United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/2-78-049a
February 1979
Air
Primary Aluminum
Draft Guidelines for
Control of Fluoride
Emissions from
Existing Primary
Aluminum Plants
-------
EPA-450/2-78-049a
Primary Aluminum Draft
Guidelines for Control of
Fluoride Emissions from Existing
Primary Aluminum Plants
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
FEBRUARY 1r'79
-------
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,
N.C. 27711, or from National Technical Information Services,
5285 Port Royal Road, Springfield, Virginia 22161.
Publication No. EPA-450/2-78-049a
-------
TABLE OF CONTENTS
Page
1. INTRODUCTION AND SUMMARY 1-1
1.1 INTRODUCTION 1-1
1.2 HEALTH AND WELFARE EFFECTS OF FLUORIDES 1-5
1.3 FLUORIDES AND THEIR CONTROL 1-6
1.3.1 Fluorides 1-6
1.3.2 Control of Fluorides: New Primary 1-7
Aluminum Plants
1.3.3 Control of Fluorides: Existing 1-9
Primary Aluminum Plants
1.4 EMISSION GUIDELINES 1-13
1.4.1 State Emission Guidelines 1-13
1..4.2 Performance of Recommended Emission Controls 1-17
1.4.3. Emission Testing 1-19
1.5 ASSESSMENTS 1-20
T.5.1 Economic 1-20
1.5.2 Environmental 1-24
1.5.3 Energy 1-24
1.6 COMPLIANCE TIMES 1-26
1.7 REFERENCES FOR SECTION 1 1-30
2. HEALTH AND WELFARE EFFECTS OF FLUORIDES 2-1
2.1 INTRODUCTION 2-1
2.2 EFFECT OF FLUORIDES ON HUMAN HEALTH 2-2
2.2.1 Atmospheric Fluorides 2-2
2.2.2 Ingested Fluorides 2-4
2.3 EFFECT OF FLUORIDES ON ANIMALS 2-4
2.4 EFFECT OF ATMOSPHERIC FLUORIDES ON VEGETATION 2-6
-------
Page
2.5 THE EFFECT OF ATMOSPHERIC FLUORIDES ON
BUILDING MATERIALS 2-7
2.5.1 Etching of Glass 2-7
2.5.2 Effect of Fluorides on Structures 2-7
2.6 RATIONALE 2-9
2.7 REFERENCES FOR SECTION 2 2-9
3. U.S. PRIMARY ALUMINUM MANUFACTURING STATISTICS 3-1
3.1 EXISTING PLANTS 3-1
3.1.1 Introduction 3-1
3.1.2 Location and Size 3-2
3.2 FUTURE TRENDS 3-8
3.2.1 Domestic Industry and Plant Growth 3-8
3.2.2 Plant Location and Cell type 3-13
3.2.3 New Producers and Technology 3-13
3.3 PRICE STATISTICS 3-16
3.4 REFERENCES FOR SECTION 3 3-17
4. PROCESS DESCRIPTION 4-1
4.1 PRIMARY ALUMINUM REDUCTION 4-1
4.2 PREBAKE PROCESS 4-10
4.2.1 Anode Bake Plant 4-10
4.2.2 Reduction Cells 4-14
4.3 SODERBERG CELLS 4-15
4.3.1 Vertical Stud Reduction Cells 4-16
4.3.2 Horizontal Stud Reduction Cells 4-18
4.4 REFERENCES FOR SECTION 4 4-19
-------
Page
5. FLUORIDE EMISSIONS $-1
5.1 POINTS OF EMISSION 5-1
5.2 UNCONTROLLED EMISSIONS—SOURCE, CHARACTERISTICS, 5-9
AND MINIMIZATION
5.2.1 Reduction Cells (All Types) 5-9
5.2.2 Anode Bake Plant 5-20
5.3 TYPICAL FLUORIDE EMISSIONS AND EXTENT OF CONTROL 5-22
5.3.1 Reduction Cells (All Types) 5-22
5.3.2 Anode Bake Plant 5-26
5.4 REFERENCES FOR SECTION 5 5-28
6. CONTROL TECHNIQUES FOR POTROOM AND ANODE BAKE PLANT
FLUORIDES 6-1
6.1 POTROOM RETROFIT PRIMARY COLLECTION SYSTEMS 6-1
6.1.1 Cell Hooding 6-2
6.1.2 Calculation of Primary Collection Efficiency 6-11
6.1.3 Primary Exhaust Rates 6-19
6.1.4 Ducting Layouts 6-23
6.2 POTROOM AND ANODE BAKE PLANT RETROFIT REMOVAL EQUIPMENT
AND ITS PERFORMANCE 6-29
6.2.1 Potroom Primary Dry Scrubbing 6-29
6.2.2 Potroom Primary and Anode Bake Plant
Wet Scrubbing 6-34
6.2.3 Potroom Secondary We't Scrubbing 6-39
6.2.4 Summary of Best Retrofit Performance 6-47
6.3 RETROFIT CASE DESCRIPTIONS 6-50
6.3.1 Plant A—HSS Cells—Primary Dry Scrubbing 6-52
Retrofit
6.3.2 Plant B—HSS Cells—Primary Wet ESP Retrofit 6-76
6.3.3 Plant C—Prebake Cells—Primary Injected
Alumina Dry Scrubbing Retrofit 6-100
i i i
-------
Page
6.3.4 Case Description Summary 6-117
6.4 DESIGN, INSTALLATION AND STARTUP TIMES FOR RETROFIT 6-127
CONTROLS
6.5 REFERENCES FOR SECTION 6 6-135
7. COSTS OF ALTERNATIVE FLUORIDE EMISSION CONTROLS 7-1
7.1 INTRODUCTION 7-1
7.2 SELECTION OF ALTERNATIVE CONTROL LEVELS 7-2
/.3 CAPITAL AND ANNUAL COSTS FOR FLUORIDE EMISSION 7-11
CONTROL OPTIONS
7.3.1 Procedure 7-11
7.3.2 Capital Costs 7-11
7.3.3 Annualized Cost 7-14
7.3.4 Cost Effectiveness 7-16
7.4 REFERENCES FOR SECTION 7 7-23
8. RATIONALE OF STATE EMISSION GUIDELINES FOR EXISTING 8-1
PRIMARY ALUMINUM PLANTS '
8.1 INTRODUCTION 8-1
8.2 FLUORIDE EMISSION CONTROL EQUIPMENT AND COSTS 8-3
8.3 RECOMMENDED STATE GUIDELINES AND COLLECTION 8-6
AND REMOVAL EFFICIENCIES OF CONTROL EQUIPMENT
FOR FLUORIDE EMISSIONS
8.3.1 State Fluoride Emission Guidelines 8-9
8.3.2 Compliance Time 8-13
8.4 EMISSION TESTING 8-14
9. ENVIRONMENTAL ASSESSMENTS 9-1
9.1 AIR 9-3
9.1.1 National Fluoride Emissions from Primary 9-3
Aluminum Reduction Cells
9.1.2 Fluoride Emission Control Systems with 9-6
Extreme Air Pollution Impacts
iv
-------
Page
9.1.3 Fluoride Dispersion 9-8
9.1.4 Particulate Emissions from Aluminum Reduction
Cells 9-14
9.1.5 National Particulate and Fluoride Emissions
from Anode Bake Plants 9-14
9.2 WATER ' 9-17
9.2.1 Effluent Limitations Guidelines for Primary
Aluminum Plants 9-18
9.2.2 Water Pollution Control Technology Required to
Meet 1983 Effluent Guidelines Standards 9-18
9.2.3 National Effluent Emissions from Primary
Aluminum Reduction Plants 9-18
9.2.4 Fluoride Air Emission Control Schemes with
Extreme Water Pollution Impacts 9-24
9.3 SOLID WASTE DISPOSAL 9-26
9.3.1 National Solid Waste Generation Due to Fluoride
Control by the Primary Aluminum Industry 9-26
9.3.2 Fluoride Emissions Control Systems with
Extreme Solid Waste Impacts 9-30
9.4 ENERGY 9-32
9.4.1 National Fluoride Emissions Control Energy
Requirements for the Primary Aluminum Industry 9-34
9.4.2 Fluoride Emission Control Systems with Extreme
Energy Impacts 9-37
9.5 OTHER ENVIRONMENTAL IMPACTS 9-37
9.6 OTHER ENVIRONMENTAL CONCERNS 9-39
9.6.1 Irreversible and Irretrievable Commitment
of Resources 9-39
9.6.2 Environmental Assessment of Delayed Standards 9-42
9.6.3 Environmental Assessment of No Standards 9-42
9.7 REFERENCES FOR SECTION 9 9-45
-------
LIST OF TABLES
Table Page
1-1 Potroom Total Fluoride Emissions in U.S., 1975 1-8
1-2 Retrofit Emission Reductions and Costs for Ten 1-10
Primary Aluminum Plants
1-3 Retrofit Controls for Eight Primary Aluminum Plants 1-11
1-4 State Guidelines for Control of Fluoride Emissions 1-14
From Existing Primary Aluminum Plants
1-5 Fluoride Emission Ranges Corresponding to State T-^8
Guidelines for Existing Primary Aluminum Plants
1-6 Primary Aluminum Control Strategies 1-23
1-7 Environmental Impact of Best Control for Total Fluoride 1-25
Emissions from the Primary Aluminum Industry
1-8 Increments of Progress for Installation of Fluoride 1-27
Emission Controls in an Existing Primary Aluminum Plant
1-9 Total Construction Time for Retrofit Emission Controls 1-28
for Primary Aluminum Plants
2-1 Examples of HF Concentrations (ppb) and Exposure 2-8
Durations Reported to Cause Leaf Damage and Potential
Reduction in Crop Values
3-1 U.S. Primary Aluminum Plants and Capacities, 1974 3-3
3-2 Distribution of Plants by Population, 1971 3-8
3-3 Estimated Distribution of Plants by Environment 3-9
VI
-------
Table
3-4 Production of Primary Aluminum in the United States 3-10
3-5 Growth in Primary Aluminum Plant Average Capacity 3-12
in U.S.
3-6 U.S. Trends in Adoption of Cell Types 3-14
3-7 Primary Aluminum Ingot Price History 3-17
5-1 Experimental Effect of Three Operating Variables 5-15
on Fluoride Generation
5-2 Potroom Total Fluoride Emissions in U.S., 1975 5-24
5-3 Extent of Potroom Control, 1975 5-25
5-4 Ring Furnace Fluoride Emissions in U.S., 1970 5-27
6-1 Primary Collection Efficiency Versus Exposed Annular 6-10
Area for Two VSS Plants
6-2 Calculation of Primary Collection Efficiency for 6-12
One Swiss-Design SUPB Plant
6-3 Calculation of Primary Collection Efficiency for 6-14
One French-Design SWPB Plant—Retrofit Case
Description C
6-4 Calculation of Primary Collection Efficiency for 6-16
Typical American-Design CWPB Plants
6-5 Calculation of Primary Collection Efficiency for One 6-17
HSS PI ant--Retrofit Case Description A
6-6 Estimate of Primary Collection Efficiency for 6-20
Typical VSS Plants
6-7 Primary Collection Efficiency Versus Standard Cubic 6-22
Foot per Ton of Aluminum Produced for Seven Primary
Aluminum Plants
6-8 Gas Volumes and Control Device Module Sizes for 6-27
Economic Impact Analysis
6-9 Capital Cost Comparison Between Courtyard and Central 6-28
Primary Collection Systems
VI1
-------
Table Page
6-10 Performance of Spray Screen Secondary Scrubbers at 6-44
Six Existing Primary Aluminum Plants
6-11 Performance of Best Retrofit Emission Controls for 6-48
Primary Aluminum Potrooms
6-12 Summary of Anode Bake Plant Best Retrofit Performance 6-49
6-13 Major Retrofit Items—Plant A—Lines 1 and 2 6-57
6-14 Major Retrofit Items—Plant A—Line 4 6-65
6-15 Before Retrofit Emissions—Plant A—Lines 1, 2, and 4 6-67
6-16 After Retrofit Emission Estimates—Plant A—Lines 1, 6-67
2, and 4
6-17 Retrofit Capital Cost Estimate—Plant A—Lines 1, 2, 6-71
and 4
6-18 Retrofit Annual Operating Cost Estimate—Plant A— 6-74
Lines 1,2, and 4
6-19 Major Retrofit I terns--Plant B—South Plant 6-80
6-20 Major Retrofit Items—Plant B—North Plant 6-85
6-21 Before Retrofit Maximum Emissions—Plant B—North 6-87
and South Plants
6-22 After Retrofit Maximum Emission Estimates—Plant B-- 6-87
North and South Plants
6-23 After Retrofit Average Emission Estimates—Plant B— 6-90
North and South Plants
6-24 Retrofit Capital Cost Estimate—Plant B 6-95
6-25 Revised Retrofit Capital Cost Estimate—Plant B 6-97
6-26 Retrofit Annual Operating Cost Estimate—Plant B— 6-99
North and South Plants
6-27 Major Retrofit Items—Plant C--PDP Design 6-107
6-28 Major Retrofit Items—Plant C—Alcan Design 6-109
vm
-------
Table Page
6-29 Retrofit Increments of Progress—Place C 6-111
6-30 Emissions Before and After Retrofit—Place C—
Lines 1, 2, and 3 6-112
6-31 Retrofit Capital Cost—Plact C—Lines 1, 2, and 3 6-115
6-32 1974 Annual Cost—Plant C—Lines 1, 2, and 3 6-116
6-33 Potroom Retrofit Emission Reductions and Costs for
Ten Primary Aluminum Plants 6-118
6-34 Sequence of Major Activities in Design and Construction 6-128
of Air Emission Control for an existing Primary
Aluminum Plant
6-35 Delivery Times for Items Required to Construct
Emission Controls for Primary Aluminum Plants 6-131
6-36 Total Construction Time for Retrofit Emission Controls
for Primary Aluminum Plants 6-133
6-37 Increments of Progress for Installation of Fluoride
Emission Controls in an Existing Primary Aluminum
Plant 6-134
7-1 Primary Aluminum Plant Capacity by Cell Type 7-3
7-2 Total Fluoride Emissions by Cell Type without
lll(d) Regulations 7~4
7-3 Primary Aluminum Control Strategies 7.5
7-4 Control Modules for Upgrading Existing Aluminum
Plants. Capital Costs 7-12
7-5 Waste Water Lime Treatment Investment Cost by Size
of Plant 7-15
7-6 Fixed Cost Components ;-17
7-7 Control Modules for Upgrading Existing Aluminum Plants. 7-18
Annualized Cost
7-8 Waste Water Treatment Plant Operating Cost 7-19
7-9 Primary Aluminum Control Strategies 7-21
-------
Table Page
8-1 Control Equipment and Costs 8-4
8-2 The Use of Capital Cost Modules 8-3
8-3 State Guidelines for Control of Fluoride Emissions 8-7
from Existing Primary Aluminum Plants
8-4 Fluoride Emission Ranges Corresponding to State 8-8
Guidelines for Existing Primary Aluminum Plants
9-1 National Total Fluoride Emissions from Primary 9-4
Aluminum Reduction Cells
9-2 Average Fluoride Emissions for Primary Aluminum
Reduction Cells 9-5
9-3 Fluoride Emission Control Systems with Extreme
Air Pollution Impacts 9-7
9-4 Fluoride Emissions at Plants A, B, and C 9-9
9-5 Maximum 30-day Average Ambient Fluoride Concentra-
tions in the Vicinity of Plants A, B, and C 9-13
9-6 National Particulate Emissions from Primary Aluminum
Reduction Cells 9-15
9-7 Average Particulate Emissions for Primary Aluminum
Reduction Cells 9-16
9-8 Anode Bake Plant Air Pollutant Emissions 9-17
9-9 Effluent Limitations for Primary Aluminum Plants 9-19
9-10 National Effluent Emissions from Primary Aluminum
Plants 9-21
9-11 Average Effluent Emissions from Primary Aluminum
Reduction Plants 9-22
9-12 Extent of Wet Controls at Alternative Levels of
Fluoride Air Emissions Control 9-23
9-13 Fluoride Air Emissions Control Schemes with Extreme
Water Pollution Impacts 9-25
9-14 Solid Waste Generation for Various Fluoride Emissions
Control Schemes 9-27
-------
Table Page
9-15 National Solid Waste Generation from Fluoride
Control by the Parimary Aluminum Industry 9-28
9-16 Average Solid Waste Generation Resulting from
Fluoride Control for the Primary Aluminum Industry 9-29
9-17 Fluoride Emissions Control Systems with Extreme Solid
Waste Impacts 9-31
9-18 Energy Requirements for Primary Aluminum Fluoride
Emission Control System 9-33
9-19 National Fluoride Emissions Control Energy Requirements
for the Primary Aluminum Industry 9-35
9-20 Average Fluoride Emissions Control Energy Requirements
for the Primary Aluminum Industry 9-36
9-21 Fluoride Emissions Control Systems with Extreme
Energy Impacts 9-38
9-22 National Criteria Pollutant Emissions Resulting from
the Electric Power Generated to Control Primary
Aluminum Fluoride Emissions 9-40
9-23 National Bituminous Coal Requirements Implied by
Primary Aluminum Fluoride Control 9-41
9-24 Environmental Impact of No Fluoride Emission
Guidelines for the Primary Aluminum Industry 9-44
XI
-------
LIST OF FIGURES
Figure Page
4-1 Aluminum Reduction Process 4-2
4-2 Aluminum Reduction Cell Diagram 4-3
4-3 Typical Plan View of a Potroom 4-6
4-4 Typical Elevation View of a Potroom 4-6
4-5 General Flow Diagram for Primary Aluminum Reduction 4-8
4-6 Flow Diagram for Preparation of Prebake Anodes 4-9
4-7 Details of Prebake Reduction Cell 4-15
4-8 Details of Vertical Stud Soderberg Reduction Cell 4-17
4-9 Details of Horizontal Stud Soderberg Reduction Cell 4-19
5-1 Prebake Plant with Anode Ring Furnace 5-2
5-2 Potroom Fluoride Emission Balance 5-3
5-3 Room Collection System, Sidewall Entry 5-6
5-4 Room Collection System, Basement Entry 5-6
5-5 Specific Prebake Potroom Fluoride Balance 5-7
5-6 Particle Size Weight Distribution of Potline 5-10
Primary Cell Emissions
5-7 Particle Size Weight Distribution of HSS Primary 5-12
Cell Emissions
6-1 Typical Prebake Cell Hooding 6-3
6-2 Typical Horizontal Stud Soderberg Cell Hooding 6-7
6-3 Typical Vertical Stud Soderberg Cell Hooding 6-9
6-4 Primary Collection Systems: Typical Ducting fi_?4
Layouts for a Single Prebake Potline with 160
Cells, 2 Rooms
6-5 Primary Collection Systems: Typical Ducting fi 25
Layout for a Single VSS Potline with 160
Cells, 2 Rooms
xii
-------
Figure
6-6 Primary Collection Systems: Typical Ducting Layouts 6-26
for a Single HSS Potline with 160 Cells, 2 Rooms
6-7 Flow Diagram for Fluidized Bed Dry Scrubbing 6-31
Process
6-8 Flow Diagram for Injected Alumina Dry Scrubbing 6-33
Process
6-9 Unpowered Roof Spray Screen 6-40
6-10 Powered Potroom Spray Screen Scrubber 6-41
6-11 Powered Spray Screen Scrubber 6-42
6-12 Powered Monitor Spray Screen Scrubber 6-43
6-13 Retrofit Layout—Plant A—Lines 1 and 2 6-56
6-14 Retrofit Layout—Plant A—Line 4 6-62
6-15 Retrofit Layout—Plant B—South Plant 6-79
6-16 Retrofit Layout—Plant B—North Plant 6-83
6-17 Retrofit Layout—Plant B—Cryolite Recovery Plant 6-91
6-18 Flow Diagram—Plant C—Injected Alumina Process 6-104
6-19 Retrofit Schematic—Plant C—POP Design 6-106
6-20 Retrofit Schematic—Plant C—A!can Design 6-108
6-21 Diagrammatic Representation of Activity Schedules 6-130
on a Major Process Industry Construction Project
7-1 Investment and Annualized Costs for Waste Water 7-20
Treatment Plants vs. Aluminum Plant Capacity
xm
-------
1. INTRODUCTION AND SUMMARY
1.1 INTRODUCTION
Section lll(d) of the Clean Air Act, 42 U.S.C. 7411(d), as
amended, requires EPA to establish procedures under which States submit
plans to control certain existing sources of certain pollutants. On
November 17, 1975 (40 FR 53340), EPA implemented section lll(d) by
promulgating Subpart B of 40 CFR Part 60, establishing procedures and
requirements for adoption and submittal of State plans for control of
"designated pollutants" from "designated facilities." Designated
pollutants are pollutants which are not included on a list published
under section 108(a) of the Act (National Ambient Air Quality Standards)
or section 112(b)(l)(A) (Hazardous Air Pollutants), but for which
standards of performance for new sources have been established under
section lll(b). A designated facility is an existing facility which
emits a designated pollutant and which would be subject to a standard
of performance for that pollutant if the existing facility were new.
Standards of performance for three categories of new sources in
the primary aluminum industry were promulgated in the FEDERAL REGISTER
(40 FR 3826) on January 26, 1976, to be incorporated into the Code of
Federal Regulations under 40 CFR Part 60. New subpart S was added to
set standards of performance for fluoride emissions from new and
modified affected facilities within primary aluminum reduction plants.
The States, therefore, are required to adopt fluoride emission
standards for existing primary aluminum plants which would be subject
to the standard of performance if they were new.
-------
Subpart B of 40 CFR Part 60 provides that EPA will publish a
guideline document for development of State emission standards after
promulgation of any standard of performance for a designated pollutant.
The document will specify emission guidelines and times for compliance
and will include other pertinent information, such as discussion of
the pollutant's effects on public health and welfare and a description
of control techniques and their effectiveness and costs. The emission
guidelines will reflect the degrees of emission reduction attainable with
the best adequately demonstrated systems of emission reduction, considering
costs as applied to existing facilities.
After publication of a final guideline document for the pollutant in
question, the states will have nine months to develop and submit plans
for control of that pollutant from designated facilities. Within four
months after the date of submission of plans, the Administrator will
approve or disapprove each plan (or portions thereof). If a State
plan (or portion therefor) is disapproved, the Administrator will
promulgate a plan (or portion thereof) within six months after the
date for plan submission. These and related provisions of subpart B
are basically patterned after section 110 of the Act and 40 CFR Part 51
(concerning adoption and submittal of state implementation plans under
section 110).
As discussed in the preamble to subpart B, a distinction is drawn
between designated pollutants which may cause or contribute to
endangerment of public health (referred to as "health-related pollutants")
and those for which adverse effects on public health have not been
demonstrated (referred to as "welfare-related pollutants"). For
1-2
-------
health-related pollutants, emission standards and compliance times in
state plans must ordinarily be at least as stringent as the corresponding
emission guidelines and compliance times in EPA's guideline documents.
As provided in Subpart B, States may apply less stringent requirements
for particular facilities or classes of facilities when economic factors
or physical limitations make such application significantly more
reasonable.
For welfare-related pollutants, States may balance the emission
guidelines, times for compliance, and other information provided in a
guideline document against other factors of public concern in establishing
emission standards, compliance schedules, and variances, provided that
appropriate consideration is given to the information presented in the
guideline document and at public hearing(s) required by subpart B and
that all other requirements of subpart B are met. Where sources of
pollutants that cause only adverse effects to crops are located in
non-agricultural areas, for example, or where residents of a community
depend on an economically marginal plant for their livelihood, such
factors may be taken into account (in addition to those that would
justify variances if a health-related pollutant were involved). Thus,
States will have substantial flexibility to consider factors other than
technology and cost in establishing plans for the control of welfare-
related pollutants if they wish. In developing and applying standards for
existing primary aluminum plants, the States are encouraged to take into
consideration, among other factors, the remaining useful life of the affected
facility. Where a facility includes both old and new cells, it may be
-------
reasonable to apply less stringent standards to the old cells provided
that they are significantly closer to retirement than the new cells.
For reasons discussed in chapter 2 of this document, the
Administrator has determined that fluoride emissions from primary
aluminum plants may cause or contribute to endangerment of the public
welfare but that adverse effects on public health have not been
demonstrated. As discussed above, this means that fluoride emissions
will be considered a welfare-related pollutant and the States will
have greater flexibility in establishing plans for the control of
fluorides than would be the case if public health might be affected.
This guideline document for primary aluminum plant fluoride emissions
provides a brief description of the primary aluminum industry and an
explanation of the four types of electrolytic cells, along with a summary
of national statistics on production, plant location, cell type, and
future trends. The causes, nature, and source of fluoride emissions from
primary aluminum reduction are discussed, and the health effects
associated with this pollutant are described. The greatest emphasis,
however, has been placed on the technical and economic evaluation of
control techniques that are effective in reducing fluoride emissions,
with particular emphasis on the retrofit of existing plants. Because of
this emphasis, EPA personnel visited nine primary aluminum plants to
gather engineering information and cost data on actual emission control
retrofits. Detailed trip reports were composed as background source
materials to support this document. Section 6.3 presents "retrofit case
descriptions" for three of the nine plants visited: these case
1-4
-------
descriptions give in depth engineering scopes of work including plant
layouts to scale; duct sizes and lengths; major items of structure,
emission control, and auxiliaries, along with cost breakdowns; and air
'emission reductions realized for the money spent.
The control equipment and the emission guidelines on which they
are based are discussed in Chapters 6 and 8. The environmental
assessment of the emission guidelines is presented and discussed in
Chapter 9. The remainder of this introductory chapter summarizes
information presented in subsequent sections.
1.2 HEALTH AND WELFARE EFFECTS OF FLUORIDES
Fluoride emissions from primary aluminum plants have been determined
to be welfare related [i.e. no demonstrated impact upon public health
for purposes of section lll(d)]. The daily intake of fluoride inhaled
from the ambient air is only a few hundredths of a milligram - a very
small fraction of the total intake of the average person. If a person is
exposed to ambient air containing about eight micrograms (yg) of fluoride
per cubic meter, which is the maximum average concentration that is projected
in the vicinity of a primary aluminum facility with only moderate control
equipment (Table 9-5), his total daily intake from this source is
calculated to be about 150 yg. This is very low when compared with the
estimated daily intake of about 1200 yg from food, water and other sources
for the average person. Also, the intake of fluoride indirectly through
standard,food chains is insignificant. Fluorides are not passed into
dairy products and are only found in farm produce in very small amounts.
1-5
-------
Fluorides do, however, cause damage to livestock and vegetation
in the immediate vicinity of primary aluminum plants. Ingestion of
fluorides by livestock frem hay and forage causes bone lesions, lameness
and impairment of appetite that can result in decreased weight gain or
diminished milk yield. It can also affect developing teeth in young
animals, causing more or less severe abnormalities in permanent teeth.
Exposure of plants to atmospheric fluorides can result in accumulation,
foliar lesions, and alteration in plant development, growth, and
yield.
1.3 FLUORIDES AND THEIR CONTROL
1.3.1 Fluorides
For purposes of standards of performance for new stationary sources,
(SPNSS) and the attendant requirements of section lll(d), "total
fluorides" means the particulate and gaseous fluorides that are measured
by test methods as set forth in Methods 13A, 13B and 14, Appendix A to
40 CFR Part 60.
Particulate fluorides emitted from primary aluminum plants include
cryolite (NaJUFg), aluminum fluoride (A1F3), calcium fluoride (CaF2),
and chiolite (Na5Al3F^4).
The principal gaseous fluoride compounds emitted during normal
operation are hydrogen fluoride (HF) and silicon tetrafluoride (Si'F.).
The intent of the SPNSS is to limit emissions of all of the above
compounds. EPA source tests have shown that if fluorides are well-
controlled, the resulting control of particulates and organics will
also be good. Control of all these pollutants requires good capture of
gases from the electrolytic cell and good fluoride removal from the
captured cell gases. Good capture requires not only good cell and
ventilation system design, but also superior equipment maintenance and
1-6
-------
careful cell working in a manner to minimize the time during which any
cell or cell door is open. Thus, the SPNSS requires control by other
methods than best practical design alone.
1.3.2 ronf"ol of Fluorides: New Primary Aluminum Plants
In accordance with section 111 of the Clean Air Act, standards of
performance were promulgated on January 26, 1976, for total fluoride and
visible air emissions from new modified, or reconstructed primary aluminum
plants. (40 CFR 60 - Subpart S).
Section 60.192 of Subpart S states that no owner or operator shall
cause to be discharged into the atmosphere from any affected facility any
gases which contain total fluorides in excess of:
(1) 1 kg/Mg (2 Ib/ton) of aluminum produced for vertical stud
Soderberg and horizontal stud Soderberg plants;
(2) 0.95 kg/Mg (1.9 Ib/ton) of aluminum produced for potroom groups
at prebake plants; and
(3) 0.05 kg/Mg (0.1 Ib/ton) of aluminum equivalent for anode
bake plants.
The owner shall record aluminum and anode production with an accuracy
of +_ 5 percent. In addition, the air pollution control system for each
affected facility shall be designed for accurate volumetric flow rate
determination and representative total fluoride sampling.
Amendments to Subpart S were proposed on September 19, 1987 (43
FR 42188) and are scheduled for promulgation in July of 1979.
Table 1-1, based on Table 5-2, gives emissions for each of the
aluminum reduction cell types, and indicates overall control efficiencies
for total fluorides. It is evident that future new plants will require much
better control.
1-7
-------
Table 1-1. POTROOM TOTAL FLUORIDE EMISSIONS IN U.S., 1975
I
oo
Cell Type3
CWPB
SWPB
HSS
VSS
All Cell Types
Controlled and Uncontrolled
Plants
Average
^
kg Al
3.15
6.3
2.85
2.6
3.5
Emissions
Ib F
ton Al
6.3
13.0
5.7
5.2
7.0
Overall Control
Efficiency, %
85
71
83
88
83
Controlled Plants
Average
kg~AT
2.3
2.4
2.85
2.5
2.5
Emissions
Ib F
ton Al
4.6
4.8
5.7
5.2
5.0
Overall Control
Efficiency, %
89
89
83
88
88
CWPB - center-worked prebake; SWPB
vertical stud Soderberg.
- side-worked prebake; HSS - horizontal stud Soderberg; VSS -
-------
1.3.3 Control of Fluorides: Existing Primary Aluminum Plants
The intent of Subpart B is to apply best adequately demonstrated
control in order to limit the emissions of designated pollutants from
designated facilities. As applied to Subpart S, this means applying
particulate and gaseous fluoride control to existing primary aluminum
plants.
Table 1-2 is based on Table 6-33 and summarizes costs and total
emission reductions for retrofit emission controls at 10 primary
aluminum plants visited and studied by EPA personnel. Plants A, B,
and C represent those that are presented as retrofit case descriptions
in Sections 6.3.1, 6.3.2, and 6.3.3, respectively. Cost and retrofit
information for plants D through M is much less complete and detailed.
Table 1-3 identifies the retrofit controls used on eight of the plants
visited.
As Table 1-3 indicates, one or two of the retrofits were not fully
installed when the costs were collected; the final costs of these were
therefore not known. Plants A and B (Table 1-2) were chosen for case
descriptions over plants D through G for several reasons other than cost
alone. Indeed, a rating system was used to choose the best three of the
seven plants for retrofit case descriptions. Additional rating factors
included amount of available engineering description, decrease in
fluoride emissions by retrofitting, and quality of emission data before
and after retrofitting.
1-9
-------
Table 1-2. RETROFIT EMISSION REDUCTIONS AND COSTS FOR TEN PRIMARY ALUMINUM PLANTS
Plant
code
G
D
F
H
C
K
A
B
E
M
Cell
type
CWPB
CWPB
CWPB
CWPB
SWPB '
SWPB
HSS
HSS
-vss
VSS
Plant capacity,
short tons/yr
250,000
115,000
32,850
130,000
265,000
35,000
80,000
210,000
91 ,000
180,000
Total emissions, Ib F/torv Al
Before ,
retrofit
19.0
7.3
5.1
6.9
9.0
7.7
9.3
5.4
4.2
5.3
After ,
retrofit0
2.7
2.6
1.2
2.5;
1.3
10. 6d
2.4
2.9
2.0
1.9
Retrof i t
capital
cost, $/ton Al
124
102
54
216
54
121
Change in
annual operating
cost, $/ton Al
NA
MA
NA
NA
3.53C
NA
141 1 -4.78e
02
64 I
115
0.81
NA
9.57
s
CWPB - center-worked prebake; SWPB - side-worked nrebake; HSS - horizontal stud Soderberg;
VSS - vertical stud Soderberg.
Average primary and secondary total fluoride emissions.
°Credit not taken for fluoride adsorbed on alumina and returned to process.
dSee Section 6.3.4.2.
'Credit taken.
-------
Table 1-3. RETROFIT CONTROLS FOR EIGHT PRIMARY ALUMINUM PLANTS
hant
code
G
D
F
C
A
H
B
E
Retrofit control ^
Primary
Collection system
Dry scrubbing
Dry scrubbing
Dry scrubbing
Dry scrubbing
Dry scrubbing
Dry scrubbing
Wet ESPa
Wet ESP
Secondary
None
None
None
No change
None
None
None
Elevated
scrubbers
Retrofit status (March 1975)
Started */73; comnlete earlv 1976
Start mid-1975; complete about 7/78
Started 9/71; complete 8/74
Completed in May 1970
Completed in Spring 1973
Started 7/72; complete ?/74
Started 5/75; complete early 1977
Started late 1970; complete about 6/75
Complete
ESP - electrostatic precipitator.
-------
Table 1-2 indicates that dry scrubbing - baghouse devices can be
retrofitted to reduce total fluorides. Dry scrubbing absorbs gaseous
fluoride on alumina. This is followed by baghouse collection of residual
particulate fluoride and alumina. Both the alumina and fluorides can then
be recycled to the electrolytic cells, thus avoiding a solid and liquid
waste problem. Wet electrostatic precipitators (ESP) can approach the
dry scrubber in performance, but they produce an aqueous waste and must
be coupled with scrubbers to absorb all of the HF.
The important message in Table 1-2 is that fluoride controls installed
in existing primary aluminum plants vary greatly in capital cost and even
in operating costs. Indeed, the actual control must be specified to the
plant and tailor-made for it. No off-the-shelf control devices are
available and retrofit emission control models are valid only for the
conditions and situations upon which they are based, when single plants are
studied.
Under the Clean Air Act Amendments of 1977, existing facilities must be
controlled to the degree of emission reduction attainable with the best
adequately demonstrated system, considering costs and any non-air quality
health, environmental impact, and energy requirements. For fluoride
emission control from existing primary aluminum facilities, the efficient
removal of fluorides from a gas stream is relatively easy. However, a
significant portion of the gaseous emissions from the reduction cells
can escape capture by the collection hoods, thus by-passing the primary
control system. This situation results in two types of emissions: those
captured by the hood system which pass through the primary control device
1-12
-------
(or primary emissions) and those which elude the hood system and exit
the building through the roof monitors (or secondary emissions).
Examination of the average fluoride removal efficiencies of primary
(98.5% removal) and secondary (75% removal) control devices reveals the
importance of good capture of reduction cell emissions by the primary
collection system. Most plants presently do not control secondary
emissions. The conclusion, therefore, is that the best system of emission
reduction, considering costs, is an effective hooding system (which
minimizes secondary emissions) in combination with wet or dry scrubbing
of the primary gases. Because of the relative cell gas volumes of
differing cell designs and plant ages; the physical layout of the plant
which in turn affects duct lengths, available space and access to areas;
the variation in nature and amount of required demolishments, movings,
and removals; and the availability of electrical power; the necessary
collection and control system must be plant-specific.
As illustration of the preceding paragraph, capital costs in dollars
per ton of aluminum produced vary up to threefold or more among the ten
primary aluminum plants shown in Table 1-2. This variation was approached
among the three plants detailed as case descriptions.
1.4 EMISSION GUIDELINES
1.4.1 State Emission Guidelines
Table 1.4 presents the State guidelines for control of total fluoride
emissions from existing primary aluminum plants. These guidelines consist of:
recommended control technologies that EPA believes can readily achieve the
stated fluoride collection and removal efficiencies; an indication
of the status of primary control on a national basis, and the conditions
under which better primary control should be installed or secondary control
1-13
-------
V"
Table 1-4. STATE GUIDELINES FOR CONTROL OF FLUORIDE EMISSIONS FROM EXISTING PRIMARY ALUMINUM PLANTS
Cell Type
VSS
SWPB
HSS
CWPB
Recommended Efficiencies
for Proposed Retrofits
Primary Collection
80
80
90 (a)
95 (a)
Primary Removal
98.5
98.5
98.5
98.5
0
Secondary Removal
75(a)
75 (a)
Guideline Recommendations
All plants now have best achievable hooding
and primary removal .
Install secondary control, but only if justified
depending on severity of fluoride problem.
Install best available hooding and primary removal
equipment.
Install secondary control wherever justified,
depending on the severity of the fluoride problem.
All plants but #26 now have the best achievable
primary collection efficiency. PI ant #26 should
install best primary control if needed.*
Secondary control does not appear to be justified,
in most locations.
Best control is best hooding and primary removal
equipment. Install where needed.
Secondary control does not appear to be justified,
in most locations.
Plant #26; see Tables 7-2 and 7-3
(a) See Section 1.4.1
-------
added.
Column 1 of Table 1-4 shows each of the four primary aluminum
electrolytic cell types. Column 2 shows the corresponding primary
collection (hooding) efficiencies chosen by EPA as readily achievable for
new retrofits. The values in Columns 3 and 4 were similarly chosen for
primary anu secondary removal efficiencies, respectively.
Column 5 of Table 1-4 gives EPA findings and recommendations. All
VSS plants and all but one HSS plant currently (1975) have best achievable
primary collection efficiency, and the VSS plants all have the best
achievable primary removal. The best available primary control systems
should be installed on SWPB and CWPB plants. Secondary controls may be
installed on VSS and on SWPB plants with severe fluoride problems, but
do not appear to be justified on HSS and CWPB plants in most locations.
It is assumed that anode butts will be carefully cleaned and their
fluoride content minimized before recycle to the anode bake plant;
otherwise, anode bake plant control should be required, depending upon
the severity of the fluoride problem.
The primary collection (hooding) efficiencies were chosen from
values calculated by EPA and compared—with good agreement—to values
measured or otherwise arrived at by plant owners. These derivations of
2
hooding efficiencies are discussed in detail in Section 6.1.2 and elsewhere.
Well-designed retrofit hoods can easily obtain the tabulated efficiencies
if properly maintained and if the cells are carefully operated. Similarly,
good retrofit dry scrubbers or spray tower-electrostatic precipitator
combinations can readily achieve 98.5 percent fluoride removal.
Some existing secondary removal units (scrubbers) may not be able to
achieve 75 percent efficiency (see Section 6.2.3). Control officials should
carefully study costs, impacts, and energy requirements before requiring
either improvements or initial retrofits.
1-15
-------
If hooding were added to, or replaced on, an existing HSS plant,
EPA believes that modern technology can achieve 90% collection efficiency,
in almost all cases. A plant may exist, however, where there is insufficient
space to install proper ducting, or it may be economically impractical to
install best hooding.
If a modern primary removal system were added to, or replaced on,
an existing HSS plant, EPA believes that modern technology can achieve
98.5% removal efficiency. A modern spray tower added ahead of existing
wet ESP's can raise primary removal efficiency to 98.5. This spray tower
addition may be economically impractical if sufficient space does not exist
within a reasonable distance of fluoride source and wet ESP.
If an HSS plant has existing primary collection efficiency of 85-90%
and primary removal efficiencies of 95 - 98.5%, control agencies should
closely study costs and benefits, before requiring retrofits. In the
above efficiency ranges, retrofit does not seem justified unless there is
a local fluoride problem.
If hooding were added to, or replaced on, an existing CWPB plant,
EPA believes that modern technology can achieve 95% collection efficiency,
in almost all cases. _A plant may exist, however, where there is insufficient
space to install proper ducting, or it may be economically impractical-to
install best hooding.
If a modern primary removal system were added to, or replaced on, an
existing CWPB plant, EPA believes that modern technology can achieve 98.5%
removal efficiency.
If a CWPB plant has existing primary collection efficiency of 90-95%
and primary removal efficiencies of 95 - 98.5%, control agencies should
closely study costs and benefits, before requiring retrofits. In the above
efficiency ranges, retrofit does not seem justified unless there is a
local fluoride problem.
1-16
-------
1.4.2 Performance of Recommended Emission Controls
Table 1-5 shows the performance to be expected by application of
the State guidelines for control of emissions from existing primary
aluminum plants. The performances are expressed as average fluoride
emissions, in Column 7; they were calculated with equation 7.1,
using the cell evolution values in Column 6 and the recommended fluoride
collection and removal efficiencies of Columns 2-4. The cell fluoride
evolution values in Column 6 correspond to the largest and smallest
values reported by industry for the given cell types. Therefore, the
average fluoride emission values shown in the last column of Table
1-5 represent expected ranges. For example, the last column shows that
the average emissions from CWPB plants will range from 1.7 - 4.2 Ibs
F/ton Al, provided that all these plants have or install controls equivalent
to 95 percent hooding efficiency and 98.5 percent primary removal efficiency.
The incremental cost effectiveness for the guidelines of Table 1-4
can be inferred from those given in Table 7-9. Secondary control of HSS
and CWPB cells has very high incremental cost effectiveness of $10 - $40
per pound of fluoride removal. Corresponding figures for VSS are $4 -
$8 per pound and are a minimum of $4 per pound for SWPB.
The States should, therefore, be guided by the principles set forth
in this discussion and in Section 8, and should set emission limits with
consideration of-severity of fluoride problem, costs, and any nonair
quality health and environmental impact and energy requirements.
It should be noted that changes in the cell bath ratio, NaF/AlF-,
will change a cell evolution rate, which will change - or tend to change -
the emissions from the fluoride control devices.
1-17
-------
Table 1-5. FLUORIDE EMISSION RANGES CORRESPONDING TO STATE GUIDELINES FOR EXISTING PRIMARY ALUMINUM PLANTS
Cell
Type
VSS
SWPB
HSS
Recommended Efficiencies
for Proposed Retrofits
Primary
Collection
80
80
90
CWPB : 95
; i
Primary
Removal
98.5
98.5
98.5
98.5
i
i
Secondary
Removal
75
75
Assumed Average
Guideline Fluoride Cell
Recommendations Evolution-lb/ton Al
All plants now have best achievable
hooding and primary removal.
Install secondary control, but only
if justified, depending on severity
of fluoride problem.
Install best available hooding
and primary revmoval equipment.
Install secondary control wherever
justified, depending on the severity
of the fluoride problem.
Install best primary control if
needed. All plants but #26 now
have best achievable primary collection
efficiency.
Secondary control does not appear to be
justified, in most locations.
Best control is best hooding and
primary removal equipment.
Install where needed.
Secondary control does not appear to
be justified, in most locations.
30 - 54
37 - 53
28 - 45
26 - 66
Average Fluoride Emission
range, Ib F/ton Al
Control
Primary 6.4* - 11.4
Secondary 1.9** - 3.4
Primary 7.8 - 11.2
Secondary 2.3 - 3.3
Primary 3.2 - 5.1
Secondary 1.1 - 1.7
Primary 1.7 - 4.2
Secondary — —
I
CO
* 30 [ 1 - .80 x .985] = 6.4 Ibs F/ton Al
** 30 [ 1 - .80 x .985 - (1 - .80) x .75] = 1.9 Ibs F/ton Al
-------
1.4.3. Emission Testing
The above guidelines are structured to give the States maximum
flexibility in use of existing emission control and of source sampling
and analytical methods. State regulations are now in place which limit
fluoride emissions from existing primary aluminum plants. These regulations
vary from State to State and differ from Federal new source performance
standards. EPA compliance test reference methods 13(a), 13(b), and 14 were
developed for use with these new source performance standards. However,
installation of Method 14 ductwork on existing plants may be very expensive in
some cases, or may even be precluded because of unadaptability to the
existing roof monitor. Therefore, in order to avoid costly and unnecessary
modifications of roof monitor sampling systems, EPA does not specify
compliance testing for existing plants.
1-19
-------
1.5 ASSESSMENTS
1.5.1 Economic
Control costs might have been derived from Table 1-2, where actual
costs for retrofit emission controls are shown. However, the plant sample
is not large, even though all four cell types are represented. Also, the
costs of greater or lesser degrees of control would be difficult to
estimate. Considerable engineering design, quotations, and drawings would
be required to make these emission level costs consistent in quality with
the actual costs shown in Table 1-2.
The model approach of reference 3 was used to investigate the
effects of various controlled fluoride emission levels on costs. Nineteen
costed control modules were chosen to represent various degrees of emission
collection and control. For example, capital cost modules for a prebake
plant included those for hooding improvement, primary collection system,
and for fluidized bed dry scrubber. The latter two items represent
different degrees of fluoride emissions control and also different costs.
The total retrofit capital cost for a given plant is estimated by determining
the items that have to be improved, installed, dismantled or replaced,
and adding the corresponding control module costs.
The control modules are based on generalized costs applicable to
courtyard control systems. They are also based on typical values of gas
1-20
-------
volumes to primary and secondary controls. In addition, control costs per
unit of production derived by using these modules do not vary significantly
with plant size. These, and other assumptions underlying the modules,
are approximations. In general, direct cost comparison with specific
plants is not justified.
The value of models lies mainly in their ability to evaluate changes,
not to estimate accurate construction costs. The latter requires very
accurate information; the former is less demanding since errors tend to
cancel out when differences are measured. This offsetting of positive
and negative cost errors should reduce the average error as the number of
plants studied increases. The cost differences among levels of control
should be more realistic than absolute cost levels.
Fluoride emission control costs were arrived at by the following
general process:
1. Data shown in Table 7-2 were obtained on fluoride emissions
from the 31 U.S. primary aluminum plants, along with current
fluoride controls at these plants. Cell evolution rates,
primary and secondary loadings, and emissions were obtained
mostly from Section 114 letter responses from plants
representing 86 percent of CWPB capacity and all of the
domestic VSS, HSS, and SWPB capacity.
The initial, or existing controls and corresponding performances
for each plant are illustrated in Table 7-2.
2. Eight plants were taken for study and illustration and
additional controls were added for logical steps from existing
to better control combinations. Specifically, these steps
were to install:
1-21
-------
a. Best primary collection for the cell type, if needed.
b. Best primary removal and water treatment, if needed.
c. Secondary control with scrubber water treatment.
d. Anode bake plant control system and water treatment, if
needed.
The cost results for all eight plants are given in September 1977 dollars
in Table 7-9, part of which is reproduced below as Table 1-6 for purposes
of illustration. Table 7-9 is too limited to allow making general
statements on costs of other plants, but the following tabulation will
serve to illustrate how model costs can be drawn up. For HSS plant 26
as illustrated:
$/Annual Ton Al
Install lime treatment of cryolite
bleed stream $ 2.43
Improve hooding 18.54
Install wet ESP 243.09
Remove spray tower 6.10
Install lime treatment for additional
cryolite bleed stream 2.43
Install spray screen 97.36
Install waste water lime treatment 15.90
$385.85
The capital cost of $386 per annual ton of aluminum results from addition
of all the cost modules, and the annualized cost can be derived in the
same way. Tables 7-4 and 7-7 contain these control module costs. As
Table 1-6 shows, water treatment of cryolite bleed plus improved hooding
1-22
-------
Table 1-6. PRIMARY ALUMINUM CONTROL STRATEGIES
IN3
Plant
Code
26
Emission Control Required for
Hooding Primary (lu)
Poor Spray tower
Average
Fluoride Uni
the Specified Average Fluoride Emission Evolution Capital
Controls Secondary (2°) Controls Ib/ton Al S/ton Al
None 41.6 0
" Install lime treatment of " " 2.43
cryolite bleed stream
Improve Hooding " "
20.97
" " Install wet ESP; remove
spray tower. " " 270.16
" " Install lime treatment of " " 070 ..,„
additional cryolite bleed ///,:»
stream
it ti n
" " Install spray screen and " 386.01
t Costs Emissions Cost-Effectiveness
Annualized Controlled Cumulative Incremental
. S/ton Al IbF/ton Al. $/lb F $/lb F
0 34.5 0 0
1.04 34.5 0 0
5.68 18.4 0.35 0.35
61.13 5.7 2.12 4.37
62.17 5.7 2116 4.45-
101.02 ' 2.5 3.i6- 12.14
waste water lima treatment
-------
costs $2.43 + $18.54 = $20.97 per annual ton, for a decrease in average
fluoride emission of 16.1 Ib F/ton Al. It costs $5.68 to remove 16.1
pounds of fluoride for a cumulative cost of $0.35 per pound of fluoride;
in this case, the incremental cost is also the same. The cost effectiveness
is based on the annualized cost.
1.5.2 Environmental
Table 1-7 gives the environmental impacts of best primary control
and of maximum control. For best primary control, each existing plant is
upgraded to the level of best cell hooding and primary control for that
particular plant. This level is not meant to be the lowest national level
of emission control improvement. However, the second case - that of
maximum or best primary and secondary control - does represent the
highest possible level of improvement. Thus, the environmental impacts
of Table 1-7 bracket the conditions that should result from applying
State guidelines. This table is based on the status of the primary
aluminum domestic plants as of the Spring of 1975.
The table shows the annual removal of particulate emissions; good
control of particulates is necessary for good fluoride removal.
There will be a negligible change in aqueous and solid wastes caused
by adoption of State guidelines.
The data on cell fluoride emissions in the final rows of Table 1-7
indicate that existing plants require much better than current controls.
1.5.3 Energy
Table 1-7 shows the national tonnage of coal that is equivalent to
the energy required for fluoride emission control. National control to
the level of best hooding and primary control will increase fluoride
control energy expenditures by as little as 120,000 megawatt hours per
1-24
-------
Table 1-7. ENVIRONMENTAL IMPACT OF BEST CONTROL FOR TOTAL FLUORIDE EMISSIONS FROM THE
PRIMARY ALUMINUM INDUSTRY
Fluoride Emissions Control Level
no
en
National Impacts
National Fluoride Air Emission (tons F/yr)
National Particulate Emissions (tons F/yr)
National Water Effluent Emissions
(tons/yr)
Fluoride
Total Suspended Solids
National Fluoride Control Energy Requirements
(Mwh/yr)
National Solid Wastes from Fluoride Control
(tons/yr)
Bituminous Coal Used for National Fluoride
Control (tons/yr)
Average Fluoride Emissions, Ib F/ton Al
cell: CWPB
SWPB
VSS
HSS
Existing
18,000
44,000
110
210
1,470,000
160,000
612,000
6.3
13.0
5.2
5.7
Best Hooding & Primary
Control
9,000
25,000
85
170
1,590,000
130,000
662,000
2.7
4.0
5.2
4.0
Best Primary & Secondary
Control
4,000
21,000
260
510
2,900,000
220,000
1,210,000
1.2
2.2
2.1
1.7
-------
year over existing control. This energy is equivalent to the generation
of an additional 14 megawatts of electrical power nationally, or to an
average incremental fluoride control power expense of 0.44 megawatts
per plant. In comparison, an average of 283 megawatts of electrolytic
power is required per domestic primary aluminum plant.
