United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
May 1982
Air
Guideline Series
Draft
Control of Volatile Organic
Compound Emissions
from Manufacture of
High-Density
Polyethylene,
Polypropylene,
Polystyrene Resins
-------
NOTICE
This document has not been formally released by EPA and should not now be construed to represent
Agency policy. It is being circulated for comment on its technical accuracy and policy implications.
Guideline Series
Control of Volatile Organic
Compound Emissions from
Manufacture of High-Density
Polyethylene, Polypropylene,
Polystyrene Resins
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carofeia 27711
May 1982
-------
TABLE OF CONTENTS
Pa.ge
List of Tables ^
List of Figures. iii
Chapter 1. Introduction 1-1
Chapter 2. Process and Pollutant Emissions 2-1
2.1 I ntroducti on 2-1
2.2 Polypropylene 2-1
2.3 High-Density Polyethylene 2-11
2.4 Polystyrene 2-17
2.5 References for Chapter 2 2-24
Chapter 3. Emission Control Techniques 3-1
3.1 Introduction 3-1
3.2 Combustion Techniques 3-1
3.3 Technical Feasibility of Retrofitting
Control Devi ces 3-20
3.4 References for Chapter 3 3-21
Chapter 4. Environmental Analysis of RACT 4-1
4.1 Introduction 4-1
4.2 Air Pollution 4-1
4.3 Water Pollution 4-3
4.4 Solid Waste Disposal 4-3
4.5 Energy 4-3
Chapter 5. Control Cost Analysis of RACT 5-1
5.1 Basis for Capital Costs 5-1
5.2 Basis for Annualized Costs 5-4
5.3 Emission Control Costs 5-4
5.4 Cost-Effectiveness 5-14
5.5 References for Chapter 5 5-20
Appendix A. Background Data and Information A-l
Appendix B. Cost Analysis B-l
Appendix C. Emission Factors C-l
-------
LIST OF TABLES
Page
2-1 End Uses of Polypropyl ene 2-3
2-2 Polypropylene (PP) Plants in Ozone Non-
attai nment Areas 2-5
2-3 Exhaust Gas Stream Characteristics
for Polypropyl ene PI ants 2-8
2-4 Components of Polypropylene Vent Streams 2-10
2-5 High-Density Polyethylene (HOPE) Plants in
Ozone Nonattai nment Areas 2-14
2-6 Exhaust Gas Stream Characteristics for High-
Density Polyethylene (Liquid-Phase) Plants 2-16
2-7 Polystyrene (PS) Plants in Ozone Non-
attai nment Areas 2-20
2-8 Exhaust Gas Stream Characteristics for Poly-
styrene (Continuous Process) Plants 2-22
4-1 Model Plant Recommendations for RACT Used As A Basis
for Environmental Analysis 4-2
4-2 Additional Energy Required After Control by RACT 4-4
5-1 Cost Factors 5-5
5-2 Cost Adjustments 5-6
5-3 Capital Costing - Thermal Incinerator 5-7
5-4 Polymers and Resins Model Plant Parameters and
Emission Control Costs 5-8
5-5 Cost Analysis for Polypropylene 5-11
5-6 Cost Analysis for High-Density Polyethylene 5-13
5-7 Cost Analysis for Polystyrene (Continuous Process) 5-15
5-8 Cost-Effectiveness of RACT for the Polymers and
Resins Industry 5-16
-------
LIST OF FIGURES
Pa^e
2-1 Simplified Process Block Diagram Polypropylene 2-6
2-2 Simplified Process Block Diagram High-Density
Polyethylene (Liquid-Phase) 2-15
2-3 Simplified Process Block Diagram of Polystyrene
Manufacture by Continuous Process Polymerization 2-21
3-1 Discrete Burner Thermal Incinerator 3-5
3-2 Distributed Burner Thermal Incinerator 3-5
3-3 Steam Assisted Elevated Flare System 3-13
5-1 Prices for Thermal Incinerators Without Heat
Exchanger 5-2
5-2 Prices for Thermal Incinerators With Primary
Heat Exchanger 5-3
B-l Thermal Incinerator Hypothetical Cost as a
Function of Residence Time B-2
B-2 Carbon Steel Straight Duct Fabrication Price per
Linear Foot vs. Duct Diameter and Plate Thickness B-3
IV
-------
1. INTRODUCTION
The Clean Air Act Amendments of 1977 require each State in which
there are areas in which the national ambient air quality standards
(NAAQS) are exceeded to adopt and submit revised State implementation
plans (SIP's) to EPA. Revised SIP's were required to be submitted to
EPA by January 1, 1979. States which were unable to demonstrate attainment
with the NAAQS for ozone by the statutory deadline of December 31, 1982,
could request extensions for attainment with the standard. States
granted such an extension are required to submit a further revised SIP
by July 1, 1982.
Section 172(a)(2) and (b)(3) of the Clean Air Act require that
nonattainment area SIP's include reasonably available control technology
(RACT) requirements for stationary sources. As explained in the "General
Preamble for Proposed Rulemaking on Approval of State Implementation
Plan Revisions for Nonattainment Areas," (44 FR 20372, April 4, 1979)
for ozone SIP's, EPA permitted States to defer the adoption of RACT
regulations on a category of stationary sources of volatile organic
compounds (VOC) until after EPA published a control techniques guideline
(CTG) for that VOC source category. See also 44 FR 53761 (September 17, 1979)
This delay allowed the States to make more technically sound decisions
regarding the application of RACT.
Although CTG documents review existing information and data concerning
the technology and cost of various control techniques to reduce emissions,
they are, of necessity, general in nature and do not fully account for
variations within a stationary source category. Consequently, the
purpose of CTG documents is to provide State and local air pollution
control agencies with an initial information base for proceeding with
their own assessment of RACT for specific stationary sources.
1-1
-------
2. PROCESS AND POLLUTANT EMISSIONS
2.1 INTRODUCTION
The polymers and resins industry includes operations which convert
monomer or chemical intermediate materials obtained from the basic
petrochemical industry and the synthetic organic chemicals manufacturing
industry into polymer products. Such products include plastic materials,
synthetic resins, synthetic rubbers, and organic fibers covered by
Standard Industrial Classification (SIC) codes 2821, 2822, 2823, and
2824. The 1979 production of the major industry polymers was 16,052 Gg.
Thirty-six percent of this total production of the industry is from
the manufacture of high-density polyethylene, polypropylene, and polystyrene.
In addition, the manufacture of these three polymers is estimated to
account for 56 percent of the total estimated industry process emissions
of 86.2 Gg/yr of volatile organic compounds (VOC).
This chapter describes the manufacturing processes for each of
these three polymers under consideration and the associated process VOC
emissions. Fabrication, blending, or formation of resin materials are
not included in the process descriptions, nor are emissions from these
operations quantified. Fugitive and storage emissions from these processes
are described in other CTG documents, "Control of Volatile Organic
Fugitive Emissions from Synthetic Organic Chemical Polymers and Resins
Manufacturing Equipment", and.hence, they are not discussed here.
The model plants in this chapter represents most of existing processes
in the ozone nonattainment areas for each particular resin. The uncontrolled
emission factors can be used as a basis for the verification of VOC
emissions developed from emission source tests, plant site visits,
permit applications, etc. These emission factors should not be applied
in cases where site-specific data are available, but rather, in instances
where specific plant information is lacking or highly suspect.
2.2 POLYPROPYLENE
2.2.1 General Industry Description
Manufacture of polypropylene, on a commercial scale, started in the
1950's when stereospecific catalysts were discovered. Polypropylene is
2-1
-------
a high-molecular weight thermoplastic crystalline polymer of propylene.
The general formula for polypropylene is as follows:
... CH9 - CH - CH9 - CH - CH0 - CH - ...
* \ ^ I * I
CH3 CH3 CH3
The polymer is lightweight, water- and chemical-resistant, somewhat
rigid, and easy to process. It exists in three different forms depending
on the geometric arrangement of the methyl groups: (1) isotactic - with
all methyl groups aligned on the same side of the chain as shown above,
(2) syndiotactic - with the methyl groups alternating, and (3) atactic -
all other forms in which the methyl groups are randomly aligned on
either side of the chain. Typically, commercial samples of polypropylene
contain about 70 percent crystalline material (isotactic), and the
remainder amorphous material (atactic).
Consumer products from polypropylene can be formed in many ways,
including solid molding, extrusion, rotational molding, powder watering,
2
thermoforming, foam molding, and fiber orientation.
Polypropylene resins are supplied in many grades for a variety of
uses. Apart from major distinctions between homopolymer, intermediate-
impact co-polymer, and high-impact co-polymer, the grades may also
differ in specific formulations. Different grades of polypropylene lend
themselves to use in different applications. Molded applications include
bottles for syrups and foods, caps, auto parts, appliance parts, toys,
housewares, and furniture components. Fibers and filaments are used in
carpets, rugs, and cordage. Film uses include packaging for cigarettes,
records, and housewares. Extrusion products include pipes, profiles,
3
wires and cable coatings, and corrugated packing sheets.
Injection molding accounts for 41 percent of polypropylene use;
fibers and filaments account for 31 percent; and other forms account for
4
28 percent. In terms of end uses, major sectors are shown in
Table 2-1.
Production of polypropylene has grown from 981 Gg in 1973 to
1,743 Gg in 1979, a 10.1 percent annual growth rate. C.H. Kline projects
2-2
-------
TABLE 2-1. END USES OF POLYPROPYLENE
Weight Percent
Sector Polypropylene Use
Consumer/Institutional 19
Furniture/Furnishings 18
Packaging 16
Transportation 12
Electrical/Electronics 7
Other 28
2-3
-------
a 9.0 percent growth rate for polypropylene from 1978 to T§83,5 and SRI
International projects an eight percent growth rate from 1977 to 1982.6
Currently, 24 plants produce polypropylene in the United States. The
existing polypropylene plants in ozone nonattafhment areas are listed in
Table 2-2.
2.2.2 Model Pla'nt
The continuous slurry process for manufacture of polypropylene is
the most widely used process commercially. Based on data from 10 existing
plants located in !n'6(nattai'nment areas, the made! plant capacity is
141 'Gg/yr.
The polypropylene resins, characterized by having a controlled
content of isotactic material, are obtained through coordination polymer-
ization, employing a heterogeneous Ziegler-Natta type catalyst system,
which typically is a combination of titanium tetrachloride and aluminum
alkyls. More recent process technology, which uses a high-yield catalyst
with improved activity, requires much less catalyst than the conventional
process. With this high-yield process, the catalyst is left in the
product. This technology results in fewer processing steps and, thus,
less emissions. This new process is incorporated in the model plant by
exclusion of several processing units, and is consistent with a proportional
reduction in the total emission factor.
2.2.2.1 Process Description
The continuous slurry processes, conventional and high-yield, are
represented in Figure 2-1. Reactor feed materials consist mainly of
monomer propylene, comonomer ethylene, monomer impurities propane and
ethane, hexane, and a stereospecific catalyst. Hexane is used as a
process diluent and acts as a heat transfer agent and polymer suspending
medium. The catalyst is usually manufactured on site to consistently
maintatn the required catalyst activity. It is mixed with necessary
solvents and metered accurately into the polymerization reactor along
with Other reactants. Process diluent is also used in catalyst preparation
and spent diluent is sent to the diluent recovery section for reuse.
The reactor is a continuously stirred jacketed vessel or a loop
reactor. During reaction, a portion of the polymer/monomer/diluent
mixture is continuously drawn from the reactor to a flash tank in which
the unreacted propylene and propane are separated, and recovered by
condensation.
2-4
-------
TABLE 2-2. POLYPROPYLENE (PP) PLANTS IN OZONE NONATTAINMENT AREA
Company
ARCO Polymers, Inc.
Amoco Chem. Corp.
Exxon Chem. .Co.
Gulf Oil
Hercules, Inc.
Northern Petrochem.
Co.
USS Novamont Corp.
Phillips Petro. Co.
Rexene Polyolefins
Co.
Shell Chem. Co.
Soltex Polymer Corp.
Location
Deer Park, TX
Chocolate Bayou, TX
Bay town, TX
Cedar Bayou, TX
Bay town, TX
Lake Charles, LA
Morris, IL
La Porte, TX
Pasadena, TX
Odessa, TX
Bayport, TX
Norco, LA
Woodbury, NJ
Deer Park, TX
Status1
NAR
NANR
NAR
NAR
NAR
NANR
NANR
NAR
NAR
NANR
NAR
NANR
NAR
NAR
Capacity
(Gg/yr)
181
125
250
181
272
376
91
159
82
23-46
136
136
91
1 .
Ozone nonattainment area requesting extension (NAR).
SOURCES: SRI International, 1980 Directory of Chemical Producers,
United States.
U.S. EPA study by Pull man-Kellogg Co., plant listing.
The BNA Environmental Reporter AQCR Listing, §121 (through
March 12, 1981).
2-5
-------
YCLE MONOMER AND SOLVENT
1
1
r
r ^
V
BYPRODUCT AND DILUENT "*"
RECOVERY BY EXTRACTION
AND DIS'TILLATION _^
' ' ^ ?E
Ft ACTOR FEED
MATERIALS
"ORAGE
B;
1 CATALYST CO
"*• PREPARATION ~^ POLY
R
^ /
?M|0 577|
\
SrHTl UNREACTED MONOMER
*^ AND DILUENT RECOVERY
BY CONDENSATION
HJ .
L
iTCH OR c. Acuiup
-------
Slurry from the flash tank is then fed to the deactivation/decanting
section for washing with an alcohol-water solution to remove most of the
catalyst residues. The diluent/crude product slurry is lighter than the
alcohol-water solution and the two phases are separated by decantation.
The alcohol-water phase is distilled to recover alcohol; whereas, the
diluent/crude product phase which is in the form of a slurry is stripped
to remove part of the diluent. The product slurry is then sent to a
centrifuge in which the isotactic polymer product solids are separated
from the diluent. The atactic polymer remains dissolved in the diluent.
The isotactic product goes through a product dryer, then is extruded,
pelletized, and sent to product storage.
In the methanol recovery section, the crude methanol streams are
refined and recycled, and the bottom streams, containing catalyst metals
are sent to the plant waste-water treatment facility.
The atactic-diluent solution is fed to the by-product (atactic) and
diluent separation unit in which the diluent is purified and dried for
recycle, and the atactic solids are recovered or burned in incinerators.
In the high-yield slurry process, the catalyst is left in the
product so deactivation/decanting and alcohol recovery sections are
unnecessary. Along with this, one of the major emission streams is also
eliminated. Figure 2-1 indicates the units that should be excluded in
this process.
In addition to the use of high-yield catalysts, other process
variations may occur. Mixtures of aliphatic hydrocarbons may replace
hexane as the process diluent, and isopropyl alcohol may replace methanol
as the catalyst deactivation agent. Also polymer dryers may vary with
the facility, but the fluid bed dryer with hot nitrogen or air is the
most common. Other types of product dryers and different operating
pressures may result in a much higher VOC emission rate. Except for
high-yield catalyst, and the product dryer type and operating pressure,
these other process variations are minor and should have little effect
on the process VOC emissions.
2.2.2.2 VOC Sources
The offgas stream characteristics for polypropylene manufacture are
shown in Table 2-3. The combined process VOC emission factor for the
2-7
-------
TABLE 2-3. EXHAUST GAS STREAM CHARACTERISTICS FOR POLYPROPYLENE PLANTS
ro
CO
Stream
Number Name
A Combined Reactor Vents
B Decanter Vents
C Neutral izer Vents
D Slurry Filter Vents
E Vacuum Jet Exhaust
F Condenser Vents
G Dryer Vents
H Diluent Recovery
H, Catalyst Prep.
H~ Diluent Recovery
Key to Table:
Cont. = Continuous.
HC = Hydrocarbon.
IPA = Lsopropyl alcohol.
Nature
Cont.
Cont.
Cont.
Cont.
Cont.
Cont.
Cont.
Cont.
Cont.
Cont.
TOTAL EMISSION FACTOR:
C, = Propylene and/or any other organic compound with three
CIQHC = A mixture of aliphatic hydrocarbons with 10-12 carbon
Emission Temperature
Factor* ( C) Pressure Composition/
4.07 54.4 Atm. C3> C10HC
11.49 37.8 Atm. Cg, IPA, C1QHC
1.82 32.2-71.1 Atm. C3 , C4HC, C1QHC, IPA
7.93 32.2 Atm. CioHC' IPA
8.72 104 Atm. CioHC' IPA
0.058 26.6-37.8 Atm. cioHC' IPA
Nil - 0.61 85-104 Atm. Air & very small % HC
.65 29.4 Atm. C1QHC, IPA
0.073 29.4 Atm. cioHC' IPA
0.577 29.4 Atm. ClnHC, IPA
34.74
carbon atoms such as propane.
atoms .
Atm. = Atmospheric pressure.
*Emission factor (kg VOC/1000 kg resin).
/Streams are diluted in 10-30 percent nitrogen.
Emission factor range, the range reflects the fluidized bed dryer emission at different operating pressures. Total
emission factor does not include the emission factor of 0.6. Dryer types other than the fluidized bed dryer may
result in higher VOC emissions.
-------
conventional slurry processes is 34.74 kg VOC/1000 kg product. For the
high-yield slurry process, Streams B and C are not present; therefore,
the combined process VOC emission factor for this process is 21.43 kg VOC/1000 kg
product. Most of the emission streams are continuous and consist mainly
of propylene, ethylene, propane, and a small amount of process diluent.
Properties of these compounds are summarized in Table 2-4. The temperature
of the streams varies from ambient to 104°C, and the pressure varies
from atmospheric to 170 psig. Each of the major VOC-containing streams
are indicated on Figure 2-1 and are described below:
1. Stream A: Combined Polymerization Reactor Vents - These emissions
are from vents of reactors from all process trains. This is a continuous
stream venting organic process offgas, consisting mainly of Co (propylene
and other hydrocarbons with three carbon atoms such as propane) and
process diluent, which could be hexane or a mixture of aliphatic hydrocarbons
with 10-12 carbon atoms.
2. Streams B & C: Decanter and Neutralizer Vents - These vents
are part of the alcohol recovery section. This is usually the largest
VOC source in the process and consists of methanol or isopropyl alcohol,
in addition to C, and process diluent. The stream is continuous and
exists in most of the existing polypropylene plants. The process using
a high-yield catalyst does not require these vents, and the reduction in
total emission factor is significant.
3. Stream D: Slurry Filter Vents - This stream is from the centrifuge
which separates the atactic and isotactic polymer. It is one of the
largest VOC emission streams venting process diluent and alcohol remaining
in the polymer. It is a continuous stream at atmospheric pressure and
exists in both the conventional and high-yield slurry process plants.
4. Stream E: Vacuum Jet Exhaust - This stream originates from the
by-product and diluent recovery section and can be the second largest
VOC emission stream in the entire process. The diluent recovery section
which consists of an evaporator, an extractor and distillation units is
part of all processes and emits process diluents and alcohol vapors.
5. Stream F: Condenser Vents - This continuous stream emits only
a small amount of hydrocarbons. It is from the diluent stripping section
and is present only in the conventional slurry process. Emissions are
process diluent and alcohol vapors.
2-9
-------
TABLE 2-4. COMPONENTS OF POLYPROPYLENE VENT STREAMS
Propylene (monomer)
Propane (monomer impurity)
n-Hexane (diluent)
Methanol or Isopropanol
(washing alcohol)
Ethylene (comonomer)
Cp-Cj- Hydrocarbons (might include
etnylene, propylene, and propane)
C-JQ H.C. (A mixture of aliphatic
Hydrocarbons with 10-12 carbon
atoms.)
MW = 42.06, 2186 Btu/cu ft
MW = 44.09, 2385 Btu/cu ft
MW = 86.17, 4412 Btu/cu ft
MW = 32.04 or 60.02
MW = 28.05, 1513 Btu/cu ft
MW = 50 (Avg)
MW = 144.0
All of these compounds are usually diluted in gases like:
Air MW = 29.0
Nitrogen MW = 28.0
Hydrogen MW = 2.0, 275 Btu/cu ft
2-10
-------
6. Stream G: Dryer Vents - This vent emits hydrocarbons diluted
in air at a relatively high temperature (104°C) and atmospheric pressure.
The emissions consist of vapors of hexane, methanol, and propane.
7. Streams HI & H2: Catalyst Preparation and Diluent Recovery -
Both of these streams are continuous and release process diluent that
is used in preparation of catalyst. Stream H2 is from the diluent
recovery section which consists of various separation units.
