Title Page
 Table of Contents
 List of Tables
 List of Figures
 Incinerator system
 Review of data from three years...
 Kinetic modelling
 Effect of chlorinated input on...
 The new combustion control...
 Experiments with old and new control...
 Discussion, conclusions, and...
 Appendix A: Detailed description...
 Appendix B: Details on the...
 Appendix C: Data calculations
 Appendix D: Vost emissions
 Appendix E: Kinetics code program...
 Appendix F: Kinetics code input...
 Appendix G: Temperature-time profile...
 Appendix H: Base case kinetics...
 Appendix I: Additional details...
 Appendix J: Additional details...
 Biographical sketch

Title: Combustion and pre-combustion control methods to minimize emissions from modular incinerators /
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00097386/00001
 Material Information
Title: Combustion and pre-combustion control methods to minimize emissions from modular incinerators /
Physical Description: xi, 239 leaves : ill. ; 29 cm.
Language: English
Creator: Wagner, John Charles, 1963-
Publication Date: 1992
Copyright Date: 1992
Subject: Mechanical Engineering thesis Ph. D
Dissertations, Academic -- Mechanical Engineering -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis (Ph. D.)--University of Florida, 1992.
Bibliography: Includes bibliographical references (leaves 227-238).
Additional Physical Form: Also available on World Wide Web
General Note: Typescript.
General Note: Vita.
Statement of Responsibility: by John Charles Wagner.
 Record Information
Bibliographic ID: UF00097386
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 001806289
oclc - 27804168
notis - AJN0122


This item has the following downloads:

PDF ( 9 MBs ) ( PDF )

Table of Contents
    Title Page
        Page i
        Page i-a
        Page ii
        Page iii
    Table of Contents
        Page iv
        Page v
    List of Tables
        Page vi
        Page vii
    List of Figures
        Page viii
        Page ix
        Page x
        Page xi
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
    Incinerator system
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
    Review of data from three years of testing
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
    Kinetic modelling
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
    Effect of chlorinated input on chlorinated organic compound emissions
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
    The new combustion control system
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
    Experiments with old and new control systems
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
    Discussion, conclusions, and recommendations
        Page 159
        Page 160
        Page 161
        Page 162
        Page 163
        Page 164
        Page 165
        Page 166
        Page 167
        Page 168
        Page 169
        Page 170
        Page 171
        Page 172
        Page 173
    Appendix A: Detailed description of the original control system
        Page 174
        Page 175
        Page 176
        Page 177
        Page 178
        Page 179
        Page 180
        Page 181
        Page 182
        Page 183
    Appendix B: Details on the instrumentation
        Page 184
        Page 185
    Appendix C: Data calculations
        Page 186
        Page 187
        Page 188
        Page 189
        Page 190
    Appendix D: Vost emissions
        Page 191
        Page 192
    Appendix E: Kinetics code program listing
        Page 193
        Page 194
        Page 195
        Page 196
        Page 197
    Appendix F: Kinetics code input file
        Page 198
        Page 199
        Page 200
    Appendix G: Temperature-time profile calculations
        Page 201
        Page 202
        Page 203
    Appendix H: Base case kinetics code output
        Page 204
        Page 205
        Page 206
        Page 207
        Page 208
    Appendix I: Additional details on the control system
        Page 209
        Page 210
        Page 211
        Page 212
        Page 213
        Page 214
        Page 215
        Page 216
        Page 217
        Page 218
        Page 219
        Page 220
        Page 221
        Page 222
        Page 223
        Page 224
    Appendix J: Additional details on the control computer program
        Page 225
        Page 226
        Page 227
        Page 228
        Page 229
        Page 230
        Page 231
        Page 232
        Page 233
        Page 234
        Page 235
        Page 236
        Page 237
        Page 238
    Biographical sketch
        Page 239
        Page 240
        Page 241
        Page 242
Full Text








To my parents


The author wishes to thank Donald Clauson, Joseph Blake,

Xie-Qi Ma, David Roskein, Britta Schmidt, and Jason Weaver

for stack sampling and analysis; Henk van Ravenswaay for

construction of the biomass feeder; Kaifa Awuma for HC1

analysis; Jamie Forester for drawing the original incinerator

circuitry; Mahadevan Sadanandan for kinetics code work;

Michael Mahon for construction of the HC1 continuous emission

monitor; Scott Quarmby and Robert Driskell for their

electrical engineering expertise; Donald Adams for setting up

the continuous emissions monitor sampling system; Jonathan

Carter, Chris Jeselson, Todd Yurchisin, Ronald Storer,

William Calhoun, Thomas Cherry, Song Mu, Wilfred Schnell,

Rodnie Barbosa, and others who helped in some way, even if it

was just feeding garbage; Craig Saltiel for getting the

kinetics code operational again; Charles Schmidt for VOST

analysis; Bruce Green for refurbishing and maintaining the

incinerator and helping install the control system; the

Tennessee Valley Authority, Florida Governor's Energy Office,

New River Waste Association, Florida Department of

Environmental Regulation, Mick A. Naulin Foundation, Gatorade

Fund, Supelco, Tacachale, and University of Florida for

project funding; and especially Alex Green, whose advice,

encouragement, and support made this all possible.



ACKNOWLEDGMENTS ........................................... iii

LIST OF TABLES ............................................ vi

LIST OF FIGURES .......................................... viii

ABSTRACT .................................................. X


1 INTRODUCTION ........................................ 1

Combustion Fundamentals ............................ 3
State of the Art in Incineration .................... 12
Some Combustion Strategies .......................... 15
The Need For a New Combustion Control System ....... 19
Development of a New Combustion Control Strategy .... 20

2 INCINERATOR SYSTEM .................................. 24

Background .......................................... 24
Incinerator Hardware ................................ 26
Incinerator Operation................................ 30
Original Control System ............................. 31

3 INSTRUMENTATION ..................................... 34

Sensors, Monitors, and Sampling Systems ............. 34
Operations .......................................... 40


CCTL Trial Burn Data ...................... ......... 44
Carbon Monoxide Data ................................ 47
Particulate Data .................................... 51
Volatile Organic Compound Data ...................... 53
Summary of Significant Relations .................... 57

5 KINETIC MODELLING ................................... 58

Inputs to the Code ................................. 64
Waste Composition .................................. 65
Pyrolysis Steps ..................................... 67
Results from the Kinetics Code ...................... 68
Conclusions from the Kinetics Code .................. 75


COMPOUND EMISSIONS ................................ 77

PVC, HC1, and VOST Data from CCTL ................... 79
Review of Statistical Procedures ................... 82
Effect of Chlorine Input on HC1 Emissions ........... 87
Review of HC1 and VOST Data ......................... 89
Conclusions from the Chlorinated Data Review ........ 118

