L.
-------
library
US EPA Region 4
AFC/9th FL Tower
61 Forsyth St. S.W.
Atlanta, GA 30303-3104
RACT DETERMINATION FOR
FIVE INDUSTRY CATEGORIES
IN FLORIDA
by
PEDCo Environmental, Inc.
11499 Chester Road
Cincinnati, Ohio 45246
Contract No. 68-02-3173
Task No. 18
PN 3450-18
Project Officer
Archie Lee
Air and Hazardous Material Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
REGION IV
ATLANTA, GEORGIA 3 03 08
November 198 0
-------
DISCLAIMER
This report was written for the U.S. Environmental Pro-
tection Agency (EPA), Region IV, by PEDCo Environmental, Inc.,
Cincinnati, Ohio, under Contract No. 68-02-3173. The contents
in this report are reproduced herein as received from the
contractor. The opinions, findings, and conclusions expressed
are those of the author and not necessarily those of EPA.
ii
-------
CONTENTS
Figures
Tables
Acknowledgment
1. Introduction
1.1 Background
1.2 Definition of RACT
1.3 Approach to determination of RACT
2. Phosphate Process Operations
2.1 Diammonium phosphate
2.2 Monoammonium phosphate
2.3 Granular triple super phosphate
2.4 Run-of-the-pile triple super phosphate
2.5 Run-of-the-pile normal super phosphate
2.6 Animal feed ingredients
2.7 Phosphate rock dryers
2.8 Phosphate rock grinding
2.9 Loading railroad cars with phosphate rock
2.10 Ships with phosphate rock loading
3 . Portland Cement
3.1 Process description
3.2 Emission sources and control options
3.3 Control costs
3.4 Recommended RACT
4. Electric Arc Furnaces
4.1 Process description
4.2 Emission sources and control options
4.3 Control costs
4.4 Recommended RACT
5. Sweat or Pot Furnaces
5.1 Process description
5.2 Emission sources and control options
5.3 Control costs
5.4 Recommended RACT
Page
v
vii
ix
1-1
1-1
1-2
1-5
2-1
2-1
2-19
2-27
2-39
2-49
2-55
2-71
2-85
2-97
2-105
3-1
3-1
3-3
3-12
3-18
4-1
4-1
4-3
4-5
4-5
5-1
5-1
5-2
5-2
5-4
iii
-------
CONTENTS (continued)
Page
6. Materials Handling, Sizing, Screening, Crushing,
and Grinding Operations 6-1
6.1 Process description 6-1
6.2 Emission sources and control options 6-3
6.3 Control costs 6-9
6.4 Recommended RACT 6-12
7. Summary 7-1
7.1 Recommended emission limits 7-1
7.2 Comparison with other existing regulations 7-4
References R-l
Appendix A-l
iv
-------
FIGURES
Number Page
2-1 Process Flow Diagram of a TVA DAP Plant 2-3
2-2 Process Flow Diagram of a Dorr-Oliver DAP Plant 2-7
2-3 Process Flow Diagram of a Typical 50 ton/h DAP
Plant Using Three Venturi Scrubbers and a
Crossflow Tail Gas Scrubber 2-16
2-4 Process Flow Diagram of a Plant Producing MAP
With a Prill Tower 2-20
2-5 Jet Nozzle Reactor Used to Produce MAP 2-21
2-6 Process Flow Diagram of Typical MAP Plant
Using Venturi Scrubber and Crossflow Tail
Gas Scrubber 2-26
2-7 Process Flow Diagram of a Dorr-Oliver Slurry
Granulation Plant 2-29
2-8 Process Flow Diagram of a Typical GTSP Plant
Using Venturi Scrubber and Crossflow Tail
Gas Scrubber 2-38
2-9 Process Flow Diagram of a Typical ROP/TSP Plant 2-42
2-10 Process Flow Diagram of a Typical ROP/NSP Plant 2-52
2-11 Process Flow Diagram of a Typical TVA AFI,
Ammonium Phosphate Plant 2-57
2-12 Process Flow Diagram of a Typical AFI Calcium
Phosphate Plant 2-59
2-13 Process Flow Diagram of a Typical AFI (Mono-
Ammonium/Diammonium Phosphate) Granulation
Plant Using Venturi Scrubbing and Crossflow
Tail Gas Scrubbing 2-64
v
-------
FIGURES (continued)
Number Page
2-14 Process Flow Diagram of AFI Production by the
Wet Phosphoric Acid Process With Fluorine
Removal by Calcining 2-67
2-15 Process Flow Diagram of Florida Phosphate
Rock Production 2-73
2-16 Process Flow Diagram of a Typical Phosphate
Rock Drying Plant 2-74
2-17 Direct-Fired, Cocurrent, Rotary Dryer 2-76
2-18 Fluidized-Bed Dryer 2-77
2-19 Typical Phosphate Rock Grinding Circuit 2-86
2-20 Roller Mill Used to Grind Phosphate Rock 2-87
2-21 Rotary Ball Mill Used to Grind Phosphate Rock 2-89
2-22 Materials Handling System for Conveying Phos-
phate Rock From Storage to a Railroad Car 2-98
2-23 Materials Handling System for Conveying Phos-
phate Rock and Granulated Fertilizer From
Storage to a Ship 2-106
2-24 Fugitive Dust Collection System Mounted on a
Movable Boom Above a Ship 2-10 7
3-1 Control Equipment Arrangement at Kiln 6 3-4
3-2 Collector Arrangement at Clinker Cooler 6 3-11
3-3 Finish Mill/Separator Fabric Filter Arrange-
ment 3-13
4-1 Electric Furnace for Steel Making 4-2
6-1 Simplified Flow Diagram of Materials Handling
Operations 6-2
7-1 Allowable Emissions From Portland Cement Plants
in Region IV 7-5
7-2 Allowable Emissions From Phosphate Processing
Operations in Region IV 7-7
vi
-------
TABLES
Number Page
1-1 Comparison of RACT, BACT, and LAER Requirements 1-4
2-1 Particulate Emission Sources and Control Options
for DAP Plants 2-12
2-2 Control Costs of a Typical DAP Plant 2-15
2-3 Particulate Emission Sources and Control Options
for MAP Plants 2-24
2-4 Control Costs of a Typical MAP Plant 2-25
2-5 Particulate Emission Sources and Control Options
for GTSP Plants 2-35
2-6 Control Costs of a Typical GTSP Plant 2-36
2-7 Particulate Emission Sources and Control Options
for ROP/TSP Plants 2-46
2-8 Control Costs of a Typical ROP/TSP Plant 2-47
2-9 Control Costs of a Typical ROP/NSP Plant 2-54
2-10 Control Costs of a Typical AFI Ammonium Phos-
phate Plant 2-62
2-11 Control Costs of a Typical AFI Calcium Phos-
phate Plant 2-69
2-12 Characteristics of Exhaust Gas From Rotary
and Fluidized-Bed Dryers 2-78
2-13 Particulate Emission Sources, Control Options,
and Achievable Emission Rates for Phosphate
Rock Dryers 2-8 0
2-14 Control Costs of Phosphate Rock Dryers 2-83
2-15 Characteristics of Exhaust Gases From Phosphate
Rock Grinders 2-91
vii
-------
TABLES (continued)
Number Page
2-16 Particulate Emission Sources, Control Options,
and Achievable Emission Rates for Phosphate
Rock Grinding Plants 2-93
2-17 Control Costs of Phosphate Rock Grinding 2-95
2-18 Particulate Emission Sources, Control Options,
and Achievable Emission Rates for Loading
Railroad Cars With Phosphate Rock 2-101
2-19 Control Costs of Loading Railroad Cars With
Phosphate Rock 2-103
2-20 Particulate Emission Source, Control Option,
and Achievable Emission Rate for Loading
Ships With Phosphate Rock 2-109
2-21 Control Costs of Loading Ships With Phosphate
Rock 2-112
3-1 Control Costs of Cement Kilns 3-14
3-2 Control Costs of Cement Clinker Coolers 3-15
3-3 Control Costs of Cement Finish Mills 3-16
4-1 Particulate Emission Controls for Electric Arc
Furnaces 4-4
4-2 Control Costs of a Typical Electric Arc Furnace
Producing Steel 4-6
5-1 Control Costs of a Typical Sweat or Pot Furnace 5-3
6-1 Particulate Emission Sources and Control Options
for Materials Handling 6-5
6-2 Control Costs of Materials Handling 6-10
7-1 Summary of RACT Recommendations 7-2
7-2 Allowable Emissions From Electric Arc Furnaces
in Region IV 7-8
7-3 Fugitive Dust Regulations in Region IV 7-10
viii
-------
ACKNOWLEDGMENT
This report was prepared for U.S. Environmental Protection
Agency (EPA), Region IV, by PEDCo Environmental, Inc., Cincinnati,
Ohio. Mr. Archie Lee was the Project Officer for the Environ-
mental Protection Agency. Mr. Charles E. Zimmer served as the
PEDCo Project Director, and Mr. Edwin A. Pfetzing was the Project
Manager. Principal authors were Messrs. Ron L. Hawks, Steve P.
M
Schliesser, Don J. Loudin and Edwin A. Pfetzing. PEDCo wishes to
thank the persons at the Florida Department of Environmental
Regulation at both the State and the regional level for their
assistance. PEDCo also wishes to thank the industries that were
visited during the project for their cooperation.
ix
-------
SECTION 1
INTRODUCTION
National Ambient Air Quality Standards (NAAQS) for total
suspended particulates are now being exceeded in portions of
Hillsborough and Duval Counties in Florida. The lowering of
particulate emissions within these areas requires that new or
modified control strategies be developed to ensure that all
reasonably available controls are used.
1.1 BACKGROUND
Two areas in Florida have been designated as nonattainment
for total suspended particulates. They are defined as follows:
the portion of Hillsborough County that falls within the
area of the circle having a centerpoint at the intersection
of U.S. 41 South and State Road 60 and a radius of 12
kilometers, and
the downtown Jacksonville area in Duval County located just
north and west of the St. Johns River and east of 1-95 and
south of Trout River.
Any particulate source that has a significant impact on
ambient particulate concentrations within the designated nonat-
tainment area are required to use Reasonably Available Control
Technology (RACT) to control particulate emissions.
1-1
-------
The application of RACT to existing stationary sources is a
required part of the particulate nonattainment corrective portion
of the State Implementation Plans (SIP). PEDCo investigated five
of the major industry categories that represent the type of
sources that are located in the two Florida nonattainment areas
to assist the state in determining specific RACT emission limita-
tions. These categories are:
Phosphate process operations
Portland cement plants
Electric arc furnaces
Sweat or pot furnaces
Materials handling, sizing, screening, crushing, and
grinding operations
1.2 DEFINITION OF RACT
Section 172(b)(2) of the Clean Air Act as amended August
1977 requires that SIP revisions "provide for the implementation
of all reasonably available control measures as expeditiously as
practicable." The use of RACT for stationary sources is defined
as "the lowest emission limit that a particular source is capable
of meeting by the application of control technology that is rea-
sonably available considering technological and economic feasi-
bility. nl
RACT is no longer defined by Appendix B, Code 4 0 of the
Federal Register, Part 51, entitled "Examples of Emission Limi-
tations Attainable With Reasonably Available Technology."
Reasonable availability is now based on the technological and
1-2
-------
economic feasibility of the control, and requires stringent and
even "technology forcing" control measures."1"
Although consistency in the application of any regulation is
important, determination of the "economic feasibility of a con-
trol may be very source specific.""'' Therefore, it is possible
that exceptions will be made to any RACT regulation on the basis
of economics. Such exceptions, however, are expected to be rare.
Every effort was made to make the recommended RACT determinations
specific for the affected plants in Florida.
The RACT emission limitations submitted in a nonattainment
SIP revision are used to calculate the emission reductions needed
to attain the NAAQS. Therefore, any deviations from these emis-
sion limitations are treated as SIP revisions. For this reason
RACT regulations should be adopted only after sufficient study to
ensure that they are indeed reasonable for the area in question.
Some of the confusion surrounding RACT stems from the compar-
ision of RACT with other control requirements. Table 1-1 gives a
comparison of RACT, Best Available Control Technology (BACT), and
Lowest Achievable Emission Rate (LAER) control requirements.
Although LAER will generally be more stringent than BACT and BACT
will generally be more stringent than RACT, in some instances the
required controls may be identical. This would occur if the
costs of installing controls at a new source were the same as
those of retrofitting controls at an existing source, or if the
cost were relatively low. The RACT control would be less strin-
gent than BACT in cases where, for technological or economic
1-3
-------
TABLE 1-1. COMPARISON OF RACT, BACT, AND LAER REQUIREMENTS
Control
requirement
Acronym
Appl icable
emi ssion
sources
Definition
Source of
definition
Stringencies
of requirement
Reasonably available
control technology
RACT
Existing sources in
nonattainment areas
The lowest emission limit that
a particular source is
capable of meeting by the
application of control tech-
nology that is reasonably
available, considering
technological and economic
feasibility
Memorandum of December 9,
1976, from the Assistant
Administrator of Office
of Air and Waste Manage-
ment
Least stringent
Best available con-
trol technology
BACT
New or modified
sources in an attain-
ment area subject to
Prevention of Sig-
nificant Deterioration
regulations
An emission limitation based
on the maximum degree of re-
duction determined on a case-
by-case basis, taking into
account several factors, in-
cluding cost, energy, and
technical feasibility
Section 169 (3) of the Clean
Air Act as amended 1977
Moderately stringent
Lowest achievable
emission rate
LAER
New or modified source
in nonattainment areas
The emission rate of the most
stringent limitation contained
in any State Implementation
Plan for such source cate-
gory or the most stringent
limitation achieved in
practice anywhere, whichever
is more stringent (with no
allowance for economic
factors)
Section 171 (3) (A) and (B)
of the Clean Air Act as
amended 1977
Most stringent
aIn some circumstances, control requirements may be equal for RACT, BACT, or LAER, but the stringency order may never be reversed.
-------
reasons, controls considered feasible for a new or modified
source would be unreasonable if an owner were required to retro-
fit them at an existing source.
The following information sources are used for guidance in
RACT determination:
New Source Performance Standards
Documents regarding particulate emission control techniques
Existing state and Federal regulations, especially
those in Region IV
Information gathered during plant visits
Information obtained from state representatives
In any SIP revision, the attainment of the NAAQS through
the application of a reasonable control strategy is the primary
objective. This requires decisions concerning which specific
sources should be controlled, based on a realistic comparison of
the available control options. Comparison of control costs must
consider both total annual costs and cost per ton of pollutant
removed. The economic justification for recommended RACT con-
trols relies heavily on such cost comparisons. Technical feasi-
bility analysis takes into account the controls required in other
states, plus information in technical publications and an assess-
ment of site specific factors that could affect the technical
feasibility of retrofitting and properly operating various tech-
nologies .
1.3 APPROACH TO DETERMINATION OF RACT
To satisfy the definition of RACT requires that controls
retrofitted at an existing facility be as stringent as possible,
1-5
-------
yet both technologically and economically reasonable. Recommended
RACT was determined by comparing control options in increasing
order of stringency until the next control option was deemed
infeasible for either economic or technical reasons. When the
economic feasibility of a control option was subject to inter-
pretation, more than one option was given, and the cost in
dollars per ton of particulates removed was calculated for com-
parison by the user. In each case, PEDCo made a judgmental
choice of which control option best represents RACT.
In general, the technological feasibility of a control was
the first parameter ascertained. Once a control was deemed
technologically feasible for retrofit, its economic feasibility
was determined. If the control was judged to be both technolog-
ically and economically feasible, its efficiency was estimated,
and an emission limitation was calculated based on this effi-
ciency. Enforceability was weighed heavily in choosing the
method of regulation. A control option that has a high cost in
dollars per ton of pollutant removed may be included if it has a
small capital cost.
Technological feasibility was based on the demonstration of
these control technologies on an identical or similar emission
source.
The economic feasibility of a control is less straightfor-
ward than its technological feasibility. The cost of retro-
fitting a control tends to be more plant-specific than the
technological feasibility of the control; therefore, economic
1-6
-------
feasibility often must be determined on a case-by-case basis.
Therefore, information gathered during plant visits was weighed
heavily in determining economic feasibilities in this report.
PEDCo used the following general tests to determine whether
a specific control could be considered RACT:
The control had a reasonable cost per ton of particulates
removed.
The control had a low overall cost.
The control has generally been applied in the industry or
within similar industries whether it was required by regu-
lation or not.
The control was reasonable in total cost and was capable of
meeting the most stringent regulations in Region IV and in
the country.
Application of the control would contribute to the attain-
ment of the NAAQS in the present nonattainment areas.
1-7
-------
SECTION 2
PHOSPHATE PROCESS OPERATIONS
This section discusses the processes involved in the phos-
phate processing industry. It also discusses the emission
sources, control options for these sources, costs of the control
options, and recommended RACT. Each phosphate product is treated
separately with subsections covering the following products:
Diammonium phosphate (DAP)
Monoarrurionium phosphate (MAP)
Granular triple superphosphate (GTSP)
Run-of-the-pile triple superphosphate (ROP/TSP)
Run-of-the-pile normal superphosphate (ROP/NSP)
Animal feed ingredients (AFI)
Also discussed in this section are RACT controls for phos-
phate rock dryers, phosphate rock grinding, loading railroad cars
with phosphate rock, and loading ships with phosphate rock. The
Appendix summarizes control and emissions data about the phos-
phate industry.
2.1 DIAMMONIUM PHOSPHATE
This subsection discusses the processes used in the produc-
tion of DAP and the monoammonium phosphate formed in the DAP
process and identifies the major particulate sources within each
2-1
-------
facility. The subsection also discusses the available control
technology and the cost of a typical plant using this tech-
nology. From this analysis RACT recommendations are made.
2.1.1 Process Description
Ammonium phosphates are produced by reacting phosphoric
acid with anhydrous ammonia. Diammonium phosphate production
combines one mole of phosphoric acid with 2 moles of ammonia to
yield a product having 21.2 percent nitrogen and 53.8 percent
available phosphorus according to the reaction:
h3po4 + 2nh3 (nh4)2hpo4
Fertilizers are identified by a three number combination that
identifies the percent N, the percent P_0r, and the percent
2 S
2
K20. Typical compositions are between 11-48-0 and 18 -
46-0. Commercial ammonium phosphates are produced by two
major processes: the TVA process (which uses a rotary drum
mixer), and Dorr-Oliver process (which uses a pugmill ammoniator).
Approximately 95 percent of the plants in the United States use
the TVA process.
TVA Process—
Figure 2-1 is a process flow diagram of a TVA DAP granula-
tion plant. Phosphoric acid is mixed in an acid mix tank with
reagent (93 percent sulfuric acid). The mixed acids typically
have a £2^5 content °f 40 to 45 percent. The phosphoric acid
used is a mixture of unconcentrated and concentrated wet process
phosphoric acid.
2-2
-------
Figure 2-1. Process flow diagram of a TVA DAP plant.
-------
The mixed acids are neutralized and premixed in a brick-
lined acid reactor. Anhydrous gaseous or liquid ammonia is
introduced with air and steam. The neutralization takes place at
atmospheric pressure and the charge ratios (NH^ to H^PO^ on a
molar basis) are between 1.3:1.0 and 1.5:1.0. The heat of reac-
tion is used to maintain the temperature of the slurry at 100° to
120°C. The heat allows the ammonium phosphate slurry to be
concentrated by the evaporation of excess water and yet to
maintain flow characteristics for pumping to the ammoniator gran-
ulator. The slurry at this temperature and molar composition is
primarily monoammonium phosphate with a solids content of 78 to
82 percent. The reactor is vented by induced draft to reduce
emissions of ammonia within the plant. Typical ventilation rates
are between 2000 and 2500 scfm, but actual rates vary with reactor
design and tightness. The tightness of a system can be improved
by an effective operation/maintenance (O&M) plan. The reactor
gases are scrubbed with a wet scrubber to remove the ammonia.
Typical scrubber solutions are phosphoric acid (30 percent P2°5^ '
The solubilized ammonium phosphate is recycled to the reactor.
The reactor slurry is pumped to the ammoniator-granulator,
in which the formation of DAP is completed and the granular
product is formed. The granulator consists of a rotary drum with
retaining rings at each end and a scraper mounted inside the
drum. A moving bed of recycled DAP fines are maintained in the
drum at all times. Slurry from the reactor is sprayed on the
recycled fine bed as ammonia is introduced under the bed. The
2-4
-------
final mixture reaches an ammonia phosphoric acid ratio between
1.8:1.0 and 2.0:1.0 (mole basis). The recycle fines are coated
with slurry and grow by agglomeration. The product is withdrawn
frijm the granulator as new fines are introduced. The recycle
rate is highly variable, but typically is 2.5 to 4.0 lb/ton of
product. The granulator is vented by induced draft to prevent
the loss of ammonia within the plant. Typical ventilation rates
are between 8,000 and 10,000 acfm. The ammoniation reaction in
the granulator is exothermic, and the reactor is maintained
between 85° and 105°C. The exhaust gases from the granulator
contain ammonia not consumed in the reaction. The off-gases are
typically scrubbed with a solution of phosphoric acid (30 percent
?20^). Ammonium phosphate is returned to the reactor as scrubber
recycle.
Moist (plastic) DAP granules are transferred to a rotary
oil- or gas-fired cocurrent flow dryer. In the dryer the mois-
ture content of granules is reduced below 2 percent. Exhaust
gases, which contain entrained particulates, are passed through a
bank of simple cyclones to remove large particulate and then
exhausted to a wet scrubber.
The temperature of the granular product at discharge from
the dryer is between 82° and 104°C. The granules are elevated by
bucket elevators and screened before cooling. The oversized
materials are transferred to cage mills for size reduction, and
the fines are recycled to the ammoniator-granulator.
2-5
-------
The product is cooled in a rotary cooler to prevent caking
and to reduce decomposition in storage. The cooler, screens, and
handling equipment are ventilated, and exhaust gases are treated
by. a bank of simple cyclones. The exhaust gases are then treated
by wet scrubbers.
The typical granule size of the product is between 1 and 4
mm. To prevent dusting in storage and transfer, some manufactur-
ers treat the granules with lubricating oil (0.5 percent by
weight). The granular DAP is placed in covered storage by over-
head belt conveyors. The product is sold in bulk form or bagged.
Dorr-Oliver Process—
Figure 2-2 is a process flow diagram of a Dorr-Oliver DAP
granulation plant. Phosphoric acid (24 to 36 percent P2°5^ "*"s
fed to a series of agitated reactors in which reaction occurs
with liquid or gaseous anhydrous ammonia. The reactants are
transferred through a series of vessels in which the slurry in-
creases in solids content and the pH is adjusted. The reactors
are vented, and the off-gases are scrubbed with phosphoric acid
(30 percent PnOc).
