United States
Environmental Protection
Agency
Industrial Environmental Research EPA-600/2-79-198
Laboratory October 1979
Cincinnati OH 4-5268 •
Research and Development
Hazardous Material
Incinerator Design
Criteria
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S Environmental
Protection Agency, have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and application ol en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields
The nine series are
1 Environmental Health Effects Research
2 Environmental Protection Technology
3 Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 "Special" Reports
9 Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution This work
provides the new or improved technology required for the control and treatment
of pollution-sources to meet environmental quality standards
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED
FROM THE BEST COPY FURNISHED US BY
THE SPONSORING AGENCY. ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE, IT IS BEING RELEASED
IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
EPA-600/2-79-198
2.
3. RECIPIENT'S ACCESSIOWNO.,,
PA JO-ft/76?
4 TITLE AND SUBTITLE
Hazardous Material Incinerator Design Criteria
3". REPORT DATE '
October 1979 issuing date
8. PERFORMING ORGANIZATION CODE
7 AUTHOH(S)
L. Hanson
S. linger
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
TRW
One Space Park
Redondo Beach, CA
10. PROGRAM ELEMENT NO.
IDC 818
90278
11. CONTRACT/GRANT NO.
68-03-2560
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Lab
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
- Cinn, OH
13. TYPE OF REPORT AND PERIOD COVERED
Final Report
14. SPONSORING AGENCY CODE
EPA/600/12
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
The objectives of this task were the evaluation of the available basis for the
prediction of destruction efficiencies of hazardous wastes in large commercial
incinerators and for scaling incinerators to much larger sizes.
A review of major commercial facilities and of waste-facilities matching criteria
led to the selection of the four incinerator types having the widest applicability
for waste destruction. These were the liquid injection, the fluidized bed, the
multiple hearth and the rotary kiln incinerators.
Any prediction of the hazardous components destruction efficiency relies on the
knowledge of the temperature/residence time requirements needed for complete
thermal decomposition. These requirements may be Influenced by the oxygen concen-
tration during the process of decomposition. In Section 4.1 a method utilizing
computerized thermal equilibrium computations and kinetic considerations is
described. Intermediate, potentially hazardous species, can be detected and the
necessary laboratory decomposition/combustion experiments pinpointed.
Recommended steps for the development of a hazardous waste_incinerator scale-up
methodology conclude this report.
17
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Croup
Incineration, incinerators, hazardous
waste, solid waste, industrial waste,
scale up
Incineration, incinera-
tors, hazardous waste,
solid waste, industrial
waste, scale up
68C
13. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (ThaReport/
Unclassified
20 SECURITY CLASS (Thispage/
Unclassified
22 PRICb
EPA Form 2220-1 (9-73)
• iis cntmiinimiinc via iwi-6>7-i4t/9490
-------
EPA-600/2-79-198
October 1979
HAZARDOUS MATERIAL INCINERATOR DESIGN CRITERIA
by
L. Manson and S. Linger
TRW
Redondo Beach, California 90278
Contract No. 68-03-2560
Project Officer
Leo Weitzman
Organic Chemicals and Product Branch
Industrial Environmental Research Laboratory
Cincinnati, Ohio 45268
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory - Cincinnati, U.S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
ii
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on our
health often require that new and Increasingly more efficient pollution con-
trol methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (IERL-C1) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
This report "Hazardous Material Incinerator Design Criteria" is the
first step of a major effort to reduce pollution of our environment by
improperly disposed solid waste. Recent events have shown that serious
problems are associated with traditional waste disposal methods. In order
to implement alternative ways of disposing of industrial wastes, such as high
efficiency incineration, more has to be known about them. The program this
study is part of is intended to take the field of hazardous waste incinera-
tion out of the category of an art and make it more of a science—to increase
its efficiency and decrease cost and energy usage. Further information may
be obtained from the Organic Chemicals and Products Branch.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
m
-------
ABSTRACT
The objectives of this task were the evaluation of the available basis for
the prediction of destruction efficiencies of hazardous wastes in large com-
mercial incinerators and for scaling incinerators to much larger sizes.
A review of major commercial facilities and of waste-facilities matching
criteria led to the selection of the four incinerator types having the widest
applicability for waste destruction. These were the liquid injection, the
fluidized bed, the multiple hearth and the rotary kiln incinerators.
Of these four devices, the liquid injection incinerator enjoys the most
abundant analytical and experimental research and development background.
Though no complete and coherent method for the prediction of specific emissions
levels exists, possible approaches to such predictions are available. Scale-up
is still an uncertain art. However, the art and science of modeling are pro-
gressing and may provide better means for emissions prediction/control and for
scale-up in the not too distant future. The liquid injection incinerator is
discussed in Section 3.2.
Fluidized bed incinerators are the newest arrival in the field of incin-
eration. The intensive development efforts expended to develop clean coal
combustion and coal derived clean fuels technologies will surely provide some
very valuable insights into the waste incineration phenomena. There are as yet
no reliable methods for emissions prediction or scale-up, but an analytical
base is in the process of development and applicable data is becoming more
abundant. Section 3.3 is devoted to the field of fluidized bed development.
Multiple hearth incinerators are widely used, but not much analytical
background was found in the literature. The reasons may well include the
difficulties in analyzing the fundamental mechanisms in this device where the
vapors, combustion gases and solids all follow singularly contorted paths. See
Section 3.4 for a description of fluid and solids flows and of an experimental
method to evaluate solids mixing.
iv
-------
The rotary kiln incinerator is probably the most widely applicable device
since it can handle solids, liquids and sludges (even explosives). Mathematical
treatment of the heat flow to the charge appears possible, as discussed in
Section 3.5. However, information on the mixing in the solids charge and heat
transfer between the charge, the walls and the combustion gases is lacking and
will have to be acquired by appropriate measurements in the field.
Any prediction of the hazardous components destruction efficiency relies
on the knowledge of the temperature/residence time requirements needed for com-
plete thermal decomposition. These requirements may be influenced by the oxygen
concentration during the process of decomposition. In Section 4.1 a method
utilizing computerized thermal equilibrium computations and kinetic considera-
tions is described. Intermediate, potentially hazardous species, can be
detected and the necessary laboratory decomposition/combustion experiments
pinpointed.
A listing of recommended steps for the development of a hazardous waste
incineration methodology conclude this report.
This report was submitted as part of the TESC 68-03-2560, T5006 by TRW
under the sponsorship of the Environmental Protection Agency, IERL ORD
Cincinnati, Ohio 45268. The work was performed between 4 January 1978 and
4 September 1978.
-------
CONTENTS
Foreword iii
Abstract iv
Figures ix
Tables x
1. Introduction 1
2. Commercial Incinerators Survey 3
2.1 Introduction 3
2.2 Incinerator Types 3
2.3 Waste-incineration facility matching criteria ..... 4
2.4 Applicability of the four selected incinerators 5
2.5 Commercial incinerator facilities 7
2.5.1 Liquid waste combustors 7
2.5.2 Fluidized bed incineration 13
2.5.3 Multiple hearth incinerators 17
2.5.4 Rotary kiln incinerators 23
3. Characterization of Incinerators for the Safe Destruction of Hazardous 30
Wastes
3.1 Generalities 30
3.1.1 Combustion and waste destruction efficiencies 31
3.1.2 Continuous, batch and periodic operation 32
3.1.3 Design of large scale incineration units 3^
3.2 Liquid injection waste incineration 3^
3.2.1 Analytical models of liquid injection incinerators 35
3.2.2 Liquid atomization and jet penetration ^
3.2.3 Droplet and spray combustion 43
3.2.4 Turbulence and the unmixedness factor 46
3.2.5 Estimation of hazardous waste destruction inefficiencies . . ^
3.2.6 Summary 49
Preceding page blank
-------
3.3 Fluidized bed incinerators 50
3.3.1 Models of oxidation of single particles 52
3.3.2 Hydrodynamic models 58
3.3.3 Attrition and elutriation 59
3.3.4 Effluent prediction 59
3.3.5 Summary 63
3.4 Multiple hearth incinerators 63
3.4.1 Temperature profiles 65
3.4.2 Turbulence and mixing 66
3.4.3 Residence time of gases and solids 67
3.4.4 Scale-up parameters 68
3.4.5 Summary 68
3.5 Rotary kiln incinerators 69
3.5.1 Combustion and heat and mass transfer 69
3.5.2 Non-dimensional parameters 73
3.5.3 Residence time of gases and solids 75
3.5.4 Scale-up parameters 76
3.5.5 Summary 76
4. Thermochemical and Kinetic Characterization of Wastes 77
4.1 Thermochemical analysis 78
4.1.1 TRW chemical analysis program . ., 78
4.1.2 Equilibrium product distribution analyses 79
4.1.3 Combustion of pesticides -. . . 80
4.1.4 Oxidation of carbon monoxide 86
4.2 The role of thermochemical equilibrium analysis in waste- 88
incinerator matching
5. Summary and Recommendations 91
References 95
Bibliography 98
viii
-------
FIGURES
Number Page
1 Waste and facilities matrix 6
2 Horizontally fired liquid waste incineration system 10
3 Typical vertically fired liquid waste incinerator 12
4 Schematic of a fluidized bed combustor 15
5 Multiple hearth incineration system 20
6 Municipal rotary kiln incineration facility 26
7 Typical major industrial rotary kiln incineration facility 27
8 Perfectly stirred reactor/plug flow reactor model of a liquid
injection incinerator 37
9 Zone model of a liquid injection incinerator 38
10 The unmixedness factor in turbulent flames 47
11 Shrinking core model 54
12 Particle elutriation 61
13 Multiple hearth cross section 64
14 Temperature profile in a multiple hearth incinerator 65
15 Schematic diagram showing the heat-flow paths and nomenclature for a
typical section in a rotary kiln 74
16 Equilibrium mole fraction of product species as a function of tempera-
ture from the combustion of 12 percent lindane emulsifiable concen-
trate with 30 percent excess air 82
17 Equilibrium HC1 concentration in combustion product gas resulting from
the incineration with 30 percent excess air 84
18 Equilibrium Clg concentration in combustion product gas resulting from
the incineration pesticide with 30 percent excess air 85
19 Comparison of experimental, kinetically determined, and equilibrium
values of CO concentrations in the incinerator 3as 89
ix
-------
TABLES
Number Page
1 Liquid Wastes Currently Burned in Liquid Waste Incineration 14
2 Wastes Currently Incinerated in Fluidized Beds 18
3 Typical Combustion Values of Waste Materials 22
4 Standard Multiple Hearth Furnace Size 24
5 Wastes Currently Incinerated in Rotary Kilns 29
6 Definition of Incineration Efficiency Terms 33
7 Summary of Calculated Incineration Efficiencies 33
8 Heat-Transfer Coefficient Correlations 72
9 Equilibrium Product Distribution from the Combustion of Pesticide
Formulations 81
10 High-Temperature Oxidation Rates of Carbon Monoxide 87
-------
SECTION 1
INTRODUCTION
At the present state-of-the-art, incineration or other thermal techniques
appear to be the most likely method available for the large-scale destruction
of hazardous organic wastes. The problem often encountered in its application,
especially by a Regulatory Agency, is how one determines safety in a new situ-
ation. For example, if a given waste is safely incinerated in a small pilot
scale incinerator, how does one determine whether a larger, supposedly similar
unit is capable of achieving the same degree of destruction. This problem
translates into the establishment of scale-up criteria for incinerators.
The development and verification of such criteria would have many benefits
beyond the above-mentioned one. Currently, it is necessary to conduct expensive
large scale tests prior to establishing the safety of an incineration procedure.
Scale-up criteria would permit these tests to be performed on smaller units with
a resultant decrease in cost and at a much greater level of safety. Such
criteria could also be useful in attempts to reduce fuel usages in an incinerator.
Task 5006 was initiated on January 4, 1978 under EPA Contract 68-03-2500
as a first step in addressing these needs. Conceptually, the Task was intended
to examine the experimental and analytical tools available for the characteri-
zation of wastes and incineration devices and recommend which may best lend
themselves to determining usable scale-up criteria. The contractor was instructed
to consider theoretical approaches such as detailed flame modeling only as a tool
to the establishment of such criteria not as an end-point to the research. Tech-
niques such as dimensional analysis were to be examined and recommendations as
to their potential for further research to be made. The task objectives,
formulated to support this goal, were to evaluate the available bases for pre-
dicting the destruction efficiency of hazardous wastes in the four most widely
used types of incineration devices and for scaling these to much larger sizes.
1
-------
The approach taken included the examination of commercial incineration
installations and of the literature pertaining to the combustion of liquids,
solids and sludges in general, and to the incineration mechanisms encountered
in four of the most widely used incinerators in particular. The dichotomy
between the methods used to design commercial incinerators and the methods for
investigating combustion and flow phenomena is readily apparent. Commercial
incinerators are by and large designed'on the basis of past designs and rely
heavily on the talents and experience of the designers. The efforts of the
scientists and engineers interested in incineration are, by contrast, directed
towards uncovering the universal laws controlling incineration and have there-
fore a limited capability to deal with the EPA's specific requirements. This
study examines the "middle ground" between these.
In this report, Section 2 is a survey of commercial incineration facil-
ities; Section 3 characterizes four major types of incinerators - liquid
injection, fluidized bed, rotary kiln and multiple hearth; Section 4 discusses
how the thermochemical properties of the wastes involved affect their incin-
eration and how the application of thermochemical analysis of the wastes can
be used; Section 5 recommends what areas should be explored by further
research into the development of incineration scale-up criteria.
-------
SECTION 2
COMMERCIAL INCINERATORS SURVEY
2.1 INTRODUCTION
In the course of prior studies pertaining to thermal destruction of
hazardous wastes, TRW has identified a large number of wastes generated by
various industries, and identified wastes which could be incinerated. A
systematic classification of wastes and of existing commercial incinerator
installations was undertaken and a matching of wastes and incinerator types
worked out. As a result of that work 50 candidate wastes have been assigned
to 14 facilities. Based on this work we have selected four types of incin-
erators and examined the present state of knowledge of the combustion in these
incinerators. A brief summary of the incineration facilities survey and of
the waste-incinerator matching effort, followed by a brief description of the
selected incinerator types is given in this section.
2.2 INCINERATOR TYPES
The survey of commercial incineration installations showed that the
eight types of incinerators most commonly used wer4
t Liquid injection incineration
• Fluidized bed incineration
• Multiple hearth incineration
t Rotary kiln incineration
• Catalytic combustion
• Molten salt pyrolysis/combustion
t Pyrolysis
• Wet oxidation
We selected for our study, the first four of that list, the most versa-
tile at accepting a wide variety of wastes.
-------
2.3 WASTE-INCINERATION FACILITY MATCHING CRITERIA
In matching different wastes with commercial incineration facilities, the
physical form (solid, liquid, etc.) of the wastes are more important than the
chemical properties. The most important chemical characteristics are the chlo-
rine and sodium contents of the wastes. Wastes containing a large concentration
of halogens can overload the scrubbing equipment on certain incinerators or
require the use of a highly hydrogenated auxiliary fuel, so that HC1, not C12>
is liberated. Sludges containing substantial amounts of sodium can cause
defluidization of fluidized beds by forming low melting eutectic mixtures
(such as NAC1 - Na2C03 or NaCl - Na2S04) . Furthermore, if the particles of
the fluidized bed are silica-sand, Na2SO. will react with the silica to form
a viscous sodium-silicate glass, which will cause rapid defluidization.
The criteria used for matching different wastes to the various inciner-
ation facilities are:
1) Physical form:
Gas, liquid, slurry, sludge, or solid
2) Temperature range required for destruction:
a) >1360 K (2000 F)
b) 1030-1360 K (1400-2000 F)
c) 640 -1030 K ( 700-1400 F)
d) <640 K ( 700 F)
3) Off-Gases:
a) Essentially oxides of carbon and nitrogen, and water vapor
b) Halogen, sulfur, phosphorus, or volatile metal species
4) Ash:
Nonfusible, fusible, or metallic
5) Heating Value:
a) >23 MJ/Kg (10,000 Btu/lb)
b) 12-23 MJ/Kg ( 5,000-10,000 Btu/lb)
c) <12 MJ/Kg ( 5,000 Btu/lb)
Additions to these criteria may become necessary when the list of haz-
ardous wastes is increased in the future.
-------
A detailed description of each of the four selected incinerators will be
found in the section on commercial incinerator facilities (Section 2.5).
2.4 APPLICABILITY OF THE FOUR SELECTED INCINERATORS
Liquid injection, fluidized bed, multiple hearth, and rotary kiln incin-
erators are all widely used to dispose of hazardous wastes. A particular
incinerator may be better suited for incineration of a parLlwular type of
waste based on the physical characteristics of the waste. Solids, sludges,
and slurries of high viscosity liquids can be disposed in rotary kiln, fluid-
ized bed, or multiple hearth incinerators, but not in a liquid injection
incinerator. If the ash resulting from incineration of a waste is fusible,
multiple hearth or fluidized bed incinerators are not well suited for its
disposal. Furthermore, the multiple hearth and fluidized bed incinerators are
not capable of operating at elevated temperatures - so that if a temperature
over 1360 K (2000 F) is needed for destruction, only rotary kiln or liquid
injection incinerators are acceptable. A matrix for matching wastes and
facilities is shown in Figure 1.
In the TRW report, "Destructing Chemical Wastes in Commercial Scale
Incinerators" , fifty wastes are prioritized on the basis of the degree of
their hazard, and the amount produced annually. Eight different incinerators
(liquid injection, fluidized beds, multiple hearth, rotary kiln, wet air
oxidation, catalytic/thermal, pyrolysis, and molten salt incinerators) were
studied as to their applicability in destroying the prioritized wastes in
prior work. A test plan was developed and commercial facilities incorporating
4-11
these incinerators tested . It was concluded that the liquid injection,
rotary kiln, and fluidized bed incinerators had widespread applicability,
whereas the multiple hearth incinerator had moderate applicability. The
liquid injection and fluidized bed offer excellent mixing so that all vaporized
waste can be effectively destroyed. The rotary kiln and multiple hearth incin-
erators utilize mechanical mixing, permitting the solid or liquid wastes to be
exposed to hot oxidizing gas for as long as needed for their destruction.
These incinerators are particularly well suited for the destruction of solids
and highly viscous sludges.