7 THE NEW COMBUSTION CONTROL SYSTEM ................... 120

The Combustion Control Strategy .....................120
Implementation of the New Combustion Control
Strategy ...........................................121
Control System Components ..........................123
The Control System Program ..........................131


Test Firings .........................................143
Experimental Firings ................................148


Specific Results .................................... 160
Recommendations for Future Work .....................167
General Projections to Other Systems ................171
Final Summary .......................................... 173



B DETAILS ON THE INSTRUMENTATION .....................184

C DATA CALCULATIONS ...................................186

D VOST EMISSIONS ...................................... 191

E KINETICS CODE PROGRAM LISTING ......................193

F KINETICS CODE INPUT FILE ............................198


H BASE CASE KINETICS CODE OUTPUT .....................204



REFERENCES .................................................227

BIOGRAPHICAL SKETCH .......... ..............................239


3-1 Typical Protocol for Experimental Burns ............. 42

4-1 Results of Measurements Taken During NHW-Only
Experimental Burns from 08-31-89 to 08-10-91 ...... 46

4-2 Results of Method 5 Particulate Measurements Taken
During NHW-Only Experimental Burns from 09-27-88
to 08-03-89 ....................................... 46

4-3 Linear Regression Results from CO Data Review ....... 49

4-4 Linear Regression Results from Particulate Data
Review ............................................ 51

4-5 Linear Regression Results from VOST Data Review ..... 55

5-1 Cottage Waste Characterization of 02-24-90 .......... 66

5-2 Kinetic Code Results ................................ 72

6-1 CCTL PVC and HC1 Test Data .......................... 80

6-2 Hydrogen Chloride, Temperature, Carbon Monoxide, and
Volatile Organic Compound Data Used for Linear
Regression and Nonlinear Analyses ................. 81

6-3 Critical Values of t, F, and R2 at the 5%
Significance Level ....................... ......... 85

6-4 Results of Linear Regression of VOST Data ........... 90

6-5 Results of Quadratic and Higher Power Regressions of
CCTL VOST data .................................... 96

6-6 Results of Logarithmic and Exponential Regressions
of CCTL VOST data ................................. 98

6-7 Results of Fits to Linearized Versions of Nonlinear
Functions ............................................ 101

6-8 Parameters for Nonlinear Equation Fits .............. 105

6-9 Results of Cross-Product Regression of CCTL VOST
Data ............................................. 113

6-10 Results of Regressions with Other Compounds ......... 117

8-1 Protocol for Control System Experimental Burns ...... 150

8-2 VOST Results for Latest Samples, Non-Blank Corrected
and Blank Corrected ............................... 156



1-1 Effect of theoretical air on temperature for burning
a typical 4500 BTU/lb waste ...................... 17

1-2 Primary chamber temperature for the 03-21-91 CCTL
burn .............................................. 21

1-3 Carbon monoxide levels during 11-23-91 burn ......... 21

2-1 Incinerator system layout .......................... 27

2-2 Functional schematic of original CCTL incinerator
control circuitry ................................. 32

3-1 Incinerator instrumentation layout ................. 35

4-1 Quadratic fit of CCTL carbon monoxide data to
primary combustion chamber temperature ............ 50

4-2 Carbon monoxide data fitted with a quadratic
function of PCC temperature and a linear function
of oxygen concentration .......................... 50

4-3 Carbon monoxide data fitted with a quadratic
function of PCC temperature and a linear function
of carbon dioxide concentration .................. 50

4-4 Particulate emissions compared to oxygen
concentration ..................................... 52

4-5 Particulate emissions data compared to natural gas
input ............................................. 52

4-6 Vost emissions compared to waste feed rate .......... 56

4-7 Vost emissions compared to PCC temperature and waste
feed rate and plotted against temperature ......... 56

5-1 Kinetics code program flowchart .................... 61

5-2 Temperature-time history for CCTL incinerator ....... 63

5-3 Kinetic aromatic and chlorination sequences ......... 69

5-4 Base case (Run 30) kinetics code output ............. 71


6-1 HC1 emissions compared to added chlorine input with
both HC1 and added Cl normalized to a 400-lb/hr
waste feed rate ................................... 88

6-2 Comparison of DCB emissions to HC1 emissions with
and without outlying DCB data point ............... 94

6-3 Replacement of the orthogonal function H(t;p,p) with
a Gaussian function J(t;u,s) ...................... 109

7-1 On/off and alarm control system circuitry ........... 125

7-2 Natural gas piping arrangements ..................... 127

7-3 Natural gas flow rate measurement orifice meter ..... 129

7-4 Control program overall flowchart ................... 133

7-5 Manual data entry screen ............................ 135

7-6 Control computer program data screen ................ 135

7-7 Incinerator warm up logic ........................... 138

7-8 Temperature maintaining logic ....................... 139

7-9 Ram feeding logic ................................... 142

8-1 CEM data from 04-08-92 burn ......................... 154

8-2 Temperature and gas data from 04-08-92 burn ......... 154

8-3 CEM data from 04-15-92 burn ......................... 157

8-4 Temperature and gas data from 04-15-92 burn ......... 157

A-i Original incinerator control system circuitry ....... 175

A-2 Original biomass control subsystem .................. 180

A-3 Original stoker control circuitry ................... 181

I-1 Power supply board with transformer ................. 214

I-2 Original modmotor circuitry ......................... 216

I-3 Modmotor modification and control circuitry ......... 218

I-4 Override relays configuration ....................... 221

I-5 Stoker control circuitry with override relay ........ 222

Abstract of Dissertation Presented to the Graduate School of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy




August 1992

Chairman: Dr. Alex E. S. Green
Major Department: Mechanical Engineering

This dissertation summarizes the results of stack

emission measurements made since 1988 using a two-stage 500

lb/hr modular incinerator located at the University of

Florida-Tacachale-Clean Combustion Technology Laboratory (UF-

T-CCTL). From the analysis of these results some

statistically significant relationships have been developed

between HC1 emissions (a surrogate for PVC in the waste) and

the emissions of a number of chlorinated organic compounds

(C6H5Cl, C6H4C12, CHC13, CH2C12, and C2C14). These compounds

appear in the stack gases even under good combustion

conditions. These type of relationships have not been

previously reported in combustion literature and have

important implications for the evolving discipline of Clean

Combustion Technology. These results contradict widely

quoted and widely accepted conclusions from other incinerator


From the analysis of the stack emissions of the UF-T-

CCTL incinerator a control strategy is developed that will

minimize emissions of carbon monoxide and volatile organic

compounds. A personal computer-based control system is

developed to implement the combustion control strategy. The

control system monitors carbon monoxide, carbon dioxide,

oxygen, and hydrogen chloride emissions and incinerator and

stack temperatures; maintains feed rate and output power

levels; and fixes temperature levels in such a way as to

operate the incinerator under desirable combustion conditions

for emissions minimization.