2. D
The ammonium phosphate slurry is transferred to a pugmill
(blunger) in which recycled fines are added. The blunger con-
tains parallel, counterrotating shafts with blades. The blades
mix the slurry and recycle fines together to form grandules. The
ratio of slurry to recycle is typically 6 to 12 lb/lb of product.
The blunger is ventilated, and exhaust gases are treated with a
scrubber.
2-6
-------
Figure 2-2. Process flow diagram of a Dorr-Oliver DAP plant.
-------
The product is transferred to a rotary, counterflow-fired
dryer in which the moisture content is reduced to less than 2
percent. Exhaust gases from the dryer containing entrained par-
ticulates are passed through a bank of simple cyclones to remove
large particulates and then exhausted to a wet scrubber. The
granular product is elevated by bucket elevator to double-deck
screens, and the oversize product is reduced in a cage mill. The
undersized product is recycled to the blunger. The screens,
mill, and transfer equipment are vented to simple cyclones and
then to a wet scrubber. The granule size is typically between
2.4 and 1.7 mm. The product is shipped in bulk or bagged form.
Review of the processes operated in the phosphate producing
area of central Florida indicated 11 facilities producing DAP by
the two processes. The production rates of the lines are between
35 and 98 tons/h. Emission rates and control devices for these
sources are listed in the Appendix.
2.1.2 Emission Sources and Control Options
A conventional DAP plant contains seven major points of
particulate emissions:
(1) Reactor(s)
(2) Ammoniator-granulator (or blunger)
(3) Dryer
(4) Product screens
(5) Cooler
(6) Cage mills
(7) Elevators, belt conveyors, etc.
2-8
-------
The volume of air required to prevent fugitive ammonia from
being lost from the reactor(s) varies greatly from plant to
plant and is typically included with exhaust from the ammoniator-
granulator (or blunger). The gas volumes reported for single TVA
process reactors range from 2000 to 2500 scfm at a temperature of
100° to 120°C. The ammoniator-granulator is exhausted at 8,000
to 10,000 acfm 85° 105°C. Gases from the reactors typically do
not contain particulate and are vented to control ammonia only.
The ammoniator-granulator contains particulate and ammonia emis-
sions. The combining of these gas streams is accomplished to
allow the recovery of ammonia in the gas stream. Typical loss
from a TVA process reactor/ammoniator-granulator process is 30 lb
of NH^/ton of DAP product in a 50-ton/h facility. The entrained
DAP dust from the ammoniator is roughly 27 lb/ton of product.
The uncontrolled emission rates make it economically feasible to
recover both ammonia and particulate for process recovery.
Without exception the preferred method of control in the
plants surveyed is a venturi scrubber. The scrubbers are operat-
ed at a wide range of static pressure drop (AP) and liquor-to-gas
ratios. Appropriate operating parameters should be specified in
the O&M plan. The scrubbing solution is phosphoric acid (typi-
cally 30 percent P2°5^ ' T^e wet Particulates anc^ ammonium
phosphate formed from the acid/ammonia reaction in the scrubber
are separated in cyclonic separator following the venturi.
Typical static pressure drop observed was between 10 and 15 in.
I^O. The liquor-to-gas ratio was approximately 12 gal/1000
2-9
-------
acfm. The venturi was followed by second-stage cyclonic spray
scrubbers, packed beds, crossflow scrubbers, or a second venturi
for final particulate control and fluoride control.
The dryer, which is used to remove moisture from the ammo-
nium phosphate after granulation, is typically gas- or oil-fired.
Heat inputs depend on moisture content of the granules and the
product recycle ratio. Typical heat inputs are in the range of
500,000 Btu/ton of product. Particulate emissions from the dryer
are controlled by a simple cyclone or bank of multiple cyclones
followed by a venturi scrubber. The scrubber is typically
followed by a cyclonic separator. Secondary collection is
provided by a cyclonic spray scrubber, a crossflow scrubber,
packed beds, or an additional venturi. Secondary collectors are
used for fluoride control. The typical exhaust volume from the
dryer is 65,000 acfm at 104°C, the typical static pressure drop
in the venturi is 14 in. and the typical liquor-to-gas ratio
is 10 gal/acfm.
The product cooler is a rotary drum cooler, which allows the
dried, screened granules to lose heat before being placed in
storage. The motion of the granules (in a cooler equipped with
flights) and of the air exhausted from the cooler results in
rapid cooling of the product. The exhaust volume required for
cooling is highly variable, depending on plant specification.
Average values of the exhaust appear to be near 50,000 acfm.
The exhaust is pretreated with a simple cyclone or bank of
cyclones before being scrubbed. Scrubbers used are venturi,
2-10
-------
packed beds and wet cyclones. The venturi is typically followed
by cyclonic spray scrubbers or crossflow scrubbers for secondary
particulate collection or fluoride control.
Several systems were observed to use simple wet cyclones
for particulate control. These systems were demonstrated to
operate at an emission rate of 0.55 to 0.83 lb/ton of product.
The cage mills, elevator, conveyors, and transport equip-
ment generate fugitive particulate emissions. These emissions
are ventilated at elevator heads, transfer points, screen head
space, and cage mills. The ventilation rate varies from plant
to plant, depending on age, tightness, and number of transfer
points. Emissions from these sources are controlled by simple
cyclones or banks of simple cyclones with recycle to granulator
and are followed by a scrubber. The scrubber is typically a
venturi followed by a packed bed, crossflow scrubber, or cyclonic
spray scrubber.
A summary of control options used in central Florida is
presented in Table 2-1. Because of the large number of com-
binations of control equipment used, stack test data do not
indicate control efficiency at individual process sources. The
major sources are controlled by venturi scrubbers with each
exhaust discharged to a common tail gas scrubber or common
stack. The uniqueness of each plant makes testing of each
subprocess component impractical. The most reasonable method of
emission measurement appears to be applying a mass emission rate
to the common exhaust or summing mass emissions from each
exhaust.
2-11
-------
TABLE 2-1. PARTICULATE EMISSION SOURCES AND CONTROL
OPTIONS FOR DAP PLANTS
Source
Control option
Reactor,
granulator/bl unger
Venturi/crossflow,
venturi/packed bed,
venturi/cyclonic spray,
packed bed/crossflow
Dryer
Venturi/crossflow,
venturi/packed bed,
venturi/cyclonic spray,
packed bed/crossflow
Screens,
cooler,
cage mills,
conveyors
Venturi/crossflow,
venturi/spray cyclone,
packed bed/crossflow,
wet cyclone
2-12
-------
A review of stack test data on the 11 process lines indicated
total controlled plant emission rates of 0.11 to 0.95 lb/ton of
product. Grouping of emission sources by control system type
indicates that the lowest emission rate is achieved with a
medium-energy Venturis followed by crossflow scrubbers at high
liquid-to-gas ratios or packed bed scrubbers followed by cross-
flow scrubbers. The highest emission rates occur at plants using
wet cyclones or cyclonic spray scrubbers. Controlled rates at
plants using medium-energy Venturis followed by crossflow scrub-
bers range from 0.19 to 0.54 lb/ton. The controlled emission
rate from systems using packed bed scrubbers followed by cross-
flow scrubbers is between 0.11 and 0.31 lb/ton. Systems using
combinations of controls (e.g., Venturis, spray cyclonic scrub-
bers, and wet cyclones) have controlled emission rates between
0.64 and 0.95 lb/ton. Data for specific plants are given in the
Appendix.
To ensure that the control equipment emission rates truly
reflect plant emissions requires complete efficient capture of
emissions from sources within the plant. The installation and
operation of enclosures, hoods, and ventilation systems in con-
junction with an O&M plan have been demonstrated to reduce fugi-
tive losses effectively within the manufacturing plant. The
maintenance of these systems has been demonstrated to be a major
problem in reducing emissions. To ensure good operating prac-
tices requires establishing an effective method of evaluating
loss to the ambient atmosphere. An opacity standard applied to
2-13
-------
fugitive loss from the building is an effective method of main-
taining control of fugitive emissions. Plants using good capture
practices can reasonably be expected to reduce fugitive losses
from building vents to less than 5 percent opacity.
2.1.3 Control Costs
An estimate of the cost of the various control options
discussed in the previous subsection is presented in Table 2-2.
The cost and emission estimates are based on a typical plant
producing 50 tons of DAP/h by the TVA process. The air volumes,
temperatures, and uncontrolled emission rates are based on mass
balances provided by the phosphate industry. The material of
construction is stainless steel. The scrubber pressure drop,
liquor rates, and fan horsepower are typical of the plants sur-
veyed. The capital cost is based on the values reported on
permit applications on file with the State of Florida Department
of Environmental Regulation, Tampa, Florida. The capital costs
have been adjusted to January 1980 dollars by use of the Chemical
3
Engineering (C-E) Plant Cost index. Because of the wide range
of plant layout and design, number of vendors available to supply
components, and degree of safety factor and redundancy designed
into individual plants, capital cost may vary by + 50 percent.
The cited examples are those for which more current data were
available (i.e., examples typical of current technology).
Figure 2-3 shows the control system arrangement and process
parameters for the first option in Table 2-2. The cost analysis
includes the total capital cost of the venturi and tail gas
2-14
-------
TABLE 2-2. CONTROL COSTS OF A TYPICAL DAP PLANT3
Emissions
Cost, $
Cost of
(credit for)
removal,
J/ ton
Source of
emissions
Control
options
Control,
%
Uncontrolled,
lb/ton
Control led,
lb/ton
Removal,
tons/yr
Total
capi tal
cost, t
Expected
1 i fe,
yr
Annual.
capital
Annual
O&M
Credits
Reactor,
dryer,
granula-
tor,
cooler,
equipment
vents
(3)
Venturl/
cross-
flow
99.37-
99.77
85.9
0.19-0.54
17,180
1,724,000
10
280,581
471,937
3,178,300
(141)
Reactor,
dryer,
granula-
tor,
cooler,
equ i pment
vents
(2)
Venturi/
spray
cyclonic
plus
(2) wet
cy-
clone
98.90-
99.25
85.9
0.61-0.94
16,992
1,302,000
10
211,914
345,024
3,143,520
(152)
aCosts in January 1980 dollars.
bBased on a 10 percent cost of capital.
-------
fO
I
0"»
REACTOR
GRANULATOR
12,000 acfm
270 F
8,100 acfm
190 F
9.9 gr/dscfm
r-& AP = 11 in. H20
1200 gal/min
22% phosphoric acid
DRYER ^>000 acfn^
225°F
3.7
rX 4P 5
1200 gal/min
22% phosphoric acid
COOLER 59>0°0 acfrU.
143°F
EQUIPMENT VENTS —n—
143 F
21,000 acfm
2.2
rK
AP =
1200 gal/min
22% phosphoric acid
gr/dscfm
11 in. H^O
CROSSFLOW TAIL GAS SCRUBBER
0--
gr/dscfm
11 in. H20
160,000 acfm, 140UF
0.0090 gr/dscfm
EFFICIENCY,^99.77%
RECYCLE POND WATER, 6,000 gal/min
Figure 2-3. Process flow diagram of a typical 50 ton/h DAP plant
using three venturi scrubbers and a crossflow tail gas scrubber.
-------
scrubber. Utilities are computed for the total system gas
volume and resistance. The water flow requirements are based on
venturi plus tail gas scrubber.
The second control system option in Table 2-2 is based on a
system using two venturi scrubbers with cyclonic separators
followed by a spray tail gas scrubber. The product cooler is
controlled by twin wet cyclones. The utilities are based on
total gas volume and water flow rates for the system. The
design volumes and production rates are not ideally matched to
those in the first option; therefore, the costs are not as
comparable as desired.
The capital cost may vary + 30 percent. The uncontrolled
emission factor is based on 85.9 lb/ton product loss from all
sources within the complex. This PEDCo estimate compares with
82 lb/ton estimated in AP-42.^ The controlled emission rate is
based on the range of controlled rates gathered from stack test
on file with the Florida Department of Environmental Regulation
(DER). The collected particulate rate is based on a maximum
production rate of 50 tons/h for 8000 h/yr. The annualized cost
is also based on 8000 hours of production.
The recovered product in the scrubbers is considered valu-
able and is typically returned to the processes in reactors or
granulators. If it is assumed that the solids collected are
composed of diammonium phosphate and the material is totally
recoverable, the credit for recycle is $3,143,000/yr, assuming a
value of $185/ton based on market value of the product.
2-17
-------
Applying the credit for product recovery to the first
control system option yields a net cost savings of $141/ton.
The cost of operation of the second option is less than the
first, but inclusion of the recovered product credit yields a net
operating credit of $152/ton removed (i.e., a net savings). In
practice scrubbers are a necessary product recovery portion of
the process and actually decrease the cost of manufacture.
2.1.4 Recommended RACT
From the standpoint of technical feasibility, a combination
of control devices (venturi scrubbers, cyclonic spray scrubbers
and crossflow or packed bed scrubbers) can achieve emission
levels in the range of 0.30 lb/ton of product. The use of low-
energy wet cyclones clearly results in higher controlled emission
rates. The variation in plant design and scrubber arrangement
does not allow the selection of one combination of control system
components or the selection of operating conditions that can
achieve this recommended level of emissions. Stack test data
indicate that at least a medium- to high-energy primary scrubber
is required.
The cost of using the lower-energy scrubber (wet cyclone) as
compared with the higher-energy venturi is not clearly defined
because capital cost is only available for plants using a venturi
in combination with wet cyclones. Because of the highly site-
specific nature of each device and control system, it was not
considered advisable to reduce the analysis to a subprocess/con-
trol system level. In either case it was observed that the
2-18
-------
primary scrubbers were used for product recovery and that the
credit from recovered product was greater than the annualized
cost of systems.
A fugitive emission limit applied to the plant process
building is recommended to ensure that the emissions measured by
the stack test represent the controlled emission level. The
fugitive emission level is based on the opacity of material
escaping the building. Based on good maintenance and good hood
capture efficiency, a 5 percent opacity limit is recommended.
The test method to be used for determining the mass emission
standard should be U.S. Environmental Protection Agency (EPA)
4
Method 5. The use of m-stack methods for scrubbers can yield
low emission rates, because of the solubilization of the diammo-
nium phosphate on the filter when collected below the moisture
5
dewpoint. Opacity should be determined by EPA Method 9.
2.2 MONOAMMONIUM PHOSPHATE
This subsection discusses the processes used in the produc-
tion of MAP by the spray tower process and identifies the major
particulate sources from the process area. Also, it discusses
the available control technology and the cost of using this tech-
nology. Based on this analysis, RACT recommendations are made.
2.2.1 Process Description
Ammonium phosphate is produced by the combination of phos-
phoric acid and ammonia to form MAP by the reaction:
h3po4 + nh3 - nh4h2po4
2-19
-------
The process used to produce the product involves the reaction of
the ammonia and acid in a jet spray nozzle under pressure. The
reaction product emerging from the reactor consists of molten
MAP suspended in a high-velocity steam jet. Suspended MAP
solidifies into tiny, round, porous particles, and steam and hot
air are exhausted from the enclosure from the top counterflow to
the falling product. The product is collected in a dry state at
the bottom of the enclosure (Prill Tower) and removed by rotating
rakes onto conveyors for transfer to a product cooler and to
storage. Figure 2-4 shows the Prill Tower method of producing
MAP.
The formation of the product occurs in the jet nozzle
reactor as the anhydrous ammonia flashes to gas in the presence
of the liquid phosphoric acid. Figure 2-5 shows the jet nozzle
reactor used to produce MAP. The MAP begins to form at the tip
of the inner nozzle and exits the outer cone at 500 miles/h.
The acid used in the process is wet process phosphoric acid (52
percent P2°5^ * T^e ammonia injection pressure is typically 120
psig. The particulate emissions from the cooler and spray tower
are typically exhausted to a scrubber.
Review of the process operated in the central Florida
phosphate producing area indicated two facilities producing MAP
by this process. The production rates were 14 and 25 ton/h.
2.2.2 Emission Sources and Control Options
In the manufacture of MAP by use of a jet nozzle reactor
and Prill Tower, there are typically two sources of particulate
2-20
-------
ANHYDROUS LIQUID AMMONIA
Figure 2-4. Process flow di
agram of a plant producing MAP with a Prill Tower.
-------
nh4h2po4
/
y!\\\v\\\\\ \ CTV
v\v\\\\\\\\y\ \ TT~
Figure 2-5. Jet nozzle reactor used to
produce MAP.
2-22
-------
emissions: the Prill Tower and rotary cooler.
The exhaust from the Prill Tower contains ammonia, moisture,
and fine particles of ammonium phosphate. The volume of gas is
dependent on the nozzle design and tower cross-sectional area.
Towers producing fine granules (powder) use less gas volume to
reduce entrainment of product. There is a minimum volume
necessary to remove heat and moisture and allow the MAP par-
ticles to solidify. Of the two plants studied, one combined the
Prill Tower and cooler exhausts into a single scrubber, and the
other did not use a cooler. The gas volumes of the two control
systems were 19,000 and 62,000 acfm, and the process weights
were 14 and 25 ton/h. The control devices used in both facil-
ities were combinations of venturi scrubbers and crossflow
scrubbers. The controlled emission rates were 0.12 and to 0.19
lb/ton.
The use of venturi scrubbers is an accepted method of
controlling particulate emissions. The tail gas scrubber is
primarily intended to control fluorides. No data are available
concerning particle size distribution or uncontrolled emission
rate from the Prill Tower or cooler. Without this data, it is
not possible to predict accurately the control efficiency of
other control option combinations. It is probable that cyclonic
spray scrubbers can also be used. Fugitive dust does not appear
to be a problem in this process because of minimal material
transport. Emission sources and control options are presented
in Table 2-3.
2-23
-------
TABLE 2-3. PARTICULATE EMISSION SOURCES AND CONTROL OPTIONS
FOR MAP PLANTS
Source
Control option
Prill Tower,
Venturi/crossflow,
rotary cooler
venturi/cyclonic spray,
venturi/packed bed
2.2.3 Control Costs
An estimate of the cost of the control method discussed in
the previous subsection is presented in Table 2-4. The cost and
emission estimates are based on a theoretical plant producing
monammonium phosphate at 25 tons/h with a jet nozzle and Prill
Tower. The uncontrolled emission rate is based on data for an
ammonium nitrate Prill Tower, because specific data on mono-
ammonium phosphate are unavailable.
The capital cost is estimated from data filed with the
State of Florida DER. The cost of operation is based on total
system pressure drop and water requirements.
Figure 2-6 presents the control system arrangement for the
control option.
The uncontrolled emission rate is estimated to be 20
lb/ton of product. The controlled emission rate is based on
stack test data filed with the Florida DER. The maximum produc-
tion rate is 25 tons/h at 8000 h/yr.
The recovered product is returned to the process for
recovery. The value of MAP phosphate is $205/ton. If the
recovered product is returned to the process and credited, the
annualized cost is $84/ton recovered.
2-24
-------
TABLE 2-4. CONTROL COSTS OF A TYPICAL MAP PLANT3
Emissions
Cost, $
Cost of
(credit for)
removal,
$/ton
Source of
emissions
Control
options
Control,
%
Uncontrolled,
lb/ton
Controlled,
lb/ton
Removal,
tons/yr
Total
capital
cost, $
Expected
1 ife,
yr
Annual.
capi tal
Annual
OSM
Credits
Prill
tower
cooler
Venturl/
cross-
flow
99-99.4
20
0.12-0.19
1981
409,836
10
66,700
172,599
406,105
(84)
aCosts in January 1980 dollars.
''Based on a 10 percent cost of capital.
-------
PRILL TOWER
»>
COOLER
p*
1200 gal/min
AP = 8 in. H^O
150 gal/min
4 62,000 acfm
120 F
0.006 qr/scf
O"
1/
CR0SSFL0W TAIL
GAS SCRUBBER
AP = 3 in. H^O
Figure 2-6. Process flow diagram of typical MAP plant using
venturi scrubber and crossflow tail ga's scrubber.
-------
2.2.4 Recommended RACT
From the standpoint of achievable emission reduction, a
combination of control methods (venturi scrubbers and crossflow
scrubbers) can achieve emission levels in the range of 0.20
lb/ton of product.
The cost of control is not considered critical because the
range of control options is not wide and the recovered product
has a value of $205/ton.
2.3 GRANULAR TRIPLE SUPER PHOSPHATE
This subsection discusses the processes used in the pro-
duction of GTSP and identifies the major particulate sources
within each facility. The subsection also discusses the avail-
able control technology and cost of a typical plant using this
technology. Based on this analysis, RACT recommendations are
made.
2.3.1 Process Description
Phosphate rock is composed of phosphate in the form of the
mineral fluorapatite [Ca.3 (PO^) 3 • CaF^. The phosphate in this
form is only slightly soluble and thus unsuitable for modern
agriculture. The phosphate is made available as a plant food by
reacting the phosphate rock with phosphoric acid by the following
reaction:
[Ca3 (PC>4) 2 ] 3 * CaF2 + 14 H3P04 + 10 H20
Fluorapatite phosphoric water
acid
-> 10 [CaH4 (P04) 2*H20] + HF
monocalcium phos- hydrogen
phate monohydrate fluoride
2-27
-------
The monocalcium phosphate is commonly referred to as a triple
super phosphate (TSP) and contains between 45 and 49 percent
. The granular form of TSP improves its storage and han-
dling properties. A granule between 1 and 4 mm in diameter is
produced by one of two processes: granulation of ROP-TSP and
direct granulation of TSP slurry.
Basic GTSP Process—
Two methods are used for the direct production of GTSP:
the Dorr-Oliver slurry granulation process and the TVA one-step
granulation process. The TVA process is not used for general
commercial production and is not considered in this study.
Figure 2-7 is a process flow diagram of the Dorr-Oliver process.
In the Dorr-Oliver process, phosphate rock is ground to a
fineness between 7 5 ym and 150 ym and is charged to a reactor
with phosphoric acid (4 0 percent p2°5^ * T^e rock anc^ acid are
reacted for 1 to 2 hours until a slurry is produced. The slurry
is pumped from the reactor tanks and sprayed onto a bed of dry
recycled GTSP fines in a rotary granulator. In the granulator
the slurry builds up on the fines by coating and agglomeration.
In some process variations, pugmills are used instead of
granulators. A pugmill is composed of a V-shaped trough con-
taining twin counterrotating shafts with blades. The shearing,
mixing, and kneading action of the mill agglomerates the slurry
into granules.
2-28
-------
K>
I
M
VD
GROUND
PHOSPHATE ROCK
TO STORAGE
AND SHIPPING
Figure 2-7. Process flow diagram of a Dorr-Oliver slurry granulation plant.
-------
The rotary drum granulator consists of an open-end, rotary
cylinder with retaining rings at each end. The rotary drum
contains a fixed scrapper mounted inside the drum to remove
material from the wall. A bed of recycled GTSP is maintained in
the granulator, and the liquid slurry is introduced through
distributor pipes set under the bed.
