico
73,000
(80,000)

McGill, Neveda
45,000
(50,000)
Magma
San Manuel, Arizona
181,000
(200,000)
Phelps Dodge
Douglas, Arizona
115,000
(127,000)

Hiladago, New Mexico
163,000
(179,000)
aProduction of blister copper.
121

-------
electrostatic precipitators (ESPs), scrubbers, and dispersion of emissions
from tall stacks.
Roasters -- The emission control technologies that can be applied to the
two types of roasters (multiple hearth and fluid bed) depend on the SO2
content of the offgas streams. Fluid bed roaster offgas streams have a high
enough SO2 concentration to be economically treated with a contact sulfuric
acid plant, which when used in conjunction with an ESP achieves a 99.9 percent
reduction in he lead content of the waste stream. The SO2 concentration of
multiple hearth roaster offgas is too low for economical treatment with a
contact sulfuric acid plant unless the stream is blended with another more
concentrated SO2 stream. ESPs are used for particulate control, and cold ESPs
can remove up to 99 percent of particulates, efficiencies for hot ESPs range
from 20 to 80 percent.^
Furnaces — The existing control equipment employed on smelting furnace
offgas streams depends on the furnace type. As with the multiple hearth
roaster, the SO2 content of reverberatory furnace offgas is too low to treat
the stream economically with a contact sulfuric acid plant, and other forms of
SO2 control, such as flue gas desulfurization, are not practiced in the
domestic primary copper smelting industry. ESPs are used exclusively within
the industry to remove particulate matter from reverberatory furnace
offgases. Contact sulfuric acid plants are used to treat streams from
electric, flash, and continuous (Noranda) furnaces because of their relatively
high SO2 concentration.
Converters -- Offgases from the converter operation that are captured by
the primary converter hooding system are treated in contact sulfuric acid
plants at 6 of the 9 copper smelters modeled for this study. Of the remaining
three plants, two use hot ESP systems for particulate control and the other
has no converter offgas pollution control equipment.
122

-------
5.11.3.2 Fugitive Emissions
Fugitive emissions may be characterized as emissions which escape directly
from the process area to the atmosphere rather than through a flue or exhaust
system. They result from leakage in and around process equipment and from
material handling and transfer operations. These emissions may be considered
as low level emissions when compared to process emissions, since they usually
leave the smelter at or near ground level, whereas process emissions are
discharged through a tall stack.
Roaster — Fugitive emissions seldom occurs during charging of multiple
hearth roasters because of the water content (8-10 percent) of the feed, and
the choke feed mechanism on the charging hoppers. Because of the enclosed
feed and discharge system, fugitives are seldom emitted from fluid bed
roasters.
Fugitive emissions from multiple hearth roasters may be emitted from leaks
that can occur at the doors located at each of the hearth levels, from holes
in the actual shell of the roaster, or from leaks around the shaft supporting
the rabble arms. Under normal operating conditions, these emissions are
minimized by operating the multiple hearth roasters under a slight negative
pressure.
Calcine Transfer — Fugitive emissions may be generated during the
discharge and transfer of hot calcine from roaster to smelting furnaces.
Smelters with multiple hearth roasters usually use larry cars (small rail
cars) to transport calcines to the furnace. When the material is dropped from
the calcine hopper located under the roaster into the covered car through a
feed opening, large quantities of dust are generated as a result of material
movement and pressure changes within the car. Some fugitive emissions can
also occur during the transportation of the roaster calcines to the smelting
furnace. In the case where larry cars are used, their feed opening is usually
covered to minimize this effect.37
Fugitive emissions are seldom emitted from a properly operated fluid bed
roaster during calcine transfer operations since it is essentially a closed
system.
123

-------
Smelting Furnaces — As noted previously, four basic types of smelting
furnaces are used; reverberatory, electric, continuous (Noranda), and flash.
Depending on the method used for charging and the type of smelting furnace,
fugitive emissions can be generated. However, by using add-on control systems
which are activated at the time the charge is dropped, effective reduction of
potential fugitive emissions can be obtained. Since most smelters are
currently using such control systems, it is reasonable to assume that fugitive
lead emissions during charging of smelting furnaces are negligible.
Reverberatory furnaces are usually operated under slight negative pressure
which helps prevent fugitive emissions from the many potential leakage points
around the furnace. The reduced requirement for providing oxygen to electric
furnaces results in these furnaces being better sealed. If positive pressure
is momentarily generated, fugitive emissions due to leakage are considerably
less. The construction of flash furnaces is such that they are virtually
gas-tight, preventing leakage of emissions. Provided they are properly
operated and maintained, it is reasonable to assume that fugitive lead
emissions due to leakage from smelting furnaces are negligible.^
Matte and slag tapping are the principle fugitive emission sources at the
smelting furnace. Smelting furnaces have from one to three matte tap holes
with associated launders on each side. The launder directs the flowing matte
to a point where it can be collected in a large ladle. Fugitive emissions are
observalbe from the point at which the matte leaves the furnace to the
location where it enters the ladle. Typically, a matte tapping operation
takes 5 to 15 minutes. Slag tap ports and slag launders have been observed to
emit less fugitive emissions than those emitted during matte tapping
operations; however, fugitive emissions are observable from the point of the
slag leaving the furnace to the location where it enters the ladle. A typical
slag tapping operation takes 10 to 20 minutes.
It has been concluded that a properly designed and operated ventilation
system applied to matte and slag tapping operations should achieve a minimum
capture efficiency of 90 percent.37 These captured emissions can be vented to
an available control devise for treatment.
124

-------
Converters -- During converter charging, visible fugitive emissions of
short duration occur when matte or other materials are poured from a ladle
into a converter. During the blow phase, some fugitive emissions are
discharged through openings between the primary hoods and the converter
openings. Skimming operations, due to equipment requirements, generally
result in significant fugitive emissions. Fugitive emissions also occur when
the blister copper is poured from the converter into the ladle. Additional
fugitives may occur during holding operations and from occassional converter
leaks.
Fugitive emissions controls for converters include air curtain secondary
hood capture systems, usually vented to a fabric filter or other existing
control device, which can achieve up to 95 percent capture efficiency.
Dust Transfer and Handling—Dust handling and transfer can generate
fugitive emissions if carelessly performed. However most smelters take
reasonable precautions to minimize fugitive emissions from dust handling and
transfer. Fugitive dust emission control techniques includes paving, traffic
control, wet suppression, and soil stabilization for unpaved haul roads and
open work areas; and wet suppression, surfactants, and enclosures for
stockpiles and conveyor systems. Table 5-21 lists fugitive emission control
techniques used at primary copper smelters.
5.11.4 Control Stragety Development
5.11.4.1 Estimation of Baseline Ambient Impacts
As noted earlier, four primary copper smelters have been closed recently,
three of them permanently. The remaining smelters currently in operation and
the one which may reopen were studied individually. The baseline ambient lead
impact for each plant was determined for each plant either by modeling, or
when available and adequate, from ambient monitoring data. Four smelters had
annual quarterly maximums below the lowest alternative and were not modeled.
When adequate ambient data was not available or indicated ambient impacts
above 0.25 ug/rn-^, annual maximum quarterly concentrations were estimated using

125

-------
TABLE 5-21. FUGITIVE EMISSION CONTROL TECHNIQUES AT PRIMARY COPPER SMELTERS
Fugitive
emission
soruce

Control
techniques
Roaster
1.
Negative pressure

2.
Hooding
Calcine Transfer
1.
Hooding
Smelting Furnace
1.
Negative pressure

2.
Capture and


collection and systems
Converters
1.
Air curtain


secondary hood


capture systems
Dust handling
1.
Paving
and transfer



2.
Wet suppression

3.
Soil stabilization
Stock piles/conveyor
1.
Surfactants
systems



2.
Enclosures
126

-------
ISCST.^2 Emission parameters for most of the remaining copper smelters were
available from Inorganic Arsenic Emissions from Primary Copper Smelters -
Background Information for Promulgated Standards.36 Wind frequency and
direction data, when available, came from the nearest state meteorological
station. When possible, coordinates of process and fugitive emission sources
were taken from plant site plans or the National Emissions Data base (NEDS)^3
in order to estimate emissions from each source, rather than colocating all
sources at plant center.
5.11.4.2 NAAQS Control Strategies and Costs-
Using the results from the calculation of the baseline ambient impacts,
individual plants were identified which would be impacted by the various NAAQS
alternatives. A plant was determined to be affected if the known or predicted
annual maximum quarterly ambient lead concentration at a point outside the
plant boundry exceeded the NAAQS level. Because of the remote locations of
the primary copper smelters, the background concentrations around them were
assumed to be negligible
Based on the results of dispersion modeling, control strategies were
developed for the nine domestic primary copper smelters modeled. Table 5-22
shows the baseline ambient lead concentrations and predicted concentrations
after control for each plant. Total installed capital and annualized control
costs associated with achieving each alternative are presented in Table 5-23.
Costs for individual controls are given in Appendix D.
Six of the nine smelters modeled attained the lowest NAAQS alternative
with add-on controls for converter and/or furnace fugitive emissions. Of
these, one will achieve the lowest NAAQS alternative with baseline controls
they have installed or will install by 1987. For remaining three smelters,
controls could not be identified to achieve the lowest NAAQS alternative.
However, given the uncertainties in the conservative emission estimates for
the technologies employed at these smelters and in dispersion modeling, it is
possible that the plants would attain the lowest alternative using the
identified controls.
127
¦j

-------
TABLE 5-22. BASELINE EMISSIONS AND PREDICTED AMBIENT LEAD
CONCENTRATIONS AFTER CONTROL
Source
Baseline
emissions
(g/sec)
Baseline
ambient
concentration3
(ug/m^)
Predicted ambient
concentration
after control3'^
(ug/m^)
1
0.020
0.45
0.068c
2
0.37
1.2
0.056
3
0.40
4.2
0.28d
4
0.14
0.81
0.051(0.25e)
5
12.
0.36
0.14
6
1.1
4.3
0. 29d
7
1.3
2.4
0.015
8
0.33
4.6
0. 26d
9
0.059
0.58
0.036
aAnnual maximum mean quarterly ambient lead concentration.
^With added converter and furnace secondary emission controls.
cSource will achieve this with baseline controls by 1987.
dGiven the uncertainties in emission estimates for new technologies and
dispersion modeling, it is possible that the smelter would attain the
0.25 ug/m^ alternative using the identified controls.
eWith added converter secondary controls only.
128

-------
TABLE 5-23. NAAQS CONTROL COSTS FOR THE PRIMARY COPPER SMELTING INDUSTRY
Lead NAAQS	Number	Number of
alternatives	Industry control costs	of plants	plants unable
(ug/m^)	Capital	Annualized	impacted	to comply
2.5
26.0
6.8
3
2.0
28.9
7.3
4
1.5
44.2
11.6
4
1.0
49.7
13.2
5
0.8
50.6
13.5
6
0.5
61.6
16.5
7
0.25
75.2
20.0
8a
aOne source has baseline controls (existing process and fugitive controls
planned or agreed to be installed before 1987^7) which will bring it below the
lowest NAAQS alternative of 0.25 ug/m^. Baseline controls have no additional
costs associated with them are therefore are excluded from the cost analysis.
129

-------
5.12 TOTAL NATIONWIDE CONTROL COSTS
Tables 5-24 and 5-25 present estimated nationwide capital and annual costs
of attaining the various alternatives for stationary sources for two scenarios
for urban mobile source background lead concentrations. Table 5-24 gives
calculated stationary source costs based on the current lead-in-gasoline
phasedown program, which is expected to result in a reduction of the average
urban mobile source impact to 0.5 ug/m^. Table 5-25 gives stationary source
costs under the proposed total ban of lead in gasoline, under which the
average mobile source lead impact is predicted at 0.01 ug/m^ (Section 3).
Most of the costs for any of the alternatives would be borne by the metals
industries, especially the lead industry. All of the costs represent costs of
controls in addition to those already in place. Especially in the case of
primary and secondary lead smelters, substantial costs already have been
incurred in the implementation of the current standard. Also, Controls
installed to meet OSHA lead standards and EPA particulate emission limits
generally act to reduce lead ambient concentrations from lead-emitting
industries. Due to the implementation of control equipment regulations, there
may be an overlap in the installation of control equipment that is considered
to be "in place" and that required to attain the various lead NAAQS
alternatives.
The costs presented in Tables 5-24 and 5-25 and throughout this analysis
are incremental costs for attainment of the various NAAQS alternatives by
1990. The incremental costs for the two alternatives in which the lead NAAQS
is relaxed do not include any cost savings for discontinuing the use of
controls that were installed to meet the current standard; thus costs for
these alternatives are biased upward. Conversely, costs for the lower
alternatives are biased downward, because controls could not be identified
that would allow some plants to attain these alternatives. Some of the plants
for which costs were developed are currently uncontrolled. Controls for these
plants may be required to meet the particulate matter NAAQS. Thus, some of
the costs for all of the NAAQS alternatives may be partially attributable to
the particulate NAAQS.
130