1.6 COMPLIANCE TIMES
The compliance times for the installation of a typical fluoride
emission control system at a primary aluminum plant is shown in Table 1-8.
This table represents a moderately complicated case involving the
improvement of hooding and the installation of fluidized bed dry scrubbers.
Such a project involves not only the installation of the very large dry
scrubbers (see Figures 6-13 and 6-14) but also of storage and surge tanks,
conveyors, fans and long lengths of huge ductwork with associated foundations,
structural steelwork, and electrical drive systems.
As explained in Section 6.5, and indicated above, the large amount
of material procurement causes an overlap of most design and construction
activities. At a given time, numerous items are in various stages of
design, procurement, and construction. For this reason, contracts for
major items cannot be awarded simultaneously, but are made over a range
of time as shown in Table 1-8.
One important step that is almost wholly beyond the control of the
customer or the control official is the delivery time. Table 1-8 is
based on delivery times as of the summer of 1974. The time required from
order to delivery increased greatly from 1973 to 1974 and passed all
former bounds for certain items used for adding retrofit controls at
1-26
-------
primary aluminum plants. Thus, delivery reports gave 35-50 weeks for
electrical switchgear, 25-60 weeks for fans and 35-65 weeks for
electrical motors. Deliveries can depend partly upon quantity ordered,
continuity of business through the years, and most favored customer
statue.
Table 1-8. INCREMENTS OF PROGRESS FOR INSTALLATION OF FLUORIDE
EMISSION CONTROLS IN AN EXISTING PRIMARY ALUMINUM PLANT
Increments of progress Elapsed time - weeks
Preliminary control plan and compliance
schedule to appropriate agency 25
Award of major contracts 35 - 55
Start of construction 60
Completion of construction 124
Final compliance 130
Table 1-9 gives the time required to retrofit eight actual primary
aluminum plants. In each case, the whole plant was retrofitted, except for
plant F. Only plants C, E, and H had secondary control originally and
only plant E improved its secondary control, at a cost of about 65 percent
of its total retrofit expenditure. Plant B built and operated a pilot
plant during two of the 4-1/2 years of retrofit activity. The completion
time of 5-1/2 years for plant G includes 3 years for improved cell hooding
and 3 years for dry scrubber installation. The 3 years for improved
hooding is due to a claimed economic advantage for modifying cells over
1-27
-------
Table 1-9 . TOTAL CONSTRUCTION TIME FOR RETROFIT EMISSION CONTROLS FOR
PRIMARY ALUMINUM PLANTS
Plant
Code
A
B
C
D
E
F
G
H
Cell
Type
HSS
HSS
SWPB
CWPB
VSS
Cl/PB
CURB
CWPB
Plant Capa-
city
(tons/yr)
80,000
210,000
265,000
115,000
91,000
32,850
250,000
130,000
Retrofit Capital
Cost ($)
11,300,000
19,300,000
14,300,000
11,800,000
5,800,000
1,800,000
31,000,000
28,000,000
Description of
Retrofit Emission Controls
Dry scrubbers-primary
Improved cell hooding and wet
ESP-primary
New cell hooding and dry
scrubbers -primary
Dry scrubbers-primary
Wet ESP-primary
Dormer tunnel -secondary
Time Required
for Retrofitting
(years)
2-1/2
4-1/2
3
1
3
2-1/2
Dry scrubbers-primary 1-1/2
(1 potline) !
Improved cell hooding and dry
scrubbers-primary
Improved cell hooding and
dry scrubbers-primary
5-1/2
i
' 2
r\j
oo
-------
the normal 3-year life of their cathode linings. Had plant G so
elected, the dry scrubber installation could have proceeded simultaneously
with cell hooding improvements, reducing the completion time to about 3
years.
The actual time requirements shown in Table 1-9 could probably have
been decreased if there had, at the time, been a requirement or incentive
to do so. With no capital return, the usual requirements for haste was
reduced, except for any time when production could have been held up.
In view of the above discussion, a reasonable total time for
retrofitting emission controls to a primary aluminum plant may be taken
as about 2-1/2 years. Because of the changing situation on equipment
manufacturing and delivery times, however, it is recommended that
enforcement officials consider each plant on a case-by-case basis. They
should require proof for the time requirements claimed for each milestone
shown in Table 1-8. Additional time allowance may be made if it takes longer
than indicated in Table 1-8 to reach compliance after completion of construction.
1-29
-------
1.7 REFERENCES FOR SECTION 1
1. Varner, Bruce A. and Crane, George B. Actual Costs for Retrofit of
Fluoride Emission Controls to Existing Primary Aluminum Plants.
Paper presented at 83rd National Meeting American Institute of
Chemical Engineers, Houston, Texas. March 20-24, 1977.
2. Varner, Bruce A. and Crane, George B. Estimation of Primary Collection
Efficiency for New or Retrofit Hooding for Primary Aluminum Cells.
Paper presented at 69th Annual Meeting of the Air Pollution Control
Association, Portland, Oregon, June 27-July 1, 1976.
3. Air Pollution Control in the Primary Aluminum Industry. Singmaster
and Breyer, New York, New York. Prepared for Office of Air Programs,
Environmental Protection Agency, Research Triangle Park, North
Carolina, under Contract Number CPA 70-21. July 1973.
l-3u
-------
2. HEALTH AND WELFARE EFFECTS OF FLUORIDES
2.1 INTRODUCTION
In accordance with 40 CFR 60.22(b), promulgated on November 17, 1975
(40 FR 53340), this chapter presents a summary of the available infor-
mation on the potential health and welfare effects of fluorides and the
rationale for the Administrator's determination that it is a welfare-
related pollutant for purposes of section lll(d) of the Clean Air Act.
The Administrator first considers potential health and welfare
effects of a designated pollutant in connection with the establishment
of standards of performance for new sources of that pollutant under
section lll(b) of the Act. Before such standards may be established,
the Administrator must find that the pollutant in question "may contribute
significantly to air pollution which causes or contributes to the endanger-
ment of public health or welfare" [see section 111(b)(l)(a)]. Because
this finding is, in effect, a prerequisite to the same pollutant's being
identified as a designated pollutant under section lll(d), all designated
pollutants will have been found to have potential adverse effects on
public health, public welfare, or both.
As discussed in section 1.1 above, Subpart B of Part 60 distinguishes
between designated pollutants that may cause or contribute to endangerment
of public health (referred to as "health-related pollutants") and those
for which adverse effects on public health have not been demonstrated
("welfare-related pollutants"). In general, the significance of the
distinction is that States have more flexibility in establishing plans
for the control of welfare-related pollutants than is provided for
plans involving health-related pollutants.
2-1
-------
In determining whether a designated pollutant is health-related
or welfare-related for purposes of section lll(d), the Administrator
considers such factors as: (1) Known and suspected effects of the
pollutant on public health and welfare; (2) potential ambient concentrations
of the pollutant; (3) generation of any secondary pollutants for which
the designated pollutant may be a precursor; (4) any synergistic effect
with other pollutants; and (5) potential effects from accumulation in
the environment (e.g., soil, water and food chains).
It should be noted that the Administrator's determination whether
a designated pollutant is health-related or welfare-related for purposes
of section lll(d) does not affect the degree of control represented by
EPA's emission guidelines. For reasons discussed in the preamble to
Subpart B, EPA's emission guidelines [like standards of performance for
new sources under section lll(b)] are based on the degree of control
achievable with the best adequately demonstrated control systems (considering
costs), rather than on direct protection of public health or welfare. This
is true whether a particular designated pollutant has been found to be
health-related or welfare-related. Thus, the only consequence of that
finding is the degree of flexibility that will be available to the States
in establishing plans for control of the pollutant, as indicated above.
2.2 EFFECT OF FLUORIDES ON HUMAN HEALTH1
2.2.1 Atmospheric Fluorides
The daily intake of fluoride inhaled from the ambient air is only a
few hundredths of a milligram - a very small fraction of the total intake
for the average person. If a person is exposed to ambient air containing
about 8 micrograms (yg) of fluoride per cubic meter, which is the maximum
average concentration that is projected in the vicinity of a primary
2-2
-------
aluminum facility with only mediocre control equipment (Table 9-5), his total
daily intake from this source is calculated to be about 150 yg. This
is very low compared with the estimated daily intake of about 1200 yg
from food, water, and other sources for the average person.
Few instances of health effects in people have been attributed
to community airborne fluoride, and they occurred in investigations
of the health of persons living in the immediate vicinity of fluoride-
emitting industries. The only effects consistently observed are
decreased tooth decay and slight mottling of tooth enamel when compared
to control community observations. Crippling fluorosis resulting from
industrial exposure to fluoride seldom (if ever) occurs today, owing
to the establishment of and adherence to threshold limits for exposure
of workers to fluoride. It has never been seen in the United States.
Even persons occupationally exposed to airborne fluoride do not usually
come in contact with fluoride concentrations exceeding the recommended
industrial threshold limit values (TLV). The current TLV for hydrogen
fluoride is 3 parts per million (ppm) while that for particulate
fluoride is 2.5 milligrams per cubic meter (mg/m ) expressed as elemental
fluorine.
There is evidence that airborne fluoride concentrations that
produce no plant injury contribute quantities of fluoride that are
negligible in terms of possible adverse effects on human health and
offer a satisfactory margin of protection for people.
Gaseous hydrogen fluoride is absorbed from the respiratory tract
and through the skin. Fluoride retained in the body is found almost
entirely in the bones and teeth. Under normal conditions, atmospheric
2-3
-------
fluoride represents only a very small portion of the body fluoride
burden.
2.2.2 Ingested Fluorides
Many careful studies, which were reviewed by the National Academy
of Sciences, have been made of human populations living in the vicinity
of large stationary sources of fluoride emissions. Even in situations
where poisoning of grazing animals was present, no human illness due
to fluoride poisoning has been found. In some of these areas much of
the food used by the people was locally produced. Selection, processing,
and cooking of vegetables, grains and fruits gives a much lower fluoride
intake in human diets than in that of animals grazing on contaminated
pasture.
In poisoned animals, fluorine levels are several thousand times
normal in bone, and barely twice normal in milk or meat. Calves and
lambs nursing from poisoned mothers do not have fluorosis. They do not
develop poisoning until they begin to graze. Meat, milk and eggs from
local animals contain very little more fluoride than the same foods
from unpoisoned animals. This is due to the fact that fluorine is
deposited in the bones almost entirely.
2.3 EFFECT OF FLUORIDES ON ANIMALS1
In areas where fluoride air pollution is a problem, high-fluoride
vegetation is the major source of fluoride intake by livestock.
Inhalation contributes only a negligible amount to the total fluoride
intake of such animals.
The available evidence indicates that dairy cattle are the
domestic animals most sensitive to fluorides, and protection of dairy
cattle from adverse effects will protect other classes of livestock.
2-4
-------
Ingestion of fluoride from hay and forage causes bone lesions,
lameness, and impairment of appetite that can result in decreased
weight gain or diminished milk yield. It can also affect developing
teeth in young animals, causing more or less severe abnormalities in
permanent teeth.
Experiments have indicated that long-term ingestion of 40 ppm
or more of fluoride in the ration of dairy cattle will produce a
significant incidence of lameness, bone lesions, and dental fluorosis,
along with an effect on growth and milk production. Continual inges-
tion of a ration containing less than 40 ppm will give discernible
but nondamaging effects. However, full protection requires that a
time limit be placed on the period during which high intakes can be
tolerated.
It has been suggested that dairy cattle can tolerate the ingestion
of forage that averages 40 ppm of fluoride for a year, 60 ppm for up to
2 months and 80 ppm for up to 1 month. The usual food supplements are
low in fluoride and will reduce the fluoride concentration of the total
ration to the extent that they are fed.
Fluoride-containing dusts can be non-injurious to vegetation but
contain hazardous amounts of fluoride in terms of forage for farm
animals. Phosphate rock is an example of a dust that seemingly has
not injured plants but is injurious to farm animals. This was made
evident forty years ago when an attempt was made to feed phosphate
rock as a dietary supplement source of calcium and phosphate. Fluoride
2
injury quickly became apparent. Phosphate rock is used for this
purpose today, but only after defluorinating by heat treatment. Phos-
phate rock typically contains up to about four weight percent fluorine.
2-5
-------
2.4 EFFECT OF ATMOSPHERIC FLUORIDES ON VEGETATION1' 3
The previous sections state that atmospheric fluorides are not a
direct problem to people or animals in the United States, but that
animals could be seriously harmed by injestion of fluoride from forage.
Indeed, the more important aspect of fluoride in the ambient air is
its effect on vegetation and its accumulation in forage that leads to
harmful effects in cattle and other animals. The hazard to these
receptors is limited to particular areas: industrial sources having
poorly controlled fluoride emissions and farms located in close
proximity to facilities emitting fluorides.
Exposure of plants to atmospheric fluorides can result in accumu-
lation, foliar lesions, and alteration in plant development, growth,
and yield. According to their response to fluorides, plants may be
classed as sensitive, intermediate, and resistant. Sensitive plants
include several conifers, several fruits and berries, and some grasses
such as sweet corn and sorghum. Resistant plants include several
deciduous trees and numerous vegetable and field crops. Most forage
crops are tolerant or only moderately susceptible. In addition to
differences among species and varieties, the duration of exposure,
stage of development and rate of growth, and the environmental condi-
tions and agricultural practices are important factors in determining
the susceptibility of plants to fluorides.
The average concentration of fluoride in or on foliage that appears
to be important for animals is 40 ppm. The available data suggest
that a threshold for significant foliar necrosis on sensitive species,
or an accumulation of fluoride in forage of more than 40 ppm would
2-6
-------
result from exposure to a 30-day average air concentration of gaseous
fluoride of about 0.5 micrograms per cubic meter (yg/m ).
Examples of plant fluoride exposures that relate to leaf damage
2
and crop reduction are shown in Table 2-1. As shown, all varieties
of sorghum and the less resistant varieties of corn and tomatoes are
particularly susceptible to damage by fluoride ambient air concentra-
tions projected in the immediate vicinity of primary aluminum facilities
(See Table 9-5).
2.5 THE EFFECT OF ATMOSPHERIC FLUORIDES ON MATERIALS OF CONSTRUCTION
2
2.5.1 Etching of Glass
It is well known that glass and other high-silica materials are
etched by exposure to volatile fluorides like HF and SiF.. Some
experiments have been performed where panes of glass were fumigated
with HF in chambers. Definite etching resulted from nine hours
exposure at a level of 590 ppb. Pronounced etching
resulted 14.5 hours exposure at 790 ppb. Such levels would,
of course, cause extensive damage to many species of vegetation.
However, ambient concentrations of this magnitude are improbable
provided that a aluminum facility properly maintains and operates
some type of control equipment for abating fluoride emissions.
2.5.2 Effects of Fluorides on Structures
At the relatively low gaseous concentrations of fluorides in
emissions from industrial processes, 1000 ppm or less, the damage
caused by fluorides is probably limited mostly to glass and brick.
Occasionally, damage to the interior brick lining of a stack has
been attributed to fluorides emitted in an industrial process.
2-7
-------
Table 2-1. EXAMPLES OF HF CONCENTRATIONS AND EXPOSURE DURATIONS REPORTED
TO CAUSE LEAF DAMAGE AND POTENTIAL REDUCTION IN CROP VALUES2
p1ant Concentration and Time
Most sensitive varieties - most resistant varieties
Sorghum 0.7 ppb for 15 days - 15 ppb for 3 days
«>o Corn 2 ppb for 10 days - 800 ppb for 4 hours
Tomato 10 ppb for 100 days - 700 ppb for 6 days
100 ppb for 120 days - 700 ppb for 10 days
00
-------
Fluoride damage occurs to the high silica brick used in the
furnaces for baking carbon anodes for aluminum reduction cells.
2.6 RATIONALE
Based :n the information provided the preceding sections of
Chapter 2, it is clear that fluoride emissions from primary aluminum
plants have no significant effect on human health. Fluoride
emissions, however, do have adverse effects on livestock and vege-
tation. Therefore the Administrator has concluded that fluoride
emissions from primary aluminum plants do not contribute to the
endangerment of public health. Thus, fluoride emissions will be
considered a welfare-related pollutant for purposes of section lll(d)
and Subpart B of Part 60.
2.7 REFERENCES FOR SECTION 2
1. Biologic Effects of Atmospheric Pollutants: Fluorides. National
Academy of Sciences, Washington, D.C. Prepared for Environmental
Protection Agency, Durham, NC, under Contract Number CPA 70-42.
1971.
2. Robinson, J. M. et al. Engineering and Cost Effectiveness Study
of Fluoride Emissions Control. Resources Research, Inc. and TRW
Systems Group, McLean, VA. Prepared for Office of Air Programs,
Environmental Protection Agency, Durham, NC, under Contract
Number EHSD 71-14. January 1972.
3. Carlson, C. E. and Dewey, J. E. Environmental Pollution by
Fluorides in Flathead National Forest and Glacier National Park.
U.S. Department of Agriculture - Forest Service. Missoula,
Montana. October 1971.
4. Peletti, E. Corrosion and Materials of Construction. In:
Phosphoric Acid, Volume I, Slack, A. V. (ed). New York, Marcel
Dekker, Inc., 1978. p. 779-884.
2.9
-------
3. U. S. PRIMARY ALUMINUM MANUFACTURING STATISTICS
3.1 EXISTING PLANTS
3.1.1 Introduction
Aluminum ranks first in production among all nonferrous metals
produced in the United States.1 The United States is the world's leading
producer of primary aluminum. Primary aluminum is produced by electro-
lytic reduction of alumina which in turn is produced by refining bauxite
ore; secondary aluminum is produced by re-working aluminum scrap.
Primary production in the U. S. in 1977 totalled 4.54 million short tons
2
compared with an estimated world total of 15.05 million short tons.
U.S. production reached a record high in 1974 of 4.90 million short tons.
Primary capacity in the U.S. at the end of 1977 was estimated at
3 4
5.19 million short tons and was accounted for by 31 plants. At
the end of 1977, there were 12 U.S. primary producers; about 65
percent of primary capacity was accounted for by three producers—
3
Alcoa, Reynolds, and Kaiser.
Aluminum produced by electrolytic reduction of alumina has an
average purity of about 99.5 percent. The largest market for aluminum
in 1977 was the building and construction field (23.1 percent), fol-
lowed by transportation (21.7 percent), containers and packaging (20.8
percent), electrical (10.0 percent), consumer durables (7.9 percent),
other - primarily for defense (4.2 percent), machinery and equipment
(6.9 percent), and exports (5.4 percent). Further refining of aluminum
can produce "super purity" aluminum, which is 99.99 percent pure. This
grade of aluminum is used as a catalyst carrier in making high octane
gasoline, for forming jewelry, and, in the form of foil, for the electronics
industry.
-------
Molten primary aluminum may be shipped directly to a customer's
plant in insulated ladles. More commonly, it is cast into ingots or
billets of varying shapes and sizes, sometimes after being alloyed.
These may be fabricated at the reduction plant site or shipped to
another site for fabrication by the primary producer or an independent
fabricator.
Fabricating may consist of: rolling the ingot into plate, sheet,
and foile; forging into special shapes; drawing the ingot into rod, bar,
wire, and drawn tube, extruding billets into tubings, rod, bar, and
special shapes; or melting ingot and atomizing into powder, paste,
and flake. Molten aluminum may also be cast, although casting proudcers
use mostly secondary aluminum for this purpose.
In 1977, 6.68 million short tons of aluminum were shipped from U.S.
primary and secondary producers at a value of $13.3 billion. In 1976,
shipments totalled 6.37 million short tons at an estimated val,ue of
$11.22 billion.6
3.1.2 Location and Size
The 31 U.S. primary aluminum plants producing alumina by electro-
lytic reduction of alumina at the end of 1977 are listed in Table
3 7-14-
3-1. * The list includes location, company ownership, and
annual capacity by the type of reduction cell in use. Table 3-1
shows that plant capacities range from 35,000 to 285^000 short tons
per year and that 20 of the 31 plants exclusively or primarily use
center-worked or side-worked prebake reduction cells. The horizontal
Soderberg reduction cell is the second most popular choice.
3-2
-------
Table 3-1. U. S. PRIMARY ALUMINUM PLANTS AND CAPACITIES, 19743' 7"14
(short tons)
State and City
Alabama
Scottsboro
Sheffield
Arkansas
Arkadelphia
Jones Mills
Indiana
Evansville
Kentucky
hawesvi lie
Sebree
Louisiana
Chalmette
Lake Charles
Maryland
Frederick
Company
Revere Copper & Brass, Inc.
Reynolds Metals Co.
Reynolds Metals Co.
Reynolds Metals Co.
Aluminum Co. of America
National-Southwire Aluminum Co.n
The Anaconda Aluminum Co.
Kaiser Aluminum & Chemical Corp.
Consolidated Aluminum Corp.
Eastalco Aluminum Co.
Annual capacity by cell type3'
CWPB
1 ,300C
125,000
280,000
180,000
120,000
SWPB
112,000
15,700C
35,000
174,0006
VSS
HSS
202,000
51,000°
260,000
CO
GO
-------
Table 3-1(continued). U. S. PRIMARY ALUMINUM PLANTS AND CAPACITIES, 1974
State and City
Missouri
New Madrid
Montana
Columbia Falls
New York
Massena
Massena
North Carolina
Badin
Ohio
Hannibal
Oregon
The Dalles
Troutdale
Tennessee
Alcoa
Company
Noranda Aluminum, Inc.
The Anaconda Aluminum Co.
Aluminum Co. of America
Reynolds Metals Co.
Aluminum Co. of America
Ormet Corp . '
Martin Marietta Aluminum Inc.
Reynolds Metals Co.
Aluminum Co. of America
Annual capacity by cell type9'
CWPB
70,000
135,000
120,000
250,000
130,000
210,000
SWPB
VSS
180,000
90,000
60,000
HSS
126,000
OJ
-------
Table 3-1 (continued). U. S. PRIMARY ALUMINUM PLANTS AND CAPACITIES, 1974
State and City
New Johnsonville
Texas
Corpus Christi
Point Comfort
Rockdale
Washington
Ferndale
Goldendale
Longview
Head
T a coma
Vancouver
w'enatchee
West Virginia
Ravenswood
Totals
Company
Consolidated Aluminum Corp,
Reynolds Metals Co.
Aluminum Co. of America
Aluminum Co. of America
Intalco Aluminum Corp.
Martin Marietta Aluminum Inc.
Reynolds Metals Co.
Kaiser Aluminum & Chemical Corp.
Kaiser Aluminum & Chemical Corp.
Aluminum Co. of America
Aluminum Co. of America
Kaiser Aluminum & Chemical Corp. •
Annual capacity by cell type9'
CWPB
285,000
206,000
115,000
180,000
163,000
2,570,300
SWPB i VSS
140,000
260,0009
649,700
185,000
HSS
115,000
i
120,000
635,000
210,000
81 ,000
1,045,000
GO
en
Total - all cell types
4,900,000
-------
Table 3-1 (continued). U.S. PRIMARY ALUMINUM PLANTS AND CAPACITIES, 1974
FOOTNOTES:
aThe design capacity of a reduction plant is not an exact figure; rather it is an estimated rate
that is often below the level at which a plant can actually produce. For this reason, some capacity
figures are more realistic than others.
3CWPB— center-worked prebake cells, SWPB-
and HSS--horizontal stud Soderberg cells.
:Based i
cells.
CWPB--center-worked prebake cells, SWPB—side-worked prebake cells, VSS—vertical stud Soderberg cells.
cBased on a total plant capacity of 68,000 short tons per year; 120 HSS cells; 37 SWPB cells; 3 CWPB
Jointly owned by National Aluminum Company and Southwire Company.
01 e!00 percent interest by Howmet Corporation.
Jointly owned by Revere Copper & Brass, Inc
950 percent interest by Howmet Corporation; 50 percent interest by Amax Aluminum Company.
Jointly owned by Revere Copper & Brass, Inc. and Consolidated Aluminum Corporation.
-------
In selecting sites for primary aluminum reduction plants,
producers have had to consider several factors that affect production
cost. Three principal factors are:
1. Costs for shipping the major raw material, alumina, to the
reduction plant site,
2. Electrical energy costs for reducing alumina to aluminum
in the reduction cells;, and
3. Costs for shipping aluminum to the fabricator or to the
market.
Alumina consumed by U. S. primary producers is manufactured in
the south central U. S., in South America, and in the West Indies and
Australia. Although principal fabricator sites are found in the
Northeast and the Midwest and in California, the availability of
low cost hydroelectric power has been the overriding factor in plant site
selection and has resulted in the location of plants in such areas as the
18
Pacific Northwest and the Tennessee River Valley. Moderate cost steam-
generated power has attracted several plants to the Ohio River Valley,
closer to the Northeast and Midwest markets.
Most of the aluminum reduction plants in the United States are
located in predominantly rural areas with a sparse population density,
as estimated by the total population of towns or cities within a 10-mile
radius of each plant given in 1960 or later census figures.
Table 3-2 shows the distribution of plants with respect to
population.
3-7
-------
Table 3-2. DISTRIBUTION OF PLANTS BY POPULATION, 1971
19
Number
of plants
13
9
2
7a
Percent
capaci ty
41.1
28.7
5.7
24.5
Surrounding 300-mi> area
Population
Less than 10,000
10-25,000
25-50,000
More than 50,000
Population/mi^
Less than 32
32-80
80-160
More than 160
One plant is surrounded by residential sections in an urban
community. The other six plants in the high density areas
are located on the outskirts of medium sized communities where
the surrounding land is utilized for dairy farming or truck
farming.
The distribution of plant capacity with respect to the type of
surrounding land use is shown in Table 3-3.
3.2 FUTURE TRENDS
3.2.1 Domestic Industry and Plant Growth
Primary aluminum production in the U. S. started in 1888. Table 3-4
shows the yearly U. S.-primary aluminum production since 1893, the date
of the-earliest recorded data. Primary aluminum production has'experienced
a steady, but somewhat cyclical, growth pattern. Average annual produc-
tion growth rates have been: 14;5 percent for'the period 1893 to 1973:,
9.5 percent for the quarter century 1946 to'1971; 7V5 percent for"
the decade 1961 to 1971 ; and 6.9 percent for the decade 1963 to .1973.'
The cyclical growth pattern has given rise to periods of excess
capacity. For the past few years, the industry has been in a period of
low profitability because of excess capacity, but indications are that
excess capacity and inventories are being eliminated. This should lead
22
to higher prices and greater profits.
3-8
20,21
-------
Table 3-3. ESTIMATED DISTRIBUTION OF PLANTS BY ENVIRONMENT
19
Environmental category
Urban
Orchard growing
Dairy farming
Truck farming
Cattle raising
Lumbering
General agriculture
Dairy plus truck farming
Dairy plus cattle
Dairy plus agriculture
Dairy plus lumber
Truck farming plus cattle
Truck farming plus lumber
Truck farming plus general
agriculture
Lumber plus general
agriculture
Truck farming plus
cattle plus general
agriculture
Total
i
\ Number
; of
| plants
j -
!
i «
i 3
i «
\ 1
j
i '
\
i
; 2
] _ 2
i
i i
1
i -1
i 1
1 1
i 4
; 1
i
i 1
i
i
1
! 3
! 31
Percent of total U. S.
aluminum capacity in
envi ronmental category
5.5
15.4
9.4
10.9
1.9
5.6
2.4
6.3
4.4
3.0
1.7
3.0
11.3
3.5
4.7
11.0
100.0
3-9
-------
Table 3-4. PRODUCTION OF PRIMARY ALUMINUM IN THE UNITED STATES
20
Year
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906a
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
Production,
IbxlO6
0.2
0.5
0.5
1.0
2.4
3.0
3.3
5.1
5.8
5.8
6.6
8.1
10.8
14.1
16.3
10.7
29.1
35.4
38.4
41.8
47.3
58.0
90.5
115.1
129.9
124.7
128.5
138.0
Year
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
Production,
lbxlQ6
54.5
73.6
128.5
150.6
140.1
147.4
163.6
210.5
228.0
229.0
177.5
104.9
85.1
74.2
119.3
224.9
292.7
286.9
327.1
412.6
618.1
1,042.2
1,840.4
1,552.9
990.1
819.3
1,143.5
1,246.9
1,206.9
1,437.2
Year
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
Production,
lbx!06
1,673.8
1,874.7
2,504.0
2,921.1
3,131.4
3,357.9
3,295.4
3,131.1
3,908.2
4,029.0
3,807.4
4,235.9
4,625.1
5,105.5
5,509.0
5,936.7
6,538.5
6,510.1
7,586.1
7,952.3
7,850.4
8,244.8
9,058.2
9,806.9
7,758.3
8,512.8
9,077.4
Data prior to 1907 represent fiscal years ending August 31
during last 4 months of 1906 totaled 5.4 million pounds.
Production
3-10
-------
At the end of 1972, U. S. primary capacity was 4.8 nillion short
tons, and it has been estimated that domestic aluminum needs will
require a capacity increase to 7.2 million short tons by 1980.
(The capacity increase will not necessarily be all domestic.) This
23
estimate is based on the following four assumptions:
1. Growth in demand averaging 7.5 percent per year between
1971 and 1980.
2. Mill imports, secondary recovery, and primary imports at
the same percentage level in 1980 as in 1972.
3. Stockpile sales of 77,000 tons per year in 1980, based on the
current General Services Administration disposal schedule.
4. Industry operating ratio of 95 percent.
The anticipated increase represents an average annual capacity
growth rate of 5.2 percent for the period 1972 to 1980. Growth in
consumption of aluminum has generally been quoted at 6 to 8 percent
over the period 1972 to 1980. The fact that capacity growth is
forecast to be less than consumption growth reflects the present
excess capacity. Applying a 5.2 percent growth rate to capacity for
the period 1980 to 1985 results in a capacity increase to 9.2 million
short tons by 1985. Applying a 7 percent growth rate results in a
capacity increase to 10.1 million short tons by 1985.
The EPA standards of performance for new primary aluminum plants
are not expected to have an adverse impact upon future growth in the
25
domestic primary aluminum industry.
3-11
-------
Table 3-5 shows the gradual increase in the average capacity of
primary aluminum plants in the U.S. since I960.3'4'26 Although the
average capacity stood at 158,000 short tons per year in 1973, the
seven plants built from 1970 to 1973 have capacities of 35,000;
70,000; 87,000; 112,000; 120,000; 120,000; and 180,000 short tons per
year. The average capacity for these seven plants is thus only 104,000
short tons per year.
Table 3-5. GROWTH IN PRIMARY ALUMINUM PLANT AVERAGE CAPACITY IN U.S.3'4'26
Year
1960
1967
1973
Number
of plants
22
24
31
Annual capacity, short tons x lO^/year
Total
2468.8
3321.0
4893.0
Average
112.2
138.4
157.8
As for future changes in average plant size, no new plants are
known to be under construction as of the end of 1974. A lead time of
about 3 years is needed for construction of a new plant.23 Hence, if
the anticipated growth rate is to be met by increased U.S. capacity,
it seems likely that much of the immediate capacity increases will
probably be in the form of expansions, thus further increasing the
average plant capacity.
3-12
-------
3.2.2 Plant Location and Cell Type
No pronounced geographical shifts in the location of primary
aluminum plants have occurred since 1966. Plants continue to be
located mainly near sources of low-cost power in the Pacific
Northwest and the Tennessee and Ohio River Valleys; however, most
of the available hydroelectric power in the U.S. has now been
harnessed. Greater use of nuclear and nonnuclear steam-generated
power could result in any future U.S. primary aluminum plants beino
located closer to the marketplace.
Table 3-6, which summarizes U. S. trends in cell design since World
%
War II, shows that the primary aluminum industry favored adoption of
Soderberg cells during the 1950s but that the present trend is toward
prebake cells. This trend is expected to continue since new prebake
cells should be able to meet EPA standards of performance for new
primary aluminum plants (see Section 1.3) more easily and less
28
expensively than new Soderberg plants.
3.2.3 New Producers and Technology
It appears that unless government regulation intervenes, aluminum
29
producers may become part of multi-material industries by the mid-1980s.
Aluminum would most likely become allied with two of its strongest competitors,
steel and plastic. One steel company has already become a primary
aluminum producer. Also, some aluminum fabricators have accomplished or
are considering backward integration into primary aluminum. Furthermore,
there appears to be a trend for foreign producers to establish primary
plants in the U. S. to gain easy access to U. S. markets.
3-13
-------
Table 3-6. U. S. TRENDS IN ADOPTION OF CELL TYPES
27
pre - 1946
1946
1947 - 1950
1951 - 1958
1959 - 1965
1966 - 1969
1970 - 1972
Prebake plants
Installations
New
-
-3a
lb
4
1
1
6
Cumulative
10
7
8
12
13
14
20
Expansions
-
-
-
2
-
1
2C
Soderberg
Horizontal stud
Installations
New
-
-
-
3
1
-
-
Cumulative
2
2
2
5
6
6
6
Expansions
»
-
-
2
-
2
-
plants
Vertical stud
Installations
New
-
-
-
3
-
-
1
Cumulative
-
-
-
3
3
3
4
Exnansinns
-
-
-
-
-
2
-
GO
I
Government-owned (DPC) plants deactivated.
5One DPC plant reactivated under private ownership.
•»
'Estimated.
-------
From its inception over 80 years ago, the U. S. primary aluminum
industry has relied upon the Bayer process for refining bauxite ore
and the Hall-Heroult process for electrolytically reducing alumina to
aluminum. The Bayer process discharges air and water pollutants;
the Hall process discharges air (notably fluoride) and sometimes water
pollutants. The Hall process uses huge amounts of electricity and
requires huge capital investments. The Bayer-Hall processes have made
the U. S. largely dependent on foreign sources of bauxite ore. In spite
of these disadvantages, technology has not changed because alternate
processes have shown higher production costs than the Bayer-Hall
processes and because bauxite ore has been in adequate supply. Improve-
ments in the Bayer-Hall processes have made it more difficult for
alternate processes to compete economically. However, it now appears
that it may not be long before other processes can compete, and the day
will come when available bauxite will no longer support the world's demand
for aluminum.
Efforts to find alternate processes have generally fallen into
three classes:
1. Production of alumina from non-bauxite ores.
2. Direct reduction of bauxite or non-bauxite ore to aluminum.
3. Conversion of alumina to aluminum by means other than
electrolytic reduction in a cryolite bath.
Non-bauxite ores that have been considered for commercial development
in the U. S. include: high-alumina clay; dawsonite; aluminous shales;
alunite; aluminum phosphate rock; igneous rock, notably anorthosite; and
saprolite- and sillimamte-group minerals. The Poles are presently
3-15
-------
using high-alumina clay commercially and the Russians are using alunite
and an alumina-containing igneous rock known as nepheline syenite. It
appears that high-alumina clay is the most likely candidate for
31
commercial development in the U. S.
U. S. and foreign producers have experimented with various
methods of directly reducing bauxite or non-bauxite ore to aluminum.
Applied Aluminum Research Corp.* has announced that it is developing
the Toth Process by which non-bauxitic ores are directly reduced to
aluminum. The process involves conversion of the ore to aluminum
chloride, purification, reaction of aluminum chloride with manganese
to produce aluminum and manganese chloride, and regeneration of the
manganese.
Alcoa has announced the development of a process involving the
conversion of conventionally made alumina to aluminum trichloride
and subsequent electrolytic reduction to yield aluminum metal and
32
recyclable chlorine. This method, known as the Alcoa Smelting
Process, reduces electric power requirements by 30 percent.
3.3 PRICE STATISTICS
Table 3-7 gives average list prices of virgin primary aluminum
33 34
ingot for selected years from 1930 to 1969 » and closing New York
1 35
cash prices for selected dates from August 1970 to February 1975. '
*Mention of specific companies or products in this document does not
constitute endorsement by the U.S. Environmental Protection Agency.
3-16
-------
Table 3-7. PRIMARY ALUMINUM INGOT PRICE HISTORY
(cents per pound)
Year
1930
1932
1937
1941
1944
1947
1950
1951
1953
1957
1959
1961
1964
1967
1968
1969
Average
List Price
23.8
23.3
20.1
16.5
15.0
15.0
17.7
19.0
20.9
27.5
26.9
25.5
23.7
25.0
25.6
27.2
Date
8/13/70
8/13/71
10/18/72
12/31/72
3/25/73
5/29/73
8/1/73
10/18/73
12/31/73
3/25/74
5/29/74
7/31/74
8/1/74
10/74
12/74
2/25/75
Closing New York
Cash Price
29.0
29.0
25.0
25.0
25.0
25.0
25.0
25.0
29.0
29.0
31.5
33.5
36.0
39.0
39.0
39.0
The table shows that the price increased dramatically from October
1973 to October 1974, and has levelled out since then. It is also
obvious that increased production costs are being passed on to the
consumer.
3.4 REFERENCES FOR SECTION 3
1. Survey of Current Business. Bureau of Economic Analysis, U. S.
Department of Commerce. Washington, D. C. 5J?:S-33, Januerv 1975.
2. Aluminum Statistical Review 1973. The Aluminum Association.
New York, N. Y. p. 52.
3-17
-------
3. Reference 2, above, pp. 32-33.
4. Reference 2, above, p. 34.
5. Reference 2, above, p. 16.
6. U.S. Industrial Outlook with Projections to 1980. Domestic and
International Business Administration, U.S. Department of
Commerce, Washington, D.C. 1974 edition, p. 77.
7. Air Pollution Control in the Primary Aluminum Industry. New York,
N.Y. Singmaster and Breyer. Prepared for Office of Air Programs,
U.S. Environmental Protection Agency, Research Triangle Park, N.C.,
under Contract Number CPA 70-21. July 23, 1973. p. 2-22, 2-23.
8. Letter from Dr. P. R. Atkins, Aluminum Company of America,
Pittsburgh, Pa. to D.R. Goodwin, Emission Standards and Engineering
Division, OAQPS, Environmental Protection Agency, Research Triangle
Park, N.C., dated June 20, 1974.
9. Personal communication from Dr. P. A. Atkins, Aluminum Company of
America, Pittsburgh, Pa. to B. A. Varner, Emission Standards and
Engineering Division, OAQPS, Environmental Protection Agency,
Research Triangle Park, N.C., June 25, 1974.
10. Letter from Dr. B. S. Hulcher, Reynolds Metals Company, Richmond, Va.
to D. R. Goodwin, Emission Standards and Engineering Division, OAQPS,
Environmental Protection Agency, Research Triangle Park, N.C., dated
June 28, 1974.
11. Letter from K. Cyr, Anaconda Aluminum Company, Louisville, Ky. to
D. R. Goodwin, Emission Standards and Engineering Division, OAQPS,
Environmental Protection Agency, Research Triangle Park, N.C.,
dated June 21, 1974.
12. Letter from J. L. Loyer, Howmet Corporation, Greenwich, Conn, to
D.R. Goodwin, Emission Standards and Engineering Division, OAQPS,
Environmental Protection Agency, Research Triangle Park, N.C., dated
October 31, 1974.
13. Letter from W. F. Boyer, Jr., Consolidated Aluminum Corporation,
Lake Charles, La. to B. A. Varner, Emission Standards and Engineering
Division, OAQPS, Environmental Protection Agency, Research Triangle
Park, N.C., dated June 25, 1974.
14. Letter from J. L. Byrne, Martin Marietta Aluminum, to D.R. Goodwin,
Emission Standards and Engineering Division, OAQPS, Environmental
Protection Agency, Research Triangle Park, N.C., dated October 10,
1974.
15. Farin, P. and G. G. Reibsamen. Aluminum, Profile of an Industry.
New York, N.Y., Metals Week, McGraw-Hill Inc., 1969. p. 153.
3-18
-------
16. Reference 15, above, p. 25.
17- Reference 2, above, pp. 37-39.
18. Reference 15, above, p. 148.
19. Reference 7, above, p. 2-28, 2-29.
20. Reference 2, above, pp. 30-31.
21. Aluminum Statistical Review 1971. The Aluminum Association.
New York, N.Y. p. 30.
22. Background Information for Standards of Performance: Primary
Aluminum Industry. Volume 1: Proposed Standards. Emission
Standards and Engineering Division, OAQPS, Environmental
Protection Agency, Research Triangle Park, N.C. October 1974.
p. 64.
23. Reference 22, above, p. 78.
24. Hamilton, W. F., Trip Report: Meetings Held in Washington, D. C.
and New York City on March 28 - 29, 1973 to discuss Aluminum
Industry. Strategies and Air Standards Division, OAQPS, OAWP,
Environmental Protection Agency, Research Triangle Park, N. C.
25. Reference 22, above, p. 93.
26. Aluminum Statistical Review 1970. The Aluminum Association.
New York, N. Y. pp. 8, 16.
27. Reference 7, above, p. 3-17-
28. Reference 22, above, pp. 74-76.
29. Reference 15, above, p. 164.
30. Reference 15, above, p. 163.
31. Reference 15, above, pp. 154-156.
32. New Processes Promise Lower Cost Aluminum. Chemical and Engineering
News. 51_(9):11-12, February 26, 1973.
33. Minerals Yearbook. Bureau of Mines, U.S. Department of the
Interior. Washington, D.C.
34. Metal Statistics, The Purchasing Guide of the Metal Industries.
The American Metal Market Company. New York, N.Y.
35. Cash Prices. The Wall Street Journal. August 16, 1971; October 19,
1973; January 2, 1974; March 26, 1974; May 30, 1974; August 2, 1974;
February 26, 1975.
3-19
-------
4. PROCESS DESCRIPTION
4,1 PRIMARY ALUMINUM REDUCTION ]"5
All primary aluminum in the United States is produced by
electrolytic reduction of alumina (AlgOg)—the Hall-Heroult
process. Alumina, an Intermediate product, is produced by the
Bayer process from bauxite, a naturally occurring ore of hydrous
aluminum oxides and hydroxides containing 45 to 55 percent AlgO-j.
The production of alumina and the electrolytic reduction of alumina
to aluminum are seldom accomplished at the same geographical loca-
tion.
Alumina is shipped to the reduction plant where it is reduced
to aluminum and oxygen by direct electric current (Figure 4-1). This
reduction is carried out in shallow rectangular cells (pots) made of
carbon-lined steel with carbon blocks that are suspended above and
extend dov/n into the pot (Figure 4-2), The pots and carbon blocks
serve as cathodes and anodes, respectively, for the electrolytical
process.
Cryolite, a double fluoride salt of sodium and aluminum
(NaJ\lFg), serves as an electrolyte and a solvent for alumina.
Alumina is added to and dissolves in the molten cryolite bath. The
cells are heated and operated between 950° and 1,000°C with heat
that results from resistance between the electrodes. During the
reduction process, the aluminum is deposited at the cathode where,
3 3
because of its heavier weight (2.3 g/cm versus 2.1 g/cm ), it
4-1
-------
REDUCTION
PLANT
i
ro
RAW MATERIAL
(SHIP OR RAIL)
ALUMINUM
SHIPMENT
POWER
Figure 4-1. Aluminum reduction process.
-------
DIRECT CURRENT
CO
MOLTEN CRYOLITE BATH,
CARBON
CATHODE
Figure 4-2. Aluminum reduction cell diagram.
-------
remains as a molten metal layer underneath the cryolite. The
cryolite bath thus also protects the aluminum from the atmosphere.
The byproduct "oxygen migrates to and combines with the consumable
carbon anode to form carbon dioxide and carbon monoxide, which
continually evolve from the cell.
Alumina and cryolite are periodically added to the bath to
replenish material that is removed or consumed in normal operation.
The weight ratio of sodium fluoride (NaF) to aluminum fluoride
(AlFo) in cryolite ts 1,50. However, it has been found that adding
excess AlFo to reduce the bath ratio to 1.30 to 1.45 will increase
cell current efficiency and lower the bath melting point permitting
lower operating temperatures. Fluorspar, or calcium fluoride, may
also be added to lower the bath melting point.
Periodically, the molten aluminum is siphoned or "tapped" from
beneath the cryolite bath, moved in the molten state to holding
furnaces in the casting area," and fluxed to remove trace impurities.
The product aluminum is later tapped from the holding furnaces and
cast into ingots or billets to await further .processing or shipped
molten in insulated ladles.
The reaction:
A12 03 + 1 1/2 C ^ 2A1 + 1 1/2 C02 (4.1)
absorbs 261.9 kcal per gram mole of alumina reacted at 1000°F, which
is equivalent to 2.56 kilowatt-hours (kwh) of energy per pound of
aluminum produced. In actual practice, however, some energy is used
to bring the reactants (including the carbon anode) up to temperature
4-4
-------
and is lost in the byproduct gas stream, with the tapped aluminum,
and to the building. The latter occurs principally through the low
temperature heat leak provided by the molten aluminum layer beneath
the cryolite bath. A small portion of the molten aluminum mixes
with the bath and is carried to the anode where it is oxidized back
to alumina, reducing some of the carbon dioxide to carbon monoxide.
(Much of the hot carbon monoxide is oxidized back to carbon dioxide
upon contacting air.) This reduction absorbs additional energy, and
practically the total cell energy requirement is from 6 to 9 kwh per
pound of aluminum produced. Furthermore, although the stoichiometric
carbon requirement by equation (4.1) is 0.33 pound per pound of
aluminum produced, the reduction of carbon dioxide to carbon monoxide
increases the carbon requirement to about 0.50 pound per pound of
7 8
aluminum produced, '
A typical late design cell may operate at 100,000 amperes and
4.5 volts (450 ktlowatts), producing 1540 pounds of aluminum per
day for an energy consumption of approximately 7 kwh per pound of
q
aluminum produced.
A large number of cells are linked together electrically in series
to form a potline, the basic production unit of the reduction plant.
The potline may be housed in one or two long ventilated buildings
called potrooms. A typical plan view of a potroom in schematic
form is shown in Figure 4-3. A typical elevation might be as shown
in Figure 4-4. The pots may be arranged end to end or side by side.
The roof "monitor" or ventilator shown in Figure 4-4 usually
4-5
-------
1000 TO 1600 FEET-
O
ss
(A
ooo
ooo
lOQiOOOO
CELLS
o
DDOI 'D(]DE:::::JD
OOO
SCRUBBERS
ooo
Figure 4-3. Typical plan view of potroom.10
SCRUBBER
\
POTROOM
CELL
| I
I PLENUM |
I J
ROOF
MONITOR
CELL
SCRUBBER
Figure 4-4. Typical elevation view of potroom.1"1
4-6
-------
runs the length of the building and serves the important function
of releasing the heat lost from the pots to the building air, thus
maintaining workable conditions around the pots. Outside air may
be introduced to the potroom through side vents, or forced through
a central floor plenum, or both.
The "process" of primary aluminum reduction is essentially
one of materials handling. It can be shown schematically as a flow
diagram such as Figure 4-5. The true difference in the various
process modifications used by the industry lies in the type of
reduction cell used. Three types of reduction cells or pots are
used in the United States: prebake (PB), horizontal stud
Soderberg (HSS), and vertical stud Soderberg (VSS). Both Soderberg
cells employ continuously formed consumable carbon anodes where the
anode paste is baked by the energy of the reduction cell itself.
The prebake cell, as indicated by its name, employs a replaceable,
consumable carbon anode, formed by baking in a separate facility
called an anode bake plant, prior to its use in the cell.
The preparation of anode materials is usually an ancillary
operation at the reduction plant site. Figure 4-6 is a typical
flow diagram for the preparation of prebake anodes. In the carbon
plant, or "green mill", coke is crushed and sized; cleaned, returned
anode butts are crushed; and both are mixed together with pitch and
molded to, form self-supporting green anode blocks. Figure 4-6 shows
solid coal tar pitch moving to a crusher. The pitch may not be coal
tar, and it may be received and handled as a liquid. The green anode
blocks are fired and baked in a pit baking furnace, or ring furnace.
Subsequently, a steel or iron electrode is bonded into a preformed hole
4-7
-------
r
Ul
II
gl
BALL
MILL
VIBRATING
SCREEN
l£
"I
BALL
WILL
AIR
CLASSIFIER
PRESS
UU--
I I
r
OVEN — — — — —
I I
NEEDED FOR PREBAKED
PROCESS ONLY
CRYOLITE
NaF, AIF3
CaF2
Figure 4-5. General flow diagram for primary aluminum reduction.12
4-8
-------
BUTTS
\-
r
-/
MIXER
u
I" "
.COOLING
CONVEYOR
MOLDING
PRESS
O
WATER
COOLING
TOWER
i
PRECIPITATOR
STACK
• TO POTLINE
Figure 4-6. Flow diagram for preparation of prebake anodes.13
4-9
-------
in each block. The electrode serves as an electrical connector and
holds the anode in place in the bath. The ring furnace operation
comprises the anode bake plant. A second typa of furnace, the
tunnel kiln, has also been developed for baking anodes.
The preparation of Soderberg anode material is similar to that
for prebake cells, except that the pitch is always liquid, the anode
paste is not molded and baked prior to cell usage, and no anode
material is returned from the cells to the carbon plant.
Since the potrooms housing the reduction cells and the prebake
anode bake plant are the facilities affected by the standards of per-
formance for new primary aluminum plants and attendant State plans for
controlling existing plants, the different cell types and the bake plant
merit further consideration. Process items specific to each are dis-
cussed in the following sections.
4.2 PREBAKE PROCESS
Prebake cells use a number of anodes suspended in the
electrolyte, in essence the original design of the Hall-Heroult
process, The anodes are press-formed from a carbon paste and are
baked in a ring furnace or tunnel kiln.
4.2.1 Anode Bake Plant
14 15
4.2.1.1 Ring Furnace ' --The ring furnace consists of compartmentalized,
sunken, brick baking pits with surrounding interconnecting flues.
Green anodes are packed into the pits, with a blanket of coke or
anthracite filling the space between the anode blocks and the walls
4-10
-------
of the pits. A 10- to 12-inch blanket of calcined petroleum coke
fills the top of each pit above the top layer of anodes. The
blanket helps to prevent oxidation of the carbon anodes.
The pits are fired with natural gas-fired or oil-fired
manifolded burners for a period of about 40 to 48 hours. The flue
system of the furnace is arranged so that hot gas from the pits being
fired is drawn through the next section of pits to gradually preheat
the next batch of anodes before they are fired, in turn, when the
manifold is progressively moved. Air for combustion is drawn through
the sections previously under fire, cooling them down. The anodes
are fired to approximately 1,200°C, and the cycle of placing green
anodes, preheating, firing, cooling, and removal is approximately
28 days.
