EPA-453/R-94-044a
Medical Waste Incinerators-Background Information for Proposed
                 Standards and Guidelines:
               Control Technology Performance
                     Report for New and
                      Existing Facilities
                          July 1994
             U. S. Environmental Protection Agency
                   Office of Air and Radiation
           Office of Air Quality Planning and Standards
             Research Triangle Park, North Carolina

-------

-------
                           DISCLAIMER
     This report is issued by the Emission Standards Division,
Office of Air Quality Planning and Standards, U. S. Environmental
Protection Agency.  It presents technical data of interest to a
limited number of readers.  Mention of trade names and commercial
products is not intended to constitute endorsement or
recommendation for use. . Copies of this report are available free
of charge to Federal employees, current contractors and grantees,
and nonprofit organizations--as supplies permit--from the Library
Services Office  (MD-35), U. S. Environmental Protection Agency,
Research Triangle Park, North Carolina 27711  ([919] 541-2777) or,
for a nominal fee, from the National Technical  Information
Service, 5285 Port Royal Road, Springfield, Virginia 22161
 ( [703] 487-4650) .
                                 iii

-------

-------
                     TABLE OF CONTENTS
1 0   INTRODUCTION
1

2
                                  n
         911  The Chemical Reactions .......           4
         212  The Combustion Process .........       6

     2'2  ^raene'rafApproacAes to ^tion

         2.2.2  ipecific Coitroi Practices  .......      10

 3 0  ADD-ON AIR POLLUTION CONTROL SYSTEMS .... •  •  •  ;      £

     3'1  f I  r^ffc^ing Meciani-s '.  '.  ----  ;  ;
          312  Venturi Scrubbers   ........           49
          o'-i'i  Packed-Bed Scrubbers .........         55
          '•ll  o?ner Wet -Scrubbing Systems  ......       |4


     32  SIT Ss^^issf-^ss' : ;       ;i
          3 '.2 '.3  Favors Affecting Performance  .  .  . - -       ^
      •3, -3,  DRY SCRUBBERS ..... . •  : :  *  ..... .    .      74
          331  Dry Scrubbing Principles  ........      77
          3*3*2  Dry Sorbent  Injection  .........      83
          3. 3 ".3   Spray Dryer  Absorbers  ..... •  •  •
                                                           89
 4 o  PERFORMANCE OF EMISSION CONTROL MEASURES     .  .  ;  -      ^
      A •}  TEST PROGRAM SUMMARY  .  •  • • , • . •
          I ? 1  Facility and Test  Condition       ^        90
                 Descriptions ..........       _      96
           412  Test Data Summary  .  • ........        10o
      4.2   I  CUSSION OF THE ^RESULTS        ^ • •

                                   SaS^        •

           4 2.,                   a Baici M«i ..
           4*2.5 Metals Partitioning  ........  ^ ^     130

                 ^S/lMllsro^CONTRoL LE^Ls" '.  '.  '.  - -     1«
           431  Particulate Matter ........     ^     151
           4* 3*. 2  Carbon Monoxide   ..........  ' ^     155
           4  3.3  Dioxins and Furans ........    ^  ^     161
           4  3.4  Hydrogen Chloride  ........    ^     16g
           4.3.5  Sulfur Dioxide .......... "...     169
           4.3.6  Lead ............... ...     173
            4.3.7  Cadmium ............. ...     176
            4.3.8  Mercury .............. .  .   •  180
            43.9  Nitrogen Oxides   ........

-------
                         TABLE OF CONTENTS
5.0  REFERENCES
                                                 Page

                                                  185
APPENDIX A.
APPENDIX B.
APPENDIX C.
APPENDIX D.
 SUMMARY OF TEST PROGRAM OPERATING DATA
TEST PROGRAM EMISSION DATA SUMMARY
GRAPHS OF POST-COMBUSTION EMISSION DATA
GRAPHS OF POST-APCS EMISSION CONCENTRATIONS
                              VI

-------
                         LIST OF FIGURES
Figure 1.    Relationship of temperature to excess air .  .

Figure 2.    Impact of temperature and fuel nitrogen on
            NO  emissions for excess air conditions
            (calculated using EER kinetic set)   	

Figure 3.    Theoretical temperature of the products of
            combustion calculated from typical municipal
            solid waste properties, as a function of
            refuse moisture and excess air or oxygen

Figure 4.    Impact ion	

Figure 5.    Diffusion	•	

Figure 6.    Absorption   	

Figure 7.    Simplified venturi scrubber configuration .

Figure 8.    Schematic of wetted-approach venturi
            scrubber   	  	

Figure 9.   Schematic of wetted-approach, variable
            throat venturi scrubber  	

Figure 10.  Venturi  scrubber  illustrating the  flooded
             (wetted) elbow and exhaust gases passing to
            a  cyclonic  separator   	

Figure 11.  Venturi  scrubber  collection  efficiency
            versus differential pressure   	

Figure 12.  Typical  medical waste  incinerator  exhaust
            gas  particle size distribution   	

Figure 13.  Schematic  of countercurrent  packed-bed
            scrubber  	

Figure 14.   Schematic  of Rotary  Atomizing™  scrubber on
             a MWI	

Figure  15.   Schematic  of rotary  atomizer module ....

Figure  16.   Schematic  of an ionizing wet scrubber™  .  .

 Figure  17.   The Hydro-Sonic® family of wet scrubbers  .

 Figure  18.   Schematic of a pulse-jet baghouse 	
Page,

   7




  17





  18

  33

  34

  35

  37



  39



  40




  41


  44



   45



   51


   57

   59

   61

   63

   67
                                vii

-------
                   LIST OF FIGURES  (continued)
                                                             Page
Figure 19.   Mechanisms of adsorption	      75
Figure 20.   Schematic of dry sorbent injection system .  .      78
Figure 21.   Schematic of spray dryer absorber system  .  .      85
Figure 22.   Effect of waste type at Facility A on
            post-combust ion Hg emissions	     102
Figure 23.   Effect on waste type on Hg emissions for all
            intermittent/continuous MWI's 	     103
Figure 24.   Effect of waste type at Facility A on
            post-combustion HCl emissions	     105
Figure 25.   Relationship between CDD/CDF and CO
            emissions	     110
Figure 26.   Relationship between CDD/CDF and PM
            emissions	     110
Figure 27.   Effect of secondary/tertiary chamber
            temperature and residence time on CDD/CDF
            emissions	     113
Figure 28.   Effect of secondary/tertiary chamber
            temperature and residence time on CO
            emissions	     114
Figure 29.   Effect of secondary/tertiary chamber
            temperature and residence time on PM
            emissions	     115
Figure 30.   24-hour real time CO and THC
            concentrations--Facility A,  test
            condition 2, run number 2	     122
Figure 31.   24-hour real time CO and THC
            concentrations--Facility J,  run numbers 5
            and 6	     122
Figure 32.   24-hour temperature plot--Facility A,
            test condition 2, run number 2  .......     123
Figure 33.   24-hour temperature plot--Facility J,
            run numbers 5 and 6	     123
                              Vlll

-------
                 LIST OF FIGURES  (continued)

Figure 34.
Figure 35.

Figure 36.
Figure 37.
Figure 38.

Figure 39.

Figure 40.
Figure 41.

Figure 42 .

Figure 43 .
Figure 44.

Figure 45.
Figure 46 .

Figure 47.

Cooldown period- -Facility A, test
condition 2, run number 2 	
Cooldown period- -Facility J, run numbers 5
and 6 	
Plot of HC1 removal efficiency versus
stoichiometric ratio for Facility A .....
PM emissions for continuous, intermittent,
and batch MWI's with combustion controls . .
PM emissions for uncontrolled pathological
MWI ' S 	 	 	
PM emissions for continuous, intermittent,
batch, and pathological MWI's with add-on
controls
CO emissions for continuous, intermittent,
and batch MWI's with combustion controls . .
CO emissions for uncontrolled pathological
MWI's 	
CDD/CDF emissions for continuous,
intermittent, and batch MWI's with
combustion controls 	
CDD/CDF emissions for uncontrolled
pathological MWI's 	
CDD/CDF emissions for continuous,
. intermittent, batch, and pathological
MWI's with add-on controls 	
HC1 emissions for continuous, intermittent,
and batch MWI's with combustion controls . .
HC1 emissions for uncontrolled pathological
MWI'S 	 	 " " * /
HC1 emissions for continuous, intermittent,
batch, and pathological MWI's with add-on
rontrols 	
Paae
125
125

136
147
149

152

153
154

1 CC
J.DD
158
T C Q
IbS
163
164

165
Figure 48
SO, emissions for continuous, intermittent,
and batch MWI's with combustion controls
167

-------
                   LIST OF  FIGURES  (continued)
Figure 49.


Figure 50.


Figure 51.


Figure 52.



Figure 53.


Figure 54.


Figure 55.



Figure 56.


Figure 57.


Figure 58.



Figure 59.


Figure 60.
 S02  emissions for uncontrolled pathological
 MWI's 	

 Pb emissions for continuous,  intermittent,
 and  batch MWI's with combustion controls

.Pb emissions for uncontrolled pathological
 MWI's 	

 Pb emissions for continuous,  intermittent,
 batch,  and pathological MWI's with add-on
 controls  	

 Cd emissions for continuous,  intermittent,
 and  batch MWI's with combustion controls

 Cd emissions for uncontrolled pathological
 MWI's ..... 	

 Cd emissions for continuous,  intermittent,
 batch,  and pathological MWI's with add-on
 controls  	

 Hg emissions for continuous,  intermittent,
 and  batch MWI's with combustion controls

 Hg emissions for uncontrolled pathological
 MWI's 	

 Hg emissions for continuous,  intermittent,
 batch,  and pathological MWI's with add-on
 controls  	

 NOy  emissions for continuous,  intermittent,
 and  batch MWI's with combustion controls

 NOX  emissions for uncontrolled pathological
 MWT's 	
Page



 168



 170



 171




 172



 174



 175




 177



 178



 179




 181



 183



 184

-------
                         LIST OF TABLES
TABLE 1.






TABLE 2.








TABLE 3.
CONDITION	
                                                               91
                                                    98
                                                               106
l>U3iJEj •» .
TABLE 5.
TZkRLiE 6
J_ f*± " -1— ' •*— 1 u *
TABLE 7.
TABLE 8.
faHTiP Q
l/^BljEi ^ •
TABLE 10 .
J_ J^JJJJA-* -^- **
TABLE 11.
TABLE 12 .
^^^^j^jij ^** •
TABLE 13.
TABLE 14.
TABLE 15.
TABLE 16.
Tatar .TT. 17 .
S^ssss rss « • •
COMPARISON OF POST -COMBUSTION EMIS^°?S FROM ^
SUMMARY OF BOTTOM ASH AND STACK GAS METALS ^
DISTRIBUTION 	
PERFORMANCE OF THE VS/PB AND FF/PB AIR
POLLUTION CONTROL SYSTEMS 	
FACILITY A HC1 PERFORMANCE SUMMARY 	
SUMMARY OF METAL EFFICIENCIES BY TEST
CONDITIONS AT FACILITY A 	 • •
____. T^/"\T^ *^^JU?
SUMMARY OF CDD/CDF PERFORMANCE DATA FOR THE ^
DI/FF AT FACILITY A 	
HC1 PERFORMANCE DATA FOR FACILITY M 	
SUMMARY OF CDD/CDF AND METALS PERFORMANCE
FOR SD/FF SYSTEM 	
TESTED MWI FACILITIES 	
COMBUSTION CONTROL DATA VS. MWI TYPES ....
120
127
129
131
134
137
138
140
141
143
144
146
151
155

-------
 TABLE  18.


 TABLE  19.


 TABLE  20.


 TABLE  21.


TABLE  22.


TABLE 23.
                    LIST  OF TABLES  (continued)
HC1 EMISSION LIMITS FOR CONTINUOUS
INTERMITTENT,  BATCH,  AND PATHOLOGICAL MWI'S

S02 EMISSION LIMITS FOR CONTINUOUS
INTERMITTENT,  BATCH,  AND PATHOLOGICAL MWI'S

Pb EMISSION LIMITS  FOR CONTINUOUS
INTERMITTENT,  BATCH,  AND PATHOLOGICAL MWI'S

Cd EMISSION LIMITS  FOR CONTINUOUS,
INTERMITTENT,  BATCH,  AND PATHOLOGICAL MWI'S

Hg EMISSION LIMITS  FOR CONTINUOUS
INTERMITTENT,  BATCH,  AND PATHOLOGICAL MWI'S
  x EMISSION LIMITS FOR CONTINUOUS
INTERMITTENT, BATCH, AND PATHOLOGICAL MWI'S
Page


 162


 166


 169


 173


 176


 182
                              xii

-------
1.0  INTRODUCTION
     The main objectives of this report are to describe the
various emission control techniques used to control emissions
from medical waste incinerators (MWI's), to summarize the
emission test data generated during a comprehensive emission test
program, and to provide an emission test data analysis that
quantifies the performance of these techniques.  Two general
types of emission control techniques are described:  combustion
controls and add-on air pollution control systems  (APCS's).
Combustion controls refer to the control of the combustion
process and the emission-generating mechanisms through proper
design, operation, and maintenance to minimize the generation of
emissions.  Add-on APCS's control or remove pollutants from  the
exhaust gas stream after the gas stream exits  the MWI  system.
This document  is  one  of a series of reports written  to provide
background  information used to develop  emission standards and
guidelines  for MWI's  under Section  129  of  the  Clean  Air  Act.
     This report  presents  technical  information on the emission
 control  techniques available  to control emissions  from MWI's.
 Section 2.0 first describes  the general approaches to combustion
 controls  and subsequently describes specific designs and/or
 operational features  that are employed to control  the combustion
 process.   Section 3.0 describes the add-on APCS's.  Section 3.1
 describes wet scrubbers including the venturi and packed-bed
 scrubbers in detail and briefly describes other types of wet
 scrubbers including the Rotary Atomizing™ scrubber. Ionizing Wet
 Scrubber™,  the collision scrubber, and the Hydrosonic® scrubber.
 Section 3.2 describes fabric filters.  Section 3.3 describes dry
 scrubbers including dry sorbent injection and spray dryer
 systems.  Finally, Section 4.0 summarizes available emission test
 data and presents a  test data  analysis that quantifies  the
 performance of APCS's  (venturi/packed-bed scrubber, fabric
 filter/packed-bed scrubber, dry sorbent injection/fabric filter,
 spray  dryer absorber/fabric filter), and  three combustion
 controls  (increased  secondary chamber  residence time,  increased
 secondary  chamber temperature,  and proper charging  procedures).
                                  1

-------
2.0  COMBUSTION:  THEORY AND CONTROL
     The emissions from an MWI are influenced strongly by the
combustion process conditions.  Consequently, combustion control
can be an important element of MWI emission control measures.
This section describes the MWI combustion process and discusses
methods that are being used to control that process.  Section 2.1
provides a brief overview of combustion theory and summarizes how
the process characteristics affect emissions.  Section 2.2
describes general approaches to combustion control and describes
specific control practices that are typically used.
2.1  COMBUSTION THEORY
     Combustion is defined generically as a chemical oxidation
reaction characterized by the evolution of energy as light and
heat.  The medical waste combustion process is characterized by a
complex combination of chemical reactions that involve the rapid
oxidation of organic substances in the waste and auxiliary fuels.
The goal of the process is to achieve complete combustion of the
organic material, while minimizing the formation and release of
undesirable pollutants.
     This section presents a brief overview of the combustion
process as it relates to medical waste incineration and describes
the key physical and chemical phenomena that occur in combustion
systems.  It also identifies the key operating parameters and
discusses their effect on the combustion process.
2.1.1  The Chemical Reactions
     Because medical waste is a heterogeneous mixture of general
refuse, pathological wastes, plastics, and laboratory wastes, the
specific chemical composition is unknown and varies with time.
However, the process can be described reasonably well by making
some simplifying assumptions about the key chemical reactions.
The organic portion of medical waste consists primarily of
carbon  (C), hydrogen (H), and oxygen  (0), with other elements
such as sulfur  (S), chlorine (Cl), nitrogen (N),  and a variety of
metals being found to a lesser degree.  Key combustion reactions
involve these major waste constituents and the 02 in the
combustion air.

-------
     The simplified equilibrium reactions that characterize the
MWT process are:
         C + .0'
         C02 + heat
2 H
                02 -» 2 H20 + heat
     1.
     2.
     3.   H + Cl •* HC1 + heat
     4.   S + 02 -» S02 + heat
     For equation 1,  carbon monoxide (GO)  also is formed if there
is incomplete combustion.   For equation 3, the available
thermodynamic equilibrium data and bench- scale test data indicate
that organic chlorine reacts almost completely to form hydrogen
chloride  (HC1)  with small quantities of elemental chlorine (C12)
also produced.   However, unless the H:C1 ratio in the feed is
very low, and external hydrogen sources (e.g., water vapor) are
unavailable, almost no C12 is formed.1  For equation 4, the
organic sulfur is oxidized during the combustion process to form
sulfur dioxide  (S02)  .  Because HC1 is a stronger acid than S02,
it will react more quickly with available alkaline compounds.
Some S02 may react with available alkaline compounds; however,
the amount of S02 involved in such reactions  is expected to be
negligible due to the high HC1 content of the flue gas.
Consequently, essentially all of the organic  sulfur  in  the waste
will leave the  combustion chamber as vapor phase S02 .
     While these simplified  reactions identify the major
combustion products, they do not present  a complete  picture of
the  combustion  process.  Because of incomplete combustion,
dissociation, and the  formation of  thermal and fuel  nitrogen
oxides  (NOX) ,  the MWI  exhaust gases will  also contain trace
amounts  of  CO,  NOX,  methane  (CH4) ,  C12, and a wide  range  of
organic  compounds.   Also, volatilization  of inorganic compounds
or entrainment  of  inorganic compounds  in  the  combustion gases
contaminate the exhaust gases with particulate matter (PM) ,
including metals.  A major goal  of combustion controls is to
minimize concentrations of compounds  other than  those identified
 in Equations 1-4 to  levels as  low as  possible.

-------
2.1.2  The Combustion Process
     The primary objective of the MWI combustion process is the
complete combustion of all organic material in the waste feed.
As described in some detail in Section 2.2, MWI manufacturers and
operators use a variety of specific design and operating
techniques to achieve this goal.  Despite this range of
techniques, the basic combustion process and the factors that
determine the ability of the process to achieve complete
combustion are reasonably consistent.  As a background for the
more detailed discussion in Section 2.2, the paragraphs below
provide a brief description of the typical MWI process and
identify key factors that affect performance.
     The basic MWI design typical of most newer systems is
comprised of two, or possibly three chambers which are referred
to as the primary, secondary, and tertiary chambers.  Medical
waste is fed into the primary chamber of these multiple-chamber
units.  This medical waste comprises a mixture of organic
material  (including volatile organics and fixed carbon),
inorganic material  (including metal compounds) , and water.  As
this waste enters the primary chamber, it is exposed to a high-
temperature heat source  (either an ignition burner or the burning
waste bed) and to combustion air.  Key events that occur in the
primary chamber are:
     1.  Moisture in the waste is evaporated from the waste and
the water vapor temperature is raised to the temperature of the
primary chamber exhaust gases;
     2.  Volatile organic materials in the waste are released and
some amount of vapor-phase combustion occurs above the waste bed;
     3.   Fixed carbon is burned via surface oxidation reactions
in the waste bed; and
     4.   Inorganic materials  (particularly metals) are
partitioned as entrained solids in the exhaust gas, vapors in the
exhaust gas, and solid materials in the ash.
     In most MWI systems, the combustion air supplied to the
primary chamber is  insufficient to complete the reactions
described in Section 2.1.1.  Consequently, a fuel-rich, exhaust

-------
gas stream passes from the primary to the secondary chamber.
Additional combustion air is supplied to the secondary chamber
and the combustion reactions are carried towards completion.
     How well this basic process achieves the goal of complete
organic combustion with minimal release of undesirable pollutants
is determined by three key factors that influence the combustion
process.  In simplest terms, these factors are usually referred
to as time, temperature, and turbulence  (or mixing).   The
combustion reactions described in Section 2.1.1 are equilibrium
reactions.  The degree to which these reactions move toward
completion is a function of reactant availability, reaction rate,
and time over which conditions are appropriate for the reaction
to occur.  The reaction rates for all combustion reactions are
temperature dependent.  The temperature  at the point of
combustion must be sufficiently high to  initiate the reaction,
and the degree of reaction  completion depends on the time/
temperature envelop to which the reactants are exposed.   The  key
combustion reactions are  oxidation reactions.  Consequently,
oxygen must be in close contact with the material  being  combusted
for the reactions to proceed quickly.   Turbulence  is  required to
expose unburned material  (in both the waste  bed and  in the
secondary chamber) to  the available  oxygen  in the  incoming air.
Without waste-bed turbulence,  much of  the unburned organic
material  (either volatile or  fixed carbon)  will be insulated
under layers of  ash  and will  not  have  sufficient  oxygen  available
 to complete  the  combustion process.  The mixing created  by
 turbulence in the waste bed ensures  that all combustible
materials contact the supply of oxygen at some point during the
 burning cycle.   Turbulence in the secondary chamber  ensures that
 volatiles and other unburned material  are adequately mixed with
 combustion air (oxygen) to complete the combustion process begun
 in the primary chamber.  Adequate secondary chamber temperature
 and gas-residence time are also required to complete combustion.
 Most combustion control measures described in Section 2.2 are
 related to time, temperature, and turbulence.

-------
     One other relationship that is an important prerequisite to
understanding combustion controls is the relationship between
combustion air and temperature.  The stoichiometric amount of
combustion air is the quantity of air needed to provide exactly
the theoretical amount of oxygen for complete combustion of
carbon and hydrogen in the waste.  Theoretically, 100 percent
stoichiometric air coincides with the condition of maximum
temperature  (>1650°C  [3000°F]) in the chamber.  A graphical
representation of the relationship between combustion temperature
and excess air level is shown in Figure 1.  As the amount of
combustion air is increased above the stoichiometric point
 (excess air), the combustion temperature is lowered because
energy is used to heat the combustion air from ambient
temperature  to the combustion chamber temperature.  The greater
the volume of the excess air, the greater the "heat loss" due to
raising the  air temperature.  As the amount of excess air is
decreased, the combustion temperature increases until it becomes
maximum at the stoichiometric point.  Below the stoichiometric
point, as the amount  of combustion air  is decreased, the
temperature  decreases because complete  combustion has not
occurred.  Because the complete  combustion reaction  (which is
exothermic)  has not occurred, the maximum heat is not generated.
Typically, the primary chamber of an MWI is operated in the
substoichiometric region of  the  curve,  while  the secondary
 chamber is operated in the excess air region.
2.2   COMBUSTION CONTROL
      Combustion controls are defined as any design features  or
 operating practices that are used by MWI manufacturers and
 operators to reduce or limit the quantities of undesirable
 pollutants  emitted with the  MWI  exhaust gases and to maximize the
 destruction, or burnout, of  organic material  in  the solids bed.
 In addition to enhancing the quality of gaseous  exhausts and
 solids discharge, the control measures  are designed to minimize
 the operation and maintenance costs of  the incineration system.
 These goals  are  often compatible in that  the  well-controlled
 combustion  conditions that lead to  low  emission  levels and good

-------
TEMPERATURE
              MAXIMUM
              TEMPERATURE
                   DEFICIENT AIR
EXCESS AIR
                        PERCENT EXCESS AIR
     Figure l.  Relationship  of temperature to excess  air.-

                                7

-------
ash quality' also tend to limit damage to the MWI system, thereby
promoting lower operation and maintenance costs.
     The remainder of this section discusses the combustion
control mechanisms that have been implemented by different
manufacturers and operators to control emissions from MWI's.  The
discussion is divided into three parts.  The first provides a
general overview of the different combustion control approaches
that have been used.  The second describes specific control
practices and presents the available information on the
effectiveness of these practices in controlling all pollutants of
interest that are emitted from MWI's.  The final subsection
provides a brief overview of how these different specific control
practices are related.
2.2.1  General Approaches to Combustion Control
     Each MWI manufacturer has developed a package of features in
its design that is aimed at controlling air emissions and ash
quality.  The mixture of controls varies among manufacturers,
making each combustion system unique in its approach to
combustion control.  While each system is unique and a wide
variety of specific practices are used, all of the systems
comprise some combination of three general approaches:
(1) controlling the rate of primary chamber chemical reactions,
thereby controlling the release rate of volatile organics and the
degree of ash burnout;  (2) controlling waste bed turbulence,
thereby limiting entrainment of particles from the waste bed; and
(3) controlling secondary chamber combustion conditions, thereby
promoting the complete combustion of volatile organic material.
These general approaches are implemented by including design or
operating features that control some combination of the following
parameters:  temperature of the reactants and reaction products,
residence time of reactants at reaction temperatures, and gas
turbulence and solids mixing of the reactants in the primary and
secondary chambers of the MWI.
     When medical waste is charged to the primary chamber and is
exposed to a high-temperature heat source, such as an already
burning waste bed or auxiliary burners, numerous physical/
                                8

-------
chemical reactions occur in and above the waste bed.  These
reactions include vaporization of moisture in the waste,  the
volatilization of organic material in the waste, both gas-phase
and surface chemical reactions that involve pyrolysis and
oxidation of organic material, volatilization of inorganic
material in the waste, and the entrainment of solid materials in
the combustion gases.  Reactions continue in the secondary
chamber for those gases and solid particles released from the
waste bed.  By controlling the key parameters of temperature,
residence time, gas turbulence, and solids mixing of the
reactants within each chamber, the rate and level of completeness
of the physical/chemical reactions are also controlled.
     Combustion control within an MWI is usually based on
maintaining temperatures in both chambers within specified limits
by controlling the combustion air rate to each  chamber, the waste
feed rate, and the auxiliary fuel burner operation.  Limiting
combustion air in the primary chamber to below  stoichiometric
conditions prevents rapid combustion, decreases the temperature,
and allows a quiescent condition within the primary chamber that
minimizes entrainment of PM.  In the secondary  chamber, however,
high temperatures can be maintained in a turbulent  condition with
excess air  (i.e., greater than stoichiometric levels)  to ensure
complete combustion of organics in the gases emitted from  the
primary chamber.
     Combustion controls are generally aimed at reducing CO and
organic emissions and limiting PM and metals emissions.
Generally,  these  combustion control measures have little or no
effect  on emissions of HC1 and S02-  In both cases, most of  the
chlorine and  sulfur will be converted to HC1 and S02 under the
entire  range  of  combustion conditions which normally occur in  an
MWI.
     There  are varied degrees  of  combustion control for MWI's,
depending not only  on the manufacturer, but also on the  type and
size of the incinerator.  Combustion control systems vary  from
simple,  manually-controlled,  manually-fed units to  highly
complex,  automatically-controlled, automatically-fed,  continuous
                                 9

-------
units.  The" specific combustion control practices used, their
interrelationships, and their effects on the emission levels of
specific pollutants are discussed below.
2.2.2  Specific Control Practices
     Manufacturers and operators of MWI's use a variety of
techniques to address the combustion control objectives
identified in Section 2.2.1.  The techniques that have been
widely applied to MWI's include feed rate control, control of
combustion air volumetric flow and distribution, temperature
control in the primary and secondary chamber, control of solids
retention time in the primary chamber and gas residence time in
the secondary chamber, enhancement of gas mixing in the secondary
chamber, control of solids mixing in the primary chamber, sizing
and location of combustion air ports in the primary chamber,
primary chamber steam injection, and operator training.  The
subsections below describe how each of these techniques is
implemented and present qualitative assessments of the effect  of
the different techniques on pollutant emissions.
      2.2.2.1  Control of Feed Rate.  Each incinerator  system is
designed for a particular thermal input rate with the  thermal
input coming from  the waste feed and, when necessary,  auxiliary
fuel  burners.  Under ideal conditions, the incinerator operates
with  a constant thermal input.  Under actual conditions, however,
the medical waste  feed is a heterogeneous mixture with variable
volatile content,  fixed-carbon content, ash  content, moisture
content, and heating value.   Incinerator operating conditions  are
varied to  the extent possible with  changing  waste characteristics
in order to promote controlled combustion and minimize pollutant
emissions.  Rates  of thermal  and volatiles release from the waste
feed are controlled by controlling  the combustion air  rates in
combination with  control  of waste  feed quantity and frequency.
This section describes waste  feed  control measures, while the
combustion air control measures are described  in Section 2.2.2.2.
      The effects  of waste feed rates  on  combustion conditions  and
emissions  are  closely  tied to waste feed characteristics.
Increasing the feed rate  with low-density, high-heating-value
                                10

-------
wastes (plastics, rubber, paper)  results in higher gas-phase and
solid-phase combustion rates, higher temperatures, higher
volatiles release rates, and higher gas velocities in the primary
chamber.   Also, the higher.temperature and resulting increase in
volatiles release rates generate gas volumes that are likely to
exceed the excess air capacity and volumetric capacity of the
secondary chamber.  All of these effects of increased waste feed
rates raise the potential for high-PM emissions from entrainment,
lower secondary chamber residence times, and the discharge of
incomplete combustion products to the environment.  Increasing
the feed rate with high-density, low-heating-value wastes
 (fluids,  tissue, bones) generally creates lower primary chamber
temperatures due to the heat required to vaporize moisture and
the low amount of heat being released from the waste.  With
wastes of this type, auxiliary fuel burners must often be
operated to maintain the temperatures in both chambers.  The
large volume of water vapor and the small amount of combustible
gases coming from the waste bed coupled with lower primary
chamber temperatures increase the likelihood of the discharge of
incomplete combustion products.
     The methods used to control the feed rate for MWI's differ
between batch-duty units and intermittent- or continuous-duty
units.  Batch-duty MWI's are designed to accept a single load of
waste at the beginning  of the incineration cycle, and feed  rate
 control is simply a matter of how much waste is loaded into the
primary chamber  at  that  time.   If the waste being charged has a
 significantly  higher heating value  than that specified for  design
purposes,  the  primary  chamber must  be loaded at less  than
volumetric capacity to decrease the thermal input from the  batch
 charge.  These units range in size  from 331 to 1,724  kg/batch
 (150  to  3,800  Ib/batch)  and  typically are  loaded  manually.   Some
 batch MWI's  also have  options for  charge door  lock-out tied to  a
 preset timer or to  the primary  chamber  temperature  to prevent  the
 operator from  opening  the door  before  the  cycle  is  complete.
      The intermittent-  and  continuous-duty MWI's  typically have a
 mechanical charging device  (ram feeder)  that permits waste
                                11

-------
charging while the MWI is operating.  Some small intermittent
units, however, do not have a mechanical loading device but are
charged manually, and some manufacturers offer a continuous
mechanical charging device  (auger feeder) for charging their
continuous-duty MWI's.  The ram feeder loading device feeds small
charges of waste material to the primary chamber at regularly
timed intervals.  The size of each waste charge is controlled by
the size of the hopper and the amount of waste the operator
places into the hopper.  The charging frequency can be controlled
manually or by a preset timer on the control panel.  Also, the
charging cycle can be locked out or overridden by a control loop
that responds to temperature in the primary chamber or the
temperatures in both the primary and secondary chambers.  To
approach a steady thermal input, manufacturers recommend a
charging procedure consisting of multiple charges at equally
timed intervals.  The recommended charge size is 10 to 25 percent
of the rated hourly capacity, charged at 5- to 15-minute
intervals.  The charging frequency may then need to be adjusted
to respond to variations in the waste composition.
     2.2.2.2  Control of Combustion Air.  For most MWI's, the
primary combustion controls are driven by control loops that
maintain the temperatures in the primary and secondary chambers
of the incinerator within specified limits.  The control
parameter of greatest importance in maintaining these
temperatures within setpoints is the combustion airflow rate in
each chamber.
     The combustion air in  the primary chamber is typically
controlled to below stoichiometric amounts.  As stated in
Section 2.1.2, limiting the combustion air in the primary chamber
to below stoichiometric conditions  (generally 40 to 70 percent of
stoichiometric)  limits the  combustion rate by limiting the amount
of oxygen available for both solid-phase and gas-phase oxidation
reactions.  By limiting these combustion reactions, the heat
release rate to  the primary chamber is reduced with a
corresponding  decrease in temperature.  Also, limiting the
combustion airflow and the  resulting combustion rate allows a
                                12

-------
quiescent condition to exist within the primary chamber that
limits the entrainment of PM in the combustion gases exiting to
the secondary chamber.
     The combustion air in the secondary chamber is typically
controlled to greater than stoichiometric amounts (generally
150 to 250 percent of stoichiometric [i.e., 50 to 150 percent
excess air]).  Consequently, the secondary chamber is said to
operate with excess air.  Combustion gases are maintained at
higher temperatures in a turbulent condition with excess oxygen
to ensure complete combustion of the volatile organic compounds
and CO emitted from the primary chamber.  Because the secondary
chamber temperature is operating in an excess air mode,
increasing the amount of combustion air to the secondary chamber
decreases the secondary chamber temperature.
     Controlling the  combustion air in an MWI permits a staged
combustion system for minimizing pollutant emissions from the
combustion process.   By limiting the combustion  air and the
temperature in the primary  chamber, the formation of fuel NOX and
thermal NOV and the entrainment of particulate matter are
          Jt
limited.  Enough combustion airflow is maintained to ensure
minimum temperatures  for the destruction of pathogens  in the
waste  and  fixed-carbon burnout  in  the primary chamber.  In  the
secondary  chamber, the  combustion  air  is controlled at  levels
that permit  enough excess air for  the  required temperature,  gas
turbulence,  and residence time of  the  combustion gases  both to
ensure the destruction  of pathogens that have been  entrained in  .
the primary chamber  exhaust and to complete  the reaction  of
organics  in the  exhaust gas stream.
      The conversion  of  fuel-bound nitrogen to NOX  (fuel NOX)
depends in large  part on the local availability of  oxygen to
 react with the volatile species,  the  amount of fuel-bound
 nitrogen,  and the chemical  structure  of the fuel-bound nitrogen.
 Because of the substoichiometrie amount of combustion air in the
 primary chamber with the fuel, the formation of fuel NOX is
 minimized.  Combustion air control affects organics destruction,
 metals volatilization,  thermal NOX generation, and ash burnout
                                 13

-------
only through the effects of combustion air on temperature and
turbulence.  The effects of temperature and turbulence on these
phenomena are described in subsequent sections.
     Combustion air control is usually based on the temperatures
of the primary and secondary chambers.  Thermocouples within each
chamber are used to monitor the temperatures continuously; this
information is then used to adjust the combustion air rate to
maintain the desired temperatures.  Many manufacturers also use
timers and information about the loading and burndown cycles to
control the combustion air.  Some manufacturers have also used
parameters such as gas flow, opacity, and oxygen concentration to
                                                             Q
control combustion air, but these mechanisms are not typical.
     The level of combustion air control varies from manually
setting the combustion airflow dampers for each chamber to fully
modulating the combustion airflow over the entire operating
range.  On some systems, the control settings are actuated by the
feedback from a monitored parameter  (e.g., temperature), by a
specified time in a cycle  (e.g., charge-door opening), or by a
control timer.  The more advanced, elaborate combustion air
control systems offer more control of the temperature and hence
the combustion rate in the incinerator, but they are also more
complicated and more costly.  The simple, manually set combustion
airflow dampers are generally set for an average charge of
medical waste with a typical composition.  As the waste
composition varies and the combustion of the waste charge
proceeds, the combustion airflow through these dampers remains
constant.  Consequently, over the duration of a burn cycle, the
percentage of stoichiometric air varies constantly in both
chambers, resulting in changes in the combustion rate,
temperature, turbulence, residence time, and pollutant emissions.
In contrast, the automatically controlled, fully modulated,
combustion air control systems permit continuous monitoring of
the chamber temperature and the control of the combustion airflow
rate according to that temperature.  These changes in the
combustion airflow rate also modulate the combustion rate and
tend to smooth out temperature swings in both the primary and
                                14

-------
secondary chambers.   The increased stability tends to minimize
the increases in pollutant emissions that can result from these
temperature swings found in less automated mi's.
     Many of the newer, automatically controlled systems also
stop the combustion airflow to the primary chamber when the
charging door is opened, thereby compensating for the amount of
combustion air that enters through the charge door with the
charge.  Many of these systems also operate with the combustion
airflow damper to the secondary chamber fully open at the time of
charging to ensure .maximum excess air to combust the initial
release of volatiles that accompanies a fresh charge.
     2.2.2.3  Prmt-.r-ol  of Primary Chamber Temperature.   Combustion
control for the primary chamber of an MWI  is typically  based  on
controlling the temperature  in that chamber.  A thermocouple
within the primary  chamber is used  to monitor the  temperature
continuously.   Feedback from the monitoring is  then used on some
systems to  control  the combustion air rate at levels that
maintain a  specific setpoint temperature within the chamber.
Control of  the  incinerator  feed rate, use of auxiliary fuel
burners, and water  injection are other  methods  that may be used
to control  the  primary chamber temperature.
      The primary chamber must  be operated at a  temperature
 sufficient to sustain combustion and to combust the fixed carbon
 in the waste bed.  The temperature in the primary chamber is
 maintained by the combustion of the fixed carbon within the waste
 bed, the combustion of a portion of the combustible gases in the
 primary chamber just above the waste bed, and,  when necessary,
 auxiliary fuel burners.  The typical operating ranges  noted  in
 the literature are 400° to 980°C (750' to  1800«F).2'4'5  The
 lower setpoint is needed to maintain fixed-carbon  combustion and
 volatiles release  from the waste bed, while temperatures below
 the upper setpoint minimize slagging and  refractory damage.
      The volatiles  release  rate and the gas- and  solid-phase
 combustion  rate  are temperature dependent.  Controlling the
 temperature  level  in  the primary chamber  also  controls the
 combustion rate  and the volatiles  release rate from the waste bed
                                 15

-------
at the time- of charging.  If the  temperature in the primary
chamber is not controlled, the resulting rapid volatiles release
immediately increases the volume  of  combustion gases leaving the
chamber, promotes the entrainment of both organic and inorganic
materials, reduces the gas residence time, and increases the
requirement for combustion air in the secondary chamber.  Taken
together, these phenomena are likely.to result in increased PM,
CO, metals, and trace organics emissions.
     The formation of NOX from nitrogen in the combustion air
(thermal NOX) is limited in both  chambers by limiting the peak
flame temperatures to less than 1649°C  (3000°F) , typically
through combustion air control.   The formation of thermal NOX is
highly temperature dependent and  proceeds rapidly at temperatures
in excess of 1649°C  (3000°F) (see Figure 2).  The theoretical
flame temperature is the maximum  flame temperature attainable
with a stoichiometric amount of air and no heat losses and is
typically in the range of 1649° to 1927°C  (3000° to 3500°F)  (see
Figure 3).  The actual flame temperatures, however, for
substoichiometric and excess air  conditions with the typical heat
losses due to radiation, convection, conduction, arid
dissociation, are expected to be  in the range of 1260° to 1371°C
(2300° to 2500°F).  Temperatures  in this range are less conducive
to thermal NOX formation than the peak temperatures associated
with stoichiometric air supply.6'7
     Primary chamber temperatures affect both the amount of
metals being volatilized and the  amount of metals being entrained
in the exhaust gases leaving the  primary chamber.   Increasing
primary chamber temperatures shifts the metals partitioning
towards the volatile fraction and the entrained particle fraction
and away from the ash fraction.   Further,  metals that volatilize
at primary combustion chamber temperatures tend to selectively
condense on small particles in the flue gas,  a phenomenon known
as fine particle enrichment.  Also,  higher primary chamber
temperatures usually occur with conditions that cause increased
particle entrainment (i.e.,  increased waste bed turbulence,
increased combustion airflow, and increased volatiles release
                                16

-------
  10,000
   1000
   100
 Q.
 Q_
 o  10
     1.0
    0.1
             3140   2813
 T(°F)

2509   2310
2112   1941
   MAX
   EXPECT
   ED
   ADIA-
   BATIC
   TEMP.
             0.5% FUEL N
          30% EXCESS AIR

          r. 0.5 SEC.

                THERMAL NO
                   FUEL N
       0.45   0.50   0.55   0.60    0.65   0.70

                           103/T(K"1)
Figure 2.  Impact of temperature and fuel nitrogen on NOX
  emissions for excess air conditions  (calculated using
                    EER kinetic set).
                            17

-------
     4000
u.
e>
D.

LU
o
Ul
§


o
     3000
     2000 -
     1000 .
                                       Moisture

                                           15%
            % Excess Oxygen:     0
                                               10
                                                       12    14
                  50
                         SA  100
                                      150
                                                 200
                                                           250
300
           Vo Excess Air:
                                       50
                                                 100
                                                           150
250
        Figure 3.   Theoretical temperature of the products of
          combustion calculated from typical municipal  solid
          waste properties,  as a function of refuse moisture
                      and  excess air or oxygen.