-------
TAT1LE 5-24. NATIONWIDE NAAQS CONTROL COSTS FOR THE UPPER LTMTT MOBILE SOURCE BACKGROUND
Costs for rioted NAAQS alternative (millions)

2.5
2.0
!. 5
1.0
0.8
0.5
0.25
iPITAT, COSTS







Primary lead smelting
17.8
27.3
44.5
46.6
53.3
53.3
53.3
Secondary lead smelting
54.2
72.4
96.1
118.7
124.4
133.3
144.0
Lead ore processing
—
—
18.3
18.3
18.3
18.3
18.3
Lead-acid battery manufacture
1.6
2.1
2.1
3.5
4.1
5.2
6.5
Lead alkyl production
—
—
--
1.0
1.0
2.1
2. 1
Cray iron foundries
2.2
2.2
6.6
11.0
11.0
16.2
21.3
Alloy steel EAFs
—
—
—
—
—
6.8
8.1
Steel EAFs
—
—

—
0.7
1.5
2.5
Iron and steel sintering
6.7
6.7
13.4
13.4
13.4
13.4
23.6
Hrass and bronze production
0.4
0.7
0.7
1.4
2.3
3.6
5.6
Primary copper smelting
26.0
28.9
44.2
49.7
50.6
61.6
75.2
TOTAL
108.9
140.3
225.9
263.7
279.1
315.3
360.3
1NUAL COSTS







Primary lead smelting
4.5
6.7
11.9
12.6
14.5
14.5
14.5
Secondary lead smelting
16.4
22.0
28.9
35.5
36.8
38.5
40.8
Lead ore processing
—
—
7.3
7.3
7.3
7.3
7.3
Lead-acid battery manufacture
0.4
0.6
0.6
1.0
1.1
1.4
1.9
Lead alkyl production
--
—
—
0.4
0.4
0.7
0.7
Gray iron foundries
0.5
0.5
1.5
2.5
2.5
3.7
4.8
Alloy steel EAFs
—
—
—

—
1.8
2.5
Steel EAFs

—
—
--
0.3
0.7
1.2
Iron and steel sintering
1.8
1.8
3.7
3.7
3.7
3.7
6.6
Brass and bronze production
0.1
0.2
0.2
0.4
0.6
1.0
1.6
Primary copper smelting
6.8
7.3
11.6
13.3
13.5
16.5
20.0
TOTAL
30.6
39.1
65.7
76.6
80.7
89.8
101.8

-------
TABLE 5-25. NATIONWIDE NAAQS CONTROL COSTS FOR THE LOWER LIMIT MOBILE SOURCE BACKGROUND

Costs for noted NAAQS alternative (millions)

2.5
2.0
1.5
1.0
0.8
0.5
1
1
O 1
• 1
NJ 1
Ui 1
CAPITAL COSTS







Primary lead smelting
17.8
27.3
44.5
46.6
53.3
53.3
53.3
Secondary lead smelting
54.2
70.3
95.7
118.6
123.4
132.7
140.3
Lead ore processing
—
—
18.3
18.3
18.3
18.3
18.3
Lead-acid battery manufacture
1.6
2.1
2.1
3.5
3.8
4.8
6.3
Lead alkyl production
--
--
--
1.0
1.0
2.1
2. 1
Gray iron foundries
2.2
2.2
4.4
8.8
11.0
16.2
20.9
Alloy steel EAFs
—
—
—
—
--
--
8.1
Steel EAFs
—
--
—
--
0.7
1.5
1.9
Iron and steel sintering
6.7
6.7
13.4
13.4
13.4
13.4
23.6
Brass and bronze production
0.4
0.4
0.7
1.4
1.9
3.6
5.6
Primary copper smelting
26.0
28.9
44.2
49.7
50.6
61.6
75.2
'total
108.9
137.9
223.3
261.4
277.4
307.4
355.5
ANNUAL COSTS







Primary lead smelting
4.5
6.7
11.9
12.6
14.5
14.5
14.5
Secondary lead smelting
16.4
21.2
28. 7
35.4
36.5
38.3
40.1
Lead ore processing
—
--
7.3
7.3
7.3
7.3
7.3
Lead-acid battery manufacture
0.4
0.6
0.6
1.0
1.1
1.3
1.8
I.ead alkyl production
—
--
--
0.4
0.4
0.7
0.7
Gray iron foundries
0.5
0.5
1.0
2.0
2.5
3.7
4.7
Alloy steel EAFs
--
--
—

--

2.5
Steel EAFs
—
--
—
--
0.3
0.7
1.0
Iron and steel sintering
1.8
1.8
3.7
3.7
3.7
3.7
6.6
Brass and bronze production
0.1
0.1
0.2
0.4
0.5
0.9
1.6
Primary copper smelting
6.8
7.3
11.6
13.3
13.5
16.5
20.0
TOTAL
23.8
30.9
53.4
62.7
66.7
71.0
80. 7

-------
The attainment costs for 1995 are expected to be the same as those for
1990 (in constant dollars) for two reasons. First, new stationary source
costs associated with attainment of the lead NAAQS after 1990 will be
attributable to new source performance standards. Second, because the
background lead concentration around stationary lead sources by 1990 is
expected to be negligible in most cases, the attainment status of a given
plants will not be affected by changes external to the plant. As noted above,
the lead-in-gasoline phasedown program will reduce mobile source contributions
around stationary sources to negligible levels. Also, because of the
relatively small number of stationary lead sources, most plants are not
subject to background concentrations from neighboring plants.
5.13 UNCERTAINTY ANALYSIS
The results of the nationwide analysis presented above are subject to a
number of limitations. These result from uncertainty in input data and
simplifying assumption made as part of the general methodology. In addition,
dispersion modeling and costing of control equipment are subject to
uncertainties.
The main source of uncertainty in the input data is in the emission
factors and emission inventories used in the study. The data used in the
analysis were the best data available at the time of the analysis (1983 to
1984). However, because of the nature of most data lead emissions, emission
data for lead are subject to more uncertainty than data for other pollutants.
High ambient lead impacts generally are the result of fugitive lead emissions,
which are difficult to quantify. In addition, for some industries included in
this study, lead is present as a trace component in particulate emissions.
The composition of lead in this particulate matter may vary by as much as a
factor of ten between different sources of the same general type. As a result
of the implementation process for the current lead standard, the data base for
lead emissions is continuously being augmented and improved.
Simplifying assumptions made in the study can be characterized briefly as
follows:
133
•j

-------
•	average source-receptor coefficients were used for secondary lead
smelting, gray iron cupolas, and alloy steel EAFs;
•	model plants were used to identify affected facilities in lead-acid
battery manufacture and lead ore processing;
•	model plants were used to develop costs for all industries except
primary lead smelting and lead alkyl manufacture; and
•	an average mobile source background concentration was used.
Source-receptor coefficients were found to vary by as much as a factor of ten
for the same type of source located in different cities. Mobile source
backgrounds may vary from 0.0 ug/m^ to about 0.13 ug/m^. NAAQS costs were
developed for two mobile source backgrounds, 0.01 ug/m^ and 0.05 ug/m^. The
uncertainty associated with the model plant approach is difficult to quantify
but is expected to have less effect than uncertainties in the input data.
In addition to the above, costing of emission controls typically is
subject to an uncertainty of i50 percent. Dispersion modeling typically is
subject to an uncertainty of a factor of two.
The sensitivity of control costs to variations in input variables and
uncertainties in costing and modeling techniques can be illustrated by a graph
of cost versus NAAQS level. Figure 5-2 gives such a plot for the 0.01 ug/m^
urban mobile source background level. A doubling of emission factor or
source-receptor coefficient would have cost result similar to dividing the
NAAQS level by two. An increase in background concentration would have the
same effect as a corresponding decrease in NAAQS.
The overall uncertainty of the analysis is difficult to quantify. It is
not expected that the uncertainties in the costs for various industries will
be in the same direction. Rather, some uncertainties would be expected to be
cancelled by uncertainties in the opposite direction for other industries.
•j
134

-------
u>
Ln
£fi
a
o
•H
•HI
&
W
U
W
O
u
4)
•Q
•HI
g
o
U
nJ
S3
360
340
320
300
280
260
2+0
220
200
1 80
1 60
1 40
I 20
1 00
80 -
60 -
40
20
0
a
X
CAPITAL

'~CK
VQ
X
X
X.
O-
-^AHHUAL
S--—

-4-..
-+
0,4
l~	-J	[	|	p—
0,8	1.2	1.6
NAAQS Alternative
2.4
Figure 5-2.
Sensitivity of NAAQS costs to standard level.

-------
REFERENCES FOR SECTION 5
1.	The International Competitiveness of the U.S. Non-Ferrous Smelting
Industry and the Clean Air Act. Everest Consulting Associates, Inc.,
Princeton, NJ, and CRU Consultants, Inc., New York, NY. April 1982.
2.	Telecon. G. Viconovic, GCA/Technology Division, with W. Woodbury, U.S.
Bureau of Mines. November 15, 1983. The current status of the Bunker
Hill primary lead smelter.
3.	Mineral Commodity Summaries - 1983. U.S. Burea of Mines. January 1983.
A. Letter and attachments from Bradley Reynolds, Missouri Department of
Natural Resources, to Timothy Curtin, GCA/Technology Division.
December 8, 1982. Modeling inventories for Missouri's lead smelters.
5.	Letter and attachments from Bradley Reynolds, Missouri Department of
Natural Resources, to Timothy Curtin, GCA/Technology Division.
December 22, 1982. Dispersion modeling inputs for Missouri's lead
smelters.
6.	Letter and attachments from Richard Daye, EPA Region VI, to Timothy Curtin,
GCA/Technology Division. December 13, 1983. Dispersion modeling input and
results for Missouri's lead smelters.
7.	State of Texas Lead SIP Analysis (Draft), U.S. Environmental Protection
Agency, Dallas, TX. October, 1981.
8.	Memo from Bob Kellam, EPA/ASB, to John Haines, EPA/ASB. Visit to Asarco El
Paso Smelter and Meetings with Region VI and TACB. October 5, 1982.
9.	Energy Technology Consultants. Identification of the Sources of Total
Suspended Particulates and Particulate Lead in El Paso Area by
Quantitative Microscopic Analysis. Texas Air Control Board, Austin, TX.
1983.
10.	Houck, J.E., et al. Particulate Source Apportionment Analysis Using the
Chemical Mass Balance Receptor Model. Prepared by NEA, Inc.,
Beaverton, OR, for the Montana Department of Health and Environmental
Sciences. 1982.
11.	Balentine, H.W., et al. Modeling Analysis of Ambient Lead Levels Around
a Primary Lead Refinery. Radian, Austin, TX. December 19, 1983.
136

