EPA/625/R-93/008
                                   October 1993
   Seminar Publication

Operational Parameters for
      Hazardous Waste
    Combustion Devices
   Center for Environmental Research Information
      Office of Research and Development
      U.S. [Environmental Protection Agency
          Cincinnati, Ohio 45268
                              Printed on Recycled Paper

-------
                                      Notice
The information in this document has been funded wholly or in part by the U.S. Environmental
Protection Agency (EPA). It has been subjected to the Agency's peer and administrative review and
approved for publication as an EPA document. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.

-------
                                 Acknowledgments
This seminar publication was developed for the U.S. Environmental Protection Agency (EPA) Cen-
ter for Environmental Research Information (CERI) in Cincinnati, Ohio. The information in the docu-
ment is based on presentations at the EPA-sponsored seminar series on Operational Parameters
for Hazardous Waste Combustion Devices. This series consisted of five seminars, held August 4-5,
1992, in Atlanta, Georgia; August 24-25,1992, in San Francisco, California; August 27-28,1992, in
Dallas, Texas; September 14-15,1992, in Philadelphia, Pennsylvania; and September 17-18,1992,
in Chicago, Illinois.

Justice Manning of CERI served as Project Director and provided technical oversight and review
for the seminar series and publication. Mara Evans  of Eastern Research Group, inc. (ERG) in
Lexington, Massachusetts, provided planning support and coordination for the seminars. Susan
Richmond, Joe Zhou, and John Bergin of ERG provided writing, editing, and production support for
the proceedings. Individual chapters of the document were developed and reviewed by the follow-
ing seminar speakers:

Chapter 1:     Overview of EPA Regulations for Hazardous Waste Combustion Devices, Sonya
              Sasseville, U.S. EPA, Office of Solid Waste and Emergency Response, Washing-
              ton, DC

Chapter 2:     Introduction to Implementation Issues, Beth Antley, U.S.  EPA Region 4, Atlanta,
              Georgia

Chapter 3:     Combustion Technologies and Operational Parameters Related to Paniculate Matter
              and Metals Emissions, Wyman Clark, Energy and Environmental Research Cor-
              poration, Irvine, California

Chapter 4:     Control Parameters and Permit Conditions, Leo Weitzman, LVW Associates,
              Inc., West Lafayette, Indiana

Chapter 5:     Toxic Metals and Paniculate Matter, Wyman Clark, Energy and Environmental
              Research  Corporation, Irvine, California

In addition to their contributions to  specific chapters of the proceedings, Ms. Sasseville and Ms.
Antley also served as peer reviewers for the entire document, as did Robert Thurnau of the U.S.
EPA Risk Reduction Engineering Laboratory in Cincinnati, Ohio.

Each seminar featured either a presentation or panel discussion on particular regional experiences
and problems associated with hazardous waste combustion devices. The issues raised by these
participants have been incorporated into the seminar publication, as appropriate. Among those
giving presentations were Ms. Antley, who presented at the Atlanta seminar; John McCarroll, U.S.
EPA Region 9, who presented at the San Francisco seminar; and William Honker, U.S. EPA Region
6, who presented at the Dallas seminar. At the Philadelphia seminar, Gary Gross, U.S. EPA Region
3, coordinated a panel discussion, and Stephen Yee, U.S. EPA Region 1, John Brogard,  U.S. EPA
Region 2, and Kate Anderson, U.S. EPA, Office of Solid Waste and Emergency Response, RCRA
Enforcement Division, participated on the panel. Gary Victorine, U.S. EPA Region 5, led a panel
discussion at the Chicago seminar. Ms. Sasseville served on each of the panels and as a resource
person for each of the seminars. Also, at the Chicago seminar, David Ullrich, Acting Deputy Re-
gional Administrator of U.S. EPA Region 5, welcomed participants and made introductory remarks.
Jack Divita, Deputy Director, Hazardous Waste Management Division, U.S. EPA Region 6, deliv-
ered a welcome and introduction at the Dallas seminar.

-------

-------
                                             Contents
           Glossary	,	ix

Chapter 1  Overview of EPA Regulations for Hazardous Waste Combustion Devices	1
       1.1  Incinerator Regulations	1
       1.2  BIF Regulations	3
       1.3  Permitting Process	8
       1.4  References	8

Chapter 2  Introduction to Implementation Issues	9
       2.1  Trial Burn Planning and Establishing Final Permit Conditions	9
       2.2  Metals Emissions: An Additional Level of Complexity	,	 10
       2.3  Monitoring and Recording	...10
       2.4  Summary	11

Chapter 3  Combustion Technologies and Operational Parameters Related to Particulate
           Matter and Metals Emissions	13
       3.1  Rotary Kilns	13
       3.2  Liquid Injectors	13
       3.3  Controlled (Starved) Air Incinerators	14
       3.4  Fluidized Bed Incinerators	14
       3.5  Boilers	15
       3.6  Cement Kilns	16
       3.7  References	17

Chapter 4  Control Parameters and Permit Conditions	..19
       4.1  Permitting Approaches	19
       4.2  Operational Parameters in General	20
       4.3  Minimization of Organic Emissions	20
       4.4  Particulate and Metals Emissions	28
       4.5  Hydrogen Chloride/Chlorine Gas Emissions	33
       4.6  Fugitive Emissions and Upsets	34
       4.7  Hourly Rolling Average Limits vs. Instantaneous Limits	36
       4.8  Indicators of Combustion Gas Velocity	37
       4.9  Control Parameters Not Subject to Regulation	37
       4.10 Air Pollution Control Equipment	38
       4.11 Trial Burn/Compliance Test Designs	45
       4.12 Monitoring Systems	51

-------
                                      Contents (continued)
                                                                                             Page

      4.13 Recordkeeping	56
      4.14 References	*	56
      4.15 Bibliography	57

Chapter 5  Toxic Metals and Particulate Matter	61
      5.1  Partitioning of Metals	61
      5.2  Trial Burns	65
      5.3  Setting Permit Conditions from Test Data	73
      5.4  Uncertainties/Research Topics	75
      5.5  References	76

Appendix A  Draft Strategy for Combustion of Hazardous Waste in Incinerators and
             Boilers and Furnaces	77
       I.  Introduction	,	77
       II.  EPA's Strategic Goals	77
      III.  The Process for Pursuing a National Strategy	78
      IV.  Actions To Implement Strategic Goals	79
       V.  Conclusion	82

Appendix B  Federal Register Notices to EPA's BIF Rule	83
                                                vi

-------
                                            List of Figures
Figure
Page
3-1    Rotary kiln	13
3-2    Liquid injector	 14
3-3    Controlled air incinerator	14
3-4    Fluidized bed incinerator	.........:	15
3-5    Boiler	....v......	15
3-6    Schematic of cement kiln, showing clinker cooler, preheater, dust collector, and stack	16
3-7    Recirculation and partitioning of lead to various exit streams in a cement kiln	17
3-8    Transient behavior in cement kilns of lead concentration in waste dust
       due to recycled particulate matter	...„„	17
4-1    Mechanics of POHC destruction in a single chamber incinerator	.......L.	23
4-2    Single source dispersion in three dimensions	32
4-3    Illustration of a wind rose	.-	.........;	33
4-4    Generalized APCE schematic for a combustipn device	L......	39
4-5    Schematic of a packed column absorber	40
4-6    Schematic of a Venturi scrubber	41
4-7    Schematic of a fabric filter	42
4-8    Cross section of a fabric filter	,	42
4-9    Schematic of the dry scrubbing process	::.:..'..:.'.~	:	..................:	 43
4-10   Schematic of an ESP	44
4-11   Factors that affect ESP collection efficiency	44
4-12   Schematic of a fan-driven condensing scrubber (tandem nozzle)	45
4-13   Schematic of an ejector-driven condensing scrubber	45
4-14   Relationship between fan power and gas flow rate for two types of fan blades	53
5-1    Partitioning of metals to exit streams	61
5-2    Principles of metals partitioning	62
5-3    Typical collection efficiency of three types of APCE as a function of particle size	63
5-4    Guide for selecting APCE, based on volatility and particle size of metals	64
5-5    BIF rule approach to setting metals feed rate limits	66
5-6    Assumptions and parameters regarding the four metals tiers	68
5-7    Hourly rolling average (HRA) of combustion chamber temperature in three trial runs	73
5-8    Upward extrapolation of metals emissions rate	74
5-9    Downward extrapolation of metals emissions rate	75
                                                    VII

-------
                                          List of Tables
Table                                                                                       Page

4-1   Advantages and Disadvantages of Permitting Strategies	20
4-2   Control Parameters for Incinerators	21
4-3   Operating Parameters for BIFs for Which Limits Are Established During Precompliance,
      Compliance, and Permit Periods	22
4-4   Purposes of the Control Parameters	23
4-5   Recommendations for Minimum Temperature Waste Feed Cutoff	26
4-6   Risk-Specific Doses (RSDs) for Carcinogenic Metals and Reference Air Concentrations (RACs) for
      Noncarcinogenic Metals, HCI, and Cl,	29
4-7   Definition of Tiers	30
4-8   Recommended Limits on System Pressure	35
4-9   Methods for Setting Minimum and Maximum Combustion Chamber Temperature Limits	37
4-10  Types and Purposes of APCDs	39
4-11   Key Elements of a Trial Burn Plan	46
4-12  Example of Target Permit Limits on Operating Parameters	48
4-13  Sample Material-and-Energy (M&E) Balance Calculation Output	49
                                               viii

-------
                                             Glossary
AAS
acfm
analyte

APCD
APCE
AWFCO
BIFs
Btu

carcinogen
CE
CEM
cfm
CFR

cfs
cms
CVAAS
DE
Atomic absorption spectroscopy.
Actual cubic feet per minute.
The element, compound, or ion the presence or amount of which is being determined as part of a
chemical analysis;
Air pollution control device.
Air pollution control equipment.
Automatic waste feed cutoff.
Boilers and industrial furnaces.
British thermal unit. The amount of energy required to raise the temperature of a 1-pound mass
of water 1 degree Fahrenheit.
A material likely to cause higher incidence of cancer in the exposed population.
Combustion efficiency, a measure of the efficiency of fuel utilization, calculated by:
               CE = -
                      CCL
                   CO2+CO
              x 100 (percent)
DF
Continuous emission monitor.
Cubic feet per minute.
Code of Federal Regulations. Available from the U.S. Government Superintendent of Docu-
ments, or may be viewed in most public libraries, especially Government Repository Libraries.
The environmental volumes (Title 40) are updated each July.
Cubic feet per second.
Cubic meters per second.
Cold vapor atomic absorption spectroscopy.
Destruction efficiency. A measure of the percentage of a given component which is destroyed by
the combustion process. This term is often confused with the ORE (see below) but it is very
different. The DE represents the fraction of the organics entering a combustor that is actually
destroyed. The ORE represents the fraction of the organics entering a combustor that is emitted.
The following equation defines the DE:
     W -W
DE = —*	""outcombustion chamber x -| QQ (percent)
             Win
where W is the weight or mass of the POHC being measured (see also  ORE).
Dispersion factor.
                                                 IX

-------
 ORE
Destruction and removal efficiency of the combustor, defined in 40 CFR 264.343(a)(1). This
value does not include the POHC remaining in the ash and captured by the APCE as part of the
       term. The following equation defines the ORE:
dscf

dscfm
dscm
EAor%EA
ECD
EP Tox.

equivalence
ratio
ESP
feed rate
FID
forced
draft

fuel

9
GC

GC/ECD

GC/FID

GC/MS

H

AHf
HLorX
DRE=W|n~Wout5tack x100 (percent)

where W is the weight or mass of the POHC being measured (see also DE).
Dry standard cubic foot. Gas volume corrected to standard conditions (see also scf) and exclud-
ing water vapor.
Dry standard cubic feet per minute (see also scf).
Dry standard cubic meter (see also dscf).
Excess air or percent excess air. The quantity of air above the stoichiometric quantity needed for
combustion. The value is equivalent to that for excess oxygen or percent excess oxygen for
combustion devices using only air as a source of combustion oxygen — no oxygen enrichment
(see also stoichiometric oxygen).
Electron capture detector (see also GC/ECD).
Extraction procedure, toxicity. A test to determine whether a waste may be classified as hazard-
ous as per 40 CFR 261 .24 and part 261 , Appendix II. This test has been superseded by the
TCLP (see also TCLP).
Equivalent to the excess air ratio. An equivalence ratio of 1 is the same as an excess air ratio
 of 1 or 0 percent excess air.
Electrostatic precipitator.
The rate of feed (Ib/hr, kg/hr, ton/hr) of a waste,  or fuel stream, to a combustion device.
Flame ionization detector (see also GC/FID).
A means of supplying air to a combustion chamber or APCD by placing the fan (or other air
mover) upstream (in front) of the device so that  the fan forces the air through the device. A forced
draft system operates at a pressure above atmospheric (see also ID).
Any combustible material fed to a combustor. The term can refer to supplemental fuel (oil, natural
gas, LP-gas, or a nonhazardous  waste) or to a combustible hazardous waste stream.
Gram or grams.
Gas chromatograph. A laboratory device used to analyze samples for organic constituents. It
must be equipped with a detector such as an FID, ECD, or MS.
GC equipped with an  electron capture detector.  The ECD commonly is used to quantitate
 halogen-organic and sulfur-bearing compounds.
GC equipped with a flame ionization detector. The FID commonly is used to quantitate
nonhalogenated organic compounds.
Gas chromatograph equipped with a mass spectrometer detector. The MS is used to identify as
well as quantitate the organic constituents present.
Virtual stack height.
Heat of combustion.
Heat of formation.
Latent heat. The heat or energy that is released by a  phase change such as evaporation, boiling,
or freezing.  For example, the latent heat of evaporation of water is approximately 950 Btu/lb,
which is the energy required to convert 1 pound of liquid water to 1 pound of vapor or steam.

-------
H.Q.
Enthalpy or sensible heat of a stream. The heat contained by a material, which manifests itself
as a temperature. It is defined as:
HCI

HHV

hr
HRA
HSWA

ID

IWS

kV
kVA


kW

L
Ib
L/G
LHV
m
M&E
MEI

min
              H.J.
where the intenral is performed between a reference temperature T° and the temperature at
which the enthalpy is required.
Hydrogen chloride or hydrochloric acid emissions regulated under RCRA 40 CFR section
264.343(b) to 99 percent removal efficiency, 1.8 kg/h (4 Ib/h) maximum emission rate, or a risk-
based level.
Higher heating value. The heat of combustion of a fuel or waste that includes the latent heat of
condensation of the water formed in the process. While this value is the one measured by
calorimetric means, the LHV is more appropriate for combustor determinations (see also LHV).
Hour(s).
Hourly rolling average.
Hazardous and Solid Waste Amendments of 1986. The law that reauthorized RCRA with a
number of changes and expansions.
Induced draft. A means of supplying air to a combustion chamber or APCD by placing the fan (or
other air mover) downstream of (after) the device so that the fan pulls the air through the device.
An induced draft system operates at a pressure below atmospheric (see also forced draft).
Ionizing wet scrubber. An air pollution control device that combines the performance of a scrub-
ber for HCI control and an ESP for particulate control.
Kilovolts,  103 volts.
Kilovolt-amperes (product of voltage and  current) is a measure of the power usage of an electri-
cal device. It is one of the parameters that is used to describe the operating performance of an
ESP or IWS. kVA is dimensionally analogous to kW, although each actually measures somewhat
different parameters.
Kilowatts, 103 watts. A measure of the power input into an electrical device such as a motor. For
DC systems, the power input in kW is equivalent to the kVA. For AC systems, the power input is
equivalent to kVA times the phase angle shift due to inductance in the circuit.
Liquid flow rate.
Pound(s).
Liquid-to-gas ratio. This is a ratio commonly used in the design and operation of wet scrubbers.
Lower heating value. The heat of combustion of a fuel or waste that does not take into account
the latent heat of water. This is usually the more appropriate value to use for most combustor
calculations. LHV is calculated by:
LHV = HHV-AHW
where AHW equals latent heat of vaporization of the water produced by the qombustion process
(see also HHV).
Meter. A measure of length.
Material and energy.
Maximally exposed individual. The hypothetical individual that a particular source of pollutants
exposes to the greatest risk.
Minute of time (60 seconds).
                                                 XI

-------
 MM5          EPA Modified Method 5. A source sampling method that is the basis for SW-846 Method 0010,
               which measures organic compounds that have boiling points greater than 100°C (212°F). Some
               times referred to as "semi-VOST" (see also semi-VOST).

 MS            Mass spectrometer (see also GC/MS).                                        ,r "

 NPDES        National Pollutant Discharge Elimination System. The permitting and regulatory program under
               the Clean Water Act that restricts discharges to waterways.

 Orsat          A type of apparatus used to measure the concentration of carbon dioxide, oxygen, and carbon
               monoxide in a gas. It operates on the principle of sequential absorption of the target gases in a
               solution.

 overfire air     Air fed to a furnace above the flame. This air may be fed at a  high velocity to improve turbulence
               in the combustion chamber. It may be aimed at the flame or simply into the post-flame combus-
               tion zone to increase turbulence and add oxygen (see also underfire air).
 oxidizing       Combustion in the presence of a stoichiometric quantity, or more, of oxygen or another oxidizing
 conditions      agent (see also reducing conditions).
 AP            Change in pressure.

 PCC          Primary combustion chamber. The initial chamber where combustion occurs such as the hearth,
               rotary kiln, or fluidized bed (see also SCO).
 PIC            Product of incomplete combustion. PICs are organic materials formed during the combustion
               process, either as equilibrium products that escaped combustion or as breakdown or recombi-
               nant organic compounds that do hot exist in the original  waste. Under RCRA,  PIC refers to
               RCRA Appendix VIII organic compounds not present in the feed that result from combustion  of
               waste.

 POHC         Principal organic hazardous constituent. The organic constituents that are measured during a
               trial burn. They are selected to be representative  of all of the organic hazardous constituents in
               the waste and typically include those constituents that are more difficult to destroy. POHC
               typically refers to RCRAAppendix VIII organic compounds present in the feed as either a compo-
               nent of the waste or added for the tests that are selected for evaluation of ORE during the trial
               burn.
 POM          Polynuclear (or polycyclic) organic materials.

 ppb, ppbv      Parts per billion, 10-9 (see also ppm and ppmv).

 ppm           Parts per million. A measure of concentration on the basis of mg of analyte per kg of sample.
               Synonyms are mg/kg, ug/g. The term ppm is sometimes used to indicate mg of analyte per liter
               of sample; however, this definition is incorrect unless the sample is reasonably pure water or
               another material with a density of 1 g/mL

 ppmv          Parts per million by volume, 10"6. A measure of concentration on the basis of volume such as
               uL/L or mL/1,000 L. This unit of measurement is normally used to specify concentrations of
               gaseous contaminants in air.
ppt, pptv       Parts per trillion, 10~12 (see also ppm and ppmv).

primary air      Air mixed with the fuel prior to the point of ignition, usually through the air nozzle or as underfire
               air through a burning solid bed (see also secondary air).

Q             Gas flow rate in the stack (for dispersion modeling).

QA/QC        Quality Assurance/Quality Control. QC is the system of activities to provide a quality product  or a
               measurement of satisfactory quality. QA is the system of activities to provide assurance that the
               quality control system is  performing adequately.

RAC           Reference air concentration. The legal air concentrations of noncarcinogens to which an MEI
               may be exposed (under RCRA).  It is determined as a fraction  of the RfD, taking into account the
               prevalence of the regulated material in emissions from other sources in the environment.
                                                 xii

-------
RCRA

RE

reducing
conditions

RfD


risk
risk
assessment
RSD
sampling
train

SCC
scf


scfm


scm


secondary air

Semi-VOST

Slo-VOST


SSU

stoichiometric
oxygen
   .
:

STP
Resource Conservation and Recovery Act of 1 976 and Amendments (see also HSWA).

Removal efficiency.

Combustion in the absence of at least a stoichiometric quantity of oxygen (see also
oxidizing conditions).

Reference dose. For toxic substances not known to display carcinogenic properties, RfD is the
assumed minimum exposure threshold below which adverse health effects do not occur.

The incremental probability of a person incurring cancer from a carcinogen or being adversely
affected by a noncarcinogenic material. The risk to the MEI from exposure to a particular
carcinogen is calculated by multiplying the predicted maximum annual average ground-level
concentration of the  substance by its unit risk.

The scientific activity of evaluating the health and environmental impact of a chemical, device,
or activity. A risk assessment ascertains the likelihood that exposed individuals will be adversely
affected by the chemical or activity and will characterize the nature of the effects they may
experience. It is a multistep process utilizing air dispersion calculations, ground-water modeling,
and other modeling efforts  to calculate probable exposures. It also utilizes health, toxicological,
and environmental data to  estimate the impact of the exposure.                       '

Risk specific dose. The dosage corresponding to a specific level of risk for a carcinogenic
chemical (see also unit risk).

Second(s) of time.

Equipment in a series including filters, absorbers, impingers and adsorbers, and gas moving
and measuring devices used to collect samples of gases from a stack or other ducts.

Secondary combustion chamber. The second chamber in a combustion device. The SCC
normally burns the off-gases from the PCC (see also PCC).

Standard cubic foot.  Gas volume corrected to standard temperature and pressure, usually 20°C,
or 70°F and 1 atmosphere.                                   :
Standard cubic feet per minute. Gas flow rate corrected to standard temperature and pressure
(see also scf).

Standard cubic meter. Gas volume corrected to standard temperature and pressure (see also
Air mixed with the fuel after ignition as in the combustion chamber (see also primary air).

A synonym for MM5 (see also MM5).

A sampling method for hydrocarbons that uses the VOST at lower sampling rates than specified
in the original VOST procedures.

Saybolt standard unit.

The amount of oxygen required to exactly react with a fuel or waste for the combustion reaction.
If the source of oxygen is air, then the term commonly used is stoichiometric air. When
no external source of oxygen is used other than air (no oxygen enrichment), the two values are
equivalent.

Standard temperature and pressure, 70°F (530°R) and 1 atmosphere (29.92" Hg) for English
units, 20°C (293°K) and 760 mmHg, 101.3kPa for metric and SI units, respectively. Other
standard conditions often are used for presentation of data in the literature. While the pressure of
1 atmosphere is virtually universal, temperatures used may be 68°F (20°C)  and 0°C. The reader
is cautioned to check the standard conditions for any thermodynamic data obtained from the
literature. The standard temperatures used in this manual for English and metric units (70°F and
20°C, respectively) are slightly different; these were chosen as the most common conditions
encountered in practice and the slight difference between them is negligible in the context of
incineration.
                                   xiii

-------
 target
TCD
TCLP
IDAS

TFE
THC

theoretical
oxygen or air
toxic or
acutely toxic
TSDF

TSHiO2

TSLoCX
turndown

underflre air
unit risk
urn
v
V
•u
VOST
The compound or category of materials for which samples are analyzed. For example, when
determining ORE, the quantity of POHCs in the wastes and flue gases must be determined. The
POHCs then are the targets for the sampling and analytical procedures.

Thermal conductivity detector. A detector used as part of a GC that relies on changes in the
cooling rate of a heated filament caused by changes in the thermal conductivity of the gas
flowing past it to detect different compounds.

Toxicity characteristic leaching procedure. A test similar to the EP Tox. test used to determine
whether a waste may be classified as hazardous as per 40 CFR 261.24 and part 261, Appendix
II. The TCLP has many other purposes, including its use to determine whether a waste satisfies
the Land Disposal Restrictions requirements. It has superseded the EP Tox. test for all
applications relating to the regulation and legal classification of wastes (see also EP Tox.).
Thermal degradation analytical system.

Tetrafluoropolyethylene. A fluorinated polymer that is highly resistant to chemical attack and has
excellent thermal stability. It is used in applications such as gasketing and lining material for
sampling trains, sampling jars, and equipment where inert materials of construction are neces-
sary. Because of its ability to withstand temperatures on the order of 400°F, it is sometimes used
for gaskets, seals, and filter-bag material (in fabric filters) in high-temperature situations. It is
commonly known by the trade name Teflon.
Total hydrocarbons. Total organic compound releases from a source such as a combustor. This
can  be monitored continuously during operation by a hydrocarbon analyzer.
Synonymous with stoichiometric oxygen or stoichiometric air (see also stoichiometric oxygen).

The  material behaves as a poison, which, in a relatively short period of time, has an immediate
effect on the health or well being of the person exposed to it.
Treatment, storage, and disposal facility. A facility regulated under RCRAthat is used to treat,
store, or dispose of hazardous wastes.

Thermal stability at high or oxygen-rich conditions. A method for estimating how readily a com-
pound will be destroyed in the presence of oxygen.
Thermal stability at low or deficient oxygen conditions. A method for estimating how readily a
compound will be destroyed in the absence of oxygen  compared with other compounds. This
ranking is being evaluated by EPA as a method of selecting POHCs for a trial burn. It is some
times referred to as the University of Dayton Research Institute (or UDRI) incinerability ranking
system.

Fraction of design capacity at which a system is operating. For example, a combustor operating
at 30 MM Btu/hr at 70 percent turndown will be operating at 30 x 0.70 = 21 MM Btu/hr.
Air fed under a bed of burning solids in a combustor (see also overfire air).

The  incremental risk to an individual exposed to ambient air containing one microgram per cubic
meter of a chemical over a 70-year lifetime. Dividing the acceptable level of additional risk to the
MEI  (which is 1 x 1Q-5) under the BIF rules by the unit risk of a substance defines the substance's
RSD (see also RSD).
                                                                                 •,'\ s> •',
Micron, equal to 1 x 10"6 meter.

Velocity or gas velocity, fl/sec, m/sec.

Volume, ft3 or m3.

Specific volume, fP/lb, fP/lb-mole, m3/g, m3/g-mole. This is the inverse of the density of a material.

Volatile organic sampling train. Sampling equipment used to capture volatile organic emissions
from a source. The captured emissions are then analyzed in the laboratory to determine the
quantity and types released from the source (SW-846 Method 0030).
                                                 XIV

-------
                                              Chapter 1
          Overview of EPA Regulations for Hazardous Waste Combustion Devices
 Note: Since the presentation of the seminars, the U.S.
 Environmental Protection Agency (EPA) Administrator
 has issued a draft Waste Combustion Strategy. The
 strategy will affect many of the areas addressed in the
 seminar, including the regulations and permitting pro-
 cedures for incinerators and boilers and industrial fur-
 naces (BIFs), as well as  other areas such as source
 (generator) reduction of hazardous waste. The seminar
 proceedings have not been updated to reflect the Waste
 Combustion Strategy. A complete copy of the strategy
 is included as Appendix A.

 Hazardous waste combustion devices are regulated un-
 der the Resource Conservation and Recovery  Act
 (RCRA) under the following provisions:

 • 40  CFR part 264, subpart Q—Permitted  Incinerator
   Standards.

 • 40  CFR part 265, subpart O—Interim Status Inciner-
   ator Standards.

•~* 40  CFR part 266, subpart H—Permitted and Interim
 • Status BIF Standards.

 * 40  CFR part 270—Permitting Requirements.

 The RCRA regulations require that all hazardous waste
 treatment, storage, and disposal facilities be permitted
 before being constructed. An exception is that existing
 facilities can continue to operate until EPA or "autho-
 rized" states make permit decisions.

 Under section 264.340,1 certain incinerators that burn
 waste defined as hazardous based  only on ignitability,
 corrosivity, or reactivity characteristics are exempted
 from most of the incinerator permit requirements. If the
 waste has either no or insignificant concentrations of
 Appendix VIII constituents, the facility can be exempted
 from all of the permit requirements except for  waste
 analysis and closure. (Appendix VIII of part 261 con-
 sists of a list of hazardous constituents, their chemical
 abstract names and numbers, and  their  hazardous
 waste numbers.)

 Three types of hazardous waste combustion devices
 are regulated under RCRA: incinerators, boilers, and in-
 1 Such references throughout the document refer to sections of the Title
  40, Code of Federal Regulations (40 CFR), parts and subparts.
dustrial furnaces. Different standards apply to incinera-
tors than to BIFs (U.S. EPA, 1992).

Only enclosed devices with a direct flame are consid-
ered incinerators and are subject to subpart O incinera-
tion standards. Thermal treatment devices that are not
enclosed or that operate without a direct flame and that
are not BIFs are regulated under subpart X, which
requires that miscellaneous units undergo an environ-
mental assessment/To be classified an industrial fur-
nace,  a  device must be  specifically listed in the
regulations. Section 260.10 of the regulations defines
the terms incinerator and boiler and provides a list of 12
devices that currently are classified as industrial fur-
naces. In general, for a facility to be classified an indus-
trial furnace,  it must produce a  marketable product
(e.g., a cement kiln must produce marketable cement).

1.1   Incinerator Regulations

1.1.1  Incinerator Performance Standards
Regulations for hazardous waste incinerators apply to
emissions of organics, hydrogen chloride, and particu-
late matter, as well as fugitive emissions.  The perfor-
mance standards for hazardous waste  incinerators
require a 99.99 percent destruction and removal effi-
ciency (ORE) for designated principal organic hazard-
ous constituents (POHCs). Since measuring  the ORE
for all organic constituents in the hazardous waste is
impractical, EPA regulations specify that the ORE must
be demonstrated on a subset of organics, POHCs, that
are considered representative of the other organic con-
stituents an incinerator will  burn. POHCs  are chosen
based on such factors as difficulty of incineration and
prevalence in the waste feed. A 99.9999 percent ORE
applies to dioxin-listed wastes.

Hydrogen chloride and paniculate emissions also are
regulated. The required removal efficiency for hydrogen
chloride is either 99 percent efficiency or a maximum of
4 pounds per hour emitted, whichever is greater. For
particulates, the emissions limit is 0.08 grains per dry
standard cubic foot (gr/dscf) corrected to 7 percent oxy-
gen. This correction is required so that regardless of the
dilution factor (the more dilution the greater the percent-
age of oxygen), the concentrations for different  com-

-------
 bustlon devices under different operating parameters
 can be compared. EPA developed a new formula to cal-
 culate the correction to 7 percent oxygen that accounts
 for oxygen enrichment by allowing substitution of the
 actual percentage of oxygen in the incoming air (56 FR
 (35):7,152; U.S. EPA, 1990).

 Fugitive emissions from the combustion zone also must
 be controlled. The two control methods are (1) maintain-
 ing negative pressure in the combustion zone so that air
 will be pulled into the device rather than allowing pollut-
 ants to escape before they go through air pollution con-
 trol equipment, and (2) totally sealing the combustion
 chamber so that no emissions can escape to the envi-
 ronment A delay time between  an exceedance of the
 maximum combustion chamber pressure limit and auto-
 matic cutoff of the waste feed generally is not accept-
 able. Any delay between a pressure exceedance and
 an automatic waste feed cutoff potentially would result
 in fugitive emissions.

 1.1.2  Incinerator Amendments

 Requirements for metals and products of  incomplete
 combustion  (PICs) were proposed in April 1990 in the
 amendments to the incinerator regulations. The emis-
 sion limits for metals are site specific and risk based,
 while the PICs regulations limit the carbon monoxide or
 hydrocarbons in stack gas. A site-specific, risk-based
 check on hydrogen chloride emissions similar in format
 to the metals requirements also was proposed. These
 incinerator amendments have been on hold. (New incin-
 eration rule amendments are planned as part of the
 Waste Combustion Strategy.) Nevertheless, incinerator
 permit writers have been implementing them since mid-
 1988 under the authority of the Omnibus provision in
 section 3005(c)(3) of RCRA. The Omnibus provision al-
 lows the permitting authority to impose permit condi-
 tions as  necessary to protect human health and the
 environment. Both site-specific risk-based metals emis-
 sion limits and PIC emission limits  have been set in in-
 cinerator permits under the authority of the Omnibus
 provision. EPA personnel initially developed guidance
 on both metals emissions and PICs, but this guidance is
 out of date and is being revised. In the meantime, the
 applicable portions of the BIF rule represent EPA's up-
 to-date approach for metals and PICs.

 A very important aspect of the regulations is that compli-
 ance with the operating conditions specified in the per-
 mit is deemed to be compliance with the performance
 standards. This provision exists  because continuously
 monitoring the concentration of emitted pollutants, with
the possible exception of hydrogen chloride, to evaluate
 compliance with the performance standards is not pos-
sible given the current state of technology. The permit,
which is site specific, is based on the results of a trial
bum in which compliance with the performance stan-
dards as well as key operating parameters, such as
temperature, are monitored. The operating conditions
under which the performance standards are met are
specified as permit conditions. The regulations specify
that a facility in compliance with the permit conditions is
deemed to be complying with the performance stan-
dards. If, during the life of the permit, EPA receives in-
formation that indicates operating conditions no longer
represent compliance with the performance standards,
EPA can require a retest or can modify the permit.

1.1.3   Operating Conditions

The regulations require the following operating condi-
tions to be specified in an incinerator permit (section
264.345(b)):

•  Carbon monoxide level in exhaust.

•  Waste feed rate and composition.

•  Combustion temperature.

•  Combustion gas velocity indicator.

•  Other requirements necessary to meet performance
   standards.

These conditions are self explanatory, except for the
requirement for a combustion gas velocity indicator and
the "other requirements." The combustion gas velocity
indicator is  important because  it indicates gas  resi-
dence time in the combustor.

To  determine the other requirements,  two questions
must be addressed:

•  What other operating conditions should be set in the
   permit to ensure long-term compliance with the per-
   formance standard?

•  How can these permit conditions be set from the trial
   burn to account for variability such as differences in
   operating conditions from one run to the next?

Additional requirements  under the incinerator regula-
tions include:

•  Automatic waste feed cutoff (section 264.345(e)).

•  Inspections and monitoring (section 264.347).

•  Removal  of hazardous  waste and residues  upon
   closure (section 264.351).

EPA developed guidance on setting permit conditions
and reporting trial burn results (U.S. EPA, 1989). EPA's
goals in developing the guidance were to  provide a
standard set of incinerator operating conditions that
would ensure maintenance of performance  standards
during incinerator operation; eliminate unnecessary or
redundant parameters that would restrict flexibility of
operation and make monitoring to ensure compliance
cumbersome; and include both the regulatory and tech-

-------
nical basis for each operating condition. The basis for
each operating condition was included for three rea-
sons. First, the guidance was intended to be used as a
training tool. Second, because all incinerators are differ-
ent, the guidance would be difficult to apply in all situa-
tions without  detailed  information.  By providing the
basis for choosing permit conditions and determining
how they are set, EPA allows the permit writer to evalu-
ate the applicability of conditions to the particular incin-
erator being evaluated and, if necessary, adapt these
operating conditions to a specific facility. Third, compli-
ance with the guidance is not required; the document is
only guidance. EPA concluded that permit writers were
more likely to implement the guidance if they under-
stood its basis.

1.2    BIF Regulations

The BIF regulations were published in the Federal Reg-
ister on February 21,1991 (56 FR (35}:7,134), and be-
came  effective on  August 27,  1991 (56  FR  (166):
42,504). Since then, several sets of technical amend-
ments  and corrections to the rule have been issued as
EPA has responded to questions. The most recent set
of technical amendments was published on September
30, 1992, providing clarifications to  the final rule pub-
lished August 27, 1991, and correcting portions of the
August 25,  1992,  amendments. (See Appendix  B for
complete listing of Federal Register notices to the BIF
rule.)

1.2.1   Performance Standards for BIFs
The performance standards for BIFs include the same
ORE standard as that for incinerators (section
266.104(a)); limits for PICs;  site-specific, risk-based
emission limits for metals, hydrogen  chloride, and chlo-
rine (sections 264.106 and 264.107); and a paniculate
matter limit (section 266.105). In addition, fugitive emis-
sions must be controlled (section 266.102(e)(7)(i)), and
testing and risk assessment must be conducted for di-
oxins and furans when emissions potential is high (sec-
tion 266.104(e)).

1.2.1.1 Limits for PICs

With certain exceptions, a facility can choose among
several alternatives for complying with the PICs emis-
sion standards:
• The 100-ppmv limit for carbon  monoxide (section
   266.104(b)).
• The 20-ppmv limit for hydrocarbons and the alterna-
   tive carbon monoxide limit (section 266.104(c)).

• The alternative hydrocarbon limit, which includes a
   baseline  hydrocarbon  and carbon monoxide limit
   (section 266.104(f)).
The alternative hydrocarbon limit applies only to indus-
trial furnaces with organics in their raw materials; these
organics contribute to the emitted hydrocarbons.  This
limit does not apply to cement kilns for which monitoring
is conducted in the bypass duct, because the bypass
duct contains emissions only from combustion hydro-
carbons and not from the raw material hydrocarbons.
The one exception is the industrial furnace that feeds
material other than solely as an ingredient, at a location
other than the "hot end." (The hot end of the kiln is
where the fuels are normally fired and, for a countercur-
rent device like a cement kiln, also where the product is
normally discharged.) This type of facility has to comply
with both the hydrocarbon limit and the 100-ppmv car-
bon monoxide limit (section 266.104(d)).

The carbon monoxide limit is a  1-hour rolling average,
on a dry basis, corrected to 7 percent oxygen. The hy-
drocarbon limit, which also is based on  a 1-hour rolling
average, is measured as propane on a dry basis cor-
rected to 7 percent oxygen. The BIF  methods docu-
ment, which is an appendix to the BIF rule, contains the
monitoring requirements for  carbon monoxide and hy-
drocarbons (U.S. EPA, 1992).

1.2.1.2 Standards for Metals,  Hydrogen Chloride,
        and Chlorine
The limits for metals, hydrogen chloride, and chlorine
are site specific and risk based. These limits are deter-
mined by one of three tiers of analysis, ranging from
Tier I, which uses all default values in the analysis, to
Tier III, which uses all site-specific values. The facility
owner/operators must decide whether to base  their
analyses on a higher tier and, in turn, have more site-
specific and somewhat less stringent limits,  or to use
the Agency's generic modeling calculations and conser-
vative assumptions, which are more easily used but re-
sult in much more restrictive limits.
The first tier is the most conservative and assumes that
all the metals fed into the device are emitted. This tier
does not allow credit for metals that partition to the bot-
tom ash and therefore are not emitted, or for removal by
air pollution control equipment.  The benefit of  using
Tier I is that testing is unnecessary, so implementation
is much easier. Also, an operator can more easily use
EPA's generic dispersion  model and substitute values
for a few parameters (e.g., terrain-adjusted effective
stack height) to calculate a  site-specific emission limit
than model the specific site. Generic dispersion models
are not applicable to all facilities, however. The regula-
tion specifies the types of facilities for which the generic
modeling is not conservative; those facilities have to do
site-specific modeling.
Under Tier II, testing is conducted to produce data to
gain credit for partitioning and air pollution control de-

-------
 vice efficiency. In addition, the Agency's generic disper-
 sion modeling is used. Under Tier III,  the modeling is
 completely site specific. The facility conducts testing
 and does its own site-specific dispersion modeling using
 EPA models.  A fourth tier, called Adjusted  Tier I, is a
 hybrid of Tier I and Tier III that uses site-specific disper-
 sion modeling but not emissions testing.

 1.2.1.3 Paniculate Matter and Fugitive Emissions
        Limits and Testing Dioxins and Furans

 The BIF regulations also specify controls  for fugitive
 emissions and paniculate matter similar to those for in-
 cinerators. They include a new requirement for testing
 and assessing the risk of dioxins and furans at BIFs with
 a high potential for emitting these compounds. The po-
 tential for emitting dioxins and furans is considered to
 be high when either an  industrial furnace is operating
 under an alternative hydrocarbon limit or a unit has a dry
 partieulate matter control device that operates in the
 temperature range of 450 to 750°F. When  conducting
 tests to support the municipal waste combustion regula-
 tions, EPA found that dioxin and furan emissions were
 greater in  that operating temperature range. EPA be-
 lieves these data also could apply to incinerators and
 will probably include this dioxin approach in the updated
 incinerator  guidance. Consequently,  permit writers
 should consider these factors when developing permits
 for hazardous waste incinerators.

 1.2.2  Special Requirements for Industrial
        Furnaces

 Special requirements for industrial furnaces  include re-
 strictions on waste feeding during interim status, an al-
 ternative hydrocarbon limit for furnaces with organic
 matter in raw material, risk assessment, and alternative
 approaches to implement the metals standards.

 1.2.2.1  Waste Feeding Restrictions

 Five waste feeding restrictions apply to  industrial
 furnaces:

 •  For cement kilns under interim status, the  hazardous
   waste cannot be fed at locations not conducive to
   good combustion (e.g., the precalciner or the pre-
   heater). Once a facility is permitted or in the permit-
   ting process, this restriction can be reconsidered. At
   that point, the ORE standard also applies, and more
   opportunity exists for interaction with the permit writ-
   er and for additional technical evaluation and testing.
•  For furnaces that cannot meet the 20-ppmv hydro-
   carbon limit because of organic matter in the raw ma-
   terial,  the owner/operator must comply with  the
   alternative hydrocarbon limit established under sec-
   tion 266.103(c)(7)(ii).
 •  The combustion gas temperature must be at least
    1,800°F where the waste is fed. EPA defines "where
    the waste is fed" as the location where the waste first
    starts releasing hydrocarbons. For example, when a
    drum is fed into a system, it starts releasing hydro-
    carbons at the point where it ruptures, not necessar-
    ily the location where the drum enters the incinerator.

 •  The owner/operator  must document that oxygen is
    adequate to burn the organic compounds and must
    retain that documentation in the operating record.

 •  A facility feeding waste at locations other than the hot
    end must meet the 20-ppmv hydrocarbon level upon
    certification of compliance, regardless of whether the
    unit meets the 100-ppmv carbon monoxide limit. This
    additional requirement was added because organics
    may simply evaporate or volatilize, allowing the facil-
    ity to  meet the carbon monoxide limit because com-
    bustion is not occurring, not because it is occurring in
    an acceptable manner.

 With one exception, the latter three waste feeding re-
 strictions apply only when hazardous waste is fed at a
 location  other than the  hot end. The exception is for
 units in which hazardous waste (1) is not used as a fuel
 (i.e.,  the waste has a heating value of no more than
 5,000 Btu/lb), but is fed solely as an ingredient; and (2)
 contains less than a total of 500 ppm of Appendix VIII
 nonmetals. The low Btu content  indicates the waste is
 not used as a fuel and the low nonmetal content indi-
 cates that the organic content of the waste is low, so
 that organic emissions will not be a problem.

 1.2.2.2 Alternative Hydrocarbon Limit for
        Furnaces with Organic Matter in Raw
        Material

 The alternative hydrocarbon limit  applies'to those units,
 such  as cement kilns, that have organics in  the raw ma-
 terial that generate hydrocarbons and, therefore, might
 not be able to meet the 20-ppmv hydrocarbon  limit.
 These devices must be operated so as to minimize
 hydrocarbon emissions from all sources (i.e., fuels, raw
 materials, slurry water).2 Because the regulation mainly
 is concerned with hydrocarbon emissions resulting from
 hazardous waste combustion, the regulation allows for
 the establishment of an alternative hydrocarbon limit as
 part of the permitting process. This limit ensures that
 when hydrocarbons are burned,  emissions are below
 baseline levels.