Firing of sections proceeds down one side of the rectangular
furnace building and back the other in a "ring" pattern. Proceeding
around the building, the pattern of sections cooling down, sections
under fire, sections heating up, and empty sections is repeated
several times.
Ring furnaces use outside flues under draft, and since the flue
walls are of dry-type construction, most volatile materials
released from the anodes during the baking cycle (principally
hydrocarbons from the pitch binder) are drawn, with the combustion
products of the firing, into the flue gases where they are burned
at about 1300°C.15
4-11
-------
Flue gases may be passed through fluoride scrubbers and perhaps
electrostatic precipitators to reduce temperature and scrub or co-
precipitate out a portion of the hydrocarbons before exhausting to
a stack.
The furnace buildings spanning the lines of baking pits are
usually open at the side and ventilated through gravity roof
monitors without emission controls.
The baked anodes are stripped from the furnace pits by means
of an overhead crane on which pneumatic systems for loading and
removing the coke pit packing may also be mounted. The packing
may subsequently become part of other green anodes in the carbon
plant.15
4.2.1.2 Tunnel Kiln --A second type of furnace, the tunnel kiln,
has been developed for application in the baking of anodes. The
kiln is an indirect-fired chamber in which a controlled atmosphere
is maintained to prevent oxidation of the carbon anodes. Green
anode blocks are loaded on transporter units that enter the kiln
through an air lock, pass successively through a preheating zone, a
firing zone, and a cooling zone, and leave the kiln through a
second air lock. The refractory beds of the cars are sealed
mechanically to the kiln walls to form the muffle chamber, and yet
permit movement of the units through the kiln.
4-12
-------
The muffle chamber is externally heated by combustion gases,
and the products of combustion are discharged through an independent
stack system.
Effluent gases from the baking anodes may be introduced into
the fire box so as to recover the fuel value of hydrocarbons and
reduce the quantity of unbunned hydrocarbon to approximately 1
percent of that coming from a ring furnace. Further reduction of
solid and gaseous emissions may be achieved by the use of heat
exchangers, scrubbers, and electrostatic precipitators.
Although the tunnel kiln presents mechanical problems in design
and operation, it is reported to have several appreciable advantages
over the ring type of furnace:
1. Baking cycle from green to finished anode is much shorter.
2. Anode baking is more uniform.
3. Space requirements for equal capacity furnaces is less.
4. Smaller gas volumes are handled through the furnace
emission control system.
The successful development of the tunnel kiln in this application
is recent, and to date only one installation is in normal operation.
Baked anodes are delivered to air blast cleaning machines
utilizing fine coke as blasting grit. Fins, scrafs, and adherent
packing is removed by this treatment, and the baked anodes are then
transferred to the rodding room where the electrodes are attached.
4-13
-------
4.2.2 Reduction Cells17'18
Figure 4-7 shows a sectional view of a typical prebake (PB)
reduction cell with a hood for collection of cell emissions.
Prebake cells use up to 26 anode assemblies per cell, which
are attached to the anode bus on the cell superstructure by means of
clamps. The anode bus is attached to the steel superstructure by
anode jacks that may be driven by an air motor or by other means,
giving a travel of from 10 to 14 inches and permitting the raising
or lowering of all 26 assemblies in the cell simultaneously. Each
of the 26 assemblies may also be raised or lowered individually
by means of an overhead crane after the anode clamp is loosened.
The anodes are lowered as they are consumed, typically at a
18
rate of about 1 inch per day. When the anodes are completely
spent, they are removed and replaced on a rotating basis, usually
a pair at a time. The total operating time before replacement is
dependent on the size of the anode blocks and the amperage of the
potline.
The anode assemblies are usually installed in two rows
extending the length of the cell. In some arrangements the two
rows are closely spaced in the center of the cell, providing a
working area on each side of the cell between the cell side lining
and the anodes (sfde-worked). In other cases, the rows are separated
and placed closer to the cell side lining, providing the working
area in the center of the cell between the rows of anodes (center-
worked) .
4-14
-------
ALUMINA
HOPPER
ELECTROLYTE
GAS COLLECTION
HOOD
SOLIDIFIED CRUST
OF ELECTROLYTt AND
ALUMINA
STEEL SHELL
INSULATION
CARBON LINING
MOLTEN ALUMINUM
GAS COLLECTION DUCT
ANODE BUS BAR
CARBON ANODE
GAS & FUME EVOLVING
CATHODE
COLLECTOR BAR
BUS BAR
Figure 4-7. Details of prebake reduction cell.18
The general trend in prebake anode design has been toward
larger anode blocks, obtaining greater effective anode/cathode
surface ratios and lower current densities at the anodes for
equivalent power inputs.
4.3 SODERBERG CELLS18'19'20
There are two types of Soderberg cells, each having a single
large carbon anode, but differing in the method of anode bus
connection to the anode mass. In both the vertical stud Soderberg
CVSS) and the horizontal stud Soderberg (HSS), a green anode paste
is fed periodically into the open top of a rectangular steel
compartment and baked by the heat of the cell to a solid coherent
mass as the material moves down the casing.
4-15
-------
In both types of Soderberg cells, the in-place baking of the
anode paste results in the release of hydrocarbon fumes and
volatiles derived from the pitch binder of the paste mixture. These
products are a component of the Soderberg cell emissions and are
essentially absent from those of the prebake cells. If not removed
from the gas stream, the pitch components will condense in and
plug subsequent doctwork and emission control devices.
Although the Soderberg cells require more electrical energy to
produce a given weight of metallic aluminum and create problems
in emission control, they were acclaimed initially because they did
away with the need for a separate anode manufacturing facility.
Partially because the volatile pitch components can condense
in the ductwork and the control device, and partially because
of .the problems of simultaneously controlling fluorides and organic
emissions, any economic advantage of the Soderberg systems is
diminishing and the trend appears to be toward the prebake cell.
Furthermore, although prebake cells may be center-worked or
side-worked, .the use of a single large carbon anode requires that
both types of SoderBerg cells be side-worked. As will be discussed
in Section 6.1, center-worked cells lend themselves to more
efficient hooding and hence more efficient emission control.
4.3.1 Vertical Stud Reduction Cells
Figure 4-8 shows a sectional view of a typical vertical stud
Soderberg reduction cell. The anode casing is stationary, the
electrical connection from, the studs to the bus bar is rigid, and
4-16
-------
STUDS
ANODE CASING
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
BURNER
GAS AND TAR BURNING
GAS EVOLVING
CATHODE
COLLECTOR
BAR
THERMAL
INSULATION
Figure 4-8. Details of vertical stud Soderberg reduction cell.18
the steel current-carrying studs project vertically through the
unbaked paste portion and into the baked portion of the anode. As
the anode is consumed and moves down the casing, the bottommost
studs are periodically extracted before they become exposed to the
bath at the bottom of the anode.
The stationary anode casing and the projection of the studs
through the top of the anode allow the installation of a gas
collection skirt between the anode casing and the bath surface.
The gases are ducted to integral gas burners where the hydrocarbon
tars are burned to gaseous fractions that do not interfere with
the operation of subsequent pollutant removal equipment.
Maintenance of the skirt system is a problem, however.
Irregularities in cell operation can extinguish the burner flame,
and the skirts may melt or be deformed by the heat. Pilot lights
can help ensure that the burners stay lighted.
4-17
-------
4.3.2 Horizontal Stud Reduction Cells
Figure 4-9 shows a sectional view of a typical horizontal
stud Soderberg reduction cell. The anode, suspended over the pot,
is contained in a rectangular compartment made of aluminum sheeting
and perforated steel channels that is raised or lowered by means
of powered (jacks. The entire anode assembly is moved downward as
the working surface is oxidized. Studs are inserted into the anode
through 3-inch perforations in the steel channels at a point about
3 feet or so above the molten bath where the paste is still fairly
soft. Electrical contact is through flexible connectors between
the studs and the bus bar. As the anode is moved downward, the
paste bakes solid and grips the stud. When the bottom channel reaches
the bath, the flexible connectors are moved to a higher row of studs,
the studs in the bottom row are pulled out, and the bottom channels
are removed.
The construction of the HSS cell prevents the installation of
an integral gas collection device such as a skirt, since the anode
casing is formed by removable channels supporting the horizontal
stud electrodes, and these channels are periodically changed as the
anode moves downward and is consumed. Hooding is restricted to
canopy suspension, resulting in so much air dilution that self-supporting
combustion in burners is not possible. The hydrocarbon tars thus
condense in the ductwork and tend to plug pollutant removal
equ i pment.
4-18
-------
ALUMINA HOPPER -
FULLY BAKED ANODE
,— REMOVABLE
\ CHANNELS
—1\
SOLIDIFIED CRL'-f
OF ELECTROLYTE
AND ALUMINA
STEEL SHELL
INSULATION
CARBON LINING
ELECTROLYTE
GAS COLLECTION DUCT
ANODE PASTE
POT
ENCLOSURE DOOR
PARTIALLY BAKED
PASTE
ANODE STUDS
GAS AND FUME
EVOLVING
MOLTEN ALUMINUM
CATHODE
COLLECTOR
BAR
Figure 4-9. Details of horizontal stud Soderberg reduction cell.18
4.4 REFERENCES FOR SECTION 4
1. Air Pollution Control in the Primary Aluminum Industry. Sinqmaster
and Breyer, New York, M.Y. Prepared for Office of Air Pronrams,
Environmental Protection Agency, Research Trianqle Park, M.C., under
Contract Number CPA 70-21." July 23, 1973. p. 3-1, 3-2, 3-6 to 3-10,
3-16, 3-18 to 3-23.
2. Background Information for Establishment of National Standards
of Performance for New Sources. Primary Aluminum Industry.
Environmental Engineering, Gainesville, Florida. Prepared for
Air Pollution Control Office, Environmental Protection Agency,
Durham, N.C., under Contract Number CPA 70-142, Task Order
No. 2. (Draft Copy dated March 15, 1971). p. 2-1 to 2-8.
3. Shreve, R. N. Chemical Process Industries. 3rd Ed. New York,
N.Y., McGraw-Hill Book Company, 1967. pp. 246 - 250.
4. Robinson, J. M. et al. Engineering and Cost Effectiveness Study
of Fluoride Emissions Control. Resources Research, Inc. and
TRW Systems Group, McLean, Va. Prepared for Office of Air
Programs, Environmental Protection Agency, Durham, N.C., under
Contract Number EHSD 71-14. January 1972. p. 3-13, 3-15, 3-17.
4-19
-------
5. Background Information for Standards of Performance: Primary
Aluminum Industry. Volume 1: Proposed Standards. Emission
Standards and Engineering Division, OAQPS, Environmental
Protection Agency, Research Triangle Park, N.C., Ockober 1974.
6. Reference 1, above, p. 3-1, 3-2.
7. Reference 3, above, p. 246.
8. Reference 4, above, p. 3-17.
9. Reference 1, above, p. 3-6, 3-8.
10. Reference 2, above, p. 2-5.
11. Reference 2, above, p. 2-6.
12. Reference 2, above, p. 2-8.
13. Reference 1, above, p. 3-20. (Modified)
14. Reference 1, above, p. 3-21.
15. Varner, B. A., Trip Report: Trip to Alcoa (Badin, N.C.)
Aluminum Plant, Standards Development and Implementation
Division, SSPCP, Environmental Protection Agency, Research
Triangle Park, N.C. August 21, 1972.
16. Reference 1, above, p. 3-22.
17. Reference 1, above, p. 3-10, 3-12.
18. Reference 2, above, p. 2-9, 2-10.
19. Reference 1, above, p. 3-12, 3-13, 3-16.
20. Reference 2, above, p. 2-11.
4-20
-------
5. FLUORIDE EMISSIONS1
Pollutants emitted from primary aluminum plants include fluorides,
particulates, hydrocarbons (organics), sulfur oxides, carbon monoxide,
and nitrogen oxides. EPA tests have shown nitrogen oxide levels to
be insignificant. Although significant levels of sulfur oxides and
carbon monoxide can be emitted, control technology has not been
demonstrated and adequate source test data defining emission levels
have not been obtained for these two pollutants. On the other hand,
fluoride control has been demonstrated and characterized through EPA
source tests. These tests have also shown that, if fluorides are"
well controlled, the resulting incidental control of particulates and
organics will be good. For these reasons , the EPA standards of per-
formance for new primary aluminum plants are stated in terms of
fluoride. Likewise, discussion of emissions, control techniques, economic
impact and emission standards in this document is restricted to fluorides
except where other pollutants have a bearing on cost or performance.
5.1 POINTS OF EMISSION1'2
The principal points of fluoride emission are the primary and
secondary emissions from the potrooms housing the reduction cells
and, in the case of the prebake cell, the emissions from the associated
anode bake plant. Figure 5-1 shows these emission points from a
prebake plant with an anode ring furnace. The anode bake plant,
together with its emissions, is not part of the Soderberg plant.
Figure 5-2 shows how the reduction cells are hooded and how the
evolved gas stream is ducted to a primary control device exterior to
b-1
-------
CJ1
I
PO
SECONDARY EMISSIONS
PREBAKE CELL POTROQM
X- ^-x^^^^^ •>.
ANODE RING
FURNACE
BAKE PLANT
EMISSIONS
EMISSION
CONTROL
EQUIPMENT
Figure 5-1. Prebake plant with anode ring furnace.
-------
PRIMARY EMISSIONS
SECONDARY EMISSIONS
cn
i
CO
Figure 5-2. Potroom fluoride emission balance.
-------
the potroom. Emissions from this device are termed primary emissions.
That portion of the evolved gas stream that escapes the hooding passes
to the monitor in the roof of the potroom, where there may or may
not be a secondary control device. Emissions from the building are
termed secondary emissions.
For potroom emissions, the overall control efficiency (OCE) may
be expressed as:
OCE = VV+ ° - V)nsc"sr (5J)
where: n = Primary collection efficiency
n = Primary removal efficiency
n = Secondary collection efficiency
s c
n = Secondary removal efficiency
Some plants in the United States employ both primary and secondary
removal equipment. However, the majority of plants do not have secondary
equipment, relying on efficient primary collection (good hooding) to
obtain high overall control efficiencies. For these plants, n =0
and equation (5.1) reduces to:
OCE= VV (5-2)
A few U. S. plants employ only secondary removal equipment.
For these plants, n = 0 and equation (5.1) reduces to:
OCE = nscnsr (5.3)
Although secondary collection efficiency might be assumed to be 100
percent in this scheme, deficiency in the design of the provisions for
5-4
-------
air intake to the buildings may bring about a reduction in the collection
efficiency. Some potline buildings have openings in the sidewalls at
working floor level through which ventilation air enters as shown in
Figure 5-3. This air is supposed to sweep past the cells and up
through the roof monitor collection system, but adverse winds may blow
through the buildings in such a way as to carry potline emissions out
through wall openings in the buildings, thus short circuiting the
collection system and reducing its efficiency. Figure 5-4 shows a
building arrangement that helps to avoid this short circuiting of the
collection system. Fresh air is drawn into the building below the working
floor level and is allowed to pass up through gratings past the cells
to the monitor collection system.
For the more general case of primary plus secondary control, if
it is assumed that all secondary emissions are from the roof monitor,
then:
nsc = 1.0 (5.4)
and equation (5.1) can be written in terms of three variables:
OCE = VV + ° " V)nsr (5'5)
Equation (5.5) is the expression of OCE that will be used in
discussing retrofit control techniques (Section 6). However, the
aforementioned limitation on secondary collection efficiency because of
short circuiting should be kept in mind.
Fluoride is lost from the potroor in other ways besides the
airborne primary and secondary emissions. Figure 5-5 shows a fluoride
5-5
-------
INDUCED DRAFT FAN
ROOF MONITOR SPRAYS
Figure 5-3. Room collection system, sidewall entry.3
ROOF MONITOR SPRAYS
INDUCED DRAFT FAN
FLOOR GRATING
Figure 5-4. Room collection system, basement entry.3
5-6
-------
RECOVERED
SOLIDS
16
44
CRYOLITE
ALUMINUM-
FLUORIDE
24
CALCIUM <
FLUORIDE
2.8
PRIMARY
AIR
EMISSION
4.2
SPRAY
SCRUBBER
WET
REMOVAL
38.4
MULTIPLE
CYCLONE
DRY
RECOVERY
54.4
SECONDARY
AIR
EMISSION
10.8
34.2
WATER
EFFLUENT
> - IMPOUNDED.
LOST, OR
RECOVERED
AS CRYOLITE
10.8
POTROOM
20
•»- POT LININGS
1.6
-»•- SPENT BUTTS
Figure 5-5. Specific prebake potroom fluoride balance (balance values
in pounds of total fluoride per ton of aluminum produced).4
5-7
-------
balance, in pounds of total fluoride per ton of aluminum produced,
A
around a specific prebake potroom. This particular plant has no
secondary removal equipment. Its primary removal equipment consists
of multiple cyclones for dry recovery of particulate fluoride followed
by spray scrubbers to remove most of the gaseous fluoride and some
particulate fluoride. This level of control is neither representative of
best demonstrated control technology for new plants nor of best retrofit
for existing plants.
For the potroom shown in Figure 5-5, approximately 65 of the 87
pounds of fluoride (or 75 percent of the total) added to the pots is
released in the potroom emissions. About 20 pounds is absorbed into the
pot cathode linings and 1.6 pounds adheres to the anode butts. The butts
are returned to the carbon plant (see Figure 4-6) and the linings may
be treated to recover cryolite after their useful life (about 3
years).
Of the approximately 54 pounds directed to the primary removal
equipment, only 16 pounds are recovered and returned directly to the
pots. About 34 pounds end up in the scrubber water discharge. This
large quantity of fluoride may be discharged directly with the plant
effluent, treated to remove most of the fluoride content and then
discharged, or sent to a cryolite recovery plant for further process-
ing. Zero fluoride water discharge is difficult to attain with any
of these alternatives.
Although the airborne primary emission is only about 4 pounds,
the relatively low primary collection efficiency (83.4 percent) and
the lack of any secondary removal equipment for the specific potroom
5-8
-------
result in a secondary emission of about 11 pounds, for a total emission
of 15 pounds and an overall control efficiency of only 77 percent
of the 65.2 pounds generated.
5.2 UNCONTROLLED EMISSIONS - SOURCE, CHARACTERISTICS, AND
MINIMIZATION
5.2.1 Reduction Cells (All Types)5'6
Fluorides are emitted from the reduction cell as particulates
and gases.
5.2.1.1 Parti oilates-~Particulates originate from the volatilization
of the cryolite bath with subsequent condensation, from mechanical
entrainment of bath material by the air sweep over the cell surface,,
and from dusting of raw materials during the raw material feeding
operations.
The largest particulate component is alumina. Fluoride
components that have been identified include cryolite (Na,AlFg),
aluminum fluoride (A1F-), calcium fluoride (CaF^), and chiolite
(Na5Al3F,.). Other non-fluoride particulates are carbon, hydrocarbon
tars, and iron oxide (Fe203). It is estimated that fluorides
comprise 10 to 25 percent of the total particulates.
Reported determinations of particle size distributions in
primary uncontrolled cell emissions are plotted in Figure 5-6.
Two plots are shown for prebake potlines, one reported as the
average of four samples of pot emissions, the other as the average
of five samples of electrostatic precipitator intake. A single plot
5-9
-------
100a
E
3.
Ul
N
co
1
95
90
WEIGHT PERCENT LARGER PARTICLES
80 70 60 50 40 30 20
10
10
90
95
100
— 0.2
20 30 40 50 60 70 80
WEIGHT PERCENT SMALLER PARTICLES
INDUSTRY REPORTED DATA
0.1
•8
Figure 5-6. Particle size weight distribution of potline primary cell emissions.8
5-10
-------
of average samples is shown for HSS. No comparable data have been
obtained for VSS emissions.
These plots are illustrative of the comparative size
characteristics of the'primary dusts from two types of cells.
The slopes of these data give an indication of the range of particle
sizes in the samples, and the placement of the curves on the plot
indicates that a substantial fraction of the prebake and HSS
particulate weight is submicron, or in the range where particulate
removal efficiencies of most equipment are low.
A more recent determination of particle size distributions
in primary uncontrolled HSS cell emissions is shown in Figure 5-7.
For this study: (1) The mass mean particle diameter was 5.5 micro-
meters (pm). (2) The geometric standard deviation was quite hioh
(around 25). (3) Thirty percent by mass of the particles were less
than 1 ym and 16 percent were less than 0.2 pm in diameter. (4) The
mass mean particle diameter and the standard deviation were lower,
and the particle mass concentration was higher when the cell crust
was unstable (gas vents). (5) Increasing the air collection flow
rate increased the mass mean particle diameter, but the particle
mass concentration remained the same. (6) The fraction of particles
less than 0.5 um decreased as the distance from the cell increased
in the primary cell gas collection duct.
Published or reported particle size distribution data are
sparse and techniques for measurement are subject to variations,
even among different investigators using similar equipment, so
caution should be exercised in drawing conclusions from these data
or in comparing data from one source with those from another.
5-11
-------
WEIGHT PERCENT LARGER PARTICLES
80 70 60 50 40 30 20
10
0.
10
20 30 40 50 60 70 80
WEIGHT PERCENT SMALLER PARTICLES
90
95
0.1
98
Figure 5-7 Particle size weight distribution of HSS primary cell emissions.9
5-12
-------
5.2.1.2 Gases—The principal gases emitted from the reduction cell
are carbon monoxide and carbon dioxide. Gaseous fluoride components
present during normal operation include hydrogen fluoride (HF) and
silicon tetrafluoride (SiF^). Other gaseous, non-fluoride components
are sulfur dioxide (S02), hydrogen sulfide (H2S), carbonyl sulfide (COS),
carbon disulfide (CS2), and water vapor. During an anode effect
(discussed below), fluorocarbons, principally carbon tetrafluoride (CF^)
o
and very small amounts of hexafluoroethane (CoFg), are known to be produced.
Thermal hydrolysis of volatized bath materials appears to be
responsible for a substantial part of the hydrogen fluoride found in
reduction cell fumes. This reaction of solid or vaporized fluorides at
elevated temperatures takes place primarily at the point where the hot
gases escape through vents in the crust at the cell surface.
A source of hydrogen is necessary for the generation of hydrogen
fluoride. Water vapor in the air is a contributor of part of this hydrogen.
Other sources include residual moisture in alumina and bath raw materials,
and hydrocarbons in the carbon anodes. Generation of HF increases with
increased cell operation temperature.
Some gaseous hydrogen fluoride is removed from the reduction cell
fumes by interaction with the contained particulate matter. Chemical
reaction is responsible for some of this pickup, and some is the result
of chemisorption, absorption, and adsorption.
5.2.1.3 Composition and Quantity—Although the determination of total
fluoride content of fumes may be quite reliable, estimates of the distri-
bution of fluoride between gaseous and particulate forms is subject to
5-13
-------
uncertainty because of such factors as the degree of thermal hydrolysis
during burning of the gases and the method of separation of gases and
particulates during sampling. One study reports that the ratio of gaseous
to particulate fluoride in reduction cell fumes varies over a range of
about 0.5 to 1.3.10 These values are given for fumes that have burned
in contact with air. Weighted average data obtained in a data
acquisition questionnaire indicate that this ratio is about 1.2 for
prebake cells, 1.7 for HSS cells, and 3.0 for VSS cells with integral
bur
10
gas burners. Unburned fumes usually show a lower rationof about
0.3.
The rate of uncontrolled total fluoride emissions (evolution or
generation) also varies over a wide range, from 25 to 65 pounds per
ton (Ib/ton) of aluminum produced. Section 5.3.1 gives the range and
average evolution for each cell type.
The effective control of emissions from an aluminum reduction
potline involves attention to:
1. Operating conditions in the cells.
2. Collection of pollutants from the cells.
3. Removal of pollutants from the collected streams.
Uncontrolled emission minimization through proper operation will now
be discussed. Items (b) and (c) will be taken up in Section 6.
The quantity and composition of uncontrolled emissions can
be strongly influenced by operating conditions such as temperature,
5-14
-------
bath ratio, frequency of anode effects, and method of crust breaking.
Moreover, it may vary with time for any given plant, because of
gradual changes that may occur in potline operations.
5.2.1.4 Normal Operation—Under normal cell operation, experimental
work established correlations between three cell operating parameters
and the level of fluoride generation for a 10,000 ampere laboratory
experimental prebake type aluminum reduction cell. It was shown that
increasing bath ratio (NaF/AlF.J, increasing alumina content of the
bath, and decreasing the bath temperature combine to effect a decrease
in the fluoride generated. These effects would be expected from the following
principles of physical chemistry: (a) A1F- is more volatile than NaF and
tends to be driven off as its relative amount increases; (b) decreased
temperature decreases the evolution of all volatiles, including A1F,;
and, (c) the addition of alumina should tend to increase aluminum ions
concentration in the melt, which may decrease fluoride ion concentration
by a mass action effect. This effect would change fluoride ions into A1F3.
Fluoride ions in the bath are probably not volatile, but A1F- is volatile.
Table 5-1 summarizes the findings of these tests.
Table 5-1. EXPERIMENTAL EFFECT OF THREE OPERATING
VARIABLES ON FLUORIDE GENERATION12
Bath ratio
(1.44 to 1.54)
1.50
1.50
Range of variable
Alumina Temperature
content, % °C
4 975
(3 to 5) 975
4 (982 to 972)
Fluoride
level,
% decrease
31
20
24
5-15
-------
The report of the above experimental work calls attention to the
fact that "determination of the effect of operating variables on the
fluoride emission from electrolytic reduction cells is difficult to
accomplish with a high degree of certainty. This is true even with
small-scale experimental cells operated by research personnel. It
appears from the work reported here, however, that cell temperature,
bath ratio, and alumina concentration are the most important variables
12
affecting total fluoride emission."
It should be noted that the researchers did not carry out an
exhaustive study of all the variables that could affect fluoride
emissions. Also, the absolute relationships reported may not hold
for full-scale cell operation.
5.2.1.5 Process Interruption—Normal cell operation is interrupted by
occasional anode effects, cell working to introduce alumina feed, and
periodic tapping of molten aluminum. Cells may also be operated at
elevated temperatures in a "sick" condition. Normal operation of prebake
cells is interrupted by the periodic changing of anodes, and normal
operation of VSS cells can be interrupted by a "stud blow."
Tapping and changing anodes cause moderate increases in fluoride
evolution, depending upon how much of the molten electrolyte is
exposed. Anode effects, operation at elevated temperatures, cell working,
and stud blows can cause significant increases in fluoride evolution and
will now be examined in some detail.
5.2.1.5.1 Anode effects—Normally a cell operates with about 2 to 5
percent of alumina in solution in the bath, but as the electrolysis
proceeds the alumina content is decreased, being intermittently replenished
•
5-16
-------
by feed additions. When this content falls to about 1.5 to 2.0 per-
cent the phenomenon of an "anode effect" occurs. It is believed that
at this alumina concentration the bath fails to wet the carbon anode
and a gas film of CF^ collects under the anode. This film causes a
high electrical resistance, the normal cell voltage increases 10
to 15-fold, and the power input to the cell increases more than 10-
fold. To correct the condition the cell crust must be broken and
more alumina added to bring the concentration back to its normal
content. The gas film under the anode is dispersed and the cell
returns to normal voltage.
The power increase to the cell is converted into heat, which in
turn raises the temperature of the cell electrolyte. At the
higher cell temperature, the fluoride evolution is increased. For
the anode effect evolution rate compared with quiet cell operation,
1 ^
one study found a 27-fold increase in solid F and a 2.7-fold
increase in HF. Another study determined that the normal fluoride
evolution from a crusted-over cell is approximately 30 pounds of fluoride
per ton of aluminum produced, but during an anode effect the fluoride
evolution rate increases to approximately 756 Ib/ton of aluminum.
Depending on the promptness with which the cell operator reacts,
this anode effect may last from 3 to 15 minutes. Occasionally
cell operators will deliberately allow anode effects to continue in or-
der to soften an unusually hard crust. Automatic crustbreakers help
to minimize the need for this practice. In normal cell operation
5-17
-------
with manual crust breaking, the frequency of anode effects is from less
than 1/2 to as many as 6 anode effects per cell-day.
The frequency of anode effects can be reduced to the range of 1/2
to 1 anode effect per cell-day by placing cells on an anode effect
suppression system. Workings of the cell are scheduled so that the
alumina content of the electrolyte is replenished before it falls below
the concentration causing the anode effect. The newer computer-controlled
potlines may operate almost free from anode effects.
5.2.1.5.2 Elevated temperature operation--The higher the bath temperature,
the more will the bath salts vaporize and be carried into the cell emissions.
Normal operating temperatures for cells with a bath ratio of approximately
1.40 are between 970 and 980 °C. Abnormal or "sick" cells operate at
temperatures in excess of 1000 °C and sometimes they do not crust over.
Under these conditions, the high-temperature molten electrolyte is
exposed, and there is a large increase in volatilization of bath salts
with a corresponding increase of fluoride. Operation of cells at
the lowest possible temperature to minimize fluoride emissions requires
trained, conscientious cell operators or computer control.
The temperature of the cell may be lowered by adding lithium-
salts to the electrolyte to lower its freezing point, but the net
benefit of these additions is the subject of controversy. One foreign
investigator reports among other advantages, a substantial decrease
of fluoride losses in waste gases, which resulted in a reduction of
fluoride emissions. In this country, experiments undertaken by a major
producer were reported to have demonstrated an increase in fluoride
emissions upon adding lithium salts.
5-18
-------
The electrolyte system is complex, and electrolyte conditions
which reduce fluoride emissions from the molten bath, but which
simultaneously destroy the ability of the bath to crust over and
carry a cover of alumina, may result in a net increase in cell em'ssions;
the alumina cover intercepts a substantial quantity of fluoride and
returns it directly to the molten bath.
5.2.1.5.3 Cell working, mechanization, and computer control--Breaking
the crust of the cell for a cell working causes the fluoride evolution
rate to rise to approximately 106 Ib/ton of aluminum. The duration
of a cell working varies according to the size and type of the cell and
whether the cell is equipped with automatic crust breakers. With the
automatic crust breaker on a prebake cell, working is accomplished quickly,
taking only 1 or 2 minutes. For a normal-size prebake cell of approximately
90,000 amperes, a manual working may be accomplished in 5 to 10 minutes
depending upon the hardness of the crust. Soderberg cells and side-
worked prebake cells are normally worked by means of a pneumatic crust
breaker similar to a paving breaker. A working may be accomplished in
approximately 5 minutes on a 90,000-ampere side-worked cell.
Mechanization of crust breaking and cell feeding allows the cell
operators time to maintain close watch over the operating cells and
to control them within narrow temperature ranges. The overall effect is
lower average operating cell temperature, fewer and briefer anode effects,
and a reduction in the fluoride content of cell off-gases compared with
normal manual cell operation.
Full mechanization of reduction cells makes it possible to apply
computer control, which incorporates the frequent scanning of operating
5-19
-------
variables on each cell and the triggering of automatic corrective
action for any variation that is outside set operating limits. Such
control makes it possible for all cells in a potline to be operated
at the lowest practical temperature and with nearly complete freedom
from upsets caused by anode effects. Cell feeding and crust breaking
operations can be cycled in response to the needs of individual cells,
and the number of abnormal or "sick" cells usually associated with
manual potline operation can be reduced. Variations in cell operation
caused by having different shift personnel tending the cells over the
24-hour period are largely avoided.
Many plants are developing various degrees of computer control in
combination with mechanization. Full automation has been satisfactorily
accomplished on at least one plant.
5.2.1.5.4 VSS stud blows—An abnormal occurrence that can increase
emissions from a VSS cell is a stud blow. (See Section 4.3.1 for a
process description of the VSS cell). This abnormality happens when the
steel, current-carrying studs are not extracted before being exposed
to the bath at the bottom of the anode. Stud blows can last up to an
hour before the unbaked paste portion of the anode eventually covers
over the exposed area. Stud blows can be prevented by proper operator
attention.
5.2.2 Anode Bake Plant
Uncontrolled fluoride emissions from anode bake plants originate from
the recycled anode butt scraps that carry absorbed or adherent bath
materials (principally cryolite) back into the anode cycle. The fluorides
5-20
-------
are incorporated into the green anode paste and released during the
subsequent baking process. (See Sections 4.1 and 4.2 for further
process information.)
5.2.2.1 Ring Furnace--Although the physical state of the fluorides
evolved from the ring furnace has not been thoroughly investigated,
it is believed that most of the fluorides evolved are gaseous at the
elevated operating temperature. (Combustion temperature is around
1300°C.)
The fluoride balance in Figure 5-5 shows 1.6 pounds of fluoride
recycled with the butts per ton of aluminum produced. An emission
level of 1.6 Ib/ton represents emissions that can be expected when
adherent cryolite is simply knocked off the butts at the potroom prior
to sending them to the anode plant. It is reported that fluoride
emissions can be maintained at less than 0.4 Ib/ton (a four-fold
reduction) by exercising particular attention to cleaning the spent
anode butts.
The principal ring furnace emissions are solid products of firing
combustion (smoke) and burned and unburned hydrocarbons derived from
the heating and carbonizing of the paste binder pitch. Some S02 and
sulfur trioxide (SO.,) is derived from the sulfur in the coke. Visible
emissions can be reduced by:
1. Using natural gas instead of oil to fire the furnace,
2. Preventing leakage of cold air into the sections under
fire, and
3. Not locating the exhaust manifold too far from the sections
under fire.
5-21
-------
5.2.2.2 Tunnel Ki"ln--A1 though the direct-fired ring furnace has been
the normally used type for prebake anodes, attention has been given to
the development of continuous tunnel kilns for this purpose. (See
Section 4.2.1.2 for a process description of the tunnel kiln.)
Combustion conditions are significantly different and zonal temperature
control closer, with one result being a reduction in the emission of
fluorides by a factor of 0.02,16 or to a level of <_ 0.03 Ib/ton of
aluminum produced. Drastic reductions in the emission levels of other
ring furnace pollutants are also achieved.
Test data on tunnel kilns are old. Emission criteria developed for
tunnel kilns should be based on new sampling data collected by prescribed
EPA methods, or other methods.
5.2.2.3 Production Bases—The above discussion of fluoride emissions
is based on weight of fluoride emitted per weight of aluminum produced.
This generally involves converting the weight of carbon anode produced
to the equivalent weight of aluminum produced. The proper method for
performing the calculation is given in the primary aluminum Standards
18
of Performance for New Stationary Sources. The weight of total
fluoride emitted per unit of time is divided by the weight of anode
produced per same unit of time. This ratio is divided by 2 or some
other proven factor representing the ratio of the weight of aluminum
produced to the weight of anode consumed.
5.3 TYPICAL FLUORIDE EMISSIONS AND EXTENT OF CONTROL
5.3.1 Reduction Cells (All Typed)
Figure 5-5 shows a total fluoride balance around a specific pre-
bake potroom that has no secondary removal equipment. The plant's
5-22
-------
primary removal equipment consists of multiple cyclones followed by
spray scrubbers. The total fluoride emission level is 15 Ib/ton of
aluminum produced. This emission level is typical of a poorly controlled
prebake, VSS, or HSS potroom.
Table ">2 gives a range of cell evolution for center-worked prebake
(CWPB), side-worked prebake (SWPB), VSS and HSS plants. The table also
shows average evolution, average total primary plus secondary emissions,
and overall control efficiencies for all 31 U.S. plants and for the 29
plants comprising the controlled segment of the industry; one CWPB and
one SWPB plant are uncontrolled. The ranges and averages in Table 5-2 are
computed directly from existing control combinations shown in Table 7-2.
Table 5-2 shows wide ranges of cell evolution, particularly for
CWPB plants. Average evolution is highest for SWPB and VSS plants
being about 10 Ib/ton higher than for HSS plants and nearly 5 Ib /ton
higher than for CWPB plants. Inclusion of the uncontrolled SWPB plant
lowers SWPB overall control efficiency much more than does inclusion
of the uncontrolled CWPB plant lower CWPB efficiency, because the SWPB
plant accounts for a much larger percentage of its respective total cell
type capacity.
In 1970, an EPA contract study determined overall control efficiencies
2
for the domestic aluminum industry. Compared to Table 5-2, efficiencies
for the controlled segment of the industry were 8 percent lower for
prebake plants, 13 percent lower for HSS plants, and 9 percent lower for
VSS plants. Overall control efficiency for all plants, controlled and
19
uncontrolled, was 8 percent lower. The improvement in overall
control from 1970 to 1975 demonstrates existence of a potential for
emission reduction.
5-23
-------
Table 5-2. POTROOM TOTAL FLUORIDE EMISSIONS IN U.S., 1975
Cell Typea
CWPB
SWPB
HSS
VSS
All Cell
Types
Range of
Evolution,
Ib F/ton Al
25.7 - 65.6
37.3 - 53.0
28.4 - 45.0
30.5 - 53.5
25.7 - 65.6
Controlled and Uncontrolled Plants
Average
Evolution,
Ib F/ton Al
40.8
44.7
34.0
44.4
40.4
Average
Emission,
Ib F/ton Al
6.3
13.0
5.7
5.2
7.0
Overall
Control
Efficiency, %
85
71
83
88
83
Controlled Plants Only
Average
Evolution,
Ib F/ton Al
40.8
43.9
34.0
44.4
40.2
Average
Emission,
Ib F/ton Al
4.6
4.8
5.7
5.2
5.0
Overall
Control
Efficiency, %
89
89
83
88
88
en
ro
CWPB -- center-worked prebake cells, SWPB — side-worked prebake cells, VSS -- vertical stud Soderberg
cells, HSS -- horizontal stud Soderberg cells.
-------
Table 5-3 shows the extent of control in the domestic industry
by cell type. The percentages are computed directly from the collected
control status information that was used to construct Table 7-2. The
table shows that CWPB plants have the greatest potential for improving
primary control. It also shows that secondary control is not employed at
any CWPB or HSS plants. By comparison, the aforementioned contract study
determined that, in 1970, only 31 percent of capacity had best primary
20
control and only 4 percent had best primary and secondary control.
Table 5-3. EXTENT OF POTROOM CONTROL, 1975
Annual
Capacity, At Least
Cell Type tons Al Primary Control
CWPB
SWPB
HSS
VSS
All Cell
Types
2,704,000
738,000
1,045,000
635,000
5,122,000
95
81 b
100
100
95
Percentage of Capacity Having:
At Least Best Best Primary Control
Primary Control + Secondary Control
61
79b
83
100
73
0
59
0
33
11
CWPB -- center-worked prebake cells, SWPB -- side-worked prebake cells,
VSS -- vertical stud Soderberg cells, HSS -- horizontal stud Soderberg
cells.
Or, secondary control with equivalent overall control efficiency.
5-25
-------
5.3.2 Anode Bake Plant
Table 5-4 gives a 1970 breakdown of evolved and emitted participate,
gaseous, and total fluorides on a pounds per ton of aluminum basis and
19
resultant overall control efficiencies for ring furnaces. The breakdown
2
is based on an EPA contract study. The data in this study are based on
reported data on furnace gases, the control equipment identified in
individual bake plants, and estimated control efficiencies ascribed to
these control systems. The evolved total fluoride emission factor of
0.86 Ib/ton lies between uncontrolled emission factors of 1.6 Ib/ton
representative of poor cleaning of the anode butts and of 0.4 Ib/ton
representative of proper cleaning. (See Section 5.2.2) According to
Table 5-4, 95 percent of the evolved and emitted ring furnace fluorides
are gaseous.
It is estimated that prebake plants representing about 43 percent
of bake plant capacity have some sort of emission control, much of it
21
experimental. It is estimated that spray scrubber control can achieve
96 percent removal efficiency on gaseous fluorides and 75 percent removal
efficiency on particulates and that 40 percent of bake plant capacity
have this level of control.
Tunnel kilns are reported to produce much lower emissions than ring
furnaces (See Section 5.2.2). However, the proportion of prebake anode
capacity baked in tunnel kilns in 1970 was small enough (about 7
percent) that to ignore them in the above discussion does not affect the
limited accuracy of the calculations.
5-26
-------
Table 5-4. RING FURNACE FLUORIDE EMISSIONS IN U.S., 1970
19
Controlled and uncontrolled
furnaces
Gaseous fluoride
Evolved, Ib/ton Al
Emitted, Ib/ton Al
Overall control efficiency,
0.816
0.483
41
Particulate fluoride
Evolved, Ib/ton Al
Emitted, Ib/ton Al
Overall control efficiency,
0.044
0.024
45
Total fluoride
Evolved, Ib/ton Al
Emitted, Ib/ton Al
Overall control efficiency,
0.859
0.507
41
Primary aluminum is a significant contributor of atmospheric
22
fluorides. One study estimated that aluminum accounted for 13.5
percent of total fluoride atmospheric emissions from major industrial
sources in 1968. Section 9.1 gives potroom and anode bake plant annual
total U.S. fluoride emissions, as well as annual particulate emissions.
5-27
-------
5.4 REFERENCES FOR SECTION 5
1. Background Information for Standards of Performance: Primary
Aluminum Industry. Volume 1: Proposed Standards. Emission
Standards and Engineering Division, OAQPS, Environmental
Protection Agency, Research Triangle Park, N.C. October 1974.
2 Air Pollution Control in the Primary Aluminum Industry. Singmaster
and Breyer, New York, N. Y. Prepared for Office of Air Programs,
Environmental Protection Agency, Research Triangle Park, N. C. ,
under Contract Number CPA 70-21. July 23, 1973. p. 3-3 to 3-6,
3-25, 5-5 to 5-7, 8-9, 8-13.
3. Reference 2, above, p. 5-6.
4. Reference 2, above, p. 3-4.
5. Reference 2, above, p. 4-2 to 4-9, 5-1 tn 5-4.
6. Background Information for Establishment of National Standards of
Performance for New Sources. Primary Aluminum Industry. Environmental
Engineering, Gainesville, Florida. Prepared for Air Pollution Control
Office, Environmental Protection Agency, Durham, N. C., under Contract
Number CPA 70-142, Task Order No. 2. (Draft Copy dated March 15, 1971).
p. 3-3 to 3-4.
7. Reference 2, above, p. 4-2.
8. Reference 2, above, p. 4-6.
9. Hanna, T. R. and M. J. Pilat. Size Distribution of Particulates
Emitted from a Horizontal Spike Soderberg Aluminum Reduction Cell.
J. Air Pol. Contr. Assoc. 2^:b33-b3b, July iy72.
10. Henry, J. L. A Study of Factors Affecting Fluoride Emission from
10 KA Experimental Aluminum Reduction Cells. In: Extractive
Metallurgy of Aluminum. New York, Interscience Publishers, 1963.
pp. 67-81.
11. Reference 2, above, p. 7-4 to 7-6.
12. Reference 2, above, p. 4-9.
13. Less, L. N. and J. Waddington. The Characterization of Aluminum
Reduction Cell Fume. In: Light Metals. New York, Proceedings
of Symposia, 100th AIME Annual Meeting, March 1-4, 1971.
14. Miller, S. V. et al. Emission of Fluorine Compounds From Electro-
lysis Cells in the Production of Aluminum. Sverdlovsk Nolich,
Issled Inst.
5-28
-------
15. Wendt, G. Operating Experiences With Electrolytes Containing
Lithium Fluoride. IMS of AIME Paper No. A70-39, February 16, 1970.
16. Reference 2, above, p. 4-11.
17. Varner, B. A., Trip Report: Trip to Alcoa (Badin, N. C.) Aluminum
Plant. Standards Development and Implementation Division, SSPCP,
Environmental Protection Agency, Research Triangle Park, N. C.
August 21, 1972.
18. Title 40 - Protection of Environment. Subpart S - Standards of
Performance for Primary Aluminum Reduction Plants. Federal Register.
39(206):37732, October 23, 1974.
19. Reference 2, above, p. 7-13.
20. Reference 2, above, p. 9-11 to 9-21.
21. Reference 2, above, p. 7-2.
22. Biologic Effects of Atmospheric Pollutants: Fluorides. National
Academy of Sciences, Washington, D. C. Prepared for Environmental
Protection Agency, Durham, N. C., under Contract Number CPA 70-42.
1971. p. 9.
5-29
-------
6. CONTROL TECHNIQUES FOR POTROOM AND ANODE BAKE PLANT FLUORIDES
Section 6.1 discusses potroom retrofit primary collection systems,
and explains in detail a method for calculating hooding efficiencies for
each of the four cell types. The calculated efficiencies are shown to
agree well witn those given by several plant operators, and are important
in later development of the State guidelines for existing primary aluminum
plants.
Section 2 discusses potroom and anode bake plant removal equipment,
both in general terms. It is difficult to generalize about retrofit
controls for this industry since costs and methods of emission control
vary widely from plant to plant. Hence, Section 6.3 gives three detailed
and seven capsule case descriptions for potroom retrofits at ten actual
plants. These descriptions cover nine best primary control retrofits
and two best secondary control retrofits, but the plants selected do not
necessarily typify all domestic plants.
Finally, Section 6.4 discusses the length of time needed to install
fluoride controls at existing primary aluminum plants.
6.1 POTROOM RETROFIT PRIMARY COLLECTION SYSTEMS1"6
For potroom emissions, overall control efficiency (OCE) was expressed
in Section 5.1 as:
OCE = VV + (1 - V)nsr (5'5)
where: r\ = Primary collection efficiency
pc
TI = Primary removal efficiency
n = Secondary removal efficiency
sr
6-1
-------
The best retrofit primary removal equiorient characteristically have
high total fluoride primary removal efficiencies, generally 98 percent or
greater. Secondary removal equipment characteristically have total fluoride
secondary removal efficiencies no higher than 75 percent and many plants
have no secondary control. Hence a high overall control efficiency greatly
depends on an efficient primary collection system.
Primary collection systems include the hooding devices installed
at the reduction cells, the individual cell ducting to common headers
serving groups of cells, and the main ducting leading to the primary
removal equipment. This section discusses hooding design, presents an
experimental method for calculating primary collection efficiency, and
discusses primary exhaust rates and ducting layouts.
6.1.1 Cell Hooding
A high primary collection efficiency greatly depends on a hood
.that is highly efficient at containing fluoride emissions and directing
them to the primary removal equipment. The characteristics of the
different cell types place various limitations on hooding design.
6.1.1.1 Prebake Cells — Figure 6-1 illustrates the hooding for a
typical prebake cell. The hooding consists of removable end doors and
a gas collection skirt on both sides made up of segmented, lightweight
aluminum doors or side covers.
As mentioned in Section 4.2.2, prebake (PB) cells may be center-
worked (CWPB) or side-worked (SWPB). CWPB cells can be worked from the
end or internally without removing the side covers. Because of this,
CWPB potlines have total fluoride primary collection efficiencies of
•
6-2
-------
ALUMINA HOPPER
I
oo
SEGMENTED
DOORS
ALUMINA
CARBON ANODES
Figure 6-1. Typical prebake cell hooding.
-------
at least 95 percent or are capable of achieving this level through
appropriate cell design changes to improve hooding. The proper approach
is to tightly seal the hood and to perform as many cell operations as
possible with the hood intact. Tight seals are achieved by having tight-
fitting end doors and side covers that fit snugly and align with each
other. Edge sealing of worn and misaligned covers is improved by
including deep 90° side edges on covers at time of fabrication.
The top seals around the anode stems are the most difficult to achieve.
This is due to the low density of the gases inside the hooding enclosure.
This top seal is particularly difficult for prebake cells equipped with
anode stems that vary in cross-sectional area, resulting in variable
clearances as the anodes are lowered. As for internal working, CWPB
cells can be equipped with automatic crust breakers and are capable
of achieving the full mechanization with computer control described in
subsection 5.2.1.5.3.
At least one CWPB plant has adopted a different hooding design
from that shown in Figure 6-1. In this design, the side covers are
flat, heavy aluminum doors hinged at the bottom with a gravity seal
at the top. In addition, a labyrinth seal provides an excellent cover-
to-cover fit. This door is expected to provide a retrofit primary
collection efficiency of 97 percent, although this has not yet been
demonstrated in full-scale potline operation.
SWPB cells must be worked manually along both sides with the side
covers removed. Hence, SWPB potlines are typically capable of achieving
a total fluoride primary collection efficiency of no higher than 85
percent, and some can do no better than 80 percent. The tight sealing
6-4
-------
and computer control discussed above are possible with SWPB cells,
but the installation of automatic crust breakers and full mechanization
are not.
To accomplish side working on all SWPB cells in a potroom within
a reasonable time period, all the side covers on one cell are normally
removed while that cell is worked. For this reason, some SWPB plants
have installed flat aluminum or steel hood doors that extend the full
length of both sides of the cell. When closed, they form an angular
gas collection skirt similar to that shown in Figure 6-1, but not seg-
mented. They are opened by air cylinder or air motor to one or more
open positions, depending on operating requirements. The opening linkages
must be precisely designed, and can be quite complicated. At each end
of the cell, the doors seal against stationary wing panels that can be
adjusted to minimize leakage. SWPB plants employing these hood designs
have cells set into the floor rather than elevated. Heat-resistant
cloth is installed around the door bottom and gravity seals the hood
when the door is closed.
SWPB potlines can seldom be converted to CWPB potlines to improve
their primary collection efficiency and reduce the need for secondary
control. This is because the steel superstructure is the most expensive
part of the SWPB cell and would have to be completely replaced in con-
verting to a CWPB design. The relatively less costly common cathode
shell is removed anyway about once every three years for lining replace-
ment. All CWPB cells and some SWPB cells are aligned side by side as
shown in Figure 4-3. However, with SWPB cells, more space is required
between the cells and less between the cells and the potroom walls. Hence
a SWPB potroom would probably be too narrow for CWPB cells, and much of
6-5
-------
its length would be wasted in a conversion. If left no alternative but
to convert to the CWPB design, the owner of a SWPB potline would probably
abandon the potline and build a new CWPB facility.
As shown in Figure 6-1, CWPB and SWPB cells generally have single-
point pickup at one end of the cell. To get uniform gas flow across
the cell, the gas duct at the top of the cell is vaned to give the
same pressure drop from various points above the cell to the single-
point pickup.
6.1.1.2 -- Horizontal Stud Soderberg (HSS) Cells — Figure 6-2 illu-
strates the total-enclosure hooding for a typical HSS cell; however,
the anode pins and the steel casing generally extend closer to the
bath than is indicated. The hood doors extend the full length of both
sides of the cell, and working a side requires opening an entire door.
Most draft systems cannot provide sufficient capture velocity to
efficiently collect emissions under these circumstances.
HSS potlines are capable of achieving total fluoride primary
collection efficiencies of 85 to 95 percent. Like CWPB potlines,
the proper approach involves tighter hood sealing and internal working
with the hood doors closed. Tighter sealing can be achieved by replacing
manual with mechanically operated doors and by eliminating
gaps on the top and the ends of the cell hooding enclosure. To minimize
door openings, at least one HSS plant (See section 6.3.1.1) has installed
a mechanized feeding system that feeds most of the alumina with
the hood doors closed. It is still necessary to manually work the cell
periodically, but the length of time per cell-day that the doors are
open is reduced. The mechanized feeding system may operate at preset
6-6
-------
TO PRIMARY CONTROL
EQUIPMENT
CTi
ANODE PINS
END DOOR
Figure 6-2. Typical horizontal stud Soderberg cell hooding.