-------
12,
13,
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Letter from W. Woodbury, U.S. Bureau of Mines, to J.	Haines, U.S.
Environmental Protection Agency. October 15, 1982.	World Directory:
Secondary Lead Plants, compiled by the International	Lead/Zinc
Conference.
Compilation of Air Pollutant Emissions Factors, Third Edition, and
Supplements 1 through 14 (AP-42). U.S. Environmental Protection Agency,
Research Triangle Park, NC. 1983. Section 7.11. 	
Preliminary Study of Inorganic Sources of Arsenic (EPA-450/5-82-005).
U.S. Environmental Protection Agency, Research Triangle Park, NC. 1982.
Schwitzgabel, K., et al. (Radian Corp.). Fugitive Emissions at a
Secondary Lead Smelter. U.S. Environmental Protection Agency, Research
Triangle Park, NC. December 1981.
National Emissions Data System (NEDS). U.S. Environmental Protection
Agency. December 1982.
Mineral Facts and Problems - Lead. 1980 Edition. U.S. Bureau of Mines.
1981.
Metallic Mineral Processing Plants - Background Information for Proposed
Standard (EPA-450/3-81-009a). U.S. Environmental Protection Agency,
Research Triangle Park, NC. 1982.
Lead - Mineral Commodity Summary. U.S. Bureau of Mines. December 1977.
Minerals and Materials, A Bimonthly Survey. U.S. Bureau of Mines,
Washington, DC.
Engineering & Mining Journal. March 1982 and January 1983.
Reference 13. Section 7.18 and Appendix E.
Lead-Acid Battery Manufacture - Background Information for Proposed
Standards (EPA-450/3-79-028a). U.S. Environmental Protection Agency,
Research Triangle Park, NC. 1979.
Telecon. W. Battye, GCA/Technology Division, with R. Burkhard, Battery
Council International Marketing Committee. April 16, 1983. Growth in
the lead-acid battery industry.
Reference 13. Section 7.15.
Reference 13. Section 5.22.
Reference 13. Section 7.10
137

-------
28.	1983 Industrial Outlook. U.S. Department of Commerce. January 1983.
29.	Federal Reserve Statistical Release: Capacity Utilization:
Manufacturing, Mining, Utilities, and Industrial Materials. G.3(402) The
Federal Reserve System, Washington, DC. November 16, 1983.
30.	Control Techniques for Lead Air Emissions - Volume II: Chapter 4 -
Appendix B (EPA-450/2-77-012). U.S. Environmental Protection Agency,
Research Triangle Park, NC. December 1977.
31.	Electric Arc Furnaces in Ferrous Foundries - Background Information for
Proposed Standards (EPA-450/3-80-020a). U.S. Environmental Protection
Agency Research Triangle Park, NC. 1980.
32.	Electric Arc Furnaces and Argon—Oxygen	Decarburization for Vessel in
Steel Industry - Background Information for Proposed Revision to Standards
(EPA-450/3-82-02a). U.S. Environmental	Protection Agency, Research
Triangle Park, NC. 1983.
33.	EPRI Journal. Electric Power Research Institute. June 1982.
34.	A Review of Standards of Performance for New Stationary Sources -
Secondary Brass and Bronze Plants. U.S. Environmental Protection Agency,
Research Triangle Park, NC. Publication No. EPA-450/3-79-011. June 1979.
35.	Personal Communication between Brass and Bronze Ingot Institute and
Michael Clowers of GCA, April 10, 1985.
36.	Compilation of Air Pollutant Emission Factors. U.S. Environmental
Protection Agency, Research Triangle Park, NC. Publication No. AP-42.
July 1979.
37.	U.S. Environmental Protection Agency. Inorganic Arsenic Emissions from
Low-Arsenic Primary Copper Smelters. Background Information for Proposed
Standards. U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Research Triangle Park, North Carolina.
April 1983.
38.	Telecom. Janice J. Jolly, Physical Scientist, Division of Nonferrous
Metals, U.S. Bureau of Mines, with Nancy Browne, GCA/Technology.
6 June 1985.
39.	Telecom. Dan Edelstein, Physical Scientist, Division of Nonferrous
Metals, U.S. Bureau of Mines, with Nancy Browne, GCA/Technology,
17 June 1985.
40.	U.S. Department of the Interior. Copper. Preprint from Bulletin 675,
Mineral Facts and Problems. Bureau of Mines, U.S. Department of the
Interior, Washington, D.C. 1985.
138

-------
41.	Compilation of Air Pollutant Emission Factors. U.S. Environmental
Protection Agency, Research Triangle Park, NC. Publication No. AP-42.
July 1979.
42.	U.S. Environmental Protection Agency. Industrial Source Complex (ISC)
Dispersion Model User's Guide, Volume EPA-450/4-79-030. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. 1979.
43.	U.S. Environmental Protection Agency. National Emissions Data System
(NEDS) Office of Air Quality Planning and Standards, Research Triangle
Park, NC.
139
¦)

-------
6. JEW SGURCE CONTROLS
The costs of meeting alternative ambient standards for lead are the
primary focus of this costs assessment. However, another perspective can be
gained by comparing these costs to the costs of meeting new source performance
standards. Costs of meeting new source performance standards (NSPS) were
assessed for the lead-acid battery manufacture industry and for alloy steel
electric arc furnaces. These are the only industry categories affected by
the NAAQS for which growth in capacity was projected (Section 4.8).
Tables 6-1 and 6-2 summarize NSPS costs for lead-acid battery manufacture
and alloy steel electric arc furnaces, respectively. These were developed
based on projections of growth in Section 4.8 and costs estimates given in the
background information documents for the NSPS. Data used to derive the NSPS
costs for lead-acid battery manufacture and alloy steel EAF's are presented in
Appendices E and H, respectively.
Comparison of Tables 6-1 and 6-2 with Table 5-14 indicates that total
NAAQS costs are higher than NSPS costs for lead-acid battery manufacture. For
alloy steel EAF's, total NSPS costs are the higher. In both cases, per-plant
NAAQS costs and NSPS costs are roughly comparable.
140

-------
TABLE 6-1. NSPS COSTS FOR THE LEAD-ACID BATTERY
INDUSTRY (OCTOBER 1984 DOLLARS)
SS = = = SSSSS = = = SS=S = SZSSS5S = SSSSS=SS=!=S==SS = S===SS = S=
Number	Total costs (thousands)
o f new		
Year	plants	Capital	Annual
1990
826
249
1991
826
249
1992
826
249
1993
826
249
1994
826
249
1995
826
249
Size of each new plant is estimated to be
10,000 batteries per day.
TABLE 6-2. NSPS COSTS FOR ALLOY STEEL ELECTRIC
ARC FURNACES (OCTOBER 1984 DOLLARS)
;ssss;====;=:=:s=;s;s:s:==ss=:ss=ss=s=s=ss:ss:5==;=:srss=!=:s=s2ssss==ss;==:=ss
Number Total costs (millions)
o f new		"	
Year
plants
Capital
Annual
1983
5
32.0
12.0
1984
4
25.7
9.1
1985
4
24.4
9.2
1986
3
18.0
6.3
1987
4
24.4
9.1
1988
3
18.0
6.3
1989
4
24.4
9.1
1990
3
18.0
6.3
1991
3
19.2
7.4
1992
2
12.8
4.6
1993
2
12.8
4.6
1994
2
10.5
3.4
1995
2
10.5
3.4
141

-------
7. COSTS OF WASTE OIL COMBUSTION REGULATIONS
In January 1985, EPA issued a proposed rule that would extend the
hazardous waste management regulations to apply to marketing of and burning
for energy recovery of hazardous wastes and used oils. This rule would
establish maximum contaminant level and minimum flash point specifications for
used oil burned in nonindustrial boilers. Burning of off-specification fuel
oil would be prohibited, because emissions from such burning are potentially
hazardous to human health and the environment. In addition to regulating
actual burning, EPA is proposing administrative requirements that create a
tracking system from initial waste fuel marketers, through distributors, to
burners, including notification, receipt of identification number, compliance
with manifest or invoice system, and recordkeeping.^
This rule is being proposed because used oil contains a number of toxic
contaminants, including toxic metals and a variety of chlorinated solvents and
other organic chemicals. One of the metals of concern is lead. Typically,
used oil contains a median concentration of 240 ppm lead, and a
90th percentile concentration of 1200 ppm; both these values are more than
two orders of magnitude greater than the analagous values for virgin fuel
oils. To determine the ambient lead contribution of used oil burning, EPA
conducted a modeling study of a hypothetical urban area where used fuel oil
containing 1000 ppm lead was being burned across the city in residential,
institutional, and commercial boilers. This study showed that the ambient
lead concentrations associated with used oil burning were over 0.7 ug/m^ in
certain sections of the study area. This value is almost three times the
lowest NAAQS being considered (0.25 ug/m^), and about 50% of the current
standard (1.5 ug/m-*).!
Based on other modeling studies, EPA believes that a lead specification
level for used oil of 100 ppm would ensure that ambient lead levels in the
vicinity of both large and small single and clustered sources would be well
below the current standard of 1.5 ug/m-*. However, a 10 ppm level is also
being discussed, because the current NAAQS is being revised downward; this
142
i

-------
10	ppm specification level would effectively preclude most burning of used
fuel oil, because only 30% of used oil can meet this lead level. If the
Agency does begin regulating the burning of used oil as fuel, the new NAAQS
will have to be considered when choosing the specification level for lead. *
EPA has estimated the national costs of implementing the hazardous
waste/used oil burning rules for nonindustrial boilers, as shown in Table 7-1;
note that the annualized costs attributable to used oil regulations alone
would be less than the total costs given in the table. The estimates are
based on worst-case conditions with respect to a number of marketers and
burners subject to regulation; thus, the costs more closely approximate
maximum rather than actual costs. However, this upward bias is negated at
least in part, because EPA did not attempt to determine and include the costs
of two important impacts of the proposed rule: (1) the cost to marketers of
treating used oil to meet the used oil fuel specification, and (2) the costs
to nonindustrial boiler owners of having to buy specification used oil fuel or
commercial fuel oil instead of hazardous waste fuel or off-specification used
011	fuel. The final specification levels set for lead and other toxic
contaminants in used oil will in part determine these costs. *
¦j
143

-------
TABLE 7-1. NATIONWIDE COSTS OF HAZARDOUS WASTE AND USED OIL COMBUSTION
REGULATIONS (1984 dollars)a>b
Nationwide costs
(thousand ofdollars)
Requirement	One-time Recurring Total
capital annual annualized
Notification
1691
0
113
Manifest system (hazardous waste)
20
379
380
Invoice system (used oil)
78
17213
. 17218
Certification to suppliers
900
307
367
Used oil analysis



Specification oil
0
328
328
Off-specification oil
0
2024
2024
Storage
3476
1061
1294
TOTAL
6165
21312
21724
aThese costs more closely approximate maximum rather than actual costs.
^These costs do not include the costs to marketers of treating used oil to
meet specifications, nor the costs to nonindustrial boiler owners of not being
able to use hazardous waste fuel or off-specification used oil fuel.
144

-------
REFERENCES FOR SECTION 7
1_ Hazardous Waste Management System: Standards for the Management of
Specific Wastes and Specific Types of Waste Facilities. Federal
Register. 50:1684-1724. January 11, 1985.
145
¦)

-------
APPENDIX A
LEAD EMISSIONS INVENTORY
146 |
¦)

-------
TABLE A-l. ESTIMATED ATMOSPHERIC LEAD EMISSIONS
FOR THE UNITED STATES, 1981
Annual emissions
Source category	(Mg/yr)
Mobile Sources
Gasoline combustion2	35,000
Stationary Sources
Waste oil combustion*3	830
Solid waste incinerationc	319
Coal combustion**	950
Oil combustione	243
Ore crushing and grinding^	325
Primary lead smeltingB	921
Secondary lead smelting*1	631
Iron and steel production1	533
Gray iron productionJ	295
Primary copper smeltingk	30
Other metallurgical^"	54
Lead alkyl manufactured	245
Lead acid battery manufacture11	85
Portland cement production0	71
Miscellaneousp	223
TOTAL	40,700
a.	References 1 and 2. Based on total gasoline sales of 100 billion gal EPA
weighted lead content requirements in effect in 1981 (0.5 g/gal), and a
lead emission rate of 70 percent of the lead in gasoline.
b.	Reference 4, and 39. Based on total annual waste oil combustion of
370 million gallons (References 3 and 4), an average lead content of
1200 ppm (Reference 39), and an emission factor of 50 percent of the lead
present in the waste oil (Reference 4).
147 i