       (Reproduced  from a paper titled "Minimizing Trace Organic
      Emissions from Combustion of MSW by Use of  Carbon Monoxide
         Monitors" by Floyd Hasselriis presented at the ASME
         Solid Waste Processing Division Conference  in  1986.)
                                   18

-------
rates).   Hence,  the amount of entrainment,  volatilization,  and
enrichment of metals in the smaller-diameter PM is expected to
increase if primary chamber temperatures are raised.
     Typically,  the primary chamber temperature is controlled by
controlling the feed rate and the combustion air as described in
previous sections.  Other alternatives include using auxiliary
fuel burners and/or water injection.  For systems that rely
primarily on airflow control, auxiliary fuel burners may be used
as necessary to maintain a minimum setpoint temperature.  The
control of the auxiliary fuel burners varies substantially from
unit to unit.  The least complex systems consist of simple on/off
switches that activate the burner or shut it off as setpoints are
reached.  More sophisticated systems fully modulate the burners
from low fire to high fire based on the primary chamber
temperature.  Some manufacturers also offer a water injection
system for rapidly decreasing primary chamber temperatures that
exceed specific maximum setpoints.  These systems are generally
simple on/off switches.
     For batch-type MWI's, manufacturers control the temperature
by switching the burners and combustion air between high/low or
                                            g n Q
on/off settings based on a series of timers.  '
     2.2.2.4  Control of Secondary  Chamber Temperature.  The key
secondary  chamber  combustion control parameter for MWI's is
temperature.  Typically, a thermocouple  is used  to monitor the
temperature  continuously near the secondary chamber outlet.   That
information  is then used  to modulate the amount  of  combustion air
supplied to  that  chamber  and, in some  cases,  to  modify  primary
chamber operations to  control the volatiles release rate.
Because the  secondary  chamber is operated  in  an  excess-air
condition, increasing  temperatures  trigger an increase  in  the
combustion airflow rate to bring temperatures back to the
setpoint.   (See  Figure 1.)   This increased combustion air  reduces
the secondary chamber  temperature  because  energy is required to
heat the excess  air from ambient to flue-gas  temperatures.
      The primary function of the secondary chamber is to complete
 the combustion process initiated in the primary chamber.  This
                                 19

-------
function is"best accomplished if the secondary chamber is
operated within a specified temperature range.  If temperatures
are too low,  the gaseous and solid organics may not be completely
destroyed.  Consequently, the system may emit excessive levels of
PM, CO, and organic emissions.  If the temperatures are too high,
refractory damage may occur, residence time may be decreased, and
auxiliary fuel may be wasted.  The limited data on pathogen
destruction indicates that pathogens can be destroyed at the
temperatures achievable in MWI's.  Experimental work at the
University of Dayton Research Institute indicates that
temperature is the primary factor that affects the decomposition
of organics and that a threshold temperature exists above which
an organic compound rapidly combusts.  The threshold temperature
found for polychlorinated dibenzo-p-dioxin (CDD) and
polychlorinated dibenzofurans  (CDF) and potential precursors
(e.g., hexachlorobenzene) is near 930°C (1700°F).7   Thermal NOX
formation is highly temperature dependent but only becomes
significant at temperatures in excess of 1650°C (3000°F).  As
stated in Section 2.2.3, thermal NOX formation is limited by
limiting the peak flame temperatures within the secondary chamber
to less than 1650°C  (3000°F).  Typically, upper setpoint
temperatures in the MWI secondary chamber  (about 1200°C [2200°F]
keep flame temperatures below those levels, thereby limiting
thermal NOX formation, and prevent refractory damage.  In
summary, a secondary chamber temperature operating range of 980°
to 1200°C  (1700° to 2200°F) promotes effective destruction of the
organics and pathogens, conserves auxiliary fuel, prevents
refractory damage, and avoids thermal NOX formation.
     The methods used to control the temperature in the secondary
chamber include controlling the feed rate and combustion air  (See
Sections 2.2.2.1 and 2.2.2.2) and the use of secondary chamber
auxiliary fuel burners.  The temperature in the secondary chamber
is typically controlled by modulating the combustion airflow.
Also, when necessary, the auxiliary fuel burners are operated to
maintain a minimum setpoint temperature.  Auxiliary burner
control varies from a single on/off switch driven by the lower
                                20

-------
setpoint to' fully modulating burners that range continuously over
low-fire to high-fire in response to secondary chamber
temperature.  Also,  since instantaneous peaks in the flow of
combustion gases generated in the primary chamber can occur, a
recommended operating practice is to keep the secondary chamber
auxiliary burner on at all times to ensure that the volatile
gases from the primary chamber combust completely during these
peak conditions.11
     2.2.2.5  Control of Primary Chamber Retention Time.  The
nonvolatile combustible portion of the waste  (fixed carbon) is
burned on the primary chamber hearth at high temperatures.  The
fixed carbon is burned in surface oxidation reactions.  This
solid-phase reaction requires higher temperatures and longer
exposure times than do the gas-phase reactions that occur above
the bed in the primary chamber and continue in the secondary
chamber.
     After the volatilization of the waste material in  the
primary chamber  is complete, sufficient  time must be provided for
the remaining fixed carbon combustion  in the  waste bed  to  be
completed.  For  batch and intermittent duty  incinerators,  there
is  typically a burndown  cycle that  is  controlled by a preset
timer and  temperature feedback  from the  primary  chamber.   During
this cycle, the  amount of fixed carbon remaining in  the waste
slowly  decreases.  To combust this  remaining fixed carbon,  a
preset  minimum  temperature  is maintained for a predetermined time
period.  The  thermal  input  needed to maintain temperature during
this  time  is  supplied by fixed-carbon combustion and,  when
necessary,  auxiliary fuel burners.   The  desired  burndown time is
preset  by a timer,  and  the  burndown cycle is initiated either
manually by a switch or through an automatic system that
 "initiates" the burndown cycle  automatically by restarting the
 burndown cycle every time the charge door is opened.   In this
 automatic system, burndown is tied to the last charge.  A typical
 burndown period for batch and intermittent units is in the range
 of 2 to 4 hours.10'12
                                 21

-------
     For continuous-duty incinerators there is no distinct
burndown period since the waste is continuously moving through
the system.  However, during shutdown, some continuous-duty
incinerators with stepped hearths and internal transfer rams use
a distinct burndown procedure.  After the last charge in a given
campaign, each ram goes into a sequential burndown mode.  The
internal ash rams, starting with the initial or drying hearth,
are operated in a mode that progressively increases the distance
of each stroke over a prescribed time period until all the waste
has been transferred to the next hearth.  At this point, the
underfire air to the initial or drying hearth is shut off, and
the same sequence of operations is applied to the next hearth
until all waste material is discharged to the ash pit.  This
burndown sequence is usually preset by the manufacturer but can
be changed if ash quality is poor.
     Three alternatives to the fixed-hearth systems described
above are the rotary kiln, the Pulse-Hearth™ and the stoker.  For
the rotary kiln the retention time of the waste for burndown  is
determined by the rate of rotation and the angle of incline or
rake of the kiln.  One manufacturer of continuous-duty MWl's uses
a Pulse-Hearth™ for moving waste material through the primary
chamber.  The retention time of the waste for burndown in this
system is controlled by the frequency of the hearth pulses that
move the waste along the hearth.  Another manufacturer provides a
stoker system to move waste through the primary chamber.  The
stoker comprises a series of overlapping, alternating stationary
and movable grates.  While a movable grate positioned over a
stationary grate is advancing, a movable grate positioned under
that stationary grate retracts to form a step 15 inches high.
This action causes waste to fall and mix as it moves across the
length of  the stoker.
     2.2.2.6  Control of Secondary Chamber Gas Residence Time.
In the modern MWI, the secondary combustion chamber serves as a
reaction vessel to complete conversion of the organic materials
released from the primary chamber.  Although some of the
carbonaceous material in entrained particulate matter may
                                22

-------
combust,  most of the chemical reactions in the secondary chamber
are gas-phase reactions that involve volatile organic materials
released from the waste and oxygen in the secondary combustion
air.  The specific reaction chains by which the organic material
is converted to C02 and H20 are not fully understood and are
likely to be quite complex.  However, they are generally assumed
to behave according to first order reaction kinetics.  Under such
an assumption, the rate of conversion of organic material to CO2
and H20 is a function of the temperature/time envelope provided
by the secondary combustion chamber.  If the residence time is
not sufficient, given the system's operating temperature,
complete combustion of the organics and conversion of CO to C02
will not be accomplished.
     The residence time in the secondary chamber is a function  of
the size of the chamber and the volumetric flow through the
chamber.  In  its simplest  form, the average  residence time can  be
calculated as:
       At.
              S
                                                            Eq.  1
 where:
     At   - secondary chamber residence time,  sec;
      S                               -3
     V   = secondary chamber volume,  m ,•  and
     O   = combustion gas flow rate at secondary chamber
      s                 ^
           conditions, irr/sec.
     The primary mechanism used to control residence time is
 appropriate sizing of the secondary chamber during the design
 process.  Information collected from manufacturers indicates that
 systems are sized to have secondary chamber residence times of
 0.25 to over 2 seconds under normal operating conditions.  The
 two approaches that have been used to increase residence time are
 to increase the size of a single chamber or to construct two
 smaller chambers in series.  For this latter approach, the two
 chambers are typically called the secondary and tertiary chamber.
 respectively.  In theory, the two approaches should provide
                                 23

-------
equivalent residence times and emission reductions.  No data that
demonstrate differences have been identified.
     Volumetric flow also affects residence time.  Typically,
this flow is not controlled directly in MWI's.  Rather, it is a
function of the feed rate, combustion air, and temperature
control systems described earlier.  Consequently, controlling
these parameters to levels that are within design specifications
is the key to achieving design secondary chamber residence times.
     Secondary chamber residence times affect the control of
organic emissions, CO concentrations, and, to a lesser extent, PM
emission rates.  Residence time does not affect metals or acid
gas emissions.  Increases in residence time at a given
temperature decreases emissions of CO and organic pollutants, if
properly mixed, by giving the oxidation reactions more time to
move to completion.  Increased residence time can also decrease
PM emissions by providing additional time for the fixed carbon in
entrained particles to combust and by allowing more complete
combustion of condensible hydrocarbons.
     2.2.2.7  Control of Secondary Chamber Mixing.  The reaction
rate for organic materials in the secondary chamber is a function
of the temperature/time envelope provided by the chamber, but the
reaction will proceed toward completion only if the gas phase
organics from the primary chamber are well-mixed with the oxygen
from the secondary chamber combustion air.  Hence, complete
combustion of the organics and CO in the secondary chamber
requires turbulent mixing of the primary chamber exhaust and
secondary chamber combustion air.  As described in the paragraphs
below, manufacturers have used a variety of techniques to promote
turbulent mixing in the secondary chamber.
     Two general types of techniques are used by manufacturers to
promote secondary chamber mixing.  First, all manufacturers use
the location, direction, and velocity of the secondary air ports
to promote turbulent flow, although specific designs vary among
the different manufacturers.  In addition, some manufacturers use
constrictions or physical barriers such as baffles in the
secondary chamber to increase gas velocity and create turbulent
                                24

-------
flow.  Note' that both types of techniques are related to the
design rather than the operation of the facility.   However,  the
system operation does play a key role in maintaining turbulent
conditions in that the different mechanisms designed to promote
mixing are related to combustion air and combustion gas
velocities.  If the system operating rate drops to the point that
gas velocities drop below minimum design specifications,
insufficient mixing can result.
     Each manufacturer has a unique system for introducing
combustion air to the secondary chamber.  Most MWI's contain a
short, narrow passage between the primary and secondary chambers
that is known as the flame port.  Many designs introduce most of
the secondary air into this flame -port at a direction
perpendicular to the flow of the primary chamber exhaust gases.
The high gas velocities in this area are purported to generate a
well mixed stream.  While many systems introduce the major
portion of secondary combustion air into the flame port, other
designs introduce the air at different locations in the secondary
chamber.  Common designs include high-velocity air jets at
multiple axial  locations that  introduce  combustion air  in a
direction perpendicular to the gas flows at velocities  sufficient
to penetrate  to the  center of  the  chamber  and airports  near the
flame  port opening that introduce  air  tangential to  the gas
stream to promote  cyclonic  flow.   One  system introduces secondary
combustion air  through a device known  as a thermal  exciter.13
Although  these  designs have  different  geometries,  they are  all
designed  to  achieve  complete penetration of the primary chamber
exhaust gas  cross-section  by the  combustion air and to promote
gas  turbulence  throughout  the secondary chamber.
      In addition to  careful introduction of secondary air,  some
manufacturers use physical restraints  in the secondary chamber to
 promote turbulence.   One  manufacturer  uses a long,  narrow,
 cylindrical secondary chamber to increase gas velocities  and
 thereby promote turbulence.  Other designs include locating
 baffles at multiple positions along the chamber and using a
 refractory choke ring about halfway along the cylindrical chamber
                                25

-------
to generate" turbulent flow.  No quantitative measures of the
effect of these physical measures on gas turbulence are
available.
     A well-mixed gas stream is needed to ensure intimate contact
of organics in the gas stream with oxygen.  Mixing also provides
a more uniform temperature profile throughout the chamber and
minimizes potential "short-circuiting" by portions of the gas
stream.  Good mixing reduces both organic and CO emissions and
may provide some reduction in PM emissions.  However, no
quantitative measures of mixing are available.
     2.2.2.8  Control of Solids Mixing in the Primary Chamber.
Solids mixing can have conflicting impacts on the control of
emissions from MWI's.  On the one hand, good mixing of the solids
in the bed is needed to promote uniform temperatures in the bed,
enable release of all volatile organic materials, and provide
contact between all solid surfaces and the primary combustion air
to ensure complete fixed carbon burnout.  In contrast, excessive
bed turbulence can lead to increased particle entrainment and,
consequently, increased PM and metals emissions.  Consequently,
bed mixing or turbulence must balance these effects.
     The two principal types of mechanisms that are used to
control solids mixing are the physical processes that are used to
move the waste along the primary chamber hearth and turbulence
generated by injection of primary combustion air into the bed.
Primary combustion air injection systems are discussed in
Section 2.2.2.9.  The paragraphs below discuss the physical
mechanisms used to promote appropriate bed turbulence in three
types of units--the fixed-hearth system  (both single and multiple
hearth) , the pulse-hearth system, and the rotary kiln system.
The basic concepts for these three systems are described in
Section 2.2.2.5; the discussion below focuses on system
parameters that affect solids mixing.
     The most common MWI configuration is the fixed-hearth
system.  Historically, these systems were constructed with a
single hearth and had either no ram  (smaller, batch-fed units) or
a single feed/ash ram.  Feed/ash ram systems are found on larger
                                26

-------
MWI's both to move waste/ash through the MWI effectively and to
provide solids mixing.  Newer designs contain multiple (three to
four) hearths with a separate ram for each hearth.  Single-hearth
systems may have only a feed ram where new waste charges provide
solids mixing.  In general, these multiple-hearth systems mix
solids in the bed better than do the single-hearth systems.  The
degree of turbulence in the bed of these systems can be
controlled by controlling ram speed and ram stroke length.  An
increase in either of these parameters is likely to increase
turbulence in the bed.
     One alternative to the fixed-hearth design is the pulse-
hearth design.  Solids on these hearths are mixed via the pulsing
action used to move the solids along the hearth.  The degree of
solids turbulence is a function of the pulse intensity.  The
manufacturer of this system claims to achieve better solids
mixing than is achieved in fixed hearth units, but no data have
been received to support this claim.
     In the past 3 years, a limited number of rotary kiln MWI's
have been installed in the United States.  The rotary action
generates substantially greater solids mixing than is achieved in
either the fixed-hearth or the pulsed-hearth system.  Within a
given kiln, the degree of turbulence in the waste bed is a
function of the rotational speed of the kiln, the angle of
incline or rake of the kiln, waste characteristics, and the waste
feed rate.
     Ideally, control of solids mixing  (or turbulence)  in  the
primary chamber should be  optimized to address conflicting
objectives--maximizing ash burnout and minimizing particle
entrainment  in the exhaust gas from the primary  chamber.   In
general,  increasing mixing increases burnout, but it also
generates  higher  concentrations  of particles  in  the exhaust  gas.
     2.2.2.9  Sizing  and Location of Combustion  Air Ports  in the
 Primary Chamber.  The sizing and location of  primary  chamber
 combustion air ports  are designed to address  three objectives
 concurrently--to  provide uniform airflows in  the bed  to control
volatiles  release,  to promote  complete  fixed-carbon burnout,  and
                                27

-------
to minimize" particle entrainment in the primary chamber exhaust
gas.  Also,  systems are designed to maintain bed temperatures at
levels that limit clinker formation and refractory damage.
     Generally, combustion air is introduced to the primary
chamber as "underfire" air, but some systems introduce the
majority of primary chamber air through ports located immediately
above the waste bed.  While most systems do introduce underfire
air into the side or bottom of the waste bed, the designs vary
widely with respect to location of the injection ports, port size
and geometry, and distribution of airflows along the combustion
zone.  Locations of the air ports include the side walls below
bed levels,  openings in the hearth itself, and the centerline of
the ash ram.  Geometries range from thin slots in the side walls
that run the length of a hearth to small-diameter, circular
openings in the hearth and the side walls to large-diameter low-
velocity openings in the side walls and ash rams.  Manufacturers
also have different methods for distributing air along the
hearth.  Some systems introduce air uniformly along the total
length of the hearth or along all hearths in a multiple hearth
system  (with the possible exception of the first hearth).  Other
systems vary the flow rates along the length of the hearth, with
some systems providing higher flows at the charging/drying hearth
and others providing higher flows at the discharge end.
     Irrespective of the specific mechanism employed, the
underfire air system design has three key impacts on emissions.
First, the degree to which the system can operate in a consistent
fashion without plugging or malfunctioning, affects how well the
release of volatiles to the secondary chamber can be controlled.
Controlling the release of volatiles to the secondary chamber
affects CO and organic pollutant emissions.  To reduce the
potential for air port plugging, some continuous MWI's are
designed with a mechanism that allows combustion air ports to be
rodded or cleaned out while the MWI is in operation.
Intermittent and batch MWI's require that the air ports be rodded
out on a daily basis prior to the start of the burn.  Second, the
degree to which the system provides good oxygen contact with ash
                                28

-------
surfaces as' the ash moves toward the discharge end of the chamber
affects ash burnout.  Finally,  the turbulence generated by the
combustion air system affects particle entrainment and PM
emissions.
     2.2.2.10  Steam In-iection in the Primary Chamber.  Some
manufacturers use steam injection into the ash bed to facilitate
burnout of fixed carbon in the ash bed and at the same time help
prevent hot spots and clinker formation on the hearth.14'15  The
steam is injected into the ash bed through the underfire air
ports.  The steam reacts with the fixed carbon in the waste/ash
bed to produce CO and hydrogen  (H2)•  This reaction is an
endothermic reaction  (absorbs heat), thus reducing the
temperature in the ash bed.  (The'normal oxidization reaction of
oxygen with carbon  is exothermic and evolves heat, causing the
ash bed temperature to rise and potentially resulting in hot
spots adjacent to the air ports).  The steam injection promotes
complete burnout of the ash, and also helps to prevent hot spots
in the ash bed adjacent to the  air ports which might promote
slagging.
     Different techniques are used by different manufacturers for
injecting steam.  One manufacturer uses a series  of  staged
hearths in its large  incineration designs.  The steam is  injected
through underfire ports  in the  last  hearth  in  lieu of air.15
Another manufacturer  of  an intermittent-duty  incinerator mixes
the  steam with the  underfire air, which is  distributed throughout
the  entire area  of  the hearth.14 During  operation,  steam
injection is  not used during the first part of the incineration
cycle.   Once  an  ash bed  is developed,  the steam injection is
activated,  and the steam is  injected with the combustion air;  the
amount of air injected  is reduced by the  volume of steam
 injected.
      2.2.2.11  Operator Training.  Previous sections identify key
 operating parameters  for the incineration systems and indicate
 the relationship between these parameters and potential
 emissions.   No studies  have been conducted to determine the
 effect of operator training (or the lack thereof) on emission
                                 29

-------
levels.  Nonetheless, it is reasonable to presume that operator
training can result in the reduction of emissions.  An operator
who understands how the key operating parameters affect emissions
and who knows the preferred operating procedures can operate and
maintain the incinerator and APCS in a manner that minimizes
emissions.
     Using proper waste charging procedures for an incinerator is
an example of an operation that can affect emissions.  Training
of the operator responsible for charging of the incinerator helps
to ensure that the incinerator is charged in such a manner as to
minimize emissions.  The operator should be trained in the proper
handling procedures and steps for charging the unit to minimize
fugitive emissions during charging.  These procedures typically
are provided by the manufacturer in the MWI operation manual.
Also, maintaining the proper frequency and size of each charge
insures that the proper charge rate  (heat input rate) is
maintained, thereby enhancing controlled combustion.  Many newer
incinerators use sophisticated controls that result in an
automated operation with few functions under the direct control
of the operator.  For example, the system may include the
automatic control of the secondary burner and combustion air
rates as a result of temperature output from the secondary
chamber thermocouple.  Also, some newer systems include an
automated feed system which controls the frequency and weight of
the charges.  However, using automated controls does not
alleviate the need for operator training; it simply refocuses the
areas in which training is required.  Although the operator may
not need to understand the theoretical aspects of how the
automated control system operates, someone responsible for the
operation of the unit must understand how the system is designed
and intended to operate, how to establish proper setpoints, and
how to recognize the telltale signs of a malfunction or need for
calibration/adjustment.
     Operator training with respect to proper maintenance and
operation is important.  If the incinerator and its combustion
control system are not properly maintained, then the system will
                                30

-------
not operate" properly.  The result is likely to be increased
emissions.  For example, if the primary chamber underfire air
ports are not maintained such that they remain clear of debris,
the combustion air distribution will be affected and will affect
the primary chamber temperature and, ultimately, emissions and
ash quality.
3.0  ADD-ON AIR POLLUTION CONTROL SYSTEMS
     Many APCS's consist of one add-on control device to control
particulate matter  (PM) and another to control acid gases
 (primarily hydrogen chloride  [HC1] gas).  These add-on control
devices can be wet scrubbers, dry scrubbers, or fabric filters.
The following sections describe these add-on control devices.
Each section contains a general description of the control device
physical characteristics  (i.e., the various pieces of equipment
and how they fit together), how the control device cleans the
flue gas of its primary pollutant(s) and its effectiveness,  the
key operating parameters,  and a qualitative discussion of how
much co-control of  other pollutants can be achieved.
Additionally, each  section contains a qualitative discussion of
the factors that affect the performance of the  control device  in
controlling PM, acid gases, metals, and CDD/CDF, including
specific  design features  and  gas  stream characteristics.
     This  section describes wet  scrubbers, dry  scrubbers,  and
 fabric filters and  their  application to MWI's.   Section  3.1
 describes venturi,  packed-bed, and  other types  of wet  scrubbers.
 Section  3.2 describes  pulse-jet  fabric filters.  Section 3.3
 describes dry  sorbent  injection  and spray  dryer dry scrubbers.
 3.1  WET SCRUBBERS
     A venturi scrubber in combination with a packed-bed scrubber
 is the most common wet scrubber  system used to  control emissions
 from MWI's.  Venturi scrubbers are  used primarily  for PM control
 and packed-bed scrubbers are used primarily for acid-gas control.
 However,  both scrubber types achieve some  degree of control of
 both PM and acid gases.  A large amount of information and
 emission test data are available for these scrubber types.  Other
 types of wet scrubbers that have found limited application to
                                 31

-------
control emissions from MWI's include the Rotary Atomizing™
scrubber,  the Ionizing Wet Scrubber™, the Calvert collision
scrubber,  and the steam-ejector scrubber.
     Section 3.1.1 describes wet scrubbing mechanisms.  Venturi
and packed-bed scrubbers are described in detail in Sections
3.1.2 and 3.1.3.  The other scrubbers mentioned above with
potential application to MWI's are briefly described in
Section 3.1.4.
3.1.1  Wet Scrubbing Mechanisms
     Wet scrubbers collect particles primarily through the
mechanisms of impaction and diffusion and remove gaseous
pollutants primarily through absorption.  The following
subsections describe these mechanisms.
     3.1.1.1  Collection of PM.  Wet scrubbers capture relatively
small dust particles with relatively large liquid droplets.
These droplets are produced by injecting liquid at high pressure
through specially designed nozzles, by aspirating the particle-
laden gas stream through a liquid pool, or by submerging a
whirling rotor in a liquid pool.  These droplets collect
particles through two primary collection mechanisms: impaction
and diffusion.  With both of these mechanisms, PM collection
increases with an increase in relative velocity  (liquid- or gas-
phase input) and a decrease in liquid droplet size.16
     In a wet scrubbing system, dust particles with diameters
greater than 1.0 micrometer  (/im) tend to follow the streamlines
of the exhaust stream.  However, when liquid droplets are
introduced into the exhaust stream, particles cannot always
follow these streamlines as they diverge around the droplet
 (Figure 4}.  The particle's mass causes it to break away from the
streamlines and impact on the droplet.  Impaction is the
predominant collection mechanism for scrubbers with a gas stream
velocity greater than 0.3 m/sec  (1 ft/sec), which is well below
the gas stream velocity, experienced in most scrubbers.17'18
     Very small particles  (less than 0.1 /xm in diameter)
experience random movement in an exhaust stream.  These particles
are so tiny that they are  "bumped" by gas molecules as they move
                                32

-------
                Gas streamlines
                 Particle
                              /  \ Droplet
                     Figure 4.  Impaction
                                          18
in the exhaust stream.  This bumping, or bombardment, causes the
particles to move first one way and then another in a random
manner, i.e., to diffuse through the gas.  This irregular motion
(or diffusion) can cause the particles to collide with  a droplet
and be collected  (Figure 5),19  The rate of diffusion depends  on
relative velocity, particle diameter, and liquid droplet
diameter.  Collection by diffusion increases as particle size
decreases.  This mechanism enables certain scrubbers to remove
the very tiny particles  (smaller than 0.1 Jim)  effectively.19
      In summary,  impaction is  the predominant  PM collection
mechanism  for particles greater than 1.0 /mi while  diffusion is
the predominant mechanism  for  particles  smaller than 0.1  jmi.   For
particles  between 0.1 and  1.0  /im, neither mechanism dominates;
particles  in this size range are  collected by both mechanisms.
However, particles in this size  range are not collected as
efficiently as  are either  larger particles by impaction or
smaller particles collected by diffusion.18'19
      3.1.1.2 Absorption of Gaseous  Pollutants.   The process of
dissolving gaseous pollutants  in a liquid is referred to as
absorption.  Absorption occurs when mass is transferred as a
 result of  a concentration difference between the liquid and the
 gas from which the contaminant is being removed.   Absorption
                                33

-------
            Particle
                                    Gas streamlines
\ Droplet
                     Figure 5.  Diffusion.19

continues as long as a concentration differential or lack  of
equilibrium exists.  In absorption, equilibrium depends  on the
solubility of the pollutant in the liquid.20
     To remove a gaseous pollutant by absorption, the  exhaust
stream must be brought into contact with  a  liquid.  Figure 6
illustrates the three step process involved in absorption.
First, the gaseous pollutant diffuses from  the bulk area of the
gas phase to the gas-liquid interface.  Second, the gas  molecule
moves (transfers) across the interface  to the liquid phase.  This
step occurs extremely rapidly once the  gas  molecules  (pollutant)
arrive at the interface area.  Third, the gas molecules  diffuse
into the bulk area of the liquid, thus  making room for additional
gas molecules to be absorbed.
     The rate of absorption depends on  the  diffusion rates of the
pollutant in the gas phase  (first step) and in the liquid  phase
 (third step).  Gas diffusion and, therefore, absorption  can be
enhanced by:
     1.  Providing a large interfacial  contact area between the
gas and liquid phases;
     2.  Providing good mixing of the gas and liquid phases
 (turbulence), and
                                34

-------
               Gaseous
               pollutant
                    Figure 6.  Absorption
                                          20
     3.   Allowing sufficient residence,  or contact,  time between
                                   20
the phases for absorption to occur.
     A very important factor affecting the amount of a pollutant
that can be absorbed is its solubility.  The solubility of the
pollutant governs the amount of liquid required and the necessary
contact time.  More soluble gases  require less liquid.  Also    ^
more soluble gases will be absorbed  faster.   (For example  HC1  ,s
absorbed more rapidly than S02  in  water because HC1 x. much more
soluble.)  Solubility is a function  of both  the temperature and,
to a lesser  extent,  the pressure of  the system.  As temperature
increases, the  amount of gas  that  can be  absorbed by  a  liquid
decreases.   For this reason,  some  absorption systems  use xnlet
quench sprays to cool  the  incoming exhaust stream,  thereby
 increasing absorption efficiency.   Pressure affects the
 solubility of a gas in the opposite manner.   When the pressure of
 a system is increased,  the amount of gas absorbed generally
           21
           *
increases .
3.1.2
              i grubbers
      This section provides a general description of a venture
 scrubber system including equipment components, the gas  cleaning
 process, and key operating parameters  followed by a discussion of
 the  factors affecting  the performance  of  the venturi.
                                 35

-------
      3.1.2.1   General  Description.  A venturi scrubber system
 typically is made  up of the following components:
      1.   Scrubber  vessel;
      2.   Cyclonic  separator;
      3.   Induced-draft (ID)  fan;
      4.   Liquid recirculation system;  and
      5.   System controllers.
 These components are described in the following paragraphs.
      Venturi scrubbers are designed to remove PM primarily by
 impaction through  high-energy contact between the  scrubbing
 liquid and the suspended PM in the gas stream.22  Figure  7
 illustrates a  simplified venturi  scrubber vessel consisting of a
 converging section,  a  throat  section,  and a  diverging  section.23
 The high-energy contact is achieved by directing the scrubbing
 liquid and the gas stream  to  the  throat section.   The  gas stream
 to be scrubbed enters  the  converging section (or quench) and  as
 the gas flows  towards  the  throat  and the cross-sectional area
 decreases,  the gas velocity and turbulence increase.23'24  The
 exhaust gas, pulled  through the throat by the system's ID fan and
 forced to  move at extremely high  velocities  in  the throat, shears
 the liquid from the  throat walls  and atomizes the  liquid into a
 large  number of  small  droplets.23'24  (Another  method of
 producing  droplets is  to atomize  the liquid  by  supplying high-
 pressure liquid  through spray nozzles  with small orifices),24
 Impaction  of the PM  on the droplets  occurs in the venturi throat
 and the exhaust  stream exits through the  diverging section where
 it is  forced to  slow down.23
      In venturi applications on MWI's, either a special quench
 section and/or the converging section  of  the  scrubber vessel acts
 as a quench, both to saturate and to cool the gases prior to the
venturi throat.  The hot MWI exhaust gas  is cooled by evaporating
 the quench liquid to an equilibrium temperature that,  for most
 incinerator applications,  lies between 66° and 85°C
 (150° and  185°F),25
     Two types of venturi  scrubbers may be installed in MWI
applications--wetted- and nonwetted-approach Venturis.   Wetted-
                                36

-------
                                    Converging
                                      section

                                  - Throat
                                  ^ Diverging
                                      section
Figure 7.  Simplified venturi scrubber configuration

                           37
                                                      23

-------
approach Venturis are the most  common.   The primary difference
between these two types  is  that in a wetted-approach venturi,
which  is illustrated in  Figure  8,  the scrubbing  liquid  is
introduced through nozzles  located at the  top  of the converging
section to wet the scrubber walls.   In a nonwet ted-approach
venturi, the liquid is injected through nozzles  that direct the
liquid directly into the throat without wetting  the walls.
Wetted-approach Venturis are used  when inlet gas streams are at
temperatures greater than 93°C  [200°F].  Coating the scrubber
walls  and throat with liquid reduces the tendency for PM to
abrade or cake on the walls or  throat.26"28
     Because of the variations  in  MWI gas  flow rate caused by
waste  charging and combustion air  modulation,  most  venturi
installations, in addition  to being wetted-approach Venturis, are
variable throat Venturis.   Figure  9 illustrates  a wetted-
approach, variable throat,  venturi  scrubber.29   Variable throat
Venturis use either a throat insert or  an  adjustable plate to
decrease or increase the cross-sectional area  of  the venturi
depending on the gas flow conditions.30  The purpose of the
variable throat is to maintain  constant  gas velocity across the
throat, thereby maintaining venturi PM  collection performance.
Although a variable throat  venturi  can  create  restrictions in gas
flow that cause the MWI  draft to become  positive, a venturi
installed with a variable speed ID  fan maintains  negative draft
in the MWI.  As shown in Figure 9,  scrubbing liquid also can be
introduced through the throat insert.
     The scrubbed gas from  the  venturi throat passes into the
diverging section where  the gas slows down.23  Typically,  the
base of the diverging section is submerged in a pool of liquid to
prevent abrasion of the metal surfaces by the collected PM.31'32
This configuration is called a wetted or flooded elbow and is
illustrated in Figures 9 and 10.
     The PM entrained in the droplets from the venturi scrubber
subsequently is collected either in the packed-bed scrubber or in
a cyclonic separator.   Typically, in MWI applications,  the gas
passes from the wetted elbow to the base of a packed-bed scrubber
                                38

-------
                                              Liquid
                                               inlet
Figure 8.  Schematic of wetted-approach venturi scrubber.

                             39
                                                           16

-------
            VARIABLE  VENTURI  SCRUBBER
                  WITH  WETTED  ELBOW
                          HOT GAS
                           IWLET
   •THIMBLE"INLET
scauwiNe LIQUOR
TO THROAT
(BOX OF LIQUID)
ALTERNATE OR SUPPLIMENTAL
NOZZLE LOCATION FOR VERY
HIGH TEMPERATURE GASES

TANGENTIAL LIQUID  INLETS
 (40X OF LIQUID)


CONVERGING INLET-WETTED
NITH TANGENTIAL LIQUID
                                              SCRUBBING LIQUOR
                                              TO THROAT

                                              THROAT INSERT
                                              THROAT-CROSS SECTION
                                              VARIES NITH INSERT
                                              POSITION

                                              EXPANDER SECTION
                                              WETTED ELBOW-FILLS
                                              WITH LIQUID
HYDRAULIC OR MECHANICAL
ADJUSTMENT FOR THROAT
  Figure 9.   Schematic of wetted-approj
                 throat venturi  scrubber."
 :h, variable
                              40

-------
                                               Cyclonic
                                               separator
                 Flooded
                  elbow
Figure 10.  Venturi scrubber  illustrating the  flooded (wetted)
  elbow and exhaust gases passing to a cyclonic  separator.-3-*
                                41

-------
used for acid gas control where the entrained PM mixed with the
scrubbing liquid is ultimately removed from the system with the
scrubber liquid discharge  (blowdown).  (Packed-bed scrubbers are
described in Section 3.1.3.)  However, in those applications
where a packed-bed scrubber is not used,  the gas passes from the
wetted elbow to the base of a cyclonic separator (see Figure 10).
The gas is introduced tangentially tp the cyclonic separator
where the particles and moisture are thrown outward to the walls
of the cylinder.33  The droplets coalesce and drop down the walls
to a central location carrying the PM with them.33
     The ID fan of a venturi scrubber system is located after the
cyclonic separator  (or packed-bed scrubber if so equipped) and
pulls the exhaust gas through the system.  Some installations
utilize a variable-speed ID fan that can be tied to the control
of the venturi throat insert and/or the MWI primary chamber draft
so that the desired draft  in the MWI primary chamber and a
constant venturi throat pressure drop can be maintained.  Other
systems utilize a single speed ID fan equipped with a damper to
control airflow.
     The scrubber liquid recirculation system comprises a caustic
solution tank, caustic addition equipment, recirculation pump,
and piping.  The liquid is recirculated through the system
continuously.  In many systems, a single recirculation pump is
used  (even on systems with packed-bed scrubbers) with a backup
pump in parallel.  A small blowdown stream is bled off to remove
collected PM and a fresh water makeup stream is added to the
system to replace the blowdown and any evaporative losses.
Because the venturi scrubber liquid also absorbs a portion of the
HC1 in the exhaust gas, the pH of the scrubber liquid decreases.
In most venturi scrubber applications, the scrubber liquid is a
caustic solution used to neutralize acid gases and to prevent
corrosion of equipment.  A pH meter monitors changes in the
scrubber liquid pH and a controller adds caustic solution to the
system to maintain a pH of 7.0 or less.22
     Venturi system controllers include those for the venturi
throat insert, caustic feed, makeup water, and emergency water
                                42

-------
quench for"high-temperature excursions.   The venturi throat
insert is controlled either pneumatically or hydraulically to
maintain constant throat velocity.  The throat insert is moved
upward or downward based on the pressure drop measured across the
venturi throat.  The rate at which caustic is added to the system
is controlled in order to maintain a preset pH, typically 7.0 or
slightly less.22  The rate at which makeup water is introduced to
the scrubber typically is set at a rate necessary to replenish
the liquid lost to blowdown and evaporation.
     3.1.2.2  Factors Affecting Performance.  The performance of
the venturi scrubber relative to PM, acid gas, metals, and
CDD/CDF emissions is affected by the venturi design and operating
variables and also by key MWI operating variables.  Key venturi
scrubber design and operating variables are the pressure  drop
across the venturi throat, the liquid/gas  (L/G) ratio, and  the
scrubber liquid pH, surface tension, and turbidity  (i.e.,  solids
content).  Key  process parameters  that  can  affect venturi
scrubber performance are the particle size  distribution  and the
variations in  temperature,  flow rate, and pollutant
concentrations  that result  from the heterogeneous nature of the
MWI process.
      Venturi  scrubbers  are  used on MWI's primarily  to control PM
emissions.  The PM collection  efficiency  in a venturi scrubber
system increases as  the pressure  drop  increases.34   Figure 11
 shows this relationship based  on  the particle size  distribution
 in Figure 12.   The static  pressure drop is a measure of the total.
 amount of energy used in the scrubber  to accelerate the gas
 stream and to atomize the  liquid  droplets.   In MWI applications.
 a common venturi scrubber pressure drop is 7.5 x 103 Pascals (Pa)
 (30 inches of water column [in.  w.c.])  but can range from
 3.7 x 103 to 1.5 x 104 Pa (15 to 60 in. w.c.).  The pressure drop
 across the venturi is a function of the gas velocity and L/G
 ratio and in practice is a surrogate measure for gas velocity.36
      The Calvert equation can be used to predict the pressure
 drop for a given throat velocity.  The Calvert equation  is:
                                 43

-------
 VERT AII5: DIFFERENTIAL PRESSURE (INCHES U.S.}

mil AIIS: Z COLLECTION EFFICIENCY
        : FROM  0 TO  100 IN STEPS OF  1
Figure 11.   Venturi  scrubber  collection  efficiency
           versus  differential pressure.35
                             44

-------
 .5
                                                              30.
100
 VERT AIIS: AERODYNAKIC PARTICLE SIZE CB1CRONS)
«u tin.
                               SIIE
      Figure 12.   Typical  medical  waste incinerator exhaust
                 gas particle size  distribution. D
                                   45

-------
                    AP  =  (5  x 10~5)  v2 (L/G)
where
     £p = pressure drop, in. w.c.
      v - gas velocity in the venturi  throat, feet/second, and
    L/G - liquid-to-gas ratio, gallons/1,000 actual cubic feet
          (gal/Macf).
The equation implies that pressure drop is equal to the power
required to accelerate the liquid to the gas velocity.  The
Calvert equation predicts pressure drop reasonably well for the
range of L/G ratios from 668 to 1,604  liters/1,000 actual cubic
meter (L/Macm) (5 to 12 gal/Macf).  At L/G ratios above
1,604 L/Macm  (12 gal/Macf), measured pressure drops are normally
about 80 percent of the value predicted by the Calvert equation.
In practice it has been found that at  L/G ratios less than
401 L/Macm (3 gal/Macf), there is an inadequate liquid supply
available to completely cover the venturi throat.  Most venturi
scrubbers are designed for liquid feed rates between 936 and
1,337 L/Macm  (7 and 10 gal/Macf), and  there is virtually no
change in performance over this range  given that the pressure
drop across the venturi throat remains constant.37
     The performance of a venturi scrubber in removing PM is
strongly affected by the size distribution of the PM.  For
particles greater than 1 to 2 /zm in diameter, impact ion is so
effective that penetration of particles in this size range
through the venturi is quite low.  However, penetration of
smaller particles, such as the particles in the 0.1 to 0.5 pirn
range, can be high.  Small particle size distribution, resulting
from the condensation of partially combusted organic compounds
and metallic vapors, is typical for combustion sources including
MWI's.38
     To attain a high PM collection efficiency, venturi scrubbers
need to achieve high gas velocities in the throat.  At normal
venturi pressure drops be.tween 5 x 103 and 1.5 x 104 Pa (20 and
60 in. w.c.), the gas velocity in the  throat section typically
lies between  30 and 120 meter/second  (100 and 400 feet/second).34
These high gas velocities atomize the  water droplets and create
                                46