Wet granules from the granulator (or pugmill) are discharged
into a rotary direct-fired gas or oil dryer, where the excess
water is evaporated and the chemical reaction of the acid accel-
erated by heat. Dried granules are elevated by bucket elevators
and screened through double-deck screens. The oversized materi-
als are reduced in chain mills, and the fines are returned to
the granulator.
Granules produced by this process are between 1 and 4 mm in
diameter and are cooled in a rotary drum cooler. The product is
then placed in storage for curing. The typical curing period is
between 3 and 5 days. The product is then screened, bagged, and
shipped.
The reactor (pugmill), dryer, cooler, chain mills, screens,
and transfer equipment are ventilated and controlled by simple
cyclones and scrubbers.
GTSP from ROP/TSP—
In this process, ROP/TSP is removed from storage and reduced
in size by pulverizers. The product is screened, and the mate-
rial is introduced into a rotary granulator. Steam and water are
introduced to the granulator to wet the product and aid in
2-30
-------
agglomeration. Granules are dried in a rotary gas- or oil-fired
dryer to produce a hard product. The product is screened, and
the fines recycled to the granulator.
The product is placed in storage for bagging and shipping.
The pulverizers, screens, granulators, and dryer are ventilated,
and particulate emissions are controlled by simple cyclones and
scrubbers.
Review of the processes operated in the central Florida
phosphate producing area indicated six facilities producing GTSP
by the two processes. Emissions and controls for five of these
facilities is given in the Appendix. The production rates of the
lines were between 31 and 72 tons/h.
2.3.2 Emission Sources and Control Options
The seven major points of particulate emissions in a con-
ventional GTSP production plant are listed below:
(1) Reactor
(2) Granulator/blunger
(3) Dryer
(4) Product screens
(5) Cooler
(6) Mills/crushers
(7) Elevators, belt conveyor, etc.
The volume of air required to collect fugitive emissions
from the reactor(s) varies from plant to plant and is included
with other sources in the plant. The primary emissions from the
reactor are fugitive particulates during rock charging and
fluorides released by the reaction. Rates of 4000 acfm at 150°F
2-31
-------
have been indicated at some facilities. Uncontrolled emission
rates of up to 10 lb/ton of product have been estimated.
The granulator/blunger is normally ventilated to reduce or
remove heat and hydrogen fluoride generated in the continued
reaction and granulation process. Typical gas volume of a
42-ton/h product granulator with a recycle ratio of eight to one
is approximately 12,000 acfm at 130°F. The estimated rate of
emissions from the granulator is 21 lb/ton of product. Control
of the reactor and granulator/blunger is typically accomplished
by a venturi scrubber followed by a crossflow or packed bed
scrubber.
The granulated GTSP is dried in a rotary dryer fired with
propane, natural gas, or light fuel oil. The heat input is in
the range of 700,000 Btu/ton of product. The particulates are
exhausted from the dryer and collected for recycle in cyclones.
The exhaust rate at a typical 42-ton/h plant is 46,000 acfm at
220°F. The uncontrolled emission rate, after fines recovery in
the process cyclone, is 16 lb/ton of product. Particulates are
typically controlled by venturi scrubbers followed by packed bed
or crossflow scrubbers.
The product is screened, with oversize crushed and placed
in storage. In some plants a cooler is included before the
product is placed in storage. A typical rate of exhaust from
the cooler is 50,000 acfm. The elevators, conveyors, and
screens are ventilated, and the product is recovered in simple
2-32
-------
cyclones. Exhaust rates from these sources depend on plant size,
layout, and tightness and are typically 20,000 acfm. Uncon-
trolled emissions from the product recovery cyclone serving the
material handling and conveying areas are estimated at .8 lb/ton.
Typical particulate control methods include venturi scrubbers
followed by either packed bed scrubbers or crossflow scrubbers.
Because of the combinations of process vents and control
devices, it is not possible to develop a source-by-source control
device emission characterization; however, a combined plant
emission evaluation has been made. The controlled emission rate
from all sources at six facilities is in the range of 0.07 to
0.37 lb/ton of GTSP produced. The highest rate was at a facility
using packed bed scrubbers (0.37 lb/ton). The lowest rate was at
a facility using a venturi scrubber followed by packed bed
scrubber (0.07 lb/ton). Facilities using various combinations of
venturi, packed bed, and crossflow scrubbers account for the
midrange data.
To ensure that the control equipment emission rate truly
reflects the plant emissions requires complete and efficient cap-
ture of emissions from sources within the plant. The installa-
tion and operation of enclosures, hoods, and ventilation systems
have been demonstrated to reduce fugitive losses effectively
within the manufacturing area. To ensure that these systems are
properly maintained and that emissions do not bypass the control
equipment requires a method of evaluating capture effectiveness.
2-33
-------
It appears that an opacity standard applied to fugitive losses
from the building is an effective method of maintaining control
of fugitive emissions. Plants with good system design and
maintenance ensured by a good O&M plan can reasonably be expected
to maintain fugitive losses from building vents to less than 5
percent opacity. Table 2-5 presents particulate emission sources
and control options for GTSP plants.
2.3.3 Control Costs
An estimate of the cost of the various control options
discussed in the previous subsection is presented in Table 2-6.
The cost and emission estimates are based on a theoretical GTSP
plant producing between 62 and 7 2 tons/h of GTSP by the Dorr-
Oliver Process. The air volumes, temperatures, and uncontrolled
emission rates are based on mass balances provided by the phos-
phate industry. The material of construction is assumed to be
stainless steel. The scrubber pressure drop, liquor rates, and
fan horsepower are typical of plants surveyed. The capital cost
is based on the values reported on permit applications on file
with the State of Florida Department of Environmental Regulation,
Tampa, Florida. The capital costs have been adjusted to January
3
1980 dollars by use of the C-E plant cost index . Because of the
wide range of plant designs, vendor specifications, and equipment
ages, the capital cost may be in error by + 50 percent. The
cited examples are those cases in which process weight, gas
volume, and design are comparable. The examples are also based
on the limited capital data available.
2-34
-------
TABLE 2-5. PARTICULATE EMISSION SOURCES AND CONTROL OPTIONS
FOR GTSP PLANTS
Source
Control option
Reactor,
granulator/blunger
Venturi/crossflow,
venturi/packed bed,
venturi/cyclonic spray/
crossflow
Dryer
Venturi,
venturi/crossflow,
venturi/cyclonic spray/
crossflow
Screens,
cooler,
cage mil 1s,
conveyors,
Venturi,
venturi/crossflow,
venturi/cyclonic spray
crossflow
Storage,
shipping
Venturi,
packed bed
2-35
-------
TABLE 2-6. CONTROL COSTS OF A TYPICAL 6TSP PLANT3
Emissions
Cost, $
Cost of
(credit for)
removal,
$/ton
Source of
emissions
Control
options
Control,
%
Uncontrolled,
1b/ton
Control led,
lb/ton
Removal,
tons/yr
Total
capi tal
cost, $
Expected
1 ife,
y
Annual,
capi tal
Annual
04 M
Credits
Reactors,
granula-
tors, dry-
er, equip-
ment vents
Venturi/
tai 1
gas
99.70
55.7
0.17
13,771a
956,284
10
155,635
336,456
2,065.650
(114)
Reactors,
granula-
tors, dryer
equipment
vents
Venturi/
cyclonic
spray
99.73
55.7
0.15
15,998b
977,272
10
159,051
289,601
2,399,700
(121)
aCosts in January 1980 dollars.
bBased on a 10 percent cost of capital.
cProcess weight of 62 tons/h.
''process weight of 72 tons/h.
-------
Figure 2-8 shows the control system arrangement for the
venturi and tail gas scrubber option. The cost analysis includes
the total capital cost for the venturi and tail gas scrubber.
Utility costs are computed for the total gas volume and resist-
ance. The water flow rate is based on venturi and tail gas
scrubber air flow rates.
The second control system option is based on a system using
a venturi scrubber with a cyclonic spray scrubber for fluoride
control. The utilities are computed assuming a pressure drop of
12 in. H20 and a water rate of 15 gal/1000 acfm. The process
rates of the two systems are not matched, but exhaust gas volumes
are comparable.
The uncontrolled emission rate estimated by PEDCo is based
on an emission factor of 55.7 lb/ton. The controlled emission
rate is based on stack test data on file with the State Agency.
The particulate removal rate is computed based on the
emission factor and a production rate of 62 and 7 2 ton/h for 8000
h/yr.
The recovered product in the scrubbers (venturi) is returned
to the process as makeup to the reactors. The value of GTSP is
$150/ton. If this credit is applied to annualized cost of
operating the systems, the scrubbers have a net profit of $114 to
$121/ton removed. The scrubbers in practice are necessary
process recovery devices that decrease the cost of manufacture.
2-37
-------
4,000oacfm
REACTORS —
14.7 gr/scf
N)
f
U>
00
GRANULATOR
1,200 acfm
130°F
9.6 gr/scf*"
DRYER
62,000 acfm
220 F
1.6 gr/scf
EQUIPMENT VENTS
21,000 acfm
120 F
22% Phosphoric acid,
200 gal/min
^31,700 acfm
3.3 gr/scf
2.1 gr/scf
1000
gal/min
22% phosphoric
acid
POND WATER,
2600 gal/min
O"
86,000 acfm
100°F
0.014 gr/dscf
EFFICIENCY,^99.6%
AP = 16 in. H^O
VENTURI/TAIL
GAS SCRUBBER
TO REACTOR
Figure 2-8. Process flow diagram of a typical GTSP plant using venturi scrubber
and crossflow tail gas scrubber.
-------
2.3.4 Recommended RACT
From the technological standpoint, combinations of control
devices (Venturis, cyclonic spray scrubbers, and tail gas cross-
flow scrubbers) can achieve an emission rate of 0.20 lb/ton of
product. Based on the cost of a venturi/crossflow scrubber and
venturi/spray cyclonic scrubber and the credit for recovered
product, option costs are not significantly different.
To ensure that the measured emission rate from the control
device represents the true emission level requires efficient and
complete capture of the emissions at the source. A fugitive
emission standard is recommended to maintain this level of
control. With the use of well-designed enclosures and hoods and
proper maintenance, the facilities can achieve and maintain a
level of uncontrolled (fugitive) emissions below an observed
opacity of 5 percent at the building vent.
The test method to be used for determining the mass emission
4
standard should be EPA Method 5. The use of in-stack methods
for scrubbers can yield low emission rates, because of the
solubilization of the diammonium phosphate on the filter when
collected below the moisture dewpoint. Opacity should be mea-
sured using EPA Method 9.
2.4 RUN-OF-THE-PILE TRIPLE SUPER PHOSPHATE
This subsection discusses the processes used in the pro-
duction of ROP/TSP and identifies the major particulate sources
within each facility. The subsection also discusses available
2-39
-------
control technology and cost of employing this technology. Based
on this analysis, RACT recommendations are made.
2.4.1 Process Description
Run-of-the-pile triple super phosphate is produced by the
chemical reaction of phosphoric acid with phosphate rock. The
rock is converted from insoluble fluorapatite to soluble mono-
calcium phosphate monohydrate. The reaction is the same as that
shown in Subsection 2.3.1. The differences between the two
methods are the physical appearance of the products and the time
required for completion of the chemical reaction.
In the ROP/TSP process, ground phosphate rock is mixed with
phosphoric acid (50 to 54 percent P2®5^ ;'"n a Pan or cone mixer.
In the cone mixer, the rock is placed in contact with the acid by
tangential flow of acid. The mixing forms a slurry of super
phosphate, which is discharged from the cone into a moving belt.
The slurry on discharge from the mixer begins to become plastic
and solidifies as it moves down the belt. The belt is referred
to as the den, and the super phosphate hardens as it moves down
the belt. The action of the acid on the fluoropatite releases
hydrogen fluoride, and the solidifying mass (matrix) becomes
porous. Some facilities use mixers or pugmills to mix the
solidifying mass, release the trapped gases, and reduce curing
time. The matrix at the end of the solidifying process has the
appearance of a honeycomb.
2-40
-------
The matrix at the end of the curing belt is not completely
cured (i.e., (chemical reactions are not completed). This
material is placed in storage for 3 to 5 weeks in the curing
building. The matrix is broken or cut at discharge from the belt
before complete solidification occurs. Some plants pass the
coarse ROP/TSP through a rotary dryer to increase the chemical
reaction rate before placement of the product in storage. The
heat reduces the curing time required before shipment. The cured
ROP/TSP product commonly is crushed and screened in the storage
area before shipping.
The cone mixer, den, dryer, and curing building are venti-
lated, and particulate and gaseous emissions are collected in
scrubbers. Review of the processes in the central Florida
phosphate producing area indicated six facilities producing
ROP/TSP. The production rates of the lines were between 18 and
48 ton/h. Figure 2-9 is a process flow diagram of a typical
ROP/TSP plant.
2.4.2 Emission Sources and Control Options
The four major points of particulate emissions in a con-
ventional ROP/TSP plant are listed below:
(1) Cone mixer
(2) Den/curing belt
(3) Dryer (optional)
(4) Curing building
The cone mixer represents the point of contact between the
phosphoric acid and ground phosphate rock. The typical gas
2-41
-------
GROUND ROCK
to
I
£>.
NJ
ROCK FCmER
SYSTEM
wn PROCESS
PHOSPHORIC
ACID
ROCK BlrJ
WEIGHER
ACID
r\
CONTROL
rwissions
*
-PRODUCT
Figure 2-9. Process flow diagram of a typical ROP/TSP plant.''
-------
volume exhausted from the cone mixer (pug mill) is not available
because the source exhaust is combined with the curing belt or
den.
The den or curing belt is used to allow the slurry from the
mixer to react and evolve gas. The den is a major source of
particulate and fluoride emissions. The belt is enclosed and
ventilated to remove the particulates and fluorides. Typical gas
rates are 20,000 to 40,000 acfm, but rates are highly variable
depending on den design, tightness, and production rate. The
tightness of the system can be maximized by a good O&M plan. No
independent data are available concerning uncontrolled emission
rates, but if compared with normal superphosphate (NSP) curing
belts, an emission factor of 9 lb/ton is a good estimate. The
options used for control of particulate emissions are cyclonic
spray scrubbers, venturi/crossflow scrubbers, and venturi scrub-
bers. Only limited data are available on the controlled emission
rate from the den (curing belt) because the source is combined
with other sources. The limited data indicate rates of 0.04 to
0.31 lb/ton.
The dryer is employed in some plants to increase the chem-
ical reaction by elevating the product temperature and removing
moisture. The dryer is a significant source of particulate
emissions; rate of flue gases from a dryer are between 25,000 and
30,000 acfm, but vary according to production requirements.
Based on limited data, the control options are cyclonic spray
scrubbers, venturi/crossflow scrubbers, and venturi scrubbers.
2-43
-------
The controlled emission rates at plants in which dryers were
controlled independently were 0.13 to 0.27 lb/ton. The cured
ROP/TSP is transferred to storage and is aged before shipment.
The storage building is typically exhausted to control
fluorides. The ventilation of the building exhausts the sus-
pended particulates, which are generated by material movement
within the building. The data collected do not indicate poten-
tial uncontrolled emissions, and only limited data are available
on controls used. The control options are venturi scrubbers,
cyclonic spray scrubbers, and venturi/packed bed scrubbers. The
controlled emission rate, based on three sources of data, is 0.08
to 0.23 lb/ton, but the effectiveness of the building capture
system is not known. Control can be maximized by a good O&M
plan. The range of data may be the result of a variance in the
capture efficiency of the collection system.
Based on all sources, the controlled emission rate is be-
tween 0.25 and 0.81 lb/ton (assuming all sources are controlled).
To ensure that the control equipment emission rate truly reflects
plant emissions requires complete and efficient capture of emis-
sions from the curing den and dryer. The installation and opera-
tion of hood enclosures and ventilation systems in conjunction
with an O&M plan have been demonstrated to reduce fugitive losses
effectively within the plant. Good maintenance of the hoods and
ductwork is necessary to maintain the level of control and
2-44
-------
minimize fugitive emissions. To ensure the continuous control of
fugitive emissions requires an effective method of evaluating
loss to the atmosphere. An opacity standard applied to fugitive
loss from the building is an effective method of maintaining
control of fugitive emissions. Plants using good capture prac-
tices can reasonably be expected to maintain fugitive losses from
building vents at less than 5 percent opacity.
Table 2-7 lists emission sources and control options for
ROP/TSP plants.
2.4.3 Control Costs
An estimate of the cost of the various control options
discussed in the previous subsection is presented in Table 2-8.
The cost and emission estimates are based on a theoretical plant
producing ROP/TSP at 20 tons/h. The air volume and water flow
rate are typical for plants in the area. The capital cost is
based on data on file as part of permit applications with the
State of Florida Department of Environmental Regulation in
Tampa, Florida. The costs have been adjusted to January 1980
3
dollars by use of the C-E plant cost index. Incomplete data are
available concerning horsepower and liquor flow rates. There-
fore, a pressure drop of 10 in. and liquor-to-gas ratio of 40
gal/1000 acfm have been used.
The recovered product has value as fertilizer material, and
the water is typically returned to the process. In some cases
the scrubbers are used as primary fluoride control, and the
2-45
-------
TABLE 2-7. PARTICULATE EMISSION SOURCES AND
CONTROL OPTIONS FOR ROP/TSP PLANTS
Source
Control options
Mixer,
den (curing belt)
Venturi,
venturi/spray cyclonic,
spray cyclonic
Dryer
Spray cyclonic,
venturi/crossflow
Storage
Venturi/cyclonic spray,
venturi,
cyclonic spray,
impingement scrubber
2-46
-------
TABLE 2-8. CONTROL COSTS OF A TYPICAL ROP/TSP PLANT9
Emissions
Cost, $
Source of
emi ss ions
Control
options
Control
%
Uncontrolled,
lb/ton
Control led,
1b/ton
Removal ,
tons/yr
To tal
capi tal
cost, $
Expected
1 i fe,
y
Annual
capi tal*5
Annual
04 H
Credits
Cost of
removal,
$/ton
Mixer,
pubmi11,
den
Cyclonic
spray
97.6-99.6
9.0C
0.04-0.31
695.2
318,000
10
50,283
134,249
90,350
135
Dryer
Cyclonic
spray
98.7-99.4
21.0d
0.13-0.27
1658
318,000
10
51,798
137,426
215,540
15
aCosts 1n January 1980 dollars. Basis: process weight of 20 tons/h.
''Based on a 10 percent cost of capital.
cUsing emissions factor for NSP production.
^Using emission factor for GTSP dryer.
-------
water is returned to the gypsum ponds. In the case of product
recovery, the annual cost has been credited with the weight of
the recovered products at $130/ton of recovered product.
Data are insufficient to determine adequately the uncon-
trolled emission rate from the dryer (optional). For purposes
of calculation, the emission rate from a GTSP rotary dryer has
been used as an upper limit. A primary cyclone is typically
used ahead of the scrubber to recover product directly into the
process.
The cost of control for a conventional den/pubmill is
estimated to be $135/ton of material removed.
2.4.4 Recommended RACT
From a technological standpoint, the use of venturi scrub-
bers, cyclonic spray scrubbers, and crossflow scrubbers can
achieve emission levels of 0.30 lb/ton of product from the
pugmill and den. The dryer is an additional (optional) source
at most plants. Where used, the dryer will not allow plant
emissions to be reduced to the 0.30-lb/ton level. The emission
level achievable by use of moderate control is 0.27 lb/ton;
however, the application of venturi scrubbers/spray scrubbers
will reduce emissions to 0.10 lb/ton.
It is recommended that allowable emissions from a con-
ventional process be 0.2-5 lb/ton. When a dryer is used, that
level should be increased to 0.35 lb/ton to accommodate the
additional source loading.
2-48
-------
The cost of particulate removal from the den is higher than
the cost of particulate removal from the dryer because of the
lower uncontrolled emission rate and the recovered product
credit. A fugitive emission limit of 5 percent opacity from the
den building is recommended to ensure a high capture efficiency.
The use of well-designed hoods and enclosures is considered
reasonable in achieving this level.
The test method to be used for determining the mass emis-
4
sion standard should be EPA Method 5. The use of m-stack
methods for scrubbers can yield low emission rates because of
the solubilization of the diammonium phosphate on the filter
when collected below the moisture dewpoint. Opacity should be
measured by use of EPA Method 9.
2.5 RUN-OF-THE-PILE NORMAL SUPER PHOSPHATE
This subsection discusses the processes used in the produc-
tion of ROP/NSP and identifies the major particulate sources
within each facility. The subsection also discusses the avail-
able control technology and cost of employing this technology.
Based on this analysis. RACT recommendations are made.
2.5.1 Process Description
Reaction of sulfuric acid (65 percent to 7 5 percent) with
ground phosphate rock (16 to 21 percent P2°5) i-s used to produce
ROP/NSP. The chemical reaction converts the insoluble fluora-
patite to the plant soluble monocalcium phosphate monohydrate.
2-49
-------
The reaction produces calcium sulfate and hydrogen fluoride as
secondary products by the following equation:
[Ca3(P04)2]3 • CaF2 + 7 H2SC>4 + 3 H20
Fluorapatite sulfuric water
(phosphate acid
rock)
¦+ 3[CaH4(PC>4)2 ~ H20] + 7 CaSC>4 + 2 HF
monocalcium calcium hydrogen
phosphate sulfate fluoride
monohydrate (gypsum)
The ground phosphate rock is reacted with the acid in a
cone mixer in which the acid is introduced tangentially. The
super phosphate is discharged from the mixer into a pugmill for
complete mixing of the rock and acid. The mixer may be either
continuous or batch depending on plant design.
The super phosphate slurry is discharged to a den for
curing. Depending on the plant design, the den may be con-
tinuous or batch. In the continuous den, the product is allowed
to react on a slow-moving belt as it is transferred to the
curing building. The movement allows about 1 hour for the
reaction of acid and rock to occur with the release of gases.
The batch den consists of a number of enclosed compartments
in which the product is stored for 1.5 to 10 hours, during which
time the material solidifies. The solid NSP product is removed
from the den and cut or broken before transfer to the curing
area.
2-50
-------
The product is a porous honeycomb in appearance when
solidified and must be cut or broken before being placed in the
curing area. The product is stored in the curing building for
2 to 6 weeks to permit the reaction to go to completion. Fol-
lowing curing, the product is ground and bagged for shipment.
The cone mixer, pugmill, and den are sources of particulate
emissions and are controlled by scrubbers.
Review of the processes operated in the central Florida
phosphate producing area indicated two facilities that have
produced ROP/NSP in recent years. The production rate of the
lines were 13 and 15 tons/h. Figure 2-10 is a process flow
diagram of a typical ROP/NSP plant.