The stream properties and VOC concentrations of Streams A to H can
vary depending on process conditions. The variation generally depends
on the product grade or type being manufactured and other variables such
as temperature, pressure, catalyst concentration or activity, and the
amount of hydrogen used for molecular weight control. The concentration
and the magnitude of each stream is, of course, highest under start-up
or shutdown conditions because of process conditions away from equilibrium.
2.2.2.3 Control Systems
No controls are routinely applied for VOC control of these continuous
sources. The polymerization reactors and the atactic separation units,
however, are generally provided with emergency relief valves leading to
a flare for safety purposes in the case of upsets. These emergency
vents usually pass through knock-out drums to separate entrained liquids
and polymer particles before the vapors are piped to the flare. Also,
in the production steps, the concentrated atactic polymer stream from
the centrifuge is led to a vessel and its liquid content is removed by
evaporation. The solid amorphous atactic polypropylene is left behind
and is then either burned in incinerators or is packed and sold as a by-
product for paper coating and other applications. For some producers,
the atactic content is very low, and it is left in the product. When
the atactic polymer is incinerated, liquid and gaseous waste streams
from the process may also be burned in the same device.
2.3 HIGH-DENSITY POLYETHYLENE
2.3.1 General Industry Description
High-density polyethylene (HOPE) resins are linear thermoplastic
polymers of ethylene with densities higher than 0.94 g/cm . HOPE resins
are typically produced by a low-pressure process in which organic solvents
are used; the solid catalyst is in suspension; and the polymer forms a
2-11
-------
slurry (e.g., the processes originated by Phillips Petroleum Company and
Solvay and Cie, sa). Although there are various solvent processes used,
the variations do not affect emissions except with respect to the solvent
recovery methods used.
HOPE is a highly (>90 percent) crystalline polymer containing less
than one side chain per 200 carbon atoms in the main chain. The typical
O Q
density range is 0.95-0.97 g/cm . It is strong, water- and chemical-
resistant, and can be easily processed. It is one of the largest volume
plastics produced in the U.S. and in the world. It is extruded into
film sheets, pipe or profiles, coated, injection molded, blow molded,
9
rotationally molded, foamed, or formed in other ways.
HDPE's primary application is blow molded bottles for bleaches,
liquid detergents, milk, and other fluids. Other blow molded forms for
which HDPE's are used include automotive gas tanks, drums, and carboys.
HDPE's also are used for injection molded forms including material
handling pallets, stadium seats, trash cans, and auto parts. Film is
used in making shopping bags. Forty percent of all HOPE is blow molded;
another 22 percent is injection molded. Film and sheet combined account
for only six percent of HOPE use. Other uses account for 32 percent.
End use sectors for HOPE include packaging (45 percent), consumer/
institutional (11 percent), building and construction (9 percent), and
other sectors (35 percent).
From 1973 to 1979, production of HOPE grew from 1,196 Gg to 2,273 Gg,
a growth rate of 11.3 percent. C.H. Kline projects growth at 7.0 percent
for 1978 to 1983.11 SRI International projected growth from 1976 to
1980 at 10 percent.12
2.3.2 Model Plant
The Phillips particle form process serves as the basis for this
model plant, but it is intended to represent all other liquid-phase
processes with high-efficiency catalysts that do not require catalyst
recovery.
This model plant specifically includes an unreacted monomer recycling
system and a closed loop nitrogen drying system. There are other similar
liquid-phase processes that do not use such systems and have larger
2-12
-------
emissions. The plant capacity for the model HOPE plant is 214 Gg/yr.
This is based on plants located in nonattainment areas. The existing
HOPE plants in ozone nonattainment areas are listed in Table 2-5.
2.3.2.1 Process Description
Referring to the schematic for this process, Figure 2-2, the feed
section includes catalyst purification and activation. The prepared
catalyst is then fed to the reactor continuously by being slurried in a
stream of process diluent (pentane or isobutane). Ethylene monomer and
comonomer (butene-1 or hexene), after purification, are also fed to the
reactor where polymerization takes place in process solvent. The reactor,
for the particle-form process, is usually a closed loop pipe reactor.
The product HOPE is separated from unreacted monomer and diluent by
flashing from a low pressure to a vacuum and by steam stripping. The
wet polymer solids are dewatered in a centrifuge and then dried in a
closed-loop nitrogen drying system prior to extrusion.
The unreacted monomer and diluent vapors are sent through a diluent
recovery unit where most of the diluent is separated and recycled back
to the reactor. The rest of the stream is then sent to the ethylene
recovery unit where ethylene is recovered and sent to recycle ethylene
treaters and back to the reactor.
2.3.2.2 VOC Sources
All the process streams, except the feed preparation stream, in
HOPE manufacture are continuous, and they consist mainly of ethylene and
process solvent diluted in nitrogen or air. Most of the streams are at
ambient temperature. An ethylene safety flare is always a part of each
system, and some plants may use it for VOC emission control. Since this
particular model plant incorporates ethylene recycle and closed-loop N~
drying systems, it has relatively small emissions, but plants which vent
unreacted monomer and use simple dryers have substantially higher VOC
emissions. The major VOC source is the flash tank where an unreacted
monomer stream (about 50 percent VOC) is released. These manufacturers
often send this stream to a boiler to recover the heat value. Table 2-6
shows the vent stream characteristics for the VOC sources; these sources
are described below:
2-13
-------
TABLE 2-5. HIGH-DENSITY POLYETHYLENE (HDPE) PLANTS IN OZONE
NONATTAINMENT AREAS
Company
Allied Chem. Corp.
ARCO Polymers, Inc.
Cities Service Co.
Dow Chemical
Amoco Chem. Corp.
E.I. Du Pont de
Nemours & Co. Inc.
Gulf Oil Corp.
Hercules, Inc.
Nat'l. Petrochem.
Corp.
Phillips Petro. Co.
Soltex Polymer Corp.
UCC
Location
Baton Rouge, LA
Port Arthur, TX
Texas City, TX
Freeport, TX
Plaquemine, LA
Chocolate Bayou, TX
Orange, TX
Victoria, TX
Orange, TX
Lake Charles, LA
La Porte, TX
Pasadena, TX
Deer Park, TX
Port Lavaca, TX
Status1
NANR
NANR
NANR
NANR
NANR
NANR
NANR
NANR
NANR
NANR
NAR
NAR
NAR
NANR
Capacity
(Gg/yr)
272
147
82
136
136
159
104
102
200
7
227
420
270
181
Ozone nonattainment area not requesting extension (NANR).
Ozone nonattainment area requesting extension (NAR).
SOURCES: SRI International, 1980 Directory of Chemical Producers,
United States.
U.S. EPA study by Pullman-Kellogg Co., plant listing.
The BNA Environmental Reporter AQCR Listing, §121 (through
March 12, 1981).
2-14
-------
IX)
_j
CJl
I
REACTOR FEED
AND CATALYST
PURIFICATION
.
UNCONTROLLED VOCSTREAMS
M| AND EMISSION FACTORS
(kg/1000 kg product)
RECYCLE
ETHYLENE TREATER
ETHYLENE
RECOVERY UNIT
RECYCLE
DILUENT
TREATER
DILUENT
RECOVERY
UNIT
CLOSED PIPE
LOOP REACTOR
UNREACTED MONOMER
AND DILUENT SEPARATION
FROM PRODUCT
CENTRIFUGE
I
CLOSED LOOP
NITROGEN
DRYING SYSTEM
POLYMER
FLUFF STORAGE
AND MIXERS
PELLETIZING
AND
PACKAGING
Figure 2-2. Simplified process block diagram high-density polyethylene (liquid-phase).
-------
TABLE 2-6. EXHAUST GAS STREAM CHARACTERISTICS FOR HIGH-DENSITY POLYETHYLENE
(LIQUID-PHASE) PLANTS
IX)
O-v
Stream
Number
A
B
C
0
Name Nature
Feed Preparation Intermittent
Dryer Nitrogen Continuous
Continuous Mixer Continuous
Recycle Treaters Continuous
TOTAL EMISSION FACTOR:
Emission
Factor* scfm
0.193 104,900.07
0.062 269.3
0.0053 10.89
12.3 166
12.56
Temperature
(Sc)
21.1
21.1
21.1
21.1
Composition
(wt. %)
100.0 Ethylene
0.3 Isobutane
99.7 Nitrogen
0.6 Isobutane
99.4 N2
61.0 Ethylene
18.0 Isobutane
20.0 Ethane
1 .0 Hydrogen
NOTES:
Pressure unknown.
*EmiSbion factor (kg VOC/1000 kg resin).
+Total volumetric flowrate of (VOC & non-VOC) stream at 0°C, 1 atm.
/scf/occurrence, 12 occurrences per year.
-------
1. Stream A: Feed Preparation - This is an intermittent stream
consisting mostly of ethylene. Assumed to vent 12 times a year, it's
sources are drying, dehydrating and other feed purification operations.
2. Stream B: Dryer Nitrogen - This closed-loop drying system has
very low emissions. It is continuous and is a dilute stream of process
solvent in nitrogen.
3. Stream C: Continuous Mixer - This is another low VOC emission
stream coming from a mixer which mixes polymer with anti-oxidants. It
is continuous and releases process solvent that is still left in the
polymer along with a large quantity of nitrogen. Usually this stream is
emitted to the atmosphere.
4. Stream D: Recycle Treaters - This is a semi-continuous VOC
emission stream with about 80 weight percent VOC. Currently this stream
is usually flared. Treaters consist of vessels containing such materials
as adsorbents, dessicants, and molecular sieves which remove water and
other impurities in the recycle ethylene stream. Emissions occur when
the vessels are purged during regeneration of the adsorber beds. This
stream is considered a continuous stream. For cost analysis purposes,
the stream continuously flows for 20 out of 24 hours.
2.3.3 Control Systems
As noted, like the other polyolefin processes, the HOPE process
generally has a flare as a part of the system for safety reasons. A
complete line of safety relief devices leading to the flare are commonly
provided to avoid accidents as a result of equipment overpressurization
or malfunction.
2.4 POLYSTYRENE
2.4.1 General Industry Description
Polystyrene offers a combination of excellent physical properties
and processibility at a relatively low price for thermoplastic materials.
It is crystal clear and has colorability, rigidity, good electrical
properties, thermal stability, and high-flexural and tensile strengths.
Polystyrene products are used in molded forms, extrusions, liquid solutions,
adhesives, coatings, and foams. The family of polymerized co-polymers
from styrene monomer and its modifications ranked third in all plastics
consumptions in the United States.
2-17
-------
Molded uses include toys, autoparts, housewares, kitchen items,
appliances, wall tiles, refrigerated food containers, radio and television
housings, small appliance housing, furniture, packages, and building
components such as shutters. Extruded sheets also are used in packaging,
appliances, boats, luggage, and disposable plates. Foamed styrene is a
good insulator and is used in construction, packaging, boats, housewares,
14
toys, and hot/cold insulated drink cups. Fifty percent of all styrene
is used in moldings. Extrusion accounts for 33 percent. Other forms
make up 17 percent.
Of end use sectors, packaging makes up 35 percent, consumer/
institutional - 22 percent, building and construction - 10 percent,
15
electrical and electronic - 10 percent, and other sectors - 23 percent.
Production of styrene has grown from 1,507 Gg in 1973 to 1,817 Gg
in 1978, a 3.2 percent growth rate. C.H. Kline projects a 6.0 percent
growth rate for 1978-198316 while SRI International projects a 4.9
percent growth rate for 1979-1982.17
Styrene polymerizes readily with the addition of either heat or
catalyst like benzoyl peroxide or ditertiary butylperbenzoate. Styrene
will homopolymerize in the presence of inert materials and co-polymerize
with a variety of monomers. Pure polystyrene has the following structure:
H H H H
till
- C - CH2 - C CH2 - C - CH2 - C -
Although polymers with molecular weights in the millions can be
made, those most useful for molding have molecular weights of about
125,000; while those used in the surface coating industry average about
35,000.
2.4.2 Model Plant
A continuous process for the manufacture of polystyrene was chosen
for developing the model plant primarily because of its significant VOC
emissions. Mass (bulk) polymerization was used as a basis for developing
the flow diagrams. However, the model plant represents all liquid-phase
2-18
-------
continuous processes. In the case of suspension polymerization, because
polymerization takes place in water, dewatering, washing, centrifuge and
dryer sections are required. These sections usually are not sources of
VOC emissions. The model plant capacity is 73.5 Gg/yr. This capacity
represents an average of capacities from polystyrene plants using batch
or continuous processes in ozone nonattainment areas. The existing
polystyrene plants in ozone nonattainment areas are listed in Table 2-7.
The list includes both continuous and batch-type processes; when the
process type is unknown the process comment is left blank. The plants
with unknown process type are included for completeness of the list.
Only the continuous processes are covered by RACT.
2.4.2.1 Continuous Process
1. Continuous Process Description - This description is for a
fully continuous, thermal co-polymerization process for the manufacture
of pelletized polystyrene resin from styrene monomer and polybutadiene.
Several grades of crystal and impact polystyrene are produced by this
process. The continuous process is represented in Figure 2-3.
Styrene, polybutadiene, mineral oil, and minor amounts of recycle
polystyrene, anti-oxidants and other additives are introduced into the
feed dissolver tank in proportions that vary according to the grade of
resin being produced. Blended feed is pumped on a continuous basis to
the reactor where the feed is thermally polymerized to polystyrene. The
polymer melt, containing some unreacted styrene monomer and by-products
is pumped to a vacuum devolatilizer where most of the monomer and by-
products are separated, condensed and sent to a styrene recovery unit.
Vapors from the styrene condenser are vented through a vacuum system.
Molten polystyrene from the bottom of the devolatilizer is pumped
through a stranding die-plate into a cold water bath. The cooled strands
are pelletized and sent to product storage.
In the styrene recovery unit, crude styrene monomer is separated in
a distillation column. The styrene vapor off the tower is condensed and
recycled to the feed dissolver tank. Noncondensibles are vented through
a vacuum system. Heavies from the column can be used as fuel make-up.
2. Continuous Process VOC Sources^ - Table 2-8 shows the vent
stream characteristics for the continuous polystyrene process. All VOC
emission streams from the process are continuous. Steam present in
2-19
-------
TABLE 2-7. POLYSTYRENE (PS) PLANTS IN OZONE NONATTAINMENT AREAS
Company
A.E. Plastik Pak Co. , Inc.
Am. Hoechst Corp.
Amoco Chemical Corp.
ARCO Polymers, Inc.
BASF Wyandotte Corp.
Carl Gordon, Ind. , Inc.
Cosden Oil & Chemical Co.
Crest Container Corp.
Dart Ind. , Inc.
Dow Chemical Corp.
Gulf Oil Chemical Co.
Mobil Chemical Co.
Monsanto
Polysar Resins , Inc.
Richardson Company
Shell Chemical Co.
Sterling Plastics Corp.
Location
City of Industry, CA
Chesapeake, VA
Leominster, MA
Joliet, IL
Torrance, CA
Willow Springs, IL
Monaca, PA
Jamesburg, NJ
South Brunswick, NJ
Owensboro, KY
Oxford, MA
Worchester, MA
Windsor, NJ
Calumet, City, IL
Saginaw, TX
Fort Worth, TX
Bayport, TX
Allyns Pt. , CT
Midland, MI
Torrance, CA
Marietta, OH
Channelview, TX
Holyoke, MA
Joliet, IL
Santa Ana, CA
Addyston, OH
Decatur, AL
Long Beach, CA
Springfield, MA
Copley, OH
Leominster, MA
Channelview, TX
Belpre, OH
Windsor, NJ
Status1
NAR
NANR
NAR
NANR
NAR
NAR
NAR
NAR
NAR
NAR
NAR
NAR
NAR
NAR
NANR
NANR
NAR
NAR
NANR
NAR
NANR
NAR
NAR
NANR
NANR
NAR
NANR
NAR
NAR
NANR
NAR
NANR
NANR
NAR
Capacity
(Gg/yr)
16
91
54
136
16
41
238
136
50
1
> 68
J
54
120
14
3.6
68
82
100
91
102
18
45
20
34
136
45
23
136
82
52
141
13.6-54.4
Process
Comment*
Continuous
Batch
Batch
Batch
Batch
Batch
Continuous
Batch
-
Continuous
Continuous
Continuous
Continuous
Continuous
Continuous
-
Continuous
Continuous
Continuous
Continuous
*0nly continuous processes are covered by RACT.
Ozone nonattainment area not requesting extension (NANR).
Ozone nonattainment are requesting extension (NAR).
SOURCES: SRI International, 1981 Directory of Chemical Producers, United States.
U.S. EPA study by PulIman-Kellogg Co., plant listing.
The BNA Environmental Reporter AQCR Listing, •il21 (through March 12, 1981).
2-20
-------
UNCONTROLLED VOCSTREAMS
AND EMISSION FACTORS
(kg/1000 kg product)
INJ
Q
REACTOR FEED
MATERIALS
STORAGE
ISOTHERMAL
FEED DISSOLVER
(Mixer)
MASS (Bulk)
POLYMERIZATION
CONTINUOUS REACTOR
Figure 2-3. Simplified process block diagram of polystyrene manufacture by continuous process polymerization.
-------
TABLE 2-8. EXHAUST GAS STREAM CHARACTERISTICS FOR POLYSTYRENE (CONTINUOUS PROCESS) PLANTS
ro
rv>
Stream
Number
A
B
C
D
Name
Tankage/
Styrene Condenser Vent
Styrene Recovery Unit
Condenser Vent
Extruder Quench Vent
TOTAL
Nature
Cont.
Cont.
Cont.
Cont.
EMISSION FACTOR:
Emission
Factor*
0.11
2.96
0.133
0.15
3.35
scfm
0.128
102.30
(acfm)
56.85
1.01
Temperature
--
98.9
98.9
21.1
Composition
Pressure (wt. %)
Styrene
Atm. 21.8 Styrene
78.2 Steam
Atm. 2.1 Styrene
97.9 Steam
99.99 Steam
Trace Styrene
NOTE_S_:
*Emission factor (kg VOC/1000 kg resin).
+Total flowrate of (VOC & non-VOC) stream.
/Combined stream from styrene monomer storage and feed dissolver tanks.
-------
Streams B and C reflects the use of a steam jet ejector in the vacuum
system used. The sources of these VOC streams are described below:
a. Stream A: The combined monomer storage and feed dissolver
vent consists of pure styrene. The VOC emission results from
breathing and washing losses. Currently, the styrene is
emitted to the atmosphere.
b. Stream B: The styrene condenser vent consists of unreacted
styrene separated from the polystyrene in a vacuum devolatilizer.
The stream is exhausted through a vacuum system (steam jet
ejector), to the atmosphere. This is the largest VOC source.
When vacuum pumps are used and followed by refrigerated brine
condenser, the emissions can be lower.
c. Stream C: The styrene recovery unit condenser vent stream
contains the noncondensible components separated in the styrene
recovery tower and is vented through a steam ejector.
d. Stream D: The extruder quench vent consists of steam and
trace of styrene vapor. The stream is usually vented through
a forced-draft hood and passed through demister-pad or electrostatic
precipitator before venting to the atmosphere.
2.4.2.2 Control Systems
No routine control is applied to continuous processes other than
normal condensation operations. One unique system, however, of vapor
condensing/recovery is used where each process vessel is equipped with
rupture discs having the respective pressure relief settings. When any
of these process vessels are overpressured, the vapors relieve to the
vapor condensing/recovery system. By flashing action and by condensation,
most of the vapors are condensed, recovered, and reused in the process.
This system also results in a single emission point in the entire process.
Unlike the polyolefins processes, no flares are used as control devices.
2-23
-------
2.5 REFERENCES FOR CHAPTER 2
1. Click, C.N. and O.K. Webber, "Polymer Industry Ranking by VOC Emissions
Reduction that Would Occur from New Source Performance Standards",
Pullman Kellogg Company, EPA Contract No. 68-02-2619, 1979, p. 174.
2. Kline, C.H. and Company, "The Kline Guide to the Plastics Indus.try",,
Fairfield, New Jersey, 1978.
3. SRI International, "Facts and Figures of the Plastics Industry", New
York, New York, 1978.
4. Ibid.
5. Kline, C.H. and Company, "Plastics and Resins: Forecast to 1983",
Chemical and Engineering News reprint, 1979.
6. SRI International, Chemical Economics Handbook.
7. Click, C.N., op. cit., pp. 177-178.
@
8. Billmeyers, F.W., "Textbook of Polymer Science", Wiley Interscience,
New York, 1971, pp. 379-386.