-------
c.	Reference 5 and 6. Based on total annual solid waste combustion of
6.76 Tg (Reference 5), an emission factor of 0.2 g/kg (Reference 6), and
an average emission control efficiency of 76 percent (Reference 6).
d.	References 6 and 10. Based on coal consumption by sector from Reference 7,
control technology application rates and lead emission factors for various
control techniques from References 8 through 10, and uncontrolled lead
emissions from Reference 6.
e.	References 4, 6, 7, and 8. Based on residual and distillate oil consumption
by sector from Reference 7; average residual and distillate oil lead contents
of 3.5 ppm (Reference 8) and 0.1 ppm (Reference 6), respectively; an emission
factor of 50 percent of the oil present in the fuel (Reference A); and control
technology application rates and control efficiencies from Reference 9.
f.	References 4, 6, 11, and 12. Based on total ore production for 1981 of
320 Tg for the copper, lead, and zinc ore processing industries. The
percentage breakdown of the type of ores processed was obtained from
Reference 4 and the appropriate Pb emission factor for each ore type was
obtained from Reference 6. All ore processed was assumed to be dry ore.
g.	Reference 13 through 21. Based on estimates of point and fugitive emissions
developed by Federal, State, and local air pollution agencies in response
to State Implementation Plan requirements.
h.	References 4, 22, and 23. Based on NEDS listing of total particulate
emissions from secondary lead smelters and a lead content of 23 percent
for the particulate (Reference 4). Total lead emissions were prorated
upward for the percent of secondary lead smelter capacity not listed in
NEDS.
i.	References 4, 5, 6, and 24 to 26. Based on raw steel production of 108 Tg
for 1981 (Reference 24) and process inputs obtained for Reference 5.
Emission control efficiencies for the various process sources were obtained
from References 4 and 6.
j. References 5, 6, and 27 to 29. Based on gray-iron shipments of 9 Tg for
1981 (Reference 27) and a production breakdown for cupola, electric
induction, and reverberatory furnaces from Reference 28. The lead emission
factors used for this industry are averages from Reference 6. Average
industry emission control efficiency of 98 percent for cupolas and 0 percent
for electric induction furnaces and reverberatory furnaces were based on
data from Reference 5 and 29.
k. References 5, 6, 24, and 30. Based on an estimated total copper concentrate
input for 1981 of 6.3 Tg (References 24 and 30). Pb emission factors for
roasting, reverberatory furnace, and converter operations were obtained
from Reference 6 and overall particulate emission control efficiency of
96 percent was based on References 5 and 6.
1. References 4, 5, 6, 31, 32, 33 and 34. Includes primary zinc smelting,
ferroalloy production, and brass and bronze production.
148 :
¦)

-------
References 13 and 35. Based on estimates of point and fugitive emissions
developed by Federal, State, and local air pollution agencies in response
to State Implementation Plan requirements.
References 6, 22, and 36. Based on uncontrolled' emission rates from
Reference 6, weighted average control technology application rates and
efficiencies from Reference 22, and a battery production rate of 70 million
units (References 32 and 36).
References 6, 22, and 37. Total lead emissions were calculated by summing
particulate emissions tabulated in the NEDS system for portland cement
plants and applying an average particulate lead content of 450 ppm
(Reference 22).
References 4, 6, 32, and 38. Includes the manufacture of glass, lead
pigments, lead oxides not produced at battery plants, type metal,
ammunition, and other lead products.
149

-------
REFERENCES FOR APPENDIX A
1.	Petroleum Supply Monthly. DOE/EIA-0109(82106), Energy Information
Administration, U.S. Department of Energy, Washington, DC, June 1982.
2.	Supplementary Guidelines for Lead Implementation Plans Revised
Section 4.3 (Projecting Automotive Lead Emissions). EPA-450/2-78-038a,
U.S. Environmental Protection Agency, Research Triangle Park, NC,
July 1979.
3.	Hess, L.Y. Reprocessing and Disposal of Waste Petroleum Oils. Noyes
Data Corp., Park Ridge, NJ, 1979.
4.	Control Techniques for Lead Air Emissions. EPA-450/2-77-012, Office of
Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC, 1977.
5.	GCA/Technology Division. Survey of Cadmium Emission Sources.
EPA-450/3-81-013, U.S. Environmental Protection Agency, Research Triangle
Park, NC, 1981.
6.	Compilation of Air Pollution Emission Factors, Third Edition (Including
Supplements 1 through 7) and Supplements 8 through 13. AP-42,
U.S. Environmental Protection Agency, Research Triangle Park, NC, 1982.
7.	State Energy Data Report, 1960 through 1980. DOE/EIA-0214(80), Energy
Information Administration, U.S. Department of Energy, Washington, DC.,
July 1982.
8.	Shih, C.C., R.A. Orsini, D.G. Ackerman, R. Moreno, E. Moon, L.L. Scinto,
and C. Yu. Emissions Assessment of Stationary Combustion Systems; Volume III:
Electricity Generation External Combustion Soruces. EPA-600/7-81-003a,
US. Environmental Protection Agency, Research Triangle Park, NC, 1981.
9.	Surprenant, N.F., W. Battye, D. Roeck, and S.M. Sandberg. Emissions
Assessment of Stationary Combustion Systems; Volume V: Industrial
Combustion Sources. EPA-600/7-81-003c, U.S. Environmental Protection
Agency, Research Triangle Park, NC, 1981.
10. Surprenant, N.F., P. Hung, R. Li, K.T. McGregor, W. Piispanen, and
S.M. Sandberg. Emissions Assessment of Stationary Combustion Systems;
Volume IV: Commercial/Insitutional Combustion Sources. EPA-600/7-81-003b,
U.S. Environmental Protection Agency, Research Triangle Park, NC, 1981.
150

-------
11. Engineering & Mining Journal. January 1982.
12.	Metallic Mineral Processing Plants - Background Information for Proposed
Standards. EPA-450/3-81-009a, U.S. Environmental Protection Agency,
Research Triangle Park, NC, 1982.
13.	State of Texas Lead SIP Analysis (Draft), U.S. Environmental Protection
Agency, Dallas, TX, October, 1981.
14.	Memo from Bob Kellam to John Haines, Visit to Asarco El Paso Smelter
and Meetings with Region VI and TACB, October 5, 1982.
15.	Letter and attachments from Bradley Reynolds, Missouri Department of
Natural Resources, to Timothy Curtin. GCA/Technology Division,
December 8, 1982, modeling inventories inputs for Missouri's lead
smelters.
16.	Letters and aattachments from Bradley Reynolds, Missouri Department of
Natural Resources to Timothy Curtin, GCA/Technology Division,
December 22, 1982. Dispersion modeling input for Missouri's lead
smelters.
17.	Letter and attachments from Richard Daye, EPA Region VI to Timothy Curtin,
GCA/Technology Division. December 13. Dispersion modeling input and
results for Missouri's lead smelters.
18.	PES, Inc. Emission Measurements at the Asarco Lead Smelter in E. Helena
MT, May 1980.
19.	ETA Engineering, Inc. Technical Support Document: Analysis of Air
Quality Levels in E. Helena, MT. November, 1980.
20.	Letter and Attachments from Gene Robinson, State of Nebraska, Department
of Environmental Control, to Timothy Curtin, GCA/Technology Division,
January 10, 1982. State Implementation Plan for Lead.
21.	Letter and Attachments from Robert Timmerman, City of Omaha Housing
and Community Development Department, to Timothy Curtin, GCA/Technology
Division, January 7, 1982, Emission Inventory for Asarco, Omaha, NB.
22.	National Emissions Data System. U.S. Environmental Protection Agency,
Research Triangle Park, NC, 1982.
23.	World Directory: Secondary Lead Plants. Received from William Woodbury,
U.S. Bureau of Mines. October 15, 1982.
24.	Mineral Commodity Summaries 1982. U.S. Bureau of Mines, Janaury 1982.
25.	Mineral Commodity Profiles: Iron Ore. U.S. Bureau of Mines, May 1978.
26.	Basic Oxygen Process Furnaces - Background Information for Proposed
Standards. U.S. Environmental Protection Agency, December 1981 (Review
Draft).
151
¦>

-------
27.	Gray Iron Foundry Shipments. Foundry Technology Management. April 1982.
28.	A Method for Characterization and Quantification of Fugitive Lead Emissions
from Secondary Lead Smelters, Ferroalloy Plants and Gray Iron Foundries
(Revised). EPA-450/3-78-003 (Revised), U.S. Environmental Protection Agency,
August 1978.
29.	Hazardous and Trace Substances Emission System. U.S. Environmental Protection
Agency, Research Triangle Park, NC, 1982.
30.	The International Competitiveness of the U.S. Non-Ferrous Smelting Industry
and the Clean Air Act. American Mining Congress. April 1982.
31.	1981 U.S. Industrial Outlook. U.S. Department of Commerce, Washington, DC,
1981.
32.	Preprint for Lead from Minerals Yearbook 1981: Volume I - Metals and
Minerals. U.S. Bureau of Mines, Washington, DC, 1983.
33.	Electric Submerged Arc Furnaces for Production of Ferroalloys - Background
Information for Proposed Standards. EPA-450/2-74-018a, U.S. Environmental
Protection Agency, Research Triangle Park, NC, 1974.
34.	Secondary Brass and Bronze Ingot Production Plants - Background Information
for Proposed Standards. APTD-1352, U.S. Environmental Protection Agency,
Research Triangle Park, NC, 1973.
35.	State of Louisiana Lead SIP Analysis (Draft). U.S. Environmental Protection
Agency, Dallas, TX, 1981.
36.	1982 Battery Council International Statistics Annual. Battery Council
International, Chicago, IL, 1982.
37.	Mineral Industries Surveys - Directory of Cement Producers in 1982.
U.S. Burea of Mines, Washington, DC, 1982.
38.	Glass Manufacturing Plants - Background Information for Proposed Standards.
EPA-450/3-79-005a, U.S. Environmental Protection Agency, Research Triangle
Park, NC, 1979.
39.	Brinkman, D.W., P. Fennelly, and N. Surprenant. The Fate of Hazardous Wastes
in Used Oil Recycling - Paper presented at the Conference on Measurements
and Standards for Recycled Oil -IV, Gaithersburg, Md. (National Bureau of
Standards), September 14-16, 1982.
152

-------
APPENDIX B
RAW DATA USED IN
PREDICTING MOBILE
SOURCE IMPACTS
153

-------
TABLE B-l. RAW DATA AND SITE PARAMETERS FOR MICROSCALE NAMS
Sita nuaber
Site
type
Probe
height
Uefcsrs)
Avg. roadway
vol. (1000
vehicles
per day)
Distance
fro# road
(aeters)
Year
Total lead
burned (Ggl
Annual
average
concsntraticn
(us/a3)
Haxiaus
quarterly
concentration
(uq/a3)
01Q380025S01
H1C
2
30
9
1983
53
0.86
1.04
(qtr, 4)
053900001101
HIC
5
25
6
1981
1982
1983
59
60
53
1.1
0.9
0.7
1.5
1.6
1.1
(qtr, 1)
(qtr. 4)
(qtr. 1)
100860024801
MC
3
100
15
1183
53
1.15
1.39
(qtr. 4)
101960084H01
HIC
3
104
14
1981
1982
1983
59
60
53
1.01
1.15
1.02
1,41
1,71
1.15
(qtr. 4)
(qtr. 4)
(qtr. 45
15J730009K01
HIC
3
75
12
1983
53
0.40
0.43
(qtr, 2)
210I20045F01
HIC
2
64
13
1982
1983
60
53
0.67
0.44
0.85
0.60
(qtr. 4)
(qtr, 4)
220240002F01
HIC
5
35
7
1983
53
0.51
0.59
(qtr. 3)
243C80004F01
HIC
2
69
7
1981
1982
1983
59
60
53
0.75
0.50
0.63
0.89 '
0.57
0,70
(qtr. 3)
(qtr. 4)
(qtrs. 3,4)
31Q9600O2F01
HIC
2
n
14
1983
53
0.73
0.87
(qtr, D
320040022H02
SIC
3
35
7
1983
53
0.74
0,95
(qtr. »
33DA60017FOI
HIC
2
78
12
1983
53
0.64
0.69
(qtr. 4)
334680052F01
HIC
6
150
10
1983
53
0.73
0.90
(qtr. 4)
372I0001SFC2
HIC
3
15.
11
1982
1983
60
53
0.40
0.44
0.40 Cqfcrs. 1,3,4)
0,50 (qtr. 4)
381460082F01
HIC
4
132
14
1982
1983
60
53
1.32
1.22
1.63
1.37
(qtr. 3)
(qtr. 2)
39714Q047HQ2
HIC
4
36
9
1903
53
0.36
0.41
(qtr. U
402140003F01
HIC
3
64
15
1983
53
1.22
1.30
(qtr. 4i
451700027F01
HIC
3
16
5
1981
1982
1983
59
60
53
1.5
1.1
1.3
1.7
1.5
1.5
(qtrs. 2,4)
(qtr, I)
(qtr. 4)
154