-------
time intervals, or the potline may be computer-controlled with alumina
being added on a demand basis.
The primary collection efficiency an HSS potline can achieve depends
upon cell age and cell geometry.8 Older, smaller cells may need to be
opened more frequently than newer ones. For example, one plant has
cells installed in the early 1940's that must be opened about 50 times
per ton of aluminum produced, and cells installed in the late
1960's that are opened only 16 times per ton. As for geometry, two
HSS plants have cells with an unusually high lengthrwidth ratio of
8:1 as opposed to a normal ratio of 5:1. Cell working constraints
require single point pickup at one end of the cell, rather than
the desirable pickup at both ends.
6.1.1.3 - Vertical Stud Soderberg (VSS) Cells — Figure 6-3 illustrates
the hooding for a typical VSS cell. The hood skirt consists of an in-
verted U- or V-shaped channel that runs around the edge of the anode
assembly at the bath level. The channel is formed by the anode itself
and the outer anode casing. The channel serves as a duct to carry the
evolved gases to integral gas burners, typically one on each end of the
cell. Hence, a substantial area of the cell surface is outside the
hood skirt. This annular, exposed area is normally covered by a crust
of cryolite and alumina, the latter adsorbing fluorides that otherwise
would escape. However, this crust is broken when the cell is worked,
exposing the molten bath until the crust reforms or the bath is covered
with alumina after the cell is worked. During the exposed period, large
quantities of fluoride escape to the potroom roof. Total fluoride pri-
6-8
-------
ORGANIC FUMES
TO PRIMARY
CONTROL EQUIPMENT
ANODE PINS
CARBON ANODE
i
vo
.GAS AND TAR
BURNING
SKIRT
ALUMINA
Figure 6-3, Typical vertical stud Soderberg cell hooding.
-------
mary collection efficiencies for VSS potlines vary from 75 to 92 per-
cent.
Table 6-1 shows the effect that exposed bath area has on primary
collection efficiency for two VSS plants.9 Both plants were built by
the same firm and have cells capable of producing about one-half ton
of aluminum per day. Table 6-1 shows that the plant with the greater
exposed surface area has a correspondingly lower primary collection
efficiency.
Table 6-1. PRIMARY COLLECTION EFFICIENCY VERSUS EXPOSED ANNULAR
AREA FOR TWO VSS PLANTS9
Plant
T
Exposed Annular
Area, ft?
37.6
62.4
Primary Collection
Efficiency. %
80 - 85
75 - 80
6.1.1.4 Effect of Hooding on Overall Control -- The aforementioned
hooding limitations mean that CWPB and some HSS plants appear capable
of achieving high overall control efficiencies without installing
secondary control if appropriate cell design changes are made to im-
prove hooding. High primary collection efficiencies (90 percent or
greater) are not achievable on SWPB, most VSS, and some HSS plants,
and secondary control would be necessary for these plants to achieve
high overall control efficiencies.
6-10
-------
6.1.2 Calculation of Primary Collection Efficiency
One primary aluminum company operating SWPB cells of Swiss-design
submitted an experimental method for calculating primary collection
efficiency. All the fluoride generated from unhooded cells in a tight
building was sent to a secondary wet scrubber, whose water was analyzed.
Analysis was repeated for all the cell functions; this resulted in the
generation severity indices shown in Table 6-2. For instance, anode
changing generates six times as much fluoride as normal cell operation.
Since all but the normal operation requires the hood to be open, the
number of minutes in 24 hours required for each function is measured,
and designated as "function time" in Table 6-2.
Using the submitted data, generation rates for each function
and an overall generation rate are calculated. Establishing the percent
of leakage for each cell function, the secondary loading (non-primary
collection) for each function and an overall secondary loading can be
calculated. For one plant employing this cell design, the leakage rate
is estimated to be 7 percent when the cell is closed. For this same
plant, the hood door opens fully for anode changing, and leakage is
assumed to be 85 percent. For all other non-normal operations, the hood
door is partially open, and leakage is assumed to be 70 percent.
Primary collection is determined as the difference between generation
rate and secondary loading, from which a primary collection efficiency
of 80 percent is calculated. As of spring 1975, the above plant had not
completed hood installation, and hence had not measured primary collec-
tion and secondary loading. However, they estimate primary collection
6-11
-------
Table 6-2. CALCULATION OF PRIMARY COLLECTION EFFICIENCY FOR
ONE SWISS-DESIGN SWPB PLANT™,11
Cell Function
Fluoride
Generation
Severity Function Time,
Index (A) Minutes (B)
Generation Rate,
A ' B Leakage
Normal Operation
Anode Effects
Anode Changing
Metal Tapping
Bath/Metal Measure-
ment .
Short Side Crust
Breaking
Long Side Crust
Breaking
Bath Addition
Alumina Addition
Other Controls
Totals
1 Xa
5.5 X
6 X
2 X
1 X
4.5 X
7 X
1 X
1 X
2.5 X
1359
6
3
6
2
10
26
1
20
2
1440
1359 X
33 X
48 X
12 X ~^
2 X
45 X !
182 X
1 X :
20 X
5 X _
1707 X
n
70%
85%
\
'-- 70%
i
Average generation rate = 1707 X / 1440 = 1.185 X
Secondary (2°) loading = Generation • Leakage = [0.07 (1359 X) + 0.85
(48 X) + 0.70 (300 X)] / 1440 = 0.240 X
Primary'collection efficiency = [(Generation - 2° loading) / Generation]
' 100% = [(1.185 X - 0.240 X) / 1.185 X] • 100% = 80%
aX is defined as the normal fluoride generation from a crusted-over cell,
expressed in units of weight of fluoride generated per weight of aluminum
produced—e.g., Ib F/ton Al.
6-12
-------
11 12
efficiency will be 81 percent, * in close agreement with the
calculated efficiency.
No other aluminum companies submitted severity indices or function
times for ^ny cell types. However, the above calculational procedure
was extrapolated to other cell types, keeping the same severity indices
but modifying function times and leakages. Other cell types chosen for
this analysis were:
a. An SWPB plant of French design that is retrofit case descrip-
tion C in section 6.3.3.
b. A typical American-design CWPB plant.
c. An HSS plant that is retrofit case description A in section 6.3.1
d. A typical VSS plant.
Table 6-3 shows the calculations for the French-design SWPB plant.
The function time for anode effects has been reduced from 6 minutes to
3 because this plant is computer-control led, while the Swiss-design
plant was not. Leakage rate is estimated to be 5 percent when the cell
is closed. The hood door opens fully for all non-normal operations.
Leakage is assumed to be 70 percent instead of the 85 percent for the
Swiss-design plant because, unlike the latter, the French-design plant
has a fixed superstructure that should cover a greater portion of the
cell with the doors open. The calculated primary collection efficiency
13
of 83 percent agrees well with a measured efficiency of 85 percent.
6-13
-------
Table 6-3. CALCULATION OF PRIMARY COLLECTION EFFICIENCY FOR ONE
FRENCH-DESIGN SWPS PLANT - RETROFIT CASE DESCRIPTION C.
Cell Function
Fluoride
Generation
Severity Function Time,
Index (A) Minutes (B)
Generation Rate,
A ' B Leakage
Normal Operation
Anode Effects
Anode Changing
Metal Tapping
Bath/Metal Measure-
ment
Short Side Crust
Breaking
Long Side Crust
Breaking
Bath Addition
Alumina Addition
Other Controls
Totals
1 X
5.5 X
6 X
2 X
1 X
4.5 X
7 X
1 X
1 X
2.5 X
1362
3
8
6
2
10
26
1
20
2
1440
1362
16
48
12
2
45
-182
1
20
5
TO-3
X
X
X
X
X
X
.X
X
;
x
X
V
5%
\
j
)
> 5%
^ 7Q%
i u :> o A
Average generation rate = 1593X / 1440 = 1.176 X
Secondary (2°) loading = Generation • Leakage = [0.05 (1376 X) + 0.70
(317 X)] / 1440 = 0.202 X
V
Primary collection efficiency = [(Generation - 2° loading) / Generation]
' 100% = Cd.176 X - 0.202 X) / 1.176 X] • 100% = 83%
6-14
-------
Table 6-4 shows the calculations for a typical modern American-
design CWPB plant. Compared to Table 6-2, the plant is again assumed
to be computer-controlled, the function time for anode effects being
reduced from 6 minutes to 3. The function time for short side crust
breaking is zero since there is none; and that for long side crust
breaking is halved since a CWPB cell has one crust break area while an
SWPB cell has two. The cell remains closed during normal operation,
crust breaking, and raw material additions. The leakage rate with the
door closed is estimated at 3 percent. The cell end covers must be
removed for metal tapping and, presumably, for bath/metal measurement
with an estimated leakage rate of 8 percent. Removal of side covers
during anode effects and anode changing increases the leakage rate to 50
percent. The calculated primary collection efficiency of 95 percent
agrees with measured efficiencies at numerous CWPB plants.
Table 6-5 shows the calculations for the HSS plant that is retro-
fit case description A in section 6.3.1. For HSS plants, the function
times for anode changing and short side crust breaking are zero since
these operations are not performed. Plant A estimates that only 20 per-
cent of crust breaking is done with the doors open. This translates to
a function time of 5 minutes with the doors open; 21 minutes with the
doors closed. Plant A estimates the door is open a total of 8 minutes
per cell day. For Table 6-5, this means that only 3 minutes open time
remain for metal tapping, anode effects, flex raising and stud pulls
combined. Function times have been arbitrarily assigned for these
functions, and anode effects do occur even though the table shows a
6-15
-------
Table 6-4. CALCULATION OF PRIMARY COLLECTION EFFICIENCY FOR
TYPICAL AMERICAN-DESIGN CWP3 PLANTS
Cell Function
Fluoride
Generation
Severity Function Time,
Index (A) Minutes (B)
Generation Rate,
A ' B
Normal Operation
Anode Effects
Anode Changing
Metal Tapping
Bath/Metal Measure-
ment
Short Side Crust
Breaking
Long Side Crust
Breaking
Bath Addition
Alumina Addition
Other Controls
Totals
1 X
5.5 X
6 X
. 2 X
1 X
4.5 X
7 X
1 X
1 X
2.5 X
1385
3
8
6
2
0
13
1
20
2
1440
1385 X
15 X
48 X
12 X
2 X
OX
91 X
1 X
20 X
5 X
IRRf) X
Leakage
3%
3%
Average generation rate = 1580 X / 1440 = 1.097 X
Secondary (2°) loading = Generation • Leakage = [O.o3 (1502 X) + 0.50
(64 X) + 0.08 (14 X)] / 1440 = 0.054 X
V
Primary collection efficiency = [(Generation - 2° loading) / Generation]
100% = [(1.097 X - 0.054 X) / 1.097 X] ' 100% = 95%
6-16
-------
Table 6-5. CALCULATION OF PRIMARY COLLECTION EFFICIENCY FOR
ONE HSS PLANT - RETROFIT CASE DESCRIPTION A.
Fluoride
Generation
Severity Function Time, Generation Rate,
Cell Function Index (A) Minutes (B) A ' B Leakage
Normal Operation
Anode Effects
Anode Changing
Metal Tapping
Bath/Metal Measure-
ment
Short Side Crust
Breaking
Long (" Door
Side j Closed
Crust ' Door
Breaking , Open
Bath Addition
Alumina Addition
Other Controls
Flex Raise &
Stud Pull
1 X
5.5 X
6 X
2 X
1 X
4.5 X
7 X
7 X
1 X
1 X
2.5 X
1 X
1386
0
0
2
2
0
21
5
1
20
2
1
1386 X
0 X
0 X
4 X
2 X
0 X
147 X
35 X
1 X
20 X
5 X
1 X
5%
-
-
70%
5%
-
5%
70%
5%
70%
Totals 1440 1601 X
Average generation rate = 1601 X / 1440 = 1.112 X
Secondary (2°) loading = Generation • Leakage = [0.05 (1561 X) + 0.70
(40 X)] / K40 = 0.073 X
V
Primary collection efficiency = [(Generation - 2° loading) / Generation]
' 100% = [(1.112 X - 0.073 X) / 1.112 X] • 100% = 93%
6-17
-------
zero function time. Also, the fluoride generation rate for flex raising
and stud pulls has been assumed equal to normal operation -- participate
generation rate would not be equal. Cell leakages are assumed identical
to the French design SWPB plant since the cell has a fixed superstructure,
The calculated primary collection efficiency of 93 percent agrees well
with a plant estimated efficiency of 95 percent that is based on proto-
type testing—see section 6.3.1.3.
This method of calculating primary collection efficiency is not
very sensitive to errors in function time or duration of hood opening.
For example: at unchanged leakage for normal operation, all other
leakages listed in Table 6-2 were increased by 10 percent - all in the
same direction. The calculated primary collection efficiency decreased
less than 2 percent from that shown.
The 3 percent hood leakage under normal operation (Table 6-4) was
assumed to double to 6 percent. All other operating leakages were held
constant. The calculated primary collection efficiency decreased less
than 3 percent. Another calculation on this same CWPB cell assumed that
all cell openings were 3 times as frequent as shown in Table 6-4. The
primary collection efficiency decreased less than 2 percent from that
s hown.
The examples shown in Tables 6-2 through 6-5 apply to specific
plants and are given to illustrate this method for estimating primary
collection efficiency. Users of the method should - as a minimum -
determine the function times for plants that they want to check.
6-18
-------
Extrapolation of the above calculational procedures to VSS cells
is probably without theoretical justification since the hooding is
radically different. However, Table 6-6 shows such an estimate for
a typical VSS plant. The function times are the same as in Table 6-2,
except there is no anode changing. Stud blows have been ignored
since their severity index is unknown. It is assumed there is no
capture of emissions from all non-normal operations, and a 3 percent
leakage from the anode channel during normal operation. The calculated
primary collection efficiency of 80 percent is within the range of VSS
cell performance of 75 to 92 percent given in section 6.1.1.3.
6.1.3 Primary Exhaust Rates
Operating the cells at the proper primary exhaust rate is important
for efficiency primary collection. Too low an exhaust rate results in
a low collection efficiency; too high a rate results in the primary removal
equipment being oversized and in solids being needlessly entrained from
the cell surface into the equipment. The proper exhaust rates for a
given cell design cannot be empirically determined because the total open
area of an operating cell's hood is virtually impossible to calculate. Instec
6-19
-------
Table 6-G. ESTIMATE OF PRIMARY COLLECTION EFFICIENCY FOR
TYPICAL VSS PLANTS
Fluoride
Generation
Severity Function Time, Generation Rate,
Cell Function Index (A) Minutes (B) A ' B Leakage
Normal Operation
Anode Effects
Anode Changing
Metal Tapping
Bath/Metal Measure-
ment
Short Side Crust
Breaking
Long Side Crust
Breaking
Bath Addition
Alumina Addition
Other Controls
Totals
1 X
5.5 X
6 X
2 X
1 X
4.5 X
7 X
1
1 X
1 X
2.5 X
1367
6
0
6
2
10
26
1
20
2
1440
1367 X
33 X
0 X
12 X
2 X
45 X
182 X
1 X
20 X
5 X
1667 X
2%
100%
-
]
|
i
i
i
i
N 100%
!
i
1
Average generation rate =1667 X / 1440 = 1.158 X
Secondary (2°) loading = Generation • Leakage = [o.03 (1367 X) + 1.00
(300 X)] / 1440 = 0.237 X
Primary collection efficiency = [(Generation - 2° loading) / Generation]
1 100% = [d.158 X - 0.237 X) / 1.158 X] ' 100% = 80%
6-20
-------
the optimum exhaust rate is usually determined from a cell prototype.
This rate is that which will continuously maintain a sliqht negative
pressure drop across all the hood openings. The pressure drop can be
measured by sensitive pitot tubes and anemometers, or proper operation
can be visually checked by releasing smoke just outside the openings
and observing the resultant travel path.
The proper exhaust rate is most difficult to maintain when the hood
door is open. Some designs maintain it through "dual range ventilation"
as explained in section 6.1.4. Consideration should also be given to the
fact that the hood will inevitably deteriorate with time as the cell
deteriorates. Such deterioration can be held to a minimum, however, by
maintaining the hoods in proper condition and makinn sure that ooerators
exercise care in handling hood doors.
EPA personnel visited seven primary aluminum plants in the Spring
of 1973 to develop data for this guidelines document. For these seven
plants, Table 6-7 shows primary collection efficiencies and primary ex-
haust rates for retrofits either underway or completed. ' ' '
The age of the plant and of the control equipment (if different) is
also given. From Table 6-7 it can be concluded that:
1. HSS potlines generally require higher primary exhaust rates
than CWPB potlines to achieve the same primary collection
efficiency.
2. Older CWPB potlines generally require higher primary exhaust
rates than newer CWPB potlines to achieve the same primary
collection efficiency.
-------
Table 6-7 PRIMARY COLLECTION EFFICIENCY VERSUS STANDARD CUBIC
FOOT PER TON OF ALUMINUM PRODUCED FOR SEVEN PRIMARY
ALUMINUM PLANTS 3,4,13,16-18
Cell
type
CWPB
CWPB
CWPB
CWPB
SWPB
VSS
HSS
HSS
HSS
HSS
HSS
Plant
code
D
F
G
G
C
E
B - south
plant
B - south
\ plant
B - north
plant
A
i
A
t
Year plant
started up
1939C
1952
1958
1958
1965
1958
1941
1941
1968
1/2 - 1942f
1/2 - 1968
1/2 - 1942f
1/2 - 1968
Before/after,
retrofit
Both
Both
Before
After
After6
Both
Before
After
Both
Before
After
Primary
collection
efficiency, %
95
98
65
95d
85
81
80
87
95
94
95
Scf/ton,
Al x 106
5.05
4,78
4.11
4.11
3,44
0.67
5.06
7.85
6.81
6.57
7.09
-
FOOTNOTES:
a. Plants have the same codes here and in Section 6.3.
b. As of May 1975, retrofits were in progress at plants D, G, and B-
south; and had been completed within the previous five years at
plants F, C, E, B-north, and A.
c. Hoods, ducts, fans, and removal equipment were installed in 1949.
d. Collection efficiency will be increased by cell design changes to
effect tighter hood sealing.
e. Plant had no primary control before retrofit.
f. Present ducts, fans, and removal equipment were installed in 1951.
6-22
-------
3. Even with higher exhaust rates, older HSS potlines may not
readily achieve primary collection efficiencies as high as
those attainable by newer HSS potlines.
VSS cells characteristically have lower primary exhaust rates than
either prebake or HSS cells.
6.1.4 Ducting Layouts
Practice varies among aluminum plants as to the number of cells
connected with a single control system. In centralized or "central"
installations, an entire potline of 150 or more cells may be ducted to
a single control system; whereas, in decentralized or "courtyard"
installations where smaller control units are usually located in
courtyards between potlines, 20 or fewer cells may be ducted to each
control system. Figures 6-4 through 6-6 illustrate schematically several
possible ducting layouts for PB, VSS, and HSS potlines.19 The manifold
ducts are generally inside the potroom and elevated above and near the
cells. However, some SWPB and VSS potrooms have basements, the primary
exhaust being directed downward into manifold ducts in the basement.
The illustrative courtyard installations are patterned after existing
installations and were selected as the bases for the cost analysis in
Section 7. With a courtyard layout, each piece of removal equipment
is a separate module. To control a larger plant, additional modules
are added. The use of courtyard layouts thus eliminates the possible
economy of scale associated with control of larger plants. Further-
more, the control costs can be presented on a cost per ton of aluminum
capacity basis because the control cost per ton does not vary with plant
6-23
-------
COURTYARD SCHEME (20 CELLS PER MANIFOLD DUCT)
20
CELLS
20
CELLS
.MANIFOLD DUCT
MAIN DUCT
JL9 "LB
20
CELLS
T r
i\
A i
i j i " i |
6 i\ ti 6 d A
CENTRAL SCHEME (80 CELLS PER MANIFOLD DUCT)
MANIFOLD DUCT.
80
CELLS
\ T
MAIN
DUCT
..4
Figure 6-4. Primary collectiorW-systems: typical ducting layouts for a iq
single prebake potline with 160 cells, 2 rooms (R indicates removal equipment).
6-24
-------
COURTYARD SCHEME (10 CELLS PER MANIFOLD DUCT )
10 10
CELLS CELLS
MAIN
DUCT-
•MANIFOLD DUCT
J?d
~n
i
flfl
10
CELLS
10
CELLS
Figure 6-5. Primary collection systems: typical ducting layout for a
single VSS potline with 160 cells, 2 rooms (R indicates removal equipment).19
6-25
-------
COURTYARD SCHEME (10 CELLS PER MANIFOLD DUCT
MANIFOLD DUCT
10
CELLS
r-^~
9
R
10
CELLS
^r-'
gg
*
b
R
N
gg
«—
ifi
R
MAIN
DUCT
W
*•
i
\b
/ ' ' 1
P\P/ gg gg gg g
R
*•
to
I 1
-GO
-0
*m
*B
fl fl flfi fld GO o
10 10
CELLS CELLS
CENTRAL SCHEME (80 CELLS PER MANIFOLD DUCT)
80
MANIFOLD DUCT . CELLS
f \ S
9 \ 9 9 9
i i
i i
I 1
i i
i A i
MAIN ,
DUCT 1
Figure 6-6. Primary collection systems: typical ducting layouts for a
single HSS potline with 160 cells, 2 rooms (R indicates removal equipment).19
6-26
-------
size. Table 6-8 shows the gas volume relationship to aluminum production
capacity that was used in Sectiqn 7 to determine the required primary
20
control equipment size, and the sizes of the courtyard primary control
device modules used as the bases for the cost estimates.21 For com-
parison, Table 6-8 also shows the gas volume relationships and equipment
capacities used for secondary control.20'21... For an equivalent production
capacity, secondary removal equipment must be larger by at lea.st an
order of magnitude.
Table 6-8. GAS VOLUMES AND CONTROL DEVICE MODULE SIZES FOR
ECONOMIC IMPACT ANALYSI$20,21
Gas volume to primary control
device, 106 acf/ton Al
Primary control device module size
Dry systems, acfm
Wet systems, acfm
Gas volume, to secondary control ...
device, 106 acf/ton Al
Secondary system equipment capacity,
106 acfm
Cell tvpe
Prebake
5.0
100,000
82,000
50
10
vss
1.0
10,000
7,000
70
10
HSS
7.0
70,000
70
10
Table 6-9 gives 1975 primary collection system capital costs for
courtyard and- central installations.22 With a central layout, the
ductwcrk is larger and longer, and thus more expensive.
For a specific retrofit, it is not possible to generalize as to
which approach is more economical. Central installations are used when
the courtyard is too narrow to install the primary removal equipment.
6-27
-------
Table 6-9. CAPITAL COST COMPARISON BETWEEN COURTYARD AND CENTRAL
PRIMARY COLLECTION SYSTEMS^
($/annual ton of Al at full capacity, new construction)
(December, 1975)
Cell hoods and
branch duct
Manifold duct
Main duct
Total
CWPB
Courtyard ! Central
8.78
15.46
7.98
32.22
8.78
35.44
3.61
47.83
HSS
Courtyard
8.78
25.45
14.09
48.32
Central
8.11
48.85
3.72
60.68
Dry scrubbing systems—like the fluidized bed and injected alumina processes--
require particularly wide courtyards. A courtyard that is 50 to 60
feet wide may be too narrow for retrofitted dry scrubbing equipment, but
wide enough for retrofitted wet scrubbing equipment. Of the nine plants that
EPA personnel visited in developing this document, the two that had courtyard-
installed dry scrubbing retrofits had courtyards that were 100 to 150 feet wide,
Other considerations, such as flexibility by provision of duct inter-
connections for continued pollution control when part of a control system
may be out of service, and the ease of cleaning deposits from the inside of
ducting, may influence the design of ducting layouts. Maximum collection
efficiencies are realized when the designs provide for continuous exhausting
of all operating cells through removal equipment even when parts of a
potline are being serviced, and when dampers are available to increase the
air flow rate from a cell that may have part of its hooding removed for
cell working or anode replacement (dual range ventilation). Duct pluggage
is a problem in HSS potlines and in poorly operated VSS potlines because
unburned hydrocarbon tars will condense in the ductwork.
6-28
-------
6 2 POTROOM AND ANODE BAKE PLANT RETROFIT REMOVAL EQUIPMENT AND
ITS PERFORMANCE
This section discusses potroom primary and secondary removal
equipment, along with anode bake plant controls. Although the intent
is to describe retrofit operation and performance, there is no known
difference in the operation and performance of a specific piece of re-
moval equipment on a new versus a retrofitted primary aluminum plant.
Hence, the descriptions should apply to both new and retrofit instal-
lations. The removal equipment considered falls into three classes:
a. Dry scrubbing equipment suitable for potroom primary control.
b. Wet scrubbing equipment suitable for potroom primary control
and anode bake plant control.
c. Wet scrubbing equipment suitable for potroom secondary control.
Finally, potroom and anode bake plant best retrofit performance
is summarized. Evolution rates in Section 5.3 are combined with best
retrofit collection and removal efficiencies in Sections 6.1 and 6.2 to
show the overall effect on potroom total fluoride emissions. Evolution
rates in Section 5.3 are combined with best retrofit removal efficiencies
from an EPA contract study to estimate controlled anode bake plant
performance.
6.2.1 Potroom Primary Dry Scrubbing
Two types of dry scrubbing systems, fluidized bed and injected
alumina, are discussed in the following subsections. These systems
have been applied to many domestic and foreign plants. In addition,
6-29
-------
one domestic company has been installing its own dry scrubbing
system on two CWPB and two HSS plants. One of these HSS retrofits
is retrofit case description plant A in Section 6.3.1. This dry
scrubbing process is proprietary. However, operating conditions are
similar to those of the fluidizied bed and injected alumina processes
described below. Total fluoride removal efficiencies are projected to
be 98-98.5 percent for the two CWPB plants and 97-98 percent for the
two HSS plants.
6.2.1.1 Fluidized Bed -- Figure 6-7 is a flow diagram of the fluidizied
.bed dry scrubbing process. The fluidized bed dry scrubber employs a
fluidized bed of sandy alumina to contact and chemically absorb HF
in the cell gas followed by a baghouse to trap particulates. Floury
alumina will not fluidize and, hence, is not suited for this process
or for the injected alumina process.
Alumina is continuously fed to the reactor bed in amounts up to
TOO percent of the potline feed requirements, and the reacted bed
material overflows and is used as cell feed. Virtually all of the cell
gas particulate is trapped in the fluid bed -- perhaps by electrostatic
agglomeration. Fugitive particulate, primarily alumina, is stopped by a
bag filter mounted over the reactor. The bags are cleaned intermittently,
po oc
and the catch drops back into the fluid reactor bed.
The vendor of the fluidized bed dry scrubber reports that, with
proper operating and maintenance procedures, this system is capable of 98
percent particulate and 99 percent HF removal efficiencies on prebake
?7
potline effluents, or about 98.5 percent on total fluoride.
6-30
-------
CRUDE
ALUMINA
TANK
ROOF
EMISSIONS
' BLENDED
ALUMINA
TANK
FLUID BED
DRY SCRUBBER:
FAN
CELL GAS
Figure 6-7. Flow diagram for fluidized bed dry scrubbing process.
-------
The fluidized bed dry scrubber has been applied in foreign plants
to VSS cell gases with pilot lights or other devices used to ensure that
27
all burners are lit. The system has not been applied to HSS cell gases.
It has been installed in one domestic VSS plant with a projected re-
moval efficiency of 98.8 percent.
Dry scrubbing processes afford much less cooling for the cell
gas than wet scrubbing processes. Since conventional filter fabrics
like Dacron or Orion deteriorate above 275°F, the cell gas is usually
25
delivered to the fluidized bed at 275°F or below. Typical pressure
drops are 8 to 10 inches of water across the fluidized bed and 4 to 5
3 28
inches of water across the baghouse. ' A typical power requirement is
3 29
4.4 horsepower per thousand cubic feet per minute (hp/Mft -min).
6.2.1.2 Injected alumina — Figure 6-8 is a flow diagram of the
injected alumina dry scrubbing process. The process is similar in con-
cept to the fluidized bed — reaction of gaseous fluoride with sandy
alumina followed by baghouse collection of particulate -- except that
the reaction occurs by injecting the alumina into the flowing gas stream
rather than by passage of the gas stream through a fluidized bed. The
3
reaction occurs in a matter of seconds.
Alumina is continuously fed to the process in amounts up to 100
percent of the potline feed requirements. The removal efficiencies
of the injected alumina process are similar to those of the fluidized
bed. One major difference, however, is that loss of feed to the
fluidized bed (Figure 6-7) will not result in a loss in removal
efficiency for 8 hours thereafter because of the large alumina inventory
in the fluidized bed. Loss of feed to the injected alumina process on
6-32
-------
01
GO
CO
ROOF
EMISSIONS
iALUMINA
Figure 6-8. Flow diagram for injected alumina dry scrubbing process.
-------
the other hand can quickly result in a loss in removal efficiency due
to a low alumina inventory. Recycling a portion of the reacted
alumina back into the cell gas stream provides some insurance against
a total feed loss.
More than one vendor markets an injected alumina process designed
for prebake potlines. An Alcan and a Prat-Daniel-Poelman design are
explained in Section 6.3.3.
Cell gases from VSS potlines have higher concentrations of HF
than prebake cell gases, and they may contain unburned tar fumes.
Here again, alumina is injected into the flowing gas stream, but from
this point on, the Alcan process is modified slightly. Provision is
made to separate the bulk of the alumina containing adsorbed HF from
the portion containing unburned hydrocarbons. The latter minor quantity
of alumina is calcined to remove the tar prior to being returned to the
cells along with the main portion of the collected alumina. This system
3D
does not require that all VSS cell burners be lit all the time.
Comments on temperature limitations for the fluidized bed also
*
apply to the injected alumina process. A typical pressure drop is
6 inches of water across bag filters operating at an air-to-cloth ratio of
3 27
about 6 ft /min per square foot of filter area. A typical power
requirement is 2.2 hp/Mft3-min.29
6.2.2 Potroom Primary and Anode Bake Plant Wet Scrubbing
A potroom primary wet scrubbing scheme that gives removal efficiencies
comparable to dry scrubbing is the combination control of spray tower
followed by wet ESP. This combination control is most frequently applied
6-34
-------
to VSS and HSS plants. Spray tower-wet ESP or wet ESP-spray tower con-
trols also effectively remove particulate and fluoride from anode bake
plant exhaust. Since most of the latter fluoride is gaseous, the spray
tower — and not the ESP -- controls fluoride emissions.
6.2.2.1 Spray Tower — The term spray tower is applied to gas scrubbing
devices in which the gas passes through an enclosure at relatively low
velocity and is contacted by water, alkaline liquor or limed water
liquor sprayed from headers usually in counterflow with the gas. In pre-
bake or HSS potline service, the units may range from 38,000 to 630,000
actual cubic feet per minute capacity and may spray from 1.7 to 10.0
gallons of liquor per thousand cubic feet of gas. A typical spray tower
in prebake service uses water or limed water and consists of an open
top redwood tower, 12 to 15 feet in diameter and 40 to 70 feet high,
with cyclonic inlet breeching and a mist eliminator at the top. Liquor
may be sprayed down from the top or at several elevations in the tower.
Spray towers operate at a low pressure drop, typically 1 inch
32 3
of water. Typical power reauirements are 0.4 to 0.9 hp/Mft -min for
prebake service, 1.0 to 1.3 hp/Mft -min for VSS service, and 0.3 to 0.5
3 29
hp/Mft -min for HSS service. Spray towers cool the cell gas stream
to near ambient temperatures.
Properly operated and maintained spray towers can achieve removal
efficiencies for potline HF in percentages ranging from the low to hioh
nineties. Compared with other types of wet scrubbing equipment, spray
towers show relatively low removal efficiency for fine particulates.
Spray towers in HSS service appear to perform less efficiently than similar
6-35
-------
scrubbers in prebake or VSS service. This has been suggested to be the
result of an interference by the hydrocarbons in the wetting of the particu-
31
lates and diffusion of HF to the spray droplets.
Typical gaseous fluoride removal efficiencies are 95 percent
for prebake potlines, 99 percent for VSS potlines, 93 percent for HSS
potlines, and 96 percent for anode bake plant ring furnaces. '33 Typical
particulate fluoride removal efficiencies are 80 percent for prebake
33
potlines, 75 percent for VSS potlines, and 64 percent for HSS potlines.
Additional information on spray towers can be found in many texts
including references 32 and 34. Two points worth mentioning are:
1. As with any mass transfer unit, the added increase in
fluoride removal efficiency drops off rapidly with each
subsequent mass transfer stage; therefore, the attainment of
fluoride removal efficiencies that are higher than those
previously given is most difficult.
2. Exhaust from a redwood spray tower that is capped with a
cone mist eliminator can be easily ducted to other control
equipment (such as an ESP) downstream of the tower. On the
other hand, it is difficult to cap redwood towers that were not
originally designed with caps. Such capping is necessary in
4
ducting the exhaust to downstream control equipment.
6.2.2.2 Wet Electrostatic Freetpttator CESP) - - The electrostatic
precipitator is a relatively large chamber through which cell gas streams
pass at low velocity, usually 3 to 5 feet per second (ft/sec). In its
usual form, high negative voltage corona discharge wires are suspended
t>-36
-------
across the air stream and grounded collector plates form parallel
passageways for the air. The ionizing field surrounding the discharge
wires ionizes part of the gas stream and imparts electric charge to most
particles, some positive but most negative. Positively charged particles
migrate toward the discharge wires and negatively charged particles
migrate to the grounded collection plates. When collected particles
lose their charges, they tend to agglomerate and collect on the surface.
The removal efficiency of electrostatic precipitators for many
kinds of particulate is improved if the entering gas is conditioned
by raising its moisture content. When applied to VSS or HSS potlines,
precipitators are usually preceded or followed by a spray tower that
removes most gaseous fluoride. Spray towers preceding precipitators
also condition the gas. However, for some HSS retrofits, space limita-
tions and requirements for balanced ducting layouts have necessitated
removal of the spray towers that would have otherwise preceded the
ESPs. In these instances, effective gaseous fluoride control and
conditioning is achieved by a scrubbing section in the ESP inlet.
Electrostatic precipitators fall into two categories: dry ESPs
where the collected particulates are knocked off the plates and wires
by mechanical rapping to be gathered dry in a hopper; and wet ESPs where
the plates and wires are washed with falling water or electrostatically
collected mist with the particulates removed as a slurry. A dry ESP
followed by a spray tower is not widely applied as primary equipment for
Soderberg cells since it does not prevent the emission of a blue
hydrocarbon haze.
Unlike many types of control equipment, electrostatic precipita-
tors may be designed for almost any selected efficiency. By using
6-37
-------
conservative design dimensions, by controlling humidity of the incoming
gas, and by operating at high voltage, both wet and dry precipitators
can achieve 98 to 99 percent removal of potline cell gas particu-
lates.35 Total fluoride removal efficiencies for scrubber-wet ESP
controls vary from 99.2 to 99.9 percent on domestic VSS plants, and
from 95 to 99 percent on domestic HSS plants.
Electrostatic precipitators operate at a pressure drop of less than
1 inch of water. Typical power requirements for the wet ESP are 0.66 to 1,
o 3 29
hp/Mft -min for VSS service and 1.4 hp/Nft -min for HSS service." Liquor
3 29
requirements are 5 to 10 gal/Mft of gas. Because wet ESPs are usually
preceded by a wet scrubbing device, they operate at near ambient tempera-
tures in potline service. For anode bake plant service, a typical
power requirement is 3.8 hp/Mft -min, and typical liquor requirements
are 0.3-0.4 gal/Mft3 of gas.29
Additional information on ESPs can be found in many texts including
references 35,36, and 37. Three points worth mentioning are:
1. The design of the ESP should insure that the plates are not
likely to warp in service. Such warping will cause the affected
sections to short out with a resultant loss in removal efficiency.
2. The design of the ESP should insure that the plates do not
develop dry spots and short out. One plant reports that
this problem was overcome by installing internal sprays to
continuously irrigate the plates.3
3. Wet ESPs in potline service are subjected to corrosive
operating conditions. For this reason, the ESP internals are
6-38
-------
usually of stainless steel construction, and the interior
steel shell walls are lined or coated. Steel ESPs are likely
to corrode rapidly unless the composition and pH of the feed
liquor are carefully controlled."3
6.2.3 Potroom Secondary Wet Scrubbing
For practical purposes, choice of potroom secondary control is
\
limited to the spray screen scrubber. The term spray screen scrubber is
applied to wet scrubbing equipment in which the liquor is sprayed into a
gas stream and on to screens or open mesh filters enclosed in a plenum
chamber. The assembly also usually includes a mist eliminator. Gas
flow may be powered by exhaust fans, or may be moved by unpowered con-
vection. Figures 6-9 through 6-12 illustrate several designs of spray
screen scrubber installations that have been used in the primary
aluminum industry. " The particulate removal mechanisms are inertial
impaction on and interception by the liquid droplets or filters. The
gaseous removal mechanism is absorption into the liquid droplets.
The low gas pressure drop across spray screen scrubbers and their
relatively low power cost recommends them for secondary, or potroom,
scrubbing service. For secondary prebake service, typical power require-
ments are 0.3 to 1.0 hp/Mft3-min and typical liquor requirements are 3 to
10 gal/Mft3 of gas. 29
Table 6-10 gives total fluoride secondary removal efficiencies for
3 13 17 42 43
spray screen scrubbers at six existing U.S. plants. ' ' ' '
Without primary control, all the fluoride generated at the cell is
directed to the secondary scrubbers that remove 80-85 percent of the total
b-39
-------
SPRAY SCREEN
\ i /
\ \ '
-.
V\\!/,
'
Figure 6-9. Unpowered roof spray screen.38
6-40
-------
CLEANED AIR
AIR
FROM
POTROOM
SCREEN
FOUR BANKS OF
FINE SPRAYS
COARSE
SPRAYS
-<
•<
<
k
-c
-<
AIR
FROM
POTROOM
Figure 6-10. Powered potroom spray screen scrubber. 39
-------
SEPARATOR
SECTION
FAN
Figure 6-11. Powered spray screen scrubber. tf°
-------
EXHAUST FAN
SPRAY NOZZLES
SPRAY CHAMBER.
PLASTIC WIRE MESH
SPRAY NOZZLES
WATER COLLECTING TROUGH
Figure 6-12. Powered monitor spray screen scrubber.41
-------
Table 6-10. PERFORMANCE OF SPRAY SCREEN SECONDARY SCRUBBERS AT
SIX EXISTING PRIMARY ALUMINUM PLANTS
Total Fluoride Secondary Removal Efficiency
Cell Typea
SWPB
SWPB
SWPB
VSS
VSS
VSS
Without Primary
Control
80
-
85.5
-
-
-
With Primary
Control
87
71b
-
75
75
80C
aSWPB - side-worked prebake; VSS-vertical stud Soderberg.
Projection based upon limited testing.
Projection based upon detailed contractor study.
fluoride. With primary control at SWPB and VSS plants, only 10-20
percent of the fluoride generated at the cell escapes the hooding and
is directed to the secondary scrubbers. At this reduced fluoride loading,
Table 6-10 shows that the scrubbers have a removal efficiency of 75-80
percent, on the average. The two secondary scrubber efficiency readings
of 80 and 87 percent for the SWPB plant - first line of table - were
taken at different times, and emission variability and sampling error
are factors to help explain why the two efficiency figures seem reversed;
i.e., the secondary efficiency should be higher without primary
control. However, the 87 percent reading was the result of 93 tests,
6-44
-------
24 hours per test, and 3 tests per week. At this same time, 92 similar
tests showed a primary collection (hooding) efficiency of 83 percent.
Thus, the value of 87 percent for secondary removal efficiency in the
presence of primary collection seems firmly supported. In addition,
one aluminum company had plans to build a new CWPB plant that included
primary control and spray screen scrubber secondary control. They
anticipated achieving a 98 percent primary collection efficiency, so that
only 2 percent of the fluoride generated at the cell would be directed to
the secondary scrubbers. At this very low loading, they projected a
secondary removal efficiency of 75 percent. Hence, based on this plant's
projected performance and Table 6-10, a secondary removal efficiency of
75 percent should be achievable at almost all plants adding on secondary
46
control to existing primary control. Although the above mentioned plant
was not built, it was proposed to a State that has extremely strict fluoride
emission limitations, and was based on designs by a major engineering firm
that is highly experienced in the design of aluminum plants and their
emission controls. It is therefore clear that both the aluminum manu-
facturer and the designer were confident that their proposed secondary
scrubber could achieve 75 percent total fluoride removal after primary
control.
In practice, a scrubber designer would balance costs of the simultaneous
addition of packing depth and wash water flow increase; both of these
design factors work to produce increased fluoride removal efficiency.
An additional factor is, that coarser particulate sizes are the easier
to remove by water scrubbing; addition or improvement of primary hooding
tends to preferentially remove the fine particulate from the secondary
6-45
-------
stream, thus shifting the size distribution to the larger sizes which
are much easier to scrub out in the secondary scrubbers.
Obviously, either increased packing depth or increased wash water
flow adds to costs. This is why secondary retrofits should not be
required in most locations (See Section 8.3) and then only if there is
a fluoride problem and costs are balanced against fluoride reduction
benefits. In addition, secondary scrubbing is rather energy intensive.
Although more sophisticated scrubbing devices, such as the cross-
flow packed bed scrubber, can achieve higher removal efficiencies for
both particulates and HF than does the spray screen, the costs are
30 to 100 percent greater and the cost effectiveness much lower when
applied to secondary treatment. It is the consensus of the industry
that, for secondary treatment in combination with primary control, the
cost differential would be more effectively invested in improved primary
collection and removal equipment. Among the alternative secondary
scrubbers only the spray screen is considered economically feasible.
Two points worth mentioning are:
1. Although many plants with secondary scrubbing use once-through
water, tighter effluent regulations will require that the water
be treated and recycled. Recycled treated water has the added
advantage of inhibiting corrosion.
2. Although Figure 6-9 through 6-12 show secondary equipment lo-
cated on the roof, the potroom roof at many plants may not
support the equipment. This may be particularly true in northern
plants that are subjected to heavy snowfalls. Installing secondary
controls in the courtyard may be time-consuming and more expensive
than installing them on the potroom roof.
6-46
-------
6.2.4 Summary of Best Retrofit Performance
6.2.4.1 Potrooms -- Table 6-11 shows the effects of various degrees of
emission control on total participate and gaseous fluoride potroom emissions.
For all four cell types, typical upper and lower evolution limits and
averages- are given, each based on actual values reported in Table 5-2.
Best retrofit primary collection efficiencies are taken from Section 6.1,
an upper and a lower limit being given for all cell types except CWPB.
Best retrofit primary removal efficiencies are taken from Sections 6.2.1
and 6.2.2, an upper and a lower limit being given for both Soderberg cell
types. Best retrofit secondary removal efficiency is taken from Section
6.2.3.
Table 6-11 shows that CWPB plants with or without secondary control
consistently achieve lower average fluoride emissions than do other
cell types. However, CWPB emissions without secondary control are
matched at those HSS plants that achieve the upper limits of primary
collection and removal; also, HSS and VSS plants having such upper
limits and additional secondary control perform comparably to CWPB plants,
but SWPB plants do not.
The emissions in Table 6-11 bracket the performance of individual
plants, but any given emission does not necessarily correspond to that
of any specific plant. Known and projected emissions for some actual
plants with best primary and secondary control are given in Table 7-3. In
addition, Section 6.3.4 gives capsule retrofit descriptions for ten
actual retrofits including the after-retrofit emissions.
6-47
-------
Table 6-11. PERFORMANCE OF BEST RETROFIT EMISSION CONTROLS FOR PRIMARY ALUMINUM POTROOMS
Cell
Type
CWPB
II
II
SWPB
II
II
II
II
II
vss
"
II
"
II
II
"
II
II
II
II
"
HSS
u
it
"
"
"
"
"
"
"
n
With or
Average
Fluoride
Evolution,
Lb F/Ton Al
25
40
65
35
M
45
II
55
II
30
II
II
II
45
II
II
II
55
II
II
II
30
n
H
II
35
II
II
II
45
"
"
Without Secondary
Primary
Collection
Efficiency. %
95
"
"
85
80
85
80
85
80
90
II
75
II
90
II
75
II
90
II
75
II
95
11
85
II
95
II
85
H
95
M
85
"
Control
Primary
Removal
Efficiency, %
98.5
"
11
98.5
II
II
II
" II
II
99.9
98.5
99.9
98.5
99.9
98.5
99.9
98.5
99.9
98.5
99.9
98.5
99.0
96.0
99.0
96.0
99.0
96.0
99.0
96.0
99.0
96.0
99.0
96.0
Without Sec-
ondary Control
Average
Fluoride
Emission,
Lb F/Ton Al
1.61
2.57
4.18
5.70
7.42
7.32
9.54
8.95
11.66
3.03
3.40
7.52
7.84
4.54
5.11
11.28
11.76
5.55
6.24
13.79
14.37
1.78
2.64
4.76
5.52
2.08
3.08
5.55
6.44
2.68
3.96
7.13
8.28
With Secondary Control
Secondary
Removal
Efficiency, %
75
"
M
75
II
II
"
II
II
75
II
»
»
II
II
II
II
"
II
II
n
75
II
II
II
II
II
II
II
II
II
II
II
Average
Fluoride
Emission,
Lb F/Ton Al
0.67
1.07
1.74
1.76
2.17
2.26
2.79
2.76
3.41
0.77
1.16
1.90
2.21
1.17
1.73
2.85
3.32
1.42
2.12
3.48.
4.06
0.66
1.52
1.38
2.14
0.77
1.77
1.61
2.50
0.99
2.27
2.07
3.22
6-48
-------
6.2.4.2 Anode Bake Plants — Table 6-12 shows the average evolution,
best retrofit removal efficiency, and resultant emissions for gaseous,
particulate and total fluoride at anode bake plants. All of the
quantities are expressed as pounds of pollutant per ton of aluminum
produced, not per ton of carbon anode produced. The evolution rates
in Table 6-12 are the same as those for ring furnaces shown in Table 5-4.
Hence, they are intermediate between uncontrolled total fluoride emission
factors of 1.6 Ib/ton representing poor cleaning of the anode butts and
of 0.4 Ib/ton representing proper cleaning. The removal efficiencies
are taken from the EPA contract study.48
By comparing Table 6-12 with Table 6-11, it can be seen that the
best retrofit anode bake plant total fluoride emission is much less
than the best retrofit potroom total fluoride emissions. Total fluoride
emissions from CWPB, VSS and HSS potrooms that have secondary control,
lowest possible evolution rates, and best possible primary collection and
removal efficiencies are about the same as uncontrolled total fluoride
emissions from anode bake plants.
Table 6-12. SUMMARY OF ANODE BAKE PLANT BEST RETROFIT PERFORMANCE
Gaseous Fluoride
Particulate Fluoride
Total Fluoride
Average
Evolution,
Ib/ton Al
0.816
0.044
0.859
Removal
Efficiency,
%
95
80
94.2
Average
Emissions,
Ib/ton Al
0.041
0.009
0.050
6-49
-------
6.3 RETROFIT CASE DESCRIPTIONS
This section contains three case descriptions of actual potline
retrofits underway or completed in the United States as of the summer
of 1973. These case descriptions are for best retrofit controls for
specific plants and are not necessarily representative of the industry.
The section contains descriptions, engineering information, performance
results, and cost data for actual retrofits. Since it is unlikely that
any two aluminum plants will face the same problems in retrofitting, one
of the objectives of this section is to make the States aware of many
of the varying problems that different plants may encounter. No attempt
has been made to match these case descriptions with the control equipment
specified in Section 6.2 or to include all types of cells and control
equipment.
Instead, engineering descriptions of actual retrofit emission
controls at three primary aluminum plants are presented. Each case
includes a description of control units, ductwork, supports, fans and
other accessories, along with practical considerations such as inter-
ferences, spatial relationships, and procurement and construction
difficulties. Capital and operating costs are accompanied by the
overall fluoride reductions obtained by the expenditures outlined.
The result is a description of some retrofit controls, each of which
is practical for its plant and for its owners and each of which will
meet the performance described. For a process as complex as a primary
aluminum plant, it is evident that a retrofit control must be tailor-
made and should not be generalized as to costs or even as to method
of emission control.
6-50
-------
•The coverage of the three detailed retrofits in this section is
primarily based upon a trip report covering visits to several primary
aluminum plants plus supplementary drawings, emission estimates, and
cost data subsequently provided by the plants visited. Most of the
drawings are considered proprietary in nature and hence are not referenced
in the case descriptions. Retrofits under construction are described
for mid-1973, but estimated construction lead times, completion dates,
emission data, and costs have been updated.
Following the three detailed case descriptions is a subsection
containing capsule descriptions of the three retrofits and of retrofits
at seven other plants. The presentation includes a summary of the actual
retrofit emission reductions and costs for the ten plants.
Although EPA conducted source tests at several retrofitted plants
in developing the data base for the standard of performance for new
primary aluminum plants, most of the detailed and capsule descriptions
are for plants other than the ones tested. Furthermore, descriptions
are not given for all the plants tested by EPA. The lettered plant
codes in the case description are not meant to correspond to those in
2
the background document for the new plant standards of performance.
Whenever possible, emission data furnished by the companies have been
included with the ten case descriptions contained herein.
6-51
-------
6.3.1 Plant A--HSS Cells—Primary Dry Scrubbing Retrofit49
This plant has three operating HSS potlines (lines 1, 2 and 4)
with wet scrubbers presently used for primary emission control. Plans
are to retrofit the primary exhaust with two dry scrubbing systems,
one for lines 1 and 2 and one for line 4. The potlines have no
secondary control and none is planned.
Dry scrubbing has not previously been installed for HSS primary
control in the United States because of the inherent plugging tendency
of the unburned hydrocarbons in the primary exhaust. Nevertheless,
tests on a company prototype have shown that the system planned for
plant A effectively removes fluoride, particulate, and hydrocarbon
from the primary exhaust. However, the long-term effect on metal
purity and potline operation when using the recovered materials in a
"closed loop" system is not yet known.
Potline operation, present controls, and the planned retrofit are
now described, first for lines 1 and 2 and then for line 4. Next the
present emissions and the emissions expected after total retrofit are
presented. Finally, capital and annual operating costs for the total
retrofit are estimated.
6.3.1.1 Engineering Description - Lines 1 arid 2
6.3.1.1.1 Potline operation — Lines 1 and 2 have a total capacity of
40,000 short tons of aluminum per year (ton/year). Each line has 120
cells set in four rows in one potroom, 30 cells per row. The lines
were built in 1942 and the present primary controls were installed in 1951.