9. Kline, -C.H., 1978, op. cit.
10. SRI International, 1978, op. cit.
11. Kline, C.H., 1979, op. cit.
12. SRI International, Handbook, op. cit.
13. SRI International, "The Story of the Plastics Industry",, 1977.
14. Kline, C.H., 1978, op. cit.
15. SRI International, 1978_, op. cit.
16. Kline, 1979, op. cit.
17. SRI International, Handbook, op. cit.
2-24
-------
3. EMISSION CONTROL TECHNIQUES
3.1 INTRODUCTION
This chapter describes the emission control techniques applicable
to the control of continuous volatile organic compound (VOC) emissions
from the production of high-density polyethylene (HOPE), polypropylene,
and polystyrene. Emission reduction effectiveness, parameters influencing
effectiveness, and mode of operation are discussed for each control
technique along with the factors likely to exist in plants which could
affect control method applicability and efficiency.
The vent streams from the manufacture of these polymers and resins
are diverse in both composition and flow. Both relatively high and low
concentration VOC streams are present within the industry, and there are
streams with both continuous and intermittent flows. These vent stream
differences extend from process-to-process; also different types of
control techniques may be applicable to controlling individual vent
streams. Condensation, absorption, adsorption, and other product recovery
techniques may be effective control techniques for many continuous
streams. A single optimum control technique applicable to every VOC-
containing vent stream, however, cannot be identified. The most common
control techniques form the basis for this chapter.
Condensers, catalytic oxidizers, and carbon adsorbers were not
evaluated in the present study because of several reasons. Condensation
is limited to reasonably concentrated VOC streams. Even with concentrated,
low vapor pressure VOC streams, the removal efficiency of condensation
may be, at the most, 95 percent. Since combustion techniques can attain
higher efficiencies with few limitations, they were considered preferable
to condensation techniques.
Carbon adsorption, as a control technique for VOC, is used on waste
gas streams of low VOC concentration where devices such as condensers or
scrubbers are ineffective or uneconomical. The VOC concentration in the
streams treated in the adsorber is usually well below the lower explosive
limit mainly to prevent bed explosions by limiting the heat of adsorption.
This is not the case in the polymer industry. Moreover, most compounds
found in emission streams considered here have high vapor pressure and,
3-1
-------
thus, are not easily adsorbable. If adsorbed, these compounds could
polymerize and cause clogging in the carbon bed;.
3.2 COMBUSTION TECHNIQUES
The three major combustion devices that can be used to control VOC
emissions from the polymers industry are thermal incinerators, boilers,
and flares. Flares are presently the most widely used control device
for both the continuous and intermittent streams from polyethylene and
polypropylene. However, boilers and thermal incinerators also can be
used in these industries to control the continuous vent streams. These
devices are presently being used in the industry but to a lesser extent
than flares. Boilers and incinerators are unable to handle large-volume
intermittent streams.
Catalytic incineration may be favored over thermal when the supplemental
fuel cost is substantially high. By catalytic oxidation, the VOC can be
destroyed at a lower temperature and therefore with less fuel. Since
the streams involved in the polymers and resins industry are rich enough
to self-combust, practically no advantage is achieved by using a catalytic
unit. On the contrary, higher capital investment and higher annualized
costs could result from its use.
3.2.1 General Combustion Principles
Combustion is a rapid oxidation process, exothermic in nature,
which results in the conversion of VOC to carbon dioxide and water in
presence of excess air. This process is summarized by the following
equation:
ccH C + 3N2 + T02 + T3.76 N2 + eH20 = (& + %*•) H20 + ayC02 + n02
+ (3 + T3.76) N2 + heat of combustion
where:
a = moles of VOC in vent stream
3 = moles of inerts in vent stream (assumed all N2)
T = moles of oxygen in combustion air (3.76 times as much N* as 0?
in air)
3-2
-------
n = moles of oxygen in flue gas after combustion = T - - ay
e = moles of water, if any, in moist vent streams
The above equation assumes that complete combustion of one type of
organic compound occurs and that the stream heating value is sufficient
to maintain a stable flame. In reality, complete combustion is hard to
achieve even in presence of excess air. Most vent streams are composed
of various types of organic compounds and inerts and destruction of
organics by thermal oxidation is a complicated process. Also, the
heating values sometimes can be inadequate to sustain combustion. The
addition of an auxiliary fuel to maintain flame stability and temperature
will result in the need for more air to combust the fuel and in the
production of more water and carbon dioxide. The inerts (Ng) from the
vent stream and combustion air comprise the major part of the matter
moving through a combustion device.
Combustion of VOC in either an incinerator, a flare, or a boiler is
influenced by time, mixing, and temperature. Any effective combustion
device must provide:
1. Intimate mixing of combustible material (VOC) and the oxidizer
(air),
2. Admission of sufficient air, within the limits of flammability,
to burn the VOC,
3. Sufficient temperature to ignite the VOC/air mixture and
sustain its combustion, and
4. Sufficient residence time at the appropriate temperature.
Combustion temperature is an important parameter in the design of a
combustion device because .oxidation rates are highly temperature-dependent.
Incineration of low heating value offgas necessitates the burning of an
auxiliary fuel to achieve the desired temperature.
Mixing is crucial in achieving good combustion device performance.
A properly designed unit will rapidly combine the offgas, the combustion
air, and, if present, the hot combustion products from the burner to
ensure that the VOC will be in contact with sufficient oxygen at a
temperature high enough to start the oxidation reaction. Improper
mixing can allow packets of waste gas to pass through the unit intact
3-3
-------
and can lead to poor temperature distributions where not all the waste
gas stream reaches or remains at the ctimbustion temperature.
To these main fundamental requirements may be added many special
requirements, depending on the application. The implementation of each
of these combustion factors is accomplished differently with varying
efficiency by the three combustion devices discussed below.
3.2.2 Thermal Incinerators
3.2.2.1 Thermal Incinerator Design
Thermal incineration systems use direct-fired burners to burn
combustible gases to carbon dioxide and water vapor. In these systems,
the vent gas stream is delivered to a refractory-lined chamber by the
process exhaust system or by a self-contained blower. Discrete dual
fuel burners and inlets for the offgas and combustion air are arranged
in the combustion chamber to thoroughly mix the hot combustion products
from the burners with the offgas and combustion air streams. The mixture
of hot reacting gases then passes into the reaction section which is
sized to allow the mixture enough time at the elevated temperature for
the combustion reaction to reach completion. Energy can then be recovered
from the hot flue gases in the heat recovery section. Preheating the
combustion air is a common method of energy recovery; however, it is
more economical to generate steam because steam generators have higher
heat transfer coefficients.
Thermal incinerators designed specifically for VOC incineration
with natural gas as an auxiliary fuel use grid-type (distributed) gas
burners instead of the conventional dual fuel, forward flame, discrete
burners. With grid-type burners the tiny gas flame jets on the grid
surface ignite the fume as it passes through the grid to ensure burning
2
of all the fumes at lower chamber temperatures. Typical configurations
are shown in Figures 3-1 and 3-2.
Packaged, single-unit thermal incinerators can be built to control
streams with flowrates in the range of a few hundred scfm to about
50,000 scfm. A typical thermal incinerator built to handle a VOC waste
stream of 850 m3 (30,000 scfm) at a temperature of 870°C (1600°F) with
0.75 second residence time probably would be a refractory-lined cylinder.
With the typical ratio of flue gas to waste gas of about 2.2, the chamber
3-4
-------
Waste Gas
Auxiliary
Fuel Burner
(discrete)
Waste Gas
Stack
Mixing
Section
Combustion
Section
Optional
Heat
Recovery
Figure 3—1. Discrete burner thermal incinerator.
Burner Plate-i Flume Jets
(natural gas)
Auxiliary Fuel
Stack
Optional
Heat
Recovery
Figure 3-2. Distributed burner thermal incinerator.
-------
volume necessary to provide for 0.75 second residence time at 870°C
(1600°F) would be about 99 m3 (3500 ft3). If the ratio of the chamber
length to the diameter is two and if a 30.5 cm (1 ft) wall thickness is
allowed, the thermal incinerator would measure 8.3 m (27 ft) long by 4.6 m
(15 ft) wide, exclusive of heat exchangers and exhaust equipment.
3.2.2.2 Parameters Affecting VOC Destruction Efficiency
The VOC destruction efficiency of an incinerator can be affected by
variations in chamber temperature, residence time, inlet concentration,
compound type, and flow regime (mixing). Test results show that a VOC
control efficiency of 98 percent can be consistently achieved by well-
designed units. The 98 percent destruction efficiency can be met under
3
various combinations of incinerator operating parameters.
The parameters, chamber temperature, residence time, and flow
regime are the most important parameters affecting VOC destruction
efficiency. Although a combustion chamber temperature of 870°C (1600°F)
was chosen for the analysis, test results show that 98 percent destruction
efficiency is sometimes achievable at temperatures of 700°C (1300°F) to
800°C (1500°F) and residence times of 0.5 to 1.5 seconds. Temperatures
higher than 870°C (1600°F) are not desirable due to the materials limitations
of metallic heat exchangers.
At temperatures over 760°C (1400°F) the oxidation reaction rate is
much faster than the rate at which mixing takes place so VOC destruction
becomes more dependent upon the fluid mechanics within the combustion
chamber. The flow regime should be such that the mixing of VOC stream,
combustion air, and hot combustion products from the burner is rapid and
thorough. This will enable the VOC to attain the combustion temperature
in the presence of enough oxygen for a sufficient period of time for the
oxidation reaction to reach completion. Chamber design, burner, and
baffle configurations provide for turbulent flow for improved mixing.
Variations in inlet concentrations also affect the VOC destruction
efficiency achievable with thermal incinerators. Kinetics calculations
describing the combustion reaction mechanisms point to much slower
reaction rates at very low compound concentrations. Therefore, at low
concentration, a greater residence time is required to achieve a high
combustion efficiency.
3-6
-------
3.2.2.3 Factors Affecting Applicability and Reliability
Thermal incinerators are applicable to a wide variety of continuous
waste gas streams. The design and applicability of the unit depend on
the composition and flow. Also, the heating value per volume of waste
gas, together with the waste gas flowrate, influences the design and
auxiliary fuel usage.
For waste gases with low heat contents, auxiliary fuel such as
natural gas or fuel oil must be added to maintain the combustion temperatures.
Heat contents of approximately 13 and 20 Btu/scf in air correspond to 25
and 40 percent of the lower explosive limit (LEL). Waste gases with
heating values of 20 to 50 Btu/scf (40 to 100 percent of the LEL) must
be diluted with inert gases to bring the concentration down to 25 percent
of the LEL or be enriched with auxiliary fuel to put the concentration
above the UEL to bring the concentration within the flammable safety
limits. Moderate (50 to 150 Btu/scf) heat content waste gases have
sufficient heat content for burning but need auxiliary fuel for flame
stability.
When the heat content is higher than ~150 Btu/scf, the waste gas
possesses enough heating value to support a flame by itself on a flame
burner if the burner is properly designed and can be considered for use
as a fuel gas or boiler feed gas. When flame temperatures resulting
from incineration of this type of waste exceeds 1200°C (2200°F), enough
excess air must be used to cool the unit to 1200°C (2200°F). Incineration
equipment such as water-wall boilers and high-temperature specialty
incinerators has been successfully designed and operated for temperatures
in excess of 1200°C (2200°F), but discussion of this equipment is beyond
the scope of this study.
Conventional thermal incinerators range in size from a unit capable
of controlling several hundred scfm of waste gas to single or multiple
units controlling waste gas in excess of 100,000 scfm. Few single
thermal oxidizers exist that are sized for more than 200,000 scfm of
flue gas.
Liquid organic wastes from a process may be a source of auxiliary
fuel for thermal incineration. However, combustion of liquid waste
streams in thermal incineration equipment can complicate the design.
3-7
-------
Inorganic compounds present in the liquids can create very difficult
particulate problems, which require additional equipment to solve.
Since the total capital cost to deal with these factors can be significantly
higher than for conventional fume thermal incineration, this study does
not address the complexities of feeding liquid organic wastes.
3.2.2.4 Advantages and Disadvantages
Following are the advantages of VOC control by thermal incinerators:
1. The VOC destruction efficiency of a thermal incinerator can be
determi ned.
2. The control efficiency of a thermal incinerator is relatively
insensitive to the specific VOC pollutant.
3. Thermal incinerator cost and efficiency determinations require
a limited amount of waste stream data (volume flow and heating value).
4. A thermal incinerator is capable of adapting to moderate
changes in effluent flowrate and concentrations.
Following are disadvantages of VOC control by thermal incinerators:
1. High capital and operating costs could result from thermal
incinerator techniques.
2. Thermal incinerators are unable to handle large volume intermittent
vent streams.
3.2.3 Boilers Used for Haste Gas Destruction
Fireboxes of boilers and fired heaters can be used, under proper
conditions, to incinerate combustible air contaminants. Boiler firebox
conditions approximate those of a well-designed thermal incinerator,
provided there are adequate temperature, retention time, turbulence, and
flame. Combustible contaminants, including smoke, organic vapors, and
gases, can be converted essentially to carbon dioxide and water in
boiler fireboxes.
Completely satisfactory adaptations of boilers for use as thermal
incinerators are not common. All aspects of operation must be thoroughly
evaluated before this method of air pollution control can be used. The
primary function of a boiler is to supply steam, and whenever its use as
a control device conflicts with this function, its purposes, will suffer.
This section describes the principles of boiler design and the factors
affecting boiler operation for VOC control.
-------
3.2.3.1 Principles of Boiler Design
The discussion here is limited to direct-fired boilers used for
steam generation. For complete combustion, the fuel to oxygen ratio
must be such that nearly all of the fuel is converted to carbon dioxide
and water vapor. If an insufficient amount of oxygen is present, not
all of the fuel will be burned, and excessive amounts of products such
as carbon monoxide and free carbon will be formed. Conversely, too much
excess air lowers the furnace temperature and becomes a major contributor
to poor boiler efficiency. The oxygen and nitrogen that pass through
the boiler is heated to the same temperature as the combustion products.
This heating consumes energy that would otherwise be available to produce
steam. Boiler efficiency is the ratio of heat output (steam and heat
losses) to the heat input (fuel, feed water, and combustion air). Flue
gas analysis and stack temperature measurements can be used to monitor
efficiency. Boiler efficiency can vary by the amount of excess air used
and the characteristics of a fuel being burned as well as the other
factors which affect combustion efficiency for thermal incinerators.
Normally, either natural gas or fuel oil is used as a fuel in industrial
boilers. When waste gas is used as a supplemental fuel in the boiler,
it becomes critical that the characteristics of the waste gas (such as
heating value) be controlled. Thus, fuel savings and boiler efficiency
can be predicted.
3.2.3.2 Parameters Affecting VOC Destruction Efficiency
VOC destruction efficiency achievable by boilers depends on the
factors affecting any combustion technique -- time, temperature, and
turbulence. The VOC must be intimately mixed with oxygen and held at
the appropriate temperature long enough for complete oxidation to occur.
For a boiler to be effective as a VOC control device, the following
conditions must exist:^
1. The volumes of contaminated gases must not be excessive or
thermal efficiency will be reduced because of reduced residence time.
2. When the waste gas stream is used to provide the oxygen for
combustion, precautions must be taken to ensure that the oxygen supply
is adequate for complete combustion. Incomplete combustion can form
tars or resins that will deposit on heat transfer surfaces and result in
reduction of boiler efficiency. When these contaminates exceed air
3-9
-------
pollution control standards for gas- or oil-fired boilers, tube fouling
will already have become a major maintenance problem.
3. An adequate flame must be maintained continuously in the
boiler firebox. High-low or modulating burner controls are satisfactory
provided that the minimum firing rate is sufficient to incinerate the
maximum volume of gases that can be expected in the boiler firebox. A
burner equipped with on-off controls would not be an acceptable control
device.
3.2.3.3 Factors Affecting Applicability and Reliability
For the plants where a steam-generating boiler is available,
retrofitting of the existing boiler to handle waste gas stream as supple-
mental fuel, manifolding the waste gas, storage, and controlling the
waste gas characteristics are important factors affecting applicability
and reliability.
Boilers can be retrofitted to serve as a control system in two
ways. In the first method, lower pressure vents are compressed and then
piped along with higher pressure vents into the boiler fuel header and
burned with the regular fuel. In the second method, one or more regular
burners in the boiler are replaced by forced draft burners capable of
accepting low pressure streams. Streams from the lower pressure vents
then are piped to the forced draft burners for combustion. Streams from
the higher pressure vents are piped directly to the fuel header or
regular burners. Burners applicable to this second method are available
as standard items for fuel pressures down to one inch of water gauge.^'^
There are two dangers in manifolding vent streams to boilers:
1. Pressure difference in the vent stream, and
2. Potential explosion hazards.
In the first case, certain vent or process streams may be at higher
pressure than others; thus, if these streams are ducted together, backflow
in the vent stream may occur. Such backflow can result in off-specification
product by introducing impurities in the vent stream into the process or
by upsetting process conditions. In the second case, potential explosion
hazards may result from ducting oxygen-rich streams to a boiler. In the
polymers and resins industry, particulate polymer in the waste gas is
also a potential explosion hazard.
3-10
-------
Designs of a system to overcome both problems have been demonstrated
in three polypropylene plants and in a high-density polyethylene
8
plant. To avoid backflow, a solution is to install equipment to
increase the pressure. Two polypropylene plants recover the waste gas,
which otherwise would be flared, for use as boiler fuel. The waste gas
from the polypropylene process is composed primarily of propane, propylene,
and butane and is considered similar in combustion properties to LPG.
The waste gas is passed through a knockout drum to remove entrained liquid
and through bag filters to. remove particulate polymer. The waste
gas then is stored in purge storages. A high-speed calorimeter monitors
and controls nitrogen addition to the waste gas line in order to control
heating value of the waste gas to the boilers. The five-psig gas is
burned in separate burners. The maximum waste gas rate is limited
either by boiler capacity or by the capacity of the recovery system.
9
One high-density polyethylene plant sends the dehydrator regeneration
gas (sweet natural gas and nitrogen) and a degassing stream from the
recycle diluent step (mostly ethylene) to steam-generating boilers as a
fuel.
3.2.3.4 Advantages and Disadvantages
1. The most apparent advantage of using existing boilers as a VOC
control device is that it can result in fuel (oil or natural gas)
saving, presuming that the waste gases have sufficient heating value to
support combustion. Depending on the heating value of the waste gases
and the capacity of VOC recovery systems, fuel savings can be significant.
2. In comparison to using thermal incinerators, using existing
boilers with minor retrofitting cost could represent a significant
capital cost saving.
3. The major disadvantage of VOC control by boilers is that
without ancilliary waste gas collection and holding equipment, boilers
are unable to handle intermittent releases of VOC.
4. The economics of boiler use for VOC control is site-specific
depending on plant steam requirements and current steam use.
3-11
-------
3.2.4 Elevated Flares
Flares are used predominantly in the polymers and resins industry
to burn waste gases which are economically unattractive to recover
either as product or for heating value. These waste gas streams are
usually streams with low flow, streams with irregular or intermittent
flow, or streams resulting from a process upset or some other emergency
condition where excess gases and vapors must be vented immediately. Due
to the preponderance of these types of streams in the polyethylene and
polypropylene industries, the elevated flare is the most commonly used
combustion device to control VOC.
3.2.4.1 Flare Design
Good combustion in a flare is governed by time, mixing, and temperature.
However, since flare combustion efficiency measurement methods are not
yet completely well-defined, the individual effects of these combustion
factors cannot be evaluated.
Elements of elevated flare systems are shown in Figure 3-3. The
VOC stream is conveyed by a transfer line from the facility release
point to the flare location. The line is equipped for purging so that
explosive mixtures do not occur in the flare system either on start-up
or during operation. The usual purge gas is natural gas, although other
fuel gases and inert gases such as nitrogen can be used.
A flare normally is equipped with ancilliary devices for safety and
maintenance reasons. Liquids that may be in the emission source gas or
that may condense out in the collection headers and transfer line usually
are removed by a disentrainment drum located close to the flare. Liquids
in a flare gas can cause smoke to form because of incomplete burning.
If the size of the droplets is greater than 150 ym, liquids may generate
a spray of burning chemicals that could reach ground level presenting a
safety hazard. A water seal is usually located between the disentrainment
drum and the flare stack to prevent flashbacks into the system. Other
devices such as flame arresters and actuated check valves, may sometimes
replace a water seal or be used in conjunction with it. For safety
reasons, a stack is used to elevate the flare so that the flame is well
above surrounding equipment. A gas barrier or a stack seal is normally
used just below the flare head to impede the flow of air into the flare
gas network, which could create an explosion hazard.
3-12
-------
Steam
Nozzles
Pilot
Burners
Gas
Barrier
GO
I
CO
Flare
Stack
Gas Collection Header
and Transfer Line
Emission
Source
Gas
1
Disentrainment
Drum
Steam
• Line
• Ignition
Device
• Air Line
• Gas Line
Drain
Figure 3—3. Steam assisted elevated flare system.