-------
TABLE B-2. RAW DATA AND SITE PARAMETERS FOR MIDDLE SCALE NAMS
Sits nuaber
Site
typs
Probe
height
(aeters)
ftva, roadway
vol. (1000
vehicles
per day)
Distance
•froa road
(aeter5>
Year
Total lead
burned (Gg)
Annual
average
concentration
(ug/«3)
Haxiaui
quarterly
concentration
(«g/i3J
030400017S01
HID
4
19
25
1983
53
0.70
1.08
(qtr. 1)
056400003F01
HID
4
40
25
1975
1980
1981
174
75
59
1.84
1.02
0.74
2.33
1.46
1.00
(qtr. 4)
tqtr, 4)
(qtr. 4i
054800004101
RID
12
&
10
1975
I960
1981
1982
174
75
59
60
• 1.01
0.47
0.56
0.45
1.51
1.22
0.90
0.81
(qtr. 1)
(qtr. 4)
(qtr. 4)
(qtr. 4)
056980004101
HID
4
10
14
1975
1930
1981
1982
1983
174
75
59
60
53
1.24
0.66
0.37
0.50
0.34
1.99
0.94
0.47
0.8S
0.45
(qtr. IS
(qtr. 1)
tqtrs. 1,41
(qtr. 4)
(qtr. 1)
057160001101
HID
4
15
19
1981
1982
1983
59
60
53
0.24
0.35
0.28
0.30
0.55
0.35
(qtr. 45
(qtr. 4)
(qtr. 4)
070420014F01
HID
13
18
12
1981
59
0.37
0.47
(qtr. I!
090020027101
HID
10
34
33
1983
.. 53
0.23
0.32
(qtr. 4)
101840001SOI
HID
4
38
37
1983
53
0.31
0.41
(qtr. 1)
141220026H01
HID
a
177
15
1983
53
0.45
0.52
(qtr. 1)
222160007F01
HID
4
81
51
1983
53
0.76
0.99
(qtr. 4)
261040002501
HIS
13
55
43
1982
1983
60
53
0.45
0.35
0.51
0.40
(qtr. 4)
(qtr. 1)
262330002H01
HID
4
14
54
1982
1983
60
53
0.22
0.21
0.28
0.33
(qtr. 1)
(qtr. 1)
402160004F01
HID
4
12
16
1983
53
0.47
0.56
(qtr. 4)
410300007F01
HID
5
90
23
1983
53
0.69
0.92
(qtr. 4)
451310029H01
HID
7
79
24
1982
1983
40
53
0.29
0.29
0.35
0.37
(qtr. 1)
(qtr. 1)
451I80003F01
HID
7
15
21
1983
53
0.6
1.0
(qtr. 1)
451SS0023F01
HID
6
70
50
1983
53
0.7
0.9
(qtr. 1)
512200099F01
HID
13
16
23
1983
53
0.29
0.32
(qtr. 1)
i
155

-------
TABLE B-3. RAW DATA AND SITE PARAMETERS FOR NEIGHBORHOOD SCALE NAMS
Avg, roadway
Probe vol. (1000
Sits height vehicles
Sits nusber typs (aetars) per day)
Distance
froa road	Total lead
(asters) Year turned (Gg)
•Annual	Haxiaui
average	quarterly
concentration concentration
(ug/»3!	(ug/a3)
030600013G01 NEI
13
IS
25
052780001101 NEI
054180103101 NEI
056535001101 NEI
OS5360004I01 NEI
101760001801 NEI
141220Q52K01	NEI
151520014H01	NEI
152040058H01	NEI
172340001F01	NEI
220380002F01	NEI
222160011F01	NEI
24226G050F01	NE!
260030001B01	NEI
281I30019S01	NEI
312320003F01	NEI
32C040019K01	NEI
13
12
4
10
15
5
12
14
4
4
11
11
24
18
53
140
25
70
11
75
46
15
33
64
33
130
87
38
15
12?
120
300
83
80
150
106
250
100
81
92
12
1975
1976
1977
1978
1979
1980
1981
1982
1983
1981
1982
'1983
1981
1982
1983
1975
1980
1981
1982
1983
19B1
1982
1983
1976
1977
1970
1979
1980
1982
1983
1981
1983
1983
1983
1983
1983
1981
1982
1983
1983
1912
1983
1981
1983
174
168
162
146
121
75
59
60
53
59
60
53
59
60
53
174
75
59
60
53
60
53
168
162
146
121
75
60
53
59
53
53
53
53
60
53
53
53
60
53
59
53
1.57
1.09
1.53
1.16
0.55
0.69
0.56
0.45
0.36
0.25
0.28
0,21
1.0
0.7
0.6
1.16
0.58
0.48
0.41
0.38
0.28
0.28
0.26
0.95
0.86
0.83
0.71
0.58
0.37
0.49
0.34
0.30
0.38
0.14
0.31
0.42
0.27
0.17
0.24
0.09
0.54
0.39
0.45
0.35
2.18
1.91
2.76
1.64
1.95
1.06
0.94
0.69
0.54
0.37
0.42
0.30
1.3
1.0
0.9
1.46
0.84
0.64
0.4?
0.53
0.35
0.37
0.29
1.21
0.99
0.99
0.84
0.79
0.40
0.57
0.45
0.39
0.47
0.15
0.40
0.55
0.39
0.20
0.29
0.10
0.70
0.42
0.64
0.49
Cqtr. 21
(qtf. 4)
(qtr. 4)
(qtr. 1!
(qtr, 4)
(qtr. 4)
(qtr. 4)
(qtr. 4)
(qtr, 11
(qtr. 4)
(qtr. 4)
(qtr. 1)
(qtr. 1)
(qtr. 4)
(qtr. 1)
(qtr.	4)
(qtr.	41
(qtr.	4)
(qtr.	4)
(qtr.	11
(qtr. 4)
(qtr. 4!
(qtr. 4)
(qtr. 3)
(qtr. 4)
(qtr. 1)
(qtr, 3)
(qtr. 1)
(qtr. 3)
(qtr. I)
(qtr, 4)
(qtr. 3)
(qtr. 3)
(qtrs. 1,3!
(qtr. 4)
(qtr. 4)
(qtr. 41
(qtr. 41
(qtr. 1)
(qtr. 21
(otr. 3!
(qtrs. 3,41
(qtr. 11
(qtr. 11
(continued!
156

-------
TABLE B-3. (con'd.)
Sits nuaber
Site
type
Probe
height
(aetars)
Avg. roadway
vol. (1000
vehicles
per day)
Distance
froi road
Uetsrs)
Year
Total lead
burned (Gg)
Annual -
average
concentration
(ug/a3)
Haxiaui
quarterly
concentration
(ug/a5)
3346B0079F01
NE!
14
30
117
1983
53
0.36
0.47
(qtr. 4)
372200033F01
m
15
9
400
1982
1963
60
53
0.14
0.13
0.19
0.14
(qtr. 1)
(qtr, 1)
381460082F01
NEI
4
19
150
1982
1983
60
53
0.19
0.21
0.23
0.28
(qtr. 1)
(qtr. 4)
390400002501
NEI
S
27
95
1981
1982
1983
59
60
53
0.25
0.29
0.22
0.31
0.40
0.23
(qtr. 2)
(qtr. 4)
(qtrs. 1,3,4)
397140049K01
NEI
8
75
150
1983
53
0.46
0.57
(qtr. 3)
402160005F01
NE!
8
49
100
1983
53
0.19
0.25
(qtr. 4)
410300017F01
NEI
14
123
152
1983
53
0.42
0.60
(qtr. 4)
451310046H01
NEI
4
59
20
1982
60
0.34
0.44
(qtr. 4)
45170001CS01
NEI
5
8
32
1982
1983
60
53
0.26
0.21
0.33
0.30
(qtr. 1)
(qtr. 1)
454570034F01
NEI
15
50
100
1981
1983
59
53
0.5
0.4
0.7
0.6
(qtr. 1)
(qtr. 1)
483240C05F01
NEI
9
45
105
1982
1983
60
53
0.19
0.17
0.30
0.24
(qtr. 4)
(qtr. 4)
157

-------
APPENDIX C
SAMPLE COST CALCULATION
158

-------
appendix c i
SAMPLE COST CALCULATION
Where possible, control costs for plants of various sizes in a given industry
were taken from published sources and simply updated to 1982 dollars. For some
industries, however, cost data were incomplete or not available. In these
cases, the needed cost components were estimated using standard cost factors
or equations. In order to illustrate the use of these factors and equations,
this appendix presents a sample cost calculation. The sample calculation is
for a fabric filter applied to a three-process operation and a paste mixing
facility at a 10,000 battery per day bpd) lead-acid battery plant. Because
no cost data were available for the 10,000 bpd plant, this example will give
an illustration of all of the various cost estimation techniques used.
AIRFLOW RATE
The first requirement to estimate control costs is the flow rate through
the control device. The required ventilation rates for the 10,000 bpd three process
operation and paste mixing facility were estimated based on typical airflow rates
for smaller facilities. Table B-l lists ventilation rates given for 500, 2000,
and 6500 bpd facilities in the background information document for lead-acid
battery manufacture new source performance standards (NSPS)It was assumed
that the required facility ventilation rate is related to hourly production
capacity by a power law:
V = aCb	(1)
where a, b = constants
3
V = facility ventilation rate (m /min)
C = production capacity (batteries/hr or bph)
Given the ventilation rates for two facility sizes, the constants a and b can be
determined as follows:
159
¦>

-------
a = V/Cb	(3)
where subscripts a and b refer to specific plant sizes
Based on the data given in Table C-lj the following relationships can be developed
for the three process operation and the paste mixing facility ventilation rates.
for 62.5 < C< 125.0 bph:
V3P = (34.16)C(0,635)	(4)
VpM = (13.07)C(0\398)	(5)
for 125.0 < C< 270.0 bph:
V3p = (7.816)C(0,-941)	(6)
VpM = (3.020)C(0,702)	(7)
where subscripts 3P and PM refer to three process and
paste mixing facilities, respectively
The ventilation rates required for 10,000 bpd facilities were estimated by
extrapolating the second set of relationships to 416.67 batteries per hour
(equivalent to 10,000 bpd for a facility operating 24 hours per day):
V10000 = 7.816(416.67)°*941 + 3.02(416.67)0,702	(8)
TOT
= 2490 m3/min	(9)
where superscript 10000 refers to the 10,000 bpd size,
subscript TOT refers to the combined three process and
paste mixing ventilation rate
CAPITAL COSTS
Purchase Cost
The relationship between the cost of control equipment and ventilation rate
2
is also assumed to be governed by a power law. For fabric filtration with an
air-to-cloth ratio of 6:1, the NSPS background document gives control equipment
160

-------
TABLE C-l. VENTILATION RATES FOR TYPICAL 'BATTERY PLANTS
Daily
battery
production
capacity
Operating
hours
per day
Hourly
battery
production
capacity
O
Ventilation (nr/min)
Three
process
Paste
mixing
500
2000
6500
8
16
24
62.5
125.0
270.8
472
733
1517
67.9
89.5
154.0
161

-------
purchase costs of $10,500 and $100,600 (1977) for airflow rates of 142 m3/min
3	3
and 1982 m /min, respectively. Using these data and the techniques outlined
in equations 1 to 3, the following relationship can be obtained for control
system purchase cost:
P(7?) = (151.2)V(0,8563)	(10)
where P(77^ = purchase capital cost in 1977 dollars
For a baghouse applied to the combined exhaust from 10,000 bpd three process
and paste mixing operations:
P(7?) = (151.2)*(2490)(0,8563)	(11)
= 122,361
Cost indexes from Chemical Engineering were used to update the 1977 purchase
cost to 1982 dollars:
P(82) = P(77) *(1'506)	(12)
• - 184,000
where (1.506) = ratio between the third quarter 1982
Chemical Engineering plant cost index and
the fourth quarter 1977 plant cost index
= purchase cost in third quarter 1982 dollars
(82)
Installed Cost
The NSPS background document gives equipment installation costs as a function
of purchase cost for control devices at new facilities and for retrofit
installations. Installation cost factors for a fabric filter are presented in
4
Table C-2. As the table shows, total installation costs are given by
2.28 times the equipment purchase cost. Because the example case is for a new
plant, this figure does not include a retrofit penalty. The retrofit cost
is normally 20 to 50 percent of the total installed cost. For the example case,
installation and total installed costs are given by:
Installation cost = 2.28 * 184,000	(13)
« 419,500	- -
Installed cost = 419,500 + 184,000	(14)
- 603,500
162