2.5.2 Emission Sources and Control Options
Four major points of particulate emissions in a conven-
tional NSP plant are listed below:
(1) Mixer
(2) Pugmill
(3) Curing belt/den (continuous/batch)
(4) Curing building
The mixer, pugmill, and den are sources of particulate emissions
because of the reaction of sulfuric acid on the phosphate rock.
The reaction generates heat and releases steam and hydrogen
fluoride. The sources are hooded, and the emissions are con-
trolled by a scrubber. The ventilation rate varies with plant
size. Data indicate typical gas rates of 15,000 to 25,000 acfm.
2-51
-------
GROUND
PHOSPHME ROCr.
Figure 2-10. Process flow diagram of a typical ROP/NSP plant.7
-------
Uncontrolled emission rates from these sources are reported
to be 9 lb/ton.^ The accuracy of the value as applied to general
sources is questionable because the range of gas rates and
production rates.
Typical control options consist of venturi scrubbers, wet
impingement scrubbers, and cyclonic spray scrubbers. Based on
State of Florida files and EPA data, 7 controlled emission rates
range between 0.02 and 0.20 lb/ton ROP/NSP.
2.5.3 Control Costs
An estimate of the cost of the various control options
discussed in the previous subsubsection is presented in Table
2-9. The cost and emission estimates are based on a theoretical
plant producing ROP/NSP at 15 ton/h. The capital cost is based
on data on file as part of permit applications with the State of
Florida Department of Environmental Regulation in Tampa, Florida.
The cost has been adjusted to January 1980 dollars by use of C-E
3
plant cost index. Incomplete data are available for fan horse-
power and liquor flow rates. A pressure drop of 10 in. and
liquor-to-gas ratio of 40 gal/1000 acfm have been estimated.
The recovered product has value as fertilizer material, and
the water is typically returned to the process. In some cases
the scrubbers are used for primary fluoride control, and the
water is returned to the gypsum ponds. In the case of product
recovery, the annual cost has been credited with the controlled
weight at $130/ton of recovered product.
2-53
-------
TABLE 2-9. CONTROL COSTS OF A TYPICAL ROP/NSP PLANT3
Emissions
Cost, $
Source of
emissions
Control
options
Control,
%
Uncontrol led,
lb/ ton
Control led,
lb/ton
Removal,
tons/yr
Total
capital
cost, $
Expected
1 ife,
yr
Annual
capital
Annual
04 M
Credits
Cost of
remova1,
$/ton
Mixer,
den
Cyclonic
spray
97.4
9.0
0.15-0.23
526.2
159,000
10
25,874
116,712
68,406
140
Mixer,
den
Impinge-
ment
99.8
9.0
0.01
539.4
c
c
c
c
c
c
tn a Costs 1n January 1980 dollars.
L
Based on a 10 percent cost of capital.
c Insufficient data.
-------
2.5.4 Recommended RACT
The use of cyclonic spray scrubbers and wet impingement
scrubbers can achieve an emission limit of 0.25 lb/ton.
The cost of product recovery is approximately equal to the
value of the recovered product. At a cost of $140/ton recovered,
the cost represents only 0.4 percent of the product value per
year.
A fugitive emission limit of 5 percent opacity from the den
building is recommended to ensure a high capture efficiency. The
use of well-designed hoods and enclosures along with a good O&M
plan are reasonable in achieving this level.
The test method to be used for determining the mass emission
4
standard should be EPA Method 5. The use of in-stack methods
for scrubbers can yield low emission rates, because of the
solubilization of the diammonium phosphate on the filter when
collected below the moisture dewpoint. Opacity should be mea-
5
sured by use of EPA Method 9.
2.6 ANIMAL FEED INGREDIENTS
This subsection discusses the processes used in the produc-
tion of AFI and identifies the major particulate emission sources
within each process. The subsection also discusses the variable
control technology and cost of using the technology. Based on
this analysis, RACT recommendations are made.
2-55
-------
The term "animal feed ingredients" refers to a number of
calcium phosphate compounds produced by use of defluorinated
phosphoric acid or wet process phosphoric acid.
Two companies manufacture these products in the Tampa area.
Because each company considers the process technology to be
confidential, limited process data are available; however, based
on general discussions and information available through the
Florida Department of Environmental Resources records, the
processes have been classified into three major categories. The
first category involves the production of granular monoammonium
and diammonium phosphate with defluorinated phosphoric acid. The
second category involves the production of calcium phosphates
with defluorinated phosphoric acid and limestone. The third
category involves the production of calcium phosphates with wet
process phosphoric acid from phosphate rock with product de-
fluorination after reaction. Although each process is described
individually, the first two processes can be accomplished on the
same process line.
2.6.1 AFI ammonium phosphate
Process Description—
Monoammonium and diammonium phosphate—The production of
monoammonium and diammonium phosphate AFI is similar to the
production of granular ammonium phosphates. Figure 2-11 is a
process flow diagram of a typical TVA ammonium phosphate AFI
plant. The ammonium phosphate slurry is produced by the re-
action of defluorinated phosphoric acid with ammonia in an
2-56
-------
Figure 2-11. Process flow diagram of a typical TVA AFI, ammonium phosphate plant.^
-------
aggitated reactor. The slurry is pumped into a rotary granulator
in which granules are formed by coating and agglomerating on
recycled product fines. The chemical composition of the product
is controlled by the temperature and rate of ammonia and acid
injection into the granulator. The product is discharged to a
rotary dryer, which reduces the moisture content to less than 2
percent.
The product is elevated by bucket elevators to a series of
vibrating screens. The oversize material is ground with cage
mills, and the fines are returned to the granulator. The
screened product is cooled in a rotary cooler and rescreened for
shipment.
The granulator, reactor, dryer, cooler, cage mills, screens,
and conveying equipment are exhausted to capture particulate
emissions and are controlled by simple cyclones followed by
scrubbers.
Mono calcium and dicalcium phosphate--Defluorinated phos-
phoric acid can be used to produce AFI consisting of dicalcium
and monocalcium phosphates in granular form. Figure 2-12 is a
process flow diagram of a typical calcium phosphate AFI plant.
The process involves the reaction of limestone and defluorinated
phosphoric acid in a pugmill to form a slurry. The slurry is
transferred to the rotary dryer and processed in a similar
manner to the ammonium phosphate products with fines recycled
and returned to the pugmill.
2-58
-------
Figure 2-12. Process flow diagram of a typical AFI calcium phosphate plant.'7
-------
The pugmill, dryer, screens, cage mills, cooler, and trans-
port equipment are exhausted, and particulate emissions are
controlled by simple cyclones followed by scrubbers.
Emission Sources and Control Options--
One plant produces monoammonium and diammonium phosphate
AFI from defluorinated phosphoric acid. The major emission
sources are listed below:
(1) Reactor
(2) Granulator (pugmill)
(3) Dryer
(4) Screens and cage mills
(5) Cooler
(6) Bucket elevator, conveyors, etc.
The general process flow of the product is similar in de-
sign to that at a TVA diammonium phosphate plant except that the
finished granules are larger.
The reactor is a source of fluorides and ammonia and is
typically vented with the ammonia granulator gases. The exhaust
rate from the reactor is about 3000 acfm at 220°F. The exhaust
from the granulator (or pugmill) is about 30,000 acfm at 180°F.
The particulate loading is 27 lb/ton, and the ammonia loading is
30 lb/ton, based on conventional DAP plant estimates. A typical
control method used for this emission source is a venturi
scrubber followed by a crossflow scrubber. The system operates
at a pressure drop of 14 in. and liquor-to-gas ratio of 20
gal/1000 acfm; phosphoric acid is used as the scrubbing liquor.
2-60
-------
The rotary dryer is a major source of particulate emis-
sions. The dryer is fired with fuel oil at a heat input of
960,000 Btu/ton of product. The typical dryer exhaust rate is
82,000 acfm at 250°F. The uncontrolled particulate emission
rate, after the reclaim cyclone, is estimated at 32 lb/ton
(based on conventional DAP plant estimates). Emissions are
controlled by a venturi scrubber followed by a crossflow scrub-
ber. The scrubber operates at a pressure drop of 14 in. H^O and
liquor-to-gas ratio of 10 gal/1000 acfm.
The cooler, screens, cage mills, and product handling
equipment are vented to a product recovery cyclone and then to
a venturi scrubber, followed by a crossflow scrubber. The
typical gas rate is 71,500 acfm at 160°F (the cooler rate is
44,500 acfm). The venturi has a pressure drop of 14 in. 1^0 and
liquor-to-gas ratio of 11 gal/1000 acfm. The emission rate from
the combined scrubber exhaust has been reported as 0.28 lb/ton.
Control Costs—
An estimate of the cost of the various control options
discussed in the previous subsection is presented in Table
2-10. The cost and emission estimates are based on a theo-
retical plant producing granular AFI at 120 ton/h. The air
volume and water flow rate are considered typical-of plants in
the area. The capital cost is based on data on file as part of
permit applications with the State of Florida Department of
2-61
-------
TABLE 2-10. CONTROL COSTS OF A TYPICAL AFI AMMONIUM PHOSPHATE PLANT3
Emissions
Cost, $
Source of
emissions
Control
options
Control
%
Uncontrolled,
1b/ton
Controlled,
lb/ ton
Removal,
tons/yr
Total
capi tal
cost, %
Expected
11 fe,
yr
Annual
capi talb
Annual
04M
Credits
Cost of
(credit for)
removal,
$/ ton
Reactor,
granulator,
pug mi 11,
dryer,
cooler,
mi 1 Is,
screens,
transfer
Venturi/
cross-
flow
99.6
85
0.28
30,500C
!, 500,000
10
406,875
626,737
5,490,000C
(146)
aCosts 1n January 1900 dollars.
bBased on a 10 percent cost of capital.
cAverage production of 90 ton/h at 8000 h/yr.
^Recovered product returned to process at a value of $180/ton.
-------
Environmental Regulation in Tampa, Florida. The cost has been
adjusted to January 1980 dollars by use of the C-E plant cost
index.^
Figure 2-13 represents the control system arrangement used
at one facility surveyed. Information on specifics at the
second facility were not available. The annualized cost is
computed based on the total system gas volume, water rate and
static pressure drop.
Because of the proprietary nature of the AFI processes and
the lack of field data, uncontrolled emission factors are
unavailable; however, the basic process used is similar to a
diammonium phosphate plant. For the purposes of emission
calculations, the emission factor is assumed to be 85 lb/ton and
that the recovered product has a value of $180 ton.
As with other ammonium phosphate and calcium phosphate
processes, the primary scrubbers are used as process recovery
devices, with the acid slurry being returned to the reactor or
granulator. A net annualized credit is estimated to be $146/ton
removed.
Recommended RACT--
Limited data on this process indicate that the use of
venturi scrubber and crossflow scrubbers can achieve an emission
limit of 0.30 lb/ton of product. The cost of control is not a
critical factor since the application of credit for product
recovered results in a recovered value greater than the an-
nualized cost.
2-63
-------
4130,000 acfm
120 F
EFFICIENCY99.8%
INPUT, 25.18 ar/scf
OUTPUT, 0.0354 ar/scf
Figure 2-13. Process flow diagram of a typical AFI (monoammonium/diammonium phosphate)
granulation plant using venturi scrubbing and crossflow tail gas scrubbing.
-------
To assure that the emissions measured by the stack test
represent the controlled emission level, a fugitive emission
limit applied to the plant process building in conjunction with
an O&M plan is recommended. The fugitive emissions are judged
based on the opacity of material escaping the building. This
level is recommended to be 5 percent opacity; based on good
maintenance and good hood capture, the level should be easily
achieved.
The test method to be used for determining the mass emission
4
standard should be EPA Method 5. The use of m-stack methods
for scrubbers can yield low emission rates, because of the
solubilization of the diammonium phosphate on the filter when
collected below the moisture dewpoint. Opacity should be mea-
sured by use of EPA Method 9.^
2.6.2 Calcium Phosphate from Phosphate Rock
Process Description—
The production of calcium phosphate AFI from phosphate rock
involves the defluorination of the rock after formation of the
complex.
The general process consists of the reaction of phosphate
rock and wet process phosphoric acid (containing fluorides) and
the aging of the product. The product is mixed with several
ingredients (salts, composition not provided) and calcined at
temperatures in excess of 2000°F in rotary kilns on fluid bed
reactors to remove the fluorides. The calcination changes the
2-65
-------
chemical structure of the calcium phosphate making it available
for animal assimilation. The calcined product is cooled in a
rotary cooler and screened and crushed for shipment.
Variations of the process can involve the further reaction
of the calcined product with defluorinated phosphoric acid and
limestone to increase the grade of the product.
Particulate emissions occur from the calcining kilns and
reactors, product cooling, and product transfer and storage.
Particulate emissions from these sources are typically con-
trolled by simple cyclones and scrubbers. Figure 2-14 is a
process flow diagram for the production of AFI using the wet
process phosphoric acid process with fluorine removal by cal-
cining .
Emission Sources and Control Options—
There are two plants which produce AFI from phosphate rock.
The major source of particulate emissions from these processes
is the defluorination kilns or reactors (fluid bed). The
general process involves the calcining of the prepared animal
feed ingredient in a rotary kiln. Data are limited on the
specific emission or process conditions of these sources since
both companies consider the process technology confidential.
Limited data, however, indicate that the typical kiln has an
exhaust volume between 40,000-45,000 acfm at 1200°F. The
uncontrolled emission rate is estimated at 5.5 gr/dscf (60
lb/ton). Typical firing rates are 484 x 106 Btu/h using re-
sidual oil. The emissions are typically controlled by packed
2-66
-------
Figure 2-14. Process flow diagram of AFI production by the
wet phosphoric acid process with fluorine removal by calcining.
-------
bed or venturi scrubbers. Few data are available concerning the
operating conditions of these scrubbers. Based on limited data,
the crossflow scrubber has been estimated to be operated at a
liquor-to-gas ratio of 55 gal/1000 acfm (2200 gpm) with the
major portion of the water being used to quench the kiln gases
before entering the scrubber.
Reported emission rates for the systems for the venturi are
0.63 lb/ton, while those for the crossflow scrubber are 0.23
lb/ton.
Control Costs—
An estimate of the cost of the various control options
discussed in the previous subsection is presented in Table
2-11. The cost and emission estimates are based on a theoretical
plant producing 14 ton/h of calcined AFI. The air volume and
water flow rate are considered typical of plants in the area.
The capital cost is based on data on file as part of permit
applications with the State of Florida Department of Environ-
mental Regulation in Tampa, Florida. The cost has been adjusted
3
to January 1980 dollars by use of the C-E plant cost index.
The uncontrolled particulate emission rate is estimated by
the industry to be 5.5 gr/dscf (60 lb/ton). If the calcine
particulate is returned to the process, the estimated credit for
removal is $4 0/ton removed; however, if the material is not
recoverable, the cost of removal is $139/ton removed. It is
2-68
-------
TABLE 2-11. CONTROL COSTS OF A TYPICAL AFI CALCIUM PHOSPHATE PLANT3
Emissions
Cost, $
Source of
emissions
Control
options
Control
%
Uncontrolled,
lb/ton
Control 1ed,
1b/ton
Removal,
tons/yr
Total
capi tal
cost, $
Expected
1 i fe,
yr
Annual
capi talb
Annual
0SM
Credlts
Cost of
(credit for)
remova 1,
S/ton
Calciner
Cross-
f lowc
99.9
60
0.23
2390
833,300
10
135,554
198,089
430,200
I39d or (40)e
Calclner
Venturl/
cyclonic
spray
99.9
60
0.63
2375
f
f
f
f
f
f
(Ti
aCosts in January 1980 dollars. Basis: process weight of 10 tons/h.
''Based on a 10 percent cost of capital.
cGas flow rate of 12,700 dscfm and 200-hp fan at water rate of 2200 gal/min.
^ 1 f collected material cannot be returned to process.
Collected material returned to process at $180/ton.
fInsufficient data.
-------
probable that the primary cyclone catch may be returned and that
the slurry may not be returned. In such case, an overall credit
of $ll/ton removed results.
Recommended RACT--
Limited data on this process indicate that the use of
venturi scrubbers and crossflow scrubbers can achieve an emis-
sion limit of 0.25 lb/ton of product.
The cost of control is assumed not to be a factor since the
recovered product is returned to the process and the credit
exceeds the annualized control cost.
To assure that the emissions measured by the stack test
represent the controlled emission level, a fugitive emission
limit is recommended for the plant process building. The
fugitive emissions are judged based on the opacity of material
escaping the building. This level is recommended to be 5
percent opacity; based on good maintenance and good hood capture,
the level should be easily achieved.
The test method to be used for determining the mass emis-
4
sion standard should be EPA Method 5. The use of in-stack
methods for scrubbers can yield low emission rates because of
the solubilization of the diammonium phosphate on the filter
when collected below the moisture dewpoint. Opacity should be
measured by use of EPA Method 9.^
2-70
-------
2.7 PHOSPHATE ROCK DRYERS
This subsection discusses the processes used in the drying
of phosphate rock and identifies the major particulate sources
within each facility. The subsection also discusses the avail-
able control technology and cost of a typical plant using this
technology. From this analysis RACT recommendations are made.
2.7.1 Process Description
The phosphate rock industry consists of rock processing
operations located near ore reserves and mining operations.
Nearly three-quarters of the national production of phosphate
rock occurs in Florida; this production is used primarily for
phosphate fertilizer. About 20 percent is converted to animal
feed ingredients, and 30 percent is exported for further proces-
sing.
The constituent of the rock that is of economic interest is
tricalcium phosphate [Ca^ (P°4)2]* commonly known in the indus-
try as bone phosphate of lime (BPL). Phosphate rock and prod-
ucts are typically graded on BPL content (e.g., 68 BPL rock
contains 68 percent by weight of BPL or tricalcium phosphate).
Final products usually contain 68 to 74 percent tricalcium
phosphate.
Fluoride-bearing material is another constituent in phos-
phate rock accounting for 4 to 5 percent by weight. The basic
structure of the fluoride ingredient is represented as fluor-
apatite (3 Ca3 (p04)2 * Ca2F^* Most phosphate ores contain a
substituted form of this structure, with fluoride and carbonate
2-71
-------
replacing some of the phosphate. Commercial rock contains 30 to
38 percent P2°5' trace amounts of iron, aluminum, magnesium,
silica, sodium, potassium, carbon dioxide and sulfates.
Florida phosphate rock deposits consist of a consolidated
mass of phosphate pebbles and clays, known as matrix. The high
clay content in Florida rock distinguishes it from other
regional phosphate deposits. Figure 2-15 is a general flow
diagram for processing of Florida phosphate rock. After the
rock is mined, two intermediate-product types, pebble and
concentrate, are produced, depending on the extent of bene-
fication. Pebble is produced after the mined ore is washed;
concentrate is the product after the ore is crushed and washed.
Final product usage determines whether the ore is processed as
pebble or concentrate. The distinction between pebble and
concentrate is made in this text because of the higher emission
rates associated with the handling and drying of pebble.
Another characteristic of Florida phosphate deposits is low
organic content. Because most Florida rock is relatively free
of organic material, calcining is not needed after benefication.
Rock reserves containing organic materials require heating up to
1600°F to volatilize organics prior to further processing for
fertilizer manufacturer.
Phosphate rock drying is accomplished in direct-fired
dryers by heating the ore to 250°F for free water evaporation.
Figure 2-16 is a process flow diagram of typical rock drying
plant. Wet phosphate ore is conveyed and charged to the dryer
2-72
-------
TO CONTROL EQUIPMENT
Figure 2-15. Process flow diagram of Florida phosphate rock production.
-------
WET PHOSPHATE ROCK
FUEL
fa:
a:
ID
SCREENS
WTr7jRAI1RnAn CARS
oo oo
Figure 2-16. Process flow diagram of a typical phosphate rock drying plant.
-------
through a feed bin or feed chute. Petroleum-based fuels (natural
gas, No. 2 and No. 6 fuel oil) are used for heating and drying
the ore. The ore is discharged when the moisture content of the
phosphate material is reduced from 14 percent down to 1 to 3
percent, depending on product use. There are two dryer designs
utilized for this process: rotary dryers and fluidized-bed
dryers. Figures 2-17 and 2-18 show typical schematics of these
two types. After drying, the phosphate material is belt-conveyed
to a screen house, where it is classified by size and transferred
to storage or to a grinding process. The exhaust gas from the
dryer contains entrained ore particles, combustion products, and
moisture from ore drying. The dryer gas effluent is first con-
trolled by a cyclone, where product recovery occurs, and then by
a final emission control device.
Inventory of Florida state files indicated the capacities of
dryers ranged from 5 to 500 tons/h with equal usage of both dryer
designs. Based on state file records, an average dryer capacity
is approximately 200 ton/h.
2.7.2 Emission Sources and Control Options
There are two major points of particulate emissions in a
conventional phosphate rock dryer process plant:
1) Dryer
2) Dried materials handling and storage, including
screen house
The characteristics of the exhaust gas from conventional
dryers are indicated in Table 2-12. Emissions from rock dryers
are dependent on several factors, including rock type (pebble or
2-75
-------
Figure 2-17. Direct-fired, cocurrent, rotary dryer.
-------
4 TO SECONDARY
CONTROL DEVICE
Figure 2-18. Fluidized-bed dryer.
-------
TABLE 2-12. CHARACTERISTICS OF EXHAUST GAS FROM
ROTARY AND FLU IDIZED-BED DRYERS
Exhaust flow rate
0.13-0.23 m^/s per Mg product/h
(250-450 scfm per ton product/h)
Temperature
394-422 K (250°-300°F)
Moisture
8-30%
Uncontrolled mass
emissions
2-9 g/kg product (4-18 lb/ton product)
Grain loading
7-11 g/dry (3-5 gr/dscf)
Particle size distri-
bution
98% <10 ym
92.9% < 5 ym
73.8% < 2"urn
39.9% < 1 ym
7.2% < 0.5 ym
2-78
-------
concentrate), fuel type, air flow rate, product moisture content,
and speed of rotation for the case of rotary dryers. The gas
rates reported for individual driers range from 4000 to 160,000
scfm, corresponding to process rates ranging from 5 to 480 ton/h.
For the considered representative case processing 200 ton/h of
dried rock, it is estimated that 80,000 scfm of gas is exhausted
from a reasonably operated and controlled source.
For all rock drying facilities processing more than 80 ton/h
of product, scrubber systems were used for particulate control.
A summary of control options used in the Central Florida area is
presented in Table 2-13. Venturi scrubbers were most commonly
used, followed by cyclonic scrubbers and then impingement-type
scrubbers. Following most scrubber units were cyclonic-type mist
eliminators to collect and remove entrained droplets. Typical
scrubber system pressure drops ranged from 10 to 25 in. 1^0, and
liquid-to-gas ratios ranged from 5 to 15 gal/1000 acfm.
A review of stack test results from several dryer facilities
indicated emission rates of 0.01 to 0.11 lb/ton of dried product.