-------
The Important elements of flare system design for VOC destruction
are the flare tip itself and the flare'burner system. The maximum and
minimum capacity of a flare tip to burn gases with a st-abTe (but not
necessarily smokeless) flame is a function of tip design. At too high a
gas exit velocity, the flame can "lift off" and extinguish; while at too
low a gas velocity, the flame can burn back into the tip or lick down
the sides of the stack. Elevated flare tip diameters are normally sized
to provide gas exit velocities at maximum throughput of about 20 percent
of the sonic velocity of the gas, in case of continuous flaring. But it
could be as high as 50 percent of the sonic velocity in case of intermittent
emergency flaring.'^ However, it represents data from the state of the
art. A large percentage of flares operates well above 50 percent of the
sonic velocity. Several systems can function very well at 100 percent
of the sonic velocity. In general, the flare burner must be designed to
provide stable burning at the maximum exit velocity.
Modern commercial flares have design capabilities to burri gases
with stable flames over flare-gas discharge velocities up to the sonic
velocity. The actual maximum capacity of a flare tip is usually limited
by the flare-gas pressure drop. In the past, the maximum pressure drop
was usually limited to approximately two psig. However, continuing
development in non-steam assisted flaring and flame holding capability
of steam smokeless burning have made this limit not applicable in modern
commercial flares. In general, the higher the allowable pressure drop
(within design limitation), the longer the life of the flare and more
importantly, the lower the utility costs and energy usage. The difference
between two psig available pressure drop from 10 to 15 psig may be tens
of thousands of Ibs/yr of steam per year.
The smokeless capacity of a steam-assisted flare is limited by the
quantity of steam that is available. There is a physical limit to the
quantity of steam that can be effectively delivered to a flare tip. The
steam requirement depends on the composition of the gas flared and on
the tip diameter. Typical values range from 0.15 to 0.5 Ib of steam per
pound of flare gas.
The use of steam injection or the use of blowers enhances mixing
and supplies air for complete combustion with minimum smoke formation.
3-14
-------
Steam can also enter into a water/gas reaction with carbon or steam
reforming reaction with hydrocarbons and further reduce smoke formation.
Steam also moderates flame temperature to prevent cracking reactions
that form carbon and smoke. A significant disadvantage of steam usage
is the increased noise problems.
In addition to the design parameters which affect combustion efficiency,
other factors in flare design must also be considered. Elevated flares
can be freestanding, guyed, or structurally supported by a derrick.
Self-supporting flares can be used for flare tower heights of up to 50
feet. The guy-supported flare is the simplest support method, but
requires considerable amounts of land space. These flares can be used
up to as high as 100 feet. Derrick-supported flares can be built as
high as required since the system load is spread over the legs of the
derrick. However, these limitations placed on support methods are very
basic in nature and serve only as theoretical guidelines. In practice,
the type of support depends on the specific application and purpose of a
flare. For instance, self-supporting and guy-wire supported flares
which are several hundred feet tall can be used.
The decision on how high to mount the flare is a function of heat
content of the gas, allowable heat flux from the flame, fraction of heat
radiated, and the distance from the center of flame to the point of
allowable radiation intensity. The last factor may be calculated from
l ?
the knowledge of flame length and deflection using Brzustowski's method.
However, this method may be a reasonable starting point when no large-scale
radiation tests are available. In practice, the height of an overwhelming
number of flares in service were and are continuing to be sized by the
vendor of the flare equipment using proprietary methods based on actual
large-scale radiation tests. These numbers are usually guaranteed by
the vendor. After knowing the distance from the center of flame to the
point of allowable radiation intensity, its merely a compromise between
the height of flare and the surrounding space (i.e., the horizontal
distance from flare to the point of allowable radiation intensity). The
flare height design is usually a function of maximum ground-level heat
radiation intensity. Minimum flare height may sometimes be determined
by the need to safely disperse the flared gas in case of flame-out. The
3-15
-------
height in these cases would be based on dispersion modeling for the
particular installation conditions.
As with most combustion devices, problems can arise with routing
the waste gas to the flare. In the polyethylene and polypropylene
industries, two specific problems may be present in manifolding vent
streams to flares:
1. Pressure differences in the vent streams, and
2. Potential explosion hazards.
Designs and equipment to overcome both problems have been demonstrated
and are straightforward to apply. The first problem can be avoided by
either segregating the low-pressure sources, or by installing equipment,
such as a compressor or an extra ejector stage, to increase the pressure.
There are distinct economic advantages in separating higher pressure
reliefs from lower pressure reliefs in terms of header sizes and overall
installed cost of the flare system. Lower operating and maintenance
costs can also result from pressure segregation. In the latter case,
the system must be able to relieve in the case of power or compressor
failure. The second problem can be avoided by using oxygen analyzers
which divert the vent streams or add inerts if unsafe oxygen levels are
detected. However, sole reliance on oxygen analyzers is not recommended
because the flare system always has a source of ignition in the pilots,
and response time of oxygen analyzers is slow enough that an explosion
may occur before the analyzer responds. Other design methods use welded
pipe to prevent air infiltration, orifices to restrict the flow from
vacuum equipment in case of leaks, and flame arresters. The use of
labyrinth flame arresters is also not recommended. These devices have
small passage-ways which easily plug with line-scale or polymerization
of waste. In addition, most flame arresters are ineffective in streams
containing hydrogen. The use of labyrinth flame arresters is questionable
when piping downstream can allow an adverse pressure gradient to develop
due to a flashback.
The minimum continuous purge-gas required is dependent on the
design of the stack-seals. These are usually proprietary devices. The
purge-gas velocities required depend on the size of the flare and may
vary from 0.1 to 1 fps. For instance, tests performed on a 30-in.
diameter by 60-in. tall stack erected for purge testing indicates that a
3-16
-------
purge of one fps is required to prevent air infiIteration due to decanting
and/or wind action on a plain, unsealed stack. The amount of purge gas
required to prevent a vacuum after a hot gas discharge is flared depends
on the condition of gas. If the gases in the system will go to dewpoint,
a very large volume may be required. Actual cases have required in
excess of 100,000 scfh of purge gas during the cool-down period.
Pilot gas-usage is a function of the number and design of pilot
units required to ensure positive ignition of the flared gas and the
mode of operation.
3.2.4.3 Parameters Affecting VOC Destruction Efficiency
Various factors affect and control flare combustion efficiency, but
the critical one is the thoroughness with which flare exit gases are
mixed with air. The key design variables affecting mixing are as follows:
1. Flare Tip Diameter: As flare diameter increases, the turbulence
required to mix air thoroughly with all exit gases becomes harder to
achieve.
2. Exit Velocity: Mixing is a function of the kinetic energy of
the exit gases, i.e., better mixing is achieved with higher velocities
until a maximum is reached when the flame blows out or the noise becomes
unacceptable. However, many flare designs will prevent flame lift-off
at velocities resulting from pressure drops in excess of 100 psig and
may not produce more noise than a comparable steam flare.
3. Steam Injection: This is another way of adding kinetic energy
and also large amounts of air. But excess steam can quench the flame
and large excesses of air drawn into the combustion zone can decrease
combustion efficiency.
4. Presence of Noncombustibles or Thermally Stable Compounds:
Use of inert gas in the process units is common in the polymers and
resins industry. It is used either for pressure control or for purging
and drying. If vent streams contain large volumes of compounds with low
heats of combustion or noncombustibles, flare combustion may be affected
with a decrease in combustion efficiency.
5. Use of One Flare for Continuous and Intermittent Streams:
Many times large-diameter flares are designed to handle emergency releases,
and these are also used to control low-volume, continuous vent streams.
3-17
-------
With such designs, optimum mixing may not be achieved since the vent gas
exit velocity is low and large flares generally cannot inject steam into
low-volume streams properly. The operation in such a case, can be
improved if a properly-designed flare system includes several steam
injection points in order to optimize steam injection for the flare at
various flowrates thus insuring proper operation.
All these factors, if not incorporated properly in the design, will
reduce the combustion efficiency. At present, no conclusive data are
available on flare efficiency. Sampling a commercial-size flare is very
difficult and is virtually impossible for the larger sizes. Results
obtained from testing small flares cannot be scaled up for commercial-
size flares. An extensive flare test program was conducted in Germany.
In this study, 1298 measurements were taken at the flame end and above
the flame for a range of test conditions; only four measurements gave
local burnout to carbon dioxide of less than 99 percent. This test
program, however, specifically was related to a refinery, and the operating
conditions selected were comparable to those found in refineries.
Hence, it is difficult to generalize the results for all industrial
flares.
Until better test methods are available and until better flare
efficiency data related to specific industries are available, an reliable
estimate of flare efficiency cannot be made.
3.2.4.4 Factors Affecting Applicability and Reliability
The reliability of using elevated flares has been established and
demonstrated by the extended use of this control technique over a long
period of time. The production of any typical polymer plant is approximately
several hundred million pounds per year and the resultant VOC emissions
due to frequent process upsets either with controlled or uncontrolled
conditions, are significant. The flares are mainly used to handle
emergency blowdowns which requires the control device to handle large
volumes of gases with variable composition. Elevated flares can handle
such a situation more effectively than other control techniques.
The reliability of flares has been further improved by a new design
concept of controlled blowdowns which can improve and save money. The
system is designed to vent gases in required time without interfering
3-18
-------
with the pressure relief system. In order to reduce the peak flowrate,
the waste gas is passed through a restriction orifice sized to set a
desired back pressure in the flare header at the beginning of blowdown.
Consequently, flow continues through the restrictive orifice and is
supplemented by flow through the main control valve which is adjusted to
maintain the desired flare header back pressure. As the pressure in the
blowdown header decays with time, a pressure indicator and controller
increases the opening of the flare header as close as possible to its
desired value. As long as the main control valve is operating, the
flowrate going to the flare is constant and the system behaves independently
with respect to time. With the controlled blowdown system, peak noise
levels can be reduced, smoke production abated, space saved, and investments
cut. But design for controlled blowdown requires extensive calculations
while its application is limited to special situations.^
3.2.4.5 Advantages and Disadvantages
VOC control by flares has several advantages:
1. A properly designed and operated flare can provide destruction
of many types of VOC. A flare designed to handle intermittent emergency
vent streams functions well for emergency streams; whereas, a flare
designed to handle continuous streams operates well for continuous
streams.
2. Elevated flares have a wide capacity range and are capable of
adapting to changes in effluent flowrates and concentrations that are
found in the polymer industry.
There are also disadvantages associated with VOC control by flares:
1. In the polymers and resins industry, most of the time a flare
designed to handle intermittent emergency vent streams is also used to
control small continuous streams. This may affect optimum mixing and
thus, the combustion efficiency. No quantitative assessment can be made
about the effects on efficiency of this practice.
2. Another disadvantage of flares is the associated inability to
determine exactly the flare combustion efficiency.
3.3 TECHNICAL FEASIBILITY OF RETROFITTING CONTROL DEVICES
Retrofitting existing boilers to accept waste gases as a supplemental
fuel has readily been demonstrated in high-density polyethylene and
3-19
-------
polypropylene plants. Depending on the type of fuel being used in the
existing boilers (fuel oil or natural gas), equipment required to incorporate
the waste gas to boiler differs. Separate burners may be required for
boilers that use fuel oil. Waste gas compressors are needed in both
cases. The nature (continuous or intermittent) and characteristic
(heating value) of waste gas stream also affect the feasibility of
retrofitting. If the waste gas stream is intermittent in nature, surge
tanks are required to store the waste gas such that a continuous supply
to the boiler would be possible. If the waste gas stream concentration
varies, nitrogen or natural gas addition may be needed to control the
heating value of the waste gas to a constant value. This means additional
equipment such as high-speed calorimeters and a nitrogen or natural gas
flow control system would be needed. It is not practical to design
these systems for a major process upset where a large volumetric flow of
waste gas is expected in a short period of time.
Thermal incinerator retrofitting technical feasibility also is
site-specific. Retrofits usually require remodeling of existing structures
and coordination of the construction efforts with process conditions.
3-20
-------
3.4 REFERENCES FOR CHAPTER 3
1. Perry, R.H., & Chilton, C.H., Chemical Engineers' Handbook, 5th
Edition, Section 9, McGraw-Hill Book Company^
2. Reed, R. J., North American Combustion Handbook, Cleveland, North
..; America Publishing Company, 1979, p. 269.
3. Memo and addendum from Mascone, D., EPA, to Farmer, J., EPA.
June 11, 1980.
4. Modern Pollution Control Technology, Volume 1, Research and Education
Association, New York, 1980, pp. 23-13-23-15.
5. Telecon. Reed, R., North American Manufacturers, with Mascone, D.,
- EPA. August 1, 1978.
6. Telecon. Miller, F., Bloom Engineering, with Mascone, D., EPA.
October 23, 7978.
7. Shell Woodbury Plant, Permit Application, New Jersey.,
8. EEA, Trip Report, Phillips Chemical Company, Contract No. 68-02-3061,
Task 2, August 8, 1980.
9. EEA, Trip Report, Phillips Chemical Company, Contract No. 68-02-3061,
Task 2, August 8, 1980.
10. Kalcevic, V., Emission Control Options for the Synthetic Organic
Chemicals Manufacturing Industry, Control Device Evaluation, Flares
and the Use of Emissions as Fuels. EPA Contract No. 68-02-2577,
August 1980.
11. Neveril, R.B., Capital and Operating Costs of Selected Air Pollution
Control Systems, EPA Contract No. 68-02-2899, December 1978.
12. Oenbring, P.R., & Sifferman, T.R., "Flare Design ...Are Current
Methods Too Conservative?" Selections from API, NPRA, and GPA
Meetings, Hydrocarbon Processing, pp. 124-129, May 1980.
13. Brzustowski, T.A., "A Model for Predicting the Shapes and Lengths
of Turbulent Diffusion Flames Over an Elevated Industrial Flare,"
22nd Canadian Chemical Engineering Conference, Toronto, Ontario
(1972).
14. Siege!, K.D., Degree of Conversion of Flare Gases in Refinery High
Flares. Pollutant Emission from Refinery High Flares as a Function
of their Operating Conditions, dissertation for the degree of Ph.D.
In Engineering Science at the Chemical Engineering Department of
the University in Karlsruhe, West German, February 1980.
15. Pariut, B., & Kimmel, W., "Control Slowdown to the Flare,"
Hydrocarbon Processing, pp. 117-121, October 1979.
3-21
-------
4. ENVIRONMENTAL ANALYSIS OF RACT
4.1 INTRODUCTION
The environmental impact of the systems considered representative
of reasonably available control technology (RACT) is essentially .two-
fold. Primary impacts are those attributed directly to the control
systems, such as reduced levels of specific pollutants. Secondary
impacts are indirect or induced in nature, such as aggravation of another
pollutant problem through utilization of a particular control system.
Both beneficial and adverse environmental impacts related to each of the.
pollution categories, air, water, and solid waste are assessed in the
following sections. Also a discussion of the additional amount and type
of energy required for control is included.
A 98 weight percent reduction in VOC emissions from continuous vent
streams from the following is considered representative of RACT:
1. Polypropylene plants using the liquid-phase process;
2. High-density polyethylene plants using the low-pressure,
liquid-phase process.
An emission limit of 0.3 kg of VOC/1000 kg of product for polystyrene
plants using continuous processes is considered representative of RACT.
Combustion control devices, such as thermal and catalytic incinerators
or boilers and process heaters, can achieve 98 percent VOC destruction
efficiency. Flare efficiency, however, cannot be quantified in absence
of adequate test data. Other control techniques such as refrigerated
condensation that can achieve the same degree of control should be
considered equivalent and acceptable.
4.2 AIR POLLUTION
The annual quantities of volatile organic compounds (VOC) from the
model plants before and after control by RACT are presented in Table 4-1.
The stream from each model plant represents a combination of continuous
emission streams from process vents excluding fugitives and raw material
and product storage facilities. The expected reduction in VOC emissions,
achieved as a result of implementation of RACT for the model plants, is
shown in Table 4-1.
4-1
-------
TABLE 4-1. MODEL PLANT RECOMMENDATIONS FOR RACT USED AS A BASIS FOR ENVIRONMENTAL
ANALYSIS
Polymer
Polypropylene
High-Density Polyethylene
CONTINUOUS STREAM
Model Plant
Uncontrolled VOC
Emission Rate
(Mg/yr)
4,898
2,687
CONTINUOUS STREAM CONTINUOUS STREAM
Projected
Control Device
Thermal Incinerator
Thermal Incinerator
Model Plant
Controlled VOC
Emission Rate
(Mg/yr)
98
54
Polystyrene (Continuous
Process)
227
Thermal Incinerator
4.54
-------
The VOC destroyed as a result of the application of RACT consists
mainly of ethylene, propylene, styrene, and certain organic diluents.
These gases are known to react in the atmosphere with oxides of nitrogen
to form oxidants, principally ozone. Reduction of emissions of these
gases will contribute to the attainment of the national ambient air
quality standard (NAAQS) for ozone.
A thermal incinerator is expected to be the major control device
used as RACT. A properly designed combustion device would lead to
minimal formation and subsequent emission of carbon monoxide. The
amount of NO products formed by incineration at 1600°F is negligible.
J\
Thus, there should be minimal generation of secondary air pollutants by
combustion techniques.
4.3 WATER POLLUTION
Incineration systems do not generate an effluent water stream.
4.4 SOLID WASTE DISPOSAL
There are no solid wastes generated as a result of control by RACT
in any model plant under consideration.
4.5 ENERGY
Table 4-2 presents the additional amount and type of energy required
after control of each model plant by RACT. The control techniques
analyzed for RACT require distillate oil and electricity.
For an incinerator, the fuel requirement is usually limited to only
the distillate oil required for flame stability since the streams encountered
in these polymer industries are rich enough to sustain self-combustion.
In a high-density polyethylene model plant, normally, there are three
streams that would be burned in the incinerator, two of which are dilute,
the other is VOC-rich. During an assumed four hours per day, when the
VOC-rich stream is not available, some supplemental natural gas is used
to enrich the two dilute streams. Electricity cost projected in
Table 4-2 is only for the fan operation. Instrumentation is assumed to
consume a negligible amount of electricity.
The possibility of fuel switching (gas to coal) is remote in case
of an incinerator.
4-3
-------
TABLE 4-2. ADDITIONAL ENERGY REQUIRED AFTER CONTROL BY RACT
Model Plant
Stream
Polymer Nature
Annual Fuel
Consumption
Distillate
Oil (Gal)
Annual
Electricity Consumption
($) (kWh) ($)
Polypropylene
Cont.
None
None 17,755 870
High-Density
Polyethylene
Cont.
9 x 10° scf
(natural gas)
None 34,470 1,738
Polystyrene
(Continuous
Process)
Cont.
7,324
7,112 4,286 210
4-4
-------
5. CONTROL COST ANALYSIS .OF RACT
This chapter describes the approach.taken to estimate the costs for
controlling volatile organic compounds (VOC) ^missions from the polymers
and resins industry. The cost analysis includes capital costs, annualized
costs, and cost-effectiveness of control. Furthermore, the control
equipment installation factors include retrofit provisions since RACT
will affect existing plants. The cost.information used in the analysis
is based on data provided in the manual.Capital and Operating Costs of
Selected Air Pollution Control Systems. The following sections separately
identify and discuss major cost components, give the basis and time
period for costs, give sources of estimates and scaling factors fbr
prorating these estimates, list assumptions made in the analysis, and
note items that are not included in the cost analysis.
5,1 BASIS FOR CAPITAL COSTS
Capital costs consist of purchased equipment costs and direct and
indirect installation costs. Thermal incinerators are the only control
device evaluated. Purchased equipment costs for this device were basically
obtained from Reference 1 with some modifications. The curves for
thermal incinerator cost calculations shown in Figures 5-1 and 5-2, were
adjusted for temperature and residence time. Reference 1 provides
relationships between residence time and purchased equipment costs and
between gas flowrate and gas temperature for incinerators. The information
required to make these extrapolations and the detailed procedures are
described in Appendix B to this document. The minimum size incinerator
considered in the cost analysis is of 1000 scfm capacity.
Purchased equipment cost for manifolding the incinerator inlet
gases was obtained from Reference 1. The quantity of manifolding required
o
was based on a single plant occupying an area of 1000 ft . The cost of
manifolding includes accessories normally associated with the conveyance
of a gas stream. These include elbows, tees, expansion joints, dampers,
and transitions. The prices of these units are a function of duct
diameter and material thickness. Since the gases involved are non-
corrosive and the temperatures involved are mild, carbon steel was
selected as the material of construction for costing the manifolding
5-1
-------
INCINERATOR PRICE ($1000)
140
en
i
ro
120-
100-
80-
60-
40-
20-
T= Temperature
RJ= Residence Time
10
T = 1600°F
R.T.= 0.75sec-
R.T0=0,5sec
NOTE.