 To obtain the alternative hydrocarbon limit, the owner/
 operator must submit a  completed Part B application
 with a request for a time extension. The application has
to include a proposed baseline hydrocarbon level based
on testing,  an interim hydrocarbon limit based on the

2 Some cement kilns did certify compliance with the 20-ppmv hydro-
 carbon limit as of the interim status compliance certification date.

-------
proposed baseline, and a time extension approved by
the Director. (The Director is defined as the EPA Re-
gional Administrator or, if an approved  state program
exists, the State Director.) Employing the alternative hy-
drocarbon limit does not exempt the facility from the
other emissions standards, and the interim hydrocarbon
limit cannot be exceeded during the compliance certifi-
cation test.
During the permitting process, the applicant must dem-
onstrate a baseline hydrocarbon level, a level of emitted
hydrocarbons  lower than the baseline  when  burning
hazardous wastes, and a risk  from  the hydrocarbon
emissions that is within an acceptable range. Cement
kiln owners that monitor in the bypass duct are not eligi-
ble for the alternative hydrocarbon limit.
The technical amendment of August 25, 1992, clarifies
the alternative hydrocarbon limit. When facilities first at-
tempted to develop their alternative hydrocarbon limits
under interim  status, they were concerned that, once
they achieved a baseline level that minimized hydrocar-
bon emissions from all sources, they would not be al-
lowed variability. The technical amendments state that
even under good combustion conditions, EPA expects
variability in operation due to variability in the raw mate-
rials. Further, variability should be considered when
developing the baseline needed to establish the alterna-
tive hydrocarbon limit.

1.2.2.3 Risk Assessment
The regulations specify the procedure  for conducting
the risk assessment. The Director specifies the stack
sampling and  analysis. The facility owner/operator has
to model dispersion to predict the level of exposure to
the  maximally exposed individual (MEI). The owner/
operator also has to show that the sum of the inhalation
risk from carcinogens (including dioxins and furans) is
less than 10'5, and that the reference air concentrations
(RACs) for noncarcinogens (found in 40 CFR 266.109,
Appendix IV) are not exceeded. The appendices to the
BIF rule contain the risk-specific doses  (RSDs) for car-
cinogens (40 CFR 266.109a-c, Appendix V).

1.2.2.4 Alternative Approaches to  Implementing
        Metals Standards

 EPA allows operators of boilers and industrial furnaces
that recycle particulate matter to use  one of three alter-
 native approaches to limiting metals emissions (40 CFR
266.102(c)(3)(ii)). The most practical  option, and there-
Jore the approach most often chosen,  calls for estab-
 lishing equilibrium (i.e., the point at which emissions
 reach 90 percent of their ultimate steady-state value)
 before developing parameters in a trial burn. According
 to this approach,  the trial burn is then a consecutive
 measure of metals emissions. For information on the
best way to achieve equilibrium and a brief explanation
of the other two approaches to limiting  metals emis-
sions, see Chapter 5 (Section 5.2.7).
         /
1.2.3  More Specific Requirements for
       Operating Conditions in the BIF
       Regulations

The operating conditions for BIFs are much more spe-
cific than they are for incinerators (section 266.102(e)—
permits,  and  section 266.103(b)(3)—interim status).
Since the incinerator regulations were promulgated, the
guidance on setting permit conditions has been com-
pleted and EPA has gained valuable experience in im-
plementing the incinerator regulations and has obtained
more specific information on which to base the BIF reg-
ulations.  The following operating conditions are more
specific in  BIF regulations than in incinerator regula-
tions:
• Overall and component (e.g., metals, chlorine) feed
   rate.
• Maximum production rate.
• Minimum and maximum temperature in the combus-
   tion chamber.
• Air pollution control system parameters.
One important difference between the two sets of regu-
lations is that individual metals were not addressed
when the incinerator regulations  were developed, so
those regulations did  not include maximum temperature
requirements. These requirements are included in the
BIF regulations and  are applied to incinerators under
the Omnibus provision. Requirements for air pollution
control system parameters also are included in the BIF
regulations and were not in the incinerator regulations.
Air pollution control system parameters are now set for
incinerators based on the permit conditions guidance.
Under interim status for BIFs, the air pollution  control
system parameters are very specific. Under permitting,
the requirement is  more general and simply requires the
setting of appropriate  parameters. EPA has provided
additional flexibility in permitting, because this process
involves more permit writer review and evaluation.
The regulation requires an automatic waste feed cutoff
for both  incinerators  and BIFs whenever permit condi-
tions are exceeded. The BIF regulatory language states
more clearly than the  incinerator regulatory language
that the operating conditions apply whenever waste is in
the unit. Therefore, even when a cutoff is in effect, the
facility is not exempt from compliance with the operating
conditions after that cutoff occurs, as long as waste is in
the unit. This clarification was made based on the find-
 ings of a task force formed to evaluate the safety, espe-
cially worker safety,  of hazardous wastes incinerators.
The task force found automatic waste feed cutoffs at

-------
  many of the incinerators. EPA is concerned that if a per-
  mit is not written so as to be consistent with the clarifica-
  tion, the automatic cutoff might be interpreted to shield
  the facility from having to comply with permit conditions
  after the cutoff goes into effect and, in addition, that the
  frequent use of automatic cutoff at a facility indicates
  that the incinerator is not operating at a steady state.

  The BIF rule also requires combustion gases to be rout-
  ed through the air pollution control system. Emergency
  vent stacks are allowed for combustion  devices, but
  opening one is considered a violation because the com-
  bustion gases will be emitted directly to the atmosphere.
  (Emergency vent stacks are discussed in Section 4.6.1.)

  On August 21,1992, owner/operators were required to
  certify  compliance with applicable interim status stan-
 dards,  which include all  of the performance standards
 except those for ORE  (i.e., particulates, metals, hydro-
 gen chloride, chlorine, carbon monoxide and/or hydro-
 carbons, and  dioxins/furans, if  applicable  [section
 266.103(c)]). EPA concluded  that the ORE  standards
 were too complicated  to implement under interim sta-
 tus, because the interim status standards are supposed
 to be self implementing. Waste feeding restrictions and
 operating conditions, however, are set based on the
 compliance  certification  test;  they  must be  complied
 with for the duration of interim status operation.

 1.2.4  Extensions of Time

 The two options for extensions are the automatic exten-
 sion and the case-by-case extension. The facility owner/
 operator can claim an automatic extension. This type of
 extension applies for a year and limits the facility to 720
 hours of hazardous waste burning during that time, or
 the length of time needed to conduct compliance test-
 ing, whichever is smaller. The Director can grant a case-
 by-case extension if a facility shows that it  could not
 comply with the standards under the schedule in the BIF
 regulation because of factors beyond its control. An ex-
 ample would be a facility that has too little  time to both
 upgrade air  pollution  control equipment and conduct
 testing. Owners or operators are required to recertify
 every 3 years, or whenever they want to change operat-
 ing conditions.

 1.2.5  Special Provisions for BIFs

 Small-quantity burners, smelters, coke ovens, and pre-
 cious metals recovery furnaces are exempt from the
 BIF rule. In addition, the rule contains the following spe-
 cial provisions:

• Waiver of ORE trial burn for boilers  operating under
   special requirements.
• Low-risk waste exemptions.
 • Direct transfer operations requirements.

 • Exclusion of Bevill residues.3

 1.2.5.1  Waiver of ORE Trial Burn for Boilers

 The ORE exemption for boilers (section 266.110(a)(1))
 is known as the special operating requirement. This ex-
 emption was a result of the testing conducted in support
 of the BIF rule. This testing, conducted on boilers oper-
 ating at non-steady-state  conditions, indicated that if a
 set of operating requirements was met for certain types
 of boilers (i.e., nonstoker  watertube boilers with a sus-
 pension or  fluidized bed), the result was a hot, stable
 flame conducive to good combustion, and therefore
 good destruction. Under these conditions, a destruction
 and removal efficiency test is not necessary.

 To be eligible for a waiver of the ORE trial burn, the boil-
 er must meet the following operating requirements:

 • A minimum of 50 percent of the primary fuel must be
   fossil fuel-based or tall  oil. (Other fuels are possible
   on a case-by-case basis.) The percentage of the pri-
   mary fuel is determined on the basis of either heat
   input (i.e., Btu/hour from the primary fuel versus all
   fuels) or the mass feed rate (i.e., Ib/hour of the prima-
   ry versus all fuels), depending on which method re-
   sults in a lower mass hazardous waste feed rate.

 • The heating value must be a minimum of 8,000 Btu/lb
   (fuels and waste). (Blending is acceptable, but fuel
   used  for blending waste is not included in the 50
   percent.)

 • The boiler load must be greater than or equal to 40
   percent of capacity on a Btu basis.

 • The hazardous waste must be fired directly into the
   primary fuel  flame zone.

 • No dioxin-listed wastes  can be burned.

 • The viscosity of the waste must be less than or equal
   to 300 Saybolt seconds (SSU).

 •  The wastes must be atomized as specified.

 •  The device must comply with the 100-ppmv carbon
   monoxide limit (on an hourly rolling average); the de-
   vice cannot comply with the alternative standards for
   hydrocarbon emissions  instead.

 •  The boiler must be a watertube type that does not
   feed fuel using a stoker  or stoker-type mechanism.

 1.2.5.2 Low-Risk Waste Exemption

The  ORE trial burn standard can be waived for certain
boilers and industrial furnaces if the applicant shows
that the waste poses low risk from organics. In addition,

3 Tom Bevill was a Congressman who listed specific residues that were
 eligible for exemption from the regulations.

-------
the paniculate standard can be waived if the facility
is exempt from the ORE trial burn and  if it complies
with either Tier I or adjusted Tier I  metals limits. EPA
regulates particulate matter emissions from hazardous
waste combustion devices because hazardous metals
or hazardous organics may be adsorbed onto the partic-
ulate. Therefore,  EPA has to be satisfied that neither
organics nor metals are a concern  before it will waive
the particulate standard.

The operating requirements for the low-risk waste ex-
emption are similar to those listed as special operating
requirements for  boilers in  Section 1.2.5.1 above, ex-
cept that fewer requirements apply to the low-risk waste
exemption. EPA wants to ensure that the device is oper-
ating with a hot, stable flame conducive  to good com-
bustion.  Both the special operating requirements and
the low-risk waste exemption are examples of where
standard procedures for testing and setting the operat-
ing conditions based on the test are not required. While
the special operating requirements and the exemption
offer a much simpler procedure because testing is not
required, the facilities must comply with very specific
operating conditions.

To be eligible for the low-risk waste exemption, the facil-
ity owner has to determine all of the part 261 Appendix
VIM nonmetal compounds that are reasonably expected
to be present in the waste, calculate reasonable "worst-
case"  emission  estimates assuming  a  99.9 percent
ORE, and  perform dispersion  modeling  to predict the
maximum annual average ground  level  ambient con-
centrations.  The  results  then  are used  in  the  risk
assessment, which is similar to other EPA risk assess-
ment approaches.  The facility  owner has to demon-
strate  that  the  predicted  ambient concentrations of
noncarcinogens are less than  the  RACs, and that the
ambient levels of carcinogens do not result in a com-
bined risk greater than 1CH5; the risk level for each com-
pound is listed in the appendices to the BIF rule. For all
carcinogens, the sum of the ratios of the predicted lev-
els divided by the RSD (part 266, Appendix V) must be
less than or equal to 1.0.

1.2.5.3 Direct Transfer Operations  Requirements

Direct transfer requirements, although not specifically
related to operating conditions, are a unique feature of
the  BIF rule (section 266.111).  Direct transfer is when
waste is fed directly  from  a transport vehicle (e.g., a
tanker truck) to a combustion unit without use of a stor-
age unit. The requirements apply on  the effective date
and apply equally to interim status facilities and permit-
ted facilities. Direct transfer is considered to be part of
the waste firing system (the combustion device), not in-
volving storage or a separate unit. Thus,  if the combus-
tion device is exempt and not regulated, direct transfer
requirements do  not apply.
The direct transfer requirements are a combination of
those for container storage and tank storage. The areas
where transport vehicles are located are subject to con-
tainment requirements similar to  those  for container
storage areas. Ancillary equipment is subject to second-
ary containment requirements like those for tank sys-
tems. If applicable, the existing facilities must install
required containment within 2  years of the effective
date.
EPA encourages facilities to switch from direct transfer
to storage. Using storage tanks is considered more pro-
tective, and it allows for better blending of waste, which
promotes better combustion. Switching to direct trans-
fer requires approval of a change under interim status or
a permit modification at a  permitted facility, but EPA be-
lieves that these are approvable modifications.

1.2.5.4 Exclusion for Bevill  Residues

The Bevill  exemption (see RCRA section 3001 (b)(3)
(A)(i-iii)) is an amendment to RCRA in which certain res-
idues (i.e., cement kiln dust, waste  from processing
ores and minerals, and waste generated from combus-
tion of coal and other fossil fuels) are not considered to
be hazardous waste, pending EPA study.

For devices processing hazardous waste, the BIF rule
retains the Bevill exclusion (section 266.112) for waste-
derived residues from three types  of devices:
•  A boiler burning more than or  equal to 50 percent
   coal. (If the boiler is burning this amount of oil or nat-
   ural gas, it is not eligible, because oil and natural gas
   do not produce a  lot of ash.  Therefore, a high per-
   centage of any ash formed will be derived from the
   hazardous waste rather than the fuel.)
•  An industrial furnace processing more than or equal
   to 50 percent ores or minerals.
•  A cement kiln  processing more than or equal to 50
   percent normal materials.
Residue excluded under the  Bevill amendment is not
considered  a hazardous waste, and  therefore is not
subject to RCRA hazardous waste management stan-
dards.
The waste-derived residue retains  the exclusion if either
the levels of toxic constituents are  not significantly high-
er than they are in the normal residues (i.e., the normal
residue when hazardous waste is not burned) or the lev-
els do not present a health risk. Most residues are likely
to retain the exclusion. The rule  contains a statistical
procedure for determining if toxics  levels are significant-
ly higher in the residues derived from hazardous waste
combustion. The BIF rule is the first regulation to estab-
lish  required site-specific procedures  to determine
whether the Bevill exclusion  continues to apply when
the device burns  hazardous waste.

-------
 1.3   Permitting Process

 The permitting process for incinerators and BIFs differs
 depending on  whether a unit is a new, as-yet-uncon-
 structed unit or an existing unit.

 1.3.1  New Units

 For owner/operators of a new unit, the first step in the
 permitting process is to submit Parts A and B of the per-
 mit application. Part A is a standard form that describes
 the types of waste management units at the facility and
 the types and  amount of waste the units will be han-
 dling. The much more detailed requirements for Part B
 are described in part 270 of the regulations. After a facil-
 ity has submitted Parts A and B, the permitting authority
 reviews the application and prepares either a draft per-
 mit or a draft denial. The proposed decision then is re-
 leased for public comment. A public hearing will be held
 if requested during the public comment period. Finally,
 the permitting  authority  incorporates  the public com-
 ments, and, if the decision is to issue the permit, issues
 a four-phase permit.
 The facility must comply with a set of  operating condi-
 tions for each of the following four phases of operation:
 • Startup/shakedown—bringing the equipment on line
   and resolving any problems.
 * The trial burn—conducting the trial burn for purposes
   of demonstrating compliance.
 • The post-trial burn period—assembling,  analyzing,
   and reviewing the results of the trial burn.
 • The final operations period—the rest of  the facility
   operation under the permit.
 Although the conditions for the final operations period
 are specified when the permit is issued, if the results of
 the trial burn are different than those expected, the con-
 ditions in that final phase  of operation may be modified.
 A transportable unit is considered new and the trial burn
 is run when the unit arrives on site. If the unit has oper-
 ated somewhere else, however, and data exist on
 wastes similar to those that will be burned at the new
 site, the facility may be able to use the previous data
 instead of conducting a new trial burn. A shakedown
 period would be allowed at a new site to ensure that the
 unit is fully operational.

 1.3.2  Existing Units

The permitting process for an existing  unit is different,
because the unit already has data. For existing  units,
the trial burn is required before the permit  is issued.
Therefore, the first step in the process for an existing
unit is to submit Part B of the permit application. (Part A
already would have been submitted when the facility ini-
 tially obtained interim status.) Either a trial burn plan or
 trial burn data must be included with Part B.

 EPA highly recommends that owner/operators of facili-
 ties submit trial burn plans, rather than performing the
 trial burn on their own and submitting data. Trial burns
 are very complex and expensive,  and operators that
 conduct trial burns on their own are likely to miss an
 important step or do something that might not be ap-
 proved by the permit writer. Consequently, the permit
 writer might require the facility to retest.

 EPA reviews the trial burn plan, analyzes the trial burn
 data, then prepares the draft permit or denial. This step
 is followed by the public comment period, and by final
 issuance or denial of the permit. The permit for an exist-
 ing facility only includes one phase  of operation,  be-
 cause the trial burn already has been run.

 The applicant and the permit writer need to establish
 before the trial burn how the operating conditions in the
 trial burn will be translated into operating conditions in
 the  permit.  When  EPA first  started issuing  permits,
 many facility owner/operators were dissatisfied with the
 operating conditions specified in their permit as a result
 of the trial burn. For example, the operating parameters
 might not have allowed for flexibility of operation. As a
 result of these experiences,  EPA encourages facility
 owner/operators to meet with the permit writers and dis-
 cuss issues such as variability and the operating param-
 eters that need to be set and monitored. In this way, the
 trial burn can be planned so that it satisfies regulatory
 objectives as well as results in a workable set of permit
 operating conditions.

 1.4  References

 When an NTIS number is cited in a reference, that doc-
 ument is available from:

  National Technical Information Service
  5285  Port Royal  Road
  Springfield, VA22161
  703-487-4650

 U.S. EPA. 1992. U.S. Environmental Protection Agency.
  Technical implementation document for EPA's boiler
  and  industrial  furnace regulations. EPA/530/R-92/
  011. NTIS PB92-154947. Office of  Solid Waste and
  Emergency Response, Washington, DC.

 U.S. EPA. 1990. U.S. Environmental Protection Agency.
  Methods manual  for compliance with the  BIF reg-
  ulations. EPA/530/SW-91/010. NTIS  PB89-120006.
  Washington, DC. December.

U.S. EPA. 1989. U.S. Environmental Protection Agency.
  Guidance on setting permit  conditions and reporting
  trial burn results. EPA/625/6-89/019. Office of Re-
  search and Development, Cincinnati, OH.

-------
                                              Chapter 2
                             Introduction to Implementation Issues
Application of Resource Conservation and Recovery
Act (RCRA) regulations to hazardous waste combus-
tion devices can be challenging, both for facility owner/
operators and for regulatory agencies. One advantage
of the performance-based RCRA standards is that the
opportunity for design and operating flexibility is maxi-
mized. On the other hand, the final RCRA permit must
restrict operating flexibility to ensure continued compli-
ance with the performance  standards after  the trial
burn. The balance between operating flexibility and en-
vironmental protection can be optimized if all parties
communicate early in the permit process. Absent this
communication, day-to-day  operating flexibility and
ability to achieve future compliance can  be adversely
affected.
A number of interrelated issues must be considered to
achieve compliance with the hazardous waste combus-
tion regulations. These issues include trial burn plan-
 ning, selection of monitoring devices, development of
waste analysis and recordkeeping protocols, and estab-
 lishment of final operating conditions. The interrelation-
 ship between these various elements is critical,  and
 future compliance can be enhanced if all of these ele-~
 ments are considered in total rather than handled in a
 piecemeal fashion.

 2.1   Trial Burn  Planning and  Establishing
        Final Permit Conditions

 As explained in Chapter 1, continuous emissions moni-
 toring  to demonstrate conpliance with RCRA perfor-
,  mance standards genera/V is not technically feasible.
 Therefore, the regulations state that  compliance with
  the operating  conditions specified in the  permit is
  deemed to be compliance yith the performance stan-
  dards. For a particular sitea trial burn must be per-
  formed  under worst-case operating conditions. The
  operating conditions that result in compliance with  the
  performance standards theJbecome the boundaries for
  continued operation  undeiihe permit. Thus, the  trial
  burn plays a very importance in the final permit condi-
  tions.
   Since final permit condiffons are established based on
   the operating conditio^/nonitored during the trial burn,
a facility owner/operator has an important opportunity to
influence final permit conditions through carefuf plan-
ning. This point cannot be overemphasized. Final per-
mits can be  unduly restrictive if the trial burn is not
designed properly. In the extreme, careless treatment of
conflicting parameters can result in permit conditions
that preclude operations altogether. Several trial  burn
configurations, all of which are for worst-case conditions
(i.e., single point approach, multiple point approach,
and universal approach),  are explored in Chapter 4.
Generally, a  more complex trial burn  is required to
achieve more flexible permit conditions. As long  as a
facility owner understands that permit conditions will be
established from the trial burn and how the numerical
limits will be determined, then the facility owner/opera-
tor can assess whether a  particular trial burn plan will
afford enough flexibility in day-to-day pperations. If the
plan appears too restrictive, then it can be reconfigured
and costly mistakes can be avoided.
To reduce the probability of misunderstandings with re-
spect to setting  permit conditions, the  trial burn plan
should include the following:
• A complete list of anticipated permit conditions.

• An indication for each condition of whether the permit
   limit will be based on the trial burn or some other in-
   formation (i.e., manufacturer specifications).
• An indication for each permit condition to be based
   on the trial burn of howihe numerical limit is to be
   established. Actual numerical values from  the trial
   burn plan  should be used as examples. If multiple
   test conditions are to be performed, an indication
   should be provided regarding which test condition will
   be used to establish the numerical limit and why that
   condition represents the worst case.
• Finally, an explanation should be provided regarding
   how the actual trial burn data will be manipulated to
   derive the permit limit. For example, if the parameter
   is monitored  on an instantaneous basis, the plan
   might state that the numerical permit limit will be de-
   termined  by a mean of the time-weighted average
   during all  test runs.  The plan also might contain a
   statement that the permit limit is a maximum, never-
   to-be-exceeded limit.

-------
  This type of explicit information in the trial burn plan can
  help permit writers and facility owner/operators identify
  misunderstandings regarding how the permit limits will
  be established. For example, a facility owner might not
  understand that the mean temperature during the trial
  burn will become the minimum temperature limit in the
  final permit Early detection of this misunderstanding al-
  lows the facility to revise the trial burn plan and conduct
  the trial bum at a lower temperature than desired for
  normal operation, avoiding the likelihood that conditions
  set for normal operations will be unduly restricted. Other
  common misunderstandings relate to treatment of con-
  flicting parameters and extrapolation of data.

  2.2   Metals Emissions: An Additional
        Level of Complexity

  Risk-based limitations on  emissions of toxic  metals
  have been imposed on incineration facilities since late
  1988, and on boiler and industrial furnace (BIF) facilities
  since the effective date of the BIF regulations in 1991.
  These limitations represent an additional  level of com-
  plexity with respect to trial burn planning, development
  of waste analysis protocols, and establishing final permit
  conditions. Depending on the "tier" chosen for establish-
  ing metal feed rate limits  (see  Sections 1.2.1.2 and
  4.4.1), a final RCRA permit may have from 12 to 38 ad-
  ditional permit conditions related to metals. Designing
  the trial burn to achieve worst-case limits for all of these
  additional parameters can be very challenging.
  As is the case with most other parameters, flexible met-
  als  permit conditions require a more complex trial burn.
 To ensure that the feed rate for each metal during the
 trial burn is as high as it might be in the future, metals
 compounds are often spiked into the trial burn feed. The
 selection of spiking compounds, as well as the physical
 and chemical aspects of the spiking process itself, can
 cause additional complications for the trial burn. Trial
 bum planning also is more challenging because metals
 limitations increase the number of conflicting  parame-
 ters. Minimum combustion temperature represents a
 worst-case condition for organic destruction, while max-
 imum temperature represents a worst-case condition for
 metals. One way to simplify the trial  burn is to establish
 the conservative Tier I or Adjusted Tier I feed rate limits
 on as many metals as possible. This can simplify issues
 related to metals spiking. Unless all metals can be limit-
 ed under the Tier I or Adjusted Tier I  provisions, howev-
 er, the conflicting parameter issue  will still exist and
 likely will require at least two test conditions for the trial
 bum.

 Metals permit limits also represent an additional level of
 complexity with respect to development of waste analy-
sis protocols. Continuous monitoring  of metal feed rates
can  be accomplished only by (1) knowing the as-fed
concentration of each metals in each feed stream, and
  (2) continuously monitoring the flow rate of each feed
  stream. Since the as-fed concentration of each metal in
  each feed stream can be variable, a waste feed plan-
  ning and analytical protocol must ensure that concen-
  trations remain at  or below  predetermined  values
  between analytical determinations. Each facility should
  be able to provide a technical justification for metals
  sampling and analytical frequencies considering waste
  stream variability and proximity to allowable feed rate
  limits. Performing a waste analysis to quantify metals is
  particularly challenging compared to analyzing  waste
  for chlorides and ash. One difference is that metals may
  be regulated at part-per-million levels, while chlorides
  and ash usually are regulated as percentages. Another
  important distinction is that traditional digestion and
  analytical methods for metals have exhibited question-
  able performance  on a number of hazardous waste
  matrices.

  Issues  related to metals limits are  discussed in sub-
  stantially more detail in Chapter 5 of this publication.

  2.3   Monitoring and Recording

  In the past, issues related  to monitoring and recording
  have received less attention than issues related to trial
  burn planning and establishing permit conditions. Moni-
  toring and recordkeeping, however, are the most impor-
  tant elements with respect to proving compliance on a
  day-to-day basis after permit issuance.

  Monitoring and  recording devices should be  selected
 prior to the trial burn. The data collected during the trial
 burn to establish permit limits and data collected after
 the trial burn to document compliance with those permit
 limits should be generated in the same manner. In se-
 lecting monitoring devices prior to the trial burn, thought
 should be given to post-trial burn operation, especially
 the ease of data retrieval. Depending on the complexity
 of a particular facility, a final RCRA permit may contain
 limits on from 20 to more than 60 parameters. Further,
 virtually all of the operating data  must be maintained
 until closure of the facility. This represents an enormous
 data storage challenge.

 A data maintenance system that allows for prompt data
 retrieval can minimize the time and inconvenience relat-
 ed to unannounced compliance inspections by regula-
 tory agencies. It also can make the  difference for
 proving compliance. Many facilities use strip charts to
 record continuously monitored parameters. Although
 strip charts meet the regulatory requirement for contin-
 uous recording, a facility should consider whether  stor-
 age and retrieval of hard copy data are  practical
 considering the number of  permit conditions that are
 likely to be imposed for that particular site. Use of strip
chart data also requires careful attention to chart speed,
inking problems, and whether overlapping traces make
the data indecipherable.
                                                   10

-------
Some final recordkeeping considerations  include the
ease with which data can be correlated and assimilated
and the desired flexibility for post-trial burn operation.
Hourly rolling average permit limits generally result in
more operating flexibility than instantaneous permit lim-
its. To receive hourly rolling average permit limits, how-
ever, a facility must plan ahead by installing necessary
data processing capabilities prior to the trial burn and by
collecting data in an hourly rolling average format. Cor-
relation of data is another important consideration. To
prove compliance with feed rate limits for metals, chlo-
rine, and ash, a facility owner/operator must be able to
correlate waste analysis records with continuous feed
rate  monitoring  records. Correlation  of continuously
monitored parameters via trend reports also might be
desirable for a facility from an operating standpoint or to
prove that all feeds were automatically cut off before a
particular parameter exceeded a permit limit.

2.4   Summary

The preceding discussion provides a glimpse into some
of the interrelated issues that must be  considered to
achieve compliance with the hazardous waste combus-
tion regulations. Careful planning, along with regular
communication between facility owner/operators and
regulatory agencies,  are essential for obtaining a final
RCRA permit that protects human  health and yet allows
for operating flexibility. These issues are explored in
considerably more detail in the chapters that follow.
                                                   11

-------

-------
                                              Chapter 3
             Combustion Technologies and Operational Parameters Related to
                            Particulate Matter and Metals Emissions
The most important factor affecting participate and met-
al air emissions from a combustion device is particle
size. Smaller particles are less likely to be captured by
air pollution control equipment (APCE) and are more
likely to  escape to the  atmosphere. Particle size de-
pends on several variables including temperature, the
form of the waste burned (i.e., solid, liquid, gas), and
many factors that are specific to the type of combustion
device. Six types of combustion devices—rotary kilns,
liquid injection  incinerators, controlled air incinerators,
fluidized  bed incinerators, boilers, and cement kilns—
are discussed in this chapter in terms of these variables
and their impact on the Effectiveness of air pollution
control devices.

3.1    Rotary Kilns
The rotary kiln  (Figure 3-1) is probably the most promi-
nent type of combustion system for incineration. These
devices are popular because they can operate in a wide
range of conditions and therefore can handle a wide
range of wastes.

             Solids
             feed
   Combustion
          air
      Waste
       feed
      Burner
             Features   -;
             • Range of teijeratures
             • Variety of w^v,
             • High entrainmeft
             • Discrete charges resulting in
               highly cyclicalbeliavior
             • Long residence time of solids
 Figure 3-1.  Rotary kiln.
Each rotary kiln  operates under different conditions
generating different emissions.  For example, rotary
kilns that devolatilize contaminated soil operate at tem-
peratures as low as 500°F, while rotary cement kilns
operate at temperatures as high as 2,800°F. (Cement
kilns are discussed in more detail in Section 3.6.) Be-
cause of this variation, no single temperature is charac-
teristic of a rotary kiln.  The operating  temperature
determines which metals will vaporize.

Rotary kilns typically have fairly high entrainment of
metals. Entrainment occurs because solids roll over and
over again inside the kiln, and are continually tumbled
and reintroduced  to the gas stream, providing multiple
opportunities for them to become entrained. In addition,
solids reside in the kiln a long time.

Many rotary kilns are charged discretely; often entire
drums are fed  into a kiln in a single charge. This means
that the temperature inside the kiln is cyclical. The met-
als may be at a much higher temperature initially as the
waste first begins to burn, and then at a lower tempera-
ture as the waste burns out before the next charge is
added. Because of the temperature variability inside the
kiln, the volatility  of metals may  be much higher than
would be expected from the average exit temperature.
As a result, a single time-averaged exit temperature is
not representative necessarily of the environment that
metals experience in a rotary kiln.

3.2   Liquid Injectors

The liquid injector (Figure 3-2) operates within a range
of temperatures and entrains metals 100 percent. Typi-
cally, the waste is atomized to relatively fine droplets,
which are entrained in the gas stream. As the droplets
evaporate and burn, the inorganics present in the waste
tend to remain behind in the gas stream. To a large ex-
tent, the size of the particles depends on the size distri-
bution of the original droplets. Larger droplets tend to
result in larger particles, and smaller droplets in smaller
particles. The fine droplets, because they tend to evap-
orate, produce finer  particles. Nevertheless, the parti-
cles are relatively large (from 5 to over 10 microns [urn]
in size).
                                                  13

-------
                  Features
                  • Range of temperatures
                  • 100% entrainment
                  • Entrained particle size
                   depends on size of droplets
                   generated by spray nozzle
                                                        Primary
                                                        burner
                                                       Waste
                                                        feed
                                                                                          Secondary
                                                                                        f    burner
                                                                      Secondary
                                                                      chamber
                     Primary
                     chamber
                                                                                            Secondary
                                                                                                air

                                                                                             Flame port
                   Primary air

            Features
            • Low temperature
            • Low entrainment
            • Air-starved primary, requires afterburner
 Figure 3-2. Liquid Injector.
 Figure 3-3. Controlled air incinerator.
 Particle size is similar for both liquid injectors and rotary
 Win incinerators. In  both  combustion devices, the en-
 trained particles are typically larger than 10 urn as com-
 pared to  particles  generated by  vaporization  and
 condensation, which are typically of submicron size re-
 gardless of whether they are from a rotary kiln or a liquid
 Injector.  Volatilization from  a liquid  injector is greater
 than volatilization from a rotary kiln, depending on such
 parameters as the temperature and the chlorine in the
 environment. All other parameters being  equal, howev-
 er, more metals probably vaporize from a liquid injector
 due to more intimate contact with the gas stream and
 better mixing, resulting in fewer limitations to diffusion.

 3.3   Controlled (Starved) Air Incinerators
 The controlled air incinerator is  popular  for medical
 waste or municipal waste incineration (Figure 3-3).  In
 this type of incinerator, the organic wastes are  intro-
 duced into the gas stream by means of  an air-starved
 primary chamber; they are subsequently burned out in a
 secondary  chamber,  an  afterburner.  The  oxygen
 present in  the primary chamber is insufficient to  com-
 pletely burn the waste. This oxygen deficiency means
that these devices typically operate at  low  tempera-
tures, resulting in relatively little metal vaporization.
Controlled air incinerators also are  characterized by rel-
atively low entrainment simply because  less  air  goes
through them. In these devices, however, an afterburner
operates at much higher temperatures, so  any en-
trained metals leaving the primary chamber will  pass
through a much hotter environment in the  afterburner,
where they may vaporize.
3.4   Fluidized Bed Incinerators

Fluidized bed incinerators are fundamentally different
from other types of incinerators (Figure 3-4). In a fluid-
ized bed incinerator, the air or combustion gas flows up-
ward through a relatively large bed of inert particles,
such as sand particles or some other inert species. The
velocity of this combustion gas is sufficient to  lift the
particles up, so that they separate from each other and
fluidize. These fluidized particles bounce around, and
any hazardous waste within this bed tends to bounce
around as well.

Fluidized beds have excellent  mixing. The bouncing
particles tend to make the composition of combustion
gas uniform within the  bed, and the temperatures and
heat transfer also are relatively  uniform within the bed.
Very high entrainment occurs for two reasons: (1) a fair-
ly high velocity of air or combustion gas goes through
the system, and (2) bouncing tends to grind down both
the waste particles and the bed particles. Eventually, a
substantial proportion of the waste and the bed parti-
cles are so finely ground that they become entrained.
One of the problems with fluidized beds is that, typically,
solids fed to a fluidized bed have to be ground up first so
that the solids are relatively fine. This pretreatment  of
waste can be expensive, and waste that is not properly
prepared can disrupt the fluidization of the bed.

The frequency with which the bed has to be loaded de-
pends on a number of factors. If the bed material is very
hard and durable, it seldom has to be replaced. The
frequency of loading the bed also depends on how the
bed is used. In some  cases, fluidized beds also are
                                                    14

-------
                       Exhaust
      Sand
      feed
Freeboard
     Waste
     feed
   Fluidizing
    air inlet
                   Fluidized sand bed
                    Burner
 Windbox
Startup
burner
        Features
        • Low temperature
        • High entrainment
        • Long solids residence time
        • Long response time to change in metal feed rate
      ,  • Bed material can act as sorbent

Figure 3-4.  Fluidized bed incinerator.
used as pollution control equipment to control, for ex-
ample, sulfur or chlorine emissions. A. calcium carbon-
ate or calcium oxide bed tends to react with acid gases,
so would have to be recharged more often, depending
upon the sulfur and chlorine content of the waste. Some
fluidized beds use the ash (i.e., the inorganic content of
the waste and  fuels) as part of the bed itself, so as
waste is incinerated, material accumulates in the bed. A
balance must be  achieved between the amount fed in
and the amount ground down and pushed out, or the
contents of the bed have to be removed periodically.
The frequency  of removal depends on  how often the
system is used  and how often the bed is recharged.

An  additional problem with fluidized beds is their long
solids residence time. As a result, the concentration of
metals in the bed responds slowly to a change in feed
concentration, especially if the feed contains nonvolatile
metals, which remain with the bed and do not vaporize
and exit with  the  gas. Because the bed is cleaned out
infrequently, the metal remains in the incinerator for a
longtime.

The long solids residence time has implications for the
design of a trial burn. In  a trial burn, the goal is to dem-
onstrate "worst-case" emissions in which the metals
emissions from the bed represent the highest expected
metals emissions. If the metals are spiked, however, as
is typically done for a trial burn, they will stay in the bed
for a long time before being emitted, because the waste
particles take a long time to grind down and become
entrained. Consequently, the concentration of metals in
the emissions takes a long time to reach equilibrium in
response to a change in the feed concentration. The
length of time required to reach equilibrium is an impor-
tant consideration in designing and conducting the trial
burn. (Trial burns are discussed in detail in Chapters 4
and 5.)

Certain  types of bed material in fluidized beds can act
as a sorbent, enhancing the capture of some metals.
Fluidized beds also can operate at lower temperatures
because of the high turbulence and the resulting excel-
lent mixing. Because fluidized beds operate at low tem-
peratures, they cause less vaporization  of metals and
thus might be more effective than higher  temperature
processes for capturing volatile metals.

3.5  Boilers

Boilers  are typically high-temperature devices (Figure
3-5) that transfer heat by  means of tubes inside  the
combustion  chamber. The heat from the combustion
                                                  Superheaters
                                                Reh eaters
                                   Overfire
                                     air
                                                                         Waterwalls
                                          Burners
                                                Features
                                                • High temperature
                                                • Soot blowing considerations
                                Figure 3-5.  Boiler.
                                                   15

-------
gas flowing on the outside of the tubes is transferred to
the water or steam on the inside of the tubes. The inte-
rior of a boiler is very dirty because the particulate mat-
ter generated accumulates on the surfaces of the tubes,
resulting in reduced heat transfer.
To counteract the accumulation of particles on boiler
tubes,  a boiler periodically blows high-velocity air or
steam  onto the tubes to clean their surfaces. This pro-
cess is known as soot blowing and is an important con-
sideration when designing a trial burn. During the short
period  of time when soot blowing is occurring, a combi-
nation of previously deposited metals, soot, and particu-
late matter surges. This surge of particulate enters the
air pollution control device. Because of soot blowing and
the consequent particulate surges, part of the trial burn
must be conducted under soot blowing conditions.
Mathematical formulas for calculating how much of the
test has to  be conducted under  soot blowing versus
non-soot blowing  conditions are provided in the boiler
and  industrial furnace  (BIF) implementation manual
(U.S. EPA, 1992).

3.6    Cement Kilns
The problems associated with metals emissions from
cement kilns (Figure 3-6) are  so different than those
from other types of boilers and industrial furnaces that
cement kilns are regulated separately in the BIF rule.

3.6.1   Operation
Cement kilns operate at very high temperatures. Typi-
cally, the solid product in  a cement  kiln is 2,500 to
2,800°F; the gas  temperature  is  several hundred de-
grees hotter. The  amount of metal in the raw materials
for cement kilns can exceed that in both the waste and
the fuel. Therefore, the metals in the raw materials have
to be taken  into account to adequately control metals
emissions.
In addition, the cement matrix in which the metals are
bound  affects their volatility. Cement is  a mixture of
compounds that forms complexes with some metals,
thereby decreasing the effective volatility of those met-
als. Therefore, such metals are much less volatile in ce-
ment  kilns than  would  be  expected  based on
thermodynamics.

An important aspect of cement kilns is that they recycle
particulate  matter. The recycling of volatile metals oc-
curs when volatile metals vaporize, condense, and re-
turn to the system.  Cement kilns  use counterflow
processes in which the fuel and the air are introduced
on one end of the cement kiln, while the raw materials
are introduced on the opposite end. As the hot burning
fuel and air go through the cement kiln, heat is trans-
ferred  to the raw materials, which eventually turn into
cement clinker product. In this intermediately high tem-
perature zone, the volatile metals tend to vaporize; but
as they continue with the gas through the process, they
tend to cool and condense onto the surface of existing
particles. When the volatile metals travel through a sus-
pension preheater, they condense or agglomerate onto
the large surface area of existing particles in the form of
raw materials. Internal recirculation occurs as the con-
densed metals are fed back into the kiln along with the
raw materials.

Any metals that escape into the gas stream and go to-
ward the stack  must  travel through  a dust collector
(e.g., bag house, electrostatic  precipitator),  which col-
lects the vast majority of the particulate matter along
with most of the metals. This collected particulate mat-
ter consists mostly of raw materials that were blown into
the preheater along with any metals that have con-
densed onto the surface of the raw material particles.
Cement kilns typically recycle most of this material back
into the preheater. As  a result of this  recirculation, ce-
ment kilns respond slowly to changes in the metals feed
rate, as do fluidized beds. Consequently, these types of
devices have to undergo a different trial burn procedure
than do other combustion devices, because the system
must reach equilibrium with respect to metals emissions
before  the trial burn can begin.
                                       Preheater
                               Kiln
     Clinker
     cooler
                                                                 Stack
     L*-A-/luvJ
                                                   Dust
                                                 collector
                        Features
                        • Very high temperature
                        • Metals in raw materials
                        • Matrix effects
                        • Recycled particulate matter
Figure 3-6. Schematic of cement kiln, showing clinker cooler, preheater, dust collector, and stack.


                                                   16

-------
 3.6.2  Case Study of Lead Incineration

 A case study of a cement kiln incinerating lead-contam-
 inated material is presented as an example of how ce-
 ment kilns work. The cement kiln in this example was
 recycling all of its  paniculate matter (so it is not repre-
 sentative of every cement kiln). The amount of lead en-
 tering the kiln in the raw materials and fuel was at least
 as much as, if not more than, the amount of lead coming
 from the waste itself.

 Since metals  are neither  created nor destroyed in this
 system, metals entering the system must  exit.  Figure
 3-7  is a schematic showing the recirculation and parti-
 tioning  to various  exit streams. The spacing between
      Raw material
          Raw mill
                       Raw material   -««i

                                     ii
                             ig-yiTTreri. i-
                   SSSSSSSS^SSS!:
                     Jt. X*"*-"«"^™*™^-~*TBI^ ]>?•<>
                     "
                                      Emissions
              External
              recirculation of
              APCS dust
      Clinker
Internal recirculation
in preheater
Figure 3-7.  Recirculation and partitioning of lead to various
          exit streams in a cement kiln (Sprung, 1985).
lines in the schematic is proportional to the mass flow of
lead in the stream. This particular system has two pos-
sible exit streams: with the clinker product or with the
stack emissions. Since lead is a volatile metal, much of
it vaporizes in the process. A portion of the metals that
vaporize or are entrained in the cement kiln are internal-
ly recirculated back into the preheater, where they tend
to condense onto the raw material particles, which are
sent immediately back into the kiln. Some of the metals
escape from the preheater into the air pollution control
device, where they are caught and subsequently added
to the raw materials. Any metal that escapes the air pol-
lution control device is emitted. This particular system
recycles all of the waste dust; therefore, most of the
lead in the system ultimately goes to the clinker, while a
small portion of the lead escapes through the stack.
                              In most cement kilns, a third possible exit stream is the
                              fraction of the waste dust collected in the bag house and
                              discarded. The elimination  of this waste dust reduces
                              the amount of recirculating volatile  metal, thereby re-
                              ducing the fraction of volatile metal that ultimately goes
                              to the clinker or the stack.