-------
the relative velocity differential between the gas and the
droplets to effect particle-droplet collision.
     The high exhaust gas velocities in the venturi result in a
very short contact time between the liquid and gas phases,
thereby limiting gas absorption into the scrubber liquid.34  The
limited residence time limits the amount of vaporous metallic and
organic compounds that may be absorbed.  However, because of the
high solubility of HC1 in water, a significant quantity of HC1
may be absorbed even though the residence time is short.  One
vendor claimed an HC1 removal efficiency of 90 percent across a
Venturi scrubber.39
     Other variables related to venturi scrubber performance are
the liquid surface tension and liquid turbidity.  If surface
tension is too high, some small particles that impact on  the
water droplet will "bounce" off and not be captured.  High
surface tension also has an adverse effect on droplet formation.
However, surface tension typically does not significantly impact
scrubber performance.  High liquid turbidity, i.e., high
suspended solids content, will  cause erosion  and abrasion of  the
venturi section and ultimately  lead to reduced performance -of  the
system.40  Therefore,  the presence of a blowdown stream to  remove
suspended solids is important.
     When hot MWI  exhaust gases are quenched  and saturated  in the
converging  section of  the venturi, an  equilibrium temperature of
between 66°.and  85°C  (150° and  185°F)  typically  is achieved.41
In this temperature  range, volatile metals  and unburned organics
condense.41 As  vaporous metals and organic constituents  such as
                                                    A O
CDD/CDF cool,  they condense,  or agglomerate,  on  PM.    The
greater the cooling,  the  greater  these condensation and
agglomeration effects.42
      The variations  inherent in the operation of MWI's translate
 to variations in gas stream characteristics including
 temperature, flow rate,  and pollutant concentrations.  In many
 MWI applications,  a waste heat recovery boiler  (WHB)  is used both
 to recover heat for use at the facility and to cool the MWI
 exhaust gases prior to an APCS.  The WHB tends to damp the
                                47

-------
 the
to
variations in the exhaust gas temperature with the result that
the WHB outlet temperature fluctuates far less than the exhaust
gas temperature.  With a L/G ratio of up to 1,337 L/Macm
 (10 gal/Macf) , this small fluctuation does not affect the
performance of the venturi significantly.
     On venturi scrubbers controlling MWI's without a WHB,
MWI exhaust gas temperature variation can range from 927°
1260°C  (1700° to 2300°F).  Venturi scrubbers in these
applications may be equipped with special quench sections ahead
of the converging section to cool the gas stream.  Emergency
quench nozzles may be.installed at the top of the converging
section that rapidly provide the. necessary quench water to cool
the gases and to prevent damage to the system in upset
conditions.  If the gases are not cooled to the typical
equilibrium temperature between 66° and 85°C  (150° and 185°F) ,
then the venturi will be less effective in condensing volatile
metals and organics.
     Variations in MWI exhaust gas flow rate are compensated  for
by a variable throat venturi.  Variations in the MWI exhaust  gas
flow rate arise during waste charging and during combustion air
modulation in the MWI.  Venturi scrubbers perform optimally when
they operate at conditions that approach steady state.  The
variable throat venturi  adjusts the throat cross-sectional area
to maintain  constant velocity across the throat.
     Variations in pollutant concentration in the MWI exhaust gas
are dependent primarily  on waste composition.  For example,
 changes in the  chlorine  content of the waste directly affect  the
HC1 concentration in the exhaust gas.  The PM and volatile
 organics  concentrations  can also change as the combustion process
 changes.  Because venturi scrubbers are designed to remove a
 certain percentage of the incoming PM based on the particle size
 distribution and inlet  loading, any increase in the inlet loading
 is likely to cause an increase  in the outlet PM loading because a
 similar percentage of PM would  be removed.  Therefore, the
 venturi scrubber system design  must be based on the maximum gas
 flow rate, PM loading,  and pressure drop considering the  particle
                                48

-------
size distribution.   Similar effects would be likely on gaseous
pollutants, i.e.,  increases in HC1, metals,  and CDD/CDF inlet
concentrations would cause increases in outlet concentrations of
these pollutants.
313  Parked-B^ figrubbers
  '  '
2
3
4
5
     This section provides a general description of a packed-bed
scrubber system including equipment components, gas cleaning
process, and key operating parameters followed by a discussion of
the  factors affecting the performance of the packed-bed.
     3.1.3.1  "-™~i T^crintion.  A packed -bed scrubber system
typically comprises  the  following  components:
     1.  Scrubber vessel;
         Packing media;
         Mist  eliminator;
          ID fan;
          Liquid recirculation system;  and
      6   System controllers.
 These components are described in the following paragraphs along
 with the packed-bed gas cleaning process.
      Packed-bed scrubbers are designed to remove acid gases
  (primarily HC1 and sulfur dioxide  [SO21 ) by absorption of the
 gases into the scrubbing liquid and subsequent neutralization.
 An  alkaline solution typically is used to maintain a constant
  scrubber liquid PH  to prevent corrosion of scrubber components.
  The scrubbing  liquid typically used is caustic solution  (NaOH)
  although sodium carbonate  (Na2C03) and calcium hydroxide
  (Ca[OH]2)  (slaked  lime) also can  be used.  When the acid gases
  are absorbed  into  the  scrubbing solution,  they react with
  alkaline  compounds to  produce neutral  chemical salts.    The rate
  of absorption of  the acid gases  is dependent on  the  solubility of
  the acid gases in the  scrubber liquid; HC1 is absorbed rapidly
  while SO, is absorbed more slowly.  The neutralization reactions
  are essentially stc-ichiometric,  i.e.,  the stoichiometric ratio of
  alkaline compounds added to the system to that required for
  complete neutralization of the acid absorbed into the scrubber
  solution is essentially 1:1 in packed-bed scrubbers.
                                  49

-------
      Packed-bed  scrubber vessels used to control acid gases from
 MWI's typically  are  countercurrent,  vertical columns where the
 gas  flows  upward through the scrubber vessel and the scrubbing
 liquid flows  downward.   Figure 13 illustrates a countercurrent,
 packed-bed absorber  with packing media,  liquid sprays,  and a mist
 eliminator.
      Scrubbing liquid is evenly distributed by the  liquid  sprays
 over the packing media and trickles  down through the bed wetting
 the  surface of the packing and,  thereby,  exposing the gas  to a
 large,  wetted surface area.   The depth of the bed depends  on many
 factors including the type of packing used,  pollutant concentra-
 tion and solubility,  the desired removal  efficiency of
 pollutants, type of  scrubber liquid  used,  liquid and gas flow
 rates,  and system temperature.   However,  a packing  depth of  1.5m
 (5 ft)  using  Intalox® saddles has been reported.35   Other  vendors
 specify packing  heights that range from 0.9m (3  ft)  to  3m  (10 ft)
 without specifying packing type.44'45  The exhaust  gas  is  forced
 to make many  changes  in direction as it winds through the
 openings of the  packing media,  thereby causing the  gas  to  mix
 with the liquid.  The large  surface  area  of  liquid/gas  interface,
 the  packing depth, and  the random packing  provide the necessary
 contact area  between  liquid  and  gas,  sufficient  residence  time,
 and  good mixing,  respectively,  for effective  absorption to occur.
      After passing through the packing, the gases flow through a
 mist  eliminator  that  removes  entrained droplets  of  liquid  that
 may  contain alkaline  salts,  absorbed acid  gases, and  PM.   In a
 combination venturi/packed-bed scrubber system,  the venturi may
 remove  90  percent of  the  PM  from the  MWI exhaust gas, and the
 remaining  particles can become entrained in droplets  from the
 packing media in the  form of  alkaline  salts and other ultrafine
 PM.   Therefore,   the appropriate mist  eliminator must be selected
 for  the packed-bed scrubber to minimize PM emissions  from the
 overall combination system.  Mesh-, chevron-, and diffusion-type
mist  eliminators are available.  At least one vendor has stated
 that a venturi/packed-bed  scrubber system equipped with mesh- and
 chevron-type mist eliminators can achieve outlet PM
                                50

-------
                                              Mist eliminator
                                              Liquid sprays
                                              Packing
Figure
13.  Schematic of countercurrent  packed-bed  scrubber



                         51
                                                                 43

-------
 concentrations of 0.03 grains/dry standard cubic foot (gr/dscf)
 while the diffusion-type mist eliminator can achieve
 concentrations less than 0.01 gr/dscf.35'46
      The ID fan,  liquid recirculation system, and system
 controllers were  discussed above in Section 3.1.2.1 on venturi
 scrubbers.  With  the exception of the variable throat and
 emergency water quench controllers,  these discussions also apply
 to packed-bed'scrubbers.
      3-1-3.2  Factors Affecting Performance.   The performance of
 the packed-bed scrubber relative to PM,  acid gases,  metals,  and
 CDD/CDF emissions is affected by the packed-bed scrubber design
 and operating variables and also by key MWI operating variables.
 Key packed-bed scrubber design and operating variables are
 scrubbing medium,  L/G ratio,  packing height (liquid-to-gas
 contact time),  suspended solids content,  pH,  and type of mist
 eliminator.   Key  MWI process  parameters  that  can affect  packed-
 bed scrubber performance are  variations  in exhaust  gas stream
 temperature,  flow rate,  and pollutant concentrations that result
 from the heterogeneous nature of the MWI  process.
      The scrubbing medium or  liquid  used  in packed-bed scrubbers
 to  remove acid  gases from MWI's typically is  a  caustic solution.
 This alkaline solution absorbs acid  gases from  the MWI exhaust
 gas  stream and  protects the scrubber components  from corrosion by
 neutralizing the  acid gases.   The high solubility of HC1  gas  in
 this solution enhances the absorption of  the  HC1 into the
 solution and thereby enhances the effectiveness  of the packed-bed
 scrubber in  removing HC1.   When insoluble alkalis such as  slaked
 lime are used as  a scrubbing  medium  in packed-bed scrubbers,  the
 suspended solids  in the  slurry deposit on the packing  and  cause
 plugging.22   Therefore,  caustic  solution  absorbs and neutralizes
 acid gases,  prevents  corrosion,  and  elifenates plugging problems
 associated with insoluble  alkalis.
      The L/G ratio of  a packed-bed scrubber is a design variable
 that  is  specified  along with  the  packing  height and packing media
 to achieve a  specific  control  efficiency  for a gaseous pollutant.
A L/G ratio between 2,674 and  3,342  L/Macm  (20 and 25 gal/Macf)
                                52

-------
                                                         •jc 4.4 45
is typical for packed-bed scrubber applications on MWI's."'**'
The countercurrent packed-bed scrubber (typically used in MWI
applications)  does not operate effectively-if .there are large
variations in the liquid or gas flow rates.47  Generally, removal
efficiency is increased by an increase in the liquid flow rate to
the scrubber.48  However, if either the .liquid flow rate or gas
                                                             47
flow rate is too high, a condition called flooding may occur.
Flooding is a condition where the liquid is "held" in the
pockets, or void spaces, between the packing media and does not
drain down through the packing.47  The gas velocity at which
flooding occurs is called the flooding velocity.49  Flooding can
be reduced by reducing the liquid flow rate.48  In packed-bed
applications to MWI's, a pump provides a constant liquid flow
rate.  The packed-bed diameter is sized so that the maximum gas
velocity to be experienced by the packed-bed lies between 50 and
75 percent of the flooding velocity.49
     The performance of a packed-bed scrubber in removing acid
gases from the gas stream depends on the effectiveness of the
absorption process that takes place on the packing media.  The
packing media provides a large surface area for liquid-to-gas
contact and promotes good mixing of the gas and liquid.  The
height of the packing determines the liquid-to-gas contact time;
the greater the height, the greater the residence time.  Packing
heights of 1.5M  (5 ft) using  7.6 cm  (3 inch)  Intalox®  saddles
have been reported.35  Manufacturers guarantee HC1 removal
efficiencies  from 99  to  99.9  percent and S02  removal efficiencies
from 90 to 99 percent when a  packed-bed scrubber  is used in
conjunction with  a venturi scrubber.35'44'45  The effectiveness
of  the  packed-bed scrubber in removing vaporous metals and
condensible organics  such as  CDD/CDF  from  the gas stream is
further enhanced  as the gas  stream is  intimately  contacted by the
liquid  film on  the packing because vaporous metals and
condensible organics  may condense  and be removed  as  PM.
     The  suspended and  dissolved solids  content and  pH of the
scrubber  liquid are monitored to maintain  scrubber performance.
Solids  accumulation at  the  entry to the  packed bed and within the
                                53

-------
                         22
bed is a problem that interferes with the absorption process by
restricting liquid flow through the bed.  To prevent solids
accumulation, an adequate scrubber liquid blowdown rate must be
maintained.  Most manufacturers have designed their systems with
a set blowdown rate that removes suspended and dissolved solids
and that does not require shutdowns for solids removal.35'44'45
The pH typically is maintained at or slightly below 7.0 by
monitoring pH and adding caustic solution as appropriate.'
Maintaining this pH level is important to prevent corrosion
damage to scrubber components.
     Selection of the appropriate mist eliminator is important to
achieve low PM emissions.  As described in Section 3.1.3.1, the
mist eliminator removes entrained droplets of liquid that may
contain alkaline salts, absorbed acid gases, and PM.  A
diffusion-type mist eliminator can achieve a three-fold reduction
in PM emissions over the traditional chevron- and mesh-type mist
eliminators.35'46
     The variations inherent in the operation of MWI's translate
into variations in gas stream characteristics including
temperature, flow rate, and pollutant concentrations.  A packed-
bed scrubber controlling emissions from an MWI typically is
located after a PM control device such as a venturi scrubber or a
fabric filter.  In both of these cases, the MWI exhaust gas
temperature  is reduced before it reaches the packed-bed scrubber.
In the case  of the venturi scrubber, the MWI exhaust gases are
cooled and saturated in the venturi and enter the packed-bed
scrubber at  a temperature between 60° and 85°C  (140° and 185°F).
In the case  of the fabric filter, the MWI exhaust gases may be
cooled initially by passing through a heat exchanger or WHB
before passing through the fabric filter and a quench prior to
the packed-bed scrubber.  In either case, the temperature
variations that occur in the exhaust gases leaving the MWI are
damped and the gas stream temperature entering the packed-bed
scrubber will be within a relatively narrow range.  Gas stream
temperatures in excess of 104°C  (220°F) can damage or melt the
54

-------
plastic packing media that typically is used in these
scrubbers.50
     Variations in MWI exhaust gas flow rate are compensated for
by the proper design and sizing of the packed-bed scrubber.  As
already described, the liquid flow rate to the packed-bed
scrubber is set at a constant rate and the packed-bed scrubber
diameter is sized such that the maximum gas velocity experienced
by the system is 50 to 75 percent of the flooding velocity.
Therefore, as the gas flow rate varies in the MWI system,
sufficient liquid is available for effective absorption without
encountering flooding.
     Variations in pollutant  concentration  in the MWI exhaust
depend primarily  on waste composition.  Because HC1  in the
exhaust gas is highly soluble in  the scrubbing solution,
variations  in HC1 concentration are not expected  to  affect the
performance of  the packed-bed in  removing HC1  if  the packed-bed
has  been properly designed  for the maximum  HC1  concentration to
be encountered.   The  effectiveness  of  the packed-bed scrubber in
 removing  CDD/CDF  and metals will  depend in  large  part on the
 effectiveness  of  the system in condensing and subsequently
 collecting the resulting fine PM.  The collection efficiency of a
 packed-bed scrubber in removing condensed metals and CDD/CDF is
 unknown.
 3.1.4  other Wet---Scrubbing Systems
      Although venturi and packed-bed scrubbers are the systems
 that historically have been  installed most frequently for MWI
 emissions control, other novel wet-scrubbing systems are
 beginning to penetrate the MWI market.  Four types of systems
 that have been applied to various waste combustion processes are
 the Rotary Atomizing™ scrubber,  the Ionizing Wet Scrubber™
  (IWS~), the Calvert collision scrubber, and the Hydro-Sonic®
 scrubber.  These systems have been used more frequently at
 municipal  or hazardous waste incineration  facilities, but they
 are being  actively marketed to MWI facilities.   Because  the
 number of  MWI  installations of these  systems  is  quite  limited,
 very  little  information has been collected to date  on  the
                                  55

-------
performance' of the systems or on the factors that affect
performance.   Consequently, the discussion below is limited to a
brief description of each of the systems.
     3.1.4.1  Rotary Atomizing™ Scrubbers.51"53  Rotary
Atomizing™ scrubbers are designed to remove PM primarily through
impaction and acid gases primarily through absorption.  Figure 14
illustrates a schematic of a rotary atomizing scrubber showing
the major components listed above.  The Rotary Atomizing™
scrubber comprises a quench vessel (on systems without a WHB), a
prespray tower, a rotary atomizer, mist eliminators, and a liquid
recirculation system.  Because the liquid recirculation system is
an integral part of the quench vessel, prespray tower, and rotary
atomizer section of this scrubber and because it contains
separate recirculation loops for each of these components, it is
described first.
     The liquid recirculation system on the rotary atomizer
scrubber is unique in that there are three scrubber liquid
recirculation loops.  The liquid is staged in a countercurrent
direction with respect to the gas flow.  There are three separate
scrubber liquid feed tanks that serve the rotary atomizer, the
prespray tower, and the quench vessel.  The fresh, clean scrubber
liquid is fed to the rotary atomizer tank.  Overflow from this
tank flows to the prespray tower tank and a purge from the
prespray tank then flows to the quench vessel.  Slowdown from the
quench vessel then flows to a sewer discharge.
     On systems without a WHB, the combustion gases from the MWI
are directed to the top of the quench vessel.  The quench vessel
serves two functions:   (1) to cool the MWI exhaust gases and
 (2) to absorb a portion of the acid gases.  The quench vessel
evaporatively cools the MWI exhaust gases  (982° to 1204°C [1800°
to 2200°F]} to saturation  (68° to 77°C  [155° to 170°F] through
the introduction of the purge from the prespray section through
spray nozzles located at the top of the vessel.  The alkaline
 (NaOH) quench water converts a large portion of the HC1 and S02
gases into sodium chloride and sodium sulfate salts.  Once the
total dissolved solids content has reached 15 percent in the
                                56

-------
57

-------
quench tank" feeding the quench vessel, a blowdown stream is
activated that removes the dissolved and suspended solids from
the system.
     The saturated gases from the quench vessel (or from the WHB
if so equipped) then enter the prespray tower where mid-size PM is
removed by impaction on droplets produced by nozzles located at
four different levels in the tower.  Acid gases are further
neutralized in the prespray tower.  This tower operates at an L/G
ratio of 8,020 L/Macm (60 gal/Macf) to completely saturate the
gases prior to entering the rotary atomizer section.  The tower is
served by a prespray tank that uses the overflow from the rotary
atomizer section as feed liquid.  Liquid that exits the base of •
the tower unevaporated flows back into the tank and is recycled to
the tower.
     The saturated gases from the prespray tower subsequently pass
to the rotary atomizer section of the system.  Figure 15
illustrates the rotary atomizer.  The rotary atomizer creates a
constantly renewed high velocity water curtain equivalent to a
2.6 x 105 mmHg (5,000 pounds per square inch  [psi]) spray nozzle.
This water curtain is generated by means of a high speed rotating
disc, that atomizes the water droplets and propels them toward the
wall of the rotary atomizer housing.  The droplets rebound off the
wall creating a dense water curtain that acts as a filter pad.
This dense spray of droplets removes PM through impaction, while
gases are collected by absorption.  The pressure drop across the
rotary atomizing scrubber is about 7.5 mmHg  (4 in. H20).
Scrubbing solution also is recirculated through the rotary
atomizer tank based on level control reward while fresh caustic
reagent is added based on pH control demand.
     The gas then passes from the rotary atomizer section through
two Chevron-type mist eliminators that are used to collect and to
separate the water droplets from the gas stream.  The liquid
collected by the mist eliminators is returned to the rotary-
atomizing section tank.  The scrubber manufacturer claims that
this 2-stage mist elimination system is rated at 99+ percent
overall in removing droplets that are 2 microns or larger in size.
                                58

-------
        HIGH-SPEED
        INDUSTRIAL
        GEARBOX
SCRUBBED OUTLET
GAS FLOW
 SCRUBBING LIQUID
 FEED PIPE
                                  HIGHLY AUTOMIZED
                                  LIQUID DROPLETS
PARTICULATE-LADEN
INLET GAS FLOW
           Figure  15.   Schematic  of  rotary atomizer  module.52
                                            59

-------
     3.1.4.2  Ionizing Wet Scrubber™.54'55  The IWS™ uses a
combination of electrostatic forces and the common wet-scrubbing
phenomena to provide both PM and acid-gas control.  The typical
system consists of a quench chamber to reduce gas temperature and
one or more of the modular "units" shown in Figure 16 in series.
Because each unit (or stage) acts to remove a nearly constant
fraction of the acid gas and PM, the control efficiency is
increased by staging these units.
     Each unit in the system has two major sections.  The gas
stream passes through the entry plenum into the ionizing section
of the unit.  Here particles are charged as they pass between
wires powered by a high-voltage DC power source and small
grounded electrode "mini plates."  The gas stream thus passes
into the charged particle scrubber, which is a horizontal,
cross-flow, packed-bed scrubber with a Tellerette® packing
material.  Acid gases are removed from the gas stream by
absorption and particles are collected in the scrubber section by
a combination of impaction and image force attraction, an
electrostatic property that causes the charged particles to
migrate to the neutral packing surfaces.  Particles with
diameters of 3 to 5 pirn and larger are collected on packing
surfaces by impaction.  Smaller particles are collected by image
force attraction.  Particles are removed from these surfaces by a
constant cross-flow of scrubbing liquor.
     3.1.4.3  Calvert Collision Scrubber.56  The Calvert
collision scrubber is designed to achieve high collection of both
PM and acid gases.  The system comprises four major components in
series: the quencher, the condenser/absorber, a collision
scrubber, and an entrainment separator.  A modified system
without the condenser/absorber is available if only moderate
control efficiencies are required.  The major components of the
system are described in the paragraphs below.
     The hot gas from the MWI enters the dual flow quench
chamber.  The unit is called dual flow because two separate
liquid streams are introduced to the gas stream at different
                                60

-------
     pJ.;reCna'S.iESe:M-.-Snor: m-enseiy cnargec High voltage DC
                           DOAS-  -o,; energy consumption
     .c,-5. £.e~s-,ts-Coni.ni!Ci:s f'usnmc, orevents soncs Dune UD
     facto'. AsseT.Diec Moames-Fast  inexpensive fteio msiallation
     AI- or::c'es cr.a-ge: as tney enter  scruDDer section
     P'as:,: Sne!  anc internals-Corrosion Resistant Minimum maintenance
mlet-To 50.000 ACFM
Collection Surtaces-ConnnuouS'v fiusnec .'
scruDDer section
All scruDDer surfaces loacKing -Aater O'ME  s"5
eiectncaiiy neutral Act as impingement sj"a:es
or coiieci particles tnrougn image force ai:rac:.o"
Tellerene packing is excellent impingement ta-;e:
ana nigniy efficient collector  Also creates
water orootets tor additional imDinge^ei: anc
collection surfaces
                                                                                            Exhaust of IWS ca- oe
                                                                                            connecteo  sirectiy :c sac-
                                                                                            fan or to anotne- IWS -o-
                                                                                            increases collection
                                                                                            efficiency
                           Seaieo High voltage DC Power Supply
                                                                                 No movm; Parts (oniy  pump)
                                                                       !       :   Recycled Liauio—For gas aosorotion soncs
                                                                                 flushing
    Figure   16.     Schematic   of  an   ionizing  wet  scrubber™.
                                                                                                         54
                                                       61

-------
points.  First, fresh makeup water is added to the inlet  to
provide a wetted approach.   Just beyond the inlet,  recycled
scrubber water is added  to  the gas stream.   These  two  liquid
streams reduce the gas stream temperature to saturation levels,
about  66° to 85°C  (150°  to  185°F),  depending on inlet  temperature
and moisture content.
     The saturated gas stream moves from the quencher  into a
horizontal, cross-flow or a vertical,  counterflow  packed-bed
absorber.  Generally, a  caustic or lime scrubbing  solution is
used in the absorber.  The  HC1 is  removed from the  gas stream and
additional gas cooling reduces the temperature to  49°  to 54°C
 (120°  to 130°F).  This temperature reduction results in
condensation of moisture from the  saturated stream.  Small
particles in the gas stream act as condensation nuclei.  The
moisture condensation on the surfaces  of the particles
effectively increases their aerodynamic diameter,  thereby
enhancing their collection  downstream in the entrainment
separator.
     The gas stream moves from the condenser/absorber  into the
collision scrubber.  There  the gas stream is split  into two equal
streams, turned, and impacted head on.   At  the point of collision
of the two streams, particles are  collected on larger droplets
via the impaction and interception phenomena described in
Section 3.1.1.  The particle-laden droplets in the  gas stream
then enter a multistage  entrainment separator.  There,  the
droplets are removed from the gas  stream by impaction.
Additional removal of HC1 also is  achieved  in  the collision
scrubber and entrainment separator.
     3.1.4.4  Hvdro-Sonic®  Scrubber.57*58   Hydro-Sonic® scrubbers
comprise a family of wet scrubbers  that  have been used to control
PM and HC1 emissions on  several waste  incineration processes.
The design most likely to be used on MWI  systems is the fan
drive, tandem nozzle shown  in Figure 17.  This design consists of
two subsonic nozzles equipped with  water  sprays in series.  As
the water spray is injected  into the high-velocity gas  stream,
very fine high-velocity water droplets are  generated.   Collisions
                                62

-------
                                                      in
                                                        to
                                                        n
                                                        QJ
                                                        U
                                                        Oi

                                                        JJ
                                                        0)


                                                        U-l
                                                         o

                                                         ^1
                                                        r-l
                                                        -i-l


                                                         U-l

                                                        ©
                                                         U
                                                         •H

                                                          O
                                                         CO

                                                          o
                                                          V4


                                                         I



                                                          I
                                                           1
                                                           en
                                                          -H
                                                           fa
63

-------
 between the" droplets and particles  take place  in the highly
 turbulent flow regimes in the mixing  tubes, as does absorption of
 HC1 into the droplets.  Cyclonic mist eliminators are then used
 to remove the pollutant-laden droplets from the gas stream.
 3.2  FABRIC FILTERS
      Historically, fabric filters  (or baghouses)  have been used
 to control PM and nonvaporous metals emissions from combustion
 sources and'other industrial processes.  Fabric filters are often
 combined with other air pollution control techniques to achieve
 control of acid gas and organic pollutant emissions in addition
 to PM control.   The two most common combinations  are a fabric
 filter downstream from a sorbent-injection system (typically dry
 sorbent injection is used,  but facilities can use spray dryers)
 and a fabric filter and packed-bed scrubber in series.   Other
 components of these combined systems are discussed later in this
 report and will not be addressed in this section.   The discussion
 below focuses on the fabric filter component of the systems with
 reference  to later discussions.as appropriate.
      Fabric filters are typically classified with respect to
 their location  relative to  the fan and to  their type of  cleaning
 mechanism.   All systems installed to date  at MWI  facilities have
 been  negative-pressure systems (i.e.,  systems  installed  upstream
 from  an induced-draft  fan).   Also,  all of  the  systems use pulse-
 jet cleaning, although some  vendors  indicate that they would use
 a  reverse-air system if a client  so  preferred.59'61  Because
 pulse-jet  systems  dominate the industry, the discussion  below
 focuses  on these systems.
     The remaining discussion is divided into three subsections.
 The first presents-a brief overview  of fabric filter PM
 collection principles.  The  second subsection presents a general
 description of  the pulse-jet fabric  filter.  It describes
 components of the  system and identifies key design and operating
variables.  The final subsection identifies primary control
system and combustor design  and operating variables that affect
performance and discusses how these factors affect the
performance relative to PM,  HC1,  metals, and CDD/CDF emissions.
                               64

-------
3.2.1  Fabric F.i 1 ter Co31 *»fit ion Principles
     Fabric filtration systems have been used for years to
control PM emissions from a wide variety of industrial processes,
to collect product materials from manufacturing processes, and to
maintain dust loadings at acceptable levels in clean room
environments.  A substantial amount of research has been
conducted on 'the performance of-these PM and dust control
systems.  The results of this research have led to an
understanding of the overall performance of fabric filter systems
in removing particles from gas streams and of the specific
collection mechanisms that apply.  For MWI exhaust streams
controlled by fabric filters  in  combination with dry  sorbent
injection, information collected to date suggests that  the fabric
filter  also plays a role  in controlling other pollutants.
However,  the mechanisms of  control  for other pollutants are not
fully understood.  This subsection  describes those mechanisms
that are  generally accepted as  the  key mechanisms  for fabric
filtration of PM.   It  also  briefly  discusses the mechanisms  that
are  postulated  to affect  the  control  of  other  pollutants.
     A fabric filter  is  a collection  of  bags  constructed of  a
fabric material (nylon,  wool, or other)  hung inside a housing.
The  combustion  gases  are drawn into the housing, pass through the
bags,  and exhaust from the housing through a stack to the
 atmosphere.   When the exhaust from the incinerator is drawn
 through the fabric,  particles are retained on the fabric
 material, while the cleaned gas passes through the material.  The
 collected particles are then removed from the filter during  the
 pulse-jet cleaning cycle.  The  dusty material cleaned  from the
 bags falls to a collection hopper and is removed from  the hopper
 for transfer to a storage area  or disposal site.
      With a new filter, .the  open areas in the fabric are of
 sufficient size that particles  easily penetrate the  bag.  Over a
 short  time, a  cake builds  on the bag surface; this cake acts as
 the primary particle  collection medium.  Particles are collected
 on  a  filter and  cake  by  a  combination of several mechanisms. The
 most  important are impaction,  direct interception, and diffusion.
                                  65

-------
     As the" exhaust gas flows through the cake and filter, it
makes numerous turns around obstructions encountered in the cake
or filter media and creates a flow pattern characterized by
curved streamlines.  In collection by impaction, the particles in
the gas stream have too much inertia to follow the gas stream-
lines around the fiber or through pores in the cake.  They leave
the gas stream and deposit directly onto the fiber or the surface
of the cake.  In the case of direct interception, the particles
have less inertia and barely follow the gas streamlines around
the fiber.  If the distance between the center of the particle
and the outside of the fiber or pore is less than the particle
radius, the particle surface will contact the surface of the
fiber and be "intercepted."  Impaction and direct interception
mechanisms account for 99 percent of the collection of particles
with an aerodynamic diameter greater than 1 pm. in fabric filter
systems.62'63
     For submicron particles, such as those generated by
homogeneous condensation of volatile metals, diffusion acts as a
key collection mechanism.  Small particles that are affected by
collisions on a molecular level behave individually through
random motions.  The particles do not necessarily follow the gas
streamlines, but instead move randomly throughout the fluid, a
phenomenon called Brownian motion.  At some point in their random
flow through the fluid, a fraction of these small particles come
in contact with a surface of the filter medium and are
collected.62'63
3.2.2  Pulse-Jet Fabric Filter Description62'63
     A schematic of a pulse-jet fabric filter is shown in
Figure 18.  The primary components of the systems are the bags
and auxiliary equipment, the housing that contains the bags, the
inlet or dirty-side plenum that receives combustion gases and
distributes them to the bags, the clean-air plenum that receives
the cleaned combustion gas from the bags before it is discharged
to the atmosphere, and the hopper and discharge system.
     The bag compartment is separated from the clean-air plenum
by a flat metal plate called a tube sheet.  In larger systems,
                                66

-------
   Clean Air Plenum.
      Blow Pipe.
    Bag Retainer.
                              '/   °/  °/
Dirty Air Inlet and Diffuser'. •
                                                    To Clean Air Outlet
                                                    and Exhauster
                                                Housing
-Tubular Filter Bags
                                                 •Dirty Air Plenum
                                         Rotary Vatve Air Lock
   Figure 18.   Schematic of  a pulse-jet  baghouse
                                                             63
                                 67

-------
the bag compartment is frequently separated into multiple,
smaller compartments.  These compartments can be taken off-line
individually for bag cleaning or maintenance.  However, most
smaller units, such as those used at MWI facilities, have a
single compartment with bags arranged in a rectangular array.
Within this compartment, bags are suspended from the tube sheet
and supported internally by rings or cages.  Bags are held firmly
in place at the tube sheet by clasps and have an enclosed bottom
(usually a metal cap).  Dust-laden gas is filtered through the
bag, depositing dust on the outside surface of the bag.  All
pulse-jet systems filter the gas from the outside to the inside
of the bag.62'63
     The dust cake is removed from the bag by a blast of
compressed air injected into the top of the bag tube.  The blast
of high-pressure air stops the normal flow of air through the
filter.  The air blast develops into a standing, or shock, wave
that causes the bag to flex as the shock wave travels down the
bag tube.  As the bag flexes, the cake fractures, and deposited
particles are discharged from the bag.  The shock wave travels
down and back up the tube in approximately 0.5 seconds.62
     The blast of compressed air must be strong enough for the
shock wave to travel the length of the bag and shatter or crack
the dust cake.  Pulse-jet units use air supplies from a common
header that feeds into a nozzle located above each bag.  In most
baghouse designs, a venturi sealed at the top of each bag is used
to create a large enough pulse to travel down and up the bag.
Alternatively, some baghouses operate with only the compressed
air manifold above each bag.  For either type system, the pres-
sures involved are commonly between 414 and 689 kPa  (60
and 100 psig).
     As the shock wave moves down the bag,  dust is released from
the bag surface into the open area on the dirty side of the bag.
A portion of the dust is immediately entrained in the incoming
gas stream and deposited on the surface of the bag just cleaned,
thereby maintaining a cake on the bag surface.  The remainder
                                68

-------
falls into the hopper at the base of the baghouse and is
subsequently discharged.62'63
     Most pulse-jet filters at MWI facilities use bag tubes that
are 10 to 15 cm (4 to 6 in.) in diameter and 1.8 to 5.8 m (6 to
19 ft) in length.59'62'64  The bags are usually arranged in rows
and are cleaned one row at a time in sequence.  The preferred
method of cleaning initiation is based on pressure differential
with a timer override to prevent extended operation without
cleaning.  Cleaning may alternately occur in a timed sequence.
3.2.3  Factors Affecting Performance
     The performance of the fabric filter relative to PM, acid
gas, metals, and CDD/CDF emissions is affected by the fabric
filter design and operating parameters and also by key MWI
operating parameters.  Key pulse-jet filter design and operating
parameters are the air-to-cloth  ratio  (or filtration velocity),
bag material, operating temperature, pressure drop across  the
filter, and cleaning frequency.  Key process parameters  that  can
affect fabric filter performance, particularly  long-term
performance, are variations  in temperature,  flow rate,  and
pollutant concentrations that result from the heterogeneous
nature of the MWI process.   Process startup/shutdown procedures
can  also  affect long-term  performance.
      The  air-to-cloth  (A/C)  ratio  is actually a measure of the
superficial gas velocity through the filter medium.   It is a
ratio of  the  flow rate of  gas  through  the fabric filter (at
actual  conditions)  to  the  area of  the  bags  and  is usually
measured in units of m3/sec/m2  (acfm/ft2).   Data obtained from
air pollution control  device vendors  indicate that pulse-jet
 fabric filters for MWI's typically are designed with an air-to-
                                            *2      ^
 cloth ratio in the range of 0.008 to 0.02 nvVsec/m
 (1.6 to 4.3 acfm/ft2)  of bag area.59'61'64   Fabric filter systems
 are designed to operate with as high an A/C ratio as feasible in
 order to limit size and cost.  However, in general, the lower the
 A/C ratio,  the lower the PM emission rate.   Short-term exceedance
 of the design ratio is not likely to have a substantial effect on
 emissions.  However, long-term exceedances will increase
                                 69

-------
particle-phase emissions from bleed-through and have the
potential to cause bag failure through abrasive action.  Under a
failure condition (e.g., tears at the collar or multiple pinhole
leaks) , emissions of PM and materials such as metals and CDD/CDF
attached to particles are likely to increase substantially.  The
fabric filter should be designed with an A/C ratio based on the
maximum expected gas flow rate.  Therefore, at lower gas flow
rates, the operating A/C ratio will be lower and the PM
collection performance would be expected to be the same or higher
than that at the design A/C ratio.
     Bags vary with respect to type of construction and material
of construction.  The three major types of construction are woven
bags, felted bags, and membranes:  A wide variety of materials
have been used to construct bags.  Usually both bag type and
material are selected based on a vendor's experience with systems
installed on comparable processes.  Key factors that are
considered in making the selection are the cleaning method,
resistance of the material to abrasion and chemical attack,
expected operating temperature, and costs.  For MWI's, both
fluctuating temperatures and resistance to acids are of concern.
The choice of bag material and construction can affect the
performance of the system in removing PM.  Felted bags typically
are recommended in pulse-jet fabric filters controlling emissions
from MWI's.  The relatively thick felt fabric provides maximum
particle impingement and collects PM more efficiently than woven
fabric at comparable gas velocities.65  Gore-Tex® membrane bags
can also be used to achieve low outlet PM loadings  (i.e.,
22.5 mg/dscm  [0.01 gr/dscf]).61  If felted bags are used,
synthetic materials that are temperature- and acid-resistant are
recommended.  Vendor recommendations include acrylic, P-84
 (Raster® Scrim), and Rylon®.60'61
     The operating temperature of the fabric filter is critical
to the long-term performance of the system.  The system must be
operated within a relatively narrow temperature range to prevent
bag failure due to chemical attack or temperature-driven
degradation.  Because the exhaust gas from an MWI can contain
                                70

-------
HC1, the unit should be operated at temperatures sufficiently
high to ensure that no surface temperatures drop below the acid
dewpoint.  Otherwise, condensation of HC1 will result in
corrosion of the housing or bags.  The boiling point of HC1
(aqueous hydrochloric acid) is 110°C (230°F); gas temperatures
should be maintained at or above 120°C (250«F) to ensure that no
surfaces are cooled below the dewpoint.  Each type of bag
material has a specified maximum recommended operating
temperature.  If the system is operated at temperatures somewhat
higher than the maximum recommended operating temperature, bags
will degrade over time; operation at substantively higher
temperatures can cause bags to fail completely.  Gas temperatures
should be within a temperature range bounded by the HC1 dewpoint
and the maximum recommended operating temperature for the
specific material to prevent bag failures and the attendant
increase in particle-phase emissions.
     In addition to  its importance in maintaining fabric  filter
integrity, temperature can also  directly influence  the  control  of
acid gases, volatile metals, and CDD/CDF.   All  of the mechanisms
described earlier related  to these pollutants are temperature
dependent.  For all  of the postulated mechanisms, improved
control  is achieved  with lower temperatures in  the  baghouse.
Consequently, the temperature should be maintained  at  as  low a
level  as possible, given the acid  gas  constraints described
above,  to enhance control  of these pollutants.   Vendors suggest
an optimal  temperature range of  120°  to  150°C (250° to 300°F) to
achieve the  best  control  in  combined  dry-injection/fabric
 filtration  systems.59"61'64
      Pressure drop  across  fabric filters  generally  is  maintained
within a narrow range by controlling  the  cleaning cycle of the
 system.  Pressure drops  below the minimum indicate  that either
 leaks have  developed or excessive cleaning is removing the base
 cake from the bags.   Either condition decrease performance
 immediately.   Pressure drops greater than the maximum indicate
 that either bags are "blinding"  or excessive cake is building on
 the bags because of insufficient cleaning.  Over time, high
                                 71

-------
pressure drops lead to bag erosion and degradation and,
subsequently,  to decreased performance.  The pressure drop on MWI
systems generally is maintained in the range of 1.5 to 5 in.
w.c.59-61,64
     As suggested by the above discussion, the fabric filter
cleaning cycle can have a significant impact on fabric filter
performance.  Typically, one of two procedures is used to
determine a cleaning cycle.  For the simpler of the two, vendor
engineers monitor facility operations carefully during equipment
shakedown.  The pressure drop pattern is observed, and a timed
cleaning sequence is established to keep the pressure drop well
within acceptable limits.  This timed cycle is then incorporated
into subsequent operating procedures.  As an alternative, the
system can be equipped with a pressure drop sensor, and a control
loop can be used to initiate the cleaning cycle when the pressure
drop reaches a specified upper level.  Either system can work
well, but a malfunction in either system or changes in the
process create performance problems.  Too frequent cleaning can
cause the cake to deplete, resulting in an increase in emissions
of PM.  On the other hand, cleaning that is too infrequent will
result in excessive cake buildup and an increased pressure drop
through the system.  This latter scenario creates two potential
problems.  If the pressure drop becomes too great, the system
will lose draft to the primary chamber of the MWI, and fugitive
emissions can be generated.  The excessive cake buildup also
results in substantial increases in local gas velocities at some .
points in the bags.  These increased velocities can generate an
immediate increase in emissions via particle seepage or bleed-
through and can result in increased abrasion that deteriorates
bag integrity.
     The variations inherent in the operation of MWI's translate
into variations in gas stream characteristics including
temperature, flow rate, and pollutant concentrations.  Of these
three gas stream characteristics, temperature is most critical.
All fabric filter control systems installed on MWI's incorporate
some type of gas-cooling system upstream of the filter.  For
                                72

