United States Office of Air Quality EPA-450/3-86-010
Environmental Protection Planning and Standards September 1986
Agency Research Triangle Park NC 27711
Air ~~ !^^~~~I~I
v>EPA Review of
New Source
Perform »nce
Standards for
Primary Aluminum
Reduction Plants
-------
United States Office of Air Quality EPA-450/3-86-010
Environmental Protection Planning and Standards September 1986
Agency Research Triangle Park NC 27711
Air
Review of
New Source
Perform »nce
Standards for
Primary Aluminum
Reduction Plants
-------
United States Office of Air Quality EPA-450/3-86-010
Environmental Protection Planning and Standards September 1986
Agency Research Triangle Park NC 27711
Air
Review of
New Source
Perforr* itice
Standards for
Primary Aluminum
Reduction Plants
-------
vvEPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/3-86-010
September 1986
Air
Review of
New Source
Perform *nce
Standards for
Primary Aluminum
Reduction Plants
-------
EPA-450/3-86-010
Review of New Source Performance Standards
for Primary Aluminum Reduction Plants
Emission Standards and Engineering Division
U5. Environmental Protection Agency
SjStfjSB K. 12th Floor
Chicago, IL 60604-3590
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
September 1986
-------
This report has been reviewed by the Emission Standards and Engineering Division of the Office of Air Quality Planning
and Standards, EPA, and approved for publication. Mention of trade names or commercial products is not intended to
constitute endorsement or recommendation for use. Copies of this report are available through the Library Services
Office (MD-35), U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711; or, for a fee, from
the National Technical Information Services, 5285 Port Royal Road, Springfield, Virginia 22161.
-------
TABLE OF CONTENTS
1. SUMMARY
1.1 Best Demonstrated Control Technology .......... 1-1
1.2 Economic Considerations Affecting the NSPS ....... 1-2
1.3 Industry Trends ..................... 1-2
1.4 Other Findings ..................... 1~2
1.4.1 Testing ..................... 1-2
1.4.2 Sulfur Dioxide Emissions ............. 1-3
2. INTRODUCTION ......................... 2-1
2.1 Background Information ................. 2-1
2.2 Scope of the Review ................... 2-2
2.3 Current Standards .................... 2"2
2.3.1 New Source Performance Standards ......... 2-3
2.3.2 State Regulations ................ 2-5
2.3.3 PSD Regulations ................. 2-12
2.4 References for Chapter 2 ......... ....... 2-14
3. THE PRIMARY ALUMINUM REDUCTION INDUSTRY ........... 3-1
3.1 The Industry ...................... 3-1
3.2 Plant Description .................... 3-1
3.3 Process Description .................. 3-9
3.3.1 Bath Ratio .................... 3-11
3.3.2 Tapping ..................... 3-12
3.3.3 Anode Effects .................. 3-12
3.4 Types of Plants in Use ................. 3-14
3.5 Aluminum Reduction Pots ................ 3-15
3.5.1 Center-worked Prebake Pots ........... 3-15
3.5.2 Side-worked Prebake Pots ............. 3-18
3.5.3 Vertical Stud Soderberg Pots ........... 3-20
3.5.4 Horizontal Stud Soderberg Pots .......... 3-22
3.6 Anode Bake Furnaces ................... 3-22
3.6.1 Ring Furnace ................... 3-22
3.6.2 Tunnel Kiln ................... 3-27
3.7 Process Emissions ................... 3-27
3.7.1 Total Fluorides ................. 3-29
3.7.2 Sulfur Dioxide ................. 3-31
3.8 References for Chapter 3 ................ 3-33
4. EMISSION CONTROL TECHNOLOGY ................. 4-1
4.1 Primary Fluoride Control Systems ............ 4-1
4.1.1 Capture/Suppression ............... 4-1
4.1.2 Primary Fluoride Removal ............. 4-6
4.2 Secondary Fluoride Controls ............... 4-8
4.3 Particulate and Sulfur'Dioxide Controls ......... 4-17
-------
4.4 Control Systems Performance 4-18
4.4.1 Total Fluorides 4-18
4.4.2 Particulate Matter 4-23
4.4.3 Sulfur Dioxide 4-26
4.5 References for Chapter 4 4-29
5. COMPLIANCE STATUS OF PRIMARY ALUMINUM PLANTS 5-1
5.1 Affected Facilities 5-1
5.2 Emissions Data 5-1
5.2.1 Total Fluoride 5-1
5.2.2 Visible Emissions 5-9
5.3 References for Chapter 5 5-11
6. COST ANALYSIS 6-1
6.1 Fluoride Controls 6-1
6.1.1 Costs for CWPB Fluoride Controls 6-1
6.1.2 Costs for VSS Fluoride Controls 6-6
6.2 Sulfur Dioxide Controls 6-6
6.2.1 Costs for CWPB S02 Controls 6-8
6.2.2 Costs for VSS SOe'Controls 6-9
6.3 References for Chapter 6 6-10
7. ENFORCEMENT ASPECTS 7-1
7.1 Comments 7-1
7.2 Emission Testing 7-1
7.3 NSPS Interpretation 7-4
7.3.1 Plant 1 7-4
7.3.2 Plant 2 7-5
7.3.3 Plant 3 7-5
7.4 References for Chapter 7 ....... 7-6
iv
-------
LIST OF FIGURES
Figure Pa9e
3-1 Block Diagram: Primary Aluminum Plant 2-6
3-2 Process Operations in a Primary Aluminum Plant 3-7
3-3 Plan View of Typical Potline 3-8
3-4 Aluminum Reduction Pot 3-10
3-5 Tapping Molten Aluminum from Primary Aluminum
Reduction Pot 3-13
3-6 Center-worked Prebake Pot 3-16
3-7 Side-worked Prebake Pot 3-19
3-8 Vertical Stud Soderberg Pot 3-21
3-9 Horizontal Stud Soderberg Pot 3-23
3-10 Ring Furnace Layout 3-25
3-11 Tunnel Kiln 3-28
4-1 Typical Center-worked Prebake Pot Hooding 4-3
4-2 Typical Vertical Stud Soderberg Pot Hooding 4-7
4-3 Injected Alumina Dry Scrubber 4-9
4-4 Fluidized Bed Dry Scrubber 4-10
4-5 Flow Diagram of the Dry Scrubbing Process for a
Primary Aluminum Plant 4-11
4-6 Fluidized Bed Dry Scrubber Used on an Anode Bake
Furnace Exhaust 4-12
4-7 Cross-section of Wet Scrubber Used to Control
Secondary Emissions from Plant E 4-13
4-8 Hood Inspection Data Sheet 4-15
4-9 Hood and Crucible Inspection Summary 4-16
-------
LIST OF TABLES
Table Page
2-1 New Source Performance Standards (NSPS) for
Primary Aluminum Reduction Plants 2-4
2-2 State Regulations for Fluoride Emissions from Existing
Primary Aluminum Plants 2-6
2-3 State Guidelines for Control of Fluorides from Existing
Primary Aluminum Plants 2-8
2-4 State Regulations for Particulate Matter Emissions from
New Plants 2-9
2-5 State Regulations for Visible Emissions from New Plants. . . 2-10
2-6 State Regulations for Control of S02, CO, NOX, and HC from
Non-fuel Burning Sources in Primary Aluminum Plants. .... 2-11
2-7 BACT Determinations for Plants Subject to PSD Regulations. . 2-13
3-1 Listing of Operating Primary Aluminum Plants in the
United States and Their Capacities - May 1986 3-2
3-2 Listing of Non-operating Primary Aluminum Plants in the
United States and Their Capacities - May 1986 3-4
3-3 Available Information on Uncontrolled Emissions of Total
Fluorides 3-30
4-1 Airflows to Individual Pots at Plants with Potlines Subject
to the NSPS 4-4
4-2 Total Fluoride Emissions from Potlines and Anode Bake
Furnaces Subject to the NSPS 4-19
4-3 Total Fluoride Emissions by Potroom Group and Type 4-21
4-4 Impacts of Changes in Primary TF Capture and Removal
Efficiencies on Overall TF Emissions 4-22
4-5 Effectiveness of Secondary Wet Scrubbers at PI ant E 4-24
4-6 Particulate Emissions from Primary Aluminum Plants Using
Dry Scrubbers to Control Fluoride Emissions 4-25
-------
Table aM
4-7 Sulfur Dioxide Emissions from a Primary Aluminum
Plant Using Wet Scrubbers for Primary and Secondary
Sulfur Dioxide Control .................... 4'27
4-8 Sulfur Dioxide Emissions from Anode Bake Plants ........ 4-28
5-1 List of Primary Aluminum Reduction Plants Subject to
the NSPS ........................... 5'2
5-2 Fluoride Emissions from Primary Aluminum Plants Subject
to the NSPS .......................... 5-3
5-3 Record of Reported NSPS Exceedances with Failure Rationale . . 5-5
5-4 Emissions from Potlines at Primary Aluminum Plants
with Fluoride Controls ............ ........ 5~6
5-5 Emissions from Anode Bake Furnaces at Primary Aluminum
Plants with Fluoride Controls
5-6 Visible Emissions from Anode Bake Furnace at PI ant J ..... 5-10
6-1 Costs of Dry Scrubbers to Control Total Fluoride Emissions
at Center-Worked Prebake Plants ................ 6-3
6-2 Effectiveness of Total Fluoride Control Systems at CWPB
Plants .................. • ......... 6'4
6-3 Cost-effectiveness of Total Fluoride Control Systems ..... 6-5
6-4 Capital and Annual i zed Costs to Control Total Fluoride and
S02 Emissions from VSS Plants ................. 6-7
7-1 Comments Received from Plants with NSPS Potlines or Anode
Bake Furnaces ......................... 7~2
-------
1. SUMMARY
The new source performance standards (NSPS) for primary aluminum
reduction plants were promulgated by the U.S. Environmental Protection
Agency (EPA) on January 26, 1976, under Section 111 of the Clean Air Act.
The standards apply to all aluminum reduction potlines and anode bake
furnaces which commenced construction or modification after October 23,
1974. The NSPS limit emissions of gaseous and particulate fluorides,
measured as total fluorides (TF).
The NSPS were amended on June 30, 1980, to permit TF to exceed
previous limits, under certain circumstances. A requirement for monthly
compliance tests was added at the same time.
The objective of this report is to document the review of the NSPS for
primary aluminum reduction plants, and to assess the need for revision on
the basis of developments that have occurred since the standard was
promulgated. This review is required under Section lll(b) of the Clean
Air Act, as amended. The following paragraphs summarize the findings of
this review.
1.1 BEST DEMONSTRATED CONTROL TECHNOLOGY
The NSPS limit emissions of TF from primary aluminum reduction potlines
and anode bake furnaces. No changes have occurred in the control techniques
defined as best demonstrated technology (BDT) for these sources. For
potlines, they are either wet scrubbers followed by wet electrostatic
precipitators (ESP's) or dry scrubbers. For anode bake furnaces, they
are either dry or wet scrubbers. However, all plants with potlines or
anode bake furnaces subject to the NSPS have elected to use dry scrubbers
for TF control, and all have demonstrated the capability to comply with
the NSPS.
-------
1.2 ECONOMIC CONSIDERATIONS AFFECTING THE NSPS
Information on the capital and annualized costs of dry scrubbers for
controlling potlines and anode bake furnaces was supplied by plants subject
to the NSPS. Additional cost data were extracted from a report published by
the International Primary Aluminum Institute (IPAI). These data show that
the installation of TF emission controls can either increase or decrease
aluminum production costs.
The cost impacts of TF emission controls are reported by the IPAI to
range from a credit of $11.60 to a cost of $10.65 for each ton of aluminum
produced. Similar information on 3 domestic plants subject to the NSPS
show net costs of $11.80 to $17.70 per ton of aluminum. Two of the NSPS
plants utilize anode bake furnaces and those costs are included. The plants
listed in the IPAI report did not include bake furnace control costs. For
the 2 domestic plants having only anode bake furnaces subject to the NSPS,
control costs are $3.60 and $4.45 per ton of aluminum produced.
Five new potlines and eight new anode bake furnaces have been placed
in service since the NSPS was proposed. , No potline or bake furnace
construction has taken place in the U.S. in the last 4 years (1983-86),
however, and forecasts indicate that none will be built in the next 5 to 10
years.
1.3 INDUSTRY TRENDS
No growth is expected in the domestic primary aluminum industry
because of the relatively high cost of power in the United States. In fact,
the domestic industry may well experience negative growth, with the less
efficient plants, or those in high power cost areas, being closed down.
Ten U.S. plants have closed in the last 5 years (1981-1985), at least six of
them permanently, and most of the remainder are operating at reduced capacity.
In December 1985, the domestic primary aluminum industry operated at 66
percent of capacity, after adjustment for permanent plant closures and capacity
reductions. There are no known instances whereby an existing facility will
become an affected facility through either modification or reconstruction.
1.4 OTHER FINDINGS
1.4.1 Testing
The principal issue raised by members of the primary aluminum industry
is the monthly testing requirements for secondary (fugitive) potroom emissions.
1-2
-------
There are no provisions in the standard for reducing the frequency of secondary
potroom emissions tests. However, the General Provisions (§60.8(b)(4)) give
to the Administrator, and subsequently to the States who have received delegation,
the authority to reduce test frequency. Data from one well-controlled plant
were used to develop formulae for determining the statistical probability of
a random failure (assumes no known changes in the level of maintenance, in
work practices, or in the frequency and thoroughness of potroom inspections).
1.4.2 Sulfur Dioxide Emissions
Sulfur dioxide (SQ2) emissions from primary aluminum reduction plants
have increased since the NSPS were proposed due to an increase in the sulfur
content of coke, and a shift to dry scrubbers for TF control.
1-3
-------
2. INTRODUCTION
2.1 BACKGROUND INFORMATION
New source performance standards (NSPS) were promulgated for primary
aluminum reduction plants on January 26, 1976, under Section 111 of the
Clean Air Act.1 The NSPS control emissions of gaseous and particul ate
fluorides, measured as total fluorides (TF), from aluminum reduction
potlines and anode bake furnaces. They apply to all facilities constructed,
modified, or reconstructed after October 23, 1974, the date of publication
of the proposed regulations. Since fluoride is a designated pollutantv,
the States were required to develop companion standards for existing
facilities. A document was, therefore, prepared to provide guidance to
the States regarding probable fluoride emissions levels which could be
expected from existing uncontrolled plants and the amounts of emission
reduction which should be achievable at those plants. It was released in
December 1979.2
Shortly after the NSPS was promulgated, petitions for review were filed
by four U.S. aluminum companies. As a consequence, additional data were
obtained and amendments to the NSPS were proposed on September 19, 1978.
On June 30, 1980, the NSPS amendment was promulgated to permit TF emissions
to exceed, under certain circumstances, the levels set initially.3 A
monthly monitoring requirement was added at the same time. This monitoring
requirement has been waived by EPA for the primary fluoride control
devices at some plants in favor of yearly tests. Measurement of secondary
(fugitive) fluoride emissions from all new potrooms is, however, required
on a monthly basis.
V A designated pollutant is one which is not included on a list
publi?hed under Section 108(a) of the Act (National Ambient Air Quality
Standards), but for which an NSPS has been established.
-------
As was discussed in the background information document and the
guidance document, effects of fluoride have been extensively documented.^»5
Fluoride does not directly affect human health but can have deleterious
effects on both plants and animals. It is, therefore, classified as a
welfare pollutant.
2.2 SCOPE OF THE REVIEW
The Clean Air Act Amendments of 1977 require that the Administrator of
EPA review and, if appropriate, revise established standards of performance
for new stationary sources at least every 4 years.6 The purpose of this
report is to document this review and to assess the need for revision of the
existing standards for primary aluminum reduction plants, based on develop-
ments that have occurred or are expected to occur within the aluminum
industry. The information presented in this report was obtained from
reference literature, discussions with industry representatives, trade
organizations, process and control equipment vendors, EPA Regional Offices,
and State and local agencies.
The review conducted to assess the current NSPS for primary aluminum
reduction plants was limited to three areas of concern, as follows:
0 technologies being used for compliance (process modifications,
maintenance, work practices, housekeeping, capture and control
equipment design, and process selection);
0 enforcement and compliance experience; and
0 State standards implemented as a result of the NSPS.
2.3 CURRENT STANDARDS
Federal NSPS for primary aluminum reduction plants regulate fluoride
emissions from aluminum reduction potrooms and, if applicable, from anode
bake furnaces. Other sources and pollutants are regulated by prevention
of significant deterioration (PSD) or State regulations. The NSPS are
summarized and discussed in Section 2.3.1 below, and the applicable
regulations for those states with primary aluminum plants are reviewed in
Section 2.3.2.
2-2
-------
2.3.1 New Source Performance Standards
2.3.1.1 Summary of New Source Performance Standards. The original
standards for primary aluminum plants (Table 2-1) were proposed on
October 23, 1974, and promulgated on January 26, 1976.7>8 They limited
Tp2_/ emissions from new, modified, or reconstructed potroom groups and (if
applicable) anode bake furnaces in primary aluminum reduction plants to a
total of 1 kilogram per megagram (kg/Mg) of aluminum produced (2.0 pounds
per ton of aluminum produced [lb/TAP]).
The NSPS limit overall fluoride emissions from potrooms, and, therefore,
require the measurement of both primary and secondary fluoride emissions.
Primary emissions are those captured by the pot hoods while secondary
emissions are fugitive emissions from the pot hoods and all emissions
generated outside the pots. An example of the latter would be outgassing
from a spent anode left beside the potline to cool.
Visible emissions regulations were set at the same time. They limit
emissions from potroom groups to less than 10 percent opacity and those from
anode bake plants to less than 20 percent opacity.
Amendments to the NSPS (Table 2-1) were proposed on September 19, 1978,
and promulgated on June 30, 1980.9»10 One major change was the addition of
higher, never-to-be-exceeded (NTBE) limits for potrooms. These NTBE limits
were added to allow for variability in fluoride emissions from the aluminum
reduction process. Emissions which exceed the original NSPS but are
below the NTBE limit are acceptable if the owner/operator can demonstrate
that the appropriate control systems^/ have been installed and are being
operated and maintained in an exemplary fashion. The other major change
was the addition of a monthly testing requirement.
2.3.1.2 Testing and Monitoring Requirements. Initial performance
tests to verify compliance with the standards for primary aluminum reduction
plants must be completed within 60 days after achieving full capacity
£/ The term "total fluoride" refers to elemental fluorine and all
fluoride compounds (gaseous and particulate) which are measured by
EPA reference methods 13A or 13B.
y The control system includes the pot hoods, the ducting, and the
primary control device.
2-3
-------
TABLE 2-1
NEW SOURCE PERFORMANCE STANDARDS (NSPS)
FOR
PRIMARY ALUMINUM REDUCTION PLANTS11.12
Affected facility
Pollutant
NSPS emission 1imita
Comments
Stud Soderberg
potroom group
(Vertical or
horizontal)
Total
Fluorides
Visible
Emissions
1.0 kg TF/Mg Al
(2.0 lb TF/ton Al)
1.3 kg TF/Mg Al
(2.6 lb TF/ton Al )
<10% opacity
Original standard
NTBE limit, Amendment13
Prebake plant
potroom group
(Center and
side-worked)
Total
Fluorides
Visible
Emissions
0.95 kg TF/Mg Al
(1.9 Ib/ton Al)
1.25 kg TF/Mg Al
(2.5 lb TF/ton Al )
<10% opacity
Original standard
NTBE limit, Amendment^
Prebake pi ant
anode bake plant
Total
Fluorides
Visible
Emissions
0.05 Kg TF/Mg Al
equivalent
(0.1 lb TF/ton Al)
<20% opacity
a kg TF/Mg Al = Kilograms total fluoride per megagram aluminum produced
lb TF/ton Al = Pounds total fluoride per ton aluminum produced.
b Compliance to this never-to-be-exceeded (NTBE) limit is acceptable
if owner/operator demonstrates that the proper control equipment was installed
and that exemplary operation and maintenance procedures were used with respect
to the emission control system.