1 Operating temperature of 1500U F.
2. Residence time for incineration is
0 5 seconds.
3. Accuracy of this curve is ±50%
4. Price includes incinerator, fan or
blower, controls, and instrumentation.
5 Prices will vary as a function of
a. Retention times
b. Materials of construction
c. Special controls
d. Heat content of pollutant
6. Inlet concentrations of 0-25%
LEL light hydrocarbons
7 Process temperatures of
70-300° F.
8. OiI fired burners.
9. Source' Ref 1.
20
I
30
40
FLOW RATE (1000 SCFM)
Figure 5—1. Prices for thermal incinerators without heat exchanger.
-------
INCINERATOR PRICE ($1000)
150
in
i
oo
130-
110-
90-
70-
50-
30-
10
T = 1600°F
R.T. = 0.75 sec
T = Temperature
R.T.= Residence Time
NOTE:
1. Residence time of 0.5 seconds,
2. Process temperature is 70° F.
3. Operating temperature is 1500° F.
4. Curve based on 35% heat recovery.
5. Accuracy of this curve is ±50%.
6. Price includes incinerator, heat
exchanger, fan or blower, damper
controls, and instrumentation.
7 Prices will vary as a function of
a. Retention times
b. Materials of construction
c. Special controls
d. Heat content of pollutant.
8. Inlet concentration of 0-25% LEL.
9. Oil fired burners.
10. Source: Ref. 1.
10
20
30 40
FLOW RATE (1000 SCFIYI)
Figure 5-2. Prices for thermal incinerators with primary heat exchanger.
-------
system. The price of the incinerator includes a fan or a blower;
however, a compressor is also added to the control system because
retrofit provisions could extend the length of ductwork up to 2000 feet.
The applicable type of compressor is based on generally applied limits
2
for operation of the reciprocating, centrifugal, and axial-flow compressors.
3
Purchased compressor cost estimates are from vendors.
Direct and indirect installation cost factors for all the systems
used in the RACT analysis are summarized in Table 5-1 and include most
of the expenses except for site preparation, facilities and buildings,
production losses during start-up and research and development costs.
Table 5-2 gives cost adjustment factors for modifying the control equipment,
direct and indirect, installation costs for the existing facility. The
retrofit provisions for the recommended RACT for the polymers and resins
industry are separately identified in Table 5-3. In evaluating the
installation costs of all systems, it is assumed that the system is
installed by an outside contractor and not by plant personnel.
The time period for most of the above costs was December 1977. An
escalation index, Fabricated Equipment Index, was used from Chemical
Engineering Journal, to update all the costs to June 1980. Table 5-1
also gives these indices. Capital costs for major components of RACT
were verified by vendor contact.
5.2 BASIS FOR ANNUALIZED COSTS
The typical annualized costs consist of the direct expenses for
labor, utilities, fuel, and materials for operation and maintenance of
the control device plus the indirect costs for overhead, taxes, insurance,
general administration, and the capital recovery charges. The indirect
costs are related to the capital investment, and these factors are also
summarized in Table 5-1. The table also gives labor, fuel, and utility
rates. All the data required in estimation of these costs were obtained
from Reference 1. The labor hours per shift for thermal incinerators
are 0.5. Three shifts a day and 365 days a year were assumed for calculating
annual labor costs.
5.3 EMISSION CONTROL COSTS
This section discusses the estimated emission control costs associated
with installation of RACT in each of the model plants. Table 5-4 summarizes
5-4
-------
TABLE 5-1. COST FACTORS
INSTALLED
Taxes and Freight (T&F): 0.08 x Purchased Equipment (PE)
Instrumentation (I): 0.10 x PE
Direct-Indirect Installation:
Thermal Incinerator
Ductwork (Manifolding)
•
Compressor
Installed Equipment Costs (IEC):
PE + T + F + I + Total Installation
Escalation Index: (Source: Chemical Engineering, Economic Indicators)
0.61 x (PE + T + F + I)
0.40 x (PE + T + F + I)
0.10 x (PE + T + F + I)
Period
December 1977
June 1980
Fabricated Equipment Index
226.2
296.1
ANNUALIZED
Pi rect
- Operating Labor: $11.10/hr
(includes overhead)
- Maintenance Labor: $10.90/hr
(includes overhead)
- Distillate Oil: $0.971/gal
- Natural Gas: $2.40/1000 scf
- Electricity: $0.049/kwh
References: 1, 4, 5, 6, and 7.
Indirect
- Interest Rate: 10%
- Equipment Life: 10 yrs
- Capital Recovery:
0.16275 x (IEC)
- Property Taxes: 0.01 x (IEC)
- Insurance: 0.01 x (IEC)
- Administration: 0.02 x (IEC)
5-5
-------
TABLE 5-2. COST ADJUSTMENTS*
A. Instrumentation: Cost Adjustment
1. Simple, continuous manually operated 0.5 to 1.0
2. Intermittent operation, modulating flow with emissions
monitoring instrumentation 1.0 to 1.5
3. Hazardous operation with explosive gases and safety
backups 3
£. facilities & Buildings:
1. Outdoor units, utilities at site
2. Outdoor units with some weather enclosures. Requires
utilities brought to site, access roads, fencing, and
minimum lighting
3. Requires building with heating and cooling, sanitation
facilities, with shops and office. May include railroad
sidings, truck depot, with parking area
01
i
CTl
1. Major metropolitan areas in continental U.S. 0.2 to 1.0
2. Remote areas in continental U.S. 1.5
3. Alaska, Hawaii, and foreign 2
C. Handling and Erection:
1. Assembly included in delivered cost with supports,
base, skides included. Small to moderate size
equipment 0.2 to 0.5
2. Equipment supplied in modules, compact area site with
ducts and piping less than 200 ft in length. Moderate
size system 1
3. Large system, scattered equipment with long runs.
Equipment requires fabrication at site with extensive
welding and erection 1 to 1.5
4. Retrofit of existing system; includes removal of
existing equipment and renovation of site. Moderate
to large system 2
0. Site Preparation:
1. Within battery limits of existing plant; includes minimum
effort to clear, grub, and level 0
2. Outside battery limits; extensive leveling and removal
of existing structures; includes land survey and study 1
3. Requires extensive excavation and land ballast and leveling.
May require dewatering and pilings 2
F. Engineering & Supervision:
1. Small capacity standard equipment, duplication of typical
system, turnkey quote 0.5
2. Custom equipment, automated controls 1 to 2
3. New process or prototype equipment, large system 3
G. Construction & Field Expenses:
1. Samll capacity systems .1
2. Medium capacity systems 1
3. Large capacity systems 1.5
H. Construction Fee:
1. Turnkey project, erection, and installation included in
equipment cost .5
2. Single contractor for total installation )
3. Multiple contractors with A&E firm's supervision 2
I. Contingency:
1. Firm process ]
2. Prototype or experimental process subject to change 3 to 5
3. Guarantee of efficiencies and operating specifications
requiring initial pilot tests, deferment of payment
until final certification of EPA tests, penalty for
failure to meet completion date or efficiency 5 to 10
*Based on data obtained in Reference 1.
-------
TABLE 5-3. CAPITAL COSTING - THERMAL INCINERATOR
1*
TYPICAL
COST
FACTOR
2*
COST
ADJUSTMENT
FACTOR
3*
SYSTEM
COST
FACTOR
DIRECT COSTS:
1. Purchased Equipment Costs:
a. Control Device
b. Auxiliary Equipment
c. Instruments & Controls
(included in 1-a)
d. Taxes
e. Freight
TOTAL
2. Installation Direct Costs:
a. Foundations & Supports
b. Erection & Handling
c. Electrical
d. Piping
e. Insulation
f. Painting
g. Site Preparation
h. Facilities & Buildings
TOTAL
INDIRECT COSTS
3. Installation Indirect Costs:
a. Engineering & Supervision
b. Construction & Field Expenses
c. Construction Fee
d. Start-up
e. Performance Test
f. Model Study
g. Contingencies
TOTAL
*Sased on Reference 1.
X
0.03 XI = 0.03
0.05 X 1 = 0.05
1.00 1.00
0.08 0.08
0.14 X 2 0.28
0.04 0.04
6.'02 — 0.02
0.01 - 0.01
0.01 - 0.01
X = o
X = Q
1.30 - 1.44
0.10 X 2 Q.2
0.05 X 1.5 = 0.075
0.10 X 2 Q.2
0.02 0,02
0-01 0.01
None
0.03 X 10 = o.3
]-51 - — 2.25
5-7
-------
TABLE 5-4. POLYMERS AND RESINS MODEL PLANT PARAMETERS AND EMISSION CONTROL COSTS
Polymer
Polypropylene
High-Density
Polyethylene
Polystyrene
Continuous
Production Nature VOC
Capacity of the Cone.
(Gg/yr) Stream (wt %)
141 Cont. 70
214 Cont. 8
73.5 Cont. 16
Gas
Flowrate
(acfm)
540
481
181
Emission
Factor
kg VOC/
1000 kg
Flow Time Product
8,000 hr/yr 34.7
8,000 hr/yr 12.4
8,000 hr/yr 3.09
Total VOC
Emissions
(Mg/yr)
4,900
2,687
227
Assumed VOC
Destruction
Projected Efficiency
Control Device (%)
Thermal Incinerator 98
Thermal Incinerator 98
Thermal Incinerator 98
Annual i zed
Cost ($)
106,000
121 ,000
71 ,100
Installed
Capital Cost
459,000
422,000
287,000
co
Minimum size (1,000 scfm capacity) incinerator.
All numbers rounded off to three significant figures.
-------
the model plant parameters used in the cost analysis and gives the
installed capital costs and annual!zed costs for thermal incinerator for
three polymers. In addition, tables presenting the in-depth cost analysis
work for each model plant are also included. These tables separately
identify the annualized costs as operating expenses and capital charges.
There are no recovery credits associated with combustion techniques.
5.3.1 Assumptions
Following is a list of all the assumptions made in the cost estimates
that are common to all model plants:
1. One compressor is sufficient to send all waste gas streams to
the control device because the streams that are combined are similar in
nature, and at, more or less, the same pressure and temperatures.
2. One, thousand-foot, main duct from the compressor .to the
control device and five, hundred-foot, ducts from any emission vent to
the compressor were used in determining the cost of manifolding. A duct
velocity of 2000 fpm was assumed for gases.
3. Pressure drop in the duct is nine inches water guage per 1000 ft.
4. Thermal incinerator cost does not include a heat exchanger.
5. Installation cost of compressor is 10 percent of the purchased
cost.
6. Ten percent heat loss from combustion chamber was accounted
for in calculating fuel requirements for incineration.
7. Installation cost of ductwork is 40 percent of the purchased
cost.
8. Three eight-hour shifts a day, 365 days/year.
9. Exactly three percent by volume of oxygen is used in flue gas
for calculating the amount of combustion air required for oxidation.
The amount of combustion air determines the amount of fuel and electricity
used in the system.
10. If heating value of the stream is above 50 Btu/scf, stream has
sufficient heating value to reach the incineration temperature of 1600°F
without additional heat recovery. Hence, there is no need for a primary
heat exchanger.
5-9
-------
11. If the stream heating value is less than 300 Btu/scf, the only
fuel required is for flame stability and the ratio of heat supplied by
waste gas stream to :heat supplied by fuel is assumed to be 10:1.
5.3.2 Polypropylene (PP.)
Polypropylene, a thermoplastic resin and a member of the olefins
family, is manufactured by an addition polymerization reaction in the
presence of a solvent and Ziegler-Natta catalyst. Methanol or isopropanol
is generally used to separate catalyst particles from the product. The
catalyst separation and subsequent alcohol recovery operations emit over
one-third of the total process emissions. Vacuum jet exhaust and slurry
filter vents are other major separation vents. These four sources
contribute over 80 percent to the total VOC emissions. Emissions from
dryers can be large depending on the type of dryers and on the operating
pressures. The cost analysis is based on a fluidized bed dryer which
emits only a small amount of VOC. Some dryers, however, are potentially
large emitters.
All the continuous emission streams consist mainly of propylene,
propane, ethylene, and solvents. Propane is present as monomer impurity
and its percentage varies depending upon the propylene source. For cost
purposes, the streams are combined together and treated as a single
stream, and it was assumed that the stream consists of 50-50 percent by
volume propylene and ethylene. Assumptions of lower molecular weight
compounds give the highest possible design volumetric flowrate for
conservative cost estimates. The combined emission stream from polypropylene
plants is very rich in VOC and is usually diluted by nitrogen. Nitrogen
is added to keep the lower heating value of the stream in the desired
range of 1000-1100 Btu/scf. This kind of stream has characteristics of
natural gas and, thus, can be used in boilers as fuel supplements. The
amount of nitrogen varies but is usually within 10-30 volume percent of
the total (VOC + N2) stream.
Table 5-5 summarizes the cost analysis for the polypropylene model
plant. The thermal incinerator is the major cost component of RACT.
For purposes of calculating the emission reduction potential for the
Q
model plant, 98 percent efficiency for incinerator was assumed.
Boiler destruction efficiency is assumed to be equally as good as the
5-10
-------
TABLE 5-5. COST ANALYSIS FOR POLYPROPYLENE
Nature of Stream: Continuous, 540 acfm. 70% VOC
Thermal
Item Incinerator
Manifolding
Reciprocating
Compressor
Installed Cost ($)
- Purchased Equipment
- Installation
TOTAL INSTALLED
94,800
118,500
213,300
120,000
48,000
168,000
70,800
7,080
77,900
Annualized Cost ($/yr)
Direct
- Operating Labor
- Maintenance Labor
- Fuel
- Electricity
- Steam
SUBTOTAL
6,080
5,970
870
12,920
Indirect
- Capital Recovery
- Tax, Insurance, &
Administration
SUBTOTAL
TOTAL*
34,715
8,532
43,247
56,200
27,342
6,720
34,062
34,100
12,678
3,116
15,794
15,800
COST-EFFECTIVENESS ($/Mg VOC) 106,000/4,800 = $22
*Rounded off to three significant figures.
-------
incinerator destruction efficiency. Since the use of emissions as fuel
supplements is very site-specific, use of a boiler is not included in
the model plant. Compressor and manifolding are the two other cost
components. All the emission streams are collected, compressed, and
discharged through the duct network to the control device. For cost
analysis, it was assumed that vent streams from sources close together
are mixed and sent to the compressor by a single 500-ft duct. Hence,
there are four 500-ft ducts that deliver the gas from nine sources.
The total installed capital cost of RACT is estimated to be $459,000.
Annual VOC emissions of 4,800 Mg could be destroyed at an annualized
cost of $106,000. Direct annualized costs are relatively insignificant
as shown in Table 5-5.
5.3.3 High-Density Polyethylene (HOPE)
The liquid-phase, low-pressure, polymerization process represents
the typical process for producing high-density polyethylene resins. The
continuous emission streams from dryer, continuous mixer, and recycle
treater vents were combined and used to size and cost an incineration
system. Table 5-6 represents in-depth cost analysis for HOPE.
The combined continuous stream (about eight weight percent VOC)
flowrate was used as a basis to determine the cost of an incineration
system. The incineration system consists of a thermal incinerator
without a heat exchanger, 2,500 ft of ductwork, and a reciprocating
compressor. The thermal incinerator represents 50 percent of total
installed cost ($422,000) of the incineration system. The total annualized
cost of the incineration system is $121,000; it could destroy 2,633 Mg of
VOC annually.
5.3.4 Polystyrene (Continuous Process)
The continuous process emits unreacted styrene monomer (VOC) because
the polymerization process approaches equilibrium before reaction completion
The remaining unreacted monomer (styrene) entrained in the product is
separated in downstream unit operations. Most of the unreacted styrene
shows up as emissions from the styrene condenser vent, the largest
source of VOC in the plant. The combination of the styrene condenser
vent and styrene recovery unit condenser vent stream flowrates is used
to cost the incineration system.
5-12
-------
TABLE 5-6. COST ANALYSIS FOR HIGH-DENSITY POLYETHYLENE
en
i
Nature of Stream:
Item
Installed Cost ($)
- Purchased Equipment
- Installation
TOTAL INSTALLED
Annuali zed Cost ($/yr)
Di rect
- Operating Labor
- Maintenance Labor
- Fuel (Natural Gas)
- Electricity
- Steam
SUBTOTAL
Indirect
- Capital Recovery
- Tax, Insurance, &
Administration
SUBTOTAL
TOTAL*
Continuous, 480 acfm, 8% VOC
Thermal
Incinerator Manifolding
94,800 103,000
118,500 41,200
213,300 144,200
6,080
5,970
21 ,600
1,740
__
35,390
34,715 23,469
8,532 5,768
43,247 29,237
57,000 29,200
Reciprocating
Compressor
59 ,000
5,900
64,900
.„
—
—
—
-~
—
10,560
2,596
13,156
13,200
COST-EFFECTIVENESS ($/Mg VOC) 121,000/2,633 = $46
*Rounded off to three significant figures.
-------
Since there are only continuous exhaust streams in a polystyrene plant,
control costs were calculated only on an incineration system consisting
of a thermal incinerator, manifolding, and a compressor. The installed
cost of the incinerator is 74 percent of the total installed cost of the
system. Annualized cost is expected to be $71,1000 to control 222.5 Mg
of VOC. The installed capital cost of the system is $287,000.
Table 5-7 summarizes the cost analysis of the continuous polystyrene
model plant.
5.4 COST-EFFECTIVENESS
The annualized cost-effectiveness values ($/Mg) for construction
and operation of a thermal incinerator are presented in Table 5-8 for
each model plant. The combined continuous streams from polypropylene,
high-density polyethylene, and continuous polystyrene plants have very
low control cost-effectiveness values of $22, $46, and $320 per ton of
VOC destroyed, respectively.
5.4.1 Cost Effectiveness on a Stream-by-Stream Basis
The cost of controlling emissions from each of the model plants was
based on combining all the continuous streams that were judged to be
reasonable to control and delivering it to a single control device. For
each model plant, the resulting "combined stream" was smaller than the
capacity of the smallest "off-the-shelf" incinerator available. Because
of this, no cost analysis was made on a stream-by-stream basis because
even if each of the streams were to be controlled completely independently,
the requisite incinerator would be the same off-the-shelf item, large
enough to handle the entire plant's needs. Although some reduction in
piping costs might result from installing several incinerators, each
close to the independent VOC source, the cost of multiple incinerators
would override savings in piping and result in much greater cost to the
owner. Further, since the amount of VOC controlled would be the same as
calculated for the model plant, the higher costs would result in a cost-
effectiveness ratio that would also be much higher.
Similarly, if any of the streams which are shown as being.controlled
in the plant were not controlled (i.e., not manifolded to the single
control device, there would be some minor savings to the owner since the
requisite piping would not be required. Since the major cost item is
the incinerator, however, the capital and operating cost to the owner
5-14
-------
TABLE 5-7. COST ANALYSIS FOR POLYSTYRENE (CONTINUOUS PROCESS)
en
CJl
Nature of Stream:
Item
Installed Cost ($)
- Purchased Equipment
- Installation
TOTAL INSTALLED
Annuali zed Cost ($/yr)
Direct
- Operating Labor
- Maintenance Labor
- Fuel
- Electricity
- Steam
SUBTOTAL
Indirect
- Capital Recovery
- Tax, Insurance &
Administration
SUBTOTAL
TOTAL*
Continuous, 180 acfm, 16% VOC
Thermal
Incinerator Manifolding
94,800 46,400
118,500 18,560
213,300 64,960
6,080
5,970
__
210
__
19,372
34,715 10,572
8,532 2,598
43,247 13,170
56,200 13,200
Gas Booster
7,730
773
8,503
__
--
—
—
—
--
1,384
340
1,724
1,720
COST-EFFECTIVENESS ($/Mg VOC) 71,100/222.5 = $320
*Rounded off to three significant figures.
-------
TABLE 5-8. COST-EFFECTIVENESS ($/Mg of VOC DESTROYED) OF RACT FOR THE POLYMERS AND RESINS INDUSTRY
Polymer
Polypropylene
High -Density
Polyethylene
Polystyrene
(Continuous
Process)
Continuous Streams
Projected
Control Device
Thermal Incin.
Thermal Incin.
Thermal Incin.