-------
those systems that do not have WHB's,  the ability of these
cooling systems to respond to rapid temperature changes in the
combustor exhaust is a key to the long-term performance of the
fabric filter.  If these systems do not cool the gas stream
properly, periodic temperature exceedances lead to bag
degradation and reduced collection efficiency.
     Variations in the MWI exhaust gas flow rate are compensated
for by the proper design of the fabric filter system.  The fabric
filter system should be designed with an A/C ratio based on the
maximum  expected gas flow rate.  Therefore, at lower gas flow
rates, the A/C ratio will be  lower and the PM collection
performance  would be expected to be the  same or higher than that
at the design A/C ratio.
      The PM  loading  and  particle size  distribution  must be
considered during the  design of  a  fabric filter  and also  during
operation; however,  within  certain limitations  (±10 to 20 percent
of design values),  changes  in these parameters  do not seriously
affect fabric filter efficiency.65 Nevertheless,  an increase in
mass loading may require more frequent cleaning of the bags as a
 result of faster filter cake buildup.65  A major advantage of a
 properly designed fabric filter system is its ability to perfortn
 well over the normal variation in MWI exhaust gas
 characteristics.
 3.3  DRY SCRUBBERS
      This section describes  the application of dry scrubbers to
 MWI's.  Dry  scrubbing techniques that could be applied to MWI's  .
 can be  grouped into two major categories:   (1) dry sorbent
 injection,  and  (2)  spray dryer absorber systems.  During  the past
 several years, many dry sorbent injection systems  have been
 installed on MWI's.   Many  spray dryer absorber  systems have  been
 installed en municipal  waste combustors (MWC's)  but only one has
 been installed  on  an MWI to date.
       Dry scrubbers use  an  alkaline sorbent  to  react with and to
 neutralize  the  acid gases  in the  MWI exhaust gas stream.
 Additionally, activated carbon can be injected to control mercury
  and CDD/CDF through adsorption.  The reaction product is a dry
                                 73

-------
 solid  that  Is  collected along with fly ash and any unreacted
 sorbent  in  a PM control device such as a fabric filter,   in  MWI
 applications,  dry  scrubbers are invariably'followed by a  fabric
 filter.   The following paragraphs describe dry scrubbers  and
 their  operation.   Section 3.3.1 describes dry scrubbing
 principles.  Section 3.3.2 describes dry sorbent injection
 systems.  Section  3.3.3 describes spray dryer absorber systems.
 3-3.1  Drv  Scrubbing Principles
     Dry sorbent injection systems use adsorption to control acid
 gases  while spray  dryer absorbers use absorption and
 adsorption.66   The addition of activated carbon also controls
 mercury  and CDD/CDF  emissions through adsorption.   The principles
 of adsorption  and  absorption as applied to dry scrubbers  are
 briefly  described  below.
     3.3.1.1   Adsorption.   During adsorption,  one or more gaseous
 components  are removed from an effluent gas stream by adhering to
 the surface of a solid.67   The gas molecules  being removed are
 referred  to as the adsorbate,  while the solid adsorbing medium is
 called the  adsorbent.67  In dry injection systems,  the adsorbents
 or sorbents typically include a finely divided alkaline material
 such as calcium hydroxide  (hydrated lime), magnesium oxide,  or
 sodium bicarbonate.   Activated carbon can be  added for the
 control of mercury and CDD/CDF.
     Adsorption occurs  in  a series  of three steps  as  illustrated
 in Figure 19.68  In  the first step,  the contaminant  diffuses from
 the bulk area  of the  gas stream to  the  external  surface of the
 adsorbent particle.68   In  the second step, the  contaminant
molecule migrates  from the relatively small area of  the external
 surface to the pores  within each  adsorbent particle.68  In the
 third  step, the contaminant molecule  adheres to the surface of
 the pore.68  in the dry sorbent injection system, the acid gas
molecules adsorb onto  the  surface of  and react with the sorbent
particles (usually hydrated  lime) to  neutralize the acid gases by
 forming neutral chemical salts.
     Effective adsorption  of acid gases, mercury, and CDD/CDF in
dry scrubber applications  requires the  following:
                                74

-------
- Q
i-a
 O
 e
 .£

 c
 o
&

I -8
"8

I1
//^>
 S8
 i-S

 U
                                             00
                                             VX)
                                              •

                                              §
                                             -H
                                              JJ
                                              a
                  •o
                  (C
                  m

                  ro
                  •H
                  c
                  rt
                  JS
                  u
                  
                                              •H
                                              Ct4
                        75

-------
     1.  A large interfacial  surface area between the  gas and  the
 sorbent;
     2.  Good mixing of  the gas  and sorbent  phases;
     3.  Sufficient residence, or contact, time between the
 phases for adsorption and subsequent neutralization to occur;  and
     4.  Sufficient sorbent  (type and amount) to  adsorb and to
 neutralize the acid gases and to adsorb mercury and CDD/CDF.
     Adsorption of vapor-phase organic and metallic compounds
 such as CDD/CDF and mercury also occurs in dry scrubber
 applications, especially when activated carbon is  injected.  The
 presence of a large interfacial  contact area formed by the
 activated carbon and the fine particles of the alkaline sorbent
 provides numerous surface sites  for condensation  and/or
 adsorption of these vapor-state  compounds  to occur.  Turbulence
 and adequate residence time also promote adsorption of these
 compounds.  The temperature of the gas stream must be  low enough
 to allow condensation of these compounds to  occur but  not so low
 as to allow condensation of HC1  gas that causes corrosive
 conditions.  Additionally, the adsorption process may  be enhanced
 by the filter cake on the filter bags  in a fabric filter.
     3-3-1.2  Absorption.  Section 3.1.1.2 described how gaseous
 pollutants are absorbed  in wet scrubber applications.   Many of
 the principles described there also apply to absorption in spray
 dryer applications.  Effective absorption of acid gases in spray
 dryer applications requires the  following:
     1.  A large interfacial  contact area between the gas and
 slurry droplets;
     2.  Good mixing of  the gas  and slurry droplets (i.e.,
 turbulence), and
     3.  Sufficient residence, or  contact, time between the
 phases for absorption and subsequent neutralization to occur.
     In the case of spray dryer absorber systems,  the  absorbing
medium is the liquid slurry droplets as opposed to the scrubbing
 liquid used in packed-bed scrubbers.  The acid gases are absorbed
 into the droplets where  they are neutralized by the lime in the
droplets.   Subsequently,   the liquid in the droplets is evaporated
                               76

-------
so that the" reaction products along with unreacted sorbents are
dried for collection in a PM control device.   After the sorbent
has been dried,  adsorption of unreacted acid gases and vaporous
metals and organics onto sorbent particles may occur as the dry
sorbent and exhaust gases pass to the PM collection device.  This
adsorption process may be enhanced by the filter cake on the
filter bags in a fabric filter.
3.3.2  Dry Sorbfnt Injection
     This section provides a general description of the dry
sorbent injection system including equipment components and gas
cleaning process followed by a discussion of the factors
affecting the performance of the system including key operating
and design parameters and gas stream characteristics.
     3.3.2.1  General Description.  A dry sorbent injection
system typically is comprised of the following components:
     1.  Sorbent storage/feed hopper;
     2.  Sorbent transport system;
     3.  Venturi/injector;
     4.  Reaction/expansion  chamber  (optional); and
     5.  Recycle system (optional).
     Figure 20  illustrates a schematic  of a  dry injection  system
showing the major  components listed  above with the exception of
the venturi and the  recycle  system.  The  following paragraphs
describe these  components along with the  dry injection system gas
cleaning process.
      The dry  injection scrubber uses injection of a  dry,  finely
divided alkaline  sorbent such as calcium hydroxide (hydrated
lime),  magnesium  oxide, or  sodium bicarbonate for the adsorption
of acid gases and powdered  activated carbon for  the  adsorption of
mercury and CDD/CDF.   The alkaline sorbent typically is stored in
a storage  bin that also acts as a feed hopper for the system.
This bin may have a cone shaped bottom to direct the sorbent to a
 rotary airlock, slide gate,  or volumetric screw feeder that
meters sorbent into the pneumatic line that feeds the MWI exhaust
 gas duct.   The bin may also be equipped with a shaker mechanism
 and/or a screw auger that help prevent clumping of the sorbent.
                                 77

-------
K III
Z O
LU <
m oe
o: o
O ^
CO CO
                                                                                                      VD


                                                                                                        0)
                                                                                                        4J
                                                                                                        m
                                                                                                        >1
                                                                                                        CQ

                                                                                                        C
                                                                                                        o
                                                                                                        •H
                                                                                                        4J
                                                                                                        O
                                                                                                        0)
                                                                                                        •i—>
                                                                                                        C
                                                                                                        •H
                                                                                                        C
                                                                                                        (!)
                                                                                                        X)
                                                                                                        )-l
                                                                                                        O
                                                                                                        CO
                                                                                                        i-l
                                                                                                        •O
                                                                                                        u
                                                                                                        -H
                                                                                                        JJ
                                                                                                        01
                                                                                                        .c
                                                                                                        u
                                                                                                        w
o
CN

 0)
 S-(

 t7>
•H
                                            78

-------
The feeding' system typically is controlled by a microprocessor
system that controls the operation of the rotary airlock,  slide
gate  or screw feeder in some systems, the sorbent injection rate
is controlled by opening and closing the airlock.  By controlling
the length of time and the number of times in a given time period
that the airlock or slide gate remains open, the sorbent
injection rate can be closely controlled.  Other systems simply
feed sorbent on a continuous basis using a controller that
controls the speed of the screw feeder.
     Typically, the sorbent is transported to the MWI exhaust gas
duct pneumatically.  As  the sorbent  is metered  into the pneumatic
line,  a blower propels the  sorbent through the  line to the  duct.
Additionally, the system's  draft  pulls the sorbent through  the
line.  The  action of the blower and  the  system draft provides the
initial  fluidization of  the sorbent  for  transport to the  duct.
Other, simpler  systems use  a gravitational  feed approach  instead
of a pneumatic  system where the  sorbent  is  fed through a  vertical
pipe from a feed hopper located  above the duct.
      in  many systems,  the sorbent is injected into the duct
 upstream of or at a venturi section of the duct.  The purpose of
 the venturi is to introduce turbulence to the gas stream and
 sorbent, thereby providing good mixing and enhancing adsorption
 of the acid gases.70'71  Some systems have an adjustable venturi
 throat that compensates for the variations in the MWI exhaust gas
  flow rate.
           70
As in a wet venturi system, adjusting the venturi
 maintains a constant gas velocity.across the throat, thereby
 preventing the  reagent  from  falling out in the duct.
       in some  dry  injection systems, a  reaction chamber  (expansion
 chamber)  is  included that allows  increased residence time  for
 adsorption and  neutralization to  occur.70  One vendor utilizes a
 specially designed heat exchanger that both  cools the gases and
 provides increased residence time.60   From the  reaction chamber,
  the gas stream, containing  the entrained sorbent particles and
  fly ash, is  ducted to a fabric filter.
       Some manufacturers include a sorbent recycle system that
  recycles a portion of the collected sorbent/fly ash back to the
                                 79

-------
sorbent injection system.71   The  primary purpose  of  the recycle
stream is to increase reagent utilization and  thereby reduce
overall sorbent costs.71  Recycle systems are  installed on a
case-by-case basis based on  the economics of the  installation.
      The buildup of a cake  on the filter bags in the fabric
filter provides additional intimate contact between  the gaseous
pollutants and sorbent for adsorption to occur.   Emission test
data suggest that the fabric filter cake plays a  role in
collecting HC1, volatile metals,  and organic pollutants such as
CDD/CDF in combination dry sorbent injection/fabric  filtration
systems.71'72  Additionally,  specific tests suggest  that HC1
control in such systems is related to cake build-up  and depletion
and that volatile mercury and CDD/CDF control is  improved when
activated carbon is used as  a sorbent in the cake.72  Testing to
investigate the effectiveness of  injecting activated carbon  (in
addition to hydrated lime),  in controlling both mercury and
CDD/CDF emissions across a dry sorbent injection/fabric filter
system, was conducted in September 1991.  The results of this
test along with the results  of the entire test program are
presented in Section 4.0 of  this  report.
     3.3.2.2  Factors Affecting Performance.  The performance of
the dry sorbent injection system  relative to acid gas, metals,
and CDD/CDF emissions is affected by the dry injection system
design and operating variables in promoting adsorption of these
compounds and also by key MWI operating variables.  Dry injection
systems control primarily acid gases by adsorption; metals and
CDD/CDF are also controlled by this mechanism but to a lesser
extent.  While PM is not controlled by dry injection, a PM
control device located downstream of the dry injection system
collects flyash from the MWI  along with the dry,  solid reaction
product and any unreacted sorbent.  The following paragraphs
describe the key dry injection system design and operating
variables that promote adsorption  including sorbent fluidization
and particle size, retention/reaction time,  stoichiometric ratio,
and gas stream moisture content and temperature.   Additionally,
key process parameters that can affect dry injection system
                               80

-------
performance" are discussed including variations in gas stream
temperature,  flow rate,  and pollutant concentrations that result
from the heterogeneous nature of the MWI process.
     Maintaining proper sorbent particle size and fluidization is
important in providing surface adsorption sites and in fully
utilizing the injected sorbent.  The sorbent (typically hydrated
lime) has particle sizes that allow 90 percent by weight of the
material to pass through No. 325 mesh screens.71  This particle
                                                             *7 T
size range is approximately the consistency of talcum powder.
Sorbent particles in this size range provide sufficient total
alkaline sorbent surface area for effective adsorption to
occur.71  Larger particles have a smaller total surface area and
would require larger fan capacity .to fluidize the sorbent.  The
dry injection fluidization system  (transport blower) capacity
must be sufficient to fluidize the sorbent adequately.  The
sorbent must be fluidized uniformly across the duct so that all
                                                               7P
portions of the gas stream are contacted by sorbent particles.
The use of a venturi constriction helps to provide the turbulent
conditions necessary for intimate contact between the gaseous
compounds and the surface of the particles.71  Insufficient
fluidization allows sorbent to fall out of the gas stream unused
and may ultimately lead to plugged lime feed pipes and/or MWI
exhaust gas ducts.
     Adequate retention/reaction time is necessary for effective
adsorption of acid gases, metals, and CDD/CDF.   Retention time in
dry  injection systems is prolonged through the use of specially
designed reaction/retention chambers to provide  intimate mixing
between the gas stream and sorbent particles.  One vendor
supplies a specially designed  heat exchanger both to extend  the
retention time and to cool the gases. °  In some cases, an
extended duct  is used to prolong retention time  while other
systems compensate for a lack  of retention time  by operating the
system at a higher stoichiometric  ratio or extending the
dirty-air side of the fabric filter.
     The stoichiometric ratio  is defined as the  molar ratio  of
calcium  in the lime  fed to  the dry sorbent  injection system to
                                81

-------
the theoretical amount of calcium required to completely react
with the HC1 and S02 in the MWI exhaust gas.  Because of mass
transfer limitations, incomplete mixing, differing rates of
adsorption and reaction (SO2 reacts more slowly than HC1),  and
the presence of other acid gases that react with the .calcium in
the lime (e.g, hydrogen fluoride, sulfur trioxide),
stoichiometric ratios in excess of 1:1 are required.  Vendors
design dry injection systems with extended retention times to
operate at stoichiometric ratios of sorbent (hydrated lime) to
HC1 that range from 1.3:1 to 2:I.59'61'64  At such stoichiometric
ratios, these vendors guarantee HC1 reductions that range from 90
to 99 percent.59'61'64  Higher removal efficiencies can be
achieved by increasing the stoichiometric ratio.59'61'64
     The gas stream moisture content and temperature are
important variables in achieving effective acid gas
neutralization.  Some vendors recommend using evaporative coolers
both to humidify and to cool the gas stream.60'61'64  This
process increases the reagent activity.61  These evaporative
coolers may be used in conjunction with or without the use of a
WHB but typically are used when no WHB is present.60'61'64  Other
vendors use combinations of evaporative coolers and heat
exchangers or heat exchangers alone.59'64  The optimum
temperature for the neutralization of acid gases ranges from 121°
to 177°C (250° to 350°F) according to vendors.59"61'64  The lower
the temperature, the higher the HC1 and S02 removal.59
Additionally, one vendor reported that lower temperatures improve
the removal efficiency of lead, mercury, and CDD.59
     The variations inherent in the operation of MWI's translate
into variations in gas stream characteristics including
temperature,  flow rate, and pollutant concentrations.  Because
dry sorbent injection systems are located after gas cooling
devices  (WHB's, heat exchangers, or evaporative coolers) and
because these devices are designed to achieve a setpoint outlet
temperature,  the effect of MWI exhaust gas temperature variation
is minimized.  The optimum temperature for gaseous pollutant
removal in dry sorbent injection systems is approximately 12l°C
                                82

-------
(250°F).   The boiling point of HC1 (aqueous hydrochloric acid)  is
110°C (230°F);  gas temperatures maintained at 121°C (250°F)
maximize gaseous pollutant removal and ensure that no surfaces
are cooled below the dewpoint.  However, condensation of HC1 on
cool surfaces (resulting in corrosion of metal components) can
occur even when measured gas temperatures are greater than 121°C
(250°F),  particularly on systems without insulation.  Insulation
around ductwork and the dirty- and clean-air sides of the fabric
filter will minimize the condensation of HC1 on internal metal
components by preventing cold spots.  Corrosion caused by
condensation of HC1 can occur on the clean side (clean air
plenum,  outlet ductwork, and outlet stack) even though most of
the HC1 has been removed from the gas stream.  Some combination
of insulation and/or corrosion resistant materials can be used to
reduce corrosion problems.
     Variations in flow rate may cause problems with fluidization
of sorbent in the duct.  Flow rate variations may be compensated
for by a variable speed fan that is tied  to the incinerator
draft.  Some systems have adjustable venturi constrictions that
maintain both a constant velocity across  the venturi and
                                           72
turbulent mixing as the flow rate changes.
     Variations in pollutant concentrations in the MWI  exhaust
gas are dependent primarily on waste composition.  The
stoichiometric ratio typically is specified based on the  maximum
expected HC1 concentration through  the  system.  Therefore, as  the
HC1 concentration varies up to the maximum amount,  the
stoichiometric ratio of lime  to  the theoretical amount  required
will  remain  high enough for effective HC1 removal.   It  is not
clear how variations in the concentrations of metals and  organics
observed  in  MWI exhaust gas streams would impact  the
effectiveness of  the dry  sorbent injection system in removing
these pollutants.
3.3.3 Spray Dryer Absorbers
      This  section provides a  general  description  of a  spray dryer
absorber  system including equipment components  and gas cleaning
process  followed by a  discussion of the factors affecting
                                 83

-------
performance" of the system including key operating and design
parameters and gas stream characteristics.
     3.3.3.1  General Description.  A spray dryer absorber system
typically is comprised of the following components:
     1.  Sorbent preparation system;
          - - storage
          - - slaker
          - - mixing tank
          - - feed tank
     2.  Atomizers; and
     3 .  Spray dryer absorber vessel.73
     The primary differences between a dry sorbent injection
system and a spray dryer absorber system are:   (1) the physical
form of the alkaline sorbent and  (2) the design of the system
used for contacting the sorbent with the acid gas stream.74  In a
dry sorbent injection system, the sorbent is dry while in a spray
dryer system, the sorbent is fed as an alkaline slurry.  In a dry
sorbent injection system, the dry sorbent is injected into a duct
and may be followed by a reaction chamber while in a spray dryer
system, the wet slurry is atomized in a spray dryer absorber
vessel.
     Figure 21 illustrates a spray dryer absorber.  In this
system, the alkaline reagent, usually pebble lime, is prepared as
a slurry containing 5 to 20 percent by weight solids by slaking
the lime.73  Slaking is the addition of water to convert calcium
oxide to calcium hydroxide.76  Proper slaking conditions are
important to ensure that the resulting calcium hydroxide slurry
has the proper particle size distribution and that no coating of
the particles has occurred due to the precipitation of
contaminants in the slaking water.76  From the slaker, the slurry
passes to the mixing tank where the slurry is thoroughly mixed
before passing to the slurry feed tank.
     The prepared slurry is atomized into the gas stream in a
large absorber vessel having a residence time of 6 to 20 seconds.
Atomization of the slurry is achieved through the use of:
(1) rotary atomizers or  (2) air atomizing nozzles.  Generally,
                                84

-------
                                                                    in
                                                                    t-
DC

H
                      UJ
                       Q.
at

D
V)
UJ
o:
o
                                CQ
0)
u
CO
>.
CO

S-i
QJ
             O
             CO
             ja
             m
              0)

              &
             TJ


              fC


              tn

              U-J
              O

              u
              -H
              JJ
              fO
              s
              Q)
              x:
              o
              CO
                                                                        ex

                                                                        0)
                                                                        i-l

                                                                        Cn
                                                                        •H
           85

-------
 only  one  rotary atomizer is  included in a spray dryer absorber.
 However,  a  few applications  have as many as  three  rotary
 atomizers.74
      In rotary atomizers,  a  thin film of slurry is  fed to  the  top
 of  the atomizer disk as  it rotates  at speeds of 10,000 to
 17,000 revolutions per minute.   These atomizers generate very
 small slurry droplets having diameters in the range of
 100 microns.  The spray  pattern is  inherently broad due to the
 geometry  of the disk.77
      High pressure air is  used  to provide the physical energy
 required  for droplet formation  in nozzle-type atomizers.   The
 typical air pressures are  483 to 621 kPa (70 to 90 psig).  Slurry
 droplets  in the range of 70  to  2'00  microns are generated.  This
 type  of atomizer generally can  operate over  wider variations of
 the gas flow rate than can be used  in a rotary atomizer.
 However,  the nozzle atomizer does not have the slurry feed
 turndown  capability of the rotary atomizer.   For these reasons,
 different approaches must  be taken  when operating at  varying
 system loads.78
      The  shape of the scrubber  vessel must be designed to  take
 into  account the differences in the slurry spray pattern and the
 time  required for droplet  evaporation for the two types  of slurry
 atomizers.  The length-to-diameter  ratio  of  a spray dryer
 absorber vessel using rotary atomizers  is  much smaller than that
 for absorber vessels using air  atomizing  nozzles.77
      3.3.3.2  Factors Affecting Performance.   The performance of
 the spray dryer system relative to  acid gas, metals, and CDD/CDF
 emissions is affected by the spray  dryer  system  design and
 operating variables in promoting  absorption of these compounds
 and also by key MWI operating variables.   Spray dryer  systems
 control primarily acid gases by absorption; metals and CDD/CDF
 also  are  controlled by this  mechanism but  to a lesser extent.
While PM is not controlled by a spray dryer, a PM control device
 collects fly ash from the MWI along with the dry, solid reaction
product and any unreacted  sorbent.  The following paragraphs
describe the key spray dryer system design and operating
                                86

-------
variables that promote absorption including stoichiometric ratio,
slurry droplet size, the approach-to-saturation and outlet gas
temperature, and retention/reaction time.  Additionally, key
process parameters that can affect dry sorbent injection system
performance are discussed including variations in temperature,
flow rate, and pollutant concentrations that result from the
heterogeneous nature of the MWI process.
     The  stoichiometric ratio  is  defined as the molar ratio of
calcium  in  the lime slurry fed to the spray dryer to the
theoretical amount  of  calcium  required  to  completely react with
the HC1  and S02  in  the flue gas at  the  inlet to the spray
dryer  79  However,  because of  mass  transfer limitations,
incomplete  mixing,  differing  rates  of reaction (S02 reacts more
slowly than HC1), and the presence  of other acid  gases  that  react
with  calcium  (e.g.,  hydrogen  fluoride,  sulfur  trioxide),  more
 than  the theoretical  amount  of lime is  generally  fed  to the spray
 dryer 79  Stoichiometric ratios in the  1.2 to 1.3 range can
 reportedly achieve  90 to 95  percent removal of HC1 and HF and 60
                              80
 to 85 percent removal of S02-
      The slurry droplet size would be expected to affect the
 performance of the spray dryer in removing acid gases, metals,
 and CDD/CDF.   Smaller droplet size increases  the surface area for
 adsorption and reaction between  lime and  acid gases and increases
 the rate of water evaporation.79  Additionally,  it is  important
 that  all of the slurry droplets  evaporate to  dryness prior to
 approaching the absorber vessel  side walls and prior to exiting
 the absorber with  the gas stream.77  Accumulations of  material on
 the side walls  or  at  the-bottom  of  the absorber  would  necessitate
 an outage  since these deposits would further  impede drying.
  Proper drying  of the slurry  requires generation  of small  slurry
  droplets and adequate mixing with the  hot flue gases.
       Drying  that is too rapid can reduce pollutant  collection
  efficiency since the primary removal mechanism is absorption into
  the droplets.77  There must be sufficient contact time for     ^
  absorption to occur and for the slurry to evaporate to dryness.
  For this reason, spray dryer absorbers controlling MWC's are
                                  87

-------
operated with an approach-to-saturation range of 32° to 82 °C
          77
                                                              (90'
                                                              77
 to 180°F).     The approach-to-saturation is the difference
 between the wet bulb and dry bulb temperatures measured by wet
 and dry bulb monitors at the outlet of the spray dryer vessel
 An increase in the approach-to-saturation is sensed by the
 automatic control system that quickly reduces the slurry feed
 rate.81
      The spray dryer outlet temperature is controlled by the
 amount of water in the slurry.  More effective acid gas removal
 occurs at lower temperatures,  but the temperature must be kept
 high enough to ensure that the slurry and reaction products are
 adequately  dried prior to collection in the fabric filter.   In
 addition, a minimum spray dryer absorber vessel outlet
 temperature of approximately 240°F is required to control
 agglomeration of PM.79
      In MWC applications,  the  amount of lime fed is generally
 controlled  by one of two means.   In one approach,  the lime-slurry
 feed rate is controlled by an  acid gas analyzer/controller
 (generally  based on S02 or HC1 emissions)  at the stack.   As the
 outlet  acid gas  concentration  increases or decreases,  the lime-
 slurry  feed rate is accordingly raised or lowered,  respectively,
 to maintain a specified outlet,  acid gas  concentration.   The
 second  approach  uses  a constant  lime-slurry feed rate  that  is
 sufficient  to react with peak  expected acid gas  concentrations.79
      For effective  absorption  of  gaseous  pollutants  to occur,
 adequate retention/reaction time  is  required.  The  retention time
 in spray dryers  is  determined  by  the  size of the spray dryer
 absorber vessel.  Retention times  in  spray  dryers controlling
 emissions from MWC's  range  from 10 to  15 seconds.80
     The variations inherent in the operation of MWI's translate
 into variations  in  gas  stream  characteristics including
 temperature,  flow rate, and pollutant concentrations.  For those
 systems equipped with a WHB or heat exchanger, the variation in
 the MWI exhaust gas temperature is damped.  The spray dryer
outlet flue gas temperature is controlled at levels well above
its saturation value by controlling the amount of water in the
                                88

-------
slurry.  This control precludes any sorbent from contacting
downstream surfaces as a wet powder leading to solids buildup.
It also assures operation well above the dew points of any acid
gases.  Because of the presence of calcium chloride (a very
hygroscopic, difficult-to-dry solid),  temperatures are typically
controlled at 110° to 160°C (230° to 320°F) by limiting the
amount of water injected.
                         82
     Flow rate variations may be compensated for by a variable
speed ID fan that is tied to the incinerator draft.
     Variations in pollutant concentrations in the MWI exhaust
gas are dependent primarily on waste composition.  A second
control loop can be provided based on the pollutant emission
levels in stack gases to regulate the addition of reagent to the
system.  In many cases, the sorbent rate is fixed at a
conservatively high rate to ensure low stack emissions, but waste
disposal plus sorbent operating costs are increased.82
4.0  PERFORMANCE OF EMISSION CONTROL MEASURES
     One component of the background information development for
MWI's was a comprehensive emission test program.  The program
generated emission test data on a wide variety of pollutants
including all of those for which emission limits are to be
established (PM, CO, HC1, SO2,  NOX, CDD/CDF, Cd, Pb, and Hg).
These test data provide information on the effects of combustion
system operating parameters on emissions and on the performance
capabilities of add-on APCS's.   This section summarizes the
results of the test program.  Section 4.1 briefly describes the
test program and facilities tested and presents a summary of the
test data.  Section 4.2 presents a discussion of the test results
including a comparison of the emission results obtained when
incinerating different waste types, a discussion of the effects
of combustion control parameters on emissions, and a discussion
of the performance of four APCS's--a combination dry injection
fabric filter (DI/FF), a combination venturi scrubber/packed-bed
scrubber  (VS/PB), a fabric filter "in series with a packed-bed
scrubber  (FF/PB), and a spray dryer/fabric filter combination
 (SD/FF)--with respect to PM, HC1, CDD/CDF, and metals emissions.
                                89

-------
Note that tests on the DI/FF and SD/FF systems were performed
with and without activated carbon injection.  Section 4.3
presents demonstrated emission control levels  (i.e., emission
limits) associated with each of the emission control techniques
described above.
4.1  TEST PROGRAM SUMMARY
4.1.1  Facility and Test- Condition Descriptions
     The emissions test program generally  consisted of  triplicate
test runs under multiple sets of operating conditions at  seven
MWI facilities, which are denoted herein by the  letters A,  B, J,
K, M,  S, and W.  At the four  facilities  equipped with add-on
APCS's (A, B,  J, and M) , simultaneous measurements were made at  .
the APCS inlet and outlet for PM, HC1,  S02/ CDD/CDF, and  metals.
Operating parameters that were, varied to achieve different
operating conditions included waste feed type,  waste  feed
charging mode, and secondary  chamber temperatures.  Table l
provides a  summary of  the  test conditions  and the test  program
conducted at  each facility,  including the  test condition numbers,
the  ,aste type charged,  the target  waste charge rates,  and the
target secondary/tertiary chamber temperatures.
      Summary  tables  for the incinerator and APCS operating data
 collected from the  test program are included in Appendix A.
 Table A-l presents  pertinent incinerator operating data for each
 of the 25 test conditions (77 test runs) that were conducted as a
 part of this program,  and Tables A-2 through A-5 present
 pertinent APCS operating data for those facilities with add-on
 APCS's.  The  paragraphs below briefly describe the systems at
 each  of the seven facilities tested and the general structure of
 the test program at each site; more detailed descriptions of the
                                                        Q *1 _ Q A
 facilities are included in the individual  test reports.
       4.1.1.1  Facility A.  The incineration unit at Facility A is
 a Cleaver-Brooks Model 780-A/31 intermittent-duty MWI  with  three
 combustion chambers.  The primary chamber operates at  starved-air
 levels, while the secondary and tertiary  chambers operate with
 excess air.   The nameplate charging  capacity of  the unit is
 295 kg/hr  (650 Ib/hr) at an assumed  waste heating value  of
                                 90

-------
   TABLE  1     MATRIX OF TEST  CONDITIONS  FOR TESTS CONDUCTED
                                AT EACH FACILITY
                                                                         Target tertiary/
                                                                       secondary chamber
                                                                           temp.. °F
Target waste charge
    rate, Ib/hr
^500 - general medical waste generated at a 500-bed hospital
 G100 - general medical waste generated at a 100-bed hospital
 RB - red bag waste
 General - general medical waste
 Pathological - pathological waste
**At facility J, the target charge rate was 750 Ib/batch.
cAt facility J, test condition 2 refers to testing during the bumdown phase of the batch where test
 condition 1 was conducted during the burn phase.
dSame hourly charge rate as test condition 2 but larger charges were charged less frequently.
                                            91

-------
19,800 kJ/kg  (8,500 Btu/lb) .   The facility is  equipped with  a WHB
for energy recovery and a  DI/FF APCS.
     Two separate test sequences were  conducted at  Facility  A.
During the first sequence,  emissions were  measured  simultaneously
at the DI/FF  inlet and outlet  under seven  sets of test conditions
based on waste feed type and tertiary  chamber  operating
temperature.  The target operating characteristics  for these test
conditions were:  Condition 1--general medical waste from a
500 bed hospital  (G500) at a tertiary  chamber  temperature of
980°C (1800°F); Condition  2--G500 waste at 870°C  (1600°F);
Condition 3--G500 waste at 1090°C (2000°F); Condition 4--general
medical waste from a 100 bed hospital  (G100) at 980°C  (1800°F);
Condition 5--red bag waste from a 500 bed  hospital  (RB) at 870°C
(1600°F); Condition 6--RB  waste at 980°C  (1800°F);  and
Condition 7--RB waste at 1090°C (2000°F).
     The second test sequence  was designed to  evaluate the effect
of activated  carbon injection  on DI/FF performance.  During  the
second test sequence, emissions were measured  simultaneously at
the DI/FF inlet and outlet.  The incinerator operating target was
identical to  Condition 1 above and the DI/FF system was operated
at three carbon levels — none  (Condition 1A) , 0.45 kg/hr  (1 Ib/hr)
(Condition 8), and 1.14 kg/hr  (2.5 Ib/hr)   (Condition 9).  The
actual incinerator operating parameters for all test runs within
both test sequences are presented in Table A-l while the actual
DI/FF operating parameters  are presented in Table A-2.
     4.1.1.2  Facility B.   The incinerator at  Facility B is  a
Basic Environmental Engineering,  Inc., Model 1500 continuous-duty
MWI with both secondary and tertiary chambers and a WHB.  The
primary chamber of this unit operates at near  stoichiometric
conditions, while the secondary and tertiary chambers operate at
excess-air conditions.  The nameplate charging capacity of this
MWI is approximately 680 kg/hr (1,500 Ib/hr) of medical waste
with an assumed heating value  of  19,800 kJ/kg  (8,500 Btu/lb).
This MWI is equipped with an Andersen 2000, Inc.,  venturi
scrubber followed by two packed-bed scrubbers in series.  The two
packed-bed scrubbers were installed in series because space
                                92

-------
limitations" prevented the installation of a single packed-bed
scrubber of suitable size.  Emissions were measured at the inlet
and outlet of the APCS at one test condition:  burning general
medical waste from the hospital at a nominal secondary chamber
temperature of 980°C  (1800°F) ..  The actual Facility B operating
parameter values for each test are included in Table A-l, and the
APCS operating parameter values for each test are presented in
Table A-3.
     4.1.1.3  Facility J.  The incinerator at Facility J is a
Simonds Model 2151B batch-duty MWI with two combustion chambers.
This unit has a nominal capacity of 340 kg/batch  (750 Ib/batch)
at an assumed heating value  of 22,400 kJ/kg  (9,600 Btu/lb).
After a batch is charged  to  the primary chamber,  the unit  is
sealed and completes  an operating cycle with three components--a
low-air phase  (burn phase),  a high-air phase  (burndown phase),
and a cooldown phase.  The unit is equipped with  an APCS
comprising an air-to-air  heat exchanger,  and a  FF/PB.  Emissions
measurements were  obtained during each of the  first  two
components of the  operating  cycle.   Simultaneous  measurements
were made  upstream from  the  heat  exchanger and downstream from
the FF/PB  under  one  set  of test conditions considered to
represent  normal  operating conditions for the  facility.   Separate
tests  were conducted during  two components of  the operating
 cycle--the low-air phase (Condition 1)  and the high-air phase
 (Condition 2).   Limited  CEM  data  were also collected during the
 cooldown phase of the process.  The values for the incinerator
 operating parameters for the six  test runs are included in
 Table A-l, and APCS operating data for facility J are presented
 in Table A-4.   The incinerator was charged with general medical
 wastes for these tests.   Note that although only one test
 condition was used at Facility J, the manual test data are
 separated into two conditions that represent the low-air and
 high-air components of the process cycle, respectively.
      4.1.1.4  Facility K.  The incineration unit at Facility K is
 an Environmental Control Products  (now Joy Energy Systems) Model
 480-E intermittent-duty, dual-chamber MWI.  The  unit nominally
                                 93

-------
operates at' starved-air conditions in the primary chamber and
excess-air conditions in the secondary chamber.  It has a
charging capacity of 145 kg/hr  (320 Ib/hr) at an assumed heating
value of 19,800 kJ/kg (8,500 Btu/lb).  The unit is not equipped
with an APCS.       '  **
     The emission test program at Facility K included a total of
eight test runs under three different test conditions defined by
charging characteristics and secondary chamber temperatures.
Condition 1 consisted of frequent charges at less than design
feed rates.  For Condition 1, waste was charged approximately
every 6 minutes at 9.1 kg  (20 lb') per charge for an average feed
rate of 91 kg/hr (200 Ib/hr).  Condition 2 represents the unit's
design feed rate with frequent charges.  For Condition 2, the
nominal charging conditions were to charge 14 kg  (30 lb) of waste
every 6 minutes for a total charging rate of 140 kg/hr
(300 Ib/hr).  Condition 3 represents the unit's design feed rate
with infrequent charges.  For Condition 3, plans called for
charging 140 kg/hr  (300 Ib/hr) waste in 23 kg  (50 lb) batches
every 10 minutes.  During all three conditions, general medical
waste was charged to the unit.  The target secondary chamber
temperatures were 1040°C  (1900°F) during Conditions 1 and 2 and
870°C (1600°F) during Condition 3.
     4.1.1.5  Facility M.  The incineration system at Facility M
was custom designed and sized by ThermAll, Inc.  The primary
components of the system are a rotary kiln, which is also called
the primary combustion chamber, followed by a fixed secondary
combustion chamber.  The system is equipped with a WHB.  The
system is designed with a heat input rate of 10.5 x 106 kJ/hr
(10 x 106 Btu/hr), which corresponds to a feed capacity of
454 kg/hr  (1,000 Ib/hr) at an assumed waste feed heating value of
21,000 kJ/kg  (10,000 Btu/lb).  Currently, the system is limited
to a feed rate 'of 354 kg/hr  (780 Ib/hr)  (permit limit) and at
that level, the rotary kiln is designed to operate at
stoichiometric conditions.  The APCS at Facility M comprises a
spray dryer followed by a pulse-jet fabric filter  (SD/FF).
                                94

-------
     The test program at Facility M comprised three runs at each
of two sets of test conditions.   The incinerator operating
conditions were identical for both conditions with the charge
rate maintained at about 354 kg/hr  (780 Ib/hr) and a secondary
chamber set point maintained at 982'C  (1800°F) .  The exhaust gas
temperature from the rotary kiln or primary chamber was typically
about 780°C  (1450°F).  During both  conditions, general medical
waste was  charged to the unit.  The primary difference between
the two conditions was that, during the second condition,
approximately  1.1 kg/hr  (2.5 Ib/hr) activated carbon was added  to
the lime  slurry to treat the exhaust gas.
      4116   T^rilitv  S.   The  incinerator at Facility  S  is  a
 Consumat Model  C-75P  intermit tent -duty,  dual -chamber MWI.   This
 unit  is designed to fire  either general  medical  waste  or
 100 percent  pathological  waste.  When general waste is fired,  the
 unit  is designed to operate at starved- air conditions  in the
 primary chamber and excess-air conditions in the secondary
 chamber.   The unit operates with excess-air conditions in both
 chambers  when pathological waste is being fired.  The nominal
 charge capacity is 80 kg/hr (175 Ib/hr)  for pathological waste
 and 110  kg/hr  (250 Ib/hr) for general medical waste.  These
 charge capacities are based on a heating value of 2,300 kJ/kg
 (1,000 BTU/lb) for pathological waste and a heating value  of
 ig'.SOO kJ/kg  (8,500 Btu/lb) for general medical waste.  This
 facility has ho APCS.
       Emissions were measured  at the  Facility  S  secondary chamber
 exhaust under  three  test conditions  defined by  waste  type,  waste
 charging  rate, and secondary  chamber temperature.  Plans for
 Condition 1 called for 45 kg/hr (100 Ib/hr)  pathological waste,
 which is  below design capacity,  to be charged in equal charges  at
  15 -minute intervals.   During  Condition 2,  pathological waste  was
  to be charged at 73  kg/hr (160 Ib/hr) in equal charges at
  15 -minute intervals.  For Condition 3,  general medical waste was
  to be charged at 73  kg/hr (160 Ib/hr) in equal charges at
  15 -minute  intervals.  Target secondary chamber temperatures were
  870°C (1600-P) for Conditions 2 and 3 and 1040 °C  <1900«P)  for
                                  95

-------
 Condition  l".  Note that for all runs actual secondary chamber
 temperatures  are  substantially below target values.   The
 relatively high primary chamber temperatures suggest that
 substantial waste combustion occurred there, and under such
 conditions, the secondary burner appeared to have insufficient
 capacity to maintain the target temperatures.
      4.1.1.7  Facility W.   The Facility W incinerator is an
 Environmental Control Products (now Joy Energy Systems)  Model-
 480  E/SR-12H  intermittent-duty,  dual-chamber MWI.  it is designed
 to operate at starved-air  levels in the primary chamber and
 excess-air levels in the secondary chamber.  The unit has a
 charge  rating of  175 kg/hr (385 Ib/hr)  at an assumed heating
 value of 19,800 kJ/kg (8,500 Btu/lb).   This  facility has no APCS.
      The test program at Facility W involved three sets of test
 conditions  defined by charging practices and secondary chamber
 temperature.  Measurements were made at the  secondary chamber
 exhaust while general  medical waste was charged to the unit.  The
 planned levels  for Condition 1,  which represents below design
 feed  rates with frequent charges and high secondary  chamber
 temperatures,  were a charge rate of 90  kg/hr (200  Ib/hr)  with
 equal charges at  6-min intervals and a  secondary chamber
 temperature of  1040°C  (1900°P).   For Condition  2,  which
 represents design  feed rate with frequent  charges  and  high
 secondary chamber  temperatures,  the charge rate  was  increased to
 136 kg/hr  (300 Ib/hr)  with the  same charging frequency and
 secondary chamber  temperature as  Condition 1.  For Condition  3,
which represents design feed rate with  infrequent  charges and low
 secondary chamber  temperatures,  the average charging rate was
maintained at 136  kg/hr (300  Ib/hr), but the frequency was
 changed to 10-min  intervals  and  the secondary chamber temperature
was decreased to 870°C  (1600°F).  For all conditions; general
medical waste was  charged  to  the unit.
4-1.2  Test Data Summary
     The emission  test program generated data on a wide variety
of pollutants  for 25 test  conditions  (77 separate test runs) at
the seven sites.  As described above, emissions measurements were
                                96

-------
made downstream from the secondary chamber and/or WHB ("post-
combustion" emissions) at all sites; measurements also were made
downstream of an APCS  ("post-APCS"  emissions)'at four sites.
Table 2 presents average post-combustion and post-APCS emission
concentrations by facility and test condition.  For those
pollutants expected to be affected by add-on APCS's, Table 2 also
presents removal efficiencies based on mass emission rates.
Tables B-l and B-2 in Appendix B present post-combustion emission
data in concentration units  corrected to a uniform excess air
level  (7 percent 02)  for each test  run conducted at the seven
sites.  Table B-l presents data on  PM, CO, HC1, S02, and NOX
emissions, while Table B-2 presents data on CDD/CDF, Cd, Pb, and
Hg  emissions.  These  data are presented with  a mixture  of English
and SI units  in accordance with most  common usage.
     Graphical summaries of  post-combustion emissions
concentrations are presented in Appendix C for  PM (Figure  C-l),
CO  (Figure C-2), HC1  (Figure C-3),  S02  (Figure  C-4),  NOX
 (Figure  C-5),  CDD/CDF (Figure C-6), Cd  (Figure  C-7),  Pb
 (Figure  C-8),  and  Hg  (Figure C-9).   Each of  the graphs displays a
 substantial amount  of information about the distribution of
 emission concentrations separately for each test condition.  The
 concentration levels  for each individual test run are indicated
 by the symbol '*'  on each graph.   In each graph and for each test
 condition, the range of emissions is shown along with a small bar
 in the range that denotes the average emission concentration.
      Post-APCS emission concentrations corrected to a uniform
 excess air level (7 percent 02)  for PM, HC1, CDD/CDF, Cd, Pb, and
 Hg for each test run are presented in Table  B-3.  Table B-3 also
 presents APCS removal efficiencies based on  mass emission rates
 for those runs for which data were sufficient to permit
 calculations.  Appendix D provides graphical summaries for  the
 post-APCS emissions  concentrations for those pollutants that  are
 expected  to be affected by  the add-on APCS's.  These graphs,
 which are analogous  to those presented  in Appendix C  for post-
 combustion pollutant concentrations  are  presented  for  PM
  (Figure D-l), HC1  (Figure D-2),  S02  (Figure  D-3),  CDD/CDF
                                 97

-------
g

CO
CQ
H
U
2

H
U
H
di

U
  H
  CO
  w
M

H
J


I
||
- 5 ~H
5 z
kl
if
e >
ec E
T- S g-
U C/3
If*.
£ >.
§ g
kl
Ii
c >
"Ii
ii
s S &

TcW
1 |
6» w
Ej «
ll
m

.«*
> 2>
H
i


S 8 5: « N 2 «

- „ „,,,.
— c* **x
^ o vi *r» •

5f S2SSJS
r* r* ^ ^ S t— S
C-* « f- o « r- ^*

>n c t- o e — e oc

ff1 O O O1 —
^r y f* w »c r* o
oc o> ^r o- oo
C4 O O C4 0v
1 i gg§
o js c j= e o o
•» «M O 00 « *
t S^g^g
6 JB o o e e> e

o o o o ee ee ee
— —
<
5?P

O O sC

O ^C f)

r*- *n »iri
SSS;
Wl «• 00
•o S S
O O 0
.f~ f*l ^
•o r-i er
— r* —
«n r< «n
•• •. r^ eo

— « m
•o r~ >o
o. O* t>
HI
o e o
HI
o e e
ill
o o o
VM *® ^^
<
fS
r-

P
O
*

^
g
lo

'§

e>
\c

0
91
i
e
§
o
"s
6
-
a
\n o
~
a«*8
"
» ,
««

ss

2g
Hg

t— m
W) ^"O
8.
^r~
A
ii
0 O
ii
o e>
S fi fi
000
--1.
-
(N O 00
r^ eo o»


V « I)
«S r* ^>
^ •"

U tl V

V «1 (I
§ 5 S3

<0
u-i O O
^~-

V U tl

«> O V
t> oo —
vi eo r4
2 — TT
e e o
"5 1 "5
000
— ts m
-
tf. ,0
S S

_

in oo
O r;

gg;
s -
Wl «O
51

u% O\
<• eo

fM 00
Sg;
ii
o e
i!
O 0
"5 "s
g g
o o
— N
^
°Si




is*

V t> V


a = i

ts
0> Cl O
r- « g

 1)
335
o o o
«-l
££o
-««
-
t^ oo o?