2-4
-------
operation, but not later than 180 days after initial start-up of the
facility. This is a uniform requirement for all affected facilities
under 40 CFR 60.8 (General Provisions). Following this initial testing,
performance tests must be conducted at least once a month during the life
of the facility to verify continued compliance. Less frequent testing of
the anode bake plant or the primary control systems for the potrooms may
be permitted, if the owner/operator can show that emissions have low
variability during day-to-day operations. The monthly test requirement
has been waived, in favor of annual testing, for the primary potline and
anode bake furnace control systems at the three plants with center-worked
prebake (CWPB) potlines subject to the NSPS. Measurement of secondary
emissions from all NSPS potrooms is required on a monthly basis.
2.3.2 State Regulations
2.3.2.1 Fluorides. Of the 17 states that now have, or have had,
operating primary aluminum plants, 14 have fluoride emissions regulations
(Table 2-2). Thirteen limit fluoride emissions directly and one
regulates atmospheric concentrations of fluorides. Another uses the PSD
permitting route to limit fluoride emissions. Comparing the regulations
listed in Table 2-2 with the recommended guidelines (Table 2-3) and the
NSPS (Table 2-1), it can be seen that one State (Oregon) imposes limitations
more stringent than the NSPS. Three other states have regulations comparable
to the NSPS maximum, and four adopted the EPA guidelines.
2.3.2.2 Particulate Matter. All 17 states have standards for
particulate matter (PM) (Table 2-4).
2.3.2.3 Visible Emissions. Thirteen of the 17 states have visible
emission limits (Table 2-5).
2.3.2.4 Other State Regulations. Three states have sulfur dioxide
(S0£) regulations which are applicable to non-fuel burning sources in
primary aluminum plants (Table 2-6). The Maryland limit of 500 parts
2-5
-------
TABLE 2-2
STATE REGULATIONS FOR FLUORIDE EMISSIONS FROM EXISTING PRIMARY ALUMINUM PLANTSl3
State
Al abama
Arkansas
Indiana
Kentucky
Louisiana
Maryland
Missouri
Montana
New York
North Carolina
Ohio
Affected facility3
Potroom groups, all types
VSS
SWPB
HSS
L CWPB
Potroom groups and
Anode bake furnace
Potroom groups with dry
scrubbers
Potroom groups with wet
scrubbers
Potroom groups, all types
HSS
PB •
Potroom group
Anode Bake Furnace
Potroom group and Anode
bake furnace
SS Potroom group
SS Potroom group
PB Potroom group
Anode Bake Furnace
PB Plant
Standardb
None
98.5% TF removal efficiency
80% capture efficiency
80%
90%
95%
90% capture efficiency
95% TF removal efficiency
1.9 Ib TF/ton Al
1.9-2.5 Ib TF/ton Al
3.25 Ib F/hr from roof monitor
1.0 Ib gas F/ton Al
0.01 gr/SCF
98.5% TF removal efficiency
90% capture efficiency
95% "
2.5 Ib TF/ton Al
0.1 Ib TF/ton Al equivalent
2.5 Ib TF/ton Al
2.6 Ib TF/ton Al
4.3 Ib TF/ton Al
4.2 Ib TF/ton Al
0.40 Ib TF/ton Al equivalent
95% capture efficiency
98.5% TF removal efficiency
None
Comments
Adopted EPA Guidelines
If design, O&M exemplary
Plant has exemption to 290 Ib/hr
SIP not yet approved by EPA
Adopted EPA Guidelines
Measurements made only at the
primary control stack
1 ton Anode production = 2 tons
Al uminum
(Si
-------
TABLE 2-2 (CONCLUDED)
State
Oregon
South Carol ina
Tennessee
Texas
Washington
W. Virginia
Affected facility
Plant
PB Potroom Group
Potroom groups, all types
CWPB
SWPB
Potroom groups, all types
VSS
SWPB
HSS
CWPB
PB Potroom groups
Standard
1.3 Ib TF/ton Al
1.0 lb TF/ton Al
12.5 tons F/month
1.02 lb TF/ton Al
1.34 lb TF/ton Al
98.5% TF removal efficiency
95% capture efficiency
80% capture efficiency
No fluoride emission limit
95% TF removal efficiency
80% Fume capture efficiency
80% "
85% "
95% "
90% fume capture efficiency
99% TF removal efficiency
Comments
Monthly average
Annual average
Total from all sources
12 month running average
Excursion (monthly average);
PSD permit requirements
Apply air quality standards
no
i
VSS - Vertical stud Soderberg
HSS - Horizontal stud Soderberg
SWPB - Side-worked prebake
CWPB - Center-worked prebake
SS - Stud Soderberg
PB - Prebake
TF - Total fluorides
lb TF/ton Al - Pounds total fluorides per ton aluminum
produced
lb F/hr - Pounds fluoride per hour
gr/SCF - Grains (particulate) per standard cubic foot
-------
TABLE 2-3
STATE GUIDELINES FOR CONTROL OF FLUORIDES FROM EXISTING PRIMARY ALUMINUM PLANTS14
Pot
typea
VSS
HSS
SWPB
CWPB
Recommended control efficiencies
Primary
Collection
80
90
80
95
Removal
98.5
98.5
98.5
98.5
Secondary
removal
75
-
75
-
Expected fluoride emission range for potlines with
recommended controls (Ib TF/ton Al )&
Primary emissions
0.4 - 0.7
0.4 - 0.6
0.4 - 0.6
0.4 - 0.9
Secondary emissions
1.5 - 2.7
2.8 - 4.5
1.9 - 2.7
1.3 - 3.3
Total emission
1.9 - 3.4
3.2 - 5.1
2.3 - 3.3
1.7 - 4.2
I
00
a VSS
HSS
SWPB
CWPB
TF
Al
Vertical stud Soderberg
Horizontal stud Soderberg
Side-worked prebake
Center-worked prebake
Total fluorides
Aluminum
Ib TF/ton Al = Pounds total fluorides per ton aluminum produced
-------
TABLE 2-4
STATE REGULATIONS FOR PARTICULATE MATTER
EMISSIONS FROM NEW PLANTS15
PO
Particulate matter (PM)
limit3
E = 4.10p°*67
= 55. Op0-11 - 40
E = 3.59p0.62
= 17.31pO.16
E = 0.24pO-67
= 0.39p0.082 _ 50
E = 0.551p
E = 0.048q°-62
E = 0.1% by wt.
0.03 gr/dscf
0.025-0.25 gr/dscf
7.0 Ib/TAP
5.0 Ib/TAP
15 Ib/TAP
E = 6.2 lb/hr
= 10.5 lb/hr
= 21.2 lb/hr
Prod. rate
(tph)
p<30
p>30
p<30
p>30
p<50
p>50
p<0.05
>200
— _ _ _
____
15
25
>50
Emission source Number
states
Misc. Process 6
Stacks
3
1
1
1
1
4
1
All sources 1
Potroom groups 1
Misc. Process 1
Stacks
States applying15
Indiana, Missouri, Montana, North
Carolina, South Carolina, Ohio
Alabama, Arkansas (option),
Tennessee
New York
Ohio
Texas
Indiana
Indiana (non-attain.), Maryland
(areas 3&4), Missouri, and New York
Tennessee
Oregon - monthly
- annual
Washington
West Virginia
b Several states have more than one type of standard.
a E = particulate matter emission limit, pounds per hour
p = production rate, tons per hour
q = ai r flow, actual cubic feet per minute
gr/dscf = grains per dry standard cubic foot
Ib/TAP = pounds per ton aluminum produced
lb/hr = pounds per hour
-------
TABLE 2-5
STATE REGULATIONS FOR VISIBLE EMISSIONS
FROM NEW PLANTS16
Opacity
1 imi t
(*)
Emission source
Number
of
states
States applying
0
10
20
Misc. Process Stacks
Misc. Process Stacks
Pri. Al urn. Potlines
Misc. Process Stacks
Misc. Process Stacks
11
Md. (areas III & IV only)
Ore.
Mont.
Ala., Ark., Md., Mo., Mont.,
NY, NC, Ohio, Tex., Wash.,
and W.Va.
Ind.
2-10
-------
TABLE 2-6
STATE REGULATIONS FOR CONTROL OF S0£, CO, NOX, and HC
FROM NON-FUEL BURNING SOURCES IN PRIMARY ALUMINUM PLANT
Pollutanta
S02
CO
NOX
HC
Standard
60 Ib/TAP
500 ppm
2000 ppm
500 Ib/day
None
None
States applying
Washington
Maryland (new plants)
Louisiana
Mary! and
a S02 = Sulfur dioxide
CO = Carbon monoxide
NOX = Nitrogen oxides
HC = Hydrocarbons
b 1 b/TAP = pounds per ton aluminum produced
ppm = parts per million
2-11
-------
per million (ppm) could be restrictive for new stud Soderberg plants.
The Washington standard of 60 lb S02/TAP limits the sulfur content of coke
used in the anodes to about 3 percent.
One state (Maryland) has a carbon monoxide (CO) standard applicable to
non-fuel burning sources in primary aluminum plants. For the one plant
affected, this limit of 500 Ib/day is roughly equivalent to 1 lb CO/TAP
at full production.
There are no applicable State regulations for nitrogen oxides (NOX)
or hydrocarbons (HC).
2.3.3 PSD Regulations
Prevention of significant deterioration regulations apply to major
sources of air pollutants subject to regulation under the Clean Air Act.18'19
A primary aluminum reduction plant is classi-fied as a major source if it emits,
or has the potential to emit, 90.7 megagrams per year (Mg/yr) (100 tons/year)
or more of a regulated air pollutant.20 Pollutants emitted by primary
aluminum plants which are regulated under the Act include: fluorides, SOg,
NOX, PM, and CO. The preconstruction review and best available control
technology (BACT) requirements of PSD apply to both new and modified plants.
Total fluoride emissions from two primary aluminum plants are controlled
under PSD regulations. One is a new plant and one is an existing plant
with a new potline. Determinations of BACT for those plants are listed
in Table 2-7. Both impose considerably more stringent TF limits than
does the NSPS. The PSD regulations also impose limits on S02 at three
plants and PM at two plants. The S02 standard at one plant would require,
in the absence of add-on S02 controls, the use of a very low sulfur coke
(about 0.7 percent).
2-12
-------
TABLE 2-7
BACT DETERMINATIONS FOR PLANTS SUBJECT TO PSD REGULATIONS21
Plant &
location
Permit
date
Source
Pollutants &
allowable emissions^/
Comments
Alumax 2/78 Potlines TF
Goose Creek, U2
S. Carolina S02
PM
Anode
bake
plant
Goldendale,
Washington
Alcoa
Wenatchee,
Washington
1,2,3
TF
Commonwealth 8/78 Potlines TF
S02
PM
2/82 Potlines S02
1,2,3
1.02 Ib/TAP 12 month running average,
excursion to 1.34 allowed
269 lb/hr/scrubber(P) Sulfur contents of coke
2.75 lb/hr/scrubber(S) & pitch limited to 3.0%
& 0.6%, respectively
5.92 lb/hr/scrubber(P)
9.07 lb/hr/scrubber(S)
0.02 Ib/TAP equivalent
(0.04 Ib/TAC)
1.3 Ib/TAP
13.97 Ib/TAP
4 Ib/TAP
46.0 Ib/TAP
To drop to 0.8 Ib/TAP
after one year
Roughly equivalent to
the use of 0.7% sulfur
coke
Sulfur content of coke
limited to 3.0%
£/ TF = Total fluorides
S02 = Sulfur dioxide
PM = Particulate .matter
(P) = Primary emissions
(S) = Secondary emissions
Ib/TAP = Pounds/ton aluminum produced
Ib/TAC = Pounds/ton anode consumed
2-13
-------
2.4 REFERENCES FOR CHAPTER 2
1. U.S. Environmental Protection Agency. Standards of Performance
for Primary Aluminum Reduction Plants. Title 40, Chapter I,
Subchapter C, Part 60, Subpart S. Federal Register, Vol. 41, No. 17.
Monday, January 26, 1976. Pages 3826-3830.
2. U.S. Environmental Protection Agency. Primary Aluminum: Guidelines
for Control of Fluoride Emissions from Existing Aluminum Plants.
EPA 450/2-78-049b. December 1979.
3. U.S. Environmental Protection Agency. Standards of Performance for
New Stationary Sources: Primary Aluminum Plants; Amendments. 40 CFR
Part 60, Subpart S. Federal Register. Vol. 45, No. 127. Monday,
June 30, 1980. Pages 44202-44217.
4. U.S. Environmental Protection Agency. Background Information for
Standards of Performance: Primary Aluminum Industry, Vol. 1.
EPA 450/2-74-020a. October 1974. Pages xvii and xviii.
5. Reference 2, Chapter 12.
6. United States Congress. The Clean Air Act as amended August 1977.
Serial No. 97-4, U.S. Government Printing Office. September 1981.
Section lll(b)(l)(B). Page 31.
7. U.S. Environmental Protection Agency. Standards of Performance for
Mew Stationary Sources, Proposed Rule for Primary Aluminum Plants.
40 CFR Part 60, Subpart S. Federal Register. Vol. 39, No. 206.
Wednesday, October 23, 1974. Pages 37730-37741.
8. Reference 1.
9. U.S. Environmental Protection Agency. Standards of Performance for
New Stationary Sources, Primary Aluminum Industry. 40 CFR Part 60,
Subpart S. Federal Register, Vol. 43. No. 182. Tuesday, September 19,
1978. Pages 42186-42198.
10. Reference 3.
11. Reference 1.
12. Reference 3.
13. Memo from W.H. Maxwell, EPA:ISB, to Primary Aluminum Docket (A-86-07).
September 18, 1986. State laws.
14. Reference 2, Table 8-4.
15. Reference 13.
2-U
-------
16. Reference 13.
17. Reference 13.
18. U.S. Environmental Protection Agency. Prevention of Significant
Deterioration of Air Quality. Title 40, Chapter I, Subchapter C,
Part 51.24. Code of Federal Regulations. U.S. Government Printing
Office. 1984. Pages 610-626.
19. Reference 18, Part 52.21, Pages 19-36.
20. Reference 18, Paragraph (b)(l)(i)(a), Page 611.
21. U.S. Environmental Protection Agency. Compilation of BACT/LAER
Determinations, Revised. EPA-450/2-80-070. May 1980.
Source Code 7.1.
2-15
-------
3. THE PRIMARY ALUMINUM INDUSTRY
3.1 THE INDUSTRY
As of May 1986, the primary aluminum reduction industry consisted of 23
plants located in 14 states (Table 3-1). They are owned by 11 companies,
•
many of which are multi-nationals and also operate plants in other countries.
The types of reduction pots used in the individual plants, and their
production capacities, are also shown in Table 3-1. Many of these plants
periodically operate at reduced capacity and some have taken part of
their capacity out of service permanently.
Industry growth has been negative in the United States in recent years
with ten plants having shut down since 1978 (Table 3-2). These plant
closings have not been offset by the new facilities noted in Chapter 5. At
this time, there are no known plans by any company to construct primary
aluminum reduction facilities in the United States.^ The economics of
aluminum production, particularly with regard to energy and labor costs, are
not conducive to the expansion of primary aluminum production in the U.S.2~4
In fact, further plant shut-downs or production cut-backs may result from these
same economic considerations. In addition, there are no known plans whereby
an existing facility will become an affected facility through either modification
or reconstruction over the next 5 to 10 years.
3.2 PLANT DESCRIPTION
The major components of a primary aluminum reduction plant are:
0 a storage area for raw materials and finished product
0 one or more potlines where alumina feed material is reduced into
alumi num
0 a cast house in which the aluminum is reheated and purified, its
characteristics are modified to meet various specifications,
and it is cast into ingots
-------
TABLE 3-1
LISTING OF OPERATING PRIMARY ALUMINUM PLANTS
IN THE UNITED STATES AND THEIR CAPACITIES - MAY 19865-7
1
1 Plant name/location
1
1
INDIANA
Alcoa, Newburgh (Warrick)
KENTUCKY
Nat'l. Southwire, Hawesville
Alcan, Sebree
MARYLAND
Eastalco (AT umax), Frederick
MISSOURI
Noranda, New Madrid
MONTANA
Arco, Columbia Falls
NEW YORK
Alcoa, Massena
Reynolds, Massena
NORTH CAROLINA
Alcoa, Badin
OHIO
OTmet, Hannibal
OREGON
Reynolds, Troutdale
SOUTH CAROLINA
Alumax, Goose Creek
1 1
1 Plant capacity (1,000 TAP/yr)a |
1 1 1
1 By pot typeb I Total I
I 1 1 1 1 1
1 CWPB | SWPB | VSS I HSS | |
298 298
190 190
180 180
176 176
225 225
180 180
226 226
126 126
127 127
270 270
130 130
200 200
(Mount Holly)
3-2
-------
TABLE 3-1 (concluded)
1
I Plant capacity (1
1
1 Plant name/location
1
1
TENNESSEE
Alcoa, Alcoa
Consolidated Aluminum,
New Johnsonville
1
1 By pot
1 1
1 CWPB I SWPB
220
146
typeb
1
1 VSS
,000 TAP/yr)a
1
1 Total
1 I
1 HSS 1
220
146
TEXAS
ATcoa, Rockdale 342
WASHINGTON
Intalco (Alumax), Ferndale 280
Kaiser, Mead (Spokane) 220
Kaiser, Tacoma
Alcoa, Vancouver 121
Alcoa, Wenatchee 226
Reynolds, Longview
Commonwealth, Goldendale
WEST VIRGINIA
Kaiser, Ravenswood 164
Totals 3,139 602
185
80
210
342
280
220
80
121
226
210
185
164
365
416
4,522
a TAP/yr = Tons (Short) aluminum production/year (1
b CWPB = Center-worked prebake
SWPB = Side-worked prebake
VSS = Vertical stud Soderberg
HSS = Horizontal stud Soderberg
ton =0.9 megagram)
3-3
-------
TABLE 3-2
LISTING OF NON-OPERATING PRIMARY ALUMINUM PLANTS IN THE
UNITED STATES AND THEIR CAPACITIES - MAY 19868'13
I
I Plant name/location
Plant capacity (1,000 TAP/yr)a
TJWB"
By pot typeb I Total I
SWPB | VS3 1 HSS | 1
ALABAMA
Severe, Scottsboro
Reynolds, Lister-hill (Sheffield)
ARKANSAS
Reynolds, Arkadelphia
Reynolds, Jones Mills
1
125
LOUISIANA
Kaiser, Chalmette
Reynolds, Lake Charles
OREGON
Martin-Marietta, The Dalles
TEXAS
"Alcoa, Palestine
Alcoa, Point Comfort
Reynolds, Corpus Christi
(San Patricio)
Totals
126
116
16
36
168
202
51
116C
1256
116f
369
90"
116
90
185 185J
114 114k
275 483 1,068
TAP/yr = Tons (short) aluminum production/year (1 ton = 0.9 megagram)
CWPB = Center-worked prebake
SWPB = Side-worked prebake
VSS = Vertical stud Soderberg
HSS = Horizontal stud Soderberg
Operations indefinitely suspended in 1982. Company filed for Chapter 11
bankruptcy; seeking buyer for facility.
Smelter shut down in 1985. Company announced permanent closure in 1986.
Smelter shut down in 1985. Company announced permanent closure in 1985.
Smelter shut down in 1983.
Consolidated Aluminum announced permanent closing in 1981. Not restarted when
purchased by Reynolds in 1983.
Plant closed in 1984. Company seeking buyer for facility.