Grouping of emission rates indicated that each control system
type achieved emission rates less than 0.10 lb/ton on a con-
sistent basis. Achievement of emission rates in the 0.03-0.04
lb/ton range was not uncommon.
Fugitive emissions from dried rock materials handling need
to be controlled. An O&M plan is also necessary. Due to the
dampness of feed material to the dryer, there are no fugitive
emissions associated with dryer charging. Rock discharged from
2-79
-------
TABLE 2-13. PARTICULATE EMISSION SOURCES, CONTROL OPTIONS,
AND ACHIEVABLE EMISSION RATES FOR PHOSPHATE
ROCK DRYERS
Source
Control options
Achievable emission
rate, lb/ton
Dryer
Venturi scrubber
0.01-0.10
Dryer
Cyclonic scrubber
0.01-0.11
Dryer
Impingement scrubber
0.034
Materials handling®
Venturi scrubber
0.0025
Materials handling
Cyclonic scrubber
0.00053-0.0014
Materials handling3
Impingement scrubber
0.00031
Materials handling3
Pulse fabric filter
0.0008-0.026
aIncludes conveyors, screens, storage.
2-80
-------
the dryers is usually conveyed to storage silos on weather-
protected conveyors. From the silos, rock is either transported
to consumers in rail cars and trucks or conveyed to grinding
mills. Provision must be made to vent the conveying airstreams
to and from the silos for fugitive emission control. Potential
emissions from typical materials handling and storage systems are
estimated at 2 lb/ton of rock handled.
Emission control options in actual use for dried rock
handling and storage include both wet scrubbers and baghouses.
Three designs of wet scrubbing systems were located in the state
file search, and are: venturi scrubbers, impingement scrubbers,
and cyclonic scrubbers. Pulse-cleaned baghouses were the only
identifiable type of baghouse application on control of mate-
rials-handling. As shown in Table 2-13, all four options showed
control levels less than 0.01 lb/ton, with achievable emission
rates less than 0.001 lb/ton not being uncommon.
To ensure that the controlled emission rate truly reflects
plant emissions requires complete and efficient capture of
emissions from sources within the plant. An O&M plan is also
necessary. The installation and operation of enclosures, hoods
and ventilation systems have been demonstrated to reduce fugitive
losses effectively within the manufacturing plant. The main-
tenance of these systems is a major problem in maintaining
effective control of plant emissions. To ensure good operating
practices requires an effective method of evaluating loss to the
ambient atmosphere. An opacity standard applied to fugitive loss
2-81
-------
from the building is an effective method of maintaining control
of fugitive emissions. Plants using good capture practices can
reasonably be expected to maintain fugitive losses from building
vents less than 5 percent opacity.
2.7.3 Control Costs
Cost estimates of the various control options presented in
the previous subsection are presented in Table 2-14. Cost and
emission estimates are based on a representative plant drying 200
ton/h of phosphate rock material. The process and control device
characteristics incorporated into the costing estimates were
derived from several existing dryers and control devices in the
central Florida area. A search of state file records and plant
visits were conducted to obtain the values used in cost estimates.
Capital cost figures were based on values reported on permit
applications on file with the State of Florida Department of
Environmental Regulation and were adjusted to January 198 0
3
dollars by use of the C-E Plant Cost Index. Due to a wide range
in plant designs, control device vendors, and the degree of
safety factor and redundancy, it is possible that capital cost
estimates may vary by + 50 percent. The cited examples represent
cases in which data was available and typical of current tech-
nology. Cost of removal per ton of product was determined by the
ratio of total annualized cost and the estimated tonnage of
emissions captured.
Cost estimate levels in Table 2-14 reveal small cost dif-
ferences between the various control options. Apparent cost
2-82
-------
TABLE 2-14. CONTROL COSTS OF PHOSPHATE ROCK DRYERS3
Emi sslons
Total
capi tal
cost, $
Expected
1 ife,
yr
Cost, t
Cost of
removal,
S/ton
Source of
emissions
Control
options
Control,
T
Uncontrolled,
lb/ton
Control led,
lb/ ton
Removal,
tons/yr
Annual,
capital
Annual
OiM
Credits
Dryer
Venturl
scrubber
99.5
11
0.05
8756
244,000
10
39,710
237,940
c
31
Dryer
Cyclonic
scrubber
99.5
11
0.05
8756
263,000
10
42,800
238,700
c
32
Dryer
Impingement
scrubber
99.5
11
0.05
8756
459,000
10
74,700
221,300
c
34
Materials.
handl1ng
Venturl
scrubber
99.75
2.0
0.005
1596
693,000
10
112,800
214,790
c
205
Mater1alsd
handl1ng
Cyclonic
scrubber
99.75
2.0
0.005
1596
224,000
10
36,500
195,880
c
146
HaterlalSj
handl1ng
Impingement
scrubber
99.75
2.0
0.005
1596
390,000
10
63,500
202,530
c
146
Materials^
handl1ng
Fabric
filter
99.75
2.0
0.005
1596
560,000
10
91,130
155,870
c
155
®Costs In January 1980 dollars.
''Based on a 10 percent cost of capital.
cData not available.
''Associated with dryer.
-------
advantages for a specific control option may not necessarily
represent actual cost advantages. As stated earlier, cost
estimate levels may vary + 50 percent, depending on several
factors. Site-specific criteria or vendor-specific costing may
override the marginal cost difference listed.
2.7.4 Recommended RACT
From the standpoint of technological feasibility, each
control option is capable of attaining emission levels in the
range of 0.10 lb/ton of product.
The cost of control is not a major factor since almost all
plants surveyed are controlled by a control method capable of
achieving the recommended emission level.
To ensure that the emissions measured by the stack test
represent the controlled emission level, a fugitive emission
limit applied to the plant process building in conjunction with
an O&M plan is recommended. The fugitive emissions must be
judged on the opacity of material escaping the building. This
level is recommended to be 5 percent opacity; based on good
maintenance and good hood capture, the level should be easily
achieved.
The test method to be used for determining the mass emission
4
standard should be EPA Method 5. The use of m-stack methods
for scrubbers can yield low emission rates, because of potential
solubilization of phosphate material on the filter when collected
below the moisture dewpoint. Opacity should be measured by use
of EPA Method 9.5
2-84
-------
2.8 PHOSPHATE ROCK GRINDING
This subsection discusses the processes used in the grinding
of phosphate rock and identifies the major particulate sources
within each facility. The subsection also discusses the avail-
able control technology and the cost of a typical plant using
this technology. From this analysis RACT recommendations are
made.
2.8.1 Process Description
Grinding is widely used in the processing of phosphate rock
to pulverize the product from the. rock drying process. Figure
2-19 shows a typical grinding circuit. Roller mills or ball
mills are typically used to pulverize the dried product to a fine
powder, usually specified as 60 percent by weight passing a 200-
mesh sieve.
The roller mill is composed of hardened steel rollers that
rotate against the inside of a steel ring, as shown in Figure
2-20. Ore is fed into the mill housing by a rotary valve that
prevents the escape of air into the feed system. The rock is
scooped up from the floor of the housing by plows and directed
into the path of the rollers, where it is ground between the
rollers and the steel ring. Ground rock is swept from the mill
by a circulating airstream. Some product size classification is
provided by the "revolving whizzers" at the top of the housing.
The size of an average particle leaving the mill can be con-
trolled by varying the speed of revolution of the whizzers.
Further size segregation is provided by the air classifier,
2-85
-------
Figure 2-19. Typical phosphate rock grinding circuit.
-------
^ PRODUCT OUTLET
I
I
REVOLVING
WHIZZERS
- WHIZZER
DRIVE
GRINDING RING
¦GRINDING ROLLER
- FEEDER
Figure 2-20. Roller mill used to grind phosphate rock.
2-87
-------
which separates oversize particles from product-size particles
and recycles the oversize portion to the mill. The product is
separated from the carrying air stream by a cyclone and conveyed
to ground-rock storage. The airstream is returned to the mill
in a closed loop, although there is a bleed stream from the
system as described below.
The ball mill is basically a drum revolving about an axis
slightly inclined to the horizontal (Figure 2-21). The drum
contains a large number of steel balls about 1 inch in diameter.
Rock is charged into the mill through a rotary valve, ground by
attrition with the balls, and swept from the mill by a circula-
ting air stream, as described above for roller mills.
Roller and ball mills are operated slightly below atmos-
pheric pressure to avoid the discharge of fugitive rock dust
into the air. As a result, atmospheric air infiltrates the
circulating streams. This tramp air is discharged from the
circuit through a dust collector to the atmosphere. Mill
capacities range from 15 tons/h of phosphate rock for a smaller
roller mill to about 260 tons/h for a large ball mill. A
typical mill has a capacity of 50 tons/h. Because roller mills
are usually limited to about 75 tons/h per unit, many operators
install several in parallel rather than a single large ball
mill. No clear trend toward either method of grinding is
evident. The volume of the tramp air discharge stream depends
more on the design and operation of the grinding circuit than on
2-88
-------
to
\
CP
v£>
*\
-------
the capacity of the mill. For example, it is not unusual for a
150-ton/h mill to discharge 19,000 dscfm, whereas a 250-ton/h
unit might discharge 10,000 dscfm.
2.8.2 Emission Sources and Control Options
There are two major points of particulate emissions in a
conventional rock grinding plant:
(1) Grinder
(2) Materials handling and storage
The characteristics of the exhaust gas from conventional
grinders are given in Table 2-15. The grinder operates under a
slightly negative pressure to minimize escaping gases containing
ground rock dust. The system is not airtight, indicating that
the drawn air must be vented and that the amount of air can vary
with design and operation criteria. Grinder exhaust rates range
from 1500 to 35,000 dscfm depending on process rate, and spe-
cific control rates can range from 60 to 160 dscfm per ton/h of
ground rock. Grinding operations are entirely mechanical, and
generate 2 to 5 grains/dscf particulate matter. For the con-
sidered representative case grinding 50 ton/h, it is estimated
that 7000 dscfm of gas volume is exhausted for a reasonable
operated and controlled source.
State records on phosphate rock grinders in the central
Florida area indicated process rates ranging 12 to 230 tons/h.
The vast majority of these grinders used fabric filter systems
to control particulate emissions. Several fabric filter vendors
were represented, but most fabric filters were pulse-cleaned
2-90
-------
TABLE 2-15. CHARACTERISTICS OF EXHAUST GASES FROM
PHOSPHATE ROCK GRINDERS
Exhaust flow rate
•3
31-83 m /s per kg product/h
(60-160 scfm per ton product/h)
Temperature
310-339 K (100°-150°F)
Moisture
Up to 9%
Uncontrolled emissions
<3.5 kg/Mg product (<7.0 lb/ton product)
Grain loading
7-11 g/dry (3-5 gr/dscf)
Dust composition
Calcium (CaO)
45.5% by weight
Phosphorous (P2^5)
32.5% by weight
Silica (Si02)
11.0% by weight
Alluminum (Al^O^)
2.0% by weight
Iron (Fe203)
0.8% by weight
Magnesium (MgO)
0.7% by weight
Other
7.5% by weight
2-91
-------
units. A minority of grinders were controlled by scrubbers;
Venturis were the only identifiable design type.
A review of stack test results from several grinding
facilities indicated emission rates of 0.003 to 0.30 lb/ton of
grpund rock. Grouping of emission rates by control system type
indicated that each type achieved emission rates less than 0.10
lb/ton; however, fabric filter control systems achieved rates
less than 0.10 lb/ton more consistently and with greater margins
(down to 0.003 lb/ton) and ease. From these results, fabric
filter control of grinder exhaust is indicated to be the pre-
ferred method of particulate control.
Fugitive emissions from ground rock materials handling need
to be controlled. Rock fed and discharged from the grinders is
usually conveyed to storage silos on weather-protected con-
veyors. From the silos, rock is either transported to consumers
in rail cars and trucks. Provision must be made to vent the
conveying assistance to and from the silos for fugitive emission
control. Potential emissions from typical materials handling
and storage systems are estimated at 2 lb/ton of rock handled.
Practiced emission control options for ground rock handling
and storage are limited to fabric filters. Pulse-cleaned fabric
filters were the only identifiable types of fabric filters used
to control emissions during materials handling. As shown in
Table 2-16, fabric filter control levels were less than 0.01
lb/ton, with achievable emission rates less than 0.004 lb/ton.
2-92
-------
TABLE 2-16. PARTICULATE EMISSION SOURCES, CONTROL OPTIONS, AND
ACHIEVABLE EMISSION RATES FOR PHOSPHATE ROCK GRINDING PLANTS
Source
Control options
Achievable emission
rate, lb/ton
Grinder
Pulse fabric filter
0.003-0.69
Grinder
Venturi scrubber
0.070-2.0
Materials handling
Pulse fabric filter
0.004-.0125
(i.e., conveyors,
screens, storage)
2-93
-------
To ensure that the controlled emission rate truly reflects
plant emissions requires complete and efficient capture of emis-
sions from sources within the plant. The installation and opera-
tion of enclosures, hoods and ventilation systems in conjunction
with a good O&M plan can effectively reduce fugitive losses
within the manufacturing plant. The maintenance of these systems
is a major problem in maintaining effective control of plant
emissions. To ensure good operating practices requires an effec-
tive method of evaluating loss to the ambient atmosphere. It
appears that an opacity standard applied to fugitive loss from
the building is an effective method of maintaining control of
fugitive emissions. Plants using good capture practices can
reasonably be expected to maintain fugitive losses from building
vents at less than 5 percent opacity.
2.8.3 Control Costs
Cost estimates of the various control options presented in
the previous subsection are presented in Table 2-17. Cost and
emission estimates are based on a representative plant grinding
phosphate rock material at 50 tons/h. The process and control
device characteristics incorporated into the costing estimates
were derived from several grinders and control device cases in
the central Florida area. State file records were reviewed, and
plant visits were conducted to obtain the specified values used
in cost estimates. Capital cost figures were based on values
reported on permit applications on file with the State of
Florida Department of Environmental Regulation and were adjusted
3
to January 1980 dollars by use of the C-E Plant Cost Index.
2-94
-------
TABLE 2-17. CONTROL COSTS OF PHOSPHATE ROCK GRINDING3
Emissions
Cost, $
Source of
emissions
Control
options
Control,
%
Uncontrolled,
1b/ton
Control led,
1b/ton
Removal,
tons/yr
Total
capital
cost, $
Expected
1 ife,
y
Annual,
capi tal
Annual
OSM
Credits
Cost of
removal,
$/ton
Rock
grinder
Venturl
99.0
5.0
0.05
990
22,500
10
3,660
105,330
c
110
Rock
grinder
Pulse
fabric
filter
99.0
5.0
0.05
990
58,000
10
9,440
102,380
c
113
Grinding
MHS
Pulse
fabric
fi1ter
99.75
2.0
0.005
399
82,400
10
13,400
104,930
c
297
aCosts In January 1980 dollars.
''Based on a 10 percent cost of capital.
cOata not available.
-------
Due to a wide range in plant designs, control device vendors,
and the degree of safety factor and redundancy, it is possible
that capital cost estimates may vary by + 50 percent. The cited
examples represent cases in which data were available and
typical of current technology. Cost of removal per ton of
product values was determined by the ratio of total annualized
cost and the estimated tonnage of emissions captured.
Cost estimate levels in Table 2-17 reveal small cost
differences between the various control options. Apparent cost
advantages for a specific control option may not necessarily
represent actual cost advantages since cost estimate levels may
vary + 50 percent. Site-specific criteria or vendor-specific
costing may override the marginal cost difference listed.
2.8.4 Recommended RACT
From the standpoint of achievable emission reduction, each
control option is capable of attaining emission levels in the
range of 0.10 lb/ton of product.
The cost of control is not considered to be a factor since
almost all plants surveyed are controlled by a control method
capable of achieving the recommended emission level.
To ensure that the emissions measured by the stack test
represent the controlled emission level, a fugitive emission
limit applied to the plant process building is recommended. The
fugitive emissions must be judged on the opacity of material
escaping the building. This level is recommended to be 5 per-
cent opacity; based on good maintenance and good hood capture,
this level should be easily achieved.
2-96
-------
The test method to be used for determining the mass emis-
4
sion standard should be EPA Method 5. The use of m-stack
methods for scrubbers can yield low emission rates, because of
potential solubilization of phosphate material on the filter
when collected below the moisture dewpoint. Opacity should be
5
measured by use of EPA Method 9.
2.9 LOADING RAILROAD CARS WITH PHOSPHATE ROCK
This subsection discusses the processes used in rail car
loading of phosphate rock materials and identifies the major
particulate sources within each facility. The section also
discusses the available control technology and the cost of a
typical plant employing this technology. From this analysis
RACT recommendations are made.
2.9.1 Process Description
Large volumes of dried phosphate rock or ground rock
products are transferred mainly by rail. After the phosphate
product is dried or ground, the material is transferred to silos
for short term storage. Later, it is transferred through
materials handling systems to shipping areas for further proc-
essing or export.
Figure 2-22 illustrates an enclosed materials handling
system for conveying phosphate rock from a storage silo to a
railroad car. Four telescopic chutes are used to load material
into standard rail cars equipped with four squared-shaped
loading hatches. An operator lowers and positions the hydraulic-
operated chutes over the rail car hatches. The telescopic
2-97
-------
Figure 2-22. Materials handling system for conveying phosphate rock
from storage to a railroad car.
2-98
-------
chutes are designed with flexible square-shaped skirts to
enclose the hatches fully and thus prevent the escape of fugi-
tive emissions. Flexible hoses are attached alongside of the
chutes to vent the emissions generated from rail car loading to
the control device.
2.9.2 Emission Sources and Control Options
The major point of particulate emissions in a conventional
phosphate rock rail car loading facility is the materials
handling system, including the conveyor belts and transfer
points from the rock storage silo to the rail car. The volume
of air required to collect the fugitive phosphate rock particles
being entrained from the rail loading facility can vary from
facility to facility. These emissions are ventilated from the
elevator heads, transfer points, conveyor belts, and the rail
car. The ventilation rate is dependent on plant age, system
tightness and number of transfer points in the rock handling
system. Particulate emissions from these sources are controlled
with a conventional scrubber or baghouse systems, identical to
the control systems used throughout the phosphate rock process-
ing industry.
Review of stack tests from rail loadout and other phosphate
rock handling and loading facilities in the central Florida area
indicated emission rates of 0.0003 to 0.0125 lb/ton of material
handled. The great majority of test results showed emission
rates in the range of 0.0006 to 0.003 lb/ton of rock handled.
Grouping of emission rates by control option categories indicated
2-99
-------
that achievement of emission levels less than 0.003 was common
for each control category. Table 2-18 lists emission rates and
control options gathered from source tests for phosphate rock
handling and loading operations.
Characteristics of a rail loading control facility include
the flexible skirt and flexible duct designs incorporated with
the telescopic chute assembly. Flexible members in the control
system hooding and ducting are required to accommodate the
movement of the telescopic chutes, and to ensure complete
capture and transfer of fugitive emissions. This type of
control facility requires an operator. Complete capture of
loading emissions is dependent on the operator's conscientious-
ness to enclose the loading hatches fully with the flexible
skirt assemblies. To enclose the hatch, the operator must lower
and position the skirt to cover the opened area of the hatch
completely. Failure to adjust the flexible skirt onto and
around the hatch will allow the escape of fugitive emissions.
To ensure that the control equipment emission rate truly
reflects facility emissions requires complete and efficient
capture of emissions from sources within the facility. The
installation and operation of enclosures, hoods, and ventilation
systems have been demonstrated to reduce fugitive losses effec-
tively within the loading operation. The maintenance of these
systems has been demonstrated to be a major problem in main-
taining effective control of emissions. To ensure good operating
practices requires an effective method of evaluating loss to the
ambient atmosphere. An opacity standard applied to fugitive
2-100
-------
TABLE 2-18. PARTICULATE EMISSION SOURCES, CONTROL OPTIONS,
AND ACHIEVABLE EMISSION RATES FOR LOADING RAILROAD CARS
WITH PHOSPHATE ROCK
Achievable
emission rate,
Source
Control options
1b/ton
Railroad car loading
Cyclonic scrubber
0.0009-0.004
Railroad car loading
Venturi scrubber
0.0065
Railroad car loading
Fabric filter
a
aData not available.
2-101
-------
loss from the facility is an effective method of maintaining
control of fugitive emissions. Plants using good capture
practices can reasonably be expected to maintain fugitive
losses from building vents and rail car hatches at less than 5
percent opacity.
2.9.3 Control Costs
Cost estimates of the various control options presented in
the previous subsection are presented in Table 2-19. Cost and
emission estimates are based on a representative rail loadout
facility handling phosphate rock at 800 tons/h. The process and
control device characteristics incorporated into the costing
estimates were derived from several facilities and control
device cases in the central Florida area. Searching of state
file records and plant visits were conducted to obtain the
specified values used in cost estimates. Capital cost figures
were based on values reported on permit applications on file
with the State of Florida Department of Environmental Regulation
and were adjusted to January 1980 dollars by use of the C-E
plant cost index.^ Due to a wide range in plant designs,
control device vendors, and the degree of safety factor and
redundancy, it is possible that capital cost estimates may vary
by + 50 percent. The cited examples represent cases in which
data were available and typical of current technology. Cost of
removal per ton of product values was determined by the ratio of
total annualized cost and the estimated tonnage of emissions
captured.
2-102
-------
TABLE 2-19. CONTROL COSTS OF LOADING RAILROAD CARS WITH PHOSPHATE ROCK9
Emissions
Cost, S
Source of
emissions
Control
options
Control,
%
Uncontrolled,
lb/ton
Controlled,
1b/ton
Removal,
tons/yr
Total
capi tal
cost, $
Expected
1 i fe,
yr
Annual,
cam tal
Annua 1
0&M
Credits
Cost of
removal,
5/ton
Railroad car
loading
Venturi
99.75
3.0
0.005
6381
51,000
10
8,300
120,730
c
20
Railroad car
loading
Cyclone
99.75
2.0
0.005
6384
58,000
10
9,400 _
121,130
c
20
Railroad car
loading
Fabric
f i1ter
99.75
2.0
0.005
6384
145,000
10
23,600
112,840
c
21
aCosts In January 1980 dollars.
bBased on a 10 percent cost of capital.
cData not available.
-------
Cost estimate levels in Table 2-19 reveal insignificant
cost differences between the various control options. Apparent
cost advantages of a specific control option may not necessarily
represent actual cost advantages. As stated earlier, cost esti-
mate levels may vary + 50 percent, depending on several factors.
Site-specific criteria or vendor-specific costing may override
the marginal cost difference listed.
2.9.4 Recommended RACT
From the standpoint of achievable emission reduction, it
appears that each control option is capable of attaining emission
levels in the range of 0.010 lb/ton of product. The cost of
control is not a major factor since almost all plants surveyed
are controlled by a control method capable of achieving the
recommended emission level.