-------
FACILITY TYPE
WASTE CHARACTERISTICS
GAS
LOW VISCOSITY *"
UQU|D (BELOW 500 SSU)
HIGH VISCOSITY
(ABOVE 500 SSU)
LOW VISCOSITY
SLURRY (BELOW 500 SSU)
HIGH VISCOSITY
(ABOVE 500 SSU)
SLUDGE
SOLID FRIABLE POWDER
TARRY
1370" K
TEMP. 1030-1370-K
j^NGE 640-1030-K
FOR 6AO°K
DESTRUCTION cw™
NON FUSIBLE
AS" FUSIBLE
METALLIC
LIQUID
INJECTION
--
MULTIPLE
HEARTH
*
\>s
ROTARY
KILN
FLUIDIZED
BED
-
,
•
I 1 NOT APPLICABLE
•• 500 SSU = 0.00011 M2/S
Figure 1. Waste and facilities matrix.
-------
2.5 COMMERCIAL INCINERATOR FACILITIES
2.5.1 Liquid Haste Combustors
Liquid waste combustors are versatile units which can be used to dispose
of virtually any combustible liquid waste with a viscosity less than 2.2 x 10
2
m /s (10,000 SSU). There are a wide variety of liquid waste combustors presently
12
marketed throughout the manufacturing industries .
Operation Principle—
Before a liquid waste can be combusted, it must be converted to the
gaseous state. This change from a liquid to a gas occurs inside the combustion
chamber and requires heat transfer from the hot combustion product gases to the
injected liquid. In order to effect a rapid vaporization (i.e., increase heat
transfer), it is necessary to increase the exposed liquid surface area. Most
commonly the amount of surface exposed to heat is increased by finely atomizing
the liquid to small droplets, usually to a 40m size or smaller. This atomi-
zation can be achieved mechanically, by two phase flow, or by a combination of
both methods. It is usually achieved in the liquid burner directly at the
point of fuel and air mixing.
Atomization is the heart of any good liquid incinerator. Mechanical
means of atomization include rotary cup and pressure atomization. The rotary
cup consists of an open cup mounted on a hollow shaft. The cup is spun rapidly
and liquid admitted through the hollow shaft. A thin film of the liquid to be
atomized is centrifugally torn from the lip of the cup and surface tension
reforms it into droplets. To achieve conical shaped flames an annular high
velocity jet of air (primary air) must be directed axially around the cup.
If too little primary air is admitted the fuel will impinge on the sides of
the incinerator. If too much primary air is admitted the flame will not be
stable, and will be blown off the cup. For fixed firing rates, the proper
adjustment can be found and the unit operated long periods of time without
cleaning.
Pressure atomizing may take many forms. The familiar garden hose nozzle
is one example. Most commonly the liquid is given a direction by internal
tangential guide slots to the center of the nozzle and then released axially
through an orifice. Good atomization can be achieved at moderate pressures
-------
0.68 MPa to 1.0 MPa (100 to 150 psi). Disadvantages include a limited variable
flow range at low pressures and, especially in the smaller sizes, a tendency to
plug with foreign matter. Large sizes are reasonably free from this problem.
Liquid burners require considerably more turbulence and time to complete
combustion than do gas burners. To complete combustion good mixing of the
liquid spray and air is needed, and the larger the particle, the greater
distance they will go before being completely vaporized and burned. Forced
draft units, if well designed, will result in higher air velocities and there-
fore will have better combustion characteristics than natural draft units.
Burners must be located to prevent flame impingement on walls and, in the case
of multi-burner units, interference with one another. While multiple atomizers
can be located within a single air register, the performance will suffer, and
combustion volume must be added to offset this characteristic. Whenever
possible, the number of liquid streams should be minimized.
Liquid streams can carry impurities of every sort. Futhermore, they
may be highly viscous, which makes handling and atomizing difficult. Liquids
o o
should generally have a viscosity of 2.2 x 10 m /s (10,000 SSU) or less to
be satisfactorily pumped and handled in pipes. For atomization, they should
have a maximum viscosity of 1.6 x 10"4 m2/s (750 SSU). If the viscosity
exceeds this value the atomization may not be fine enough, and the resultant
droplets of unburned liquid may cause smoke or other unburned particles to
leave the unit. Viscosity can usually be controlled by heating with tank
coils or in-line heaters. Should gases be evolved in any quantity before
the desired viscosity is reached, they may cause unstable fuel feed and
burning. If this occurs, the gases should be trapped and vented safely,
either to the incinerator or elsewhere. If preheating is not feasible, a
lower viscosity and miscible liquid may be added to reduce the viscosity of
the mixture.
Prior to heating a liquid waste stream, a check should be made to
insure that undesirable preliminary chemical reactions such as polymer-
ization, nitration, oxidation, etc., will not occur. Should these occur, it
may be more desirable to fill disposable containers with the liquid and
treat them as solids. Other preparatory steps may include filtration, de-
gassing, pressurizing, neutralizing, storage, mixing, etc. In every one of
8
-------
these steps care must be employed to see that undesired and harmful results
do not occur. Pump and piping materials of construction must be suitable
for the liquids encountered. Liquids that can solidify or become too viscous
should have jacketed or traced piping. Provision should be made to clean out
the piping and equipment when long shutdowns occur. This Is usually done by
purging with steam. Certain atomizing nozzles should always be blown clear
with steam whenever flow 1s stopped. If not, the residual heat In the Incin-
erator may cause thermal cracking of the liquid remaining in the nozzles,
resulting 1n partial or complete pluggage.
Process Design-
Liquid waste incinerators can be vertically or horizontally fired units.
Their operating temperatures range from 920 K (1200 F) to 1920 K (3000 F)
(most units operate around 1140 K (1600 F) and residence times range from one
half to two seconds. Most units have combustion chamber volumes which provide
for a heat release of approximately 0.25 MW/m (25,000 Btu/hr-ft ), however,
the vortex type liquid combustor has an unusually high heat release of about
1.0 MW/m3 (100,000 Btu/hr-ft3).
A typical horizontally fired liquid waste incineration system is pre-
sented (Figure 2). This particular system is the one operated by the Dow
Chemical Company at their Midland, Michigan facility. The unit is a 24 MW
(81 million Btu/hr) incinerator which has a combustion chamber 10.6m (35 ft.)
2 2
long and 0.93 m (10 ft ) in cross section. Wastes are fed to the unit
through a combination of four dual-fired nozzles. Combustion gases are
quenched in a spray chamber, followed by a high-pressure-drop venturi scrubber,
o 3
and a cooler/mist-eliminator. About 6.3 x 10 m /s (1000 gpm) of water is
recycled from the primary tanks to the wastewater treatment facilities to
furnish scrubbing water. This water flows back to the wastewater plant for
treatment. About 1,100 hp. is required for this unit.
The majority of the liquid wastes treated in the Dow unit are solids at
room temperature and must be kept hot in order to remain liquid. Many residues
are chlorinated and can contain as high as 50 percent chlorine, plus several
percent of ash in the form of ash in the form of Fe, Ca, Mg, Na, oxides and
chlorides.
-------
LIQUID WASI '• FROM PLANT
STRAINER
SEPARATE TANKS FOB
HIGH AND LOW
MELTING-POINT LIQUIDS
STACK 100 FT HIGH
6 fl. 6 IN I. D.
4 FT 6 IN. I. D OUILFT
UMED WITH ACID-RI LISTING
PLAST 1C
WASTE-TAR
FEED
NATURAL
GAS
ATOMIZING
BLOWER
RELIEF
STACK
(CLOSED
DURING
OPERATION)
UMPERING
AIR BLOWER
/^^ 10,000
fOJ CU. FT./MIN.
IT "HP
VEMtURI SCRUBBER LINED WITH
ACID RESISTING PIAS11C
RECYCLED
WASTE
WATER
FRESH WATER \ 1.300 GPM.
300 GPM.
WASTE
WATER
1.000 GPM.
)
COMBUSTION AIR BLOWER
13,000 CU. FT./MIN.
75 HP.
TOTAL AIR, 76 LB./LB WASTE
TEMPERING
AIR BLOWER
10,000
CU. FT MIN.
25 HP.
75F-
WAIER
?. 300 GPM.
pH I.Q
INDUCED-DRAFT FAN
7,600 IB. /MIN
15,000 CU. FT /MIN.
600 HP
WASTE TAR FEED AVG. 10 GPM.
13,00 BTU. LB.
TEMPERATURE 80-1 (HOC
VISCOSITY 150 SSU
5PSI FIFO
4 BURNERS. COMBUSTION
GAS AND TAR MOZZLIS
5'16 - IN ORIFICt
Figure 2. Horizontally fired liquid waste incineration system.
(English systems of units in the actual commercial
incineration system).
-------
A typical vertically fired liquid waste incinerator is presented (Figure
3). This particular unit is designed and marketed by the Prenco Division of
Pickands Mather and Company. It is claimed to be a versatile system in that
it can be brought up to operating temperature in one to two hours with minimal
fuel requirements. This quick warm-up permits periodic rather than continual
operation.
i
The Prenco vertical combustor operates in the following manner. A
mixture of auxiliary fuel (usually natural gas) and high pressure air are
first fed into the vertical retort to bring it up to proper waste decomposi-
tion temperature. When the retort reaches the correct temperature, as
determined by the temperature measuring instruments, fuel flow is modulated
and waste is admitted to the air-waste entrainment compartment. From there
the aerated waste is fed into a turbulence compartment where it is mixed with
more high pressure air and injected into the high-temperature vertical retort.
Here the process breaks down the waste by molecular dissociation, oxidation,
and ionization. The gases and any inert particles produced flow vertically
through the air cone and out of the top of the retort.
Decomposition efficiency is greatly increased through the injection of
pressurized air at a point near the top of the retort through ports in a
specially designed refractory module. The air cone, which serves as a fuel
saver, increases decomposition efficiency by increasing heat retention. It
also provides additional air for an afterburner effect. In addition the
air cone reduces the temperature of the decomposed effluent to about 620 K
(650 F). As a result, scrubbers and effluent test equipment can be utilized
if desired.
The Prenco unit utilizes air preheat. Intake of air from the top of the
upper nacelle causes it to be preheated as it travels down the outer wall of
the decomposition chamber to both the turbo-blower and afterburner fans.
The use of preheated air significantly increases decomposition efficiency and
economy of operation.
Process Applicability-
Liquid waste incinerators are generally applicable to the ultimate dis-
posal of most forms (including dilute) of combustible liquid waste materials
11
-------
FREE STANDING
INTERLOCKING REFRACTORY
MODULES
TEMPERATURE MEASURING
INSTRUMENTS
EFFLUENT DIRECTLY TO ATMOSPHERE FPF<;H AIR IWTAKF
OR TO SCRUBBERS AND STACK . FQR TURBO -BLOWER
AND AFTERBURBER FAN
AIR CONE
TURBO-BLOWER
IGNITION CHAMBER
HIGH VELOCITY
AIR SUPPLY
AIR-WASTE ENTRAINMENT
COMPARTMENT
WASTE LINE
UPPER NACELLE
DECOMPOSITION CHAMBER
DECOMPOSITION STREAM
AFTER-BURNER FAN
FLAME SENSITIZER
TURBULENCE COMPARTMENT
LOWER NACELLE
AUXILIARY FUEL LINE
TUBULAR SUPPORT COLUMNS
ELECTRICAL POWER LINE
Figure 3. Typical vertically fired liquid waste incinerator.
-------
and represent proven technology. Some of the materials currently being dis-
posed with this type incinerator are presented (Table 1).
2.5.2 Fluidized Bed Incineration
Fluidized bed incinerators are versatile devices which can be used to
dispose of solid, liquid and gaseous combustible wastes. The technique is
a relatively new method for ultimate disposal of waste matei ia'is. It was
first used commercially in the United States in 1962 and has found limited
use in the petroleum industry, paper industry and for processing nuclear wastes.
In addition, fluidized bed combustion has been applied for the disposal of sani-
12
tary sludge .
Operating Principle—
A typical fluidized bed incinerator is shown schematically (Figure 4).
Air driven by a blower enters a plenum at the bottom of the combustor and
rises vertically through a distributor plate into a vessel containing a bed
of inert granular particles. Sand is typically used as the bed material. The
upward flow of air through the sand bed results in a dense turbulent mass which
behaves in a way that is similar to that of an air flow through a liquid.
Waste material to be incinerated is injected into the bed and combustion occurs
within the bubbling bed.
Air passage through the bed produces strong agitation of the bed particles.
This promotes rapid and relatively uniform mixing of the injected waste material
within the fluidized bed.
The mass of the fluidized bed is large in relation to the injected material,
Bed temperatures are quite uniform and typically in the 1030 to 1140 K (1400 to
1600 F) range. At these temperatures, heat content of the fluidized bed is
3 3
approximately 600 MJ/m (16,000 Btu/ft ) thus providing a large heat reservoir.
By comparison, the heat capacity of flue gases at similar temperatures is three
orders of magnitude less than a fluidized sand bed.
Heat is transferred from the bed into the injected waste materials to be
incinerated. Upon reaching ignition temperature (which takes place rapidly)
the material combusts and transfers heat back into the bed. Continued bed
agitation by the fluidizing air allows larger waste particles to remain
13
-------
TABLE 1. LIQUID WASTES CURRENTLY BURNED IN LIQUID
WASTE INCINERATION
Separator sludges
Skimmer refuse
Oily waste
Detergent sludges
Digester sludges
Cutting oils
Coolants
Strippers
Phenols
Wine wastes
Potato starch
Vegetable oils
Washer liquids
Still & reactor bottoms
Soap & detergent cleaners
Animal oils & rendering fats
Cyanide & chrome plating wastes
Lube oils
Soluble oils
Polyester paint
PVC paint
Latex paint
Thinners
Solvents
Polymers
Resins
Cheese wastes
Dyes
Inks
Off-specification isoprene
Hexachlorocyclopentadi ene
Organophosphate pesticides
Waste from polymer polyol production
Dodecyl mercaptan wastes
Fluorinated herbicide wastes
Ethylene glycol manufacture residue
Waste residues from alkyl benzene
production
Perch!oroethylene manufacture
still bottoms
Alkyl and oryl sulfonic acid wastes
Still bottom from acetaldehyde
production.
14
-------
FLUE GAS
MAKEUP SAND
ACCESS DOOR
:-:• •'
ISAND BEDl
AUXILIARY
BURNER (OIL OR GAS)
LJ—U—LI—LJ—'•
i i
V
ASH REMOVAL
WASTE INJECTION
FLUIDIZING AIR
Figure 4. Schematic of a flin'dized bed combustor.
15
-------
suspended until combustion is completed. Elutriated fines are carried off the
bed by the exhausting flue gases at the top of the combustor. These gases are
subsequently processed and/or scrubbed before atmospheric discharge.
Process Design-
In specifying or designing a fluidized bed combustor, primary factors to
be considered are: gas velocity; bed diameter; bed temperature; and the type
and composition of waste to be incinerated.
Gas velocities are typically low, in the order of 1.5 to 2.4 m/s (5 to 8
ft/s). Maximum gas velocity is constrained by the terminal velocity of the bed
particles and is therefore a function of particle size. Higher velocities re-
sult in bed attrition and an increased particulate load on downstream air
correction equipment. Relatively low velocity reduces pressure drop and there-
fore lowers power requirements, but increases the size of the equipment.
The largest fluidized beds are on the order of 15 m (50 ft) in diameter.
At nominal values of gas velocity and temperature, the maximum volumetric flow
would be approximately 1200 m3/s (2.5 x 10 acfm).
Bed depths range from about 0.4 m (16 inches) to several feet. Variations
in bed depth affect waste particle residence time and system pressure drop.
One therefore desires to minimize bed depth consistent with complete combustion
and minimum excess air.
The type and composition of the waste is a significant design parameter
in that it will impact storage, processing and transport operations (prior to
incineration), as well as the combustion. If the waste is a heterogeneous
mixture such as municipal refuse and has a relatively low <19 MJ/kg (<8000
3tu/lb) heating value, processing (shredding, sorting, drying, etc.) operations
will be more complex and auxiliary fuel addition to the combustor will be
required.
Process Applicability-
Many of the fluidized bed incineration applications involve the disposal
of sludges or slurried wastes. This may necessitate a dewatering step in pro-
cessing the water prior to incineration if combustion gases are to be used for
steam-electric or gas turbine power generation. If power generation is a
16
-------
desired by-product of the incineration process, then waste moisture content
values less than approximately 60 percent are required. Moisture values in
excess of this value, or heavy concentrations of inert matter will require
auxiliary fuel burners to preheat the waste and ensure sufficient heat content
in the flue gases. Pre-drying of the sludge may be accomplished by aeration
or more sophisticated mechanical systems involving the addition of heat.
Table 2 shows a list of wastes currently incinerated in fluidized bed facili-
ties.
Waste material is pneumatically, mechanically or gravity fed into the
fluidized bed. Normally, inhomogeneous waste material must be reduced in
size (shredded, pulverized, etc.) to facilitate the feed system operation
and permit injection, distribution and combustion within the fluidized bed.
Advantages
1) The combustor design concept is simple and does not require
moving parts in the elevated temperature regions of combustion.
2) Designs are compact due to high volumetric heating rates 1.0
MW/m3 to 2.0 MW/m3 (100,000 to 200,000 Btu/hr-ft3) resulting
in lower capital investment.
3) Comparatively low gas temperatures and excess air requirements*
minimize formation of nitric oxide and the cost of equipment.
Limitations
1) Bed diameters are limited with present design technology;
therefore, maximum volumetric flow rates per unit are limited.
2) Removal of inert residual material from the bed is a potential
problem area.
2.5.3 Multiple Hearth Incinerators
The multiple hearth incinerator (commonly called a Herreshoff furnace)
is a versa!ite unit which has been utilized to dispose of sewage, sludges,
tars, solids, gases, and liquid combustible wastes. This type of unit was
initially designed to incinerate sewage plant sludges in 1934. In 1968,
there were over 125 installations in operation with a total capacity of
For example, excess air requirements as low as 5 percent have been re-
ported in the combustion of coal in fluidized bed reactors. Low excess
air requirements reduce the size and cost of gas handling equipment.
17
-------
TABLE 2. WASTES CURRENTLY INCINERATED IN FLUIDIZED BEDS
Off-specification phenol
Ami ben manufacture liquid wastes
Corboryl manufacture waste
Ethylene manufacturing wastes
Waste from toluene diamine production from
dinitrotoluene
Tetraethyl orthosilicate wastes
Organic wastes from pharmaceutical manufacture
Organic peroxide manufacturing wastes
Ethylene bromide manufacturing wastes
Urethane manufacture wastes
18
-------
180 kg/s (17,000 tons per day) twet basis) for this application alone. There
are currently numerous industrial installations in operation which are pri-
marily utilized for chemical sludge and tar incineration as well as activated
12
carbon regneration .