\fi it m
r- eo £>

« « «


o o o

o
r- ui o>


« a ti

V V tl
o O t--
O ^ "^
o o o
III
O O O
-«„
-
C
I
u
I
C
1
to
i
Js
S
'u
i£
e
_c
1
U
1

8
£
1
.2
V
aa
e
C
V

.1
U
j;

e
I
|
I
intrationa corrected to 7 percent Oj. Polk
measured.













^
|
C
J!
£
u
1
1
^
§
*o
•5
|
1
E
i
1
i
y not calculated because either data not me
town based on CEM data.






















m
ob
1 f
a £
"I -
| CO
V "c
" V
« ,£.

S at this facility.
lues higher than inlet values; negative effic
lore nutlet values helow detection.
i based nn at least one value below detectii
                                                                       z6o<
                                                                            -      '" -
                                           98

-------
99

-------
 (Figure  D-4) ,  Cd (Figure D-5),  Pb (Figure D-6),  and Hg (Figure
 D-7).
 4.2  DISCUSSION OF THE TEST RESULTS
     This  section summarizes general findings based on an
 examination of the MWI and APCS operating parameters and  emission
 test data.   Substantial insight about the performance of  the
 combustion systems and add-on APCS's tested can  be gained by  a
 careful  examination of the information presented in Table 2 and
 Appendices  B through D.
     Table  2 presents emission  concentrations for PM,  CO,  HC1,
 S02, NOX/  CDD/CDF,  and three metals.   With the exception  of CO
 and NOX, emissions of each of these pollutants can be affected
 substantially  by waste characteristics.   Emissions of four of the
 pollutants--PM,  CO,  CDD/CDF,  and to a lesser extent,  N0x--are
 likely to be affected by the combustion system characteristics.
 Section  4.2.1  discusses the effects  of waste types on post-
 combustion  emissions.   In particular,  emissions  measured  for  test
 conditions  burning red bag waste and  general medical  waste are
 compared and emissions measured for  test  conditions burning
 pathological waste and general  medical waste are compared.
 Section  4.2.2  discusses the effects of variations  in  combustion
 parameters  on  measured emissions.  First,  the relationships of
 CDD/CDF  emissions  to  CO and PM  emissions  are discussed.   Second,
 the relationships  between emissions of CDD/CDF,  CO, and PM and
 combustion  process  parameters are  discussed.  In Section  4.2.3,
 the variations  in  CO  and THC emissions during the
 burndown/cooldown  phases  of the batch  and the intermittent MWI
 operating cycles are  described.  In Section  4.2.4,  the operating
 cycle of batch MWI's  is  discussed.  In Section 4.2.5, the effect
 of combustion parameters  on metals partitioning  is described.
 Finally,  Section 4.2.6 presents  a discussion  of  the performance
 of the four APCS's:  DI/FF, SD/FF, VS/PB, and FF/PB.
 4.2.1  Effects of Waste Types on Emissions
     The test programs conducted at Facilities A and S were
designed to evaluate the  effect of different waste types on
emissions.   The test program at Facility A was designed to
                               100

-------
characterize emissions during combustion of two general medical
wastes,  (general waste from a 500-bed hospital [G500]  and general
waste from a 100-bed hospital [G100]),  and red bag infectious
wastes.   The test program at Facility S was designed to compare
emissions from the combustion of pathological waste and general
waste.
     4.2.1.1  Emissions from Combustion of General Medical Waste
Versus Red Baa Waste.  In general, emission concentrations at
Facility A showed substantial overlap between general waste runs
and red bag runs.  Neither waste type produced substantially
greater emissions for the pollutants measured.  Additionally, for
the pollutants most likely to be affected only by waste
characteristics and not combustion conditions  (HC1, S02, and
metals), the concentrations associated with red bag waste
generally are within the ranges obtained when general waste was
fired at Facility A and the other four intermittent-duty units.
The results of this test program  indicate that uncontrolled
emissions from the combustion of  red bag and general medical
waste are comparable.
     While the emissions from red bag and general waste
combustion are generally comparable, two additional observations
 (one related to Hg and one related to HC1) on  the  red  bag/general
waste comparison at Facility A  are noted below.  The principal
difference in emissions resulting from burning red bag and
general waste at Facility A  is  in the Hg emissions.  Figure  22
shows the effects of waste type on emissions  of Hg at  Facility A.
As  indicated in  Figure 22, Hg emissions from burning G100  (waste
obtained  from another hospital)  are one to two orders  of
magnitude lower  than Hg emissions measured while burning  G500 and
red bag wastes.  Also, with  the exception  of  the  first run of
 test condition  6  (RB  at 1800°F),  the Hg concentrations associated
with burning G500 waste are  higher than those from burning red
bag waste.  Figure  23  compares  the Hg  concentrations measured at
 Facility  A  with the Hg  concentrations  measured at  other
 facilities.  The Hg concentrations  found  during the  G500  (general
 medical  waste)  runs at  Facility A are  substantially higher than
                                101

-------
  II
 12112
 4§4i
nn--
* 8 -O * -9
*llsl
llill

is i
r °
                                   a>
                                   co
                                           co
                                           C
                                           o
                                           •H
                                           to
                                           co
                                           -H

                                           I
                                                            o
                                                            •H
                                                            JJ
                                                            CD
                                           O
                                           u
                                            m
                                            o
                                            n
                                            o
                                   CJ
                                           •H

                                           U
                                            £
                                            AJ
                                           AJ
                                           n
                                            0
                                            0)
                                           W
(2O %Z P) (u»sp/Bn) BH
                                                            0)
                                                            b
                                                            3
                                                            O)
              102

-------
lljjls
» E _ E _ & *
ulill!
111! Ill

Si  i  U
  c  -  5
 *  CO CD






-f  CO OJ


 +  CO -r-






4-  CD --
                                                                     O
           O
                                                                       o
                                                                              CO

                                                                              H




                                                                              CO
                                                                             c
                                                                             •H
                                                                             J_)
                                                                             c
                                                                             o
                                                                             u
                                                                             0)
                                                                             JJ
                                                                             JJ
                                                                             -H
                                                                             0)
                                                                             AJ
                                                                             c
                                                                             •H
                                                                  CM
                                                                      §
                                                           !-*
                                                                  < O)

                                                                  < 00

                                                                  < i- ra

                                                                  < co  S

                                                                  < CM

                                                                  < -



                                                                  < 1^

                                                                  < 
-------
those typically found at other facilities burning general medical
waste.  Although the source of the Hg in the Facility A waste has
not been identified, the data indicate that the general waste
stream at Facility A has an Hg source that differs from those at
the other facilities tested.  Overall, these data indicate that
Hg emissions may vary substantially from facility to facility and
between waste types, with neither RB nor general medical waste
consistently having higher Hg emissions.
     Figure 24 shows the effects of waste type on emissions of
HC1 at Facility A.  It is clear from Figure 24 that HC1
concentrations measured during two of the three test condition 7
runs  (RB at 2000°F) were substantially higher than those during
the other red bag runs.  These higher concentrations are the
result of spiking hexachlorobenzene as a cytotoxic surrogate
during these runs and are not attributed to the red bag wastes.
Additionally, two general waste runs  (test condition 3, run
number 2 and test condition 8,,run number 1) had HC1
concentrations somewhat higher than the other general waste runs.
However, no differences in waste feed conditions were noted
during these runs, and the increased HC1 concentrations for these
runs are assumed to be attributable to normal waste variability.
     4.2.1.2  Emissions from Combustion of General Medical Waste
Versus Pathological Waste.  Table 3 presents the post-combustion
emissions measured at Facility S during the combustion of general
and pathological wastes.  At Facility S, for the majority of
pollutants  (PM, CO, HC1, CDD/CDF, and Pb), the emissions from the
pathological waste runs were substantially less (often as much as
an order of magnitude less) than those from general waste
combustion.  The emission characteristics for the other
pollutants were more varied as described below.
     The concentrations of Cd emissions were generally lower for
pathological runs than for general waste runs.  The lone
exception was test condition 1, run no. 3, which had a Cd
concentration  (504 fig/dscm at 7 percent 02) that was more than an
order of magnitude higher than the next highest concentration
(26 /jg/dscm at 7 percent O2) from the pathological runs.  A
                               104

-------
   «  •
   I  !•
„•?* •
£ .5s • s *
§1-1 =
* C ^ c ^
II  I
•8 -  w
                                                                              8
                                                                 tn
                                                                 c
                                                                 o
                                                                 -H
                                                                 tn
                                                                 to
                                                                 -H

                                                                 0)
                                                                  c
                                                                  o
                                                                 •T-t
                                                                  JJ
                                                                  m
                                                                  3
                                                                             oo
                                                                                              O
                                                                                              u
                                                                                              I
                                                                                              JJ
                                                                                              ro
                                                                                              O
                                                                              J
                                                                               O
                                                                  c
                                                                  o
                                                                                              -i-f
                                                                                              U
                                                                                              ftf
                                                                                              re

                                                                                              Q>
                                                                                              OJ
                                                                                              4-1
                                                                                              to
                                                                                              44
                                                                                               o
                                                                                               a
(30 %^ P)
                                              IOH
CM

(U
h
3
W
-rf
fo
                                                  105

-------








rt
•
CO E"1



3 *
AT FACI
HOLOGIC/
gs
cvl

03 B
DQ ^*
i!
H *
Wi-3
tJ O
M W
Q

O S
00 Si
m E8
0 g
7
^o
O 53
£)
H
FH

§S

H
§


=

•3? ^

3.
e

E 3
$

*%
U 2
BO
a.
CCD/CDF,
ng/dicm
. >
X"C

z £

rt

c "i
« c.


c" 1
^ £

sl
tfi


«


1 s.
JS.
?1

r*i
•^


M




-



£

U"l
r-.


i

ec



t-


r^
C
O


-

3
1
1



0-
c
V

TT

ec


S



*

vn
oo


t-

oc


O
§


O
O


«

5
1
s
I




s£


o




V



1

5*



5

e-



g


o
o


o

8





r*i


O1
n



12



r-

oo
r-


S

^
—


IS
O


oo
S
0


„


1
£

"

t^
e
V



M


S



r-

eo



5

OO



»*-.
o-


i
e


-

R
1
£

cs

Vi
c
V


oe



a



§

f*".



i

***.
"-


w*.
O


oo
S
o


-

S
B6
I


-------
review of the charging records provided no indication that the
wastes used for this run were different from those used for the
other pathological runs.
     The Hg concentrations from three pathological runs and one
general waste run at Facility S were below the detection limit.
The other two general waste runs had concentrations of 10 ng/dscm
at 7 percent 02 or less.  However, two of the pathological runs
 (10 and 5) had Hg concentrations of 116 and 182 jtg/dscm at
7 percent O2, respectively.  Review of test records provided no
additional information  that could explain the higher levels.
Consequently, they are  assumed to be within the normal range of
Hg emissions from combustion of pathological waste.  Mercury
emissions from the combustion of pathological waste are low in
comparison to test runs conducted at other facilities  where
general medical waste and  red bag waste were combusted.
Additionally, Hg emissions from the  combustion  of  general medical
waste  at  Facility S  were significantly lower than  the  Hg
 emissions measured at other facilities. Overall, Hg  emissions
 from Facility S during  combustion of pathological  waste  and
 general  medical waste fall into the  low end  of  the range of Hg
 emissions measured  from all facilities during  this test  program.
      The two pollutants that showed  consistently higher
 concentrations  when pathological  wastes were burned than when
 general  wastes  were burned at Facility S  are NOX and S02.  A
 significant amount  of heat input  to the primary chamber from
 auxiliary fuel is required during the combustion of pathological
 waste.  Because pathological waste combustion requires the use of
 substantially more natural gas than the combustion of general
 waste, the higher NOX  concentrations are not unexpected.
 Initially, the higher  SO2 levels were somewhat surprising.
 However, a review of the biological literature suggest that the
 sulfur level in pathological tissue is high enough to explain the
 S02 concentrations  found  at Facility  S.91  Consequently,  both the
 NOX and S02 concentrations  found at Facility S are assumed to be
 reasonable uncontrolled levels for pathological waste combustion.
 Overall, the waste-related pollutant  emissions from pathological
                                107

-------
wastes are significantly lower than  from general medical wastes.
The figures in Appendix C provide a  comparison of Facility S data
with data from all of the other  tests.
4.2.2  Effects of Combustion Controls
     A wide variety of combustion control measures were described
in Section 2.2  As a part of the test program, two of those
parameters--waste feed rate and  secondary chamber temperature--
were varied intentionally between different test conditions.
Additionally, facilities with different design secondary chamber
residence times were selected for testing.  Variations in primary
chamber temperature and combustion air rates as indicated by
excess air levels were monitored during each test run.  The post-
combustion emission data were analyzed to assess the effects of
variation in these combustion parameters on PM, CO, and CDD/CDF
emissions.  These pollutants were selected because they are the
pollutants most likely to be affected by combustion controls.
     The subsections below summarize the results of these
analyses.  First, the interrelationships among CO, PM, and
CDD/CDF emissions are described.  Then, the relationships among
emissions of each of these three pollutants and combustion
process parameters are discussed.
     Initial examination of the data in Table 2 suggest that the
batch unit at Facility J operates quite differently than the
other units tested and that emissions for the Facility S
pathological waste combustion were different than those for the
Facility S general waste condition.  Because of these
differences, analyses of combustion controls and PM,  CO,  and
CDD/CDF emissions were conducted using only the data from
continuous and intermittent units firing general waste.
     4.2.2.1  Relationship of CDD/CDF to PM and CO Emissions.
One of the goals of the data analysis was to determine whether
relationships between emissions of CDD/CDF and CO and between
emissions of CDD/CDF and PM could be established.   If strong
relationships between these pollutants exist,  then CO and/or PM
could be used as surrogates for CDD/CDF emissions.   The results
show that, overall,  the relationship between CDD/CDF and CO is
                               108

-------
strong over" the full range of data and that large (order of
magnitude)  changes in CO are indicative of changes in CDD/CDF
emissions.   However, small changes in CO at low CO levels are not
related to changes in CDD/CDF emissions.  These results are
described in more detail below.
     Examination of the complete set of test data from continuous
and intermittent units showed CDD/CDF emissions to be strongly
related to both CO and PM emissions.  Figures 25 and 26 show
plots on a log-log scale of CDD/CDF versus CO and PM,
respectively.  Note that these plots suggest a nearly linear
relationship on a log-log scale between CDD/CDF emissions  and
both  PM and CO emissions over  the  full  range of the  test data.
This  visual observation is  confirmed by the estimated
correlations among  the data which  yield correlation  coefficients
of  0.74 between the logs  of CDD/CDF and PM and 0.85  between the
logs  of CDD/CDF and CO.
      While  these  results  suggest  reasonably  strong relationships,
further insight  into  the  relationships can be  obtained by looking
at  several  subsets  of the data.   First, the  data are divided into
two groups--those from newer facilities with better overall
 combustion control  measures (A,  B, and W) and those from older
 facilities (S and K).  The results for the two groups are quite
 different.   For the newer facilities a strong correlation  (0.90)
 between the logs of CDD/CDF and CO still exists, but the
 correlation between the logs of CDD/CDF and PM is substantially
 weaker (0.55).  However, for older facilities both  correlations .
 remain strong (0.90 for PM and 0.93 for  CO).
      On balance, these data show  a strong relationship between
 CDD/CDF and CO over a wide range  of emissions represented by the
 test data as a whole.  However, the results should  be  interpreted
 cautiously for any smaller range  of data.  Examination of the
 relationship between the  logs of  CDD/CDF and  CO  at  Facility A,
 which has  emissions  at  the low end of the emission  ranges, showed
 essentially  no  relationship between  CO concentrations and CDD/CDF
 emissions.
                                 109

-------
     100000
   jf 10000
   a.
   1
      1000.
                           ?  E-
      100
                                  •  100
                                                1000
                                                            10000
                            M**n CO Dry (ppnrtv) {ft 7% O2)

Figure 25.   Relationship  between CDD/CDF and CO  emissions,
    100000
   §
     10000,
      1000
      100
          0.01
                                   E    5

                                   0.10
                                                            1.00
                                      (•! 7% O2)
Figure 26.   Relationship between CDD/CDF and PM emissions,

                                 110

-------
     Taken"together,  the data suggest quite different
interpretations,  depending on the magnitude of changes in CO
levels.   Large (order of magnitude)  changes in CO concentrations
appear to be indicative of changes in CDD/CDF emissions.
However, because small changes in CO concentrations at low levels
such as those found at Facility A are not related to changes in
CDD/CDF emissions, CO does not provide a good surrogate for
CDD/CDF emissions at low concentrations.
     4.2.2.2  Effect of Combustion Parameters on PM. CO.  and
rnryrnF Emissions.  The three pollutants that are most affected
by combustion control parameters are PM, CO, and CDD/CDF.  In
developing the matrix of test conditions for this test program,
several combustion control parameters were considered potentially
to have an impact on emissions of PM, CO, and CDD/CDF.  These
parameters include secondary/tertiary chamber residence time,
secondary/tertiary chamber temperature, primary  chamber
temperature, and  waste  charge rate.  One of  the  goals  of  the test
program was  to measure  the impact of as many of  these  parameters
as possible  on emissions by  selecting MWI  facilities and  test
conditions that would allow  the  appropriate  comparisons  of
emission  results  to  be  made.  The evaluation of  the effect of
combustion parameters  on measured emissions  are  presented in two
sections. The  effects  of  variations in the  secondary/tertiary
combustion  chamber control parameters--secondary/tertiary chamber
temperature  and secondary/tertiary  chamber gas residence time-on
emissions are discussed in Section  4.2.2.2.1.  The effects of
variations  in the primary combustion chamber control
 parameters--primary chamber temperature and waste charging
 procedures--on emissions are discussed in Section 4.2.2.2.2.
      4.2.2.2.1.  Secondarv/tertiarv rhamber combustion control
 parameters.  In the secondary/tertiary chambers of MWI's,
 combustion gases from the primary chamber containing volatiles
 are further combusted under excess-air conditions.  Effective
 combustion  in the secondary/tertiary chambers is accomplished
 when sufficiently high temperatures are attained,  turbulent
 conditions  are achieved in the presence of  sufficient oxygen
                                 ill

-------
 (good mixing of combustion air with  volatiles) , and sufficient
residence time at high temperature is  allowed.  Because all MWI's
are designed to operate the  secondary/tertiary  chambers in an
excess-air mode and because  all MWI's  incorporate secondary
chamber designs that introduce some  degree of turbulent mixing
between combustion air and primary chamber exhaust gases, the
focus of the test program was  to  select MWI facilities with
specific features, to operate  them at  specified conditions, and
to measure post-combust ion emissions to determine the impact of
secondary/tertiary chamber temperature and residence time on
emissions.  This section first summarizes the variations in
secondary/tertiary combustion  control  parameters integrated into'
the test program design and  subsequently presents an evaluation
of the emission results with respect to these combustion control
parameters.
     Testing conducted at Facilities A, W, and  K included test
conditions where general medical  waste was charged to the MWI and
target secondary/tertiary chamber temperatures  were 870°C
(1600°F) and 980° to 1040°C  (1800° to  1900°F).   (Facility A also
included a test condition where general medical waste was charged
and the target secondary/tertiary chamber temperature was 1090°C
[2000°F].)  At Facilities B  and M, testing comprised charging
general medical waste with a target  secondary/tertiary chamber
temperature of 980°C (1800°F).  Testing at Facility S included
one test condition where general  medical waste  was charged and
the target secondary/tertiary  chamber  temperature was 870°C
(1600°F).  These MWI facilities were selected for testing in part
because they had different design secondary/tertiary chamber
residence times.  Facility A has  a design secondary/tertiary
chamber residence time of 1.34  seconds at 1800°F.  The design
secondary chamber residence  time  at  Facility W  is 1.0 second;
Facility K is 0.33 second; Facility  B  is 1.75 seconds; and
Facility M is 2.1 seconds.   In designing the test program in this
manner,  the effect of secondary/tertiary chamber temperature and
residence time on the emissions of CDD/CDF,  CO, and PM could be
evaluated.  Figures 27, 28,  and 29 present the data from these
                                112

-------
                                                                 c
                                                                 o

                                                                 0)

                                                                 -H
                                                                 JJ

                                                                 0)
                                                                 o
                                                                 c
                                                                 Q)
                                                                 "O
                                                                 -H
                                                                 CD
                                                                 0)
                                                                 i-l

                                                                 •O
                                                                 c
                                                                 OS
                                                                  AJ
                                                                  m
                                                                  v-<
                                                                  Q)
                                                                  Q)
                                                                  JJ
                                                                  (1)
                                                                                CO

                                                                                O
                                                                              Q -H
                                                                              E tn
                                                                              rt en


                                                                              SB
                                                                                0)
                                                                   ro Q
                                                                  -H a
                                                                   4J *»v.
                                                                   S-i Q
                                                                   0) Q
                                                                   aJ U
                                                                   m
                                                                  T3
                                                                   c
                                                                   o
                                                                   u
                                                                   0)
                                                                   en
                                                                   u
                                                                   cu
                                            -II"
(so
(aosp/Bu) jao/aao
I
Dl
•H
Pu
                    113

-------
                                             §


                                             i
                                             •H
                                             JJ

                                             0)
                                             CJ
                                             C

                                             •S
                                             •H
                                             CQ
                                             0)
                                             rt
                                             i-l
                                             3
                                             jJ

                                             2
                                             0)
                                             0)
                                             Jj
e
                                                CQ

                                                CJ

                                                O
                                                •H

                                                CO

                                                CO
                                             (C
                                             •H

                                             JJ


                                             
B   o   o   g    o   o   o
71   o   o>   oo    h»   w   u>


       (SO %Z P) (Apiudd) OO
                  8
                                             •a
                                             o
                                             u
                                             0)
                                             m

                                             M-l
                                             o
                                             u
                                             0)
                     eo

                     «N


                     0)
                     h


                     CD
114

-------
                                  I

                                  c
                                  o


                                  8
                                  •H
                                  OJ

                                  0)
                                  O
                                  c
                                  o
                                  !°
                                  m
                                  a)
                               CO
                            <  s
                                    •O
                                    C
                                    rt

                                    0)
§
HiniSS3S9!3918SS8S§
         (30 %* P)
                                   OJ
                                   rt
                                   j-i
                                   0)
                                   0)
                                   4J
                             §
e
                                   (C
                                   •H


                                   S-l
                                   0)
                                   4J
                                     m
                                     c
                                     o
                                     •t-1
                                     m
                                     m
                                     •H

                                     I

                                   rt
                                   T3
                                   C
                                   O
                                   U
                                   0)
                                   m
                                   o
                                   0)
                                   U-l
                  115

-------
tests for CDD/CDF, CO, and PM, respectively, as a function of
both secondary/tertiary chamber temperature and residence time.
The data presented in these figures are the averages of the three
test runs conducted at each test condition.  The data for each
facility are plotted in order of increasing residence time (x-
axis).  The data for the 980°C  (1800°F) test conditions conducted
at five of the six facilities are represented by solid triangles,
The data for the 870°C  (1600°F) test conditions conducted at
three of the six facilities are represented by open squares-.  For
each figure, to highlight the general trend of the results,
separate lines have been added to connect the 980°C  (1800°F) data
points and the 870°C  (1600°F) data points.  Note that although CO
and CDD/CDF emissions were measured during the 870°C  (1600°F)
general medical waste test condition at Facility A, PM emissions
were not measured.  Therefore, the PM results from the 870°C
(1600°F) RB test condition are shown because, as discussed
earlier, there is little difference in the general medical and RB
emission rates.  Also, note that the PM emissions are not shown
for Facility M.  The results of the PM test at Facility M showed
that, as expected, a rotary kiln MWI has a higher post-combust ion
PM emission rate than other MWI's because of the increased
turbulence of the ash bed in rotary kilns.  Therefore, the PM
data from Facility M were not comparable to the data from fixed
hearth MWI's.  In reviewing the test data, the following general
observations can be made:
     1.  An increase in the secondary/tertiary chamber
temperature from 870°C  (1600°F) to 980°C  (1800°F) decreases the
emissions of PM, CO, and CDD/CDF for the tests conducted on units
with residence times of less than or equal to 1 second.  As the
secondary chamber residence time increases, the effect on
emissions of increasing the secondary chamber temperature
decreases.  Secondary chamber temperature has much less effect on
emissions for the facilities which had longer residence times and
better combustion air control.  At Facility A (with a design
secondary chamber residence time of 1.34 seconds),  there is
                               116

-------
essentially' no decrease in emissions as a result of increasing
the secondary chamber temperature;
     2.  As indicated by the downward trend in the lines
connecting the 870°C (1600°F)  data and the 980°C (1800°F)  data,
as the secondary/tertiary chamber residence time increases,  the
emissions of CDD/CDF, CO and PM decrease;
     3.  The emissions measured at Facilities A, B, and M at a
nominal secondary/tertiary chamber temperature of 980°C (1800°F)
and a secondary/tertiary chamber residence time of at least 1.34
seconds are significantly lower than the emissions measured under
operating conditions with 870°C  (1600°F) secondary chamber
temperatures and shorter residence times.
     4.2.2.2.2.  Primary chamber combustion control parameters.
In the primary chamber of most MWI's, waste is  combusted under
substoichiometric air conditions.  The purpose  of maintaining a
substoichiometric air condition  in the primary  chamber  is to
maintain control of  the waste combustion rate.  The objective is
to completely combust the waste  in the primary  chamber  at a rate
that does not generate more combustible  gases than the
secondary/tertiary combustion chambers  can efficiently  and
completely  combust.  The  rate at which  combustion  proceeds  in the
primary  chamber  can  be  controlled  in several  ways.  These primary
chamber  combustion control  methods were described  in  Section  2
and  include:   (1) proper  waste  charging procedures;  (2)  control
of combustion airflow  to  the primary chamber; and  (3)  use  of  a
water  spray.   Proper waste  charging procedures  are operational
practices  that include charging waste in charges of .similar
weight at  regular intervals.   The  other two  methods  incorporate
MWI  combustion control system designs that use  primary chamber
 temperature as an indicator of the rate of combustion in the
 primary chamber.  In these designs,  when primary chamber
 temperature exceeds a specified setpoint temperature, a control
 loop is triggered that either decreases the combustion airflow
 into the primary chamber or activates a water spray system to
 help reduce temperatures below the setpoint.
                                117

-------
     All MWI's control combustion in the primary chamber through
proper waste charging procedures.  The combustion air rate and
distribution to the primary and secondary combustion chambers
also are controlled to some degree for all MWI's; however,
sophistication of the design of the air control system varies and
the level and degree of air control achieved varies widely.  Some
MWI systems have combustion control designs that allow the
combustion air rate and distribution to be varied very little, if
at all, from predetermined settings during the combustion
process.  As a result, these MWI's must rely solely on proper
charging procedures to maintain operation of the unit within the
limits of the preset conditions to achieve proper combustion.
     Combustion control at Facility A is achieved through use of
a combination of techniques.  First, the primary air combustion
rate is set at a level which will maintain the primary chamber at
a substoichiometric condition when the unit is charged at or near
the design rate.  Second, a water spray system is provided in the
primary chamber to help quench the primary chamber combustion if
the temperature begins to exceed the desired level; the water
spray system is activated at a setpoint temperature of 730°C
(1350°F).  Third, the waste charge amount and frequency is
carefully monitored and controlled.  Secondary combustion air is
provided via a separate system which is independently controlled;
the secondary combustion air is fully modulated.
     At Facilities K and W, a single combustion air blower
provides the air for the primary and secondary combustion
chambers.  Combustion control is accomplished through the use of
a modulating damper that controls the distribution of the
combustion air to the primary and secondary chambers.  The
control system at Facility W is fully modulated whereas at
Facility K, the modulating damper does not move very much about a
preset position.
     One of the objectives of this test program was to determine
the impact on emissions of alternative waste charging practices,
At Facilities K and W,  (MWI's with the same capacities but
different secondary chamber residence times),  the following waste
                               118

-------
charging practices were employed:  test condition 1 comprised
undercharging waste,  i.e.,  charging at a rate (200 Ib/hr)  less
than design capacity at 6-minute intervals; test condition 2
comprised charging waste at the MWI capacity charge rate
(300 Ib/hr) at 6-minute intervals; and test condition 3 was
comprised of charging waste at the MWI capacity charge rate
(300 Ib/hr) at 10-minute intervals (larger, less-frequent
charges) . '
     Table 4 presents the average emission rates of PM, CO, and
CDD/CDF  for the test conditions  conducted at Facilities K and W.
It is clear that at Facility K,  emissions of PM, CO, and CDD/CDF
increase from test condition 1 to test condition 2 to test
condition  3.  The differences in emissions between test
conditions l and 2 are  relatively small.  The differences  in
emissions  between test  conditions 2 and 3 are dramatic with a
four-fold  increase in  PM emissions, an order of magnitude
increase in CO emissions, and a  five-fold  increase  in  CDD/CDF
emissions.  Note also,  that the  average primary  chamber
temperature for test conditions  2 and 3 conducted  at  the  design
capacity charge conditions  (1820°F and 1835°F,  respectively)  also
were significantly higher  than  for test condition  1 (1620°F).
These data indicate  that at Facility  K,.infrequent waste  charging
 (large  charges  every 10 minutes  as opposed to  smaller charges
every  6 minutes)  coupled with poor combustion  control in  the
primary chamber (as  indicated by higher primary chamber
 temperatures),  and a lower secondary  chamber temperature,  results
 in a significant  increase  in  emissions.
      The data obtained at  Facility W also show an increase in
 emissions of PM,  CO, and CDD/CDF from test condition 1 to test
 condition 2.   The average primary chamber temperature for test
 condition 2 was also higher than for test condition 1.  However,
 when the size and frequency of  the charges were changed (larger.
 less frequent charges--test condition 3)  no increase in emissions
 was noted.  For test condition  3, the primary chamber temperature
 was actually lower than for condition 2,  indicating that
                                119

-------
w
erf
W
Cu
b
H
Q

tt

g
8
   CO
   H
S



w
   04
   o
Vb
1
H
U




I
            §1
            U
            •s*  8
            S T3
            a. t
                00
         •g.fi
a.


u

I  U.
         60

                   in O
                      00 —
                      oo ts
                   O O O
                   >c r-- \c
                   o o o
                   ts cs n
                   vC OS «
                   VO
                       S
                                   *
                                §9^ O^
                                — —
                             COO
                             oo o
                                 o  o
                                 —  >c
                                 oo  >c
                              \D  vo  O
                                 88
                                           &
                                           I
                                          •8
                              a.