                              The concentration of lead emitted is proportional to the
                              concentration of lead collected in the kiln dust. If lead is
                              spiked into the kiln at temporarily high feed rates, as in a
                              trial burn, the concentration of lead in the collected kiln
                              dust  is initially  quite low.  But the  lead  emissions in-
                              crease as the lead builds up in the recirculating system,
                              and they continue to increase until reaching a steady-
                              state equilibrium. The time required to reach equilibrium
                              is system specific and depends on such variables as the
                              design of the cement kiln and the fraction of the dust
                              that is recycled. Data show that cement kilns take from
                              3 to more than 15 hours to reach steady state. As shown
                              in Figure 3-8, one facility took 8 hours to reach 90  per-
                              cent of the ultimate steady-state concentration of lead in
                              the waste dust (Clark et al., 1991). (The actual numbers
                              are not shown in the figure, at the facility's request. The
                              numbers used for the lead concentration in waste dust
                              are normalized concentrations proportional to the abso-
                              lute concentrations.)
                                                                100
                                                       Percent of
                                                       steady-state
                                                       concentration
                                                                 20-
                                           0          5          10         15
                                                      Spiking time (hr)
                             Figure 3-8.  Transient behavior in cement kilns of lead
                                       concentration in waste dust due to recycled
                                       particulate matter (adapted from Clark et al., 1991).
                             3.7  References

                             When an NTIS number is cited in a reference, that doc-
                             ument is available from:
                               National Technical Information Service
                               5285  Port Royal  Road
                               Springfield,  VA22161
                               703-487-4650
                                                   17

-------
Clark, W.D., R.G. Rizeq, G.P. Franklin, and R.H. Sim-
  mons. 1991. Spiking time required for metals emis-
  sions to approach  steady  state  in cement kilns
  burning  hazardous waste. Presented at the Second
  International  Congress on Toxic  Combustion  By-
  products: Formation and Control, University of Utah.

Sprung, S. 1985. Technological problems  in pyropro-
  cessing  cement clinker: Cause and solution. Dussel-
  dorf: Breton-Verlag.
U.S. EPA. 1992. U.S. Environmental Protection Agency.
  Technical implementation document for EPA's boiler
  and industrial  furnace  regulations. EPA/530/R-92/
  011. NTIS PB92-154947. Office of Solid Waste and
  Emergency Response, Washington, DC.
                                                  18

-------
                                              Chapter 4
                           Control Parameters and Permit Conditions
 To ensure that harmful levels of pollutants are not emit-
 ted from combustion devices burning hazardous wastes
 (incinerators, or boilers and industrial furnaces [BIFs]),
 the Resource Conservation and Recoveiy Act (RCRA)
 and the regulations written under it require facilities that
 burn hazardous waste  to obtain a permit.  The permit
 includes conditions that place limits on such operation-
 al parameters  as combustion chamber temperature,
 gas flow rate, and waste feed rates. This approach to
 controlling  pollutants is well established for incinera-
 tors, but is relatively new for BIFs, which traditionally
 have been operated to  maximize energy generation or
 production  rate. Using  BIFs for hazardous waste  de-
 struction can necessitate restrictions on operating con-
 ditions that, while minimizing the release of pollutants
 from the combustion of the hazardous waste, might af-
 fect the "combustor's"  main function. The establish-
 ment of protective,  realistic, and  enforceable control
 parameters, therefore, is crucial to ensuring that incin-
 erators and BIFs safely destroy waste while continuing
 to meet their intended industrial purpose. For this dis-
 cussion, combustor is used to mean any combustion
 device,  whether incinerator or BIF, that burns hazard-
 ous waste.

 Permit conditions are set  on  the basis of a  series of
 tests—termed the trial  burn or compliance test—con-
 ducted under the "worst-case" operating conditions for
 the combustor, the conditions under which the greatest
 amounts of harmful emissions  are  generated. If  the
 combustor meets all environmental requirements under
 worst-case conditions in the trial burn, the U.S. Environ-
 mental Protection Agency (EPA) assumes that it also
 will satisfy the  requirements  during  operation under
 less severe operating conditions.  For example, if  the
 combustor maintains the required metals emission rate
 at a maximum combustion chamber temperature {one
 of the conditions that is considered a worst case  for
 metals emissions), the emission rate is assumed to be
 no higher at a lower operating temperature. Yet, since
 more than one operational parameter affects each type
 of environmental impact, the  trial  burn must be per-
formed under worst-case operating conditions for each
operational parameter.

This approach requires defining the operational param-
eters that affect each of the  performance standards
 (e.g., destruction and removal efficiency [ORE], particu-
 late emissions) required under the RCRA incinerator
 and BIF regulations (40 CFR parts 264-272).  These
 regulations further specify that limits must be  set on
 certain operational parameters. EPA's BIF (U.S. EPA,
 1992) and incinerator (U.S. EPA, 1989a) guidance doc-
 uments define additional control parameters and speci-
 fy how limits are set. This general guidance forms the
 basis  for setting  control parameters. Because each
 combustion  system has unique operational require-
 ments, however, specific permit conditions have to be
 established. When gathering data  for the permit, the
 owner/operator and the permit writer must:

 • Identify the specific operational parameters that af-
   fect each regulated environmental impact.

 * Establish the worst-case value for those parameters
   (i.e., maximum or minimum).

 • Ensure that all  permit conditions are physically con-
   sistent with one another and with the operational re-
   quirements for the system.

 4.1    Permitting Approaches

 The EPA incinerator guidance  document (U.S. EPA,
 1989a) recommends that limits be set using one of the
 three approaches listed in Table 4-1, which are based
 on the premise that the more complex the trial burn, the
 more flexible the resulting permit conditions.

 The  single point approach is the simplest. It requires
 that one set of operating conditions be set on the worst-
 case conditions required for burning a particular waste
 stream. Although more than one waste category (e.g.,
 solid or liquid) can be included in  the waste stream,
 each must be specifically described. This approach is
 appropriate for combustors that burn wastes with well-
 defined and consistent  properties.  Examples  include
 combustors that receive wastes from one or a few spe-
 cific industrial processes or combustors that extensive-
 ly blend waste to achieve a relatively narrow range of
 properties.

The multiple point approach is appropriate for facilities
that burn a few well-defined  waste types for which one
set of operating  parameters applies.  For example,
where the solid waste stream can be either a high-Btu,
                                                  19

-------
Table 4-1. Advantages and Disadvantages of Permitting
         Strategies
Permitting Method   Advantages
                Disadvantages
Single waste/one
operating condition
(stngh point)

Multiple waste/
multiple operating
conditions (multiple
paint or campaign
burning)
Trial burn is
relatively simple.
Limited operational
flexibility.
The greater the     Limited operational
number of tests, the  flexibility within
more operating     campaign.
modes.
                Multiple trial burns are
                required.
Multiple waste/
single operating
conditions
(universal)
This method can
result in more
flexible operation.
Trial burn can be very
complex, with
numerous tests
representing different
worst-case conditions.
Source: U.S. EPA, 1989a.
readily combustible solid or a contaminated soil, main-
taining the maximum waste feed rate to both the prima-
ry and secondary combustion chambers during the trial
burn may prove otherwise impossible.  Each mode of
operation (i.e., high-Btu solid  waste mode and  soils
mode) would require its own unique set of operating
conditions. To make the multiple point permit enforce-
able, each mode of operation would need to be tied to
well-defined waste types, and operating records would
need to document that the waste type fits the particular
set of operating conditions at all times.
While the universal approach is the most complex per-
mitting  strategy  for setting  parameters, once estab-
lished, the facility  can  burn a wide variety of wastes
without changing operating conditions. The trial burn for
this approach is complicated because  all parameters
normally cannot be established for a worst case during
one set of tests. When setting permit conditions  from
multiple trial burns, one must keep in mind that many
parameters are physically  interrelated.

4.2  Operational Parameters in General
The operational parameters for incinerators and  BIFs
correspond to the control parameters for which permit
conditions are set on the basis of the trial burn and en-
gineering judgment. The EPA incinerator guidance doc-
ument identifies 14 control parameters for which permit
conditions should be set (Table 4-2). The BIF guidance
document identifies the same  control parameters but
expands the list to include regulation  of toxic metal
emissions (Table 4-3). (These parameters also are used
in incinerator permits to implement metals limitations.)
Parameters are divided into three groups in Table 4-2,
depending on whether they are interlocked with the au-
tomatic waste feed cutoff (AWFCO) and whether they
are set from the trial burn or from design specifications.
Although the BIF guidance document does not segre-
gate parameters into these categories, the groupings
generally can be applied to BIFs as well.

Each of the permit conditions in Tables 4-2 and 4-3 en-
sures that the combustor minimizes one or more of the
following:

•  Emissions of hazardous (regulated) organic  com-
   pounds.

   - Principal organic hazardous constituents (POHCs)
     via ORE.
   - Products of incomplete combustion (PICs) via car-
     bon monoxide  or  hydrocarbon and air  pollution
     control equipment (APCE) temperature.

•  Hydrogen chloride  (HCI) and chlorine (Cl2) emis-
   sions.

•  Paniculate and toxic metals emissions.

•  The likelihood of fugitive emissions and system up-
   sets.
The purpose of each control parameter is identified in
Table 4-4, where many conditions are  shown to serve
more than one purpose. The  method  used to set the
limits for each control parameter is discussed below in
the context of the parameter's environmental purpose.

4.3  Minimization of Organic Emissions
The release of organic emissions from a combustor can
occur in two ways: (1) the release of unburned organic
compounds from the waste and (2) the formation of new
organic compounds during the combustion  process.
The first type of occurrence is regulated by specific re-
quirements for the combustor's ORE. As discussed in
Chapter 1, regulations require that a combustor's effi-
ciency be rated at s 99.99 percent for most hazardous
wastes and s=  99.9999 percent for dioxin-listed wastes.
Emissions of PICs are regulated on the basis of empir-
ical data. Numerous tests show that whenever carbon
monoxide concentrations in the flue gas of a combustor
are less than 100 ppm (dry, corrected to 7 percent oxy-
gen), few PICs are emitted. At higher carbon monoxide
concentrations, PICs might form. If the combustor's
carbon monoxide emissions exceed the 100 ppm limit,
then PIC emissions are regulated by limiting total hy-
drocarbon (THC) emissions.

Although the ORE of a combustor is closely related to
its destruction efficiency (DE), the terms are not equiva-
lent. ORE limits emissions only; for example, 99.99 per-
cent ORE corresponds to an emission limit of 100 mg of
POHC  per kg fed to the combustor. In principle, the
compound could be destroyed, remain in the ash, or be
captured in the APCE.  DE,  however,  represents the
fraction  of the organic  compound that is actually de-
stroyed. Beyond this theoretical  difference, the DRE
and DE generally correlate well, especially because the
                                                  20

-------
 Table 4-2. Control Parameters for Incinerators

 Group                          Parameter
 Group A

 Continuously monitored parameters
 are interlocked with the automatic
 waste feed cutoff. Interruption of
 waste feed is automatic when the
 specified limits are exceeded. The
 Parameters are applicable to all
 facilities.
                                 1. Minimum temperature measured at each combustion chamber exit

                                 2. 100 dppmv (corrected to 7% O2) if CO > 100 ppmv, then set at trial burn value and add limit for
                                   hydrocarbons of 20 dppmv (as propane, corrected to 7% O2)

                                 3. Maximum flue gas flow rate or velocity measured at the stack or other appropriate location

                                 4. Maximum pressure in PCC and SCC

                                 5. Maximum feed rate of each waste type to each combustion chamber

                                 6. The following as applicable to the facility:
                                   • Minimum differential pressure across particulate Venturi scrubber
                                   • Minimum liquid-to-gas ratio and pH to wet scrubber
                                   • Minimum caustic feed to dry scrubber
                                   • Minimum kVA settings to ESP (wet/dry) and kV for ionizing wet scrubber (IWS)
                                   • Minimum pressure differential across bag house
                                   • Minimum liquid flow rate to IWS
 Group B

 Parameters do not require
 continuous monitoring and thus are
 not interlocked with the waste feed
 cutoff systems. Nevertheless,
 operating records are required to
 ensure that trial burn worst-case
 conditions are not exceeded.

 Group C

 Limits on these parameters are set
 independently of trial burn test
 conditions. Instead, limits are based
 on equipment manufacturers'
 design and operating specifications
 and thus are considered good
 operating practices. Selected
 parameters do not require
 continuous monitoring and are not
 interlocked with the waste feed
 cutoff.
                                 7. POHC incinerability limits

                                 8. Maximum total halides and ash feed rate to the incinerator system

                                 9. Maximum size of batches or containerized waste

                                10. Minimum particulate scrubber blowdown or total solids content of the scrubber liquid



                                11. Minimum/maximum nozzle pressure to scrubber

                                12. Maximum total heat input (capacity) for each chamber

                                13. Liquid injection chamber burner settings:
                                   • Maximum viscosity of pumped waste
                                   • Maximum burner turndown
                                   • Minimum atomization fluid pressure
                                   • Minimum waste heating value (only applicable when a given waste provides 100% heat input to
                                    a given combustion chamber)


                                14. APCE inlet gas temperature

Source: U.S. EPA, 1989a.
majority of organic material burned by a combustor is
destroyed rather than captured.

Although waste can contain many different types of or-
ganic constituents, only compounds listed in Appendix
VIII of 40 CFR part 261 are specifically regulated under
RCRA incinerator and BIF regulations. Recognizing the
impossibility of measuring the ORE for all organic com-
pounds that could be present in a given waste, 40 CFR
section 266.104 specifies that the ORE be measured
and set for POHCs, namely, organic compounds that
are chosen for testing  because they are likely to be
present in the waste in "significant quantities" and that
are considered difficult to destroy.

4.3.1   Mechanism of Organic Compound
        Destruction

Consideration of the mechanisms of organic compound
destruction is necessary  for understanding  how the
                                                          permit conditions (e.g., temperature and gas flow rate)
                                                          relating to emissions of hazardous organic compounds
                                                          are set. The process of combustion can be viewed as
                                                          taking place in three primary zones (Figure 4-1): (1) the
                                                          volatilization or pyrolysis zone, referred to here as the
                                                          pre-flame zone,  (2) the flame zone,  and  (3) the post-
                                                          flame or burnout zone.  In the pre-flame zone, the or-
                                                          ganic material in the gaseous, liquid, or  solid fuel is
                                                          vaporized and mixed with air or another source of oxy-
                                                          gen. Organic compounds that do not vaporize typically
                                                          pyrolyze,  forming  a combustible  mixture of  organic
                                                          gases. In the flame  zone,  the chemical reactions are
                                                          rapid (on the order of milliseconds), and temperatures
                                                          are high (on the  order of 3,000°F+). Although the vast
                                                          majority of the organic compounds in the waste are de-
                                                          stroyed in the flame zone,  the level  of destruction is
                                                          usually below that  required for hazardous waste incin-
                                                          eration.
                                                       21

-------
Table 4-3. Operating Parameters for BIFs for Which Limits Are Established During Precompliance,
           Compliance, and Permit Periods
                                                                  Operating Limits
Total feed rate of hazardous waste X
Total feed rate of pumpable hazardous waste" X
Feed rate of each of the 1 0 BIF-regulated metals in:
• Total feed streams X
• Total hazardous waste feed streams0'11 X
• Total pumpable hazardous waste feed streams'1'6 X
Total feed rate of chlorine and chloride in total feed streams X
Total feed rate of ash in total feed streams' X
Maximum production rate when producing normal product X
CO concentration in stack gas
HC concentration in stack gas, if necessary^
Maximum combustion chamber temperature"
Maximum flue gas temperature entering the paniculate matter
control device"
Various APCS-specific operating parameters"'1
Minimum production rate when producing normal product,
if applicable
Minimum combustion gas temperature!
Maximum emission rate for each metalk
Maximum emission rate for HCI and CI2
Feed rate of other fuels
Appropriate controls of the hazardous waste firing system
Appropriate indicator of combustion gas velocity
Allowable variation in boiler and industrial furnace system
design or operating procedures
Other operating requirements as are necessary to ensure
that ORE is met
X
X

X
X
X
X
X
X
X
X
X
X
X









X
X

X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
•See 40 CFR section 266.102(c) for complete listing and description of permit operating requirements.
6 Not applicable if complying with Tier I or Adjusted Tier I metals feed rate screening limits.
°Not applicable during compliance or permit period if complying with Tier I or Adjusted Tier I metals feed rate
screening limits.
"The final BIF Rule specifies that facilities complying with Tier I or Adjusted Tier I metals feed rate screening
limits must establish limits for these parameters during interim status (precompliance or compliance, as
noted). EPA is considering amending the rule to rescind the requirements for facilities complying with Tier I or
Adjusted Tier I metals feed rate screening limits to establish limits on these parameters.
• Not applicable during precompliance or permit period if complying with Tier I or Adjusted Tier I  metals feed
rate screening limits.
'Not applicable for cement and lightweight aggregate kilns.
fl HC limit necessary if operating under Tier II controls for PICs or if feeding wastes at locations other then the
hot end.
h Parameters are specified in 40 CFR section 266.103(c)(2)(ix-xiii).
1 Limits not applicable if complying with Tier I or Adjusted Tier I for metals and total chlorine and  chloride.
i During compliance, minimum combustion chamber need only be maintained following a waste  feed cutoff, for
the time that the waste remains in the chamber.
"Not applicable if complying with Tier I or Adjusted Tier I total  chloride and chlorine feed rate screening limits.
Source: U.S. EPA, 1992.
                                               22

-------
Table 4-4. Purposes of the Control Parameters
Partic-
POHCs/ ulatesi/
Parameter PICs Metals
Group A
1. Min.T X
1A. Max.T
2. Max. CO X
3. Max. gas flow X
4. Max. pressure
5. Max. waste feed I
6. Particulate APCD
6A. Acid gas APCD
14. Max. APCE inlet T
Group B
7. Incinerability X
8. Max. halides feed
8. Max. ash feed
8A. Max. metals feed
9. Max. size of batches X
10. Min. blowdown
Group C
1 1 . MinVMax. scrubber


X

X


X

X


I
X
X

X

X
Fugitive
Emis-
HCI sions




X X
X
I I

X



X


X


X
              Nozzle pressure change
              (if applicable)
          12.  Max. heat input
          13.  Burner settings:
              • Max. viscosity
              • Max. burner turndown
              • Min. atcmization fluid
               pressure
              • Min. waste heating value
    X
    X
    X

    X
          X = Direct impact
          I = Indirect impact
          Note: Parameter numbers correspond to those in Table 4-2.
          The gases from the flame zone flow into the post-flame
          zone of the combustor, where they mix with additional
          air. Along  with  providing additional oxygen that im-
          proves the destruction of organics, the mixing with air
          also cools the gases. In this zone, the destruction of
          organics is determined by temperature, time, and turbu-
          lence factors  (commonly referred to as the three Ts).
                    TIT
                                 Liquid waste
                                 storage tank
Atomizing
medium
(air,  _j,
nitrogen,
or steam)
I
                        Supplemental
                        fuel, if
                        required
I/-"Pre-flame zone   1    TO ajr
I*x-  - Flame zone ^  pollution
                     control
                     device
                    Combustion air
                                       Post-flame zone
                                                  r
Temperature is critical because it determines trie rate of
organics destruction. Time refers  to the length of time
that the gases are present in the combustion chamber,
frequently called the residence time. Turbulence is the
most conceptually complex of the three terms, because
it describes the ability of the combustion system to suf-
ficiently mix the gases with oxygen to oxidize the organ-
ics released from the fuel.

The gas temperature in the post-flame zone of a com-
bustor is typically in the 1,200 to  2,200°F range.1 The
chemical reactions that result in the destruction  of the
organic compounds continue to occur in the post-flame
zone, but because of the lower temperatures, the reac-
tion rates are much slower than in the flame zone (typi-
cally on the order of tenths of a second). Given the
longer reaction times, the temperature of the gases in
the post-flame zone needs to be maintained for a rela-
tively long time (on the order of 1  to 2 seconds)  to en-
sure  adequate destruction. Successful design  of  a
combustion chamber requires that it be capable of
maintaining the combustion gases at a high  enough
temperature for a long enough period to complete the
destruction of the hazardous organic constituents.

Turbulence is an  important factor  because the process
of combustion consumes  oxygen in the immediate vi-
cinity of pockets  of fuel-rich vapor. Because the de-
struction of organic compounds occurs far more rapidly
and cleanly under oxidizing conditions, the combustion
gases moving away from the oxygen-poor pockets of
gas need to be mixed with the oxygen-rich gases in the
bulk of the combustion chamber.

Thus, turbulence  can be thought of as the ability of the
combustor to keep the products of combustion  mixed
with oxygen at an elevated temperature. The better this
ability (up to a point), the higher the degree of destruc-
tion of the organics. Conditions of poor turbulence are
often  characterized  by high carbon  monoxide  emis-
sions. The lack of mixing prevents the hot organic ma-
terial  from  coming  in  proper contact with oxygen,
resulting in localized pockets where pyrolytic conditions
exist. Consequently, low carbon monoxide concentra-
tion is indicative of both good organics destruction and
low PIC formation.

The three Ts of combustion can be directly related to
the conditions for control  parameters given in  Tables
4-3 and 4-4. Because combustion gas flow rate is di-
rectly related to  residence  time,  the greater the flow
rate, the lower the residence time. Technically, the per-
mit condition should place a maximum limit on the com-
bustion gas flow rate; however, because gas flow rate is
            Figure 4-1.  Mechanics of POHC destruction in a single
                      chamber incinerator.
                                1 Temperature ranges and reaction times are intended to provide
                                 a sense of orders of magnitude. This discussion should not be inter-
                                 preted to mean that 1 or 2 seconds are adequate for complete de-
                                 struction or that a shorter residence time or lower temperature is not
                                 acceptable.
                                                              23

-------
relatively difficult to measure, the limit is set on an indi-
cator of this parameter. As discussed below, the limit
only needs to be such that the combustion gas flow rate
is lower than that measured during the trial burn.
In theory, the minimum  combustion gas flow rate can
result in turbulence that is insufficient to allow proper
mixing; however,  this condition is highly unlikely be-
cause difficulties in maintaining a minimum tempera-
ture will be encountered before this low a gas flow rate
is reached. At such poor turbulence, carbon monoxide
limits likely will be exceeded.

While the three Ts probably are the most important fac-
tors affecting organic compound emissions from haz-
ardous waste combustors, they are not the only ones.
Other factors that  can influence  ORE are  waste feed
rate, the types of  organic compounds being  burned,
and burner settings that govern waste atomization.
Waste feed rate affects  the emission of organics indi-
rectly. The effect is minimal as long as the heat input to
the combustion chamber is maintained,  but excessive
waste feed rates can "smother" the flame and tax the
waste feed system's ability to vaporize and pyrolyze the
organic constituents. A parameter that is closely related
to waste feed rate is the size of batches or containers of
waste fed into the system. Permit limits for combustors
that are fed solid waste  in discrete batches rather than
continuously should be  set on the maximum size and
frequency of batches to prevent  instantaneous over-
charging.
The type of organic compound to be burned also is a
factor in the destruction of organics. The more difficult a
given compound is to destroy, the more likely it is to
pass through the combustion system. The difficulty of a
given compound's destruction is estimated by its Incin-
erability ranking and its quantity in the waste. Quantity
is a POHC selection criterion, but concentration corre-
lates inversely with destruction. (Incinerability ranking
Is discussed in Section 4.3.2.2.) Parameters relating to
the combustor's burners are important for ensuring that
liquid waste is properly atomized.  Factors such as the
waste's viscosity, burner turndown (the fraction of the
design capacity at which the burner is operating), and
the pressure of the atomizing fluid (where  applicable)
influence the level  of waste atomization, which can in
turn influence the destruction of POHCs or the forma-
tion of PICs. The limits are based on the design specifi-
cations for the particular waste nozzles.

The droplets of fuel will  be larger  and will stay in sus-
pension for a shorter time if not properly atomized. The
partially burned droplets can fall  into the combustion
chamber, strike a cooler surface, and vaporize, rather
than remaining in the flame zone where the majority of
the organic compound destruction occurs. Inadequate
atomization can be caused by any of the following:
•  The nozzle has corroded, eroded, or clogged.

•  The fluid's viscosity is outside the nozzle's operating
   range.

•  The atomizing fluid is being delivered at the wrong
   pressure.

•  The fuel is being delivered to the nozzle at a flow rate
   that is outside its operating range.

•  The fuel is nonhomogeneous.

A final parameter for which permit limits are commonly
set is the minimum heating value of wastes that provide
all of the heat input to the combustion  chamber. This
limit is set to ensure that the temperature in the com-
bustion chamber will be maintained at a  reduced waste
feed rate. In  principle, this limit is unnecessary because
a  reduced heating value will manifest  itself as a re-
duced  combustion chamber temperature. In practice,
however, such a limit is desirable because  too low a
heating value for the waste stream can result in an un-
stable flame. Although minimum combustion chamber
temperature can be maintained by reducing the amount
of excess air introduced to the system, the  net result
would be a substantial change in operating conditions
as established in the  trial burn,  making the system's
ORE behavior less predictable.

Thus, the combustion process is highly interrelated,
and the failure to achieve adequate combustion can be
detected by monitoring changes in  readily measured
parameters.  In particular, many failures in the combus-
tion process manifest  themselves as increases in  car-
bon  monoxide.  As a result,  interlocking the carbon
monoxide monitor with the hazardous waste feed is a
requirement for all hazardous waste combustion devic-
es. Systems that produce such elevated levels of car-
bon  monoxide from  their raw materials  that  this
parameter may not be an adequate indicator of com-
bustion problems (e.g., cement kilns and other types of
industrial furnaces) are required to operate using an in-
terlocked THC monitor.

Upsets in the system also can manifest themselves as
changes in the furnace temperature. Such fluctuations
are rarely caused by failure of the combustion process
itself and typically  indicate an unexpected change in
the composition of certain critical waste streams. Such
occurrences trigger an automatic cutoff  of the hazard-
ous waste feed.

4.3.2  Incinerability Considerations

4.3.2.1 POHC Selection

Because of  the wide range  of organic compounds
present in most wastes, testing for each one is not pos-
sible. Thus, the approach taken when testing for worst-
case conditions is to select the most difficult to destroy
                                                  24

-------
compounds (i.e., POHCs) that are likely to be present in
the waste and demonstrate during the trial burn that
they can be properly destroyed. The POHCs are cho-
sen  based on the types of organic hazardous com-
pounds  (i.e., those listed in Appendix VIII) that will be
present  in the waste and the incinerability rankings (dis-
cussed below) that are applicable.

The POHC selection process begins with an examina-
tion of the waste streams that will be fed into the com-
bustor and an identification of Appendix VIII organics
that occur in the waste. Once identified, the Appendix
VIII organics are evaluated against various factors such
as quantity in the waste feed; structural category, such
as aliphatics, aromatics, and chlorinated aliphatics; tox-
icity; and incinerability ranking. This procedure estab-
lishes the  POHCs for  representing the worst-case
conditions of organic hazardous  compounds in  the
waste feed.
Because compounds such as chloroform, carbon tetra-
chloride, and  methane can be products of combustion,
a useful rule of thumb is "Don't choose a POHC that's a
PIC." Because the formation of such compounds ap-
pears to be  dominated by a quasi-equilibrium, their
presence in a waste cannot be reduced without radical-
ly changing the  combustor. (See Dellinger et al., 1988,
for information on  identifying compounds that could be
PICs.) If a given POHC compound forms in the com-
bustor or its gas path, its presence in the stack would
decrease its apparent ORE.  As a result, such com-
pounds  often are avoided. If, due to sampling and anal-
ysis,  compound availability,  or other constraints,  a
POHC must be selected that also can be a PIC, the
compound's concentration should  be spiked to levels
high enough to override the "PIC effect" on ORE.

The amount of each POHC fed to the combustor during
a trial burn is determined in part by the sensitivity of the
measurement method that will be used. The quantity
should be sufficient to enable the measurement method
to show 99.99  percent ORE (or 99.9999 percent  for
polychlorinated  biphenyls  [PCBsJ   or  dioxin-listed
wastes). The amount of the compound fed into the com-
bustor must be greater than 104 (106 for PCBs and diox-
in-listed  wastes) times the method's detection limit to
ensure that the test does show greater than 99.99 per-
cent ORE (99.9999 percent for PCB or dioxin-listed
wastes). The concentration of POHC also should be in-
dicative  of the maximum levels of POHC that the incin-
erator will routinely feed once permitted.

For example, if the sampling and analytical methods
used to  measure the emissions of a given POHC have
a detection limit of 8 g  emitted from the stack, then a
minimum of 8 x  104 g (80 kg) of that POHC must be fed
into the  combustor for each run of the trial burn to es-
tablish a ORE of 99.99 percent for POHC, or 8 x 106 mg
(8,000 kg) to establish 99.9999 percent ORE for PCBs
or dioxin-listed wastes. Because the sampling train only
removes a small fraction of the total gas emitted, the
determination must be based on the minimum sensitiv-
ity of the combined sampling and analytical methods for
measuring the combustor's total emissions, rather than
exclusively on the sensitivity of the analytical methods.

Finally, the regulations state that a compound is more
likely to be chosen as a POHC if it is found in the waste
feed at high concentrations. Incinerator and BIF guid-
ance  documents  also suggest looking at  compound
structure so that the various structural classes of com-
pounds to be found in the wastes are  represented by
structurally similar POHCs. Toxicity also should be con-
sidered; the desire to ensure that a particularly toxic
compound is destroyed must be balanced with any
safety concerns associated with testing and handling.

4.3.2.2  Incinerabiiity Ranking

Incinerability ranking is a concept that  was developed
for comparing the difficulty of destroying various organ-
ic compounds.  The ranking  affords  the  combustor
owner/operator the flexibility to burn wastes that are
less difficult to destroy than those tested. Without this
flexibility, combustor operation might be limited to the
specific compounds that were burned  during the trial
burn.  In this  regard, the RCRA regulations (40 CFR
section 170.62(b)(4)) require that the:
  Director will specify as trial Principal Organic Haz-
  ardous Constituents  (POHCs) . . . based on his
  estimate of the difficulty of incineration of the con-
  stituents identified in the waste analysis, their
  concentration or mass in the waste feed, and for
  wastes  listed in Part 261, Subpart D  (listed
  wastes) the hazardous waste organic constituent
  or constituents identified in Appendix VIII of that
  part as the basis for listing.
A number of  incinerability  rankings have  been pro-
posed, and any one might be appropriate for a given
application. The objective of each ranking is to correlate
a measurable property of the compound to its incinera-
bility—how readily it can be destroyed  in a combustor.
Each  ranking is based on a different property, such as
the  compound's heat of combustion or  how readily it is
destroyed under substoichiometric oxygen  conditions
(TSLoO2). The difficulty in using the ranking lies in the
fact that while each ranking method can be tied to a
specific destruction mechanism, any properly operating
combustor subjects the waste to a combination  of de-
struction mechanisms. As a result, each ranking system
lists specific compounds in somewhat different order.
As discussed in EPA guidance documents (refer to per-
mit  conditions guidance  and BIF implementation guid-
ance), no single ranking  system is recommended. EPA
is continuing its investigation of the various ranking sys-
tems, but has offered the following general guidelines:
A variety of considerations should be  used to select
                                                  25

-------
 POHCs, including  incinerability  rankings, compound
 structure (i.e., choose compounds structurally similar to
 those in the waste), concentration,  and toxicity. (Bal-
 ance the desire to show directly that a particularly toxic
 component is destroyed versus the need to guarantee
 the safety of handling and emissions during the trial
 burn.)

 An element of subjectivity is associated with the POHC
 selection process, because the advantages of  a rela-
 tively simple test (i.e., only a few POHCs) must be bal-
 anced against the need to select POHCs that represent
 the range of hazardous organics  likely to be burned in
 the combustor. (For additional guidance, see U.S. EPA,
 1992, 1989a,b.)

 In applying the incinerabiiity ranking concept, the incin-
 erability of compounds in the waste is compared, and
 the most difficult to incinerate organics are used for the
 trial burn. In many cases, the most difficult compounds
 to incinerate from different rankings are tested in the
 trial burn, and POHCs also are chosen based on other
 considerations such as compound structure. Typically,
 when this has been done, the permit places no restric-
 tions on the types of organics that can be burned in the
 combustor, except that the facility  can  burn  only the
 wastes and constituents identified in the trial burn plan
 and considered in the POHC selection process.

 4.3.3  Setting Limits for Parameters That
        Affect Organic Compound Destruction

 4.3.3.1  Group A and Group B Parameters

 The regulations  covering  the burning  of  hazardous
 waste require that the permit specify a minimum tem-
 perature at which an incinerator must be operated and
 that suitable interlocks be provided to cut off the haz-
 ardous waste feed to a combustor if the temperature
 within the combustion chamber drops below the speci-
 fied limit (see 40 CFR  section 264.345(f)). The mini-
 mum temperature must be one of the following:

 *  The lowest mean temperature at which a successful
   test (i.e., minimum of three runs) occurred, or

 •  A rolling average  minimum temperature limit.

When establishing these limits, the minimum tempera-
tures for both the primary combustion chamber (PCC)
and the secondary combustion chamber (SCC) must
be determined from the same test, even though it is
possible for the SCC to achieve its minimum tempera-
ture during one test while the PCC achieves its mini-
mum during another.

All waste feeds do not need to be cut off when the tem-
perature in only one chamber drops below the mini-
mum.  (The  recommended  guidelines  for  specifying
 when to stop the hazardous waste feed to a combustor
 are shown in Table 4-5.) Rather, the hazardous waste
 feed to both chambers should be stopped if the temper-
 ature in the SCC drops below the cutoff value. If only
 the PCC temperature drops below the cutoff value, con-
 tinuing to feed hazardous waste  to the SCC while stop-
 ping the feed to the PCC may be appropriate.
 Table 4-5. Recommendations for Minimum Temperature Waste
         Feed Cutoff
 Gas Temperature Below
 Minimum in:	Waste Feed Cutoff in:
 SCC only

 PCC only
 SCC and PCC
Both PCC and SCC

PCC only
Both PCC and SCC
 Note: Does not apply to maximum temperature.
 Source: U.S. EPA, 1989a.
 As discussed above, carbon monoxide is regulated to
 ensure proper mixing of oxygen and the organic constit-
 uents in the combustion chamber and thus to inhibit the
 formation of PICs. The limit on carbon monoxide, which
 now is set on the basis of regulation and guidance rath-
 er than the trial burn, is established as follows:

 • Carbon monoxide must be ^ 100 ppmv dry, correct-
   ed to 7 percent oxygen  (either instantaneous or
   hourly rolling average), or

 • Carbon monoxide is > 100 ppmv, and the limit is set
   on THC as follows (40 CFR section 266.104(f)), THC
   :£ 20 ppmv dry, calculated as propane, corrected to 7
   percent oxygen.

 The regulation at 40 CFR section 264.345(b)4 requires
 that "the permit [for a hazardous waste combustion de-
 vice] will specify acceptable operating limits" for "an
 appropriate indicator of combustion gas velocity." To
 satisfy this requirement, parameters that can be related
 to the  combustion  gas flow  rate must be measured.
 Upper  limits on flow need to be established for the fol-
 lowing  reasons:

 • To control the gas residence time in each combus-
   tion  chamber.

 • To control the gas throughput of the entire system so
   that back pressure is minimized at joints and seals
   (e.g., at the SCC inlet of a rotary kiln).

 • To control the gas flow through the APCE so that the
   equipment is not overloaded.

 While the measurement of "an indicator of gas velocity"
 is required by the RCRA regulations, it is not an inde-
 pendent variable. The flue gas flow rate is interrelated
with the temperature such that limits on gas flow rate
cannot be set without affecting the limits on combustion
                                                 26

-------
chamber temperature. As a result, the limit on combus-
tion gas flow rate must be set from the same trial burn
that serves as the basis for setting the permit condition
on minimum temperature.

The regulation at 40 CFR section 264.345(b) requires
that a limit be set on the rate at which hazardous waste
is fed to the combustor. The limits on this parameter
serve several purposes. First, they prevent overload of
the combustion chamber, which can reduce combustor
performance. If low heating value (LHV) waste is added
to  the combustor too quickly, the waste may cool the
flame and  inhibit combustion.  Second, limits on the
waste feed rate keep the residence time above the min-
imum level required to destroy the POHCs. Because
the greater the fuel and waste feed, the higher the flue
gas flow rate, decreasing the fuel and waste feed rate
lowers the residence time. Also, limiting the waste feed
rate limits a Group C parameter somewhat—the mini-
mum  heat  released per unit volume. Further, limits to
the waste feed rates often are necessary to accommo-
date other parameters, such as chlorine or ash feed
rates.

Two types of limits  should be placed on the total waste
feed. The first is total waste feed per unit time for each
waste stream  or waste stream  type, such  as solid
wastes, aqueous wastes, or organic liquids. This limit is
based on  the average waste feed rate achieved over
time during the trial burn. The second factor that should
be controlled is the instantaneous waste feed rate. This
Group B parameter, which is referred to in Table 4-2 as
the maximum size of batches or containerized waste to
the PCC, has a close relationship to the maximum
waste feed rate. The instantaneous waste feed rate is
not significant if the waste is fed by nozzles, a continu-
ous conveyor, or a  screw feed.
The volatile content of a waste influences the rate at
which such a sudden release of emissions can occur.
For instance, an excessive amount of volatile material
in a waste will result in a rapid release of hydrocarbons,
the combustion of which will lead to a rapid increase in
the PCC pressure and possibly an increase in the car-
bon monoxide level.  These conditions could trigger a
shutdown.  Drum feeding of waste is particularly sus-
ceptible to such an occurrence. Consider, for example,
a combustor that is fed a 55-gallon drum of waste every
15 minutes in  addition to other wastes that are fed to
the combustor on  a continuous basis.  Because the
drum is cold when it reaches the kiln or grate, it will
quench the flame and the burning waste in the PCC.
Then, as the drum heats up, it will melt, if it is a metal
drum, or burn, if it is a fiber drum, and release flamma-
ble material, causing a rapid heat and gas release that
might overload both the combustion  air capacity of the
incinerator and the capacity of the downstream air han-
dling system. Such an event can result in puffing and
fugitive emissions at the kiln seafe or the joint between
the PCC and SCC or cause a significant change in the
temperature and the residence time of the gases in the
SCC. Ultimately, this change can affect the combustor's
destruction efficiency and, hence, the  ORE  for the
POHCs.

Such a scenario can be regulated by designing the trial
burn to match both the mean and instantaneous waste
feed rates for the combustor. For example, the permit
condition could  specify that a given waste stream feed
rate may not exceed 300 kg of waste per hour, with indi-
vidual batches not exceeding 30 kg of waste at no less
than 6-minute intervals. Other ways of specifying this
type of limit (e.g., by maximum drum size) can be used
depending on the unique requirements of a given sys-
tem. Although maximum volatile content is an important
factor, establishing it as a limit for containerized waste
is not  recommended as a permit  condition because
measuring  this  parameter is impractical during con-
tinued operation. Nonetheless, during the trial burn, op-
erators  should  not  inject containerized waste that
includes an amount  of volatiles equal to the greatest
amount anticipated during typical operation.

This approach,  however, would need to be modified to
account for different waste streams, especially  be-
cause a combustor that burns all the wastes at a fixed
feed ratio will  be the exception. When a  variety of
wastes  will  be  fed to the combustor,  the trial burn
should incorporate a combination of waste feeds to en-
sure sufficient operational flexibility.  Limits on the waste
feeds should be such that the combination of wastes
fed at any one time would result in a total heat release
rate in each chamber that matches the conditions in the
test. Although the waste feed  limits should not allow
data from a test burning of one set of wastes to be used
for a different category of wastes, the limits can be used
to allow a certain amount of flexibility in the waste flows.
Sometimes wastes can be fed in different combinations
of feed  rates as long as a cap on total thermal input is
not exceeded.  In this case, the high  heating value
(HHV) of waste does not need to be monitored.
During operation, the heating value of the waste typical-
ly does not have to be tracked with a high degree of
accuracy to adhere to the variations in the waste feeds.
Typically, the operator will feed the waste with the low-
est heating value first, then control the temperature at
the combustion  chamber outlet by varying the feed rate
of wastes with higher heating values. If the temperature
cannot be maintained in this  manner, the operator can
either lower the  feed rate of wastes with a lower heating
value, use supplemental fuel, or vary the air feed rate.
This procedure  will translate into a reasonably constant
heat release rate and, under most circumstances, a
reasonably constant flue gas flow rate.
                                                  27

-------
 4.3.3.2 Group C Parameters

 Group C parameters generally are set independently of
 trial burn conditions, are not required to be monitored
 continuously,  and are not interlocked with the waste
 feed cutoff. Most limits are set according  to accepted
 operating practices (i.e., the equipment manufacturer's
 specifications).

 Burner settings are the primary Group C parameters
 that affect organics emissions. In liquid injection and
 afterburner  chambers particularly,  this  parameter
 should be consistent with manufacturer design and op-
 erating specifications. Settings also should take into
 account the ability of a burner to properly  atomize the
 liquid waste and to promote efficient mixing. Specifica-
 tions vary according to the waste burned, burner and
 nozzle type, and method of atomization. To restrict the
 operation of these burners to trial burn settings could
 overly constrain the operation of the facility and limit the
 types of wastes that can be incinerated (or overly com-
 plicate trial burn procedures) without a significant bene-
 fit  with regard  to  minimizing emissions. The permit
 should allow sufficient operational flexibility in  waste
 viscosity,  burner pressures,  and turndown  limits, as
 long as these settings are compatible with burner man-
 ufacturer recommendations.
 Additionally, a minimum waste heating value should be
 set in a permit for burners providing  100 percent of the
 heat input to a  liquid or afterburner chamber. A liquid
 waste with a high heating value of 5,000 Btu/lb should
 be sufficient to maintain a stable flame consistent with
 good operating practices.

 Although APCE-inlet temperature primarily is set ac-
 cording to  its effect on particulate emission,  the pro-
 posed temperature should be evaluated with respect to
 dioxin formation criteria (U.S. EPA, 1992).