The control system's strategy is based upon a review of

previous combustion conditions and emissions data collected

from the UF-T-CCTL incinerator, kinetic modelling of

emissions, regulatory constraints, incinerator size and

thermal characteristics, and desired feed rate and output

power levels. The control system regulates the flow rates of

natural gas into the two combustion chambers, primary chamber

overfire and secondary chamber air flow, and waste charging


Experiments at an average 350-lb/hr waste feed rate with

the control system installed and operating show a significant

reduction in carbon monoxide emissions over those from before

the installation of the control system. Also, the

temperature versus time profile during incinerator operation

has been considerably smoothed out.


Modular-type incinerators have been used for many years

for on-site medical and institutional waste disposal. These

shop-built units usually have two combustion chambers. In

the first chamber the waste is burned under substoichiometric

conditions, i.e., with insufficient air for complete

combustion. In the second chamber additional air and support

fuel are introduced to burn out the combustion gases from the

first chamber. This starved-air, two-stage method was

developed to reduce the smoke generated from earlier one-

stage units.

Increasing regulation and public awareness of

incinerator emissions, particularly organic compounds, have

brought incinerators under great scrutiny. The current

method of operating modular incinerators in a starved-air

mode is inadequate for controlling organic and acid

emissions. The regulators' response has largely been to

require post-combustion air pollution control devices. These

devices can be many times the cost of the incinerator unit

itself. A scrubber to control acid gas emissions could add

$200,000-$800,000 to the the cost of an $100,000 modular


The objective of this research is to use the extensive

experience of the Clean Combustion Technology Laboratory

(CCTL) at the University of Florida in modular waste

incineration to determine relations between emissions and

operating conditions useful for control strategies to

minimize emissions from modular incinerators. These

strategies can be used as an alternate method of pollution

control to requiring expensive post-combustion air pollution

control devices by serving as a guide for upgrading,

operating, and designing this general class of incinerators.

Another objective of the research is to develop a combustion

control system to implement some of these strategies and

minimize emissions of carbon monoxide and volatile organic

compounds from a two-stage modular incinerator.

The CCTL has operated a two-stage modular incinerator

(see Chapters 2 and 3) at the Tacachale center in

Gainesville, FL, for over 4 years. This institutional

incinerator has been modified to burn hotter and run in an

excess-air mode. Careful attention is given to what is

burned in the CCTL unit. Metallic and chlorinated materials

are mostly avoided. These low-cost combustion and pre-

combustion strategies of burning hotter, operating in an

excess-air mode, and avoiding toxic producing materials in

the input waste stream should alleviate the need for

expensive post-combustion control devices.

Data collected from three years of stack testing at the

CCTL facility is reviewed here to determine desirable

operating conditions for minimal emissions for this type of

modified incinerator (see Chapter 4). The combustion process

in the incinerator is examined using simple chemical kinetic

models (see Chapter 5). Emissions of chlorinated organic

compounds are analyzed to determine if relationships exist

between the level of these emissions and the level of

chlorine in the input (see Chapter 6). Some widely quoted

and accepted studies [1-2] have suggested that no such

relationships exist. Studies by Green et al. [3-8]

contradict this "conventional wisdom".

A simple combustion control system is developed to

regulate the operation of the CCTL incinerator (see Chapter

7). The control strategy is based on the desirable operating

conditions determined from the data review and kinetic

modelling. Tests with the control system show lower

incinerator emissions than without the control system (see

Chapter 8).

Some combustion fundamentals and emissions control

strategies are presented in this chapter.

Combustion Fundamentals

Pyrolysis is the process of using high temperatures in

an atmosphere of little to no oxygen to drive off, but not

combust, lighter organic compounds from a more heavy or

complex organic compound or mixture of such compounds.

Combustion is the chemical reaction of a fuel with air

(or oxygen). The reaction is usually exothermic (releases

heat energy), which increases the temperature of the product

gases substantially. Fuels normally contain carbon and

hydrogen, which if combusted completely form carbon dioxide

(CO2) and water (H20).

Incineration is the combustion of waste. Waste and most

other fuels also contain oxygen, chlorine, sulfur, nitrogen,

metals, and ash incombustiblee matter such as minerals and

other metallic oxides). Chlorine and sulfur in the fuel

usually generate hydrogen chloride (HC1) and sulfur dioxide

(SO2) as combustion products. Nitrogen in the fuel, and

nitrogen in the air at high combustion temperatures, may

generate nitrogen monoxide (NO) and nitrogen dioxide (NO2).

Mixtures of NO and NO2 are referred to as NO Metals

usually oxidize and remain with ash residue. Some of this

residue may be drawn up with the combustion gases as

particulates. Volatile (low boiling point) metals such as

mercury, cadmium, and lead may vaporize and leave with the

combustion gases as gases or on particulates. Organic

compounds in the waste also can escape the combustion process

and directly become stack gas emissions.

Products of incomplete combustion (PICs) form when there

is insufficient oxygen (a substoichiometric amount) and/or

insufficient mixing and time to fully burnout the carbon in

the fuel. Very substoichiometric combustion can lead to

sooting conditions. The principal PIC is carbon monoxide

(CO). The PICs may be gaseous, volatile, or semi-volatile.

Gaseous compounds include CO and light alkanes. Volatile

organic compounds (VOCs) are those with boiling points below

100 OC and/or vapor pressures greater that 10-1 mm Hg.

Sampling methods for VOCs usually can measure those with

boiling points up to 130 OC. The VOCs include light

aromatics and chlorinated alkanes and alkenes. Semi-volatile

organic compounds (SVOCs) are those with boiling points above

100 OC and/or vapor pressures between 10-7 mm Hg and 10-1 mm

Hg. The SVOCs may condense into particulates or coagulate

onto other particulates. When chlorine is present in the

fuel, chlorinated PICs are also formed. Chlorinated alkanes

are usually volatile, while most chlorinated aromatics are

semi-volatile. The most toxic PICs are probably the

polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated

dibenzofurans (PCDF). These are SVOCs with the most toxic

being 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD) [9].