Operation Principle—
The multiple hearth furnace consists of a refractory-lined circular steel
shell with refractory hearths located one above the other (Figure 5). Sludge
and/or granulated solid combustible waste feeds through the furnace roof by
a screw feeder or belt and flapgate. A rotating air-cooled central shaft with
air-cooled rabble arms and teeth plows the waste material across the top hearth
to drop holes. The waste falls to the next hearth and then the next until ash
discharged at the bottom. The waste is agitated as it moves across the hearths
to make sure maximum surface is exposed to hot gases. Waste grease and tars
are generally fed into the furnace through side ports.
Liquid and gaseous combustible wastes may be injected into the unit
through auxiliary burner nozzles. This utilization of liquid and gaseous
waste represents an economic advantage since the secondary fuel (e.g., natural
gas, fuel oil) requirements will be reduced thus lowering oprating costs.
The system has three operating zones: the top hearths where feed is
dried to about 48 percent moisture; the incineration/deodorization zone,
which has a temperature of 1030 to 2260 K (1400 to 1800 F); and the cooling
zone, where the hot ash gives up heat to incoming combustion air. Exhaust
gases exit at 530 to 860 K (500 to 1100 F).
Incineration ash is sterile and inert. Volume discharged from the
bottom hearth is about 10 percent of the furnace feed, based on sludge cake
with 75 percent moisture and 70 percent volatile content in the solids.
The ash usually has less than 1 percent combustible matter, which is normally
fixed carbon. Discharge can be moved hydraulically, mechanically, or
pneumatically, and used as landfill or roadfill.
Current systems include gas cleaning devices on exhaust air. A number
of multiple hearth incinerators are operating without difficulty in areas
19
-------
WASTE AIR TO
ATMOSPHERE
CLEAN GASES TO
ATMOSPHERE
VACUUM
FILTERS
SLUDGES
FILTRATE -•
ro
o
GREASE AND TARS
BURNERS
(FUEL OIL, GAS,
LIQUID AND GASEOUS WASTE) *"IT=i.
AIR
AIR RECYCLE
T w^ , BYPASS
DAMPER
AIR
INDUCED
DRAFT FAN
SCRUBBERS
WATER
ASH TO
BLOWER DISPOSAL
ASH SLURRY TO FILTRATION AND
ASH DISPOSAL
Figure 5. Multiple hearth incineration system.
-------
with strict air pollution codes. Although the exhaust does not violate opacity
codes, existence of steam plumes has on occasion caused adverse public reaction.
Process Design--
Most multiple hearth incinerators are primarily designed for sludge dis-
posal. The other forms of waste which are simultaneously fed to the system are
usually considered a heat source to be utilized during sluage incineration. A
heat balance across a multiple hearth furnace must consider the heat absorbed
by: latent heat in free moisture and combustion moisture, sensible heat in
combustion gases, excess air, ash, radiation and shaft cooling. These quanti-
ties are balanced against the heat evolved from the combustibles in sludge
solids and the fuel. Below is a typical analysis of sewage sludge combustibles.
C 59.8 percent
H2 8.5
02 27.5
N2 4.2
100.0 percent
Calorific value of this sludge is 23 MJ/kg (10,000 Btu/lb).
Sludge parameters that have the most influence over incineration are
moisture content, percent volatiles and inerts, and calorific value.
Moisture is the principal one over which the plant operator has some
control. Minimum moisture is important because of its thermal load on the
incinerator.
Volatiles and inerts, which affect the calorific value of the sludge,
can be controlled to some extent by treatment processes such as degritting,
mechanical dewatering and sludge digestion. Almost all combustibles are
present as volatiles, much in the form of grease. Volatile percentage
can vary a great deal, so equipment must be designed to handle a range of
values.
The sizing of a multiple hearth incinerator is dependent upon waste
combustion characteristics (Table 3) and water content. Incineration
burning rates vary from 0.009 kg/in2-s to 0.016 kg/m2-s (7 to 12 Ib/ft -hr)
for sewage plant sludges with the value 0.01 kg/m2-s (7.5 Ib/ft -hr) generally
21
-------
TABLE 3. TYPICAL COMBUSTION VALUES OF WASTE MATERIALS
Material Combustible.%
Grease & scum
Fresh sewage solids
Fine screenings
Ground garbage
Rags
Digested sewage and garbage
solids
Digested sludge
Grit
88.5
74.0
86.4
84.8
97.5
49.6
59.6
33.2
ASM
11.5
26.0
13.6
15.2
2.5
50.4
40.4
69.8
MJ/kg (Btu/lb.)
39.0 (16,750)
23.9 (10,285)
20.9 (8,990)
19.2 (8,245)
18.7 (8,050)
18.7 (8,020)
12.3 (5,290)
9.3 (4.000)
Note: Where organic polymers can be utilized to condition sludges, rather
than ferric chlorides and lime, the heat value of the sludge cake
can be increased from 3.5 to 9.3 MJ/kg (1,500 to 4,000 Btu/lb) of
dry solids. The ash from the furnace will also he reduced by 5 to
10 percent.
22
-------
accepted as typical. The area referred to in the burning rate is the total
hearth area of the unit. Standard multiple hearth incineration sizes range
from 7.9 m2 (85 ft2) of hearth to greater than 278 m2 (3000 ft2) of hearth
(Table 4). The secondary fuel requirement is dependent upon the water content
of the waste being incinerated. For instance, a waste sludge with a heating
value of 23.2 MJ/kg (10,000 Btu/lb) of volatile solids which is composed of
3
60 percent volatile solids, will require about 3 m of natural gas per thou-
sand kg (100 ft /ton) of wet feed when the moisture content of the sludge is
3
75 percent. The same sludge will require about 37 m of gas per thousand kg
(1200 ft /ton) of wet sludge when the moisture content is 82.5 percent.
The multiple hearth incinerator is usually operated so that the top
hearth temperature is in the 590 to 810 K (600 to 1000 F) range, the combus-
tion hearths are in the 1030 to 1260 K (1400 to 1800 F) range, while the
cooling hearths are maintained in the 480 to 590 K (400 to 600 F) range.
2.5.4 Rotary Kiln Incinerators
Rotary kiln incinerators are versatile units which can be used to dispose
of solid, liquid and gaseous combustible wastes. They have been utilized in
both industrial and municipal installations. In addition, applications of
rotary kiln incineration to the disposal of obsolete chemical warfare agents
12
and munitions have been reported .
Operation Principle--
The rotary kiln incinerator is a cylindrical shell lined with firebrick
or other refractory and mounted with its axis at a slight slope from the
horizontal. It is a highly efficient unit when applied to solids, liquids,
sludges and tars because of its ability to attain excellent mixing of unburned
waste and oxygen as it revolves. Its use as a concentrated waste gas combustor
is considered a secondary application. This is due to the fact that although
proper conditions are present for efficient gas combustion (i.e., long resi-
dence time at elevated temperatures) there is no need for the cylinder to be
rotating. Therefore rotary kiln incinerators are used for gaseous waste
combustion only in conjunction with solid or liquid waste incineration.
Rotary kiln incinerators used in municipal applications are generally
desgined to handle large volumes of solid combustible waste (refuse) along
23
-------
TABLE 4. STANDARD MULTIPLE HEARTH FURNACE SIZE
Outsiae
diameter
«.£ ft*
(ID)
7.0 ft
8.5 ft
10.0 ft
13.5 ft
16.0 ft
18.0 ft
19.5 ft
21.5 ft
Hearth ares.sq ft
Column height, ft-ln.
Shell height, ft-in.
Overall height, ft-in.
Hearth area, sq ft
Column heignt, ft-in.
Shell height, ft-in.
Overall heignt, ft-in.
Hearth area, sq ft
Column height, ft-in.
Shell heignt, ft-in.
Overall heignt, ft-in.
Heartn area, sq ft
Column height, ft-in.
Snell height, ft-1n.
Overall height, ft-1n.
Hearth area, sq ft
Column height, ft-in.
Shell height, ft-in.
Overall height, ft-in.
Hearth area, sq ft
Column height, ft-in.
Snell height, ft-in.
Overall heignt, ft-in.
Hearth area, sq ft
Column height, ft-in.
Shell height, ft-1n.
Overall height, ft-1n
Heartn area, sq ft
Column heignt, ft-1n.
Shell height, ft-1n.
Overall height, ft-in.
Hearth area, sq ft
Column height, ft-in.
Shell height, ft-in.
Overall height, ft-in.
4
Hearth
130
5-0
10-10
16-7
188
6-6
10-8
18-8
390
6-6
11-8
20-8
b73
7-0
13-2
22-11
727
7-0
H-3
24-3
863
8-0
14-4
25-8
1077
8-0
16-1
27-9
6
Hearth
85
4-0
10-6
15-7
125
4-0
11-10
16-1
193
5-0
15-5
21-2
276
6-6
15-1
23-0
575
6-6
16-7
25-6
845
7-0
18-7
28-4
1068
7-0
20-2
30-1
1268
8-0
20-2
31-7
1580
8-0
22-9
34-6
8
Hearth
112
4-0
13-8
18-9
166
4-0
15-5
20-6
256
5-0
20-0
25-9
364
6-6
19-5
27-5
760
6-6
21-5
30-5
1117
7-0
24-1
33-10
1410
7-0
26-0
36-0
K60
8-0
26-1
37-5
2084
8-0
29-6
41-2
10
Hea-tn
143
4-0
16-10
21-11
208
4-0
19-0
24-1
319
5-0
24-7
30-4
452
6-6
23-10
31-9
944
6-6
26-4
35-3
1305
7-0
29-6
3d-3
1752
7-0
31-11
41-10
2060
8-0
31-11
43-4
2570
8-0
36-2
47-11
12
Hes-th
1128
6-6
31-2
40-2
1550
7-0
35-0
44-9
2090
7-0
37-9
47-9
2464
8-0
37-10
49-2
3046
8-0
42-11
54-7
English units are used in this table because the furnaces sizes in the United
States are quoted in these units. See Unit Conversion Table.
24
-------
with any entrained liquid. In this type of facility, the kiln actually serves
as a secondary combustion unit since the waste material is ignited on traveling
grates prior to entering the kiln (Figure 6). In this instance, the kiln
serves mainly as an efficient mixer of the burning waste with combustion air.
Rotary kiln incinerators when applied to industrial applications are
generally designed to accept both solid and liquid feed. A typical major
industrial installation is operated by the Dow Chemical Company at Midland,
Michigan (Figure 7). This particular unit consists of a 19 MW (65 million
Btu/hr) kiln that is used for the incineration of solid chemical refuse,
liquid residues, paper, wood and other solids of varying calorific content. A
pack-feed mechanism is used to feed packs and drums of solid waste chemicals
into the incinerator.
Liquid wastes transported to the incinerator are transferred to des-
ignated receiving tanks that contain compatible wastes. All drums of
liquid wastes are also transferred to the receiving tank by the way of a
drum-dumping dock. The waste is strained as it is pumped from the receiving
tank into a burning tank, where it is blended for optimum burning charac-
teristics. All liquid residues are burned in suspension by atomization with
steam or air.
Drum quantities of solid tars are destroyed by feeding them into the
rotary kiln incinerator via a hydraulically operated drum and pack-feeding
mechanism. All refuse, except full drums and packs of material, is dumped
into the refuse pit. An overhead crane is used to mix the refuse and
raise it to the charging hopper of the rotary kiln (see Figure 7).
While the solid refuse is being fed, liquid tars are fired horizontally
into the rotary kiln. As the refuse moves down the kiln, organic matter
is destroyed, leaving an inorganic ash. This ash is made up primarily of
slag, and other nonburnables such as drums and other metallic material.
The ash discharges from the end of the kiln into a conveyor trough that
contains water. After quenching, the material is conveyed into a dumping
trailer, and then to a landfill.
25
-------
CHARGING
CHU1E
ro
cr.
TO EXPANSION CHAMBER
AND GAS SCRUBBER
RESIDUE CONVEYORS
Figure 6. Municipal rotary kiln incineration facility.
-------
j TAR PUMPING
FACILITY
PACK STORAGE AND
DEEDING FACILITY
SCRAP METAL
FLY ASH
RESIDUE
Figure 7. Typical major industrial rotary kiln incineration facility.
-------
After leaving the kiln, the products of combustion enter the secondary
combustion chamber and impinge on refractory surfaces that cause a swirling
action. No secondary fuel or afterburners are used. Downstream of the
secondary combustion chamber, the gases pass through several banks of water
sprays in which the flyash in knocked down and slurried into the ash-conveyor
floor. Cooled gases pass under a stack damper and then a 61 m (200 ft) stack.
Process Design-
Specific data on rotary kiln incinerator design parameters are scarce.
This is due to the fact that incineration is a relatively new application for
rotary kilns. Additionally, information of this type is generally considered
proprietary by manufacturers.
Information sources indicate that rotary kiln incinerators generally
have a length to diameter ratio (L/D) between 2 and 10. Smaller L/D ratios
result in less particulate carry over. Rotational speeds of the kiln are
usually much slower than those for kilns which are utilized as calciners
or dryers and are on the order of 5 to 25 millimeter per second (1 to 5
feet per minute) measured at the kiln periphery. Both the L/D ratio and the
rotational speed are strongly dependent upon the type of waste being com-
busted. In general, larger L/D ratios along with slower rotational speeds
are used when the waste material requires longer residence times in the
kiln for complete combustion.
The residence time and combustion temperature required for proper in-
cineration is totally dependent upon the waste materials combustion char-
acteristics. Combustion temperatures usually range froir 1140 to 1920 K
(16?0 to 3000 F). Required residence times vary from seconds to hours.
For instance, a finely divided propellent may require 0.5 seconds while
wooden boxes, municipal refuse, and railroad ties may require 5, 15 and 60
minutes respectively.
28
-------
Process Applicability—
The rotary kiln incinerator is generally applicable to the ultimate dis-
posal of any form of combustible waste material and represents proven technology.
It can incinerate combustible solids (including explosives), liquids (including
chemical warfare agents), gases, sludges and tars. Table 5 gives a partial
list of wastes currently incinerated in rotary kiln.
TABLE 5. WASTES CURRENTLY INCINERATED IN ROTARY KILNS
Epichlorohydrin manufacturing wastes
Steam still bottoms from aniline and
alkylated phenol production
Acryonitrile manufacturing wastes
Reactor tar bottoms from adrysonitrile manufacture
Phenolic tar from 2,4-D manufacture
Chlorotoluene production wastes
Phenylamine tar wastes
PCB wastes in capacitors
Evaporate residue from the cumere process for phenol
manufacture
Nitrochlorobenzene tars
Catch basin grease, nitrile pitch from production
of surface active agents
TDI manufacture reactor tar bottoms
Diphenylamine manufacture reactor for bottoms
Mercaptobenzothiazale tars
29
-------
SECTION 3
CHARACTERIZATION OF INCINERATORS FOR THE
SAFE DESTRUCTION OF HAZARDOUS WASTES
3.1 GENERALITIES
The safe destruction of hazardous wastes by incineration depends on the
exposure of the hazardous components to oxygen at sufficiently high tempera-
tures for a sufficient time to promote either combustion or decomposition
These are also the conditions that lead to complete combustion of fuel.
Thus, safe waste destruction, exactly like all combustion, depends on
temperature, residence time and turbulence (or mixing of reactants). Tempera-
ture levels and residence time needed for the destruction of any given specie
can be determined in the laboratory. The capability of any given incinerator
to provide an adequately high temperature and residence time (including
generous safety margins) can also be established, through measurements and
calculations. When necessary, auxiliary fuel will be burned to insure
desired temperature levels, and waste, fuel and air flow rates will be
adjusted to meet the residence time requirements, on the average. Evaluation
and control of mixing are more difficult, because the overdesign or "safety
factor" approach is not readily applicable. More sophisticated approaches
than overdesign are needed to ensure the desired destruction efficiency for
a given waste or waste component in a given incinerator, and to permit scale-
UD of proven designs to a larger capacity system. Mathematical modeling of
the heat and mass exhange mechanisms and of reactant flows has been attempted
in some cases, with limited success. Collection of voluminous data and
application of regression analysis and other statistical methods have also
been proposed. The sheer size of a program necessary to obtain models
adequately representing a system as complex as an incinerator burning non-
gaseous material can be imagined by recalling the massive effort expended in
modeling liquid propellent combustion in rocket motors - a much simpler
system than a waste incinerator. On the other hand, a "data crunching" approach,
30
-------
when not supported by some understanding of the underlying laws of physics,
demands an inordinately large amount of data in order to have a chance of
succeeding.
So far the most promising avenue is a judicious mixture of simple
mathematical models and data to "tune" these models. By gradually developing
the analytical tools (in conjunction with experimental R&D) and acquiring a
data base from measurements obtained on industrial systems one can expect to
arrive at useful scale-up laws and reliable prediction of hazardous waste
destruction efficiencies.
The development of analytical tools mainly consists of the study and
description of physical phenomena in terms of appropriate non-dimensional
parameters, involving the expression of basic laws or the establishment of
correlations. For instance, the combustion of liquid fuels and wastes depends
on liquid atomization and jet mixing. Correlations in this field are avail-
able and can be used to evaluate liquid destruction inefficiency and to
estimate liquid jet penetration for a scaled-up system. On the other hand,
correlations permitting the evaluation of the degree of mixing in a fluidized
bed are not as yet established. In the following discussion of the four
incinerator types, the present state of knowledge is summarized and the
possibilities of using what knowledge is available to evaluate the incin-
erator effectiveness and the approach to scale-up of existing designs are
indicated.
3.1.1 Combustion and Waste Destruction Efficiencies
The goal of hazardous waste incineration is the complete destruction of
hazardous species, which is related to, but not identical with, the complete
combustion of the fuel and of the combustible waste components. Sufficient
information on the destruction efficiency of hazardous species in an incin-
erator is most often not available. Therefore, in the absence of directly
applicable data, the overall combustion efficiency and other, experimentally
accessible, destruction efficiencies may be useful for the purpose of
predictive modeling or other approaches to prediction of hazardous species
destruction. Since the ability to define a destruction efficiency in terms
of measurable quantities is crucial, and since hazardous waste components
31
-------
are many and varied, the definition of destruction efficiencies will also vary
to some extent.