To ensure that the emissions measured by the stack test
represent the controlled emission level, a fugitive emission
limit applied to the plant process building is recommended. The
fugitive emissions must be judged on the opacity of material
escaping the building. This level is recommended to be 5
percent opacity; based on good maintenance and good hood capture,
this level should be easily achieved.
The test method to be used for determining the mass emission
4
standard should be EPA Method 5. The use of m-stack methods
for scrubbers can yield low emission rates, because of potential
solubilization of phosphate material on the filter when col-
lected below the moisture dewpoint. Opacity should be measured
by use of EPA Method 9.^
2-104
-------
2.10 SHIPS WITH PHOSPHATE ROCK LOADING
This subsection discusses the processes used in the ship-
loading of phosphate rock materials and identifies the major
particulate sources within each facility. The subsection also
discusses the available control technology and the cost of a
typical plant employing this technology. From this analysis
RACT recommendations are made.
2.10.1 Process Description
Large quantities of phosphate rock materials are exported
for use or further processing by ship. Cost-effective ^transfer
to foreign ports is accomplished by transoceanic tanker ships.
Figure 2-23 shows a typical materials handling schematic,
including materials unloading, storage, and shiploading for an
export terminal.
The phosphate rock material is received by rail or truck
shipments. After the material is unloaded, it is transferred
through a series of weather-protected belt conveyors to storage
buildings. From storage, the phosphate material is conveyed
with similar equipment to the shiploader. Typical capacities of
the handling systems range from 300 to 3000 tons/h of phosphate
material.
The shiploading system depicted in Figure 2-24 consists of
weather-protected conveying equipment, mounted on tracks to
allow for movement across the full length of the stationary,
docked ship. Each of several compartments in the ship are indi-
vidually loaded. Before compartment loading, the telescopic
2-105
-------
Figure 2-23. Materials handling system for conveying phosphate
rock and granulated fertilizer from storage to a ship.
-------
Figure 2-24. Fugitive dust collection system mounted on a
movable boom above a ship.
-------
chute is lowered several feet below the level of the deck.
Several large canvases are stretched across the compartment open-
ing and overlapped to minimize the escape of fugitive emissions.
A well-designed shiploader chute has canvases attached "directly
to the chute assembly to minimize escaping emissions. Potential
emissions from the conveyors, transfer points, and ship hold are
vented through ducting to the control hardware. Shiploading
capacities ranged from 800 to 3000 tons/h of phosphate material.
2.10.2 Emission Sources and Control Options
The major point of particulate emissions in a conventional
phosphate rock ship loading facility is the materials handling
s.ystem, including the conveyor belts and transfer points from the
conveyor belt to the ship's hold tanks. The volume of air
required to collect the fugitive phosphate rock particles being
lost from the ship loading facility can vary from facility to
facility. These emissions are ventilated at elevator heads,
transfer points, conveyor belts, and the hold tanks. The venti-
lation rate is dependent on equipment age, tightness and number
of transfer points in the rock handling system. Tightness of the
system can be maximized with a good O&M plan. Particulate emis-
sions from these sources are controlled with conventional bag-
house systems, identical to the control systems used throughout
the phosphate rock processing industry.
Review of source tests from ship loading facilities in the
Tampa bay area indicated emission rates of 0.0001 to 0.007
lb/ton of material handled. The test results showed handling
2-108
-------
rates in the range of 800 to 3000 tons/h. All the surveyed ship
loading facilities used fabric filters for control of particulate
emissions. Most fabric filter systems were pulse-cleaned units;
they were preferred due to advantages in cost, operation, flow
handling capacity, and physical size. Well-designed control
systems incorporated two fabric filters for recovering ship-
loading emissions. As shown in Figure 2-24, one fabric filter
system controls fugitive emissions from the conveyor belt and
transfer points before the phosphate material is loaded into the
ship's hold tanks. The second fabric filter system controls
emissions generated during loading of the ship's tanks, through
flexible ducting attached to the shiploader's telescopic chutes.
Table 2-20 shows the typical control method and emission rates
achieved from source tests conducted on several shiploading
facilities.
TABLE 2-20. PARTICULATE EMISSION SOURCE, CONTROL OPTION, AND
ACHIEVABLE EMISSION RATE FOR LOADING SHIPS WITH PHOSPHATE ROCK
Achievable
emission rate,
Source
Control option
lb/ton
Shiploading
Fabric filter
0.0001-.007
Characteristics of a shiploading control facility include
the use of canvases and flexible ducts as an integral part of
the control system. Canvases are required to minimize the
escape of fugitive emissions from the hold tank during ship-
loading. To minimize escaping emissions further, the canvas
2-109
-------
must be attached to the telescopic chute and similarily stretched
across the ship's hold tank, and overlapped with the adjacent
canvases. Flexible ducting is used to accommodate movement of
the telescopic chute in and out of the hold area. Capture
efficiency of loading emissions is dependent on the loading
crew's conscientiousness to enclose the ship's opening by using
and overlapping large-sized canvases as specified in a good O&M
plan. Dimensions of hold tanks vary with the different tanker
ships. Use of canvas is the practical method of covering the
large and various-sized tanks. Tank dimensions range from 4 0 to
60 feet across for both the length and width.
To ensure that the control equipment emission rate truly
reflects facility emissions requires complete and efficient
capture of emissions from sources within the facility. The
installation and operation of enclosures, hoods, and ventilation
systems effectively reduce fugitive losses within the loading
operation. The maintenance of these systems is a major problem
in maintaining effective control of emissions. To ensure good
operating practices requires an effective method of evaluating
loss to the ambient atmosphere. An opacity standard applied to
fugitive loss from the facility is an effective method of main-
taining control of fugitive emissions. Plants using good capture
practices can reasonably be expected to maintain fugitive losses
from vents and ship hatches at less than 5 percent opacity.
2-110
-------
2.10.3 Control Costs
Cost estimates of the control method presented in the previ-
ous subsection are presented in Table 2-21. Cost and emission
estimates are based on a representative facility loading phosphate
rock at 1500 tons/h. The process and control device character-
istics incorporated into the costing estimates were derived from
several shiploading and control device cases on the Florida
coast. State files were reviewed and plant visits were conducted
to obtain the specified values used in cost estimates. Capital
cost figures were based on values reported on permit applications
on file with the State of Florida Department of Environmental
Regulation and were adjusted to January 1980 dollars by use of
3
the C-E plant cost index. Due to a wide range in plant designs,
control device vendors, and the degree of safety factor and re-
dundancy, it is possible that capital cost estimates may vary
by + 50 percent. The cited examples represent cases in which
data were available and thus cases typical of current technology.
Cost of removal per ton of product values were determined by the
ratio of total annualized cost and the estimated tonnage of
emissions captured.
2.10.4 Recommended RACT
From the standpoint of achievable emission reduction, the
control method is capable of attaining emission levels in the
range of 0.010 lb/ton of product. The cost of control is not a
major factor since almost all plants surveyed are controlled by
a control method capable of achieving the recommended emission
level.
2-111
-------
TABLE 2-21. CONTROL COSTS OF LOADING SHIPS WITH PHOSPHATE ROCK3
Emlsslons
Cost, t
Source of
emissions
Control
options
Control,
I
Uncontrolled,
lb/ton
Control led,
lb/ton
Removal,
tons/yr
Total
capi tal
cost, $
Expected
1 i fe,
yr
Annual^
capi tal
Annual
0&M
Credits
Cost of
removal,
S/ton
Shlpload-
Ing
Fabric
filter
99.75
2.0
0.005
11,970
495,000
10
80,550
217,820
c
19
aCosts In January 1980 dollars.
''Based on a 10 percent cost of capital.
cData not available.
-------
To ensure that the emissions measured by the stack test
represent the controlled emission level, a fugitive emission
limit applied to the plant process building is recommended. The
fugitive emissions must be judged on the opacity of material
escaping the building. This level is recommended to be 10
percent opacity; based on good maintenance and good hood capture,
this level should be easily achieved.
The test method to be used for determining the mass emis-
4
sion standard should be EPA Method 5. The use of in-stack
methods for scrubbers can yield low emission rates, because of
potential solubilization of phosphate material on the filter
when collected below the moisture dewpoint. Opacity should be
measured by use of EPA Method 9.5
2-113
-------
SECTION 3
PORTLAND CEMENT
This section discusses the processes employed in the produc-
tion of portland cement, available control technology, the cost
of control an finally recommended RACT. Since there is only one
Portland cement producer in the TSP nonattainment areas in
Florida, this process description and technical analysis of
control options has been limited to specifics of this plant.
3.1 PROCESS DESCRIPTION
The plant is a wet process rotary kiln operation, having
three kilns with a maximum capacity of 140 tons/h clinker produc-
tion rate.
The process flow consists of raw material receiving and
storage, raw material grinding, calcining of slurry, clinker
cooling and storage, finish cement grinding, cement storage,
bagging and shipment.
The raw materials received for slurry formation are clay,
koalin sand, slag and aragonite (Caribbean CaCO^). The raw
materials are reduced in rotary mills and fed into the kilns as
a slurry.
The plant operates three kilns which are fired with pulver-
ized coal. The coal is pulverized by Raymond Mills and is
3-1
-------
injected by solid fuel burners in the burner end of the kilns.
Carbon dioxide rich gases are recycled from the kiln hoods to the
mills to dry the coal and to maintain an inert atmosphere.
The calcination process consists of heating the raw mate-
rials to approximately 1800°F at which point carbonaceous mate-
rials are oxidized and alkaline components are vaporized. The
calcium carbonate in the slurry is converted to calcium silicates
(CaO • SiC>2) by combination with the Silicon oxides in the sand
and clay and fused to a clinker. The fusion takes place near the
burner end of the kiln at a temperature between 2700 and 3000°F.
The clinker is discharged from the burner end of the kilns to
clinker coolers.
The entrained particulates and vaporized alkaline materials
which are exhausted from the kilns are controlled by electro-
static precipitators.
The clinker at discharge from the kiln is iridescent and
must be cooled before being placed in storage. The cooler con-
sists of a moving grate on which the hot clinker is cooled.
Cooling air is forced through the bed of clinker and discharged
through an induced draft fan. The air entrains particulates from
the clinker bed and potential emissions are as high as 20 lb/ton.
The plant uses gravel bed filters to collect and separate these
emissions from the gas stream.
The clinkers are placed in silo storage until they can be
introduced into the finish mills. In the finish mills, the
clinker is ground to a fine powder 95 percent less than 200 mesh.
3-2
-------
The ground cement is mixed with gypsum and is classified in sepa-
rators. The oversize is returned to the mills with the feed
material. Particulate emissions from the conveyors, elevators,
separator, and finish mills are controlled by fabric filters.
The ground cement is transferred pneumatically to finish
silos for storage before bagging or bulk shipment.
For the purposes of this analysis, the general materials
transport, and storage areas of the plant are discussed under the
general requirements of material transfer discussed in Section 6.
The kilns, clinker coolers, and finish mills are addressed in
Section 3.
3.2 EMISSION SOURCES AND CONTROL OPTIONS
3.2.1 Kiln 6
Kiln 6 is a wet process kiln in which a slurry of aragonite,
sand, and koalin are calcined to form a calcium-silicate clinker.
The slurry (sand and koalin) is produced by rotary wet raw grind
mills and pumped to the feed end of the kiln. Unground aragonite
is mixed with the slurry at the feed end of the kiln. The slurry
typically has a moisture content of 29 percent and a carbonate
content of 76 percent +0.2. The feed rate of dry solids is 148
tons/h and produces clinker at 7 3 ton/h.
Particulate emissions from the kiln are controlled by a
multicyclone and an electrostatic precipitator (ESP). The
control system arrangement is shown in Figure 3-1. The ESP is a
Western Precipitation unit installed in 1962. The plate area is
2
69,984 ft and at design gas volume, has a specific collection
3-3
-------
U)
I
CLINKER,
74 tons/h
SLURRY,
148 tons/h
350,000 acfm
0.01 gr/acfm
25 lb/h
12,375 1b/h
4.8 gr/acfm
ESP
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
52,398 lb/h
(20 gr/acfm) I
O"
DRAG CONVEYORS
77777/
Figure 3-1. Control equipment arrangement at Kiln 6.
-------
2
area (SCA) of 200 ft /1000 acfm. The superficial velocity is 4.5
ft/s. The unit has three chambers and six fields. The original
design consisted of 7 T-R sets (field 1-4, 1 each; field 5, 3
units). The collected dust was discharged by three drag convey-
ors with removal at the inlet end.
In the past the collection of particulate in the unit has
been less than that required by the regulatory agency. The com-
pany has initiated and implemented modifications to the original
design to reduce the deficiencies. The inlet to the ESP has had
severe gas distribution problems and, based on rapper/opacity
correlations, was experiencing reentrainment from the drag chains.
The plate area and power supply are within accepted design ranges
for wet process kilns, but because of design problems, did not
achieve the desired removal efficiency.
The changes which were made were:
1. Isolation of the number six field from the drag chains
by installation of curtain and hoppers to reduce
hopper sneakage and reentrainment.
2. Installation of timer sequencing of the three drag
chain conveyors.
3. Installation of a primary collector before the ESP.
4. Installation of inlet turning vanes and separate nozzles
for each chamber, to correct gas maldistribution.
5. Insulation of the ESP shell and hoppers.
i
6. Sectionalization of the fourth field with three T-R
sets and the installation of additional power to
fields 2 and 3.
7. Replacement of original rappers with air activated im-
pulse rappers in all fields. The installation of peg
board rapper controls in all fields.
3-5
-------
8. Installation of digital power control circuits to
provide maximum power input to all sections.
The changes incorporated into the design have allowed the kiln
to achieve a mass emission rate of 25.1 lb/h. This is equivalent
to 0.156 lb/ton of feed (dry solids). The inlet grain loading
to the multicyclone precleaner is estimated to be 20.0 gr/acfm
(52,398 lb/h).
The plant monitors the power levels of the ESP each shift
and records the values in a log book. In addition, an air load
test is performed on the unit each time the unit is brought down
for repairs or routine maintenance. Review of operating logs
indicated that the power levels have been stable for extended
periods and have not deviated from the levels during the last
performance test.
The kiln stack is equipped with a transmissometer and
review of the strip charts indicates that the reentrainment
spikes have been eliminated by the design changes. The opacity
is stable between 6 and 10 percent.
The ESP is energized during kiln start up to minimize emis-
sions. The digital controller allows the power input to be con-
stantly maintained below the spark limit. The start up is
initiated on oil firing until the ESP has achieved operating
temperature (maximum power). The kiln is then changed over to
pulverized coal firing and after a stable flame has been estab-
lished, the kiln speed increased to maximum rate. The slurry is
3-6
-------
introduced to the kiln at programmed rate until maximum speed has
been achieved. The start up time varies between 16 and 24 hs
depending on the condition of the kiln.
Because of the high moisture content of flue gas from the
kiln the control options available are limited. The accepted
control technique for wet process kilns is a precleaner followed
by an ESP. The plate area, sectionalization, power control, and
power levels have increased the removal efficiency of the pre-
cipitator to a level near that of an NSPS source (0.30 lb/ton dry
feed).
The opacity of the stack before moisture condensation is
low, however, under certain conditions, the opacity after mois-
ture dissipation is high (80-100 percent). The increased opacity
is believed to be the result of secondary aerosol formation in
the plume after moisture condensation. The aerosol is believed
to be composed of sulfate or ammonium chloride radicals. The
aerosol normally occurs in wet process kilns firing coal (high in
nitrogen and chlorides) and having low primary particulate emission
rates. More research is needed to completely quantify reactions.
Because of the unpredictable character of the stack plume, the
opacity which reflects the primary particulate level is in the
stack. The transmissometer should be installed, calibrated and
maintained in a manner consistent with part 61 of the Federal
Register.
It is the opinion of the investigators that the proper
maintenance and operation of the ESP can consistently maintain
the emissions below 0.30 lb/ton of feed (dry solids).
3-7
-------
3.2.2 Kilns 4 and 5
Kilns 4 and 5 are similar to Kiln 6 with the exception that
the aragonite is introduced to the raw grind mills with the sand
and koalin before introduction to the kilns.
Kilns 4 and 5 operate at a dry solids feed rate of 50 ton/h
each and produce clinker at a rate of 25 ton/h each. The moisture
content of the slurry is 29 percent and the carbonate is approxi-
mately 76 percent.
The particulate from the kilns is controlled by a Kopper's
ESP. The ESP has two chambers and four fields. The plate area
2
is 82,080 ft and with both kilns operating, has a SCA of 373
2
ft /1000 acfm. The superficial velocity is 4.3 ft/s.
The ESP has four T-R sets and digital power input controls.
The plates and discharge wires are rapped by electric vibrators.
The collected dust is removed through eight pyramidal hoppers.
The kiln stack is equipped with a transmissometer to measure
plume opacity.
Based on stack test data provided by the Florida Department
of Environmental Regulation, the combined emission rate of Kilns
4 and 5 can be reduced to 24.0 lb/h (0.]9 lb/ton of dry feed
solids).
The gas volume being handled by the unit when a single kiln
is fired is less than that for which the ESP was designed. The
collection efficiency is greater due to the higher SCA and lower
superficial velocity.
3-8
-------
Based on the the design of the inlet and the offset required
to enter the ESP, it is suspected that the outer chamber is
carrying an increased load of particulate.
It has been noted that the secondary current in Fields 2 and
3 has been supressed for an extended period (90 days log re-
viewed) and that the supression was not occurring during the
stack testing. The reduction in secondary current indicates
reduced collection efficiency. The reason for the supression is
not known, but is suspected to provide complete discharge wire
and plate rapping in these fields. The maintenance of continuous
compliance with 0.30 lb/ton dry feed solids would require correc-
tion of this deviation in optimum power input. It is estimated
that a minimum power input of 60,000 watts is necessary to comply
with the proposed standard at 220,000 acfm gas volume with both
kilns operating.
3.2.3 Clinker Coolers
The plant uses oscillating grate clinker coolers to cool the
clinkers after discharge from the kilns. The clinker is cooled
from a maximum temperature of 2700°F to approximately 350°F by
forcing air through the moving bed. The cooling air removes fine
particulate from the bed. Uncontrolled emission rates for
oscillating grate clinker coolers are in the range of 30 lb/ton
of clinker.
There are three coolers, one on each kiln. The process
rates are approximately 140 tons/h (80 tons/h for Cooler 6, 30
tons/h for Cooler 4, and 30 tons/h for Cooler 5). The particulate
3-9
-------
emissions are removed from the gas stream by passing the gases
through a bed of gravel. The collector arrangement at Clinker
Cooler 6 is shown in Figure 3-2. The collector is divided into a
number of individual compartments, each separately cleaned on a
preset cycle. The collection compartment pressure drops are
monitored and recorded. The gravel bed is cleaned by passing
reverse-air through the bed. Stack test data provided by the
State of Florida Department of Environmental regulation indicate
the controlled emission rates for these systems are between 0.08
lb/ton and 0.4 6 lb/ton. All tests except one were below 0.20
lb/ton. It appears that the Cooler 6 control has deteriorated
since installation from 0.14 to 0.46 lb/ton.
Options for particulate control include multicyclone,
fabric filters, and ESP's.
3.2.4 Finish Mills
Cement is produced from clinker by grinding to 200 mesh and
mixing with gypsum. The clinker is ground in rotary ball mills.
The clinker is fed into the mill at one end and removed at the
other end via a screw conveyor. The ground clinker is elevated
to an air separator in which the particles are classified by
size. The oversize is returned to the finish mills. The product
is collected and transported pneumatically to finish cement
silos. The finish mills are vented to remove the gas volume
generated by the evaporation of water sprays used to cool the
mills. The finish mill/separator fabric filter arrangement
3-10
-------
GRAVEL BED FILTERS
¦ i i
i*. m ¦>*«
FORCED AIR
Figure 3-2. Collector arrangement at Clinker Cooler 6.
-------
is shown in Figure 3-3. The exhaust serves the mill and screw
conveyor transporting the cement to the bucket elevator. The
exhaust is approximately 13,000 acfm and contains approximately
35 gr/acfm of entrained particulate. The exhaust is controlled
by a reverse-air-type fabric filter operating at an air-to-cloth
2
design ratio of 3.64 acfm/ft .
The separator is exhausted to a fabric filter. The gas
volume is approximately 34,000 acfm and the inlet grain loading
is 35 gr/acfm. Fabric filters are the accepted method for con-
trolling emissions from these sources. More modern plants
combine the sources into a common collector. Test data on
finish mills using one pulse-jet collector have demonstrated the
capability to reduce the emissions to 0.015 gr/dscf.
Because of the high grain loading and abrasive nature of
cement dust, the process equipment and conveying systems must be
adequately ventilated and maintained to reduce fugitive emis-
sions. The proper operation and maintenance of the system
should reduce the fugitive emissions to a level at which visible
emissions from the finish mill building can be maintained below
5 percent opacity.
3.3 CONTROL COSTS
The estimated cost of the control methods discussed in the
previous sections is presented in Tables 3-1, 3-2, and 3-3. The
cost is based on data supplied by the Company and typical values
for the industry.
3-12
-------
SEPARATOR
FABRIC FILTER
Figure 3-3. Finish mill/separator fabric filter arrangement.
3-13
-------
TABLE 3-1. CONTROL COSTS OF CEMENT KILNS9
Emissions
Cost, $
Total
Expected
Cost of
Source of
Control
Control,
Uncontrolled,
Controlled,
Removal,
capital
1 i fe.
Annual.
Annual
removal,
emissions
options
%
lb/ton
1 b/ton
tons/yr
cost, J
y
capital
OfcM
Credits
S/ton
Kilns 4
ESP
99.2
302.7
0.23
23,988c
745,000
20
87,507
144,908
d
2
and S
Kiln 6
Mechan-
99.95e
354
0.19 205,S92f
,000,000
20
117,460
527,472
d
3
ical
col-
lector.
ESP
aCosts In January 1980 dollars.
bBased on a 10 percent cost of capital.
cBased on operation for 8000 h/yr at dry feed rate of 102 tons/h.
"^Product recovered Is not currently being returned to the kilns.
eMechan1cal collector is 76 percent efficient; ESP Is 99.8 percent efficient.
^Based on operation for 8000 h/yr at dry feed rate of 148 tors/h.