Problems with Chlorine in Combustion

The destruction of organic compounds involves free

radicals such as H, O, and OH. When the carbon and hydrogen

atoms in the organic are oxidized, heat energy is liberated,

increasing the temperature of the combustion gases. Higher

temperatures lead to more free radicals and faster

destruction of the organic. The conversion from CO to CO2

is usually the step that limits the overall speed of the

combustion. Chlorine acts as a flame inhibitor and retards

combustion reactions by removing hydrogen atoms from the

combustion gases [10-12]:

H + HC1 --> H2 + Cl. (1-1)

This has the effect of limiting the number of hydroxyl (OH)

radicals and oxygen atoms, which are generated by the major

chain branching reaction:

02 + H --> OH + 0. (1-2)

With less OH available the chemical reaction processes in

combustion are slowed [10-11,13), which leads to lower

temperatures and less organic destruction.

The Deacon reaction [14-16]

2 HC1 + 02 ----> C12 + H20 (1-3)

produces free chlorine from HC1, especially in the presence

of CuCl2, which is found on particulates. The free chlorine

is then available to chlorinate organic compounds found in

the stack gas and on the particulates. To reduce this

effect, HC1 and particle emissions should be minimized, and

02 levels should be decreased. Decreasing 02 levels has the

side benefit of reducing the amount of air used.

Residence Time Versus Temperature

Rules for incinerator operation, such as 1800 F for 1

second in the last combustion chamber, are written for

regulatory purposes. These rules seem somewhat arbitrary

since, for good toxic destruction, there is much more

flexibility due to the chemical and physical processes in

combustion. Whereas a 1-second residence time may be needed

for four 9's (99.99%) destruction of an organic compound at

1800 OF, much less time is needed at 2000 F.

Data for the time required for four 9's reduction in

2,3,7,8-TCDD at temperature in the range 1340-2240 F [17]

can be modelled as

1800 F T
t = exp( 77 F ) sec (1-4)

where t is time in seconds and T is temperature in OF. This

equation can be solved for the temperature needed for a

specific residence time:

T = 1800 OF 77 F ln(-) (1-5)

At 1800 OF, 1 second is required for four 9's destruction.

At 1900 OF, only 0.27 seconds are needed. For a half-second

residence time the temperature need only be raised to 1853


A time versus temperature regulatory rule would allow

older hospital incinerators with small afterburners to

continue to operate, provided the last combustion chamber's

minimum temperature meets this rule.

For a fixed mass flow rate through a chamber, assuming

the gas mixture can be modelled as an ideal gas, the

residence time in the chamber is proportional to the

reciprocal of the absolute temperature:

t = V p M (1-6)
Q Ru Ta Tbs T + 460
u abs abs

where Q is the volumetric flow rate, V is the volume of the

chamber, P is the absolute pressure in the chamber, M is the

gas mixture's average molecular weight, m is the mass flow

rate through the chamber, Ru is the universal gas law

constant, Tabs is the absolute temperature (T + 460 for

U.S.C.S. or T + 273 for S.I.), and K is a constant defined by

Equation 1-6 (P and M are taken as constants). Equations 1-5

and 1-6 can be combined and solved iteratively for the

minimal operating temperature for four 9's destruction:

T = 1800 77 n ( 460 (1-7)
min T + 460

For example, choosing a starting value for Tmin of 1800 in
Equation 1-7, will converge to a solution for T in a

couple of iterations.

Emissions Control

Particulate emissions control can be achieved with dry

scrubbers--baghouses, electrostatic precipitators, and

cyclones. Baghouses use fabric filters and capture

particulate matter much like a vacuum cleaner. Electrostatic

precipitators capture particulates by charging the particles

as they pass by one series of plates and attracting them to

another series of oppositely charged plates. Cyclones use

centripetal forces from swirling the exhaust gases to remove

the particles from the gas stream.

Acid gas and organic compound emissions control can be

achieved with dry scrubbers--spray dryers--and wet scrubbers

--spray towers, packed towers, and venturi-type scrubbers.

In dry scrubbers a powder such as lime is sprayed into the

exhaust gases to neutralize the acid gases. In wet

scrubbers, the spray in a solution of the neutralizing

compound in water. Water spray is used to cool the

combustion gases before they enter certain air pollution

control devices.

Organic emissions control also can be achieved through

controlling the combustion process. Combustion temperature,

residence time, turbulence, and oxygen level are the

parameters of interests here. Higher levels of each of these

four parameters generally leads to more complete combustion

and reduced emissions.

Types of Incinerators

There are three types of incinerators used in municipal

waste combustion: mass burn facilities, modular combustors,

and refuse-derived fuel (RDF) fired units [18]. The most

common incinerator for medical waste incineration is the

modular combustor. These may be broken down into two

classes: starved- (or controlled-) air types and excess-air

types. Both types usually have a primary combustion chamber

(PCC) and a secondary combustion chamber (SCC) or

afterburner. In a starved-air unit, the waste is burned in

the PCC with a substoichiometric amount of air generating

many PICs and pyrolysis products. Combustion gases from the

PCC pass into the SCC where more air is added to finish

burning out the PICs and pyrolysis products generated in the

PCC. Typical PCC and SCC operating temperatures are 1400 F

and 1700 OF, respectively, in a starved-air unit. In an

excess-air unit, the waste is burned in the PCC with more air

than is needed for combustion. No additional air is added to

the SCC in this case. In either type, natural gas or oil

burners may be added to the PCC and SCC to aid in igniting

and burning the waste and in bringing the temperatures in the

chambers up to levels needed to sustain combustion.

Another type of incinerator, the rotary kiln, is used

mainly in hazardous waste incineration. The PCC of an rotary

kiln incinerator slowly rotates about its long axis to

thoroughly burn out the waste and residue inside. Fluidized

bed combustion for waste disposal is also becoming more

popular [19].

Advantages of Starved-Air Incineration

Starved-air units have a few advantages over excess-air

units. One advantage is that, with the lower temperature and

turbulence, less particulates and metals are carried into the

SCC and up the stack. Another advantage is that waste takes

longer to burn, which distributes out over time the heat

energy released from batch-fed waste. This allows for a

steadier flow of steam from those units with heat recovery


Problems with Starved-Air Incineration

Because starved-air units operate with insufficient air

in the PCC for complete combustion, the SCC must burn out the

large quantity of PICs being generated in the PCC. A

tertiary combustion chamber may be required to sufficiently

destroy the PICs. Starved-air units cannot handle high

moisture content waste [20], such as in hospital red bags,

since there is insufficient heat release to sustain

combustion with substoichiometric burning. Starved-air units

are usually designed for paper wastes (trash), and have small

PCC and SCC burners for the low supplemental fuel

requirements of paper wastes. Starved-air units are least

able to adapt to changes in the waste constituents [20].

The destruction of organic at high temperature is a

pyrolytic and oxidative process [13]. In a nonoxidative

(starved-air) environment the pyrolysis process leads to a

breakdown from large organic molecules to smaller and more

unsaturated species. At some point these become highly

carbonized structures and soot. In an oxidative (excess-air)

environment carbon atoms become bonded to oxygen in CO and

CO2 and thus removed from the sooting process. Molecules

that are not fully oxidized have a greater chance of being

chlorinated in a starved-air environment [13].