-------
TABLE C-2. COMPONENT CAPITAL COST FACTORS FOR A FABRIC
FILTER AS A FUNCTION OF EQUIPMENT COST, E

Direct costs

Component
Material
Labor
Equipment
1.00E
0.25E
Ductwork
0.04E
0.21E
Instrumentation
0.04E
0.006E
Electrical
o. he
0.16E
Foundations
0.Q3E
0.05E
Structural
0.03E
0.05E
Sitework
0.02E
0.02E
Painting
0.004E
0.02E
Total direct costs
1.27E
0.77E


Indirect costs

Component
Measure of costs
Factor
Engineering
10% material and labor
0.204E
Contractor's fee
15% material and labor
0.306E
Shakedown
5% material and labor
0.102E
Spares
1% material
0.013E
Freight
3% material
0.038E
Taxes
3% material
0.038E
Total indirect costs

0.696E
Contingencies - 20% of direct and
indirect costs
0.547E
Total capital costs

3.28E
163

-------
ANNUALIZED COSTS
Electricity
The NSPS background document estimates the electricity requirements for
fabric filters applied to lead-acid battery production operations at 20 watts
3
per m /min of airflow, and gives an operating schedule of 6000 hours per year
for a large plant.^ An electricity cost of 4.9 cents per kilowatt (kW)-hour
3
was used in this study. Using the total ventilation requirement of 2490 m /min
for the three process operation and the paste mixing facility, the electricity
requirements and costs are given as follows:
Electricity =	(2490 m^/min) (20.8W/m^/min) (6000 hr/yr) (15)
=	311 MW-hr/yr
Cost =	(311,000 kW-hr/yr) (0.049 $/kW-hr)	(16)
=	$15,200/yr •
Maintenance and Supplies
Maintenance labor and material costs for a baghouse at a lead-acid battery
plant are estimated in the NSPS background document at 6 percent of total
installed capital costs, and supplies are estimated at 15 percent of labor
and material.^ Based on the total installed cost of $603,500, maintenance
and supply costs for the example case can be calculated as follows:
Maintenance/supplies = (0.06) (1.15) ($603,500)	(17)
= $ 41,600/yr
Operating Labor
The NSPS background document estimates the direct labor requirement for
a baghouse at a lead battery plant at 1 operator hour per 10 hours of operation.
Supervision labor is estimated at 15 percent of direct labor.^ Hourly operating
labor costs for the lead-acid battery manufacturing industry were estimated in
this study at 12 dollars per hour (including benefits).^ Using the baghouse
operating schedule of 6000 hours per year, total operating labor requirements
and costs for the example case are given as follows:
164

-------
Operating labor = (1.15) (1 hr/10 hr) (6000 hr/yr)
= 690 hr/yr
Cost = (690 hr/yr) (12 $/hr)
(18)
(19)
$8280/yr
Capital Charges
A standard 10 percent interest rate was assumed in calculating capital
recovery factors for this study. Based on an equipment life of 15 years,^
this interest rate results in a capital recovery charge of 13.2 percent per
year. In addition, an additional 4 percent capital charge was added to cover
state taxes, insurance, and overhead costs. Thus, using the total installed
capital cost of $603,500, annual capital charges are calculated as follows:
Annual capital = (0.132/yr + 0.04/yr) ($603,500)
(20)
$103,800/yr
165
"5

-------
REFERENCES
Lead-Acid Battery Manufacture - Background Information for Proposed Standards.
EPA-450/3-79-028a, U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina. November 1979. pg. 8-14.
Reference
1.
pg. 8-24.
Reference
1.
pg. 8-19.
Reference
1.
pg. 8-60.
Reference
1.
pg. 8-31.
Reference
1.
pp. 8-28 and 8-29.
U.S. Bureau of Labor Statistics. Monthly Labor Statistics. 1982.
166

-------
APPENDIX D
COSTS OF INDUSTRY-SPECIFIC
CONTROL OPTIONS AND
TECHNIQUES
167

-------
TABLE D-l. CONTROL TECHNIQUES AND COSTS FOR PRIMARY LEAD SMELTERS AND REFINERIES
Emission source
Overall emissions
Control technique control efficiency (7.)
Control costs®
Capital
Annual
Small ore storage pile
Large ore storage pile
Open work areas
Product handling/
unloading
Concentrate storage
building
Sinter plant
Blast furnace
Dross plant
Refinery
Zinc Oxide plant
Slag pile
Continuous wet duBt	90
suppression with surfactants
Continuous wet dust	70+
suppression with surfactants
Paving and sweeping^	90
Railcar and truck covers	75
Dust control system
60 meters (200 feet) stack	99+
Ducting and fan
Hooding and ventilation	70
Replacement sinter crushing
circuit
Air filtration Bystem
Hooding and ventilation	75
Replacement baghouse system
Automatic tuyere punching
Plenum building ventilation	50-70
Hooding and ventilation
Building ventilation	50-60
Hooding and ventilation
Hooding, ventilation, and covers 75
for material handling
Chemical surfactants	90
$ 342,600	$ 261,000
1,469,300
105,100
3,463,900
473,100
2,583,000c
2,034,700c
2,981,600c
627,900
949,800
377,000
45,100
1,027,800
118,600
6,702,700c	1,842,500c
449,550c
719,900c
1,373,900c
190,100
204,200
(continued)

-------
TANT.E D-l (CONTINUED). CONTROL TECHNIQUES AND COSTS FOR PRIMARY LEAD SMEI.TERS
AND REFINERIES
Overall emissions	Control costsil	
Emission source	Control technique control efficiency (%)	Capital	Annual
Ducon scrubber sysrem	Replacement baghouse system	93	$ 169,200	$ 93,500
Baghouse dumping	Conveyor and hopper system	99	1,135,700	215,700
Baghouse stacks	New central stack, ducting	99+	941,200	214,300
and fans
a0ctober 1984 dollars.
kflased on a total paved surface area of 11,200 meters^ (120,000 feet^).
cControl costs are averages based on combinations of control options.
References 1, 2, 3, and 4.

-------
TABLE D-2. FUGITIVE EMISSION CONTROL TECHNIQUES AT
SECONDARY LEAD SMELTERS5>6
Fugtive	Estimated
emission	Control	control
source	techniques	efficiencies (7o)a
1) Battery breaking
2) Charing
3) Tapping
a.	Building enclosure
b.	Process modifications
a.	Building enclosure
b.	Process modifications
a.	Hooding and automated
slag tapping
b.	Hooding for metal tapping
c.	Automated tuyere punching
75
75
75
aThese average control efficiencies are based on fugitive emission control
efficiency data from References 2 and 5, and good engineering judgement.
170

-------
TABI.E D-3. CONTROI. COSTS ESTIMATES FOR FUCITIVE EMISSION SOURCE AT EXISTINC
SECONDARY I.EAD SMKl.TliRS (OCTOBER 1984 DOJ.J.ARS)6
Plant size	(Mg/hr)
1.5	4.5	9.1
I)B	Clirg	Tpg	BB	Clirg	Tpg	nB	Chrg	Tpg
Capital costs
Total system lnatalled $ 404.000 $ 833,000 $ 177,000 $ 950,000 $	1,960,000 $ 417,000 $ 1,650,000 $ 3,400,000 $ 723,000
capital cost
Annualized costs
Operating maintenance0	40,400	83,300	17,700	95,000	196,000	41,700	165,000	340,000	72,300
Capital chargesb	76.600	51.000	48,600	180,000	363,000	114,000	313,000	629,000	198,000
Total annual costs	$ 117,000	$ 134,400	$ 66,300	$ 275,000	$ 559,000	$ 155,700	$ 478,000	$ 969,000	$ 270,300
BB - Battery breaking
Chrg - Charging
Tpg - Tapping
^Operating and maintenance cost baaed on 10 percent of total installed capital cost.
''Capital recovery factor based on 10 percent Interest rate and control technique lives of 5 to 25 years depending upon technique.


-------
TABLE D-4. CONTROL COST ESTIMATES FOR EXISTING LEAD ORE PROCESSING
PLANTS (OCTOBER 1984 DOLLARS)7"11
		Plant size
270 Mg/hr	540 Mg/hr
Capital costs
Control system purchase cost3
$
299,000
$
421,000
Control system installation cost

636,000

898,000
Retrofit cost**

186,000

264,000
Total capital cost
1
,121,000

1,583,000
Annualized costs




Utilities
$
137,000
$
203,000
Maintenance0

21,700

38,600
Operating laborc

65,200

65,200
Subtotal (direct operating costs)
$
223,900
$
306,800
Capital charges^

227,000

321,000
Gross annualized costs
$
450,900
$
627,800
Product recoverye

(1,030)

(1,130)
Net annual cost
$
449,870
$
626,670
aControl system consists of 3.75 kPa (15 in. W.G.) Venturi scrubber (95
percent efficiency).
^20 percent of total installed capital cost.
cMaintenance based on 2 manhours per 8-hour shift and operating labor based on
6 manhours per 8-hour shift and a labor rate of $15.00 per hour. Operating
schedule of 5,820 hours per year.
^Capital recovery factor based on an equipment life of 10 years.
eBased on a recovered product value of $25.11/Mg (October 1984). Also, the
parentheses indicate a cost savings.
172

-------
TABLE D-5. CONTROL COSTS FOR EXISTING LEAD-ACID BATTERY
PLANTS—PASTE MIXING (OCTOBER 1984 DOLLARS)12
Plant size (batteries per day
500	2,000	6,500
Capital costs
Control system purchase cost3
$ 8,590
$ 11,000
$ 17,600
Control system installation cost
19,800
25,100
40,000
Retrofit cost^
5,730
7,260
11,600
Total capital cost
34,120
43,360
69,200
Annualized costs
Electricity0
$ 143
$ 376
$ 975
Maintenance and supplies0
2,360
2,990
4,770
Operating labore
2,840
5,690
8,530
Subtotal (direct operating costs)
5,343
9,056
14,275
Capital charges^
5,890
7,450
11,900
Total annual costs
11,233
16,506
26,175
aControl system consists of a baghouse (99 percent efficiency).
^20 percent of installed capital cost.
cOperating schedules are 250 days per year at 8 hours per day for 500 battery
per day plants, 16 hours per day for 2,000 battery per day plants, and
24 hours per day for 6,500 battery per day plants.
^6.9 percent of total capital costs. Includes bag replacement.
eBased on 1 operator manhour per 10 hours of operation and supervision
manhours at 15 percent of operator manhours. The labor rate was $12.35 per
hour.
^Capital recovery factor based on an equipment life of 15 years.
173

-------
TABLE D-6. CONTROL COSTS FOR EXISTING LEAD-ACID BATTERY PLANTS—
THREE-PROCESS OPERATION (OCTOBER 1984 DOLLARS)12
Capital costs
Control system purchase cost3
$
45,700
$
66,600
$ 124,000
Control system installation cost

104,000

152,000
283,000
Retrofit cost*5

30,200

44,000
81,900
Total capital cost
$
179,900
$
262,000
$ 488,900
Annualized costs
Electricity0	$ 933	$ 3,080	$ 9,850
Maintenance and suppliesc	12,400	18,100	33,800
Operating labore	2,840	5,700	8,530
Subtotal (direct operating costs)	16,233	26,880	52,180
Capital charges^	31,000	45,100	84,100
Total annual costs	$ 47,233	$ 71,980	$ 136,280
aControl system consists of a baghouse (99 percent efficiency).
^20 percent of installed capital cost.
c0perating schedules are 250 days per year at 8 hours day for 500 battery per
day plants, 16 hours per day for 2,000 battery per day plants, and 24 hours
per day for 6,500 battery per day plants.
^6.9 percent of total capital costs. Includes bag replacement.
eBased
174