A useful example of a generally applicable set of definitions is shown
in Tables 6 and 7. These are the definitions adopted by TRW for the incin-
eration at sea of the Herbicide Orange . The overall combustion efficiency
is defined in terms of CO,, and CO; both are measurable quantities. Hydrogen
and sulfur analysis were not included, because neither of these combustibles
were important in the case of Herbicide Orange (HO) incineration. Note that
the destruction efficiencies of HO as well as that of the two toxic con-
taminants chlorinated hydrocarbon (CHC) and of 2,3,7,8 - tetra chlorodibenzo-
p-dioxin (TCDD), were found to be so high that their emissions were only
marginally measurable, as shown in Table 7.
In the case of solid wastes incineration, solids carry over in the
stream, as well as solids in the liquid effluents and in the solids residue,
must be evaluated and measured. In most cases, the destruction efficiency
can be defined as [(material fed - (material emitted)]/[material fed]. The
exceptions are cases were hazardous species are created in the course of
incineration. Then alternate, ad hoc definitions must be developed.
A more convenient form of the above expressions would be a destruction
inefficiency, i.e., the ratio of quantity of material not destroyed to the
material fed. This would replace the 99.999 ... percent by 100 - 99.999 =
OlO"n percent, or ppm, ppb etc., which speak more readily to the imagina-
tion.
3.1.2 Continuous. Batch and Periodic Operation
Destruction efficiencies, as defined above, could be applied to
operations which include transient and cyclic operations, provided measure-
ments of materials fed and emitted are taken over an appropriate period of
time. Note that batch introduction of waste into an incinerator, such as
dumping big drums into a rotary kiln, may result in a temporary overloading
of the incinerator and of the effluent scrubbing systems. The resulting
destruction efficiency will be much lower than if the material were introduced
at a steady rate. Reduction of pollution could be achieved by an operating
32
-------
TABLE 6. DEFINITION OF INCINERATION EFFICIENCY TERMS
Efficiency term
Method of calculation
Overall combustion efficiency DECE « 100 X [ % C02 ] - [ % CO ]
Total hydrocarbon (THC) DE
destruction efficiency
Herbicide Orange (HO) DE
destruction efficiency
TCDO destruction efficiency DE
THC
HQ
jCDD
Chlorinated hydrocarbon (CHC) DE
destruction efficiency
CHC
C0
100 X [ THC fed ] - [ THC found ]
[ THC fed ]
• 100 X [ HO fed ] - [ HO found ]
' [ HO fed ]
* 10° X E TCDD fed ] - C TCDD found ]
[ TCDD fed ]
- 100 X C CMC fed ] - [ CHC found ]
[ CHC fed ]
TABLE 7. SUMMARY OF CALCULATED INCINERATION EFFICIENCIES
DECE
DETHC
DEHO
OETCDD
DECHC
First
burn
99.992
99.982
> 99. 999
>99.99
> 99. 999
Second
burn
99.989
99.992
>99.999
>99.88
> 99. 999
Third
burn
99.983
.(a)
> 99. 999
>99.96
> 99. 999
Combined
3 burns
99.990
99.985
> 99. 999
>99.93
> 99. 999
Analyzer was inoperative during third burn
33
-------
procedure modification and would not require modification of the incinerator
proper. In the following work only steady state operation is considered.
3.1.3 Design of Large Scale Incineration Units
The manufacturers of incinerators and furnaces have generally been very
conservative in their approach to scaling up their units; this is not sur-
prising since designs were largely based on past experience with little
analytical background to give confidence in an economically risky endeavor.
Even today there are no scaling laws applicable to any one of the major
incinerator types, not even for the liquid injection furnace, the device
for which much prior research is available. It is somewhat misleading to talk
about overall scaling laws, since larger units will not be geometrically scaled
small units, but may utilize multiple subsystems. For instance, a large
furnace will have multiple burners of the same size as used in a smaller unit,
and a large fluidized bed will need several waste feed points rather than
larger feed ports. The scaling laws, still to be developed, will most likely
address the reproduction in a larger unit of favorable operating conditions
developed in smaller ones. If good analytical tools are developed permitting
this to be accomplished, then very large units could be designed without going
through the design and operation of intermediate size devices. The economic
advantages of skipping to large systems, such as are needed to address the
ever growing quantity of hazardous wastes, are clear.
3.2 LIQUID INJECTION WASTE INCINERATION
Liquid injection waste incinerators are furnaces fired with liquid fuels
which can be the waste itself or an auxiliary fuel, or a combination of both,
depending on the heat content and combustion characteristics of the waste.
0 great deal of work, experimental and analytical, has been done to enhance
the art of furnace design. This work should be very valuable in the pre-
diction of pollutant emissions levels from hazardous waste incineration.
The overall temperature level and residence time in the incinerator are
reasonably well known, both from measurements and analysis, and can be so
chosen as to result in very high hazardous waste destruction levels. The
effects of turbulence are more elusive. Analytical examination of the over-
all phenomena in furnaces relies on modeling of the overall heat and gas
34
-------
flows, and yields accurate prediction of gas temperature histories, but gener-
ally does not permit evaluation of the pollution levels of stack emissions.
There are two sources of incomplete destruction of hazardous components:
• Carry over of liquid droplets
• Carry over of unreacted vapor
The survival of liquid droplets depends on the atomization and penetration
of the liquid jet, as well as the droplet or spray combustion, while carry
over of unreacted vapors depends on recirculation and turbulence phenomena.
In the following subsections, the use of furnace models to establish the
temperature and residence time is described, followed by discussions of the
work on fuel atomization, droplet and spray combustion and turbulence. Approx-
imate prediction of probable emission levels of hazardous species can be
attempted for a given furnace and a given waste material by combining all of
the above elements.
3.2.1 Analytical Models of Liquid Injection Incinerators
Destruction of hazardous waste components in any given incinerator de-
pends on the same "three T's" as combustion in general, i.e., on Temperature,
residence Time and Turbulence. If a detailed characterization of the hydro-
dynamic and chemical phenomena in the incinerator, i.e., if a complete
mathematical model for the combustion in the incinerator were available,
then the degree of waste destruction could be predicted. Mathematical
modeling of liquid injection furnaces or combustion chambers has not reached
such a state of development. Traditionally these devices are designed and
operated on the basis of past experience guided by engineering insight.
Some of the larger manufacturers of boilers and furnaces have initiated
mathematical modeling efforts and it would be interesting to determine to
what extent this is also true of incinerator manufacturers.
Most work on combustion chamber modeling applicable to liquid injection
incinerators pertains to boilers and gas turbine combustors. The simplified
models described in the literature are essentially one-dimensional and
utilize either a system comprised of a stirred reactor (PSR) followed by
35
-------
a plug flow reactor (PFR) or the zone method . The PSR-PFR system has been
particularly useful in predicting such combustor performance characteristics
as flame temperature, combustion intensity, extinction limits and combustion
efficiency in radiant boilers.
The PSR-PFR Model —
In this model, the combustion chamber is divided into two sections. The
first section, the PSR, includes the fuel injector and air inlet reqion, and
the volume occupied by the flame. In this portion, all reactants are assumed
to be perfectly mixed and are at a uniform temperature. The gas stream leaves
the PSR with that same composition and temperature. The incoming stream to the
PSR is composed of air and fuel, as shown in Figure 8. The reaction between
fuel and air releases a major part of the heating value of the fuel in the PSR
sections; combustion is complete in the second section, the PFR, in which a
uniform flow of gases is postulated. The gas velocity and temperature vary
from entrance to exit as part of the heating value is released and heat is lost
to the walls.
Heat transfer to the combustion chamber walls is predominantly by radia-
tion in the PSR, and by radiation and convection in the PFR sections. A
simplication that may be acceptable in refractory lined furnaces is to neglect
heat transfer to the walls. The function of the PSR section is to insure
ignition and stable flame holding, that of the PFR to allow complete fuel com-
bustion. The division into the two sections is made on the basis of visible
flame boundary observation or from past experience.
Liquid fuel atomization characteristics and the velocity and direction of
the incoming air influence the extent of the zone that should be included in
PbR volume; penetration of the largest droplets, as well as the velocity of the
atomizing fluid (air or steam) determine the length of liquid spray and there-
fore influence the flame length, though the PSR assumption is that all the
incoming material is instantaneously and uniformly distributed and hence does
"not permit to include explicitly any such influence. The plug flow reactor
section must allow for enough residence time to complete the combustion.
36
-------
The Zone Method--
The zone method provides a more detailed model of combustion. However,
the initial point is again an assumed knowledge of the fluid flow patterns, of
the chemical reactions and of the radiation characteristics of the gases and
solid particles (if any). The above assumptions are derived from experience
acquired in similar systems and from simple empirical theories on the mixing
behavior of confined jets.
The model is established as follows: The furnace volume and its walls are
divided into a number of gas and surface "zones"; in each zone the temperature
and properties are assumed to be uniform and the gas flows in and out of each
zone known. A schematic of the models shown in Figure 9. Energy balances are
then drawn for each zone, taking into account all forms of energy transport and
including heat released or absorbed by chemical reactions. The solution of the
many simultaneous energy conservation equations yields the temperature distri-
bution in the furnace. Since the flow patterns, including perfect mixing inside
AIR »-
FUEI/ ^
WASTE
PSR PFR
fqWALL f QWAM
•
Q
PSR
UNIFORM
TEMPERATURE
-*- PFR ^
TEMPERATURE &
GAS VELOCITY
VARY
GAS
Figure 8. Perfectly stirred reactor/plug flow reactor
(PSR/PFR) model of a liquid injection incinerator.
37
-------
qWl qW2 qW3 q'
4-
CO
00
t
t
Q
I
Q
/
o.
Q.
Figure 9. Zone model of a liquid injection incinerator.
-------
each zone, were assumed to be known a priori, all three T's are now known, hence
the combustion efficiency and the waste destruction efficiency can be derived.
The difficulties stem from the underlying assumptions more than from the computa-
tional difficulties encountered in solving a large number of strongly nonlinear
simultaneous equations (e.g., the radiation equation involving the fourth power
of the temperature and Arrhenius type reaction rate equations involving exponen-
tial functions of the reciprocal of the temperature), though a great deal of
effort is now devoted to developing time saving computational methods. The
initial assumption that recirculation patterns in the furnace are known is weak,
and so is our knowledge of the radiative properties of gases during combustion.
Analysis of radiation from nonluminous flames, i.e., from gas containing CCL, CO,
H20, H2 etc., is soundly based. However, soot is formed during combustion of
most liquid and solid fuels and the presence of soot drastically increases the
gas emissivity . Soot concentrations are known to depend on temperature, the
presence of water vapor and other hydrogen compounds and on the molecular weight
of fuel constituents, but correlation between gas emissivity and the above para-
meters are not as yet available. Therefore the development of a good, predictive
zone model for any furnace depends to a large extent on the amount of data (gas
temperatures, emissivities, recirculation patterns, turbulent mixing etc.)
available from similar devices and on the skill of the analyst in using that
data to "tune" the model.
Data Utilization in Modeling-
There are various ways of using experimental results to tune a mathematical
model, but the underlying logic is essentially the same for most. It consists
of writing some of the equations to be solved in a form that reflects the
governing mechanisms and includes terms (coefficient, powers etc.) which will
be adjusted to experimental results.
For instance, in a PSR-PSF model, the volume occupied by the flame, the
flame emissivity and the portion of the total heating value of the fuel re-
leased in the flame zone (the PFR) are not known. However, the influence
of atomization on the wall heat flux may be available from measurements
effected with several injector orifice sizes. It will be shown in a
39
-------
subsequent portion of this report that the droplet diameters are approximately
proportional to the square root of the injector diameter and that the length
of the flame depends on the droplet diameter. By judicious use of appropriate
adjustable coefficients, power laws etc., and of regression analysis of
experimental data, the model can be adjusted to correctly calculate one or
more quantities that can be measured. In our example, the split between PSR
and PSF can be adjusted to yield correct wall heat fluxes. The validity of
the model must then be checked in comparing computations to data not previously
used to adjust the model. The validity of the model and its range of appli-
cability will depend on the amount and quality of the data available for tuning
and validation. Under the most favorable conditions, the data will include a
well defined independent parameter whose effect on one of the mechanisms
mathematically expressed in the model can be tracked. The cleanest approach
to tuning obtains when each data parameter influences one of the tuning
factors. Such an ideal condition rarely exists, but can be approximated by
examining and adapting the analytical formulation of the model in the light
of available (or obtainable) data.
The model can be developed with data obtained with nonhazardous fuels
reasonably similar to the hazardous fuel and be used to predict hazardous
materials destruction efficiency, provided, of course, that the fuel pro-
perties are adequately represented in the model formulation (e.g., heating
values, viscosity, etc.). The temperature distribution and residence time
computed by the model are in many cases sufficient to estimate the complete-
ness of the decomposition of the waste component of interest (assuming of
course that the decomposition rate as a function of temperature and time is
known). Difficulties could be encountered if a waste decomposition rate
varies very steeply in the range of temperatures encountered in the incin-
erator. For instance, there are difficulties in the prediction of nitrogen
oxide production because the reaction rate is extremely sensitive to
temperature. However, Bueters et al succeeded in developing a zone model
for a tangentially fired utility boiler which is accurate enough to predict
nitrogen oxide formation . It is probable that these authors had access
to a great deal of data.
40
-------
Effects of turbulence (or unmixedness) are not included in the models
discussed here. In turbulent flow, adjoining eddies found inside a "zone"
may not have the same composition (contary to the assumption of uniformity
inside each zone). The probability of incomplete decomposition of waste
vapor due to turbulence will have to be evaluated separately as long as the
simplified PSR/PSF and zone models are used.
3.2.2 Liquid Atomization and Jet Penetration
Liquid jet atomization has been studied extensively and droplet size
and penetration have been usually correlated in terms of the Weber Number
(We = «/Do/*gVg2). the droplet Reynolds Number (Re - pAVl^ and of
the ratio of liquid and gas velocities18'19- Ingebo has developed correla-
tions for the volume-mean droplet diameter, D3Q, and for the maximum
droplet diameter, Dm *, for the case of a simple, round injector orifice
(TlaX
of diameter DQ. He has also verified experimentally a drop size distribu-
tion equation proposed by Nukiyama-Tanasawa (in a somewhat modified form).
20
Jet penetration into an air stream was also shown by Ingebo to
depend on the Weber and Reynolds Numbers and on the liquid to gas velocity
ratio. The maximum penetration distance xmax is related to the maximum
drop diameter
Wmax • °-08 Re «e-0-4'
-------
More recently M. C. Simmons21 has established non-dimensional correlations
for the drop sizes/volume fraction distribution and the drop size/number distri-
bution for drop diameters normalized to a Mean Median Diameter or to a Sauter
Mean Diameter. The correlations were developed for aircraft and industrial gas
turbine nozzles, including pressure atomizers, air assisted and air-blast
nozzles. (Industrial, steam atomizing nozzles may produce a different droplet
size distribution.) The correlation plot is linear when a square root scale
is used for the normalized drop diameter, and a normal probability scale for
the cumulative volume less than the stated value of drop diameter. This fact
allows computation using statistical tables of the relationship of l^D/DM and
the volume fraction. Similarly, Simmons has shown that the cumulative number
of drops is a simple exponential function of the normalized drop diameter.
He has also found evidence that for each fuel nozzle and operating condition
there is a maximum drop size, above which the probability of survival is
"vanishingly small", and that
Dmax = 3 °MMD °r " 3-6 DSMD <2>
for any given spray. Note that Ingebo also had found that such a maximum
droplet diameter existed. More generally it has been found" that secondary
break-up of droplets occurs at high relative velocities, the criterion for
break-up being a critical Weber number of the order of 6 to 30 (depending on
the Reynolds number). This finding is very important, since the residence
time needed to evaporate a spray is a strong function of droplet diameter.
The evaporation of the largest droplets influences the degree of complete-
ness of combustion (or of any other reaction) of the injected liquid
components.
Atomization of liquids by rotating cups has been examined by Hinze and
Mil born , and found to depend on a Weber and a Reynolds number; though
the authors chose somewhat different groups of properties to develop their
correlations, their groups can be rearranged into Weber and Reynolds
numbers.
The importance of the existence of the various correlations lies in
showing how we can manipulate, for any given injector, the operating con-
ditions so as to obtain the same atomization characteristics with several
42
-------
fuels; for instance, by heating a viscous fuel, altering the atomizing gas
velocities or by using fuel mixtures, so as to maintain invariant the Weber
and Reynolds Numbers. Thus, invariant liquid spray characteristics can be
insured for fuels with different physical properties.
Modeling and scaling of the atomization is relatively straightforward,
since cold flow experiments suffice to develop suitable correlations, or to
verify the applicability of the existing ones, for the candidate injectors.
These correlations then can be used in designing the incinerator. Jet
penetration can also be experimentally investigated in cold flow. However,
the influence of the internal gas flows, which depend on the furnace con-
figuration and on the combustion-induced gas flows, is not fully represented
by cold flow.
3.2.3 Droplet and Spray Combustion
The combustion of a spray of liquid fuel involves the combustion of
fuel vapors and of the discrete droplets that make up the spray. The
classical shrinking-sphere model of diffusion-controlled combustion in an
oxidizing atomosphere is not sufficient to describe spray combustion; never-
theless, that model is fundamental to understanding spray combustion phenomena.
Droplet Combustion—
The simplest model of droplet combustion is the diffusion-controlled
model of an isolated spherical droplet, when gas velocities are relatively
low. This model is applicable to dilute sprays only. In dense sprays,
complete evaporation of the droplets may precede burning, which then is con-
trolled by the diffusion of the vapor into the air or by gas-gas mixing.
For dilute sprays, i.e., isolated droplets in quiescent air, a droplet
evaporates and the fuel vapor and the air burn in a diffusion flame which
surrounds the droplet. The mass burn rate m, is related to the decrease in
ni *•
droplet diameter D such that
<' IT PL> (3)
43
-------
which can be rewritten as
A •> 4 m.
-3t <°2> ' - -5^ ««)
It has been shown experimentally that m, is proportional to D,
so that the above equation can be integrated to yield
D2 = DQ2 - Kt (5)
In the above equation, the constant K must be obtained from experiment
or by solving conservation equations for global mass, species mass, energy,
and momentum.
Solution of the conservation equations is possible provided some major
assumptions are made. These assumptions include:
• Infinitely rapid chemical reaction rate
• Simplified chemical reactions, such as,
Fuel + 0- •* products, or
Fuel + Oxygen •* CO followed by CO + Oxygen •*• C02
The solutions depend on whether fixed or variable transport properties
are used. A widely used model, assuming fixed properties and a Lewis Number
of one,* results in a burning rate coefficient K
K = 4-£- In (1+B) (6)
CPPL
where the transfer number B is
and
X = thermal conductivity of the gases
Q - fuel heating value
L = latent heat of evaporation
C = specific heat of gases
* = equivalence ratio (air/fuel)/(air/fuel)stoichiometric
The Lewis Number is the ratio of thermal diffusion to gas diffusion co
efficients.