Advantages of Excess-Air Combustion in Primary Chamber

Burning overstoichiometric in PCC will greatly reduce

the amount of PICs passing into the SCC. In this case, when

air is added to the SCC, the SCC acts like a tertiary

combustion chamber in a starved-air unit.

The excess-air mode of operation more completely

combusts the waste in the PCC, though it may increase

particulate emissions. The SCC and hotter PCC of excess-air

incinerator allows for longer high temperature residence time

for toxic destruction than just the SCC of a starved-air


Problems With Burning Too Hot

Burning above 1800 OF in the PCC can lead to ash

slagging problems [20]. At 1880 OF in a reducing (oxygen-

starved) atmosphere ash will start to deform. When the ash

moves to a cooler or oxygen rich section of the furnace, it

will harden into a slag (or clinker). This can have

detrimental effects on refractory lined walls and can clog

air ports and burners. However, in an oxidizing (excess-air)

environment ash may not deform until the temperature reaches

2030 OF.

With high PCC temperatures some heavy metals can

volatilize and be carried into the combustion gas stream. If

the SCC temperature is too high, excessive NO emissions can

State of the Art in Incineration

Control systems for early modular incinerators were

primarily concerned with safety and the operation of the

incinerator [21-22], without regard to optimizing the

combustion process. Even into the mid-1980s, systems used

on/off control of support-fuel burners to regulate the

combustion temperature. These control systems slowly evolved

from using electro-mechanical timers and relays to using

programmable logic controllers [23-24]. Temperature-based

control of the waste charging frequency [25] and the

combustion air flow [26-28] followed. Eventually, computers

were used to monitor and control the incinerator operations

[29], especially when air pollution control devices are

installed. These new control systems are still mainly

concerned with regulating feed rate and operating

temperatures. Air pollution control devices have been (and

still are) added only when required by regulations.

The Basic Environmental Engineering, Inc. incinerator

[30] uses a four stage unit with Basic's patented pulsed

hearth" primary stage, excess air in the second and third

stages, reburn tunnels, and some recirculation from the

outlet of the heat recovery boiler back to the fourth stage.

Emissions of hydrocarbons and CO are typically less than 35

ppm, which is well below the 100 ppm CO level of typical

state regulations.

State-of-the-art mass-burn [31-32] and large refuse-

derived fuel (RDF) [33-34] incinerators incorporate

travelling grate stokers, air preheaters, zone control of

combustion air, dry gas scrubbers to neutralize acid gases,

and fabric filters to capture particulates. Dry gas

scrubbers and fabric filters have replaced electrostatic

precipitators as the desired post-combustion pollution

control devices. Electric power generation is usually

included. Continuous monitoring and regulation of combustion

parameters is used to optimize combustion efficiency and

control the formation of organic compounds. Homogenization

of the waste by careful mixing is used to maintain a

consistent quality of waste and steadier incinerator


Incineration is the best way to deal with hazardous

wastes [35]. Some hazardous waste incinerators use oxygen to

supplement or replace the combustion air to achieve higher

temperatures and avoid NOx [36]. Medical waste is sometimes

labelled bio-hazardous waste by regulatory agencies [37-38].

Incineration is the most widely used technology for the

treatment of medical waste [39]. For medical waste

incineration, the state of the art is a starved-air multi-

stage modular unit. Most medical waste incinerators do not

have air pollution control devices [39]. Most of those that

do use wet scrubbers, which do not achieve a high degree of

particulate control.

Several "State-of-the-Art-Reviews" have been written

recently on medical waste incineration [40-43]. The

recommendation of these reviews is to incorporate post-

combustion pollution control devices as the principal means

of minimizing toxic emissions.

Shands Hospital at the University of Florida,

Gainesville, FL was going to replace its old starved-air

modular incinerator with a $3.6 million, state-of-the-art,

starved-air modular incinerator with post-combustion air

pollution control devices, but is now reviewing other options

for waste disposal.

Some Combustion Strategies

The U.S. Environmental Protection Agency (EPA) lists

"Good Combustion Practices" (GCPs) for minimal organic

emissions from mass burn, RDF, and starved-air incinerators

[18]. For a starved-air incinerator, the GCPs include an

1800-F secondary chamber, 6-12 % oxygen in the flue gas,

less than 50 ppm CO, and use of auxiliary fuel during warm up

and prolonged high levels of CO. Recently, GCPs for modular

excess-air incinerators have been added [44] These new

rules, which also apply to modular starved-air incinerators,

include operating with less than 50 ppm CO (4-hour average)

and not exceeding 110% of the average load (feed rate) used

during the most recent PCDD/PCDF compliance test.

Regulations in Australasia (Australia, New Zealand, and

nearby islands) require the afterburner to maintain a minimum

of 1100 OC (2012 OF) for at least 1 second at 6-10% excess

oxygen under turbulent conditions [45-47]. Danish

regulations require the afterburner to be at least 850 C

(1562 OF) for a minimum of 2 seconds at 6% oxygen at a

Reynolds number of 60,000 or more [48]. Bulley [45,47] also

recommends avoiding the use of PVC and heavy metal pigments

in medical plastics.

Green [3-5] advocates the 3 laws of Clean Combustion

Technology--pre-combustion, combustion, and post-combustion

measures to minimize emissions. The pre-combustion measures

include avoiding or removing toxic materials and toxic

producing materials from the input waste stream. The

combustion measures include optimizing temperature and CO

levels, using stokers, and coburning with natural gas. The

post-combustion measures include hot gas clean up and use of

scrubbers only as needed. Green's modelling [7-8] of the

data from the Pittsfield-Vicon municipal waste incinerator

emission test study [1] shows the benefit of lowering

chlorinated inputs on incinerator emissions.

Wagner et al. [49] studied the emissions from the CCTL

incinerator and found that for minimum CO and organic

emissions when feeding institutional waste at a typical rate

of 400 lb/hr in a modular incinerator the PCC temperature

should be maintained in the range 1750 OF to 1800 oF, the SCC

temperature should be maintained in the range 1825 F to 1875

F. This incinerator was a 500 lb/hr excess-air unit with

additional air added to the SCC.

Hasselriis [50-52] also analyzed data from the

Pittsfield-Vicon municipal waste incinerator emission test

study [1]. Hasselriis found minimum CO and total PCDD and

PCDF emissions for a furnace temperature range of 1600-1700

OF, and minimum CO at 8% oxygen. Minimum total PCDD and PCDF

emissions were found at 9% oxygen and 15 ppm CO. The

Pittsfield-Vicon incinerator [1] uses exhaust gas


recirculation to control the temperature of the gases

entering the heat recovery boilers.