-------
The cells in lines 1 and 2 are unique in that their anodes are
channel-less. The casing remains stationary, much as in a VSS cell.
The current-carrying studs move downward with the anode by means of
vertical slots in the casing. Anode weight limits usage of this
design to small HSS cells.
The cells in lines 1 and 2 have total-enclosure hooding with hood
doors extending the full length of both sides of the cell. A mechanized
time-feed system adds alumina to the cells with the hood doors closed.
This system consists of four sets of vertical crust punchers and
alumina feeders, two sets on each side of the cell. At preset time
intervals, the puncher makes a hole in the crust and the feeder
immediately dumps alumina into the hole. The result of this system
is that over 90 percent of the alumina is fed with the doors closed
and the cell doors are only open an average of 8 minutes per cell-
day. The cell doors do have to be opened to work both sides of each
cell every 24 hours and to raise the flexible current connectors and
pull the bottom row of studs every 10 to 12 days. They also have to
be opened every 24 hours to tap the molten aluminum from beneath the
cryolite bath.
6.3.1.1.2 Present controls ~ Two ducts, one on each end of each cell,
pick up the primary exhaust from the top of the cell hoodinq enclosure
and carry it to a circular manifold duct. Primary exhausts from 15
cells (half of one row) are manifolded together. Total manifold
exhaust is 30,000 acfm at 150°F, or 2000 acfm per cell.
6-53
-------
Each manifold originates in the middle of the potroom and grows
in size from a diameter of 2 feet to a diameter of 4 feet as it proceeds
to a fan that is outside on the end of the potroom. Each fan is driven
by a 100-hp motor and is upstream of a redwood spray tower. Hence,
lines 1 and 2 have a total of 16 towers, 4 on each end of each potroom.
Each tower is capped with an inverted cone that serves as a mist elimina-
tor. The capping was strictly for design purposes, and there was never
any intention of adding on control equipment downstream of the towers
at a later date.
The plant has experienced emission control problems on lines 1
and 2 because of improper hood sealing and duct pluqgage. The cell
cathode shells tend to bow down in the center of the sides due to lack
of proper structural support. This tendency makes effective sealing
of the hood doors difficult. Duct pluggage is caused by the hydrocarbons
present in an HSS primary exhaust and by the fact that the ducts in
lines 1 and 2 are retrofitted and contain design flaws.
6.3.1.1.3 Aqueous waste -- Water passes once through the spray towers
and, along with the water discharged from the scrubbers on line 4,
goes to water treatment. At the water treating facility, water from
the scrubbers is fed to a circular, ground-level, open-top reactor
tank about 25 feet in diameter. Lime is added as a slurry from an
elevated 14-foot by 25-foot tank. The water is continuously mixed
and bled off to a second circular, ground-level, open-top tank for
continued mixing and reaction but no lime addition. This second
6-54
-------
tank discharges to a large porid for solids settling. About three-
fourths of the water is pumped out of the pond and recycled, and about
one-fourth is discharged to a nearby waterway. The effluent discharged
to the waterway has a pH of 6.8 to 7.0.
Upon conversion of the potlines to dry scrubbing, no scrubbing
water will be discharged from the plant. The existinq water treating
facility, consisting of reactor tanks and a settling pond, will remain
to handle the plant's cooling water, but there will be no lime treating.
6.3.1.1.4 Planned retrofit — The two potrooms comprising potlines 1
and 2 are oriented in a north-south direction. The 60-foot width
of the courtyard between the potrooms will not permit courtyard
installation of the control equipment. The planned retrofit consists
of ducting all the primary exhaust from lines 1 and 2 to 18 dry scrubbers
located together in an area north of the potrooms. This is termed
a central, as opposed to a courtyard, installation.
Figure 6-13 is a layout of the retrofit for lines 1 and 2, Table
6-13 lists the major retrofit items. The ductwork inside the potrooms
will remain unchanged, and the existing fans and spray towers will be
bypassed. The two circular ducts listed as item 1 in Table 6-13 will
pick up the exhaust from the eight manifolds servicing the south halves
of both potrooms and convey it north between the potrooms. A new damper
will be required at the end of each manifold.
6-55
-------
en
cr>
1000-ton
ALUMINA BIN
X 10 ton/hr
C AIR SLIDES
DRY SCRUBBER
UNIT
150,000 acfm FAN
WITH 800-hp MOTOR
ALUMINA AIR SLIDE
UNLO
ADING
STATION
20 ton/hr
AIR SLIDE
Jv_ 20 ton/hr
vJ>*--AIR SLIDE
EXISTING 900-ton
ALUMINA BIN
FROM SOUTH
END MAIM I-
HOLDS
Figure 6-13. Retrofit layout -- plant A — lines 1 and 2.
-------
Table 6-13. MAJOR RETROFIT ITEMS—PLANT A—LINES 1 AND 2
1. Two circular elevated mild steel ducts, each 8 feet in diameter
and 700 feet long, convey primary exhaust from the south halves
of potlines 1 and 2 to the retrofit area north of the ootlines.
Each duct is designed to handle 150,000 acfm of exhaust at 180°F.
2. Four fans, each driven by an 800-hp motor, designed to handle
150,000 acfm of exhaust at 180°F and 26 inches of water total
pressure drop. Each fan has two inlet dampers and one outlet
damper.
3. Two ground-level rectangular mild steel ducts, each about 200
feet long. Each duct feeds nine reactor-baghouse dry scrubbinq
units, reducing in size from a height of 13 feet and a width
of 6 feet to a height of 4 feet and a width of 2 feet.
4. Eighteen mild steel reactor-baghouse dry scrubbers, set in two
rows of nine each. Each scrubber is a rectangular box, 11 feet
by 42 feet and 15 feet high. The top of each scrubber is 40
feet above the ground. Each scrubber has four compartments. Each
compartment has a gas inlet section shaped like an inverted
rectangular pyramid on the bottom and a stack on the top, or a
total of 72 stacks for the 18 scrubbers. The stacks are 15 feet
high, discharging 55 feet above the ground. Each scrubber is
designed to handle 40,000 acfm of exhaust at 180°F. The bag-
houses on each scrubber are cleaned by air pulse, requiring 90
psig compressed air. Each scrubber requires one damper in the
inlet gas line, air activated gravity alumina feed and discharge
devices, and five manually operated alumina shut-off gates.
5. Alumina unloading station containing a hopper with screen to
receive alumina from railroad dump cars. The station is located
immediately north of the potrooms.
6. Combination mild steel air slide about 400 feet long. It is
designed to simultaneously convey 50 ton/hr of fresh alumina
from the new unloading station to the new fresh alumina storage
bin and 20 ton/hr of reacted alumina from the scrubbers to the
four existing 900-ton reacted alumina storage bins that are
located at the north ends of the potrooms. Each end of the air
slide is preceded by an equivalently-sized air lift.
7. Mild steel 1000-ton fresh alumina storaae bin located near the
18 dry scrubbers with high, intermediate, and low level bin
indicators. The bin is circular, 38 feet in diameter, with conical
top and bottom. Straight side height is 19 feet. The bottom of
the bin is 45 feet and"the top is 95 feet above the ground.
6-57
-------
Table 6-13 (continued). MAJOR RETROFIT ITEMS—PLANT A—LINES 1 AND 2
8. Two 10 ton/hr mild steel air slides, each slide conveying alumina
to nine dry scrubbers and equipped with a flow control valve and
a manually'operated shut-off gate. Total length of each slide is
about 230 feet.
9. Two 10 ton/hr mild steel air slides, each slide conveying reacted
alumina from nine scrubbers to an activator tank that services
both air slides and feeds the 20 ton/hr air slide in item 6. Total
length of each slide is about 190 feet.
10. Three small cylonic dust collectors for alumina transfer and
storage operations.
11. A 30- by 4U-foot control building with a power substation.
At the north end of the potrooms, exhaust from the eight manifolds
servicing the north halves of both potrooms joins the two central ducts
as depicted in Figure 6-13. A new damper will be required at the
end of each manifold. The central ducts cross a plant roadway and lower
into rectangular ducts 7 feet wide and 13 feet high. Each of these ducts
handles the total primary exhaust from one potroom of 300,000 acfm at
180°F. Each duct splits and feeds two of the four fans installed on a
north-south axis. As shown in Figure 6-13, the two north fans move
exhaust from potline 2 to feed the nine dry scrubbers on the north side;
while the two south fans handle potline 1 and feed the dry scrubbers
on the south side.
The per-cell primary exhaust rate on lines 1 and 2 should increase
from 2000 acfm at 150°F to 2500 acfm at 180°F. Hence, the primary
collection efficiency on lines 1 and 2 should increase. The ducts
inside the potrooms are undersized for handling the increased flowrate,
resulting in a high fan pressure drop requirement and a resultant
increased power cost. However, it was considered to be more economical
to leave the internal ducts unchanqed.
b-58
-------
Dry scrubbing involves reaction of gaseous fluoride with alumina
followed by baghouse collection of fluoride and non-fluoride particulate.
The scrubbers are sized so that any one in a set of nine can be off-
line at a given time and the remaining eight will still handle the 300,000
acfm exhaust.
Considerable solids handling is involved in the retrofit. Presently
each of the four existing 900-ton alumina storage bins has its own small
alumina unloading station. A new larger alumina unloading station
is needed to supply fresh alumina to the dry scrubbers. The existing
stations will be left as a backup. The percentage of the alumina cell
feed that will have to pass through the dry scrubbers is not known, but
all of the solids handling equipment is befno designed for 100 percent
feed. Alumina will normally pass once through the scrubbers before
being fed to the cells, although it will be possible to recycle alumina
from two of the four reacted alumina storage bins to the fresh alumina
storage bin. It will also be possible to unload fresh alumina directly
to the four reacted alumina storage bins. All of the air slides will
be operated by blowers.
The fans, the fresh alumina storage bin, and the dry scrubbers
will occupy an area roughly 350 feet long by 150 feet wide. To
accomodate the equipment, a 25-by 100-foot engineering building had to
be torn down, an existing well had to be covered, and some power lines
had to be moved. The existing control equipment will be left in place
until the tie-in to the dry scrubbers is made and will then be removed.
Retrofit items that are common to lines 1 and 2 and to line 4 are:
1. A 25- by 100-foot bag rehabilitation buildino.
6-59
-------
2. A 45- by 50-foot compressor building to supply compressed
air to the baghouse (cleaned by air pulse).
3. A crane to remove the baghouse internals for maintenance.
The funding for the retrofits for lines 1, 2 and 4 was
approved in July 1972. The retrofit for lines 1 and 2 just discussed
was operating in September 1974. The retrofit for line 4 to be
discussed in the next section was operating in July 1974. Hence
total installation took 26 months.
6.3.1.2 Engineering Description - Line 4
6.3.1.2.1 Potline operation — Line 4 has a capacity of 40,000
ton/year for a total plant capacity of 80,000 ton/year. Line 4 has
four rows of 160 cells in two potrooms, or two 40-cell rows per pot-
room. The line was completed in 1969 with a centralized primary control
system. The cells in line 4 are larger than those in lines 1 and 2,
and the anodes have channels.
The cells in line 4 have total-enclosure hooding with hood doors
exteriding the full length of both sides of the cell. A mechanized
feeding system operates with the hood doors closed. However, rather
than operating at preset time intervals, the potline is computer-
controlled, alumina being added on a demand basis. There are four
crust punchers and four alumina feeders, two to a side, and the puncher
and feeder are both in one vertical mechanism. Seventy percent of the
alumina is fed with the doors closed and the cell doors are only open
an average of 8 minutes per cell-day. As in lines 1 and 2, the cell
doors have to be opened to work the cells, pull the studs, and tap the
aluminum.
6-60
-------
6.3.1.2.2 Present controls — Every cell has a duct at each end that
carries the primary exhaust from the top of the cell hooding enclosure
to a circular manifold duct. Primary exhaust from 10 cells -- 25 percent
of one cell row ~ is manifolded together. These are a total of 16
cell gas r^nifolds for line 4. The exhaust rate is 4400 acfm per cell
at 180°F.
The two potrooms in line 4 are west of lines 1 and 2 and are also
oriented in a north-south direction. The control equipment for line 4
is southwest of the line 4 potrooms and constitutes a central
installation. Two elevated circular steel ducts running south alonq
the west side of the westernmost potroom pick up the sixteen 4-foot
manifold ducts, four at a time. Each set of four pickups consists of
one manifold from each of the four cell rows in the potline. The
elevated ducts increase in size from a diameter of 6 feet to a diameter
of 10 feet as they proceed south. The two ducts .jointly handle all of
the exhaust from the preceding manifolds.
Near the south end of the potroom, each elevated circular duct
divides in two, turns west, and lowers to a pair of ground-level fans.
The four fans are installed on a north-south axis as shown in Figure 6-14.
Each fan is driven by a 500-hp motor and is rated at 218,000 acfm, a
capacity exceeding the 4400 acfm per cell exhaust rate. Flow is dampered
during normal dual-fan operation, and if one fan is not operating, the
exhaust rate only drops 30 to 40 percent in the respective feeder duct,
not 50 percent.
Exhaust from each pair of fans is recombined into a rectangular
steel duct. The two rectangular ducts proceed west, then north
to four cement blockhouse scrubbers in one building. One rectangular
6-61
-------
1000-ton
ALUMINA Bl
1
i
1-
f
f
1
1
F
f
t
11
CONTROL
BUILDING
^
I
I
I
I
\
\
\
%
1
AIM OL
^S
\
\
f
RY SCRUBBER
UNIT
0 15 30 60
J I
SCAL E, ft
150.000 acfm
FAN WITH 700-hp
MOTOR
EXISTING DAMPER
AND DUCT-
,WORK
FROMPOTLINE4
Figure 6-14. Retrofit layout -- plant A - line 4.
6-62
-------
duct feeds two scrubbers from the east side of the building and the
other feeds two scrubbers from the west side.
Each of the four scrubbers has three gas inlets to a single spray
chamber, the inlets being on the same horizontal plane. There is a bank
of nine countercurrent sprays at each inlet. The qas flows upward
through the chamber and out the top through a single exit. Near the
top is another bank of countercurrent sprays, with a mist eliminator
above this bank. Gas enters the scrubbers at 150 to 160°F and leaves
at about 5°F above ambient temperature. Hence scrubbing causes a
loss of thermal stack lift.
Rectangular wooden ducting conveys the exhaust from the four
scrubbers to a common inlet on a single stack that discharges to the
atmosphere 500 feet above grade.
The plant has experienced less emission control problems on line 4
than on lines 1 and 2. The cell cathode shells do not bow down in the
center of the side, making tight hooding possible. The ducts are better
designed to handle the hydrocarbons in the primary exhaust.
Water passes once through the cement blockhouse scrubbers and,
along with the water discharged from the scrubbers on lines 1 and 2,
goes to water treatment. The operation of the water treating facility is
described in subsection 6.3.1.1.3.
6.3.1.2.3 Planned retrofit — The planned retrofit consists of rerouting
the line 4 primary exhaust downstream of the fans. The primary exhaust
will go to 18 dry scrubbers located together in an area west of the
blockhouse scrubbers — again a central installation. The existing
scrubbers will be bypassed.
6-63
-------
Figure 6-14 is a layout of the retrofit for line 4, and Table 6-14
lists the major retrofit items. The ducting to the f?.ns will remain
unchanged. The existing fans will be modified to handle the increased
pressure drop requirement. Primary exhaust from all four fans is
ducted together as shown. Hence, four fans handle both potrooms of
line 4 and feed all 18 dry scrubbers as pictured in Figure 6-14.
The per-cell primary exhaust rate on line 4 should not increase
as a result of the retrofit. Hence, there should be no increase in
the primary collection efficiency.
The dry scrubbers are sized so that at least two of the 18 can be
off-line at a given time and the remaining scrubbers will still handle
the 600,000 acfm exhaust. Originally the plant planned to manifold
the 72 stacks and convey all the scrubbed line 4 primary exhaust to the
existing 500-foot stack. However, a private firm did a meteorological
study that indicated that it was not necessary to go to the stack in order
to achieve ambient air quality standards. The plan now is not to use
the stack. One advantage in not using it is the ability to pinpoint
broken bags. The 72 stacks could be tied in to the existing stack at a
later date.
As in the retrofit for lines 1 and 2, considerable solids
handling is involved in the line 4 retrofit. The existing alumina
unloading station is adequate to supply fresh alumina to the dry
scrubbers. All of the solids handling equipment is being designed
for 100 percent feed. Although one-pass feeding to the scrubbers is
planned, it will be possible to recycle alumina from the reacted alumina
storage bin to the fresh alumina storage bin. It will also be possible
6-64
-------
Table 6-14. MAJOR RETROFIT ITEMS—PLANT A—LIKE 4
1. Four fans, each modified to handle 150,000 acfm at 180°F and 20
inches of water total pressure drop and driven by a 700-hp motor.
Modification will include a new damper on the outlet of each fan.
2. Two ground-level rectangular mild steel ducts, each about 150
feet long, feeding into one rectangular duct about 200 feet lonq.
The latter duct feeds all 18 reactor-baghouse dry scrubbing units,
reducing in size from a height of 13 feet to a height of 4 feet.
3. Eighteen mild steel reactor-baghouse dry scrubbers, set in two
rows of nine each. Each scrubber is a rectangular box, 11 feet
by 42 feet and 15 feet high. The top of each"scrubber is 40 feet
above the ground. Each scrubber has four compartments. Each
compartment has a gas inlet section shaped like an inverted
rectangular pyramid on the bottom and a stack on the top, or a
total of 72 stacks for the 18 scrubbers. The stacks are 15 feet
high, discharging 55 feet above the ground. Each scrubber is
designed to handle 40,000 acfm of exhaust at 180°F. The baghouses
on each scrubber are cleaned by air pulse, requiring 90 psio com-
pressed air. Each scrubber requires one clamper in the inlet aas
line, air activated gravity alumina feed and discharge devices,
and five manually operated alumina shut-off gates.
4. Combination mild steel belt conveyor-air slide about 500 feet
long. The 24-inch belt conveyor is designed to handle 100 ton/hr
of fresh alumina. An existing belt conveyor transports alumina
uphill from the existing unloading station east of the line 4
potrooms to the top of the existing 2750 ton reacted alumina
storage bin. This bin is located between and above the two line 4
potrooms, centered along the length of the potrooms. The new
conveyor transports alumina from the existing conveyor up a slight
grade to the top of the new fresh alumina storage bin — item 5.
The 20-ton/hr air slide returns reacted alumina from the scrubbers
to the existing reacted alumina storage bin. The air slide is
preceded by a 20-ton/hr air lift.
5. Mild steel, 1000-ton fresh alumina storage bin located near the
18 dry scrubbers with high, intermediate, and low level bin
indicators. The bin is circular, 38 feet in diameter, with
conical top and bottom. Straight side height is 19 feet. The
bottom of the bin is 45 feet and the top is 95 feet above the
ground.
6. Two 10-ton/hr mild steel air slides, each slide conveying alumina
to nine dry scrubbers and equipped with a flow control valve and
a manually operated shut-off gate. Total length of each slide
is about 190 feet.
6-65
-------
Table 6-14 (continued). MAJOR RETROFIT ITEMS-PLANT A—LINE 4
7. Two 10-ton/hr mild steel air slides, each slide conveying reacted
alumina from nine scrubbers to an activator tank that services
both air slides and feeds the 20-ton/hr air slide in item 4.
Total length of each slide is about 230 feet.
8. Four small cyclonic dust collectors for alumina transfer and
storage operations.
9. A 20- by 25-foot control building.
to unload fresh alumina directly to the reacted alumina storage bin. All
of the air slides will be operated by blowers.
The fresh alumina storage bins and the dry scrubbers will occupy
an area roughly 220 feet long and 110 feet wide. Nothing had to be
torn down or moved to accomodate the equipment. The existinq cement block-
house scrubbers, located east of the dry scrubbers, will continue to
operate until the tie-in to the dry scrubbers is made and will not be torn
down afterwards.
The fate of the existing waste treating facility is explained in
subsection 6.3.1.1.3.
Retrofit items common to line 4 and to line 1 and 2, and estimated
installation times are given at the end of subsection 6.3.1.1.4.
6.3.1.3 Emissions Before and After Retrofit
Tables 6-15 and 6-16 present data on before and after
retrofit emissions provided by the operating company in mid-
1973 and re-submitted in October 1974. Table 6-15 shows the
quantities of fluoride and particulate generated at the cells, the
6-66
-------
Table 6-15. BEFORE RETROFIT EMISSIONS-PLANT A--LINES 1,2, AND 4
(Ib/ton AT)
Fluoride (as F")
Gaseous
Particulate
Total
Particulate
Dry solids
Condensibles
Total
Generation
19.0
11.0
30.0
114.3
12.8
127.1
Emissions
Primary
5.0
-fc.5
7.5
20.1
9.0
29.1
Secondary
0.6
1.2
1.8
9.5
0.9
10.4
Total
5.6
3.7
9.3
29.6
9.9
39.5
Removal
13.4
7.3
20.7
84.7
2.9
87.6
Table 6-16. AFTER RETROFIT EMISSION ESTIMATES—PLANT A—LINES 1,2, AND 4
(Ib/ton Al)
Fluoride (as F")
Gaseous
Particulate
Total
Particulate
Dry solids
Condensibles
Total |
Generation
19.0
11.0
30.0
114.3
-12.8
127.1
Emissions
r Primary
0.4-0.6
0.2-0.4
0.6-1.0
2.0-3.0
T. 5-2.5
3.5-5.5
Secondary
0.5
1.1
1.6
6.0
0.5
6.5
Total
0.9-1.1
1.3-1.5
2.2-2.6
8.0-9.0
Z.0-3>0
10.0-12.0
Recovery
18.0
9.6
27.6
105.8
10.3
116.1
6-67
-------
quantities emitted from the present primary control equipment and the
secondary building roof monitors, and the quantities removed by the
primary control equipment that eventually become solid and liquid
waste. All of the quantities are expressed as pounds of pollutant per
ton of aluminum produced (Ib/ton Al). The fluoride values are expressed
as fluoride ion. Dry solid particulate includes particulate fluoride as
well as alumina and carbon. Condensibles could be alternately labeled
"CgH6 Solubles" or "Hydrocarbon Tar Fog and Gas."
The generation and emission levels in Table 6-15 correspond to a
primary collection efficiency of 94 percent, a primary removal efficiency
of 73 percent, and an overall control efficiency of 69 percent on total
fluoride for plant A. These levels also correspond to a primary collection
efficiency of 92 percent, a primary removal efficiency of 75 percent, and
an overall control efficiency of 69 percent on total particulate. Overall
control efficiency on hydrocarbon condensibles is only 23 percent.
Table 6-16 shows the quantities of fluoride and particulate that are
expected to be emitted after the dry scrubbing retrofit is installed by
late 1974, and the quantities recovered and recycled to the cells. The
emissions are preliminary estimates based on prototype tests mentioned in
the introduction of this case description. Although the retrofit does not
include secondary control, secondary emissions should be reduced through
increased primary collection. This improved collection should be brought
about by the increase in the per-cell primary exhaust rate on lines 1 and 2,
mentioned under subsection 6.3.1.1.4, and by better hood sealing and
improved operating practices throughout the plant.
6-68
-------
The emission levels in Tables 6-15 and 6-16 are averages.
However, the preliminary nature of the data upon which the
primary emissions in Table 6-16 are based necessitates stating
these emissions in ranges.
The generation and emission levels in Table 6-16 correspond to a
primary collection efficiency of 95 percent, an average primary removal
efficiency of 97 percent, and an average overall control efficiency of
92 percent on total fluoride for plant A. They also correspond to a
primary collection efficiency of 95 percent, an average primary removal
efficiency of 96 percent, and an average overall control efficency of
91 percent on total particulate. Average overall control efficiency on
hydrocarbon condensibles should increase to 80 percent.
The 105.8 Ib/ton Al of dry solids that are expected to be recovered
after retrofit includes 52.4 Ib/ton Al of alumina. This alumina and the
27.6 Ib/ton Al of fluoride are considered to be the only valuable
materials recovered.
Three conclusions that can be drawn from Table 6-16 are:
1. The expected total fluoride emissions for this existing plant
are somewhat higher than the EPA standard of performance for
new primary aluminum plants of 2.0 Ib/ton Al.
2. After retrofit this plant should be well within the existing
State emission standard of 15 pounds of soVid particulate
per ton of aluminum produced. The retrofit is being installed
6-69
-------
to bring the plant into compliance with the State particu-
late standard. The State has no fluoride emission standard.
According to company personnel, plant emissions have not
resulted in any fluoride vegetation damage. The plant is
located in an industrial region.
3. The retrofit should simultaneously bring about low emission
levels of fluoride, solid particulate, and hydrocarbon.
The generation and emission levels in Tables 6-15 and 6-16 show
that the ratio of total particulate to total fluoride at plant A is
4.24 for generation, 4.25 for total (primary and secondary) emissions
before retrofit, and 4.58 for average total emissions after retrofit.
6.3.1.4 Capital and Annual Operating Costs of Retrofit
6.3.1.4.1 Capital costs — Table 6-17 presents actual capital costs
and estimates for the total retrofit furnished by the company in
December 1974 and broken down into the major retrofit Items. Al-
though the installation is complete, not all of the final figures are
known. Assuming an annual aluminum capacity of 80,000 tons, $11,313,000
is equivalent to a capital cost of $141 per annual capacity ton.
The largest cost item in Table 6-17 is ductwork. Of the $1,819,000,
$1,600,000 is estimated for the collector ducts on lines 1 and 2.
Equipment purchase costs for fans, reactors, and baghouses amount
to $200,000, $500,000, and $987,000 respectively. Costs of the
seven small cyclonic dust collectors listed in Tables 6-13 and
6-70
-------
Table 6-17. RETROFIT CAPITAL COST ESTIMATE—PLANT A—LINES 1,2, AND 4
Direct—Capital
Ducts $1,819,000
Fans 341,000
Reactors 1,775,000
Baghouses 1,269,000
Alumina transfer 1,196,000
Alumina storage 415,000
Electrical 975,000
Instrumentation and sampling 320,000
Bag maintenance 670,000
Compressed air 458,000
Capital spares 60.000
Subtotal $9,298,000
Direct—NonCapital
Preoperating expense $ 150,000
Equipment testing 35.000
Subtotal $ 185,000
Indirect—Capital
Engineering «. $1,830,000
Contingency
Escalation -
Subtotal $1,830,000
Project total $11,313,000
6-71
-------
6-14 are covered under alumina transfer and storage in Table 6-17.
Site preparation costs total $367,000 and are covered under ducts, fans,,
reactors, and alumina storage in Table 6-17; the costs include
building and equipment demolition both before and after retrofitting.
Of the $320,000 for instrumentation and sampling, the control buildinas
are estimated at $68,000, instrumentation at $220,000, and sampling at
$30,000. All of the sampling cost is for gas sampling; part of it for
sample ports and part for sampling equipment. The $670,000 for bag
maintenance covers the cost of the aforementioned (see subsection
6.3.1.1.4) bag rehabilitation building, a mobile crane and associated
equipment. The $458,000 for compressed air covers the cost of the
aforementioned compressor building and associated equipment. Of the
$1,830,000 for engineering, engineering performed by the operating
company is estimated at $200,000; plant engineering at $60,000; con-
struction management at $200,000; and contract engineering, fee, and
procurement at $1,160,000 for a total actual engineering cost of
$1,620,000. There are no contingency and escalation costs since total
installation is complete.
All of the ducts, reactor-baghouses, bins, and conveyors are mild
steel construction. This is a cost advantage that a dry control retrofit
enjoys over a wet ESP retrofit, as the latter normally requires 316
stainless steel construction.
The book value of all assets to be retired as a result of the
retrofit was $801,214 as of December 31, 1971. Approximately
50 percent of these assets will be demolished or abandoned in place.
6-72
-------
The remaining assets, including fans, motors, pumps, and steel
ductwork, may have some salvage value if they can be sold. Because
of the uncertainty of the disposition and recoverable value of
these assets, their salvage value has been ignored by the company
in its capital cost estimate.
The company has done considerable development work at this plant
and at other locations. These development costs are not reflected
in Table 6-17.
6.3.1.4.2 Annual operating costs — Table 6-18 is a company estimate
of what the gross and net annual operating costs of the total retrofit
should be during the first year of operation. Net annual operating cost
for the before-retrofit control is estimated to be $292,800.
Assuming a daily aluminum production of 205 tons equivalent to an
annual production of 74,825 tons, the gross annual operating cost in
Table 6-18 of $741,450 amounts to $9.91 per ton. Making the same
assumptions, the net annual operating cost of -$65,128 amounts to
-$0.87 per ton. The negative net annual operating cost does not
represent profit because capital-related charges are not included.
Most of the items under gross annual operating cost in Table 6-18
are self-explanatory. The electric power rate is equivalent to 2.99
mills per kilowatt-hour, which is very low for the United States. Part
of the power requirement is for producing compressed air for bag clean-
ing.
6-73
-------
Table 6-18. RETROFIT ANNUAL OPERATING COST ESTIMATE—PLANT A--
LINES 1, 2, AND 4
Gross Annual Operating Cost
Based on 1975 (first year of operation) cost levels.
Operating Supplies
Bags: 28,080 installed + 5% damaged = 29,484 replaced
once per 18 months or 19,656 per year
19,656 replaced @ 5.50 each = $108,108
Other supplies (15% of operating labor) 30,500
$138,608
Operating Labor
Bag change-out @ 3 bags/manhour @ $8.55/manhour
19,656 X $8.55 $ 56,020
3
Operating and control: 1 operator/shift = 8,760
manhours at $10.31/manhour 90,316
Fan and duct cleaning: 7,500 manhours @ $7.60/manhour 57,000
$203,336
Maintenance
Labor: 11,484 manhours @ $12.47/manhours $143,205
Material: 57% of labor 81,627
Outside contract: Painting @ $140,000/5 years 28.000
$252,832
Power
49,056 megawatt-hours @ $2.99 $146.677
Total Gross Annual Operating Cost $741,450
Value of Recovered Material
Alumina Recovered: 1960 ton/year @ $96.80/ton $189,728
Aluminum Fluoride Recovered: 1690 ton/year @ $3657ton 616.850
Total Value of Recovered Material $806,578
Net Annual Operating Cost -$ 65,128
6-74
-------
As mentioned in subsection 6.3.1.3, only the recovered alumina
and fluoride are considered to be valuable materials. The amount of
recovered alumina is estimated from a recovery rate of 52.4 Ib/ton Al
and an annual production of 74,825 tons. The amount of recovered
aluminum fluoride is estimated from a recovery rate of 27.6 Ib/ton Al
(Table 6-16) equivalent to 45.2 pounds of aluminum fluoride containing
61.1 percent fluoride, and an annual production of 74,825 tons. The
latter estimate assumes that both gaseous and particulate fluoride
ion will be returned to the cells as aluminum fluoride. An alumina
cost of $96.80 per ton is equivalent to 4.8 cents per pound. A cost
of $365 per ton for aluminum fluoride containing 61.1 percent fluoride
is equivalent to 29.9 cents per pound of fluoride. By comparison, an
EPA contract study gives 1971 recovered alumina and fluoride values
50
of 3.2 and 25 cents per pound, respectively. The value of recovered
materials has increased due to rather significant increases in the
value of alumina which reflects the recent changes in bauxite prices
around the world, as well as some increase in the value of fluorides.
Net annual cost includes the above operating costs along with
capital-related charges. Such charges include depreciation, interest,
administrative overhead, property taxes, and insurance. These were
not furnished by the company and, hence, are not included in Table 6-18.
Based oh a "capital recovery" factor of 11.683 percent, an "administrative
overhead" factor of 2 percent, and a "property taxes and insurance"
factor of 2 percent, capital related charges would amount to 15.683
percent of capital cost for this retrofit. The "capital recovery"
factor covers depreciation and interest and is based on a 15-year equip-
6-75
-------
ment life and 8 percent interest. Capital related charges for
this retrofit thus amount to 15.683 percent of $11,313,000 — or
$1,774,218. Adding these charges to Table 6-18 would result in a nross
annual cost of $2,515,668 and a net annual cost of $1,709,090.
6.3.2 Plant B—HSS Cells—Primary Wet ESP Retrofit51
This plant's reduction facilities include two HSS plants — a
south plant and a north plant. Primary control presently consists
of wet scrubbers, but plans are to retrofit the primary exhaust with
31 wet electrostatic precipitators, 10 for the south plant and 21 for
the north plant. The plants have no secondary control and none is
planned.
Wet ESPs are being installed because:
1. The presence of a cryolite recovery plant to handle the
scrubber-ESP effluents makes dry scrubbing less attractive.
2. Aluminum product purity at plant B is high, among the
highest in the nation. According to plant personnel, dry
scrubbing with attendant recycle would lower this purity.
3. High energy scrubbers would require excessive power inputs
to achieve the desired control.
4- The cross flow packed bed scrubber with TelleretteR packinci
applied to an HSS potline will plug after only 30 minutes of
operation.
5. Although the floating bed scrubber does not plug, it cannot
attain as high a level of control as the ESP.
6-76
-------
Potline operation, present controls, and the planned retrofit
are now described, first for the south plant and then for the north
plant. Next the present air emissions and the emissions expected after
Urtal retrofit are presented. Then the plant's present water treating
facility and the changes to it necessitated by the retrofit are
explained. Finally, capital and annual operating costs for the total
retrofit are estimated.
6.3.2.1 Engineering Description -- South Plant
6.3.2.1.1 Potline operation -- The south plant has three potlines and
a total capacity of 70,000 ton/year. Each potline has 124 cells set in
four rows in one potroom for a plant total of 372 cells. The ootrooirs
have sidewall ventilation. The plant was built in 1941 and expanded
in 1952.
The cells have total-enclosure hooding with manually operated steel
roll-down hood doors extending the full length of both sides of each
cell. Pollutants continuously escape from the top of the cell enclosure
and also from the hood doors when they are open. The doors have to be
opened frequently to add alumina to the cryolite bath by working the
cell, to tap the molten metal layer from beneath the bath, and to
insert and remove studs from the anode block while raising the flexible
current connectors.
6-77
-------
6.3.2.1.2 Present controls -- Four 8-inch ducts, two on each end
of each cell, pick up the primary exhaust from the top of the cell
hooding enclosure and carry it to a circular manifold duct. Each
manifold handles primary exhaust from 15 or 16 cells. The primary
exhaust rate is 2000 to 2500 acfm per cell at 200°F.
Each manifold leads to a fan that is driven by a 50-hp motor
and is located just outside the potroom. Originally each fan was
upstream of one Douglas fir spray tower, each potroom having eight
spray towers apiece. However, four of the 24 towers have been replaced
by two larger towers, the larger towers each handling exhaust from two
fans or manifolds. Each tower is capped with an inverted cone. The
20. smaller towers are each about 8 feet in diameter and 35 feet high,
and the two larger towers are each about 15 feet in diameter and 50 feet
high. By way of comparison, the peaks of the adjacent potrooms are 39 feet
high and the tops of the roof monitors are 22 feet above the peaks, for
a total building height of 61 feet.
The towers are equipped with dual sprays that are fed with a cir-
culating alkaline solution that contains 2 grams of fluoride per liter of
solution. Because fine sprays plug, plant personnel consider it essential
to use a coarse spray and thoroughly wet the walls of the tower in order
to maximize gaseous and particulate fluoride removal efficiency.
6.3.2.1.3 Planned retrofit — Figure 6-15 is a layout of the south
plant retrofit and Table 6-19 lists the major retrofit items. The
three potrooms are oriented in a northeast-southwest direction. The
6-78
-------
0 25 50 100
SCALE, ft.
POTROOM L
POTROOM K
POTROOM J
\
14 ft.
DUCT DI-
AMETER
5 ESP'S
7 ft.
Figure 6-15. Retrofit layout - plant B - south plant.
-------
Table 6-19. MAJOR RETROFIT ITEMS—PLANT B—SOUTH PLANT
1. Several circular elevated mild steel ducts conveying primary
exhaust to two retrofit areas southeast and southwest of the
plant. For each area, 4-foot ducts combine and grow into
one 14-foot duct that reduces in size to 7 feet as it feeds
5 ESPs.
2. Ten fans, each driven by a 300-hp motor and upstream of an ESP.
Each fan is designed to handle 100,000 scfm of exhaust at about
4 inches of water total pressure drop.
3. Ten steel ESPs each designed to handle 100,000 scfm of exhaust.
Each ESP is a rectangular box 29 feet square and 29 feet high with
a stack discharging about 80 feet above the ground. Each ESP
has a gas side-inlet section of flattened rectangular pyramidal
shape.
4. A 20- by 50-foot control building.
planned retrofit consists of ducting all the primary exhaust from
potroom J and half from potroom K to five ESPs located together as shown
in Figure 6-15, and ducting all the primary exhaust from potroom L and
the other half from potroom K to five ESPs also located together. Each
set of five ESPs is termed a central installation.
The per-cell primary exhaust rate will be increased from 2000-2500
acfm at 200°F to 3500 acfm at 200°F, increasing the south plant's
primary collection efficiency. The ducts inside the potrooms are
presently oversized, so they will not have to be modified to handle
the increased flowrate. Primary collection efficiency will also be
improved by installing new motorized doors on the cells and sealing
the top of each cell's hooding enclosure with glass wool.
6-80
-------
Redirecting the south plant's primary exhaust from courtyard
to central controls will require a balanced ducting layout design
that ensures equal pressure and flowrates in all of the 12 manifolds
that are serviced by each set of 5 ESPs.
The ducting changes for this central retrofit will be external
to the potrooms. The existing fans and spray towers will be bypassed.
All of the spray towers will eventually be torn down, but many will
have to be torn down during the installation to make room for the
ducting shown in Figure 6-15. This will mean that portions cf the
plant will run uncontrolled for varying periods of time during the
installation. Nothing but the spray towers will have to be torn
down as a result of the south plant retrofit.
Removal of the existing spray towers will force the wet ESPs
to act as absorbers for gaseous fluoride and require that liauor be
fed to the inlet sections of the ESPs. Plant personnel hope to control
corrosion of the ESP steel internals by controlling the composition and
pH of this feed liquor. Even so, they anticipate having to rebuild the
internals every 10 years.
Estimated installation times are given at the end of Subsection
6.3.2.2.
6.3.2.2 Engineering Description - North Plant
6.3.2.2.1 Potline operation — The north plant also has three potlines
and has a capacity of 140,000 ton/year for a total plant capacity of
210,000 ton/year. Each potline has four rows of 168 cells in two pot-
6-81
-------
rooms, or two 42-cell rows per potroom, for a plant total of 504
cells. The potrooms have sidewall and basement ventilation. The
plant was built in 1968.
The cells are elevated slightly above the floor and have total -
enclosure hooding with mechanically operated aluminum doors extending
the full length of both sides of each cell. Comments on emissions from
the top of the cell enclosure and on door opening for the south plant
also apply to the north plant.
6.3.2.2.2 Present controls — Four ducts, two on each end of each
cell, pick up the primary exhaust from the top of the cell hooding
enclosure and carry it to a circular manifold duct. One manifold
handles primary exhaust from 14 cells. The primary exhaust is 3600
scfm per cell.
Each manifold proceeds to a 50,000 scfm fan that is driven by a
125-hp motor, is located outside the potroom, and is upstream of a
spray tower. Figure 6-16 shows the general location of the 36 spray
towers at the north plant. Each tower is 13 feet in diameter and is
capped with an inverted cone that connects to a 5-foot stack. This
stack discharges to the atmosphere about 70 feet above the ground.
By way of comparison, the peaks of the potrooms are 54 feet high
and the tops of the roof monitors are 8 feet above the peaks, for
a total building height of 62 feet. The towers are fed with the same
alkaline solution as the towers at the south plant.
6-82
-------
o o
POTROOM
o
o
CO
CO
D
SPRAY TOWERS -<^O
Qi ESP
o
o
D
POTROOM
0
O
D
O
o
D
o
D
POTROOM
O
o
D
8
D
o
o
D
POTROOM
O
o
D
o
o
a
o
o
D
POTROOM
O
o
D
o
o
a
o
o
D
POTROOM
o a
o o
0 50 100 200
111 -1
SCAL f., ft.
Figure 6-16. Retrofit layout — plant B — north plant (location and position
of control equipment not exact).
-------
6.3.2.2.3 Planned retrofit — Figure 6-16 is a layout of the north
plant retrofit and Table 6-20 lists the major retrofit items. The
south end of the north plant is 1100 feet northwest of the north end
of the south plant. The planned retrofit consists of adding 21 ESPs
downstream of the existing fans and spray towers. The plan is to install
three 50,000 scfm wet ESPs on the outside of each of the two end
buildings, one per spray tower, and three-100,000 scfm wet ESPs in each
of the five 51-foot wide courtyards between potrooms and between pot-
lines (see Figure 6-16). Each of the latter ESPs will handle the exhaust
from two spray towers, one from each adjacent building.
Primary collection efficiency should not increase at the north
plant as a result of the retrofit because the north plant already in-
corporates all of the same modifications that are expected to increase
collection efficiency at the south plant. The per-cell primary exhaust
rate will remain at 3600 scfm. The existing fans and spray towers and
all the ductwork upstream of the spray towers will not be changed, and
nothing will have to be torn down as a result of the north plant retrofit.
Downstream of each tower a 5-foot duct will carry the tower exhaust
from the tower's inverted cone to the inlet section of the adjacent ESP.
There will also be a valving arrangement to vent the tower exhaust to
the atmosphere if the ESP is inoperative.
At the north plant, liquid will be fed to the inlet section of the
ESPs and will pass through an ESP before passing through its associated
spray tower(s). As at the south plant, plant personnel hope to control
corrosion by controlling the composition and pH of the liauor, but
anticipate rebuilding the ESP internals every 10 years.
6-84
-------
Table 6-20. MAJOR RETROFIT ITEMS—PLANT B—NORTH PLANT
1. Six steel ESPs, each designed to handle 50,000 scfm of exhaust
Each ESP is a rectangular box 29 feet by 14 feet and 29 feet high
with a stack discharging about 80 feet above the ground. Each
ESP has a gas side-inlet section of flattened rectangular
pyramidal shape.
2. Fifteen steel ESPs, each designed to handle 100,000 scfm of
exhaust. Each ESP is a rectangular box 29 feet square and 29
feet high with a stack discharging about 80 feet above the ground.
Each ESP has a gas side-inlet section of flattened rectangular
pyramidal shape.
3. Seven 8- by 20-foot control buildings, one for each set of three
ESPs.
It was necessary to have the wet ESP tailor-made to the plant.
It was also necessary to prove its operability before makinci a total
plant commitment. For these reasons, and because of the limited
availability of funds and manpower, the retrofit was completed in
phases.
In late 1970, design was started on a pilot 50,000 scfm unit on
the north plant. As of September 1973, one more 50,000 scfm unit and
three 100,000 scfm units were operating on the north plant. These
five units comprise Phase I of the retrofit. Phase II involves
installing the remaining 16 ESPs on the north plant, and Phase III
involves installing the 10 ESPs and accompanying fans and ductwork on
the south plant.
Plant B completed Phases I and II in January 1975 and, as of
March 1975, planned to complete all three phases by June 1975. It
6-85
-------
will then have been about 4-1/2 years from the start of development
through total plant installation. Had the pilot unit not proven
operational, this time could have been much longer.
Because the plant's engineering design capabilities increased
with operating experience, each subsequent phase has taken less time
to design than the former. By necessity, work on a subsequent phase
begins before the installation of the former phase is completed.
6.3.2.3 Emissions Before and After Retrofit
Tables 6-21 and 6-22 reflect estimates of emissions before and
after retrofit provided by the company. All of the quantities are
expressed as pounds of total fluoride ion per ton of aluminum pro-
duced (Ib/ton Al). The tables show the quantities generated at the
cells, the quantities emitted from the applicable primary control equip-
ment and the secondary building roof monitors, and the quantities removed
by the primary control equipment that eventually become either cryolite
or liquid waste. The overall plant average is a weighted average
based on the north plant accounting for 67 percent of plant B's pro-
duction.
The generation estimates in Tables 6-21 and 6-22 are based on a
statistical analysis for the 10-month period beginning June 1, 1972,
and ending April 1, 1973. Plant personnel selected this time interval
because the total plant was at full production and had the fewest
in-process variables to distort the results. Data from other time
periods would, of course, be somewhat different.
6-86
-------
Table 6-21. BEFORE RETROFIT MAXIMUM EMISSIONS-PLANT B--NORTH
AND SOUTH PLANTS
(Ib total F'/ton Al)
North plant
South plant
Overall plant
average
Generation
38.2
50.0
42.1
Emissions
Primary
3.6
4.0
Secondary
1.9
10.0
Total
5.5
14,0
8.3
Removal
32.7
36.0
33.8
Table 6-22. AFTER RETROFIT MAXIMUM EMISSION ESTIMATES—PLANT B--
NORTH AND SOUTH PLANTS
(Ib total F"/ton Al)
North plant
South plant
Overall plant
average
Generation
38.2
50.0
42.1
Emissions
Primary
0.7
1.4
Secondary
1.9
5.1
Total
2.6
6.5
3.9
Removal
35.6
43.5
38.2
6-87
-------
The method of analyzing the data was taken from Probability and
f o
Statistics for Engineers.' Sample averages and standard deviations
were calculated from data derived from the plant's standard monthly
sampling program. In this program, the plant samples the input to and
the output from 6 to 10 spray tower fume control units and the output
from 4 to 6 roof monitor locations. Different locations are sampled
each month. The plant's objective is to sample every location across
the plant once every 6 months.
The average pounds of total fluoride generated per day was computed
on a monthly basis by adding the average daily emissions from the
monitors to the average daily spray tower inputs per month per plant.
The monthly average pounds of total fluoride generated per ton of aluminum
produced was then computed for each of the 10 months by dividing each such
average generation per month per plant by that plant's average daily
production rate for the month. The average and the standard deviation
for the 10-month period in each plant was then computed from the 10
monthly averages. The Kolomogorov-Smirnov Test was conducted on the data
derived for each plant and it was determined that a normal distribution
provided a good fit for each plant and for the total plant. The generation
estimates in Tables 6-21 and 6-22 represent 95 percent tolerance limits
at a 95 percent confidence level.
The primary and secondary emissions in Tables 6-21 and 6-22 are
computed by applying estimated primary collection and removal efficiencies
to the above generation estimates. Estimated primary collection and
6-88
-------
removal efficiencies for Table 6-21 are 95 and 90 percent, respectively,
for the north plant, and 80 and 90 percent, respectively, for the south
plant. Estimated primary collection and removal efficiencies for Table 6-22
are 95 and 98 percent, respectively, for the north plant, and 90 and 97
percent, respectively, for the south plant. Assuming a zero percent
secondary removal efficiency, the above primary collection and removal
efficiencies correspond to overall control efficiencies before retrofit
for the north and south plants of 86 and 72 percent, respectively; and to
overall control efficiencies after retrofit for the north and south plants
of 93 and 87 percent, respectively.
Although not shown in Tables 6-21 and 6-22, the retrofit should
increase the total particulate primary removal efficiency from 55 to
98 percent, and the hydrocarbon primary removal efficiency from a control
level of 8 to 10 percent up to a control level of 92 to 94 percent, the
latter being a ten-fold increase. The hydrocarbons comprise a substantial
portion of the small-diameter particulate that the present scrubbers are
incapable of removing.
Table 6-23 contains revised June 1974 company estimates of emissions
after retrofit. The sampling methods and statistical treatment are the
same as for Table 6-22. However, the data are averages rather than 95
percent tolerance limits and are computed over a 14-month period of full
production that includes the 10-month period used for Table 6-22. Also,
the primary emission estimates at both the north and south plants are 95
percent confidence level estimates based on actual testing of primary
emissions at the north plant. Twelve months of emission'testing in 1972
yielded an average total emission before retrofit of 5.4 Ib F/ton Al.
6-89
-------
Table 6-23. AFTER RETROFIT AVERAGE EMISSION ESTIMATES—PLANT B--
NORTH AND SOUTH PLANTS (Ib total F/ton Al)
North plant
South plant
Overall plant
average
Generation
32.1
31.1
31.8
Emissions
Primary
0.25
1.10
Secondary
1.6
4.0
Total
1.85
5.1
2.9
Removal
30.25
26.0
28.9
The primary and secondary emissions in Table 6-23 correspond to
primary collection, primary removal, and overall control efficiencies
of 95, 99.2, and 94 percent, respectively, for the north plant; and of
87, 96, and 34 percent, respectively, for the south plant.
From Table 6-22 it can be seen that the maximum expected total
fluoride emissions of 3.9 Ib/ton Al for this existing plant after
retrofit is about twice that of the EPA standard of performance for
new primary aluminum plants of 2.0 Ib/ton Al. The average expected
total fluoride emission of 2.9 Ib/ton Al in Table 6-23 is somewhat
higher than the average expected total fluoride emission for plant A
of 2.4 Ib/ton Al shown in Table 6-16. However, as can be seen by com-
paring the efficiencies for plant B with those for plant A in Section
6.3.1.3, the total fluoride primary removal efficiency for the wet ESP
retrofit at plant B is the same or higher than the 96 percent primary
removal efficiency for the dry scrubbing retrofit at plant A. The
expected total fluoride emissions at plant B are higher than those
expected at plant A because the primary collection efficiency of the
6-90
-------
south plant at plant B after retrofit is estimated to be only
87-90 percent. The primary collection efficiency at plant A and at
the north plant of plant B are both estimated to be 95 percent after
retrofit.
6.3.2.4 Water Treatment
6.3.2.4.1 Current practice-- Fluoride is removed from the primary exhaust
in the north and south plant spray towers into a recirculatinq liquor
stream. Fluoride is recovered from this liquor as standard grade (90
percent) cryolite in a cryolite recovery plant. This recovery is
illustrated in Figure 6-17:
LIQUOR
FROM
SCRUBBING TOWERS
THICKENERS
ALKALINE LIQUOR
OVERFLOW
SLUDGE OR
UNDERFLOW
NaOH
BLEED
TO
RIVER
DIGESTER
C02
PRECIPITATOR
TANK
ALKALINE LIQUOR
LIQUOR TO
SCRUBBING TOWERS
(30 grams/liter TOTAL SODA)
CRYOLITE
Figure 6-17. Flow diagram - plant B -- cryolite recovery plant.
6-91
-------
The lowering of the pH in the precipitator tank causes cryolite
precipitation.
At the spray towers, the recirculating liquor nicks up some of
the sulfur dioxide that is generated at the reduction cells. This
causes a sulfate buildup in the liquor, and it is necessary to bleed a
small portion of the liquor to control the sulfate level. The bleed
is controlled by a constant volume regulator. It goes directly to a
waste water discharge sump where it is thoroughly mixed before being
discharged to a nearby waterway.