44
-------
The results markedly depend on the values assigned to the thermal con-
ductivity X and the specific heat C of the gases; these depend on gas
temperature and composition, neither of which is known a priori. Usually
A and C are computed on the basis of air or nitrogen properties at the log
mean temperature between the liquid and the flame temperature, though somewhat
more sophisticated approaches have been proposed.
Incinerator fuels are mostly mixtures of hydrocarbons and other more or
less volatile compounds, rather than a single component liquid. An expres-
sion for the burning rate coefficient of binary mixtures was derived by Wood,
et al and numerical calculations have been carried out for multicomponent
26
heavy fuel oils • The method proposed in the latter work could be used
to compute the evaporation and combustion of a given compound (hazardous
waste) contained in a liquid fuel. The compound could be liquid or a sus-
pended solid-, or a solid formed by fuel cracking, such as occur in heavy
fuel oils. However the computation is cumbersome and a large computer is
needed.
Equation (5) is valid for a droplet burning in quiescent air, whereas
in most furnaces air velocities must be high enough to promote mixing. A
correction of the burning rate to take into account forced convection has
77
been proposed*1'
Kforced conv. = Ko H+0.278 ReL1/2 Pr1/3(l+1.237/ReLPr4/3)"1/2] (8)
v
This equation is possibly valid for 10104
45
-------
None of the proposed predictive methods can take into account transient
burning effects or change in combustion mode, such as the transition, above a
critical gas velocity, from a diffusion flame all around the droplet to a wake
flame (i.e., no burning on the droplet surface and combustion in the wake be-
hind the droplet). Still more important is the fact that the single droplet
combustion model does not really represent spray combustion.
Spray Combustion-
It has been shown that droplets exist in a relatively small volume, close
to the injector nozzle and that most droplets do not burn individually, but
that the fuel vapor from the droplets burns in a jet, essentially as a gas
diffusion flame. This was proven in experiments conducted with a light distil-
late oil and with heavy fuel oil . The "mixed is burned" principle applies
well to a diffusion flame, i.e., the turbulent mixing of fuel vapor and air is
the combustion rate controlling mechanism. Therefore, one would consider
characterizing the droplet shrinking shown in equation (5), by heating and
evaporation only and omit the fuel heat release time term Q x * in equation (7)
when modeling the incineration mechanisms as a spray, i.e., a gas-diffusion
f1ame.
3.2.4 Turbulence and the Unmixedness Factor
Hawthorne, et al31 have shown that in turbulence controlled combustion
of gas sampling and concentration measurements yield data that permits cal-
culation of an unmixedness factor yC*2 , which in turn can be used to used
to predict emissions of unburned or incompletely burned fuel components.
Figure 10 from Reference 31 shows the correlation between the unmixedness
factor and the ratio of oxygen measured in a gas sample to the stoichiometric
oxygen required to complete the combustion of fuel as measured in the sample
(CO, H2 or UHC).
Thus the unmixedness factor can be obtained at a point at which the gas
is sampled. Often it is important to calculate the unmixedness factor at a
point upstream of the point of sampling, which is usually far downstream of
the flame region. For instance, if it is known that the destruction of a
hazardous waste component cannot take place below a certain temperature, then
46
-------
AC
30
20
10
S
6
1.0
.8
.6
.4
2 4 6 8 10 20 40 60 80 100 200 400
ACTUAL OXYGEN IN UNBURNED GAS, OR ITS RELIPROCAL
O2 EQUIVALENT OF UNBURNED GAS
Figure 10. The unmixedness factor in turbulent flames.
the contribution to emissions of that component due to unmixedness will depend
on the unmixedness factor in the region of the incinerator where the tempera-
ture drops to that level.
Unmixedness at points upstream of the sampling point can be inferred,
if it is assumed (Reference 32) that unmixedness decay follows the same laws
as turbulence decay - i.e., y C'2 = exp (-Bt)
47
-------
wherein s is defined by a semi-empirical equation
B = 0.19 (S/L2)1/3
where
L = characeristic length of system
L= rate of energy input of mixing energy per mass of system
The rate 2 of energy input is not directly accessible, though there are
some theoretical correlation applicable to some devices. Alternate approaches,
still based on the unmixedness factor, have been used. For instance the pro-
bability of undecomposed vapor carry over in the incineration of Herbicide
Orange was evaluated by the Arthur D. Little teanr^ by assuming a typical
Gaussian distribution (i.e., based on experience with other furnaces) of the
air/fuel equivalence ratio at the inlet to the plug flow section and estima-
ting the time constant for the decay on the basis of measured carbon monoxide
content at the furnace exit. They also used a theoretical approach wherein
the air/fuel equivalence ratio initial distribution and characteristic mixing
time are based on the specific power delivered to the gases in the main
chamber. The two methods were in satisfactory agreement.
However, calculation of et by this means seems to be rather arbitrary
and could be avoided by taking two gas samples - one in the flue gas the
other closer to the flame zone; then B could be calculated as
,2U
^j^l = exp [-B(trt2)]
since (t-j-tg) is known from the gas velocities and the distance between the
two sampling locations.
3.2.5 Estimation of Hazardous Haste Destruction Inefficiencies
The three sources of liquid waste destruction inefficiencies are:
• carry over of unevaporated liquid droplets
• carry over of unreacted vapor due to inadequate temperature
history
• carry over of unreacted vapor due to unmixedness
48
-------
The following steps would lead to an estimation of an upper bound of
destruction inefficiencies:
• Obtain the temperature and residence time required to achieve
complete decomposition of the hazardous species.
• Characterize the liquid atomization and jet penetration of
and fuel. Needed are the physical properties of the fuel/
waste and cold flow experiments.
• Develop a PSF-PFR model of the incinerator. If measurements
of flame temperature and wall heat flux are not available,
engineering estimates based on fuel spray characteristics
(cold flow tests) are possible. The model will yield the
average temperature.
From the above, the probability of a drop surviving passage through
the entire furnace and the probability of some (low) part of the vapor not
reacting completely can be established.
• Obtain data on the carbon monoxide and oxygen content in the
stack when burning a nonhazardous fuel with equivalent pro-
perties.
From these data, the portion of unreacted vapor due to unmixedness
(adjoining eddies of fuel vapor and oxidizer) and to bypass of the main
reaction zones (boundary layer flows and recirculation) can be estimated,
3.2.6 Summary
Despite the scope of the efforts expended to date in investigating the
combustion of atomized liquid fuels in furnaces there is no generally appli-
cable method to predict the composition of the flue gases with the precision
needed to insure safe incineration of hazardous waste components. An
approach to the estimation of an upper bound of hazardous component destruc-
tion inefficiencies has been proposed here. The success of the estimation
depends to a large degree on the availability of experimental data because
the theoretical base and analytical tools alone are insufficient.
49
-------
3.3 FLUIDIZED BED INCINERATORS
Fluidized bed incinerators are reactors in which a bed of inert particu-
lates (sand) is supported by a distribution plate through which air is flowing.
Waste and auxiliary fuel are injected into the bed above the distribution
plate. Fluidization occurs when the frictional forces between gas and a
particle equal the weight of the particle. At this flow rate the bed is con-
sidered incipiently fluidized and the voidage of the bed is equal to the
voidage of the most loosely packed fixed bed. An increase in fluid flow rate
for liquid-solid systems results in a progressive expansion of the bed. A
gas-fluidized bed does not expand uniformly with increasing fluid flow rate.
Rather, bubbling of the gas is observed, and the bed does not expand much
beyond its volume at minimum fluidization; this bed is referred to as a
bubbling fluidized bed, a gas fluidized bed, or a dense phase fluidized bed.
If the fluid velocity is increased further, the terminal velocity of the
solid particles is exceeded, the upper surface of the bed disappears and the
solids are carried out of the bed; this bed is known as a disperse, dilute
or lean-phase bed.
As discussed in more detail later, the quality of fluidization affects
the performance of the fluidized bed incinerator. In dense-phase fluidi-
zation, the gas is distributed between the bubble-phase and the emulsion-
phase of the bed. The gas carried in the bubble-phase does not directly
participate in the solid-gas reactions. However, the bubbles promote solids/
solids and emulsion gas/solids mixing. Furthermore, in the gases in the
bubble-phase and in the emulsion-phase are continuously exchanged as the
bubbles rise through the bed. Several factors such as bed geometry,
particle size, gas flow rate, gas distributor plate, and vessel internals
afreet the quality of fluidization. For example, a sintered distributor
will give a better quality of fluidization than a single orifice plate
because the sintered plate will produce many small bubbles wereas the
single orifice plate will promote larger bubbles which leads to channeling
and slugging. Better gas distribution is acquired at the cost of higher gas
pressure losses, so that the design of the distributor plate is crucial in
combining performance and economics. Despite identification of the factors
50
-------
Involved In promoting a good quality of fluldlzatlon, no successful mathematical
relationships have yet been developed which describe the effects of each factor
on the quality of fluidization.
Recent efforts in mathematical modeling of fluidized beds have resulted
in the predictions of their efficiency as chemical reactors rather than the
definition of scale-up parameters. In developing these models> various assump-
tions are made concerning bubble size and velocity, coalescence, gas velocity
in the emulsion phase, particle size and shape, etc. So far, the relationship
of these parameters with bed size is not established, so scale-up procedures
are still based on previous experience.
Models predicting performance of fluidized beds are based on conditions
more uniform and ideal than usually obtained operationally. Such factors as
wall effects, distributor plate effects, number and location of waste feed
points, etc., are rarely considered in developing hydrodynamic models of
fluidized beds (although heat transfer to the walls has been modeled). For
this reason, and because of transient upsets suffered in normal operations,
models of "ideal" fluidized beds would most likely predict better efficiencies
than are actually achieved.
Wastes contain combustible and noncombustible components and the com-
bustible fraction is composed of volatiles and nonvolatiles. The volatiles
are those compounds which volatilize at the relatively low operating
temperatures achieved in the burning process in fluidized beds and are com-
posed mostly of hydrocarbons. The nonvolatile combustibles are mostly
carbonaceous and polymeric materials. The noncombustibles consist of
moisture and inorganic ash, which influence the effluent concentrations and
the behavior of the bed. They act as a thermal sink and thus can lower bed
temperatures. Furthermore, some inorganics such as silicates and sodium
compounds can defluidize the bed at high temperatures or appear in the
effluent stream as flyash or metal oxides (see Section 2.3).
There are two sources of waste incineration inefficiency: 0) incom-
plete oxidation of the volatiles, and (2) loss of solids which contain
unoxidized combustibles. In the design and operation of fluidized bed
incinerators, the incomplete oxidation of solids presents the greater
difficulty in attaining complete incineration because solids generally
51
-------
require a longer time for complete oxidation than gases at a specific temper-
ature. Although the hydrodynamics of gas flow through a fluidized bed are
extremely complicated, the residence time of vapors at a specific temperature,
which can be calculated, basically determines the extent of oxidation of the
volatiles. Computations of the extent of loss in solid form is more difficult.
The loss of incompletely oxidized solids can occur by elutriation or by re-
moval with the bed material. The bed material must be removed and regenerated,
continuously or periodically, because of build-up of noncombustibles or
attrition of the inert heat carrier (usually sand). Inadequate solid residence
time is a major cause of inefficiency of fluidized bed incineration.
The efficiency of waste incineration in fluidized beds is influenced by
several factors: (a) the physical form of the waste, (b) the chemical com-
position, and (c) the size of solid wastes. Furthermore, the performance of
the incinerator, and hence, the prediction of effluents composition is
dependent on factors such as inert particle size, temperature, feed mechanism,
fluidizing air flow rate, size of the bed, etc., as well as on the kinetics
of incineration. A complete model would involve the coupling of the kinetics
of waste incineration with the factors which predict bed performance, and
because neither is well understood, attempts to predict the composition of
effluents from fluidized bed incinerators have not been successful.
Single particle models permit a first cut at analyzing the residence time
requirement and/or determining what kind of waste preparation (shredding,
crushing etc.) is needed. These models also yield some information for the
evaluation of waste destruction efficiencies.
Several hydrodynamic models have been proposed, describing the gas
'nterchange rate between the bubbles and the dense phase through which the
bubbles rise. Correlations for solid particle attrition and elutriation of
fines have also been proposed and will be discussed below.
3.3.1 Models of Oxidation of Single Particles
Fluidized bed reactors are, at present, designed on the basis of
kinetic models for the oxidation of single particles, which determine the
residence time needed complete combustion in the bed and in the freeboard.
34
Three kinetic models cover most cases:
52
-------
The continuous reaction/constant size model is applicable to
reactions in which a solid is consumed uniformly throughout the
particle. The gas diffusion into the particle is much faster
than the gas-solid reaction.
The unreacted core/constant size model (also known as the shell
model) is applicable to cases where both or either trie gas-solid
reaction and the gas diffusion through the residual solid shell
are controlling.
The shrinking particle model is applicable to the cases where
oxidation leaves no solid residue, i.e., a case very similar
to droplet combustion. Here either gas-through-gas diffusion
or surface reaction can be controlling.
For the continuous reaction model, the progress of conversion of solid
reactant B is independent of particle size as a first approximation (a
uniform concentration of oxygen is implicit in the fast gas diffusion assump-
tion). The rate of conversion of the solid reactant B is
dXR
P _ v f, /i y \
dt ~ vw U~V
where
Xg = fraction of B converted
C. = concentration of oxygen (constant)
K = rate coefficient based on the volume of solid
The continuous reaction model usually applies to very small particles.
These are very likely to have been elutriated and at least part of the
reaction is taking place in the freeboard. Note that the residence time
required to completely oxidize the particle is independent of particle
size.
The shrinking core/constant size model includes several mechanisms:
1. Oxygen diffuses through the boundary layer surrounding the
particle to its surface.
53
-------
2. Oxygen diffuses through the product layer to the reaction
front.
3. Oxygen reacts with the solid in a narrow reaction zone.
4. Gaseous products diffuse to the main gas stream.
Figure 11 illustrates the diffusion of oxygen.
GASEOUS
BOUNDARY
LAYER
Figure 11. Shrinking core model.
54
-------
In a vigorously bubbling bed, the diffusion resistances in steps 1 and 4
are not as great as the resistance in step 2 when there is a solid product
layer. Consequently, either the reaction rate, the diffusion through the
solid product layer (or a combination of the two) controls. Generally, when
the reaction rate is strongly temperature-dependent, resistance to gas dif-
fusion does not control the rate.
The chemical kinetics for heterogeneous oxidation of a spherical particle
34
are represented as first-order in oxidant . Thus, for an unreacted core of
radius rc> the rate of reaction of A (oxygen) is represented as:
dNA 7
- """ K
Ac
where
N« = moles of oxygen
t = time
rc = radius of particle at the reaction zone
K = kinetic rate constant (first-order irreversible
reaction)
(C.) = concentration of oxygen at r
The rate equation for the disappearance of A due to diffusion through
the solid product layer is:
dN. dC
where
__ . 4, T( De
De = effective diffusivity of oxygen through the solid product
layer
A steady-state concentration gradient through the product layer leads
to a mass balance such that
J_ (r2 De
55
-------
The expression is integrated twice with the boundary conditions:
CA = (CA)$ at r = rs
CA - (CA)C at r = rc
where
rg = radius of spherical particle
CA - (C.) 1 - r /r
" ft C = _ C (A\
(CA>s - c ' - rc/rs W
so,
A - (CA>S - (CA>C
^
dr rc " rc (1 • rc/rs)
Substituting equation (5) into equation (2) gives
dN,
can be eliminated from equation (1) to yield a value for (CA)C in
terms of (CA)S:
Substituting equation (7) into equation (1) yields
r.K .
- H ' r)]'1 (8)
According to the spherical geometry of the particle, the time rate of
disappearance of the particle is:
dNB P
- — •
56
-------
where
NB
PB
MB
= moles of particle B
= density of B
= molecular weight of B
Thus,
4ir r
dr.
HT
dt
The stoichiometry is such that
dt"
J_ ^i .
b dt
bMB
(10)
where b = stoichiometric constant.
Thus,
dT
bMBK(CA);
^ + DirT
(11)
Thus, the time rate of particle shrinkage can be evaluated if (CA)S is
known. For fluidized bed incinerators operating with excess (L, the assumption
of (CA)S as constant and equal to the mean oxygen concentration is reasonable.
Equation (11) can be integrated to yield the time necessary for a particle
to completely disappear. For cases where both the reaction kinetics and diffu-
sion through the product layer represent rate limiting steps, Equation (11)
must be solved using numerical methods. When either the chemical kinetics or
diffusion controls the rate of reaction, an analytical solution is possible.
Kinetic controlled: r K«De
drc
dt~
bMB K (CA),
57
-------
so,
pBrc
bMB K
i.e., the residence time needed to completely react a particle is proportional
to the particle radius.
Diffusion control case: r,.K»De
dr. bMR (C.) De rc
C _ p ft 5 S
dt PB rc rs'rc
(CA)S De
In this case the residence time for complete particle oxidation is pro-
portional to the particle radius squared.
The shrinking particle model is very similar to the model for liquid
fuel drop burning discussed in Section 4.2.3, and differs from the unreacted
core only in that the controlling diffusion rate is now that of the oxygen
diffusing through a gas layer. For small particles, at temperatures below
1300 K, the reaction rates are low, and are controlling. Equation (1) is
applicable, with ^representing the ever diminishing radius of the particle.
The time needed to consume a particle is given by equation (12) when the
reaction rate is low. When gas diffusion is controlling (highly reactive
solids) equations (13) applies, with the diffusion coefficient De representing
diffusion of oxygen through the gaseous boundary layer rather than thru the
spend solid layer.
3.3.2 Hydrodynamic Models
The fluid mechanical properties of the bed have important consequences
on the attrition, elutriation, and the gas interchange between phases.
Efforts in modeling fluidized beds have usually accepted the two-phase model
first proposed by Toomey and Johnstone35: a bubble phase in which the
gases, essentially free from solids, rise through the bed and a dense phase
consisting of both solids and gas. Bubbles form at an unpredictable rate,
grow, coalesce and may even split. There is clearly a gas flow through the
58
-------
bubbles, so that there is gas interchange between the phases. Particles move
up and down and around the bubbles, and some rise in the wake of each bubble.