 4.4   Particulate and Metals Emissions

 While the RCRA regulations limit particulate emissions
 from combustors to less than 0.08 grains per dry stan-
 dard cubic foot (gr/dscf) corrected to 7 percent oxygen,
 Individual states may enforce more stringent standards.
 Particulate  emissions  are  considered any solid and
 condensible matter at  standard conditions emitted to
 the atmosphere. Such  emissions from combustion are
 composed of varying amounts of soot, unburned drop-
 lets of waste or fuel, and ash. Soot contains unburned
 carbonaceous residue, consisting in part of the  high
 molecular weight portion of polynuclear organic materi-
 als (POM). POM can condense onto other  particles of
 soot or other particulate matter.
As  discussed in Chapter 5, toxic metals emitted from
the incineration of metal-bearing wastes now are regu-
 lated using a risk-based approach. The concepts that
 apply to toxic metal emissions from burning hazardous
 waste also apply to the incineration of mixed wastes
 (i.e., combined hazardous and radioactive wastes,
 which can result in the emission of fine particles con-
 taining radioactive isotopes). In both cases, the opera-
 tional problem associated with the resulting particulates
 is that they tend to be very fine, less than 1 micron (urn)
 in diameter, and thus more likely to cause health risks
 because they enter the respiratory system more readily
 than coarser  particulates. In addition, special design
 considerations are required for APCE to enable it to
 capture submicron particulates.

 The formation of particulate matter in a combustor is
 closely related to the physical and chemical character-
 istics of the wastes  and fuels, as well as combustion
 aerodynamics, the mechanisms of waste/fuel/air mix-
 ing, and the effects of these factors on combustion gas
 temperature-time history (U.S. EPA, 1992).  Particulate
 matter can form by three fundamental mechanisms:

 • Particulate  entrainment

 • Volatilization/Condensation
 • Abrasion

 The first two mechanisms occur in the combustion of
 both liquids and solids; the third is characteristic of the
 combustion of solids only.

 When  ash-containing  liquid  wastes  and  fuels  are
 burned,  the organics are destroyed and the inorganic
 ash remains behind. The size of the resulting particu-
 late matter is a function of the concentration of the inor-
 ganics in the fuel, the size of the droplets formed by the
 nozzle, and physical  characteristics of the ash. When a
 droplet of atomized liquid leaves the nozzle and reach-
 es the combustion chamber, the organic portions va-
 porize and burn, leaving the ash behind in suspension.
 Depending on the composition of the ash, a fraction
 may volatilize as well (Barton et al.,  1988). Regardless,
 nearly the entire inorganic fraction typically is entrained
 by the gas flow.

 Another source of  particulate matter entrained in flue
 gas is the dissolved salts in quench water and, on occa-
 sion, in the scrubber. Particulate matter forms in the
 quench when  the water is evaporated by the hot flue
 gases; the solids dissolved and suspended in the water
 can form particulate matter. In the scrubber, particulate
 formation is less common but can occur if the quench
 does not cool the flue gas completely or if gas velocities
 in the scrubber are high enough to entrain liquid drop-
 lets. The salts, such as sodium chloride, can escape to
the flue gas along  with entrained mist. Although mist
elimination equipment provides a measure of control,
some of the salts  can  escape to  the stack and be
measured as  particulates by conventional sampling
methods. While the contribution of these salts to the
                                                  28

-------
overall paniculate loading is not large, this contribution
could result in a failure to meet paniculate emission
standards.

The particulate matter formed by volatilization/conden-
sation is especially relevant to the emission of metals
(U.S. EPA, 1992). Many metals and their salts will form
vapors at temperatures reached by the flame and in the
post-flame zones of a combustion chamber. When the
vapors cool, they condense to form veiy fine particu-
lates (< 1 urn in  diameter), which tend to be relatively
difficult to capture in APCE. Volatilization of metals and
other inorganics  can occur whether the waste or fuel is
a solid or a liquid; as long as the waste or fuel contains
an inorganic fraction, metals and other inorganics  may
volatilize. This phenomenon is of special concern when
the inorganic fraction includes toxic metals (e.g., cad-
mium and chromium) or their salts.
Abrasion, which  is the simplest mechanism, forms par-
ticles via a straightforward action that involves the solid
mass rubbing against itself and against the walls of the
combustion chamber. Because abrasion tends to form
coarse particles, which can be removed readily by a
reasonably well-designed air pollution control device
(APCD), it is not  usually a primary concern when evalu-
ating combustors.

The size of the entrained particles can range from less
than 1 urn to over 50 urn in diameter, but the particles
typically  do not exceed 20 urn  in diameter (Petersen,
1984; Goldstein and  Siegmund, 1976).  Particle  size
distribution can affect the collection efficiency of APCE
significantly and, consequently, the ability of combus-
tors  to meet RCRA paniculate  standards. Generally,
particles less than 1 urn in diameter are more difficult to
collect, requiring higher energy control devices.
The permit limit for particulates is 0.08 gr/dscf adjusted
to 7 percent  oxygen.  The limits for toxic  metals are
based on health  and environmental considerations and
are given as a maximum allowable contribution by the
source to the ambient concentration for  that metal
(Table 4-6).

4.4.1   Establishing Emission Limits for Metals

In Table 4-6, regulated metals are listed in two catego-
ries: carcinogenic metals, which are regulated cumula-
tively to achieve  a combined incremental cancer risk of
£10'5; and noncarcinogenic metals, which are regulat-
ed individually. The  10 noncarcinogenic constituents
are regulated through their reference air concentration
(RAC), which is the maximum annual average concen-
tration of the constituent to which a hypothetical maxi-
mally exposed individual (MEI) may be exposed.  The
RAC has been determined based on health consider-
ations. The level of allowable emissions of that metal is
derived by dividing the RAC by the dilution factor, which
Table 4-6. Risk-Specific Doses (RSDs) for Carcinogenic Metals
         and Reference Air Concentrations (RACs) for
         Noncarcinogenic Metals, HCI, and CI2
BIF-Regulated Carcinogenic Metals
Arsenic
Beryllium
Cadmium
Chromium (hexavalent)
BIF-Regulated Constituents
Antimony
Barium
Lead
Mercury
Nickel
Selenium
Silver
Thallium (oxide)
Hydrogen chloride (HCI)
Chlorine gas (CI2)
RSD (jig/m3)
2.3X10-3
4.1X10-3
5.5 x10'3
8.3 x 10-4
RAC (jig/m3)
0.3
50.0
0.09
0.08
20
4.0
3.0
0.3
7.0
0.4
Note: Ambient concentration limits for HCI and CI2 gas are high
enough that they will not affect most incinerators or BIFs. If in
specific cases these limits might be significant, then their limits can
be determined by using the same procedure used for the noncarci-
nogenic metals.
Source: U.S. EPA, 1992.

is calculated using  dispersion modeling and a disper-
sion factor. (See Section 4.4.3 for a discussion of dis-
persion modeling.)

The carcinogenic metals are regulated through their
risk-specific dose (RSD) in a manner analogous to that
used  for setting the limits on the noncarcinogenic met-
als. The RSD is defined as the annual average concen-
tration for the constituent that will result in a 10'5 risk of
cancer to the MEI.  The risk is assumed to be propor-
tional to the metal's ambient concentration.  Moreover,
the risk from all of the carcinogenic compounds must be
added and the total must be below the requisite cancer
risk level of 10'5.

The regulations allow a facility to determine the maxi-
mum  allowable feed rate for each of the 12 regulated
metals by one of four different methods: the Tier I, Ad-
justed Tier I, Tier II, or Tier  III approach. The tiers  are
described  in the BIF  regulations, in the BIF guidance
document (U.S. EPA,  1992), and in Section 5.2.3 of this
report; they are summarized in Table 4-7. The differenc-
es among the tiers relate to the following two concepts:

• Whether site-specific dispersion modeling  is per-
   formed or highly conservative look-up tables are used.

• Whether credit is taken for the metal's capture in  the
   ash and/or in the APCE (in which case a trial burn
   must be performed), or all of  the regulated metal is
   assumed to be emitted from the stack (in which case
   this metal's emissions do not have to be  measured
   during the trial burn).
                                                    29

-------
 Table 4-7. Definition of Tiers

Tier 1 (most
conservative)
Adjusted Tier I
T}or II
Tier 111 (least
conservative)
Method of
Dispersion
Estimation
Look-up tables
Site specific
Look-up tables
Site specific
Credit Taken for
Metals Capture in
Ash (Partitioning)
and APCE
No
No
Yes
Yes
 Note: For look-up tables, refer to BIF regulations (part 266,
 Appendices I, II, and III) (U.S. EPA, 1992).
 To calculate the allowable emission rate estimated to
 yield the maximum allowable ambient concentration for
 antimony, calculate:
 Allowable emission rate
             = RAG -r DF
             = Mg/m3 * (ug/m3)/(g/sec) = g/sec
              (g/sec)
                        11b
                      453.6 g
(3,600 sec/hr) = Ib/hr
            = 0.3-s-0.05 x-
                           1
                         453.6
-x 3,600
             • 47.6 Ib/hr
 where the RAG is 0.3 ug/m3. Assume for this example
 that dispersion modeling (see Section 4.4.3) for the site
 resulted in an annual average dispersion factor (DF) of
 0.05 ug/m3 for each g/sec of the metal emitted from the
 combustor's  stack. Then the allowable emission rate
 (47.6 Ib/hr) for antimony is determined by dividing the
 RAG value by the dispersion factor and correcting  for
 the units as shown.

 Once the allowable emission rate has  been estab-
 lished, the operator must calculate the amount of anti-
 mony that may be  fed  to  the furnace.  The permit
 condition is set on the maximum feed rate of the metal,
 which can be established in two  ways. The simplest
 approach  (using Adjusted Tier I) is to assume that all
 antimony fed to the furnace is emitted by the stack and
 that none of it partitions to the ash  or is captured by the
 APCE. With this method, the maximum allowable feed
 rate of antimony to the combustor would be equal to its
 allowable emission rate of 47.6 Ib/hr. Testing for antimo-
 ny during the trial burn is not  necessary in this case
because the emissions rate cannot be greater than the
antimony's feed rate.
 The second method for setting the permit condition on
 antimony is to feed the metal to the combustor at a giv-
 en level (presumably greater than the calculated maxi-
 mum emission level of 47.6 Ib/hr) during, the trial burn
 and measure the emissions. If at this higher feed rate
 the emissions are  less than 47.6 Ib/hr, the maximum
 allowable feed rate of antimony is established. This is
 the concept behind the Tier II and Tier III approaches.

 A dispersion factor of 0.05 is used in this discussion
 only for purposes of illustration. The actual value of a
 dispersion factor determined through  site-specific dis-
 persion modeling must be used if this procedure is fol-
 lowed.  Performing site-specific dispersion  modeling
 may be avoided in cases where the feed rates of regu-
 lated metals are very low. The BIF regulations (part
 266, Appendices I,  II, and III) provide  reference tables
 based on  conservative  dispersion  assumptions  that
 can be used to calculate allowable emissions (under
 the Tier II procedure) and health-based allowable feed
 rates for chlorine gas and hydrogen chloride.

 The allowable emission rate for the carcinogenic metals
 is calculated in a similar manner, except that the risk
 calculation is based on the fact that the total cumulative
 risk of all the carcinogenic metals cannot exceed the
 specified value of 10'5  under  the RCRA regulations.
 RSDs specified in  Table 4-6 are  based on a risk for
 each metal of 10'5 that is assumed to be linearly depen-
 dent on the concentration of the metal. By apportioning
 the risk among the metals, the applicant can reduce the
 allowable ambient levels for each metal. Further, if the
 desired risk based on state requirements is lower than
 10-5, the allowable concentration is reduced proportion-
 ately. For example, for a 10~6 cancer risk, each value is
 divided by 10.

 To illustrate, consider a facility that must meet a 10'5
 cumulative risk and that has made a determination,
 based on the composition of the wastes it receives, that
the cancer risk should be distributed as shown below.
The DF assumed for this calculation is the same as the
factor used above, 0.05.
                        Assumed risk distribution
                        (Established by
                        owner/operator):
                         As = 25%   Be=1%
                                                 Cd = 25%  Cr+^49%
                                                 £Risk=1x10-5
                        In this example for Cr+6,

                        Incremental cancer risk = 0.49 x 10'5
                                               (based on apportionment)
                        To calculate the contribution of chromium to the allow-
                        able emission rate that is estimated to yield the allowable
                        maximum ambient concentration for the MEI:
                                                  30

-------
Allowable emission rate for Cr+6
    = BSD (from Table 4-6) -r DF (from dispersion
      modeling)
    = 0.49 x (8.3 x 10"4 ug/m3) -r 0.05 (Mg/m3)/(g/sec)

    = 8.134 x 10-3 g/sec 4- 453.6 g/lb x 3,600 sec/hr

    = 0.0646 Ib/hr of Cr+6

A similar approach would be used for the other metals.

4.4.2   Setting Control Parameters for
        Particulate and Metals Emissions

Control parameters for limiting particulate and metals
emissions are divided between those that affect the for-
mation of particulate in the combustor and quench, and
those  that affect the capture of particulates in the
APCE:
Combustor and Quench Parameters
• Maximum combustion chamber temperature (metals
   only).
• Maximum combustion gas flow rate.
• Maximum halides feed rate (indirect parameter).

• Maximum ash feed rate.

APCE Parameters
• Operating conditions of the particulate air pollution
   control device.
• Minimum scrubber blowdown or  quality of scrubber
   liquids.
• Scrubber nozzle pressure (if applicable).
• Maximum APCE inlet temperature.
The APCE parameters are discussed in Section 4.10.
The primary consideration is that the parameters must
be related to the specific method of operation for the
particular control device(s) used. The other parameters
are set from the trial burn. The maximum combustion
chamber temperature is set as a permit condition be-
cause metals are more volatile at higher temperatures;
hence, more of them may be emitted as fine particu-
lates. For incinerators and industrial furnaces, the oper-
ator usually needs to perform one or more tests as part
of the trial burn to measure ORE and other related fac-
tors, then a separate test to measure metal emissions.
The DRE-related  factors  require minimum tempera-
tures,  while the metals emissions-related factors re-
quire  a maximum  temperature. (The recommended
method of setting  the  maximum temperature  is de-
scribed in Section 4.7.)
A limit on maximum combustion gas flow rate is set be-
cause a high flow rate is  likely to entrain particulates
and to tax most types of APCE. The maximum  limit is
set from the trial  burn during which particulates and
ORE are measured. (If ORE and metals are measured
in  separate tests, combustion gas velocity should be
the same for both  tests.)
The maximum halides feed rate limit is set primarily to
control hydrogen chloride emissions. This limit also has
an  indirect effect  on metals emissions, based on test
data showing maximum metals emissions under high
chlorine loadings  (U.S. EPA, 1989a). Thus, maximum
chlorine should be fed during the metals test.

Limits are set on  the  maximum  amount of ash  in the
waste that may be fed to a combustor to avoid emis-
sions of excessive particulate matter and overloading
of the APCE. For the purpose of this discussion, ash is
defined as any constituent of the waste that, when
properly burned, forms particulates. These can include
a number of inert materials (e.g., sand), dissolved com-
pounds (e.g., sodium  chloride or inorganic elements),
metals, and metal salts.

In general, the total amount of ash that may be burned
in a combustor is limited by specifying  the maximum
total ash feed rate at which an acceptable level of par-
ticulate emissions was achieved  during the trial burn.
Sometimes  restrictions must be placed  on specific
components of the ash content. For example, consider
the following hypothetical waste feed:
 Waste Stream

       A
       B
Ash Components       kg/h
Silicon dioxide         10
Sodium chloride         5
 Silicon  dioxide (as opposed to  siianes or silicones,
 which are  organosilicon compounds that can form a
 fine particulate fume) is not volatile under the condi-
 tions in  a typical combustor. Sodium chloride, however,
 can volatilize and form fine particulate. The trial burn
 can be assumed to demonstrate acceptable particulate
 releases at the maximum feed rates shown for these
 two compounds.
 In most cases, based on these tests, the total ash feed
 rate could be limited in the permit to 15 kg/h. If, howev-
 er, the amount of sodium chloride is increased and sili-
 con dioxide decreased during operation,  an increase
 could occur in the fine particulate loading to the APCE
 and in the total released particulates. If this release ap-
 pears to be of concern, the permit condition might spec-
 ify a total maximum sodium chloride feed  of 5 kg/h as
 well as  a limit on the total ash feed rate. If only waste
 stream  B is likely to contain sodium  chloride, that limit
 could be converted to a maximum feed rate and sodium
 chloride concentration for  that stream. While this con-
 version  reduces an operator's flexibility to blend
 wastes, it also reduces the need for waste analysis dur-
 ing operation and makes the limit easier to enforce.
                                                   31

-------
 The following categories of material typically form fine
 paniculate in a combustor:

 • Sodium salts, especially sodium chloride.
 • Volatile metals such as antimony, arsenic, chromi-
   um, lead, mercury, and tin.

 • Inorganics  with oxides that are volatile under the
   conditions of the combustion chamber, such as ar-
   senic trioxide.

 • Silicon organic compounds such as silanes  or sili-
   cones.

 Another important parameter related to  paniculate
 emissions is the APCE inlet gas temperature. Certain
 types of particulate, especially the fine particulate dis-
 cussed above, form in the incineration process as gas-
 es in the combustion zone and condense when the
 temperature decreases downstream.  The amount of
 condensed  particulate matter is a function of tempera-
 ture: as the temperature to the  inlet of the APCE in-
 creases, less  of this "condensible fraction" enters the
 APCE as particulate matter subject to collection.  Matter
 also can condense as particulates downstream of the
 APCE, typically in the stack. According to this scenario,
 as the inlet temperature to the APCE increases, particu-
 late emissions also would increase.  As  a result, the
 maximum APCE inlet temperature should be the maxi-
 mum measured during  the trial burn.  To protect the
 APCE from  damage from excessive temperatures, the
inlet temperature should not be higher than the manu-
facturer's maximum specification.

4.4.3   Dispersion Modeling

The concept behind dispersion calculations using the
Gaussian model (Turner, 1972) is illustrated in Figure
4-2. When a pollutant is emitted from a stack or other
source, it is carried by the wind. As it moves, it mixes
with the ambient air in both horizontal and vertical direc-
tions. The amount of mixing is a Gaussian (the same as
a normal distribution used in statistics) function related
to the distance from the source. The basis for the typi-
cal  method of performing a dispersion calculation is to
assume that the wind is moving along the direction of
the x-axis and that dispersion occurs along the y-axis
(horizontal) and the z-axis (vertical) as the wind contin-
ues downwind. The ellipses in the figure illustrate how
the plume spreads according to this model.

Inevitably, changes in the weather near the source will
affect wind speed and direction and cause shifts in the
stability classes. Under one set of conditions, however,
the pollutant concentration at one of the receptors will
be greater than at any other time for any of the other
receptors. This concentration is the maximum possible.

A particular receptor, however, is unlikely to register the
maximum concentration of the contaminant at any one
time, because each receptor is exposed to a range of
concentrations of pollutants from the source. The annu-
                                                                                           x,-y,o
Figure 4-2. Single source dispersion In three dimensions (Turner, 1972).

                                                  32

-------
 ticular year. Still, because one receptor will be exposed
 to the greatest annual mean concentration, the annual
 mean concentration at that receptor is the maximum
 annual mean concentration.

 Four mean concentrations (instantaneous, 24-hour, an-
 nual mean, and maximum annual mean) are important
 concepts to keep in mind when evaluating the impact of
 a given source on a population and on the surrounding
 environment.  Relevant data must be analyzed using
 statistical methods. When  evaluating the impact of a
 given source on a receptor or a series of receptors, an
 examination of meteorological conditions in the area is
 essential. For example, the frequency, as a percentage,
 with which the wind blows from the source (the combus-
 tor stack) to a given receptor, such as a hospital or
 school,  must be considered.  Data  on  the historical
 probability of wind directions, stability, and other meteo-
 rological conditions are available for this purpose. The
 GEMS User's Manual (U.S. EPA, 1979b) provides guid-
 ance for using such data and sources of additional in-
 formation. A person with experience and competence in
 performing dispersion modeling should be  consulted
 for answers to questions.
 The statistical nature of meteorological data is illustrat-
 ed in Figure 4-3 using variations in wind direction  and
 speed in a graphical representation called a wind rose.
 In a wind rose, wind direction is given  as one of 16
 points on the compass. The probability that the wind will
 blow in a particular direction is represented  by an arc
 placed along the radius within that point on the com-
 pass. The distance of the arc from  the center of the
 compass is proportional to the probability of the wind
W270
                                               090 E
blowing from that direction. A number placed by each
arc represents the annual mean velocity of the wind
from that direction.  The wind rose is a convenient
graphical method of showing annual wind conditions at
a location. It is especially useful for identifying factors
such as prevailing winds.

4.5   Hydrogen Chloride/Chlorine Gas
      Emissions

Halogens  often  are present  in  hazardous wastes,
where typically they take the primary chemical form of
halogenated hydrocarbons (e.g., carbon tetrachloride,
tetrachloroethanes, and fluoro-trichloromethane). Halo-
gens also can be introduced into a combustor in inor-
ganic form as salts (e.g., sodium chloride). Generally,
however, they are never found in hazardous waste in
their free state (e.g., chlorine, fluorine, bromine, and io-
dine). When  halogenated hydrocarbons  are burned,
the halogen is transformed primarily to  its vapor phase
acid form  (i.e., hydrogen chloride, hydrogen fluoride,
hydrogen  bromide, and hydrogen iodide),  but also
small quantities  of free  halogens can  be formed. Be-
cause  hydrogen  halides  are  much  more   easily
scrubbed from the combustion gas than free halogens,
the ability to  ascertain the potential emissions of  the
halogen in its free form is important.
The quantity of free halogens compared to halogens in
acid form  in the flue gas depends on  the amount of
available hydrogen in the feed and on the thermochem-
ical equilibrium between the acid and its free halogen
form. For example, a mixture  of 32 Ib of methane and
154 Ib of carbon tetrachloride  provides 8 mol of hydro-
gen for 4 mol of chlorine, or twice the number of moles
of hydrogen needed to react with chlorine to form  hy-
drogen chloride. For this scenario, the amount of free
chlorine formed  is negligible and can be  disregarded.
Consider the case, however, where only 1  mol of meth-
ane reacts with 2 mol of carbon tetrachloride. In such
cases, the amount of hydrogen is insufficient to com-
pletely react with chlorine to form hydrochloric acid. The
chemical equation for this reaction is given by

   CH4 + 2CCI4 + 4O2 + 15.03N2 =
     3CO2 + 4HCI + 2CI2 + O2 + 15.03N2        (4-1)

For cases in which sufficient hydrogen is present to
consume all feed chlorine, the relative amounts of hy-
drogen chloride and free chlorine can be  estimated
from an equilibrium calculation.  The thermochemical
equilibrium among water, chlorine, hydrogen chloride,
and oxygen is given by

                 K
   CI2 + H2O <	> 2HCI + 1/aO2           (4-2)

where  K equals the standard equilibrium constant.
  Figure 4-3. Illustration of a wind rose.
                                                   33

-------
 Fortunately, because a combustor that achieves condi-
 tions sufficient for adequate destruction efficiencies of
 organic compounds is unlikely to be operated with less
 than stoichiometric hydrogen present, the formation of
 large quantities of chlorine gas is highly unusual. None-
 theless, the possibility of this occurrence should be
 considered  when designing trial burns  because  the
 combustor might be operated at some point with syn-
 thetic wastes  and unusual  settings to achieve worst-
 case operating conditions. Because chlorine might not
 be removed completely from emissions by an acid gas
 scrubber and it is much more toxic than hydrogen chlo-
 ride, pains must be  taken to avoid a situation where it
 could form.
 Operational parameters that influence hydrogen chlo-
 ride emissions are (1) the amount of total chloride fed to
 the combustor, (2) the amount of total halides fed to the
 combustor, and  (3)  the operating parameters for  the
 acid gas APCE. The permit condition on the maximum
 amount of chlorides  that can be fed to the combustor is
 set based on the amount burned in the trial burn.
 The amount of acid-forming compounds (usually  ha-
 lides) burned in the combustor also will affect hydrogen
 chloride emissions by consuming the absorptive capac-
 ity of the APCE. As a result, the permit places upper
 limits on the amount of total halides (in addition to  the
 amount of chlorides) that can be burned. The maximum
 amount of acid-forming compounds that can be burned
 in the combustor during normal operation is based on
 the trial burn.
 Although no specific limit is set for chlorine emissions
 from combustors, the BIF regulations specify that BIF
 operators burning chlorinated  wastes must perform
 a risk  assessment  for both hydrogen chloride and
 chlorine emissions and establish limits using a Tier I,
 Adjusted Tier I, Tier II, or Tier III procedure, similar to
 the approach used for metals (Section 4.4.1). Measur-
 ing or estimating chlorine emissions in addition to  hy-
 drogen chloride emissions might  be  necessary to
 perform this type of analysis.
 Limits on operational parameters that influence hydro-
 gen chloride emissions do not need to be set for com-
 bustors that include a hydrogen chloride monitor that is
 interlocked to the appropriate waste feed cutoffs.

 4.6   Fugitive Emissions and Upsets

The  draft or pressure in the chambers of a hazardous
waste combustion device is regulated to minimize the
 release of partially burned POHCs and other untreated
products of  combustion  as fugitive emissions  and to
 minimize the opening of emergency vent stacks. Fugi-
tive emissions, which are regulated under 40 CFR sec-
tion 264.345(d), are of concern in multichamber sys-
tems but especially in rotary kiln systems that partially
degrade wastes into gaseous components in the PCC.
Such systems feed off-gases, which  contain large
amounts of POHCs, PICs, acids, and particulates, into
the SCC, where the PICs and  POHCs are destroyed,
then into the APCE, where pollutants are removed to
concentrations below the required level. The release of
these intermediate gases typically is  prevented by set-
ting limits on the maximum pressure at which the PCCs
and SCCs can operate.

In most systems, gases from the PCC are forced into
the SCC by the pressure differential between the two. If
the gas production rate in the PCC suddenly increases
or the draft in the SCC decreases, partially  burned
POHCs, PICs, particulates, and acid gases can be re-
leased. Increases in the pressure in the PCC can be
caused by an explosion or when a drum of exceptional-
ly flammable waste ignites. Decreased draft in the SCC
can be caused by a fan  failure or an increase in the
burning rate. Any condition that results  in the  sudden
release of more gas than the upstream system  can ac-
cept will result in an overpressure condition. The por-
tion of the combustion system  between the PCC  and
SCC typically is equipped with seals that can  contain
the gas from a specified  level of overpressure. When
this level is  exceeded, however, untreated gases are
released.

Fugitive emissions are particularly common in rotary
kiln combustors, where the kiln must rotate against a
seal that separates it from the secondary chamber. Typ-
ically, these  units are  operated at a  sufficient  draft to
ensure that normal fluctuations in the burning rate will
not result in pressure above atmospheric conditions. In
addition, because the waste  and supplemental  fuel
guns are mounted in openings at the upstream end of
the kiln, an  overpressure is not likely to result in hot
gases "backfiring" past the guns—a  dangerous situa-
tion that indicates the unit is poorly designed or is being
improperly operated.

A relatively uncommon rotary kiln design is particularly
sensitive to  overpressures and the  resulting  fugitive
emissions. In the typical rotary kiln, the kiln enters the
SCC without any constrictions in the gas stream follow-
ing the rotating seal. If the kiln is attached to a seal
leading to a hot gas duct that is followed by an elbow, a
diameter reduction, or another restriction to the hot gas
flow between the two chambers, the likelihood of  fre-
quent overpressures increases dramatically.

For the majority of hazardous waste combustors,  fre-
quent fluctuations in pressure at the outlet of the PCC
usually indicate highly heterogeneous wastes that burn
                                                 34

-------
unevenly or periodic overfeeding of waste to the com-
bustor. If  the  overpressures are sufficiently  high  to
cause  fugitive emissions through the seals or other
openings between the PCC and SCC, they should be
considered an upset condition that requires cutting off
the hazardous waste feed to the PCC.

Factors to consider when setting limits on the PCC draft
or pressure include whether the combustor is designed
to operate under positive pressure and is  thus much
more likely to tolerate a short overpressure without re-
leasing fugitive emissions, or whether the combustor
operates under negative pressure and relies on draft
to contain the combustion gases. Recommended limits
for pressure in the PCC and SCC are summarized in
Table 4-8.
Table 4-8. Recommended Limits on System Pressure
                 Forced Draft        Induced Draft
                 (Positive Pressure)   (Negative Pressure)
Primary chamber
Time-averaged
pressure from trial
burn
Slightly below
atmospheric
Secondary chamber   Time-averaged      Always below PCC
                 pressure from trial
                 burn
4.6.1  Emergency Vent Stacks
A design feature of many hazardous waste combus-
tors, primarily solid waste incinerators, is an emergency
vent stack, which typically is located at the point in the
system where sudden overpressures might occur. Pos-
sible locations of vents include:

• The waste feed end of a kiln, if explosions in the solid
  waste being fed are a concern.
• The downstream  end of the primary combustion
  chamber,  if the system is prone to explosion (e.g., a
  furnace designed to  burn munition waste or rocket
  propellant—a popping furnace).

« Immediately after the SCC,  if a thermal siphon effect
  might cause a sudden overpressure.

• In the quench,  if a sudden release of steam might
  cause overpressure.

An emergency vent stack is a necessary component of
a combustor because without it the system could be
destroyed  during improper operation, endangering life,
property, and the  environment.  The emergency vent
stack typically is configured as a short, refractory-lined,
carbon-steel stack with  a caplike lid designed to open
when pressure at the chamber outlets exceeds a  pre-
determined level. The pressure limit can be set using
one of the following:
•  A weighted lid

•  A column of water that can be blown into a receiver

•  A metal membrane designed to rupture

To illustrate  the purpose of an emergency vent stack,
assume that a container holding 70 Ib (about 9 gallons)
of a high-Btu liquid (e.g., waste acetone) is fed into an
incinerator inadvertently. If the  combustion occurs sto-
ichiometrically (i.e., without excess air), the combustion
will generate approximately 1,200 ft3 of gas at a temper-
ature of about 3,350° F. If the material burns  in 30 sec-
onds, the net effect will be  an  instantaneous surge of
combustion  gases of approximately 144,000 actual cu-
bic feet per  minute (acfm).  A surge  of this magnitude
can  overwhelm the capacity of the air handling equip-
ment, resulting in release of combustion gas either via
openings in the combustion  chamber  (puffing)  or
through emergency vent stack opening (venting).

The emergency vent stack must open rapidly and be
failsafe.  Typically, it is triggered directly, rather than
through a control system, whenever a problem occurs
that jeopardizes the system's  integrity.  Whenever an
emergency vent stack opens, it releases flue gas that
has not passed through the APCE, or possibly even the
SCC. Such  releases can contain particulates, metals,
POHCs, and PICs. Regulation of emergency vent stack
releases reflects the concern  about such emissions.
Clarifications to the incinerator regulations published in
the Federal Register (55(82): 17,890) on April 27,1990,
state EPA's  position that "no emergency release stack
openings are allowed while hazardous waste is in the
combustor unless the applicant has demonstrated dur-
ing the trial burn that the performance standards of [40
CFRj section 264.343 will be met while an emergency
vent stack is being used." The regulations further state:
   During the opening of an emergency vent stack,
   emissions of metals and HCI could pose unac-
   ceptable health risk. In addition, if temperatures at
   the  inlet  to the  emergency vent stack are  not
   maintained at permit levels,  HC emissions could
   also pose substantial health risk. While it is under-
   stood that there  can  be mitigating circumstances
   which  require the use of emergency relief stacks,
   these instances must be minimized.
The regulations specify that the Preparedness, Preven-
tion, and Contingency Plan that is part of the permit ap-
plication must include details concerning measures the
applicant will take to prevent emergency vent stack re-
leases under all but the most critical situations. Such a
plan requires preparing a detailed fault  analysis that
identifies conditions that might cause the emergency
vent stack to open and showing how the conditions are
either avoided or, if avoidance is not possible, how the
probability of their occurrence  is minimized. Although
details on the development of such an analysis—called
                                                  35

-------
a fault tree analysis—are beyond the scope of this doc-
ument, a brief description of the approach is provided.

4.6.2   Fault Tree Analysis

The analysis begins with an examination of conditions
that might cause the emergency vent stack to open,
such as a loss of electrical power, a loss of induced
draft, a significant increase of pressure within the com-
bustion chamber due to unexpectedly rapid combustion
or an explosion, an extraordinarily high gas tempera-
ture within the various combustor system components,
or a loss of process control instrumentation. Any one of
these operational upsets could necessitate the immedi-
ate cutoff of all waste and fuel feeds. The next step is to
consider the causes of each condition to determine the
likelihood  of reducing its occurrence and the feasibility
of designing alternative systems into the process.
For example, the loss of electrical power to the fans that
produce the system's draft could be caused by:
• A failure in the local utility system.
• An overloading of the power lines leading to the fans.
• A short circuit in the motors driving the fans.
• A rapid release of gas from an  explosion or from
   rapid burning.
• A failure in the gas cooling system, which can result
   in hot  combustion gases destroying downstream
   heat-transfer surfaces, ducts, APCE, as well as the
   fans.
Each of these causes then would be examined to es-
tablish various corrective modes. For example, backup
circuits or motors for powering the fans might be  in-
stalled. Multiple fans operating in parallel fashion could
be added so that the failure of one fan would not knock
out the  entire system,  or steam ejectors that would
serve in a backup capacity could be installed.
As a further example, if the power to the plant fails fre-
quently or the wastes burned are prone to explosion,  an
estimate of  the frequency of occurrence should  be
developed. Mitigation measures (e.g., backup power
supplies, backup ejectors, changes in operating proce-
dures) should be factored into this analysis to demon-
strate that the appropriate steps have been taken  to
avoid the emergency vent stack opening unnecessarily.
Such an analysis yields data on the anticipated fre-
quency of emergency vent stack openings. The analy-
sis serves two purposes:  first, and  most important, it
formalizes the process by which the frequency of the
emergency vent stack opening is minimized; second, it
establishes a basis  for discussions with regulatory
agencies.
4.7  Hourly Rolling Average Limits vs.
      Instantaneous Limits

Both the BIF and incinerator guidance documents (U.S.
EPA, 1992,1989a) allow operational limits on combus-
tors to be set on the basis of either an hourly  rolling
average (HRA) or an instantaneous limit. An instanta-
neous limit, which is based on the mean measured val-
ue of the control parameter in question during the trial
burn, is easy to determine and monitor but is conserva-
tive to the extent that it may be overly restrictive for
many combustors. In such cases, the HRA limit should
be used.

The HRA is the average of the instantaneous measure-
ments taken over a past hour of operation. At every
minute in that hour, this average will have been comput-
ed by  dropping the minus 60 minute  measurement
(counting the  present time as  zero) and adding the
most recent measurement to the calculation.

For purposes of illustration, this discussion considers
the HRA as it is used to set a combustor's temperature
limit. Setting operational limits for control parameters
other than temperature is analogous.

The BIF regulations require that a maximum tempera-
ture be  set to limit the formation of metal fumes, while
both the incinerator and  BIF regulations require that a
minimum temperature limit be set as well for the pur-
pose of minimizing organics emissions. The incinerator
guidance recommends that the mean temperature of
the trial  burn be used as the HRA limit for the minimum
temperature permit condition. Because metals were not
regulated at the time the incinerator guidance  docu-
ment was issued, a maximum temperature limit was not
specified. Metals emissions from incinerators now are
regulated, so establishing a maximum temperature limit
is necessary. The general consensus at present is that
until formal guidance is issued for incinerators, the BIF
regulations should be followed.

The BIF regulations specify that the maximum tempera-
ture limit be set as the mean of the maximum HRA from
each trial burn. To illustrate, consider a compliance test
consisting of three runs.  The maximum values of the
HRA (not the same value as the maximum temperature
observed) during the runs were 1,900°F, 2,100°F, and
2,030°F. The limit on the maximum HRA temperature
would be the  mean of these three values, which  is
2,010°F.

The various methods that can be used to set minimum
and maximum temperature limits based on the regula-
tions and guidance are summarized in Table 4-9. While
the BIF guidance document says that the HRA limit on
minimum temperature should be set using the "mean of
the three highest HRA"  from the trial  burns (as de-
scribed  above), the BIF  regulations specify only that
                                                 36

-------
Table 4-9. Methods for Setting Minimum and Maximum
         Combustion Chamber Temperature Limits

Minimum Combustion Chamber Temperature (from ORE Test)
BIF guidance
Absolute limit
                 HRA limit
Time weighted average (T)
for test
Mean of lowest HRA from
each run
Incinerator guidance  Absolute limit    Average of mean T for
                              each run (i.e., {STmsan}/n)
                 HRA limit       Average pf mean T for
                              each run (i.e., {ZTmean}/n)

Maximum Combustion Chamber Temperature (from Metals Test)

BIF guidance       Absolute limit    Time weighted average (T)
                              for test
                 HRA limit       Mean of highest HRA from
                              each run
Incinerator guidance  No guidance limit given

Sources: U.S. EPA, 1992, 1989a.

this method be used for setting the maximum tempera-
ture. Thus, the permitting authority should be contacted
to determine the proper procedure for a given situation.

4.8    Indicators of Combustion Gas
       Velocity

Although combustion gas flow rate can be monitored in
several ways, the preferred method is a direct gas flow
monitor at the outlet of the SCC. In some systems, how-
ever, conditions such as high temperature, high particu-
late loading, and high acid  gas loading can limit the life
and  performance of the monitor. For such cases, a
practical alternative approach to monitoring gas flow in
the combustor is  to place  the  monitor just before the
stack,  even though this practice increases the likeli-
hood of introducing errors due to  air  infiltration  or
changes in the water content of the gas stream, which
can be difficult to predict. If this location  is chosen for
the combustion gas monitor, the permit  writer should
place constraints on the water content and the amount
of permissible air  infiltration upstream of the monitor to
maintain conditions consistent with those achieved dur-
ing the trial burn.

When neither of these  alternatives is practical, measur-
ing the flow rate of combustion air instead of combus-
tion gas may be possible. For a given temperature, the
flow rate of the combustion gas  (i.e., the products of
combustion) can  be approximated within reasonable
accuracy based on the flow rate of the combustion air
(i.e., the primary  and  secondary air being fed to  the
combustor). Typically the combustion air constitutes 95
percent or more of the  combustion gas. In many cases,
especially in forced-draft combustors, the primary and
secondary combustion air can  be  measured fairly easi-
ly. In this case, monitoring combustion air is an attrac-
tive alternative to monitoring combustion gas.
Limits on flue gas velocity set by the permit conditions
should be based on the maximum combustion gas flow
rate that was measured during the trial burn. This flow
rate measurement should be taken at the minimum ob-
served temperatures during the test to ensure that the
conditions include the lowest temperature and shortest
residence time  necessary to achieve acceptable com-
bustor performance. (For detailed guidance concerning
flue  gas measuring devices,  see  U.S.  EPA,  1981,
1989a.)

4.9   Control Parameters Not Subject to
      Regulation

The following parameters were considered but not se-
lected to serve as control parameters by which the reg-
ulations would limit the  release of pollutants from
combustion devices:

«  Minimum oxygen concentration.
•  Maximum gas volumetric flow rate, maximum veloci-
   ty, or minimum residence time in each combustion
   chamber.

•  Maximum volatile content of containerized waste.
•  Minimum total  heat  input to  each  combustion
   chamber.
•  Maximum kiln slope.
•  Maximum kiln rotational speed.
•  Minimum liquid flow to the Venturi scrubber.
The permit writer should be aware that imposing addi-
tional permit conditions on a hazardous waste combus-
tion device might  not  improve compliance, because
many parameters interrelate and each condition  has an
effect on operational flexibility. If the combustion device
is unique in some manner, however, some restriction on
parameters other than those specified by regulation
may be appropriate. The permit writer is urged to seek
assistance from the EPA's Office  of Solid Waste when
an unusual design appears to require setting additional
conditions on operating parameters.

4.9.1   Oxygen Concentration

While complete combustion of POHCs and PICs re-
quires the presence of sufficient oxygen, several strong
arguments can be made against setting minimum limits
on oxygen. The most persuasive argument is that pick-
ing one oxygen level that is satisfactory for the com-
bustion of a wide variety of wastes is difficult. While
determining a suitable worst-case feed to test oxygen
demand during  a trial burn may be theoretically possi-
                                                  37

-------
bla, doing so for facilities that burn a range of wastes is
extremely speculative because of a lack  of detailed
knowledge about mechanisms of waste destruction.
Other reasons for not setting minimum limits on oxygen
include:

•  Insufficient oxygen results in a rise in carbon monox-
   ide concentration. Because carbon monoxide is al-
   ready a permit parameter, a limit on oxygen would be
   redundant.

•  Continuously and reliably measuring oxygen  con-
   centration at combustion chamber outlet conditions
   is difficult; thus, oxygen measurements are normally
   made at the stack. Often, air leaks out between the
   combustion chamber outlet and the stack. The oxy-
   gen in this leakage air can mask oxygen deficiencies
   in the combustion chamber, thus limiting  or negating
   the value of such measurements.

•  Several combustion chambers are designed to oper-
   ate under oxygen-starved (pyrolytic) conditions with
   additional air  supplied in  downstream combustion
   equipment. Minimum oxygen requirements for these
   pyrolytic chambers would be inappropriate and un-
   enforceable.

4.9.2   Gas  Volumetric Flow Rate,  Velocity,
        and Residence Time in the Combustion
        Chamber
Maximum gas volumetric flow rate, maximum velocity,
or minimum residence time are far easier to monitor at
the stack than in the combustion chambers. Moreover,
such monitoring in the stack—or in the duct leading to
the APCE—correlates reasonably well with conditions
in each chamber (especially in the SCC). Thus, even
though monitoring actual conditions in the  chamber is
conceptually preferable, this was not chosen as a con-
trol parameter subject to regulation.

4.9.3   Volatile Content of Containerized Waste

Maximum volatile content of containerized waste was
not selected as a control parameter because measur-
ing this characteristic during system  operation  is diffi-
cult While the guidelines do not  recommend that the
maximum volatile content of the waste be set as a per-
mit condition, they do recognize  the undesirability of
feeding a combustor excessive amounts of  highly vola-
tile materials in  containers.  The  guidelines suggest
choosing a containerized  waste for the trial burn with
the largest amount of volatile compounds expected
during continuous operation.

4.9.4   Total Heat Input

The minimum heat input to each combustion chamber
was not selected as a control parameter because this
value is difficult to measure during routine combustor
operation. It also is specified already by the minimum
temperatures of each combustion chamber—if the heat
input to a combustion chamber is reduced, so is the
temperature, and if the heat input is reduced too much,
a temperature cutoff will occur. Thus, regulating both
parameters is not needed in  most cases.

4.9.5  Kiln Slope and Rotational Speed

Kiln slope and rotational speed also were not chosen as
control parameters.  The slope is fixed at the time  of
construction and cannot be changed (except, conceiv-
ably, in very unusual  designs) without rebuilding the
combustor, which would require a new permit or a mod-
ification of the  existing one. Although kiln rotational
speed has a major influence on the quality of ash, no
incinerator regulations address the quality of ash. Ash
quality is addressed by the land  disposal restrictions
and, if requested by the permittee, by the delisting pro-
cess. It may be appropriate in some cases to establish
a limit on maximum kiln rotational speed and ensure
proper implementation of the above requirements.