Plant B also recovers fluoride from its spent potliner. It used
to buy potliner from other plants, but no lonner does this due to
stricter water effluent standards.
53
Plant B water effluent loadings are presented in an EPA study.
Plant B in this document is plant J in the EPA study. Plant B
net effluent loadings include fluoride and suspended solids loadings
of 2.2 and 3.8 Ib/ton Al, respectively.0 By comparison, the recom-
mended 30-day effluent limitations for the primary aluminum industry
to be achieved by July 1, 1977, are 2 and 3 Ib/ton Al for fluoride and
suspended solids, respectively; and the recommended daily effluent
limitations are 4 and 6 Ib/ton Al, respectively.55 These limitations
are considered to be attainable through the application of the best
practicable control technology. For wet scrubbing systems, best
practicable control technology is defined as cryolite precipitation
with recycle as practiced at plant B, or lime treatment with either re-
cycle or subsequent adsorption on activated alumina.06
6-92
-------
Table 6-21 shows that 33.8 pounds of total fluoride are removed
by the spray towers per ton of aluminum produced. Plant personnel
did not provide an estimate of the fluoride recovered from potliners.
If we assume the latter to be 20 Ib/ton Al (see Figure 5-5), then the
cryolite recovery plant handles 53.8 Ib/ton Al of fluoride. If it is
assumed that there is 100 percent recovery of fluoride from the pot-
liners and that the plant's net effluent fluoride loading of 2.2 Ib/ton
Al is all attributable to cryolite recovery, then it can be concluded
that the cryolite plant fluoride recovery efficiency for plant B is
95.9 percent.
6.3.2.4.2 Changes due to retrofit—Presently the hydrocarbons collected
in the recirculating scrubber liquor cause foaming in the cryolite
recovery plant when treating the resultant sludge. Such foaming can
make it extremely difficult to operate sludge treating equipment and
can result in airborne fluoride emissions. The plant can process in
spite of the present foaming, but, as noted in subsection 6.3.2.3,
the retrofit is expected to result in a ten-fold increase in the hydro-
carbon collected. If foaming is a direct function of the hydrocarbon
content in the sludge, then something must be done.
The plant has investigated three possible solutions. The first
two involve oxidizing the hydrocarbons and the third involves con-
trolling cryolite plant process variables so foaming does not occur.
The oxidation methods considered are direct calcination in a
rotary kiln and the Zimpro wet oxidation process. Direct calcination
is difficult to operate, has high energy requirements and high operating
6-93
-------
costs, and probably will require one more wet ESP to control emissions.
The performance of the Zimpro process on this type of sludge is not
well known and the capital costs are quite high.
The plant hopes to be able to identify the process variables that
affect foaming and thus accomplish a solution with considerably lower
capital costs. As of March 1975, it was not known if oxidation
would be required.
The plant has no immediate plans to reduce the effluent loadings
associated with its bleed stream.
6.3.2.5 Retrofit Capital and Annual Operating Costs
6.3.2.5.1 Capital costs—Table 6*-24 is a spring 1973 estimate of the
total retrofit capital costs for the north plant, south plant, and the
sludge treatment project broken down into the major retrofit items.
Assuming an annual aluminum capacity of 210,000 tons, $23,457,500 is
equivalent to a capital cost of $112 per annual capacity ton.
Plant B furnished the direct costs in Table 6-24. Reduction cell
sealing, new motorized doors, and new fans at the south plant will
increase primary collection efficiency. Two new thickeners, one for
each plant, are included under phases II and III in Table 6-24; but
in reality, they are part of cryolite recovery. New thickeners are
needed to handle the higher flowrate and higher fluoride loading re-
sulting from the retrofit and to remove smaller particulate. Smaller
particulate removal is required because the ESPs will have finer
spray nozzles than the present spray towers, and finer nozzles are more
likely to plug.
6-94
-------
Table 6-24. RETROFIT CAPITAL COST ESTIMATE—PLANT B
.
Direct costs
Reduction cell
sealing
Motorized doors
Ducts
Fans
Electrostatic
preci pita tors
Foundations
Drains, pumps,
and piping
Thickeners
Electrical
Control buildings
Equipment sales tax
Subtotal s
North plant
Phase I
Not
broken
down
Phase II
-
-
$520,000
-
4,810,000
200,000
200,000
400,000
380,000
42,000
328,000
$1,480,000 i $6,880,000
South plant
Phase III
$200,000
280,000
1,395,000
160,000
3,430,000
300,000
100,000
300,000
1,115,000
30,000
365,000
$7,675,000
Sludge treatment costs
Site preparation and foundation
Slurry tank and pumps
Centrifuge, kiln, feed screw,
afterburner and scrubber
Treated solids handling equipment
_Electrical
Wet oxidation equipment, including
foundations and electrical
Subtotal— sludae treatment including sales tax
$127,000
33,500
705,000
101,000
156,000
887.500
$2,010,000
6-95
-------
Table 6-24 (continued). RETROFIT CAPITAL COST ESTIMATE—PLANT B
Subtotal Phase I $1,480,000
Phase II 6,880,000
Phase III 7,675,000
Sludge treatment 2.010.000
Subtotal direct costs $18,045,000
Indirect costs
Engineering $1,804,500
Contingency 1,804,500
Escalation 1,804,500
Subtotal indirect costs 5,412,500
Subtotal direct costs 18.045,000
Project total cost $23,457,500
Sludge treatment costs are shown for the equipment associated with
both direct calcination and wet oxidation because plant personnel believed
that, regardless of the alternative selected, they will probably spend
$2 million for suitable sludge treatment equipment. The sludge treat-
ment equipment will be installed on land that is presently used to store
used potliners. The site preparation costs for sludge treatment in Table
6-24 represent the funds necessary to prepare this land.
Plant B did not furnish indirect costs. Engineering, contingency,
and escalation costs in Table 6-24 are each based on arbitrary factors
of 10 percent of direct capital.
6-96
-------
The plant B retrofit would have been more costly had the ESPs
been constructed of 316 stainless steel, the normal material of con-
struction for wet ESPs applied to an aluminum plant. As mentioned on
subsection 6.3.2.1.3, the plant hopes to control corrosion of the
asl internals but still anticipates rebuilding the internals every
10 years.
The assets to be retired as a result of the planned retrofit have
essentially no book value.
Table 6-25 is an October 1974 update of the total retrofit capital
costs for the north plant, south plant, and the sludge treatment pro-
jects. Assuming an annual aluminum capacity of 210,000 tons, $19,300,600
is equivalent to a capital cost of $92 per annual capacity ton. The
direct costs in Table 6-25 are from the company. The indirect engineering
cost is based on an arbitrary factor of 10 percent of direct capital.
There are no escalation and contingency costs since installation is
nearing completion.
Table 6-25. REVISED RETROFIT CAPITAL COST
ESTIMATE—PLANT B
Direct costs
North plant $8,871,000
South plant 7,675,000
Sludge treatment 1,000,000
Subtotal $17,546,000
Indirect costs
Engineering $1,754,600
Contingency
Escalation -_
Project total cost $19,300,600
-------
6.3.2.5.2 Annual operating costs— Table 6-26 is a company estimate
of year-to-year additional gross and net annual operating costs for
operating the ESPs after the plant has been retrofitted. Assuming
an annual aluminum production equal to the annual capacity of 210,000
tons, $171,000 for operating the ESPs amounts to $0.81 per ton. Plant
personnel stated that, even though these additional costs are quite
modest, annual operating costs for plant B's present system are
substantial. However, plant personnel are not able to break out the
present emission control annual operating costs. Also, although
an estimate of annual operating costs for sludge treatment was not
obtained, plant personnel stated that its operating costs should be
considerably smaller than that shown for the ESPs in Table 6-26.
The planned retrofit will not directly recover any valuable
material; hence, the zero credit. Generally, the value of the
fluoride recovered in an aluminum plant that has a wet scrubbing
system and cryolite recovery is offset by the operating costs of
recovering the fluoride.
The capital-related charges that are part of net annual cost were
not furnished by plant B and are not included in Table 6-26. Based on a
"capital recovery" factor of 14.903 percent, an "administrative over-
head" factor of 2 percent and a "property taxes and insurance" factor
of 2 percent, capital related charges would amount to 18.903 percent of
capital cost for this retrofit. Since the plant anticipates rebuilding
the ESPs every "10 years, the "capital recovery" factor covering
6-98
-------
Table 6-26. RETROFIT ANNUAL OPERATING COST ESTIMATE—PLANT B—
NORTH AND:SOUTH PLANTS
Gross annual operating cost^
Operating labor and materials $56,000
Utilities
Fuel -0-
Electricity 40,000
Water 5,000
Maintenance labor and materials 70,000
Total gross annual operating cost $171,000
Value of recovered materials -0-
Net annual operating cost $171,000
depreciation and interest is based on a 10-year equipment life and 8
percent interest. Capital related charges for this retrofit thus amount
to 18.903 percent of $19,300,600—or $3,648,000. Adding these
charges to Table 6-26 would result in gross and net annual costs of
$3,819,000.
6-99
-------
6.3.3 Plant C--Prebake Cells—Primary Injected Alumina Dry
Scrubbing Retrofits?
This plant has three side-worked prebake potlines. As built,
the potlines had only secondary control consisting of wet scrubbers,
but all three potlines were recently retrofitted with primary control
systems. Hoods were installed on the cells, and the primary cell gas
exhausts were directed to injected alumina dry scrubbers.
In the following sections, potline operation, secondary controls
with associated water treatment, and the primary retrofit are described;
Next the emissions before and after total retrofit are presented; and
capital and annual costs for the total retrofit are estimated,
6.3.3.1 Engineering Description
6.3.3.1.1 Potline operation--The plant was built in 1965 using
European technology. Its three computer-controlled potlines have a
total capacity of 265,000 ton/year. Each potline has 240 cells in
4 rows of 60 cells per row, installed in 2 buildings, for a plant total
of 720 cells. Each building consists of 2 single-row potrooms with
side-wall ventilation on the outside walls and a corridor between the
center walls, so that there are 4 potrooms per potline or 12 potrooms
for the whole plant. The cells are set into the potroom floor, bat
the potrooms have no basements.
Each cell has "18 anode assemblies, 9 to a side. Each assembly
consists of three small rectangular carbon anode blocks, two copper
branch rods to a block - six rods to an assembly. The six branch rods
are connected to a center rod that introduces electrical current. The
6-100
-------
cells are designed for 4 volts and 130,000 amperes, and are amonq
the lowest-voltage cells in the industry.
6.3.3.1.2 Secondary control system—The as-built secondary controls
consisted of 30 fiberglass Ceil cote scrubbers per building, on a floor
on top of the corridor between the potrooms but under the building
roof. Each scrubber thus handled four cells. To reduce water effluent
discharges, only 17 scrubbers per building are now operating — about
every other one -- and each thus handles emissions from about 7 cells.
Each scrubber consists of a horizontal spray section with 80 co-
current spray nozzles, a 40-hp fan on the inlet, and a slat mist
eliminator on the outlet. Each scrubber handles 104,000 scfm at 20 to
22°C and discharges through a 12- by 18-foot rectanoular stack 18 inches
above the peak of the potroom. This peak is 52 feet above the ground.
A 40-hp pump recirculates the scrubbing water at 1200 gal/min from a
hold tank beneath the scrubber. A small amount of water is bled off
this scrubbing loop to a water treatment plant.
The secondary control system as installed cost $10 million.
6.3.3.1.3 Water treatment—In the water treatment plant, water from the
scrubbers is treated with sodium aluminate to form cryolite. The
cryolite is filtered on a vacuum drum filter and then dried in a kiln
and recycled to the cells. This cryolite is of poor quality.
The water treatment plant was installed in 1971 for $1.45 million.
6-101
-------
6.3.3.1.4 Primary control retrofit—Overall control efficiency with
just the secondary system was low, and plant management started
investigating improved control in 1970. After experimenting with
small hoods over the gas holes in the crust, they constructed a
Quonset hut over a cell to find out what was being emitted. They
determined that fine particulates were the most difficult to control,
installed a 20-cell ESP, and ran it for 6 months. The level of control
obtained by ESP was not satisfactory. They also considered the venturi
and determined that it lacked the proper level of control. At about
that time, the plant designers had been investigating bag collectors,
and the plant decided to abandon wet scrubbing for the injected alumina
dry scrubbing system. The retrofit that was completed in the Spring of
1973 consisted of installing primary collection systems and injected
alumina primary removal equipment on all three potlines.
6.3.3.1.5 Primary collection retrofit—Side-worked prebake cells must
be worked manually along the entire side of a cell. Hence, the gas
collection skirt at plant C consists of two nonsegmented doors, one
on each side of the cell. Some of the cells have doors operated by air
cylinder, others by air motor.
The doors have to be open about 10 percent of the time to change
anodes and to add alumina by manually working the cells. The 20
cells whose primary exhausts were directed to the ESP prototype
have doors that must also be opened to tap aluminum. The remaining
700 cells have small tapping doors in the cell doors, so the cell
doors remain closed during tapping.
6-102
-------
A door closes with its bottom edge on the potroom floor. This
edge has an asbestos cloth seal. The closinq doors hit the floor
forcefully, generating considerable dust. The doors are made of steel
because gas jets from the cells can cause holes in aluminum doors.
In the middle of the cell superstructure, two circular ducts
pick up the primary exhaust and direct it upwards to a horizontal
12-inch circular branch duct that runs along the centerline of the
cell. The primary exhaust rate is about 3000 acfm per cell at 200°F-
Duct headers run along the corridors between the potrooms, each
header picking up branch ducts from 30 cells per potroom in line 1
and 15 cells per potroom in lines 2 and 3. The headers are rectangular
and increase in size as they pick up more branch ducts. An average
size is about 2 feet by 4 feet. Two headers join at the top of the
corridor, one from each potroom, and the common duct passes through
the roof to the control equipment in the courtyard. A secondary
scrubber has been removed to accomodate each common duct. Line 1 has
two common ducts, and two scrubbers were removed per buildinq. Lines 2
and 3 have four common ducts, and four scrubbers were removed per buildinq.
Nothing else had to be torn down or moved to accomodate the retrofit
equipment.
6.3.3.1.6 Primary removal retrofit—Each potline has two injected
alumina units located in the courtyards between its two buildings,
each unit servicing half of each building. Figure 6-18 is a oer.eral
flow diagram for the injected alumina process at plant C. The process
6-103
-------
2 j
FUMES
n-77] CONTAINING
[••'I FLUORIDES
DAY BIN FOR
FLUORINATED ALUMINA
STORAGE
CLEAN
AIR
CLEAN
AIR
FLUORINATED ALUMINA
D TO POTS
FLUORINATED ALUMINA
) DUCTING
/ TO COLLECT
/ FUMES
.HOODS COVERING
INDIVIDUAL POTS
FILTER BAG
HOUSES
FRESH ALUMINA INJECTED
TO ADSORB FLUORIDES IN FUMES
Figure 6-18. Flow diagram — plant C — injected alumina proces's.
-------
involves reaction of gaseous fluoride with the alumina to be fed to
the'cells followed by baghouse solids collection. Alumina is injected
into the flowing gas stream and the reaction occurs in a matter of
seconds. As designed, 100 percent of the alumina fed to the cells
should pass through the units. 97 percent passage is actually achieved.
Plant C's retrofit is somewhat unusual in that the fans are located
downstream of the baghouses. This location helps keep the fan blades
from scaling. Scale deposits will cause the fans to lose their dynamic
balance. However, a downstream fan location requires that the baq-
houses operate under negative pressure, and a negative operating pressure
requires a stiffer baghouse structure.
Potline 1 has injected alumina control units designed by Prat-
Daniel -Poel man (POP) and potlines 2 and 3 have control units designed
by Alcan. The order of retrofit was 2-3-1. Both designs are unique
in this country and will now be described in detail.
6.3.3.1.7 PDP design—Figure 6-19 is a schematic of the retrofit and
Table 6-27 lists the major retrofit items for one of the two control
units on potline 1. Each unit has a total of 12 Venturis, 12 baghouses,
6 fans, and 6 stacks.
6.3.3.1.8 Alcan design—Figure 6-20 is a schematic of the retrofit
and Table 6-28 lists the major retrofit items for one of the four
control units on potlines 2 and 3. There are no Venturis in the
Alcan design. Each unit has a total of 22 baghouses, 22 fans and
22 stacks.
6-105
-------
EXHAUST
FROM
POTROOMS
FRESH
ALUMINA-
FEED
REACTED
ALUMINA
TO EXISTING
DAY BINS ON
BOTH BUILDINGS
EXHAUST
FROM
POTROOMS
8ft.
FAN
CONTINUOUS
LOW PRES-
SURE CON-
VEYING
SYSTEM
Figure 6-19. Retrofit schematic -- plant C -- POP design (Venturis, ducts,
fan, stack, and solids handling are depicted for one pair of baghouses on.ly).
6-106
-------
Table 6-27. MAJOR RETROFIT ITEMS—PLANT C--PDP DESIGN
1. Two 8-foot ducts, one from each building (see Figure 6-19),
join opposite ends and sides of an 8-foot horizontal duct.
This horizontal duct runs underneath the control unit and feeds
six paired venturi-baghouse subsections.
2. Six pairs of vertical Venturis. Gas flows upward through the
Venturis each of which has two injection ports - one for fresh
alumina and one for reacted alumina that is recycled from the
baghouse. The ratio of recycled to fresh alumina is fixed at
20:1.
3. Six pairs of baghouses. Each baghouse handles 30,000 acfm of
200°F primary exhaust. Gas flow'entering and leaving a baghouse
is horizontal. Gas leaving one of the two baghouses in each pair
passes horizontally through the opposite baghouse but not through
the bags—there is no process connection—and the two baghouse
exhausts join downstream of the opposite baghouse. Each baghouse
is a rectangular box 18 feet square and 20 feet high with an
inverted pyramid bottom gas inlet. The top of each baghouse is 40
feet above the ground. The bags are cleaned by shaking with reversed
air flow. At the end of every 30 seconds, one baghouse is shaken
for 4 seconds. Thus the total cycle time for all 12 baghouses is
6 minutes.
4. Six 250-hp fans located at ground level. Each fan exhausts a pair
of baghouses (60,000 acfm) and discharges to a 60-foot stack.
5. A 100-ton fresh alumina bin.
6. A continuous low-pressure conveying system to convey reacted and
unrecycled alumina to the existing day bins on top of both pot-
line 1 buildings.
7. Local controls mounted on the baghouse structure.
8. A roof of simple truss design covering the Venturis, baghouses,
and local controls. The roof is about 20 feet above the tops
of the baghouses and is supported by 14 I-beams, 7 to a side.
6-107
-------
FAN WITH
STACK
en
i
o
oo
EXHAUST FROM
POTROOMS
NOT SHOWN ARE
11 DUPLICATE
BAG HOUSES AND A
FEEDER DUCT, IN-
STALLED AT
THE RIGHT
FEEDER DUCT
6ft.
REACTED ALUMINA TO
EXISTING DAY BINS
ON BOTH BUILDINGS
BATCH WISE
HIGH PRES-
SURE CONVEY-
ING SYSTEM
Figure 6-20. Retrofit schematic -- plant C - Alcan design (side view -- 5
baghouses in a farther plane are not shown).
-------
Table 6-28. MAJOR RETROFIT ITEMS—PLANT C--ALCAN DESIGN
1. Two 6-foot ducts, one from each building (see Figure 6-20),
join an 8-foot duct that runs underneath the control unit
and feeds 11 baghouses. The horizontal duct reduces in diameter
after each pair of take-offs. There are three alumina injection
ports -two for recycled alumina and one for fresh—in the 8-foot
duct upstream of the control unit. The ratio of recycled to fresh
alumina is variable from one to 10.
2. Ductwork identical to that just described under item 1 to feed
a control unit that is installed to the right, and in reverse,
of the control unit shown in Figure 6-20.
3. Twenty-two baghouses in 2 groups of 11 each as shown in Figure 6-20.
Each baghouse except the eleventh (nearest the bin) has a twin
beyond the plane of the paper. Each baghouse handles 16,400 acfm
of 200°F primary exhaust. Gas flow entering and leaving a baghouse
is vertical. Each baghouse is a rectangular box 10 feet square
and 18 feet high with an inverted pyramid bottom gas inlet. The
top of each baghouse is about 50 feet above the ground. The bags
are cleaned by a variable 15- to 30-second high pressure jet air
pulse.
4. Twenty-two 60-hp fans, one for each baghouse (16,400 acfm).
Each fan sets on top of its respective baghouse and discharges
to a stack. The stacks discharge to the atmosphere 60 feet above
the ground.
5. A 100-ton fresh alumina bin.
6. A batchwise high-pressure conveying system to alternately convey
reacted and unrecycled alumina to the existing day bins on top
of each potline building.
7. Local controls mounted on the baghouse structure.
6-109
-------
6.3.3.1.9 Retrofit increments of jirgqress--Table 6-29 presents the
time increments of progress for each of the three ootline retrofits.
As Table 6-29 shows, the four major contracts were awarded at different
times for lines 2 and 3, but for line 1 they were awarded all at once.
Compliance testing for line 2 started after several weeks of shakedown
operation, approximately on September 15, 1972. The line 1 retrofit was
operational when EPA personnel visited plant C on May 9, 1973.
For each of the potline retrofits, only 9 to 10 months elapsed
between the date that the first contract was awarded and the date that
both control units on that potline were operational. However, as mentioned
in subsection 6.3.3.1.4, the plant started investigating improved control
in 1970, which was 3 years prior to all control units being operational.
6.3.3.2 Emissions Before and After Retrofit
Table 6-30 shows average emissions before and after retrofit
furnished by the company in October 1974. All of the quantities are
expressed as pounds of total fluoride ion per ton of aluminum produced
(Ib/ton Al). The table shows the quantities generated at the cells;
the quantities directed to the injected alumina primary removal equip-
ment after retrofit (primary collection); the quantities escaping
collection (secondary loading); the primary, secondary, and total
emissions; the quantities removed by the secondary equipment that are
sent to the cryolite recovery plant; and the quantities recovered by
the dry primary retrofit and recycled to the cells.
6-110
-------
Table 6-29. RETROFIT INCREMENTS OF PROGRESS—PLANT C
Portion of control unit
contract awarded
Treatment system and
baghouses
Hood fabrication and
installation
Ducting fabrication and
installation
Electrical wiring
Total project
Construction started
Construction completed
Unit operational
Line 2
11/19/71
12/28/71
2/11/72
2/17/72
-
January 1972
Unit 1 Unit 2
6/27/72 8/4/72
7/5/72 8/15/72
Line 3
1/26/72
2/2/72
2/16/72
2/17/72
-
NAa
Unit 3 Unit 4
9/29/72 10/27/72
*1 0/1 5/72 *1 1/1 5/72
Line 1
-
-
-
7/7/72
8/1/72
Unit 5 Unit 6
3/1/73 4/10/73
NAa NAa
t
C
I
IT;
Not available.
-------
Table 6-30. EMISSIONS BEFORE AND AFTER RETROFIT—PLANT C-
LINES 1, 2, AND 3
(15 total F"/ton Al)
Emissions
Generation
Primary collection
Primary emission
Secondary loading
Secondary emission
Total emission
Secondary removal
Primary recovery
Before
retrofit
45.5
-
-
44.5
9.0
9.0
35.5
-
After
retrofit
45.5
37.8
0.4
6.7
0.9
1.3
5.8
37.4
Monthly average inlet loadings to the primary and secondary con-
trol systems were 37.76 and 6.72 Ib F/ton Al in September 1974. The
generation level of 45.5 Ib/ton Al is the sum of these loadings, plus
a rough approximation that building leakage is 1.0 Ib/ton Al. These
loadings and the secondary emission of 9.0 Ib/ton Al before retrofit
were measured with the plant operating at capacity. The primary and
secondary emissions of 0.4 and 0.9 Ib/ton Al are based on 92 and 93
tests, respectively, during January-September 1974 when the plant
was at or near full production. For these nine months, testing typi-
cally consisted of three tests per week on both the primary and secondary
fi-119
-------
systems. Each of the three tests was for emissions from a different
potline and lasted 24 hours. Plant personnel have not been able to
determine any difference between the performance of the POP units and
that of the Alcan units.
The emissions before retrofit in Table 6-30 correspond to a secon-
dary removal efficiency of 80 percent and an overall control efficiency
(including building leakage) of 78 percent on total fluoride for plant
C. The emissions after retrofit correspond to a primary removal
efficiency of 99 percent, a secondary removal efficiency of 87 percent,
a primary collection efficiency of 83 percent and an overall control
efficiency (including leakage) of 95 percent on total fluoride for
plant C.
Two conclusions that can be drawn from the above efficiencies and
Table 6-30 are:
1. Without secondary control, a primary collection efficiency
of 83 percent would result in a secondary emission of 7.7
Ib/ton Al and a total emission of 8.1 Ib/ton Al for total
fluoride after retrofit.
2. The retrofit reduced by 84 percent the quantity of total
fluoride that is removed by the secondary control system and
sent to the water treatment plant. This in turn has reduced
the plant water effluent discharges.
6^113
-------
6.3.3.3 Retrofit Capital and Annual Costs
6.3.3.3.1 Capital costs—Table 6-31 presents the total retrofit
capital cost for the three potlines broken down into the major retrofit
items. Assuming an annual capacity of 265,000 tons, $14,300,000 for
retrofit is equivalent to a capital cost of $54 per annual capacity ton.
The duct costs in Table 6-31 include all ductwork from the cells
to the 8-foot horizontal ducts underneath the control units. The
control unit costs include the remaining ductwork, Venturis, baghouses,
fans, stacks, and solids handling for all the Alcan and PDF units.
Nondistributed costs are primarily, but not exclusively, related to
the control units and include such things as utilities (primarily
compressed air) and instrumentation. Research and development (R&D)
costs include only the development work that eventually became part
of the retrofit. Hence, the costs are included for the hoods on'the
20 cells whose primary exhausts were directed to the ESP prototype,
but not for the ESP itself. Plant personnel are unable to determine
the remaining R&D costs from their records. All contractor engineering
and the plant R&D engineering costs that pertain to the installed
retrofit are included in the Table 6-31 costs. Plant personnel are
unable to determine the remaining plant engineering costs from their
records.
The secondary scrubbers were the only assets that were retired
as a result of the retrofit. They were installed for $1,166,000,
were being depreciated over a 20-year life, and when retired had a
book value of $907,000.
6-114
-------
Table 6-31. RETROFIT CAPITAL COST-PLANT C-
LINES 1, 2 AND 3
Hoods $3,160,000
Ducts 1,190,000
Emission control units 7,970,000
Nondistributed costs 1,250,000
Research and development 730.000
Total $14,300,000
6.3.3.3.2 Annual costs—Table 6-32 gives annual costs for both the
primary injected alumina retrofit and the secondary scrubbers as
furnished by the company in March 1975. Assuming an annual aluminum
production equal to the annual capacity of 265,000 tons, the total
retrofit annual cost amounts to $12.99 per ton; the total secondary
scrubber annual cost amounts to $7.07 per ton; and the plant's ool-
lution control an.nual cost amounts to $20.06 per ton. The total retro-
fit annual operating cost of $936,000 amounts to $3.53 per ton.
The cost of producing compressed air for the retrofit is included
in maintenance materials. The plant pays no royalty costs for the
Alcan or the Prat-Daniel-Poelman designs. The secondary scrubbers are
leased. Hence the depreciation cost of $1,190,000 is rent, and there
are no charges for interest or taxes.
6-115
-------
In Table 6-32, no credit is given for the alumina and fluoride
recovered by the retrofit because the plant accounting system does
not credit these recovered materials. Assuming a fluoride recovery
rate of 37.4 Ib/ton of aluminum produced (see Table 6-30), an annual
aluminum production of 265,000 tons, and a fluoride cost of $0.25 per
pound,50 the value of the recovered fluoride would be $2,477,750
per year.
Table 6-32. 1974 ANNUAL COST—PLANT C—LINES 1, 2, AND 3
Operating costs:
Labor incl . dir. supv.
Supplies
Electricity
Mater
Maintenance materials & labor
Bag replacement
Subtotal
Capital -related charges:
Depreciation
Interest
Insurance
Taxes
Administrative & overhead
Subtotal
Total
Injected
Alumina
Retrofit
413,000
20,000
130,000
-
152,000
221 ,000
936,000
973,000
1,317,000
9,000
176,000
32,000
2,507,000
3,443,000
Secondary
Scrubbers
141,000
38,000
104,000
21 ,000a
355,000
_
659,000
l,190,000a
-
7,000a
-
17,000
1,214,000
1,873,000
Both
554,000
58,000
234,000
21 ,000
507,000
221 ,000
1,595,000
2,163,000
1,317,000
16,000
176,000
49,000
3,721,000
5,316,000
a
Estimated.
6-116
-------
6^.3.4 Case Description Summary
Table 6-33 shows actual retrofit emission reductions and cost
for potroom retrofits at ten primary aluminum plants.3"6'43 EPA
personnel had visited seven of these plants (A-G) at the time the
detailed case descriptions were developed. Since any one of these
seven could have served as a retrofit case description, a comparison
table and a rating sheet were prepared to select the best cases.
The three cases that have been selected (plants A, B and C)--
together with the discussion of primary collection, primary removal
and secondary removal systems—are believed to adequately cover
primary aluminum fluoride retrofit control techniques.
The emission numbers in Table 6-33 are average total primary
and secondary total fluoride emissions expressed as pounds of fluoride
ion per ton of aluminum produced. The increase in plant K emissions
after retrofit is explained in subsection 6.3.4.2. The capital costs
include direct and indirect costs. The indirect costs include
engineering and, where a retrofit is underway, contingency and
escalation costs. However, as noted in Section 6.3.3, not all the
engineering costs are included in the plant C retrofit. Except for
plants G and M, the retrofit costs are final or, where the retrofit
is still underway, are the customary accurate appropriation request
estimates. Plant G costs are based on written vendor quotations and
should thus be reasonably accurate. Since the accuracy of plant M
costs is questionable, this plant is separated from the others in
Table 6-33. The capital costs are also shown adjusted to April 1974
using plant cost indices from Chemical Engineering magazine.
6-117
-------
Table 6-33. POTROOM RETROFIT EMISSION REDUCTIONS AND COSTS FOR TEN PRIMARY ALUMINUM PLANTS
Cell
type
CWPB
CWPB
CWPB
CWPB
SWPB
SWPB
HSS
HSS
VSS
VSS
Plant
code
D
Fd
G
H
C
K
B
A
E
M
Plant capacity
short tons/yr
115,000
32,850
250,000
130,000
265,000
35,000
210,000
80,000
91 ,000
180,000
Total ootroom emissions,
Ib F/ton Al
Before .
retrofit
7.8
5.1
19.0
6.9
9.0
7.7
5.4
9.3
4.2
5.3
After .
retrofit
2.6
1.2
2.7
2.5
1.3
10.6
2.9
2.4
2.0
1.9
Retrofit capital cost,
$/annual ton Al
Actual
102
54
124
216
54
121
92
141
64
115
Adjusted to
Anril 1974
117
71
108
188
62
105
98
157
81
121
Increased net
annual operating
cost, $/ton Al
NAC
NA
NA
NA
3.53e
NA
0.81
-4.78f
MA
9.57
-------
Table 6-33 (continued). POTROOM RETROFIT EMISSION REDUCTIONS AND COSTS FOR TEN PRIMARY ALUMINUM PLANTS
FOOTNOTES:
aCWPB - center-worked prebake cells, SWPB - side-worked prebake cells, VSS - vertical stud
Soderberg cells, HSS - horizontal stud Soderberg cells.
Average primary and secondary total fluoride emissions.
CNA = Not available.
Results shown for only the one potline retrofitted.
Increased gross annual operating cost; net not available. Net annual operating cost includes
credits for recovered alumina and fluoride; gross does not.
f
"Negative sign means decreased net annual operating cost.
-------
Plants A and B furnished the additional net annual operating cost and
plant C furnished the additional gross annual operating cost for their
retrofits as shown in Table 6-33. Additional net annual operating cost
is also shown for plant M. These annual operating costs do not include
capital-related charges.
Table 6-33 shows as much as a three-fold variation in cost for
actual retrofits. Real-life differences between plants that can
affect the cost include: the need to tear down varying types of
existing control; the possible need to tear down other equipment and
buildings; the extent to which support structure for the retrofit
already exists; and, the need for installing or modifying primary
collection systems. The latter includes the extent of modification
that is dictated by potroom layout and by cell geometry and operating
requirements.
To further illustrate the complexity of real-life situations,
the vendor of the fluidized bed claims that the installation cost of
fluidized bed removal equipment can vary greatly, from as low as about
$30 per annual ton on some new prebake installations to levels such
as shown for Plant D 1n Table 6-33 ($117 per annual ton). This four-
58
fold variation 1n cost 1s largely determined by the following factors:
1. The volume of cell gas to be treated per ton of metal
produced. Smaller and older design prebake cells, such as
those of Plant D, genertte as much as twice the gas volume
of some newer cell designs on a cubic foot per ton basis.
6-120
-------
2. The physical layout of the existing plant, which affects:
a. The length of the duct svstem.
b. The availability of alumina storage tanks for storage
both preceding and following the fume treatment system.
c. The available space for locating the fume treatment svstem
d. The access to the area by large construction equipment
used to erect the reactors, baghouses, etc.
e. The availability of sufficient electrical power close to
the site chosen for the control equipment.
This section Illustrates the point that for a process as complex
as a primary aluminum plant, a retrofit control must be tailor-made
and should not be generalized as to costs or even as to method of
emission control.
In the following subsections, capsule descriptions of each of
the ten actual retrofits are given by cell type.
6.3.4.1 Center-worked Prebake Cells
Plant D completed a central primary fluidized bed dry scrubbing
retrofit in July 1974. A total of 25 reactor-baghouses units were
installed, along with supporting equipment, to replace 30 courtyard
rotoclone-to-spray tower fume control units on the five plant
potlines. Total system capacity is 1,250,000 acfm. The retrofit
did not improve primary collection efficiency, although the capital
cost included replacing the side shields on all 650 cells with new
identically-designed covers. There was no secondary control before
or after retrofit. Total retrofit capital cost was $11,766,900
which included the cost of removing and relocating the former control
equipment.
6-121
-------
Plant F completed a courtyard primary fluidized bed dry scrubbing
retrofit on one of its five potlines in May 1970. A total of 10
reactor-baghouse units were installed, along with supporting equip-
ment, to replace three courtyard dry ESP-to-dual spray tower fume
control units. Total system capacity is 400,000 acfm, and the
retrofit did not improve primary collection efficiency. There
was no secondary control before or after retrofit. Total retrofit
capital cost was $1,772,000. This did not include the cost of
removing the six spray towers. The three ESPs were left in
place because it was considered too costly to remove them.
Plant G has 12 courtyard dual multiclone-to-quadruple spray
tower primary control units and plans to install a primary courtyard
fluidized bed or injected alumina dry scrubbing retrofit on all six
potlines by July 1978. As of January 1975, the retrofit capital
cost estimate for the dry scrubbers was $28 million, the median of
three vendor preliminary estimates. The retrofit also includes im-
proved primary collection efficiency by modifications to the plant's
1032 cells. These modifications included tighter sealing between the
hood side shields and around the anode stems, replacement of the curved
side shields with braced, flat side shields, and installation of new
end doors. Cost of these hooding modifications is estimated at $3 mil-
lion for a total retrofit cost of $31 million. Table 6-33 shows the
combined emission reductions and costs for the hooding-dry scrubbing
retrofits. Plant G has no secondary control before or after retrofit.
6-122
-------
• Plant H plans to install a central primary control injected alumina
dry scrubbing retrofit on all five potlines by November 1976. Present
controls on four of the five potlines are courtyard primary multi-
clone-to-spray tower units and 50 secondary cyclone scrubbers per
potline along one edge of the potroom roof. Present controls on the
other potline are primary central Venturis with no secondary controls.
The retrofit includes new hoods on all 700 cells which will improve
primary collection efficiency to such a degree th*at the company plans
to abandon all secondary controls. As of February 1975, the total
retrofit capital cost was estimated at $28,046,000, equivalent to
$216 per annual ton in Table 6-33. This figure does not include any
costs for dismantling existing equipment. In addition, plant H plans to
install two parallel sets of spray cyclone scrubber-to-wet ESP
controls on its uncontrolled anode bake plant at a cost of
$2,150,000.
6.3.4.2 Side-worked Prebake Cells
The Plant C retrofit is described in detail in Section 6.3.3.
In April 1973, plant C completed a courtyard primary injected
alumina dry scrubbing retrofit on all three potlines. There are
six control modules with a total system capacity of 2,160,000 acfm.
Former control consisted only of 180 roof-mounted secondary spray
scrubbers. By necessity, the retrofit included the hooding of all
720 cells. The total retrofit capital cost of $14,300,000 included
the removal of 20 secondary scrubbers.
6-123
-------
Plant K plans to retrofit its one potline with a central
primary injected alumina dry scrubbing system and to abandon the
present spray screen secondary controls along the entire peak of
the potroom roof. The retrofit also includes hooding all 90 cells and
oversizing the removal equipment to handle primary exhaust
from an additional 48 cells that are part of a possible plant expan-
sion. The total retrofit capital cost was estimated to be $4,250,000
in March 1975. The plant plans to abandon secondary control because
they consider a projected capital cost of $56 per annual ton for water
treatment of their once-through scrubbing water to be economically ex-
cessive. The before retrofit emission of 7.7 Ib F/ton Al is an average
for the first five months of operation in 1974. The after-retrofit
emission of 10.6 Ib F/ton Al is based on an average generation level
of 53 Ib F/ton Al for the same five months and a projected
plant overall control efficiency of 79.86 percent. The actual emis-
sion level will not be known until the retrofit has been completed in
the summer of 1975 and then operated for several months. Plant personnel
are hopeful that emissions will average 6-7 Ib F/ton Al.
6.3.4.3 Horizontal Stud Soderberg Cells
The Plant B retrofit is described in detail in Section 6.3.2.
Former controls were courtyard primary spray towers. The plant is
installing fifteen 100,000 scfm and six 50,000 scfm courtyard pri-
mary spray tower-to-wet ESP units on the six potrooms comprising
two-thirds of its capacity, and ten 100,000 scfm central primary
wet ESP-only units on the three potrooms comprising the other one-
third. The former does not include improved primary collection while
the latter does. Improved collection on the latter includes an in-
creased exhaust rate, new doors, and better sealing on 372 cells.
6-124
-------
Estimated retrofit completion date is June 1975. There was no secon-
dary control before or after retrofit. An October 1974 total retrofit
capital cost estimate of $19,300,000 does not include costs of removing
any of the 22 existing spray towers for the central retrofit.
The Plant A retrofit is described in detail in Section 6.3.1.
The plant has installed a central primary dry scrubbing retrofit in
two locations, each having 18 reactor-baghouse units and handling
two potrooms, or half the plant's capacity. For half the capacity,
the retrofit involves bypassing the 16 spray towers at the ends of the
potrooms and improving primary collection efficiency on all 240 cells
by an increased primary exhaust rate. For the other half, the
retrofit involves using the central ductwork of the existing cement
blockhouse scrubbers and not improving primary collection efficiency.
Total system capacity for the whole plant is 1,200,000 acfm. The
retrofit was operational in September 1974. There was no secondary
control before or after retrofit. A December 1974 total retrofit
capital cost estimate of $11,313,000 includes demolition costs for
half the retrofit. The bypassed spray towers ana a 25- by
100-foot building were torn down, but the cement blockhouse scrubbers
were not.
6.3.4.4 Vertical Stud Soderberg Cells
Plant E completed a secondary retrofit in November 1970 and a
primary retrofit in February 1972 on all five of its potrooms. The
secondary retrofit consisted of abandoning previously retrofitted
roof monitor spray screen scrubbers and installing a new dormer-
tunnel design that is shown in Figure 6-10, one dormer tunnel along
one enttre edge of each potroom roof. The primary retrofit
6-125
-------
consisted of adding eight 12,000 acfm and four 6,000 acfm wet
ESPs downstream of 20 previously retrofitted courtyard bubbler-
scrubbers. The ESP retrofit did not improve primary collection
efficiency. Table 6-33 shows the combined emission reduction and
costs for the dormer-tunnel and ESP retrofits. Retrofit capital
costs were $4,155,078 and $1,662,701 for the secondary and primary
retrofits, respectively. The primary retrofit included removal of
the plant's 20 multiclones.
Plant M has ten potrooms, courtyard multiclone-to-
venturi primary controls and no secondary control. It has been
developing a foam scrubber secondary control system. If this scrub-
ber proves too ineffective or costly, the plant will revert to in-
stalling spray screen secondary controls. An EPA contract study
estimated that, in December 1973, roof mounted powered spray screen
scrubbers would cost $20,688,000 or $115 per annual ton to reduce total
fluoride emissions to 1.8 Ib F/ton Al. There would be 60 scrubbers
and 60 fans per potroom, or 600 apiece for the plant. Total system
capacity would be 25,800,000 acfm with a Hquid-to-gas ratio of 5
gallons per thousand acfm. The retrofit would also Include 20 redr-
culating pumps, 10 recirculating tanks, six miscellaneous pumps, and
one clarifier. The scrubber water would be lime treated. The con-
tractor estimated that final installed costs for other systems, such
as a foam scrubber, would not vary more than about 30 percent from
that of the spray screen. A December 1973 annual1zed operating cost
estimate of $1,723,000 1s equivalent to $9.57 per ton.
6-126
-------
6.4 DESIGN, INSTALLATION AND STARTUP TIMES FOR RETROFIT CONTROLS
The emission control retrofit cases studied and described in
Section 6.3 have shown that the upgrading of fluoride emission controls
or the initiation of control for a primary aluminum plant is a major
engineering undertaking. Such a project does not involve the instal-
lation of one simple item of control equipment, but instead involves
complex controls which may be several in number. Associated with the
controls are storage and surge tanks, conveyors, fans, and long lengths
of huge ductwork along with the necessary foundations, structural
steelwork and electrical drive systems.
Table 6-34 shows the approximate sequence of activities which are
necessary to design and install an improved air emission control system
in a primary aluminum plant. The sequence of work outlined is not
necessarily normal, but it should apply to periods such as the summer of
1974, when structural steel had particularly long delivery time.
Obviously, such steel would be ordered as soon as possible—in fact,
even before the full requirement is known. Thus, some parts of item 5
may not be firm until item 7 and item 11 are done. Similarly, it will
be understood that other items of Table 6-34 may overlap in time.
Figure 6-21 illustrates that the activities in a big engineering
job—such as retrofitting controls to a primary aluminum plant—tend to
progress in a continuous, non-stepwise manner. This is because there is
so much to do; at a given time, numerous items are in various stages of
design, procurement, and construction. The four curves in Figure 6-21
show the typical progress for the named activites throughout
6-127
-------
Table 6-'34- SEQUENCE OF MAJOR ACTIVITIES IN DESIGN AND CONSTRUCTION
OF AIR EMISSION CONTROL FOR AN EXISTING PRIMARY ALUMINUM PLANT
1. Process design and flow diagram.
2. Engineering flow diagram and preliminary plot plans.
3. Specification and procurement of major items such as dry scrubbers,
and fans. Long delivery items first.
4. Ductwork and piping arrangements, specification, and procurement.
5. Structural steel design.
6. Foundation design.
7. Specification of minor items, obtainable without complete drawings,
such as pumps and materials handling equipment.
8. Design of electrical starters, switchgear and distribution system.
9. Specification of instruments.
10. Receipt of certified dimension drawings of dry scrubbers, storage
tanks, conveyors, fans.
11. Dimension drawings for ductwork.
12. Release of foundation and structural steel drawings.
13. Start construction. Site preparation, necessary removals or
relocations will have already taken place.
14. Complete the pipe and ductwork takeoffs, and drawings for field
supports.
15. Release drawings and material listings for construction.
16. Complete underground installations.
17. Complete foundations.
18.. Delivery of structural steel and major items of equipment.
19. Erect major items of equipment.
20. Install ductwork and conveyors.
21. Install piping.
6-128
-------
Table 6-34(continued). SEQUENCE OF MAJOR ACTIVITIES IN DESIGN AND
CONSTRUCTION OF AIR EMISSION CONTROL FOR AN EXISTING PRIMARY ALUMINUM
PLANT
22. Install electrical.
23. Install instrumentation.
24. Startup.
25. Source testing and analytical.
26. Compliance with air pollution control regulations.
the job. The relative positions of the curves vary with the actual job
and the graph is diagrammatic only. However, each line tends to approach
linearity in the 25-75 percent completion interval. This figure shows
that process design usually continues into the early stages of procurement
Engineering also continues well into the construction period. For this
reason, total time requirements are best estimated from experience and
cannot be derived by adding the time requirements for design, ordering,
manufacture, delivery, installation and startup as can be done for one
simple control.
One important step that is almost wholly out of control of the
customer or the control official is the construction item delivery
time. Table 6-35 gives some historical delivery times for items which
are very important in installing emission controls at primary aluminum
plants. The historical variation is somewhat obscured because data
extending back to the Korean war period (when deliveries were very
long) is not available. However, deliveries greatly increased
from 1973 to 1974, and many lead times passed all previous
6-129
-------
TOO
CO
o
START
(a) PROCESS DESIGN
(b) PROCUREMENT
(c ENGINEERING
(j) CONSTRUCTION
FINISH
TIME
Figure 6-21- Diagrammatic representation of activity schedules on a major
process industry construction project.
-------
Table 6-35. DELIVERY TIMES FOR ITEMS REQUIRED TO CONSTRUCT EMISSION
CONTROLS FOR PRIMARY ALUMINUM PLANTS'59,60
CO
Construction Items
Structural steel
Ductwork
Fans & blowers
Airslides
Motors
Electrical controls
Electrical switch gear
1960
17-21
10-14
26
Delivery Times (weeks)
1966
23-30
10-14
30
1969
21-28
17-22
17-19
12
13
23
1971
23-31
18-23
25
17-19
12
13
21
1973
27-35
18-23
26
20-22
14
15
21
May 1974
34-50
32-37
26
28-30
36
25
35
Remarks
>500 tons
Large
No specials
AC standard
<_ 600 volt
-------
bounds. Table 6-3E represents the experience of an aluminum company
doing its own purchasing. Another company reports up to 52 weeks for
delivery of switchgear. 61 A contractor reports 61 weeks for blowers;
62
and about 65 weeks for motors over 40 HP. Deliveries may depend
partly upon quantity bought, continuity of business through the years,
and most-favored customer status. Fabrication and shipping consumes a
significant fraction of the total time required to design and install
emission controls.
The actual time in years that was required to add retrofit controls
to eight aluminum plants is given in Table 6-36. Plant codes are the
same as in Section 6.3. Except for plant F, the whole plant was retrofitted
in each case. Only plants C, E, and H had secondary control and only
plant E improved its secondary control, at a cost of about 65 percent of
its total retrofit expenditure. Plant B built and operated a pilot
plant during two of the 4-1/2 years of retrofit activity. The completion
time of 5-1/2 years for plant G includes 3 years for improved cell
hooding and 3 years for dry scrubber installation. The 3 years for
improved hooding is due to a claimed economic advantage for modifying
cells over the normal 3-year life of their cathode linings. Had plant 6
so elected, the dry scrubber installation could have proceeded simultaneously
with cell hooding improvements, reducing the completion time to about 3
years.
The actual time requirements shown in the last column of Table 6-36
are probably greater - on the average - than needed for enforcement
purposes. In spite of the large capital tied up, there is no return,
and the usual economic incentive for haste in startup is lacking. Any
interferences with production during installation of controls are
6-132
-------
Table 6-36. TOTAL CONSTRUCTION TIME FOR RETROFIT EMISSION CONTROLS FOR
PRIMARY ALUMINUM PLANTS
CO
GO
Plant
Code
A
B
C
D
Cell
Type
HSS
HSS
SWPB
CWPB
IE ; vss
Plant Capa-
city
(tons/yr)
80,000
210,000
265,000
115,000
1 91,000
Retrofit Capital
Cost ($)
11,300,000
19,300,000
H, 300, 000
1 1 ,800^000
5,800,000
; F • Ck'PB : 32,850 I 1,800,000
i ; '
! G
H
CWPB
CWPB
250,000
130,000
: 31,000.000
28,000,000
Description of
Retrofit Emission Controls
Dry scrubbers-primary
Improved cell hooding and wet
ESP-primary
New cell hooding and dry
scrubbers-primary
Dry scrubbers-primary
Wet ESP-primary
Dormer tunnel -secondary
Time Required
for Retrofitting
(years)
2-1/2
4-1/2
3
3
2-1/2
Dry scrubbers-primary 1-1/2
i (1 potline)
i
Improved cell hooding and dry
scrubbers -primary
Improved cell hooding and dry
dry scrubbers-primary
5-1/2
2
-------
normally few and brief and, of course, the plant will make haste when
these occur.
In view of the above discussion, a reasonable total time for
retrofitting fluoride emission controls to a primary aluminum plant may
be taken as about 2-1/2 years. Table 6-37 shows the approximate lead
times required to reach a few important milestones in providing emission
controls for an existing primary aluminum plant. The first item in the
table can require a year or two if piloting must be done or if con-
siderable cost study has not already been done. Also, for reasons shown
by Figure 6-21, the time for item 2 must be given as a range.
In practice, enforcement officials should consider each plant on a
case-by-case basis and they should require proof for the time requirements
claimed for each milestone.
Table 6-37. INCREMENTS OF PROGRESS FOR INSTALLATION OF FLUORIDE
EMISSION CONTROLS IN AN EXISTING PRIMARY ALUMINUM
PLANT
Increments of Progress Elapsed Time, weeks
Preliminary control plan and compliance 25
schedule to appropriate agency
Award of major contracts 35 - 55
Start of construction 60
Completion of construction 124
Final compliance 130
6-134
-------
6.5 REFERENCES FOR SECTION 6
1. Air Pollution Control in the Primary Aluminum Industry. Sing-
master and Breyer, New York, NY. Prepared for Office of Air
Programs, Environmental Protection Agency, Research Triangle
Park, NC under Contract Number CPA 70-21. July 23, 1973.
2. Background Information for Standards of Performance: Primary
Aluminum Industry. Volume 1: Proposed Standards. Emission
Standards and Engineering Division, OAQPS, Environmental
Protection Agency, Research Triangle Park, NC. October 1974.
3. Varner, B. A., Trip Report: Six Pacific Northwest Primary
Aluminum Plants. Emission Standards and Engineering Division,
OAQPS, Environmental Protection Agency, Research Triangle
Park, NC. June 11, 1973.
4. Varner, B. A., Trip Report: Ormet Hannibal, Ohio Primary
Aluminum Plant. Emission Standards and Engineering Division,
OAQPS, Environmental Protection Agency, Research Triangle
Park, NC. July 2, 1973.