-------
TABLE 3-2. CONTROL COSTS OF CEMENT CLINKER COOLERS3
Emissions
Cost
. t
Cost of
(credit for)
removal,
J/ton
Source of
emissions
Control
options
Control,
%
Uncontrolled,
lb/ton
Control led,
lb/ton
Removal,
tons/yr
Total
capital
cost, $
Expected
1 ife,
yr
Annual.
capital
Annual
OSM
Credits
CI inker
cooler 4
MultIclone
85
30
4.5
3060d
50,000
10
B, 136
41,685
133,171
(?7)
CI Inker
cooler 4
Fabric filter
99
30
0.30
3564d
558,000
10
90,803
154,053
155,105
25
Clinker
cooler 4
Gravel bed
99.3
30
0.20
3576d
500,000e
10
55,000
107,884
155,627
2
CI Inker
cooler 5
Multlclone
85
30
4.5
3060d
50,000-
10
8,136
41 ,685
133,171
(27)
CI Inker
cooler 5
Fabric filter
99
30
0.30
3564d
558,000
10
90,803
154,053
155,105
25
CI Inker
cooler 5
Gravel bed
99.3
30
0.30
3576d
500.0006
10
55,000
107.884
155,627
2
Clinker
cooler 6
Multiclone
85
30
4.5
7548f
80,000
10
13,018
62,603
328,488
(33)
CI Inker
cooler 6
Fabric filter
99
30
0.30
8791f
1,250,000
10
203.412
223,519
382,584
5
CI Inker
cooler 6
Gravel bed
99.3
30
0.30
3820f
1,000,000
10
110,000
195,123
383,846
(8)
aCosts In January 1980 dollars.
bBased on a 10 percent cost of capital.
cBased on 80 percent of finished cement value of $54.40/ton. Source: Material Prices, Engineering News Record, 205(10)'36-37, September
1980. (M1U price Alabama.)
dBased on operation for 8000 h/yr at 30 tons/h.
eEst1mate.
fBased on operation for 8000 h/yr.
-------
TABLE 3-3. CONTROL COSTS OF CEMENT FINISH MILLS3
Emissions
Total
capital
cost, $
Expected
11 fe,
yr
Cost, i
Cost of
(credit for)
removal,
$/ton
Source of
emissions
Control
options
Control,
%
Uncontrol1ed,
1b/ton
Control led,
1b/ton
Removal ,
tons/yr
Annua 1^
capita1
Annual
OiM
Creditsc
Finish
mill
separator
Fabric
filter
99
10,200e
s.s?f
40,776
306,000
10
49.801
113,386
2,218,214
(51 )
Finish
mill
Fabric
fi1ter
99
3,900e
2.2f
15,591
117,000
10
19,039
93,580
848,15C
(47)
aCosts in January 1980 dollars.
''Based on a 10 percent cost of capital.
cBased on finished cement price of $54.40/ton. Source Material Prices. Engineering News Record, 205(10) 36-37, September 1980. (Mill price
Alabama.)
''Each mill requires two collectors. Four finish mills are operted at the plant.
eBased on 35 gr/acfm.
^Based on 0.02 gr/acfm.
-------
3.3.1 Kilns
The cost of particulate removal in kiln operations is based
on power input to the ESP required, fan horsepower, maintenance
costs, and capital charges. The cost of achieving the standard
is not directly related to the mass emitted because the basic
structure (plates, shell, etc.) are in place. The difference in
achievable mass emission rates is related directly to the avail-
able power input.
The changes incorporated by the company have been successful
in allowing the continuous power levels necessary to achieve the
standard at other times than that required for performance
tests. The cost of control for the two ESP's is different
because of the difference in power inputs, gas volume, and SCA.
The values, however, are comparatively low for particulate
removal.
3.3.2 Clinker Coolers
The method of controlling clinker coolers at the plant is by
gravel bed filtration. For comparison the costs of control of
fabric filters and multicyclones have been provided.
The last cost option is the multicyclone. The lower cost is
reflected in lower pressure drop and maintenance costs. The
highest cost is fabric filters. This is caused by the higher
maintenance cost and capital investment required. The method
chosen by the plant is lower in cost than fabric filtration even
though the system has a higher pressure drop.
3-17
-------
The value of recovered clinker has been calculated in the
cost/ton removed and this credit substantially reduces the
anualized cost.
3.3.3 Finish Mills
The finish mill collectors used at the plant are fabric
filters. Since the product must be recovered in a dry state and
has market value, the accepted control method is fabric filters
irrespective of cost.
The annualized cost estimates include the value of recovered
product. The credit for this recovery substantially reduces the
annualized cost.
3.4 RECOMMENDED RACT
Based on the design and operation of the ESP serving the
cement kilns, it is the opinion of PEDCo that the units are cap-
able of achieving an emission limit of 0.30 lb/ton dry feed on a
consistent basis. The opacity as measured in the stack can be
consistently maintained below 20 percent.
The control devices employed to control emissions from the
clinker coolers have demonstrated the ability to achieve an
emission level of 0.20 lb/ton of clinker. The cost of control
appears to be less than that of fabric filtration but substan-
tially higher than mechanical methods. Observation of the
stacks indicates that the opacity can be maintained below 20
percent.
3-18
-------
The finish mills are used to grind the cement clinker and
produce the market product. The collected particulates are re-
turned to the process and typically represent 20 percent of the
mill's throughput. The recovery of the material is necessary for
economic operation of the process.
The proper maintenance and operation of fabric filters on
the mill and separator exhausts as specified in an O&M plan can
reduce the outlet loading to 0.02 gr/acfm.
The cost of recovery of the dust is low compared to the
value of the material. The proper operation of the enclosures,
hoods, and ventilation system is necessary to collect the dust
from the mill, separator, elevator, and conveyors. The proper
use of the system can consistently maintain the fugitive dust
level from the finish mill enclosure below 5 percent opacity.
To ensure proper maintenance of the fabric filters and pro-
vide continuous compliance with the mass standard, an opacity
limit of 5 percent is required. The mass standard cannot be
achieved if an opacity of less than 5 percent is consistently
observed.
3-19
-------
SECTION 4
ELECTRIC ARC FURNACES
This section describes the operation, emission sources,
control options and costs, and reasonably available control
technology (RACT) for electric arc furnaces.
4.1 PROCESS DESCRIPTION
Electric arc furnaces are widely used to produce steel.
Figure 4-1 presents a schematic of a typical electric arc furnace.
Furnace operation is initiated by swinging the furnace roof aside
to permit charging of the furnace with scrap steel. After charg-
ing, the roof containing carbon electrodes is returned to the top
of the furnace and the electrodes are moved down into the scrap.
Electric power is then introduced to the electrodes to heat and
melt the charge of scrap steel. By alternate charging and melt-
ing, the furnace eventually is filled to its capacity with molten
steel. Lime is also added with the scrap steel charges to act as
a flux in removal of impurities present in the scrap.
When the charging and melting is completed, the composition
of the melt is modified as desired in a refining operation in
which alloying agents such as ferromanganese, silicon manganese,
and ferrosilicon are added. The refined melt is then tapped
4-1
-------
Figure 4-1. Electric furnace for steel making.
-------
from the furnace into ladles which are taken by crane to an
adjoining work area when the melt is poured into an appropriate
mold for shaping and solidification into the final desired steel
product. After discharge of the melt, the furnace is cooled for
refractory lining repair if needed before another cycle of scrap
charging and melting is repeated.
Furnace capacities vary from as small as one ton to as much
as several hundred tons and the length of a furnace cycle (as
measured from tap to tap) varies as a function of furnace capacity
and electric power input.
4.2 EMISSION SOURCES AND CONTROL OPTIONS
Particulate emissions from an electric arc furnace occur
throughout the cycle but are particularly noticeable during the
charging and tapping steps of operation. The particulate emis-
sions consist in large part of metal oxides and are usually small
in size (i.e. , less than 10 micrometers diameter) . The amount of
emissions varies from 4 to 40 lb/ton of produced steel. The
quantity depends upon the cleanliness and type of scrap employed,
the manner of scrap charging to the furnace, and the method of
tapping. An average uncontrolled level of 10 lb/ton of steel
produced is representative for the industry.
Several options of control exist for particulate emissions
with respect to the control device used and the collection con-
figuration employed. These options are outlined in Table 4-1.
As can be noted from the table, the use of a canopy plus ductwork
at the furnace venting to a fabric filter is the option that has
the advantage of both good capture and removal efficiencies.
4-3
-------
TABLE 4-1. PARTCULATE EMISSION CONTROLS FOR ELECTRIC ARC FURNACES
Control option
Comments
Emissions capture
Direct evacuation at
furnace roof
Does not capture emissions during
charging and tapping
Direct evacuation plus
canopy hood over arc-furnace
Commonly used
Total Building Evacuation
Normally used for small buildings only
or where a large number of sources
exist in building
Control device
Electrostatic precipitator
Not commonly used because gas stream
conditioning is required
Fabric filter
Commonly used because of inherent high
efficiency
High-energy scrubber
Rarely used because capture of small
particles is difficult
4-4
-------
4.3 CONTROL COSTS
Representative costs for the use of the various dust control
devices are presented in Table 4-2 for a canopy and direct evacu-
ation configuration of dust collection. The canopy and direct
evacuation method of gas collection is recommended as it is the
most widely employed configuration in the industry. As shown by
the costs in Table 4-2, the fabric filter is the most economic
control device with respect to both capital investment and direct
operating cost. The use of a high energy scrubber is penalized
at the high control efficiency required by the proposed regula-
tion both by the high energy level required and by the difficulty
of fine particulate capture. The use of an ESP for control
requires gas conditioning and a large plate area to comply with
the emission level proposed. By comparison, a fabric filter has
an inherently high capture efficiency for fine particulate
matter and the gas needs no preconditioning for efficient filter
operation.
4.4 RECOMMENDED RACT
On the basis of the relative investment and operating costs
required and its widespread employment in the industry, the use
of a fabric filter with a canopy and direct evacuation collection
system is the recommended RACT for the control of particulate
emissions from an electric arc furnace producing steel.
A fabric filter on an electric arc furnace can reasonably be
expected to reduce emissions from the control device to the level
4-5
-------
TABLE 4-2. CONTROL COSTS OF A TYPICAL ELECTRIC ARC FURNACE PRODUCING STEEL3
Emissions
Total
capi tal
cost, I
Expected
life,
yr
Cost, $
Cost of
removal,
S/ ton
Source of
emissions
Control
options
Control
X
Uncontrolled,
1b/ton
Controlled,
lb/ton
Removal,
tons/yr
Annual,
capital
Annual
0«H
Electric arc
furnace
Electrostatic
precipitator
99.9
10
0.01
8392
4,500,000
10
732,375
480.000
144
Electric arc
furnace
High-energy
venturl
scrubber
99.9
10
0.01
8392
4.100.000
10
667.275
1.000.000
199
Electric arc
furnace
Fabric filter
99.9
10
0.01
8392
3.400.000
10
565.335
330,000
107
aCost In January 1980 dollars,
biased on a 10 percent cost of capital.
-------
of 0.006 gr/dscf at a reasonable cost. In addition, the use of a
canopy over the furnace should permit capture such that there
should be no visible emissions from the building openings except
during charging and tapping periods. Therefore, it is recom-
mended that no visible emissions be permitted from the building
openings except during charging and tapping periods when a higher
opacity would be permitted.
Emissions from the control device should be measured by use
4
of EPA Method 5, and opacity should be measured by use of EPA
Method 9.5
4-7
-------
SECTION 5
SWEAT OR POT FURNACES
This section identifies the major particulate sources from
sweat or pot furnaces. Also discussed are the available control
technology and the cost of a typical plant employing this tech-
nology. From this analysis, RACT recommendations are made.
5.1 PROCESS DESCRIPTION
Sweat or pot furnaces are used to melt metals which have
melting temperatures less than 1400°F. Furnace capacities range
from 1 to 50 tons and are usually indirectly heated by gas fir-
ing. As employed in melting and sweating practice, the furnace
is cylindrical in shape and is built of refractory-lined steel or
iron.
Sweating, as applied to metallurgical practice, is the pro-
cedure whereby a material containing metals of different melting
points is heated to liquify the metal of lower melting point and
thus separate it from the higher melting metal. Lead is commonly
recovered from scrap metal and storage batteries by this pro-
cedure.
The furnace process for melting - consists of the sequential
steps of charging the pot with solid metal, heating and melting
the charge, adjusting its composition as required, and then
5-1
-------
discharging it from the furnace in liquid form into a ladle for
transport to a casting operation. Lead, zinc, bismuth, and
antimony are among the metals that are processed in this manner.
5.2 EMISSION SOURCES AND CONTROL OPTIONS
Particulate matter in the form of metal and metal oxides are
emitted from the mouth of the furnace especially during charging
and tapping operations. The emissions are a function of the
composition of the charged materials and of the furnace tempera-
ture. The particulate emissions are characteristically small in
diameter, 10 microns or less, and range from 0.1 to 14 lbs/ton of
metal processed.
Because of the relatively small size of the furnaces used,
total enclosure of the top of the furnace is easily accomplished,
thus ensuring effective capture of particulate matter, especially
charging and tapping. Captured particulates can be removed by
scrubbing or by the use of a fabric filter. The use of an ESP
is impractical since the gas flows involved are small, 20,000 cfm
or less. Of the two options available for control, fabric
filters are more widely used for particulate emission control^
than are venturi scrubbers.
5.3 CONTROL COSTS
Investment and direct operating costs for scrubber and
fabric filter systems are presented in Table 5-1 for several gas
flow rates. The data shows that for systems in this range,
fabric filter systems are less expensive to install and operate
5-2
-------
TABLE 5-1. CONTROL COSTS OF A TYPICAL SWEAT OR POT FURNACE3
Emissions
Total
capltal
cost, I
Expected
1 i f e.
y
Cost, $
Cost of
removal,
S/ton
Source of
emissions
Control
options
Control,
X
Flow
rate, scfm
Uncontrolled,
lb/ton
Controlled,
lb/ton
Removal,
tons/yr
Annual^
capital
Annual
O&H
Furnace
High-energy
wet scrubber
99
10,000
10
0.1
1782
555.000
10
90,330
88.000
100
Furnace
High-energy
wet scrubber
99
20,000
10
0.1
3564
691,000
10
112.460
108,000
62
Furnace
High-energy
wet scrubber
99
32,500
10
0.1
5792
805,000
10
131.000
124.000
44
Furnace
Fabric filter
99
10,000
10
0.1
1782
273,000
10
44.430
27,000
40
Furnace
Fabric filter
99
20,000
10
0.1
3564
441,000
10
71,770
43,000
32
Furnace
Fabric filter
99
32,500
10
0.1
5792
592,000
10
96,348
60,000
27
aCosts In January 1980 dollars.
bBased on a 10 percent cost of capital.
-------
than are venturi scrubber systems. The auxiliary equipment
required for slurry and water handling in a scrubber system are
responsible for the higher capital investment. The higher operat-
ing costs for a scrubber are due to electric power requirements
of the system for removal of fine particulate matter. By con-
trast, a fabric filter has an inherently high collection effi-
ciency and has modest utility requirements.
5.4 RECOMMENDED RACT
The use of a fabric filter with a complete enclosure of the
top of the pot furnace is the recommended RACT. This recommenda-
tion is based upon its relative economy and widespread use in
industry.
Since a fabric filter can reasonably be expected to reduce
emissions by 99 percent at a reasonable cost, a RACT emission
limit of 0.05 gr/dscf from the control device should be attain-
able by sweat or pot furnace sources. In addition, the use of
complete enclosure of the top of the furnace should control
emissions such that there should be no visible emissions from
building openings. It is also recommended that emissions from1
4
the control device be measured by use of EPA Method 5, and that
opacity be measured by use of EPA Method 9.^
5-4
-------
SECTION 6
MATERIALS HANDLING, SIZING, SCREENING,
CRUSHING, AND GRINDING OPERATIONS
This section discusses the general processes employed in
material handling, sizing, screening, crushing and grinding
operations and an estimate of emissions from these sources. Also
discussed is the available control technology and the cost of
employing this technology. From this analysis, RACT recommenda-
tions are made.
6.1 PROCESS DESCRIPTION
Materials handling begins upon receipt of a particular raw
material commodity and continues through operations of crushing,
classifying, grinding, and storage, and culminates in the ship-
ment of a final product. The type and sequence of process opera-
tions vary according to the specific commodity to be handled.
Products handled include but are not limited to cement, clinker,
fly ash, coke, gypsum, shale, lime, sulfur, phosphatic materials,
slag, and grain or grain products. Figure 6-1 is a simplified
flow diagram of materials handling operations.
Raw material unloading operations frequently associated with
materials handling include: dumping by trucks; crane-clamshell
and bucket ladder removal from vessels; and side, rotary, or
bottom dumping from railcars. Depending upon the nature of the
6-1
-------
Figure 6-1. Simplified flow diagram of materials handling operations.
-------
raw material, initial stockpiling may be open or enclosed.
Transfer and conveying of materials is usually accomplished
through use of belt or screw conveyors, bucket elevators, or
vibrating conveyors and pneumatic equipment. Primary crushers
are often jaw or gyratory crushers, set to act upon rocks larger
than about six inches and to pass smaller sizes. Depending on
the ultimate size requirements of the product, material from the
primary crusher may be screened with the undersize going directly
to the screening plant and the oversize to secondary crushing, or
all material from primary crushing may be routed to the secondary
crusher. Secondary crushers are often of the cone or gyratory
type. The material at this point may either be conveyed to open
storage or transferred to silos or enclosed bins. Grinding,
which reduces the material to specification product size, is
commonly conducted through use of ball or hammermills. For
example, in gypsum processing, finish grinding of rock, which
reduces the size to approximately 100 mesh, is accomplished al-
most exclusively by hammermill. For gypsum though, the mineral
is then sent on to vertical kilns or kettles for calcination and
further processing to meet the desired product specifications.
Final product is stored to await bulk shipment by either rail,
vessel, or truck.
6.2 EMISSION SOURCES AND CONTROL OPTIONS
There are six major points of particulate emission in
general material handling:
6-3
-------
(1) Unloading
(2) Conveying and transfer
(3) Storage
(4) Crushing and grinding
(5) Screening and sizing
(6) Loadout
Because the proposed regulation covers the handling of
numerous materials, the control options discussed are general in
nature. Control methods available for reducing fugitive emis-
sions from material handling activities are specific to the site
of emissions (i.e, the site of unloading) during conveying and at
points of transfer. Therefore, discussions of controls and later
of control costs are addressed by the individual sites of dust
generation. Table 6-1 summarizes the available control options.
The minimization of dust from unloading activities can be
accomplished through the total or partial enclosure of the un-
loading facility and the removal of the particulate to a bag
filter system, an enclosure without a fabric filter system, or
8 9
a water or chemical spraying system. '
The control of fugitive dust from truck dumping activities
can be accomplished with either the enclosure or spray system
techniques. The application of control practices ~to truck dump-
ing sites are dependent largely on the industry or material
involved. A 90 to 95 percent reduction of fugitive dust from
truck dumping activity can be accomplished when the site is en-
closed and the captured particulate is vented to a control
6-4
-------
TABLE 6-1. PARTICULATE EMISSION SOURCES
AND CONTROL OPTIONS FOR MATERIALS HANDLING
Source
Control option
Unloading
Partial enclosure,
total enclosure/vent to a fabric
filter, water/chemical spray
Conveying and transfer
Partial enclosure,
total enclosure/vent to fabric
filter, water/chemical spray
Storage (in structure)
Controls on transfer of material,
enclosure,
enclosure/vent to fabric filter,
water spray
Screening and sizing
Enclosure,
enclosure/vent to fabric filter
Loadout
Partial enclosure,
total enclosure/vent to fabric
filter, water spray
6-5
-------
q
device. A 50 percent control efficiency can be achieved with a
water spray system.^
Fugitive dust emissions can be controlled through the en-
closure of rail car unloading stations accompanied by dust col-
lection with bag filters. This method of control can effectively
reduce 99 percent of the fugitive dust. Depending on the type of
material involved, fugitive dust from rail car unloading opera-
tions can also be controlled using spray systems. This measure
results in an effective control efficiency of 80 percent. The
use of chemical stabilizers may improve the efficiency of this
control measure. The addition of chemicals to the spray system,
however, increases the cost of operation. The control of dust
from conveying and transfer operations can be accomplished
through methods similar to those used during unloading operations.
Conveying or transfer emissions can be minimized through the use
of enclosures or spray systems. Enclosure of conveying systems
can be either partial (top) or total. The control efficiency of
a partial enclosure system is rated at 80 percent. The total
enclosure of a conveying system, which includes the use of a dust
collection system (e.g., bag filter) can result in a control
efficiency increase to 95 percent.
Transfer stations located along the course of a conveying
operation can be significant sources of fugitive dust. The
control of dust from these sources is also accomplished using
enclosures. The total enclosure of a transfer point can effec-
tively reduce fugitive emissions by 70 percent. The addition
6-6
-------
of a bag filter to a transfer point enclosure can raise the
control efficiency to approximately 99 percent. Effective con-
trol of dust from transfer stations can also be accomplished
using water and chemical spray systems. The spray system has an
added advantage in that the aggregate subject to chemical spray
is adequately treated to effect dust suppression throughout the
entire material handling system. The control efficiency of spray
systems at transfer points is estimated to be between 70 and 95
percent.
RACT for material handling operations must, of course, be
site specific and material specific. In most cases, where the
material characteristics will not suffer from increased moisture
content, water, oil, or chemical sprays offer good control effi-
ciencies at reasonable costs. However, where material character-
istics or specifications preclude wetting, the emissions should
be controlled by enclosure and ventilation to a fabric filter.
Again a case-by-case assessment must be made to ascertain the
severity of the emissions and the relative economics of control.
During PEDCo plant visits emissions from storage of mate-
rials did not appear to be a major contributor to nonattainment
of the NAAQS. Therefore, the only emissions discussed here are
from storage facilities that are enclosed. The only emissions
from these sources are associated with the transfer of material
in and out of storage and are treated in the conveying and trans-
fer section.
Primary crushing operations can constitute a significant
emission source. As material is crushed, its surface area is
6-7
-------
greatly increased. If incoming material has a high internal
moisture content, the new surfaces will be moist and nondusting;
however, if the material has a low internal moisture content, the
crushing greatly increases the potential for generation of air-
borne dust. These emissions can be controlled by spraying with
water, oil, or chemical dust suppressants. This method can be
supplemented by venting the crushing area to a fabric filter. A
fabric filter should be used if the material is screened after
being crushed, as screening requires a low moisture content to
avoid blinding of the screens.
Secondary crushing or grinding operations also generate a
significant amount of particulate emissions. This operation is
similar to that of primary crushing, and the control options are
the same.
The control of emissions from screening and sizing is the
same as for crushing and grinding with the use of a water, oil,
or chemical spray precluded in many cases because the moisture
can cause the blinding of the screens for many materials.
The control of emissions from loadout is the same as that
for unloading except that material that is loaded out is gener-
ally finer than that loaded in and therefore may require a higher
degree of control. Also if the moisture content of the product
is important, the use of water, oil, or chemical sprays may be
precluded.