The  effect of kiln rotational speed on particulate emis-
sions is insignificant. While an increase in kiln rotation
will result in the "grinding" of ash and an increase in the
amount of ash released  into the  flue gas, this type  of
particulate-forming mechanism results in large particu-
lates that can be removed  easily with well-designed
APCE. In almost all cases, the kiln would have to rotate
much faster than prudent operation dictates to gener-
ate enough particulate matter to overload the APCE.
Thus, the argument for restricting kiln rotation rate for
this purpose is  not persuasive, unless combustor par-
ticulate emissions during the trial burn are close to the
maximum level  allowable.

4.9.6   Liquid Flow to the Venturi Scrubber

A final parameter considered but not chosen as a con-
trol value is the minimum  liquid flow  to the Venturi
scrubber;  minimum  liquid-to-gas (L7G)  ratio,  which  is
closely related,  was selected instead. Venturi scrubber
efficiency can be related to the L/G ratio and the veloc-
ity through the Venturi. Because the  pressure drop
across a Venturi is a function of the liquid and gas flow
and because the gas flow and velocity can be related to
the flue gas flow  rate, permit limits on the minimum
pressure drop across a Venturi and maximum flue gas
flow rate are sufficient. Thus, a limit does not need to be
set on the minimum liquid flow to the Venturi scrubber.

4.10  Air Pollution Control Equipment

The  APCE on a hazardous waste combustor removes
acid gases (commonly hydrogen  chloride) and particu-
lates. In all cases, the performance of the APCE is es-
                                                 38

-------
tablished during the trial burn, so the purpose of setting
permit conditions on the APCE is to ensure that the sys-
tem's removal efficiency during operation is  at least
equivalent to its  demonstrated  efficiency. A  permit's
conditions are set according to control parameters that
indicate whether the particular system is working prop-
erly. Because different types of APCDs  operate on dif-
ferent principles,  each  type of device requires that
permit conditions  be set according to the appropriate
control parameters.

For controlling hydrogen chloride emissions, monitors
provide  an alternative to establishing limits on  the acid
feed rate and the acid gas APCE parameters. Because
hydrogen chloride monitors measure the concentration
of hydrogen chloride in the effluent  gas stream, and
because the permit limits the combustion gas flow rate,
hydrogen chloride  emissions are effectively limited by
setting conditions on the readout of the  hydrogen chlo-
ride monitor, which must be interlocked  with the appro-
priate waste feeds.

Generally, hazardous waste combustors use APCDs in
combination. Such systems for incinerators typically in-
clude a  quench for cooling combustor gases, followed
by acid gas control and paniculate control devices (Fig-
ure 4-4). In some combustion units, the quench is pre-
ceded by a heat recovery system; however, in boilers
used for steam generation, the quench is likely to  be
unnecessary. Similarly, if the unit is a furnace with  an
insignificant amount of halogens in its waste feed, the
APCE might not include an acid gas removal device.
Specific APCE varies according to the type and opera-
tion of the particular combustion unit. Conventional de-
vices (listed in Table 4-10 along with their respective
functions) are described below.
                                  Table 4-10. Types and Purposes of APCDs

                                  APCD                 HCI Control    Participate Control
Packed bed absorber
Venturi scrubber
Fabric filter
Dry scrubber
Electrostatic precipitator
Ionizing wet scrubber
Condensing scrubber
Ejector scrubber
Quench
X
X

X

X
X
X


X
X

X
X
X
X
X
                                  4.10.1  Quench and Scrubber Water
                                  Particulates can be produced in both the quench and
                                  the scrubber. The greater the amount of solids present
                                  in the quench and  scrubber water, the greater the po-
                                  tential  for  particulates  to be  produced.  From  the
                                  quench, particulates are produced when the water is
                                  evaporated and the dissolved solids are released; from
                                  the scrubber, particulates can be carried out with mist.
                                  The  operator can  control  the quality of the scrubber
                                  water by varying the fraction of water leaving the scrub-
                                  ber that is  recycled  and the fraction  being "blown
                                  down." To control the quality of the scrubber water in the
                                  reservoir, a fraction of it is sent to the sewer or other-
                                  wise discharged (the blowdown), and the remainder is
                                  recycled to the combustor. The larger the blowdown,
                                  the cleaner the scrubber and quench water tend to be.
                                  The permit limit on the acceptable degree of contami-
                                  nation for the scrubber and quench water is set by
              Fresh
              water
     1,800-
     2,200'F	

     From —
     incinerator
Quench
180-
220'F
                                      Sump
                                                              Paniculate
                                              Blowdown
Figure 4-4.  Generalized APCE schematic for a combustion device.

                                                   39

-------
specifying the minimum amount of blowdown  during
operation.

Obviously, blowdown is an adequate indicator of scrub-
ber and quench water quality only when the process
recycles the water. In cases where once-through water
enters the scrubber, blowdown rate has no meaning,
and no limit needs to be set for this parameter. If panic-
ulate matter is a problem, however, limits on dissolved
solids might need to be set.
The minimum blowdown rate for a combustor cannot be
determined easily from the trial burn. If the operator
starts the trial burn with clean water and then uses the
standard blowdown rate, the scrubber water will be
clean initially  and then become contaminated  with
dissolved and  suspended solids.  Thus, the trial burn
can show a satisfactory level  of particulate removal
as a transient phenomenon associated with the trial
technique.
One method of reducing the probability of such  an oc-
currence is to design the trial burn so that the system is
operated for a sufficient time before the test to ensure
that the quality of the water in the sump has reached a
steady state. For example, the applicant could  be re-
quired to refrain  from cleaning the sump or changing
the water during a specified pretest period. Another ap-
proach would be to specify the blowdown rate such that
the combined dissolved and suspended solids in the
scrubber and quench water pond or sump  do not ex-
ceed the mean determined during the successful trial
burn with the highest solids in the quench and scrubber
water.

A third alternative for determining the operational quali-
ty of the scrubber and quench water is to measure its
density, because the salts that dissolve in the water will
increase the density and can be monitored. The read-
out of the density monitor can be used to control the
amount of blowdown and/or the addition of fresh water.
If the readout is correlated with the dissolved solids
content of the aqueous stream, then the permit limit can
be set based on the solids monitoring rather than on the
scrubber blowdown.

4.10.2  Packed Column Absorber

The packed column absorber is  the single most com-
mon technology  in use for controlling acid  gas emis-
sions from combustors, even though it is not suitable for
particulate control.  The device (Figure 4-5) removes
acid by contacting the cooled combustion gases with an
                                          Flue gas out
                        Rue gas in
                            pH controller
                 Caustic
                 storage
&§&£ccz3§t!£&

.T

T *
SVi
XJL


JLLd

rYVVY-rrvY^ i

_ 5
JLl,
AJLI,
jOL

M»Mv
jt u
^Ili J
jC „
JL .
ALJ


VVWV*V b1


TTT
VrV
SJn
Jtj*




IjLJ^J .JL-3t.-3C J 31 jL J 3L J. J
NiiWlrMranrWiiWwn
»-

7

Water
tank
Le
cc
J— L.
"t—r


•JLi—. Makeup
>ve
nt
^.
O
^^
1
oiler
)
                                 Metering
                                 pump
                                                          Blowdown
                                              Sludge removal
  Figure 4-5. Schematic of a packed column absorber.
                                                 40

-------
aqueous stream.  The packing provides the surface
area needed for mass transfer between the gaseous
and liquid phases. Other types of columns contact the
liquid and gas in  a cocurrent or cross-current, as op-
posed to a countercurrent, flow pattern. In the cocurrent
flow configuration, both liquid and gas enter the column
at the top of the device and flow down together. In a
cross-current flow configuration, liquid flows down the
packing while the gas flows hprizontally across it. The
flexibility of design is possible because hydrogen chlo-
ride gas is readily soluble in water to very low pH. Thus,
a packed  column absorber requires only a few stages
or transfer units to achieve high hydrogen chloride  re-
moval efficiencies.

The key parameters on which permit conditions need to
be set to maintain hydrogen chloride removal efficiency
ensure that the absorber has a sufficient amount of
scrubbing liquid to properly wet the packing and to  ab-
sorb the acid. Additionally, the liquid feed system needs
to be monitored to ensure that sufficient liquid is being
distributed over the packing  to wet it thoroughly. Typi-
cally, permit conditions are set on the following  param-
eters:
•  Maximum flue gas flow rate, which  also is set for the
   purpose of maintaining combustion chamber resi-
   dence time.
•  Minimum L/G, which is the ratio of the scrubbing liq-
   uid flow (i.e., gallons per minute) to the flue gas flow
   (i.e., acfm).
•  pH of the scrubber's incoming liquid and effluent.
•  Pressure drop across the scrubber and demister,
   which is used to identify channeling, packing deterio-
   ration,  or solids accumulation. In  most cases,  this
   parameter does not need to be interlocked.
•  Nozzle pressure, if nozzles are used to distribute the
   liquid over the packing. This parameter is not neces-
   sary if the liquid is being distributed by a perforated
   plate  or  other  nonpressurized  system.  In most
   cases, this  parameter does  not  need to  be inter-
   locked.

4.10.3  Venturi Scrubber

A Venturi scrubber (Figure 4-6) can control both acid
gas and particulates, although acid gas removal is sup-
plemented with a packed column in some systems. The
Venturi scrubber consists of a  converging section of
duct followed by a diverging section.  Scrubbing liquid is
injected into the gas stream immediately before or at
the throat (the narrowest point) of the device.  As the
gas passes  into the converging section, it  accelerates
and the impact of the accelerated  gas and  the injected
liquid results in droplets capturing particulates and acid
gas. The smaller the droplets,  the larger  the surface
area of contact and the better the acid gas removal effi-
ciency. To a degree, paniculate removal efficiency is
also a function of droplet size.
Many variations of system design are possible.  For ex-
ample, a cone that moves back and forth can be placed
in the throat so that the pressure drop can be adjusted.
The removal efficiency for acid gases and  particulates
can  be correlated to the  pressure drop  across the
scrubber. For a given design, the higher the pressure
drop, the greater the removal  efficiency.  Varying the
throat size allows the operator  to maintain the scrub-
ber's removal efficiency  at lower gas flows. Because a
Venturi scrubber only collects acid gas and particulates
into an aerosol mist, it must be followed by a collection
device such as a demister to remove the aerosol.
Figure 4-6. Schematic of a Venturi scrubber (U.S. EPA, 1973).
                                                   41

-------
 Typically, permit conditions for a Venturi scrubber are
 set on the following parameters:

 •  Minimum L/G, which is the ratio of the scrubbing liq-
    uid flow (i.e., gallons per minute) to the flue gas flow
    (i.e., acfm).

 •  pH of the scrubber's incoming liquid and effluent, if
    the scrubber is used for hydrogen chloride removal.

 •  Change  in pressure drop across the scrubber and
    demister.

 •  Nozzle pressure, if nozzles are used to atomize the
    scrubbing liquid.

 4.10.4 Fabric Filter

 A fabric filter—a "bag house"—is one of the most effi-
 cient types of particulate APCDs and is in common use
 on incinerators, boilers, cement kilns, and other indus-
 trial furnaces (Figures 4-7 and  4-8). When  properly
 sized and operated, a fabric filter can achieve particu-
 late removal efficiencies of approximately 99  percent.
 Combined with a dry scrubber, a fabric filter also can be
 used to remove acid gases.

 Fabric filters consist of independent modules  contain-
 ing fabric bags that are hung such that dirty gas flows
 through them and is filtered. The modules are defined
 as having a "clean side"  and "dirty  side," referring to
 whether the side contains the filtered or unfiltered gas.
 On the dirty side of the bags, a filter cake forms, contrib-
 uting significantly to the efficiency of filtration. Depend-
 ing on  the design of the particular system,  gas flows
 from the outside to the inside of the bags and the filter
 cake is deposited on  the outside of the bag,  or  gas
 flows from the inside out and the filter cake is deposited
 on the inside of the bag.

 As filter cake builds up on the bags, the pressure drop
 across them increases. At a preset pressure drop, or at
 a preset time interval,  the bags are cleaned by one of
 various mechanisms,  e.g., shaking  them, sending a
 blast of air through them, or sounding a horn near them
 to create a sound wave that dislodges the particle cake.
 In all cases, the method of cleaning is built into the sys-
 tem. During cleaning, filter cake drops down into the fil-
 ter's hopper, from where it is removed subsequently for
 disposal.

 Some dust may bleed through the  bags during cleaning
 and be  discharged with the  cleaned  gas to the atmo-
 sphere. Such discharges can  be reduced in installa-
tions using multiple modules by cleaning the bags in
sequence and isolating each module from the system
during the cleaning. Systems with  multiple modules are
designed to accommodate brief downtimes for individu-
al modules.
         Top access hatches
                                         Diaphragm
                                         valves
                                            ,Air
                                            manifold
 Gas
 outlet
                                    Hoppers
  Fan
Figure 4-7. Schematic of a fabric filter (Richards and Quarles,
          1986).
 Shaker
 mechanism
 Outlet
 pipe
Clean
air side
Dusty
air side
                                                      Figure 4-8. Cross section of a fabric filter (U.S. EPA, 1973).
                                                   42

-------
The primary cause of failure in fabric filter systems is
bag rupture (i.e., tearing or bursting). A less-common
cause of failure is bleed-through of particulates. As a
result, the incinerator (U.S. EPA, 1989a) and BIF (U.S.
EPA,  1992) guidance  documents  recommend that a
permit condition be set on the minimum pressure drop
for  fabric filters because in the case of bag  rupture,
pressure drop across the filter will decrease.  In large
installations, however, the pressure change caused by
a small hole in a bag may not be measurable. A prefer-
able method for monitoring for bleed-throughs might be
to install specially designed dust sensors on the clean
side of the fabric filter that measure opacity or triboelec-
tric potential.

4.10.5  Dry Scrubber
A dry scrubber (Figure 4-9) collects acid gas directly
from the hot flue gas. The term dry scrubber is some-
                                I
Final collector
(bag house of
electrostatic
precipitator
 Figure 4-9. Schematic of the dry scrubbing process.

what of a misnomer because the system actually does
not operate in the dry state but by injecting a slurry of
lime (calcium hydroxide) into a chamber through which
the flue gas passes. Contact of the slurry with the acid-
laden flue gas occurs in suspension. The hot flue gas
then dries the slurry, and in the process hydrogen chlo-
ride and other acid gases are absorbed by the lime. Be-
cause of this drying action, a dry scrubber is also
sometimes called a spray dryer. Numerous dry scrub-
ber designs are  in use, some  of which  utilize  one or
more contacting chambers to optimize lime usage; oth-
ers simply inject the slurry into the duct (generally at a
wide point in the system): The suspended lime is car-
ried with the flue gas to the particulate collection device,
which is usually a fabric filter. Lime is collected along
with dust, and the lime in the filter cake removes addi-
tional acid gas.
Dry scrubbers typically do not achieve a high enough
removal efficiency of  hydrogen chloride to be used  as
stand-alone control devices. Rather, they typically are
used in conjunction with other acid gas removal devices
such as a packed column. Permit conditions usually are
set on the caustic feed rate (U.S. EPA, 1992) and the
gas flow rate. Alternatively, limits can be set based on
readouts from a hydrogen chloride monitor.

4.10.6 Electrostatic Precipitator

Electrostatic precipitators (ESPs) are used rarely in in-
cinerators but are a common component in boilers, ce-
ment kilns,  and other industrial furnaces.  Although
ESPs typically  are used to collect particulate  matter
only, the technology can be combined with an acid gas
collector for acid gas removal.

In ESPs (Figure 4-10), particles are electrically charged
while passing through  a strong, nonuniform electrical
field generated, by a transformer-rectifier set, which
supplies pulsed direct-current power to a set of small-
diameter electrodes suspended between grounded col-
lection plates. Corona discharges on these electrodes
generate electron flow, which in turn leads to the forma-
tion of negative ions as the electrons travel the lines of
the electrical field toward the grounded plates. The neg-
ative ions also continue  on the-field lines toward the
plates. Within the corona itself, positive ions form, and
these travel to the high-voltage electrodes.
Particle charging occurs by several different physical
mechanisms: for particles  larger than  0.5  urn,  field
charging is dominant; for particles smaller than 0.1 um,
diffusion charging  is dominant. In the range between
0.1 and 0.5 urn, both mechanisms contribute to modest
charging of the particles. Field charging occurs due to
the distortion of electrical field lines around the particles
and continues until the electrical charge on particle sur-
faces is sufficient to deflect the electrical field lines. For
this condition, termed saturation charge, the magnitude
of the charge is related to the square of the particle's
diameter.
Collection efficiency is a function of particle size,  elec-
tric field strength, particle resistivity, secondary current,
and construction factors such as plate area.  Collection
.efficiency increases rapidly as particle size increases.
Also, minimal improvements in collection efficiency for
very small (i.e., 0.1  to 0.2 urn) aerosols can result from
diffusion charging  related  to the  random  Brownian
movement of negative gas ions.  If an ion approaches
the surface of a particle, the ion can be captured by the
particle, causing the particle to become charged.
Because of the combined  effect of field charging and
diffusion charging,  collection efficiency versus particle
size curves for  electrostatic precipitators have the gen-
eral form shown in Figure 4-11. The shape of the curve
is important,, because hazardous  waste incinerators
can  have a large fraction of inlet dust in the submicron
range where performance is weakest.
                                                    43

-------
                       Rappers
             High tension
             support frame
                    Corona
                    wires ~
                   Collectim
                   plates
                      Shell
                                                                                    High tension
                                                                                    cable from
                                                                                    rectifier
                                                                                     Gas inlet
                                                                                     Hoppers
                                                                    Wire-tensioning
                                                                    weights
                                                              Hopper baffles
              Figure 4-10. Schematic of an ESP (U.S. EPA, 1973).
 Permit conditions on ESPs are set on minimum kilovolt
 amperes charged to the precipitator's plates, as mea-
 sured at the secondary side of the transformer, and on
 the maximum gas flow rate to the ESP.

 4.10.7 Ionizing Wet Scrubber
 An  ionizing wet scrubber (IWS), which is similar in ap-
 pearance to  an  arrangement in  which an ESP is fol-
Efficlency
(%)
       93 >
      94

      90

Particulata size  0.9
(microns)
                                   Plate area
                                   per unit gas flow
Figure 4-11.  Factors that affect ESP collection efficiency
           (Barton etal., 1983).
 lowed by a wet scrubber, is used to control both particu-
 late matter and acid gases. Along with its appearance,
 the operating principles of an ionizing wet scrubber are
 quite similar to those of a conventional electrostatic
 precipitator. The ionizing section ionizes the particulate
 matter, which then is collected in the packed tower sec-
 tion along with the acid gas. Because the particulate is
 collected on wet packing,  its resistivity is not a major
 factor. Permit conditions for an  IWS are the same as
 those for an ESP and a packed bed absorber.

 4.10.8  Condensing Scrubbers

 Two types of scrubbers—known as condensing scrub-
 bers but commonly referred to as Hydro-Sonics scrub-
 bers—are in wide use. One type uses a fan as a source
 of motive power (Figure 4-12), a second incorporates
 steam or compressed air that acts as an ejector (Figure
4-13). Both types of condensing scrubbers include a
cyclonic  pretreatment chamber, one or more converg-
 ing section nozzles for flue gas, a ring of liquor-spray
 nozzles around the flue gas  converging sections, a
gas-liquor mixing section, a long contact throat, and a
mist eliminator. Because the  technology relies on a
combination of particle condensation  growth and parti-
cle  impaction,  pressure drop and  scrubber perfor-
mance relationships apply.
                                                   44

-------
     Gas
   V inlet
 eWater injected
      Turbulent mixing
      'particulate wetted
   Subsonic
   nozzle
Free-jet mixing
              Agglomeration
 Figure 4-12. Schematic of a fan-driven condensing scrubber
           (tandem nozzle) (John Zink Co., 1988).
   Steam or
   compressed
                       Turbulent mixing
                       particulate wetted
                                   • ^Agglomeration
        /     Water injected"
Small ejector nozzle
(supersonic)
                  Free-jet mixing
Figure 4-13. Schematic of an ejector-driven condensing
          scrubber (John Zink Co., 1988).

In an ejector-driven scrubber, a compressed air nozzle
operating at supersonic velocities effects the initial at-
omization of scrubber liquor  and generates  suction.
Flue gas and atomized liquor then pass through a sub-
sonic nozzle, where a ring of  spray nozzles injects
a concurrent stream of liquor. Next, the flue gas is ac-
celerated in  a long throat, where particle growth by
condensation and particle capture by impaction occur.
Water droplets are collected in a low-pressure-drop cy-
clonic collector or in a demister vessel. If low-pressure
steam  is available from a waste heat boiler or other
source, it can be used instead of compressed air in the
supersonic nozzle.  Permit  conditions for condensing
scrubbers are similar to  those  for Venturi scrubbers,
with the following exceptions:
• The condensing scrubber also should have a limit
  set on the steam or compressed air-to-gas ratio, if
  either of these fluids is  used by the particular design.

• A condensing scrubber that includes an ejector will
  not have a pressure drop associated with it. Rather, it
  will cause a pressure increase in the flue gas. As a
  result, a minimum pressure drop limit has  no rele-
  vance for this design.

4.11 Trial Burn/Compliance Test Designs

This section contains a summary of an approach to de-
signing a trial burn, or compliance test, and a discus-
sion of the common sampling and analytical methods
used to measure emissions from BIFs and incinerators.
Descriptions are general in nature and are intended
only as background; specific information on various as-
pects of trial burns is available in the literature. (See the
references at the end of this chapter, especially U.S.
EPA, 1986, a multivolume work on sampling and analyt-
ical methods that is regularly expanded and updated to
reflect the  latest EPA procedures and is incorporated
into the RCRA regulations by reference.)

The trial burn serves two main purposes:

•  It demonstrates that the combustor  can meet all
   applicable  regulations.

•  It establishes the conditions  under which the com-
   bustor can meet the applicable regulations.

Assuming that the combustor is capable of meeting the
applicable  regulations, the trial  burn should use mea-
surement methods that are adequate to demonstrate
the requirements.  Methods have been developed for
the vast majority of measurements that are required to
be taken in the trial burn. If these methods and the ap-
propriate  quality  assurance/quality control (QA/QC)
procedures are observed, the trial burn should demon-
strate whether the unit can operate in compliance with
the regulations. Achieving the second purpose requires
that the combustor be operated during the test under
conditions that represent the worst case anticipated for
normal operation. If the combustor satisfies the regula-
tory, health, and safety demands under these worst-
case conditions, it will satisfy them under less severe
operation.

4.11.1  Trial Burn Operating Conditions
In most cases, the trial burn must be performed under
the most severe conditions that the combustor will en-
counter for waste destruction  and removal,  and for
POHC,  particulate,  hydrogen  chloride, and  metals
emissions. Thus, an appropriate set of operating condi-
tions must be developed. The worst-case conditions for
a combustor are defined by a series of limits (e.g., ab-
solute or rolling, maxima or minima) on the control pa-
rameters, many of which are reasonably independent
of one  another. Changes in parameters that  are not
likely to affect other parameters to a significant extent
can be set as extremes with little or no complication.

Certain parameters, however,  are highly interdepen-
dent, including:

•  PCC and SCC temperatures.
•  Flue gas velocity.

•  Waste feed rates.
•  Waste composition.

•  Oxygen or excess air (a permit condition set by some
   regulatory  agencies).
                                                   45

-------
 Values for these parameters must be determined by
 performing a series of iterative combustion calculations
 to identify waste compositions, waste feed rates, and
 air flow rates that will result in the desired temperature
 and flue gas flow rates. Extensive computation can be
 necessary if calculations are not performed in an order-
 ly sequence.

 When  both metals  emissions  and  ORE  compliance
 must be  demonstrated, the  trial burn must include at
 least two sets of tests:

 • At minimum temperature to demonstrate ORE—the
   ORE test.

 • At maximum  temperature to demonstrate metals
   emissions—the metals test.

 For each of these tests, waste combinations and oper-
 ating conditions must be selected that:

 • Maximize combustion gas flow rates.

 • Maximize feed rates for each category of waste as
   defined by the permit writers.

 • Minimize (ORE test) or maximize (metals test) the
   temperatures in each combustion chamber.

 • Maximize the chlorine feed rate. (This condition may
   make the ORE test easier to perform because the
   higher the chlorine content of organic  compounds,
   the lower their heating value and, hence, their flame
   temperature.  Conversely, this condition may pose a
   challenge for the metals test because the needed
   maximum  combustion chamber  temperatures are
   more difficult to achieve  with lower heating value
   organics.)

 • Maximize the ash feed rate.

 • Minimize oxygen  concentration  (required  by some
   regions and states).

 A trial burn plan should be submitted to the regula-
 tory agency for approval before the triai burn is per-
 formed. Key elements of a plan are presented in Table
 4-11 and expanded in the following list.

 • A detailed description of the  incinerator or BIF sys-
   tem—including the air pollution control equipment;
   waste receiving, blending, and feed system compo-
   nents; and all ancillary systems. While the descrip-
   tion does not need to include a full set of engineering
   specifications and drawings, such detail may be re-
   quired for certain portions of  the permit application.
   In general, information should be sufficient to show
   how the test plan relates to the system, because the
   description in part must show that the trial burn will
   represent worst-case operation of the system.

• A description of the waste. The description  should
   profile both wastes that will be burned during typical
 Table 4-11. Key Elements of a Trial Burn Plan

 Trial Burn Plan

  • Detailed engineering description of the combustor:
    - Manufacturer's name, model number.
    - Type.
    - Maximum design capacity.
    - Description of the feed system for the hazardous waste, fuel,
      and other feedstocks.
    - Capacity of hazardous waste feed systems.
    - Description of automatic waste feed cutoff system(s).
    - Description of stack gas monitoring and any pollution control
      monitoring systems.
  • Description of each feed stream and waste that will be burned dur-
    ing the trial burn and a discussion of how they represent the worst-
    case conditions for the combustor:
    - Heating value.
    - Source, composition, and chemical analysis, if possible.
    - Levels of antimony, arsenic, barium, beryllium, cadmium, chro-
      mium, lead, mercury, silver, thallium, total chlorine/chloride, and
      ash.
    - Viscosity or description of physical form.
    - Identification of organic hazardous constituents (40 CFR part
      261, Appendix VIII) present in the feed stream.
    - An  approximate quantification of 40 CFR part 261, Appendix
      VIII, hazardous constituents in the hazardous waste.
    - Description of blending procedures, if applicable, prior to firing.
  • Operating conditions during the trial burn, a discussion of how they
    represent the worst-case conditions for the combustor, proposed
    permit operating conditions, and anticipated results from these
    conditions.
  • Description of the air pollution control system, its operating condi-
    tions,  and a discussion of how the  test conditions represent the
    worst-case conditions for the combustor.
  • Test protocol:
    - Operating conditions for emission control equipment.
    - Sampling and monitoring procedures, equipment,  frequency,
      analytical procedures,  and proof that they will  satisfy the re-
      quirements of the tests.
    - Quality Assurance/Quality Control (QA/QC) Plan.
    - Test schedule.
    - Shutdown procedures  in the event of equipment malfunction,
      including hazardous waste feed cutoffs and emissions controls.
    - Identification of ranges of hazardous waste feed, feed rates of
      other fuels and feedstocks, and other parameters affecting the
      ability of the unit to meet emissions standards.
    - Other necessary information.	

 Source: Adapted from U.S. EPA, 1992.


   operation and wastes planned to be burned during
   the trial burn.  The description should show that the
   trial burn wastes are representative of wastes that
   will be burned in worst-case conditions.

•  A list  of operating conditions planned for the trial
   burn. The test conditions and  the plan  must  explain
   the parameters selected for establishing worst-case
   operating conditions.

•  A detailed sampling and analysis protocol describing
   the types of measurements made. All measurement
   methods should be defined, either by EPA guidance
   documents (U.S. EPA, 1989b; U.S. EPA, 1992; U.S.
                                                      46

-------
  EPA, 1990) or by the regulations themselves. The
  sampling and analysis plan must include a complete
  QA/QC plan consistent with the measurement meth-
  odology and must be sufficiently rigorous to demon-
  strate  that  the  methods  used  will  measure the
  desired parameters to at least the required level of
  accuracy, sensitivity, and reproducibility.

• A list of the proposed permit conditions that are the
  focus of the trial burn plan. Such a list is useful for
  focusing the plan and for facilitating discussion be-
  tween the applicant and the regulators. (A sample
  list that includes footnotes clarifying individual points
  is provided in Table 4-12.)

Once the target permit limits and the general operating
conditions for the combustion system have been devel-
oped, the specific conditions for the trial burn must be
established using the material-and-energy balance, as
shown in Table 4-13. Typically, the first step is to deter-
mine the proposed target temperatures of each com-
bustion chamber.  In a  multichamber  system, the
temperature of the SCC usually is determined first be-
cause  it is more critical than the temperature in the
PCC.  Maximum temperatures  are determined from
equipment limits and metals emission considerations;
minimum temperatures are set on the basis of operat-
ing experience or the system designer's recommenda-
tions for successfully destroying organic constituents.

The next step is to determine maximum gas flow rates,
coming as close as is practical to the capacity of the
fans, ducts, and APCE. The  most desirable maximum
gas flow rate for actual operation might be lower than
the theoretical maximum based on equipment capacity.
Because the permit will place an upper limit on the gas
flow rate, maximizing it will give the system's operators
as much flexibility as possible.
Waste feed rates for the trial burn also should be con-
sidered as a relatively inflexible condition. Because the
feed rates must meet the facility's operating needs, and
the maximum feed rates will be those  used in the trial
burn, the trial burn should be  performed at the maxi-
mum waste feed rates that are required for operation.

Although the permit  should set an upper limit on the
rate of feed to each combustion  chamber for each
waste category, many permit applications overcatego-
rize waste. Waste types are categorized as HHV liquid
waste (greater than  5,000 Btu/lb); LHV liquid wastes
(less than 5,000 Btu/lb); and solid wastes. Waste feed
rate is determined by waste type in combination with
feed mechanism and chamber type. Some examples of
waste categories are:
• HHV organic liquids via atomizing nozzles (PCC and
   SCC).
• Aqueous liquids via atomizing nozzles (PCC).
• Sludges via lance (PCC).
• Solids via bulk conveyor (PCC).

Heating values are maximized by varying the composi-
tion of each waste stream, typically by establishing the
type and amount of POHCs required and then adding
ash (e,.g., as a soil or as a fly ash). Heating values are
increased by adding fuel oil and decreased by replac-
ing a portion of the fuel oil with an oxygenated fuel,
such as methanol, or by adding water. Combustion cal-
culations are useful for identifying waste compositions
and  quantities that yield the required temperatures,
halogen inputs, and gas flow rates.

Operating conditions for the trial burn must include the
desired maximum or minimum values for the Group A
and B control parameters. For example, EPA guidance
(U.S. EPA, 1989a) specifies that the permit condition on
PCC and SCC outlet gas temperatures must be set at
the mean temperature measured during the trial burn.
Assuming that the mean is 1,800°F (±100°F), if the per-
mit condition is determined by the first, simpler condi-
tion,  the absolute minimum  operating temperature
specified by the permit condition would be 1,800°F. If
the incinerator is intended to burn hazardous waste at
a mean temperature of 1,800°F with a variation  of
±100°F, the trial burn would need to be conducted at a
lower mean temperature (e.g., 1,700°F), so that during
normal operation the typical 100°F temperature fluctua-
tion  would not result in frequent waste feed cutoffs.
Similarly, other parameters may need to be adjusted for
the trial burn to anticipate  operating conditions (e.g.,
lower excess air [oxygen concentration],  higher gas
flow rate, higher waste feed rates).

4.11.2 Conflicting Parameters
One  of the difficulties associated with  (designing a trial
burn is that some furnace or incinerator operating pa-
rameters that will provide worst-case conditions might
conflict. Examples include:
• Worst-case organics destruction (closely linked to
   ORE) is achieved at the lowest combustion chamber
   temperature. Worst-case metals emissions are con-
   sidered to be achievable at the highest combustion
   chamber temperature.

• Worst-case flue gas flow rate for demonstrating ORE
   is the maximum (minimum gas residence time), but
   many types of scrubbers  (e.g., Venturi scrubbers)
   achieve the lowest particulate removal efficiencies at
   minimum gas flows.
 A number of techniques are available for addressing
 the  difficulty posed  by conflicting parameters. While
 special mixtures of wastes can minimize conflicts be-
 tween certain types of parameters, a facility might need
 to conduct tests at two or more sets of  conditions under
                                                   47

-------
  Table 4-12.  Example of Target Permit Limits on Operating Parameters
Operating Mode
Parameter"
1a. Min. PCC temp. (°F)
1b. Max. PCC temp. ("Ff
1a. Min. SCO temp. (°F)
1b. Max. SCO temp. (0F)"
2. CO concentration
3. Max. flue gas flow rate (acfm)
4. Max. pressure in PCC
5. Max. waste feed rates (Ib/hr):
Pumpable waste to PCC
Non-pumpable waste to PCC
Pumpable waste to SCC
6. Applicable APCE parameters:
Fabric filter/baghouse
• Min. pressure change across baghouse (in.
• Max. baghouse inlet temperature (°F)e
Packed tower
• Min. L/G (ga!/1,000 ft3)'
• Min. pH of scrubbing liquid
7. Allowed POHCss
8. Max. halides feed rate
8. Max. ash feed rate (Ib/hr)
9. Max. size of batches or containerized waste
10. Max. scrubber liquid solids content
11. MinJmax. liquid pressure to scrubber nozzles
12a. Max. total heat input to PCC (Btu/hr)
12b. Max. total heat input to SCC (Btu/hr)
13. SCC burner settings:
• Max. waste viscosity (SSU)
• Min. waste higher heating value (Btu/Ib)
• Min. atomization steam pressure
* Max. burner turndown (%)
Parameter Group
A
A
A
A
A
A
A
A



A

WC)




B
B
B
B
B
C
C
C
C




Mode-1
Value From Test
DRE-1
Metals
DRE-1
Metals
100ppm Guidance0
DRE-1
e?ifies that th9 value shou!d be  set at less than atmospheric for a negative draft system and that the trial burn should
  verity that the system is capable of maintaining a negative draft under maximum gas flow conditions.

 *P'Su Efo/ll-9,8^ SP60'1193 tnts to be a GrouP c parameter; however, recent information on metals and on dioxin and dibenzofuran formation
  in tno APCE Indicates that this parameter needs to be continuously monitored and interlocked with the waste feed cutoff system Thus it is
  treated as a Group A parameter here. The value will  be set at the highest of the three average temperatures of the three trial burns.   '
 'The liquid flow for the UG ratio is set on the basis of the minimum liquid to gas flow ratio; the actual parameter that is interlocked with the
  automatic waste feed cutoff is liquid flow.

 oBased on the trial burn having been performed using worst-case POHCs under both the Thermal Stability (UDRI) and heat of combustion
  ranking schemes.

 Group A:   Determined from the trial burn, continuously monitored, and interlocked with the waste feed cutoff system. Interruption of waste
             feed is automatic when the specified limits are exceeded. The parameters are applicable to all facilities.
 Group B:   Determined from the trial burn but do not require continuous monitoring and are thus not interlocked with the waste feed cutoff
             systems. Operating records are nevertheless required to ensure that trial burn, or worst-case conditions, are not exceeded.
 Group C:   Limits on these parameters are set independently of the trial burn from either manufacturer's or design specifications.
Source: U.S. EPA, 1989a.
                                                              48

-------
Table 4-13.  Sample Material-and-Energy (M&E) Balance Calculation Output
 Vl
  2
  3
  4
  5
 .6
  7
  8
: 9
 10
 11
 12
 13
C      D
                                          J
                                              L
                                                         K       X
                                                                                                                             AF        All
                                Spreadsheet I, Haste feed rates and croposition
Heat of Forration
Streas
                                      E13
(caT/g-iole>
   Ib/hr
S
                               24,200 >
                              Ib/hr
                                                                   FC-H3
                 %
                                                    31,050 >
                                                    lo/hr
                                                                  I
                                                                              I
                     51,036
                     Ib/hr
                                                                                                    water
                Ib/hr
%       Ib/hr
                                                                                                                                       Fuel
                                                                                                                                    Natural Gas
                                                                                                                         HIKED SOLIDS
                                                                                               Ib/hr
                                                                                          Ib/hr
     iolid Haste  (PCC)      1,5
IffllBTULiq. (PCC)
High8TUli(|.(SCC)
Suppl. Fuel - gas
                          2,000
                          1,200
                             1
                        3,0*
                        l.«
                        3.B
                        0.0%
          24.0
           0.0
          36.0
           0.0
           0.0
                 8.0%
                 4.0%
                 0.0%
                 4.0%
                                                     120.0
                                                      32.0
                                                             1.B
                                                             O.IR
                                                             3.0%
                                                                      15.0
                                                                      8.0
                                                                      0.0
                                                                      36.0
                                                                      0.0
          16.0%    240.0
          54.0%    432.0
           0.0%      0.0
          54.$    648.0
                                                                  3.01
                                        45.0
                                         O.Q
                                                                           0.0
                                                                           0.0
                            33.W
                             0.0%
                            36.0%
                                                            0.0
                                                                                                                                      .0%
                                                                                                  0.0
                                                                                                  0.0
                                                                                                  1.5
                                                                                                                                     60.0%     900.0
                                                                                                                                                0.0
                                                                                                                                                0.0
                                                                                                                      0.01      0.0
Total  (lb/hr)-->
             M
             12
              1
             16
             14
             36
             32
             19
             31
             18
       	0
      B BASIS N
              C1
               S
               F
                          5,501
                          Ib/lir
                         6
                       33J
                         1
                        18
                                10.04%
                                 0.84%
D.00%
                                            24.24?      48
               71.72%     143
                               12.80%
                                                      56.80%
                                                       0.001
                                                      30.40%
                                                                       37.50%
                                                                       12.50%
                                                                                              0.00%
     Hotalar Height,
                          3,4
                     total moles
                                noles
       Forwla  C
       oriole  H
      fraction  0
     (rater as  N
     eleaents)  Cl
               S
               F
                    121.73
                    350.92
                     67.70
                      0.46
                      9.50
                      fi.03
                      0.94
                                      119.5

                                    soles
                                    per hr
liber   iBles
isles   per hr
          1.51
          1.51
   3      4.52
                                            1.51
                               Number
                               moles
                                                2     4.04
                                                4     8.08
                                                0     MO
                                                0     0.00
                                                2     4.04
                                                           0.00
                                                       2.02
     34
     0
     18
     0
     0
     0
  187.5

 moles
perhr
                                                                      Number
                                                                       noles
                                                          2
      0.63
      0.00
0     0.00
3     0.94
0     0.00
3 '    0.94
                                         0.31
                                                           32

                                                        Dies
                                                       per hr
                                                                                1

                                                                    71.90%       1
                                                                    22.95%       0
                                                                     0.00%       0
                                                                     5.15%       0
                                                                     0.00%       0
                                                                     0.00%       0
                                                                     0.00%       0
                                                                     0.00%       0
                                                                     0.00%       0
                                                                     0.00%       0
                                                                    (avg)    19.115
                                                                                                                        27.80%
                                                                                                                         3.70%
                                                                                                                        23.00%
                                                                                                                         0.70%
                                                                                                                         0.00%
                                                                                                                         0.10%
                                                                                                                                                          22.00%
                                                                                                                                                          22.70%
                                                                                                                                                                  33
                                                                                                                                                                 207
                                                                                                                                                                   6
                                                                                                                                                                   0
                                                                                                                                                                   1
                                                                                                                                                                   0
                                                                                                                                                                   0
                                                                                                                                                                 198
                                                                                            Uer    noles
                                                                                             noles    per hr
 41.25
165.00
 41.25
                                                                           0     0.1
                                                                                 41.25

                                                                                 8,584
                                                                                                                    rale      Dies
                                                                                                                  fraction    per hr
                                                                                                   5.00
                                                                                                   2.50
                                                                                                        0.00
                                                                           2.50
                                                                   tole
                                                                 fraction
                                                                    31.57%     53.36
                                                                    68.43%    115.67
                                                                     0.00%      0.00
                                                                     0.00%      0.00
                                                                     0.00%      0.00
                                                                                                                                           per hr
                                                                 ole      soles
                                                              fraction    per hr
                                                                                                      78.30%    0.34
                                                                                                       0.00%    OJ
                                                                                                       1.26%    0.01
                                                                                                       0.00%    0.00
                                                                                                       0.00%    0.00
                                                                                                                                                      18.69%     20.85'
                                                                                                                                                      49.57%     55.30
                                                                                                                                                      21.46%    '23.94
                                                                                                                                                       0.40%      0.45
                                                                                                                                                       0.00%      0.00
                                                                                                                                                       0.03%      0.03.
                                                                                                                                                       9.86%      0.00
                        (BIP)
                   P/lb-itole)
                   VALUE  (BTU/lb)
                   (BTU/lb-uole)
                                      1,179
                                    140,837
                                     78,243
                                      1,179
                                                                    87,949
                                                                       844
                                                         9,777
                                                                          (1,060)<
                                                                         (19,080))
                                                                         (10,600)>
                                                                                                                          18,450 <     950 BTU/scf
                                                                                                                    19,9
                                                                                                       1,050 BTU/scf
                                                                                                                                                                 (585)


                                                                                                                                                                    1
                                                                                     49

-------
Tabla4-13. (continued)
      *       «      AT
A?
AX
AZ
BB
BF      HI
3 immi
4
5
C
7
8
§
19
11 C02
12 m
13 Dei
K H2
15 02
IS S02
17 HF
is m
1902req'.d
29
21
22
23
24
25
26
27
21
2f
30
31
32
33 .
34
35
35
37
39
39
41
41
42
43
44
45
4S
4?
43
49
59
51 _»»
11
1 OS excess air
ioles/6r nole I
121.73 12.8!
170.69 17.9!!
9.59 1M
651.75 68.3X
0 O.ffl
0.023125 9.01!
0.94 0.1)!
0 Q.OI
173.25 -
953.68
Spreadsheet 2. Flue Gas Properties
Exit froa Coiiwstion Chaiter 8 Heat Exchanger
? i excess air-->
ACFH- 45,807
SCFH wles/hr
785
1,101
61
7,213
800
0
6
0
121.73
170.69
9.50
1,118.28
124.01
0.03
0.94
OJ
297.26
m
1 net 5! dry Ib/hr
7.« 8.9* 5,356
ll.Ci 0.(» 3,072
0.6X 0.7JS 347
72.4J 81.41 31,312
8.01! 9.01 3,968
O.OS 0.0)! 2
0.1! 0.1!! 0
O.Oi 0.09! 0
9,966.45 1,545.19 • 0.99 44,057
nuauKEiair- 28.5 IT
pew wins commons nw/nrgpor TABLE *mc«iH.ii?
IK THIS TABLE
* Smtter Eff {daw- 	 	
Solid Haste (PCC) (
High Bra Liq.(PCC) (
iHiDLiq. (PCC) (
Hi§li 81 liq.(SCC) (
Stjipl. Fwl - gas (
Stea (
E ojqp ifi SCC exit fa
!os6. Ctaber Jeff. oF-
Heat Excl. Tkml Duty
Jtet loss sr Boiler Ik
ToUl ispfit to kifi.feii
ACFH t Sec Exit 	

U*\. 	
Ib/hr) 	
Ib/iir) —
scfa) 	
fyhr) —
:t) 	
(BUflirJ-i
y "'
cl losses)
EM SOfflfflS (10«38Tl!/iir) 	
mini iHMtniini i! M" ***'**"**
Teip. Ideal Exd. Untie
Tap J pncfi outlet (oF
Kater Evaporated in Own
Sas flw rate leaving qu
Sas flw rate leaving qu
X loistiire 	 	 	
t(oF)-->
ch 	
ench (ACFH,
nek (SCFH,
I wisttire ? saturation 	

INPUT
> 99.55!
> 1,500
> 800
> .2,000
> '1,200
> 0.5
> 0
> 71.58t
8.03!
1,930
10.00&
1
45,807
CALCULATED
<
< 2,000
<
( 71.SB
: 1,930
b/hr
IH ^
23343.99
800 <
180 < 160
8,829 Ib/hr
Net) 	 	
let) 	







3.31E+06 BTU/hr)
8.77E+06 (BTU/hr)
O.OffitOfl (BTU/hr)
l:38Et07 (BTU/hr)
2.85E+04 (BTU/hr)
O.OOEtOO (BTll/hr)
by tei^erature
1,328 I!
O.OOE400
> 2.59E+06 BTU/hr)
2.59E+07 BTU/hr)
OUT iiiiiini
23344.01
niirmimrriiiiii
700 oK
356 oC
1,059 al/hr
15,855
13,130
32.5X
51.K
Exit of quench
SCFH * net
785 6.K
4,265 32.55!
61 0.551
7,213 54.9X
800 6.15!
0 O.OX
6 O.K
0 0.01
13,130.11
»- 26.0
Exit of scrubber
SCFH Suet .Ib/ftr
785 6.B 5,356
4,265 32.5X 11,901
0.31 * O.U 2
7,213 54.9X 31,312
800 6.11 3,968
0 Q.OS 2
0.03 * 0.0)! 0
0.00 * 0.« 0
13,063.07
«- 25.8
* Based on scrubber efficiency input
m M
(m/]\i) (BTU/lb)
2,208 2,856
10,956 12,090
0 0
11,527 12,704
950 BTU/scf 1,050 BTU/scf
0 1,060
AV - By feed (including losses), Spreadsheet 4ft
M by gas flw, fron spreadsheet 3
990 BTU/lb latent of heat tier* quench T
                                                           50

-------
the same operating mode. The first test run would be
used to set limits for all parameters, except those in
conflict.  The additional tests would be run for setting
limits on conflicting parameters. Operating conditions
for the additional tests must be the same as for the first
test, with the exception of adjustments for the conflict-
ing parameters. The purpose of this approach  is to
demonstrate the combustor's performance within an
"envelope"  of conflicting conditions.