                              ^


                              •o

                              w
                              V
                              =c

                              O


                              °3
                                           I
                                           I
                                          I
                                                    120

-------
Facility w"incorporates more effective primary chamber combustion
controls than Facility K.
     At Facility A, the primary combustion chamber temperature
was controlled and maintained at a temperature of less than 730°C
 (1350°F) for all runs, indicating good combustion control in the
primary chamber.  Proper waste charging procedures were followed
for all test conditions.  Charge rates.were adjusted based on the
target secondary/tertiary chamber temperature  (i.e., lower charge
rates were used for the 870°C  [1600°F] secondary chamber test
 condition  than for the 980°C  [1800°F]  test condition).  The
 emission results obtained during the  test at  Facility A indicate
 that proper waste  charging  procedures coupled with  good control
 of primary chamber temperatures, a  longer residence time  (than
 Facilities K and W),  and good control of  secondary/tertiary
 chamber temperature produce the lowest emissions of PM,  CO,  and
 CDD/CDF between Facilities  A,  W,  and K.
 423   Kmissiorg During Burndown and Cnnldown Phases
    " Manual  emission  test  data were collected at the seven
 facilities during  only the normal combustion component of the
 operating cycle.   However,  CEM data were collected during the
 burndown/cooldown phase at Facility A and during the cooldown
 phase at Facility J.   Figures 30 and 31 show example plots of CO
 and THC concentrations for one 24-hour period at Facilities A and
 j  respectively.  Corresponding plots of the'primary and
 secondary chamber temperature profiles over  the same 24-hour
 periods are shown in Figures 32 and  33.  The data  from the other
 test runs are comparable to  those  shown in the  example plots.
      These 24-hour CEM data  indicate that at the two facilities
 tested, CO and THC emissions during  the cooldown period  are
 substantially higher than  those during "normal" operations.
 Also,  given the very high  levels  of  CO emissions measured (always
 over'l 000 ppmv and  sometimes as high as  50,000 pprrtv)  and the
  strong relationship  between CO emissions and CDD/CDF found at
  these high CO levels,  CDD/CDF emissions during cooldown could be
  substantial.
                                 121

-------
        WOOD-
        13000
                       wghgOparaHo
 • Bumdown
                                                                  . Cooktown
        12000 ^ -
           i
        TIOOO-;

      O 10000-
      I 8000 •


      fc 700°:

      i eooo:


      ! MOO]

      8' 4000 -
            i
        3000
        2000-
                                        End Charging
                                            Bumdown
            Begin Charging
                                                           Buminan/
                                                         Begin Cooidown
Sec. Chamber
AuxBurneroB
Burner on




J™^;J.
           09:30 11:00  1230 14:00  15:30 17:00  18:30 20:00 21.30 23:00 00:30 02:00 03:30 06:00 06:30

                                   CO  TU4E (24 HOUR)               ^



                  Figure  30.    24-hour  real  time  CO  and

THC concentrations--Facility A,  test condition  2,  run  number 2
       60000 •>
                    . UwMr-
                                    . High AK.
                                     (Sundown)
       soooo'


     §    !
     Jf 40000-
     §    I
      . 30000-
      1 20000]
     8
       10000-
                                                   ,'. t:
                                               CO
                                                       THC
          0*30  1030   13:30
                               1*30  13:30  SD20  22:30  00:30  02:30  04:30  06:30
                                  co
                  Figure 31.   24-hour  real time  CO and
       THC concentrations--Facility J,  run numbers  5  and
                                          122

-------
1900
1800
1700
1800
            -Charging Operator
   iSCTEMP
                            •ifml f'.
                                 ,v..
                                          Sundown
                                                          . Coottown
                                        Stoke
                                           Stoke
 1200 -;   Ji,
| 1100 '- PC TEMP
                                                          SCTEMP
  800 :
  TOO
  eooj
  500 J
  400
  300
  200
  100
                                                  PC TEMP
                                End Charging

Begin
Bumdo
1
wn
Sec. Chamber
Aux Burner off
Sec. Pilot
Burners OKI
Begin Cootdomm
Boiler Aux.
1 Burner on
1 i
    06:30  11:00  12:30
                      15.30  17:00 18:30  20:00 21.30  23:00 00:30 02:00  03.30 05:00  06,30
   Figure  32.   24-hour temperature plot--Facility A,
               test  condition  2,  run number 2.
 2300 '•'
 2200•'
  2WO-
 2000 :
  1800-
tr1800 ISCTEMP
  1700'
               i~- **
                                                   Cooktown
                               (Bumdown)
                                                                 sc TEMP
      OfrSO  10:30  1230   14:30   W:30
                                                                       i
                                                                       ,!
                                     20:30  22:30  OOJO  0230  0430  06.30  08:30
                                     (24 HOUR)
     Figure 33.   24-hour  temperature  plot--Facility J,
                        run  numbers  5 and  6.
                                     123

-------
     Further examination of the  CEM data  for the cooldown period
provides insight into potential  control options.  Figures 34 and
35 provide an expanded view of the cooldown period for the data
presented earlier.  These figures overlay the CO concentration
curve and the temperature plots.  Note that at both facilities,
primary chamber temperatures remain elevated for some period
after the cooldown period starts.   (In fact, for some runs not
shown at Facility J, the primary chamber  temperature actually
increased during cooldown).  These elevated temperatures indicate
that waste burnout at these facilities has not been completed
before the start of the cooldown period.  Consequently, unburned
organic material may still be released to the secondary chamber.
However, during the cooldown period the auxiliary fuel combustion
in the secondary chamber is substantially reduced or completely
shutdown, thereby accounting for the reductions in secondary
chamber temperature shown in the graphs.  This secondary chamber
temperature decrease is accompanied by increases in CO
concentrations.
     The available CEM data suggest that primary chamber
temperatures may be indicative of complete burndown,  Generally,
CO concentrations appear to remain at high levels until primary
chamber temperatures fall to the 120° to  150°C (250° to 350°F)
range.  These data suggest,  therefore, that the significant
emissions during the cooldown period may be controlled by
maintaining the secondary chamber temperature at normal operating
levels until burndown is complete,  as indicated by primary
chamber temperature.
4.2.4  Operating Cycle of a Batch MWI
     Batch MWI's operate on a different cycle than intermittent
and continuous MWI's.  In batch MWI's, all of the waste to be
burned during a complete batch cycle is loaded into the primary
chamber prior to operation with no additional waste charging
during operation.  The batch operating cycle normally takes l to
2 days depending on the size of  the MWI and the amount of waste
charged.  In intermittent and continuous MWI's,  waste is charged
                               124

-------
    14000 ;
                                            .Cootdc
 12000:
 •nooo;
 10000
O 9000
£
a 8000
        :SCTEMP
1>  I
1i
                         'l tl
               t
   I 7000 :
   x    i       '
   S 	!
   > 8000 < pc TEMP !
   O    I *	
   ' 9000 i     " ~1 -
                         I
                         i
                        •4
                                                          PC opocwd for
                                                          Art Removal
        1  B*gto Coottown
     2000 \
     1000-
        :CO
        01:30
                                                 CO
(. 1900
 •woo
 1700
 1600
 1500
 1400
 1300
 1200
 1100
 WOO
  900
  800
  700
  800
  500
  400
                                            «;TEMP
                                            SCTHMP
                                                                    200
                                                                    100
                    02:30
                               03:30
                                           04:30
                                                       06:30
                                                                   06:30
                                     (24HOUR)
             Figure  34.    Cooldown period--Facility  A,
                   test condition 2,  run  number  2.
60000 -





SOOOO;





g 40000
^
*
§
-DHY(ppnw)
§20000



WOOD

0

I




I
i





SCTEMP

._
"~- ^
PC TEMP ""'-^ v



i .-"*'

-"
CO







V



;
V
\
(
\
\
\
"" r U
- ' *\£









i
I










\
\
\
\
-:T>- -,
• ^--^r ~ —
^ ^— _ —
x~._ "" ' "- -""- ": ^~_ - .. SCTEMP
--~^ PC TEMP
^1, 	 -.
1 ^-- - -
1!""--
• CO
jnyyi
2200

2100
2000

1900
1800

1700

1600
1500

1400
1300
1200
1100
WOO
900
800
700
600
500
400
300
200
100








£




i
1
§
1

1
*




         2230
                23:30    00:30
                            01:30    02:30    03:30
                                TIME (24 HOUR)
                                                    0430
                                                                   0&30
Figure 35.   Cooldown  period--Facility J,  run numbers  5 and  6,
                                         125

-------
in small charges on a regular basis throughout much of the
operating cycle.
     In addition to the waste charging differences, the burn
cycle of a batch MWI proceeds differently than that of a
continuous or intermittent MWI.  The burn cycle of a batch MWI
comprises three distinct phases--a low-air phase, a high-air
phase, and a cooldown phase.  During the low-air phase,
combustion of the waste proceeds very slowly and primary chamber
temperature increases gradually over several hours.  Combustion
proceeds more rapidly during the high-air phase and the primary
chamber reaches and maintains maximum operating temperatures.
Continuous and intermittent MWI's do not incorporate a low-air
phase and, therefore, reach maximum operating temperatures in
only a fraction of the time required for a batch unit.
     Table 5 presents the average post-combustion emission
concentrations for test conditions conducted at the batch MWI at
Facility J during the low-air phase (test condition 1) and the
high-air phase  (test condition 2).  Additionally, Table 5
presents post-combustion emission concentrations for the test
conditions indicated for Facilities A, B, M, and W where the
target secondary/tertiary chamber temperatures were the same or
similar to the tests conducted at Facility J.  In comparing the
emission concentrations for the low-air phase at Facility J with
the emission concentrations at the other facilities, the post-
combustion emission levels for all pollutants except Hg are
significantly lower for the batch MWI than for the intermittent
and continuous MWI's,,  The emission concentrations for the high-
air phase at Facility J, however, are comparable to the
concentrations from the other facilities.  These comparisons
suggest that during the high-air phase of operation, batch MWI's
achieve a somewhat steady-state operating level that is
equivalent to the steady-state levels of continuous and
intermittent units.
                               126

-------

g
Bi
en

om  ,
H m|
cc •
2
w
o w

H EH

EH EH



Bg
8
w 03

O D
ft O
 0S

 §§
 CO U
 H
 ft

 !
LE
       E
      . o
     oo 45

     fl
          R
     If
     u
      U
S
        T3

      O. ^:
              s
                    R
                    \C
        I
                              00
                              vO
                              8
8
                        VC

                              in
                              «^.
                               cs
                                   2 •- •-


                                   « s= — 2
                                   11 '5 -s

                                   8tis
                                   CA ed 0} uJ



                                   111 i
                                   S ""5 'S =
                                    - ^ ^ 1

                                    <^^l
                                    eo & v-9
                                   127

-------
4.2.5  Metals Partitioning
     In general,  combustion control measures are not used to
limit metals emissions from MWI's.  However, some combustion
parameters,  particularly primary chamber temperature and
combustion air controls in the primary chamber, can affect the
partitioning of metals among the different incinerator exhaust
streams.  The data from the EPA test program were examined to
determine whether the relative fractions of Cd, Pb, and Hg that
were partitioned to the bottom ash and exhaust gas varied
substantially across test runs, and if so, whether the
differences could be related to process operating conditions.
     Data from the emission tests were sufficient to estimate
relative fractions of metals distributed to the bottom ash and
stack gas for a total of 15 test conditions at 4 "-.cilities.
Table 6 summarizes the average fraction of Pb, Cd, and Hg
partitioned to each of the two streams by facility and test
condition.  Note that for Facility A, the ash samples collected
for each run were composited across the different runs by test
condition prior to metals analysis.  Consequently, only one
sample was obtained per test condition and this one sample was
compared to average stack gas emissions over the runs that
constituted the test conditions.  For other facilities, the
values in Table 6 are based on distinct ash and stack gas samples
for each of three runs per test condition.
     With the exception of the data from Facility S, the
partitioning data are relatively consistent and follow expected
patterns.  Generally, almost no Hg remains in the bottom ash,
with the percentage ranging from less than 1 percent at
Facilities A and W to about 6 percent at Facility K.  Cadmium
partitions primarily to the stack gas with the percentage in the
bottom ash ranging from about 14 percent at Facility W to about
37 percent at Facility A.  Finally, Pb partitions approximately
equally in the bottom ash and stack gas with the percentage in
the bottom ash ranging from 44 percent at Facility K to
59 percent at Facility A.  Among these three facilities, the
                               128

-------
      TABLE  6.   SUMMARY OF BOTTOM ASH AND STACK GAS METALS
                           DISTRIBUTION
Facility
A
K
M
S
w
Test
condition
1
3
4 '
5
6
7
1
2
3
1
">
1
2 .
3
1
2
3
Cadmium
B-Ash
fraction,
mean
0.245
0.140
0.258
0.297
0.235
0.389
0.366
0.378
0.355
0.150
0.104
0.793
a
0.722
0.163
0.156
0.109
Stack
fraction,
mean
0.755
0.860
0.742
0.703
0.765
0.611
0.634
0.622
0.645
0.850
0.896
0.207
a
0.278
0.837
0.844
0.891
Lead
B-Ash
fraction,
mean
0.559
0.489
0.671
0.584
0.541
0.700
0.331
0.430
0.523
• 0.301
0.324
a
a
0.937
0.416
0.898
0.431
Stack
fraction.
mean
0.441
0.511
0.329
0.416
0.459
0.300
0.669
0.570
0.477
0.699
0.676
a
a
0.063
0.584
0.102
0.569
Mercurv
B-Ash
fraction.
mean
0.000
0.000
0.003
0.000
0.000
0.001
0.065
0.063
0.063
a
a
a
a
0.663
0.003
0.024
0.000
Stack
fraction.
mean
1.000
1.000
0.997
1.000
1.000
0.999
0.935
0.937
0.937
a
a
a
a
0.337
0.997
0.976
1.000
  aAll B-ash values at detection limit. Fraction could not be determined.

  B-ash: Bottom ash
  Stack: Incinerator exhaust gas
distribution showed some variation, but the variation could not
be related to waste type or primary chamber temperature.
     The distribution at Facility S differed  substantially from
those at the other facilities with significantly greater
fractions of metals remaining in the bottom ash.   The percentages
of metals in the bottom ash were Cd--75 percent and
Pb--90 percent.   The results appeared to  be consistent across
waste types.   In general, primary chamber operating data appeared
to be reasonably consistent across facilities,  and no explanation
was found for  the difference between Facility S and the other
facilities.  However, the data from Facility  S should be
interpreted  cautiously in that ash levels were near detection
limits.
                                 129

-------
4.2.6  APCS' Performance
     The emission test program generated performance data for
four APCS's--a DI/FF system at Facility A, a VS/PB system at
Facility B, a FF/PB system at Facility J, and a SD/FF system at
Facility M.  At Facilities B and J, performance data were
collected for only a single test condition at each facility;
consequently parametric analyses of APCS performance at these
facilities were not feasible.  However, data were collected for a
range of operating conditions at Facility A, making detailed
analysis of the DI/FF system performance possible.
     The performance of these four APCS's is described in three
subsections below.  The first briefly summarizes the performance
of VS/PB system and FF/PB system found at Facilities B and J,
respectively.  The second presents the results of the evaluation
of the effect of system operating parameters on the DI/FF
performance at Facility A including the effect of activated
carbon injection.  The third subsection describes the performance
of the SD/FF system at Facility M and the effect of activated
carbon injection on system performance.
     4.2.6.1  Performance of the VS/PB and FF/PB Systems.
Pollutant specific removal efficiencies for PM, HC1, CDD/CDF, and
three metals  (Cd, Pb, and Hg) are shown for each run at
Facilities B and J in Table 7.  Note that for Facility J, the
test runs represent different components of the operating cycle.
Those test runs designated by BR are from the low-air component
of the cycle., while those designated as BD are from the high-air
component of the cycle.
     The data from Facility B indicate that the VS/PB system
performs very well with respect to HC1 emissions with a
consistent removal efficiency of 99.8 percent or greater and
reasonably well with respect to CDD/CDF  (approximately 70 percent
removal).  Given the moderately high VS pressure drop (30 in.
water gauge), the PM removal efficiency was quite low, about
60 percent on the average.  The particle size data collected at
the inlet to the VS/PB suggest that the particles at this
facility are relatively fine  (about 60 percent of PM in the
                               130

-------
         TABLE 7.  PERFORMANCE  OF THE VS/PB AND FF/PB AIR
                     POLLUTION CONTROL SYSTEMS

Facility
B


J






APCD
VS/PB


FF/PB






Test
MM 1-1
MM 1-3
MM 1-4
Jl-BR
J2-BD
J3-BR
J4-BD
J5-BR
J6-BD
Removal efficiency, percent
PM
48.3
21.2
72.4
>94.2
97.6
73.7
94.4
56.8
95.0
HC1
100
100
99.8
92.4
82.7
90.4
69.1
89.1
97.1
CDD/CDF
66.3
69.4
77.8
b
b
b
b
b
64.1
Cd
34.9
34.2
54.9
93.6
98.9
92.0
98.9
92.4
98.1
Pb
40.6
44.7
50.1
99.3
99.3
98.2
99.4
98.9
99.0
Hg
62.5
a
44.0
65.3
59.7
81.9
79.7
87.6
69.5
  aOutlet Hg emissions measurements were higher than inlet Hg emissions measurements.
  ^Outlet emissions were higher than inlet emissions measurements by factors of 5 to 2.500.

 submicron range),  which may in part explain the low efficiencies.
 Due  in  part to a low average  inlet  loading of 0.1 gr/dscf at
 7 percent 02,  Facility B did  achieve  moderately low outlet PM
 concentrations with the outlet concentration averaging
 0.046 gr/dscf at 7 percent 02.   The VS/PB  system provided some
 removal of Cd (41 percent) and Pb  (42 percent), but the
 performance relative to Hg varied widely (no control up to
 62.3 percent removal) suggesting that on the average, little Hg
 control is achieved by the system.
      The FF/PB at Facility J  generally  performed quite well with
 respect to PM and metals emissions.   The PM removal efficiencies
 were generally 94 percent or  higher with the exception of two
 low-air component  (74 and 57  percent) runs that had low inlet PM
 loadings  (0.004 gr/dscf  for both runs).   For all six runs, the
 outlet PM loadings were  less  than  0.004 gr/dscf at 7 percent 02
.with an average outlet loading  for the  six runs of 0.0018 gr/dscf
 at 7 percent 02-  When metals inlet concentrations were at
 "typical" levels such as  those  achieved during the high- air
 component of the cycle,  Cd removal efficiencies exceeded
 98 percent while Pb  removal efficiencies exceeded 99 percent.
 For Hg, which showed no  real  variation  between the two operating
 cycle components,  the  overall removal efficiency was about
                                 131

-------
73 percent with no apparent relationship between removal
efficiency and inlet loading.  The HC1 removal efficiencies at
Facility J ranged from 69 to 97 percent with an average of
87 percent.  The variation in efficiency had no apparent
relationship to inlet HC1 concentrations, and no apparent process
differences were found to account for the one extremely low
efficiency (69 percent on Test J4) or the generally poor
performance of the system.  Subsequent to the test program, a
representative of the engineering firm that designed the facility
visited the facility to examine the APCS.  He found that
approximately 2 months after the test, six of the eight scrubber
spray nozzles were plugged with material commonly encountered
during start-up.92   Although there had been a 2-month interval
since the test, the nature of the plugging material convinced the
engineering firm representative that it had probably been there
during the test and prevented adequate scrubbing liquor from
entering the scrubber and may have contributed to the low HC1
removal efficiencies.
     For all but one test run at Facility J, the levels of
CDD/CDF were substantially higher at the APCD outlet than at the
heat exchanger inlet.  An understanding  of the system
configuration at Facility J helps in interpreting these results.
The exhaust from the incinerator first passes through an indirect
air-to-air heat exchanger into the fabric filter before passing
through the packed bed scrubber and out  the stack.  The post-
combustion sample location was upstream  of the heat exchanger and
the post-APCS sample location was downstream from the packed bed.
The temperature drops from over 590°C  (1100°F) at the inlet of
the heat  exchanger to 85°C  (185°F) at the outlet of the packed
bed scrubber.  Although the exact amount of time required to
accomplish the drop in temperature is unknown, this range spans
temperatures over which de novo synthesis of ODD/CDF is.known to
occur.  Consequently, CDD/CDF formation  potentially occurred in
the heat  exchanger, and/or fabric filter, and is not unexpected.
These data indicate that  the CDD/CDF that is formed in the system
is not retained in the fabric filter or  packed bed scrubber.
                               132

-------
     4.2. 6." 2  Performanr-P of the DI/FF rontrol System.   As
outlined in Section 4.1.1.1, two test sequences were performed on
the DI/FF system at Facility A.   During 1990,  tests were
conducted under 7 different operating conditions.   Generally,
these different conditions were related to the incinerator
operation and were not.expected to affect the APCS performance.
However, lime rates were varied during those tests and factors
such as baghouse inlet temperatures and inlet gas moisture
content, which could affect the DI/FF performance, were monitored
during all  tests.  Subsequently, during 1991, additional tests
were conducted under three  test conditions to assess the effects
of carbon injection on APCS performance.  The paragraphs below
describe the performance of the DI/FF relative to  PM, HC1,
CDD/CDF, and metals emissions and summarize the results of the
evaluation  of the  effects  of operating parameters  on performance.
     The data on PM concentrations at  the outlet  of  the DI/FF
system  at Facility A that  were  presented  in Table  B-3  indicate
that the system  typically  achieves outlet PM  levels  corrected to
7  percent O2  of  0.004  gr/dscf  or  less.  The  four  exceptions  to
these  low  readings were  test  condition 1, run 1  (0.068 gr/dscf},
test condition  5,  run  2  (0.056  gr/dscf),  test condition 5,  run 3
 (0.088  gr/dscf),  and  test  condition  7,  run  1  (0.020 gr/dscf), all
 of which were obtained during the first series of tests.   Such
 large  variations in outlet concentrations are unusual for fabric
 filter systems,  and examination of plant operating data showed no
 readily apparent differences in operating conditions during the
 runs with high PM loadings.  Also,  the data showed no
 deterioration in metals performance during those runs.  Between
 the 1990 and 1991 test series, all of the filter bags and the
 entire outlet plenum were replaced.   During the 1991 test series,
 PM concentrations were consistently low with all  runs measuring
 0.004 gr/dscf or  less @ 7 percent 02.  The inconsistent results
 obtained during the first  test series indicate that the  fabric
 filter system might have been  in need of repair at  that  time.
 Consequently, the high  PM levels found during a  few runs at  the
                                 133

-------
 first series of tests are not considered representative of FF
 performance.
      Two measures of HC1 performance were considered for
 Facility A,  mass emission rate-based removal efficiency and
 concentration-based removal.  Because the two measures were
 highly correlated (r = 0.99) and concentration-based measures
 were available  for more test runs than were mass emission rate-
 based measures,  concentration-based efficiencies were used in the
 analyses.   Table 8 presents average removal efficiencies and
 inlet and outlet concentrations for each of the nine test
 conditions.   The average removal efficiencies range from
 93.2 percent for Condition 5 to 98.6 percent for Condition 8.
 The  average  for Condition 5 was strongly influenced by a low
 efficiency of 88 percent on one of the three runs.   This
 efficiency was  substantially below the next lowest  value of
 92 percent.

           TABLE 8.  FACILITY A HC1 PERFORMANCE SUMMARY
Test condition
1
2
3
4
5
6
7
1A
8
9
l^^g^^gg^8fgB-J- - •_ Jl 	 ,L MB
Inlet HC1 (ppmdv)
(@ 7% O7), mean
1,770
1,370
1,740
1,050
1,600
1,770
2,450
1,670
2,230
1,940
iBfea. Li^Lii... a. M m «... ••..! ..^ ,',„!..;., _ ^-™»
Outlet HC1 (ppmdv)
(©7% O2), mean
47.9
40.5
82.4
18.7
108
72.8
67.2
69.5
34.4
40.8
Efficiency concentration-
based, mean
97.5
97.1
95.3
98.3
93.2
95.4
97.0
96.0
98.6
97.9
     The HC1 performance data were examined further graphically
and with a combination of correlation and regression analyses to
                               134

-------
assess the"affects of carbon injection,  stoichiometric ratio,  gas
temperature,  and gas moisture on HC1 performance.  No
relationship was found between performance and either gas
moisture or gas temperature.  However, HC1 performance of this
system appears to be related strongly to stoichiometric ratio and
may be related to carbon injection, although the evidence related
to carbon injection is much weaker than that related to
stoichiometric ratio.  These relationships are depicted
graphically in Figure 36, which shows a plot of HC1 removal
efficiency as a function of stoichiometric ratio with the carbon
injection runs circled.  The HC1 efficiency has a strong
nonlinear relationship to stoichiometric  ratio.  For all runs
having a  stoichiometric ratio  of 6  or more, efficiencies of at
least 97  percent  are  achieved, while  efficiencies for runs with
stoichiometric ratios of less  than 6  range between  88 and
99 percent.   Note that in the  lower regime, the  carbon  runs
showed much higher efficiencies  on average than  did the noncarbon
runs.  However, because performance did not appear  to improve
with increased carbon usage,  the effect of  carbon usage on HC1
performance  is considered  to be inconclusive.
      Metals  emissions data  were obtained for  only  seven of the
 nine test conditions at  Facility A.  Table 9  summarizes the
 removal  efficiencies achieved by the DI/FF system for Cd,  Pb, and
 Hg for these seven conditions.  The removal efficiencies for Cd
 and Pb were 99 percent or greater irrespective of process
 conditions.   However, the Hg removal efficiencies were obviously
 affected by carbon addition.  During the six conditions that
 involved no carbon injection  (Nos. 1, 3, 5, 6, 7,  and  1A),
 essentially no Hg removal was achieved.  At a carbon injection
 rate of  0.45 kg/hr  (1 Ib/hr)  (Test Condition 8), the average
 removal  efficiency was 86 percent, and at a carbon injection  rate
 of  1.13  kg/hr  (2.5  Ib/hr)  (Test Condition 9), an average Hg
 removal  efficiency  of about 95  percent was achieved.   Taken
 together, these  data indicate that the DI/FF system  can
 consistently achieve Cd and Pb  removal  efficiencies  of
                                 135

-------
                                                                               (0
                                                                              I
                                                                              i
                                                                                                           u
                                                                                                           ft
                                                                                                                   o
                                                                                                                  -H
                                                                                                                   J-J
                                                                                                                   fO
                                                                                                  -  CO
                                                                                                           u
                                                                                                           •H
                                                                                                           i-l
                                                                                                           iJ
                                                                                                           
                                                                                                           i— I
                                                                                                           U


                                                                                                           14-1
                                                                                                           O

                                                                                                           iJ
                                                                                                           o
                                                                                                           r-l
                                                                                                           Ol
8
1
odd
        (pesBg
                                                   O

m

 0)
 )-i
 3
 Cn
•r-l
fc.
                                                           136

-------
99 percent,"and with carbon injection,  the system can achieve Hg
removal efficiencies in  excess  of 85 percent.

    TABLE  9.   SUMMARY OF METAL EFFICIENCIES BY  TEST CONDITIONS
                           AT FACILITY A
Test condition
1
3
5
6
7
1A
8b
9°
Cd removal efficiency, %
99.5
99.5
98.8
99.4
99.1
99.6
99.6
98.8
Pb removal efficiency, %
99.5
99.3
99.7
99.6
99.4
99.7
99.7
99.5
Hg removal efficiency, %
9.0
42.7
a
10.7
a
a
86.3
94.7
  aMeasured outlet Hg concentrations were higher than measured inlet Hg concentrations.
  k Activated carbon at 1.0 Ib/hr.
  cActivated carbon at 2.5 Ib/hr.

     The run-specific CDD/CDF inlet and outlet  concentrations for
Facility A  are presented in Appendix B.  Analysis  of these data
suggest  that  in  the absence of carbon injection, the DI/FF system
at Facility A achieves essentially no reduction in CDD/CDF.  The
CDD/CDF  mass  flux data averaged across test  conditions,  which are
presented in  Table 10, provide further insight  into the
performance of the system relative to CDD/CDF emissions.
     Table  10 contains average CDD/CDF mass  flux rates by
conditions  in units of /tg total CDD/CDF per  hour for three
different DI/FF  streams--the postcombustion  gas stream to the DI
unit,  the stack  gas stream exhausted from  the FF (post-APCS), and
the  FF catch discharge.  Table 10 also presents the ratio of
CDD/CDF stack gas mass flux to the postcombustion  CDD/CDF mass
flux and the ratio of the total APCS CDD/CDF discharge rate
 (stack gas  plus  FF catch) to postcombustion  CDD/CDF mass flux.
     For the test conditions involving no  carbon injection
 (Conditions 1,  2, 3,  6, 7, and 1A), the post-APCS  to
postcombustion ratio  ranges from 0.47  to 1.24 with an average of
                                 137

-------
   TABLE 10.   SUMMARY OF CDD/CDF  PERFORMANCE DATA FOR THE DI/FF
                           AT FACILITY A
CDD/CDF mass flux, /xg/hr
Test
condition
lb
2
3
6
7
1AC
8c,d
9c,e
Post-
combustion
759
615
623
745
786
325
591
576
Post-
APCS
943
514
520
353
884
201
23.8
10.2
FF
catch
285
467
892
625
502
245
540
1,140
Post-APCS to
postcombus-
tion ratio
1.24
0.84
0.83
0.47
1.12
0.62
0.040
0.018
Total APCS
discharge to
postcombus-
tion ratioa
1.62
1.60
2.27
1.31
1.76
1.37
0.95
2.00
      of the CDD/CDF measured in the post-APCS and FF catch streams.
  "Data from 1990 test series.
  jData from 1991 test series.
  QActivated carbon at 1.0 Ib/hr.
  Activated carbon at 2.5 Ib/hr.
0.85.  These ratios  are  consistent with essentially zero CDD/CDF
control in the absence of carbon injection.  However, when carbon
was injected at a  rate of 1 Ib/hr, the ratio fell to 0.04 which
is equivalent to a 96 percent reduction in CDD/CDF emissions.
When the carbon injection rate was increased to 2.5 Ib/hr, the
ratio fell to 0.018, which represents a 98.2 percent CDD/CDF
removal efficiency.  On  balance, these data suggest that a DI/FF
system with carbon injection can achieve average CDD/CDF removal
efficiencies of 95 percent or greater.
     While the data  in Table 10 indicate that DI/FF systems with
carbon injection can reduce stack emissions of CDD/CDF, they also
indicate that CDD/CDF formation occurs within the DI/FF system.
With the exception of the 1 Ib/hr carbon injection conditions,
which comprised only two runs with variable results, the ratio of
total CDD/CDF discharged to postcombustion CDD/CDF range from 1.3
to 2.3.  These ratios suggest that at this facility with
                                138

-------
incinerator'APCS inlet levels in the range of 300 to 500 ng/dscm
at 7 percent 02 the amount of CDD/CDF formed in the system can be
50 to 100 percent of that entering with the combustion gas.   To
date, no relationship between this formation rate and DI/FF
operating conditions has been found.
     4.2.6.3  Performance of the SD/FF Control System.  The SD/FF
system at Facility M was tested under two sets of operating
conditions.  During Condition 1, normal facility operating
conditions for both the incinerator and APCS were used.  The same
operating levels were used during Condition 2, except that
activated carbon was added to the lime slurry to produce a carbon
rate of about l.l kg/hr  (2.5 Ib/hr).  In general, most process
parameters were maintained at nearly constant levels for all runs
except stoichiometric ratio.  Because the lime feed rate was
nearly constant, the stoichiometric ratio fluctuated somewhat
with inlet HC1 loadings.  The only major operating difference
occurred during Run 1.  For that run the lime concentration  in-
the  slurry was maintained at 6 percent by weight in contrast to
9 percent by weight for the other  five tests.  Consequently, the
lime feed rate and stoichiometric  ratio were  lower for  Run 1 than
for  the other  five runs.  Because  all other parameters  varied  so
little, the only factor  that was examined in  detail was the
effect of carbon addition.  The paragraphs below provide a
general discussion of the performance of the  system with respect
to  PM, HC1, CDD/CDF,  and metals and discuss  the  effect  of  carbon
addition  on CDD/CDF and metals  performance.
     The  SD/FF system yielded outlet PM  concentrations  that
ranged  from 0.0006  to 0.0098 gr/dscf at  7 percent  02,  and with
the exception  of Run  2,  all  concentrations were  below
0.002 gr/dscf  at 7 percent  02-  Average  outlet PM  concentrations
were 0.0038 for Condition 1  and 0.0011 for Condition  2, and
carbon  appeared to have  no effect  on emissions.
      Because analytical  data were  outside  allowable  quality
control  limits for two  runs, HC1  performance data  are available
for only four  test runs.   These data are shown in  Table 11.
Although the data  are somewhat  limited,  they do  indicate that
                                139

-------
stoichiometric ratio is important  to  the  system's performance.
Additionally, the data indicate  that  if a stoichiometric ratio  of
two or greater is maintained, HC1  control efficiencies of at
least 99 percent were achieved by  the system.

         TABLE 11.  HC1 PERFORMANCE DATA  FOR FACILITY Ma
Run No.
3
4
5
6
Lime rate,
Ib/hr
37
34
31
32
SR
3.39
1.91
2.29
2.11
Activated
carbon
rate,
Ib/hr
0
2.8
2.5
2.5
HC1 concentration,
ppmdv at 7V O?
Inlet
724
1,220
946
1,030
Outlet
4.50
26.1
<0.049
<0.045
Removal
efficiency
99.3
97.8
>99.99
>99.99
  aHCl analytical data for run number 1 and 2 were outside control  limits
  and are not reported.

     The SD/FF performance  data  for metals  and CDD/CDF are
summarized in Table 12.  These data indicate that for Cd and Pb,
the SD/FF is highly efficient  (99.7 percent or greater)  under
both test conditions.  These findings  were  consistent across test
runs within each condition.  However,  for both Hg and CDD/CDF,
the system yielded substantially greater reduction with  activated
carbon  addition than without.  For Hg,  the  average removal
efficiency without activated carbon was about 30 percent with the
efficiencies on individual  runs  ranging from no control  to
50 percent.  When  activated carbon was added not only was the
average removal efficiency  higher (90  percent),  but system
performance also was more consistent with efficiencies for
individual runs ranging  from 84  to 96  percent.   These data show
that an overall Hg removal  efficiency  achieved by a SD/FF system
with activated carbon  injection  is similar  to that achieved by a
DI/FF system with  activated carbon injection.
     With no activated carbon added to the  system, the average
CDD/CDF removal efficiency  was  84 percent with a range of 56 to
95 percent.  As was the  case with Hg,  addition of activated
carbon  both increased  average efficiency and provided more
consistent performance.  The average  removal efficiency  was about
                                140

-------
      TABLE 12.   SUMMARY OF CDD/CDF AND METALS PERFORMANCE
                        FOR SD/FF SYSTEM5
Pollutant
Cd
Pb
Hg
CDD/CDF
CDD/CDF
(balance)
Parameter
postcombustion, jtg/dscm
post-APCS, /ig/dscm
Removal efficiency. %
postcombustion, /ig/dscm
post-APCS, /tg/dscm
Removal efficiency, %
postcombustion, jtg/dscm
post-APCS, /ig/dscm
Removal efficiency, %
postcombustion, jig/dscm
post-APCS, /tg/dscm
Removal efficiency, %
postcombustion, ftg/hr
post-APCS, /ig/hr
SD catch, ;tg/hr
FF catch, fig/hr
post-APCS: postcombustion ratio
Total out:Total in ratio
Condition 1
(no carbon)
471
1.77
99.7
3,590
5.7
99.9
2,590
1,980
29.6
192
32.2
94.0
820
131
12.8
589
0.16
0.89
Condition 2
(with carbon)
492
0.84
•99.8
4,410
2.5
99.9
2,630
284
90.0
199
3.30
98.3
848
14.5
5.27
484
0.017
0.59
98 percent with individual runs having efficiencies in the range
of 95 to 99 percent.  Two additional observations concerning the
CDD/CDF data in Table 12 are worth noting.  First, even without
activated carbon addition a substantial reduction in CDD/CDF
emissions was achieved by the system, a dramatic contrast to the
absence of control found with the DI/FF system.  Second, unlike
the results obtained for the DI/FF system, the mass balance
around the control system shows no evidence of CDD/CDF formation
across the system.  Although no definitive explanation was found
for why CDD/CDF formation occurred in the DI/FF system but not in
the SD/FF system, the system characteristics that may have the
greatest effect are the quick cooling/adsorption obtained with
the slurry spray  in the SD/FF, the longer gas/lime residence time
prior to the FF provided by the SD vessel, and the lower FF
temperatures of the SD/FF system  (146°C  [295°F])  compared to the
DI/FF system  (163°C  [325°F]).  In addition, it should be noted
                               141

-------
that the MWI preceding the SD/FF system is a rotary kiln design
with significantly higher post-combustion levels of PM.  This
increased level of PM would indicate that significantly more
carbon existed in the gas stream entering the SD/FF than in the
gas stream entering the DI/FF, which may have aided the
performance of the SD/FF.  The results from these tests indicate
that an overall CDD/CDF removal efficiency of at least 98 percent
can be achieved by this SD/FF system with activated carbon
injection.
4.3  DEMONSTRATED EMISSION CONTROL LEVELS
     The purpose of this section is to present demonstrated
emission control levels  (i.e., emission limits) associated with
each of the emission control techniques discussed earlier.  Each
of nine pollutants is presented in a separate subsection.  Each
subsection includes tables showing the emission limits and
figures presenting the data used to establish these limits.
     The emission limits were developed from actual test data
from EPA-sponsored emission tests conducted at seven MWl's.
Table 13 presents a summary of information on the type and size
of each MWI tested.  Other data were also considered  (e.g.,
emission test reports submitted to State agencies), but none of
these data were used in establishing emission limits because the
test reports were incomplete  (i.e., lacked process or design
information and/or lacked information on sampling techniques).
     In establishing the emission limits for all pollutants, the
amount of data available and variation in that data were taken
into consideration.  Except for HC1 controlled by DI/FF systems,
the emission limits were set above the highest 3-run average
shown by the data.  For HC1 controlled by DI/FF systems, the
emission limit is set above the highest individual runs with SR's
above 6:1.  For pollutants and/or MWI types where limited test
data are available, appropriate data from related pollutants
and/or MWI types were considered.
     The numerical emission limit only has meaning when coupled
with an averaging time.  During each of the tests conducted at
the seven MWI facilities, emissions were measured over three
                               142

-------
                TABLE 13.  TESTED MWI FACILITIES
Facility
A
B
J
K
M
S
w
Description
650 Ib/hr, intermittent, ram-fed; 2-sec
secondary chamber; DI/FF system tested
activated carbon injection
1,500 Ib/hr, continuous, ram-fed; 2-sec
secondary chamber; VS/PB system
750 Ib/batch, batch, manually fed; 1.75
in secondary chamber; FF/PB system
300 Ib/hr, intermittent, manually fed;
time in secondary chamber
residence time in
with and without
residence time in
-sec residence time
0.33 -sec residence
800 Ib/hr, continuous, ram-fed; 2-sec residence time in
secondary chamber; SD/FF system tested with and without
activated carbon injection
250 Ib/hr, intermittent, manually fed; 0. 2-sec residence
time in secondary chamber; conditions 1 and 2 =
pathological waste, condition 3 = mixed medical waste
300 Ib/hr, intermittent, ' ram- fed; 1-sec residence time in
secondary chamber
4-hour periods.  Therefore, the emission limits are based on a
12-hour average consisting of three 4-hour test runs.
     Continuous, intermittent, batch, and pathological MWI's are
four different MWI types  (based on their physical design
characteristics, operating characteristics, and overall emission
profiles).   However, for continuous, intermittent, and batch
units, there is a period in the combustion cycle when the
emission profiles are similar.  In continuous and intermittent
units, this period occurs- during waste-charging.  In a batch
unit, this period occurs in what is sometimes referred to as the
"high-air" or  "burndown" phase.  In all three MWI types, this
period can be  distinguished by the temperature in the primary
chamber.  This is also the period of highest emissions.
Consequently,  data taken during this period from these three MWI
types have been combined  in establishing one set of emission
limits for continuous, intermittent, and batch MWI's.  There is
no  corresponding period in a  pathological  unit during which
emissions are  comparable  to the other types of units.  Therefore,
data from only pathological units have been used  to  set  emission
limits for pathological MWI's.
                               143

-------
     The nine pollutants discussed in this section fall into two
categories:   those dependent upon the composition of the waste
being burned and those dependent upon the MWI combustion process.
Waste-related pollutants (HC1, S02, Cd, Pb, and Hg) are formed as
a result of the presence of components within the waste and are
unaffected by combustion controls.  Combustion-related pollutants
(PM, CO, and CDD/CDF) are emitted at different levels from
different MWI's depending on the waste-charging patterns, the
temperature maintained in the secondary chamber, and the gas
residence time in the secondary chamber.  Nitrogen oxides can be
considered both waste-related and combustion-related.  However,
neither the combustion controls nor the add-on controls evaluated
in this analysis achieved any NQX reduction  (see Section 4.3.9).
     Because combustion conditions do not affect waste-related
pollutants,  applying only combustion control options result in
uncontrolled emission levels of these pollutants.  As a result,
combustion control limits have not been established for these
pollutants.
     Establishing emission limits for the combustion-related
pollutants was difficult because not all possible combinations of
combustion control and MWI type were tested.  Of the four MWI
types, only intermittent units were tested at each of the three
combustion control levels  (see Table 14).

         TABLE 14.   COMBUSTION CONTROL DATA vs.  MWI TYPES

Continuous
Intermittent
Batch
Pathologies]
Uncontrolled

K, S (condition 3)

S (conditions 1 and 2)
1-sec

W


2-sec
B, M
A
J

     For continuous, intermittent, and batch MWI's, data from
facility K  (conditions 1 and 2) and facility S  (condition 3)
represent uncontrolled emission levels; data from facility W
 (conditions 1 and 2) were used to establish the emission limits
                               144

-------
for l-sec combustion;  and data from facilities A,  B,  J
(condition 2),  and M were used to establish the emission limits
for 2-sec combustion.   Data from test condition 3  at facilities K
and W were not considered in establishing emission limits based
on combustion control because both MWI's were purposely
overcharged during these test conditions to show the effects of
poor MWI operation.
     For pathological MWI's, only one level of combustion control
was tested  (uncontrolled at facility S) .  To establish emission
limits for pathological MWI's at each combustion control level,
the relationships  of the emission limits established  for the
other types of MWI's were applied to the data  for the
pathological MWI's to  calculate  emission limits for  the untested
conditions.
     'control options  reflecting  add-on controls combine  2-sec
combustion  with  an add-on  control  device.   Emission limits  for
these  control  options were  established based on inlet/outlet
tests  conducted  on MWI's with add-on control.   Typically,  the
emission reduction capability is expressed as a percent reduction
 relative to the  2-sec combustion emission limit.   Exceptions to
 this approach are PM, Pd,  and Cd emissions from fabric filter
 systems, which are capable of achieving constant  outlet levels
 for these pollutants.  For pollutants that are unaffected by
 combustion controls  (waste -related pollutants for all MWI's and
 PM for pathological MWI's) the 2-sec combustion emission limit  is
 equivalent to the uncontrolled emission level  (the highest 3-run
 average of uncontrolled emission data) .
      The remainder of this section is  divided into nine
 subsections,  one  for  each pollutant.   These subsections present
 the emission  limits  for each MWI type  and  show the  data used to
 determine  the emission  limits.
 4.3.1   Pfrrtieulate Matter
       Table 15 presents the PM emission limits under each control
  option for continuous, intermittent,  batch,  and pathological
  MWI's.  Figures 37 through 39 present the data used to develop
                                 145

-------
these emission  limits.   At the facilities  tested,  the PM
emissions were  reported in gr/dscf.  However,  to be  consistent
with emission limits for other pollutants,  the PM  emission limits
in this  section are expressed in metric units  (milligrams per dry
standard cubic  meter [mg/dscm]).

   TABLE 15.  PM EMISSION LIMITS FOR CONTINUOUS, INTERMITTENT,
                   BATCH, AND PATHOLOGICAL MWI'S
Control option
1-scc
2-sec
Wet systemsb
FF/PBb
DI/FFb
SD/FFb
Fabric filter systems with carbon
injection13
PM emission limits,
mg/dscm at 7 percent 02
Continuous, intermittent, batch
600
• 300
150
30
30
30
30
Pathological
NAa
• NAa
58
30
30
30
30
  aNo applicable limit. Highest 3-run average of uncontrolled
  data =115 mg/dscm (facility S condition 2).
        2-sec combustion.
     4.3.1.1   Combustion Controls--PM.  Figure 37 presents  PM
emission data  for  continuous,  intermittent, and batch MWI's
burning general medical waste.  Data from facilities K and  S
reflect uncontrolled PM emission levels.  The PM emission limit
for 1-sec combustion is based on facility W, and was set
relatively close to the majority of individual test runs because
1-sec combustion will exhibit less variation in PM emissions  than
uncontrolled levels.   Test data show that at higher residence
time, combustion-related emissions are less affected by the MWI
operation.  The PM emission limit for 2-sec combustion is based
on facilities  A, B,  and J.  In this case, the emission limit  has
been set very  close to the highest 3-run average and very close
to most of the individual test runs because the large amount  of
                                146

-------
 2200-
  2100 -;
  2000:
  1900:
  1800:
  1700:
  1600:
  1500:
  1400^
S 1300:
   1100 -i
900:
800:
700:
600:
500:
400^   *
300 :~
200^*
 100 ^
      0-
              *
     K  K  S
     1  2  3
      Uncontrolled
  Figure  37.
                        W  W
                         1   2
                       1-sec comb.
AAAAAAAAABJJ
134567189112
                    a
           2-sec combustion
                  PM emissions for continuous,  intermittent, and batch
                       MWI's with  combustion controls.
                                      147

-------
data available for these units  indicate  that  this limit can
consistently be achieved.  Facility M was not considered in
developing the PM limit for 2-sec  combustion because the unit at
this facility is a rotary kiln  MWI that, as expected, had higher
PM emissions due to the turbulence of the waste in the primary
chamber.  This turbulence results  in entrainment of non-
combustible ash, which will not be affected by combustion
control.
     Figure 38 presents PM emission data for pathological MWI's
(facility S, uncontrolled) .  Compared to PM emissions from an MWI
firing general medical waste, PM emissions from a pathological
MWI will be lower, more stable, and less affected by combustion
control.  Pathological MWI's operate with excess air in the
primary chamber, which limits the  generation of incomplete
combustion products but increases  the entrainment of ash.  The
result is a higher fraction of  non-combustible material (not
affected by combustion control) in the PM emissions.  Emissions
of PM from pathological MWI's will be very stable because of the
homogeneous nature of pathological waste and the absence of
incomplete combustion products.  Because combustion controls do
not reduce PM emissions from pathological MWI's,  no limits were
established for the 1-sec and 2-sec control options.
     4.3.1.2  Add-On Controls--PM.  Figure 39 presents PM removal
efficiencies and outlet levels  achieved with the add-on control
technologies tested.  The PM outlet concentrations achievable
with wet systems depend on the  inlet loadings and therefore are
based on a percent reduction rather than a constant outlet value.
Facility B data indicate that wet  systems are capable of
achieving a PM removal efficiency  of at least 50  percent.   This
performance is applicable to all MWI types.
     The remaining control technologies use a fabric filter to
control PM emissions.  The emission limit for these technologies
is based on a constant outlet value achievable with a fabric
filter as demonstrated at facilities A,  J,  and M.   Test
conditions l through 7 at facility A were not considered in
establishing the PM emission limit for fabric filter systems.
                               148

-------
               23M-"
               2100-
               2000-
               1900-
               1700-
               1500-
               1400-
               1300-
               1200-
             JIOOO
             I  900
                700-
                500-
                400^
                300
                200-
                100^
4
                              SI             82
                                  lU.ii.Jptiliirl
                                  unconBonu
Figure 38.   PM  emissions  for uncontrolled pathological MWI's,
                                  149

-------
 I"
 I-
                           2100-
                           2000
                           1900-
                           1700-
1-
£| !
S*
r
1500
1400
1900
?sr1200
o
g tO
©1000
1 900
|>800
nr Tnft
                           600
                           500
                           400
                           300
 I-
                            0-
Ala
Jl
J2
                                                   Ml
                                                                       M2
Figure  39.
                               ftWcFtor Systems
PM emissions for  continuous,  intermittent,  batch,  and
pathological  MWI's with add-on controls.
                     150

-------
Outlet PM concentrations varied significantly for these tests,
which were obtained during the first series of tests at
facility A.  Prior to the second series of tests, the fabric
filter bags were replaced.  The consistently low outlet PM levels
achieved during this second series of tests, along with data  from
facilities J and M, indicate that the fabric filter was not
functioning properly during the first series of tests at
facility A, which invalidates PM runs 1 through 7 at facility A.
Based on facility A  (conditions 1A, 8, and 9), and on facilities
J and M, PM concentrations of 30 mg/dscm are achievable by fabric
filter systems.
4.3.2  Carbon Monoxide
     Table 16 presents  the CO emission limits under  each  control
option for continuous,  intermittent, batch, and pathological
MWI's.  Figures 40 and  41 present the data used  to develop  these
emission limits.