Several hydrodynamic models have been proposed and they were reviewed by
Horio and Wen . A number of assumptions must be made in specifying a model
such as uniform bubble size, uniform void fraction, etc.
Horio and Wen worked out a code to classify fluidized-bed reactor models
including a system that classifies the assumptions and factors included in the
model; they also classify the models by levels of sophistication. But, to
•3C
quote Horio and Wen , "Inspite of the many improvements attempted by many
researchers, the accuracy of the fluidized bed reactor models is still insuffi-
cient for general use in design and scale-up."
3.3.3 Attrition and Elutriation
Attrition and elutriation have been modeled by Merrick and Highley37
based on data obtained in a fluidized bed coal combustion plant. They
developed a new form of correlation for the elutriation, superseding the
widely used Wen and Hashinger correlation which exhibits improper behavior
for small particles (see Reference 34 p. 316 for this correlation). It
remains to be seen if the work Merrick and Highley is applicable to waste
incinerators with sand beds with very different particle size distribution
and reaction kinetics than those describing pulverized coal in a bed of
limestone.
3.3.4 Effluent Prediction
Pollutants can arise from three sources in fluidized bed incineration:
(1) elutriation, (2) continuous or periodic removal of bed material con-
taining uncombusted wastes, and (3) gaseous hydrocarbon emissions.
First, elutriation is considered and the following assumptions are
used to develop the model:
1) There is a critical size, r , such that particles smaller
than this size are elutriated to the freeboard immediately
from the bed. Merrick and Highley37 present a correlation
showing that r <_ T-6^ where rt is the size particle whose
59
-------
terminal velocity is equal to the fluidizing velocity Uf. Con-
sequently, TC, and hence the amount of bed material elutriated is
dependent on the operating conditions of the bed (because the
larger r , the larger proportion of bed material that is finer than
the critical size). The well-known expression for terminal velocity
is:
„ 4g(pp.og)dp
1 " 3pgCD
where
Ut = terminal velocity, m/s.
;i(
.3
2
g = gravitational acceleration, m/s .
p = density of particle, kg/m
p = density of fluid, kg/m .
d = particle diameter, m.
Crj = drag coefficient, dimensionless.
Of course, the drag coefficient is dependent on the Reynolds
number and the shape of the particle.
2) Particles which are larger than rQ stay in the bed until they
are reduced to a size smaller than the critical radius by
simultaneous attrition and chemical reaction. They are then
elutriated to the freeboard.
3) As the particles rise through the freeboard, they may be further
reduced in size by combustion. It is assumed that the rate of
reaction is controlled by diffusion such that the time for complete
conversion is
where
t- = time needed for incineration, s.
Qn = diameter of particle, m.
2
K = effective diffusion rate constant, m/s.
4) No attrition, abrasion, or agglomeration occurs in the freeboard.
60
-------
5) Upward particle velocity is assumed to be (U- - U.).
Although IL decreases with decreasing size at low Reynolds
numbers, the velocity change due to decreasing size is small
when compared to the overall velocity. Hence, the particle's
upward velocity is assumed to be constant. Consequently,
"f- (Uf - Ut)
where
tf = residence time in the freeboard, s.
L = height of freeboard, m.
U- = fluidizing velocity, m/s.
Ut = terminal velocity, m/s.
Small size combustible particles which are elutriated to the freeboard
will burn sufficiently quickly such that they are not carried out of the
freeboard. Consequently, for a given particle size distribution in the bed,
only a narrow cut of particles will be elutriated as shown in Figure 12.
u.
o
PARTICLES
LEAVING THE
FREEBOARD
PARTICLES
ELUTRIATED^
TO THE V
FREEBOARD /
PARTICLES IN
THE BED
PARTICLE SIZE
Figure 12. Particle elutriation.
61
-------
The larger sized particles are not elutriated to the freeboard, giving the
narrow cut of elutriated material.
Incompletely Oxidized Vapors--
Unburned hydrocarbons and CO are effluents which are indicative of incom-
plete oxidation of the vaporized material. Although mixing in the bed is
generally adequate for complete oxidation, some hydrocarbon vapors are gener-
ated near the top of the bed and swept to the freeboard before oxidation is
completed. Furthermore, bubbles which are lean in oxygen will contain unburned
hydrocarbons. Although the gas interchange coefficient describes the rate of
gaseous interchange between the bubbles and emulsion gas, the residence time of
a particular bubble depends on its place of origin in the bed. Assuming that
all bubbles originate near the distributor plate, and the bubbles have a
velocity on the order of the gas velocity near the distributor plate or that
the bubble velocity can be determined experimentally, then
where
I = the number of times that the bubble is completely
replaced with emulsion gas.
L = height of fluidized bed, m.
K * gas interchange coefficient, s
Ub = velocity of bubble, m/s.
Kunii and Levenspiel have proposed a correlation for a gas interchange
rate constant which allows for gas interchange by both convection and dif-
fusion.
UMF D01/2 9V4
K = 4.5 [^-F] + 5.85 [ 9d5/4 ]
b
where
UMF = minimum fluidizing velocity, cm/s.
dK = diameter of bubble, cm.
2
D = gas diffusion coefficient, cm/s.
2
g = gravitational acceleration, cm/s .
62
-------
A high value of I means that the gas between the bubble and emulsion has
interchanged many times, and hence the mixing is good. However, a corre-
lation between I and incineration efficiency has yet to be formulated.
Unreacted vapors entering the freeboard will have additional time to
react with oxygen. The oxidation completness depends on the mean residence '
time in the freeboard (i.e., flow rates and freeboard height) and on the
turbulence level in the flow; thus the conditions are similar to those
encountered in by vapors in liquid injection incinerators beyond the flame
zone.
3.3.5 Summary
Despite intensive efforts to analyze, correlate and model the hydro-
dynamic, heat and mass transfer and the reaction mechanisms in fluidized
bed reactors, there are today no scaling laws and very few proved design
and operating parameters which would permit the evaluation of the hazardous
waste destruction capability of a proposed fluidized bed reactor.
3.4 MULTIPLE HEARTH INCINERATORS
Multiple hearth furnaces are particularly well suited for the incin-
eration of solid wastes and sludges. A cross section of a multiple hearth
incinerator is shown in Figure 13. The wastes are fed to the top hearth
while air is fed to the bottom hearth. A shaft which is positioned
vertically through the furnace swings the arms across each hearth. The
angle of the rabble teeth on the arms determines whether the solid material
moves outward or inward across the hearth. The waste material which spirals
inward on the hearth drops through a hole in the center of the hearth, while
the material which spirals outward passes through a series of drop holes on
the circumference of the hearth. The hearths are arranged vertically with
alternating in-hearths and out-hearths. The flow of solid material through
the reactor can be controlled by the rotational speed of the shaft, the
spacing of the rabble teeth, and the distance between the bottom of the
teeth and the hearth.
The kinetics of solid waste incineration are described in the fluidized
bed section. The proper residence time for complete sludge or solid incin-
eration must be determined experimentally. Then various operating parameters
63
-------
Figure 13. Multiple hearth cross section.
64
-------
such as feed rate, shaft rotational speed, and the spacing, angle and size of
the rabble teeth, can be adjusted to the particular Incineration requirements.
The multiple hearth Incinerator Is characterized by different zones, each
comprised of one or more hearths. The waste is fed near the top of the
reactor where any water associated with the waste is driven off. This top
zone is termed the "drying zone". Some multiple hearth furnaces have a
hearth above the drying zone hearths which serves as an afterburner. The
waste is fed to the drying zone, while air is fed to the afterburner to
facilitate the oxidation of unburned CO hydrocarbons and vapors. Below the
drying zone is the combustion zone. Air is fed to the bottom of the com-
bustion zone while the solids fall into it from the drying zone. Below the
combustion zone is the ash cooling zone through which the ash drops before
removal from the furnace.
3.4.1 Temperature Profiles
The temperature profile is such that the temperature drops through
each hearth as the gas rises from hearth to hearth. The temperature in the
afterburner, however, may be higher than the temperature in the drying zone.
Typically, the temperature is controlled by using excess air or auxiliary
fuel which are both fed at the bottom of the combustion zone.
A typical temperature profile is sketched in Figure 14.
HOT GAS
FEED
DISTANCE FROM
BOTTOM OF
FURNACE
AIR
FUEL
TEMPERATURE
ASH
Figure 14. Temperature profile in a multiple hearth incinerator.
65
-------
The use of either excess air or auxiliary fuel shifts the temperature
profile to the left or right respectively from the profile derived from the
incineration of feed without the use of excess air or auxiliary fuel. The
temperature difference across each hearth is dependent on such factors as gas
flow, the solid mass in each hearth, and the furnace's heat release
3.4.2 Turbulence and Mixing
The gas turbulence in the furnace is difficult to characterize. The
design of the furnace is such that the gas travels from the inside of one
hearth to the outside of the next hearth. Turbulence is promoted by the
countercurrent contacting of the gas with the solid waste as the gas rises
through the dropholes. There is no satisfactory model to predict the degree
of turbulence from the various operating parameters.
The mixing of solids in the furnace is also difficult to characterize a
priori. Mixing occurs as the solid waste drops through a center drophole in
an "in-flow" hearth. The degree of mixing can be defined using experimental
data, and is usually described in statistical terms. Various analytical
methods such as gravimetric, volumetric, electrometric, particle counts, and
optical have been used to describe the degree of mixing. A general procedure
for determining the degree of mixing is to add a tracer material to the waste
material. Let p be the fraction of tracer in the waste. A number, N, of
small samples are taken from various locations in the multiple hearth and the
fraction of tracer, X.., in each sample is determined. The average value of
the measured tracer concentration in the N samples is designated by X. As N
increases, X ->• v. If the waste material is well mixed, every value of X.
would equal X. Thus, the standard deviation of X- from X is a measure of the
quality of mixing.
However, S is a valid measure only for a set of tests under specific
conditions. A more general parameter is the mixing index, I. I is the
ratio of S/aQ. where a is the standard deviation before the onset of mixing
Smith and Van Ness38 propose the equation
66
-------
oQ = /w (1 - u)
Thus,
I . -S .
ao
In a batch mixing process, I is unity before mixing begins and becomes
progressively smaller. Theoretically, I would approach zero at long mixing
times, but mixing is never complete and analytical methods are not precise
enough to give measured values of X^ equal to each other. Typically the
low values of I fall in a range of 0.1 to 0.01.
3.4.3 Residence Time of Gases and Solids
The simplest method for determining the residence time of the gas phase
is to assume plug flow through the incinerator. Thus,
where
t = residence time.
V = volume of multiple hearth incinerator.
W = volumetric flow rate of air.
Vapors which are generated from the solid wastes are assumed to be swept
through the multiple hearth incinerator at the same flow rate as the air fed
to the incinerator. Therefore, the residence time of vapor generated in a
given hearth will depend on the vertical location of that hearth.
As mentioned previously, the solid waste moves across each hearth in a
spiral. The expression describing the movement of the solid material across
the hearth in polar coordinates is:
r = + a 9
67
-------
where
r = Distance from center of the hearth.
a = Distance of teeth from the center of hearth.
(Assuming they are evenly spaced)
9 = Angle
+_ = Describes whether the material is moving towards
or away from the center of the hearth (+ for an
"out-hearth"; - for an "in-hearth")
If r = radius of hearth and 92 = angular velocity, then the time for the
solid to traverse across the hearth, t, is:
t =
Ea
here E is a factor describing the slip of solid matter with respect to the
movement of the rake teeth. For liquid wastes E depends on viscosity and
can be estimated. For solid wastes the slippage depends on several factors
such as size, shape, and moisture content of the material and is less readily
estimated. The height of the dead-bed, which is the height of material on
each hearth, also affects the slippage of material. The slippage of material
is greater the further the material is from the rake teeth such that the
material on the floor of the hearth moves across the hearth more slowly than
the material closer to the rake teeth.
3.4.4 Scale-up Parameters
Scale-up of rotating hearth incinerators will be limited by the neces-
sity of keeping angular velocity low, so as not to increase velocity of
the outermost rake teeth beyond acceptable values. The relationship between
incinerator capacity and hearth diameter, incinerator height and rotational
speed is complicated by the dependence of solids and gas residence times not
only on incinerator height and diameter but also on the rake teeth con-
figuration (angle, spacing etc.) and the properties of the solids. The
probability of developing useful generalized scaling laws for multiple
hearth incinerators is not high.
3.4.5 Summary
Multiple hearth incinerators are widely used for waste incineration.
Nevertheless, there seems to have been very little effort made to develop
68
-------
an analytical basis for the design of these furnaces. The development of a
general treatment, i.e., one applicable to a wide range of wastes, would
involve the expenditure of much effort - including the collection of data
(temperatures, gas and solids compositions) and the development of solids
mixing - rabble teeth geometry correlations.
3.5 ROTARY KILN INCINERATORS
Rotary kiln incinerators are long, cylindrical rotating furnaces in
which solids and slurries are heated by combustion of an auxiliary fuel.
The fuel, as well as the solids can contain hazardous components. The
axis of the kiln makes an angle with the horizontal. The feed is intro-
duced at the upper end of the kiln and the hot product discharged at the
lower end. Fuel and air inlets are located either at the lower end,
resulting in a countercurrent gas/solids flow, or in the upper end, yielding
a cocurrent flow. The rotation of the kiln provides continuous mixing of
the solids and continuously renewed contacts between solids and the hot
walls as well as direct contact with the hot gases. Exposure to hot gas is
often enhanced by the use of hanging chains, which break up any solid chunks,
mix and stir slurries and permit wet, sticky material to cling to the chains
until dry. In direct fired kilns, the solid materials are not showered
through the gas stream, as they are in dryers, but are retained in the
lower part of cylinder, except in the feed section in which the hanging
chains (if any) do lift the material into the gas stream.
Often the kiln is preceeded by drying or preheating installation such
as moving grates, and is followed by an afterburner. The effect of these
devices on hazardous waste destruction must, of course, be accounted for
when examining the thermal decomposition of hazardous wastes in the incin-
eration facility. Here we will address only the process taking place in
the rotary kiln itself.
3.5.1 Combustion and Heat and Mass Transfer
Rotary kiln incinerators are heated by natural gas, oil or pulverized
coal. The kilns are usually very long, so that the combustion zone
occupies a small portion of the incinerator. Most of the heating of the
charge is due to exchange with the combustion product gases and the walls
69
-------
of the kiln. Mass transfer, drying and volatilization of the charge, is very
important in the case of solid or liquid waste incineration, but only drying
has been analyzed in the literature. Moreover, the correlations for drying
of the charge pertain to low temperature rotary driers with "flights", i.e.,
scoops on the walls of the kiln; these scoops scoop up the charge and shower
it through the combustion gas so that the charge receives heat by direct
contact with the gases. The correlations for rotary dryers are of the form
Qch = Ua v (AT)m
Qch = heat transferred to the charge (J/s)
Ua = volumetric heat transfer coefficient (J/s-K-m )
v = dryer volume (m )
(AT) = mean temperature difference between charge material
m and hot gases (K)
A correlation for the volumetric heat transfer coefficient has been
developed. It is of the form
Ua = KGn/D
where D is the kiln diameter and G is the mass flowrate of the combustion
o
gas (kg/s-m of kiln cross section).
P. Y. Me Cormick39 has correlated data for single shell direct heat
dryer's; he found that
n = 0.67
and proposed that, in order to segregate independent design parameters, the
correlation should take the form
Qch = bLDGn (AT)m f^f^M) f3(A) f4(N)
where b is a proportionality constant, L and D the length and the diameter
of the kiln, and fj, f2. f3 and f. functions of the number of flights Nf,
the radial flight depth M, the flight load A, and the rotation speed of the
kiln N, respectively. Note that direct radiative heating from flame and
70
-------
walls, and direct wall contact heating are not represented, which is justified
in the relatively low temperature rotary dryer applications, but not for incin-
erators. The heat exchange mechanisms in the latter are too intricate to
allow for simple correlations, and a more sophisticated approach e.g., model
development is needed.
The mathematical treatment of high temperature rotary if"ins found in the
literature is limited to specific applications, such as cement kilns or ore
dryers and do not lend themselves to generalization. Exceptions are a model
presented by Sass*0 which includes a preheat section wherein the solids
are heated to the boiling point of the liquid, and an isothermal section where
the liquid is evaporated, followed by a final section where the solids are
heated to a desired discharge temperature; and a very useful treatment by
Imber and Pashkis in which dimensionless parameters were used. A schematic
diagram of the heat flow paths in a cross-section of a rotary kiln is shown in
Figure 15 taken from the work of Imber and Pashkis. Sass includes some heat
loss to the ambient air. The heat transfer equations from gas to kiln wall
to the charge used by Sass (Reference 40) are shown in Table 8, taken from
that reference.
The most salient feature of these equation is the use of the emprical
correlation
h = 0.05 (Cg/Sx)0'67
for gas-to-wall heat transfer and
h = 0.25 (Cg/Sx)°-67
for the corresponding wall to charge heat transfer. These correlations are
at best applicable as a rule of thumb, useful to calculate an approximate
kiln length.
Solids and gas emissivity calculations also present some difficulties,
because of gas-borne particle radiation and uncertainties of the emissivities
of solids in the charge. If and when better documented heat and mass transfer
correlations are developed, an analytical treatment of the heat transfer in
71
-------
TABLE 8. MEAT-TRANSFER COEFFICIENT CORRELATIONS
Heat-transfer path
Gas to inner kiln wall
Gas to solid
Inner kiln wall to solid
Inner kiln wall to outer kiln
wall
Outer kiln wall to ambient air
hl =
h2 =
h3 =
h,
3cc
h3RAD
hd =
*\
h5 ~
0.05 (Gg/Sx)0'67
0.05 (Gg/Sx)0'67
h + h
3cc 3RAD
= 0.25 (G /S )0>(
= 0.173 x 10"8
k
ro-ri
*
2.5
hi
+ 0.173 x ID'8 Eg(Tg4 - Tw4)/(Tg - Tj
+ 0.173 x ID'8 Eg(Tg4 - Ts4)/(Tg . TS)
57
fEs
-------
the rotary kiln incinerator (and hence a good basis for scale-up) will becone
possible because the problem is then reduced to solving a periodic heat con-
duction problem which is defined by classical non-dimensional parameters: a
Fourier number and two Biot numbers, as discussed below.