Plant used an experimental chloride reduction process. Company wrote off
investment in 1985.
Smelter temporarily closed in 1978; shut down in 1980. Company announced
permanent closing in 1982.
Smelter shut down in 1981. Company announced permanent closing in 1984.
c
d
e
f
9
3-4
-------
0 a power source for the direct current (DC) voltage used in
the reduction process
0 maintenance and repair facilities
0 an anode bake plant (optional) where the anodes used in some
types of pots are prepared
A simplified diagram of a typical plant showing material flow patterns is
provided as Figure 3-1. Figure 3-2 is a somewhat more detailed schematic
showing many of the process operations performed in a typical plant.
An aluminum reduction potline is typically housed in one or two long,
narrow buildings called potrooms (Figure 3-3). It usually consists of
150 to 200 aluminum reduction pots (cells). Aluminum reduction pots are
shallow, rectangular vessels which may be lined up side-by-side or end-to-
end in one or more rows down the center of the potroom. All of the pots
in a potline are electrically connected, in series, with a typical DC
voltage drop across each pot of 4 to 5 volts. The current flow through
each pot may range from 40,000 to 280,000 amperes (150,000 amperes or
more in newer designs). The pots are large heat sources, so the potrooms
are ventilated to maintain reasonable working conditions and to ensure
proper pot operation. Usually this ventialation air enters at the sides
of a potroom and exits through roof vents (roof monitors).
Alumina and other raw materials are delivered to the plant by ship or
railcar and stored. Alumina is transferred to the aluminum reduction
pots as needed by airslide or crane-mounted hopper. Aluminum fluoride,
sodium carbonate, and fluorspar are added to the pots manually or by
hopper. Coke and pitch are mixed and either delivered to the bake plant
for forming and baking, or transferred directly to the pots, depending on
plant type.I/ Periodically, the aluminum is removed from the pots by a
process called "tapping" and transferred, still molten, to the cast house
in crucibles or ladles. There, it is placed in holding furnaces or cast
furnaces, alloying materials (iron, silicon, magnesium, and manganese)
are added, and the aluminum alloy is fluxed with mixed gas or solid fluxes
or with argon or chlorine to remove impurities. The purified alloy, still
The principal differences between primary aluminum plants are in the
types of pots (cells) they use. Pot descriptions are presented in more
detail in Section 3.5.
3-5
-------
FIGURE 3-1
BLOCK DIAGRAM
PRIMARY ALUMINUM PLANT
^Product
Raw Material
I Shipping & I
1 Receiving |
I Area I
I
I Raw Mat'l.I
I Storage I
I I
I Anode I
I Bake !•
I Furnace I
I—
I Aluminum I
•I Reduction j"
I Potline I'
I I
I I
I Cast I
I House f
I I
Product
Storage
I Power I
"I PI ant I
"I I
3-6
-------
Figure 3-2
PROCESS OPERATIONS IN A PRIMARY ALUMINUM PLANT
Product
Raw Materials
Shipping & Receiving
Coke
Pitch
Crushed Anode I Anode Paste I
Butts I Preparation I
— •»! I
Anode
Paste
I
I Anode Forming !<«--
I & Baking I Baked Anodes
I I
Alumina
I Dry
. [~*1 Scrubber |
Fluorspar
Cryolite
Aluminum
Fluoride
r
Anode Butts I
I
Aluminum Reduction
•I Anode Butt I*-1
I Crushing I
I I
Molten
Aluminum
Molten Aluminum
1
Aluminum I
Tapping I
I
I III Aluminum
I Aluminum | Molten Aluminum I Aluminum j Ingots
I Fluxing & Alloying | I Casting j
Legend: Dotted lines (—) indicate operations performed only in
a prebake plant
3-7
-------
Figure 3-3
PLAN VIEW OF TYPICAL POTLINE
Potroom
i r
i r
s
Potroom
A I
View A-A
Cross-Section of Potroom
Roof Vent
and Hopper
or
ina Airslide
Alumina Storage Hopper
Anode Assembly
Anode Bus Bar
f /-Side
/ Shield
Side Vent-
3-8
-------
molten, is then direct chilled, cast into ingots, billets, or slabs, or is
poured into molds to set. After cooling, the aluminum ingots are transferred
to storage or shipped.
3.3 PROCESS DESCRIPTION
From its inception in 1886, the primary aluminum industry in the United
States has used the Hall-Heroult process to electrolytically reduce aluminum
oxide (alumina) to aluminum.^/ Alumina, an intermediate product, is refined
from bauxite ores using the Bayer process.
The reduction of alumina to aluminum is carried out in shallow rectangular
pots, or cells. A pot consists of a shell supported by a pot cradle, lined
with insulating material and having an electrically conductive bottom and sides
made of carbon. It is filled with molten cryolite.^/ One or more carbon blocks
are suspended above the pot and extend down into the cryolite bath
(Figure 3-4).
A low voltage direct electric current is passed through the cryolite
bath, which serves as an electrolyte and a solvent for the alumina, from
the carbon blocks (anodes) to the molten aluminum on the bottom of the pot
(cathode). Heat produced by resistance to this current flow keeps the
cryolite molten and at a temperature of about 950°C (1740°F).^/ A crust
is allowed to form over the cryolite in the pot. This crust contains
alumina and cryolite. It helps reduce heat loss and protect the pot
lining and is broken only to add fresh alumina or to allow the escape of
generated gases.
£/ One small, experimental plant in the United States used the aluminum
chloride process.
^f Cryolite is a double fluoride salt of sodium and aluminum (Na3AlF6).
It is formed by the chemical mixing of two salts, sodium fluoride (NaF) and
aluminum fluoride (A1F3).
4/ pure cryolite has a freezing temperature of about 1008°C (1846°F).
3-9
-------
GO
I—•
O
FIGURE 3-4
ALUMINUM INDUCTION POT
DIRECT CURRENT 1
I CARBON ANODE
I ALUMINUM PAD
MOLTEN CRYOLITE BATH
-4. C^E
rv-
-------
Alumina is periodically added to, and dissolves in, the molten
cryolite bath. Cryolite, in its molten state, has the capability to
dissolve up to 8 percent alumina.14 The alumina then disassociates into
its components,' the molten aluminum settles to the bottom of the pot, and
the oxygen migrates to the carbon anode. There, it reacts with the
carbon, sulfur, and other impurities in the anode to form carbon dioxide
*
(C02), carbon monoxide (CO), sulfur dioxide (S02), etc. The anodes are
lowered as they are consumed, which occurs at the rate of about 0.23 kilogram
(kg) (0.5 pounds [lb]) of carbon per 0.45 kg (1 Ib) of aluminum produced.
The theoretical energy requirements for extracting aluminum from
alumina are 20.3 megajoules per kilogram (MJ/kg) (2.56 kilowatt-hours/Ib
[kWh/lb]) of aluminum produced.15 In practice, however, energy is required
-to bring the reactants up to temperature, is lost in the exhaust gases or
through radiation into the potroom, and is removed from the pot when the
molten aluminum is tapped. The increase in energy costs in recent years
has fueled efforts to reduce energy losses, with some success. In the
early 1970's, a modern pot consumed approximately 56 MJ/kg (7 kWh/lb) of
aluminum produced, while more recent pot designs require only 48.4 MJ/kg
(6.1 kwh/lb).16'19 The newer pot operates at 185 kiloamperes (kA) and
4.1 volts.
3.3.1 Bath Ratio
Cryolite is added to the bath periodically to replenish material
that is removed or consumed in normal operation, as is aluminum fluoride.
The bath (weight) ratio of sodium fluoride to aluminum fluoride required to
form a pure cryolite is 1.50. However, it has been found that adding excess
aluminum fluoride to reduce the bath ratio increases pot current efficiency
and lowers the freezing temperature of the bath, thus permitting lower
3-11
-------
pot operating temperatures. Bath ratios in use range from 1.05 to 1.50.
Calcium fluoride, or fluorspar, may also be added to lower the melting
point of the cryolite.
3.3.2 Tapping
The molten aluminum which collects in the bottom of the pot is
periodically removed by "tapping". This involves the use of a ladle or
crucible with a long snout, which is lowered through the cryolite bath
into the layer of molten aluminum (Figure 3-5). Then, aspiration air is
used to create a suction and the aluminum is sucked up into the ladle.
The ladle is then moved to the next pot and the cycle is repeated. A
tapping cycle takes about 5 minutes for a center-worked prebake
(CWPB) pot, of which 1.5 to 2 minutes may be actual siphon time.20 when
full, the ladle may be transported directly to the cast house with the
cover still in place. There, any cryolite that is accidentally siphoned is
recovered as part of the dross skimmed from the surface. Alternatively,
the cover may first be removed and placed on an empty ladle. In the
latter case, any cryolite picked up with the aluminum quickly rises to
the surface and freezes. The chunks of cryolite are scraped onto the
potroom floor and the ladle with the still molten aluminum is routed to
the cast house.
3.3.3 Anode Effects
An anode effect can be caused by either a shortage or (rarely) an
excess of alumina in the pot. It takes 2 to 5 minutes to correct either
type of anode effect.21 The cryolite bath normally contains 5 to 8
percent alumina at saturation.22 If too much alumina is added to the pot
bath, the excess does not dissolve. Instead, it settles to the bottom of
the pot and increases resistance. This condition is corrected by shutting
off the flow of alumina and rowelling (stirring) the pot with a steel rod
to put the alumina in suspension where it can more readily dissolve.
3-12
-------
FIGURE 3-5
TAPPING MOLTEN ALUMINUM
FROM
PRIMARY ALUMINUM REDUCTION POT
Crucible Cover
CO
I
ALUMINUM REDUCTION POT
-Alumina-Cryolite Crust
Cryolite Bath
r Molten Aluminum
-------
If insufficient alumina is added to the pot, a gas film forms on the
surface of the anodes and creates a barrier to the flow of electrical current.
The pot voltage then increases from 4 to 4.5 volts to 50 to 100 volts in
seconds. This condition is corrected by adding alumina and changing the
height of the anodes, or sticking a green wooden pole under one or more
anodes and stirring (usually from an end door). For pots under computer
control, this condition can usually be corrected without human intervention,
by adding more alumina, by adjusting the height of the anode, or by shaking
or swaying the anodes. If these actions are unsuccessful, the computer calls
for assistance. A worker then sticks a green wood pole under one or more
anodes and stirs to dissipate the gas layer. This can usually be accomplished
from an end door, but sometimes one or more side shields must be removed.
Anode effects resulting from underfeeding are much less objectionable
than those from overfeeding, so plants may purposely underfeed alumina--using
the results as an analytical tool to determine when alumina is needed.23
These plants either get one anode effect per day or they reduce the alumina
feed.24 Other facilities operate with less than one anode effect per week.25
3.4 TYPES OF PLANTS.IN USE
Primary aluminum reduction plants are characterized by the type of
reduction pots (cells) they contain. There are two major types: prebake
and stud Soderberg. A majority of the primary aluminum plants in the
U.S. currently use prebake technology (18 of 23, or 78 percent). Also,
three of the four plants which have potlines subject to the new source
performance standards (NSPS) use prebake pots.
3-14
-------
The pots in prebake plants use multiple anodes which are formed and
baked prior to use, while the stud Soderberg pots use a single, continuous
anode which is shaped and baked in place. Each of these pot types has,
in turn, two variations. The pots in prebake plants are classified as
CWPB or side-worked prebake (SWPB), depending on where the pot working
(crust breaking and alumina addition) takes place. Stud Soderberg pots,
on the other hand, are differentiated by the positioning of the current-carrying
studs in the anodes. They may be inserted vertically (VSS) or horizontally
(HSS).
The anode bake plants which produce the anodes used in prebake pots
are of two basic types. One is the ring furnace and the other the tunnel
kiln.
3.5 ALUMINUM REDUCTION POTS
As noted above, primary aluminum reduction plants are characterized
by the type of reduction pot (cell) they use. Each of these pot types is
discussed in the following sections.
3.5.1 Center-Worked Prebake Pots
In the mid 1970's, 16 primary aluminum plants used CWPB pot technology.
Two of these plants have since ceased operation while two have added new
potlines. In addition, one new CWPB plant has been constructed. These
15 plants represent 65 percent of the total U.S. plants and 69 percent of
domestic production capacity.
3.5.1.1 Design and Operation. A cross-sectional view of a CWPB pot
is shown in Figure 3-6. Each CWPB pot may hold from 18 to 26 closely-spaced
anode assemblies^/ in two parallel rows running the length of the pot.
Alumina is delivered to CWPB pots by a crane-mounted hopper or by air-slide
and stored in hoppers located atop the pot superstructures. The hoppers
run the full length of the pots, between the anode bus bars. The anode
assemblies, which are suspended on these bus bars, are positioned close
^/ An anode assembly consists of an anode and a hanger. The hanger is
positioned in a recess in the top of the anode after baking and molten iron
is poured around it to hold it in place. The hanger serves both to support
the anode and to transfer electricity from the bus bar to the anode.
3-15
-------
FIGURE 3-6
CENTER-VKDRKED PREBAKE POT
Alumina
hopper
Anode beam
Gas collection hoods
Gas off take
Frozen flux and
alumina
Steel shell
Iron cathode bar
3-16
-------
to the sides of the pot to provide an area in the center for "pot working".
All the anodes in a pot can be raised or lowered simultaneously by moving
the anode bus bars, which have a vertical travel of 25 to 36 centimeters
(cm) (10 to 14 inches [in.]).26 Additionally, each anode assembly can be
adjusted individually by releasing its latch and repositioning it on the
bus bar. The anode assemblies are lowered as the carbon anodes are consumed.
The spent anodes (butts) are replaced on a rotating basis, usually at the
rate of about one per day for each pot.
The pot superstructure has a number of crustbreakers (punchers) mounted
on the underside of each alumina hopper that serve a dual function. When
activated, they extend down to punch holes in the crust over the molten
cryolite. Then, as they retract, they release a metered amount of alumina
into the holes. At the newer plants, the crust-breaking frequency of each
pot and, thus, its alumina feed rate is monitored and controlled by computer.
In this way, the frequency and severity of anode effects and other pot
malfunctions can be minimized.
To prepare for tapping, an end door on the hood is opened and the crucible
spout is inserted. After tapping, this operation is reversed.
3.5.1.2 Anode Replacement and Reclaiming. The anode replacement
process usually takes 3 to 4 minutes.27 To remove a spent anode assembly
(anode butt), two to three side shields are removed and the crust around the
anode is broken with a jackhammer. Then, the anode butt is clamped to a
crane, the latches holding it to the bus bar are released, and the anode
butt is extracted and placed on the potroom floor to cool. At some plants
the spent anode assemblies are removed within 30 minutes and transferred to
a holding area to cool. A fresh anode assembly is then clamped in place,
after first removing any floating chunks of cryolite or anode which could
prevent it from seating properly. Finally, a layer of recycled, crushed
bath is spread over the top of the anode and the side shields are replaced.
After cooling, anode butts are cleaned, crushed, and recycled. First,
jackhammers and brushes are used to remove most of the caked-on cryolite and
alumina. This cryolite/alumina mix (typically about 25 percent alumina)
3-17
-------
is crushed and used to cover and insulate the fresh anodes. The hangers
are then removed and refurbished and the butts are crushed and recycled
to the green anode mix.£/
3.5.2 Side-Worked Prebake Pots
At the time the NSPS was proposed there were six plants in the
United States with potlines using SWPB pots. Three of these plants have
since ceased operations, leaving a total of three (13 percent of U.S.
total and 13 percent of U.S. capacity). No SWPB potlines have been
constructed since the NSPS was proposed in 1974.
A cross-sectional view of a SWPB pot is shown in Figure 3-7. Late
model SWPB pots differ from CWPB pots in the placement of the anodes, the
type of side-shield used, and the method of alumina addition. Alumina is
added along the sides of the SWPB pot, rather than down the center as in
a CWPB pot, so the two rows of anodes are set close together near the
center-line of the pot. The side shields are two one-piece covers (one per
side) which are hinged at the bottom and motor-driven. Alumina addition
is typically accomplished using a gantry crane. This crane carries an
alumina hopper and two jackhammers. On a predetermined cycle (usually 3
to 4 hours), the pot covers swing open and the gantry crane straddles the
pot, one jackhammer per side. The crane moves the length of the pot and
the jackhammers break the crust between the anodes and the sides of the
pot. It then cycles up and down the pot, adding alumina until a wire
positioned just behind the alumina spout provides a cut-off signal. The
crane then rises, the covers swing closed, and the crane moves on to the
next pot.
£/ About 25 percent of the carbon used in green anodes is recycled
anode butts. This is the source of the fluoride emitted during anode
baking.
3-18
-------
FIGURE 3-7
SIDE-WDRKED PREBAKE EOT
Anode beam
Gas collection hoods
Frozen flux and
alumina
L'teel shell
Iron cathode bar
3-19
-------
3.5.3 Vertical Stud Soderberg Pots
Six primary aluminum plants utilized VSS pot technology at the time
the NSPS was proposed (two of these plants also used prebake pots).
Since then, the VSS potlines at two plants have ceased operation and one
new VSS potline has been constructed at an existing plant. These plants
constitute almost 9 percent of the domestic primary aluminum plants (2 of
23) and 8 percent of domestic capacity.
A cross-sectional view of a typical VSS pot is shown in Figure 3-8.
It utilizes a single large anode. A green anode paste is periodically fed
into the open top of a rectangular compartment, or casing, which serves to
shape the anode. As the bottom face of the anode is consumed, the paste
moves down inside this stationary casing, is compressed by the weight of
the material above it, and is gradually hardened and baked by the heat of
the pot.
Steel studs are positioned vertically in the green anode paste and
move down with it. They are rigidly connected to the bus bar and form an
electrical interface between the anode and the bus bar. Those studs
which project the farthest down into the anode are disconnected from the
bus and extracted, so that they will not become exposed to the bath at
the bottom of the anode. At the same time, a fresh stud is inserted and
clamped to the bus.
The green anode paste is composed of coke and a pitch binder. The
in-place baking of the anode results in the release of sulfur oxides from
the coke and hydrocarbon fumes and volatiles from the pitch binder. If
not removed from the gas stream, the fumes and volatiles tend to condense
in, and plug, control system hoods and ductwork.
Vertical stud Soderberg pots, because of their single anode, must be
sideworked. That is, crust-breaking and alumina addition take place
along the sides of the pots.
3-20
-------
STUDS -\
BUS BAR
RISERS
ANODE PASTE
BAKED ANODE
SKIRT
SOLIDIFIED CRUST
OF ELECTROLYTE
AND ALUMINA
STEEL SHELL
CARBON LINING
ELECTROLYTE
MOLTEN ALUMINUM
TO GAS
. TREATMENT
PLANT
ry—
1^ BURNER
#W— GAS AND TAR BURNING
GAS EVOLVING
CATHODE
COLLECTOR
BAR
THERMAL
INSULATION
FIGURE 3-8
VERTICAL STUD SODERBERG POT
3-21
-------
3.5.4 Horizontal Stud Soderberg Pots
The number of plants with HSS pots has declined from seven in the
early 1970's, to three by May 1986. No new HSS potlines have been built.
Although 13 percent of domestic plants use HSS pot technology (3 of 23),
it accounts for only 9 percent of domestic capacity.
The HSS pot, like its VSS counterpart, utilizes a single, large,
formed-in-place anode. The principle difference is the horizontal placement
of the studs. A typical HSS pot is shown in cross-section in Figure 3-9.
The anode casing is made of either steel or aluminum sheeting and removable
steel channels. The anode and its casing are suspended over the pot and
are moved downward as the anode is oxidized. The current-carrying studs
are inserted horizontally into the anode through perforations in the
steel channels at a point where the anode paste has not baked out.
Electrical contact with the bus bar is through flexible connectors. When
the lower channel reaches the bath, the flexible connectors are moved up
to the next row of studs, the bottom row of studs is extracted, and the
steel channel is removed.