-------
TABLE D-7. CONTROL COSTS FOR EXISTING LEAD-ACID BATTERY PLANTS--
THREE-PROCESS OPERATION AND PASTE MIXING (OCTOBER 1984)2
Plant size (batteries per day)
500	2,000	6,500
Capital costs
Control system purchase cost3
$
1,300
$
73,500
$
135,000
Control system installation cost

117,000

168,000

308,000
Retrofit cost*5

33,800

48,500

89,000
Total capital cost
$
202,100
$
290,000
$
532,000
Annualized cost
Electricity0
$
1,140
$
3,460
$
10,500
Maintenance and supplies0

13,900

20,000

36,700
Operating labore

2,840

5,690

8,530
Subtotal (direct operating costs)
$
17,880
$
29,150
$
55,730
Capital charges^

34,700

49,800

91,400
Total annual costs
$
52,580
$
78,950
$
147,130
aCor.trol system consists of a baghouse (99 percent efficiency).
^20 percent of installed capital cost.
cOperating schedules are 250 days per year at 8 hours day for 500 battery per
day plants, 16 hours per day for 2,000 battery per day plants, and 24 hours
per day for 6,500 battery per day plants.
^6.9 percent of total capital costs. Includes bag replacement.
eBased on 1 operator manhour per 10 hours of operation and supervisions
manhours at 15 percent of operator manhours. The labor rate was $12.35 per
hour
^Capital recovery factor based on an equipment life of 15 years.
175

-------
TABLE D-8. CONTROL COSTS FOR EXISTING LEAD-ACID BATTERY PLANTS-
GRID CASTING (OCTOBER 1984 DOLLARS)1'2
Plant size (batteries per day)
500	2,000	6,500
Capital costs
Control system purchase cost3
$
15,500
$
19,000
$
25,500
Control system installation cost

40,900

49,900

67,400
Retrofit cost*3

9,480

11,500

15,600
Total capital cost
$
65,880
$
80,400
$
108,500
Annualized cost
Electricity0
$
103
$ 515
$
1,650
Waterc

412
1,240

4,220
Maintenance and suplies0

4,530
5,560

7,520
Operating labore

2,880
5,670

8,550
Subtotal (direct operating costs)
$
7,925
$ 12,985
$
21,940
Capital charges^

11,300
13,800

18,600
Total annual costs
$
19,225
$ 26,785
$
40,540
aControl system consists of a baghouse (99 percent efficiency).
^20 percent of installed capital cost.
cOperating schedules are 250 days per year at 8 hours day for 500 battery per
day plants, 16 hours per day for 2,000 battery per day plants, and 24 hours
per day for 6,500 battery per day plants.
d6.9 percent of total capital costs. Includes bag replacement.
eBased on 1 operator manhour per 10 hours of operation and supervisions
manhours at 15 percent of operator manhours. The labor rate was $12.35 per
hour
^Capital recovery factor based on an equipment life of 15 years.
176

-------
TABLE D-9. CONTROI. COSTS FOR NEW F.KAD-ACID HATTERY PI.ANTS--
10,000 llATTKIU KS I'F.H DAY (OCTOIIKK 1904 U0I.1.AKS)12
Taste mixing,
Lead oxide	Grid	three-process	Lead
production9 casting operation reclamation
Capital cost
Control system purchase cost''	$ 2,970	$ 30,000	$ 190,000	$ 23, 300
Control system Installation cost	6,760	79.3U0 432.000	61,400
Total capital cost	$ 9,730	$109,300	$ 622,000	$ 84,700
Annual costs
Electricity0	---	$ 2,920	$ 15,700	$ 769
Wotcrc	---	6,480 ---	2,030
Maintenance and supplies''	672	7,540 42,900	5,850
Operating labore 		8,530 8,530	5,160
Subtotal (direct operating costs)	$ 672	25,470	$ 67,130	$ 13,859
Capital charges^	1,670	18,800 107,000	14,600
Total annual costs	$ 2,342	$ 44,270	$ 174,130	$ 28,459
"Lead oxide production control costs are the difference the costs of a bnghouse normally
used for product recovery (3 to 1 air to cloth ratio) and one used for recovery and
emission control (2 to 1 air to cloth ratio).
''Controls are baghouses for lead oxide production, paste mixing and the three process
operation) and Impingement ncruhbers for grid casting nnd lend reclnmat1 on. Airflow
rates for the 10,000 battery plant wer extrapolated from the airflows given in
reference 2 for the 6500 and 2000 battery plants. Exponents used were 1.0 for lead
oxide production and grid casting, 0.71 for paste mixing, and 0.94 for the three-process
operatIon.
c0perating schedules are 250 days per year at 8 hours per day for 500 battery per day
plants, 16 hours per day for 2,000 battery per day plants, and 24 hours per day for
6,500 battery per day plants.
d6.9 percent of total capital coats. Includes bag replacement for baghosues.
cBased on 1 operator manhour per 10 hours of operation and supervision manhours at
15 percent of operator manhours. The labor rate was $12.00 per hour.
^Capital recovery factor based on an equipment life of 15 years.

-------
TABLE D-10. CONTROL COST ESTIMATE-BAGHOUSE FOR SMALL
ALKYL LEAD PLANT (OCTOBER 1984 DOLLARS)13
Capital Costs
Control system purchase cost3	$ 35,000
Control system installation cost	43,300
Retrofit cost*3	23, 700
Total capital cost	$102,000
Annualized Cost
Utilities	$ 5,200
Maintenance0	39,100
Operating labor	58,700
Bag replacement	1,850
Subtotal	$104,850
Capital charges^	17,000
Total annual cost	$121,850
aControl system consists of baghouse (99 percent efficiency).
^30 percent of total installed capital cost.
cMaintenance based on 2 manhours per 8-hour shift and operating labor based on
3 manhours per 8-hour shift and a labor rate of $17.85 per hour; 24 hours per
day, 365 days per year.
^Capital recovery factor based on an equipment life of 15 years.
178
¦)

-------
TABLE D-ll. CONTROL COST ESTIMATE-BAGHOUSE FOR LARGE
ALKYL LEAD PLANT (OCTOBER 1984 DOLLARS)13
Capital Costs
Control system purchase cost3	$ 361,000
Control system installation cost	439,000
Retrofit costb	240,000
Total capital cost	$1,040,000
Annualized Cost
Utilities	$ 36,100
Maintenance0	39,400
Operating labor	59,200
Bag replacement	65,900
Subtotal	$ 200,600
Capital charges'*	174,000
Total annual cost	$ 374,600
aControl system consists of baghouse (99 percent efficiency).
b30 percent of total installed capital cost.
cMaintenance based on 2 manhours per 8-hour shift and operating labor based on
3 manhours per 8-hour shift and a labor rate of §17.85 per hour; 24 hours per
day, 365 days per year.
^Capital recovery factor based on an equipment life of 15 years.
"5
179

-------
TABLE D-12. CONTROL COST ESTIMATE FOR LARGE ALKYL LEAD PLANT STACK
(OCTOBER 1984 DOLLARS)13
Capital Costs
Control system purchase cost3	§ 35,300
Control system installation cost	43,300
Retrofit cost*3	23,600
Total capital cost	$ 102,200
Annualized Cost
Utilities	$ 5,970
Maintenance
Operating labor	9,890
Subtotal	$ 15,860
Capital charges0	17,100
Total annual cost	$ 32,960
aControl system consists of stack and auxiliary equipment.
^30 percent of total installed capital cost.
cCapital recovery factor based on an equipment life of 15 years.
180

-------
TABLE D-13. CONTROL COST ESTIMATES FOR EXISTING GRAY IRON FOUNDRIES-
UNCONTROLLED CUPOLA FURNACE (OCTOBER 1984)14-17


Furnace
melt capacity (Mg/hr)


3.6

9.1

20.0
Capital costs






Control system purchase cost3
$
163,000
$
357,000
$
778,000
Control system installation cost

190,000

418,000

911,000
Retrofit cost*5

106,000

233,000

507,000
Total capital cost
$
459,000
$1
,008,000
$
2,196,000
Annualized cost






Utilities $

8,030
$
21,200
$
76,700
Maintenance and supliesc

7,210

8,650

12,700
Operating laborc

10,800

13,000

19,000
Bag replacement

10,200

22,100

44,400
Subtotal (direct operating costs)
$
36,240
$
64,950
$
152,800
Capital charges'3

72,200

159,000

346,000
Total annual costs
$
108,440
$
223,950
$
498,800
aControl system consists of side draft hood and baghouse. (88.6% efficiency).
^30 percent of installed capital cost.
Maintenance based on 2 manhours per 8-hour shift and operating labor based on
3 manhours per 8-hour shift and a labor rate of $18.00 per hour.
Operating schedules: 3.6 Mg/hr plant - 1,600 hr/yr
9.1 Mg/hr plant - 1,920 hr/yr
20.0 Mg/hr plant - 2,800 hr/yr
^Capital recovery factor based on an equipment life of 20 years.
181


-------
TABLE D-14. DESCRIPTION OF CONTROL OPTIONS FOR ALLOY STEEL ELECTRIC
ARC FURNACES18
Control Option Process Controls	Fugitive Controls
1
Direct-shell evacuation
control and fabric filters
Fabric filters; single canopy
hood, open roof monitor
enclosed dust handling
equipment.
2
Direct-shell evacuation
control and fabric filters
Fabric filters; segmented
canopy hood, scavenger duct,
cross-draft partitions (or
single canopy hood with
separate tapping and slagging
hoods), closed roof (over
furnace)/open roof monitor
elsewhere; enclosed
dust-handling equipment.
3
Direct-shell evacuation
control and fabric filters
Fabric filters, segmented
canopy hood, scavenger duct,
cross-draft; enclosed
dust-handling equipment.
182

-------
TAU1.E U-15. CONTROJ, COST ESTIMATES FOR EXISTING A MAY STF.EI.
EI.ECTIUC ARC KUKNACES (OUTOIIEK 1984 DOI.I.ARS ) 1U" 2 1
Furnace capacity (Hg/heat)	23 91 272
Control option8	123 1 2312 3
Capital costs
Control system purchase cost	NA NA $ 777,000 $ 1,980,000 $ 2,050,000 NA NA NA $ 2,530,000
Control system Installation cost	1,170,000 2,970,000 3,080,000 NA NA NA 3,790,000
Retrofit costb	583,000 1,490,000 1.540,000 1,900.000
Total Installed capital cost	$ 2,530,000 $ 6,440,000 $ 6,670,000 $ 8,220,000
Annualized costs
Ut1111lea
$
63,300
$
345,000
$
345,000
$
526,000
Maintenance

17,500

19,000

19,000

23,200
Operating labor

27,300

29,900

29,900

36,300
Monitoring

20,500

28,400

20,500

20,500
Solid waste disposal

62,600

269,000

272,000

999,000
Filter replacement

6,590

25,100

25,100

45,300
Subtotal (direct operating costs)
$
197,790
$
716,400
$
711,500
$
1,650,300
Capital charges0

433,000

1,110,000

1,140,000

1,410,000
Total annual costs
$
630,790
$
1,826,400
$
1,851,000
$
3,060,300
NA - Control option not applicable to this size model plant.
aControl option 1 - 80.07. fugitive emission control
Control option 2 - 90.0% fugitive emission control
Control option 3 - 97.57. fugitive emission control
b30 percent of total annualized costs
cCapltal recovery factor based on an equipment life of 15 years.