3.5.2 Non-dimensional Parameters
Heat is received from the gases by the kiln wall, ana transported to the
charge. If we follow an element rdO of the kiln wall from the position marked
"origin" on Figure 15, as it emerges from contact with the charge and is rotated
through the ange (2ir-9Q), that wall element will be progressively heated by
the combustion gases. Initially the temperature distribution in that element
was TQ(r), after a rotation of (2ir-90), i.e., after a time period of (2ir-90)/
seconds (N is the number of revolutions per second), the temperature
distribution in the element is T, (r). Neglecting axial conduction, the
equations describing the heat transfer are:
_|T_ . 0,2T . a(lr_ + _L _|I_ +fr
3t arz r 3r 30Z
_3]_ = h,(T -T), r = r.
3r ' s '
T = T0(r,9), t = o
where o is the the thermal diffusivity of the brick. If (3 T/39 ) is small
compared to the radial derivatives, the solutions will be of the form
T-T.
-r-4 = f(Fo, Bi)
o 'g
_2
where Fo is the Fourier number Fo = at/r and Bi is the Biot number Bi =
h,r/k. Closed form solutions are cumbersome and require simplifying
assumptions because radiation causes the heat transfer coefficient to
depend on temperature*.
Imber and Paschkis (Reference 41 ) give closed form solutions for two
limiting cases.
73
-------
HEAT TRANSFERRED
FROM GAS TO WALL
DIRECTION OF
ROTATION
KILN WALL
HEAT TRANSFERRED
FROM GAS TO CHARGE
HEAT TRANSFERRED
FROM WALL TO CHARGE
Figure 15. Schematic diagram showing the heat-flow paths and
Nomenclature for a typical section in a rotary kiln.
74
-------
Numerical solutions are quite readily accessible. At the end of the time
period (2ir-90)/2nN, the kiln wall element, rd9, which we are following,
reaches the upper border of the charge. The temperature distribution in
the element is known, and heat is transferred from the element to the (cooler)
charge. The equations for the wall in contact with the charge is:
-ff - ^
_3J_ = h (T-T ). r = r.
8r * b ]
T = T(FVBis), t--J
where
Fo] = (
-------
solids, which depends on the slope of the rotary kiln axis and on the varying
friction-related properties of the charge. The latter seem to have little
influence since the empirical correlation for "time of passage" given in
Perry's Handbook does not include charge properties:
9 = 0.19(L/D)/(SN) (min.)
where
L, D = kiln length and diameter
s = slope of kiln (m/m)
N = RPM
Evaluation of the residence time of combustion gases of the auxiliary
fuel is straightforward, that of gases evolving from the charge more dif-
ficult. Mass transfer from the charge to gas stream will have to be
addressed, as well as the chemical reactions in the charge and in the gases.
3.5.4 Scale-up Parameters
Despite all the uncertainties and gaps in the definitions of heat and
mass transfer, limited scale-up of existing kilns can be undertaken by attempt-
ing to keep the values of the Fourier, Biot and Reynolds numbers unchanged
or as close as possible. The limiting assumptions are that the heat and mass
transfer coefficients will not change greatly because of scale-up, that an
increase in charge depth will not drastically change its behavior (particle
size in the charge will not scale-up, but the height will, hence'charge mixing
may change), and that the rotational speed decrease needed to limit stress
levels will not impair heat transport.
3.5.5 Summary
Rotary kiln incinerators are the most versatile of all devices for
hazardous waste incineration. Mathematical examination of the processes
within the kiln does not present insurmountable difficulties; however, a
sufficient experimental basis is not available at present. A concerted
effort towards the development of an experimental basis and of an analytical
description of the thermochemical phenomena in the rotary kilns should be
undertaken.
76
-------
SECTION 4
THERMOCHEMICAL AND KINETIC CHARACTERIZATION OF WASTES
Temperature - residence time relationships obtained in the laboratory have
been the basis for establishing the requirements for the thermal destruction of
hazardous wastes. The University of Dayton Research Institute has developed a
special laboratory technique, incorporating a two stage quartz system, which
was successfully used to determine the thermal decomposition properties of poly-
chlorinated biphenyls (PCB's) and of Kepones ' and permitted the selection
of the appropriate incineration systems and of their operating conditions to
safely dispose of the highly toxic wastes containing Kepones or PCB's.
In this system the pesticide was first converted to the gas phase, then
exposed to the high-temperature destruction conditions. Critical parameters
of temperature and residence time were accurately measured. Thermal destruc-
tion testing was conducted with three pesticides: Kepone, Mi rex, and DDT.
Both the Kepone and DDT molecules, at a residence time of ~1 second, were
essentially destroyed at 500°C; however, Mirex, at the same residence time,
required 700°C for destruction. The thermal destruction properties of PBC's
and related compounds were similarly determined. Initial decomposition occur-
red at about 640°C; 99.995% molecular destruction was found at 1000°C. Also,
it was determined that PCB's (and certain related compounds) thermally decompose
to low molecular weight products. An extension of the above technique to in-
clude thermal destruction in the presence of oxygen is being planned and should
be very valuable.
The experimental techniques are essential and irreplaceable. However, the
effort and time involved in conducting the experiments are substantial and can
be greatly assisted by also conducting a thermochemical kinetic analysis of the
combustion of the wastes in air. Equilibrium analysis cannot eliminate the
need for the experimentally acquired reaction rate information on some of the
77
-------
hazardous compounds, but can serve to decrease the number of experiments re-
quired and provide essential information on the formation of dangerous
intermediate products.
TRW has developed a method for examining the combustion of pesticides and
other organic compounds by determining the equilibrium product distribution
44
under various operating conditions . A way to analytically determine the
upper limit for the residence time required was also developed during pesticide
incineration tests. It was based on the finding that the slowest and control-
ling step in high temperature combustion is the oxidation of the initially
formed carbon monoxide.
The following sections describe the TRW approach to the thermal equili-
brium and kinetic analysis for pesticide incineration.
4.1 THERMOCHEMICAL ANALYSIS
4.1.1 TRW Chemical Analysis Program
Determination of the theoretically expected equilibrium products of
pesticide combustion requires appropriate thermochemical data for each
potential product. In addition, a mathematical model is required to combine
these data in order to ascertain the equilibrium quantities of each product.
The data and model are used under various conditions of temperature, pressure,
and reactant combinations to simulate each chemical system and reaction stage.
The primary data base used by the TRW Chemical Analysis Program included
the JANAF thermochemical tables prepared by the Dow Chemical Corporation and
a previous TRW effort to characterize the equilibrium product distribution of
waste plastics combustion/pyrolysis. The JANAF tables include the potential
oroducts methane, acetylene, ethylene, ethylene oxide, formaldehyde, and the
various methyl chlorides among the organics; but no higher organics are
represented. The TRW addition include data for the following classes of
compounds:
Higher saturates, e.g., alkanes
Higher unsaturates, e.g., alkenes
Alcohols
78
-------
Glycols
Higher aldehydes
Ketones
Ethers
High Aliphatic chlorides
Amines
High expoxides
Aromatics
Alicyclics, e.g., cyclopentadiene
Polynuclear aromatics, e.g., naphthalene
Organic acids, e.g., acetic acid
Chlorinated aldehydes, e.g., chloral
Chlorinated aromatics, e.g., chlorobenzenes and chlorophenols
Chlorinated alicyclics, e.g., hexachlorocylopentadiene.
4.1.2 Equilibrium Product Distribution Analyses
The purpose of the equilibrium product distribution analyses is to provide
a sound thermochemical basis for the determination of the species resulting
from the combustion or pyrolysis of a hazardous waste. For example, in the
combustion studies, for pesticide incineration the equilibrium product
distributions were examined for the temperature range 800 K (1000 F) to 1650 K
(2500 F) and for three air/fuel ratios:
Case 1: Stoichiometric amount of air
Case 2: 130 percent of Stoichiometric amount of air
Case 3: 70 percent of Stoichiometric amount of air.
In the pyrolysis studies, the thermochemical computer program was applied
to determine not only the equilibrium product distribution, but also the
secondary thermodynamically feasible reaction products. This was accomplished
simply by eliminating from consideration at each stage those products that
were thermodynamically favored but less likely to be formed from the kinetic
standpoint. The general conclusions from the thermochemical analysis and the
effects of temperature and the type of pesticide formulation on the equilibrium
product distribution are discussed, in part, in the following sections.
79
-------
4.1.3 Combustion of Pesticides
The distribution of the equilibrium products from the combustion of chlori-
nated hydrocarbon pesticide formulations depends on the temperature, the air/
fuel ratio, and the carbon-hydrogen-chlorine-oxygen ratio in the original
formulation. A brief summary of the primary equilibrium products and their
relative concentrations as obtained from the thermochenrical calculations is
presented in Table 9.
Effects of Air/Fuel Ratio—
As indicated in Table 9, the product species 0, NO, N02> CIO and HOC1
are only found under excess air conditions, where as the product species CH^,
NH,, and H are only found under oxygen deficient conditions. In addition, the
formation of C02, Cl, C12, HgO, 02, and OH is favored at higher air/fuel ratios,
and the formation of CO and H2 is favored at lower air/fuel ratios.
Effects of Temperature—
The effects of temperature on the relative concentrations of the equilib-
rium product species are illustrated in Figure 16 for the case of combustion
of the 12 percent lindane emulsifiable concentrate with 30 percent excess air.
In general, the formation of CO, Cl, CIO, NO, N02, H, and OH is favored by
increasing the reaction temperature, whereas the formation of CH4> C02> C12,
and NH3 is favored by decreasing the reaction temperature. Increasing the
reaction temperature also favors the formation of H2 under stoichiometric or
excess air conditions, and the formation of H20 under oxygen deficient con-
ditions. In addition, analyses of the results of the thermochemical calcula-
tions led to the following specific conclusions:
0 Thermochemical analysis predicts that HCL formation is highly
favored at pesticide incineration temperatures (1100 to 1650 K
range), and that Cl?. CIO and HOC! are found in only trace
quantities. Cl is found in concentrations above 100 ppm at
the higher incineration temperatures, but should react readily
with the OH radical to form HC1 and 02 at lower temperatures.
On the other hand, although the equilibrium concentration of
Cl2 increases to as much as 5000 ppm at the lower temperatures
for the highly chlorinated pesticide formulations, the relative
slowness of the reaction:
2HC1 + 0.502 5 H20 + C12
80
-------
TABLE 9. EQUILIBRIUM PRODUCT DISTRIBUTION FROM THE
COMBUSTION OF PESTICIDE FORMULATIONS
Case 1: Combustion with stoichiometric amount of air
Major equilibrium products: COg, HgO, HC1, Ng.
Minor equilibrium products: CO (< 1 to 460 ppm), H2 (< 1 to 108 ppm),
NO (<1 to 71 ppm), Cl (<1 to 391 ppm),
C12 (<8 ppm), 02 (<1 to 180 ppm),
OH (<1 to 47 ppm).
Case 2: Combustion with 130% of stoichiometric air
Major equilibrium products: C02> HgO, HC1, N2, 0^.
Minor equilibrium products: NO (1 to 1150 ppm), CO (<1 to 21 ppm),
N02 (<2 ppm), 0 (<6 ppm), OH (<1 to
165 ppm),
Cl M to 1433 ppm), C12 (<1 to
5073 ppm),
CIO (<7 ppm), HOC1 (<1 to 12 ppm),
H2 (<6 ppm)
Case 3: Combustion with 70% of stoichiometric air
Major equilibrium products: CO, C02, H2> H20, HC1, N2
Minor equilibrium products: Cl<4 (<1 to 8500 ppm), OH (<2 ppm),
NH3 (<1 to 55 ppm), Cl (< 1 to 25 ppm),
H (<1 to 20 ppm)
81
-------
1000 1100 1200 1300
TEMPERATURE (K)
1400
1500
1600
Figure 16. Equilibrium mole fraction of product species as a
function of temperature from the combustion of 12
percent Undane emu!sifiable concentrate with 30
percent excess air.
82
-------
means that the reaction products will not have sufficient time
to equilibrate, and that under actual incineration conditions
HC1 will still be the only predominant chlorinated compound
found in the combustion products. Analyses of the grab samples
taken during pesticide incineration tests confirmed that HCL
was the only major chlorinated species found in the combustion
products.
• CO formation increases with increasing reaction tempt.dLures.
At lower temperatures, the water-gas shift reaction:
H20 + CO 5 H2 + C02
becomes the principal mechanism for equilibrating the amounts of
CO and C02 present in the combustion products. To minimize the
CO concentration in the incinerator effluent, the initial cooling
of the incinerator gases should therefore be slow enough to allow
for equilibration of all product species. The measured CO con-
centrations during pesticide incineration tests were found to be
considerably higher than the corresponding calculated equilibrium
CO concentrations, indicating that equilibrium for the water-gas
shift reaction was not attained under test conditions.* From the
equilibrium point of view, the concentrations of almost all the
undesirable combustion products (with the exception of Cl2) such
as NO, N02, Cl, CIO, and CO, increase with increasing reaction
temperature. Pesticide incineration should therefore be con-
ducted at temperatures sufficiently high to cause complete
combustion, and yet not excessively high as to lead to unaccept-
ably large amounts of nitrogen oxides and carbon monoxide in the
incinerator effluent.
Effects of Types of Pesticide Formulations—
The principal effects of different types of pesticide formulation on
the equilibrium product distribution are the relative amounts of chlorinated
species formed, which depend primarily on the Cl/C mole ratio of the formu-
lation combusted. For example, the equilibrium concentrations of CIO and HOC1
are found to be less than 1 ppm at Cl/C mole ratios of less than 0.0318
(corresponding to the mixture of 2,4-D ester, 2,4,5-T ester, and No. 2 fuel
oil). The effects of Cl/C mole ratio on the equilibrium HC1 and C12 con-
centrations are presented in Figures 17 and 18, respectively, and indicate
that the highly chlorinated pesticide formulations (such as the 72 percent
chlordane emulsifiable concentrate and the 2,4,5-T ester) should be diluted
with a fuel oil prior to incineration to minimize potential HC1 corrosion
and to safeguard against C12 formation.
The oxidation of CO is discussed in greater detail in Section 4.1.4.
83
-------
10
-3
0 0.05 0.10 0.15 0.20 0.25 0.30
Cl/C MOLE RATIO FOR PESTICIDE FORMULATION
Figure 17. Equilibrium HC1 concentration in combustion product gas
resulting from the incineration with 30 percent excess air.
84
-------
10
0.05 0.10 0.15 0.20 0.25 0.30
C1/C MOLE RATIO FOR PESTICIDE FORMULATION
Figure 18. Equilibrium Cl2 concentration in combustion product gas
resulting from the incineration pesticide with 30 percent
excess air.
85
-------
4.1.4 Oxidation of Carbon Monoxide
During the pesticide incineration tests, it was found that the slow and
controlling step in the high temperature combustion process appeared to be the
oxidation of the carbon monoxide initially formed. The rate of oxidation of
carbon monoxide is therefore of fundamental importance in specifying the
operational criteria for pesticide incineration. To determine the applica-
bility of the previously reported rate equations in the case of pesticide
incineration, a computer program was developed to calculate the carbon monoxide
concentrations along the pesticide incinerator, utilizing the reported rate
equations. The calculated carbon monoxide concentrations were then compared
with the experimentally measured carbon monoxide concentrations under the
same set of operating conditions.
Since there has been general disagreement among various investigators on
the high-temperature oxidation rates of carbon monoxide, the overall rate
expression
A exp (-E/RT)
was incorporated into the computer program with different values of the
pre-exponential factor A, the activation energy E, and the pressure, oxygen
concentration and water concentration dependence exponents p, m, and n as
determined by six groups of investigators (see Table 10).
Data collected in tubular reactor at various stations along the length
of the reactor was used to verify the six proposed rate equations. The
results of these computations for 25 test cases have indicated that:
1) The calculated CO concentrations according to Sobolev are in
good agreement with the measured CO concentrations in the
pesticide incinerator, for incinerator temperatures above
1200 K.
2) The calculated CO concentrations according to the other five
groups of investigators are far lower than the measured CO
concentrations in the pesticide incinerator.
This conclusion is not too surprising as the Sobolev rate expression is the
only one derived from direct measurements of the rate of oxidation of CO
86
-------
TABLE 10. HIGH-TEMPERATURE OXIDATION RATES OF CARBON MONOXIDE
d C,
A exp (-E/RT)
/ p V
H
fco fo.
Investigator
Fen i more and Jones
Williams, et al.
Hottel, et al.
Friedman and Nugent
Sobolev
Kozlov
Applicable
\
0 to 1.0
0 to 1.0
0 to 1.0
0 to 1.0
0.05
0.05
0.05
0.05
. /liteAP-1 1*
" ^mole 1 sec
1.2 x 109
1.8 xlO10
1.2 x 1011
1.85 x 107
6.44 x 108
5.86 x 107
1.90 x 1013
2.01 x 1012
E
cal/mole**
24000
25000
16000
20000
27000
27000
32000
32000
P
2
2
1.8
1.5
2
2.0
2.5
2.5
m
1
0.5
0.3
0
1
0.2
1
0.25
n
0
0.5
0.5
0.5
0
0
0.5
0.5
00
*1 liter = 10"3 m3
**
1 calorie/mole = 4.19 joule/mole
-------
in the afterburning zone instead of the flame front. The rate constants of the
carbon monoxide reaction taking place in the flame front are generally several
orders of magnitude larger than the rate constant of the chemical reaction
taking place in the afterburning zone. However, at incinerator temperatures
below 1200 K the Sobolev rate equation predicted a much slower carbon monoxide
oxidation rate than that measured experimentally. As illustrated in Figure 19.
the CO concentrations from both experimental measurements and determined by
the Sobolev rate equation are considerably higher than the equilibrium CO
concentration.
4.2 THE ROLE OF THERMOCHEMICAL EQUILIBRIUM ANALYSIS IN WASTE-INCINERATOR
MATCHING
The number and quantities of chemical wastes are increasing, their dis-
posal in bodies of water or in landfills is becoming less and less acceptable.
As a consequence, clean-stack incineration will have to accommodate a fast
growing list of chemical compounds. Selection of an incineration system
capable of insuring complete and safe destruction can be aided by further
developing the methodology which was used on chlorinated hydrocarbons and
described in this section. The approach to selecting the means for the complete
destruction hazardous species would include the following steps:
• Obtain the chemical composition and proportions of the waste.
• Determine the equilibrium composition of the products of
combustion in oxygen rich and oxygen deficient mixtures in
wide range of temperatures using the previously discussed
computer model.
• Examine the initial equilibrium products distribution and
determine those products that are not kinetically favored.
• In the next stage of calculations, determine another equilib-
rium product distribution by eliminating from consideration
those products that are thermodynamically favored but less
likely to be formed from the kinetic standpoint.