3.6 ANODE BAKE FURNACES
Anode bake furnaces produce the anodes used in CWPB and SWPB pots.
They are located in carbon plants, which also contain pitch storage, coke
storage, green anode or paste production, and a rod shop. Two basic
types of furnaces are used in the United States, the open top ring furnace
and the tunnel kiln. A short description of each furnace type is provided
in the following sections.
3.6.1 Ring Furnace
Essentially all of the anodes produced for prebake plants in the
United States are baked in open-top ring furnaces. Since the advent
of the NSPS, eight ring furnaces have been built for anode bake plants at
five locations.
3-22
-------
ALUMINA HOPPER
FULLY SAKED ANODE
SOLIDIFIED CRUST
OF ELECTROLYTE
AND ALUMINA
STEEL SHELL
INSULATION
CARBON LINING
,—- REMOVABLE
\ CHANNELS
-f\
GAS COLLECTION DUCT
ANODE PASTE
POT
ENCLOSURE DOOR
PARTIALLY BAKED
PASTE
ANODE STUDS
GAS AND FUME
EVOLVING
MOLTEN ALUMINUM
CATHODE
COLLECTOR
BAR
FIGURE 3-9
HORIZONTAL STUD SODERBERG POT
3-23
-------
Ring furnaces vary greatly in size and production rate, but all have
the same basic layout and operating parameters. Each ring furnace consists
of a large number of indirectly fired sunken ovens, or pits, arranged in
rows as shown in Figure 3-10. The pits are open-topped and made of brick.
Some of the spaces between the bricks are mortared, while others are inten-
tionally left open. The pits sit in, and are surrounded by, a flue which
is split down the middle by a wall. The wall is slightly shorter than the
flue, to permit the flue gases to pass from one side to the other at each
end. A large pipe, or duct, circles the ring furnace and leads to an
exhaust fan. Double-sealed manholes are spaced along the top of this duct,
with at least one manhole per furnace section. Each one-half row of pits,
from the center wall out, is called a section.
An operating furnace will have one or more "fires" operating continuously
A fire, as will be discussed in the following paragraphs, has three phases:
preheat, bake, and cool-down. Each fire gradually traverses the length of
the furnace on one side in a series of steps, one section per step. It
then returns on the other side. Ahead of the fire(s), pits are filled
with green anodes to within about 0.9 meter (m) (3 feet [ft]) of the surface.
Petroleum coke is then dumped into the pits from an overhead hopper and
packed around the anodes. The anodes are then covered with coke, petroleum
coke, or some other insulating material to slightly above the tops of the
pits. After the fire has passed by and the baked anodes have cooled, the
packing coke is removed from the pits by vacuuming or other means, and
reused. The baked anodes are then removed and necessary pit repairs are
performed while the pits are empty. Both the coke placement and removal
operations can be very dusty.
As previously noted, a "fire" (sometimes called a firing cycle) has
three phases: preheat, bake, and cool-down. Ambient air is drawn or
forced?/ into the flue and around pits containing just-baked anodes. In
the process, the air is heated and the anodes are cooled down. Usually,
11 Some furnaces do not have a forced draft air supply. In those cases
all the draft is supplied by an exhaust fan.
3-24
-------
Exhaust Duct
Tapered-
Flue|Gas
Exhaust
Header
Manifold
0
Ovens(pits)
O-Double-sealed
Exhaust Port
o o
Pi==4l===P===P===n===n======-Movab1e
Fi ring Frame
oooo
-ELOJ3
0
-Exhaust Duct
9
0
0
0
fi
—
—
~~"
y
0
0
0
fl
—
I
~
y
0
0
0
1
—
u
0
0
0
—
y
0
0
0
y
0
0
0
flnfinl
F
0
0
Ambient
Ai r
—Fan
0
*»«.
0
-9
0
'^Am
Ai
He
^Ex
Duct
FIGURE 3-10
RING FURNACE LAYOUT
3-25
-------
the air preheat (anode cool-down) zone encompasses three to five sections.
The preheated air "then enters the firing zone (anode bake zone), usually
under slightly negative pressure .8_/ There, natural gas or other fuel injected
into the flue through movable firing frames is ignited by the high air
temperature in the flue, increasing the flue gas temperature to 1225 to
1250°C (2237 to 2282°F). In the process, the anodes are heated to about
1150°C (2100°F), partly by the heat from the flues and partly by the
calcining of the binder pitch in the anodes. One source reported that a
substantial percentage of the total energy used in the baking process
comes from the anode binder pitch.28 Another source reported that the
sulfur content of packing coke drops by 50 percent (from 4 to 2 percent)
during the bake process.29
The flue gases leaving the anode bake zone pass around the pits in the
anode preheat (flue gas cool-down) zone, transferring heat to the green
anodes, and become progressively cooler as they approach the movable
exhaust manifold. A typical exhaust temperature is about 300°C (570°F).
The negative pressure in the flue also increases with proximity to the
exhaust manifold. This pressure difference tends to draw fumes generated
in the pits through cracks and seams in the pit walls and into the flues.
There, if flue temperatures are adequate, and there is sufficient oxygen,
the fumes are incinerated. The movable exhaust manifold extracts the exhaust
gases from the flue at the end of the last gas cool-down section through
ports in its upper surface. It then vents them into the large duct
circling the furnace through one of the manholes located atop the duct.
From these, the gases are routed to either a control system or to the
atmosphere through a large exhaust fan.
8/ The negative pressure in the flue will be considerably higher in the
firing zone if air is not supplied under forced draft.
3-26
-------
3.6.2 Tunnel Kiln
Tunnel kilns are installed at only one plant in the United States.
Twelve such kilns were installed at that plant between 1956 and 1974 but
are currently being used only to provide "swing" capacity. There appears
to be little potential for additional installations. Problems cited are:30
0 Poor energy efficiency - heat requirements are greater than for
a modern ring furnace;
0 Poorer quality anodes - anodes are less dense than those baked
in ring furnaces and so are more apt to crumble, break, or fracture
during handling and have a lower life expectancy; and
0 Anode quality is not consistent - those on the tops and sides of the
stacks tend to be fired better than those in the center.
A kiln is a long, narrow, indirect-fired enclosure in which a controlled
atmosphere is maintained to prevent oxidation of the anodes. Each kiln
will hold either 44 or 54 railcars stacked with anodes (Figure 3-11).
There is a vestibule or air lock at each end which is large enough for
one car. Each vestibule has an inner and outer door, and an exhaust fan
ducted to the atmosphere. The outer door will not open until the vestibule
has been cleared of fumes which collect while the inner door is open.
3.7 PROCESS EMISSIONS
The principal emission points in a primary aluminum plant are the
potrooms housing the aluminum reduction pots and, for prebake plants, the
anode bake plant. Pollutants emitted include gaseous and particulate
fluorides, other particulate matter (PM), and other gases. Non-fluoride
PM emissions include alumina, carbon, hydrocarbon tars, and iron oxide
(FeOs). Gaseous non-fluorides include C02, CO, S02, hydrogen sulfide (^S),
carbonyl sulfide (COS), carbon disulfide (C$2), nitrogen oxides (NOX),
and water vapor.
Fluoride evolution from aluminum reduction pots and anode bake
furnaces were quantified during EPA source tests conducted during develop-
ment of the NSPS. These data are summarized here, as no additional data
were found during this study.
3-27
-------
Vent Stack
22 30
Positions
r
44
/v
Vestibule
Anode Preheat Section
Bake Zone Anode Cool-down
Section
FIGURE 3-11
TUNNEL KILN
Side View of a Tunnel Kiln
3-28'
-------
The amount of SOg emitted during the production of primary aluminum
is a function of the sulfur contents of the coke and pitch used in the
anodes. Sulfur levels have been rising in recent years. In the early
1970's, a typical purchase specification for petroleum coke contained a
1.5 percent sulfur limit. By 1978, this limit had been driven up to 3.0
percent.31*32 In 1985, sulfur contents of available cokes range from 2
to 7 percent with an average of about 3 percent.33 One source has projected
that the trend to increased sulfur in coke will continue until the sulfur
limits in petroleum coke purchase specifications reach 4 to 6 percent.34
Other sources, however, project that 3 percent sulfur petroleum coke will be
available for the next 10 years.35"37
3.7.1 Total Fluorides
The NSPS limits emissions of total fluorides (TF), which includes
both gaseous and particulate fluorides, from potrooms and anode bake
furnaces. Gaseous fluorides present in potroom emissions during normal
operation are reported to include hydrogen fluoride (HF) and silicon
tetrafluoride (SiF4). During an anode effect, fluorocarbons, principally
carbon tetrafluoride (CF4), and small amounts of hexafluoroethane (C2Fs)
are also known to be produced.38 Particulate fluorides identified include
cryolite (^AlFs), aluminum fluoride (A1F3), calcium fluoride (CaF2), and
chiolite (NasAl ^14) .
The ratio of gaseous to particulate fluorides in the TF emitted from
uncontrolled potrooms varies with pot type and operating conditions. One
study cited in a previous document reported that this ratio varied from
0.5 to 1.3.39 Uncontrolled TF emissions from anode bake furnaces are believed
to be mostly gases.
As previously noted, no data are available on TF evolution from any
potrooms or anode bake furnaces subject to the NSPS. However data are
available from other sources. Table 3-3 contains the results of tests
conducted by EPA for the NSPS and guidance documents, plus more recent
information supplied by manufacturers and users of primary aluminum
3-29
-------
TABLE 3-3
AVAILABLE INFORMATION ON UNCONTROLLED EMISSIONS OF TOTAL FLUORIDES43'46
Data source
BID
Guidance Doc.
IPAI Report
ALCOAC
BID
Guidance Doc.
Document
date
1974
1979
1985
1985
1974
1979
Pot
type3
CWPB
CWPB
CWPB
CWPB
VSS
VSS
No.
tests
2
• «
7
» _
2
—
TF emissions (Ib TF/TAP)b
Potline
Range
49.3-62.6
25.7-65.6
26.4-56.0
65-70
39.3-47.3
30.5-53.5
Avg.
__
40.8
43.5
__
__
44.4
Bake plant
Range
—
0.4-1.6
—
__
Avg.
--
0.86
--
--
a CWPB = Center-worked prebake pot
VSS = Vertical stud Soderberg pot
b Ib TF/TAP = pounds total fluoride per ton aluminum produced.
(1 Ib TF/TAP = 0.5 kilogram/megagram aluminum produced)
c ALCOA is a major primary aluminum reduction pot manufacturer. They
produced all the pots used on the CWPB potlines subject to the NSPS,
3-30
-------
reduction pots. Based on this information and the current trend to
maximize efficiency by reducing bath ratio,jV current TF evolution rates
are expected to be 33 kilograms per megagram (kg/Mg) (66 Ib/ton) of
aluminum produced and 25 kg/Mg (50 Ib/ton), respectively, for CWPB and
VSS potlines. There are no data available to either support or modify
the estimated TF emission rate of 0.43 kg/Mg (0.86 Ib/ton) of aluminum
produced for anode bake furnaces.
3.7.2 Sulfur Dioxide
Little test data are available on S02 emissions from potlines or
anode bake plants. It is generally recognized, however, that S02 i s
evolved from electrolytic pots in direct proportion to the anode consumption
rate and the sulfur content of the anode materials. Bake plant S02
emissions correlate to anode weight loss during baking and to any change in
sulfur content. At those anode bake plants where a lightly calcined coke
is used for packing material, the packing coke is an additional source of
S02.
Considering only the direct contribution of the anode materials, S02
emissions to the atmosphere are calculated to'be 53 kg/Mg (106 Ib/ton) of
anode consumed.^ This calculation is based on an anode composition of 85
percent coke and 15 percent pitch, where the coke and pitch have sulfur
contents of 3 percent and 0.6 percent, respectively. For an anode consumption
rate of 0.23 kg of anode per 0.45 kg (0.5 Ib/lb) of aluminum produced,
this correlates to 27 kg of SC£ per megagram of aluminum produced (53 Ib/ton).
If the sulfur content of the coke were to increase to 5 percent, the S02
emissions would increase to 44 kg/Mg (87 Ib/ton).
£/ Reducing the bath ratio of a pot tends to increase TF evolution.4°
Bath ratios of 1.30-1.45 were common in the 1970's.41 Since then, however,
large increases in the cost of power have forced plants to increase efficiency.
One of the ways this was accomplished was to lower bath ratio. Some CWPB
pots now operate at around 1.12 or 1.13 and the VSS pots at around 1.25.
Pot operating efficiency may also be improved through the use of proprietory
bake additives.42
3-31
-------
A 90,700 Mg (100,000 ton) per year primary aluminum plant using a 3
percent sulfur coke will generate 2,400 Mg (2,650 tons) of S02 per year
(274 kg S02/hr [605 lb/hr]). If the plant has HSS or VSS pots, all the S02
will be released by the potlines. For plants using CWPB or SWPB pots,
the S02 emissions are split between the potlines and the anode bake
furnaces. If it is assumed that 80 percent of the S02 emissions are
released into the pots, then the SO2 emissions distribution within the
primary aluminum plant is 1,925 Mg/yr (2,120 tons/yr) S02 from the
potlines and 480 Mg/yr (530 tons/yr) SO^ from the anode bake furnaces.48
(Another source indicates the split may be 95 percent from the potline
and 5 percent from the anode bake furnace.49)
Sulfur dioxide emissions from the packing coke used around the anode
in the anode baking pits were not considered in these calculations because
it is reused. However, it often has a higher initial sulfur content than
anode coke and must be replenished periodically with makeup coke.50 10/
10/ Fugitive emissions during the loading and emptying of the pits
account for some packing coke losses. Some coke also passes through
cracks in the pit walls and is burned in the furnace flue.
3-32
-------
3.8 REFERENCES FOR CHAPTER 3
1. Letter and attachments from Goldman, J.H., The Aluminum Association,
to Maxwell, W., EPA:ISB. May 21, 1986. Comments on draft document.
2. Letter and attachments from Dickie, R.C., Al umax of South Carolina, to
Noble, E.A., EPA:ISB. April 16, 1986. Comments on draft document.
3. Letter and attachments from Tropea, L.C., Jr., Reynolds Aluminum, to
Noble, .E.A., EPA:ISB. May 9, 1986. Comments on draft document.
4. Reference 1.
5. U.S. Department of the Interior. Preprint from the 1983 Bureau of
Mines Minerals Yearbook, Aluminum. Superintendent of Documents,
Washington, D.C. 20402. Page 5.
6. U.S. Department of the Interior, Bureau of Mines. Mineral Industry
Surveys. Aluminum Industry in January 1985 through Aluminum
Industry in April 1986, inclusive. Prepared in the Division of
Nonferrous Metals. Dated April 9, 1985, through July 3, 1986.
7. U.S. Environmental Protection Agency. Primary Aluminum: Guidelines
for Control of Fluoride Emissions From Existing Aluminum Plants.
EPA-450/2-78-049b. December 1979. Table 3-1.
8. Reference 1.
9. Reference 2.
10. Reference 3.
11. Reference 5.
12. Reference 6.
13. Reference 7.
14. Ravier, E.F. (Al uminum Pechiney, Paris). Technology of Al uminum
Reduction. In: Health Protection in Primary Aluminum Production,
International Primary Aluminum Institute. Proceedings of a Seminar,
Copenhagen, June 28-30, 1977. Page 17.
15. Reference 7, Page 4-4.
16. Reference 7, Page 4-5.
17. Andrade, C.M. Energy and Environmental Conservation at Valesul
Aluminum Smelter. In: ISSN.0378-9993 Industry and Environment,
Htun, N. United Nations Environment Programme. July/August/September
1983. Page 3.
3-33
-------
18. Reference 1.
19. Reference 2.
20. Memo from Noble, E.A., EPA-.ISB, to Durkee, K.R., EPA:ISB. September 4, 1985.
Report of June 1985 trip to Noranda Aluminum, New Madrid, Missouri. Page 2.
21. Reference 20, Page 3.
22. Reference 20, Page 2.
23. Memo from Noble, E.A., EPA:ISB, to Durkee, K.R., EPA:ISB. July 16, 1985.
Report of May 1985 trip to Alumax of South Carolina, Goose Creek,
South Carolina. Page 5.
24. Reference 20, Page 3.
25. Reference 3.
26. Reference 7, Page 4-14
27. Reference 7, Page 4-14.
28. Memo from Noble, E.A., EPA:ISB, to Durkee, K.R., EPA:ISB. November 26, 1985,
Report of June 1985 trip to Aluminum Company of America, Newburgh,
Indiana. Page 2.
29. Letter from Dickie, R.C., Alumax of South Carolina, to Farmer, J.R.,
EPA:ESED. August 27, 1985. Enclosure 1, Page 3.
30. Telecon. Boyt, J.S., Aluminum Company of America, with Noble, E.A.,
EPA:ISB. September 16, 1985. Information on tunnel kilns in use at
Alcoa's Warrick plant.
31. International Primary Aluminum Institute. Review of the Petroleum
Coke Situation and its Potential Impact on Sulfur Dioxide Emissions
from Primary Aluminum Plants. December 1978.
32. Tropea, L.C. (Reynolds) and Atkins, P.R. (Alcoa). The Proper
Perspective on Sulfur Dioxide Emissions from the Primary Aluminum
Industry. (Presented at Air Pollution Control Association Meeting,
Houston. June 25-30, 1978.) APCA Report #78-61.4. Pages 7 and 8.
33. Reference 23, Page 6.
34. Reference 31.
35. Reference 1.
36. Reference 2.
3-34
-------
37. Letter and attachments from Boyt, J.S., Aluminum Company of America,
to Noble, E.A., EPA:ISB. April 29, 1986. Comments on draft document.
38. Reference 7, Page 5-13.
39. Reference 7, Page 5-14.
40. Reference 23, Page 4.
41. Reference 7, Page 4-4.
42. Reference 3.
43. U.S. Environmental Protection Agency. Background Information for
Standards of Performance: Primary Aluminum Industry. Volume I: Proposed
Standards. EPA 450/2-74-020a. October 1974. Pages 23, 36.
44. Reference 7, Pages 5-21, 5-24, 9-17.
45. International Primary Aluminum Institute. Fluoride Emissions Control:
Updated Costs for New Aluminum Reduction Plants. February 1985.
Chapter 3, Table 2.
46. Reference 28, Page 4.
47. Memo from Maxwell, W.H., EPA:ISB, to Primary Aluminum Docket (A-86-07).
May 30, 1986. Green anode composition and calculated SOg emissions.
48. -Reference 31, Table A.
49. Reference 3.
50. Letter from Givens, H., Alcan, to Noble, E.A., EPA:ISB.
September 26, 1985. Attachment, Page 3, Items 2.A.11 and 2.A.15.
3-35
-------
4. EMISSION CONTROL TECHNOLOGY
The current new source performance standards (NSPS) limit all fluoride
emissions from aluminum reduction potlines, not just those emissions which
pass through the control device. Given the high efficiency of current primary
control devices, the fluoride emissions released to the atmosphere consist
mostly of emissions which bypass the primary control system. Thus, the pot
hood, or enclosure, must prevent the emission of fluorides evolved within the
pot. Also, fluoride evolution outside the pot must be kept to a minimum.
The first is accomplished by utilizing the types of pots which can be tightly
enclosed; by reducing the frequency, extent and duration of pot openings to
an absolute minimum; and by developing a maintenance program to ensure that
the integrity of the pot hoods does not deteriorate. The latter is achieved
through operating procedures and by good housekeeping.
4.1 PRIMARY FLUORIDE CONTROL SYSTEMS
The primary control system for a potroom consists of the pot hood (or
enclosure), necessary ducting, and a fluoride control device. For a bake
furnace it includes the movable header(s) connecting the furnace flue to the
exhaust duct, the exhaust duct, possibly an exhaust gas conditioning tower,
and a fluoride control device.