Continuous Streams
Metric Tons
VOC Destroyed/yr
4800
2633
222.5
Continuous Streams
Artnuali zed
Costs ($1000)
106
121
71.1
Contiguous Streams
Cost-Effectiveness
($/Mg VOC)
22
46
320
-------
would remain substantially the same. The amount of VOC control would
decrease and the resultant cost-effectiveness ratio would increase.
The cost-effectiveness ratio of $46 per Mg of VOC for high-density
polyethylene was based on the operation of the incinerator while it was
being fed a high concentration stream (the recycle treater vent) and two
low-concentration streams (the dryer vent and the mixer vent). In some
plants the recycle treater may be down for as much as 4 out of 24 hours.
During such times it would be necessary to add natural gas to keep the
incinerator up to an efficient temperature. The cost of this gas would
cause the cost-effectiveness ratio to rise to $155/Mg of VOC during these
periods. The cost-effectiveness ratio for operating the incinerator
continuously would average $64/Mg of VOC over an entire year where the
recycle treater was down as much as 4 out of 24 hours on an annual average.
5.4.2 Cost-Effectiveness Considerations in Requiring Incineration of
Streams Now Being Flared
Most, if not all, polypropylene and polyethylene plants already have
flares on the site for safety purposes. Typically, the process is
instrumented so that any potentially dangerous abnormality in an
operating parameter, such as extremes in temperature or pressure, causes
the flammable gases and liquids inventoried in the process to be relieved
to a flare where they can be safely burned. This precludes the danger of
those materials drifting or falling within the battery limits of the
process where they could be inadvertently ignited, resulting in an
explosion or fire.
The availability of the flare already within a plant would appear to
offer the most cost-effective alternative for destroying the VOC in the
combined continuous streams. Unfortunately, it is not possible to verify
this because of the uncertainty of the combustion efficiency of a flare and
the absence of any generally accepted method for measuring flare efficiency.
Available flare studies have been evaluated in an effort to develop
destruction efficiency estimate for calculating emission impacts from
flares. The number of flare studies is very limited, and no data specifically
addresses conditions present within vent streams from polymer plants. Four
flare studies provide information on flare gas composition, flowrate, and
5-17
-------
destruction efficiency. The flare studies present flare destruction effici-
encies of: ethylene ranging from 98 to 99.99 percent,9 natural gas ranging
from 70 to 98 percent,10 propane ranging from 94 to 99.91 percent,11 and a
hydrogen dominated stream of 99 percent.12 These studies indicate that
destruction efficiencies of flares can be very good under experimental
conditions. Destruction efficiencies of 99 percent or greater can be expected
with the right combination of flare gas composition, gas flowrate, flare head
design, and weather (which affects mixing of exit gas and air). However, even
under the experimental conditions under which the studies were examined, flare
destruction efficiency covered a wide range from 70 to greater than 99 percent.
Based on the available studies and a comparison of vent stream
characteristics between these studies and process emission sources, the EPA
concluded that the destruction efficiency of the flares currently in place
within the polymer plants would be greater than the 70 percent, but lower than
the 99 percent projected by the flare studies. Furthermore, there are reasons,
other than the very noticeable black plumes sometimes observed while a flare is
in service, to question their efficacy. For example, flares are generally
sized by assuming that a catastrophic failure instantaneously releases to the
flare all combustable liquids and gases inventoried by the process. For this
reason, flares used in polyethylene and polypropylene plants are large. Use
of large flares for low-volume continuous streams may allow the waste to
channel from the flare tip resulting in poor mixing of air and incomplete
combustion. It is also conceivable that variations in flow and heat content of
the waste stream could extinguish the flame thereafter allowing emissions to
escape unabated.
The uncertainty associated with flare combustion contrasts starkly with
our knowledge of incinerators and boilers. Evidence to show the thoroughness
of combustion efficiency in these devices is ponderous. Furthermore, test
methods for determining the efficiency of a specific unit are readily available
and can be conducted at reasonable cost.
In order to assess the reasonableness of requiring incinerators for
these plants, the incremental cost effectiveness between a flare of
90 percent efficiency and an incinerator of 98 percent efficiency was
calculated for each plant. A flare on a polypropylene plant which emits
5-18
-------
4900 Mg of VOC/yr would destroy 4410 Mg VOC/yr while an incinerator would
destroy 4800 Mg/yr, an increment of 390 Mg/yr. The annualized cost of the
incinerator would be $56,200/yr. The incremental cost effectiveness would
be $56,200 7390, or $144/Mg. A flare of a high-density polyethylene plant
which emits 2690 Mg of VOC/yr would destroy 2420 Mg/VOC, while an incinerator
would destroy 2640 Mg/yr, an increment of 220 Mg/yr. The annualized cost
of the incinerator would be $57,000/yr. The incremental cost effectiveness
would be $265/Mg. These cost-effectiveness figures make it reasonable to
require that an incinerator be used to destroy the emissions from these plants.
5-19
-------
5.5 REFERENCES FOR CHAPTER 5
1. Neveril, R.B., Capital and Operating Costs of Selected Air Pollution
Control Systems, EPA Report No. EPA-450/5-80-002, Gard, Inc.,
Niles, Illinois
2. Compressor Handbook for the Hydrocarbon Processing Industries,
Prepared by Gulf Publishing Company Book Division, P.O. Box 2608,
Houston, Texas 77001, 1979.
3. Norwalk Air and Gas Compressors, Norwalk Company, Inc., P.O. Box 548,
N. Water Street, South Norwalk, Connecticut 06856
4. Vatavuk-Porter, EPA, memorandum, "Guidance for Developing CTGD Cost
Chapters,"September 17, 1980.
5. Chasko-Porter, EPA, memorandum, "Guidance for Developing CTGD Cost
Chapters," September 17, 1980.
6. Richardson, Process Plant Construction Estimating Standard, Process
Equipment, Vol. 4.
7. Chemical Engineering, Economic Indicators, Vol. 87, No. 16,
August 11, 1980.
8. Mascone, D.C., EPA, Draft memorandum concerning incinerator efficiency,
April 25, 1980.
9. Palmer, P.A. A Tracer Technique for Determining Efficiency of an
Elevated Flare, E.I. DuPont de Nemours & Co., 1972
10. Straitz, J.F., III, Flaring for Gaseous Control in the Petroleum
Industry, National Air Oil Burner, June, 1978.
11. Lee, K.C., and G.M. Whipple, Waste Gaseous Hydrocarbon Combustion in a
Flare, Union Carbide Corporation, June, 1981.
12. Seigel, K.D., Degree of Conversion of Flare Gas in Refinery High Flares,
Karlstrohe University, February, 1980.
5-20
-------
APPENDIX A: EMISSION SOURCE TEST DATA
Introduction
The material in this appendix is the Appendix A which was compiled
to provide source test data for the development of the control techniques
guideline (CTG) document for air oxidation orocesses of the synthetic
organic chemicals manufacturing industry (SOCMI). Since the nature and
characteristics of VOC streams in the polymer and resins industry are
similar to those of the air oxidation industries, the tests can be
used as a basis for the control level recommended as achievable by RACT
for VOC streams in polymerization processes also.
The size of streams involved in the polymer industry varies around
500 scfm, much smaller than those in the tests from which the data was
obtained. However, one test on a small lab-scale unit achieved a
minimum of 99.6 weight percent destruction efficiency at 1600°F and
0.75 second residence time. Thus, for properly designed incinerators,
it appears that the size of the stream does not adversely affect
destruction efficiency and the results of these tests should be a
satisfactory basis for control levels to be expected in the polymers and
resins industry.
A-l
-------
EMISSION SOtlfiCE TEST DATA
The purpose of this appendix is to describe results of tests of
volatile organic compound (VOC) emissions reduction by thermal incineration.
These test results were used in the development of the control techniques
guideline (CTG) document for air oxidation processes of the synthetic
organic chemicals manufacturing industry (SQEMI). Background data and
detailed information which support the emission levels achievable are
included.
Section A.I of this appendix presents the VOC emissions test data
including individual test descriptions. Section A.2 provides a summary
of NO emissions from some of the tests. Section A.3 consists of comparisons
/\
of various test results and a discussion exploring and evaluating the
similarities and differences of these results,
A.I VOC EMISSIONS TEST DATA
The tests were aimed at evaluating the performance of thermal
incinerators when used under varied conditions on the air oxidation
process waste streams. The results of this study indicate that 98
percent VOC reduction or 20 ppmv by compound exit concentration, whichever
is less stringent, is the highest control level currently achievable by
all new incinerators, considering available technology, cost, and energy
use. This level is expressed in both percent reduction and ppmv to
account for the leveling off of exit concentrations as inlet concentrations
drop. This level can be achieved by incinerator operation at conditions
which include a maximum of 1600°F and 0.75 second residence time. The
98 percent level can frequently be achieved at lower combustion temperatures.
Three sets of test data are available. These sets consist of field
unit data from tests conducted by EPA and by chemical companies and of
lab-scale incinerator data from tests by Union Carbide.
A.1.1 Chemical Company Test Data
These data are from tests performed by chemical companies on incinerators
at three air oxidation units: the Petro-tex oxidative butadiene unit at
Houston, Texas, the Koppers maleic anhydride unit at Bridgeville, Pennsylvania,
and the Monsanto acrylonitrile unit at Alvin, Texas.
A-2
-------
2
A. 1.1.1 Petro-Tex Test Data
1. Facility and Control Device - The Petro-tex incinerator for the
'Oxo1 butadiene process is designed to treat 48,000 scfm waste gas containing
about 4000 ppm hydrocarbon and 7000 ppm carbon dioxide. The use of the term
hydrocarbon in this discussion indicates that besides VOC, it may include non-
VOC such as methane. The waste gas treated in this system results from air
used to oxidize butene to butadiene. The waste gas, after butadiene has been
recovered in an oil absorption system, is combined with other process waste
gas and fed to the incinerator. The waste gas enters the incinerator between
seven vertical Coen duct burner assemblies. The incinerator design incor-
porates flue gas recirculation and a waste heat boiler. The benefit achieved
by recirculating flue gas is to incorporate the ability to generate a constant
100,000 Ibs/hr of 750 psi steam with variable waste gas flow.3 The waste gas
flow can range from 10 percent to 100 percent of design production rate.
The incinerator measures 72 feet by 20 feet by 8 feet, with an average
firebox cross-sectional area of 111 square feet. The installed capital cost
was $2.5 million.
The waste gas stream contains essentially no oxygen; therefore, significant
combustion air must be supplied. This incinerator is fired with natural gas
which supplies 84 percent of the firing energy. The additional required
energy is supplied by the hydrocarbon contamination of the waste gas stream.
Figure A-l gives a rough sketch of this unit.
2. Sampling and Analytical Techniques
Waste Gas
The waste gas sampling was performed with integrated bags. The analysis
was done on a Carle analytical gas chromatograph having the following columns:
1. 6-ft OPN/PORASILR (80/100).
2. 40-ft 20 percent SEBACONITRILER on gas chrom. RA 42/60.
3. 4-ft PORAPAKR N 80/100.
4. 6-ft molecular sieve bx 80/100.
A.-3
-------
(Supplemental)
Air Duct
Recirculation
Air Duct
RECIRCULATION
AIR FAN
Figure A-1. Petro-Tex oxo unit incinerator,
A-4
-------
Stack Gas
The stack gas samples are collected via a tee on a long stainless
steel probe which can be inserted into the stack at nine different
locations. These gas samples are collected in 30-50 cc syringes.
The gas samples are then transferred to a smaller 1 cc syringe via
a small glass coupling device sealed at both ends with a rubber grommet.
The 1-cc samples can then be injected into a chromatograph for hydro-
carbon analysis. A Varian 1700 chromatograph is used, having a 1/8-in.
x 6-ft column packed with 5A molecular sieves and a 1/4-in. x 4-ft
column packed with glass beads connected in series with a bypass before
and after the molecular sieve column, controlled by a needle valve to
split the sample. The data are reported as ppm total HC, ppm methane,
and ppm non-methane, hydrocarbons (NMHC). The CO content in the stack is
determined by using a Kitagawa sampling probe. The 0^ content in the
stack is determined via a Teledyne 02/combustible analyzer.
3. Test Results - Petro-tex has been involved in a modification
plan for its 'Oxo1 incinerator unit after startup. The facility was
tested by the company after each major modification was made to determine
the impact of these changes on the VOC destruction efficiency. The
incinerator showed improved performance after each modification and the
destruction efficiency increased from 70 percent to well above 98 percent.
Table A-l provides a summary of these test results. The type of modifications
made in the incinerator were as follows:
November 1977
Test data prior to these changes showed the incinerator was not
destroying hydrocarbons as well as it should (VOC destruction efficiency
as low as 70 percent), so the following changes were made:
1. Moved the duct burner baffles from back of the burner to the
front.
2. Installed spacers to create a continuous slot for supplemental
air to reduce the air flow through the burner pods.
3. Installed plates upstream of the burners so that ductwork
matches burner dimensions.
4. Cut slots in recycle duct to reduce exit velocities and improve
mixing with Oxo waste gas.
A-5
-------
TABLE A-l. THERMAL INCINERATOR FIELD TEST DATA
cr>
...
Company & Location Type of Process
Petro-tex Chemical Butadiene
Corp. , Houston,
TX
Koppers Co., Inc. Maleic Anhydride
Bridgeville, PA
Monsanto Chemical Acryloni trile
Intermediates Co. ,
Alvin, TX
Denka, Houston, TX Maleic Anhydride
Rohm & Haas, Acrylic Acid &
Deer Park, TX Esters
Union Carbide Acrylic Acid &
Corp., Taft, LA Acrylate Esters
Production Rate
During Test
Waste Gas Flow
(Inlet) scfm
7,250
15,617
20,750
15,867
12,500
Avg. Combustion
Air: 49,333
33,200
Air: 8000
24,200
Air: 2000
75,000 (Avg.)
33,000
(70% of total
capacity)
Each 52,500
(12,500 tank
fflTITl VGflt
(TVF))
(40,000 oxidi-
zer vent (OXV))
20,600
Number
of Tests
or Sets
Set 1
Set 2
Set 3
Set 4
Set 5
Set 1
Set 2
Unit 1
Unit 2
3
Set 1
3
Set 2
Set 3
Set 1
6
Set 2
3
Test Date
5/25/77
9/09/77
12/01/77
4/19/78
9/27/78
11/02/77
11/16/77
12/16/77
12/1 6/.77
3/21/78
3/22/78
3/23/78
3/78
3/78
3/78
12/78
12/78
Supplemental
Fuel Residence Incineration
& Amount Time Temperature
Used (scfm) (Seconds) (°F)
Natural
Gas
1400
1467
900
1175
1176
Natural
Gas
Natural
Gas
1060 (gas)
1060
1060
900 (gas)
900
900
Natural
Gas
0.6
0.6
0.6
0.6
0.6
0.6
0.6
N/A
N/A
0.6
0.6
0.6
1.0
1.0
1.0
2-3
2-3
1400
1400
1400
1400
1400
"Below
2000"
Confiden-
dential
1400
1400
1400
1425 TVF
OXV
1510 TVF
OXV-
1545 TVF
OXV
1160
1475
Inlet
VOC
(ppmv)
10,300
10,650
10,650
10,300
10,300
834
834
Outlet
VOC
( ppmv )
1000
215
215
10
10
7
8
Confiden- 25
dential
950
950
950
2,580
11,600
2,600
12,800
2,410
12,200
11,900
ii.goa
47
13
13
13
1330
150
25
Z43
10
VOC
Destruction
Efficiency
by Weight
70.3
94.1
94.1
99.6
99.6
98.96
98.96
>99
>99
98.5
98.5
98.5
82.6
98.3
99.7
96.-. 1
9^9
-------
5. Installed balancing dampers in augmenting (supplemental) air
plenums, top and bottom.
6. Installed balancing dampers in three of the five sections of
the recycle duct transition.
7. Cut opening in the recirculation duct to reduce the outlet
velocities.
March 1978
After the November changes were made, a field test was made in
December 1977, which revealed that the incinerator VOC destruction
efficiency increased from 70.3 percent to 94.1 percent. However, it
still needed improvement. After much discussion and study the following
changes were made in March 1978:
1. Took the recirculation fan out of service and diverted the
excess forced draft air into the recirculation duct.
2. Sealed off the 5-1/2-in. wide slots adjacent to the burner
pods and removed the 1/2-in. spacers which were installed in
November 1977.
3. Installed vertical baffles between the bottom row of burner
pads to improve mixing.
4. Installed perforated plates between the five recirculation
ducts for better Oxo waste gas distribution.
5. Cut seven 3-in. wide slots in the recycle duct for better
secondary air distribution.
July 1978
After the March 1978 changes, a survey in April 1978, showed the
Oxo incinerator to be performing very well (VOC destruction efficiency
of 99.6 percent) but with a high superheat temperature of ~850°F. So,
in July 1978, some stainless steel shields were installed over the
superheater elements to help lower the superheat temperature. A subsequent
survey in September 1978, showed the incinerator to still be destructing
99.6 percent VOC and with a lower superheat temperature (~750°F).
This study pointed out that mixing is a critical factor in efficiency
and that incinerator adjustment after startup is the most feasible and
efficient means of improving mixing and thus, the destruction efficiency.
A-7
-------
4
A.1.1.2 Koppers Test Data
1. Facility and Control Device - The: Koppers incinerator is
actually a boiler adapted to burn gaseous wastes from maleic anhydride
unit. The boiler is designed to operate at a temperature of 2000°F and
a residence time of 0.6 second. Current operating parameters have not
been measured, but it is the company's judgement that the boiler now
operates somewhat below 2000°F. The flowrate of waste gas to the boiler
is usually 32,000 scfm and contains 350 Ibs/hr benzene, 2850 Ibs/hr
carbon monoxide, 22,100 Ibs/hr oxygen, 6434 Ibs/hr water, and 105,104
Ibs/hr nitrogen. While these values are typical for the system, they
vary throughout the production cycle. The boiler is fired with natural
gas.
2. Sampling and Analytical Techniques - Different methods were
used for inlet and outlet sampling. Although integrated samples were
used for the outlet, gas bottle samples were used for the inlet. Such a
sampling technique would likely give a low bias to the measured inlet
VOC concentration.
The inlet concentration was taken to be the average of all maleic
reactor offgas measurements, made. There were four samples taken, and
the results were 600 ppmv, 1172 ppmv, 600 ppmv, and 964 ppmv for an
average of 834 ppmv benzene. (These values are not boiler inlet values
since they were collected prior to the introduction of the additional
combustion air.) This wide range of benzene values indicates the great
deal of variability inherent in efficiency calculations employing such a
sampling technique.
For the June 1978 tests, samples of stack gas were taken in glass
bottles by plant chemists and analyzed at Koppers' Monroeville Research
Center by direct injection to a gas chromatograph with flame ionization
detector. The November 1977, method used specially-designed charcoal
adsorption tubes, instead of impingers, in a United States Environmental
Protection Agency-type sampling train. The charcoal was eluted with CS9
and the eluent analyzed by gas chromatography.
3. Test Results - One test run of the Koppers data indicates 97.2
percent efficiency at 1800°F. However, the entire Koppers test is
disregarded as not demonstrably accurate because of the poor sampling
A-8
-------
technique. Grab samples employed in obtaining inlet gas could give a
low bias to the measured, inlet VOC concentration. Therefore, the calculated
VOC destruction efficiency would be artificially low. Table A-l provides
a summary of these test results.
A.1.1.3 Monsanto Test Data
1. Facility and Control Device - The Monsanto incinerator burns
both liquid and gaseous wastes from the acrylonitrile unit and is termed
an absorber vent thermal oxidizer. Two identical oxidizers are employed.
The primary purpose of the absorber vent thermal oxidizers is hydrocarbon
emission abatement.
Acrylonitrile is produced by feeding propylene, ammonia, and excess
air through a fluidized, catalytic bed reactor. In the process, acrylonitrile,
acetonitrile, hydrogen cyanide, carbon dioxide, carbon monoxide, water,
and other miscellaneous organic compounds are produced in the reactor.
The columns in the recovery section separate water and crude acetonitrile
as liquids. Propane, unreacted propylene, unreacted air components,
some unabsorbed organic products, and water are emitted as a vapor from
the absorber column overhead. The crude acrylonitrile product is further
refined in the purification section to remove hydrogen cyanide and the
remaining hydrocarbon impurities.
The organic waste streams from this process are incinerated in the
absorber vent thermal oxidizer at a temperature and residence time
sufficient to reduce stack emissions below the required levels. The
incinerated streams include (1) the absorber vent vapor (propane, propylene,
CO, unreacted air components, unabsorbed hydrocarbons), (2) liquid waste
acetonitrile (acetonitrile, hydrogen cyanide, acrylonitrile), (3) liquid
waste hydrogen cyanide, and (4) product column bottoms purge (acrylonitrile,
some organic heavies). The two separate acrylonitrile plants at Chocolate
Bayou, employ identical thermal oxidizers.