5. Varner, B. A. and G. B. Crane, Report on 1975 Primary Aluminum
Pacific Northwest Trip. Emission Standards and Engineering
Division, OAQPS, Environmental Protection Agency, Research
Triangle Park, NC. March 17, 1975.
6. Crane, G. B. and B. A. Varner, Trip Report to the Primary Aluminum
Plant of Conalco at Lake Charles, LA. Emission Standards and
Engineering Division, OAQPS, Environmental Protection Agency,
Research Triangle Park, NC. March 28, 1975.
7. Personal communication from D. Rush, Singmaster and Breyer, New
York, NY, to B. A. Varner, Emission Standards and Engineering
Division, OAQPS, Environmental Protection Agency, Research
Triangle Park, NC. February 25, 1974.
8. Letter from C. A. MacPhee, Reynolds Metals Company, Richmond,
VA, to D. R. Goodwin, Emission Standards and Engineering
Division, OAQPS, Environmental Protection Agency, Research
Triangle Park, NC, dated November 19, 1974.
9. Letter from Dr. P. R. Atkins, Aluminum Company of America,
Pittsburgh, PA, to B. A. Varner, Emission Standards and Engineering
Division, OAQPS, Environmental Protection Agency, Research
Triangle Park, NC, dated April 21, 1975.
10. Memorandum from G. B. Crane to S. T. Cuffe. Meeting with Conalco
and Revere on Proposed Primary Aluminum lll(d) Standards.
September 26, 1974. (Also meeting notes of Crane and B. A. Varner.)
6-135
-------
11. Boyer, W. F., Jr., Arguments in Opposition to State Guidelines for
Standards of Performance for Existing Primary Aluminum Plants (Second
Draft of August 1974). Consolidated Aluminum Corporation, St.
Louis, MO. September 23, 1974. pp. 6-8.
12. Letter from C. L. Keigley, Consolidated Aluminum Corporation, Lake
Charles, LA, to G. B. Crane, Emission Standards and Engineering
Division, OAQPS, Environmental Protection Agency, Research Triangle
Park, NC, dated April 1, 1975.
13. Letter from J. L. Loyer, Howmet Corporation, Greenwich, CT,
to D. R. Goodwin, Emission Standards and Engineering Division,
OAQPS, Environmental Protection Agency, Research Triangle Park,
NC, dated October 31, 1974. pp. 1-2.
14. Letter from Dr. P. R. Atkins, Aluminum Company of America,
Pittsburgh, PA, to G. B. Crane, Emission Standards and
Engineering Division, OAQPS, Environmental Protection Agency,
Research Triangle Park, NC, dated January 30, 1975. p. 9.
15. Reference 14, above, p. 3.
16. Letter from Dr. P. R. Atkins, Aluminum Company of America,
Pittsburgh, PA, to D. R. Goodwin, Emission Standards and
Engineering Division, OAQPS, Environmental Protection Agency,
Research Triangle Park, NC, dated June 20, 1974.
17. Letter from J. L. Byrne, Martin Marietta Aluminum, The Dalles, OR,
to D. R. Goodwin, Emission Standards and Engineering Division,
OAQPS, Environmental Protection Agency, Research Triangle Park,
NC, dated October 10, 1974.
18. Letter from Dr. B. S. Hulcher, Reynolds Metals Company, Richmond,
VA, to D. R. Goodwin, Emission Standards and Engineering Division,
OAQPS, Environmental Protection Agency, Research Triangle Park,
NC, dated June 28, 1974.
19. Reference 1, above, pp. 8-10 to 8-12.
20. Reference 1, above, p. 8-8.
21. Reference 1, above, pp. 8-27 to 8-29.
22. Reference 1, above, p. 8-22.
23. Cook, C. C. and L. L. Knapp. Treatment of Gases Evolved in the
Production of Aluminum. U.S. Patent No. 3,503,184. March 31, 1970.
24. Cook, C. C. and G. R. Swany. Evolution of Fluoride Recovery
Processes Alcoa Smelters. In: Light Metals 1971. New York, NY.
Proceedings of Symposia 100th AIME Meeting, March 1-4, 1971.
6-136
-------
25. Cook, C. C., G. R. Swany, and J. N. Colpitts. Operating
Experience with the Alcoa 398 Process for Fluoride Recovery-
Journal of the A1r Pollution Control Association. 21_, August 1971.
26. Cochran, C.N., W. C. Sleppy, and W. B. Frank. Chemistry cf
Evolution and Recovery of Fumes in Aluminum Smeltinq. TMS of
AIME Paper No. A70-22. February 16, 1970.
27. Reference 1, above, p. 5-27.
28. Varner, B. A., Trip Report: Trip to Alcoa (Badin, N.C.) Alumi-
num Plant, Standards Development and Implementation Division,
SSPCP, Environmental Protection Agency, Research Trianole Park,
N.C. August 21, 1972.
29. Reference 1, above, pp. 5-14 to 5-17.
30. Reference 1, above, p. 5-29..
31. Reference 1, above, p. 5-34.
32. Control Techniques for Particulate Air Pollutants. National
Air Pollution Control Administration, Washington, D. C. Pub-
lication Number AP-51. January 1969. p. 54~.
33. Reference 1, above, p. 5-40.
34. Reference 1, above, pp. 5-34 to 5-40.
35. Reference 1, above, pp. 5-29 to 5-33.
36. Reference 32, above, pp. 81-96.
37. A Manual of Electrostatic Precipitator Technology. Southern
Research Institute, Birmingham, Alabama. Prepared for National
Air Pollution Control Administration, Cincinnati, Ohio under
Contract Number CPA 22-69-73. August 25, 1970.
38. Callaioli, G., et al. Systems for Gas Cleaning in Electro-
lytic Cells of Montecatini Edison Aluminum Plant. TMS of
AIME Paper No. A70-57. February 16, 1970.
39- Byrne, J. L. Fume Control at Harvey Aluminum. (Presented at
Annual Meeting Pacific Northwest Section, Air Pollution Control
Association. Spokane, Wash. November 16 - 18, 1970).
*°- Calvez, C. et al. Compared Technologies for the Collection of
Gases and Fumes and the Ventilation of Aluminum Potlines.
(Presented at the AIME Meeting, Chicago. December 11 - 16, 1967.)
6-137
-------
41. Moser, E. The Treatment of Fumes from Primary Aluminum Reduction
Plants. (Presented at International Conference on Air Pollution
and Water Conservation in the Copper and Aluminum Industries.
Paper Number 12. October 21-23, 1969).
42. Letter from W. F. Boyer, Jr., Consolidated Aluminum Corporation,
Lake Charles, La., to B. A. Varner, Emission Standards and Enaineerina
Division, OAQPS, Environmental Protection Agency, Research Triangle
Park, N.C., dated June 25, 1974.
43. Ponder, T. C. and R. W. Gerstle. Particulate and Fluoride Emission
Control, Anaconda Aluminum Company, Columbia Falls, Montana.
PEDCo-Environmental, Cincinnati, Ohio. Prepared for Division of
Stationary Source Enforcement, Environmental Protection Agency
under Contract Number 68-02-1321. February 1974.
44. Air Contaminant Discharge Permit Application. Test Reports.
Ammax Pacific Aluminum Corporation, Warrenton, Oregon. June 1974.
45. Letter from J. J. Miller, Amax Aluminum Company, San Mateo, Calif.
to D. Trerice, Kaiser Engineers, Oakland, Calif., dated June 19,
1974. In: Air Contaminant Discharge Permit Application. Test
Reports. Amax Pacific Aluminum Corporation, Warrenton, Oregon.
June 1974.
46. Memorandum from G. B. Crane to S. T. Cuffe. Retrofitting Secondary
Scrubbers at Existing Primary Aluminum Plants. Scrubber Efficien-
cies. Emission Standards and Engineering Division, OAQPS, Environ-
mental Protection Agency. July 22, 1975.
47. Reference 1, above, p. 5-49.
48. Reference 1, above, p. 7-13.
49. Reference 3, above, pp. 33-39. (Corrected).
50. Reference 1, above, p. 8-31.
51. Reference 3, above, pp.3-8. (Corrected).
52. Miller, I., and J. E. Freund. Probability and Statistics for
Engineers. Englewood Cliffs, N. J., Prentice-Hall, 1965. pp.
349, 350, 413.
53. Development Document for Effluent Limitations Guidelines and New
Source Performance Standards for the Primary Aluminum Smelting
Subcategory of the Nonferrous Metals Manufacturing Point Source
Category. Effluent Guidelines Division, Office of Air and Water
Programs, Environmental Protection Agency, Washington, D.C. Report
Number EPA-440/l-74-019-d. March 1974. '
0-138
-------
54. Reference 53, above, p. 43.
55. Reference 53, above, p. 3.
56. Reference 53, above, p. 99.
57. Reference 3, above, pp. 22-27.
58. Letter from Dr. P. R. Atkins, Aluminum Company of America,
Pittsburgh, Pa., to B. A. Varner, Emission Standards and Engineering
Division, OAQPS, Environmental Protection Agency, Research Triangle
Park, N.C., dated July 23, 1973.
59. Personal communication from Dr. P. R. Atkins, Aluminum Company
of America, Pittsburgh, Pa., to B. A. Varner, Emission Standards
and Engineering Division, OAQPS, Environmental Protection Agency,
Research Triangle Park, N.C. June 25, 1974.
60. Purchasing Lead Time Reports. Aluminum Company of America,
Pittsburgh, Pa. August 1969, August 1971, August 1973, May 1974.
61. Letter from A. L. Morgan, PPG Industries, Inc., Pittsburgh, Pa.,
to B. A. Varner, Emission Standards and Engineering Division,
OAQPS, Environmental Protection Agency, Research Triangle Park,
N.C., dated May 30, 1974.
62. Personal communication from R. Davis, Chemical Construction
Corporation, New York, N.Y., to B. A. Varner, Emission Standards
and Engineering Division, OAQPS, Environmental Protection Agency,
Research Triangle Park, N.C. May 31, 1974.
6-139
-------
7. COSTS OF ALTERNATIVE FLUORIDE EMISSION CONTROLS
7.1 INTRODUCTION
This chapter presents the status of fluoride emission control for all
domestic primary aluminum plants as of early 1975. It also illustrates
the application of available control strategies to two selected plant
examples from each of the four cell types. The effect on emissions of
each successive control is shown. Also, cost models are developed and
used to estimate the capital and annualized costs of the above control
strategies. Other plants than those illustrated here can be investigated
for emission reductions and costs in an analogous matter. In general,
cost modules cannot apply closely to any actual plant: they are
approximations, and are especially useful in showing cost comparisons
among various degrees and kinds of control.
Plant code numbers will be spoken of and tabulated in various tables
in this Section. This is done because EPA wishes to avoid identifying
plants, production rates, and other items which may be proprietary, but
are unnecessary to the mission of this document. No meaning should be
sought in the ordering of the code numbers, nor in the numerous uppercase
letters used.
7-1
-------
7.2 SELECTION OF ALTERNATIVE CONTROL LEVELS
Table 7-1 presents the structure of the domestic primary aluminum
industry by cell type. Slight differences in plant capacity exist
between this table and Table 3-1. This is because Table 7-1 includes
planned capacity additions.
Table 7-2 presents the total gaseous plus particulate fluoride
emissions that each domestic plant would be expected to have in the
absence of lll(d) regulations, organized by cell type and with existing
controls. Existing plant emission control, average cell evolution
rates, and collection and removal efficiencies were obtained for the
combinations that describe existing plant control situations in Table
7-2. Most of the evolution rates, primary and secondary loadings, and
emissions were taken from Section 114 letter responses received from
plants representing 100 percent of domestic VSS, HSS, and SWPB capacity,
and 86 percent of CWPB capacity. These responses were supplemented and
modified as necessary with trip reports, letters, phone memoranda and
other EPA file information.
Table 7-3 adds alternate control systems for successive steps from
existing to better control combinations. The following illustrates
the general procedure that has been made specific by Table 7-3 with two
plants from each of the four primary aluminum cell types,
a. First: install best available hooding (primary collection) for
cell type, if needed.
7-2
-------
Table 7-1. PRIMARY ALUMINUM PLANT CAPACITY BY CELL TYPE
(thousands of annual tons)
Company VSSa HSSa CWPBa SWPBa Total
Alcoa 245 -- 1390 — 1635
Reynolds — 704 255 17 976
Kaiser — 341 369 -- 710
Martin Marietta 210 — -- -- 210
Anaconda 180 -- 120 -- 300
Conalco — — — 175 175
Eastalco — -- — 174 174
Intalco -- -- -- 260 260
Revere -- — -- 112 112
Noranda -- — 140 — 140
Ormet -- — 250 — 250
Nat'l Southwire ^ ^ ]80 ==_ 180
Total 635 1045 2704 738 5122
aVSS vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake.
bSpring 1975
7-3
-------
Table 7-2. TOTAL FLUORIDE EMISSIONS BY CELL TYPE WITHOUT lll(d) REGULATIONS
Total Fluoride Emissions, 1b F/ton A1
Cell Type Plant No.
VSS
Weighted Average VSS
SWPB
Weighted Average SWPB
19
20
5A
9
3B
18
17B
23B
11
26
24
HSS
Weighted Average HSS
CWPB
Weighted Average CWPB
31C
16B
13B
25B
31D
30B
28
26
8B
15B
2C
14B
1
29B
10
6B
22B
7B
7A
21B
4B
8A
5B
12
4A
21A
2B
27
Cell Average
42
42
30.5
53.5
42.9
44.5
45.6
53
37.3
41.6
48
32.1
30
36.7
30
31.1
45
28.4
41.6
45.5
25.7
50
33.9
42
40.1
43.2
40
38
50
31.7
43
53
43.5
44.7
41.5
65.6
43
43.2
40.3
Primary
0.03
0.03
0.2
0.8
0.4
0.4
0.4
0.4
0.6
0
4.3
0
0.4
0.3
0.8
0.8
1.3
1.1
1.2
3.5
4.3
1.4
0.2
0.5
0.5
0.5
0.5
0.5
0.6
0.4
0.8
0.8
0.8
0.6
2.1
3.6
2.2
1.8
8.7
11.9
4.4
0
1.8
+ Secondary
2.0
2.0
4.0
4.5
8.6
478
0.9
1.7
10.0
10.6
30.2
48.2
TO"
1.6
1.6
2.4
3.3
4.0
4.5
4.3
30.2
~473
1.0
1.4
1.5
1.5
1.5
2.0
2.0
2.2
1.9
2.1
2.1
2.2
2.8
1.5
3.1
8.3
3.8
2.1
9.9
40.3
4.5
= Total
2.03
2.03
4.2
5.3
9.0
0
1.3
2.1
10.6
10.6
34.5
49.0
13.0
1.9
2.4
3.2
4.6
5.1
5.7
7.8
34,5
~5T7
1.2
1.9
2.0
2.0
2.0
2.5
2.6
2.6
2.7
2.9
2.9
2.8
4.9
5.1
5.3
10.1
12.5
14.0
14.3
40.3
6.3
7-4
-------
Table 7-3. PRIMARY ALUMINUM CONTROL STRATEGIES
VSS CELLS
en
Plant
Code
5A
3B
18
24
Emission Controls Required for the Specified
Hooding Primary (1°) Controls
Best Available Spray tower + wet ESP
" Install waste water
lime treatment
' ' No change (water treat-
ment handled by 2°)
' Fluidized bed dry scrubber
n i n n n
" ' Injected alumina dry
scrubber
None None
Install primary Install injected alumina
Average Fluoride Emission
Secondary (2°) Controls
None
"
Install spray screen and
waste water lime treatment
1° and 2°
None
Install spray screen and
waste water lime treatment
SWPB CELLS
Spray scrubber
Install lime treat-
ment of cryolite bleed
stream
None
None
Average
Fluoride
Evolution
Ib/ton Al
30.5
II
II
42.9
11
44.5
1
48
'
Average
Fl uori de
Efficiencies Emission
1° collection 1° removal 2" removal Ib/ton Al
87 99.2 0 4.2
II II II II
75 1.2
80 98.8 0 9
75 2.6
85 99 87 1.3
II II II
00 0 48.0
80 98.5 " 10.2
collection dry scrubber
system
Install spray screen &
waste water lime treatment
75
3.0
-------
Table 7-3. PRIMARY ALUMINUM CONTROL STRATEGIES (Continued)
HSS CELLS
Plant Emission Controls Required for the Specified Average Fluoride Emission
Code HoodingPrimary (1°) Control?Secondary (2°) Controls
31C Best Available
Spray tower + wet electro- None
static precipitator (ESP)
Average
Fluoride
Evolution
Ib/ton Al
32.1
Average
Fluoride
Efficiencies Emission
1° collection 1° removal 2° removaT Ib/ton AT
Install lime treatment
of cryolite bleed stream
Install spray screen and
waste water lime treatment
spray tower; install lime
treatment of cryolite bleed
stream
Install spray screen and
waste water lime treatment
95
99
75
75
1.9
0.7
26
Poor Spray tower
Install lime treatment of
cryolite bleed stream
Improve hooding ' "
' Install wet ESP; remove
None
I
1
II
41.6 27
II II
90
1
62
1
II
96
0
II
II
1
34.5
II
18.4
5.7
2.5
CWPB CELLS
8B Best Available Fluidized bed scrubber
i H ii ii H
None 45.5
Install spray screen '
and waste water lime
treatment
98 99.5 0
75
1.2
0.5
4A
Dry ESP + spray tower
Install waste water lime
treatment
None
65.6
94
86
12.5
Install fluidized bed dry "
scrubber; remove dry ESP
& spray tower; lime treatment
unnecessary
Install spray screen &
waste water lime treatment
-2°
98.5
75
4.;
1.9
-------
b. Second: install best available primary control (fluoride removal)
with water treatment, if needed; and,
c. Third: install spray screen or spray scrubber secondary control
with water treatment.
Plants without Initial primary control obviously require both steps
a and b. It is assumed that all plants will, as a minimum, maintain
their present control combinations. It is further assumed that all
companies will add new retrofits that are the best that their cost
class allows.
Average emission rates are calculated as:
/ \
EM = EV
1 -
100
100
fn
sr
/
100 100
(7.1)
where:
EM = average emission rate, Ib F/ton Al
EV = average evolution rate, Ib F/ton Al
npc = primary collection efficiency, percent
npr = primary removal efficiency, percent
"sr = secondary removal efficiency, percent
A removal efficiency of 75 percent is assumed for secondary control
retrofits to all cell types. This is based on performance with primary
control as reported in Section 6.2.3.
Facilities to meet'1983 effluent guidelines are included for plants
with wet primary or secondary control. The two water treatment systems
considered are:
a. Waste water lime treatment of a bleed stream off the scrubber
loop, with total recycle.
7-7
-------
b. Cryolite recovery with lime treatment of the cryolite bleed
stream.
System b. is considered only when the plant already has cryolite recovery.
If a plant has a suitable water treatment system for secondary control,
it is assumed that this system can additionally handle wet primary
control effluent; but a suitable water treatment system for primary
effluent is assumed to be undersized for handling secondary effluent and
a new secondary effluent treatment system would be required. If a plant
has cryolite recovery for primary effluent and adds secondary control, it
is assumed that the secondary effluent will be lime treated. Addition of
water treatment systems should be considered for plants not undergoing
air pollution retrofits, because the plants might choose to abandon
effluent-generating control systems in the absence of lll(d) regulations.
Costs for lime treatment of the cryolite bleed stream are taken from the
2
EPA effluent guidelines document.
In Table 7-3, it is believed that each VSS plant presently has the
highest primary collection efficiency achievable for that plant. Both
plants have best available primary control. Plant 5A has no water treatment.
Plant 3B has no need for water treatment with present controls.
In Table 7-3, it is assumed that no SWPB plant can achieve a primary
collection efficiency higher than 80 percent unless it is already doing
so. Plants achieving higher efficiencies are of French design, while
Swiss-design plants are not capable of higher efficiencies for reasons
•
detailed in Section 6.1.2. All SWPB primary retrofits would probably be
dry scrubbing systems; these already predominate SWPB plants with primary
controls. It is assumed that plant 24 would install injected
alumina since it is operated by a small company, and injected alumina has
7-8
-------
slightly lower capital and operating costs than the fluidized bed. A
primary removal efficiency of 98.5 percent is assumed for the dry scrubbing
retrofits, based on past performance at CWPB and SWPB plants. An efficiency
of 75 percent is assumed for secondary removal based on demonstrated
retrofit performance reported in Section 6.2.3. Plant 18 has cryolite
recovery that will require lime treatment of the bleed to meet 1983
effluent guidelines. Plant 24 has no need for water treatment with present
controls, or lack of controls.
It is assumed that all HSS plants except plant 26 have the highest
primary collection efficiency achievable within existing cell constraints.
Cell age and geometry affect the ability of an HSS plant to achieve high
collection efficiencies, as explained in Section 6.1.1.2. Geometry
restricts plant 26. Courtyard space limitations, or the necessity for
balanced ducting layouts in central intallations, necessitates removal of
the scrubbers that would otherwise precede the ESPs at plant 26. Gaseous
fluoride control is--or would be—achieved by a scrubbing section in the
ESP inlets. A primary removal efficiency of 96 percent is
assumed for a primary retrofit for plant 26, based on experience at
similar plants 25B, 31D, and 30B. Plants 31C, and 26 are believed to
•3
have cryolite recovery and to require lime treatment of the bleed to
meet 1983 effluent guidelines with present controls.
In Table 7-3, it is assumed that no CWPB plant can achieve a primary
collection efficiency higher than 95 percent unless it is already doing
so. All CWPB primary retrofits would probably be dry scrubbing
systems, following the general practice of the industry. However, it
is assumed that primary retrofit at 4A would be fluidized bed, the
system marketed by the company operating these plants.
7-9
-------
A primary removal efficiency of 98.5 percent is assumed for all dry
scrubbing retrofits, based on past performance at CWPB and SWPB plants.
Replacement of fluidized bed with fluidized bed should improve
primary removal efficiency because there have been different generations
of fluidized beds, with newer beds achieving 98.5 percent removal. For
primary retrofits, all former control equipment is considered to be
removed from service except multiple cyclones. Plant 4A has no water
treatment. Hence, this plant will have to install lime treatment with
recycle to meet 1983 effluent guidelines. All other CWPB plants except
one have no need for water treatment with present controls.
The "best hooding + best primary control" option requires only two
plants in Table 7.3 to retrofit primary controls, while the "secondary
control" option additionally requires all plants but plant 18 to retrofit
secondary controls. For this reason, intermediate levels are established;
levels which would require some plants to install secondary control.
All "existing + water treatment" control options afford no improve-
ment in emission control over levels expected without lll(d) emission
guidelines and thus are not considered in measuring economic impact.
The added water treatment is that which is adequate to meet 1983 effluent
guidelines; it is assumed that these will universally have to be met.
The above analysis options do not consider CWPB and SWPB anode bake
plant total fluoride emissions. Table 6-12 shows controlled bake plant
emissions to be only 0.05 Ib F/ton Al. Capital cost for such control is
estimated at $10.49/annual ton Al, and annual cost at $5.79/ton Al.
7-10
-------
Since these costs are small compared to potroom retrofit control costs
and no bake plants are known to have effective fluoride control, it is
assumed that these costs will be incurred at all prebake plants. Within
the accuracy of the emission data, a controlled bake plant emission of
0.05 Ib F/ton Al is so small that it was not considered.
7.3 CAPITAL AND ANNUAL COSTS FOR FLUORIDE EMISSION CONTROL OPTIONS
7.3.1 Procedure
Since primary aluminum plants consist of a grouping of modules
(potlines), the control devices are also modular. This reduces the
economies of scale advantage for the larger plants and also allows the
calculation of capital and annualized costs on a per ton of capacity
basis for use with any size of plant. However, the number of much
greater interest is the cost in dollars per ton of aluminum actually
produced. This can be calculated for each plant using either a historical
or a forecasted operating ratio and dividing it into the annualized cost
at capacity. The only exception to this modular approach is the waste
water lime treatment facility. This was estimated from an EPA design.
7.3.2 Capital Costs
The module approach for capital costs is presented in Table 7-4.
Modules are segregated by cell type. For instance, since all the
vertical stud Soderberg plants have acceptable primary controls, only a
spray screen secondary control to capture and remove emissions eluding
primary control systems is presented. The center-worked and side-worked
pre-baked plants can be modified by means of thirteen modules. One
module provides water treatment for the cryolite bleed stream. Two
modules can be used to improve the collection system. Two modules will
7-11
-------
Table 7-4
I
IN:
Control Modules for Upgrading Existing Aluminum Plants. Capital Costs
(September 1977 Dollars)
($/Ton Annual Capacity
Control Module
VSS
1. Install spray screen secondary
CWPB and SWPB
Basic ,
Capital Cost8
$ 51.94
Capital Cost
Adjustment
Factor5
1.0
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
HSS
15.
16.
17.
18.
19.
20.
Lime treatment of cryolite bleed stream
Improve hooding
Install primary collection system
Install Inj. Alum, dry scrubber-primary
Install spray screen secondary
Remove dry ESP primary
Remove floating bed wet scrubber-primary
Install Fluidized bed dry scrubber-primary
Remove coated bag filters-primary
Remove fluidized bed dry scrubber-primary
Remove multiple cyclone-secondary
Remove spray tower-primary
Install anode bake plant controls
Lime Treatment of cryolite bleed stream
Improve hooding
Install wet ESP primary
Remove spray tower-primary
Install spray screen-secondary
Remove floating bed scrubber-secondary
—
6.18
22.69
32.41
37.10
5.46
5.33
37.10
10.00
14.27
0.87 (35.67/5.33)
2.47
--
— g—
6.18
129.68
4.34
51.94
7.99 (35.67/5.33)
—
1.6
1.0
1.08e
1.0
0.75
0.75
1 .084e
0.75
0.75
0.75
0.75
—
1.6
1.0
0.75
1.0
0.75
Old Plant
Factor13
1.15
15
15
15
15
15
15
15
15
1.15
1.15
1.15
1.15
1.15
1.15
1.15
1.15
1977 Adj.
Factorc
1.63
1.63
1.63
1.63
1.63>
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
1.63
Adjusted
Capital
$ 97.36
2.43U
18.54
42.53
65.86
69.54
7.68
7.49
75.39
14.06
20.06
8.19
3.47
20.49
2.43°
18.54
243.09
6.10
97.36
75.17
aTaken from reference 4. Significant figures were retained for identification in original reference. Removal cost is taken as 75% of
direct installation change of original equipment.
Reference 5.
cAdjust costs from January 1971 to September 1977.
Taken from reference 6 and adjusted to September 1977.
eWeighted average gas flow adjustment factor from reference 7.
-------
provide improvement of the primary removal system. The installation of
a spray screen will improve the removal of secondary emissions. The
final module represents control of the anode bake plant. Horizontal
stud Soderberg (HSS) modules are similar to the pre-bake modules.
The numbers in the second column are the basic costs taken from
. ^ference 4. In the two cases where a ratio is used to multiply a base
cost, the cost is listed as a primary module, but not as a secondary
module in reference 4. However, the floating bed wet scrubber is listed
both as a primary and as a secondary module. The cost ratio between the
two is $35.67 for the secondary module divided by 5.33 for the primary
module. This ratio of 35.67/5.33 is used to convert "Remove multiple
cyclone" and "Remove floating bed wet scrubber" from primary to secondary
modules.
The next column is the capital cost adjustment factor. The factor
1.6 for the hooding modules is derived from reference 5 which assumes
that, in most cases, modifications will be made and new ducting added to
the primary collection systems amounting to an arbitrary 160 percent of
the estimated cost for main ducting in courtyard systems. It is further
assumed that, when elements of existing control systems are changed from
one type to another, the original element will be either bypassed without
cost, or will be removed to provide physical space for the new element.
In the latter case, the net cost of demolition, including salvage credit,
is estimated to be 75 percent of the direct installation cost of equipment
removed. The factor 1.08 represents the weighted average of the flow
adjustment between the model used in reference 4 and a sampling of
conditions existing in actual plants. The column labeled "1977 Adj.
Factor" is the inflation adjustment using the Chemical Engineering Plant
7-13
-------
Cost Index to convert reference 4 costs to September 1977. The final
column represents the conversion of the basic costs to current costs
with all the adjustments included.
Reference 6, the source of costs for the water pollution treatment
of the cryolite bleed stream, did not contain costs for a wastewater
lime treatment unit. Therefore, cost estimates were prepared for a
treatment unit designed by the Emission Standards and Engineering Division
of EPA. Since this treatment unit is not susceptible to the modular
treatment, units were estimated for three sizes of primary aluminum
plants, 78,000, 156,000, and 312,000 tons of aluminum per year. These
costs are shown in Table 7-5.
7.3.3 Annualized Cost
Annualized costs are based on data in reference 4 just as were the
capital costs. However, the updating adjustments are considerably
different from those for capital costs. Operating labor is adjusted
using the Department of Commerce Index of Hourly Earnings - Manufacturing.
Electric power is adjusted using the Wholesale Price Index for Industrial
Power. Circulating water, since most of its cost is in electricity for
pumping plus a small amount for treating chemicals, is adjusted by a
factor 10 percent higher than the electric power adjustment to account
for this extra expense. Lime costs are adjusted using the Chemical
Market Reporter quotations. Product recovery credits are based on data
found in reference 4 updated from 1972. For aluminum returned to the
8 Q
cells, the ratio of current prices to 1971 prices for aluminum ingot
was determined to be 1.8. This was applied to the unit price for the
credit. The same publications were used to obtain the ratio of fluorspar
7-14
-------
prices. This ratio was 2.0. Royalties are often tied to Wholesale
Price Index (Industrial). The few cases where royalties are involved,
this index was used.
In the cases where the modules specified the removal of control
equipment, the annualized cost represents the savings resulting from not
operating the equipment or the loss incurred by not realizing the recovery
credits gained by operating the equipment.
Table 7-5. WASTE WATER LIME TREATMENT INVESTMENT3 COST BY SIZE OF PLANT
(September 1977 Dollars)
78.000 TPY 156.000 TPY 312.000 TPY
Installed Major Equipment $828,000 $1,101,000 $1,477,000
Contingencies and Fee (? 20% 166.000 220.000 295.000
TOTAL $994,000 $1,321,000 $1,772,000
Unit Cost, $/ton capacity $12.74 $8.47 $5.67
a
Process and instrumentation design by the Emission Standards and
Engineering Division of EPA. Cost estimates prepared by vendor
contacts by the Economic Analysis Branch, SASD, EPA.
7-15
-------
In order to bring the treatment of fixed cost into line with EPA's
current practice of combining interest and depreciation into a capital
recovery factor, the fixed cost components of the annualized costs were
changed from those found in the contract report. These changes are
detailed in Table 7-6.
The annualized costs, recalculated as outlined above are found in
Table 7-7. The annualized costs for the waste water lime treatment
appear in the next table - Table 7-8.
Since the waste water lime treatment unit is not calculated in the
module basis, the investment and annualized costs for the three plant
sizes are plotted in Figure 7-1 so that interpolation can be made for
individual plants.
7.3.4 Cost-Effectiveness
Cost-effectiveness is defined as the annualized cost of operating a
given control system divided by the number of pounds of pollutants captured
by the system per year. When several systems are installed in succession,
the total overall cost divided by the overall weight of pollutant captured
is the "cumulative cost-effectiveness", shown in the next-to-last column
of Table 7-9. Sometimes it is of interest to examine the stepwise effect
of the addition of each of the several systems. This is referred to as
the "incremental cost-effectiveness" shown in the last column in Table 7-9.
This is obtained by dividing the annualized cost of each individual system
by the additional pollutants captured by it. From this table, it appears
that the installation of spray screen secondary control with its attendant
waste water lime treatment is much less cost-effective than primary controls.
In fact, it adds from 1 to 2
-------
Table 7-6. FIXED COST COMPONENTS
Percent of Investment
Component Reference TO EPA
Taxes and Insurance 2% 2%
Administration 5% 5%
Depreciation 8%
Interest 8%
Capital Recovery — 13%
(15 yrs @ 10*)
TOTAL 23% 20%
7-17
-------
Table 7-7. CONTROL MODULES FOR UPGRADING EXISTING
ALUMINUM PLANTS ANNUALIZED COSTa
($/ton Aluminum Produced, in 1977 $)
Net Annualized Cost
Control Module (Credit)
VSS
1. Install spray screen-secondary $31.71
CWPB and SWPB
2. Lime treatment of cryolite bleed stream 1.04
3. Improve hooding 4.64
4. Install primary collection system 11.56
5. Install injected alumina dry scrubber-primary (1.76)
6. Install spray screen-secondary 22.84
7; Removal dry ESP-primary 1.69
8. Remove floating bed wet scrubber-primary (7.94)
9. Install fluidized bed dry scrubber-primary 3.16
10. Remove coated bag filters - primary 1.14
11. Remove fluidized bed dry scrubber - primary (3.16)
12. Remove multiple cyclone - secondary (5.78)
13. Remove spray tower-primary (4.12)
14. Install anode bake plant 5.79
HSS
15. Lime Treatment of cryolite bleed stream 1.04
16. Improve hooding 4^54
17. Install wet ESP - primary 61.93
18. Remove spray tower - primary (6.48)
19. Install spray screen - secondary 31.70
20. Remove floating bed scrubber - secondary (77.79)
aSource: Reference 4 data updated as described in text.
7-18
-------
Table 7-8.
WASTE WATER TREATMENT PLANT OPERATING COST
(By Size of Aluminum Plant)
(September 1977 Dollars)
350 Days Operated/Year
Annualized Cost
Operating Costs
A. Direct
1. Supplies
CaO 135 Ib/ton Al. P
FeCl3 0.135 Ib/ton Al.
Separan (AP-30)
0.00675 Ib/ton
2. Operating Labor
3. Supervision, 15% of 2
4. Utilities
a. Electricity
b. Process Water
5. Maintenance
6. Laboratory 30% of 2
7. Total Directs
B. Indirects
8. Taxes and Insurance
(2% of capital)
9. Administration
(4% of capital)
10. Capital Recovery
(13% of capital)
11. Total Indirects
Total Annualized Cost
Unit Cost $/ton Al.
Unit Cost
78,000 TRY
156,000 TPY
312.000 TPY
l%t/lb.
4$/lb.
300
-------
Figure 7-1. Investment and Annualized Costs for Waste
Water Lime Treatment Plants vs. Aluminum
Plant Capacity
It)'
Plant Capacity, Tons Al./Yr.
7-20
-------
Table 7-9. PRIMARY ALUMINUM CONTROL STRATEGIES
VSS CELLS.
F\3
Plant
Code
5A
3B
18
24
Emission Controls Required for the Specified
Hooding Primary (1°) Controls
Best Available Spray tower + wet ESP
Install waste water
lime treatment
" No change (water
treatment handled
by 2»)
Fluidized bed dry scrubber
" " Injected alumina dry
scrubber
None None
Install primary Install injected alumina
Average Fluoride Emission
Secondary (2°) Controls
None
Install spray screen and
waste water lime treatment
1° and 2°
None
Install spray screen and
waste water lime treatment
SWPB CELLS
Spray scrubber
Install lime treatment
of cryolite bleed stream
None
None
Average
Fluoride
Evolution
Ib/ton Al
30.5
1
II
42.9
I
44.5
48
Unit Cost
Capital
$/ton Al
0
14.70
84.24
0
77.19
20.49*
22.92
20.49*
128.88
Annual 1 zed
$/ton Al
0
6.50
29.34
0
26.89
5.79*
6.83
5.79*
15.59
Average
Fluoride
Emissions
Ib F/ton Al
4.2
4.2
1.2
9.0
2.6
1.3
1.3
48.0
10.2
Cost-Effectiveness
Cumulative Incremental
$/lb F $/lb F
9.78 7.61
4.20 4.20
0.41
collection dry scrubber
system
Install spray screen &
waste water lime treatment
207.42
42.88
3.0
0.95
3.79
Anode bake controls required on all prebake plants
-------
Table 7-9. PRIMARY ALUMINUM CONTROL STRATEGIES (Continued)
ro
ro
Plant
Code
31C
26
8B
4A
HSS CELLS
Average
Fluoride
Emission Controls Required for the Specified Average Fluoride Emission Evolution
Hooding Primary (1°) Controls Secondary (2°) Controls Ib/ton Al
Best Available Spray tower + wet electro- None 32.1
static precipitator (ESP)
Install lime treatment
of cryolite bleed stream
' Install spray screen and
waste water lime treatment
Poor Spray tower None 41 . 6
Install lime treatment of '
cryolite bleed stream
Improve Hooding ' '
Install wet ESP; remove '
spray tower; install lime
treatment of cryolite bleed
stream
Install spray screen and
waste water lime treatment
CWPB CELLS
Best Available Fluidized bed scrubber None 45.5
Install spray screen
and waste water lime
treatment
Dry ESP + spray tower None 65.6
" Install waste water lime '
Unit Cost
Capital Annuall zed
$/ton AT $/ton AT
0 0
2.43 1.04
108.60 37.14
0 0
2.43 1.04
20.97 5.68
272.59 62.17
386.01 101.02
20.49* 5.79*
102.43 34.21
20.49 5.79*
29.39 10.19
Average
Fluoride
Emissions
1b F/ton Al
1.9
1.9
0.7
34.5
34.5
18.4
5.7
2.5
1.2
0.5
12.5
12.5
Cost-Effectiveness
Cumulative Incremental
$/1b F $/lb F
0 0
0 0
30.95 30.95
0 0
0 0
0.35 0.35
2.16 4.45
3.16 12.14
0 0
48.87 40.60
0 0
0 0
treatment
Install fluidized bed dry ' 107.03 6.52 4.7 0.84
scrubber; remove dry ESP
and scrubber; lime treatment
unnecessary
" Install spray screen and 185.74 33.76 1.9 3.18 9.72
waste water lime treatment
- 2°
Anode bake controls required on all prebake plants.
-------
7.4 REFERENCES FOR SECTION 7
1. Memorandum from B. A. Varner to 6. B. Crane. Estimation of
Secondary Removal Efficiencies for Two Side-Worked Prebake Plants
Adding Primary Control of Existing Secondary Control. Emission
Standards and Engineering Division, OAQPS, Environmental Protection
Agency. June 2, 1975.
2. Development Document for Effluent Limitations Guidelines and
New Source Performance Standards for the Primary Aluminum Smelting
Subcategory of the Nonferrous Metals Manufacturing Point Source
Category. Effluent Guidelines Division, Office of Air and Water
Programs, Environmental Protection Agency, Washington, D. C.
Report Number EPA-440/l-74-019-d. March 1974. p. 110.
3. Air Pollution Control in the Primary Aluminum Industry. Singmaster
and Breyer, New York, N.Y. Prepared for Office of Air Programs,
Environmental Protection Agency, Research Triangle Park, N.C. under
Contract Number CPA 70-21, July 23, 1973. p. 2-31.
4. Ibid. pp. 8-22, 8-27 thru 8-31.
5. Ibid, page 9-12.
6. Reference 2
7. Singmaster and Breyer op. cit. p. 9-10.
8. Chemical Marketing Reporter, January 2, 1978.
9. Minerals Yearbook, Volume I, 1973 United States Department of the
Interior, Bureau of Mines, pp. 137, 155.
10. Singmaster and Breyer op. cit. p. 8-25.
7-23
-------
8. RATIONALE OF STATE EMISSION GUIDELINES
FOR EXISTING PRIMARY ALUMINUM PLANTS
8.1 INTRODUCTION
The recommended State fluoride emission guidelines in Section 8.3
are not expressed in terms of emission limitations, but are presented as
recommended control technologies that will achieve certain average
fluoride control efficiencies when applied as new retrofits to existing
plants. The relative performances of the recommended controls are
calculated from known cell fluoride evolution rates.
The data base underlying the State guidelines has been derived from
State and industry test methods that often differ from EPA methods of
emission measurement. Therefore, significant differences among the
accuracy of these methods is possible. Because of the varying design of
existing roof monitors, the use of source test method 14 may be precluded.
The different roof monitor configurations may also prohibit the deter-
mination of the relationship between the emission test method used and
Method 14.
State regulations are now in force that limit fluoride emissions
from existing primary aluminum plants. The terms of the regulations and
their compliance test requirements tend to differ among States and from
-------
the Federal standards of performance for new primary aluminum plants.
The guidelines are therefore structured to give the State maximum flexibility
to utilize existing emission control and source sampling and analytical
methods, and to avoid the requirement for unnecessary modifications of
roof monitor sampling systems.
Good operation and good maintenance of potrooms is essential to
good control, and the States should take steps to insure such objectives
in their plans for implementing the guidelines.
. All hood covers should be in good repair and properly positioned
over the pots. The amount of time hood covers are removed during
pot working operations should be minimized.
. Some hooding systems are equipped with a dual low and high hood
exhaust rate. This should be conscientiously used whenever hood
covers are removed and returned to the normal exhaust rate once
the hood covers are replaced.
. A fuming pot often indicates a sick cell or clogged hooding ductwork.
Either case represents poor potroom operation and should not be
allowed to continue.
. Some tapping crucibles are equipped with hoses which return
aspirator air under the hood. The hoses should be in good repair
and the air return system should function properly.
. Dust entrainment should be minimized during the sweeping of work
aisles. Some plants utilize vacuum sweepers which collect floor
sweepings in fabric bags.
-------
8.2 Fluoride Emission Control Equipment and Costs (September 1977)
Table 8-1 gives some typical costs for certain model operations
pertaining to fluoride control at existing aluminum plants. Conservative
values for capture and removal efficiencies are also included as
percents. The given cost values are taken from Tables 7-4 and 7-7
and represent only three of the several plant construction operations
that may take place in any real situation. To illustrate the use of such
modules refer to plant 26 in Table 7-9.
Table 8-2. The Use of Capital Cost Modules
Fluoride Emission Capital Cost
Control Module $/Annual ton Al
Install lime treatment of
cryolite bleed stream $2.43
Improve hooding 18.54
Install wet ESP 243.09
Remove spray tower 6.10
Install lime treatment for additional
cryolite bleed stream 2.43
Install spray screen 97.36
Install waste water lime treatment 15.90
$385.85
As shown, the final cost involved in any selected degree of control simply
involves a determination of the construction scope of work (new equipment,
deletions, etc.) followed by addition of the respective module costs. The
seven cost modules shown for plant 26 add up to a total capital cost of
$386 per annual ton of aluminum produced. The annualized costs could be
derived in an analogous manner.
8-3
-------
Table 8-1. CONTROL EQUIPMENT AND COSTS (September 1977)
CO
Fluoride Emission
Control Module
Improve Hooding
CWPB
SWPB
HSS
Install Primary Removal
CWPB
SWPB
HSS
Install Secondary Removal
VSS
CWPB
SWPB
HSS
Capital
$/annual ton
18.54
18.54
18.54
75.39
75.39
243.09
97.36
69.54
69.54
97.36
Cost
Annual i zed
$/ton
4.64
4.64
4.64
3.16
3.16
61.93
31.71
22.84
22.84
31.71
Efficiency,
Percent
95
80
90
98.5
98.5
98.5
75
75
75
75
-------
The cost modules are estimates and may be used when actual
engineering cost estimates or retrofit costs are not available. They
also allow cost comparisons among degrees of control at the same plant
or of costs among different plants. Some actual retrofit costs are
given in Table 6-33.
8-5
-------
8.3 RECOMMENDED STATE GUIDELINES AND COLLECTION AND REMOVAL
EFFICIENCIES OF CONTROL EQUIPMENT FOR FLUORIDE EMISSIONS
The recommended State guidelines haye been developed as described
in Section 8.3.1 and are summarized in Table 8-3. The table may be
explained in the following manner: Column 1 shows each of the four cell
types. Column 2 gives the recommended average primary collection efficiencies
(hooding) and Column 3 the primary removal efficiencies that EPA believes
are readily achievable with new retrofits. An achievable secondary
removal efficiency is also presented in Column 4. Column 5 shows the
recommended technology for control of total fluoride emissions. Included
is an indication of the status of primary control on a national basis
and the conditions under which better primary control should be installed
or secondary control added.
The fluoride emission ranges corresponding to the State guidelines
are presented in Table 8.4. The recommended minimum fluoride collection
and removal efficiencies have been used in equation 7-1 to estimate
average fluoride emissions after the various controls are applied. Two
cases are worked out for each of the four cell types: these cases
correspond to the smallest and greatest evolution rates shown for each
cell type in Table 7.2, and these extreme evolution rates are displayed
in Column 6 of Table 8.4. The calculated cell average emissions
corresponding to the evolution rates are arranged in the last column to
show the range of emissions caused by the variations in cell fluoride
evolution at the various plants from which EPA received cell evolution data.
The guideline primary collection efficiencies of Column 2,
8-6
-------
Table 8-3. STATE GUIDELINES FOR CONTROL OF FLUORIDE EMISSIONS FROM EXISTING PRIMARY ALUMINUM PLANTS
Cell Type
VSS
SWPB
HSS
CWPB
Recommended Efficiencies
for Proposed Retrofits Guideline Recommendations
Primary Collection
80
80
90 (a)
95 (a)
Primary Removal
98.5
98.5
t
98.5
98.5
Secondary Removal
75 (a)
75(a)
All plants now have best achievable hooding
and primary removal .
Install secondary control, but only if
justified depending on severity of fluoride
problem.
Install best achievable hooding and primary
removal equipment.
Install secondary control wherever
justified, depending on the severity of
the fluoride problem.
All plants but #26 now have the best
achievable primary collection efficiency.
Plant #26 should install best primary
control if needed.
Secondary control does not appear to be
justified, in most locations.
Best control is best hooding and primary
removal equipment. Install where needed.
Secondary control does not appear to be
justified, in most locations.
•
oo
-•j
(a) See Section 8.3.1
-------
Table 8-4. FLUORIDE EMISSION RANGES CORRESPONDING TO STATE GUIDELINES FOR EXISTING PRIMARY ALUMINUM PLANTS
Cell
Type
VSS
SWPB
HSS
CWPB
Recommended Efficiencies
for Proposed Retrofits
Primary Primary Secondary
ol lection Removal Removal
80 98.5 75
V
/
80 98.5 ' 75
90 98.5
(a)
95 98.5
(a)
Guideline
Recommendations
All plants now have best achievable
hooding and primary removal.
Install secondary control, bat only
if justified, depending on severity
of fluoride problem.
Install best available hooding
and primary revmoval equipment.
Install secondary control wherever
justified, depending on the severity
of the fluoride problem.
InstaVI best primary control if
needed. All plants but #26 now
have best achievable primary collectic
efficiency.
Secondary control does not appear to t
justified, in most locations.
Best control is best hooding and
primary removal equipment.
Install where needed.
Secondary control does not appear to
be justified, in most locations.
Assumed Average
Fluoride Cell
Evolution-lb/ton Al
30 - 54
37 - 53
n
28 - 45
e
26 - 66
Average Fluoride Emission
range, Ib F/ton Al
Control
Primary 6.4* - 11 .4
Secondary 1.9** - 3.4
Primary 7.8 - 11.2
Secondary 2.3 - 3.3
Primary 3.2 - 5.1
Secondary 1.1 - 1.7
Primary 1.7 - 4.2
Secondary — —
00
CO
* 30 [ 1 - .80 x .985] = 6.4 Ibs F/ton Al (a) See Section 8.3.1
** 30 [ 1 - .80 x .985 - (1 - .80) x .75] = 1.9 Ibs F/ton Al
-------
Table 8.3 are based on methods of calculation given in detail in Section
6.1.2. These calculated efficiencies agree well with those that were
directly measured or otherwise arrived at by the owners of similar
plants. Hooding efficiency depends on several factors such as cell
design, age, operation and maintenance, and hood exhaust rate; and Table
7.3 ;hows cases where claimed efficiencies vary from those of the guide-
lines. There are various reasons for the cell hooding efficiencies for
given cell types. The VSS cell has exposed molten electrolyte bath area
around the hood skirt that varies with cell design. This factor limits
cell collection efficiency and also causes fluoride escape according to
the amount of cell bath area exposed. Hood efficiency also depends on
the number of times a cell hood has to be opened to produce a ton of
aluminum; fluoride escapes during openings. The HSS cell varies from 16
to 50 openings per ton of aluminum produced, depending on cell design.
The plants considered in Table 8-4 are hypothetical plants, and
therefore, costs have not been derived for these specific cases. Instead,
the cost of Table 7-9 for analogous cases will be discussed in the
guidelines. The costs derived in Table 7-9 are applied to actual plants,
but are model plant costs. Considering their uncertain accuracy applied
to real situations, they will be sufficient to illustrate the cost
effectiveness of State guidelines.
8.3.1 State Fluoride Emission Guidelines
The following State emission guidelines for control of total fluoride
emissions from existing primary aluminum plants are restated from Table
8.3. The range of average fluoride emissions, according to the last
column of Table 8-4, is given after the guideline for each cell type.
8-9
-------
As explained above, the range of the average emissions reflects the
range of known cell evolutions.
VSS CELLS
The primary collection efficiencies for all existing plants are
only about 80 percent, but are essentially the best achievable for this
type of cell. The primary removal efficiencies for all existing plant
VSS cells are high, in the range of 98.4 to 99.9 percent. Therefore,
secondary controls should be installed only if justified by the severity
of the fluoride problem. The expected emission ranges are as
follows:
Average emissions from primary control (calculated):
= 6.4 to 11.4 Ib F/ton Al
Average emissions from primary plus secondary control (calculated):
= 1.9 to 3.4 Ib F/ton Al
Incremental cost effectiveness for secondary control is in the
approximate range of $4 to $8 per pound of fluoride removed, as indicated
in Table 7-9. Essentially, no expenditures are required for primary
control. The cases in Table 7-9 were all chosen to represent—for each
cell type—the least and the greatest emission rates after installation
of best primary and secondary control.
SWPB CELLS
SWPB cells must be worked along both sides with the side covers
removed, and for longer times than other cells (Tables 6.2-6.6). Therefore,
there is an inherent limitation to the primary collection efficiency of
these cells.
The best achievable hooding for this type of cell has about 80%
8-10
-------
collection efficiency. In addition to installing the best available
primary hooding and removal equipment, secondary controls should be
installed, if justified, depending on the severity of the fluoride
problems. The emission ranges are as follows:
Average emissions from primary control (Calculated):
= 7.8 to 11.2 Ib F/ton Al
Average emissions from primary plus secondary control (calculated):
= 2.3 to 3.3 Ib F/ton Al
The cost effectiveness of secondary control is $3.79 per pound of
fluoride removed for plant #24. It will be somewhat higher for other
SWPB plants because they are initially better controlled.
HSS CELLS
All plants except plant #26 presently have essentially the best
achievable primary collection efficiencies of about 90 percent. If the
primary removal systems for HSS cells are upgraded to 98.5 percent, this
should be suitable for best retrofit control technology. The emission
ranges are as follows:
If hooding were added to, or replaced on, an existing HSS plant,
EPA believes that modern technology can achieve 90% collection efficiency,
in almost all cases. A plant may exist where it is difficult or
economically impractical to install best hooding.