   TABLE 16.  CO EMISSION LIMITS FOR CONTINUOUS,  INTERMITTENT,
                  BATCH, AND  PATHOLOGICAL MWI'S

Control option
1-sec
2-sec
Wet systems3
FF/PBa
DI/FF8
SD/FFa
Fabric filter systems with carbon
injection*
CO emission limits, ppmdv at 7 percent O2
Continuous, intermittent, batch
500
50
50
50
50
50
50
Pathological
50
10
10
10
10
10
10
  alncludes 2-sec combustion.

      4.3.2.1  Combuation Cnntrols--CO.  Figure 40 presents CO
 emission data for continuous, intermittent, and batch MWI's
 burning general medical waste.  Data from facilities K and S
 reflect uncontrolled CO emission levels.  The CO emission limit
                                151

-------
  2800-~


  2600-


  2400-


  2200-


  2000-


  1800-


0 1600-


  1400-


  1200-


  1000-


   800-
•5*
a
8
   600-



   400-
       4
       i

   200^
          T
                        1

                        *
                                *******
                                                                     *  *
       K  K  S
       1  2  3

        Uncontrolled

      Figure 40.
                     W  W
                     1  2
              AAAAAAAAAABJ  JMM
              123456718911212
                                  a
                             2-sec combustion

CO emissions  for continuous, intermittent,  and batch
     MWI's iwth combustion controls.
                                      152

-------
for l-sec combustion is based on facility W,  and was set
relatively close to the majority of individual test runs,  because
l-sec combustion will exhibit less variation in CO emissions than
uncontrolled levels  (as previously discussed for PM emissions).
The CO emission limit for 2-sec combustion is based on
facilities A, B, J, and M.  In this case, the emission limit has
been set very close to the highest 3-run average and very close
to each of the individual test runs because the large amount of
data available for these units indicate that this limit can-
consistently be achieved.
     Figure 41 presents CO emission data for pathological MWI's
(facility S, uncontrolled).  Compared to CO emissions from an MWI
firing general medical waste, CO emissions from a pathological
MWI will be lower and more stable, but improved combustion
control will further reduce CO emissions.  Pathological MWI's
operate with excess  air in the primary chamber, which limits the
generation of incomplete  combustion products.  However,
combustion control will further reduce these  emissions.
Emissions of CO from pathological MWI's  will  be very stable
because of the homogeneous nature of pathological  waste.  The CO
limit for l-sec combustion control  is calculated using  the  ratio
of the l-sec CO limit  to  the uncontrolled  CO  emission  level  for
general-waste units  (500/1,200) and applying  this  ratio to  the
uncontrolled CO emission  level  for  pathological MWI's  (130).
Uncontrolled emission  levels are  represented  by  the highest  3-run
averages  of  the uncontrolled emissions data  (facility  S
condition 3  for continuous,  intermittent,  and batch MWI's and
facility S  condition 2 for pathological  MWI's).   The CO limit  for
2-sec combustion control  is  calculated using  the ratio of the
2-sec CO limit  to the uncontrolled CO emission level for general-
waste units (50/1,200) and applying this ratio to the
uncontrolled CO level for pathological MWI's  (130).
      4.3.2.2  Add-On Controls--CO.   Add-on control devices  are
 not effective in further removing CO emissions.   As a result,  the
 limits for all add-on control techniques are the same as those of
 2-sec combustion control.
                                153

-------
              2400

              2200

              2000

              1800 -;

              1600-
           81000
                         4-
                          SI            S2
                              Uncortrokd
Figure  41.  CO  emissions for uncontrolled pathological MWI's,
                                 154

-------
4.3.3  ninx'ins  and Furans
     Table  17 presents the  CDD/CDF emission  limits under each
control  option  for continuous,  intermittent,  batch,  and
pathological MWI's.  Figures  42 through 44 present the data used
to develop  these emission  limits.
        TABLE  17.   CDD/CDF EMISSION LIMITS FOR CONTINUOUS,
           TNTERMITTENT, BATCH, AND PATHOLOGICAL MWI'S
                                CDD/CDF emission limits, ng/dscm at 7 percent
                               Continuous, intermittent, batch
    Fabric filter systems with carbon
   alncludes 2-sec combustion.
   bLamits reflect CDD/CDF generation across FF/PB system.
   cThe CDD/CDF emission limits achievable with a FF/PB system
   with activated carbon injection are not known.

      4.3.3.1   rnmhnstion  rv^i-rols- -CDD/CDF.   Figure 42 presents
 CDD/CDF emission data  for continuous, intermittent, and batch
 MWI's  burning general  medical waste.  Data  from facilities  K and
 S reflect uncontrolled PM emission levels.   The CDD/CDF emission
 limit  for l-sec combustion is based  on  facility W, and was  set
 relatively close to the majority of  the individual test runs
 because l-sec  combustion will exhibit less variation  in CDD/CDF
 emissions at uncontrolled levels  (as discussed earlier for PM and
 CO).   The CDD/CDF  emission limit  for 2-sec. combustion is  based on
 facilities A,  R, J, and M.  In this  case, the emission limit has
 been set very  close to the highest 3-run average  and  very close
 to each of the individual test runs  because the large amount of
                                   155

-------
  45000-"
  40000-
  35000-
  30000-
S- 25000-
&
I
o
  20000-
   15000-
   10000-   *

           *

        _*
   5000-7
      0-
                                                 : * *
        K  K  S
        1  2  3

         Uncontrolled
     W  W
     1  2

      1-see comb.
AAAAAAAAABJJMM
12346718911212
                 a
            2-sec combustion
     Figure 42.
CDD/CDF emissions for continuous,  intermittent, and
 batch MWI's  with combustion controls.

                   156

-------
data available for these units indicate that this limit can
consistently be achieved.
     Figure 43 presents CDD/CDF emission data for pathological
MWi's  (facility S, uncontrolled).  Compared to CDD/CDF emissions
from an MWI firing geneal medical waste, CDD/CDF emissions from a
pathological MWI will be lower and more stable, but improved
combustion control will reduce CDD/CDF emissions.  Pathological
MWI's  operate with excess air in the primary chamber, which
limits the generation of incomplete combustion products.   .
However, CDD/CDF  emissions can be further reduced by combustion
control.  Emissions  of  CDD/CDF from pathological MWI's will be
very stable because  of  the homogeneous  nature  of pathological
waste   The CDD/CDF  limit for 1-sec combustion control  is
calculated using  the ratio of the 1-sec CDD/CDF limit  to  the
uncontrolled  CDD/CDF emission level for general-waste  units
 (7,000/25,000)  and applying  this ratio to  the  uncontrolled
CDD/CDF  emission  level  for pathological MWI's  (910).
Uncontrolled  emission levels are represented by the highest  3-run
averages of  the uncontrolled emissions data (facility S condition
 3 for continuous, intermittent,  and batch MWI's and facility S
 condition 1 for pathological MWI's).   The CDD/CDF limit for 2-sec
 combustion control is calculated using the ratio of the 2-sec
 CDD/CDF limit to the uncontrolled CDD/CDF emission.level for
 general-waste units  (1,500/25,000)  and applying this ratio to the
 uncontrolled CDD/CDF emissions for pathological MWI's  (910).
       4.3.3.2  Arid-on Consols- -CDD/CDF.  Figure 44 presents the  .
 CDD/CDF removal  efficiencies and outlet concentrations of the
 add-on control technologies  tested.  The CDD/CDF emission limit
  for the wet  systems is based on the 70-percent  removal efficiency
  demonstrated at  facility B.  The three fabric filter  systems
  without activated carbon injection achieved varying levels of
  CDD/CDF control.  In fact,  formation  of CDD/CDF occurred in  one
  system.
       Formation of CDD/CDF occurs when there is intimate  contact
  between a gas stream containing CDD/CDF precursors and fly  ash,
  which acts as a  catalyst.   The optimum temperature window for fly
                                 157

-------
               45000
               40000
               35000
               30000
               25000
             $20000-
            g 15000-
            8
               10000-
               5MO
                 0-
                                         82
                                Urratofed
Figure 43.
CDD/CDF emissions for uncontrolled pathological
                MWI's.

                  158

-------
1-
•Si'
1
•o
$
C .
!;
w
n
I-
8-
t

g":
IE :
® •
i :
1 ':
U.5B-
j)
8 ;
^B
i 1
t |






' i
i

                                                   I
                                                      K    .  IB
     Fiqure 44     CDD/CDF emissions for continuous,  intermittent,
         batch, and pathological MWI's with add-on controls.
                                 159

-------
 ash catalyzed CDD/CDF formation is between 300°F and 600°F.   The
 CDD/CDF formation is minimized when using combustion control or
 wet systems  because these options provide: (1)  rapid cooling of
 the gas stream through the temperature window and (2)  quick
 dispersion (or removal in the case of wet systems)  of  CDD/CDF
 precursors and fly ash.   In DI/FF and SD/FF systems,  the  presence
 of  an acid gas sorbent (lime,  for example) also limits the
 formation  of CDD/CDF.  • The fabric filter in a FF/PB- system,  on
 the other  hand,  can provide those conditions conducive to CDD/CDF
 formation.   In fact,  test data have shown CDD/CDF formation  in a
 FF/PB system.
      The emission limit  for the FF/PB system is based  on  test
 data from  facility J,  which shows a generation  of CDD/CDF across
 the system.   Figure 44 shows the outlet CDD/CDF emissions at
 facility J.   Because  facility J burns general medical  waste,  the
 actual  test  data  could not be  used directly in  establishing  an
 emission limit for pathological MWI's.   As a result, the  limit
 for the pathological MWI  has been calculated based  on  the CDD/CDF
 generation rate at facility J.   The generation  rate  is based on
 the FF/PB  outlet  limit (7,000  nanograms [ng]/dry standard cubic
 meter  [dscm])  and the  2-sec combustion  limit  (1,500  ng/dscm)  at
 facility J.   This generation rate was applied to the 2-sec
 combustion limit  for pathological MWI's to determine the  FF/PB
 limit for  this MWI.
     The CDD/CDF  emission limit for the DI/FF system is based  on
 test data  from facility A.   This  test data showed no consistent
 CDD/CDF removal from the  DI/FF  system without carbon injection.
 Therefore, the limit for  this  technology is the  same as that of
 2-sec combustion  control.
     Test data from facility M  indicate  that a SD/FF system on a
 rotary kiln MWI is  capable  of achieving  80-percent CDD/CDF
 reduction.  However, this significant reduction may be due to the
 rotary kiln design  (the higher  PM emissions may aid in the
 removal of CDD/CDF).   It  is  not known if a SD/FF system (without
activated carbon  injection)  can achieve this same level of
 control on the other MWI designs.  On the other hand, a SD/FF
                                160

-------
system is not expected to generate CDD/CDF for the reasons
described above.   Therefore,  the limit for this technology is the
same as that of 2-sec combustion control.
     Test data from facilities A and M indicate that the DI/FF
and SD/FF systems with activated carbon injection are effective
in significantly reducing CDD/CDF emissions (see Figure 44).
These two fabric filter systems (with activated carbon injection)
are able to achieve at least 95-percent CDD/CDF control.  Because
data are not available from a FF/PB system with activated carbon
injection, and because the FF/PB system tested without carbon
injection generated CDD/CDF, it is not known what CDD/CDF
reductions are achievable with the activated carbon in a FF/PB
system.
4.3.4  Hydrogen  Chloride
     Table 18 presents the HC1 emission limits under each control
option for continuous, intermittent,  batch, and pathological
MWI's.   Figures  45 through 47 present the  data used to  establish
these emission limits.
     4.3.4.1  combustion  Controls--HC1.   Figure  45  presents HC1
emission data  for  continuous,  intermittent, and  batch MWI's
burning  general  medical  waste,  and Figure 46  presents  these data
for pathological MWI's.   Because  emissions of HC1 are  waste-
related,  combustion  controls do not reduce HC1 emissions.
      4.3.4.2  Add-on Controls--HC1.   Figure 47 presents the HC1
 removal  efficiencies of  the add-on control technologies tested.
The emission limit for wet  systems is based on test data from
 facility B,  while the emission limit for fabric filter systems is
 based on data for a DI/FF system from facility A and a SD/FF
 system from facility M.   The data show that each of these systems
 can reduce HC1 emissions by at least 97 percent.  This
 corresponds to an emission level of 42 ppmdv for a typical
 continuous, intermittent, or batch unit  (97 percent reduction
 from a typical uncontrolled level of  1,400 ppmdv); and 4 ppmdv
 for pathological units  (97 percent reduction from a typical
 uncontrolled level of 120 ppmdv).
                                161

-------
  TABLE  18.   HC1 EMISSION LIMITS  FOR  CONTINUOUS,  INTERMITTENT,
                  BATCH,  AND PATHOLOGICAL MWI' S
Control option
1-sec
2-sec
Wet systems**
FF/PBb
DI/FFb
SD/FFb
Fabric filter systems with carbon
injection*5
HC1 emission limits, ppmdv at 7 percent O2
Continuous, intermittent, batch
NAa
NAa
42C
42C
42C
42C
42C '
Pathological
NAa
NAa
4C
4C
4C
4C
4C
  *No applicable limit.
  ^Includes 2-sec combustion.
  cThese emission limits correspond to a reduction of 97 percent.
     For the DI/FF  system without a retention chamber, as  at
facility A, greater than 97 percent reduction is achieved  when
the stoichiometric  ratio is at least 6:1.   (In some cases,
similar reductions  can also be achieved with lower stoichiometric
ratios).  The 3-run test averages at facility A were not used in
establishing the HC1 emission limit because, as shown on
Figure 47, the stoichiometric ratio was below 6:1 for at least
1 run at each of the test conditions, except test condition 4.
Test data for a DI/FF with a retention chamber are unavailable.
However, as noted earlier in this report, vendors indicate that  a
50 percent lower stoichiometric ratio can be used in a system
with a retention qhamber to achieve the same reduction as  a
system without a retention chamber.
     For the SD/FF  system at facility M, test runs were conducted
with stoichiometric ratios between 1.9 and 3.4, and the resulting
emission reductions were greater than 98 percent for each  run.
                                162

-------
  2800 -~
  2600-
  2400-
  2200-
  2000-
  1800-
1600-
                                    *

                                    *
  1400-
r 1200-
o

  1000-1
   800-
   600
   400
                                        *
                                     *
                                     *    *
    200-
     0-
        K K  K  S
        1233

        Uncontrolled
                                                                *
                                                                ¥
                     WWW
                     1  2  3

                      1-sec comb.
AAAAAAAAABJ  JMM
12345618911212
                a
           2-sec combustion
   Figure 45.   HC1  emissions for continuous,  intermittent,  and batcn
                     MWI's with combustion controls.

                                   163

-------
               2400-




               2200




               2000
                1600-
                1400-
             S.
                1200



                iooo
                400
                                        T
                  0-
                            SI
S2
Figure 46.  HC1 emissions for uncontrolled pathological MWI's,



                                 164

-------
  100
   90-
             t-s
                           -o-

                           0
                         tx   8

                         o
8
   80-
   70-
§  60
   50
•  Data used to establish emission limit:
   - -  DI/FF runs with SR >6
      All wet system runs
   - -  All SD/FF runs

O  Data not used to establish  emission limits;
      DI/FF runs with SR <6
   40-
   30-
   20-
   10-
    0-
        B1
          A1  A2  A3  A4  A5  A6  A7  Ala AS  A9
                                                                   M1  M2
      Wet Systems
                       DI/FF Systems
               SD/FF Systems
       Figure 47.  HC1 emissions for continuous,  intermittent,  batch,
                and pathological MWI's with  add-on controls.

                                      165

-------
The SD is a retention chamber,  which allows the stoichiometric
ratio to be lower at Falicity M than at Facility A.
     The HC1 removal efficiencies at facility J were relatively
low and were also not considered in establishing the HC1 emission
limits for the fabric filter systems.   During the tests conducted
at this facility, nozzle plugging resulted in poor performance,
thereby invalidating these  tests.
4.3.5  Sulfur Dioxide
     Table 19 presents  the  SO2  emission limits under each control
option for continuous,  intermittent, batch, and pathological
MWI's.  Figures 48 and  49 present the data used to establish
these emission limits.

   TABLE 19.   SO2 EMISSION LIMITS  FOR CONTINUOUS,  INTERMITTENT,
                  BATCH, AND PATHOLOGICAL MWI'S
Control option
1-sec
2-sec
Wet systems15
FF/PBb
DI/FFb
SD/FFb
Fabric filter systems with carbon
injection"
SO? emission limits, ppmdv at 7 percent O%
Continuous, intermittent, batch
NAa
NAa
NAa
NAa
NAa
NAa
NAa
Pathological
NAa
NAa
NAa
NAa
NAa
NAa
NAa
  aNo applicable limit.
  "Includes 2-sec combustion.
      4.3.5.1   Combustion Controls- -.SO^•   Figure 48 presents S02
 emission data for continuous,  intermittent, and batch MWI's
 burning general medical waste, and Figure 49 presents these data
 for pathological MWI's.  Because emissions of S02 are waste-
 related,  combustion controls do not reduce SO2-
      4.3.5.2   Add-On Controls- -SO.-..  At the low inlet S02 levels
 associated with MWI's,  test data indicate that add-on controls
                                166

-------
 150 -~~



 140-



 130 :



  120 :



  110-



  100-
sr 90
o
70-



60-



50-



40-



30:      *



20-

      *  *

   ~ * *

   "» *  *
   I*
                                                                  *
                                                                  *
                      *  * *
                                      *
        *

        *
                                                                        *
                                        *
                                        *
                                                *  * -
                                                        *  *
       K K  K  S
       1233


        Uncontrolled
                   WWW
                    1  2  3

                    1-sec comb.
AAAAAAAAAABJ JMM
123456718911212

                  a
              2-sec combustion
    Figure 48.  SO, emissions  for continuous, intermittent,  and batch
                     MWI's with combustion controls.
                                    167

-------
150
140
130
120
no
100
90
^80
j*
^*J5-
|GO
s
"50
30
20
10-

1
t
*

*




;
                0-
                          Sl           S2
                             Urartofed
Figure 49.   S02 emissions for uncontrolled pathological  MWI's
                             168

-------
are not  effective in  reducing these  emissions.  Therefore, S02
emission limits have  not been established for these remaining
control  technologies.
4.3.6  Lead
     Table 20 presents the Pb emission limits under each  control
option for continuous, intermittent, batch, and pathological
MWI's.  Figures 50  through 52 present the data  used to develop
the  emission limits.
    TABLE 20    Pb EMISSION LIMITS FOR CONTINUOUS,  INTERMITTENT,
    i^iouj-i *-w.       	  _.__  —-.nTTrvT r\i~<-rr
-------
  8
s

       *  *  *
          *
  0
                                                                 *
                                                                 *
                                   I ±
                    a

                    *
                                                            *

                                                            *
                   *     *

                      *
                                                    *

                                                    *
                                          *

                                          f
     K K  K  S
     1233

     Uncontrolled
WWW
1  2  3

 1-sec comb.
AAAAAAAAABJJMM
13456718911212
                 a
           2-sec combustion
   Figure  50.   Pb emissions for continuous,  intermittent,  and batch
                    MWI's  with combustion controls.
                                   170

-------
                8-
                7-
              54
                2-
                1-
                0-
                          SI           S2
                             UrxoHrated
Figure 51.  Pb emissions  for uncontrolled pathological MWI's,
                              171

-------
I
         IW WH^V
                    9"
                    8-
                    7-
'8 i
    |
                  15
                     24
                       M   A3   A5
                                        Jl    J2   Ml   M2
   Figure 52.  Pb emissions for continuous,  intermittent,  batch,  and
               pathological MWI's with add-on controls.
                                  172

-------
remaining control technologies use  a fabric filter to  control Pb
emissions.  As  is the case for PM,  the Pb emission limit is based
on a constant outlet achievable with a fabric filter.   Test data
from facilities A,  J, and M indicate that fabric filter systems
can achieve a Pb level of 0.10 mg/dscm.
4.3.7  Cadmium
     Table 21 presents the Cd emission limits under  each control
option for continuous, intermittent, batch, and pathological
MWI's.   Figures 53 through 55 present the data used  to develop
the emission  limits.

   TABLE 21.  Cd EMISSION LIMITS  FOR CONTINUOUS, INTERMITTENT,
                   BATCH, AND PATHOLOGICAL MWI'S
Control option
1-sec
2-sec
Wet systems0
FF/PBC
DI/FF0
SD/FF0
Fabric filter systems with carbon
injection0
Cd emission limits, mg/dscm at 7 percent O2
Continuous, intermittent, batch
NAa
NAa
1.1
0.05
0.05
0.05
0.05
Pathological
NAb
NAb
0.10
0.05
0.05
0.05
0.05
  aNo applicable limit. Highest 3-run average = 1.9 mg/dscm (facility A condition 8).
  bNo applicable limit. Highest 3-run average = 0.17 mg/dscm (facility S condition 1).
  °Includes 2-sec combustion.
      4.3.7.1  Combustion Controls--Cd.  Figure 53 presents the  Cd
 emission data for continuous,  intermittent,  and batch MWI's
 burning general medical  waste, and Figure  54 presents the data
 for pathological MWI's.   Because emissions of Cd are waste-
 related, combustion  controls do not reduce Cd emissions.
                                 173

-------
3.6-
3.4-
3.2-
3.0-
2.8-
2.6-
2.4-
£22.
^ 2.0-
3.
•§• 1.8-
§ 1.6-
P ^
S 1.4;
1.2-
1.0-
0.8-
i
0.6-
0.44
*
H
0.2-
1
!
0.0^

















fi 4
JL *
* 4
* 1 - * i *
* * T
*
KKKS WWW AAAAAA
1233 123 134567
*




•
i
i

-!-

i
j
I
T
t

	

T :* ~
t 1 ^r * *
*
Jt
4
*
AAABJJMM
18911212
    Uncontrolled
1-sec comb.
     a
2-sec combustion
Figure 53.  Cd  emissions for continuous, intermittent,  and batch
                 MWI's with  combustion controls.
                                174

-------
             3.6
             3.2;
             3.0-
              1.6:
              1.4-
              \2-
              1.0-
              0.8 ;
              0.6 -
              0.4
              02
              0.0
                           SI            82
                               Unoonboled
Figure 54.   Cd emissions  for uncontrolled pathological  MWI's.
                                 175

-------
     4.3.7.2  Add-On Controls--Cd.  Figure 55 presents Cd removal
efficiencies and outlet  levels  achieved with the add-on control
technologies tested.  The  limit for wet systems is based on test
data from facility B, which indicate that wet systems are capable
of achieving at least 40 percent reduction in Cd emissions.  The
remaining control technologies  use a fabric filter to control Cd
emissions.  As is the case for  PM and Pb, the Cd emission limit
is based on a constant outlet achievable with a fabric filter.
Test data from facilities  A,  J, and M indicate that fabric filter
systems can achieve a Cd level  of 0.05 mg/dscm.
4.3.8  Mercury
     Table 22 presents the Hg emission limits under each control
option for continuous, intermittent, batch, and pathological
MWI's.  Figures 56 through 58 present the data used to develop
the emission limits.

   TABLE 22.  Hg EMISSION  LIMITS FOR CONTINUOUS, INTERMITTENT,
                  BATCH, AND PATHOLOGICAL MWI'S
Control option
1-sec
2-sec
Wet systems'3
FF/PBb
DI/FFb
SD/FFb
Fabric filter systems with carbon
injection
Hg emission limits, mg/dscm at 7 percent O2
Continuous, intermittent, batch
NAa
NAa
NAa
NAa
NAa
NAa
0.47C
Pathological
NAa
NAa
NAa
NAa
NAa
NAa
0.01°
  *No applicable limit.
  ^Includes 2-sec combustion.
  cThis emission limit corresponds to an 85 percent reduction.
      4.3.8.1  Combustion ContrplsL-j^Hg.  Figure 56 presents the Hg
emission data for continuous, intermittent, and batch MWI's
burning general medical waste, and Figure 57 presents this data
for pathological MWI's.  Because emissions of Hg are  waste-
related,  combustion controls do not reduce Hg emissions.
                                176

-------
I
  si
                     3,6"
                     3.4;
                     32;
                     3,0;
                     2,s;
                     2.6;
                     2.4;
                     22;
                   sr>°i
                   O  i
                   t*\
                     1.41
                 5
                 *UM
                     0.4-
                     02
                       A1A3A5/16A7A1a
                                                           J1J2MlM2
Figure 55.
              Cd emissions for continuous,  intermittent,  batch, and
               pathological MWI's with add-on  controls.
                                  177

-------
26 ~
24-
22-
20-
18-
8
|14!
I12"
.s5
X 10 ->
f
8-
s
6- *
•i
4-- :
H ji
2-j -
t i
i '
1* - JL * | '
0-»- •*• * «• * *
K K K S W W V
1233 12C
* !
' 4
j
!
t
i . i
!
i
i
t
i
i
*

> i
t * ji
* '
V A A A A t
3 1 3 4 5 (


*
t
• • • •
v
• 	 :['• "l:;, 	 ! 	 • 	 " • 	 ' 	 ::! 	 " 	 "" 	 ",
' ,-l ! r '
* :
* *
~ i - *
i 	 , . * • _
i 	 -.."• 	
' * :• * *
I i * *
kAAAABJ JMM
5718911212
    Uncontrolled
1-see comb.
     a
2-sec combustion
 Figure 56.  Hg emissions for continuous, intermittent,  and
                 MWI's with combustion controls).
                                          batch
                                 178

-------
              26-"
                <

              24;


              22-


              20;
                H
              18 J
           S   i
              0-
                         81           82
                             Urcortrcted
Figure 57.   Hg emissions for uncontrolled pathological MWI's.
                              179

-------
     4.3.8.2  Add-On Controls--Hg.  Test data from facilities A,
B, J, and M (wet systems and fabric filter systems without carbon
injection) showed no consistency in removing Hg emissions from
MWI's.  Control of Hg is dependent on the presence of carbon at
relatively low temperature for a sufficient period of residence
time.  There is not enough carbon in an MWI exhaust gas stream to
consistently reduce Hg.  Therefore, emission limits have not been
established for these devices.  On the other hand, as shown in
Figure 58, test data from facilities A and M show that injection
of activated carbon into fabric filter systems can achieve at
least 85-percent control of Hg.  This reduction corresponds to
0.47 mg/dscm for a typical continuous, intermittent, or batch
unit; (85-percent reduction from a typical uncontrolled level of
3.1 mg/dscm) and an emission level of 0.01 mg/dscm for
pathological units  (85-percent reduction from a typical
uncontrolled level of 0.05 mg/dscm).  It is not known whether
carbon injection in a wet system will reduce Hg.
4.3.9  Nitrogen Oxides
     Table 23 presents the NOX emission limits under each control
option for continuous, intermittent, batch, and pathological
MWI's.  Figures 59 and 60 present the data used to establish
these emission limits.
     4.3.9.1  Combustion Controls--NO.^.  Figure 59 presents NOX
emission  data for continuous, intermittent, and batch MWI's
burning general medical waste, and Figure 60 presents this data
for pathological MWI's.  Because emissions of NOX are not
affected  by increased residence time in the secondary chamber,
the 1-sec and 2-sec control options will not reduce emissions of
NOX.
     4.3.9.2  Add-On Controls--NO^.  None of the control systems
evaluated reduced NOX emissions. Therefore, NOX emission limits
have not  been established for the add-on control technologies.
                                180

-------
            100 -"
             90-
                                         f

                                         *
             80-
             704
             60-
          rr  so
         S.  40
             30-i
             20
             10-
              0-
                A8
                      A9
M2
Figure 58.
              Fabric Filter w/ carbon

Hg emissions for  continuous,  intermittent, batch, and
pathological MWI's with add-on  controls.
                                181

-------
 TABLE  23.  NOX EMISSION LIMITS FOR CONTINUOUS,  INTERMITTENT,
                  BATCH, AND PATHOLOGICAL MWI'S
Control option
1-sec
2-sec
Wet systems13
FF/PBb
DI/FFb
SD/FFb
Fabric filter systems with carbon
injection"
NO^ emission limits, ppmdv at 7 percent C>2
Continuous, intermittent, batch
NAa
NAa
NAa
NAa
NAa
NAa
NAa
Pathological
NAa
NAa
NAa
NAa
NAa
NAa
NAa
aNo applicable limit.
''Includes 2-sec combustion.
                                 182

-------
  1000--
   900
   800
   700-
«
?7  600-
O
   500
   400^
   300
   200
   100
    0-
  *!*
  K  K  K S
  1233

   Uncontrolled

Figure  59.
                                                  * t
                    ***
                                                           *
WWW
1  2  3

 1-sec comb.
                                AAAAAAAAAABJ JMM
                                123456718911212
                                                a
                                         2 - sec combustion
               NOX emissions for continuous, intermittent, and batch
                    MWI's with combustion controls.
                                 183

-------
                1000
                 900
                 700-

                 500-
                 400 J
-*_
 *
                 3004
                 200
                             SI
S2
Figure  SO.   NOX emissions for uncontrolled pathological MWI's.
                                184

-------
5.0  REFERENCES
 2.
 3.
 4.
 5.
 6.
 7.
 8.
 9.
10.
11.
12.
 13.
Chang, D., R. Mournighan, and G. Hoffman.  Thermodynamic
Analysis of Post Flame Reactions Applied to Waste
Combustion.  Land Disposal, Remedial Action, Incineration,
and Treatment of Hazardous Waste:  Proceedings of the
Thirteenth Annual Research Symposium.  EPA/600-9-87-015.
(NTIS PB87-233151).

McRee, R.  Operation and Maintenance of Controlled-Air
Incinerators.  Ecolaire Combustion Products.   (Undated.)
Letter and attachments from T. Kendron, Simonds
Manufacturing Corp., to J. Farmer, EPA/ESD.  April 9,
Response to Section 114 information request to MWI
manufacturers.
                                                           1990.
Ontario Ministry of the Environment.  Incinerator Design and
Operating Criteria, Volume II - Biomedical Waste
Incineration.  October 1986.

Barbeito, M., and M. Shapiro.  "Microbiological Safety
Evaluation of a Solid and Liquid Pathological Incinerator."
Journal of Medical Primatology.  pp. 264-273.  1977.

U. S. Environmental Protection Agency.  Hospital Waste
Combustion Study:  Data Gathering Phase, EPA-450/3-88-017.
December 1988.

U. S. Environmental Protection Agency.  Municipal Waste
Combustion Study:  Combustion Control of Organic Emissions,
EPA/530-SW-87-021C.   (NTIS PB87-206090)  June 1987.

Hasselriis,  Floyd.  "Minimizing Trace Organic Emissions from
Combustion of MSW by Use of Carbon Monoxide Monitors,"
presented at the ASME Solid Waste Processing Division
Conference,  1986.

Simonds Incinerators.  Operation and Maintenance Manual for
Models 751B, 1121B, and 2151B.  January 7, 1985.

John  Zink Company.  Standard Instruction Manual:
John  Zink/Comtro A-22G General Incinerator and One-Halt
Cubic Yard Loader.

Letter from  K. Wright, John Zink Company, to J. Eddinger,
U. S. Environmental Protection Agency.  January 25,  1989.

Ecolaire Combustion Products, Inc.  Equipment Operating
Manual for Model No.  480E.

Basic Environmental Engineering, Inc., Product literature.
Undated.
                                185

-------
14.  Joy Energy Systems, Inc.  Controlled-Air Incineration.
     Product Literature.  Undated.

15.  Cleaver-Brooks.  Operation, Maintenance, and Parts Manual,
     Pyrolytic Incinerator.  CBK_6826.  September 1988.

16.  Joseph, G. and D. Beachler.  APTI course SI:  412C, Wet
     Scrubber Plant Review - Self Instructional Guidebook.
     EPA 450/2-82-020.  March 1984.  p. 1-3.

17.  Perry, J.H. and C.H. Chilton,  eds. Chemical Engineers'
     Handbook, 5th ed.  New York:  McGraw-Hill Book Company.
     1973.  p. 20-95.

18.  Reference 16, p. 1-4.

19.  Reference 16, p. 1-5.

20.  Reference 16, p. 1-7,

21.  Reference 16, p. 1-8.

22.  Andersen 2000, Inc.  Emission Control Systems for
     Incinerators.  Bulletin TR 89-900239.  February 1989.
     p. 15.

23.  Reference 16, p. 3-2.

24.  U. S. Environmental Protection Agency.  Operation and
     Maintenance of Hospital Medical Waste Incinerators.
     EPA-450/3-89-002.  March 1989.  p. 3-9.

25.  Reference 22, p. 14.

26.  Andersen 2000, Inc., Andersen Equipment Brochure.  Bulletin
     TR 81-90001 32 REV. B.  Undated.  p. 18.

27,  Andersen 2000, Inc.  Venturi and Packed Tower Scrubbing
     Equipment.  Engineering Manual with Operating and
     Maintenance Instructions.   [Model No. WAV-118 (variable
     throat venturi)].  October 7, 1988.  P. 2.

28.  Reference 16, p. 3-3.

29.  Brady, J., Andersen 2000, Inc., Economically and
     Operationally Attractive Incinerator Emission Controls,
     Undated,  .p.. 21,

30.  Reference 16, pp. 3-4, 3-5.

31.  Reference 27, p. 7.

32.  Reference 16, p. 3-6.
                               186

-------
33.   Reference 16,  p.  2-3.

34.   Reference 16,  p.  3-7.

35.   Letter and attachments from D. Sanders,  Andersen 2000,  Inc.
     to W. Maxwell, EPA/ISB.  January 29,  1990.   Response to
     request for control equipment design and cost information.

36.   Reference 24,  pp. 3-9, 3-12.

37.   Andersen 2000, Inc.  Venturi Scrubbers for Fine Particulate
     Control.  Bulletin 78 900075 REV. B.   December 1982.
     pp. 6-7.

38.   Reference 24,  p.  3-9.

39.   Telecon.  M. Turner, MRI, with D. Sanders,  Andersen 2000,
     Inc.  March 16, 1990.  Design and operating parameters for
     APCS.

40.   Reference 24,  p.  3-12.

41.   Reference 22,  p.  14.

42.   Letter and attachments from H. Marschall, Emcotek Corp., to
     W. Maxwell, U. S. Environmental Protection Agency.  Product
     literature.  December 13, 1989.

43.   Reference 16,  p.  5-3.

44.   Letter and attachments from S. Sheppard, The Ceilcote Co.,
     to W. Maxwell, EPA/ISB.  January 31,  1990.  Response to
     request for control equipment design and cost information.

45.   Letter and attachments from A. Dozier, Advanced Concepts,
     Inc., to W. Maxwell, EPA/ISB.  January 21, and February 8,
     1990.  Response to request for control equipment design and
     cost information.

46.   Reference 22,  p.  16.

47.   Reference 16,  p.  5-3.

48.   Reference 16,  p.  5-2.

49.   Joseph, G. and D. Beachler.  APTI Course 415, Control of
     Gaseous Emissions  - Student Manual.  EPA 450/2-81-005.
     December 1981, p. 4-22.
                               187

-------
50.  Andersen 2000,  Inc., Technical Description, Performance
     Information,  Material Balance and Flowsheet Data, Typical
     Guarantees and Turnkey Installed Pricing for a 16.2 MM
     BTU/hr (1,700 Ib/hr) State-Of-The-Art Medical Waste
     Incinerator With Wet Scrubber Emission Controls.  Workshop
     on Medical Waste Management, Sacramento, CA.  April 18-19,
     1990.  p> 10.

51.  Marschall, H.L., et al, Emcotek Corp., New Oregon
     Incinerator Demonstrates Process Innovations to Protect Both
     Air and Water,  presented at the AWMA Specialty Conference -
     Thermal Treatment of Municipal, Industrial, and Hospital
     Waste II  (Pittsburgh, PA) and at the Pacific Northwest
     International Section Annual Meeting  (Spokane, WA).
     November 1989.

52.  Marschall, H.L., et al, Emcotek Corp., Comparison of
     Commercial Gas Cleaning Systems Advanced Scrubber vs.
     Baghouse For Hospital and Infections Waste Incinerators,
     presented at the Second National Symposium on Infections
     Waste Management.  San Francisco, CA   (September 19-23,
     1988) and in Boston  (October 3-7, 1988).

53.  Marschall, H.L., et al, Emcotek Corp., Retrofitting Air
     Pollution Controls to Existing Incinerators, presented at
     the 82nd Annual Meeting of the Air and Waste Management
     Association, Anaheim, CA.  June 1989.

54.  Ceilcote.  IWS™ System, Air Pollution Control Division,
     Technical Bulletin 12-19.  December 1986.

55.  Ensor, D. S., Ceilcote.  Ionizing Wet-Scrubber Evaluation.
     EPA-600/7-79-246.  U. S. Environmental Protection Agency,
     Research Triangle Park, North Carolina.  November 1979.

56.  Shifftner, K. and G. W. Burke.  Incinerator Emissions
     Control Using Wet Scrubbers.  Presented at the Second Annual
     National Symposium on Infectious Waste Management and
     Incineration of Infectious Waste, San Francisco, California.'
     September 19-23, 1988.

57.  Holland, 0. L. and J. D. Means.  Utilization of Hydro-Sonic®
     Scrubbers for the Abatement of Emissions from Hazardous,
     Industrial, Municipal, and Bio-Medical Wastes.  Technical
     Paper 7802.  John Zink Company.  Tulsa, Oklahoma, 1988.

58.  Freeman, H. M. and R. A. Olexsey.  Evaluation of Pilot-Scale
     Pollution Control Devices for Hazardous Waste Incineration.
     Paper 85-77.5.  Presented at the 78th Annual Meeting of the
     Air Pollution Control Association.  Detroit, Michigan.  June
     1985.
                               188

-------
59.
60.
61.
62.




63.




64.





65.




66.





67.

68.

69.

70.
     Letter and attachments  from E.  Mull,  Jr.,  Interel  Corp.,  to
     W.  Maxwell,  U.  S.  Environmental Protection Agency.
     January 30,  1990.   Response to  request for information from
     air pollution control device vendors.

     Letter and attachments  from W.  Wiley,  Consumat  Systems,
     Inc.,  to W.  Maxwell, U.  S.  Environmental  Protection Agency.
     March 19,  1990.   Response to request  for  information from
     air pollution control device vendors.

     Letter and attachments  from P.  Finnis,  Procedair Industrie
     Corporation,  to W.  Maxwell,  U.  S.  Environmental Protection
     Agency.  February 1, 1990.   Response  to request for ,
     information from air pollution  control device vendors.

     Beachler,  D.,  and M. Peterson.   APTI  Course SI:412A,
     Baghouse Plan Review Student Guidebook.  EPA 450/2-82-005. .
     U.  S.  Environmental Protection  Agency.  April 1982.

     U.  S.  EPA.  Control Techniques  for Particulate  Emissions
     from Stationary Sources,  Volume I. EPA 450/3-81-005a.
     U.  S.  Environmental Protection  Agency.  September 1982.