Each thermal oxidizer is a horizontal, cylindrical, saddle-supported,
end-fired unit consisting of a primary burner vestibule attached to the
main incinerator shell. Each oxidizer measures 18 feet in diameter by
36 feet in length.
The thermal oxidizer is provided with special burners and burner
guns. Each burner is a combination fuel-waste liquid unit. The absorber
A-9
-------
vent stream is introduced separately into the top of the burner vestibule.
The flows of all waste streams are metered and sufficient air is added
for complete combustion. Supplemental natural gas is used to maintain
the operating temperature required to combust the organics and to maintain
a stable flame on the burners during minimum gas usage. Figure A-2
gives a plan view of the incinerator.
2. Sampling and Analytical Techniques
Feed Stream and Effluent
The vapor feed streams (absorber vent) to the thermal oxidizer and
the effluent gas stream are sampled and analyzed using a modified analytical
reactor recovery run method. The primary recovery run methods are Sohio
Analytical Laboratory Procedures.
The modified method involves passing a measured amount of sample
gas through three scrubber flasks containing water and catching the
scrubbed gas in a gas sampling bomb. The samples are then analyzed with
a gas chromatograph and the weight percent of the components is determined.
Stack Gas
Figure A-3 shows the apparatus and configuration used to sample the
stack gas. It consists of a line of the sample valve to the small
water-cooled heat exchanger. The exchanger is then connected to a
250 ml sample bomb used to collect the unscrubbed sample. The bomb is
then connected to a pair of 250 ml bubblers, each with 165 ml of water
in it. The scrubbers, in turn, are connected to another 250 ml sample
bomb used to collect the scrubbed gas sample which is connected to a
portable compressor. The compressor discharge then is connected to a
wet test meter that vents to the atmosphere.
After assembling the apparatus, the compressor is turned on and it
o
draws gas from the stack and through the system at a rate of ~0.2 ft /min.
Sample is drawn until at least 10 ft have passed through the scrubbers.
After 10 ft has been scrubbed, the compressor is shutdown and the
unscrubbed bomb is analyzed for CH^, C2's, C3Hg, and C3Hg, the scrubbed
bomb is analyzed for N-, air, 0-, CO^, and CO, and the bubbler liquid is
analyzed for acrylonitrile, acetonitrile, hydrogen cyanide, and total
organic carbon. The gas samples are analyzed by gas chromatography.
A-10
-------
'•>••
PLAN VIEW
Figure A-2. Off-gas incinerator, Monsanto Co., Chocolate Bayou Plant.
-------
3=
I
HEAT
EXCHANGER
Cooling Watei
IN
Cooling Water
TEMPERATURE
INDICATOR
2SOHL
SAMPLE BOMB
( Unsciubbed Sample )
TOO 250 ML BUBBLERS
With 165 ML Distilled Water in Each.
« ^ ;fi
J>>J^>x>7»]>
Wti
s%? '/
I
Bucket Contains Wet Ice Slush
VENT TO
ATMOSPHERE
10 Feel Above Grade
PORTABLE
COMPRESSOR
2SOML
SAMPLE BOMB
( Scrubbed Sample )
TEST
WTEH
NOTE: From Exchanger Process Outlet, All Lines are Vacuum Tubing.
Figure A—3. Thermal incinerator stack sampling system.
-------
For the liquid samples, acrylonitrile and acetonitrile are by gas
chromatography; hydrogen cyanide (HCN) is by titration; and total organic
carbon.(TOC) is by a carbon analysis instrument.
3. Test Results - Monsanto's test results show efficiencies well
above 98 percent, however, the parameters at which it is achieved are
confidential. All other known conditions are presented in Table A-l.
A.1.2 Environmental Protection Agency (EPA) Test Data
The EPA test study represents the most in-depth work available.
These data show the combustion efficiencies for full-scale incinerators
on air oxidation vents at three chemical plants. Data includes inlet/outlet
tests on large incinerators, two at acrylic acid plants, and one at a
maleic anhydride plant. The tests measured inlet and outlet VOC by
compound at different temperatures, and the reports include complete
test results, process rater., and test method descriptions. The three
plants tested, are the Denka, Houston, Texas, maleic anhydride unit and
the Rohm and Haas, Deer Park, Texas, and Union Carbide, Taft, Louisiana,
acrylic acid units. The data from Union Carbide include test results
based on two different incinerator temperatures. The data from Rohm and
Haas include results for three temperatures. In all tests, bags were
used for collecting integrated samples and a GC/FID was used for organic
analysis.
A.1.2.1 Denka Test Data6
1. Facility and Control Device - The Denka maleic anhydride
facility has a nameplate capacity of 23,000 Mg/yr (50 million Ibs/yr).
The plant was operating at about 70 percent of capacity when the sampling
was conducted. The plant personnel did not think that the lower production
rate would seriously affect the validity of the results.
Maleic anhydride is produced by vapor-phase catalytic oxidation of
benzene. The liquid effluent from.the absorber, after undergoing recovery
operations, is about 40 weight percent aqueous solution of maleic acid.
The absorber vent is directed to the incinerator. The thermal incinerator
uses a heat recovery system to generate process steam and uses natural
gas as supplemental fuel. The size of the combustion chamber is 2195 ft2.
There are three thermocouples used to sense the flame temperature, and
A-l 3
-------
these are averaged to give the temperature recorded in'the control room. A
rough sketch of the combustion chamber is provided in Figure A-4.
2. Sampling and Analytical Techniques
THC. Benzene, Methane, and Ethane
The gas samples were obtained according to the September 27, 1977,
D
EPA draft benzene method. Seventy-liter aluminized Mylar bags were used
with sample times of two to three hours. The sample box and tag were
heated to approximately 66°C fl50°F) using an electric drum heater and
insulation. During Run 1-Inlet, the variac used to control the temperature
malfunctioned so the box was not heated for this run. A stainless steel
probe was inserted into the single port at the inlet and connected to
the gas bag through a "tee". The other leg of the "tee" went to the
R
total organic acid (TOA) train. A Teflon line connected the bag and
the "tee". A stainless steel probe was connected directly to the bag at
the outlet. The lines were kept as short as possible and not heated.
The boxes were transported to the field lab immediately upon completion
of sampling. They were heated until the GC analyses were completed.
A Varian model 2440 gas chromatograph with a Carle gas sampling valve,
3
equipped with two cm matched loops, was used for the integrated
bag analysis. The SP-1200/Bentone 34 column was operated at 80°C. The
instrument has a switching circuit which allows a bypass around the
column through a capillary tube for THC response. The response curve
was measured daily for benzene (5, 10, and 50 ppm standards) with the
column and in the bypass (THC) mode. The THC mode was also calibrated
daily with propane (20, 100, and 2000 ppm standards). The calibration
plots showed moderate nonlinearity. For sample readings which fell
within the range of the calibration standards, an interpolated response
factor was used from a smooth curve drawn through the calibration points.
For samples above or below the standards, the response factor of the
nearest standard was assumed. THC readings used peak height and column
readings used area integration measured with an electronic "disc" *
integrator.
CO
Analysis for these constituents was done on samples drawn from the
integrated gas bag used in THC, benzene, methane, and ethane. Carbon
A-14
-------
12 ft
FLOW
SIDE VIEW
(Inlet)
23ft-3Jin
17ft-Sin
I
(Outlet)
There are Three Thermocouples Spaced Evenly Across the Top of the Firebox.
The Width of the Firebox is 6ft-6 in.
Figure A—4. Incinerator combustion chamber.
A-15
-------
monoxide analysis was done following the GC analyses using EPA Reference
Method 10 (Federal Register, Vol. 39, No. 47, March 8, 1974). A Beckman
Model 215 NDIR analyzer was used to analyze both the inlet and- outlet
samples.
Duct Temperature, Pressure, and Velocity
Duct temperature and pressure values were obtained, from the existing
inlet port. A thermocouple was inserted into the gas sample probe for
the temperature while a water manometer was used for the pressure readings.
These values were obtained at the conclusion of the sampling period.
Temperature, pressure, and velocity values were obtained for the
outlet stack. Temperature values were obtained by thermocouple during
the gas sampling. Pressure and velocity measurements were taken according
to EPA Reference Method 2 (Federal Register, Vol. 42, No. 160,
August 18, 1977). These values also were obtained at the conclusion of
the sampling period.
2. Test Results - The Denka incinerator achieves greater than 98
percent reduction at 1400°F and 0.6 second residence time. These results
suggest that the recommended 98 percent control level is achievable by
properly maintained and operated new incinerators, for which the operating
conditions are less stringent than 1600°F and 0.75 second. Table A-l
provides a summary of these test results.
A.1.2.2 Rohm and Haas Test Data
1. Facility and Control Device - The Rohm and Haas plant in Deer
Park, Texas, produces acrylic acid and ester. The capacity of this
facility has been listed at 400 million Ibs/yr of acrylic monomers.
Acrylic esters are produced using propylene, air, and alcohols, with
acrylic acid produced as an intermediate. Acrylic acid is produced
directly from propylene by a vapor-phase catalytic air oxidation process.
The reaction product is purified in subsequent refining operations.
Excess alcohol is recovered and heavy end by-products are incinerated.
This waste incinerator is designed to burn offgas from the two absorbers.
In addition, all process vents (from extractors, vent condensers, and
tanks) which might be a potential source of gaseous emissions are collected
in a suction vent system and normally sent to the incinerator. An
A-16
-------
organic liquid stream generated in the process is also burned, thereby
providing part of the fuel requirement. The remainder is provided by
natural gas. Combustion air is added in an amount to produce six percent
oxygen in the effluent. Waste gases are flared during maintenance
shutdowns and severe process upsets. The incinerator unit was tested
because it operates at relatively shorter residence times (0.75-1.0
seconds) and higher combustion temperatures (1200°-1560°F) than most
existing incinerators.
The total installed capital cost of the incinerator was $4.7 million.
The estimated operating cost due to supplemental natural gas use is $0.9
million per year.
2. Sampling and Analytical Techniques - Samples were taken
simultaneously at a time when propylene oxidations, separations, and
esterifications were operating smoothly and the combustion temperature
was at a steady state. Adequate time was allowed between the tests
conducted at different temperatures for the incinerator to achieve
steady state. Bags were used to collect integrated samples and a GC/FID
was used for organic analysis.
3. Test Results - VOC destruction efficiency was determined at
three different temperatures: 1425°F, 1510°F, and 1545°F. Efficiency
is found to increase with temperature and, except for 1425°F, it is
above 98 percent. Test results are summarized in Table A-l. These
tests were for residence times greater than 0.75 second. However,
theoretical calculations show that greater efficiency would be achieved
at 1600°F and 0.75 second than at the longer residence times, but lower
temperatures represented in these tests.
A. 1.2.3 Union Carbide (UCC) Test Data8
1. Facility and Control Device - The capacities for the UCC
acrylates facilities are about 200 million Ibs/yr of acrolein, acrylic
acid, and esters. Acrylic acid comprises 130 million Ibs/yr of this
total. Ethyl acrylate capacity is 90 million Ibs/yr. Total heavy ester
capacities (such as 2-ethyl-hexyl acrylate) are 110 million Ibs/yr. UCC
considers butyl acrylate a heavy ester.
The facility was originally built in 1969 and utilized British
Petroleum technology for acrylic acid production. In 1976 the plant was
converted to a technology obtained under license from Sohio.
A-l 7
-------
The thermal incinerator is one of the two major control devices
used in acrylic acid and acrylate ester manufacture. The UCC incinerator
was installed in 1975 to destroy acrylic acid and acrolein vapors. This
unit was constructed by John Zink Company for an installed cost of $3
million and incorporates a heat recovery unit to produce process steam
at 600 psig. The unit operates at a relatively constant feed input and
supplements the varying flow and fuel value of the streams fed to it
with inversely varying amounts of fuel gas. Energy consumption averages
52.8 million Btu/hr instead of the designed level of 36-51 million
B'tu/hr. The operating cost in 1976, excluding capital depreciation, was
$287,000. The unit is run with nine percent excess oxygen instead of
the designed three to five percent excess oxygen. The combustor is
designed to handle a maximum of four percent propane in the oxidation
feed.
Materials of construction of a non-return block valve in the
600 psig steam line from the boiler section requires that the incinerator
be operated at 1200°F instead of the designed 1800°F. The residence
time is three to four seconds.
2. Sampling and Analytical Procedures - -The integrated gas samples
were obtained according to the September 27, 1977, EPA draft benzene
method.
Each integrated gas sample was analyzed on a Varian Model 2400 gas
chromatograph with FID, and a heated Carle gas sampling valve with
o
matched 2 cm sample loops. A valved capillary bypass is used for total
hydrocarbon (THC) analyses and a 2 m, 1/8-in., 00 nickel column with
D
PORAPAK P-S, 80-100 mesh packing is used for component analyses.
Peak area measurements were used for the individual component
analyses. A Tandy TRS-80, 48K floppy disc computer interfaced via the
integrator pulse output of a Linear Instruments Model 252A recorder
acquired, stored, and analyzed the chromatograms.
The integrated gas samples were analyzed for oxygen and carbon
dioxide by duplicate Fyrite readings. Carbon monoxide concentrations
were obtained using a Bec.kman Model 215A nondispersive infrared (IR)
analyzer using the integrated samples. A three-point calibration (1000,
3000, and 10,000 ppm CO standards) was used with a linear-log curve fit.
A-18
-------
Stack traverses for outlet flowrate were made using EPA Methods 1
through 4 (midget impingers) and NOX was sampled at the outlet using EPA
Method 7.
3. Test Results - VOC destruction efficiency was determined at
two different temperatures. Table A-l provides a summary of these test
results. Efficiency was found to increase with temperature. At 1475°F,
the efficiency was well above 99 percent. These tests were, again, for
residence times greater than 0.75 second. However, theoretical calculations
show that greater efficiency would be achieved at 1600 F and 0.75 second
than at the longer residence times but lower temperatures represented in
these tests.
All actual measurements were made as parts per million (ppm) of
propane with the other units reported derived from the equivalent values.
The values were measured by digital integration.
The incinerator combustion temperature for the first six runs was
about 1160°F. Runs 7 through 9 were made at an incinerator temperature
of about 1475°F. Only during Run 3 was the acrolein process operating.
The higher temperature caused most of the compounds heavier than propane
to drop below the detection limit due to the wide range of attenuations
used, nearby obscuring peaks, and baseline noise variations. The detection
limit ranges from about 10 ppb to 10 ppm, generally increasing during
the chromatogram, and especially near large peaks. Several of the minor
peaks were difficult to measure. However, the compounds of interest,
methane, ethane, ethylene, propane, propy.lene, acetaldehyde, acetone,
acrolein, and acrylic acid, dominate the chromatograms. Only acetic
acid was never detected in any sample.
The probable reason for negative destruction efficiencies for
several light components is generation by pyrolysis from other components.
For instance, the primary pyrolysis products of acrolein are carbon
monoxide and ethylene. Except for methane and, to a much lesser extent,
ethane and propane, the fuel gas cannot contribute hydrocarbons to the
outlet samples.
A sample taken from the inlet line knockout trap showed 6 yg/g of
acetaldehyde, 25 yg/g of butenes, and 100 yg/g of acetone when analyzed
by gas chromatography/flame ionization detection (GC/FID).
A-19
-------
A.1.3 Union Carbide Lab-Scale Test Data
Union Carbide test data show the combustion efficiencies achieved
on 15 organic compounds in a lab-scale incinerator operating between 800°
and 1500 F and .1 to 2 seconds residence time. The incinerator consisted
of a 130 cm, thin bore tube, in a bench-size tube furnace. Outlet
analyzers were done by direct routing of the incinerator outlet to a FID
and GC. All inlet gases were set at 1000 ppmv.
In order to study the impact of incinerator variables on efficiency,
mixing must first be separated from the other parameters. Mixing cannot
be measured and thus, its impact on efficiency cannot be readily separated
when studying the impact of other variables. The Union Carbide lab work
was chosen since its small size and careful design best assured consistent
and proper mixing.
The results of this study are shown in Table A-2. These results
show moderate increases in efficiency with temperature, residence time,
and type of compound. The results, also show the impact of flow regime
on efficiency.
Flow regime is important in interpreting the Unio'n Carbide lab unit
results. These results are significant since the lab unit was designed
for optimum mixing and thus, the results represent the upper limit of
incinerator efficiency. As seen in Table A-2, the Union Carbide results
vary by flow regime. Though some large-scale incinerators may achieve
good mixing and plug flow, the worst cases will likely require flow
patterns similar to complete backmixing. Thus, the results of complete
backmixing would be, relatively, more comparable to those obtained from
large-scale units.
A.2 NITROGEN OXIDES (NO) EMISSIONS
/\
Nitrogen oxides are derived mainly from two sources: (1) from
nitrogen contained in the combustion air called thermal NO. and (2)
X
from nitrogen chemically combined in the fuel, called fuel NO . In
X
addition, combustion of waste gas containing high levels of nitrogen-
containing compounds also may cause increases in NO emissions. For
/\
fuels containing low amounts of nitrogen, such as natural gas and light
distillate oils, thermal NOV is by far the larger component of total NO
A-20
-------
TABLE A-2. RESULTS OF DESTRUCTION EFFICIENCY UNDER STATED
CONDITIONS (UNION CARBIDE TESTS*)
Residence Time/Compound
0.75 second
Flow . Temperature
Regime0 (°F)
Two-stage
Backmixing
Complete
Backmixing
Plug Flow
1300
1400
1500
1600
1300
1400
1500
1600
1300
1400
1500
1600
Ethyl
Aery late
99.9
99.9
99.9
99.9
98.9
99.7
99.9
99.9
99.9
99.9
99.9
99.9
Ethanol
94.6
99.6
99.9
99.9
86.8.
96.8
.99.0
99.7
99.9
99.9
99.9
99.9
Ethyl ere
92.6
99.3
99.9
99.9
84.4
95.6
98.7
99.6
99.5
99.9
99.9
99.9
Vinyl
Chloride
78.6
99.0
99.9
99.9
69.9
93.1
98.4
99.6
90.2
99.9
99.9
99.9
.5 & 1.5 sec
Ethyl ene
87.2/27.6
98.6/99.8
99.9/99.9
99.9/99.9
78.2/91.5
93.7/97.8
98.0/99.0 ..
99.4/99.8
97.3/99.9
99.9/99.9
99.9/99.9
99.9/99.9
The results of the Union Carbide work are presented as a series of. equations. These
equations relate destruction efficiency to temperature, residence time, and flow
regime for each of 15 compounds. The efficiencies in this table were calculated
from these equations.
Three flow regimes are presented: two-stage backmixing, complete backmixing, and
plug flow. Two-stage backmixing is considered a reasonable approximation of actual
field units, with complete backmixing and plug flow representing the extremes.
A-21
-------
emissions. By contrast, fwel NO predominates for heavy oils, coal, and
J\
oth'er high-nitrogen fuels such as coal-derived fuels and shiHe oils.
Table A-3 provides a summary of N€L emissions data obtained from
A
industrial-scale thermal oxidizers. Results from three different units
indicate that emissions range from 8 to 200 ppmv, although these values
could increase by several orders of magnitude in a poorly designed or
operated unit. NO samples were obtained according to the EPA Reference
Method 1,
A.3 COMPARISON OF TEST RESULTS AND THE TECHNICAL BASIS OF THE SOCMI
AIR OXIDATION EMISSIONS LIMIT
This section compares various test results, discusses data and
findings on incinerator efficiency, and presents the logic and the
technical basis behind the choice of the above control level.
Published literature indicates that any VOC can be oxidized to
carbon dioxide and water if held at sufficiently high temperatures in
the presence of oxygen for a sufficient time. However, the temperature
at which a given level of VOC reduction is achieved is unique for each
VOC compound. Kinetic studies indicate that there are two slow or rate-
determining steps in the oxidation of a compound. The first is the
initial reaction in which the original compound disappears. It has been
determined that the initial reaction of methane (CH») is slower than
that of any other organic compound. Kinetic calculations show that, at
1600°F, 98 percent of the original methane will react in 0.3 seconds.
After the initial step, extremely rapid free radical reactions occur.
Each carbon atom will exist as carbon monoxide (CO) before oxidation is
complete. The oxidation of CO is the second slow step. Calculations
show that, at 1600°F, 98 percent of an original concentration of CO will
react in 0.05 second. Therefore, any VOC would be expected to be 98
percent destroyed at 1600°F in about 0.35 second. The calculations on
which this conclusion is based have taken into account the low mole
fractions of VOC and oxygen which would be found in the actual system.