If a modern primary removal system were added to, or replaced on,
an existing HSS plant, EPA believes that modern technology can achieve
98.5% removal efficiency. A modern spray tower added ahead of existing
wet ESPs can raise primary removal efficiency to 98.5. This spray tower
addition may be economically impractical if free space does not exist
within a reasonable distance of fluoride source and wet ESP.
8-11
-------
If an HSS plant has existing primary collection efficiency of
85-90% and primary removal efficiencies of 95 - 98.5%, control agencies
should closely study costs and benefits, before requiring retrofits.
In the above efficiency ranges, retrofit does not seem justified unless
there is a local fluoride problem.
Average emissions from primary control (calculated):
= 3.2 to 5.1 Ib F/ton Al
The lowest incremental cost effectiveness to improve the hooding is
about $0.35 per pound of fluoride, and to add best primary removal is
about $4 per pound of fluoride removed.
Secondary control does not seem justified at an incremental cost
effectiveness ranging from $12 to $30 per pound of fluoride removed,
depending on the plant. No HSS plants now have secondary control.
CWPB CELLS
Retrofit primary hooding can be added to achieve a collection
efficiency of 95 percent for CWPB cells, while primary fluoride removal
systems of 98.5 percent are common. A primary collection and removal
system should therefore suffice for best retrofit control technology.
If hooding were added to, or replaced on, an existing CWPB plant,
EPA believes that modern technology can achieve 95% collection efficiency!
in almost all cases. A plant may exist where it is difficult or
economically impractical to install best hooding.
If a modern primary removal system were added to, or replaced on,
an existing CWPB plant, EPA believes that modern technology can achieve
98.5% removal efficiency.
If a CWPB plant has existing primary collection efficiency of
90-95% and primary removal efficiencies of 95 - 98.5%, control agencies
8-12
-------
should closely study costs and benefits, before requiring retrofits.
In the above efficiency ranges, retrofit does not seem justified unless
there is a local fluoride problem.
The emission ranges are as follows:
Average emissions from primary control (calculated)
= 1.7 - 4.2 Ibs F/ton Al
Secondary control does not seem justified at an incremental cost effect-
iveness of $10 to $40 per pound of fluoride removed, depending on the
plant. No CWPB plants now have secondary control.
Secondary Removal
Some secondary removal units (scrubbers) may not be able to achieve
75 percent efficiency (See Section 6.2.3). Control officials should care-
fully study costs, impacts, and energy consumption before requiring either
replacements or initial retrofits.
8.3.2 Compliance Time
Section 6.4 has discussed at some length the design, construction,
and startup time requirements for retrofit air emission controls on
existing primary aluminum plants. The historical variation of delivery
times for supplies and equipment has been shown, and it was pointed out
that these deliveries are often completely outside the control of either
customer or control official.
Because of the nature of plant construction and emission source
testing, compliance times for either activity tend toward a case-by-case
basis for the primary aluminum industry. However, most air pollution
retrofits can be designed and installed in not more than 2-1/2 years.
States should add to this time any compliance demonstration time in excess
of that indicated in Table 6-37. In all cases, States should require
proof for the time requirements claimed for each milestone.
8-13
-------
8.4 EMISSION TESTING
EPA Reference Methods 13(a), 13(b), and 14 were developed for use
in the determination of total fluoride emissions from primary aluminum
plants, and are adapted for use with new sources. However, it is
recognized that installation of Method 14 ductwork on existing sources
will result in variable costs, depending on the plant. In some existing
plants, unreasonable costs may be incurred. For these reasons, EPA
does not specify compliance testing for existing plants: such testing
is to be decided by each State on a case-by-case basis, taking into
account economic feasibility.
8-14
-------
9. ENVIRONMENTAL ASSESSMENT
An environmental assessment for emission guidelines for existing
plants is unique in that the exact number of affected facilities is or
can be known. Further, for the primary aluminum industry, individual
capacities, existing or proposed control schemes, and fluoride emissions,
are known for each plant. From this degree of knowledge and specificity,
the national environmental impacts of alternative emission control
systems or levels were evaluated by simply summing individual plant
impacts for the entire population of United States primary aluminum
plants.
A vast number of fluoride emissions control scheme permutations
exist within the primary aluminum industry. There are four basic
aluminum reduction cell types, two fundamental levels of control--
primary and secondary—and many possible primary control schemes. For
this reason, an environmental impact assessment of every control
scheme permutation was not attempted. Instead, national environmental
imapcts—the sums of 31 individual plant impacts—of a few fundamental
levels of fluoride control were analyzed to illustrate the methods
that are applicable to all such analyses. The levels of fluoride
emissions control considered were:
1. Initial - The level of fluoride emissions control which would
9-1
-------
be expected in the absence of lll(d) State guidelines emissions (Table 7-2)
limitations for existing plants. At this level, all plants are expected
to install water treatment systems which comply with 1983 effluent
guidelines for all existing or proposed fluoride emissions control
schemes.
2. Best Hooding with Best Primary Control - Each existing plant
is upgraded to the level of best cell hooding and primary control for
that particular plant.
3. Best Cell Hooding with Best Primary and Secondary Control -
Each existing plant is further upgraded to the level of best cell
hooding and best primary and secondary control for that particular
plant.
National air pollution, water pollution, energy, and solid waste
disposal impacts were estimated for each of these alternative levels
of fluoride emissions control. Future experience will probably show
that the national degree of control achieved by the States will lie
somewhere within the boundaries of levels 2 and 3, above. However,
step 2 is not meant to be the lowest national level of emission control
improvement, even though step 3 does represent the highest possible
improvement. Levels 2 and 3 may be thought of as examples of two
national levels of fluoride emission control.
Although it was not possible to illustrate individual impacts for
all fluoride control scheme permutations, examples of control schemes
with extreme impacts have been provided. For example, in the water
pollution impact assessment, examples are given of plants with fluoride
9-2
-------
emissions control schemes which both maximize and minimize water pollution.
Thus, for each of the four major impact areas (air, water, energy, and
solid waste), specific plant examples showing maximum and minimum impacts
have been provided, along with national impact assessments for each of
the three fundamental levels of fluoride emission control.
9.1 AIR POLLUTION ASSESSMENT
The air pollution assessment for emission guidelines progresses
from a generalized evaluation of fluoride emissions impact to more
specific impacts. The first three sections of the air pollution assessment
discuss national fluoride emission impacts, extremes in source strengths,
and fluoride dispersion calculations. The last two sections of the
assessment discuss particulate emissions from aluminum reduction cells
and anode bake plant emissions.
9.1.1 National Fluoride Emissions from Primary Aluminum Reduction Cells
National fluoride emissions from primary aluminum reduction cells
have been calculated for all plants by methods indicated for plant 5A,
etc., in Table 7-3. For each alternative level of fluoride emissions
control, the products of individual plant capacities and average emissions
were added to yield national fluoride emissions for each of the four
reduction cell types. The results of these calculations, along with
average fluoride emissions by cell type and weighted by capacity, are
presented in Tables 9-1 and 9-2.
The dramatic effects of emission control are illustrated by the
mass fluoride emissions figures in Table 9-1. Implementation of the
lowest illustrated degree of control would result in a 50 percent
reduction in mass emissions of fluorides from primary aluminum reduction
9-3
-------
Table 9-1. NATIONAL TOTAL FLUORIDE EMISSIONS FROM PRIMARY
ALUMINUM REDUCTION CELLS
National Fluoride Emissions by
control Level
Initial
Best Hooding and
Primary Control
Best Primary and
Secondary Control
vssa
1,700
1,700
660
tei i
HSSa
3,000
2,100
890
lype uons
CWPBa
8,500
3,700
1,600
r/rr;
SWPBa
4,800
1,500
810
Total
18,000
9,000
4,000
VSS - vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake.
9-4
-------
TABLE 9-2. AVERAGE FLUORIDE EMISSIONS FOR PRIMARY
ALUMINUM REDUCTION CELLS
Average Fluoride Emissions
Control Level
Initial
Best Hooding
and Primary Control
Best Primary and
Secondary Control
vssa
5.2
5.2
2.1
(Lb F/T<
HSSa
5.7
4.0
1.7
)n Al)
CWPB3
6.3
2.7
1.2
SWPB3
13
4.0
2.2
VSS - vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake
9-5
-------
cells. National fluoride emissions from this source would thus be
reduced from 18,000 tons/yr to 9,000 tons/yr. Although the average
fluoride emissions for all cell types under this degree of control are
similar (Table 9-2), the fundamental level of control varies somewhat
among plants. Secondary control exists at plants 18, 19, and 20.
9.1.2 Fluoride Emission Control Systems with Extreme Air Pollution
Impacts
To illustrate the extremes in effectiveness of fluoride emissions
control systems, the overall environmental impact for two individual
plants is presented in Table 9-3. Plant 6B represents the most effective
fluoride emissions control system applied to the most controllable cell
type--center-worked prebake. One of the least effective emission
control systems is typified by Plant 26. Table 9-3 compares the overall
environmental impacts of the two extremes in fluoride emissions control
effectiveness.
The range of control scheme effectiveness is best indicated by the
difference in average fluoride air emissions; 1 Ib F/ton Al for Plant
6B, and 34.5 Ib F/ton Al for Plant 26. Although the more effective
emission control scheme reduces fluoride emissions by more than a
factor of thirty over the poor control scheme, only twice as much
energy is required per ton of aluminum produced. More solid waste is
generated by the poor control scheme compared to the effective control
scheme, mainly because of the higher percentage of mass removed in a
wet primary versus a wet secondary control system. Effluent emissions
per ton of aluminum produced must be the same for both examples, due to
the applicability of 1983 Effluent Guidelines Standards.
9-6
-------
Table 9-3. FLUORIDE EMISSION CONTROL SYSTEMS WITH EXTREME AIR
POLLUTION IMPACTS
Plant Code
Cell Type
Capacity (Tons Al/Yr)
Primary Control Scheme
Secondary Control Scheme
Air Fluoride Emissions
(Tons F/yr)
Average Fluoride Air
Emissions
(Lb F/ton Al )
Effluent Emissions
1. Fluoride: (Tons F/yr)
2. Total Suspended Solids
(Tons/Yr)
Fluoride Control
Energy Requirements
(Mwh/yr)
Average Fluoride Control
Energy Requirements
(Kwh/ton Al)
Solid Waste Generated from
Control of Fluoride
Emissions (Tons/yr)
Average Solid Waste Gen-
eration from Control of
Fluoride Emissions
(Lb/ton Al)
6B 26
CWPB HSS
115,000 51,000
Fluidized bed Spray tower
dry scrubber
Spray screen None
scrubber
57.5 880
1 34.5
5.75 2.55
11.5 5.1
58,800 13,200
511 259
2300 3,930
40 154
9-7
-------
9.1.3 Fluoride Dispersion
Dispersion estimates were prepared comparing ground level con-
centrations before and after the retrofit of emission controls described
under cases A, B, and C in Section 6.3. The purpose of those estimates
is to demonstrate the improvement in air quality that may result from
the retrofit. The estimates pertain to specific primary aluminum
plants located in the northwestern United States.
Receptors
As stated in Section 2.3, the most sensitive receptors are dairy
cattle grazing on forage that has a fluoride accumulation of more than
40 ppm. Such an accumulation can be caused by a 30-day average ambient
air concentration of gaseous fluoride of about 0.5 micrograms per cubic
3
meter (yg/m ). Hence, dispersion calculations should be concerned with
the 30-day average concentration out to distances where the concen-
tration is normally diluted to 0.5 yg/m3.
Source Characteristics
Emissions (Table 9-4) and ambient air concentrations (Table 9-5)
are expressed in terms of total fluoride, because emission breakdowns
into gaseous and fine particulate forms were not available. Both forms
are harmful and EPA New Source Performance Standards are in terms of
total fluoride.
9-8
-------
Table 9-4. FLUORIDE EMISSIONS AT PLANTS A, B, AND C
Plant Source
A Potrooms
" Scrubbers
11 Fume Control Units
500-foot Stack
B Potrooms
C Potrooms
" Fume Control Units
Fluoride Emissions (g/s)
Before Retrofit After Retrofit
2.08 1.84
4.3
0.92
4.3
16.3 8.6
(including ESPs)
34.3 3.4
1.5
Table 9-4 presents the before and after retrofit fluoride emissions
at the three plants studied. The fluoride emission rates are annual
averages, and are treated as constant on a year-round basis. The
following paragraphs describe the basic emission characteristics at
each facility.
At Plant A, potroom, scrubber, and fume control unit emissions are
from stacks and rooftop monitors about 55 feet above grade; effluent
temperatures are slightly above ambient temperatures. The emissions
from the 500-foot stack are at a temperature of 90°F, with a flow rate
of 700,000 CFM.
9-9
-------
At Plant B, all fluoride emissions before retrofit and about 90
percent of the emissions after retrofit are from the potroom areas.
Potroom area emissions are from roof monitors and from spray towers
near roof-top level. All such emission are at or near ambient temperature
and should be characterized by negligible plume rise. The remaining 10
percent of emissions after retrofit are from nearby ESPs, also near
roof-top level. The proportion of emissions from the ESPs is small
enough to permit the simplifying assumption that all emissions are from
the potroom areas.
At Plant C, all fluoride emissions before retrofit and about 70
percent of emissions after retrofit are from the potroom roof monitors.
Those roof monitor effluents, as at the other plants, are near ambient
temperature and are characterized by negligible plume rise. The
remainder of emissions after retrofit are from fume control units near
roof-top level at a temperature between 150°C and 200°F.
Generalized illustrations of roof monitor and stack emissions are
contained in Figures 5-1 to 5-4, Section 5.1. In Section 6.3, retrofit
layouts are given for Plant A in Figures 6-13 and 6-14, for Plant B in
Figures 6-15 and 6-16, and for Plant C in Figures 6-19 and 6-20.
Dispersion Estimates
For this analysis, a computerized dispersion model (CRS-1) recently
developed by the Meteorology Laboratory, NERC, was utilized. The CRS-1
is a Gaussian point source dispersion model. The model generates, for
any given year, maximum 1-hour, 24-hour, and annual ground level con-
centrations. Maximum concentrations for other averaging times (e.g., a
30-day averaging time in this case) can be obtained through special
analysis of the model output.
9-10
-------
Maximum 30-day fluoride concentrations were estimated for distances
of 0.75, 2, 10, 20, and 40 kilometers from the center of each facility.
The 0.75 kilometer distance is assumed to be approximately that of the
plant boundary in each case. All three facilities were assumed to be
isolated from other sources of fluorides.
The CRS-1 is designed to accept one year of hourly atmospheric
stability-wind data as input. For these analyses, the meteorological
input to the model consisted of one year of 3-hourly data, specially
preprocessed to generate estimated hourly values. The data were obtained
from a National Weather Service Station characterized by generally
restrictive dispersion conditions and reasonably representative of
conditions at the facilities studied. Probably the most important
characteristic of the dispersion conditions, as they pertain to this
analysis, is the high frequency of wind from a few directions, resulting
in higher 30-day concentrations than would occur with less restricted
air flow. Possible impaction of the effluent plumes on elevated terrain
was not considered.
The characteristics of the facilities, discussed earlier, required
modifications to the CRS-1 model itself. The facilities have certain
area-source characteristics and are affected by aerodynamic downwash of
plant effluents much of the time. Except for the 500-foot stack in the
before-retrofit case at Plant A, all fluoride emissions at the facilities
were of a low-level, area source configuration. The height and areal
extent of low-level emissions were similar at all three facilities.
For modeling purposes, low-level fluoride emission at all three facilities
were approximated by a uniform circular area source 500 meters across
9-11
-------
and 20 meters above grade. The area source was handled by approximating
low-level emissions as a "virtual point source" upwind of the facility
for each of the hourly CRS-1 computations.
Aerodynamic downwash of plant effluents was simulated by assuming
an effective stack height of 10 meters (one-half the average height of
low-level emissions) for wind speeds greater than or equal to 2 m/s.
For wind speeds less than 2 m/s, downwash was assumed to have a lesser
influence, and an effective stack height of 20 meters was assumed.
The 500-foot stack (in the before-retrofit case at Plant A) was
modeled through a separate CRS-1 analysis, which included a plume rise
estimate. Ground level ambient fluoride concentrations were superimposed
on concentrations due to low-level emissions to determine total impact
of Plant A before retrofit.
The highest 30-day average ground level ambient fluoride concen-
trations that were estimated for the year of data are presented in Table
9-5 for several downwind distances. The given concentrations are the
maximums for each of the specified distances. Note that best adequately
demonstrated control technology does not preclude undesirably high
fluoride concentrations, although the improvements in air quality are
still significant. Ambient fluoride concentrations may still exceed 0.5 p
up to 14, 24, and 20 Km downwind of plants A, B, and C respectively
after retrofit. Close to the source (e.g., at the plant boundary),
where the greatest impact on air quality occurs, the ambient ground
level fluoride concentrations resulting from controlled emissions can
be reduced further by directing most of the emissions up stacks tall
enough to avoid aerodynamic downwash of the effluent. Usually, a stack
9-12
-------
Table 9-5. MAXIMUM 30-DAY AVERAGE AMBIENT FLUORIDE
CONCENTRATIONS IN THE VICINITY OF PLANTS A, B, AND C
Plant
A
ii
ii
it
ii
it
B
n
n
n
n
ii
C
ii
n
n
n
Downwi nd
Distance (km)
0.75
(plant boundary)
2
10
14a
20
40
0.75
(plant boundary)
2
10
20
24a
40
0.75
(plant boundary)
2
10
20
40
o
Fluoride Concentration (yg/m )
Before Retrofit
41
13
2
-
0.6
0.2
104
34
5
2
-
0.5
219
72
10
3
1
After Retrofit
18
6
0.8
0.5
0.3
0.1
55
18
3
0.7
0.5
0.2
31
11
1
0.5
0.1
aThese values were interpolated.
9-13
-------
height about 2-1/2 times the height of any nearby buildings, or other
obstacles to wind flow, is sufficient to avoid such downwash problems.
With increasing distance from the source, the benefits of a taller
stack diminish. To significantly reduce ground level fluoride concen-
trations beyond a few kilometers from the source, there is no choice
but to further reduce emissions.
9.1.4 Particulate Emissions from Aluminum Reduction Cells
Particulate emissions will be significantly reduced by emission
controls. Reference (2) gives the particulate removal efficiencies for
various fluoride emissions control equipment. These efficiencies were
used with Tables 7-2 and 7-3 to calculate national particulate emissions
and average emissions for all plants for the three alternative levels of
fluoride emissions control as illustrated by Table 7-3. The results of
these particulate emissions calcualations are presented in Tables 9-6
and 9-7. The particulate fluoride emissions are reduced by about 50
percent.
As shown in Table 9-6, national particulate emissions from primary
aluminum reduction cells will be reduced from 44,000 to 25,000 tons/yr,
or to 21,000 tons/yr at the lowest level technically feasible. The
effectiveness of good controls in controlling particulate emissions is
not coincidental; a substantial percentage of the total fluorides
emitted from primary aluminum reduction cells is in particulate form.
Thus, in order to effectively control total fluoride emissions, total
particulate emissions must also be well controlled.
9.1.5 National Particulate and Fluoride Emissions from Anode Bake
Plants
Fluoride and particulate emissions for anode bake plants are
9-14-
-------
Table 9-6. NATIONAL PARTICULATE EMISSIONS FROM PRIMARY ALUMINUM
REDUCTION CELLS
National Participate Emissions by Cell Type
Control Level
Initial
Best Hooding and
Primary Control
Best Primary and
Secondary Control
vssa
5,200
5,200
4,500
HSSa
8,700
6,400
5,200
(Tons/Yr)
CWPBa
18,000
8,300
6,600
SWPB3
12,000
5,400
5,000
Total
44,000
25,000
21 ,000
VSS - vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake
9-15
-------
Table 9-7. AVERAGE PARTICIPATE EMISSIONS FOR PRIMARY ALUMINUM
REDUCTION CELLS
Control Level
Initial
Best Hooding and
Primary Control
Best Primary and
Secondary Control
Average Participate Emissions
by Cell Type (Lb/Ton Al )
vssa
16
16
14
HSSa
17
12
10
CWPBa
13
6.1
4.9
SWPB3
32
15
13
aVSS - vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake
9-16
-------
listed in Table 9-8, assuming that all anode bake plants retrofit with
spray tower and wet electrostatic precipitator (WESP) control. Accordingly,
before and after retrofit emissions are given.
Table 9-8. ANODE BAKE PLANT AIR POLLUTANT EMISSIONS
Air Pollutant
Emissions
Parti cul ate
Fluoride
Before Retrofit
Average
(Ib/ton Al)
5
0.86
National
(ton/yr)
8600
1500
After Retrofit
Average
(Ib/ton Al)
0.5
0.05
National
(ton/yr)
860
86
Table 9-8 shows that fluoride and particulate mass emissions from anode
bake plants can be reduced by 94 and 90 percent respectively through
application of the indicated control scheme. This degree of control
will reduce fluoride emissions from anode bake plants by 1,400 tons/yr
and decrease national particulate emissions by 7,700 tons/yr.
9.2 WATER POLLUTION IMPACT
As stated in Section 7.2, all plants with either wet primary or
secondary control were expected to meet 1983 effluent limitations
guidelines for primary aluminum plants. The only two factors which
could influence mass effluent emissions were capacity and the percentage
9-17
-------
of capacity with wet controls. In the following sections, 1983 effluent
guidelines are given, effluent emissions control schemes are outlined,
and national effluent emissions have been estimated for the alternative
levels of fluoride air emissions control. A more detailed discussion
of water pollution control can be obtained from Reference (4).
5
9.2.1. Effluent Limitations Guidelines for Primary Aluminum Plants
Table 9-9 lists 1977 and 1983 effluent limitations guidelines for
existing primary aluminum plants. As shown by the table, 1983 standards
require fluoride effluent emissions to be reduced by a factor of twenty
over the 1977 standards. 1983 effluent guidelines will affect all
existing primary aluminum plants by the date indicated; consequently,
they have been applied in calculating effluent emissions for all alter-
native levels of fluoride air emissions control.
9.2.2 Water Pollution Control Technology Required to Meet 1983
Effluent Guidelines Standards
The two basic water treatment schemes which will be practiced by
the primary aluminum industry are scrubber water recycle with lime pre-
cipitation of a bleed stream, and cryolite recovery with lime treatment
of the bleed stream. The two methods differ only in that alumina and
fluoride are recovered from the aqueous stream when cryolite recovery
is practiced. Success of both these methods require lime treatment of
a low volume, high concentration bleed stream; the bulk of the scrubber
water is recycled. This water pollution control approach allows fluoride
effluent emissions to be adequately controlled at a feasible cost.
9.2.3 National Effluent Emissions from Primary Aluminum Reduction Plants
National effluent emissions from primary aluminum plants have been
estimated by applying 1983 effluent guidelines standards to the individual
9-18
-------
Table 9-9. EFFLUENT LIMITATIONS FOR PRIMARY ALUMINUM PLANTS
vo
To Be
Achieved
by July 1,
1977
Effluent
Characteristic
Fluoride
Total Suspended
Solids
pH
Effluent Limitations
Maximum for Average of Daily Values for 30
any 1 Day Consecutive Days Shall Not Exceed
Metric Units English Units Metric Units English Units
(Kg/1000 Kg Al) (Lb/Ton AT) (Kg/1000 Kg AT) (Lb/Ton AT)
2.0
3.0
4.0
1.0
6.0 1.5
Within the range of 6.0 to 9.0.
2.0
3.0
To Be
Achieved
by July 1 ,
1983
Metric Units
(Kq/1000 Kg Al)
English Units
(Lb/Ton Al)
'letric Units
(Kg/1000 Kg Al)
English Units
(Lb/Ton Al)
Fluoride
Total Suspended
Solids
PH
0.1
0.2
0.2
.05
0.4 0.1
Within the range of 6.0 to 9.0.
0.1
0.2
-------
plant control options contained in Table 7-2, illustrated in Table 7-3,
and employed in Section 9. In estimating national effluent emissions,
all aluminum plants with wet controls at a particular fundamental level
of fluoride control were assumed to discharge aqueous wastes at the
maximum level allowed by 1983 standards. Thus, by a simple summation
procedure, similar to the one used to calculate national fluoride air
emissions, national effluent emissions at alternative levels of control
were derived.
Table 9-10 gives national effluent emissions by cell type for the
two pollutants regulated under the 1983 standards. Average effluent
emissions by cell type—equal to national cell effluent discharges
divided by total cell capacity--are contained in Table 9-11. As mentioned
in the introduction to this section, the applicability of 1983 effluent
limitations guidelines to the alternative levels of fluoride emissions
control allows the percent of capacity with wet controls to be determined
by inspection of the table of average effluent emissions. The ratio of
average emissions to 1983 effluent standards for a particular level of
control represents the fraction of capacity with wet controls at that
level. For example, if a particular cell type had average fluoride
effluent discharges of 0.050 Ib/ton Al, then the degree of wet control
would be 0.050/0.100 or 50 percent of capacity. Table 9-12 presents the
percent of capacity employing wet controls for the various cell types
and levels of air pollution control.
Tables 9-10 and 9-12 show that the example controls will not sig-
nificantly alter effluent emissions compared to the initial level of
air pollution control. Returning to the impact analysis procedure,
Q- 20
-------
Table 9-10. NATIONAL EFFLUENT EMISSIONS FROM PRIMARY ALUMINUM PLANTS
Air Pollution
Control Level
Initial: Fluoride
Total Suspended Solids
Best Hooding and Primary
Control :
Fluoride
Total Suspended Solids
Best Primary and Secondary
Control :
Fluoride
Total Suspended Solids
Nati
b
vssa
23
45
23
45
32
64
onal Efflu
y Cell Typ
HSSa
35
70
35
70
52
104
ent Emissio
e (Tons/Yr)
CWPBa
21
42
0
0
135
270
ns
SWPBa
28
56
27
55
37
74
Total
no
210
85
170
260
510
VSS - vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake.
9-21
-------
Table 9-11. AVERAGE EFFLUENT EMISSIONS FROM PRIMARY
ALUMINUM REDUCTION PLANTS
Air Pollution
Control Levels
Initial :
Fluoride
Total Suspended Solids
Best Hooding and Primary
Control :
Fluoride
Total Suspended Solids
Best Primary and Secondary
Control :
Fluoride
Total Suspended Solids
A\
vssa
0.07
0.14
0.07
0.14
0.10
0.20
/erage Eff
(Lb/1
HSSa
0.07
0.13
0.07
0.13
0.10
0.20
uent Emij
fon Al)
CWPBa
0.02
0.03
0
0
0.10
0.20
;sions
SWPBa
0.08
0.15
0.07
0.15
0.10
0.20
VSS - vertical stud Soderberg; HSS
CWPB - center-worked prebake; SWPB
horizontal stud Soderberg;
side-worked prebake.
9-22
-------
Table 9-12. EXTENT OF WET CONTROLS AT ALTERNATIVE LEVELS
OF FLUORIDE AIR EMISSIONS CONTROL
Air Pollution
Control Level
Initial
Best Hooding and
Primary Control
Best Primary and
Secondary Control
vssa
71
71
100
P
HSSa
67
67
100
ercent of Ce"
with Wet C(
CWPBa
16
0
100
1 Capacity
Dntrols
SWPBa
76
74
100
Tota
42
33
100
VSS - vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake.
9-23
-------
this can be interpreted to mean that effluent emissions in the absence
of emission guidelines would be comparable to effluent emissions with
additional emission controls. Thus, installing better air pollution
controls will not significantly increase water pollution.
9.2.4 Fluoride Air Emissions Control Schemes with Extreme Water
Pollution Impacts
To illustrate fluoride air emissions control systems with extreme
water pollution impacts, the overall environmental impact for two
individual plants is presented. Plant 14B represents the air pollution
control system with the most beneficial water pollution impact, primary
dry scrubbing. A typical wet air pollution control system is illustrated
by Plant 25B. It should be emphasized that the applicability of 1983
effluent guidelines standards makes the mass effluent emissions a
function only of plant capacity for a plant with wet controls. As a
result, all wet control syterns have been assumed to have average effluent
emissions equal to the 1983 standards.
Table 9-13 shows the overall environmental desirability of primary
dry scrubbing versus primary wet control. The CWPB plant generates no
fluoride control-related solid waste, nor does it discharge any effluent
emissions. Plant 25B generates 15,600 tons of solid waste resulting
from fluorides control and discharges 10.1 tons of aqueous fluorides
per year. Air fluoride emissions can also be controlled as--or more—
effectively with primary dry scrubbing, versus wet control for most
cell types. However, SWPB and VSS cells generally require secondary
control to achieve an acceptable level of fluoride emissions, and dry
scrubbing is not technically or economically feasible as a secondary
9-24
-------
Table 9-13. FLUORIDE AIR EMISSIONS CONTROL SCHEMES WITH EXTREME
WATER POLLUTION IMPACTS
Plant Code
Cell Type
Capacity (Tons Al/yr)
Primary Control Scheme
Secondary Control Scheme
Air Fluoride Emissions
(Tons F/yr)
Average Fluoride Air Emissions
(Lb F/ton Al)
Effluent Emissions
1. Fluoride: (Tons F/yr)
2. Total Suspended Solids:
(Tons/yr)
Fluoride Control Energy Require-
ments (mwh/yr)
Average Fluoride Control Energy
Requirements (kwh/ton Al)
Solid Waste Generated from Control
of Fluoride Emissions (Tons/yr)
Average Solid Waste Generation
from Control of Fluoride Emissions
(Lb/ton Al)
14B
CWPB
206,000
Dry Scrubbing
None
206
2.0
0
0
43,500
211
0
0
25B
HSS
202,000
Wet Electrostatic
Preci pita tor
None
465
4.6
10.1
20.2
20,200
100
15,600
154
9-25
-------
control strategy. Because of the necessity of secondary control in some
instances, the abandonment of all wet control schemes is precluded.
9.3 SOLID WASTE DISPOSAL IMPACT
All fluoride control related solid waste produced by the primary
aluminum industry is a direct result of wet, fluoride air emissions
control and the accompanying water treatment. Dry scrubbing techniques
do not generate any solid wastes, because all captured solids are returned
to the process. National and average solid waste generation from wet
fluoride control by the primary aluminum industry has been estimated by
applying the average solid waste generation figures in Table 9-14 to the
individual plant control options listed in Tables 7-2 and 7-3, or
derived as for plant 5A, etc. As with other impact evaluations, a
summation procedure was used to obtain national impacts. Solid waste
generation quantities listed with an asterisk in Table 9-14 were estimated
from probable effluent loadings based upon average fluoride generation.
All other solid waste entries in the table were obtained from Reference
(6).
9.3.1 National Solid Waste Generation Due to Fluoride Control by
the Primary Aluminum Industry
Tables 9-15 and 9-16 present national and average solid waste
generation caused by fluoride control in the primary aluminum industry.
Emission limitations will have an effect upon national solid waste
production similar to that for water pollution, namely: installation of
best controls will not significantly increase the solid waste over the
amount produced at the initial level of air pollution control. As shown
in Table 9-15, the national generation of solid waste by the aluminum
9-26
-------
Table 9-14. SOLID WASTE GENERATION FOR VARIOUS FLUORIDE
EMISSIONS CONTROL SCHEMES6
Fluoride Emissions Control Scheme
Primary Dry Scrubbing
Primary Wet Scrubbing with Cryolite Recovery
Secondary Wet Scrubbing with Cryolite Recovery
Primary Wet Scrubbing with Lime Treatment
Secondary Wet Scrubbing with Lime Treatment
Primary Wet Electrostatic Precipitator (WESP)
with Lime Treatment
Primary WESP with Cryolite Recovery
Solid Waste
Generation
(Lb/Ton Al)
0
154
160
120
40*
150*
154*
*These values were estimated from probable effluent loadings.
9-27
-------
Table 9-15. NATIONAL SOLID WASTE GENERATION FROM FLUORIDE CONTROL
BY THE PRIMARY ALUMINUM INDUSTRY
Air Pollution
Control Level
Initial
Best Hooding and
Primary Control
Best Primary and
Secondary Control
Natioi
Resull
vssa
35,000
35,000
44,000
lal Solid I
ting from
(Ton;
HSSa
54,000
54,000
75,000
daste Generc
Huoride Cor
3/Yr)
CWPBa
25,000
0
54,000
ition
itrol
SWPBa
45,000
44,000
48,000
Total
160,000
130,000
220,000
VSS - vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake
9-28
-------
Table 9-T6. AVERAGE SOLID WASTE GENERATION RESULTING FROM FLUORIDE
CONTROL FOR THE PRIMARY ALUMINUM INDUSTRY
Air Pollution
Control Level
Initial
Best Hooding and
Primary Control
Best Primary and
Secondary Control
AVI
vssa
110
110
140
2rage Solic
by C(
(Lb/-
HSSa
100
100
140
J Waste Gene
2ll Type
Fon Al)
CWPBa
19
0
40
jration
SWPBa
120
120
130
VSS - vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake
9-29
-------
industry from fluoride control is equal to 160,000 tons per year at the
initial, and 220,000 tons per year at the best, control levels of air
pollution control.
Average fluoride control solid waste generation is lowest for
center-worked prebake cells at all levels of fluoride emissions control.
Table 9-16 illustrates the large difference in average solid waste
generation for CWPB cells as compared to the other three cell types.
Soderbergs and side-worked prebake cells generate at least 100 pounds of
sludge per ton of aluminum produced at all levels of air pollution
control; CWPB cells produce from 0-40 pounds of sludge per ton of
aluminum over the same range of control levels. This difference in
solid waste generation can be attributed to the almost universal use of
primary dry scrubbing in controlling fluoride emissions from CWPB cells.
9.3.2 Fluoride Emissions Control Systems with Extreme Solid Waste
Impacts
Extremes in solid waste generation resulting from fluoride control
are illustrated by two individual plant examples. Table 9-17 compares
the overall environmental impacts of a plant with both primary and
secondary wet control and a plant with dry primary control only. The
two plants chosen as illustrative examples are both vertical stud
Soderbergs. Plant 3B employs primary dry scrubbing only, while Plant 20
utilizes primary wet scrubbers and wet electrostatic precipitators as
well as secondary spray screen scrubbers. These two examples also show
the large fluoride air emissions reductions which can be realized with
judicious application of secondary control.
9-30
-------
Table 9-17. FLUORIDE EMISSIONS CONTROL SYSTEMS WITH EXTREME SOLID
WASTE IMPACTS
Plant Code
Cell Type
Capacity (Tons Al/Yr)
Primary Control Scheme
Secondary Control Scheme
Air Fluoride Emissions
(Tons F/Yr)
Average Fluoride Air Emissions
(Lb F/Ton Al )
Effluent Emissions:
Fluoride (Tons F/Yr)
Total Suspended Solids (Tons/Yr)
Fluoride Control Energy Requirements
(Mwh/Yr)
Average Fluoride Control Energy
Requirements (Kwh/Ton Al)
Fluoride Control Related Solid Waste
Generation (Tons/Yr)
Average Fluoride Control Related
Solid Waste Generation (Lb/Ton Al)
3B
VSS
185,000
Fluidized Bed
Dry Scrubbing
None
833
9.0
0
0
39,000
211
0
0
20
VSS
120,000
Wet Scrubber and
Electrostatic
Preci pita tor
Spray Screen
Scrubber
122
2.0
6
12
55,600
463
11,400
190
9-31
-------
As shown in Table 9-17, Plant 20 generates 190 pounds of fluoride
sludge per ton of aluminum produced. This sludge generation implies an
annual production of 11,400 tons of solid waste containing CaF2, CaS04
(generated from sulfur in carbon anodes) and other insoluble compounds.
Sludge of this type can usually be safely landfilled.. Plant 3B does not
discharge any effluent emissions or produce solid waste, but this plant
has average fluoride air emissions 4.5 times greater than Plant 20. For
vertical stud Soderberg cells and other cells with similar emissions
characteristics, the increase in solid waste and effluent emissions
resulting from wet primary and secondary control is justified by the
dramatic decrease in fluoride emissions. Plant 3B with primary dry
control only, discharges 833 tons of fluoride per year into the atmos-
phere; Plant 20 emits only 122 tons of fluoride per year.
9.4 ENERGY
The energy assessment has been prepared through application of a
procedure similar to the ones employed for the other impact area analyses.
Average energy consumption requirements for the numerous fluoride emissions
control schemes (Table 9-18) have been applied to the individual plant
control options contained in Tables 7-2 and 7-3 or defined in Section
9. National and average fluoride emissions control energy requirements
are presented in Section 9.4.1, and examples of fluoride control systems
with extreme energy impacts are outlined in the section immediately
following. Energy consumption figures in Table 9-18 and throughout the
energy assessment represent the amount of energy required for both air
and water pollution control and have been obtained from References (6)
and (7). In some instances, the energy required for water treatment
9-32
-------
Table
9-18. ENERGY REQUIREMENTS FOR PRIMARY ALUMINUM
FLUORIDE EMISSIONS CONTROL SYSTEMS6'7
Fluoride Control
System
Dry Scrubbing
Primary Wet Scrubbing with
Cryolite Recovery Water Treatment
Secondary Wet Scrubbing with
Cryolite Recovery Water Treatment
Primary Wet Scrubbing with Lime Water
Treatment
Secondary Wet Scrubbing with Lime Water
Treatment
Electrostatic Precipitator Incremental
Power when Used in Series with Another
Primary Control Device
Primary Venturi Scrubbing
Primary Multiclone
Primary Wet Electrostatic Precipitator
with Lime Water Treatment
Average Energy Requirements
(kwh/ton Al)
Electrical
211
78
357
76
300
87
600
75
100
Thermal3
0
181
181
0
0
0
0
0
0
aThermal energy supplied by fossil fuels is required to operate rotary
kilns and generate steam when practicing cryolite recovery water
treatment.
9-33
-------
exceeds that for fluoride air emissions control.
t
9.4.1 National Fluoride Emissions Control Energy Requirements for
the Primary Aluminum Industry
National energy requirements for the three basic levels of fluoride
emissions control are listed in Table 9-19. As shown in the last
column of Table 9-19, emission control will increase fluoride control
energy expenditures for the primary aluminum industry by as little as
120,000 megawatt-hours (Mwh) per year over the initial level of control.
This increase in energy demand would require an additional 14 megawatts
of electrical power generation nationally. With 31 existing domestic
primary aluminum plants, the average incremental fluoride control power
expense resulting from best hooding and primary control will amount to
only 0.44 megawatts per plant. In comparison, an average of 283 megawatts
of electrolytic power is required per domestic primary aluminum plant.
Thus, incremental fluoride control power expenditures for the case given
are equivalent to only a 0.15 percent increase in average electrolytic
power.
Table 9-20 presents the average fluoride control energy require-
ments by cells type for the alternative levels of emissions control. On
an absolute basis, the average energy requirements per ton of aluminum
for each cell type is primarily a function of the percentage of that
cell type with secondary control. By examining the row entitled "Best
Primary and Secondary Control", it is clear that when all cell types are
upgraded to the level of best secondary control the average energy
requirement for all cell types is 600 - 90 kilowatt-hours (kwh) per ton
of aluminum. In contrast, the other two levels of control require a wide
range of average fluoride control energy expenditures: these range from
9-34
-------
Table 9-19. NATIONAL FLUORIDE EMISSIONS CONTROL ENERGY REQUIREMENTS
FOR THE PRIMARY ALUMINUM INDUSTRY
Air Pollution
Control Level
Initial
Best Hooding and
Primary Control
Best Primary and
Secondary Control
National Fluoride Control
vssa
270,000
270,000
400,000
Enet
HSSa
280,000
280,000
590,000
-gy Requirei™
(Mwh/Yr)
CWPBa
520,000
590,000
1,400,000
?nts
SWPBa
400,000
450,000
510,000
Total
1,470,000
1,590,000
2,900,000
VSS - vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake.
9-35
-------
Table 9-20. AVERAGE FLUORIDE EMISSIONS CONTROL ENERGY REQUIREMENTS
FOR THE PRIMARY ALUMINUM INDUSTRY
Air Pollution
Control Level
Initial
Best Hooding and
Primary Control
Best Primary and
Secondary Control
;
vssa
420
420
620
Average Fli
Energy R(
(Kwh/-
HSSa
260
270
570
joride Coi
squiremen'
fon Al)
CWPBa
190
220
520
itrol
ts
SWPBa
540
610
690
VSS - vertical stud Soderberg; HSS - horizontal stud Soderberg;
CWPB - center-worked prebake; SWPB - side-worked prebake
9-36
-------
a low of 190 kwh per ton of aluminum for CWPB cells to a high of 610 kwh
per ton of aluminum for SWPB cells. The wide range of fluoride control
energy requirements reflects the varying percentages of cell capacities
that would be required to install or maintain secondary controls.
9.4.2 Fluoride Emissions Control Systems with Extreme Energy Impacts
Table 9-21 shows the extremes in energy requirements of the various
fluoride emissions control schemes, and lists the overall environmental
impact for two individual plants. Plant 21A employs CWPB cells and is
equipped with primary multiclones only. Plant 18 is one of the best
controlled, but the most control-energy-intensive of primary aluminum
plants. Although Plant 21A has no effluent emissions or solid waste
production, its average energy expenditure of only 75 kwh per ton of
aluminum yields high average emissions of 14 pounds of fluoride per ton
of aluminum produced. Plant 18 uses ten times as much energy to control
fluoride emissions; however, its average emissions are 1.3 pounds of
fluoride per ton of aluminum produced. One significant aspect of Plant
18 is that nearly one-third of the energy utilized in control of fluorides
is a result of cryolite recovery water treatment. Substitution of lime
treatment only, in place of the cryolite recovery systems, would decrease
average energy consumption by 240 kilowatt-hours per ton of aluminum
produced.
9.5 OTHER ENVIRONMENTAL IMPACTS
Because of the electrical power required to control fluorides
emitted by the primary aluminum industry, each alternative level of
fluoride control has a unique indirect pollution penalty. Although many
primary aluminum plants are supplied with hydroelectric base load power,
9-37
-------
Table 9-21. FLUORIDE EMISSIONS CONTROL SYSTEMS WITH EXTREME ENERGY IMPACTS
Plant Code
Cell Type
Capacity (Tons Al/Yr)
Primary Control Scheme
Secondary Control Scheme
Air Fluoride Emissions
(Tons F/Yr)
Average Air Fluoride Emissions
(Lb F/ton Al)
Effluent Emissions:
Fluoride (Tons F/Yr)
Total Suspended Solids
(Tons/Yr)
Fluoride Control Energy
Requirement (Mwh/Yr)
Average Fluoride Control Energy
Requirements (Kwh/Ton Al )
Solid Waste Generated from
Control of Fluoride Emissions
(Tons/Yr)
Average Solid Waste Generation
Factor from Control of Fluoride
Emissions (Lb/Ton Al)
21 A
CWPB
70,000
Mul tic! ones
None
490
14
0
0
5,250
75
0
0
18
SWPB
260,000
Injected Alumina
Dry Scrubbing
Spray Scrubbing
with Cryolite
Recovery
169
1.3
13
26
195,000
750
20,800
160
9-38
-------
swing loads are often handled by coal-fired steam generators. In order
to determine the maximum indirect pollution penalty associated with the
alternative levels of fluoride control, it has been assumed that all
fluoride control power requirements are supplied by coal-fired steam-
electric plants. Table 9-22 lists incremental SOp, nitrogen oxides
(NO ), and particulate emissions which would be emitted by the power
A
generation required for the alternative levels of fluoride control. In
preparation of Table 9-22, it has been assumed that the steam generators
employed can meet the applicable standards of performance for new stationary
sources.
Comparing the first and last rows of Table 9-22, the incremental
indirect pollution penalty of the fluoride emissions controls is apparent.
Adoption of the proposed emission limits would cause an incremental
minimum of 700 tons per year of S09, 500 tons per year of NO , and 60
£ A
tons per year of particulates to be discharged to the atmosphere from
the impacted power plants. Although the national mass emissions of
indirect pollutants seems to be high, only a negligible effect on ambient
air quality will result.
9.6 OTHER ENVIRONMENTAL CONCERNS
9.6.1 Irreversible and Irretrievable Commitment of Resources
The only major irreversible impacts of the alternative fluoride
emissions control levels are the amounts of ultimate fossil-fuel required
to generate the necessary electrical power. Continuing with the assumptions
made in Section 9.5, national bituminous coal requirements have been
calculated for the same alternative levels of fluoride control and are
presented in Table 9-23. Although the probability is not high that all
9-39
-------
Table 9-22. NATIONAL CRITERIA POLLUTANT EMISSIONS RESULTING FROM
THE ELECTRIC POWER GENERATED TO CONTROL PRIMARY
ALUMINUM FLUORIDE EMISSIONS
Fluoride Air
Pollution Control
Level
Initial
Best Hooding and
Primary Control
Best Primary and
Secondary Control
National Criteria Pollutant
Emissions from Impacted Power
Plants (Tons/Yr)
so2
8,800
9,500
17,400
NOX
5,100
5,600
10,200
Parti cul ate
740
800
1,500
9-40
-------
Table 9-23. NATIONAL BITUMINOUS COAL REQUIREMENTS IMPLIED BY PRIMARY
ALUMINUM FLUORIDE CONTROL
Fluoride Emissions
Control Level
National Bituminous
Coal Requirements
(Tons/Yr)
Initial
612,000
Best Hooding and
Primary Control
662,000
Best Primary and
Secondary Control
1,210,000
For calculational purposes, bituminous coal was assumed to have a
high heating value of 12,000 Btu/lb.
9-41
-------
fluoride control energy requirements would be supplied by bituminous
coal, incremental fluoride control energy consumption will increase
fossil-fuel consumption directly or indirectly.
As shown in Table 9-23, the best fluoride primary control will
increase national bituminous coal (or other equivalent fossil-fuel)
consumption by 50,000 tons per year. Any additional drain upon fossil -
fuel reserves would be negligible compared to overall national fossil-
fuel energy requirements. In fact, an equivalent amount of energy would
be required if only one average size aluminum plant increased its capacity
by about 3 percent.
9.6.2 Environmental Impact of Delayed Action
Postponement of any fluoride emission limits would have some
deleterious effect upon the environment. Although some states have
strict fluoride emissions standards for primary aluminum plants, many do
not. Without the impetus for proper maintenance of existing control
systems, current fluoride emissions could conceivably increase significantly.
Federal effluent guidelines for primary aluminum plants will go into
effect regardless of air emissions limitations, and consequently, several
wet control systems could be abandoned if adequate air emissions limits
are not implemented. Thus, the suggested air emissions limitations
procedures will not only require poorly controlled plants to upgrade or
install new control systems, they will also force well-controlled plants
to maintain their existing control systems.
9.6.3 Environmental Impact of No Action
The environmental impact of failing to implement any additional
emission limits is represented by the "Initial" level of fluoride control
9-42
-------
used throughout the environmental impact assessment. As stated in the
previous section, the environmental impact of no fluoride emission
limits could be considerably more severe than that for the initial level
of control. Poor maintenance and the abandonment of existing control
systems could further increase fluoride air emissions. Thus, it
mist be realized that the initial level of fluoride control is a conservative
estimate of the effect of not implementing additional controls for
existing plants. With this limitation in mind, Table 9-24 compares the
overall environmental impact of the "Best Hooding and Primary Control"
to that for the initial level of control.
As discussed in Section 9.1, failure to adopt any of the suggested
emission controls for the primary aluminum industry would result in
national fluoride emissions of at least 18,000 tons per year, versus
9,000 tons per year for the improved control. With or without State
action on air emissions limits, effluent discharges and solid waste
generation by the primary aluminum industry will remain at nearly the
same level. The only beneficial impact of failing to adopt additional
emission limits would be an energy savings of 120,000 megawatt-hours per
year for the case shown in Table 9-24. This energy savings is negligible
in comparison to the amount of energy used in normal primary aluminum
plant operation.
9-43
-------
Table 9-24. ENVIRONMENTAL IMPACT OF NO ADDITIONAL STATE FLUORIDE
AIR EMISSIONS LIMITATIONS FOR THE PRIMARY ALUMINUM
INDUSTRY
National Impacts
National Fluoride Air
Emissions (Tons F/Yr)
National Effluent
Emissions:
.Fluoride (Tons F/Yr)
Total Suspended Solids
(Tons/Yr)
National Fluoride Control
Energy Requirements
(Mwh/Yr)
National Fluoride Control
Related Solid Waste
Generation (Tons/Yr)
Initial
18,000
110
210
1,470,000
160,000
Fluoride Emissions
Control Level
Best Hooding and Primary Control
9,000
85
170
1,590,000
130,000
9-44
-------
9.7 REFERENCES FOR SECTION 9
1. Memorandum from Laurence J. Budney to Stanley T. Cuffe. Primary
Aluminum Retrofit Controls; Impact. Monitoring and Data Analysis
Division, OAQPS, EPA. February 4, 1974.
2. Air Pollution Control in the Primary Aluminum Industry. Singmaster
and Breyer, New York, N.Y. Prepared for Office of Air Programs,
Environmental Protection Agency, Research Triangle Park, N.C. under
Contract Number CPA 70-21. July 23, 1973. pp. 5-14 to 5-16.
3. Reference 2, above, p. 4-11.
4. Development Document for Effluent Limitations Guidelines and
Standards of Performance. Nonferrous Metals Industry. Part II -
Primary Aluminum Smelting. Battelle Columbus Laboratories,
Columbus, Ohio. Prepared for Office of Air and Water Programs,
Environmental Protection Agency, Washington, D.C. under Contract
Number 68-01-1518. (Draft Contractor's Report, dated June 1973).
5. Reference 4, above, pp. 3 and 4.
6. Reference 4, above, p. 115.
7- Reference 2, above, pp. 5-14 to 5-16.
9-45
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA 450/2-78-0493
3. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
Primary Aluminum: Draft.Guidelines for Control of
Fluoride Emissions From Existing Primary Aluminum
Plants ^_____
5. REPORT DATE
2/14/79
6. PERFORMING ORGANIZATION CODE
AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This document serves as a text to State agencies in planning for control of
fluoride emissions from existing primary aluminum plants. Achievable fluoride
capture and removal efficiencies are given for new retrofit hooding and for
primary and secondary removal devices, respectively. Methods for deriving
capital and annualized costs are illustrated by a few examples. Costs and
fluoride emission reductions achieved by actual retrofits are given for ten
plants, and the construction scope of work is described in detail for three of
these plants. The guidelines are presented as recommended control technologies
that will achieve certain average control efficiencies when applied as new
retrofits to existing plants.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COS AT I Field/Group
Primary aluminum plants, existing
Fluorides
State Guidelines for emission control
Hooding efficiency
Fluoride removal efficiency
Retrofit control costs
Air Pollution Control
13B
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
343
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
No Charge
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
------- |