-------
TABLE D-16. CONTROL COST ESTIMATES FOR EXISTING STEEL ELECTRIC ARC
FURNANCES.22 (OCTOBER 1984 DOLLARS)
Furnace capacity (Mg/heat)	3.6	9.1	22.7
Capital cost
Control system purachase and
installation costs3
$
352,000
$
278,000
$
527,000
Retrofit costs*5

106,000

83,000

158,000
Total installed capital costs
$
458,000
$
361,000
$
685,000
lual costs






Utilities
$
8,960
$
7,280
$
25,800
Maintenance

3,500

2,800

5,320
Operating labor

13,400

16,100

23,500
Bag replacement

11,600

7,560

18,900
Subtotal (direct operating costs)
$
37,460
$
33,740
$
73,520
Capital charges0

78,500

61,900

117,000
Total net annual costs
$
115,960
$
95,640
$
190,520
aControl option consists of side draft hood and baghouse for 3.6 Mg/heat plant
and direct furnace evacuation and baghouse for 9.1 and 22.7 Mg/heat plants.
*>30 of total purchase and installation costs
cCapital recovery based on an equipment life of 15 years.
184

-------
TABLE D-17. CONTROL COST ESTIMATES FOR EXISTING IRON ORE
SINTERING OPERATIONS (OCTOBER 1984 DOLLARS).23>24
Material	Sinter
handling	plant operations
Capital costs
Total system installed
capital costs3	$ 3,460,000	$ 6,700,000
Annual Costs
Operating and
maintenance	$ 440,000	$ 690,000
Capital charges**	590,000	1,150,000
Total annual costs	$ 1,030,000	$ 1,840,000
aMaterial handling - Railcar and truck cover, plus dust control system
(75% control efficiency)
Sinter plant operations - Hooding and ventilation, replacement of sinter
crushing circuit, and air filtration system.
^Based on an equipment life of 15 years.
185 ,

-------
TABLE D-18. CONTROL COST ESTIMATES FOR EXISTING BRASS AND BRONZE
INGOT PRODUCTION FACILITIES (OCTOBER 1984 DOLLARS)24
Furnace size
(Mg/day output)
Capital costs
System installation cost3
Retrofit cost*5
Total installed capital cost
18.1
$ 176,000
52,800
45.4
78,600
68.1
$ 262,000 $ 310,000
93,000
$ 228,000 $ 340,600 $ 403,000
Annual costs
Operating and maintenance0
Capital charges**
Total annual cost
$ 22,900
39,200
$ 62,100
$ 34,100 $ 40,300
58,400
69,100
$ 92,500 $ 109,400
3Replacement of existing particulate control systems with 95% + efficient
baghouse control system.
b30 percent of total installed capital cost.
cOperating and maintenance based on 10 percent of total installed capital cost.
^Based on an equipment life of 15 years.
¦i
186

-------
TABLE D-19. CONTROL OPTIONS ANNUALIZED FOR EACH SOURCE
Source Control Option Capital g Annualized
				,	($ x 10 )
1

Secondary
Converter
4.6
1.3
2
a)
Secondary
Converter
10.1
2.9

b)
Secondary
Furnace
0.95
0.27
3
a)
Secondary
Converter
8.3
2.2

b)
Secondary
Furnace
0.92
0.27
4

Secondary
Converter
9.0
2.3
5
a)
Secondary
Converter
8.9
2.0

b)
Secondary
Furnace
7.9
1.9
6
a)
Secondary
Converter
10.1
2.9

b)
Secondary
Furnace
1.8
0.53
7
a)
Secondary
Converter
6.9
1.8

b)
Secondary
Furnace
0.92
0.27
8
a)
Secondary
Converter
4.6
1.3

b)
Secondary
Furnace
0.92
0.27
187

-------
REFERENCES FOR APPENDIX D
Primary Lead Smelting and Refining
1.	Charles River Associates. Economic and Enviornmental Analysis of the
Current OSHA Lead Standard. Occupational Safety and Health
Administration. Washington, DC. 1981.
2.	Technical Guidance for Control of Industrial Process Fugitive Particulate
Emissions (EPA-450/3-77-010). U.S. Environmental Protection Agency,
Research Triangel Park, NC. March 1977.
3.	Cost Files. Chemical Engineering. 1980 to 1983.
4.	Uhl, Vincent. A Standard Procedure for Cost Analysis of Pollution Control
Operations. U.S. Environmental Protection Agency, Research Triangle Park,
NC. 1979.
Secondary Lead Smelting
5.	Schwitzgebel, K., et al. (Radian Corp.). Fugitive Emissions at a
Secondary Lead Smelter. U.S. Environmental Protection Agency,
Philadelphia, PA. December 1981.
6.	Ref. 2.
Lead Ore Processing
7.	Metallic Mineral Processing Plants—Background for Proposed Standards.
(EPA-450/3-81-009a). U.S. Environmental Protection Agency, Research
Triangle Park, NC. August 1982.
8.	The International Competitiveness of the U.S. Non-Ferrous Smelting
Industry and the Clean Air Act. American Mining Congress. April 1982.
9.	Reference 4.
10.	Neveril, R.B. Capital and Operating Costs of Selected Air Pollution
Control System (EPA-450/5-80-002). U.S. Environmental Protection Agency,
Research Triangel Park, NC. December 1978.
11.	Engineering and Mining Journal. September 1982.
12.	Lead-Acid Battery Manufacture - Background Information for Proposed
Standards (EPA-450/3-79-028a). U.S. Environmental Protection Agency,
Research Triangle Park, NC. 1979.
188
T

-------
Lead Alkyl Plants
13.	Reference 10.
Gray Iron Foundries
14.	Electric Arc Furnaces in Ferrous Foundries - Background Information for
Proposed Standards (EPA-450/3-80-020a). U.S. Environmental Protection
Agency, Research Triangle Park, NC. 1980.
15.	Reference 4.
16.	Reference 10.
17.	Control Techniques for Lead Emissions (EPA-450/2-77-012). U.S.
Environmental Protection Agency, Research Triangle Park, NC. 1977.
Alloy Steel Electric Arc Furnace
18.	Electric Arc Furnaces and Argon-Oxygen Decarburization for Vessels in
Steel Industry - Background Information for Porposed Revision to Standard!
(EPA-450/3-020a). U.S. Environmental Protection Agency, Research TriangL
Park, NC. 1983.
19.	Reference 4.
20.	Reference 10.
21.	Reference 17.
Steel Foundries - Electric Arc Furnaces
22.	Reference 14.
Iron Ore Sintering
23.	Reference 1.
24.	Reference 4.
Brass and Bronze Ingot Production
25.	Secondary Brass and Bronze Ingot Production Plant — Background
Information for Proposed Standards (APTD 1352a). U.S. Environmental
Protection Agency, Research Triangle Park, NC. 1973.
189
¦>

-------
I
APPENDIX E ;
i
IDENTIFICATION OF BATTERY PLANTS AFFECTED UNDER ALTERNATIVE
NAAQS LEVELS
190

-------
APPENDIX e'
IDENTIFICATION OF BATTERY PLANTS AFFECTED UNDER
ALTERNATIVE NAAQS LEVELS
Table E-1 gives the baseline control status of the lead-acid battery
manufacturing industry. Existing plants were divided into 12 groups by size
and control status based on a NEDS survey. Each of the plant groups shown
in Table E-l were studied to determine the number of plants which will be
affected in the group under the NAAQS alternatives. This was done by
calculating the probability that a model plant representative of the group
will be affected, and multiplying by the number of plants in the group. This
appendix presents the algorithms used to estimate the numbers of affected
plants. In addition, the inputs to the algorithms are given and a sample
calculation is presented.
E.l. CALCULATION METHOD
The source-receptor coefficient, defined as the ratio between the maximum
ambient impact for a given source to the emission rate for the source, was
used extensively in determining the number of battery plants affected under
the NAAQS alternatives. For lead-acid battery manufacture, the source-
receptor coefficient was found to be dependent only on the plant location
and distributed uniformly between a maximum and minimum observed value. The
coefficient was found to be independent of the individual emission source
under study. Given a uniform distribution of this coefficient, the number
of plants affected by a given NAAQS in a given size/control status category
can be calculated as follows:
SRCA(i,j)
P(i,j)
n(i,j)
191
NAAQS - MSB
ER(i,j)
= 0	if
= 1	if
= SRCUL - SRCA(i.l) .f
SRCUL - SRCLL 1
= P(i,j)*N(i,j)
SRCA(i,j) > SRCUL
SRCA(i,j) < SRCLL
SRCLL < SRCA(i,j) > SRCUL

-------
TABLE E-l. ESTIMATED BREAKDOWN OF BATTERY PLANTS
. BY SIZE RANGES AND CONTROL STATUS
_ .... . , Estimated number of plants
Facilities with							c	:	j
controls	Small	Medium	Large
PbO; Rec
5
7
4
PbO; Rec; PM
1
2
1
PbO; Rec; PM; 3-P
7
10
6
PbO; Rec; PM; 3-P; GC
15
23
13
Total
28
42
24
a"PbO" refers to lead oxide production; "Rec" to lead reclamation; "PM" to
paste mixing; "3-P" to the three-process operation; and "GC" to grid
casting.
^400-1,000 batteries per day (bpd).
cl,000-3,500 bpd.
d0ver 3,500 bpd.
192
.... -j

-------
where: SRCA(i,j)
NAAQS	=
MSB	=
ER(i,j)	=
P(i,j)	=
SRCUL	=
is a variable used for calculation purposes equivalent
to the value of the source-receptor coefficient above
which a model plant representing size class i and
control status j would be affected,
the NAAQS alternative level under consideration,
the mobile source background,
the emission rate for the model plant representing size
class i and control status j,
the probability that a plant in size range i and with
control status j will cause an exceedence,
the upper limit of the source-receptor coefficient as
determined by dispersion modeling,
SRCLL
r-(i, j)
N(i,j)
E.2 INPUTS
the lower limit of the source-receptor coefficient,
= the number of plants affected by the NAAQS in size
range i and with control status j,
= the total number of plants in size range i and with control
status j,
A size of 500 batteries per day (bpd) was chosen to represent plants in
the 0-1000 bpd size range, and sizes of 2000 and 6500 bpd, respectively, were
selected to represent the 1000-3500 bpd and 3500-and-over ranges. Operating
hours used in this study were 8 hours per day for small plants, 16 hours per
day for medium-sized plants, and 24 hours per day for large plants, at 250 days
per year. Uncontrolled and controlled emission factors used in this study
are given in Table E-2. The lower and upper limits of the source-receptor
3
coefficient were 0.089 and 0.51 (ug/m )/(Mg/yr), respectively. A mobile
3.
source background concentration of 0.2 ug/m was assumed.
E .3 SAMPLE CALCULATIONS
3
Take as an example the 1.0 ug/m NAAQS alternative applied to the group
of large battery plants with no controls for lhe paste mixing, three process
operation, or grid casting facilities. From Tables E~1 and E-2 and the inputs
listed in the previous section:
1
193

-------
TABLE E-2. EMISSION FACTORS FOR LEAD-ACID BATTERY MANUFACTURE
Emission factors (g/battery)
uncontrolled	controlled
Grid casting
Paste mixing
Lead oxide production
Three process operation
Lead reclamation
0.35
1.13
4.79
0.63
0.026
0.05
0.072
0.006
Control of grid casting was not found to be required to meet
any of the NAAQS alternatives.
5Lead oxide production facilities are equipped with baghouses
for product recovery.
"In a survey of NEDS, lead reclamation was found to be controlled
in all cases.
194

-------
Emission factor —	(0.35 + 1.13 + 0.05 + 4.79 + 0.006) g/battery
=	6.33 g/battery
ER(i,j) =	(6.33 g/b)*(6500 b/day)*(250 day/yr)
=	10.28 Mg/yr
SRCUL =	0.51 ug-yr/m3-Mg
SRCLL = 0.89 ug-yr/m3-Mg
NAAQS = 1.0 ug/m3
MSB = 0.2 ug/m3
N(i,j) = 4
Substituting into the equations given in Section E.l.
SECA(i.j) - 1-°	- °'2 u«/m3
10.28 Mg/yr
= 0.078 ug-yr Mg/yr
SRCA(i,j) < SCRLL
P(i,j) = 1
n(i,j) = 1*4
3
Thus, all of the plants in this group will be affected under the 1.0 ug/m
NAAQS and ould need to control at least the paste miing facility. To
determine how many plants in the group would also need to control the
three-process operation, the controlled emission rate must be used
for paste mixing:
Emission factor = (0.35 + 0.026 + 0.05 +4.79 + 0.006) (g/battery
= 5.22 g/battery
ER(i,j) = 5.22 * 6500 * 250
= 8.49 Mg/yr
1.0 - 0.2
SRCA(i,j) =
8.49
3
0.094 ug-yr/m -Mg
= 0.51 ~ 0.094
0.51 - 0.089
195

-------
= 0.988
n(i,j) = 0.988 * 4
= 4 plants
Thus, all plants in the group would require control of both paste mixing
and the three-process operation.
196

-------