They have also provided for the great decrease in concentration per unit
volume due to the elevated temperature. But the calculations assume
.perfect mixing of the offgas and combustion air. Mixing is therefore
identified from a theoretical viewpoint as the crucial design parameter.
A-22
-------
TABLE A-3. SUMMARY OF RESULTS: TO DATA
A
Company
Number of Sets
and/or
Number of Runs
Outlet NO
in Flue Gts
(ppmv)
Union Carbide
Set 1
(6)
Set 2
(3)
27
30
Denka
Set 1
Set 2
Set 3
9.3
10.2
8.0
Monsanto
Unit 1
Unit 2
200
8
A-23
-------
The test results both indicate a?n achievable control level of 98
percent at or below 1600°F and illustrate the. importance of mixing.
Union Carbide results on lab-scale incinerators indicated a minimum of
98.6 percent efficiency at 1400°F. Since lab-scale incine*?aiors primarily
differ from field units in their excellent mixing, these results verified
the theoretical calculations. The tests cited in Table A<-1 are documented
as being conducted on full-scale incinerators controlling offgas from
air oxidation process vents of a variety of types of plants. To focus
on mixing, industrial units were selected where all variables except
mixing were held constant or accounted for in other ways. It was then
assumed any changes in efficiency would be due to changes in mixing.
The case most directly showing the effect of mixing is that of
Petro-tex incinerator. The Petro-tex data show, the efficiency changes
due to modifications on the incinerator at two times after startup.
These modifications included (1) repositioning baffles, (2) adjusting
duct slots and openings in the mixing zone to improve exit velocity, (3)
installing new dampers, baffles and perforated plates, and (4) rerouting
inlet combustion air. These modifications increased efficiency from 70
percent to over 99 percent, with no significant,change in temperature.
A comparison indirectly showing the effect of mixing is that of the
Rohm and Haas test versus the Union Carbide Tab test as presented in
Table A-4. These data compare the efficiency of the Rohm and Haas (R&H)
incinerator in combusting four specific compounds with that of the Union
Carbide lab unit. The lab unit clearly outperforms the R&H unit, The
data from both units are based on the same temperature, residence time,
and inlet stream conditions. The more complete mixing of the lab unit
is judged the cause of the differing efficiencies. The six tests of in-
place incinerators do not, of course, cover every feedstock. However,
the theoretical discussion given above indicates that any VQC compound
should be sufficiently destroyed at 16QQ°F. Mare critical than the type
of VOC is the VOC concentration in the offgas. This is true because the
kinetics of combustion are not exactly first-order at low VQC concen-
trations. The Petro-tex results are for a bi-tadiene plant, and butadiene
offgas tends to be lean in VOC. Therefore, test results support the
validity of the standard for lean streams.
A-24
-------
TABLE A-4. RESULT COMPARISONS OF LAB INCINERATOR vs. ROHM & HAAS
INCINERATOR3
Compound
Propane
Propyl ene
Ethane
Ethyl ene
TOTAL
Rohm & Haas
Inlet
(Ibs/hr)
900
1800b
10
30
2740
Incinerator
Outlet
(Ibs/hr)
150
150b
375
190
865
Union Carbide
Inlet
(Ibs/hr)
71.4
142.9
0.8
2.4
217.5
Lab Incinerator
Outlet
(Ibs/hr)
0.64
5.6
3.9
3.4
13.54
% VOC Destruction: 68.4% 93.8%
Table shows the destruction efficiency of the four listed compounds for the
Rohm & Haas (R&H) field and Union Carbide (UC) lab incinerators. The R&H
results are measured; the UC results are calculated. Both sets of results
are based on 1425 F combustion temperature and one second residence time.
In addition, the UC results are based on complete backmixing and a four-step
combustion sequence consisting of propane to propylene to ethane to ethylene
to COo and H20. These last two items are worst case assumptions.
Are not actual values. Actual values are confidential. Calculations with
actual values give similar results.
A-25
-------
The EPA, Union Carbide, and Rohm and Haas tests were for residence
times greater than 0.75 second. However, theoretical calculations show
that greater efficiency would be achieved at 1600°F and 0.75 second than
at the longer residence times but lower temperatures represented in
these two tests. The data on which the standard is based is test data
for similar control systems: thermal incineration at various residence
times and temperatures. If 98 percent VOC reduction can be achieved at
a lower temperature, then according to kinetic theory it can certainly
be achieved at 1600°F, other conditions being equal.
Four tests at temperatures less than 1600°F are relied upon to
support the 98 percent reduction requirement.
A-26
-------
A-4. REFERENCES FOR APPENDIX A
1. Mascone, D.C., EPA, Draft memorandum concerning incinerator efficiency,
April 25, 1980.
2. Letter from Towe, R., Petro-^Tex Chemical Corporation, to Farmer, J., EPA,
August 15, 1979.
3. Broz, L.D. and Pruessner, R.D., "Hydrocarbon Emission Reduction Systems
Utilized By Petro-Tex", paper presented at 83rd National Meeting of AIChE,
9th Petrochemical and Refining Exposition, Houston, Texas, March 1977.
4. Letter from Lawrence, A., Koppers Company, Inc., to Goodwin, D., EPA,
January 17, 1979.
5. Letter from Weishaar, M., Monsanto Chemical Intermediates Co., to Farmer, J.,
EPA, November 8, 1979.
6. Maxwell, W., and Scheil, G., "Stationary Source Testing Of A Maleic
Anhydride Plant At The Denka Chemical Corporation, Houston, Texas," EPA
Contract No. 68-02-2814, March, 1978.
7. Blackburn, J., Emission Control Options For The Synthetic Organic Chemicals
Manufacturing Industry, Trip Report, EPA Contract No. 68-02-2577,
November 1977.
8. Scheil, G., Emission Control Options for The Synthetic Organic Chemicals
Manufacturing Idustry, Trip report, EPA Contract No. 68-02-2577,
November 1977.
9. Lee, K., Hansen, J., and Macauley, D.," Thermal Oxidation Kinetics Of
Selected Organic Compounds," paper presented at the 71st Annual Meeting of
the APCA, Houston, Texas, June 1978.
A-27
-------
APPENDIX B: COST ANALYSIS
B.I DEVELOPMENT OF MODIFIED INCINERATOR COST CURVE FOR VOC WITH
COMBUSTION CHAMBER TEMPERATURE OF 1600°F AND INCINERATION
RESIDENCE TIME OF 0.75 SECOND
B.I.I Method and Sample Calculation
Step I: Adjust Residence Time from 0.5 Second as in Reference 1
to 0.75 Second
The relationships between residence time (RT) and cost were given
in Reference 1 as follows:
Reduction of RT from 0.5 to 0.2 second - cost reduced by 25 percent
Increasing of RT from 0.5 to 1 second - cost increase by 25 percent
From this relationship a curve of cost vs. residence time can be
shown (Figure B-l) i.e., Points 1, 2, and 3 can be plotted (using reference
cost of $50,000). From this figure at 0.75 second residence time the
cost increasing is equivalent to:
cTJ- - 15% (as an approximation)
Thus, the new curve can be drawn as shown in Figures 5-1 and 5-2.
Step II: Adjust Operating Temperature from 1500°F to 1600°F
1. At a- flowrate (F-,), a new adjusted flowrate (F^) can be calculated
to accomodate the increased temperature by converting the temperature to
the Rankine scale.
the factor of or -, .051
F2 = 1.051 F]
2. At the new volumetric flowrate (F2) a new cost can be read for
the same volumetric flow (F, ).
3. Steps 1 and 2 can be repeated and a new curve for 160°F can
be drawn as shown in Figures 5-1 and 5-2.
B-l
-------
80
.1
.2
3
.6
.7 0.75 .
.4 .5
RESIDENCE TIME (SEC)
Figure B-l. Thermal incinerator hypothetical cost as a function of residence time.
.9
1 C
-------
CD
I
CO
600
550
500
a:
*s*
350
a 300
a:
« 250
»AJ
£150
0
TrmTnn:rrrrrrTr."t'' ri~ r i ;:; 11 IT mm 1111111II111111 ITrrrrTTTT
NOTE: Estimates include flanges every 40 feet.
*
Range of duct diameters for each plate thickness are based
on typical industrial practice. Extrapolating costs to
larger diameters may result in ductwork that is not self-
supporting for thin-wall plate thicknesses.
CURVE EQUATIONS
THICKNESS
EQUATION
$=-3.86+3.010
=-2.93+2.380 :
=-2.05+1.770
=-1.76+1.370
=-1.46+1.090
; I
1/2" Thick
HPlate
1/8" Thick Plate
I
t
I :
I !
3/8" Thick
Plate
•••-' 1/4" Thick Plate
3/16" Thick Plate..
I
Source: Costs developed from
data from Fuller Co.
11111II H 111i'Ii " 11:ii j I i j I j I i;
10 20 30 40 50
70 80 90 100 110 120 130 140 150 160 170 180 190 200
0, DUCT DIAMETER, INCHES
Figure B-2. Carbon steel straight duct fabrication price per linear foot vs. duct diameter
and plate thickness (Chapter 5, Reference 1).
-------
B.2 ESTIMATION OF ELECTRICITY COST FOR INCINERATOR FAN"
Sample Calculation:
Eleccost = 1.7 * Elecprice* PDROJ> * Outflow * FTurat'io
= 0.5 (F)
where:
PDROP = Pressure drop = 6" WG
Elecprice = Cost of electricity in $/KWHR = 0.049'
Outflow * FTur'atio = Flu6 gas flowrate scfm
Eleccost = $
(F) = Flue gas flowrate scfm
A. Polystyrene (Continuous) Model. Plant:
Continuous Stream:
$ = 1.7 x .049 x 6 x F
= 0,5 x F
= 0.5 x 420 * $210 mm - -~~ = 4,286'
B. HOPE Model Plant:
Continuous Stream:
$ = 1.7 x 0,049 x 6 x F
= 0.5 x 3,476 = $1,738 KWH'R = i = 35,470
C. Polypropylene Model Plant:
$ = 1.7 x .049 x 6 x F
= 0.5 x 1,740 = 870 KWHR = - = 17,755
B.3 FUEL CONSUMPTION CALCULATIONS
B.3.1 Thermal Incinerator
1600°F Incineration
Fuel: #2 distillate oil: 133,375 - (NET)
B-4
-------
Assume it is C15H26: MW (206); MOR (21.5)
MOR = Molar Oxygen Ratio
Design: >3% 02 in flue gas
0% or 35% heat recovery
10% heat loss from combustion system
In range of offgas heating value
0 -^r < HV <300 1^-, minimum fuel required
Stream:
Flow scfm
269.3
10.89
166.0
is 10% of total heating value of offgas (for flame stability)
Temperature = 70°F
Standard Temperature = 32°F
3 combined continuous streams
Composition by Wt %
0.3 Isobutane
99.7 Nitrogen
.0.6 Isobutane
99.4 Nitrogen
61.0 Ethylene
18.0 Isobutane
20.0 Ethane
1.0 Hydrogen
446.2
Composition by Volume %
0.15
99.85
0.29
99.71
59.6
8.5
18.24
13.66
C4H10
C4H10
C2H4
C4H10
C2H6
Component
Isobutane
Ethyl ene
Ethane
Hydrogen
Nitrogen
Btu/lb
Heating
Value
19,468
20,276
20,416
51,571.4
--
M.W.
58
28
30
2
28
Vol. %
3.26
22.17
6.79
5.08
62.7
scfm
14.55
98.94
30.28
22.67
279.75
Btu/scf
Heating
Value
3,145.2
1,581.4
1,760.1
287.3
100.0
446.19
B-5
-------
Heating value of the stream = 102.5 + 350.6 + 115.8 + 14'.6
= 583.5 Btu/scf
Since H.V. >300 Btu/scf; no fuel at all is required (not even for flame
stability). Also, no heat recovery is required. Determination of
combustion air (C) is needed:
C Ciso + CC2H6 + CC2H4 + CEX + CH2
where:
C. = Amount of air required for combustion of C..W-.
Cr u = Amount of air required for combustion of C0H, .
C2H6 ^ b
Cr u = Amount of air required for combustion of C~H/,
^24
Cu = Amount of air required for combustion of H0.
rip f.
C = Excess combustion air.
For compound CH, MOR = X + y/4
MOR = Molar Oxygen Ratio
Compound
Isobutane
Ethyl ene
Ethane
H
MOR
6.5
3.0
3.5
0.5
Ciso = 446<2 x (°-0326) x (6'5) x (
= 450 scfm
Cr u = 446.2 x (0.2217) x (3.0) x (4.76)
C2H4
= 1.412.6 scfm
Cr u = 446.2 x (0.0679) x (3.5) x (4.76)
C2H6
= 504.7 scfm
B-6
-------
Cu = 446.2 x (0.0508) x (0.5) x (4.76)
H2
=53.9 scfm
C£x = 4.76 x .03 x F = 0.143 F
where:
F = flue gas flowrate
C = 450 + 1,412.6 + 504.7 + 53.9 + 0.143 F
C = 2,421.2 + 0.143 F
Equations:
C4H1C
•C2H4
C2H6
H2
Offgas
(446.2 scfm)
, + 6.5 0
+ 3.0 0
+ 3.5 0
i
, + 0.5 0
N2 = 279.75
Isobutane =
Ethyl ene =
Ethane =
Hydrogen =
Combustion
Air = 2,421 +
N2 = 1,906 + 0.
02 = 515.2 + 0.
14.55
98.94
30.28
22.67
.143F
113F
03 F
2 >- 4 C02 + 5 H20
2 *• 2 C02 + 2 H20
2 > 2 C02 + 3 H20
2 »• H20
00 Used Flue Gas'
(F scfm)
N2 C02 H20
279.75
94.58 - 58.2 72.75
296.82 - 197.88 197.88
105.98 - 60.56 90.84
11.34 - - 22.67
1,906 + .113F -
F = 2,185.75 + 0.113F + 316.64 + 384.14 + .03F
= 2,886.53 + 0.143F
0.857F = 2,886.53
F = 3,368.2 scfm
B-7
6.5 + 0.03F
-------
C = 2,421.2 + 0.143 (3,368.2)
= 2,903 scfm
F/w * Flue Gas Flow _ 3,.3'68.2 _ 7 cc
7 Off gas Flow 446.2. /*00
F/u + C = Outlet flow = 1
h/W L Inlet Flow 1<
Flue gas composition:
Component
Mole
scfm
N2 76.18 2,566
C0'2 9.41 317
H20 11.4 384
0, 3 ,00 TOT .2
Heat capacity
Component
Nitrogen
Oxygen
Water
Carbon Dioxide
Heat capacity
t»
100
constants (°K)
a b
6.529 0.
6.732 0.
6.97 0.
6.393 1.
equation: C = a + b(T)
3,368.2
x 102 ex 105
1488 -0.02271
1505 -0.01791
3464 -0.0483
01 -0.3405
+ c(T2)
Energy required =
4.7491 x
Flue gas heat
6.
in ptn 9 mole /F\
IU DLU f-.f r-.} \' )
capacity coefficients:
a b x 102
5719 0.2524
(aT + l/2bT2 + l/3cT3)
c.
c x 10°
-0.05538
1144°K
273°K
B-8
-------
n
e.g., a = £ a. X,
1=1 1 ^
where:
a. = coefficient of i— component.
X. = mole fraction of i— component.
n = number of components in flue gas.
Let:
I = aT + 1/2 bT2 + 1/3 cT3
114°K
273°K
= 6.5179 M144 - 273 ] + 1/2 (.2524 x 10"2) [(1144)2 - (273)2J
-1/3 (0.05538 x 10"5) [(1144)3 - (273)3]
= 7,009
Energy required = 4.7491 x 10"3 x (3,368.2) x (7,009)
= 1.1212 x 105 Btu/min
Energy available = 0.9 (Heating value of offgas) (offgas flow)
= 2.3432 x 105 Btu/min
Energy available » Energy required
Assumptions of no fuel required and of no heat recovery are valid.
NOTE: In cases where fuel is required for flame stability, the calculations
would be as follows:
Ratio of heat supplied by offgas stream to heat supplied by fuel is
10:1,
If stream heating value = X Btu/scf,
then fuel required = (0.1X) (W) (60) (8,000) Btu/yr
where:
W = offgas flowrate in scfm
Fuel required in gal/yr = (0.1X) W j^ffl Btu/yr
B-9
-------
APPENDIX C: EMISSION FACTORS
C.I INTRODUCTION
The following emission factors and sample calculations are included
to form a basis for the verification of VOC emissions inventories developed
from emission source tests, plant site visits, permit applications, etc.
These factors and procedures should not be applied in cases where site-
specific data are available, but rather in instances where specific
plant information is lacking or highly suspect.
C.2 VOC EMISSION FACTORS FOR EXISTING MODEL PLANTS (UNCONTROLLED)
Continuous Streams
Polypropylene Model Plant (Liquid-Phase Process) kg VOC/1000 kg Resins
A.
B.
C.
D.
E.
F.
G.
Polymerization Reactor Vent Stream
Decanter Vent Stream
Slurry Filter Vent Stream
By-Product & Diluent Recovery Vent Stream
Neutral izer Vent Stream
Condenser Vent Stream
Catalyst Preparation Vent Stream
TOTAL
4.07
11.49
7.93
9.3
1.82
0.058
0.073
34.74
High-Density Polyethylene Model
Plant (Liquid-Phase Process) kg VOC/1000 kg Resins
A. Product Dryer Vent Stream 0.062
B. Product Mixer Vent Stream 0.0053
C. Ethylene Recycle Treater Vent Stream 12.3
TOTAL 12.37
C-l
-------
Polystyrene Model Plant
(Continuous Process) kg VOC/1000 kg Resins
A. Styrene Condenser Vent Stream 2.96
B. Styrene Recovery Unit Condenser Vent Stream 0.13
TOTAL 3.09
C.3 VOC EMISSION FACTORS FOR MODEL PLANTS
Continuous Stream after applied RACT:
Polypropylene Model Plant: 0.69 kg VOC/1000 kg resins
High-Density Polyethylene Model Plant: 0.25 kg VOC/1000 kg resins
Continuous Polystyrene Model Plant: 0.06 kg VOC/1000 kg resins
C.4 VOC EMISSION FACTORS AS APPLIED TO MODEL PLANTS PROJECTED
FOR UNCONTROLLED VOC EMISSION RATE
A. Sample Calculation, Polypropylene Model Plant
Existing uncontrolled VOC emission rate:
(Combined Emission Factor) x (Production Rate) = (VOC Emission Rate)
(34 74 kc: v?c , ) v (141 v in6 kg polypropylene^ _ . Rqfi ,.6 kg VOC
^'/4 1000 kg polypropylene' x U41 x IU year ' " 4'896 x 10 year
= 4896 Mg VOC/year
B. Sample Calculation, High-Density Polyethylene Model Plant
Existing uncontrolled VOC emission rate:
(Combined Emission Factor) x (Production Rate) = (VOC Emission Rate)
) x (214 x 105 kg HDPE) = 2 650 x 106 kg VOC
' x u'* x IU ' ^ • "7° * lo5
= 227.0 Mg VOC/yr
C-2
-------
C.5 VOC EMISSION FACTORS PROJECTED AFTER APPLICATION OF RACT TO
MODEL PLANTS
A. Sample Calculation, Polypropylene Model Plant
*
* ("1 * 'O ) - 9.796 X 10
= 98 Mg VOC/yr
B. Sample Calculation, High-Density Polyethylene Model Plant
(0 02) (12 37 kg VOC \ v (214 x 106 kg HDPE) = 5 294 x 104 kg VOC
(U.ut) u^-J/ 1000 kg HDPE; x ^14 x lu year ; s.^ x lu year
= 53 Mg VOC/yr
C. Sample Calculation, Continuous Polystyrene Model Plant
x 106 ) - 4.542 X 103
= 4.54 Mg VOC/yr
C.6 MODEL PLANT'S VOC REDUCTION EFFICIENCY
Total Annual Plant VOC Emission Reduction
Total Annual Emissions Total Annual Emissions Total Annual Emission
from Existing Process - From RACT Equipment = Reduction
Stream
A. Polypropylene Model Plant
4896 Mg VOC/yr - 98 Mg VOC/yr = 4798 Mg VOC/yr
B. High-Density Polyethylene Model Plant
2650 Mg VOC/yr - 53 Mg VOC/yr = 2597 Mg VOC/yr
C. Continuous Polystyrene Model Plant
227 Mg VOC/yr - 4.54 Mg VOC/yr = 222 Mg VOC/yr
C-3
------- |