Hasselriis [53] shows the temperature and oxygen levels

at which to operate a starved-air incinerator on a graph such

as that in Figure 1-1. The PCC would be operated at 1500 F

and 70% theoretical air, while the SCC would be operated at

1800 oF and 160% theoretical air. Hasselriis's approach is

to reduce the air to the PCC when the overall air to the

incinerator is insufficient to keep CO and other emissions

down. This approach makes the PCC even more starved, which

reduces the rate that the waste is being pyrolyzed or

gasified and sent to the SCC. With fewer waste gases

entering the SCC, the SCC can more completely combust them.

Temperature (OF)
2500 I I I






50% 100% 150% 200%

Theoretical Air

Figure 1-1. Effect of theoretical air on temperature for
burning a typical 4500 BTU/lb waste (adapted
from Hasselriis [53]).


Chung and Tsang [54] find that soot production from the

combustion of polystyrene could be reduced by adding air to

the pyrolysis zone of the flame, improving mixing, and using

metal salts.

Nihart et al. [55] found that higher destruction

efficiency generally correlated with low levels of CO, total

hydrocarbons, and methane emissions. In a study on

correlating incinerator operating parameters with

performance, Stately et al. [56] found that principal organic

hazardous constituents (POHCs) and products of incomplete

combustion (PICs) generally increased with higher levels of

CO, but the relationship depended on the compound studied.

Weldon and Beacher [57] find than CO, combustion temperature,

and residence time can be used as surrogates to monitor

dioxin emissions from municipal waste combustors.

With sufficient mixing and oxygen it is generally

believed [42] that 1800 F is sufficiently high [58-59] to

destroy 99.99% of most thermally stable compounds including

chlorinated benzenes, phenols, dioxins, and furans. Carbon

monoxide is a good index of combustion efficiency [58,60].

Most authorities agree than operation at less than 100 ppm CO

minimizes organic emissions [61-62]. Lee [63] also advocates

the 3-T (time, temperature, and turbulence (and oxygen))

approach to good combustion.

Acharya et al. [16] expects that minimal formation of

dioxins will occur in hazardous waste incinerators with an

oxidative, high-temperature, turbulent SCC with a 2-second

residence time if the combustion gases quench in air directly

without being sent through a heat recovery boiler.

The Need For a New Combustion Control System

The 1990 Clean Air Act Amendments [64] mandate the

eventual measurement of and limitations for the emissions of

up to 189 pollutants in 250 source categories including

incinerators. New Florida Department of Environmental

Regulation and EPA regulations [37-38,65] for biohazardous

and municipal incinerators have minimum temperature and

residence time levels requirements; emission limits for

particulates, CO, and HC1; and 02 and temperature monitoring

requirements. For biohazardous waste incinerators under 500-

lb/hr capacity in Florida the new regulations require that

the final chamber's temperature be maintained above 1800 F

with a residence time of at least 1 second, particulate

emissions be less than 0.10 gr/dscf (grains per dry standard

cubic foot) corrected to 7% 02, CO emissions be less than 100

ppm by volume dry basis corrected to 7% 02, and HC1 emissions

be less than 4 lb/hr.

Improving systems to automatically monitor and control

combustion parameters is one of the research needs for

municipal solid waste seen by the Solar Energy Research

Institute [66].

The current control system of the CCTL incinerator is

completely manually set. Blowers, feed systems, and stokers

have to be turned on by hand when needed. The burner

controllers do not reset themselves after a failed ignition

or a flame out. The control system does not maintain steady

or even minimum temperature and oxygen levels. The system

provides no alarms for burner outage, stoker breakdown, empty

feed hooper, etc. The on/off control of the burners causes

rapid, high-amplitude cycling of the primary combustion

chamber (PCC) and secondary combustion chamber (SCC)

temperatures. The system uses no feedback except, for PCC

and SCC temperature, to regulate the combustion process.

Figure 1-2 illustrates some of these problems. The

high-frequency, high-amplitude cycling of the PCC temperature

and three burner outages (at 14:55, 15:45, and 16:05) are

very evident. The PCC temperature reached only 1230 F

before garbage was fed in at 13:00, and did not reach 1800 F

until the fourth feed. Burns of July 21, 1990, and August

10, 1991, also had similar problems. The temperature

controller and stoker also broke down during the August 10,

1991, burn. Figure 1-3 shows the carbon monoxide spikes that

often occur following ram-feed charges into the incinerator.

If a control system could reduce the amount of support fuel

before waste is fed into the incinerator, these spikes could

be minimized.

Development of a New Combustion Control Strategy

A versatile control system should be able to control

emissions of CO, NOx, SO2' HC1, volatile and semi-volatile

organic compounds (including dioxins), heavy metals, and

Temperature (F)










Time of day (hours)

Figure 1-2.

Primary chamber temperature
CCTL burn.

for the 03-21-91

CO (ppmv, dry)



Time of Day (hours:minutes)

Carbon monoxide levels during 11-23-91 burn.












- ~Ni A---- I-

Figure 1-3.

particulates. It also should maintain temperature, 02, and

minimum feed rate and output power levels. Not all emissions

can be minimized by controlling the combustion conditions.

The waste stream itself should be monitored to avoid

halogenated materials, metals, toxic producing materials,

non-combustibles, and excessive liquids. The stack gas may

still need to be scrubbed for acid gases and filtered for

particulates even under ideal combustion conditions.

One objective of this research is to develop a

combustion control system that minimizes emissions of carbon

monoxide and volatile organic compounds from a two-stage

modular incinerator while maximizing the feed rate of

garbage. Development of a new combustion control strategy

should incorporate an understanding of the configuration and

operation of the incinerator (Chapter 2), examining what

instrumentation is available (Chapter 3), reviewing previous

combustion conditions and emissions (Chapter 4), modelling of

emissions (theoretical and/or empirical) and reaction

mechanisms (Chapter 5), and reviewing regulatory constraints

(Chapter 1), desired feed rate, and/or output power levels.

Also other combustion control strategies should be reviewed

(Chapter 1).

Certain surrogate pollutants that have simple relations

to most other pollutants will be identified from the data

review and modelling. Those pollutants that are representa-

tive of the effectiveness of the overall combustion process

and are able to be monitored will serve as a basis for the

feedback control system. Carbon monoxide has been tradition-

ally used as an indicator of combustion conditions. Analysis

of the data review and modelling will show preferred

operating and combustion conditions for various inputs.

These conditions will include parameters such as PCC and SCC

temperatures, oxygen, and feed rate.