The trial burn plan should identify any anticipated con-
flicts among parameters and indicate the following:

•  The reason for the conflict, the interactions between
   the conflicting parameters, and the way in which the
   parameters are mutually exclusive.

•  The test methods used to avoid the conflict.

•  A description of all operating conditions for the vari-
   ous test  burns and a discussion of how the results of
   each test relate to those of the others.

4.12 Monitoring Systems

RCRA regulations require that once parameters have
been written into the permit for operation of a hazard-
ous waste  combustion device, the unit must be moni-
tored during operation so that its waste feed is cut off if
a  parameter is exceeded. Although operational charac-
teristics that should be  monitored vary (U.S.  EPA,
1979a), they generally include:
•  Combustion chamber outlet temperature

•  Combustion gas flow rate

•  Combustion system pressure
•  Waste feed and production rates

e  Carbon  monoxide and oxygen concentrations
Monitoring systems must meet  certain specifications
that cover measurement range, response time, and rel-
ative accuracy. These specifications and  recommend-
ed monitoring methods are described elsewhere (U.S.
EPA, 1989b, 1989c, 1990,1992; Gable, 1974; Shanklin
etal., 1990).
The waste feed cutoff is critical  to all  monitoring sys-
tems. Some hazardous waste combustion devices are
equipped with  monitoring systems that electronically
cut off the waste feed when the absolute value of a pa-
 rameter has been exceeded. Such systems should in-
clude  an  alarm  that sounds when a  parameter  is
approaching its cutoff threshold so that  the operator
 can initiate corrective action. For example, if a monitor-
 ing system is programmed to cut off the waste feed
 when   the combustion  chamber  temperature  drops
 below  1,800°F, the alarm might  sound when the tem-
 perature reaches 1,900°F. Additionally, an  electronic
monitoring system should be programmed to activate
the cutoff in the case of a power outage.
A waste feed cutoff is distinct from a combustion device
shutdown because it allows operations to continue on
auxiliary fuel or a nonhazardous waste until the prob-
lem is corrected. Except in extreme situations, contin-
ued operation is desirable so that the unit's temperature
can be maintained.
Monitoring systems for combustion  devices  that are
permitted to operate under rolling average conditions
must be capable of taking operational variations into
account through constant computation. Given the com-
plexity of electronic monitoring systems,  periodic in-
spections that include  the testing of the alarm system
should be conducted.
Sensors represent another key aspect of monitoring
systems. For a sensor to function properly, it must be
located where parameter measurements are represen-
tative, interference is minimal, and service can be per-
formed readily. For example, if a sensor is located too
close to the combustion device's flame, or other radiant
heat source, it will not measure gas flow accurately and
may be difficult to service.
System components and configurations for monitoring
the primary operational characteristics of hazardous
waste combustors are presented in this section.

4.12.1  Combustion Chamber Outlet
        Temperature
The temperature of a hazardous waste combustor must
be monitored on a continuous basis to ensure that the
minimum demonstrated temperature for waste destruc-
tion is maintained. To effectively monitor the combustor
temperature, sensors  must be located at the outlet of
each combustion chamber.  Two factors  to  consider
when evaluating temperature  measurement systems
are sensor type and sensor placement.
Most temperature sensors measure either radiant heat
(e.g., optical or infrared pyrometers) or changes in elec-
trical properties  associated  with temperature  (e.g.,
thermocouples or thermistors). Pyrometers  generally
are not acceptable for measuring the temperature of a
gas. Because a pyrometer will measure the  tempera-
ture of the refractory, the flame, or any other solid object
 in its line of sight, this type of sensor should not be used
to measure the temperature of gas unless the combus-
tion device system has been thoroughly tested and cal-
 ibrated against appropriate gas measurement devices.
The most appropriate use of a pyrometer is  as an in-
 strument for measuring the temperature of a radiating
 material at a remote point in the combustor, such as the
 burning bed or the flame.
                                                   51

-------
 Nonetheless, pyrometers sometimes are used as an
 alternative to a thermocouple to measure the tempera-
 ture of the gas in a combustor. In  this use,  the py-
 rometer is  inserted into a well that is placed at an
 advantageous point in the gas stream.  The sensor is
 aimed at the inside of the well, and its readout is cali-
 brated for the emissivity of the well's  surface.

 Several types of devices exist for measuring electrical
 properties  associated with temperature. The shielded
 thermocouple, however, is the type of device most com-
 monly used for combustion gas temperature monitor-
 ing. A thermocouple consists of two wires, each made
 of a different type of metal or alloy, that are welded to-
 gether at a  particular point,

 Regardless of the type of sensor used, placement is
 critical because temperature can vary widely from point
 to point. Ideally, a  sensor will measure temperatures in
 the bulk gas flow after the gas has traversed the com-
 bustion chamber volume that provides the specific res-
 idence time needed to achieve the desired destruction
 of organic constituents. Thus, the appropriate location
 for the sensor is likely to be at the outlet of the combus-
 tion chamber, well into the gas flow. When using a ther-
 mocouple  or other  temperature-sensing device, the
 sensor must be placed in a location where the tempera-
 ture Is consistent with  the temperature of the material
 being monitored, because such sensors read only their
 own temperature.

 If the sensor is intended to measure gas temperature at
 the outlet of the combustion chamber, it must not be
 placed in direct line with the flame.  Because a flame
 gives off significant  radiation, a temperature reading
 that is higher than that of the surrounding gases could
 result. Thus, a thermal  sensor should be shielded from
 the combustor's flame. Ensuring that the shield itself
 does not glow when it is subjected to radiant  heat is
 equally important.

 Often, placing multiple sensors in a combustion system
 is practical, because most are relatively inexpensive
 and multiple readouts are useful for identifying hot and
 cold spots within the system as well as for checking ac-
 curacy. Frequently, multiple sensors are used to moni-
 tor bed temperatures in different zones  of a fluidized
 bed  furnace. In such  combustors, exceptionally hot
 spots can fuse the bed material, while cold spots can
 result in tar deposits on the bed material. Also, temper-
 ature extremes can increase the particle size of the bed
 material, which can lead to defiuidization.

 4.12.2 Combustion Gas Flow Rate

 Regulation 40 CFR section 264.345(b)(4) requires that
"the permit [for a hazardous waste combustion device]
will specify acceptable  operating limits" for "an  appro-
priate Indicator of combustion gas velocity." Thus, up-
 per limits placed on parameters that can be related to
 the combustion gas flow rate must be measured to:

 • Control the gas residence time in each  combustion
   chamber.

 • Control the gas throughput of the entire system so
   that back pressure is minimized at joints and seals
   (e.g., at the SCC inlet of a rotary kiln).

 • Control the gas flow rate through the APCE so that
   the equipment is not overloaded.

 Combustion gas flow rate can  be monitored in several
 ways. The preferred method is to use a direct gas flow
 monitor at the outlet of the SCC.  This method can be
 undesirable if the duct work is not long enough to ac-
 commodate flow pattern development, the combustor is
 vertically oriented such that access to the duct is limit-
 ed, or the system's configuration requires breeching the
 duct at a high-temperature point. (High temperatures in
 the duct might require that the sensor be made of spe-
 cial materials, such as Inconel.) Conversely, monitoring
 gas flow in the stack avoids concerns associated with
 high temperatures, accessibility, and flow pattern devel-
 opment. Additionally, concerns about particulates and
 other contaminants that  can plug or foul sampling lines
 or sensors are minimized. A minor disadvantage asso-
 ciated with monitoring gas flow at this  location  is the
 increased possibility that ambient air leaks into the sys-
 tem, upstream  of the induced-draft  fan, can bias the
 flow measurement.

 In some systems, however, conditions  such  as high
 temperature,  high  particulate loading, and high acid
 gas loading can result in unsatisfactory life  and perfor-
 mance of the monitor. A practical alternative approach
 is to place the  monitor  immediately before the stack.
 This placement, however, increases the likelihood of
 introducing related errors  (usually small) in the mea-
 surement  because of air infiltration or changes in the
 water content of the gas stream.

 Gas flow rate can be measured or approximated in sev-
 eral ways:

 •  By placing a flow sensor in the  combustor's  duct
   (e.g., a pitot tube, a thermal conductivity indicator, or
   a sonic flow indicator, though not commonly used for
   this application).

 •  By measuring the drop in pressure across a constric-
   tion in the gas flow (e.g., baffle plate, Venturi section,
   or orifice).

 •  By measuring the power consumption, current us-
   age, or  kilovolt amperes of the fan(s).

 Placing a  pitot  in the  flow is one of the most direct
 means of monitoring gas velocity. Pitots used for such
applications are similar to the type used for source test-
                                                  52

-------
ing in which the sensor is attached to an electronic
pressure transducer rather than a gauge, which is used
for manual measurements. Because the gas flow in a
combustor duct varies from point to point, the pitot must
be located at a point  in the duct system where the ve-
locity is reasonably proportional to the gas flow over a
range of flows. A multiple-tap  pitot (one type on the
market is called an annubar) takes flow ranges into ac-
count. Multiple-tap pilots in a straight configuration can
be placed across the width, diameter, or radius of a duct
so that pressure differences—proportional to the mean
velocity pressure—can be correlated to the total gas
flow across the pitot. For such applications,  the pitot
must be properly sized for the duct; however,  because
the sensor is not  used for primary gas flow measure-
ment, the  sizing can be approximate. Round,  multiple-
tap pilots,  which typically have evenly spaced openings
around the exterior of the instrument, are used only for
round ducts. They are placed so that the center of the
pitot corresponds to the center of the duct.

The application of a pitot is limited in a combustor duct,
because the gas can  be dirty or corrosive—and usually
wet—causing the pilot's orifices to corrode or plug such
thai Ihe accuracy of readings can be undermined.

Of the instruments available to measure  gas flow in
closed conduils, pressure  or velocity head meters  (a
variation on the pitot) are among the oldest and most
widely used. Head-type flow meters incorporate prima-
ry elements that interact direclly with the gas stream to
induce velocity changes, and secondary elements that
sense the  resulling pressure perturbations. The flow
rate of interesl is a function of the detectable differential
in pressure. The principal shortcomings of such  instru-
menls include: Ihe need for elemenls to be inserted di-
rectly into flow paths (to make conlact with the gas
slream), making Ihem susceptible to  corrosion, ero-
sion, and  fouling;  Ihe requirement for seals; the likeli-
hood that  the conduit may have lo  be  opened  for
inspection  or  service; and Ihe permanenl  pressure
losses caused by  reslriclions placed in  Ihe channels.

A related  approach for monitoring the  velocity of gas
involves Ihe sensing of thermal conductivity by placing
an electrically heated wire, or anemometer, into Ihe gas
flow. The rale at which the wire loses heat can be corre-
lated lo Ihe rale al which gas flows past Ihe wire. Such
sensors work reasonably well when Ihe gas slream is
dry,  ils  lemperalure  is conslanl, and  ils composilion
does nol vary. Because Ihe gas often is wel, however,
sensors often are used in less lhan ideal silualions.

A second  melhod involves measuring Ihe  pressure
drop of gas across a conslriclion in Ihe flow,  using an
orifice,  or  Venluri, inserted into the duel.  Addilional
pressure drops in Ihe combustor can  be avoided  by
measuring Ihe pressure drop across a portion  of Ihe
combustor unil lhal has conslanl geomelric properties,
such as a constriction in the diameter of the duct. Be-
cause gas can change composition and temperalure al
different points in the combustion device system, care
must be taken to ensure that the gas being monitored
has the same properties at bolh pressure laps. For ex-
ample, monitoring Ihe pressure drop across a quench is
nol recommended because Ihe gas changes bolh lem-
peralure and composilion in belween Ihe two sensors
at Ihis poinl in Ihe combustion system.

Yel anolher method of monitoring combustion gas flow
rate is to measure Ihe power used (vollage and currenl
may be adequate when Ihe fluclualions in Ihe load are
small) by Ihe induced-draft fans. When using fan power
consumption as an indicator of combustion gas veloci-
ty, Ihe fan musl be operating in a region of Ihe power
curve where Ihe sensor's readoul will be representative
of Ihe overall system. As shown in Figure  4-14, for in-
stance, a large change in gas flow can be associated
with  a small, possibly  immeasurable, change in the
power curve.
                Backward curved
                fan blade
Fan power
(hp or kW)
                           Straight or forward
                           curved fan blade
                        Gas flow rate (CFM)

Figure 4-14.  Relationship between fan power and gas flow
           rate for two types of fan blades.
For some forced  draft combustors in which  Ihe inlel
poinls for combustion air are well defined and easily
measured, measuring Ihe flow rale of combustion air
instead of combustion gas may be possible. For a given
temperature, the flow rale of combustion gas (i.e., Ihe
producls of combustion) can be approximated wilhin
reasonable accuracy by Ihe flow rale of combustion air
(i.e., Ihe primary  and secondary air being fed to Ihe
combustor). In mosl cases, combustion  air conslilules
95 percenl or more of a device's combustion gas.

4.12.3  Combustion System Pressure
Pressure  in combustion devices is monitored almosl
exclusively by using eleclronic pressure Iransducers
lhal  utilize piezoelectric materials. Pressure is mea-
sured bolh in terms of pressure drop across equipmenl
(i.e., lo measure flow or lo monitor APCE performance)
and  total  pressure (i.e., to minimize Ihe likelihood of
                                                  53

-------
 puffing or opening of the emergency vent stack). For
 both types of measurements, the main concerns are:
 •  Pressure taps must be situated and sized to mini-
    mize the risk of openings becoming plugged. If plug-
    ging is a likelihood, provisions for back-blowing the
    lines and testing for plugging must be made.
 •  The pressure taps  must be  situated according  to
    whether static  or dynamic pressure will be mea-
    sured. Flow measurements with pitot tube devices
    require that one pressure measurement be taken  of
    dynamic pressure. In most cases, however, espe-
    cially when monitoring  for puffing, readings taken
    must be of static pressure. The pressure tap for stat-
    ic pressure measurements must be placed  in the
    flow at a point where smooth fluid flows in a direction
    parallel to the mouth of the tap.

 4.12.4 Waste Feed and Production Rates
 Regulations for the operation of hazardous waste com-
 bustors require the monitoring of waste feed rates and
 production rates. For boilers, production rate refers  to
 steam; for other industrial furnaces, it refers to the rate
 of manufacture of the intended product. Because mon-
 itoring these  rates often requires measuring both solid
 and liquid flow streams, the following discussion on
 feed rate monitoring includes information applicable  to
 monitoring production rates.
 The feed rate of solid streams can be  monitored using
 continuous weighing conveyors, auger rotational speeds,
 or weigher clamshells. Because of the variability  in
 types of solids that can be fed into a hazardous  waste
 combustion device, the choice of a monitoring system
 Is specific to particular situations.
 The feed  rate of liquid flow streams is monitored using
 systems specific to liquid properties, pressure drop, re-
 quired accuracy, and the flow range to be measured.
 For liquid flow monitoring, the compatibility of the flow
 meter and the liquid being measured is critical.  Flow
 rate can be measured using the following technologies:
 •  Orifice meter
 •  Flow tube meter
 •  Positive-displacement meter
 •  Vortex shedding meter
 •  Turbine meter
 •  Rotameters
 •  Mass flow meter

4.12.4.1   Orifice Meter and Flow Tube Meter
Both the orifice meter and the flow tube meter monitor
fluid velocity by measuring the pressure drop across an
 Obstruction in the flow stream. Typically, the pressure
 drop varies with the  square of  the  velocity. Orifice
 meters measure the differential pressure between the
 upstream and downstream side of a plate with a hole
 (the orifice) in it by utilizing pressure taps on either side
 of the plate. The flow tube utilizes a tapered constriction
 in the line rather than an orifice (the front half of a Ven-
 turi) and monitors velocity by measuring the pressure
 drop across the constriction in the flow.

 A disadvantage of the orifice meter is that suspended
 matter can build up on the plate, affecting the instru-
 ment's accuracy and thus limiting its application to com-
 bustion systems in which the fluid has a low suspended
 solids content.  Because the flow tube meter does not
 have sharp edges,  like those on the orifice plate, it is
 better suited to applications with moderate suspended
 particulate content.  Like the orifice meter, however, the
 flow tube meter is susceptible to plugging,  thus its use
 in high-solids applications is  limited.

 4.12.4.2  Positive-Displacement Meter
 Positive-displacement  (p-d)  meters measure  cumula-
 tive volume rather than flow and can be combined with
 systems that continuously measure the time derivative
 of volume to generate a flow rate.  The p-d  meter func-
 tions like a pump operating  in reverse that traps flow
 with a constant-volume element, such  as a reciprocat-
 ing piston, rotary piston, rotary vane, or nutating disk.
 The meter measures the number of constant volume
 elements that pass through the device using mechani-
 cal or electronic counters. Because of  their moving
 parts, p-d meters are hampered by high-solids concen-
 trations.  Certain  designs, however, especially those
 with flexible membranes, can handle some amount of
 suspended solids.

 A common type of p-d meter  is the nutating disk meter,
 which consists of a movable disk mounted on a concen-
 tric sphere. The disk is contained in a working chamber
 with spherical sidewalls and top and bottom surfaces
 that are conic in shape. A radial partition that extends
 across the entire height of the working  chamber keeps
 the disk from rotating on its axis. Each complete move-
 ment of the disk displaces a fixed volume of liquid. The
 liquid enters through an inlet port and  fills the spaces
 above and below the disk, which fits closely and pre-
 cisely in the  measuring chamber.  The advancing vol-
 ume of liquid moves the piston in a nutating motion until
the liquid discharges from the outlet port. The major lim-
 itation of this type of flow meter is its sensitivity to grit.

4.12.4.3  Vortex Shedding Meter

In a vortex shedding meter, the fluid stream is forced
past an obstruction (shedding bar) that sets  up vortices
(eddies) in the fluid. The vortices cause vibrations in the
shedding bar that are proportional to the flow, and a
                                                  54

-------
.piezoelectric crystal (or another type of sensor) con-
verts the vibrations into voltage. The voltage is ampli-
fied and then transmitted to an electronic device that
converts the signal into a flow rate. Because these
meters have no moving parts, they can tolerate a mod-
erate amount of grit. They provide sufficiently accurate
measurements for most liquid  flow monitoring situa-
tions.

4.12.4.4 Turbine Meter

The turbine meter is a small turbine mounted in the flow.
The blades of the turbine are driven by the passing flu-
id, while a  mechanical sensor measures the flow rate.
This type of flow meter rarely is used for waste applica-
tions because it is not well suited for liquids containing
suspended solids.

4.12.4.5 Rotameter

The rotameter consists of a vertically mounted, tapered
tube in which a ball or tapered plug of known weight is
mounted. Either the  tube is made from a transparent
material or it has a window through which the plug or
ball can be seen. The fluid to be measured flows up the
rotameter  such that the faster the flow, the higher the
ball or plug rises. The volumetric flow rate of the fluid is
determined by reading the level of the ball or plug on an
attached scale and using a calibration curve for the fluid
inside. The rotameter can tolerate moderate amounts
of fine grit, but it cannot accommodate large  solids or
large amounts of grit.

4.12.4.6  Mass Flow Meter
The mass flow meter consists  of identical twin tubes,
either U-shaped or coiled. As the fluid passes through
the tubes,  they resonate at frequencies that are func-
tions of the mass flow rate of the fluids. The tubes' res-
onances are fed to an electronic circuit that computes
the difference and displays the  result as a flow rate. A
 unique feature of these meters is that they measure the
actual mass flow rate rather  than the velocity or volu-
 metric flow rate of the liquid.

 4.12.5 Carbon Monoxide and Oxygen
        Concentrations

 Hazardous waste combustion devices are required by
 regulation  to be equipped with continuous emission
 monitors (CEMs) that  measure the concentration of
 carbon monoxide and oxygen. Oxygen is monitored to
 ensure that measured carbon monoxide concentrations
 are normalized to 7  percent oxygen. Combustors with
 carbon monoxide concentrations greater than 100 ppm
 (dry, corrected to 7 percent oxygen) in the stack also
 must monitor total hydrocarbons. For particular operat-
 ing conditions, some combustors also  are required to
monitor opacity, sulfur dioxide/sulfate, and  nitrogen
oxides.
Minimum  performance standards have  been estab-
lished for  systems that monitor carbon monoxide and
oxygen concentrations in hazardous waste combustors
in lieu of requirements for specific types of monitors (40
CFR part 266, Appendix IX). These standards are sum-
marized in the BIF  guidance document (U.S.  EPA,
1992).  Additional information is available in the litera-
ture on the following topics:

• Operation and installation of CEMs for opacity, car-
  bon monoxide, carbon dioxide, sulfur dioxide, and
  nitrogen oxides (U.S. EPA, 1979a).

• Test information on carbon monoxide and total hy-
  drocarbon  monitors.
Carbon monoxide and  oxygen concentrations in com-
bustion gas typically are monitored at the stack, down-
stream of the  APCE, unless significant dilution or air
infiltration occurs in the APCE. Although APCE compli-
cations are rare, such situations might require the loca-
tion of monitors at the outlet of the last combustion
chamber.  This location offers the benefit of reducing in-
strument  response time in case a significant combus-
tion upset takes  place in  any of the chambers.  A
disadvantage, however, is that the monitor or extractive
tap will be subjected to high temperatures and often
acidic, dusty conditions at this outlet.
CEMs are operated in one of three modes:
• Cross-stack. The sensing portion of the analyzer is
   mounted on the duct,  and the analyzer projects a
   sensing beam across the stack through the flue gas.
• In situ. The sensing portion of the analyzer is located
   in the flue gas stream, and the flue gas flows through
   or over the sensor.
• Extractive. The sample  gas  is extracted  from the
   flue, conditioned, and transported to a remote ana-
   lyzer.
Cross-stack monitors  are limited to systems that use light
or other forms of ionizing or electromagnetic  radiation.
Examples include nondispersive infrared  monitors for
carbon monoxide and various types of opacity monitors.

In situ monitors have a faster response time than ex-
tractive monitors, which cannot sense the condition of
the gas stream until the sample travels down the sam-
pling line, through the  conditioning system, and to the
sensor. Because in situ monitors must rely on sensors
with enhanced durability to withstand harsh conditions,
however,  they are generally unsuitable for use at points
upstream of the quench. Typically, they are installed af-
ter the APCE, immediately before the fan or stack.
                                                   55

-------
 More detailed information on GEM technologies will be
 available in the revised Handbook: Thermal Combus-
 tion of Hazardous Waste, to be published  by  EPA's
 Center for Environmental Research Information  in the
 future.

 4.13  Recordkeeping

 Once a hazardous waste combustion device is brought
 on line, records must be kept that demonstrate that the
 cornbustor is operating within the parameters estab-
 lished  in the permit  Along with tracking operating
 modes and conditions, records also must be kept on
 waste types fed into the cornbustor.

 The information related to complying with permit condi-
 tions that must be retained includes:
 Group A Parameters

 • Combustion chamber temperatures
 • Combustion gas flow rates

 • Carbon monoxide and oxygen monitor readouts
 • APCE inlet temperatures
 • APCE parameters (as applicable)
 • Changes in regulated operating parameters
 • Amounts and types of wastes burned
 Group B Parameters

 • Amount of chlorine feed
 • Amount of ash fed
 • Metal types and feed rates
 • Scrubbing liquid solids content
 Group C Parameters
 •  Nozzle pressures

 •  Changes in burner settings
 •  Heating value of primary wastes used as primary fuel
 Because RCRA requires "cradle-to-grave" recordkeep-
 Ing on hazardous waste,  operators also should retain
 manifests for wastes generated off site and develop a
 manifest system for tracking wastes generated on site.
 This tracking should include information on waste type,
 quantity, storage, and analysis data, as well as data on
 which wastes were received and burned.

 Information on the  maintenance and calibration of
 CEMs and other monitoring systems also should be re-
tained along with records on upset conditions and other
emergencies, including:

•  Dates, times, reasons for AWFCOs and emergency
   vent stack openings.
 •  Summary reports of all incidents.

 •  Records affecting time limitations (e.g., 720 hours
    limit for shakedown).

 •  Unusual facility conditions related to waste manage-
    ment:

    - Leaks

    - Spills

    - Observed puffing

 The data format must be designed so that related infor-
 mation can be matched quickly. For example, a facility
 that has  a multiple-point/multiple-operating condition
 permit must be able to present data on waste types and
 on operating conditions that demonstrate that the sys-
 tem was  operating in the proper mode at any given
 time. Multiple waste feed streams must be clearly iden-
 tified,  and  units of measurement must be "readily con-
 vertible" to permit conditions.

 All  records must be available  for inspection  by ap-
 proved authorities and retained until facility closure, ex-
 cept smelters and  small quantity burners,  for which
 records must be kept for 3 years.

 4.14  References

 When an NTIS number is cited in a reference, that doc-
 ument is available from:

   National Technical Information Service
   5285 Port Royal Road
   Springfield, VA22161
   703-487-4650

 Barton, R.G., P.M. Maly, W.D. Clark, W.R. Seeker, and
   W.S. Lanier. 1988. Prediction of the fate of toxic met-
   als in hazardous waste incinerators. Prepared for the
   U.S. Environmental Protection Agency, Contract 68-
   03-3365, Work Assignment No. 2. Energy and Envi-
   ronmental Research Corp., Irvine, CA.

 Dellinger, B.,  D.A. Tirey, P.M. Taylor, J. Pan,  and C.C.
   Lee. 1988.  Products of incomplete combustion from
  the high  temperature pyrolysis of chlorinated meth-
  anes. Third Chemical Congress of North America and
  the 195th National Meeting of the American Chemical
  Society. Environmental Chemistry Preprints 28(1 ):81.

Gable, L.W. 1974. Installation and calibration of ther-
  mocouples. ISA Transactions. 13(1):35-39 (January-
  March).

Goldstein, H.L., and C.W. Siegmund. 1976. Influence of
  heavy fuel  oil  composition  and boiler combustion
  conditions on particulate emissions. Environ.  Sci. &
  Technol. 10(12):1109.
                                                 56

-------
John  Zink  Co. 1988.  Hydro-Sonics Systems, Gas
  Cleaning  Equipment  Tandem Nozzle—Series  TN.
  Bulletin HSS 003A, Tulsa, OK.

Petersen, H.H. 1984. Electrostatic precipitators for re-
  source recovery plants. National Waste Processing
  Conference.  American Society of Mechanical Engi-
  neers, Orlando, FL.

Richards, J., and P. Quarles.  1986.  Fabric filter mal-
  function evaluation.  Final Report.  EPA/68-02/3960.
  Prepared  for the  U.S. Environmental  Protection
  Agency by Perrin Quarles  Associates, Charlottes-
  ville, VA  Available from J. Richards at Air Control
  Techniques,  Durham, NC. September.

Shanklin, S., L. Cone, and S. Steinberger. 1990. Evalu-
  ation of HCI measurement techniques at a hazardous
  waste incinerator. Prepared for U.S. EPA Atmospher-
  ic Research and  Exposure Assessment Laboratory,
  Research Triangle Park, NC.
Turner, B.D. 1972. Workbook of atmospheric dispersion
  estimates. AP-26. Office of Air Programs, Research
  Triangle Park, NC. Out of print.
U.S. EPA. 1992. U.S. Environmental Protection Agency.
  Technical implementation document for ERA'S boiler
  and industrial  furnace regulations.  ERA/530/R-92/
  011. NTIS PB92-154947. Office  of Solid  Waste,
  Washington, DC.
U.S. EPA. 1990. U.S. Environmental Protection Agency.
  Methods  manual for compliance with the BIF  regula-
  tions: Burning hazardous waste in boilers and indus-
  trial furnaces. EPA/530/SW-91/010. NTIS  PB91-
  120006. Office of Solid Waste, Washington, DC.
U.S. EPA. 1989a. U.S. Environmental Protection Agen-
  cy.  Guidance on setting permit conditions and report-
  ing  trial burn results. Volume 2. EPA/625/6-89/019.
  Center for Environmental Research Information, Cin-
  cinnati, OH.
U.S. EPA. 1989b. U.S. Environmental Protection Agen-
  cy.  Hazardous waste incineration measurement guid-
  ance manual.  Volume  3. EPA/625/6-89/021. NTIS
  PB90-182759. Center for Environmental Research
  Information, Cincinnati, OH.
U.S. EPA. 1989c. U.S. Environmental Protection Agen-
  cy.  Proposed methods for measurement of CO, O2,
  THC, HCI, and metals at hazardous waste incinera-
  tors. In:  Volume 4: Hazardous  Waste Incineration
  Guidance Series. Office of Solid Waste, Washington,
  DC. Out of print.
U.S. EPA. 1986. U.S. Environmental Protection Agency.
  Test methods for, evaluating solid waste: physical/
  chemical  methods  (3rd  ed.).  EPA/SW-846. NTIS
  PB88-239223. Office of Solid Waste, Washington,
  DC. Continuously updated.
U.S. EPA. 1981. U.S. Environmental Protection Agency.
  Engineering handbook for hazardous waste incinera-
  tion. EPA/SW-889. NTIS PB81-248163. Office  of
  Solid Waste, Washington, DC.

U.S. EPA. 1979a. U.S. Environmental Protection Agen-
  cy.  Continuous air pollution  source monitoring sys-
  tems. EPA/625/6-79/005. Center for  Environmental
  Research Information, Cincinnati, OH. Out of print.

U.S. EPA. 1979b. U.S. Environmental Protection Agen-
  cy. Graphical exposure modeling system (GEMS) us-
  er's manual. Office of Toxic Substances, Washington,
  DC. Unpublished.
U.S. EPA. 1973. U.S. Environmental Protection Agency.
  Air  pollution  engineering manual (2nd ed.). AP-40.
  NTIS PB-225132. Office of Air Quality Planning and
  Standards, Research  Triangle Park, NC.

4.15  Bibliography
        ^
ASME. Thermocouple handbook. New York, NY: Amer-
  ican Society of Mechanical  Engineers.  Periodically
  updated.
ASTM. 1986. American  Society for Testing  and Materi-
  als. Standard test method for water  in halogenated
  organic  solvents and their admixtures.  D-3401-85.
  Philadelphia, PA.
Bruce, K.R., L.O. Beach,  and  B.K. Gullett. 1990. The
  role of gas-phase CI2 in the formation of PCDD/PCDF
  in municipal  and hazardous  waste  combustion. Pro-
  ceedings of an Incineration Conference. University of
  California at Irvine (May 14-18).
Buonicore, A.J., and W.T. Davis, eds. 1992. Air pollution
  engineering manual. Air and  Waste Management As-
  sociation. New York, NY: Reinhold.
Grumpier,  E.P., E.J. Martin, and G. Vogel.  1981. Best
  engineering  judgement for permitting  hazardous
  waste incinerators. Paper presented  at  ASME/ERA
  Hazardous Waste Incineration Conference, Williams-
  burg, VA (May 27).
Cudahy, J.J., and W.L. Troxler. 1983. Autoignition tem-
  perature as an indicator of thermal oxidation stability.
  J. Haz. Mat. 8(59).
Dellinger, B., and P.H. Taylor. 1988. Development of a
  thermal stability based index of hazardous waste in-
  cinerability: Status report. Dayton,  OH: University of
  Dayton Research Institute.
Dellinger, B., W.A.  Rubey, D.L. Hall, and J.L. Graham.
  1986. Incinerability of hazardous waste.  Hazardous
  Waste and Hazardous Materials 3(2):139-150.
                                                  57

-------
 Delltnger, B., D.L. Hall, J.L. Graham, S.L. Mazer, W.A.
   Rubey, and P.H. Taylor. 1986. The theory of thermal
   zone chemistry and its influence on hazardous waste
   Incineration. Proceedings of the 79th Annual APCA
   Meeting. Paper #86-84.1. Minneapolis, MN (June).

 Dellinger, B., M. Graham, and DATirey. 1986. Predict-
   ing emissions from the thermal processing of hazard-
   ous wastes.   Hazardous  Waste  and Hazardous
-  Materials 3(3):293-307.
 Dellinger, B., J. Torres, W.A. Rubey, D.L. Hall, J.L. Gra-
   ham, and R. Carnes. 1984. Determination of the ther-
   mal  stability  of  selected   hazardous  wastes.
   Hazardous  Waste  and  Hazardous   Materials
   1(2):137-157.
 Denbigh,  K. 1964. The principles of chemical equilibri-
   um. London: Cambridge University Press.
 Graham, J.L., D.L. Hall, and B. Dellinger. 1986. Labora-
   tory investigation of the thermal degradation of a mix-
   ture  of hazardous organic wastes.   Environ.  Sci.
   Technol. 20(7):703-710.
 Graham, J.L., W.A. Rubey, and B. Dellinger.  1982. De-
   termination of the thermal decomposition properties
   of toxic organic substances. Proceedings of the Sum-
   mer National Meeting of AIChE. Paper #17d. Cleve-
   land, OH (August).
 MacKinnon, D.J. 1974. Nitric oxide formation at high
   temperature. JAPCA 24(3):237-239.
 Miller, D.L., V.A. Cundy, and R.A. Matula. 1983. Pro-
   ceedings of the EPA/I ERL Ninth Annual  Research
   Symposium—Incineration and Treatment of Hazard-
   ous Waste. EPA/600/9-84/015. NTIS PB84-234525.
   Cincinnati, OH.
 Nelson, J.  1987.  Continuous  measurement of  HCI
   emissions from municipal solid waste incineration fa-
   cilities. Presented at the Air Pollution Control Associ-
   ation, International Specialty Conference on Thermal
   Treatment of Municipal,  Industrial, and  Hospital
   Wastes. Pittsburgh, PA (November 3-6).
 Niessen, W.R. 1978. Combustion and incineration pro-
   cesses, pollution and technology. New York, NY: Mar-
   cel Dekker, Inc.

 Perry, R., and D. Green. 1984. Perry's chemical engi-
   neering handbook (6th ed.). New York, NY: McGraw-
   Hill.
 Seeker, W.R., J.C. Kramlich, and M.P. Heap. 1983. Pro-
   ceedings of the ERA/I ERL Ninth Annual  Research
   Symposium—Incineration and Treatment of Hazard-
   ous Waste. EPA/600/9-84/015. NTIS PB84-234525.
   Cincinnati, OH.
Taylor, PH., and B. Dellinger. 1988. Thermal degrada-
  tion characteristics of chloromethane mixtures. Envi-
  ron. Sci. Technol. 22(4):438.

Taylor, PH., and J.F. Chadbourne. 1987. Sulfur hexaflu-
  oride as a surrogate for monitoring hazardous waste
  incinerator performance. JAPCA 37(b):729.
Theodore, L., and J.. Reynolds.  198/.  Introduction to
  hazardous waste incineration.  New York, NY: John
  Wiley & Sons.
Tsang, W., and W. Shaub. 1982. Chemical processes in
  the incineration  of hazardous  materials: Detoxifica-
  tion of hazardous waste (J. Exner ed.). Ann Arbor, Ml:
  American Chemical Society.
U.S. EPA. 1992. U.S. Environmental Protection Agency.
  Catalogue of hazardous and solid waste publications
  (6th ed.). EPA/530/B-92/001. Office of Solid Waste,
  Washington, DC.
U.S. EPA. 1989. U.S. Environmental Protection Agency.
  Handbook on quality assurance/quality control (QA/
  QC) procedures for hazardous waste incineration.
  EPA/625/6-89/023.  NTIS PB91-145979. Center for
  Environmental   Research  Information,  Cincinnati,
  OH.
U.S. EPA. 1988. U.S. Environmental Protection Agency.
  Air dispersion  modeling as applied to hazardous
  waste incinerator evaluations; an introduction for the
  permit writer. Office of Solid Waste, Washington, DC.
  Unpublished. Obtainable by calling the RCRA Hotline
  at 800-424-9346.
U.S. EPA. 1986. U.S. Environmental Protection Agency.
  Practical guide:  Trial burns for hazardous waste in-
  cinerators.  EPA/600/2-86/050. NTIS  PB86-190246.
  Washington, DC.
U.S. EPA. 1985. U.S. Environmental Protection Agency.
  Laboratory  and  field evaluation of the semi-VOST:
  Volumes 1 and 2. EPA/600/4-85/075. Environmental
  Monitoring  Systems Laboratory, Research  Triangle
  Park, NC.
U.S. EPA. 1985. U.S. Environmental Protection Agency.
  Modified Method 5  Train and Source Assessment
  Sampling System:  Operator's manual. EPA/600/8-
  85/003.  NTIS  PB85-169878.  Research   Triangle
  Park, NC.
U.S. EPA. 1984. U.S. Environmental Protection Agency.
  Protocol for the collection and analysis of volatile
  POHCs using VOST. EPA/600/8-84/007. NTIS PB84-
  170042. Environmental Monitoring Systems Labora-
  tory, Research Triangle Park, NC.
                                                  58

-------
U.S.EPA. 1984. U.S. Environmental Protection Agency.
  Investigation of the thermal destructibility of hazard-
  ous wastes using the  IDAS. EPA/600/2-84/138.
  NTIS PB84-232487. Cincinnati, OH.

U.S. EPA. 1984. U.S. Environmental Protection Agency.
  Introduction to  boiler operation: Self-instructional
  guidebook. EPA/450/2-84/010. Air Pollution Training
  Institute, Course SI:428A. Research Triangle Park,
  NC.

U.S. EPA. 1983. U.S. Environmental Protection Agency.
  Guidance manual for hazardous  waste incinerator
  permits. EPA/SW-966. NTIS PB84-100577. Office of
  Solid Waste, Washington, DC.
Vandell, R.D., and L.A. Shadoff. 1984. Relative rates
  and partial combustion products from the burning of
  chlorobenzenes   and   chlorobenzene   mixtures.
  Chemosphere 13(11): 1,177.

Wallace, D., and A. Trenholm. 1986. Products of incom-
  plete combustion from hazardous waste combustion.
  Midwest Research Institute. Kansas City, MO.

Weast, R.C., ed. 1971. CRC handbook of chemistry
  and physics (51st ed.). Cleveland, OH: The Chemical
  Rubber Company.
                                                 59

-------

-------
                                              Chapter 5
                               Toxic Metals and Particulate Matter
5.1    Partitioning of Metals

The control of toxic metals and participate matter are
addressed in this chapter. Most of the discussion is fo-
cused on toxic metals but applies equally well to partic-
ulate matter,  because the same mechanisms control
them both. The choice of control  methods depends on
the physical location of the metals in the combustion
process. The partitioning of metals to different places is
the key to how metals react in combustion.

5.1.1  Fundamental Principles

Unlike organics, metals are not destroyed during com-
bustion. All of the metals fed into a combustion device
are conserved, and thus partition into one or more of the
following places (Figure 5-1):

•  The bottom of the combustion device with the ash.
•  The bottom of the scrubber with the scrubber ash.
•  The bottom of the particulate control system with the
   fly ash.

•  The air when they are emitted from the stack.

Most metals, especially those that enter the system as
solid waste, end up in the bottom ash. The fundamental
principles that influence the behavior of metals in com-
bustion devices are discussed  in this section. A more
detailed discussion is provided by Barton et al. (1990).

Metals behavior in combustion is complex (Figure 5-2).
Metals follow one of three pathways:

•  Remain with the waste as a solid

•  Become entrained in the gas stream

•  Vaporize

In the first pathway, the metals remain with the waste
throughout the entire process and end up as residuals
or bottom ash. This is the path of least resistance.
                        Combustor
Figure 5-1.  Partitioning of metals to exit streams.
                                                  61

-------
                Burning spray
          Reducing
          environment
          Ash particle
       Homogenous
       condensation
                                                                                  Fume nuclei
Figure 5-2.  Principles of metals partitioning.