     Letter and attachments  from J.  Childress,  United McGill
     Corporation,  to W.  Maxwell,  U.  S.  Environmental Protection
     Agency.  February 9, 1990.   Response  to request for
     information from air pollution  control device vendors.
     U.  S.  Environmental Protection Agency.
     Maintenance Manual for Fabric Filters.
     June 1986.
Operation and
EAP/625/1-86/020.
     Richards Engineering.   Air Pollution Source Field Inspection
     Notebook;  Revision 2.   Prepared for the U.  S.  Environmental
     Protection Agency,  Air Pollution Training Institute.
     June 16, 1988.   p.  97.

     Reference 49,  p. 5-1.

     Reference 49,  p. 5-2.

     Reference 24,  p. 3-23.

     Bundy,  R.  P.,  United McGill Corporation.  Fabric Filter
     Systems for Dry Acid Gas and Particulate Removal, presented
     at the Incineration of Infectious Waste Conference.  San
     Francisco,. California (September 19-23, 1988)  and in Boston
     (October 3-7,  1988).  p. 3.

71.  Reference 66,  p. 102.
                               189

-------
72.
73.

74.

75.

76.

77.

78.

79.





80.




81.

82,




83,


84
 85
Letter'and attachments from J. Hsieh, Research Cottrell Air
Pollution Control Division, to D. Wallace, Midwest Research
Institute.  August 31, 1989.  Information regarding the
performance of a dry injection fabric filter system
installed on a Skovde, Sweden, MWI.
Reference 66,

Reference 66,

Reference 24,

Reference 66,

Reference 66,
p. 99.

p. 98.

p. 3-24.

p. 101.

p. 100.
Reference 66, pp. 100-101.

U. S. Environmental Protection Agency, Municipal Waste
Combustors--Background Information for Proposed Standards:
Post-Combustion Technology Performance.  EPA-450/3-89-27C.
August 1989.  p. 6-2.

Jorgensen, C., Acurex Corp., Assessment of Flue Gas Cleaning
Technology for Municipal Waste Combustion.  September 1986.
pp. 16-31.

Reference 66, p. 115.

U. S. Environmental Protection Agency, Municipal Waste
Combustion Study:  Flue Gas  Cleaning  Technology.
EPA/530-SW-87-021d.  June 1987.  p. 4-14.
Memo from D. Wallace, MRI,  to R.  Copland,  EPA/ESD/SDB.
Emissions from  Pathological MWI's.  March  30,  1992.
                                          SO.
Michigan Hospital  Incinerator  Emissions Test  Program; Volume
II:  Site  Summary  Report,  Borgess Medical  Center
Incinerator.   Energy  and Environmental Research Corporation.
Prepared for Michigan Department of  Commerce  and U.  S.
Environmental  Protection Agency, Research  Triangle Park, NC.
August  13,  1991.

Medical Waste  Incineration Emission  Test Report for  Borgess
Medical Center,  Kalamazoo,  MI;  Volume I, Volume II:
Appendices A-C,  and Volume III:  Appendices D-I.  Radian
Corporation.   Prepared  for U.  S. Environmental Protection
Agency, Research Triangle Park, NC.  January  1992.
                                190

-------
86.   Michigan Hospital Incinerator Emissions Test Program;
     Volume III:   Site Summary Report,  University of Michigan
     Medical Center Incinerator.  Energy and Environmental
     Research Corporation.  Prepared for Michigan Department of
     Commerce and U. S. Environmental Protection Agency,  Research
     Triangle Park, NC.  August 13, 1991.

87.   Medical Waste Incineration Emission Test Report for Jordan
     Hospital, Plymouth, MA; Volume I-III.  Radian Corporation.
     Prepared for U. S. Environmental Protection Agency,  Research
     Triangle Park, NC.  EMB Report 90-MWI-6.  February 1991.

88.   Medical Waste Incineration Emission Test Report for Lenoir
     Memorial Hospital, Kinston, NC; Volumes I-II.  Radian
     Corporation.  Prepared for U. S. Environmental Protection
     Agency, Office of Solid Waste, Washington, DC.  May 1990.

89.   Medical Waste Incineration.  -Emission Test Report for
     Morristown Memorial Hospital, Morristown, NJ; Volume I and
     Volume II:  Appendices.  Radian Corporation.  Prepared for
     U. S. Environmental Protection Agency, Research Triangle
     Park, NC.  February 1992.

90.   Medical Waste Incineration Emission Test Report for AMI
     Central Carolina Hospital, Sanford, NC; Volumes I-II.
     Radian Corporation.  Prepared for U. S. Environmental
     Protection Agency, Office of Solid Waste, Washington, DC.
     EMB Report 90-MWI-5.  December 1991.

91.   Medical Waste Incineration Emission Test Report for Cape
     Fear Memorial Hospital, Wilmington, NC; Volumes I-II.
     Radian Corporation.  Prepared for U. S. Environmental
     Protection Agency, Office of Solid Waste, Washington, DC.
     EMB Project No. 90-MWI.  November 1990.

92.   Telecon.  Maxwell, W., U. S. Environmental Protection
     Agency, with Hadley, H., B.G. Wickberg Co., Inc.  April 23,
     1992.  MWI emission test at Jordan Hospital, Plymouth, MA.
                               191

-------

-------
             APPENDIX A.




SUMMARY OF TEST PROGRAM OPERATING DATA

-------

-------
O

§
Cn

EH
co
H
EH
O
fc
H
EH
g
s
CO
CO

o

§
CM

H
      is
      ll
      (S cT
      « =5
      £?£
      CQ

      u
                                     cs
                                           VO
                                              Ov
                                        es
                                                 en
                                                    en
                                                        en
                                                           oo
                                                           es
                                                                    vo
                                                                                 VO
                                                                                       00

                                                                                       tn

                     vo
                                     VO
                                              s.
                                                     V)
                                                     en
                                                              §

                                                                    §
                                                                       00
                                                                          VO
                                                                              es
                                                  i
8«

ll
            a
                  en
                        VI
                           VO
                                                                 o\
                                     5?
                                                                                 a
                                                                                    5?
               00
                     VO
                        vo
                           vo
                           en
                           vo
ondary

chamber
                            o\
                            §
                                        8
                                                        vo
                            00
                            vo
                                                                    a
                                                                    00
Sec

                  00
                  £

                        00
                        es
                        VO

   1
                                           vo
                                              00
                                                 i
                                                     00
                                                        v>

                                                                 00

                                                                                 SI
                                                            s
                                                                                              VO
      •
                     O
                     en
                                                 3
                                                     es
                                                        en
                                                              en
                                                                    en
                                                                    a

                                                                              a
                                                            a
                     es
                               en
                                  en
CO
w
I
            O
            8
                     O
               i
               i1
                        0
                        s
                      0
                            £
O
                                  O
                                     en
1
0
                                  00
                                  §
                                        en
i
O
         §1
                                           en
i
0
                                           8
                                                              vo
                                                                 \O
                                                                    VO
                                                                        \O
                                                                                              00

o
CQ
OA

o
ral
G
                                     BQ
                                     Ul
                                                  CO
                                                        &
                         s
                         ta
                         I
                         *-«

                         s'
                                                                 a
                                                                 £
                                                                 u
                                VO

                                ?Q

                                                                              oo
                                                                                       s
                                                                                              Ov
                                                                                              00

                                                 A-l

-------
U


•5^
ia O
           o

           8  _


           §> S
          05 -X
It
                         VO
                         0\
                             o\
                             00

                                 VO
                                        ts

                                                00
                                                    Os
                                                        S.
                                                        t-
                                                            Os
                                                                               o\
                                                                    s
                                                                                          en
                                                                                                  SO
                                                                                                      CM
                                                                                                     so
                                                                                00
                                                                                              SO
                                                                                                      so
•8

 i
•H
JJ

 §
 O
H

 i
W
J



I
       CO
           5?
                  g
          i
          £
               *

                                                                               VO
                                                                                   00
    o,

    E
          £
                         Os
                             Os
             '

                  o
                     £
G


ral
G
ral
G
                              o
G
G
eral
G
                                                                O
                                                                                          O

                                                                                    O
O
O
                                                                                                  oo

                                         CO
                                                          A-2

-------
            bo
           J
            x ,H
           U3 O
                                           00
                                                                                                             es
            9j

            «
            S

           6
                                               00

                                                   C-J
                       fS
              •-
•o
 0)

 B
-H
4J
 a
 o
 u

                                           00

                                                                                                     00
                                                                            VO
                                                                                                 00
                                                                                                 VO
                                                                                                                 VO


         5  h  tt-
 E fc
 nt o

*§   r
           £
                                   00
                                           00

                                                                            00
                                                                                        0\
                                                                                                         VO
               •a
                                                                                                 ts
                                                                                                         tn
                                                                                                                 «n

eral
                                                                                    O
                                                                             O

                   00
                                            en

                                                VI
                                                    tn
                                                        co
                                                            CO
                                                                CO
                                                                    CO
                                                                        co
                                                                            CO
                                                                                                 U
                                                                                          u
U


                                                              A-3

-------
                   £

                   I

                   I'
TJ
(U

§
-H
XJ

g
O
<
w


I
                    « •£
.   5s
                                           A-4

-------
o

I 2
.- .^
35 a
u
c £j
I £
•e ~
o'
1 J5

E -21
J
2 .
i <=- 6"

o- p ac
tt. T3
B.
il
— ^o
.£2 U«
o
3
CO
O rT
"e °
**
S D.
'5 S
!s "
tt, &
tt. §
"5 ^
1°.

p
j

2


VO
vo


o
o
m
VO


CM



r-
r-
m

oo



Xi



CM
CO



CO
-
CO
CN
CO
s


00
CO


0
0
VO
Tt

OS
CM



00

CM
OO
fN



Xi



^
CO




-
8
IN
00


CO
CO


0
o
00
t-


CM



«

CO



XI



^
CO



CO
-
CO
00


CO
vO


o
o
oo
(—


«M



o
oo

s



XI



^
CO

Is-

CO
"CN
1
1
CO


VO
in


o
o
o
OS


CM



EJ

8



Xl



m
CO

00
t-»
CO
tN
B,
i


Os
m


o
o
VO
t-


CM



CO
CO

S
CO



XI



•n
CO



CO
-
1
1


VO
•*


o
o
CO
VO


CO



i

CM



XI



VO
CO



CO
CO
t—
5
i



XI


o
0
t-
in

CM
co



XI

XI



XI



Os
CO

Os


CO
oo
§


CN
VO


o
0
o
00

VO
CM



CO
CM

8



Xl



tN
CO

Is-
r»
CO
to
K
s
ta



X)


o
0
^
CM

o
co



•o
m

XI



x>



JN
CO



CO
to
s
CO
01



x>


o
0
m


0
CO



%

s



XI



Tt
CO



CO
CO
8
s
w


vO
o


0
o
^
CN

CO
CN



CO

XI



Xi



X
CO

Xl


-
1
CO
00
S


r-
r-


o
o
o
CN

CM
CO



g

X)



x>



oo
CO



CO
Tt
«n
«n
S
00
s


oo
•n


0
o
o
o

Tt
CM



1=
CO

x>



XI



^
CO




-
1
1



CO
Tt


0
o
m
Tt

Tt
CM



lo
5?

JS



XI



Tt
CO
CO

XI


-
i,
s
w


Os
to


0
o
Os
m

Tt
CM



T?

XI



Xl



ON
CO

Xl


-
s
tt)


VO
•n


o
0
CO
CN


CM



T?
CN

t-
VO
CO



X)



2
to


oo
to
vo
1
s
ta


..
in


o
0

Pi

Tt
tN



VO
CO

1



Xl



o
CO

t-

co
VO
s
§


Os
Tt


o
o



Tt
CN



CO
fO

XI



XI



in
CO


r-
to
vo
Tt
§


in
00


O
o

CM

tN
CO



VO
OO

g



XI



8
CO

00
r—
to
VO
VO
i


in
«n


o
o
oo
r-

eo
CM



T?

5
CO



Xl



VO
CO

CN
oo
CO
Is-
1
00
ta


00
CO


o
o
CM
00

in
CM



1

in
CO



Xl



T?
CO

00


t-
o
s,
00
00
01


0
Tt


o
o
Tt
oo

in
CM



tN
oo

s
CO



XI



00
CO
to




Is-
i
*"l
3
ttl


Os
•*


0
o
Tt
00

oo
o



§

CM
OS
CN



t-
r-
0

CO
CO

XI


1
Os
s'
A-5

-------
0)
§

O
o?
.°1
•5 o
'5 '£
35 |
il
Q .D
a ~-
U 2
I-
« *5h
a"*
£ .
1 o^
1 %•£
6*
»l
(x. «
B. 2
£ U.
u
co
t> ft.
•gr
jo 6.
'5 §
CQ •=
is "
£|
I?1
If
§
|



2

•sf

o
o

ON
ON


M

§

VO


§

<*
en
*


-

s
c3N
s
NO
VO

0
0


ON


O

8
«—

§


•**•
S

«
^


"

s
ON.
a

en

O
—


00


en

NO
oo

vo
oo


oo
1«M

e«i
*


'oo

ON
ON
C

rj-

O
•-•

OO
ON


O

00

ON
«S


cn

en
en
A


00

S
ON
s

>r>

V,
cs


ON


O

ON

ON


ON
S;

oo
ts
^


ON

ON
en
ON
5
ON
m

in
es


oo


C-4

f-

1


V*

s
en
JS


ON

§
ON
OO
«

en

«
ts

en
oo


r»

s
VO

1


t-
5?

t-
en
en
.a


»

NO
ON
§
                                     8
                                     g

                                     I
                                           A-6

-------
      TABLE A-3.   APCD  OPERATING PARAMETERS FOR FACILITY B
Test
MM1 1622
MM1 3624
MM1 4626
MM1 5627
Condition
1
1
1
1
Scrubber
inlet
temp., °F
545
543
541
a
Scrubber
outlet
temp. , °F
136
138
142
a
Pressure
drop in H00
31
31
31
31
aData not measured.
                               A-7

-------
       TABLE: A-4.  APCD OPERATING PARAMETERS AT FACILITY  j
Test
Jl BR305
J2 BD305
J3 BA307
J4 BD307
J5 BR309
J6 BD309
Condition
1
2
1
2
1
2
FF inlet
temp., °F
372
373
371
375
374
376
Scrubber
inlet
temp., °F
148
134
136
135
133
a
Scrubber
pressure drop,
in H,0
0.45
0.36
0.46
0.40
a
a
aData not available.
                                A-8

-------









S
£
H
^^
H
O
g
CO
i
i
04
r1)
&
H
PERAT
O
Q
U
«
in
rtl
S
9
H





||
"« 1)
o
.S «
-1 2
o
*l
a w
Q
•n
o -a
2 g
u
o
c/3
li
3
a"
2 ^
a

^
f*
u.
r
tfl
r
e
5?
b.
fr.
- fc
•3°
.9 r
g a
*
1



I

o




CO
f>


CS




0
d


8




oo



oo

oo

-
00
*«• «
1

1
r-
o




1
c<<


CS




o
d


w




00
(S



^

i

-
2
1

i
00
o




5?
pn


w




O
d


£




i



oo

1

-
0
1

i

0




9
tr>


2




00


rt




1



i

i

.«
-
«»«
•«*•'
S

o




^
en


S




c4


S




a



oo
ts

i

-
gj
— ,
VI
i

o




1
<«^


ci




S


w




S



i

!

-
a
i
VO
1
A-9

-------

-------
            APPENDIX B.




TEST PROGRAM EMISSION DATA SUMMARY

-------

-------
 M
 .J
 J
 O
 04

 <
 H
 tt!
 W
 E-«
 H
 Pd
 U

 rt
 O
 fa



1
. H
 H

 Srt
 EH CO
• & H
 W CO

-^1

 8Q
   H
 sy
 O rt!
 M
 CO Q
 CO Z
 H i«
 S
 O
 H
 E-"

 S
 O
 U
  I
 EH
 CO
 O
 CU
 m
1
«•
X
O
z
1
n.
D.
cT
00
i
p.
Q.
U
X
"1
i
o
u
<*.
00
S
a.
e
•.2
•o
S
u

S
"1
B-

t^



in
c-4



oo
ON
NO


O
8


£
o



—




s
s
<


00
00
' '


NO
ON



ON


VO
NO


in
o
o



—




S



«n
fl



00
ON



1


ON
NO


1
O



—




1



2



r-J
ON
^~*


00



oo



•°



ts




**l
<


00




in



i


o
«



x>



r<



t-
i'



^




NO
' '


NO
f>


NO
S



XI



«s




i1
<


8




oo



i



-


i
o



w




a



NO




O



R
t-


o
ON



X)



n




S
w
,


9




«*•>



i



o


oo
o








s
ta



R




NO



u
—
•
ON
^



•°



W




1
a



s




00
""


ON
00


00
8


rf
S
O



•*




5



s;




^
n


oo



8


r~
o



•



00
fS



NO
oo
M


8


i
o



f-




1
<


s




•-«



-
«


o


o
o



t~-




1


f~
JC




in



NO
NO


>n
NO


i
O



•<
*-^



ON
<

NO
tS



oo
f)



ON
g



oo


i
o



<
~^



a1
<


ON



^
O



§


ON
"*


00
g
O



<
••*



i'


ON
r-
\o



in
ts



1


NO
Tf


1
O



00




ON
1



NO



fS
c<



o
oo


ON
en


ON
O



00




1


                                             B-l

-------
•d
a)
•H
4J
fl
O
U
H
n
s
I
i
°:
X
0
z
1
i.
a.
1
£
O
3C
i
a.
0
U
1
u*
oo
s
a.
e
1

1
t
a
00
s


v>
TT
VO


NO
en


g
o
o



ON

ON
en
S5


VO
s



es
-j.


vq


i
o



ON

ON


00
s



VO
i
cs


en


i
0



ON

ON
VO
O\
8



£


oo

s
f.


en
en


I
o




S
VO



vO



O
1


V)


i
o




s
s
ea


u




o
m


a


en
O




8
3
i
09


s


VO

u


en
O
en



JO



*•*

i
CO


8




8


GO


g
O





ffl
*">


8


ON
en
en
VO
a


5


VO
o
0




v>
0
!
Ira,


OO


(S

00
n


en


g
O





o
i



jq



VO
5


6


i
o





CO



o3



VO
c-
00


s


g
o





v>



s



VO
en
CO
VO
en


i


O
O



cs

oa



S


o

1


oo
VO



ja





|



oo
VO




-


i


3
O





|



S


VI

1


00
s


t
o





i



VO


VO

|


VO
s


i
o





R



oo
oo


00
VI
VO


i


ON
vo
O





5?



g>


00

r?
en
cf


en
vo


22
O





s
en



»



VI
|


•*
i


5
o



en

1



g>


VO
OO
o
**!


o


8
O



en

s
go



5


ON

O


§


8
o



en
oo
s
ON


en
3


ON
v>
•a


V)
V)


3
o




00
1


o
n


V)
ON
•o


V5
00


3
O




ON
i



VI
VO




s
r-


3


V)
O




R
en


V)
en
VO


oo

!


VO


VI
V)
VI
o




es
|


en
vo


O
vo
I


t—
OO


en
o



es
es
i'


                                        B-2

-------
0)
•ri
4-)

g
U
H
i
m
w
•o
1
X
U
Z
1

g
a.
D.
U
a
>
&
^
o
u
T§
&
cu
a
•2
•o
U

1
1

VO
on


O
VC

1


r-
*



oo
0





•t
i



9
o


en


S
oo
8
CO


00

o
S

N



&



S
o


—


oo
BU
U
-




„
—

5



i



i
0


—



U
'


VO


O

s


0
s



i
o


—


r*
U
*





-

i


OS
a



§
o


ts


00
00
s
-


00


00
en

g
en


Ov
oS



oo
o


«s


00
s
'


Ov

CJ
a

r>



I



2
o


CJ


a
U
»


SI


2

g


00
so
I'M



oo
0


en


s
U
'


$


ts
en

n
n


00
i



s
0


en


^
&
U
1


Os


N

1


en
§



i
o


en


00
8'
-
                                            2
                                            8
                                            o
                                            1
                                            8
                                            Is
                                            'I
                                            I

                                            i
                                        c?
&
r-
S
1
                                        o  ^j S
                                        ° • S •"
                                         J"2 CO -a
                                         1 U 1

                                        nil
                                        l§|i
                                        — a i -2
                                        ^ Q OQ Q
                                        es"^ O T3
                         B-3

-------
TABLE B-2.  POST-COMBUSTION EMISSION CONCENTRATIONS  FOR
                  CDD/CDF AND METALSa
Facility
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
B
B
B
B
Test
MB1 1523
MB1 2524
MB1 3531
MB3 1505
MB3 2507
MBS 3508
EB1 1517
EB1 2518
EB1 3521
EB1 4522
EB1 5525
MB131502
MB132515
EB6 1426
EB6 2427
EB6 3430
EB7 1509
EB7 2511
EB7 3514
EB7 4516
EB8 1529
EB8 2530
EB8 3601
B2 90791
B3 90991
B4 91091
B5 91191
B6 91291
B7 91391
B8 91491
B9 91691
MM1 1622
MM1 3624
MM1 4626
MM1 5627
Condition
1
1
1
2
2
2
3
3
3
3
3
4
4
5
5
5
6
6
6
6
7
7
7
1A
1A
1A
8
8
9
9
9
1
1
1
1
CDD/CDF, ng/dscm
665
410
394
229
531
505
250
b
485
b
318
b
271
b
b
b
691
317
b
553
319
484
574
158
274
279
306
517
471
429
347
1,554
1,388
1,411
. b
Cd, ng/dscm
365
283
358
b
b
b
343
b
729
603
b
143
420
144
102
190
260
329
263
211
190
201
239
396
364
446
346
3,520
241
1,210
690
498
285
373
b
Pb, ng/dscm
3,441
2,928
3,530
b
b
b
3,730
b
3,509
2,994
b
2,155
2,549
4,911
4,625
6,534
6,855
7,642
8,629
3,927
4,685
5,156
5,766
952
2,180
1,377
2,600
1,620
459
2,280
1,760
3,810
2,620
3,520
b
Hg, ng/dscm
18,007.9
11,715.5
5,992.2
b
b
b
14,848.5
b
25,708.3
8,403.2
b
363.5
43.1
402.0
3,012.5
238.9
9,960.8
154.6
842.7
4,816.3
1,891.0
1,906.1
116.7
9,580.0
7,300.0
3,770.0
6,300.0
8,460.0
8,479.8.
6,830.0
13,200.0
167.0
414.0
516.0
b
                           1-4

-------
TABLE B-2.  (continued)
Facility
J
J
J
J
J
J
K
K
K
K
K
K
K
K
K
M
M
M
M
M
M
S
S
S
S
S
S
S
S
S
W
w
W
w
Test
Jl BR305
J2 BD305
J3 BR307
J4 BD307
J5 BR309
36 BD309
LM4 602
LM4R 604
LM6 605
LM1 530
LM2 531
LM3 601
LM7_606
LM8 607
LM9 608
MW1 1118
MW2_1119
MW3 1120
MW4_1121
MW5 1122
MW6 1123
CC1_920
CCS 922
CC10 102
CCS 924
CC6 925
CC9 928
CC2 921
CC4 923
CCS 927
CF1 815
CF5 821
CF6_822
CF2_818
Condition
1
2
1
2
1
2
1
1
1
2
2
2
3
3
3
1
1
1
2
2
2
1
1
1
2
2
2
3
3
3
1
1
1
2
CDD/CDF, ng/dscm
4
5
3
599
94
3,036
4,818
b
6,325
3,247
8,336
10,213
48,571
27,769
30,243
144
129
301
242
292
63
279
1,994
468
427
157
1,144
12,728
40,969
20,457
3,084
3,562
539
4,070
Cd, /tg/dscm
44
152
23
252
22
216
b
163
173
110
392
140
430
478
202
494
369
548
524
402
549
16
504
<1
13
26
22
55
42
67
296
336
209
333
Pb, /ig/dscm
718
5,187
634
5,028
1,063
6,412
b
7,012
1,549
1,316
1,855
2,439
3,482
2,121
2,416
3,801
3,570
3,590
4,250
4,730
4,260
512
814
0
329
202
389
2,380
1,210
2,050
4,754
6,372
2,864
4,476
Hg, /tg/dscm
527.0
654.0
2,862.0
8,620.0
383.0
908.0
b
208.8
75.6
152.3
124.5
139.7
5,926.0
85.4
172.4
2,200.0
643.0
4,930.0
2,420.0
2,240.0
3,220.0
4.3
<0.9
116.2
182.5
<0.7
<0.5
10.7
5.2
-0.7
286.3
120.0
146.4
613.0
          B-5

-------
                           TABLE  B-2.   (continued)
Facility
W
W
W
W
W
Test
CF3 819
CF4 820
CF7 826
CF8 827
CF9 828
Condition
2
2
3
3
3
CDD/CDF, ng/dscm
8,102
5,671
5,215
2,679
7,198
Cd, fig/dscm
461
361
731
619
401
Pb, ng/dscm
7,537
6,741
4,316
7,710
4,008
Hg, ng/dscm
880.7
35.7
6,543.4
939.4
432.7
aAll concentrations corrected to 7 percent
"Data not measured.
                                         B-6

-------
rt
 CO

 H


 g
 fa-
 H
 U
 H
 fa
 fa
 Ptl

 >4
 O
 g
 1
 H
5" "5 £
Sg S.
U E
•? a 3
>N ~
Igi
< =*•
s I*
£
8 i
a, -a
3"!
CDD/CDF
efficiency,
percent
Post-APCS
CDD/CDF,
ng/dscm
t-
£•§§
e s.
4)
| |
< o-
^0
£ X
fs
£ Oa
||
•5
a


i
i
ON
S
O
NO
vT
i
8
NO
ft
A
V
x .
E

s
VI
—
NO
3
o
-
en
s
CO
<
ON
fn
f*l
o

i
ON
£
A
V
en
8
t-
s
p-
s

5
6
-
„.
S,
CO
'«
-
*n
wi
cS
NO"
ft
s
i
ON
X
1
NO
S
ON
§

*
6
-

en
a
<
O

o
«
0
o
u
oo
£
o
*

a

0
o
es
VI


<
o

o
o
o
»
o
*
i

8

s

0
o
-
^
VI
ts

<
o

o
o
0
u
u
00
*
o
5
en
8

0
o
-
00


<
e^
^
o"
Tt
C^
A
oo
V
-
1

S

o

*
0
en

v>
\

<
O

0
»
o
u
o
o
»
^
s
00
—

o
o
en

5,
S
<
r-
R
ON
5
r-
ON
ON
i
A
p
V
2
i
oo
£
00
s

8;
0
en

S(
§
<
3
_^
1
i
a
00
00
u
u

O

O

s
o
en

S,
§
<
o

o
u
o
o
«•
00
§

o

0

o
0
en

S,
§
<
U

o
u
o
u
»
o
u
•o
oo

—

o
0
-

1
CO
«
«

u
0
0
»
0
o
u
,
8;

*

o
o
»
VI
s
03
<
4)

U
«>
o
-
.
-
o
T3
OO

?3

e
•
•n

|
U)
-
-
^.
5-
«
NO
a
TT
oo
A
NO
V
o
u

s
oo
-
ON
5?
1
o
V,

s,
i
-
-
V,
a
oo
Si
ON
si
A
oo
V
o
0

s
ON
ON
e
00
en
00
O
«


'i
<
•

O
•
u
0
.
«n
p
en

&,

-

§
S
<
-
en
VI
ON
NO.
ON
a
en
ON
ON
2
-
s




00
£
o
-

1
S
<
X
ON
§
r-
Si
ON
a
i
A
V
-
5
^
R

"
NO
«.
O
-

!•
s
<
J3
O
o
oT
t-
ft
ON
ri
P-;
ON
d
-
f-
fn
OO
8!
r-
g

S
d
^

i
&
<
-
o
I
oo"
ON
ft
O
en
§1
p;
00
en
S
o
S
o
s
ON
5;
d
S

i
s
<
-
o
1
en"
i
oo
SO
1
en
cs
00
s
?
en
ft
^
00
ON
S
d
S

§
ON
S
<
m
*o
oo
C4
?=
i
NO
en
ft
ts
i
S
^
S;

2
en
§
d
oo

ON
ON
m
<
oo
o
p.
00
§1
CO
i
T
ts
t-;
oo
ON
en
<0
ft
^
NO
00
*
d
oo

51
a
<
 Pel

 g
 O
 o

 CO
 o
 CO
 o
 04

 fa
 o
 CO
 n
  \
                                                             B-7

-------
•8
§
•H
4J
g
O
CO
m
w
^

I
N
*!8
i8-
y e
S. . o
^=t
1 S
£_
£§ 8
I ^
fc o-
U
i« i
255
3 =£
.J
S-8*
1
co £
P
*S
CDD/CDF
efficiency,
percent
§S"s
!^s
l§ =
f«
*|!
U) >
< 0.
Js
cu S
f *
g|i
|l
? &
* ..-
<2g
|
£

1
r^
G
iS
vO
00
OO
Ov
CO
s
n
g;
o

o
8
eo

O
S;
c~
S
o
s
1
o
Ov
»•*
2
e>
&


CS
oo
Ov
	
vi
o
>*•
&
(S

en
s
01
CO
cs
M
oo
Ov
t~

o
g;
»»•
Ov.
00
s;
|
0
o
^^
1
o\
s



g
*n

eo
VI
a
•^

en
oo
r-
»A
S
S

Ov
»
<*-
«n


00
I
0
~
9
m


c-
m
if
r»
S
•»»•
*n
Tf
O
1
o
$
•«•
ft
1?
«n
tn

oo
8!
"oo
C4
so
S
f-
s
e
-
1
4



0

0

O

o

o

o
u>
o


V

o

u

o
~4
s
uS


*n
VI
vO
O
>o
5
«n
S
o>
—
t~
S
NO
VO
^
§
VO
oa
e>
eo

&
t~-
9
oo
*o
oo
00

^r
2
*r
SJ

r4
.0
VO
Wl
V)
t~
g
00
•-
00
VO
in
§
O
-
s
S3


r~
o»
»n
O
U1
S
en
s
V*
fa
0<
oo
CJv
•<»•
«n
Xi
§
en
oo
oo

00
«
s;
8
o
n

=
•*
s
§
0
tM
?;
i
S


vri
S
O
VD
OO
'T
O
&
O
VO

IS
ts
r-
i
oo
00
^
0
S;
r~
CO
O
s
I
0
«
s
s
ca
VO
1^1


00
s
o
5
Ov
§:
•SJ-
o\
VO
8;

cs
00
s
!2


u
0
t—
O
s
I
o
-
00
1



^
Ov
S
o
1
d

oo
g:
Ov
O
>n
5
2


u
Ov
VO
00
S
i
o
-
o
$


•*
s
f>
V
t~
S
Tf
—
VO
8!
t
«s
t^
s
VO

V
g;
>n
•*

^
§
o
-
8
i


e*
S3
\
S
oo
CN
Ov
g;
ON
O
•0
s
m

_
g

S
Ov
g;
i
o
cs
IN
1
S


Ov
XI
Ov
O
s
Ov
g;
0
% • •** •
                                                   I § P S « § 1

                                                   liiliij
                                                   « ^ u *o «> (*!T ec
                               B-8

-------
              APPENDIX C.




GRAPHS OF POST-COMBUSTION EMISSION DATA

-------

-------
                                                            =5

                                                            CM
                                                            m
 CO
 fl
 o
-H
 4J

 2
                                                            CO
 0)
 u
 §
 u

§
                                                                    o
                                                                   •H
                                                                    AJ
                                                                    CO

                        y  I yy
                        A  IAA
                                                            5!
                                                                    O
                                                                    o
                                                                     I
                                                                    JJ
                                                                    m
                                                                    o
                                                                    M-l
                                                                    o

                                                                    4J
                                                                    o
                                                                    o
                                                                   -H
                                                                    JJ
3
 u
w
                                                            s

                                                            S?
 i
u

 0)

                                                            CO
(SO %Z IB) (UJOSP/B)
                   C-l

-------
                         +  "5
I

A •• £:
•^
4"
*

.
T

i
s

qj

-------
                                                         =3


                                                         CM
                                                         &




                                                         5
                                                         CO
                                                      I   5!

                                                         5
          I*   x


        *	iH	*

                                                y ___ l\L[
                                                         3
                                                         55
 CD

 §
•H
.U

 2

 c
 0)
 o

 §
 o
                                                                 U
                                                                 a

                                                                 s
                                                                 o
                                                                 -H
                                                                 OJ
                                                                 03
 O
 U

 .U
 CO
 O
 a

 14-1
 o
 o

"a

 o
-H
4J


 0)

 O
CO
 m

 6

 2
(3O %Z IB)
                  C-3

-------
                                                                                                            CO
V'tf 1 i

H**

>*!*, .
\w

yy[ y

.
T

x x| y

X IKX

* — }*•
*j*
H*

x X)t x *

Y i ..

*j-»
}v y
X A
CM
5
i

m

Cft
<
„

-------
                                                 ***
                                                  T("""T')O(
                                                                     i   5
                                                                         (0
                                                 yy 1  y
4	»
                                                                        CO
 CO
 c
 o
-H



s
 c
 0)
 CJ
 a
 o
 u

  x
                                                                                CJ

                                                                                o
                                                                               -H

                                                                               -U

                                                                                CO
                                                                                O

                                                                                o
                                                                                I

                                                                                4J

                                                                                CQ

                                                                                O
                                                                                (Ct
>4-l

 O



4J

 O

iH

 &



 O

-H

4J
                                                                               3
                                                                                O
                                                                               CO
                                                                                in

                                                                                6
•H

PCI
            (SO %Z IB) (AUJdd) XQN
                              C-5

-------
                                                             C\J
                                                         f   «5

                                                         j_   w


                                                        +   i

                                                             5
                                                             to
D/CDF concentrations
                                                             2!
                                                   •XX-j  X-
                                              *+*
                                                             %


                                                             &
ust ion
o
o
 i
4J
m
O
a,

M-l
O

4J
o
rH
G.

U
•H
4J
                                                                      0)

                                                                      O
                                                                      CO
VO


O

 0)
(so %JL VB} (u»sp/8u)
                       C-6

-------
                                                     cvj
                                                     CD
                                                      03
                                                     2!
2
        n

        §
        •H
        4J
        rt
        C
        0)
        o
        fl
        O
        u

        s
        fl
        O
        •H
        AJ
        tn
        O
         I
        4J
        tn
        o
        O,

        14-1
        o

        4J
        O
        rH
        CU

        U
        •H
                                                              O
                                                              CO
                                                *
S3
                                                 +   &

                                              +-*.  35
                                                              (1)
(2O %/ IB) (iuosp/Bn) po
                  C-7

-------
                                              -3



                                              C\J
                                              5
                                               CO
            CD
            fl
            O
            -H
            4J
            (ti
            H
            4J
            c
            <1)
            O

            8
            O


            fi
                                                         O
                                                        •H
                                                        .U
                                                         m

   ff-1	*
             o
             O
              I
             0->
             CQ
             O
  •«(-)	X-
i   5
             <4-4
             O


             4J
             O
                                                         U
                                                        •H
i   2
   2
                                                         0)
                                                         J3
                                                         o
                                                         w
                                                         GO
                                                          I
                                                         s
                                           Li   co
C-8

-------
                                           T



                                           =5


                                           CJ

                                           CO
                                                   0}
                                                   G
                                                   o
                                                   -H
                                                   fi
                                                   0)
                                                   O
                                                   S
                                                   o
                                                   o
                                                   §
                                                   -H
                                                   AJ
                                                   CO
                                      +
                                                    u
                                         .  5!
CD
O
O,

4-1
O

JJ
O
                                                    0
                                                   -H
                                                   OJ
                                       •*:   2
                                            S3
                                                    o
                                                   CO
                                                   O
                                                    g.
                                                   •H
(u»sp/6n)
       C-9

-------

-------
                APPENDIX D.




GRAPHS OF POST-APCS EMISSION CONCENTRATIONS

-------

-------
                                                      s
                                               -
      m
      a
      o
      -H
      4-)

      S
      4J
      fl
      0)
      o
      c
      o
      u
                                                             CO
CO
                                                              CO
                                                              o
                                                              ft
                                                             JJ
                                                             o
                                                             rH
                                                             a

                                                             o
                                                             -H
                                                              O
                                                             CO
                                                              I

                                                             Q

                                                             0)
                                                             h
(SO %/ © uosp/6)
                D-l

-------
                                                        CJ
                                                               en
                                                               C
                                                               o
                                                              •H
                                                               4-)
                                                               g
                                                               O
                                                               §
                                                               o
                                                         CO
                                                               CJ
                                                               S!
CD
O
                                                               4J
                                                               O
                                                               O
                                                               •H
                                                                O
                                                               CO
*•#-
(SO
                                               o   o
5

2
 I
Q

0)
                                                                Cn
                                                               •H
                 D-2

-------
                                                     C\J
                                                     CM
                                                     m
                                                      CO
                                                   I


                                            •*-]  *  i

                                                             m
                                                              i

                                                             Q



                                                             &
                                                             3
                                                             cn
(SO %L IB) (u»sp/6u)
                    D-3

-------
                                                 t
                                                   I    C\J




                                                  f   i
 tn
 C
 o
-H
 JJ

                                                              0)
                                                              o
                                                              a
                                                              o
                                                              o

                                                              s
                                                              CO
                                                   i '    eo
                                                  T   5
 4J
 CO
 O
                                                              M-l
                                                              o
                                                              o
                                                              rH
                                                              &

                                                              o
                                                              •H
                                                              .U


                                                              (1)

                                                              O
                                                              CO
                                                               I
                                                              Q

                                                              0)
                        8
(ZO %JL © luosp/Sn) po
                D-4

-------
                                               CNJ
                                                      CQ
                                                      G
                                                      O
                                                      •H
                                                      4J
                                                      rt
                                                      H
                                                      4J
                                                      c
                                                      0)
                                                      u
                                                      c
                                                      O
                                                      O

                                                      O)
(so
(iuo«p/ony
                                                      CO
                                                      u
                                                      o<
                                               co
                                                      0)
                                                      O
                                                      .u
                                                      O
                                                      iH
                                                      a

                                                      o
                                                      -H
                                                      4J
                                                      (1)
                                                      X!
                                                      o
                                                      CO
                                                      in
                                                       i
                                                      Q

                                                      0)
                                                      01
      D-5

-------
(SO
uespyfin)
                                          CM
                                      f  i
                                                 i   QQ




                                                   I





                                                 f  I
                                     4
                                          3
                                        !  5!
m
a
o
•H
4J

c
OJ


-------
                                    TECHNICAL REPORT DATA
                               (Please read Instructions on reverse before completing)
 i. REPORT NO.
    EPA-453/R-94-044a
                                                                   3. RECIPIENT'S ACCESSION NO.
 4. TITLE AND SUBTITLE
   Medical Waste Incinerators - Background Information for
   Proposed Standards and Guidelines:  Control Technology
   Performance Report for New and Existing Facilities
                 5. REPORT DATE
                    July 1994
                 6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)
                                                                   8. PERFORMING ORGANIZATION REPORT NO.
 9. PERFORMING ORGANIZATION NAME AND ADDRESS

    Emission Standards Division (Mail Drop 13)
    Office of Air Quality Planning and Standards
    U.S. Environmental Protection Agency
    Research Triangle Park, NC 27711
                 10. PROGRAM ELEMENT NO.
                 11. CONTRACT/GRANT NO.

                     68-D1-0115
 42. SPONSORING AGENCY NAME AND ADDRESS

    Director
    Office of Air Quality Planning and Standards
    Office of Air and Radiation
    U.S. Environmental Protection Agency
    Research Triangle Park, NC 27711
                                                                   13. TYPE OF REPORT AND PERIOD COVERED
                     Final
                  14. SPONSORING AGENCY CODE

                     EPA/200/04
 15. SUPPLEMENTARY NOTES
         Published in conjunction with proposed air emission standards and guidelines for
         medical waste incinerators
 16. ABSTRACT
     This report describes the various emission control technniques used to control emissions from
 medical waste  incinerators (MWI's), summarizes the emission test data generated during a
 comprehensive EPA emission test program, and provides an emission test data analysis that quantifies
 the performance of these techniques.  This is one in a series of reports used as background information
 in developing air emission standards and guidelines for new and existing MWI's.
 17.
                                      KEY WORDS AND DOCUMENT ANALYSIS
                    DESCRIPTORS
                                                  b. IDENTIFIERS/OPEN ENDED TERMS
                                                                                      c. COSATI Field/Group
   Air Pollution
   Pollution Control
   Standards of Performance
   Emission Guidelines
   Medical Waste Incinerators
Air Pollution Control
Solid Waste
Medical Waste
Incineration
 18. DISTRIBUTION STATEMENT

    Release Unlimited
19. SECURITY CLASS (Report)
   Unclassified
21. NO. OF PAGES
      233
                                                  20. SECURITY CLASS (Page)
                                                     Unclassified
                                                                                      22. PRICE
EPA Form 2220-1 (Rev. 4-77)   PREVIOUS EDITION IS OBSOLETE

-------

-------