Parts of a modular incinerator system that can be

controlled include the flow rate of natural gas (or other

supplemental/ignition fuel) into the two chambers, flow rate

of air supplied to the burners and chambers, waste and/or

other solid fuel charging frequency, stoking frequency, and

ash removal rate.



The Interdisciplinary Center for Aeronomy and (other)

Atmospheric Sciences (ICAAS) has been engaged in a program of

Research and Development on co-combustion since 1980. Early

work focused on the co-combustion of pulverized coal and

natural gas at a level of 1 MMBTU/hr [67-71]. In 1985 a 535-

hp, 20 MMBTU/hr boiler at the steam plant at Tacachale (then

called Sunland Training Center (STC)) was made available for

use by ICAAS. Tacachale is a state-funded center for

training the developmentally and intellectually impaired.

About 800 residents live at 48 cottages, which are organized

into 8 facilities, at the center. The 1954-vintage boiler

was originally oil-fired and was converted to use natural gas

in 1965. Experiments with this boiler involved coburning

natural gas with a slurry of pulverized coal in water or oil.

Early experiments focused on high levels of coal-water slurry

(25% to 60% by power input) to eliminate the boiler derating

(decrease in output power) caused by the conversion to

natural gas, and to replace oil by a cheap domestic fuel

(coal) [72-78], while later experiments focused on low levels

of coal-water and coal-oil slurries (5% to 20%) to enhance

the radiation of the natural gas flame [79]. The radiation

enhancement experiments at this industrial level, and with a

laboratory-scale setup led to a masters thesis [80] and two

patents [81-82].

In 1987 a 500 lb/hr Environmental Control Products (ECP)

incinerator was donated to ICAAS. The Clean Combustion

Technology Laboratory (CCTL) was set up in front of the STC

steam plant with the installation and refurbishing of the

incinerator. The incinerator was made operational in

December 1987 and since January 1988 has been used as an

experimental facility for trial burns. It has not been used

as an routine operating facility to burn cottage waste from

the center.

Until April 1990 the weekly waste from 6 to 16 selected

cottages was collected in dumpsters located at each cottage.

The dumpsters were hauled to the CCTL site once or twice a

week for experimental burns. The dumpsters were emptied with

pitchforks onto a conveyor belt leading to the incinerator's


In May 1990 the dumpsters were removed from the cottages

in a beautification effort. The waste from all of the

cottages is now collected daily and delivered to a compactor.

The compactor compresses and stores the waste. The compactor

is emptied twice a day, and the waste is hauled to the Archer

landfill. Since May 1990 one day's waste collected from all

48 cottages has been used in experimental burns once every

two to eight weeks.

Early experiments have involved coburning cottage waste

with locally available biomass [3,83-90], while later

experiments have involved spiking the incinerator with

polyvinyl chloride (PVC) to measure the effect of hydrogen

chloride (HC1) emissions on volatile organic compounds (VOCs)

emissions [4-6,91]. Other experiments have involved

coburning with coal, paper, or tire chips.

A laboratory-scale combustor has also been set up

simulate combustion and emission characteristics of waste

disposal and energy recovery systems [92-93].

One of the goals of ICAAS is to avoid chlorinated

materials in the waste stream so as to minimize emissions of

chlorinated organic compounds. Sociotechnical motivational

efforts, such as scoring each cottage or facility based on

its waste quality, are used to try to maintain a consistent,

low-toxic-producing, free-of-excess-liquids-and-

noncombustibles waste stream. Before the dumpsters were

removed, a Silver Dumpster award was presented to the cottage

with the best scoring garbage. The regular rusty green

dumpster for that cottage was replaced the week following a

burn with an aluminum painted dumpster with the UF-STC-CCTL

logo on it.

Incinerator Hardware

The incinerator system [94] (see Figure 2-1) consists of

a lower or primary combustion chamber (PCC), an upper or

secondary combustion chamber (SCC), a dual ram feeding


blower bicnass




board board

stoker ram

Figure 2-1. Incinerator system layout.

Method 5


system, a waste hopper, and a biomass feeding system. A

conveyor belt brings waste bags into the hopper. A water-

cooled stoker with rake fingers pushes unburned garbage back

toward the PCC burner and drags ash toward the ash door at

the end of the PCC. An external blower (big blower) provides

underfire air that comes through eight ports in the bottom of

the PCC. The biomass feeding system has a blower (bio

blower) for educing the biomass chips into the top of PCC, a

main hopper with 12-inch auger, and two side hoppers with 6-

inch augers that feed into the main hopper. The bio blower

also provides overfire air. Each combustion chamber has a

natural gas burner. A main blower provides air to the two

burners. An auxiliary blower provides air to the secondary

chamber. A butterfly valve (damper) at a "T" in the

horizontal portion of the stack allows ambient air to cool

and dilute the stack gases. The cooling is necessary since

the glass sampling probes used for emissions testing cannot

withstand the high temperatures directly from the

incinerator. Data from emissions testing is corrected to a

standard 7% oxygen content or is normalized to the waste feed

rate. Either procedure negates the effect of the dilution

air. The gases in the horizontal stack are assumed to be

well-mixed by the time (0.9 seconds) they reach the sampling

ports. A sampling platform runs the length of the horizontal

portion of the stack to facilitate emissions testing. The

100-foot tall chimney provides sufficient draft for the

combustion gases and the dilution air.

The incinerator was originally installed at the Veterans

Administration hospital in Gainesville, FL, in 1972. In 1986

it was replaced with a larger unit. In 1987 it was donated

to ICAAS and set up in front of the Tacachale (then called

Sunland Training Center) steam plant. The incinerator is

rated at 500 lb/hr of Type 1 waste. Type 1 waste or rubbish

is classified as having approximately 25% moisture, 10% ash,

and a higher heating value of 6500 BTU/lb. Type 2 waste or

refuse is classified as having approximately 50% moisture, 7%

ash, and a higher heating value of 4300 BTU/lb. Tacachale's

cottage waste typically falls in the Type 1 and Type 2

categories. The two burners are rated at 1.15 MMBTU/hr each,

though the PCC burner has been operated consistently at up to

1.8 MMBTU/hr without problem. The PCC and SCC chambers are 6

ft and 5 ft in diameter and 7 ft and 8 ft in length,

respectively. The inside walls of the chambers are covered

with an approximately 6-inch thick layer of refractory,

except at the top of the SCC where it thins out considerably.

The 36"-diameter horizontal stack is lined with 2" of


The additions of the big blower, biomass feeder and

blower, stoker, dilution damper, and sampling platform are

all modifications made to the incinerator by the CCTL group.

The extra blowers convert the incinerator from a starved-air

(in the PCC) type to an excess-air type.