In the second pathway, the metal is entrained into the
gas stream. If the metal is in liquid waste, entrainment is
probably the primary pathway. In the combustor, liquid
waste and the metals in it atomize into the gas stream,
where most of the metal  will remain. If the hazardous
waste is solid waste, the amount that becomes entrained
depends on many variables such as the velocity of the
combustion gas. The higher the velocity, the more likely
the particles are to be swept into the gas stream. The
degree to which this particle entrainment occurs de-
pends on the size of the particles in the solid waste and
how those particles behave  as the organics  burn—
whether they break up and form friable, dusty particles
that can easily become entrained, or whether they stay
with the residuals. The smaller paniculate tends to be-
come entrained, although the particles are still relatively
large (i.e., 10 microns [urn] in size and larger). Particles
in this size range are generally relatively easy to remove
from the gas stream, downstream in the process. Since
they are easily controlled, these entrained particles are
not the ones of most concern in metal emissions.
The third, and most important, pathway is the vaporiza-
tion route. Some toxic metals are volatile and  have a
significant vapor pressure. In the hot, burning environ-
ment of the combustion device, a portion of the metals
in the waste vaporizes, changing from a liquid or solid
state to a gas. This process depends on the tempera-
ture and on the chemical environment. A nonvolatile
metal vaporizes if temperature increases or if the metal
comes into contact with another species and reacts to
form a volatile compound. If chlorine or other halogens
are present in the local surroundings, for example, the
metals may undergo a chemical reaction and form chlo-
rides, which are typically more volatile than the oxides
of the metals. Those metal chlorides will subsequently
vaporize.
Lead is the most dramatic example of how chlorides are
more volatile than oxides of the same metal. In the ab-
sence of chlorine, the most stable form of lead is lead
oxide (PbO), which has medium volatility. In the pres-
ence of chlorine, however, the most stable form of lead
is lead chloride, which vaporizes at very low tempera-
tures and can remain in the flue gas in vapor form as it
goes through the air pollution control device. Fluorine
and chlorine act in 3 similar manner; metal chlorides
and metal fluorides have similar volatility and both are
typically more volatile than the corresponding metal ox-
ides.
Volatile metals do not always vaporize. Those that are
enclosed in a larger particle may have difficulty surfac-
ing and subsequently vaporizing; so diffusion limits how
quickly metals  vaporize. Chemical  kinetics also can
prevent vaporization, even if chlorine is present. If the
rate of the reaction to form the volatile chloride is slow,
the metal might not undergo the reaction in the combus-
tion process and not vaporize.
The hot combustion  gas stream cools as it exits the
combustion chamber, causing vaporized metal to con-
dense into  particulate matter. The vaporized metal can
                                                   62

-------
condense in one of two ways: homogeneously or heter-
ogeneously. In homogeneous condensation, gas forms
tiny fume particles, all by itself, in the tens-of-angstrom
size  range. These particles stay this size only briefly,
very quickly growing to larger particles (typically 0.1 to
0.5 urn) by colliding with each other and coagulating, as
well  as by condensing with other vapor. Unfortunately,
most types of air pollution control equipment (APCE)
are least effective in this submicron size range (see Fig-
ures 4-13 and 5-3). In heterogeneous condensation, the
vapors condense onto the surfaces of existing particles,
such as condensed fume or entrained fly ash. Hetero-
geneous condensation  occurs on whatever  surface
area is most prominent. Since the available surface
area is dominated by very small particles, heteroge-
neous  condensation also  tends to occur in the  very
small size particle range (i.e., mostly less than a mi-
cron).
     99.99

      99.9


       99
Collection
efficiency g5

(%)     90

       80

       50
                                  Fabric filter
                    Venturi
                    scrubber
                    (20 in HaO)
           0.05 0.01        0.05   1
                  Particle diameter (microns)
                                              10
Figure 5-3.  Typical collection efficiency of three types of APCE
          as a function of particle size.
Regardless of whether the metal condenses by the het-
erogeneous or homogeneous route, in most systems,
vaporized metals ultimately end up as submicron parti-
cles that are difficult to control. The operational parame-
ters of most concern from the viewpoint of controlling
toxic metals are those that influence the formation of
fine particles, which are generally the same parameters
that influence volatilization.

5.1.2   Parameters That Influence Metals
        Behavior

A number of parameters influence the behavior of met-
als including:

•  Physical and chemical form

•  Feed rate and characteristics
•  Propensity to fragment

•  Chlorine or halogen concentration
•  Combustion zone parameters
•  Air pollution control equipment parameters

•  Other parameters of secondary importance

EPA's strategy for controlling metals emissions from in-
cinerators and boilers and industrial furnaces (BIFs) has
been to develop regulations and guidelines for setting
permit limits  on these and  related  parameters (U.S.
EPA, 1992). A more detailed discussion of specific per-
mit limits with regard to specific types of combustion
devices is found in Chapter 4.

The parameters of most concern are specific to the
types and forms of metals in the waste. The physical
and chemical forms of the metal are not regulated per
se, but they are indirectly regulated by the guidelines
describing how to spike metals or design a test. When a
test is being designed and the metals are spiked, ideally
the metals used in the test should be in a physical and
chemical form that most closely resembles the actual
form in the waste. This resemblance is  important be-
cause the form affects the volatility and the propensity
for entrainment of the waste.

Since  individual metals react differently, an individual
limit for the feed rate of each toxic metal is required. As
more metals are put into  the incinerator, more can be
emitted. The size distribution of the metals in the waste
is  important, because the smaller the particulate, the
more likely it is to become entrained into the gas stream
and the less prominent are the diffusional restraints on
vaporization.
Other  factors to consider are  propensity to fragment
and the concentration of chlorine or other halogens in
the waste. For many metals, chlorinated compounds
are more volatile than nonchlorinated compounds; thus,
the more chlorine in the waste, the more likely that met-
als emissions will occur.

Metals emissions are most influenced by parameters
that affect formation and control of fine particles, such
as combustion zone parameters. The most important of
these is the temperature  of the combustion zone; the
hotter the zone, the more likely that metals will vaporize
to produce fine particles. Numerous air pollution control
device parameters also must be considered including
those that control fine particle capture (which are specif-
ic to the type of device) and temperature.

A number of other parameters influence metals emis-
sions but are of secondary importance. These include
combustion gas velocity and  oxygen  concentration.
Combustion gas velocity influences entrainment and
must be evaluated because it is an organics-destruction
related parameter. Oxygen concentration also typically
is not considered a metals-related parameter. Although
some metals under certain circumstances  are some-
what more volatile in a reduced-oxygen environment,
usually this is a secondary effect. Some sulfides  are
                                                  63

-------
more volatile than oxides or other forms of the metals,
but this effect is dwarfed by the effect of halogens.
A host of very technology-specific parameters, such as
the rotation speed of a rotary kiln, also could have some
small effect on  metals emissions, but generally are not
considered metals-related parameters.

5.1.3 Air Pollution Control Devices
Two aspects of APCE are important with respect to met-
als: (1) the overall effectiveness of the device, and  (2)
the effectiveness of the device in catching fine particu-
late. Vaporized metals tend to form fine paniculate,
which is difficult to capture in most APCE. The degree
to which the device captures fine particulate can deter-
mine  its effectiveness in controlling metals emissions.
Although the efficiency of APCE is system dependent,
generalizations can be made regarding the  effective-
ness of different types of devices in capturing fine par-
ticulate. Fabric filters and other filtration-type devices
are usually the  most effective types of APCE for captur-
ing all sizes of  particulate, including fine particles (see
Figure 5-3). In  general, the efficiency of fabric filters is
least affected by particle size. Electrostatic precipitators
(ESPs) are somewhat more influenced by the particle
size; they are less effective than fabric filters at captur-
ing particles in the  1  to 0.5 urn-size  range. Venturi
scrubbers are very effective at capturing coarse parti-
cles (i.e.,  particles greater than 5 urn in size), but are
ineffective at capturing submicron particles.
To choose an appropriate air pollution control device,
the owner/operator must consider the kinds of metals to
be  incinerated  and their form (vapor, fine particulate,
coarse particulate) (see Figure 5-4). For example, if the
primary metal is  mercury, which is so volatile that it
could be a vapor when it enters the APCE, particulate
capturing control devices will be ineffective. Instead,
APCE such as scrubbers, which operate at very low
temperatures, could be chosen so that as much mercu-
ry as possible is cooled to the point where it condenses.

An  adsorption technique such as  carbon  adsorption
also would be appropriate for mercury. The literature in-
dicates that activated carbon injection effectively en-
hances mercury removal (Durkee and Eddinger, 1992;
Shoner, 1992). Tests conducted in medical incineration
facilities show that activated carbon is 85 percent effec-
tive in removing  mercury. Ongoing investigations are
evaluating possible additives so that a combined filtra-
tion/adsorption system might effectively  remove mer-
cury in  the future. For example, injection of activated
carbon into or upstream of the bag house may enhance
collection of mercury vapor through adsorption and mer-
cury particulate through filtration.
EPA recently has published  reference air  concentra-
tions (RACs) for nickel and selenium, and the guidance
now will include nickel and selenium as toxic metals that
have to be considered in the trial burn (U.S. EPA, 1992).
Although EPA is investigating the volatility of both nickel
and selenium, the preliminary findings are that selenium
is quite volatile,  in the range of antimony  (Sb), while
nickel is less volatile, probably in the  range of barium
(Ba) (see Figure 5-4).
Removing hexavalent chromium from emissions is a
topic of current research. Preliminary work  on its ther-
modynamics  suggests that hexavalent  chromium  is
very volatile, especially in its chloride,  oxychloride, and
                Increasing
                 velocity
Metal
Hg
As
Sb
Tl
Cd
Po
Ag
Ba
Be
Cr
Likely Form at APCE



	 1


—

__
I 	 Vapor

	 Fine particles





Most Effective APCE
Adsorber,
scrubber

Filter, ESP


Any APCE for
particulate matter
control

Figure 5-4.  Guide for selecting APCE, based on volatility and particle size of metals.

                                                    64

-------
 oxyhydroxide forms. In nature, the trivalent oxidation
 state of chromium seems to be the most stable form.
 But significant formation of hexavalent chromium is like-
 ly to occur at high combustion temperatures. Moreover,
 equilibrium calculations suggest that significant hexava-
 lent chromium formation may occur in the presence of
 chlorine even at very  low temperatures, although the
 reaction rates may be too slow to be of concern. If chlo-
 rinated hexavalent chromium is present in the flue gas,
 it likely will be in vapor form as it passes through the
 APCE. Consequently, dry air pollution control devices,
 such as bag houses, do not control hexavalent chromi-
 um very effectively, while wet devices, such as  scrub-
 bers and packed towers, are generally more effective.

 5.2   Trial Burns

 The objectives of the trial burn  are (1) to demonstrate
 that a facility can meet emission standards, and (2) to
 establish under what conditions those standards can be
 met. These conditions  established in the trial burn be-
 come the permit conditions under which the facility must
 operate. To meet these objectives, the test should be
 conducted under the "worst-case" conditions that will
 allow the unit to pass the emissions standards.

 5.2.1  General Considerations
 In planning the trial burn, the owner/operator needs to
 consider:
 • The facility's operation under normal conditions.
 « The facility's operation under worst-case conditions.
 • The appropriate testing tier for metals.
 • The form of the metal in  different stages and its
   teachability.
 The owner/operator should first define the desired per-
 mit conditions (i.e., operating ranges and waste feed
 variations) for the facility by determining how the facility
 operates under normal  circumstances, and how the fa-
 cility will be operated in the future. In this  stage, the
 owner/operator should fully characterize the  facility and
 develop a body of historic data that  accurately  repre-
 sents the facility's operation.
 The goal is to run the test under worst-case conditions
 such that the resulting permit will be flexible enough to
 allow the owner/operator to run the facility as he or she
 desires, while complying with the appropriate regula-
 tions. To accomplish this, the owner/operator must de-
 termine what controls will be needed to  achieve
 worst-case conditions and how they can be achieved for
 all parameters simultaneously. In many  situations, all
 parameters cannot be operated under worst-case con-
 ditions at once. Therefore, the owner/operator must de-
velop a plan describing how many tests must be  run to
demonstrate all of the various worst-case conditions,
and which conditions will be demonstrated in each test.

The next step is to select the appropriate testing tier for
metals. (Metals tiers  are discussed  in more detail  in
Section 5.2.3.)  Then, based  on the testing tier, the
amount and form of metal to be spiked and the neces-
sary feed rate limits can be determined. The form of the
metal used in these tests should be as similar as possi-
ble to the typical form of metal  in the waste. Depending
on the metal, and the form of the metal needed, spiking
of metals can be expensive, even  prohibitively so.
Some metals cannot be found  at any price in the quan-
tity needed. Therefore, the owner/operator must take
this into account when designing the test to ensure that
the test plan is reasonable.

Usually, metals that are difficult  to find for spiking are
ones that are relatively rare. When facility owner/opera-
tors project the amount of a metal they will be incinerat-
ing in the future, they tend to overpredict the burning of
rare metals, and then try to  purchase excessive quanti-
ties of these metals for use in the trial burn.

The owner/operator must decide the best way to re-
solve tradeoffs such as those among the expense of the
trial burn, the conservativeness of the high feed rate lim-
its, and the flexibility permit writers will allow concerning
the forms of metals used in  the test. For example, spik-
ing an organometallic waste  into an organic waste
stream is not always  necessary. Organometallic com-
pounds are typically very expensive and quite toxic.
Usually, the metal in the waste  stream is not an organo-
metallic, so an alternative spiking technique might be
found. If the metal is not obtainable, the owner/operator
should consult the permit writer to develop an alterna-
tive.

The form and teachability of the metal in the fly ash, kiln
dust, or solid residual product  are additional consider-
ations. Although teachability does not directly affect the
trial burn  (no intrinsic teachability limit is set),  it does
affect subsequent disposal  of solid residue and ash. If
the facility owner/operator plans to have the solid resi-
due  delisted to avoid disposing of  it as  hazardous
waste, the residue has to meet certain criteria, including
those concerning the  teachability of the metals in that
waste. The solid residual waste must  meet a similar re-
quirement to  qualify as nonhazardous under the Bevill
amendment (RCRA section 3001  (b)(3)(A)(i-iii))  (see
Section 1.2.5). Cement kiln dust and cold boiler fly ash,
for example,  are typically considered  nonhazardous,
but that status  depends on, among  other things, the
teachability of metal in the waste. The limiting factor on
the amount of metal to spike in the trial burn might de-
pend on how much can be spiked while still retaining the
nonhazardous designation for the residual, rather than
on the  metal feed concentration expected in the future.
                                                  65

-------
5.2.2   The Test Plan

5.2.2.1  Developing the Test Plan
Once the owner/operator has considered  all of the
above issues, he or she must prepare the test plan. One
of the most important first steps is to discuss the test
plan with the permit writer. The permit writer needs to
review and  understand the test plan, and the facility
owner/operator and the permit writer must reach an
agreement before the test is conducted. If those condi-
tions are met, the permit writer's test requirements will
be satisfied, and the facility's permit will contain the ex-
pected conditions.
The metals feed rate permit parameter is treated differ-
ently in the new rules than in  the old. The BIF rule uses
a box approach (Figure 5-5), which establishes sepa-
rate limits for each metal in pumpable waste, all hazard-
ous waste, and all feed streams based on the results of
the trial burn or compliance test. First, a limit is estab-
lished on the feed rate for each of the 12 hazardous
metals  in the facility's pumpable hazardous waste,
based on the feed rate used in the  trial burn.  Second, a
feed rate is specified for each metal in all hazardous
waste, which includes pumpable hazardous waste and
nonpumpable or solid hazardous waste. Third, a feed
rate limit is set on the combined amount of each metal
from all of the feed streams to the facility, including pum-
pable hazardous waste, nonpumpable hazardous waste,
fuels, raw materials, and anything else that goes into
the unit.
The most conservative way  to spike metals in a trial
bum is to spike them all into the pumpable waste. Using
pumpable hazardous waste  is the most conservative
spiking method because all metals in the pumpable haz-
ardous waste are atomized.  Thus, metals  emissions
from pumpable  hazardous waste will be somewhat
                  Box In a box in a box
                 Metals feed rate limits


Pumpable waste
All hazardous waste

                   All feed streams
Rgure 5-5.  BIF rule approach to setting metals feed rate limits.
greater than those from other feed streams. Some facil-
ities, however, might be unable to meet the emissions
standards if they spike all of the metals into their pump-
able hazardous waste. The owner/operator needs to
plan the trial burn with respect to the box limits.  The
feed rate limit for each metal in each "box" needs to
allow for the facility to burn its current and future waste
load under reasonable operating conditions. If metals
emissions are of little  concern for a facility, and it will
pass emissions standards easily, the best approach is
to spike all of the metals into the pumpable hazardous
waste.

5.2.2.2 Case Study: Developing  a Plan for a
       Fluidized Bed Unit

Planning a trial burn for a particular fluidized bed unit is
described in this section. The ash content of this facili-
ty's waste was relatively high, and as more waste was
fed into the system, the ash tended to stay with the  bed.
As a result, the ash in  the bed tended to build up,  until
eventually the ash in  the bed became  the bed itself.
Once a week, the operators removed material from the
bed, and once a year they recharged the bed by install-
ing a new sand bed. Thus, this fluidized  bed unit had a
very low recharge rate.

The permit limits  were somewhat different for this
facility. The permit specified a  maximum temperature
limit, which is standard; the higher the temperature, the
more vaporization. Thus,  this facility's maximum  tem-
perature limit was a worst-case condition. Maximum flu-
idizing air flow also was considered a permit parameter,
because of its impact on metal entrainment.The greater
the airflow, the greater the entrainment, and, thus, the
greater the emissions. The facility needed to conduct
tests to determine the maximum chlorine feed rate and
the maximum waste feed rate, although  the latter is an
organics-related, not a metals-related, parameter.

The Venturi scrubber settings also were permit parame-
ters for this facility, including the pressure drop across
the Venturi, the liquid-to-gas ratio, and  the blowdown
rate for the Venturi. The maximum metals feed rates,
another standard permit condition, also are specified.
These feed rates are very important, because the emis-
sions of volatile metals relate directly to the feed rate of
those metals. Volatile  metals tend to vaporize quickly
and escape from the bed.  Thus, as the volatile metal
feed rate increases, the amount of metal  available to be
vaporized increases, leading to increased emissions.

The maximum metals feed rate, however, does not ad-
dress the nonvolatile metals. If these metals are spiked
into the fluidized bed at high concentrations,  the  con-
centration of those  metals in the  bed may not reach
equilibrium for weeks.  This situation happens because
nonvolatile metals fed into the system tend to stay in the
                                                   66

-------
bed and build up, until eventually some of the entrained
metal particles become so small that they are entrained
in the gas stream. The  amount of metals entrained de-
pends on the concentration of metals in the bed. Most of
these entrained particles would be captured by the Ven-
turi, but a Venturi is not 100 percent effective. There-
fore, the emission of nonvolatile metals is a concern.

On a short-term basis, the concentration of metals in
the bed does not necessarily reflect the metals concen-
tration in the feed, because the feed metals will build up
until, after a long time, they reach equilibrium. This facil-
ity, however, may take weeks  to reach an equilibrium
concentration of nonvolatile metals in response to a
change in the metals feed  rate, yet a trial burn is too
expensive to run for weeks. Therefore, an additional
constraint was placed on this facility. The permit set a
limit  for the maximum  concentration of metals  in the
bed, which would be measured periodically. As long as
the facility does not exceed its permit limit as demon-
strated in the trial burn, the facility is in compliance.

The facility may have  difficulty, however, in achieving
the desired concentration of metals in the bed in a trial
burn, especially if the facility is limited to 720 hours of
hazardous waste burning. Owners and operators were
required to certify compliance with applicable interim
status standards, which are all of the performance stan-
dards except those for destruction and removal efficien-
cy (ORE)  (i.e.,  particulate, metals, hydrogen chloride,
chlorine, carbon monoxide, and hydrocarbons and diox-
ins/furans, if applicable), by August 21, 1992. Those
unable to meet this deadline became eligible to receive
an automatic extension, which applies for a year and
limits the facility to 720  hours of hazardous waste burn-
ing during that time or the amount of time needed to
conduct compliance testing, whichever is smaller. In
720 hours, however, the metals concentration in the bed
probably cannot be increased to the maximum concen-
tration projected by  merely  spiking the feed streams.
Spiking the bed might be the easiest approach, assum-
ing the permit writer has determined that maximum con-
centration must be demonstrated.
If an owner/operator can present a good scientific argu-
ment that shows that the concentration of metals in the
bed can be credibly related to the concentration of met-
als in emitted particulate, the permit writer may not re-
quire further testing.  One  approach to relating the
metals concentrations  in the particulate emissions to
those in the bed is to assume that metals in the bed are
released by a grinding mechanism (not volatilization)
that creates relatively coarse particulate.  The particu-
late emission rate then can be related to the maximum
emission rate from  the bed  for those metals, and the
removal efficiency for the particulate control equipment
can be calculated. Another approach is to assume the
metals concentrations in emitted particulate are similar
to the concentrations in the bed. In that case, particulate
can be assumed to be emitted at the emissions limit,
and the bed concentration can be calculated. Permit
writers, however, rarely accept the latter approach.

5.2.3  Metals Tiers

The four metals tiers are Tier I, Tier II, Tier III, and Ad-
justed Tier I. As one moves from Tier I to Tier III, the
analysis is progressively more complex, while the re-
sults are less and less conservative. The three parame-
ters of concern in choosing the metals tier are:

• The degree of sophistication in the dispersion model-
  ing.

• Whether the facility can take credit for the effective-
  ness of the incinerator and air pollution device in re-
  moving metals or particulate matter.

• The degree 6f conservativeness of results that the
  facility can tolerate.

Tier I is the simplest tier. No credit is given for removal
of metals in either the incinerator or the air pollution con-
trol device. All of the metals that go into the device are
assumed to come out of the device via the stack. Dis-
persion from the stack to the maximally exposed individ-
ual (MEI) is based on a very simple  and  conservative
dispersion model. EPA has provided the results of this
model in the form of tables in an appendix to the BIF
rules  (U.S. EPA,  1992). (This model is discussed in
more detail in Section  4.4.1.) The calculations are very
simple to perform, but the results are very conservative.
The resulting limits, in most cases, will not be accept-
able for some of the carcinogenic metals.
For Tier II, a trial burn is conducted to determine the
removal efficiency of the system in either the combustor
or the air pollution control device. The facility then can
take credit for this removal efficiency. The emission lev-
els are measured; all the metals in the waste that go into
the unit are  not assumed to come out the stack. The
conservative EPA tables still are used, however, to de-
termine the dispersion from the stack and the expected
dose to the MEI. This tier is more expensive to use than
Tier I in that a trial burn is run and the concentrations of
all of the metals are measured. The results are some-
what less conservative than those achieved using Tier I.

Tier III provides the complete picture. A trial burn is con-
ducted to generate data on which to base the credit for
metals removal in the combustor and in the air pollution
control device. Site-specific dispersion modeling also is
conducted to determine the concentration to which the
MEI is exposed. Tier III takes many  more factors into
account and the result is more realistic and less conser-
vative, allowing for more flexible, higher feed rate limits.
It is the most expensive tier, however, because both a
trial burn and dispersion modeling have to be conducted.
                                                   67

-------
 In the fourth tier, Adjusted Tier I, a trial burn is not neces-
 sary because credit for air pollution control is not al-
 lowed. A detailed site-specific dispersion model is used,
 however, so credit is awarded for more dispersion.

 In Figure 5-6, the assumptions and parameters involved
 in choosing the four metals tiers are summarized.

 Each metal can be evaluated using a different tier. Most
 facilities, unless they have very low emissions, will need
 to use a detailed dispersion model for at least one met-
 al. Tier III probably would  need to be run for the most
 toxic metals, while Tier I would be suitable for others.
 Tier III would require a more limited trial burn in which
 certain metals would have  to be spiked and measured.

 ATier II or III approach is appropriate for boilers even if
 they have no air pollution control devices. Metals parti-
 tioning can occur within the system through deposition
 onto the surfaces of the tubes and in the bottom ash, so
 the emissions could be lower than the feed rates. EPA
 will accept such data, however, only if the facility has a
 discernable removal mechanism for lowering emissions
 (e.g., a bottom ash mechanism).

 5.2.4  Multiple Tests

 In many cases, trial burns have to be run under multiple
 sets of test conditions. This requirement is necessary to
 test different sets of permit conditions for different oper-
 ating modes or to test multiple worst-case  situations
 that cannot be created simultaneously.

 5.2.4.1  Operating Modes

 Each mode of operation yields its own independent set
 of permit parameters.  For  example, some incinerators
 operate either in the liquid injection mode or in the solid
                                      waste burning mode, but not in both modes at the same
                                      time. The waste feed rate limits required for the liquid
                                      injection mode could be much lower than those needed
                                      for  the  solids  burning mode.  Under these circum-
                                      stances, the owner/operator might decide to run two in-
                                      dependent trial burns and to obtain two different sets of
                                      permit conditions, giving the  facility more flexibility in
                                      operating the system. One disadvantage of this ap-
                                      proach is that two trial burns are more expensive to run
                                      than one. The recordkeeping  also is more difficult, be-
                                      cause at any given time the mode of operation has to be
                                      documented and the appropriate permit limits applied.

                                      5.2.4.2 Conflicting Parameters

                                      Conflicting parameters are those that cannot be operat-
                                      ed simultaneously at worst-case conditions. An obvious
                                      example  is  setting  a maximum temperature limit for
                                      metals emissions and a minimum  temperature limit for
                                      organics  emissions;  a facility cannot  operate at both
                                      maximum temperature and minimum temperature simul-
                                      taneously. If the permit limits are set from the same test
                                      condition, only a very narrow range of  operating condi-
                                      tions will be allowed. Facilities, therefore, are allowed to
                                      run additional tests for conflicting parameters.

                                      Two sets of tests usually are run. The limits are  deter-
                                      mined for as many different parameters as possible us-
                                      ing  one set of operating conditions. Then the limits for
                                      conflicting parameters  are set at  an additional  set  of
                                      operating conditions in which only the conflicting param-
                                      eters are varied.

                                      The constraint to separate testing is that certain param-
                                      eters must remain the same for both sets of tests. The
                                      feed rate of metals,  chlorine, and  ash cannot  be
                                          Tier II: Emissions
                                          Trial burn required;
                                          conservative general
                                          dispersion tables
Tier I: Feed
Assumes all metals escape;
no trial burn required;
conservative general
dispersion tables
                        d
                 a
                                                                      Tier III: Dispersion
                                                                      Trial burn required;
                                                                      site-specific dispersion
                                                                      model
                                                                      Adjusted Tier I:
                                                                      Assumes all metals escape;
                                                                      no trial burn required;
                                                                      site-specific dispersion
                                                                      model
Figure 5-6.  Assumptions and parameters regarding the four metals tiers.


                                                    68

-------
 changed from one set of tests to the next. Other param-
 eters should be varied as little as possible.

 5.2.5  Waste Sampling and Analysis

 Waste and gas stream emissions both need to be sam-
 pled and analyzed. To sample and analyze waste, the
 sample is  collected and then digested to remove the
 metals from the waste. Then the metals are put into a
 solution for injection into an analytical device, such as
 an inductively coupled argon plasma (ICAP) emission
 spectrometer  or  an atomic  absorption spectrometer
 (AAS).  For mercury and  a few other compounds, a
 variation of atomic absorption called cold vapor atomic
 absorption spectroscopy (CVAAS) is used. A lot of con-
 troversy surrounds the best way to digest various types
 of  waste.  The EPA-approved  digestion techniques
 specified in SW-846 (U.S.  EPA, 1986) are not totally ef-
 fective at digesting all of the metals in the waste or feed
.streams. Some researchers have claimed that ASTM
 hot-acid extraction techniques, and especially a micro-
 wave extraction variation, are more effective than EPA-
 approved  methods for  recovering  metals.   These
 techniques, however, utilize hydrofluoric  acid, which
 presents safety concerns.  The  BIF  rule (40  CFR
 266.100 (c)(1)(ii)) allows  the use  of such alternative
 methods so long  as they meet or exceed the SW-846
 method performance capabilities.
 Sampling and analysis can have a significant effect on
 trial burns  in which the waste is being spiked. When
 planning a  trial burn, the owner/operator must ensure in
 advance that the metals can be measured in the form in
 which they will be  introduced into the system. If the
 amount of metal measured is less than the amount add-
 ed, for example, the feed rate limits for the metals in the
 operating permit will be much lower than those used in
 the trial burn.  To  ensure proper quantification, owner/
 operators should work with the analytical lab conducting
 the sampling and analysis to ensure that pretesting is
 conducted  and that the sampling  and analysis  tech-
 niques are accurate.

 The sampling  and analysis of gases are more straight-
 forward. The techniques are defined clearly in the meth-
 ods manual published as part of the BIF rule, in the BIF
 implementation document (U.S. EPA, 1992), and in
 SW-846 series 7000 methods  (U.S. EPA, 1986) (40
 CFR part 60, AppendixA). Particulate matter is sampled
 isokinetically using  Method 5, which involves use of a
 filter/impinger train (U.S. EPA, 1985).

 All metals, except for hexavalent chromium, are  sam-
 pled using  the new multiple  metals train, which is dis-
 cussed  in the  BIF methods manual (40 CFR part 266,
 Appendix IX).  In this modification of the old Method 5
 train, impingers in the train are filled with acid and oxi-
 dizer solutions to capture  the various kinds of metals.
Analysis is by ICAP or AAS except for mercury, which is
analyzed by CVAAS.

Hexavalent chromium is sampled using a completely
different probe—a recirculating impinger train filled with
potassium hydroxide solution. The sample is analyzed
using a different analytical technique—ion chromatog-
raphy with a post-column reactor (40 CFR part 266, Ap-
pendix IX).

5.2.6   Metal Spiking

Permit conditions are based on the metal feed rates
demonstrated in a trial burn. To ensure that the feed rate
for each metal in the trial burn is as high as it might be in
the future, the metals are generally spiked to obtain arti-
ficially high concentrations in the feed rate. The BIF im-
plementation document contains guidelines for spiking
metals (U.S. EPA, 1992).

The spiked metals should be in a form as similar as pos-
sible to that found in the waste. For solids, the spiked
metals should occur in a particle size as fine as the typ-
ical particle size in the waste. Ideally, the  metals should
be mixed thoroughly with the waste.  In most situations,
however, such  mixing is not practical,  and the usual
practice is to spike solids with discrete packets of metals.

Liquid wastes typically are spiked with aqueous solu-
tions of metals. Ideally, organic wastes should be spiked
with organic solutions of metals, but finding metal forms
soluble  in organics is often difficult. In these situations,
an aqueous solution of waste can be spiked  into an or-
ganic stream, if it is spiked and metered continuously. A
batch cannot be mixed in advance because it will sepa-
rate between the streams, making determination of the
feed rate at any given time very difficult. In addition,
good in-line mixing between the point of spiking and the
point of injection is essential.

If an aqueous metal solution is needed, the appropriate
form will vary from metal to metal, but typically the chlo-
rides and nitrates of the various metals are relatively
soluble in water. Oxides typically are not  soluble in wa-
ter. (Oxides also are the cheapest form of most metals,
so they are frequently used for solid  spiking.) Often
a  metal that is  soluble in the liquid  a testing situation
requires cannot be found, so  many spiking suppliers
have developed suspensions of metals in both inorganic
and organic matrices. Typically, these suspensions are
formed by suspending an oxide of the metal in the me-
dium. For example, paint is  a stable suspension con-
taining metals that can be injected into the system as a
liquid. Organometallics are generally soluble in organic
solutions, but they are very expensive and often dan-
gerous to handle.  Unless unavoidable, organometallics
should not be used. Some inorganic compounds are
soluble in mineral spirits  or other organic  solvents.
                                                   69

-------
 Chromium is spiked differently than all the other metals,
 because it occurs in two forms: the trivalent form, which
 is more stable and generally more prevalent in both the
 feed and the emissions; and the hexavalent form, which
 is a highly toxic carcinogen. EPA limits chromium emis-
 sions based on the assumption that all emitted chromium
 is in the hexavalent form, unless the facility operator spe-
 cifically measures emissions for hexavalent chromium.
 Measuring for hexavalent chromium is often advanta-
 geous, because hexavalent chromium emissions typi-
 cally are much lower than total chromium emissions. If
 hexavalent chromium emissions are measured in the
 trial burn, two different types of chromium feed rate lim-
 its will result:  one for hexavalent chromium based on
 the amount of hexavalent chromium in the feed in the
 trial burn, and the other for total chromium based on the
 amount of total chromium in the feed in the trial burn.
 Thus, spiking hexavalent chromium in the trial burn may
 be necessary. Spiking  can cause problems, however,
 especially in an organic waste where hexavalent chro-
 mium tends to break down very quickly.

 5.2.7  Alternate Metals Approaches for
        Facilities That Recycle Collected
        Particulate Matter
 Another aspect of planning a trial burn is the treatment
 of recycled particulate matter. In a cement kiln or other
 type of industrial furnace with similar processes, the
 device recycles the captured particulate matter into the
 process. Consequently, the change in metals emissions
 lags behind a change in the metals feed  rates. For this
 reason, EPA treats facilities that recycle particulate mat-
 ter differently from other boilers and industrial furnaces.
 EPA allows facilities to use one of three options (40 CFR
 part 266.102(c)(3)(ii)), the first two of which are general-
 ly impractical.
 The first option is to measure the concentration of metals
 in the waste kiln dust daily, and to run a series of tests
 during the trial burn to correlate the concentration in the
 waste kiln dust with the actual metals emissions. On the
 basis of this correlation, a limit is set on the allowable
 concentration of each metal in the waste kiln dust. Ev-
 ery day the concentration of metals in the dust must be
 measured to ensure that the limit is not exceeded. In
 addition, quarterly stack sampling is required to verify
 that the correlation has not changed. Because numer-
 ous  additional constraints  also  are attached  to this
 methodology,  most facilities  have not used this  ap-
 proach.
The second alternative is to measure the concentration
of metals emitted from the stack by taking multiple met-
als samples of stack emissions for 4 hours every day
when hazardous waste is being burned. The goal is to
document that emissions are below the emissions lim-
 its. This simple-to-understand methodology is very ex-
 pensive, and very few facilities have elected to use it.

 The third alternative, and the one most often used, is to
 establish equilibrium prior to running the trial burn. The
 facility is run under trial burn conditions long enough in
 advance to  ensure that  the  metal  emissions have
 reached steady state when the trial burn is run. Then
 the trial burn will be a conservative measure  of metals
 emissions. The constraint on this method is that the fa-
 cility must be operated at the trial burn test conditions,
 including all the metals feed rates, for a period of time
 sufficient to establish equilibrium prior to the test run.

 The best way to determine  how long the combustion
 device needs to run to reach equilibrium is to conduct a
 simulated trial burn for an estimated time, collect sam-
 ples over the course of  the burn, and send them to a
 laboratory for analysis.  When  the sample  results are
 available (after about 2  weeks), the length  of time the
 unit took to reach equilibrium can be  estimated based
 on the metal emissions  or the dust collected  in the air
 pollution  control system. If the determination  is based
 on dust collected by the air pollution control system, the
 owner/operator needs to provide EPA with  data docu-
 menting the correlation between the dust and  the stack
 emissions. This correlation demonstrates that the con-
 centration in the dust would reach equilibrium at the
 same time as would the  concentration in the stack.

 In the BIF rule, equilibrium is defined as the point at
 which  emissions  reach  90  percent  of their ultimate
 steady-state value. This point can be determined based
 on data collected over time and subjected to a least
 squares fit regression to generate a curve (see Figure
 3-8). The functional form and shape of the curve may
 vary from one facility to another. From the curve, the
 ultimate concentration and the time required to reach it
 can  be determined. In  general,  equilibrium  does not
 have to be determined for all the metals, usually only for
 those metals expected to take the longest time to reach
 equilibrium, which typically are the most volatile. Own-
 er/operators, however, should come to  an agree-
 ment with the permit writer on this subject.

 The time required for a unit to reach equilibrium depends
 on the particular type of device and its operating practic-
 es. A wet process cement kiln, for example, takes con-
 siderably less time to reach equilibrium than does a dry
 process kiln, because the latter has an internal recircu-
 lation route where condensation forms on the preheated
 particulate or raw materials. The time required to reach
 equilibrium also depends on what fraction of the waste
 dust is recycled;  the more that is recycled, the longer
the process takes. If the feed stock is variable,  the equi-
 librium is established statistically. The more variable the
feed stock, the more "noise" will be in the data, and the
 more samples will be needed.
                                                   70

-------
5.2.8   Case Studies

In this series of case studies, the test design maxims
discussed above are applied to particular hypothetical
facilities.

5.2.8.1  Case Study #1

In this first case study, a Tier I model is used. Based on
the results of the modeling and on the amount of metals
to be fed into this system, only lead needs to be evaluat-
ed beyond Tier I at this facility. The next step is to run a
detailed site-specific dispersion model (Adjusted Tier I),
the results of which indicate that lead remains a problem.
The operator of the facility concludes that a trial burn in
which  lead emissions are measured is necessary. To
maximize the lead feed rate, the lead is spiked in the
trial burn. First, however, a pretest is  run without any
spiking. (This step is not  mandatory, but is advisable.)
The feed rate and lead emissions are measured in the
pretest, and the system removal efficiency for lead is
determined as is the proportion of lead emitted from the
stack. The leachability of the bottom ash and fly ash are
evaluated to determine whether they  are  suitable for
delisting. At the same time, the economics of delisting
are reviewed to determine whether delisting should be
pursued.

The amount of lead to be  fed in the trial burn is chosen
based  on the expected concentration of lead in wastes,
emissions (based on system removal efficiency), and
bottom ash. Next, the operation of the system is charac-
terized by reviewing historical data and talking to opera-
tors. Consideration also is given to how the system will
be operated in the future. The goals are to determine
the desired operating envelope for all permit parameters
and to maximize the parameters that need to be maxi-
mized and minimize those that  need to be minimized.

On the basis of the characterization, this facility clearly
has conflicting parameters. A maximum temperature
conflicts with a minimum temperature, and also conflicts
with maximum combustion gas velocity. In this particu-
lar system, the high combustion gas velocity is achieved
by running high excess  air levels, which decreases
temperature. Therefore, the system cannot be run at
maximum combustion gas velocity and maximum  tem-
perature at the same time.

The facility owner/operator decides to run trial burns
under two conflicting conditions.  The first condition is
the maximum temperature condition in which the worst-
case values for all metals-related permit parameters are
run. The second condition is the minimum temperature
condition, which also is the maximum combustion gas
velocity condition and the worst-case condition for or-
ganics.
This two-condition strategy is based on the assumption
that maximum combustion gas velocity is not a metals-
related parameter, but is solely an organics-related pa-
rameter as implied by the BIF rules. If the permit writer
were to consider it a metals-related parameter, tests
might need to be run under three sets of conditions.

5.2.8.2  Case Study #2

The second case study is the first of two presented in
this section that focus on planning a trial burn and apply-
ing the collected data. In this case study,  the trial burn
plan originally submitted by operators of a rotary  kiln
facility with a Venturi scrubber was inappropriate.

The permit writers noted several issues of concern in
reviewing the trial burn plan. The first issue concerned
the selection of metals to test. This facility operator de-
termined what metals to spike in the trial burn on the
basis of the metals concentrations in the scrubber water
(blowdown). Only metals with high concentrations in the
scrubber blowdown were selected. No means were giv-
en, however, of relating the concentration of metals in
the scrubber blowdown to the feed rate or to the emis-
sions of the metals, and no determination was made as
to whether the expected emissions would be above risk-
based limits. This facility operator had already run a
site-specific dispersion model. A more appropriate ap-
proach would have been to use an Adjusted  Tier 1
approach, in which  any  metals having  a  desired feed
rate limit above the Adjusted Tier 1 feed rate limit would
need to be tested.

The second issue involved the use of surrogate metals.
This facility operator wanted to measure the emissions
of a subset of all of the toxic metals and use the results
as surrogates for the emissions of other  metals. This
approach raises two questions:

•  Is the use of surrogate metals acceptable?
•  Which metals can be used reliably as surrogates for
   other metals?
In this case, the permit writer did not approve the use of
surrogate metals and, in fact, permit writers rarely ap-
prove this procedure. Even if the permit writers had
approved surrogate metal use,  the metals the  owner/
operators proposed as surrogates were not necessarily
conservative surrogates of  the  metals that they were
not going to measure. For example, the owner/opera-
tors proposed cadmium as a surrogate for mercury, but
because mercury is  much more volatile than cadmium,
this choice would not yield conservative results.

The third problem arose in the determination of the met-
als feed rate. In the  trial burn plan, the facility operator
did not perform the calculations needed to demonstrate
adequately that the projected emissions would be below
                                                  71

-------
 the emissions limits. Once these calculations were per-
 formed, the trial burn was shown to be adequate.
 Finally, the trial burn plan did not provide detailed infor-
 mation as to how the metals would be spiked, an essen-
 tial component.  The trial  burn plan  always should
 include information on which metals will be spiked, their
 form, and the extent to which they represent the metals
 in the waste.

 5.2.8.3 Case Study #3

 The third case study is of a liquid injector system for
 which only volatile metals are a concern. In this facility,
 all the metals with feed rates above Tier I limits were
 volatile metals expected to totally vaporize in the injec-
 tor. The issue for the permit writers regarded the limits
 required of this facility. Based on thermodynamics, most
 of the metals would be expected to totally volatilize over
 the entire range of operating temperatures. Since tem-
 perature would have no further effect on volatility and
 hence on emissions,  setting a maximum temperature
 limit for the combustion chamber for this facility would
 seem to be unnecessary. Due to uncertainties associat-
 ed with thermodynamic predictions, however,  maximum
 combustion chamber temperature was retained as a
 permit limit. Thus,  the facility operator was required to
 run the metals portion of the trial burn at maximum tem-
 perature conditions.
 The next issue concerned the  maximum  combustion
 gas velocity limit. For a liquid injector with 100  percent
 entrainment, combustion gas velocity should not affect
 entrainment  or emissions; therefore, gas  velocity
 should not be required as a metals-related permit pa-
 rameter. Although maximum  combustion gas velocity is
 not required for metals, it is an organics-related permit
 parameter. Because it controls the minimum residence
 time for organics, this parameter should be retained as
 a permit limit, and the facility should be required to run
 the organics portion of the trial burn at maximum com-
 bustion gas velocity.