The most common fluoride control device currently in use for both these
applications is the dry scrubber. All plants subject to the NSPS utilize dry
scrubbers. Wet scrubbers are also used at some plants.
4.1.1 Capture/Suppression
The effectiveness of a hood (or enclosure) depends not only on how
much of the pot area it covers and how tightly it can be sealed, but also on
-------
how frequently it must be opened to perform process functions, the extent to
which it must be opened, and the time it must stay open. Pot hooding, as
discussed herein, includes those steps taken to minimize fluoride evolution
from the pots, as well as those to maximize fume capture. Hoods for side-worked
prebake (SWPB) pots and horizontal stud Soderberg (HSS) pots will not be
discussed here, since neither type of potline has been built in the U.S. since
proposal of the NSPS. They are, however, discussed in detail in the guidance
document.!
4.1.1.1 Center-Worked Prebake Hoods. Center-worked prebake (CWPB) pots
have a superstructure which supports the anode bus bars and the alumina
storage hopper. Hoods are formed using curved metal side shields which
extend from the outside edges of the pot sides to this superstructure
(Figure 4-1). At each end of a pot, the space between the pot and the hopper
is closed and fitted with a door. Usually, there is one side shield per
anode and the side shield may be notched to fit tightly around the anode
hanger. The shields and doors are removed and replaced manually. Together,
the superstructure, side shields, and end pieces form an enclosure.
The fumes evolving from the pots are captured by enclosing the whole
pot bath area; by sealing the pot enclosure to the maximum extent possible;
by maintaining an airflow through the pot high enough to prevent fumes from
escaping through apertures in the hood (such as between side shields) without
entraining excessive alumina; and by minimizing the frequency, number, and
duration of side shield and door removals. For CWPB pots, hood airflows have
been optimized at around 1.89 cubic meters per second (mVs) (4,000 actual cubic
feet per minute [acfm]). Some plants also increase airflow whenever the hood is
opened by moving a damper set in the pot hood exhaust duct. Table 4-1 lists
the airflows for individual pots at plants with NSPS potlines, for both
normal and high flow conditions.
4-2
-------
CO
STATIONARY
END SHIELD
REMOVABLE
END DOOR
ALUMINA HOPPER
FIGURE 4-1
TYPICAL CENTER-WORKED PREBAKE POT HOODING
REMOVABLE
SIDE SHIELD
TO PRIMARY
CONTROL SYSTEM
-------
TABLE 4-1
AIRFLOWS TO INDIVIDUAL POTS AT PLANTS WITH POTLINES SUBJECT TO THE NSPS2-5
Plant
codea
Pot
typeb
Pot airflows, nP/s (acfm)c
Normal High f1
A
B
C
D
E
CWPB
CWPB
CWPB
CWPB
VSS
1.89 (4,000)
1.89 (4,000)
1.84 (3,900)
1.77 (3,750)
0.26 (550)
1.89 (4,000)
1.89 (4,000)
1.89 (4,000)
2.64 (5,600)
0.26 (550)
a Plants are coded for simplicity. Information is non-confidential.
b CWPB = Center-worked prebake.
d m^/s = cubic meters per second
acfm = actual cubic feet per minute.
d Airflow controlled by damper. At some plants, airflow is
increased whenever hoods are opened.
4-4
-------
The evolution of fluoride from CWPB pots is generally believed to have
been reduced by using three to four crust breakers and point feeders located
down the center!ine of the pot to punch small holes in the crust and to drop
alumina and bath additive into the bath through these openings. Older designs
dumped enough alumina for several hours on the crust and then used a breaker
bar to open the crust along the full length of the pot. The crust breakers
operate more frequently than the breaker bars but open much smaller holes in
the crust, so the potential for fluoride evolution is reduced.
However, one source believes that just the opposite may be true; that is,
the use of point feeders over breaker bars may lead to increased fluoride
evolution.6 This is because more than one hole may now be open in the crust
continuously, leading to more air being drawn in under the crust. A constant
evolution of fluoride gases would then be permitted without the "crust scrubbing"
effect of alumina sitting on the bath. In addition, the addition of cold,
moist alumina directly onto the molten bath surface could lead to increased
fluoride evolution through the hydrolysis of the moisture and the particulate
fluoride.
The advent of computer control for many pot functions, while not a control
panacea, has helped reduce the need to open the hood to correct overfeeding
problems, to add bath additives, or to correct anode effects (see Section 3.3.3).
Computer-controlled point feeders add precisely metered amounts of activated
alumina directly into the bath, minimizing the potential for overfeeding and
reducing the frequency and severity of anode effects. The computer can also
correct most anode effects, by cycling the anodes up and down in the pot.
Thus, side shields or end doors must be removed only to correct the serious
anode effects, which occur relatively infrequently.
The degree to which hoods are opened and the time they remain open are,
to some extent, at the discretion of the plant management (or plant operators).
Side shields are removed primarily for anode replacement and to correct the
occasional serious anode effect. End doors are opened for inspections, to
4-5
-------
measure the depth of the aluminum produced, and for tapping. Typically, at
NSPS plants, only two to three side shields are removed to replace an anode.
Usually, these side shields are replaced before those for the next anode
change are removed.7 Also, at NSPS plants, the time the end doors remain open
is kept to a minimum and monitored frequently. During tapping, for example,
one plant allows no more than three end doors to be open at a time.8 While
tapping is proceeding at one pot, the door to the preceding pot is closed and
that to the following pot is opened. Also, the aspirator air used to draw
the molten aluminum up into the ladle is vented into the door opening during
tapping, thus minimizing fume escape through the opening.
4.1.1.2 Vertical Stud Soderberg Hoods. The hood of a vertical stud
Soderberg (VSS) pot does not cover the total bath area. Rather, it forms a
skirt around the anode, leaving the sides open for crust breaking and alumina
addition (Figure 4-2). The hood captures most of the fumes evolving from the
consumable anode, plus the fluoride emissions from that portion of the bath
which it covers. Since the VSS hood area is much smaller than that for a
CURB hood, the optimum airflow is also much lower. The one VSS plant subject
to the NSPS has selected an airflow of 0.26 m3/s (550 acfm) (Table 4-1).
The fumes from the anode are burned at the entrance to the exhaust duct to
prevent carbon buildup in the ducts. Emissions from the bath area which is
not covered by the hood are contained, except during pot working, by a crust
of cryolite and alumina.
4.1.2 Primary Fluoride Removal
Either dry or wet scrubbers may be used for potroom primary control
and for anode bake furnace control.9 However, all plants subject to the NSPS
have selected dry scrubbers for fluoride removal from both potlines and bake
furnaces. These dry scrubbers are fully integrated into the feed material
delivery process. They act as material handling equipment in the transfer of
feed alumina from storage silos to the potlines. Most, if not all, of the
alumina for the pots passes through the dry scrubbers at these plants.
4-6
-------
-J
ANODE PINS
.1 OBGANIC FUME!
TO PRIMARY
CONTROL EQUIPMENT
CARBON ANODL
GAS AND TAR
' BURNING
SKIRT
FIGURE 4-2 TYPICAL VERTICAL STUD SODERBERG POT HOODING
-------
Two basic types of dry scrubbers are in use, the injected alumina
and the fluidized bed. Cutaway views of these types are shown in Figures 4-3
and 4-4, respectively. Both introduce alumina feed material from the storage
silos into the gas stream from the pots or furnace(s). There the alumina
adsorbs total fluorides (TF) (i.e., gaseous and particulate fluorides) from
the gas stream. 'Then, the gases pass through a baghouse where the alumina is
removed and routed to the potline. Thus, much of the evolved fluoride is
returned to the pots. A flow diagram of the dry scrubbing process is shown
i n Figure 4-5.
The inlet air to dry scrubbers used on anode bake furnaces must be
cooled before entering the baghouse. This is accomplished with dilution
air, with a water spray in a conditioning tower, or by injecting air or water
directly into the fluid bed.10 A cutaway view of a fluidized bed dry scrubber
for a bake furnace, showing fume and water injection points, is shown in
Figure 4-6.
4.2 SECONDARY FLUORIDE CONTROLS
Secondary controls include both add-on control units and the actions
taken to suppress or eliminate the sources of emissions generated outside
the pots.
Add-on potroom controls are not used at any new or existing CWPB
plants. They are used at one plant with a new VSS potline and on some existing
VSS and HSS potlines. Wet scrubbers are used in this application to control
particulate and gaseous fluorides.
The wet scrubbers used for secondary fluoride control on potlines 1
and 2 at Plant E are configured as shown in Figure 4-7. Each potline has a
scrubber (or mist eliminator) with five sections. Each section has five
4-8
-------
Emissions to Atmosphere
Crude
Alumina
Tank
Blended
Alumina
Tank
Crude Alumina
Blended Alumina
—Alumina Injection into Gas Stream
Primary Potline Emissions
FIGURE 4-3
INJECTED ALUMINA DRY SCRUBBER
4-9
-------
Emissions to Atmosphere
Crude
Alumina
Tank
Crude
Alumina
I I
Baghouse
Fluidized Bed
I I
Fan
Blended
Alumina
Tank
I !
Blendea
Alumina
Primary Potline Emissions
FIGURE 4-4
FLUIDIZED BED DRY SCRUBBER
4-10
-------
Crude
Alumina
Tank
Primary
Emissions
T
Dry
Scrubber
Blended
Alumina
Tank
Secondary
(fugitive)
Emissions
Potroom
Primary Pot Emissions
Pot
Hood
Pot
FIGURE 4-5
FLOW DIAGRAM OF THE DRY SCRUBBING PROCESS
FOR A PRIMARY ALUMINUM PLANT
4-11
-------
i iuur\c t-o
FLUIDIZED BED DRY SCRUBBER
USED ON AN ANODE BAKE FURNACE EXHAUST
FILTER
BAGS
WASTE GAS
MANIFOLD
ALUMINA
FLUIDIZED
SED
WATER
SPRAY
WASTE GAS
MANIFOLD
WATER
SPRAY
FLUIDIZING
AIR FAN
4-12
-------
FIGURE 4-7
CROSS-SECTION OF WET SCRUBBER USED TO CONTROL SECONDARY EMISSIONS
FROM PLANT E
scrubber stack
\
-water sprays
scrubber Intake
••'filter media
exhaust fan
fumes
potroom roof
4-13
-------
stacks and six fume intake ducts, and each stack contains an exhaust fan.
Fumes from the potroom roof area are drawn in through the intake ducts,
sprayed with water, and captured on filter media. Fumes which escape the
filter are drawn into the scrubber stacks and expelled to the atmosphere.
The scrubber on line 3 operates in the same way, but is configured somewhat
differently. It has 2 rooms, each with 14 fans. Total airflow through the
line 1 and 2 scrubbers is 3,540 m3/s (7.5 million acfm) each, while the line
3 scrubber has an airflow of 3,445 m3/s (7.3 million acfm). All have negligible
pressure drops.H
To minimize the release of activated alumina to pot rooms when it is
dumped from overhead conveyors into the alumina storage hoppers atop CWPB
pots, most CWPB plants now use sandy alumina (a fairly coarse grade) to feed
the pots. Also, some conveyer hoppers are equipped with baghouses to minimize
particulate emissions during alumina transfer. One CWPB plant enclosed the
pot hoppers and delivers the alumina by airslide.^
Most plants with NSPS potlines require good housekeeping. Small particles
of alumina, solidified bath material, and crust tend to collect beside the
pots during anode changes, etc., and, if not removed, can be picked up as
dust and passed through the roof monitors. Also, a buildup of crust and bath
material on the outer rim of the pot can prevent the side shields from seating
properly, increasing fugitive emissions from the pots. One CWPB plant also
makes it a point to remove hot anode butts from the potroom within 30 minutes.13
They are carried to a storage area which is vented to the bake plant scrubber.
Plants with NSPS potlines and bake furnaces have instituted work practices
designed to prevent the escape of fumes from the pots and to suppress emissions
generated outside the pots. They also conduct periodic inspections to ensure
that these practices are being followed. Figures 4-8 and 4-9 are inspection
forms modelled after those used by some plants subject to the NSPS. Figure 4-8
is a form used to record the condition of a single pot, while Figure 4-9 is
handy for summarizing and tabulating the condition of many pots and crucibles.
4-14
-------
FIGURE 4-8
HOOD INSPECTION DATA SHEET14
Date
Potroom No.
Pot No.
1
3
5
7
9
11
13
15
17
19
21
23
25
DUCT SIDE
1 27
END PANEL
-HOOD COVERS-
END PANEL
1
28
2
4
6
8
10
12
14
16
18
20
22
24
26
Deficiency codes:
1. Bottom vent in or out
2. Damage to side of hood
3. Damage to top of hood
4. Hole burned in hood
other than at step
5. Hole burned in hood
step
6. Damaged end door
S Not sealed
0 Open not tending
OT Open tending
TENDING SIDE
4-15
-------
FIGURE 4-9
HOOD AND CRUCIBLE INSPECTION SUMMARY^
POT
NO.
1
2
3
4
5
etc
END
DOORS
OPEN
POOR
SIDECOVER
PLACEMENT
SIDECOVER
PARTIALLY
OPEN
HIGH
DRAFT
SMOKING
POT
BROKEN
DAMPER
ARM
LEAKING
TRANSITS
CUT-OUT
POTS NOT
BLANKED
OFF
CRUCIBLE CONDITION
CRUCE
NO.
1
2
3
4
5
etc
CRUCE
ACCEPTABLE
DAMAGED
FLEX
PIPE
FLEX
PIPE
DISCONNECTED
FLEX
PIPE
MISSING
4-16
-------
4.3 PARTICULATE AND SULFUR DIOXIDE CONTROLS
Particulate matter (PM) emissions from potrooms and anode bake furnaces
are controlled by the equipment and work practices used for TF control. No
additional equipment has been installed at any NSPS plant solely to control
PM from these sources. Other emission sources in primary aluminum plants
do have controls specifically for PM, but these will not be addressed here.
As noted in Chapter 3, sulfur dioxide (SOg) emissions are a function of
the sulfur content of the coke and pitch used in the reduction pot anodes.
Thus, SOg emissions can be controlled either by limiting the sulfur content of
the coke and pitch used in the anodes, or by using add-on controls such as a
wet scrubber.
Petroleum coke with a low sulfur content (e.g., less than 3 percent) can
be purchased at a premium price for use in the anodes. Coke of this type is
available under contract for periods up to 10 years.^
The dry scrubbers now employed for TF control at most primary aluminum
plants have no long-term effect on S02 emissions. The S02 in the gas stream
from a potroom is adsorbed onto the surface of the alumina cycled through the
dry scrubber, along with the fluorides. It is then returned to the reduction
pots with the feed alumina. There, the S02 is converted back to the vapor
phase and returned to the scrubber, where the cycle is repeated. This continues
until the S02 content of the gas stream exceeds the adsorption capability of
the alumina. At that point, S02 is emitted from the dry scrubber at the same
rate that it is evolved from the pot.17 If the dry scrubber treats an anode
furnace exhaust, the net effect is the same. The S02 emissions are merely
redistributed, with the S02 captured by the anode furnace scrubber being
emitted by the potroom scrubber.
4-17
-------
Wet scrubbers have been installed for S02 control at one primary aluminum
plant with VSS pots.18 These scrubbers are located downstream of the dry
scrubbers on the primary fluoride control system and use either sodium hydroxide
or sodium carbonate as the scrubbing medium. The wet scrubbers used to control
secondary TF also remove S02.
4.4 CONTROL SYSTEMS PERFORMANCE
Emissions test data have been received from plants known to have
potlines and/or anode bake furnaces subject to the NSPS. This data base
consists of one new CWPB plant, with two potlines and two anode bake furnaces;
two existing CWPB plants, each with an NSPS potline and furnace; one VSS
plant with a new potline; and two existing CWPB plants with anode bake furnaces
subject to the NSPS. It should be noted that the emissions data provided by
plant E are unusual in two respects. The data are for three lines, not one,
and cannot reasonably be separated. All three lines have comparable primary .
and secondary control systems and all are required to meet the same emission
limits, even though only line 3 is subject to the NSPS. Also, the plant
normally performs only one test run per month on each line. Therefore, each
monthly test is the average of three runs, as usual, but each run is on a
different line.
4.4.1 Total Fluorides
All four plants with potrooms subject to the NSPS have demonstrated
the capability for meeting the NSPS for TF, as have the five plants with anode
bake furnaces (Table 4-2). The three plants with CWPB pots have average
TF emissions between 0.43 and 0.64 kg/Mg (0.86 and 1.27 Ib/ton) of aluminum
produced.
4-18
-------
TABLE 4-2
TOTAL FLUORIDE EMISSIONS FROM POTLINES AND ANODE BAKE FURNACES SUBJECT TO THE NSPS19
PI ant
code*
A
B
C
D
E
F
G
H
J
K
Plant Emission
i
typeb source
CWPB Potroom
CWPB
CWPB
CWPB
VSS
CWPB Furnace
CWPB
CWPB
CWPB
CWPB
Number
monthly
tests
51
48
34
16
22
20
16
10
" 10
7
Measured TF
emissions (lb/TAP)c
Range Average
0.51-1.32 0.90
0.36-1.46 0.86
0.67-3.48 1.27
0.71-1.49 1.02
0.88-3.11 1.49
0.003-0.113 0.017
0.001-0.043 0.008
0.002-0.017 0.007
0.003-0.056 0.014
0.003-0.038 0.010
NSPS
limit
(Ib/TAP)
1.9
n
H
M
2.0
0.1
II
II
II
II
Number
exceed-
ances
0
0
3d
0
36
If
0
0
0
0
a Plants coded for simplicity.
b CWPB = center-worked prebake
VSS = vertical stud Soderberg
c Measured emissions of total fluorides (TF) includes both primary and
secondary emissions. Ib/TAP = pounds per ton aluminum produced
(1 Ib/TAP = 0.5 kilogram per megagram).
d Two failures occurred in same month, one on retest.
6 Specific reason for failures not reported. Plant conducts one test run/line
each month, so each reported test is the average for all three lines.
f Failure occurred in the first month test results reported.
4- 19
-------
The plant with VSS pots had average emissions which were somewhat higher, at
0.75 kg/Mg (1.49 1b/ton). The potlines at two plants have, however, exceeded
the allowable emissions limit on occasion, as has the bake furnace at one
plant.
It should be noted that the numbers presented in Table 4-2 include
both primary and secondary TF emissions, as required by the NSPS. As mentioned
earlier, TF emissions generated in the pots and captured by the pot hoods are
called primary TF, while those which escape from (or are generated outside)
the pot are classified as secondary TF. The split between primary and
secondary TF emissions is shown in Table 4-3. As can be seen by comparing
the data in this table, secondary TF emissions account for between 90 and
93 percent of total TF emissions to the atmosphere at both CWPB and VSS
plants. This is significant, because it illustrates the importance of proper
hood design and maintenance, and good work practices. In the absence of
secondary TF controls, relatively small increases in emissions escaping the
hoods can have dramatic impacts on emissions to the atmosphere. This effect
is evident in Table 4-4, which tabulates the effect of changes in primary TF
capture and control efficiencies on overall TF emissions for two TF evolution
rates. It indicates, for example, that a drop of only 1 percent in capture
efficiency can increase TF emissions to the atmosphere by 30 to 90 percent,
depending on initial control efficiency. (Secondary emissions can also
increase as a result of process changes that result in higher evolution rates,
even through the hooding efficiencies remain the same.)