• Reiterate the above two steps
• Examine all of the above equilibrium products distributions
to determine which initial compounds and possible intermediate
species should be subjected to a laboratory test series (UDRI).
For many compounds laboratory tests will not be needed because the
necessary conditions to achieve thermal decomposition and oxidation of
88
-------
1000
1000 1100 1200
TEMPERATURE (K)
1300
1400
1500
Figure 19. Comparison of experimental, kinetically determined, and
equilibrium values of CO concentrations in the incin-
erator gas.
89
-------
hazardous soecies can be safely established from prior experience. On the
other hand, new waste formulations may introduce new compounds and new
intermediate products; the existance of the latter could be forecast by
the above described equilibrium computations.
The development of the proposed analytical approach hinges on the intro-
duction of reaction kinetics into a thermodynamic equilibrium analysis. The
science of reaction kinetics is not as well developed or securely based as
that of equilibria. The success of the proposed appraoch will depend on the
talents of the kineticists in using all available kinetic and thermodynamic
information to predict the rate-dependent behavior of hazardous compounds.
The selection of an incineration system for a candidate waste will
proceed as follows:
1) Chemical analysis of the waste.
2) Thermodynamic and kinetic analysis on the waste components
incineration.
3) Laboratory determination of the required temperature history
of species, selected in step 2.
4) Selection of the incinerator type.
5) Pilot incineration (if needed).
6) Full-scale Incineration.
90
-------
SECTION 5
SUMMARY AND RECOMMENDATIONS
Hazardous waste Incineration in large commercial facilities is of growing
importance because of the increasing volume of chemical wastes that must be
safely destroyed. The study reported herein started with the selection of
four incinerator types, based on their applicability to the largest volume and
variety of wastes, and proceeded to identify the available experimental and
analytical methods which could be used to predict the waste destruction effi-
ciency and facilitate the scale-up of each incinerator type. Previous TRW
studies and the open literature surveyed showed that neither the accurate pre-
diction of waste incineration efficiencies, nor the establishment of adequate
scale-up methods was possible today. Analytical methods and experimental
correlations describing some of the important combustion and heat and mass-
transfer mechanisms are available.
Analysis and modeling of incinerators have been most widely applied to
liquid and fluidized bed incinerators; the work was not specifically directed
towards, but is to a large degree applicable to, hazardous waste destruction.
The continuing development of analytical methods and increasing volume of data
pertaining to fluidized beds and liquid injection furnaces should be followed
and exploited for that purpose. Not much analytical work or evidence of data
collection was found for the rotary kiln or multiple hearth incinerators, so
that the expectations of finding ongoing analytical work for these devices
are unrealistic.
The mathematical modeling of rotary kilns does not appear to present
unsurmountable difficulties. Still, some very basic information on the be-
havior of the solid charge and heat transfer between charge and kiln walls
and charge and combustion gases is lacking. The multiple hearth incinerator
is the device least amenable to analysis, hence in this case prediction will
91
-------
have to rely heavily on the acquisition of data, and on purely empirical
correlations, at least in the near future.
The sheer size of the waste disposal problem warrants development of
methodology for hazardous waste incineration. While the many gaps in the
available information do not permit detailed modeling of incinerators, the
pressing need to establish better design criteria and reduce the costs and
risks involved in the testing of hazardous waste incinerators, a program to
develop scale-up offers promise of success. For example, while the detailed
equations governing the behavior of an incineration system are complex and
often poorly understood, techniques such as dimensional analysis can be used
to identify key parameters which govern the similarity of the various mecha-
nisms of the destruction of hazardous waste components.
The term "similarity" needs to be qualified in order to be useful. We
know that similarity of fluid flow in two geometrically similar systems is
achieved if the Reynolds numbers are the same in both systems, assuming that
the fluid is incompressible and gravitational forces small. If these assump-
tions do not apply, the Mach number and Froude numbers must also be replicated
(45). It has also been shown that to achieve a similarity in convective heat flux
between a flowing fluid and a surface the Nusselt number must be reproduced.
The relationship Nu = f(Re, Pr) can be derived through dimensional analysis
or by examination of the Navier Stokes and the energy equations (45, p 253).
In the case of incineration of hazardous waste, the similarity should be
viewed in light of the objective of the incinerator. A potential definition
of similarity is as follows:
Two geometrically similar incinerators of differet.t capacity are
deemed similar if, when burning the same type of waste, they produce
the same level of destruction and their stack gases and residues
respectively have the same composition.
Since incinerators will be handling wide varieties of wastes it is not
possible to use emission levels directly. However, the degree of waste de-
composition or oxidation depends almost entirely on the three T's of
incineration: time, temperature and turbulence. The alliteration is somewhat
misleading, as turbulence mostly stands for mixing and dispersion (e.g.,
exposing fuel to oxidizer, and contact between solids and gases) and time
92
-------
refers to residence time of a reactant or fuel above a specified temperature
level. It appears, from previous research, that if two geometrically similar
devices have identical residence times at specified temperature levels and
provide for equal levels of mixing at those temperatures, the same level of
waste destruction (through thermal decomposition and oxidation) will be
achieved. A research program can be directed to determine which dimensonless
parameters need to be taken into account in order to insure incinerator
similarity. For example, in a rotary kiln similarity may require equivalent
Reynolds. Biot and Nusselt numbers in different zones of the kiln. Because of
material restrictions and because of contradictory demands that may be pre-
sented by each mechanism (e.g., high velocity for equivalent mixing vs low
velocity for equivalent residence time) full similarity may not be achieved.
Analytical and experimental work will be needed to determine which parameters
influence similarity of emissions most.
We recommend that the following tasks be included in the development of a
hazardous waste incineration program:
• Continue the thermal decomposition experiments of hazardous
components of waste.
• Initiate similar experiments on decomposition in the presence
of oxygen.
• Canvass the major manufacturers of incinerators, furnaces and
fluidized beds to determine the extent and applicability of
in-house and government funded development programs.
• Collect existing data on gas composition CO, 03, unburned hydro-
carbons etc., upstream of pollution control equipment, and data
on the effectiveness of control equipment installed on incin-
erators .
• Develop plan and recommendations for instrumentation installations
on commercial incinerator and for data gathering.
• Develop an analytical model for rotary kiln incineration. (Imber
and Paschke's paper is a good starting point.)
93
-------
• Initiate the development of analysis using a statistics approach,
leading to the evaluation of the probability of emissions ex-
ceeding preimposed levels. A statistical approach is appropriate
since the waste decomposition and oxidation take place in a
distribution of temperature and residence time, and mixing is not
uniform.
• Investigate a systematic approach to combining furnace models
(for example by the zone method) with experimental results to
build predictive analytical tools.
The development of a waste incineration methodology is by its very nature
a continuing effort as the composition of the wastes will certainly change
with time.
94
-------
REFERENCES
1. Destruction of Chemical Wastes In Commercial Scale Incinerators. Final
Report, Phase II (to be published under NTIS).
2. Wall, C. J., J. T. Graves, and E. J. Roberts. How to Burn Salty Sludges.
Chem. Engrg., April 14, 1975. (p. 77)
3. Destructing Chemical Wastes in Commercial Scale Incinerators. Technical
Summary, Volume I, NTIS No. PB257 710/4WP.
4. Destructing Chemical Wastes in Commercial Scale Incinerators. Facility
Test Plans, Volume II, NTIS No. PB257 710/4WP.
5. The Marquardt Company. Destroying Chemical Wastes in Commercial Scale
Incinerators, Facility Report No. 1, NTIS No. PB265541.
6. Surface Combustion Division, Midland-Ross Corp. Destroying Chemical
Wastes in Commercial Scale Incinerators, Facility Report No. 2, NTIS
No. PB268232.
7. Systems Technology, Destroying Chemical Wastes in Commercial Scale
Incinerators, Facility Report No. 3, NTIS No. 265540.
8. Zimpro Inc. Destroying Chemical Wastes in Commercial Scale Incinerators,
Facility Report No. 4, NTIS No. 267987.
9. 3M Company. Destroying Chemical Wastes in Commercial Scale Incinerators,
Facility Report No. 5 (to be published under NTIS).
10. Rollins Environmental Services. Destroying Chemical Wastes in Commercial
Scale Incinerators, Facility Report No. 6, NTIS No. PB270897.
11. Laboratory Analysis Results from the Chem-Trol/St. Lawrence Cement
Facility Test. Destroying Chemical Wastes in Commercial Scale Incinerators,
Facility Report No. 7.
12. Disposal Process Descriptions, Ultimate Disposal, Incineration and Pyrolysis
Processes. Volume III Recommended Methods of Reduction, Neutralization,
Recovery or Disposal of Hazardous Wastes. Prepared by TRW for EPA Contract
No. 68-03-0089.
13. At-Sea Incineration of Herbicide Orange Onboard the M/T Vulcanus,
600/2-78-086, April 1978.
95
-------
14. Palmer, H. B., and J. M. Beer, editors. Combustion Technology. Academic
Press. New York and London, 1974. (p. 374)
15. Hottel, H. C., and A. F. Sarafim. Radiative Transfer. McGraw-Hill Book
Co., 1976. (p. 243)
16. Bueters, K. A., J. G. Gogole, and W. W. Habelt. Performance Prediction
of Tangentially Fired Utility Furnaces by Computer Model. 15th (Intl)
Symp. on Comb., Tokyo, 1974. (p.1245)
17. Habelt, W. W., and A. P. Selker. Operating Procedures and Prediction
for NOX Control in Steam Power Plants. Presented at the Central States
Section of the Combustion Institute Meeting, Madison, Wisconsin, March
26 & 27, 1974.
18. Ingebo, R. D., and H. H. Foster. Drop-Size Distribution for Crosscurrent
Break-up of Liquid Jets in Airstreams. NACA TN 4087, 1957.
19. Ingebo, R. D. Vaporization Rates and Drag Coefficients for Isooctane
Sprays in Turbulent Air. NACA TN 3265, 1957.
20. Ingebo, R. D. Penetration of Drops into High-Velocity Airstreams.
NACA TM X-1363, 1966.
21. Simmons, H. C. The Correlation of Drop-Size Distributions in Fuel
Nozzle Sprays, Parts I and II. Engineering for Power, July 1977.
22. Faeth, G. M. Current Status of Droplet and Liquid Combustion. Prog.
Energy Combust. Sci., Volume 3, 1977. (pp. 191-224)
23. Hinze, J. 0., and M. Mil born. Atomization of Liquids by Means of
Rotating Cups. J. Applied Mechanics, June 1950.
24. Williams, A. Combustion of Droplets of Liquid Fuels: A Review.
Combustion and Flame 21. 1973. (pp. 1-31)
25. Wood, B. J., H. Wise, and S. H.Inami. Combustion and Flame 4, 1960.
(p. 235)
26. Shyu, R. R., C. S. Chen, G. 0. Goudie, and M. M. El-Wakil. Behavior
of Combustible Fuel Drops at High Temperature. Fuel, 1972, Volume 51,
April, (p. 135) ^
27. Faeth, G. M., and R. S. LazaV. A.I.A.A.J 9, 1971. (p. 2165)
28. Dickerson, R. A., and M. D. Schuman. J. of Spacecraft 2, 1956. (p. 99)
29. Onuma, Y., and M. Ogasawara. Studies on the Structure of a Spray Com-
bustion Flame. 15th Symposium (Intl) on Combustion, Tokyo, Japan, 1974.
(p. 453)
30. Prof. Bee'r, Y. M. Comment to the above paper.
96
-------
31. Hawthorne, W. R., D. S. Weddel, and H. C. Hottel. Mixing and Combustion
in Turbulent gas Jets. Third Symposium on Combustion, Flame and Ex-
plosion Phenomen. (p. 266)
32. Corrsin, S. Simple Theory of an Idealized Turbulent Mixer. The Johns
Hopkins University, Baltimore, Maryland, 1957. (p. 329)
33. Review of proposed action to dispose of Orange Herbicide by Incineration.
Final Report prepared by A.D.L. Inc., for the U.S. Air ccrce, Contract
F33615-75-C-5116, June 30, 1975.
34. Kunii, D., 0. Levenspiel, John Wiley and Sons, Inc. Fluidization
Engineering, 1969.
35. Toomey, R. D., and H. F. Johnstone. Chem. Engrg. Progr., Volume 48,
1958. (p. 220)
36. Hario and Wen. An Assessment of Fluidized-Bed Modeling. Fluidization
Theories and Applications, AIChE Symposium Series, No. 161, Volume 73,
1977.
37. Merrick and Highley. Particle Size Reduction and Elutriation in a
Fluidized Bed Process. Recent Advances in Air Pollution Control, AIChE
Symposium Series, No. 137, Volume 70, 1974.
38. Me Cabe, W. L., and J. C. Smith. Unit Operations in Chemical Engineering
McGraw-Hill, 1967. (p. 855)
39. Me Cormick, P. Y. Gas Velocity Effects on Heat Transfer in Direct Heat
Rotary Dryers. Chem. Engrg. Progr., No 6, Volume 58, June 1962. (p. 57)
40. Sass, A. Simulation of the Heat Transfer Phenomena in a Rotary Kiln.
I&EC Process Design and Development, No. 4, Volume 6, October 1967.
41. Imber, M., and V. Paschkis. A New Theory for a Rotary-Kiln Heat Ex-
changer. Intl J. Heat Mass Transfer, Volume 5, 1962. (pp. 623-638)
42. Laboratory Evaluation of High-Temperature Destruction of Polychlorinated
Biphenils and Related Compounds. EPA Report 600/2-77-228, December 1977.
43. Laboratory Evaluation of High-Temperature Destruction of Kepone and
Related Pesticides. EPA Report 600/2-76-299, December 1976.
44. Thermal Degradation of Military Standard Pestidice Formulations TRW
Report No. 24768-6018RU-00 prepared for the U.S. Army Medical R&D Command,
1 December 1974.
45. Schlichting, H. Boundary Layer Theory. Pergamon Press. New York, London
and Paris, 1955.
97
-------
BIBLIOGRAPHY
FLUIDIZED BED
Baron, R. E., J. L. Hodges, and A. F. Sarofim. Mathematical Model for
Predicting Efficiency of Fluidized Bed Steam Generators. Industrial
Liaison Program, '"September 1977.
Be'er, J. M. The Fluidised Combustion of Coal. Industrial Liaison Pro-
gram, August 15. 1976.
Bergougnou, M.A. et C. G. J. Baker. Etude Sur L'Extrapolation et Le
Calcul Des Grands Lits Fluidisc's. Chimie et Industrie, Genie Chimique,
Volume 106 No. 4, Fevrier 1973.
Borghi, G., A. F. Sarofim, and J. M. Beer. A Model of Coal Devolatili-
zation and Combustion in Fludized Beds. Presented at AIChE 70th Annual
Meeting, New York, November 1977.
Bowling, K. McG., and P. L. Waters. Fluidized Gasification of Solids
Fuels. Mechanical & Chemical Engrg Trans, November 1968.
El-Kaissy, M. M., and G. M. Homsy. Instability Waves and the Origin of
Bubbles in Fluidized Beds. Intl J. Multiphase Flow, Volume 2. (pp 379-
395)
Grace, J. R. Fluidization and Its Application to Coal Treatment and
Allied Processes. 'AIChE Symposium Series: Fluidization and Fluid-
Particle Systems, No. 141, Volume 70.
Halow, J. S. Fluidized-Bed Applications. AIChE Symposium Series:
Fluidization and Fluid-Particle Systems, No. 141, Volume 70.
Kehoe, P. W. K. Ph.D. Thesis, 1969 (Cambridge: University of Cambridge),
Montagna, J. C., F. F. Nunes, G. W. Smith, G. J. Vogel, and A. A. Jonke.
High Temperature Fluidization and Agglomeration Characteristics of
Limestone and Coal Ash Particle Systems. Presented at the 70th Annual
AIChE meeting.
Pereira, F. J., and J. M. Beer. A Mathematical Model of No Formation
and Destruction in Fluidized Combustion of Coal. Fluidization,
Cambridge University Press. 1978.
98
-------
Pyle, D. L. Fluidized Bed Reactors: A Review. Adv. Chem. Series 109, 1972.
(p. 106)
Rao, C. S. R. Fluidized-Bed Combustion Technology: A Review. Combustion
Science and Technology, Volume 16, 1977. (p. 215-227)
Shaw, I. D., T. W. Hoffman, A Orlickas, and P. M. Reilly. The Hydrogenolysis
of n-Butane on a Nickel on Silica Catalyst: II Fluidized Bed Studies. The
Canadian Journal of Chemical Engineering, Volume 50, October 1972.
Squires, A. M. Clean Power from Coal. Number 3948, Volume 169, 28 August
1970.
Wen C. Y., and S. Dutta. Research Needs for the Analysis, Design, and Scale-
up of Fluidized Beds. AIChE Symposium Series: Fluidization Theories and
Applications, No. 161, Volume 73, 1977.
INCINERATION AND COMBUSTION
Cox, Jr. G. B. Multiple Jet Correlations for Gas Turbine Engine Combustor
Design. Journal of Engineering for Power, April 1976. (p. 265)
Ivanov, Yu. V. Effective Combustion of Overfire Fuel Gases in Furnaces.
Estonian State Publishing House, Tallin, 1959.
Onuma, Y., and M. Ogasawara. Studies on the Structure of a Spray Combustion
Flame. 15th Symposium (Intl) on Combustion, Tokyo, 1974. (p. 453)
Stephens, R. H., and C. M. Mohr. Reduction of Combustible Emissions in
Municipal Incinerators Using Overfire Jets. AIChE Symposium Series, No. 147
Volume 71.
Swithenbank, J., I. Poll, M. W. Vicent, and D. D. Wright. Combustion Design
Fundamentals. 14th Symposium on Combustion, Flames in Furnaces and Combustor.
(P. 627)
Thring, M. W., and M. P. Newby. Combustion Length of Enclosed Turbulent Jet
Flames. 4th Symposium (Intl) on Combustion, MIT September 1952. (p. 789)
Williams, F. A. Combustion Theory. Addison-Wesley Publishing Company, 1965.
Fisheries and Environment Canada. Burning Waste Chlorinated Hydrocarbons in
a Cement Kiln, EPS-4-WP-77-2, March 1977.
Igham County Solid Waste Recycling Project. Igham County Board of Public
Works, August 27, 1974.
Incinerator Overfire Mixing Study-Demonstration of Overfire Jet Mixing.
Prepared by Arthur D. Little, Inc., prepared for Office of Research and
Monitoring, June 1975.
99
------- |