 The combustion gas velocity also may affect the partic-
 ulate  removal equipment. This facility has  a  Venturi
 scrubber. Maximum combustion gas velocity results in
 maximum pressure across the Venturi scrubber, which
 would be a best-case rather than a worst-case situation
 for capturing metals, and, therefore, not a parameter of
 concern for this facility. If the  particulate removal equip-
 ment consisted of an electrostatic precipitator, however,
the combustion gas velocity might be of concern.

The maximum chlorine feed rate limit also was ques-
tioned by the permit writers, again  based on the logic
that the chlorine would not be expected to affect volatil-
ization if the metals already were vaporized. The argu-
 ment for testing for chlorine is the same one used for
 testing maximum temperature—that the thermodynam-
 ics are uncertain. In addition, the chlorine could prevent
 some  volatile metals from condensing before entering
 the particuiate control system, which might affect sys-
 tem efficiency. Therefore, the maximum chlorine feed
 rate should be included as a test parameter. Further-
 more,  this parameter affects organics destruction effi-
 ciency, because chlorine typically is a flame inhibitor.
 Therefore, maximum chlorine concentration can be as-
 sociated with minimum organics destruction efficiency.

 For this facility, the results of the  test in which all four
 carcinogenic metals were maximized simultaneously
 were very close to the risk-based emissions limit for the
 sum of those metals. Each of the four metals was run at
 about  one-quarter of its individual risk-based limit, and
 the sum of the risk associated with these four limits was
 slightly less than all of the available risk. (The risks of all
 carcinogenic metals have to be added together.)

 After the test was run, the facility operator discovered
 that, to achieve the desired permit limits, the feed rate
 for chromium should have been doubled, and the feed
 rate for beryllium halved. Instead of running another
 test, the operator wanted to extrapolate from the results
 generated  during the first test  to  estimate emissions
 under  the new conditions. The  permit writers decided
 not to  allow extrapolation because of uncertainties in
 measurement techniques. Theoretically, carcinogenic
 metals emissions could be traded, but no  credible ex-
 trapolation technique is available  (see Section 5.3.2).
 At this facility, in particular, a high  degree  of accuracy
 was required because the results of the first test indicat-
 ed  that the facility was almost at 100 percent of its risk-
 based  aggregate limits.

 The facility may not have been required to test in the
 trial burn all carcinogenic metals, such as beryllium, that
 had very low anticipated feed rates. To achieve an ag-
 gregate risk-based limit for metals, the operator could
 assess each carcinogenic metal using a different tier.
 For example, a Tier III approach could be  used for chro-
 mium,  and an Adjusted Tier I approach  for beryllium.
The chromium emissions limit and feed rate would be
 based  on the emissions measured in the trial burn re-
quired  by Tier III. The risk calculated for beryllium, how-
ever, is based on Adjusted Tier I, which assumes that all
the beryllium that goes into the incinerator is emitted.
The assumed feed rate for beryllium, therefore, would
be the  basis for and identical to the assumed emission
rate of  beryllium.The risk based on these assumed feed
and emission rates then would be added  to that gener-
ated by the other metals in the feed—in this case, chro-
mium,  cadmium, and arsenic—to derive  an aggregate
risk-based limit.
                                                   72

-------
5.3   Setting Permit Conditions from Test
      Data
The three major topics regarding setting permit condi-
tions from test data are averaging, extrapolation, and
the use of surrogate metals. Extrapolation and the use
of surrogate metals are very controversial, as they are
ways of using test data to set permit limits at conditions
not evaluated in a trial burn.

5.3.1  Averaging
Permit levels are based on two types of averages: 1-
minute averages and rolling averages. According to the
BIF rules, parameters such as incinerator temperatures
have to  be monitored continuously. The data  points
must be collected at least every 15 seconds  and be av-
eraged on a 1-minute basis. For most parameters, how-
ever, the permit  limit is based  on  the hourly  rolling
averages achieved in three test  runs. For these runs,
data are recorded each minute. The 1-minute measure-
ments of the combustion chamber temperature for three
test runs are shown in Figure 5-7, for a hypothetical trial
burn. The temperature is nominally 1,800°F. The 1-minute
data contain a fair amount of noise, which is typical of
real data. In addition, a substantial drift occurs overtime
as the incinerator conditions change slightly.

The hourly rolling  average for any point in  time is  the
average of all of the data taken at that time and over the
previous 59 minutes. If the time is 2 p.m., the data gen-
erated from 1:01 p.m. through 2 p.m. would be used to
calculate the hourly rolling average. The rolling average
changes every minute, because in the minute just after
the period for which the rolling average was calculated,
new data are generated. To calculate the hourly rolling
average at 2:01 p.m., for example, data from 1:01 p.m.
are dropped from the calculation and the data generat-
ed from 1:02 p.m. through 2:01  p.m. are  used. The
                                       hourly rolling average tends to dampen the noise, be-
                                       cause it averages measurements taken over an hour,
                                       allowing general trends to become apparent.

                                       For each test run, rolling averages are calculated for
                                       each minute once the first hour has passed. The highest
                                       rolling average for each of the three tests then are aver-
                                       aged. That average is used to set a maximum  tem-
                                       perature limit. In Figure 5-7, the highest hourly rolling
                                       average for Run 1 was about midway through the test.
                                       in the second and third runs, the highest hourly rolling
                                       average was at the beginning of the tests. When using
                                       rolling averages, the more  data the better; if data are
                                       collected only for 1 hour, only one hourly rolling average
                                       can be taken.

                                       For parameters that are not linearly related, the hourly
                                       rolling average is not useful. For example, if the temper-
                                       ature of a combustor increases, organic destruction will
                                       improve only slightly. On the other hand, decreasing the
                                       temperature could drastically increase pollution or emis-
                                       sions. The hourly rolling average can be misleading in
                                       that it gives equal weight to each temperature, even if
                                       more metals emissions occur at high temperatures than
                                       at low temperatures.

                                       While the use of hourly rolling averages is challenged
                                       for scientific  reasons,  instantaneous  measurements
                                       (taken at 15-second intervals) have even more prob-
                                       lems in terms of practicality. Because of noise in the
                                       data and in instrument use, averages may be prefera-
                                       ble.  An in-between approach, such as the use of short-
                                       er term rolling averages for some parameters, might be
                                       the best method. As the regulations are implemented,
                                       EPA will determine the flexibility that can be used  in
                                       interpreting averages, but using a method in-between
                                       instantaneous and hourly rolling average is being con-
                                       sidered.
    Temperature
2,000


1,900


1,800


1,700


1,600


1,500
                    1-minute
                    temp.
                              Run 1
                                                         Max HRA
                                                       Run 2
                                                                                     Max HRA
     Run 3
                          60      120     180  0 '
                          Time (minutes)
                                       60       120     180 0
                                      Time (minutes)
60      120
Time (minutes)
               180
 Figure 5-7.  Hourly rolling average (HRA) of combustion chamber temperature in three trial runs.


                                                   73

-------
 5.3.2  Extrapolation
 Extrapolation, which is discussed in detail in the BIF im-
 plementation document, is not allowed under BIF inter-
 im status with one exception—to justify test conditions.
 The system removal efficiency based on data from one
 set of test conditions can be extrapolated to another set
 of test conditions only to justify running a test at those
 conditions. It cannot be used, however, to actually set
 permit conditions until the facility is beyond the interim
 status phase and in the permit phase, at which point the
 permit writer determines the use of extrapolation.
 Extrapolation can be executed in an upward or down-
 ward  direction. Upward extrapolation is used when the
 emissions from the trial burn are  well below the risk-
 based emissions limits, and  a higher feed rate is desir-
 able.  Scientifically, upward extrapolation makes sense,
 because the shape of the curve relating the metals feed
 rate to the metals emissions rate  is typically concave
 and facing downward (Figure 5-8).  At very low feed
 rates, the vapor pressure of the metal is sufficient to
 vaporize all of the metal fed into the system. Conse-
 quently, the more metal fed  into the  system, the more
 metal will vaporize, and the higher the emissions. Even-
 tually, most metals reach a saturation point  where the
 combustion gas cannot hold  any more metal  vapor and
 an increased metals feed  rate no  longer results in in-
 creased  emissions  rates.  As a result, the  emission
 curve plateaus and is concave. For nonvolatile metals,
 saturation is reached quickly; for volatile metals it oc-
 curs much later.
 For example, suppose a trial burn for a particular metal
 is conducted at a metals feed rate indicated by Point 1
 in Figure 5-8. At this point, the metal is unsaturated, and
Metals
omissions
rate
                                   Actual metals
                                   emissions curve
                Extrapolation
          Trial burn
          point
Emissions limit
                        3     5
   _                Metals feed rate
   Features
   • Theoretically conservative
   * Recommended for
    -Justifying test conditions to establish new limits outside of
      present permit limits
    - Situations where estimated emissions at extrapolated permit
      limit «risk-based emissions limit


Rgure 5-8. Upward extrapolation of metals emissions rate.
 its emissions are well below the metals emission limit.
 Theoretically, the emissions to feed rate ratio estab-
 lished in the trial burn could be used to extrapolate up-
 ward to Point 2, where the emissions would be equal to
 the emissions limit. A conservative metals feed rate limit
 thus would be established at Point 3. This limit would be
 especially conservative if the extrapolation went beyond
 the saturation point (i.e., the  break in  the emissions
 curve),  and the emissions actually reached the emis-
 sions limit at Point 4 (where the metals feed rate is at
 Point 5). This value would exceed the metal feed rate
 limit established at Point 3 by upward extrapolation.

 Because the data on which extrapolation is based are
 generated at lower feed rates (by definition since the
 extrapolation  is an upward extrapolation), the point to
 which the extrapolation is made always will be either on
 the curve or  above the curve. Consequently,  upward
 extrapolation results are conservative estimates, in the-
 ory. So much noise is associated with the data, howev-
 er, that extrapolation magnifies the uncertainty already
 associated with any given data  point. Therefore, extrap-
 olation should not be used with inherently noisy data to
 set an exact emissions limit.

 Extrapolation over a smaller range of values does make
 sense, however, even if the data contain noise. Extrap-
 olation would be a useful tool to a facility that has such a
 low feed rate of a certain metal that emissions are not a
 concern. Nevertheless, the feed rate is variable, and in
 3 months the waste might contain more of this metal
 than was contained in the waste used in the trial burn.
 As a result, the low-concentration waste must be spiked
 in the trial burn, even though the  emissions still will be
 well below the emissions limit. In this situation, extrapo-
 lation of data from one  low point to another may be
 allowed  to account for future variations of a very low-
 concentration metal, and thus relieve a facility owner/
 operator from running additional metals in a trial burn.

 Downward extrapolation raises different concerns, how-
 ever, because the curve  has a concave-downward
 shape (Figure 5-9). Usually a  facility owner/operator
 would want to extrapolate downward when the emis-
 sions in the trial burn exceed the risk-based limits, and
 he or she would want to determine the metals feed rate
 that will result in acceptable emissions levels. The prob-
 lem is that the characteristic concave-downward shape
 of the curve results in an estimate that is not conserva-
 tive. If the test were run at the new predicted feed rate,
 the emissions still would be above the emissions limit.
 Downward extrapolation, therefore, is not acceptable.

 For example, suppose a trial burn for a particular metal
 is conducted at Point 1 in Figure 5-9. At this point, the
 metal is saturated, and its emissions are above the met-
al emissions limit. The emissions to feed rate ratio es-
tablished in the trial burn could  be used to extrapolate
downward to Point 2, where the  emissions would be
                                                    74

-------
Metals
emissions
rate
                              Actual metals
                              emissions curve
               Trial burn point
                                  Emissions limit
                      5    3
                     Metals feed rate

               Features
               • Not theoretically conservative
               • Not recommended

Figure 5-9.  Downward extrapolation of metals emissions rate.
equal to the emissions limit, and thus establish a metal
feed rate limit at Point 3. However, this, would not be
conservative because of the concave down shape of
the emissions curve. As illustrated by Point 4, the actual
emissions at this extrapolated feed rate limit would be
above the emissions  limit. In addition,  the metal feed
rate at which the metal emissions are equal to the limit
(Point 5) would be well below the feed rate limit estab-
lished by downward extrapolation.

5.3.3  Surrogate Metals
Surrogate metals are metals used as conservative indi-
cators of emissions for other metals. Use of a metal as a
surrogate requires a reliable ranking system that indi-
cates the relative likelihood of a metal being emitted. At
present, the most reliable way of ranking metals is vola-
tility. More volatile metals are more likely to be emitted
than  less volatile  metals. At any given  temperature,
some metals, such as  barium,  are  quite nonvolatile,
while other  metals, notably mercury, are very volatile.
These metals can be ranked according to their volatility,
from  the least volatile to the  most volatile. Ideally, a
more volatile metal should be used as a conservative
indicator of a less volatile metal.

Using surrogate metals is a sensible approach for three
reasons:
• Decreased cost of  trial burns.  A trial  burn is less ex-
   pensive when surrogate metals are used, because
   fewer metals will have to be spiked, thereby decreas-
   ing spiking costs.
• Decreased environmental impact during trial burns.
   Using fewer metals also reduces the environmental
   impact of a trial burn, especially if nontoxic metals
   are used as surrogates for toxic metals.
•  Increased development of data. From the point of
   view of research and data collection, use of surro-
   gate metals promotes the development of data for a
   standardized metal "soup," if certain metals are al-
   ways used  in trial burns.  Such standardized data
   could be used to compare results from one trial burn
   to another and improve the ability of researchers  and
   EPA to compare facilities to determine how metals
   behave.
Some arguments against using surrogate metals, how-
ever, override arguments for their use. First, insufficient
data exist to justify a ranking scheme for metals based
on volatility. Some metals, most notably arsenic, are not
consistently volatile. Arsenic's volatility depends on  oth-
er factors, such as the presence of other metals. In con-
trast, barium has been used successfully as a surrogate
for beryllium, which is very expensive and not normally
available in the large quantities required for a trial burn.
The behavior of barium and beryllium is fairly predict-
able.
A second argument against the use of surrogate metals
is that most permit writers will not approve their use. A
great deal of documentation and justification  is required
for the approval of surrogate metal use. Any  use of sur-
rogate metals requires the collection of sufficient data to
support assumptions made regarding which  metals are
conservative indicators of other metals.

5.4   Uncertainties/Research Topics

Many uncertainties remain  concerning the behavior of
metals, including such crucial phenomena as the kinet-
ics of metals transformation. New information may af-
fect future regulations, changes, and guidance.
This chapter is based on the current understanding of
how metals behave thermodynamically. Although as-
sumptions based on thermodynamics probably are ac-
curate at very high temperatures, other factors  may
control the outcome as temperatures decrease. Situa-
tions may arise in which the reactions do  not occur
quickly enough, so that thermodynamic equilibrium is
never reached.
Failure to reach thermodynamic equilibrium is especial-
ly important with respect to chromium, where the  con-
cern  is not only how much chromium is emitted, but
whether it is emitted as hexavalent or trivalent chromi-
um. For example, current guidance, which is based on a
very limited set of EPA data, suggests that  hexavalent
chromium emissions depend on the amount of hexava-
lent chromium fed into the system; the more  hexavalent
chromium fed into a system, the more emitted.  This
suggests a  kinetic limitation on the conversion  of
hexavalent chromium to trivalent chromium, and  vice
versa.
                                                   75

-------
 Methods for measuring hexavalent chromium in solid
 and liquid matrices are in a state of flux. EPA is conduct-
 ing research to determine the best methods. To find out
 more about the status of these measurement methods,
 contact Larry Johnson,  Research Triangle Park, NC,
 919-541-7943.

 Another area of research concerns the behavior of ar-
 senic, which does not behave as predicted by thermo-
 dynamics.  Recent research, however,  indicates that
 arsenic behavior  can  be predicted more accurately if
 sufficient thermodynamic data on arsenic speciation are
 included in the thermodynamics data base  used for
 equilibrium calculations.

 The binding of metals  by earth elements (elements
 most common in the Earth's crust) such as calcium, alu-
 minum, and silicon is  another area of research. Earth
 elements also are the  most common metals in fly ash-
 es, slags, and cement.  The  volatility of many metals
 (e.g., arsenic) in the presence of earth elements is sig-
 nificantly lower than their volatility predicted by thermo-
 dynamics, or the volatility observed in the absence of
 earth elements.

 Clearer, more defined  guidelines for metal spiking are
 needed and require further research.

 Research is under way to develop on-line capability to
 monitor metals using lasers and other devices. If a de-
vice was available that could be inserted into a stack to
directly measure  metals emissions, operating limits
would not need to  be set and permit limits would be un-
necessary;  only a limit on metals emissions would be
required.Thus, permitting and testing procedures would
be greatly simplified. Such devices, however, will take
many years to develop. In the meantime, research will
continue to explore current waste sampling and analy-
sis methods in attempts to improve them.
 5.5   References

 When an NTIS number is cited in a reference, that doc-
 ument is available from:

   National  Technical Information Service
   5285 Port Royal Road
   Springfield, VA 22161
   703-487-4650

 Barton, R.G., W.D. Clark, and W.R. Seeker. 1990. Fate
   of metals in waste combustion systems. Combustion
   Science and Technology 4:327-342.

 Durkee, K.  and J. Eddinger. 1992. Status of EPA Regu-
   latory Development Program for Medical Waste Incin-
   erators. 1992 Incineration Conference Proceedings.
   Albuquerque, NM, pp. 447-456.

 Shoner, P. 1992. An innovative system for the emission
   control of heavy metals and dioxins.  1992 Incinera-
  tion Conference Proceedings. Albuquerque, NM, pp.
   129-132.

 U.S. EPA. 1992. U.S. Environmental Protection Agency.
  Technical implementation document for EPA's boiler
  and  industrial furnace regulations. EPA/530/R-92/
  011. NTIS PB92-154-947. Office of Solid Waste and
  Emergency Response, Washington, DC.

U.S. EPA. 1986. U.S. Environmental Protection Agency.
  Test methods for evaluating solid wastes: Physical/
  chemical  methods, third edition. EPA/SW-846. NTIS
  PB88-239223. Office of Solid Waste, Washington,
  DC.

U.S. EPA. 1985. U.S. Environmental Protection Agency.
  Modified  Method 5 Train  and Source Assessment
  Sampling System. Operator's manual. EPA/600/8-
  85/003. NTIS PB85-169878. Research Triangle Park,
  NC.
                                                 76

-------
                                           Appendix A
                    Draft Strategy for Combustion of Hazardous Waste in
                            Incinerators and Boilers and Furnaces
I.    Introduction

During the last decade, a dramatic transition in hazard-
ous waste management has occurred. Untreated haz-
ardous waste ceased to be placed on the land, and
widespread use of waste treatment technologies, in-
cluding combustion, ensued. We also began to under-
stand that even vigorously regulated and  enforced
hazardous waste management requirements cannot to-
tally solve our long-term waste problems.  Rather, our
long-term  national waste management strategy must
have reduction of waste as its first and primary goal.

EPA, the States, industry, and  the public have learned
much about the concept of waste reduction over the last
decade. Our challenge for the next decade is to take
these lessons and develop a strategy to accomplish our
goal of source reduction.
Source reduction is and will continue to be at the top of
our waste management  hierarchy and must be more
aggressively pursued to reduce the long-term demand
for waste management facilities. EPA intends to take a
fresh look at hazardous waste management issues as
part of moving towards the promise of pollution preven-
tion and source reduction. Specifically, in this  effort,
EPA's goal is to develop an integrated and  balanced
program for source reduction and waste management.
EPA will examine the appropriate roles of source reduc-
tion and waste treatment  in  the nation's hazardous
waste management system. EPA also  intends to reex-
amine its  existing  regulations and policies on waste
combustion.
This evaluation will be led by a committee of EPA and
State officials. This EPA-State Committee will first be
asked to address the relationship between hazardous
waste combustion facilities and source reduction of haz-
ardous waste, and to make recommendations on addi-
tional source reduction  opportunities  that should be
pursued. The Committee's second charge will be to ad-
dress how EPA could improve  its technical and permit-
ting rules  for  hazardous waste combustion facilities to
ensure that such facilities reflect the state-of-the-art as
well as  continued technological innovation. The Com-
mittee will also be asked to explore the development of
alternative waste treatment technologies, as well as the
need for better science in evaluating combustion tech-
nologies and monitoring emissions from combustion fa-
cilities.
As a starting point for this effort, EPA is issuing this
Draft Combustion Strategy. This document will serve as
a catalyst for discussion with and input from all interest-
ed parties on how best to  integrate source reduction
and waste combustion and on ways by which we can
better assure the public of safe operation of hazardous
waste combustion facilities.

This draft combustion strategy consists of a discussion
of the goals and objectives for this project and a series
of short and longer-term actions that can be taken to
achieve our goals. These actions are intended as the
starting point for discussions with the public and indus-
try. The list of actions in this document are presented for
debate and additional ideas.  However, while that dis-
cussion is taking  place,  EPA intends to aggressively
pursue several of the interim activities.

II.   EPA's Strategic Goals

A.   Background for the Goals

Combustion  is currently a large component of hazard-
ous waste management in the United  States. It has be-
come a large component as the  nation moved away
from land disposal in the 1980's and into treatment to
reduce the volume and toxicity of hazardous waste. As
this shift occurred  in the 1980's, citizens in areas where
incinerators or boilers and industrial furnaces (BIFs) are
located have increasingly challenged the need for these
hazardous waste combustion facilities. Citizens evi-
dence concern that waste combustion is too frequently
used where source reduction may be the preferred al-
ternative. Citizens also raise concerns regarding facility
siting and potential health risks posed by waste man-
agement facilities.
Hazardous wastes being burned today are generated
by major segments of American industry, and represent
a spectrum of commonly-encountered wastes, includ-
ing spent solvents, sludges and distillation bottoms, and
off-spec organic chemicals and products. About 5  mil-
lion tons of these highly organic wastes are being com-
                                                 77

-------
 busted each year—some 40% in incinerators and 60%
 in BIFs. Based on our most recent data, it appears that
 substantial excess capacity  exists  for combustion  of
 hazardous waste, particularly liquid wastes. It should
 also be recognized that  although some  additional
 wastes are untreated today, these wastes will soon be
 subject to treatment requirements mandated under the
 Hazardous and Solid Waste  Amendments of  1984
 (HSWA). These treatment  requirements could use up
 much of today's surplus capacity over the next several
 years.

 Incinerators and BIFs burning hazardous waste are reg-
 ulated by EPA and authorized states under the Re-
 source Conservation and Recovery Act (RCRA). EPA's
 incinerator permit regulations, first promulgated in 1981,
 control emissions of principal organic hazardous constit-
 uents (POHCs), hydrochloric acid (HCI), and particulate
 matter at incinerators. For interim status incinerators,
 only general facility regulations are in place. In 1988, the
 Office of Solid Waste (OSW) issued guidance to the
 ERA Regional Offices directing that, on a case-by-case
 basis under the omnibus provision in RCRA section
 3005(c), incinerator permits should be issued with major
 substantive improvements including controls on metals
 and products of incomplete combustion (PICs) and im-
 proved controls on HCI and acid gases.

 BIF facilities burning hazardous waste are all currently
 in interim status. These facilities—such as cement and
 lightweight aggregate kilns—are subject to EPA regula-
 tions adopted in 1991. These  regulations, among other
 things, impose emission controls for  metals, PICs, and
 HCI and acid gases that remain in effect until final per-
 mits are issued for these facilities. Currently, there are
 about 160 interim status BIFs, which are pending final
 determinations on their permits.

 Waste combustion has  been viewed as  a  means to
 detoxify  many hazardous  wastes,  particularly those
 containing high levels of organics. EPA's  position has
 been that, if conducted in compliance with regulatory
 standards and guidance, combustion  can be a safe and
 effective  means of disposing  of hazardous waste. As
 new information has come to light, improvements to the
 regulations governing BIFs and incinerators have been
 and will continue to be pursued.

 EPA believes that our task now is to better  integrate
 source reduction with the regulatory approach to com-
 bustion of hazardous waste, and further ensure that na-
 tional  rules  reflect  the best  combustion  controls
 possible. For example, we should broaden our approach
to include consideration of how an aggressive source
 reduction program should factor into national policy on
the permitting of hazardous waste combustion facilities.
Of course, remediation wastes present a different cir-
cumstance than newly generated wastes and, given the
 finite  set of options for dealing with historic  cleanup
 sites, combustion may be the most appropriate remedy.
 In addition, waste minimization opportunities at cleanup
 sites are usually severely limited. The EPA-State Com-
 mittee will focus on these and other similar issues as
 part of the national dialogue on integration of source
 reduction and waste management.

 B.   EPA's Goals

 The foundation of this draft strategy are the following
 five goals:

 • To establish a strong preference for source reduction
   over waste  management, and thereby reduce the
   long-term demand  for combustion and other waste
   management facilities.

 • To better address public participation in setting a na-
   tional source reduction agenda, in evaluating techni-
   cal combustion issues, and in reaching site  specific
   decisions during the waste combustion permitting
   process.

 • To develop and impose implementable and rigorous
   state-of-the-art safety controls on hazardous waste
   combustion facilities by using the best available tech-
   nologies and the most current science.

 • To ensure that combustion facilities do not pose an
   unacceptable risk, and use the full extent of legal au-
   thorities in permitting and enforcement.

 • To continue to advance scientific understanding with
   regard to waste combustion issues.

 These goals address the major issues surrounding haz-
 ardous waste combustion today and provide an appro-
 priate framework for a broad assessment of how source
 reduction and combustion of hazardous waste can be
 integrated into a national waste management program.
 This assessment will be comprised  of many different
 activities, many of which will be led by the EPA-State
 Committee.  The Committee and other interested par-
 ties  are encouraged to examine these goals critically
 and to consider whether and how they can be improved.

 Hi.  The Process for Pursuing a National
     Strategy

 Under RCRA, ERA and the States are partners and co-
 regulators of the generation, transportation, treatment,
 storage, and disposal of hazardous waste. EPA there-
 fore is firmly committed to the view that any evaluation
 of the role of hazardous waste combustion in our haz-
 ardous waste management strategy must be undertak-
 en as a joint federal and state  effort. To that end, an
 EPA-State Committee will be formed under the aegis of
the EPA-State Operations Committee. As mentioned
earlier, the initial charge to this Committee includes
                                                 78

-------
components dealing with aggressive source reduction,
improvements to technical and permitting standards, al-
ternative treatment technologies, and a better scientific
foundation for decision making.

In each of these areas, this draft strategy lays out a se-
ries of short and longer-term actions for public discus-
sion.  EPA intends to involve all stakeholders  in this
dialogue. EPA is  providing these ideas as a  starting
point for discussion on needed source reduction actions
and regulatory changes that must be pursued, and en-
courages all  interested parties to comment upon and
contribute additional ideas. In addition, however, EPA
believes that we must immediately pursue a number of
actions to ensure that existing combustion facilities are
operated safely and without unacceptable risks to hu-
man health and the environment. Accordingly, while we
implement the elements of this strategy, EPA is direct-
ing its Regions to immediately take actions to pursue
aggressive source  reduction programs at hazardous
waste facilities, and to ensure that waste combustion is
closely controlled through  permitting and aggressive
enforcement.
Both  EPA and  the EPA-State Committee will  seek to
engage the widest  range of interested parties in our
evaluation of source reduction and waste combustion.
This will include federal, state, and local officials, waste
generators and treaters, the waste combustion industry,
environmental and citizen groups, and members of the
public at large. Meaningful participation by, and commu-
nication among, all affected parties is a cornerstone of
EPA's federal hazardous waste program. We intend to
take all steps necessary to foster this participation and
communication.
EPA is also keenly aware that, ultimately, we serve the
public. Our mission  under RCRA, and that of the autho-
rized states, is explicit—we must ensure adequate pro-
tection of  human health  and the environment. EPA
fulfills this responsibility in the light of full public scrutiny.
We will continue to do so during this revaluation of the
role of combustion  in our national waste management
strategy.

IV.   Actions To Implement Strategic Goals

All waste management technologies must assure full
protection of human health and the environment. EPA
will not tolerate operation  of waste management facili-
ties that present unacceptable risks to human  health
and the environment. Accordingly, EPA will engage in a
series of short and longer-term actions designed to pur-
sue aggressive source reduction, to enhance controls
on existing combustion facilities, and to promote public
participation in permitting and source reduction efforts.

The short-term actions include:
• An aggressive source  reduction program that inte-
  grates waste  combustion  with waste management
  decision making.
• Direction to EPA Regions and states to:

  -  Perform site-specific risk assessments, including
      indirect exposure, at incinerator and BIF facilities
      in the permitting process.
  -  Use omnibus  permit authority in new  permits at
      incinerator and BIF facilities as necessary to pro-
      tect human health,  to impose upgraded particu-
      late matter standards and if necessary additional
      metal emission controls, and to impose limits on
      dioxin/furan emissions.
  -  Establish a priority for reaching final permit deci-
      sions for incinerators and BIF facilities.

  -  Enhance public participation in permitting of incin-
      erators and BIFs.
  -  Enhance inspection and enforcement for inciner-
      ators and BIFs.

 The longer-term actions include:
•  Continued efforts to build an aggressive  source re-
   duction program, including exploration of the useful-
   ness and  feasibility of setting a national capacity
   reduction goal for generation of combustible waste.

•  Investigation  of feasibility and risks associated with
   alternative waste treatment technologies.
•  Upgrades to EPA's rules  on emission  controls at
   combustion  facilities and  on continuous emission
   monitoring techniques.

•  Upgrades to EPA's rules on the permitting and public
   involvement process for combustion facilities.

A.   Short-Term Actions
 1.  Integration of Aggressive Source Reduction and
    Waste Combustion
•  Use of permit priorities to stimulate source re-
    duction
 Over the next 18 months, as the national dialogue on
 source reduction is held, EPA will give low priority to
 permit-related requests for additional combustion ca-
 pacity except where that capacity offsets the retirement
 of existing combustion capacity.

 The Agency will consider such requests for additional
 combustion capacity only if the new capacity would pro-
 vide a substantial reduction in emissions. These admin-
 istrative measures will allow the Agency to  focus as a
 priority matter on assuring the safety of currently oper-
 ating  facilities.  Furthermore, to the extent any  new
 capacity is considered, it will be state-of-the-art  com-
                                                   79

-------
  bustlon units designed to achieve more efficient com-
  bustion and lower emissions. These measures will ex-
  tend to new  permit  applications,  modifications  to
  existing permits to expand combustion capacity, and ex-
  pansion of interim status combustion operations.
  •  Publication of final "Waste Minimization Program
     In Place" guidelines

  These guidelines identify the elements of a waste mini-
  mization program for generators and facilities to make a
  proper certification to EPA that they have a waste mini-
  mization program in place, as  required by the RCRA
  statute. EPA will also pursue compliance with the en-
  hanced certification  requirements to the maximum ex-
  tent permitted  under RCRA authority. EPA is also
  considering publication of lists of non-compilers to alert
  the public and the waste treatment industry. Where le-
  gally appropriate, EPA may also use enforcement or-
  ders and permits to incorporate the elements of a good
  waste  minimization  program into the set of require-
  ments that a facility must meet.
 •  Work with  the waste treatment industry as a
    means to get more aggressive action on source
    reduction from  the generators of combustible
    waste

 EPA will ask treatment companies to consider accepting
 wastes only from customers  that  have conducted
 source reduction audits and have an enhanced waste
 minimization program in place (per ERA'S "Program in
 Place" notice). In doing so, we hope that a working part-
 nership can be established among the regulatory agen-
 cies, the treaters, and the generators such that we can
 achieve, as a national priority, the maximum amount of
 source  reduction possible. All interested parties must
 pursue an aggressive source reduction program. EPA
 will work closely with the treatment industry to identify
 additional opportunities for source reduction.
 • Target generating industries that produce com-
   bustible wastes both for source reduction inspec-
   tions and for requiring generators to conduct
   waste minimization audits
 EPA will give top, priority to ensuring compliance with
 waste minimization requirements/guidance at those fa-
 cilities that are driving the demand for waste combus-
 tion. In addition,  at the same facilities, EPA will to the
 maximum extent possible include audit requirements in
 enforcement settlement agreements,  permits, and as
 part  of corrective action  orders.  The audits will allow
 these companies to investigate the maximum possible
 use of source reduction to the extent that they are not
 already doing so  in partnership with EPA and the states.
•  Maximum public involvement and  information
   regarding source reduction and its integration
   with waste combustion
  EPA will also establish a program to more effectively
  provide information to the public on the types of wastes
  going  to combustion units  and the  sources of those
  wastes. First, EPA will compile information from the Bi-
  ennial Report and will collect information from commer-
  cial combustion facilities. This information—such as the
  specific types and volumes of wastes being sent for
  combustion as well as the generators of these wastes
  —will be complied in a report and be provided to the
  public. This information will apprise citizens of those in-
  dustries that rely on combustion of their wastes and will
  allow the public to better focus their attention on the ap-
  propriate generating facilities.

  2.  Immediate Actions in Combustion Facility
     Permitting

  The Agency's goal is to continuously improve the regu-
  lation of hazardous waste combustion to reflect advanc-
  es in  scientific  understanding  so  that  adequate
  protection of human health and the environment is as-
  sured.  During the  time it takes to propose and finalize
  updates to  national regulations, EPA will use its omni-
  bus authority on a case-by-case basis as necessary to
  protect human health and the environment to include
  the appropriate conditions in  permits being issued.
 At this time,  EPA believes that regions and  states
 should use the  RCRA omnibus provision and  RCRA
 permit modification regulations to add permit conditions
 as necessary to protect human health and the environ-
 ment whenever a combustion facility owner/operator is
 seeking issuance of a new permit or reissuance of an
 expiring permit, or, in appropriate circumstances, when
 10 existing permits are reopened for modification.  The
 following will be addressed during the permitting  pro-
 cess.

 o  Risk assessments

 EPA is directing that site-specific risk assessments be
 conducted at incinerators and BIFs during the permit-
 ting process. These should  be done in  accordance
 with EPA's  draft indirect risk assessment guidance.
 EPA is currently developing updated, final guidance on
 conducting risk assessments at combustion facilities,
 including consideration of the risks from indirect expo-
 sures. Until  this national risk assessment guidance is
 completed, all risk assessments at combustion facilities
 will be done on a site-by-site basis. EPA and State tech-
 nical experts will be available to serve on risk assess-
 ment teams  to assist regions and states in conducting
 these risk assessments (particularly with regard to indi-
 rect risks).

• Upgraded particulate  matter standard and
   supplemental controls on metal emissions
Hazardous waste combustion units should be required,
through appropriate use of the omnibus permit authori-
                                                 80

-------
ty, to meet the more stringent particulate matter stan-
dard that is now applicable to municipal waste combus-
ters—0.015 mg/dscm. This technology-based standard
operates to provide a major control on metals emissions
from combustion units. In addition, each combustion fa-
cility should be required to consider, as part of its facili-
ty-specific risk assessment, whether the upgraded PM
standard affords adequate protection against the risks
posed by metals. If additional metal controls are found
to be necessary, the regions and states should impose
these controls through use of the omnibus permit au-
thority.

The upgraded PM standard will be used for BIFs unless
another protective standard is applicable under state or
federal law. These upgraded PM standards will continue
to be  used until  an alternative PM standard has been
promulgated for incinerators and BIFs. It may be that
the upgraded PM standard is sufficient for many com-
bustion facilities. However, some combustion units may
be emitting  metals above de minimis quantities,  in
which case additional controls may be warranted.
•  Dioxins and furans

Site-specific  risk assessments  at hazardous waste
combustion facilities may reveal the need for additional
controls on dioxin and furan emissions. Through appro-
priate use of the omnibus permit authority, the regions
and states should impose as an interim measure emis-
sion limits of 30 ng/dscm  (based on the sum of all tetra
through octa  dioxin and furan congeners). This stan-
dard is the same as the New Source Performance Stan-
dard for new municipal waste combusters. Regions and
states should supplement this with more stringent emis-
sion limits if the site-specific risk assessment warrants.
•  Permit controls on incinerators and BIFs
EPA regions and states should bring incinerators and
BIFs under permit controls as soon as possible. This will
be implemented through  establishment of a schedule
for calling in  all BIF permits for final determinations.
Each  region will develop  a plan that provides for com-
mercial  BIF permits to be called in within the next 12
months and for all other BIF permits to be called in with-
in the succeeding 24 months. Permits represent one of
the most effective means by which EPA and the autho-
rized states can  develop and enforce conditions on the
operation of incinerators and BIFs. At this point, no BIFs
have had final permit decisions. Thus, permit determi-
nations should be made as expeditiously as feasible to
effectively control those operations  that can be operat-
ed safely as well as deny permits at those facilities that
can not be operated safely.
•  Enhanced public participation

Public participation is one of the major cornerstones of
EPA's environmental programs.  EPA is committed to
meaningful public  involvement  in its permitting pro-
grams. Local citizens must be given the opportunity to
assure themselves that facilities in their neighborhoods
will be operated safely.

EPA will immediately provide for greater public partici-
pation in the permitting of BIFs and incinerators, and will
initiate amendments to its rules to reflect new avenues
for public participation. Prior to these amendments be-
ing finally adopted, EPA willdirect all regions and states
to provide immediately for additional public participation
opportunities during  permitting of combustion units -
particularly at earlier stages than now provided for un-
der EPA's current permitting regulations. These should
include, but are not limited to, public comment on the
trial burn plan. EPA will also direct that local citizens be
given the opportunity to  participate during the risk as-
sessment process at combustion facilities.

•  Enhanced inspection and enforcement

EPA will continue and enhance its current enforcement
efforts regarding combustion units through aggressive
inspection and enforcement at both BIFs and incinera-
tors and through use of specialized combustion inspec-
tors. Based on our experience and the  level of public
concern about the compliance  record  of commercial
combustion units, the use of aggressive enforcement
and special inspectors will ensure the maximum timeli-
ness and extent of compliance. In particular, if an event
occurs that results in non-compliance, EPA or the state
will be in a position to take the appropriate enforcement
or permitting action, including abatement of the problem
or, if necessary, shutdown  of combustion operations.
Whenever appropriate, Regions and States are encour-
aged to use permanent on-site inspectors at commer-
cial incinerators and BIFs.

B.  Longer-Term Actions

EPA will also immediately pursue a number of longer-
term actions to continue the progress towards our goals
of source reduction, balancing the amount of combus-
tion capacity with the actual needs, ensuring combus-
tion safety, and providing for greater public participation.

•  Continue to build an aggressive source reduc-
   tion program

EPA will conduct a national round table on source re-
duction opportunities for hazardous wastes. The nation-
al round table on source reduction will seek to highlight
avenues for reducing the amount of waste being com-
busted, and will explore the appropriate  balance be-
tween source reduction  and use of combustion as a
waste management tool. The round table will explore
both regulatory and non-regulatory options to encour-
age  and/or require source reduction. Generating and
treatment industries will be asked to participate actively
                                                  81

-------
 in this effort. Results will also be used by the EPA-State
 Committee to establish a national dialogue among the
 interested parties on the proper integration of source
 reduction and waste combustion.

 •  Establishment of a percent/target year program
    for reduction of combustible hazardous wastes

 EPA will work with the states towards establishing a pro-
 gram in which industry is challenged to reduce by a se-
 lected percentage and by a target year the amount of
 process wastes going to combustion units. EPA will dis-
 cuss with all interested parties the appropriate percent-
 age reduction to be used as a goal and the appropriate
 time frame for this reduction.

 •  Upgrade  EPA's rules to reflect state-of-the-art
    advancements

 EPA will  initiate a rulemaking to upgrade our combus-
 tion rules. In doing so, EPA will explore the feasibility of
 a technology-based approach, particularly with respect
 to setting emission controls on metals,  dioxins and
 furans, acid gases, particulate matter, and products of
 incomplete combustion, in addition, EPA will continue to
 refine its risk assessment guidelines to  ensure that all
 risks are effectively addressed by national regulations
 or site-specific permit conditions.
 •  Upgrade EPA's rules on permit process for com-
    bustion units
 While EPA is directing regions and states to immediate-
 ly afford greater public participation on a permit-by-per-
 mit basis, we will seek to modify our rules to reflect
 expanded public participation. EPA will  initiate a rule-
 making to codify our goal of increased public participa-
tion at earlier stages  in the permitting  process  for
incinerators and BIFs. In particular, EPA will address the
trial burn process and the public's role in that process.
 EPA also believes there is a need to explore a rulemak-
ing to reform the permit appeal process for combustion
units whose permit applications have been denied by
the Regional Administrator or State Director. In particu-
lar, where the unit has been burning waste under interim
status, EPA will seek to establish rules that prevent the
continued burning of waste during administrative ap-
 peals of a permit denial decision. EPA will also explore
 additional guidance or a rulemaking to clarify the num-
 ber  of  permissible trial  burns allowed  before permit
 denial.

 • Use and feasibility of a  long-term national ca-
   pacity reduction goal

 EPA will explore the  usefulness of developing a long-
 term reduction goal (e.g., a 25% reduction in combus-
 tion  capacity  over  the  next 10 years)  to  reduce
 combustion capacity beyond  that  which  can  be
 achieved through source reduction efforts. The purpose
 of such a goal would  be to give more concrete national
 guidance on  how best to mesh combustion demand
 with capacity.

 • Conduct research on continuous monitoring for
   organics, including dioxins and metals

 EPA will use its research resources to continue and en-
 hance scientific inquiry  on ways to better determine
 what constituents are in emissions  from combustion
 units and to develop the technology needed to  monitor
 these emissions on a continuous basis. EPA will work
 cooperatively with the waste combustion  industry to ad-
 dress these research areas.

 • Investigate innovative waste treatment technolo-
   gies that provide protection to human health and
   the environment

 EPA will continue and enhance its efforts to foster the
 development of innovative technologies for the safe and
 effective treatment of hazardous waste. Such  actions
 are essential to our national waste management system
 and to our global competitiveness.

 V.   Conclusion

 EPA is committed to evaluating the role that source re-
 duction  and combustion of hazardous  waste  should
 play in our  national waste management program. EPA
will work in  full partnership with the States in this effort.
 EPA and the States will embark upon a full and open
discussion with all stakeholders,  including affected citi-
zens and industries, on the issues and actions detailed
in this Draft Combustion Strategy.
                                                  82

-------
                                          Appendix B
                         Federal Register Notices to EPA's BIF Rule
Burning of Hazardous Waste in Boilers and Industrial Furnaces (40 CFR part 260 et seq.)
Final Rule
Technical Amendments
Technical Amendments
Technical Clarifications, Amendments, and Corrections
Technical Amendments and Corrections
    Federal Register 56(35):7,134 (February 21,1991)
      Federal Register 56(137):32,688 (July 17,1991)
    Federal Register56(166):42,504 (August 27,1991)
    Federal Register57(165):38,558 (August 25, 1992)
Federal Register 57(190) :44,999 (September 30,1992)
                            OU.S. GOVERNMENT PRINTING OFFICE: 1994-550-001/80345
                                                83

-------

-------