6-8
-------
6.3 CONTROL COSTS
An estimate of the cost of the various control options dis-
cussed in the previous subsection is presented in Table 6-2. The
universe of materials that may be handled each have different
costs of control. The costs and control efficiencies given in
this section are given for a specific case and therefore are not
applicable to all materials. The estimates should give a ball
park estimate that is applicable to most cases.
Material handling operations move what is usually considered
to be a "valuable" commodity from one point to another within a
given industrial setting. Because the material has been acquired
at some cost to the industry, the loss of a portion of this
material constitutes a waste. In some cases the cost of instal-
ling collection devices can be partially offset by the market
value of the material which has been captured. This type of side
benefit associated with collection devices have applications in a
number of industries. Because of the wide variety of materials
handled, no credits are taken in the cost of control calculations
in this section.
Costs in Table 6-2 are calculated assuming 500 tons per hour
of material handled 8 hours per day, 250 days per year. All
costs are in 1980 dollars. Specific cost calculations should be
done on a site-specific basis to determine if the cost per ton
removed is reasonable.
6-9
-------
TABLE 6-2. CONTROL COSTS OF MATERIALS HANDLING3
Emissions
Total
capital
cost, J
Expected
life,
yr
Cost, i
Source of
emissions
Control
options
Control,
I
Uncontrolled,
1b/ton
Control led,
lb/ton
Removal
tons/yr
Annual
capi tal
Annual
OAH
Cost of
removal,
I/ton
Unloading,
by truck0
Partial
enclosure
90
1. 5e
0.150
675.0
50,000d
10
8,140
12,500d
31
Unloading,
by truck
Total enclo-
sure/fabric
filter
95
1. 5e
0.075
712.5
76,000d
10
12,370
17,000d
41
IMoadlng.
by truck0
Hater spray
50
1. 5e
0.750
375.0
2,b00f
10
410
80f
1
Unloading
by ship"
Total enclo-
sure/fabric
filter
95
0.2e
0.010
95.0
51,600d
10
8.400
11,600d
210
Unloading
by raild
Partial enclo-
sure
70
1.5e
0.450
525.0
9
9
9
9
9
Unloading
by rail"
Total enclo-
sure/fabric
filter
99
1. 5e
0.015
675.0
120,000d
10
19,530
9
9
Unloading
by rail"
Chwtical spray
BO
1.5e
0.300
600.0
37,000d
10
6,020
9
c
Conveying
and trans-
fer'
Partial enclo-
sure
80
o.2d
0.040
80.0
2.500f
10
410
Negligible'
5
Conveying
and trans-
fer'
Total' enclo-
sure/fabric
filter
95
o.2f
0.010
95.0
20,000f
10
3,260
800f
43
Conveying
and trans-
fer'
Water spray
70
0.2f
0.060
70.0
2,500f
10
410
80f
7
(continued)
-------
TABLE 6-2 (continued)
Emissions
Total
capltal
cost, S
E»pected
1 i fe,
y
Cost, 1
Source of
emissions
Control
options
Control,
X
Uncontrolled,
lb/ton
Control led,
1 b/ton
Removalb,
tons/yr
Annual
capital1
Annual
0SH
Cost of
removal,
J/ton
Crushing
and grlnd-
Ingd.f
Partial enclo-
sure
95
0.5f
0.025
237.5
1 ,ooof
10
160
Negltglble'
1
Crushing
and grlnd-
Ingd.f
Total enclo-
sure/fabric
filter
99
0.5f
0.005
247.5
20,000f
10
3.260
8O0f
16
Crushing and
and grind-
1ngd, f
Hater spray
70
0.5f
0.015
242.5
2,500f
10
410
80f
2
Screening
and .
sljlng
Partial enclo-
enclosure
95
0.5f
0.025
237.3
1,000f
10
160
NeglIg1blef
1
Screening f
and siring
Total enclo-
sure/fabric
filter
99
0.5f
0.005
247.5
20,000f
10
3.260
800f
16
loadout
h
h
h
h
h
h
h
h
h
h
aCosts in January 1980 dollars.
''Based on operation for 200 h/yr and a feed rate of 500 tons/h.
cBased on a 10 percent cost of capital.
^Reference 10.
'Reference 11.
fReference 12.
'Data not available.
hSame as for unloading.
-------
6.4 RECOMMENDED RACT
The controls listed in Tables 6-1 and 6-2 are considered
most representative of RACT based on technological and economic
feasibility. In general, the application of current technology
can completely eliminate visible emissions from materials handling
at a reasonable cost. Which specific control will depend on the
size of the facility, the material being handled and the use of
the material. The use of water sprays is precluded if the
moisture content of the material content is important. Because
the cost will be higher at small sources, specific economic
feasibilities should be considered by enforcement agencies on an
individual basis. If the source is vented to a control device,
the use of a fabric filter should be capable of controlling
emissions to less than 0.03 gr/dscf.
The test method to determine mass emissions should be EPA
4 5
Method 5. Opacity should be measured by use of EPA Method 9.
6-12
-------
SECTION 7
SUMMARY
Section 172(b)(2) of the Clean Air Act as amended August
1977, requires that SIP revisions "provide for the implementation
of all reasonably available control measures as expeditiously as
practicable." The use of RACT for stationary sources is defined
as "the lowest emission limit that a particular source is capable
of meeting by the application of control technology that is
reasonably available considering technological and economic
feasibility."1 The purpose of this report has been to identify
control techniques that best represent RACT for particulate
emission sources in TSP nonattainment areas in the State of
Florida. These sources include phosphate process operations;
Portland cement plants; electric arc furnaces; sweat or pot
furnaces; materials handling, sizing, screening, crushing, and
grinding operations.
7.1 RECOMMENDED EMISSION LIMITS
The RACT emission limits recommended in Sections 2 through 6
are presented in Table 7-1. In support of these limits, the
preceding sections have provided the following information for
each of the industry categories:
7-1
-------
ss
%
lit
5
5
5
5
5
5
5
5
5
5
5
a
a
7-1. SUMMARY OF RACT RECOMMENDATIONS
Source
DAP production
ROP/TSP produc-
tion
GTSP production
NSP production
MAP production
AFI (granulation)
AFI (defluorizing)
Phosphate rock
dryers
Phosphate rock
grinding
Ship loading
Rail loading
Kilns
Clinker coolers
Mass emission
1 imit
0.30 lb/ton
of product
0.30 lb/ton
of product
0.20 lb/ton
of product
0.25 lb/ton
of product
0.20 lb/ton
of product
0.30 lb/ton
of product
0.25 lb/ton
of product
0.1 lb/ton
rock handled
0.1 lb/ton
rock handled
0.01 lb/ton
handled
0.01 lb/ton
handled
0.30 lb/ton of
feed
0.20 lb/ton of
feed
7-2
-------
TABLE 7-1 (continued)
Source
category
Mass emission
1 imit
Visible emission
limit, %
Source
Stack
Fugitive
Portland cement
plants
Finish mills
0.02 gr/dscf
5
5
Sweat or pot fur-
naces
Furnace
0.05 gr/dscf
10b
5
Electric arc fur-
naces
Furnace
0.006 gr/dscf
5b
5
Materials handling,
sizing, screening,
crushing, and
grinding operations
All sources
0.03 gr/dscf
5b
5
aNo fugitive opacity standard.
bA higher limit is allowed for one 6-minute period per hour.
7-3
-------
A description of the equipment, operations, and products.
A summary of the various sources of particulate emissions
and control options at these sources.
An estimate of the costs of various control options (both
capital costs and cost in dollars per ton of particulate
controlled are considered).
Discussion of various factors, including technological
advantages and disadvantages, relative costs, and operation
and maintenance considerations that justify the choice of
RACT.
7.2 COMPARISON WITH OTHER EXISTING REGULATIONS
One method to judge the strictness and reasonableness of a
RACT limit is to compare the proposed regulations with existing
regulations. The proposed regulations should be at least as
stringent as generally existing regulations. This-subsection
compares proposed and existing regulations in Region IV for port-
land cement plants, phosphate process operations, electric arc
furnaces, sweat or pot furnaces, and materials handling, sizing,
screening, crushing, and grinding operations.
7.2.1 Portland Cement Plants
I
Four states (Alabama, Florida, North Carolina, and South
Carolina) have specific regulations for control of particulate
emissions from existing portland cement plants. In the other
states, these plants would be subject to control under process
weight rate curves for general processes. New sources in all
states are subject to New Source Performance Standards (NSPS).
Figure 7-1 shows allowable emissions from portland cement plants
in Region IV.
7-4
-------
70
t
Or
60
t/i
S i
tn
v»
ui
-j
eo
3
O
—J
*
tA
0
20
30
5°
"XWCTIW MIt> tons/h
60
70
. i\n i C » lons/h
Figure 7-1. Allowable emissions from portland cement plants
in Region IV.
80
-------
7.2.2 Phosphate Process Operations
None of the states in Region IV has existing regulations
specifically addressing particulate emissions from phosphate
processing operations. Therefore, any plant of this type located
in these states would be subject to process weight rate curves
for general process sources. Figure 7-2 compares these general
regulations with the one proposed for Florida.
7.2.3 Electric Arc Furances
Of the states in Region IV, only Kentucky has a regulation
specifically for existing electric arc furnaces. New electric
arc furnaces in all states are subject to NSPS. A comparison of
these two regulations with the one proposed for Florida is
presented in Table 7-2.
7.2.4 Sweat or Pot Furnaces
No state in Region IV has regulations specifically for sweat
or pot furnaces. Any new sweat or pot furnace at secondary lead
smelters in the region would be subject to NSPS, which prohibit
the discharge of gases with greater than 10 percent opacity from
pot furnaces of more than 550-pounds charging capacity. In
comparison, the proposed emission limits for Florida are 0.05
gr/dscf from existing sweat or pot furnaces and 10 percent
opacity.
7.2.5 Materials Handling, Sizing, Screening, Crushing, and
Grinding Operations
Current regulations in Region IV states do not set specific
emission limits for materials handling, sizing, screening,
crushing, and grinding operations. Rather, they require that
7-6
-------
70
50
8 40
- <$7 IK kFNTUCKY• NORTH CAROLINA.
^ISTlHGigWCpj^
SOUTH
EE <™U
vl
I
v»
(O
<5
30
EXISTING SOURCES IN FLORIDA AND ALABAMA (GENERAL)
20
10
pv.w*2XL
id
PROPOSED
PROPOSED FLORIDA (DRYERS,
- KlU£U
afL
HAP)
oO"»
GRINDERS. ANO MILLS)
50 —=tr—
Mrt, to„s/|i W
'*"•"tsa-SM-
80
-------
TABLE 7-2. ALLOWABLE EMISSIONS FROM ELECTRIC ARC FURNACES
IN REGION IV
Regulation
Emission limitation
Opacity
Kentucky
0.01 gr/dscf
3 percent from control
device,
0 percent from shop
NSPS
0.0052 gr/dscf
3 percent from control
device,
0 percent from shop
Proposed Florida
0.006 gr/dscf
5 percent from control
device, shop
7-8
-------
appropriate control measures be used. These control measures are
referred to in regulations by such phrases as "reasonable pre-
cautions" and "measures to reduce" and are generally followed by
a list of "reasonable precautions." In comparison, the proposed
Florida regulations limit emissions to 0.03 gr/dscf from any en-
closed operation vented through a stack and a 5 percent opacity.
Table 7-3 presents regulations in Region IV.
7-9
-------
TABLE 7-3. FUGITIVE DUST REGULATIONS IN REGION IV
State
Capsulized regulation
Alabama
No person shall . . . [list of activities] without taking
reasonable precautions to prevent particulate matter from
becoming airborne.
Florida
No person shall . . . allow the emissions of particulate
matter from any source whatever . . . without taking rea-
sonable precautions to prevent such emission, except
[emissions covered by other regulations] ....
Georgia
All persons responsible for any operation . . . which may
result in fugitive dust shall take all reasonable pre-
cautions to prevent such dust from becoming airborne.
Kentucky
No person shall . . . [list of activities] without taking
reasonable precautions to prevent particulate matter from
becoming airborne.
Mississippi
a) No person shall cause or permit the handling or
transporting or storage of any material in a manner which
allows or may allow unnecessary amounts of particulate
matter to become airborne.
b) When dust . . . escape(s) from a building or equipment
in such a manner and amount as to cause a nuisance to
property other than that from which it originated . . . the
Commission may order . . . that all air and gases . . .
leaving the building or equipment are controlled ....
c) No person shall . . . allow . . . particulate fallout
to exceed background levels by 5.25 grams/meter squared/
month.
North Carolina
a) Particulates
from mica
or feld-
spar
processing
plants
a) No person shall . . . allow . . . particulate matter
caused by processing of mica or feldspar to be discharged
from any stack ... in excess of ... .
(P < 30), E = 4P0,677
(30
-------
TABLE 7-3 (continued)
State
Capsulized regulation
b) Particulates:
sand,
crushed
stone oper-
tions
a) No person shall . . . [list of activities] without
taking measures to reduce to a minimum any particulate
matter from becoming airborne, and in no case shall
established ambient air quality standards be exceeded at
the property line.
b) The owner . . . shall direct control of the plant
premises and access roads.
c) All stone crushing operations shall employ a water
spray over the crusher.
South Carolina
a) All nonenclosed sources shall be operated in such a
manner, that a minimum of particulate matter becomes
airborne.
b) The owner ... of all sources shall maintain dust
control of the premises and roads ....
c) All crushing, drying, classification and like oper-
ations shall employ a suitable control device ....
Tennessee
No person shall . . . allow [list of activities] without
taking reasonable precautions to prevent particulate
matter from becoming airborne.
Florida
(proposed regu-
lation)
a) £5 percent opacity, except for one six-minute period
per hour which shall not exceed 20 percent.
b) 0.03 gr/dscf from the stack of an enclosed operation.
7-11
-------
REFERENCES
1. U.S. Environmental Protection Agency. Workshop on Require-
ments for Nonattainment Area Plans. Revised ed. April
1978.
2. "Fertilizer." Kirk-Othrner Encyclopedia of Chemical Tech-
nology. 3d ed. Vol. 10. Wiley Interscience, 1978.
3. "Economic Indicators." Chemical Engineering, August 1980.
4. U.S. Environmental Protection Agency. "Method 5, Determina-
tion of Particulate Emissions From Stationary Sources."
Standards of Performance for New Stationary Sources. 1977.
Appendix A, Reference Test Methods.
5. U.S. Environmental Protection Agency. "Method 9, Visual
Determination of the Opacity of Emissions From Stationary
Sources." Standards of Performance for New Stationary
Sources." 1974. Appendix A, Reference Test Methods.
6. U.S. Environmental Protection Agency. Compilation of Air
Pollutant Emission Factors. 3d ed. AP-42. Research
Triangle Park, North Carolina, July 1979. Including
Supplements 1-9.
7. Monsanto Research Corporation. Source Assessment: Phos-
phate Fertilizer Industry. EPA 600/2-79/019c, May 1979.
8. Bohn R., T. Cuscino, Jr., and C. Cowherd, Jr. Fugitive
Emissions From Integrated Iron and Steel Plants. Midwest
Research Institute, Kansas City, Missouri. EPA 600/2-78-050,
March 1978.
9. PEDCo Environmental, Inc. Evaluation of Fugitive Dust
Emissions From Mining. Prepared for U.S. Environmental
Protection Agency. Cincinnati, Ohio.
10. PEDCo Environmental, Inc. Reasonably Available Control
Measures for Fugitive Dust Sources. Prepared for the Ohio
Environmental Protection Agency. Cincinnati, Ohio, March
1980.
R-l
-------
REFERENCES (continued)
11. PEDCo Environmental, Inc. Technical Guidance for Control of
Industrial Process Fugitive Particulate Emissions.
EPA-450/3-77-010, March 1977 .
12. PEDCo Environmental, Inc. RACT Determination for Selected
Process Weight/Fugitive Emissions Categories in Region I.
Prepared for the U.S. Environmental Protection Agency,
Region I. Cincinnati, Ohio, September 1979.
R-2
-------
APPENDIX
SUMMARY OF CONTROL AND EMISSIONS DATA
FOR PHOSPHATE INDUSTRY
A-l
-------
DAP
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
rate
lb/hr
Total
plant
emission
rate
lb/ton
1
50
reactor/granulator
cooler/screens
dryer
venturi
venturi cross flow
venturi
9.50
0.19
2
55
reactor
screens/mills
dryer
venturi/cyclonic
venturi/cyclonic
venturi/cyclonic
2.41
1.75
7.19
0.20
3
70
reactor/granulator
screens/mi 11s/dryer
cooler
venturi/cyclonic
venturi/cyclonic
wet cyclone
1.28
18.88
32.50
0.75
4
35
granulator
screens/dryer
venturi/packed bed
11.00
0.31
5
50
granulator
screens/dryer
venturi/packed bed
12.00
0.24
6
30
dryer/screens
cooler, saturator
blunger
venturi/cross flow
16.00
0.53
7
30
dryer/screens
cooler, saturator
blunger
venturi/cross flow
16.8
0.54
8
98
dryer, reactor
screens, cooler
packed bed/cross
flow
11.00
0.11
9
35
reactor, screen
cooler, dryer
venturi/cross flow
4.70
0.13
10
25
reactor/screens
reactor/screens
cooler
venturi/cyclonic
venturi/cyclonic
wet cyclone
2.19
2.69
21.71
0.64
11
24
reactor/screens
reactor/screens
cooler
venturi/cyclonic
venturi/cyclonic
wet cyclone
2.96
¦ 3.40
'31.8
0.95
A-2
-------
MAP
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
rate
lb/hr
Total
plant
emission
rate
lb/ton
1
25
prill tower
venturi/cross flow
3.0
0.12
cooler
2
14
prill tower
venturi/cross flow
2.69
0.19
A-3
-------
GTSP
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
rate
lb/hr
Total
plant
emission
rate
lb/ton
1
63
reactor, dryer
screens,
granulator
venturi/cross flow
10.88
0.17
2
31
dryer, screens
granulator,
cooler
venturi/cross flow
2.4
0.07
3
66
dryer
8.50
0.12
4
33
dryer/screens
mill/blunger
reactor
venturi/packed
12.4
0.37
5
72
dryer/screens
mill/reactor
venturi/packed
11.00
0.15
A-4
-------
ROP/TSP
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
rate
lb/hr
Total
plant
emission
rate
lb/ton
1
18
be! t
dryer
cyclonic
cyclonic
1.40
4.55
0.32
2
48
belt
dryer
cyclonic
cyclonic
6.50
6.40
0.43
3
40
dryer/screen
belt
venturi/cross flow
8.00
0.27
4
45
den
venturi
2.1
0.04
5
45
den
cyclonic
6.2
0.14
A-5
-------
NSP
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
rate
lb/hr
Total
plant
emission
rate
lb/ton
1
15
curing den
wet impingement
1.90
0.01
2
13
curing den
cyclonic
2.0
0.15
3
15.7
curing den
cyclonic
0.36
0.023
-------
PHOSPHATE ROCK DRYING
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
ra te
Ib/hr
Emission
rate
lb/ton
1
475
f1uidi zed bed dryer
impingemet
scrubber
16.0
0.034
2
350
fluidized bed dryer
cyclonic scrub-
ber
17.8
0.050
3
230
fluidized bed dryer
venturi scrub-
ber
19.2
0.084
4
230
rotary dryer
venturi scrub-
ber
16.4
0.071
5
85
unknown type dryer
venturi scrub-
ber
25.5
0.30
6
270
rotary dryer
venturi scrub-
ber
8.35
0.031
7
470
fluidized bed dryer
cyclonic scrub-
ber
9.10
0.019
8
470
fluidized bed dryer
cyclonic scrub-
ber
14.9
0.032
9
200
unknown type dryer
venturi scrubber
8.2
0.041
10
520
unknown type dryer
cyclonic scrubber
18.6
0.036
11
200
unknown type dryer
venturi scrubber
16.6
0.083
12
500
unknown type dryer
venturi scrubber
28.5
0.057
13
3
rotary dryer
cyclone & scrub-
ber
0.84
0.28
14
330
rotary dryer
venturi scrubber
5.2
0.016
A-7
-------
PHOSPHATE ROCK DRYING MATERIAL HANDLING
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
rate
Ib/hr
Total
plant
emission
rate
lb/ton
1
2700
dry rock transfer
pulse baghouse
2.3
0.00085
2
2700
storage transfer
venturi scrubber
6.84
0.0025
3
420
dry rock transfer
impingement scrub-
ber
0.13
0.00031
4
950
dry rock transfer
cyclonic scrubber
0.50
0.00053
5
296
dry rock transfer
cyclonic scrubber
0.42
0.0014
-------
PHOSPHATE
ROCK GRINDING
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
rate
lb/hr
Total
plant
emission
rate
lb/ton
1
1.1
grinding mill
venturi scrubber
2.2
2.0
2
40
grinding mill
pulse baghouse
27.8
0.69
3
36
grinding mill
pulse baghouse
0.20
0.0055
4
230
grinding mill
pulse baghouse
4.3
0.019
5
230
grinding mill
pulse baghouse
4.3
0.019
6
168
grinding mill
pulse baghouse
0.50
0.003
7
12
Raymond mill
pulse baghouse
0.77
0.063
8
12
Raymond mill
pulse baghouse
1.4
0.11
9
12
Raymond mill
pulse baghouse
0.77
0.063
10
115
Raymond mill
pulse baghouse
2.7
0.023
11
40
ball mill
pulse baghouse
4.17
0.10
12
120
ball mill
pulse baghouse
32.6
0.27
13
120
ball mill
pulse baghouse
30.0
0.,25
14
100
ball mill
pulse baghouse
27.0
0.27
15
35
grinding mill
pulse baghouse
2.0
0.057
16
60
grinding mill
pulse baghouse
6.4
0.11
17
90
ball mill
pulse baghouse
11.0
0.12
18
24
Raymond mill
venturi scrubber
1.7
0.070
A-9
-------
PHOSPHATE ROCK GRINDING MATERIAL HANDLING
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
rate
lb/hr
Total
plant
emission
rate
lb/ton
1
150
ground rock transfer
pulse baghouse
1.9
0.013
2
15
ground rock unloading
pulse baghouse
0.14
0.0093
3
70
ground rock storage
pulse baghouse
0.28
0.004
A-10
-------
PHOSPHATE ROCK RAIL LOADOUT
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
rate
lb/hr
Total
plant
emission
rate
lb/ton
1
750
rail loadout
cyclonic scrubber
1.2
0.0016
2
900
rail loadout
venturi scrubber
18.5
0.021
A-ll
-------
PHOSPHATE ROCK SHIPLOADING
Plant
number
Process
weight
ton/hr
Controlled
sources
Control
devices
Emission
rate
lb/hr
Total
plant
emission
rate
lb/ton
1
2700
shiploader
pulse baghouse
9.85
0.0036
2
800
shiploader
pulse baghouse
0.10
0.00013
3
BOO
shiploader
pulse baghouse
0.20
0.00025
A-12
------- |