4-20
-------
TABLE 4-3
TOTAL FLUORIDE EMISSIONS BY POTROOM GROUP AND TYPE20
Measured TF emissions (Ib TF/TAP)&
Plant Potroom PrimarySecondaryTotal Secondary
code9 group Range Avg. Range Avg. Range Avg. %
A 1 0.02-0.26 0.09 0.44-1.15 0.77 0.51-1.28 0.88
2 0.02-0.08 0.05 0.57-1.29 0.87 0.59-1.32 0.91
1 & 2 0.07 0.82 0.90 91
B 1 0.02-0.16 0.07 0.41-1.12 0.73 0.36-1.17 0.79
2 0.04-.22 0.10 0.37-1.31 0.85 0.48-1.46 0.94
1 & 2 0.08 0.79 0.86 92
C 1 & 2 N/AC 0.50-3.41 1.18 0.67-3.48 1.27 93
D 1X2 0.04-0.13 0.07 0.35-1.40 0.92 0.71-1.49 1.02 90
E 1-3
(3 Lines) 0.01-0.64 0.13 0.86-3.01 1.36 0.88-3.11 1.49 91
a = Plant code same as Table 4-2. Coding is for simplicity.
b = Ib TF/TAP = pounds total fluoride per ton aluminum produced.
(1 Ib/TAP =0.5 kilogram per megagram)
c = Not available.
4-21
-------
TABLE 4-4
IMPACTS OF CHANGES IN PRIMARY TF CAPTURE AND REMOVAL
EFFICIENCIES ON OVERALL TF EMISSIONS21
TF
evolution Hood
from CWPB capture
potsa efficiency
(lb TF/TAP) (%)
66 99
98
97
96
55 99
98
97
96
Primary
TF
removal
efficiency
(%)
99.9
99.8
99.6
99.9
99.8
99.6
99.9
99.8
99.6
99.9
99.8
99.6
99.9
99.8
99.6
99.9
99.8
99.6
99.9
99.8
99.6
99.9
99.8
99.6
Overall
TF
control
efficiency
(%)
98.9
98.8
98.6
97.9
97.8
97.6
96.9
96.8
96.6
95.9
95.8
95.6
98.9
98.8
98.6
97.9
97.8
97.6
96.9
96.8
96.6
95.9
95.8
95.6
TF emissions to the
atmosphere (lb TF/TAP)
Primary
0.065
0.131
0.261
0.065
0.129
0.259
0.066
0.128
0.256
0.063
0.127
0.253
0.055
0.110
0.220
0.054
0.108
0.216
0.053
0.107
0.213
0.053
0.106
0.211
Secondary
0.660
ii
n
1.320
n
n
1.980
II
II
2.640
II
II
0.550
II
II
1.100
II
II
1.650
II
II
2.200
II
II
Total
0.725
0.791
0.921
1.385
1.449
1.579
2.046
2.108
2.236
2.703
2.767
2.893
0.605
0.660
0.770
1.154
1.208
1.316
1.703
1.757
1.860
2.273
2.306
2.411
lb TF/TAP = pounds total fluoride per ton aluminum produced.
(1 Ib/TAP = 0.5 kilogram per megagram)
4-22
-------
The capture efficiencies of the hooding systems and the TF removal
capabilities of the primary control equipment cannot be determined for the
NSPS plants because no data are available on uncontrolled fluoride evolution
from the pots at those plants, nor on the magnitude of the emissions routed
to their dry scrubbers. The average capture efficiency of a hood is a composite
number which takes into account the effects of opening the hood for pot
working, inspection, and tapping.
Primary TF emissions are controlled by dry scrubbers at all primary
aluminum plants subject to the NSPS. Secondary TF emissions are not controlled
at any CWPB plants, but are removed by wet scrubbers at the one VSS plant
subject to the NSPS. Information provided by this plant on secondary TF
control efficiencies (monthly averages) are summarized in Table 4-5. Their
secondary wet scrubbers, which use a calcium additive and have negligible
pressure drops, usually recovered over 68 percent of the secondary TF.
4.4.2 Particulate Matter
The control of PM from potlines and anode bake furnaces occurs as a
side benefit of TF control. Data have been received on PM emissions from TF
control devices on four potlines and three anode bake furnaces (Table 4-6).
No data are available on uncontrolled primary PM emissions, so the effectiveness
of dry scrubbers in controlling primary PM cannot be determined. One plant,
however, provided information on the PM control efficiency of the wet scrubbers
installed for secondary TF control. A summary of this information is presented
in Table 4-5. The average monthly efficiency of these scrubbers in reducing
PM was over 56 percent during the reporting period.
4-23
-------
TABLE 4-5
EFFECTIVENESS OF SECONDARY WET SCRUBBERS AT PLANT
Pollutant Reduction (%)a
Pollutant RangeAverage
Total Fluoride 56.26 - 80.94 68.73
Sulfur Dioxide 40.33-82.99 62.05&
Particulate Matter 35.05 - 72.72 56.93
a Monthly averages
b Inlet concentration averaged 1.0 ppmv (Reference 23)
4-24
-------
TABLE 4-6
PARTICULATE EMISSIONS FROM PRIMARY ALUMINUM PLANTS
USING DRY SCRUBBERS TO CONTROL FLUORIDE EMISSIONS24
Plant
code3
Emission
source'3
Test
year
No.
monthly
tests
Emissions (Ib PM/TAP)c
Primary
Min.
Max.
Avg
Secondary
Min. Max.
Avg
Total
Min. Max. Avg
A CWPB
Potline,
Subgroup
1
2
B CWPB
Potline,
Subgroup
1
2
D CWPB
Pot! i ne
E VSS
Potlines
1&2
F Anode
Bake
Furnace
J Anode
Bake
Furnace
K Anode
Bake
Furnace
a Plant code
b rwPR = Ont
1984
1984
1984
1984
1984
1985
all
1984
1985
all
1985
1984
1985
all
1984
1985
all
same as
pr-worki
1
1
1
1
1
1
2
12
10
22
1
5
6
11
11
4
15
Tabl
ed or
--
— _
0,
0
0
0
.018 -- -- 0.576
.057 -- — 0.542
.057 -- — 0.66
.091 -- -- 0.84
0.531
1.870
-- 0.594
— 0.599
-- 0.72
-- 0.93
1.200
0.05
0.04
--
__
0.102
0.176
--
0.023
0.210
- —
e 4-2.
•ebake
0.39
0.29
--
__
0.223
2.034
--
1.061
0.728
— —
Codi
0
0
0
0
0
0
0
0
0
0
ng
.14 4.85 11.84
.14 2.94 6.30
.14
.003 --
.169 —
.603 --
.405 —
.458 --
.442 --
.454 --
is for simplicity.
6.43 5.01
5.03 3.12
5.85 --
__
-- 0.102
— 0.176
— —
-- 0.023
-- 0.210
_ _ — —
11.94
6.36
—
--
0.223
2.034
—
1.061
0.728
— ™
6.68
5.17
6.00
0.006
0.169
0.603
0.405
0.458
0.492
0.454
VSS = Vertical stud Soderberg
Ib PM/TAP = pounds particulate matter per ton aluminum produced.
(1 Ib/TAP = 0.5 kilogram per megagram)
4-25
-------
4.4.3 Sulfur Dioxide
The one plant with SOg controls (Plant E) reports SCfc emissions
averaging 3.6 kg/Mg (7.2 Ib/ton) (Table 4-7).25 This is equivalent to
the use of coke with a sulfur content of 0.36 percent. In addition to
listing total S02 emissions from the three VSS potlines at Plant E,
Table 4-7 shows how this total is split between the primary and secondary
control systems. It also shows the amount of variation experienced
during the reporting period. This plant, which uses VSS pot technology,
added SO2 controls to meet a prevention of significant deterioration
(PSD) limit of 6.99 kg S0£/Mg (13.97 1 b/ton). They control both primary
and secondary S02. Lines 1 and 2 have a single sodium (wet) scrubber for
primary S02 control, located downstream of the two dry scrubbers. The
primary SOg control for Line 3 is provided by a sodium scrubber located
downstream of the dry scrubber. Secondary SOg emissions from all three
potlines are passed through calcium (wet) scrubbers which are provided
primarily for TF control. Information provided by the plant and summarized
in Table 4-5 indicates that these scrubbers generally remove about 62
percent of the secondary S02. However, the data provided by the plant
are not adequate to determine S0£ removal efficiencies of the primary S02
scrubbers, since the S02 content of the uncontrolled potroom primary exhaust
gases was not measured and the sulfur content of the coke used at the
plant during the test reporting period was not provided.
Data were also obtained on S02 emissions from two anode bake plants.
These data are presented in Table 4-8. Average emissions range from 0.59 to
1.55 kg S02/Mg (1.18 to 3.10 1 b/ton). Sulfur contents of the anode
constituents ranged from 0.45 percent in the pitch to 2.0 percent for fresh
coke at these two plants.
4--26
-------
TABLE 4-7
SULFUR DIOXIDE EMISSIONS FROM A PRIMARY ALUMINUM PLANT
USING WET SCRUBBERS FOR PRIMARY AND SECONDARY
SULFUR DIOXIDE CONTROL26
Plant
code9
Emission
source'3
Test
year
No.
monthly
tests
Emissions (Ib S02/TAP)C
Primary
Min.
Max.
Avg
Secondary
Min. Max.
Avg
Total
Min. Max. Avg
E
VSS 1984
Potlines 1985
1-3 all
12
10
22
1.09 11.71 3.49 2.12 7.43 3.52 3.21 14.87 7.02
1.55 7.86 3.33 2.14 6.22 4.01 3.96 12.75 7.34
1.09 11.71 3.42 2.12 7.43 3.73 3.21 14.87 7.17
a Plant code same as Table 4-2. Coding is for simplicity.
b VSS = vertical stud Soderberg. Data are for 3 potlines. Potlines 1
and 2 use same S02 scrubber.
c Ib SOo/TAP = pounds S02 per ton aluminum produced (1 Ib/ton = 0.5
kilogram per megagram).
4-27
-------
TABLE 4-8
SULFUR DIOXIDE EMISSIONS FROM ANODE BAKE PLANTS27
Plant Emission
codea source
No.
Test monthly
year tests
jc Anode 1984 5
Bake 1985 6
Plant all 11
K.d Anode 1984 10
Bake 1985 4
Plant all 14
Max.
5.15
4.39
5.15
1.63
1.74
1.74
a Plant code same as Table 4-2. Coding
b lh SOo/TAE = n
nunds SOo oer
ton al umi r
Emissions (Ib
Total
Min.
2.31
1.38
1.38
0.39
1.23
0.39
is for simplici
mm eaui valent (
S02/TAE)b
Avq.
3.62
2.67
3.10
1.07
1.46
1.18
ty.
1 Ib/ton = 0.5
kilogram per megagram).
Sulfur contents ranged from 0.45 percent for pitch to 1.95 percent
for packing coke (Reference 28).
Sulfur contents ranged from 0.60 percent for pitch to 2.0 percent
for fresh coke (Reference 28).
4-28
-------
4.5 REFERENCES FOR CHAPTER 4
1. U.S. Environmental Protection Agency. Primary Aluminum: Guidelines for
Control of Fluoride Emissions from Existing Aluminum Plants.
EPA-450/2-78-049b. December 1979.
2. Letter and attachments from Dickie, R.C., Alumax of South Carolina, to
Farmer, J.R., EPA:ESED. August 27, 1985. Response to Section 114
information request.
3. Letter and attachments from Givens, H.L., Alcan, to Noble, E.A., EPA:ISB.
September 26, 1985. Response to Section 114 information request.
4. Letter and attachments from Hurt, R.E., Noranda Aluminum, to Farmer, J.R.,
EPA-.ESED. September 25, 1985. Response to Section 114 information request
5. Letter and attachments from Casswell, S.J., Commonwealth Aluminum, to
Farmer, J.R., EPA:ESED. September 6, 1985. Response to Section 114
information request.
6. Letter and attachment from Tropea, L.C., Jr., Reynolds Aluminum, to
Noble, E.A., EPA:ISB. May 9, 1986. Comments on draft chapters.
7. Reference 4.
8. Reference 1.
9. Reference 1, Page 6-29.
10. Memo from Noble, E.A., EPA:ISB, to Durkee, K.R., EPA-.ISB. September 4,
1985. Report of June 1985 trip to Noranda Aluminum, New Madrid, Mo.
11. Reference 5.
12. Reference 4. Attachment 1.
13. Reference 10, Page 3.
14. Reference 2.
15. Letter and attachments from Givins, H.L., Alcan, to Noble, E.A.,
EPA:ISB. Received June 28, 1985. Emissions data.
16. Letter and attachments from Dickie, R.C., Alumax of South Carolina,
to Noble, E.A., EPA:ISB. April 16, 1986. Comments on draft chapters.
4-29
-------
17. International Primary Aluminum Institute. Review of the Petroleum Coke
Situation and Its Potential Impact on Sulfur Dioxide Emissions from
Primary Aluminum PI ants. December 1978.
18. Reference 5.
19. Memo and attachments from Maxwell, W.H., EPA:ISB, to Primary Aluminum
Docket (A-86-07). June 18, 1986. Fluoride emissions from NSPS primary
aluminum plants.
20. Reference 19.
21. Memo from Maxwell, W.H., to Primary Aluminum Docket (A-86-07). June 17,
1986. Impact of capture and removal efficiencies.
22. Memo and attachments from Maxwell, W.H., EPA-.ISB, to Primary Aluminum
Docket (A-86-07). June 17, 1986. Control device efficiencies.
23. Memo and attachments from Maxwell, W.H., EPA.-ISB, to Primary Aluminum
Docket (A-86-07). June 17, 1986. S02 emission concentration from
Commonwealth Aluminum.
24. Memo and attachments from Maxwell, W.H., EPA:ISB, to Primary Aluminum
Docket (A-86-07). June 17, 1986. Particulate and S02 emissions from
NSPS primary aluminum plants.
25. Letter from Casswell, S.J., Commonwealth Aluminum, to Noble, E.,
EPArlSB. November 21, 1985. Emissions data.
26. Reference 24.
27. Reference 24.
28. Letter and attachments from Boyt, J.S., Aluminum Company of
America, to Farmer, J.R., EPA:ESED. October 7, 1985. Response to
Section 114 information request.
4-30
-------
5. COMPLIANCE STATUS OF PRIMARY ALUMINUM PLANTS
5.1 AFFECTED FACILITIES
Six plants have been built or have added aluminum reduction potlines
or anode bake furnaces since the primary aluminum new source performance
standards (NSPS) were proposed in 1974. Information on their locations and
startup dates is presented in Table 5-1. The one new plant installed two
center-work prebake (CWPB) potlines and two anode bake furnaces. Two plants
replaced only their anode bake furnaces. The three existing plants which
installed potlines utilized the same types of pots as were already in use on
their existing potlines. Two of these were CWPB facilities and one was a
vertical stud Soderburg (VSS) facility. Each of the CWPB plants also added
an anode bake furnace at the same time.
5.2 EMISSIONS DATA
The NSPS limit emissions of total fluoride (TF) from potlines and
bake furnaces. In addition, particulate and sulfur dioxide (S02) emissions
are regulated by some states. The following sections present the available NSPS
compliance emissions data, organized by pollutant.
5.2.1 Total Fluoride
The NSPS require monthly testing of primary TF emissions from anode
bake furnaces and both primary and secondary TF emissions from aluminum
reduction potrooms. However, the requirement for monthly testing of the
primary control systems has been waived at three NSPS plants in favor of
yearly tests and at one plant in favor of twice yearly testsJ~3 In those
cases, total TF emissions for potroom groups are calculated using the results
of the last available test. Available data on TF emissions from potroom
groups and anode bake furnaces are summarized on Table 5-2.
-------
TABLE 5-1
LIST OF PRIMARY ALUMINUM
REDUCTION PLANTS SUBJECT
TO THE NSPS^-8
Plant name & location
NSPS Facility
Plant Line
i / /•* \
Bake furnace
Start
up
date(s)
Pot
type3
NSPS
potline
capacity,
tpy A!6
Total per plant,
new and existing
Potlines
Bake furnace
i
no
Mount Holly Plant
Goose Creek,
South Carolina
Alcoa
Warrick Plant
Newburgh, Indiana
Alcan Aluminum
Sebree Plant
Henderson, Kentucky
Noranda Aluminum Co.
New Madrid, Missouri
Commonwealth Aluminum
Goldendale Plant
Goldendale, Washington
Alcoa
Rockdale Works
Rockdale, Texas
(2)
6/80
CWPB
100,000
X 2
1/84 CWPB
6/79 CWPB 60,000
7/83 CWPB 85,000
3/82
VSS 65,000
12/77, CWPB
12/81, &
8/82
3
3
b
c
CWPB = Center-worked prebake
VSS = Vertical stud Soderberg
TPY Al = tons per year aluminum (1 ton = 0.9 megagram)
Plant also has 12 tunnel kilns, 4 of which would be used when operating at full capacity,
-------
TABLE 5-2
FLUORIDE EMISSIONS FROM PRIMARY ALUMINUM PLANTS
SUBJECT TO THE NSPS9
Plant
code3
Plant
type*3
Emission
source
Test
peri od
No.
tests
Emissions of total
fluorides (Ib TF/TAP)C
Min.
Max.
Avg
No.
exceed-
ances
A
B
C
D
E
F
G
H
J
K
CWPB
CWPB
CWPB
CWPB
VSS
CWPB
CWPB
CWPB
CWPB
CWPB
Potline 1
Potroom
Group 1
Group 2
Potline 2
Potroom
Group 1
Group 2
Potline 3
Potroom
Grpsl&2
Potline 3
Potroom
Grpsl&2
Potl ines
1,2,43
Anode
Bake Pit
Anode
Bake Pit
Anode
Bake Pit
Anode
Bake Pit
Anode
Bake Pit
1/81-3/85
1/81-3/85
3/81-2/85
2/81-2/85
l/82-12/84d
3/84-6/856
1/84-10/85
8/81-5/85
3/80-3/85
3/84-1/85
9/84-6/85
9/84-4/85
51
51
48
47
34
16
22
20
16
10
10
7
0
0
0
0
0
0
0
0
0
0
0
0
.51
.59
.36
.48
.67
.71
.88
.003
.001
.002
.003
.003
1
1
1
1
3
1
3
0
0
0
0
0
.28
.32
.17
.46
.48
.49
.11
.113
.043
.017
.056
.038
0
0
0
0
1
1
1
0
0
0
0
0
.88
.91
.79
.94
.27
.02
.49
.017
.008
.007
.014
.010
0
0
0
0
3
0
3
1
0
0
0
0
a Plants are coded for simplicity
b CWPB = Center-worked prebake
VSS = Vertical stud Soderberg
c Ib TF/TAP = pounds total fluoride per ton aluminum produced (1 Ib/ton =
0.5 kilogram per megagram)
d Secondary TF data available from 9/79 to 5/85
e Secondary TF data available from 1/84 to 6/85
5-3
-------
Table 5-2 is organized by plant and emission source (potroom group or
furnace) and shows the test period, the total number of tests in the data
base, the range of emissions reported in the test period, and the. mean (or
average) emission rate. It also indicates the number of times the NSPS has
been exceeded during the reporting period. As can be seen, all the affected
facilities (potroom groups and bake furnaces) have the capability to meet the
NSPS, since average TF emissions are considerably below the NSPS. Average TF
emissions for potroom groups range from 0.40 to 0.64 kilograms per megagram
(kg/Mg) (0.79 to 1.27 Ib/ton) aluminum produced for CWPB plants and are 0.75
kg/Mg (1.49 Ib/ton) for the VSS plant, compared to their respective NSPS
limits of 0.95 and 1.0 kg/Mg (1.9 and 2.0 Ib/ton).' Average anode bake plant
emissions range from 0.004 to 0.009 kg/Mg (0.007 to 0.017 Ib/ton) equivalent
(NSPS limit is 0.05 kg/Mg [0.1 Ib/ton]).
As indicated in Table 5-2, two of the five potlines subject to the NSPS
reported TF emissions which exceeded the NSPS limits during the periods for
which data are available. One of these used CWPB pots and the other VSS pots.
The former reported 3 exceedances over a 3-year period (January 1982 to
December 1984) and the latter recorded 3 in 20 months (January 1984 to
August 1985). The dates of the exceedances and the reasons cited for their
occurrence are provided on Table 5-3.
Tables 5-4 and 5-5 expand on the information provided in Table 5-2 by
giving a calender year breakdown of that data. Most plants seem to show
relatively little variation in emissions from year to year. Table 5-4
also indicates the relative contributions of primary and secondary sources to
total potline emissions. The data show that 83 to 96 percent of total TF
emissions are emitted through the roof monitors as secondary TF.
5-4
-------
TABLE 5-3
RECORD OF REPORTED NSPS EXCEEDANCES WITH FAILURE RATIONALE10'12
Plant
code3
Emission
source^3
Date of
exceedance
Reported
emission0
(Ib TF/TAP)
Emission
1 imit
(Ib TF/TAP)
Comments
CWPB 7/82
Potline 3
2.1
1.9/2.5 Exceeded original NSPS
1imit but not amended
not-to-be-exceeded (NTBE)
limit. Failure attributed
to work practices and
potroom conditions.
E VSS
Potlines
1,2,43
F CWPB
Anode
Bake Pit
7/82
10/83
2/84
5/84
7/84
8/81
2.07
3.38
2.45
2.15
3.11
0.113
1.9/2.5
1.9/2.5
2.0/2.6
2.0/2.6
2.0/2.6
0.1
Retest. Failed for same
reason.
Exceeded NTBE limit due to
damper control malfunction.
Retested at 1.20 after
damper travel corrected.
Exceeded original NSPS
but not NTBE limit. Reason
for exceedance not
documented.
Same as above.
Exceeded NTBE limit. Reason
not cited.
Exceeded NSPS in first
month data reported. Reason
not cited.
a Plant code same as Table 5-2. Coding is for simplicity.
b CWPB = Center-worked prebake
VSS = Vertical stud Soderberg
c Ib TF/TAP = pounds total fluoride per ton aluminum produced
(1 Ib/ton = 0.5 kilogram per megagram)
5-5
-------
TABLE 5-4
EMISSIONS FROM POTLINES AT PRIMARY ALUMINUM PLANTS
WITH FLUORIDE CONTROLS13
Plant Plant Potroom Test No. Monthly
codea type13 group year tests c
(Pri /Sec/Total)
A CWPB 1 1981 12/12/12
1982 5/12/12
1983 1/12/12
1984 1/12/12
1985 O/ 3/ 3
all 19/51/51
2 1981 12/12/12
1982 5/12/12
1983 1/12/12
1984 1/12/12
1985 O/ 3/ 3
all 19/51/51
8 CWPB 1 1981 10/10/10
1982 5/12/12
1983 1/12/12
1984 1/12/12
1985 O/ 2/ 2
all 17/48/48
2 1981 10/10/9
1982 5/12/12
1983 1/12/12
1984 1/12/12
1985 O/ 2/ 2
all 17/48/47
C CWPB 1&2 1979 --/ 3/--
1980 --/III—
1981 — III/—
1982 --/13/13
1983 — /12/12
1984 --/ 9/ 9
1985 — / 5/--
all --/64/34
Average TF
(lb/TAP)d
Primary
0.08
0.10
0.14
0.06
--
0.09
0.05
0.03
0.05
0.06
--
0.05
0.07
0.06
0.05
0.05
--
0.07
0.13
0.08
0.04
0.08
--
0.10
__
--
--
--
--
--
--
--
Sec.
0.82
0.71
0.79
0.77
0.66
0.77
0.89
0.85
0.86
0.87
0.87
0.87
0.69
0.70
0.79
0.77
0.57
0.73
0.99
0.85
0.80
0.85
0.53
0.85
1.14
1.15
1.11
1.12
1.34
1.15
1.25
1.18
Total
Min.
0.66
0.51
0.62
0.60
0.66
0.51
0.69
0.59
0.73
0.80
0.76
0.59
0.56
0.36
0.48
0.60
0.46
0.36
0.79
0.54
0.48
0.64
0.50
0.48
__
--
--
0.68
0.90
0.67
--
0.67
TF (Ib/TAP)
Max.
1.28
1.11
1.29
1.05
0.83
1.28
1.32
1.15
1.08
1.06
1.11
1.32
1.0
0.97
1.17
1.03
0.77
1.17
1.46
1.30
1.22
1.35
0.72
1.46
—
--
--
2.13
3.48
1.66
--
3.48
Avg.
0.90
0.83
0.93
0.89
0.72
0.88
0.94
0.89
0.90
0.93
0.93
0.91
0.76
0.75
0.84
0.82
0.62
0.79
1.01
0.95
0.89
0.90
0.61
0.94
—
--
--
1.17
1.40
1.24
--
1.27
5-6
-------
TABLE 5-4 (concluded)
EMISSIONS FROM POTLINES AT*PRIMARY ALUMINUM PLANTS
WITH FLUORIDE CONTROLS13
PI ant Plant Potroom
code3 typeb group
D CWPB 1&2
E VSS 1&2
(lines)
( 1-3 )
Test
year
1984
1985
all
1984
1985
all
No. monthly
tests c
(Pri /Sec/Total)
10/11/10
O/ 6/6
9/17/16
12/14/12
10/10/10
22/24/22
Average TF
(lb/TAP)d
Primary
0.07
--
0.07
0.19
0.05
0.13
Sec.
0.90
0.95
0.92
1.60
1.04
1.36
Total 11- Ub/lAP)
Min.
0.71
0.83
0.71
1.27
0.88
0.88
Max.
1.49
1.37
1.49
3.11
1.37
3.11
Avg.
1.02
1.03
1.02
1.83
1.09
1.49
a Plant code same as Table 5-2. Coding is for simplicity.
b CWPB = Center-worked prebake
VSS = Vertical stud Soderberg
c Pri = primary
Sec = secondary
d All plants utilize dry scrubbers to control primary total fluorides (TF). Only
one plant (Plant E) controls secondary emissions (using wet scrubbers).
Ib/TAP = pounds per ton aluminum produced (1 Ib/ton = 0.5 kilogram per megagram)
5-7
-------
TABLE 5-5
EMISSIONS FROM ANODE BAKE FURNACES AT PRIMARY ALUMINUM PLANTS
WITH FLUORIDE CONTROLS*4
Plant
code3
Plant
typeb
•
Year
No.
monthly
tests
Emissions (Ib TF/TAP)c
Minimum
Maximum
Average
F
G
H
J
K
CWPB 1981
1982
1983
1984
1985
all
CWPB 1980
1981
1982
1983
1984
1985
all
CWPB 1983
1984
1985
all
CWPB 1984
1985
all
CWPB 1984
1985
all
5
12
1
1
1
20
9
3
1
1
1
1
16
1
9
1
10
4
6
10
3
4
7
0.010
0.003
--
--
-_
0.003
0.003
0.002
-_
—
—
—
0.001
_ «
0.002
—
0.002
0.005
0.003
0.003
0.003
0.005
0.003
0.113
0.040
__
-'-
--
0.113
0.043
0.016
--
__
--
__
0.043
_ —
0.017
__
0.017
0.012
0.055
0.056
0.004
0.038
0.038
0.043
0.009
0.011
0.006
0.003
0.017
0.011
0.007
0.003
0.001
0.004
0.001
0.008
0.011
0.006
0.006
0.007
0.008
0.019
0.014
0.003
0.016
0.010
a
b
c
Plant code same as Table 5-2. Coding is for simplicity.
CWPB = center-worked prebake.
All NSPS plants utilize dry scrubbers to control total fluorides (TF).
Ib/TF/TAP = pounds total fluorides per ton aluminum produced (1 Ib/ton
0.5 kilogram per megagram)
6-8
-------
5.2.2 Visible Emissions
Visible emissions data were reported on one anode bake plant subject to
the NSPS (Table 5-6). 15 The opacity readings were generally zero, but
ranged up to 32 percent (6 minute average) on individual stacks.j/ Readings on
individual stacks met or exceeded 20 percent on 6 test runs. The NSPS limit
is 20 percent. No reason was given for the high readings.
The dry scrubber at this plant has 4 sections, each with 3 stacks
(12 stacks total).
5-9
-------
TABLE 5-6
VISIBLE EMISSIONS FROM ANODE BAKE FURNACE AT>LANT Jl6
Reactor Stack
Date
08/01/84
09/09/84
08/24/84
08/27/84
09/26/84
09/28/84
10/2/84
10/10/84
10/11/84
10/12/84
11/6/84
11/7/84
11/9/84
12/3/84
12/6/84
12/7/84
01/08/85
01/09/85
01/10/85
02/20/85
02/21/85
02/22/85
02/26/85
02/27/85
02/28/85
04/17/85
04/18/85
04/19/85
05/21/85
05/22/85
05/23/85
07/01/85
07/02/85
07/03/85
164
N
0.4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
14.8
0
0
0
0
C-l
C
13.8
0
0
0
0.4
0
0
0
0
0
0
0
0
0
0
0.4
0
2.2
0
0
0
0
0
0
0
0
0
1
0
0
s
0
0
1
0
0.4
0
0
0
0
0
0
0
0
0
0
0.2
0
2.2
0
0
0
0
0.4
0
3.4
1.6
5.6
0
0
0
164
N
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0.2
0
0
0
C-2
C
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2.8
32
30
32
S
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.6
25
30
26
164 C-3
N
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
C
0
0
0
0
0
0
0
0
0
0
0
0
0
1.2
0
0
0
0.
s
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6 4
164 C-4
N
0
0
0
0
0
0
0
0
0
0
2.2
1.2
0
0
0
0
0
. 0
0
0
0
0
0
0
0
C
0
0
0
0
0
0
0
0
0
0
2
1.67
0
0
0
0
0
0
0
0
0
0
0
0
0
s
0
0
0
0
0
0
0
0
0
2.2
1.2
0
0
0
0
0.4
0
0
0
0
0
0
0
0
5-10
-------
5.3 REFERENCES FOR CHAPTER 5
1. U.S. Environmental Protection Agency. Code of Federal Regulations.
Title 40, Chapter I, Subchapter C, Part 60. Washington, D.C.
Office of the Federal Register. June 30, 1980. Pages 44202-44217.
2. Letter and attchments from Hurt, R.E., Noranda Aluminum, to Noble, E.,
EPA:ISB. December 10, 1985. Information related to alternate test
frequency.
3. Letter and attachments from Boyt, J.S., Aluminum Company of America, to
Farmer, J.R., EPArESED. October 7, 1985. Response to Section 114
information request.
4. Letter and attachments from Dickie, R.C., Al umax of South Carolina, to
Farmer, J.R., EPA:ESED. August 27, 1985. Response to Section 114
information request.
5. Reference 3.
6. Letter and attachments fromGivens, H.L., Alcan, to Noble, E.A., EPArlSB.
September 26, 1985. Response to Section 114 information request.
7. Letter and attachments from Hurt, R.E., Noranda Aluminum, to
Farmer, J.R., EPArESED. September 25, 1985. Response to Section 114
information request.
8. Letter and attachments from Casswell, S.J., Commonwealth Aluminum, to
Farmer, J.R., EPArESED. September 6, 1985. Response to Section 114
information request.
9. Memo and attachments from Maxwell, W.H., EPArlSB, to Primary Aluminum
Docket (A-86-07). June 18, 1986. Fluoride emissions from NSPS primary
aluminum plants.
10. Letter and attachments fromGivens, H.L., Alcan, to Noble, E.A., EPArlSB.
Received June 28, 1985. Emission test data.
11. Letter and attachments from Casswell, S.J., Commonwealth Aluminum, to
Noble, E., EPArlSB. November 21, 1985. Emission test data.
12. Letter and attachment from Dickie, R.C., Al umax of South Carolina, to
Noble, E., EPArlSB. May 30, 1985. Emission test data.
13. Reference 9.
14. Reference 9.
15. Reference 3.
16. Reference 3.
5-21
-------
6. COST ANALYSIS
Of the four types of primary aluminum reduction pots, only two types
have been built in the years since the new source performance standards
(NSPS) were promulgated. Those two are the center-worked prebake (CWPB)
and the vertical stud Soderberg (VSS) pots. Industry sources advise that
if any more primary aluminum plants are built they will contain potlines
with CWPB pots. The one new VSS potline is unique in that it is the only
one subject to the NSPS that has a sulfur dioxide (S02) scrubber.
The cost analysis will center on four areas. The first area will deal
with the fluoride control costs for CWPB plants and will include both the
potlines and the anode bake furnaces. The second area will cover fluoride
control costs for a VSS plant and will include potline controls as well as
controls for the fugitive fluorides leaving the building (primary and
secondary controls). The third area will deal with the costs of S02 controls
for the CWPB plants and the fourth with S02 controls for the VSS plants.
6.1 FLUORIDE CONTROLS
Only one level of fluoride emission control has been designated.
Since no better technology has been developed, the dry alumina scrubber
remains the best demonstrated technology for the control of total fluorides
(TF). However, there is a difference in the level of control between the
CWPB and the VSS units. The overall fluoride control efficiency for the
CWPB potlines is assumed to be 98.5 percent, and that for the VSS potlines
to be 89.6 percent based on information submitted by industry.
6.1.1 Costs for CWPB Fluoride Controls
Costs are drawn from two sources. The first source is a study of the
cost for fluoride controls published in 1985 by the International Primary
Aluminum Institute (IPAI).1 The second source is information sent in from
U. S. plants currently subject to the NSPS. The IPAI study covers the
control costs for ten plants that are either new or recently retrofitted.
Six of these are CWPB plants controlled by dry scrubbers which use fresh
-------
alumina as an adsorbing material.^/ Fabric filters in the dry scrubbers
collect the alumina which is then routed to the potline for use. Cost data
for the six CWPB plants with dry scrubbers are shown in Table 6-1, along with
the data from four NSPS installations. One other plant submitted data but a
confidentiality request was made.
The IPAI costs are quoted in January 1981 dollars. These costs were
r\
escalated to August 1985 dollars using the Chemical Engineering Plant Index.L
The IPAI data did not specify any anode bake furnace costs, so it was assumed
that they were not included in the overall costs. The costs for the four
NSPS installations have also been escalated to August 1985 dollars. Using
an August 1985 price quotation of $1.047 per kilogram ($0.475 per pound) of
aluminum from the Mineral Industries Surveys, the potential percentage
increase in the price of aluminum resulting from the installation of controls
was calculated and ranged from a credit of 1.3 percent (due to fluoride values
recovered) to a positive 2.0 percent.^
Table 6-2 lists the fluoride control efficiency data and is presented to
facilitate the cost effectiveness calculations shown in Table 6-3. Table 6-2
does not contain any emissions data for anode bake plants at the number-coded
plants, since the IPAI document does not list bake plant emissions. Nor
does it contain any emission information for the potlines at plants J and
K, since only the anode bake plants at these facilities are subject to the
NSPS.
The cost effectiveness in Table 6-3 varies from a negative $476 per
megagram total fluoride ($476/Mg TF, $431/ton TF), to a positive $935/Mg
($848/ton) TF for the potlines. Cost effectiveness ratios for the anode
bake plants vary from $9,400 to $ll,500/Mg ($8,500 to $10,500/ton) TF.
This might be attributable to an incorrect estimate of TF evolution from
the furnace (no data are available on TF evolution rates at plants subject
to the NSPS).
V One of the four exceptions is a side-worked prebake (SWPB) plant that was
retrofitted to a CWPB unit. A second is a SWPB plant, unchanged. A third is
a CWPB plant that utilizes electrostatic precitators before and after the
dry scrubber to remove the fluoride laden alumina from the exhaust gases. The
fourth utilizes a wet scrubber after the dry scrubber.
6-2
-------
TABLE 6-1
COSTS OF DRY SCRUBBERS TO CONTROL TOTAL FLUORIDE EMISSIONS
AT CENTER-WORKED PREBAKE PLANTS5'11
Plant
code3
3
4
7
8
9
10
C/G
D/H
J
K
Capacity
Mg Al/yrb
40,000
60,000
106,000
98,000
171,000
230,000
55,000
77,000
311 ,000
271,000
Control
investment, $xlQ6
1/81$
8.00
7.75
14.32
9.66
19.75
28.000
8.509
44.48"
3.54i
2.84k
8/85$
9.40
9.11
16.83
11.35
23.21
32.90
12.67
45.81
3.86
2.95
Net annual TF control costs, $/Mg Al produced
Potlines
1/81$
4.0
<7.6>f
10.0
<7.8>
<10.9>
6.0
-
-
--
__
8/85$c
4.7
<8.9>
11.8
<9.2>
<12.8>
7.1
-
-
--
--
Furnace
L_8/85$
_ e
-
-
-
-
-
-
-
4.9
4.0
Total cost
8/85$
4.7
<8.9>
11.8
<9.2>
<12.8>
7.1
13.0
19. 5i
4.9
4.0
Cost as %
of price^
0.4
<0.8>
1.1
<0.9>
<1.2>
0.7
1.2
1.9
0.5
0.4
Arabic numerals represent plant designations from the IPAI report. Letters
refer to designations of plants subject to the NSPS. coding is for simplicity.
b
c
d
e
f
g
h
i
J
k
Mg Al/yr = megagram aluminum per year = 1.1 short tons Al/yr.
IPAI potline costs in January 1981 dollars were updated to August 1985
using the Chemical Engineering Plant (CEP) Index (325.0/276.6 = 1.175)
The U.S. Market price of aluminum, $0.475/lb or $950/ton ($1047/Mg), August 1985.
Not reported or not applicable
Credit
CEP Index 1978 annual to August 1985 (325.0/218.8 = 1.49)
CEP Index 1984 annual to August 1985 (325.0/322.7 = 1.01)
CEP Index 1983 annual to August 1985 (325.0/316.9 = 1.03)
CEP Index 1981 annual to August 1985 (325.0/297.0 = 1.09)
CEP Index 1982 annual to August 1985 (325.0/314.0 = 1.04)
6-3
-------
TABLE 6-2
EFFECTIVENESS OF TOTAL FLUORIDE CONTROL SYSTEMS AT CWPB PLANTS12'19
Plant
name/
code3
3
4
7
8
9
10
C/G
D/H
J
K
Fluoride evolution
kg/Mg of Alb
Potline
_ c
25.0
13.2
28.0
28.0
27.8
33. Qd
33. Qd
-
-
Furnace
-
-
-
-
-
-
0.436
0.43S
0.43
0.43
Overall fluoride
removal efficiency, %
Potline
-
95.8
95.6
96.6
96.0
97.3
98.1
98.5
-
-
Furnace
-
-
-
-
-
-
99.1
99.2
98.4
98.8
Fluoride removed,
kg/Mq of Al
Potline
-
23.95
12.62
27.05
26.88
27.05
32.37
32.51
-
-
Furnace
-
-
-
-
-
-
0.426
0.427
0.423
0.425
Total
-
23.95
12.62
27.05
26.88
27.05
32.80
32.94
0.423
0.425
a Arabic numerals represent plant designations from the IPAI report. Letters
refer to designations of plants subject to the NSPS. Coding is for simplicity
b kg/Mg of Al = kilograms per megagram of aluminum (1 kg/Mg = 2 pounds per ton).
c Not reported or not applicable.
d Estimated, based on review of available literature and information provided
by the manufacturer. No data are available on TF emissions entering the
primary control systems of potlines subject to the NSPS
e Estimated from the guidance document for primary aluminum reduction plants.
No data are available on uncontrolled TF emissions from anode bake furnaces
subject to the NSPS.
6-4
-------
TABLE 6-3
COST-EFFECTIVENESS OF TOTAL FLUORIDE CONTROL SYSTEMS
Plant
code3
3
4
7
8
9
10
C/G
D/H
J
K
Fluoride removal ,
kg/Mg of Alb
Potline
_c
23.95
12.62
27.05
26.88
27.05
32.37
32.48
-
Furnace
-
-
_
_
_
_
0.426
0.427
0.424
0.425
Total
-
_
_
_
—
_
32.80
32.91
-
Removal costs,
$/Mg of Al
Potline
4.7
<8.9>d
11.8
<9.2>
<12.8>
7.1
-
-
-
Furnace
-
„
_
.
.
.
-
-
4.9
4.0
Total
-
_
13.0
19.5
-
Cost-effecti veness ,
$/Mg of TF ($/ton of TF)
Potline
-
<372>
(<337>)
935
(848)
<340>
(<308>)
<476>
(<432>)
262
(238)
-
-
Furnace
-
_
-
-
11,500
(10,500)
9,400
(8,500)
Total
_
396(359)
592(537)
-
cri
i
en
a Arabic numerals represent plant designations from the IPAI report. Letters refer to designations
of plants subject to the NSPS. Coding is for simplicity.
b kg/Mg of Al = Kilograms per megagram of aluminum (1 kg/Mg = 2 pounds per ton).
c Not reported or not applicable.
d Credit.
-------
6.1.2 Costs for VSS Fluoride Controls
Fluoride controls for a VSS plant are somewhat different from controls
for a CWPB plant because the pots are more difficult to hood. One VSS
potline has been installed since the NSPS was promulgated. This potline is
equipped with primary controls consisting of a dry scrubber and a baghouse
to collect fluoride emissions. Emissions which escape the primary hoods rise
to the top of the potroom building and pass through a secondary scrubber
consisting of screens that are continuously s.prayed with a calcium solution.
This secondary control system has a pressure drop of only about 25 pascals
(0.1 inch of water). However, it removes over 65 percent of the TF present.
It also removes about 62 percent of the S02 and 57 percent of the non-fluoride
particulate. The resulting calcium fluoride/calcium sulfite sludge is flushed
off of the screens and pumped to a lagoon for settling. Table 6-4 presents
the capital investment and annualized costs for the VSS plant controls as
reported by Plant E .20 The-total capital cost for TF and S02 controls is $233/Mg
of annual capacity ($211/ton). The annual ized costs are reduced by a large
credit for fluorides captured by the dry scrubbers. With the plant operating
at full capacity (161,000 Mg/yr, 177,600 tons/yr), the cost per Mg of
aluminum produced is $14.99 ($13.60/ton). At the aluminum price of $l,047/Mg
($950/ton) used earlier, the $14.99/Mg ($13.60/ton) annualized control cost
is 1.3 percent of the selling price. At 75 percent of capacity, the numbers
are $19.99/Mg ($18.13/ton), and 1.7 percent, respectively.
6.2 SULFUR DIOXIDE CONTROLS
Two approaches are possible for the control of S02- The first is to
limit the sulfur content of the anode constituents (the source of the SOg).
As noted in Chapter 2, three plants in two states are operating under
prevention of significant deterioration (PSD) regulations limiting the
sulfur content of their cokes to 3.0 percent in two cases and 0.7 percent
in the other.21 Plants not operating under PSD regulations also report
using anode components with sulfur contents less than 3 percent.22 Other
sources indicate that cokes having a sulfur content of 3 percent or less
will be available for the next 10 years.23-25 AS ^e practice of using less
than 3 percent sulfur anodes is fairly widespread, no costs were developed
for this method of S02 control.
6-6
-------
TABLE 6-4
CAPITAL AND ANNUALIZED COSTS TO CONTROL TOTAL FLUORIDE AND
SOo EMISSIONS FROM VSS PLANTS^b>^7
($000)
Capital Costs
Lines I X II
Line III
Annuali zed Costs
Primary
$ (yr)
8,431(79)3
7,341(80)
340(84)C
$ (8/85)
11,500
9,100
342
Secondary
$ (yr)
2,420(70)
8,510(80)b
2,058(84) .
$(8/85)
6,300
10,600
2,072
Total
$ (8/85)
17,800
19,700
2,414
a Chemical Engineering Plant Index Factors:
1979 Annual Index to August 1985: 325.0/238.7 = 1.36
1970 " " " " : 325.0/125.7 = 2.59
1980 " " " " : 3Z5.0/261.2 = 1.24
1984 " " " " : 325.0/322.7 = 1.007
b Line III secondary scrubber costs include equipment to recycle the
scrubbing medium for all three lines. Lines I and II originally
utilized once-through scrubbing in their secondary scrubbers and
the costs reflect that.
c Annualized costs of $5,963,000 less $5,873,000 recovery credits, plus
$250,000 reporting costs
6-7
-------
The second approach to SOe control is the use of add-on control technology
(i.e., wet scrubbers). The use of CWPB versus VSS pots pose different problems
for add-on S02 control. The CWPB plant uses much greater volumes of air to
capture the pollutants from the pots (about 1.89 cubic meters per second [m3/s],
4,000 actual cubic feet per minute [acfm]) compared to the 0.24 to 0.28 m3/s
(500 to 600 acfm) volume from VSS pots. For instance, a CWPB plant producing
90,700 Mg (100,000 tons) a year of primary aluminum and using coke with 3
percent sulfur will have a stack concentration of about 100 parts per
million volume (ppmv) S02. On the other hand, a VSS plant producing 54,400 Mg
(60,000 tons) a year of primary aluminum and using coke with 3 percent sulfur
will have a stack concentration of about 350 ppmv of S02.
6.2.1 Costs for CWPB SO? Controls
The only add-on control in use on an'NSPS potline is a spray tower on
a VSS line. Therefore, estimating the costs of S02 controls when applied
to CWPB potlines required several assumptions arid cost data were lacking
for some aspects of the analysis. Nevertheless, a rough cost analysis was
performed to estimate the costs of this technology when transferred to a
CWPB potline.28"30 The resulting cost effectiveness values ranged from a
credit up to a cost of $6,200/Mg ($5,645/ton).
Some of the assumptions made in performing the rough cost analysis
include:
0 coke of 3 percent sulfur content would become increasingly unavailable;
0 S02 reductions of 65 to 85 percent would be achievable on CWPB pots
using sodium-alkali scrubbers;
0 costs from a prior document could be updated directly without considering
any cost additions/deletions; and
0 the use of a wet scrubber (spray tower) would be sufficient for fluoride
and particulate control.
Further consideration of this analysis and comments received from the industry
indicates that the resulting costs, and subsequent cost effectiveness
values, are low.
6-8
-------
6.2.2 Costs for VSS S02 Controls
Gases leaving the baghouse in the primary dry scrubber at Plant E are
routed to a wet scrubber (spray tower) which captures the S02 in a sodium
alkali spray. The cost analysis noted earlier also included VSS potlines.31
However, the cost data submitted by industry did not allocate primary
control costs between fluoride, controlled by the dry scrubber, and SOg,
controlled by a wet scrubber with a water treatment facility. In addition,
the results of the analysis understate the costs of wastewater treatment
and disposal, and, thus, the resulting control costs and cost effectiveness
values are low.
6-9
-------
6.3 REFERENCES FOR CHAPTER 6
1. Fluoride Emissions Control: Updated Costs for New Aluminum Reduction
Plants. International Primary Aluminum Institute Environmental
Committee Report. February 1985. Pages 6,7,11,16.
2. Economic Indicators. Chemical Engineering. 88(10):7. May 18, 1981.
3. Economic Indicators. Chemical Engineering. 92_(25):7. December 9, 1985.
4. U.S. Department of the Interior, Bureau of Mines. Mineral Industry
Surveys. Aluminum Industry in January 1985 through Aluminum Industry in
April 1986, inclusive. Prepared in the Division of Nonferrous Metals.
. Dated April 9, 1985, through July 3, 1986.
5. Reference 1.
6. Reference 2.
7. Reference 3.
8. Reference 4.
9. Letter and attachments from Givens, H., Alcan, to Farmer, J.R., EPA:ESED.
September 6, 1985. Response to Section 114 information request.
10. Letter and attachments from Hurt, R.E., Noranda Aluminum, to Farmer, J.R.,
EPA-.ESED. September 25, 1985. Response to Section 114 information request.
11. Letter and attachments from Boyt, J.S., Aluminum Company of America,
to Farmer, J.R., EPA:ESED. October 7, 1985. Response to Section 114
information request.
12. Reference 1.
13. Reference 9.
14. Reference 10.
15. Reference 11.
16. U.S. Environmental Protection Agency. Background Information for
Standards of Performance: Primary Aluminum Industry. Volume I:
Proposed Standards. EPA-450/2-74-020a. October 1974. Pages 23 and 36.
17. U.S. Environmental Protection Agency. Primary Aluminum: Guidelines
for Control of Fluoride Emissions from Existing Aluminum Plants.
EPA-450/2-78-049b. December 1979. Table 3-1.
18. Memo from Noble, E.A., EPA:ISB, to Durkee, K.R., EPA-.ISB. November 26,
1985. Report of June 1985 trip to Aluminum Company of America,
Newburgh, Indiana. Page 4.
6-10
-------
19. Memo and attachments from Maxwell, W.H., EPA:ISB, to Primary
Aluminum Docket (A-86-07). June 18, 1986. Emission control
effectiveness.
20. Letter and attachments from Casswell, S.J., Commonwealth Aluminum,
to Farmer, J.R., EPA:ESED September 6, 1985. Response to Section 114
information request.
21. U.S. Environmental Protection Agency. Compilation of BACT/LAER
Determinations, Revised. EPA-450/2-80-070. May 1980. Source Code 7.1
22. Memo from Maxwell, W.H., EPArlSB, to Primary Aluminum Docket (A-86-07).
May 30, 1986. Green anode composition and calculated S02 emissions.
23. Letter and attachments from Dickie, R.C., Alumax of South Carolina,
to Noble, E.A., EPArlSB. April 16, 1986. Comments on draft document.
24. Letter and attachments from Boyt, J.S., Aluminum Company of America,
to Noble, E.A., EPA:ISB. April 29, 1986. Comments on draft document.
25. Letter and attachments from Goldman, J.H., The Aluminum Association, to
Maxwell, W.H., EPA:I SB. May 21, 1986. Comments on draft document.
26. Reference 3.
27. Reference 20.
28. Memo and attachments from Maxwell, W.H., EPA:ISB, to Primary Aluminum
Docket (A-86-07). June 19, 1986. Cost for add-on S02 control.
29. Singmaster and Breyer. Air Pollution Control in the Primary Aluminum
Industry. EPA-450/3-73-004A. July 23, 1973. Pages 8-22 and 8-26.
30. Reference 3.
31. Reference 28.
6- 11
-------
7. ENFORCEMENT ASPECTS
7.1 COMMENTS
EPA Regional Offices, State agencies, the Aluminum Association, and
companies subject to the new source performance standards (NSPS) were
contacted to determine whether there were any problems with either enforcing
the NSPS or complying with it.
Discussions with EPA offices and State agencies revealed no problems
in enforcing the NSPS. Personnel at the plants contacted reported no
problems in testing, monitoring, or recordkeeping (Table 7-1). One plant
contact did suggest that alternate test methods be considered for secondary
potroom testing.1 He also claimed that the NSPS requirement for installing
Method 14 sampling manifolds and stations is not interpreted consistently.
He noted that his plant was required to install two sampling monitors and
stations per potline, while other NSPS plants had to install only one.
Another respondent noted that his plant is still operating under a consent
decree (no operating permit has been issued). This plant is having
difficulty in consistently achieving emission limits imposed for total
fluorides, nonfluoride particulates, and sulfur dioxide.2 Another commented
about the high cost of testing and the time required to get waivers of the
monthly testing requirements for primary control systems.3
7.2 SECONDARY EMISSION TESTING
Industry contacts have suggested that primary aluminum plants be
allowed to petition for a reduction in the emission test schedule for
secondary emissions from potrooms. As noted in Chapter 2, the 1980
amendment to the NSPS added a requirement for monthly compliance tests.''
At the same time, provisions were made for establishing an alternative (less
frequent) test schedule for primary potroom control systems and for anode
bake plants. No provision was made for reducing the frequency of secondary
potroom testing. However, Section 60.8(b)(4) of the General Provisions gives
-------
TABLE 7-1
COMMENTS RECEIVED FROM PLANTS
WITH NSPS POTLINES OR ANODE BAKE FURNACES
Plant
Alcoa-
Comment Rockdale4 Alumax5 Commonwealth6
° NSPS unclear regarding the number - x
of Method 14 sampling manifolds/
stations required per potline
(One plant required to install
2 per line, others needed only
one)
° Excessive time required to x x
conduct monthly testing of primary
and secondary emissions
» Consider alternate test methods - x
o Consider reducing frequency of - x
secondary tests
Excessive time to get variance
on monthly test requirement for
primary control systems
Excessive time required to get
operating permit (still operating
under consent decree)
7-2
-------
to the Administrator, and subsequently to the States whose delegation requests
have been approved, the authority to evaluate on a case-by-case basis whether a
reduced test frequency is reasonable.^
There are two aspects to the question of whether a specific alternative
test schedule should be made part of the standard for potroom secondary
emissions. One is the possiblity that the normal variability of potroom
emissions will result in periodic exceedances of the NSPS. The other is the
very real possibility that plants granted a less stringent test schedule
might cut back on maintenance activities and relax work practices. The first
possibility can be quantified with an adequate data base; the second is more
subjective.
Four years of data were evaluated from a plant which meets a State
limitation of 0.51 kilograms total fluoride per megagram aluminum produced
(kg/Mg) (1.02 pounds per ton [lb/ton]), much more stringent than the NSPS.9
This plant is the only new "greenfield" plant subject to the NSPS. It uses
sophisticated computer control techniques for potline operation and monitoring.
The test results from this plant showed average TF emissions of 0.43 and 0.45
kg/Mg (0.86 and 0.90 lb/ton) (2 potlines). The data revealed that, 'if one
assumes that plant operation and maintenance practices remain unchanged, the
probability of exceeding the NSPS at this plant due to random variation alone
is extremely remote (less than once every 100,000 years). The risk of an
exceedance would be greater for plants with higher emissions or greater
emissions variability. A procedure was developed for making similar assess-
ments for other plants.
As noted earlier, the principal risk involved in reducing test frequency,
aside from the possibility of a random failure, is that plants might take
this opportunity to reduce their maintenance efforts and relax work practices.
This risk might be reduced to a more acceptable level by mandating the develop-
ment and use of work practices, housekeeping, and hood inspection programs to
supplement less frequent testing. As noted in Chapter 4, informal inspection
programs (which include hood inspections) have already been developed by some
plants with NSPS potlines, to serve two purposes:
0 to help in the allocation of maintenance dollars, and
0 in the event of a test-failure, to support the claim of having a viable
and continuing maintenance program.
7-3
-------
The value of such programs cannot be determined with any certainty, because
there has been no attempt to correlate such programs to the emission levels
experienced.
States can use the statistical procedure developed to assess on a case-by-
case basis the appropriateness of reduced secondary emission test frequency.
The procedure documents the amount of test data needed to make an accurate
assessment of the purely statistical probability of failure and the formulae
to be used in making this assessment. Procedures for ensuring adequate
operation and maintenance practices would need to be tailored to plant
specific conditions.
7.3 NSPS INTERPRETATION
The affected facilities covered under the standards are each potroom group
and each anode bake furnace. A potroom group can be an uncontrolled potroom, a
potroom which is controlled by a single primary control device, or a group of
potrooms or potroom segments ducted to a common primary control device.
Typical potlines built since proposal of the standards have been housed in
two potrooms. Each of these potrooms have been divided in half by crane- and
traffic-ways. Thus, there are four potroom segments per potline. The typical
ductwork configuration has taken the primary emissions from two of these
segments to one primary control device, forming two potroom groups per potline.
The NSPS, and Reference Method 14, also require one secondary emissions
sampling manifold per potroom group. Under the typical configuration noted
above, two secondary monitors would be required per potline.
It has been determined that interpretation and application of the NSPS is
not consistent, at least with regard to the number of secondary monitoring
stations required and the application of the NSPS emission limits, at prebake
plants. The three NSPS prebake facilities are discussed below.
7.3.1 Plant 1
This plant has two potroom groups per potline as per the affected facility
designation of the NSPS. Each primary control device is tested and the
results reported separately. The plant also has one secondary emission
monitoring system per potroom group, in accordance with the standards, and
each is tested monthly.
7-4
-------
7.3.2 Plant 2
This plant also has two potroom groups per potline. However, the two
primary emission test values are added together to provide one primary value
per potline. In addition, the plant only has one secondary monitoring system
for the entire potline instead of the two specified by the NSPS. Data from
this one secondary station are combined with the one primary emission value
to determine compliance with the NSPS emission limit. That is, two primary
emission tests results and one secondary emission test result are added to
determine compliance with the NSPS limit. Thus in this case, the NSPS emission
limit is applied to the entire potline, but with only one secondary emission
value, rather than to the individual potroom groups.
7.3.3 Plant 3
This plant also has two potroom groups per potline. In addit-ion, it has
two secondary monitoring stations per potroom group instead of the one
required. However, only one of the stations is used during the monthly test,
instead of one per potroom group. It cannot be determined from the data
submitted how these test results are combined, but it appears that this
plant also is meeting the NSPS emission limit applied to, the potline rather
than to the potroom groups individually but with only one secondary monitor
data value.
7-5
-------
7.4 REFERENCES FOR CHAPTER 7
1. Letter and attachments from Dickie, R.C., Alumax of South Carolina,
to Farmer, J.R., EPA:ESED. August 27, 1985. Response to Section 114
information request.
2. Letter and attachments from Casswell, S.J., Commonwealth Aluminum,
to Farmer, J.R., EPA:ESED. September 6, 1985. Response to Section 114
information request.
3. Letter and attachments from Boyt, J.S., Aluminum Company of America,
to Farmer, J.R., EPA:ESED. October 7, 1985. Response to Section 114
information request.
4. Reference 3.
5. Reference 1.
6. Reference 2.
7. U.S. Environmental Protection Agency. Standards of Performance for
New Stationary Sources: Primary Aluminum Plants; Amendments. 40 CFR
Part 60, Subpart S. Federal Register, Vol. 45, No. 127. Monday,
June 30, 1980. Pages 44202-43?I7:
8. U.S. Environmental Protection Agency. Code of Federal Regulations.
Title 40, Chapter I, Subchapter C, Part 60, Subpart A. Washington,
D.C. Office of the Federal Register. July 1, 1984. pp 180-181.
9. U.S. Environmental Protection Agency. Primary Aluminum:
Statistical Analysis of Potline Fluoride Emissions and Alternate
Sampling Frequency. EPA-450/3-86-012. October 1986.
7-6
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing/
1. REPORT NO.
EPA 450/3-86-010
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Review of New Source Performance Standards for
Primary Aluminum Reduction Plants
5. REPORT DATE
September 1986
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Office of Air Quality Planning and Standards
Environmental Protection Agency
Research Triangle Park, NC 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
DAA for Air Quality Planning and Standards
Office of Air and Radiation
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA 200/84
15. SUPPLEMENTARY NOTES
16. ABSTRACT
As required by Section 111 (b) of the Clean Air Act, as amended, a four year
review of the new source performance standards for primary aluminum reduction
plants (40 CFR Subpart S) was conducted. This report presents a summary of the
current standards, the status of current applicable control technology, and the
ability of plants to meet the standards. No revision to the standards are recom-
mended, but EPA should make available a procedure upon which a decision to reduce
the frequency of secondary monitoring can be made.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air pollution
Aluminum industry
Fluorides
Standards of performance
Pollution control
Air Pollution Control
13B
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS /ThisReport)
Unclassified
21. NO. OF PAGES
124
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
------- |