Incinerator Operation

Operation of the incinerator system is as follows: the

incinerator is turned on, which fires the PCC and SCC

burners; the cooling water to the stoker is turned on;

garbage bags are weighed, placed on the conveyor belt, and

fed into the hopper while the incinerator is warming up; the

auxiliary blower turns on when the SCC temperature reaches

1500 OF; the ram-feed timer is turned on at the end of the

45-minute warm up phase; the ram-feed timer is set at the

desired feed interval, usually 7 minutes, and manually held

in to trigger the first feed; the stoker control subsystem is

turned on and its timer set to operate using the same

interval as the ram feeder, but set to trigger 3 minutes

after each ram feed; the overfire and underfire blowers are

turned on after the second ram feed; the hopper is kept at

approximately 2/3 full while garbage remains to be fed; the

PCC burner turns on and off as the PCC temperature falls

below and rises above 1800 F; the ram feed timer is turned

off after the last bags of garbage are fed in; the stoker is

left at the same triggering interval; the incinerator,

stoker, and external blowers are shut off at the end of the

2-hour burndown period. Two hours has proved to be a

sufficient time to burn out the waste in the incinerator

after feeding is done. The stoker cooling water is left on;

the ash is removed a day or so later.

The biomass feeder can be operated (turned on and off)

manually, or be turned on once and be set to turn off for a

specified length of time with each ram feed. The biomass

feeder requires several minutes for the material being fed to

first reach the top of the 12-inch auger. The port in the

top of the PCC must be open when feeding biomass. The

biomass hopper being used is kept about half full and is

checked for bridging (biomass arching over the auger)


Original Control System

The original control system for the CCTL incinerator

consists of temperature-based on/off burner controllers, a

temperature-based on/off switch for the auxiliary (SCC)

blower, manual on/off switches for the underfire (big) and

overfire (biomass or bio) blowers, a timer-based ram-feed

charging cycle with PCC upper and SCC lower temperature

limits, a timer- and ram-feed-based biomass feed cycle, and a

timer and/or ram-feed-based stoker cycle. The ram feeder,

biomass feeder, and stoker all have manual on/off switches as

well. Most modular incinerators have the same type of

electro-mechanical timers and switches found in the CCTL

control system.

The original control system circuitry is contained in

three panels mounted on the burner side of the incinerator

system. A functional schematic of the control circuitry is

shown in Figure 2-2. A schematic of the entire control

circuitry is shown in Figure A-i in Appendix A. A detailed


too hot


too cold

turn off
upper burner
leave upper
burner on

leave upper
burner on

leave ram-
feeder on
turn on

turn off


too hot


too cold


leave lower
burner off
turn off lower
burner & throttle
air to it
leave lower
burner on

pull acc ram back
push acc ram forward
pull feed ram back
pull acc ram back
push acc ram forward
open feed door
push feed ram forward
pull feed ram back
close feed door

turn off

leave ram-
feeder on
leave ram-
feeder on

feed door


Figure 2-2. Functional schematic of original CCTL
incinerator control circuitry.

description of the original control system in also in

Appendix A.

The lower right panel contains Honeywell model R4795D

and R4795A flame safety controllers for the upper and lower

burners, respectively. These burner controllers are powered

on and off via Compack API temperature limit switches in the

upper right control panel. When powered on, a burner

controller opens the valves for its burner. The burner

controllers continually read flame rod currents to make sure

the burners remain lit. If the flame rod current falls too

low, the controller assumes the flame must have gone out and

the controller goes into a lockout mode. The controller must

then be reset manually by pressing a switch on its outer

cover. During normal operation, the SCC temperature is

between its two API temperature limits (1500 OF and 2500 OF)

and the PCC temperature cycles between being below and

between its two limits (1800 OF and 2500 OF).

The CCTL's ECP incinerator has a dual-ram feeding system

[94]. Many modular incinerators have only a single ram for

feeding the waste, and incorporate an air-lock between the

waste and the ram. The dual ram of the CCTL's ECP system

negates the need for an external air-lock by using its upper

ram as the air-lock.


Sensors, Monitors, and Sampling Systems

The data acquisition system (see Figure 3-1) consists of

a Zenith Z-158 personal computer; a QuaTech expansion board

with a 16-channel digital input/output (I/O), 16-channel 8-

bit analog to digital (A/D), and 8-channel 8-bit digital to

analog (D/A) modules installed in the computer; and a Doric

data logger with remote front end module (FEM). The data

logger is connected to the computer through a standard RS-232

serial cable. A circuit board containing 6 relays, which

convert on/off 110-volt AC signals to 5-volt DC signals, is

connected to the digital I/O module. The FEM is installed in

one of the incinerator control panels and is connected with a

shielded cable to the data logger. The computer, data

logger, and relay board are installed in an instrument

trailer in front of the incinerator.

The continuous monitoring system consists of

temperature, control setting, air flow rate, and stack gas

composition monitors. One type-K thermocouple is located in

each of the following places: the primary combustion chamber

(PCC), the secondary combustion chamber (SCC), the stack

before the dilution "T", and the stack at the middle sampling

port 260" downstream of the "T". These thermocouples have

c~r 1U'Q
u (a 0 40 4
00 N ON uxdd g5i

E se s eB

0 0 -

3r U)I
0 90

4-4 0
0 00
Fx e, eTll H LO
ca T c 5 "a"

0 0 0 4) 4-
~5~B=I r. c 0 0C I YY

0a a)


v)O )


00 >1C O
N C,l r

0 E
0 coulI
0 co 4-)

co (

OQ f (

rc O

IZ4 0 ca 4

RlA 4 IrJ

1/8"-thick leads, which makes them fairly rugged for the

environment they are in. The thermocouples are connected to

the FEM with type-K thermocouple wire. A thermocouple was

installed in the continuous emissions monitoring (CEM) port

100" downstream of the "T" before the CEM system was

installed. Another thermocouple has been occasionally

installed in the PCC just above the ash door where the

combustion gases exit the PCC. The thermocouples in the PCC

and SCC are also read by the API temperature limit switches.

The incinerator control panel signal lines for the PCC

burner, the SCC burner, the auxiliary blower, the ram feed

door, the stoker, and the biomass feeder are connected to the

relay board by a 7-wire cable.

Orifice meters are installed in the air lines that run

from the main and auxiliary blowers to the two natural gas

burners, to the secondary chamber, and to the closed under-

fire air port. Honeywell Micro Switch pressure sensors are

mounted on the taps around the orifice plates. These 0-1 psi

sensors measure total and differential pressure and are wired

through 2:5 voltage dividers to the FEM. The pressure

sensors have been calibrated by a water U-tube manometer.

These orifice plates are located too close (less than 8

diameters downstream and 2 diameters upstream) to bends and

disturbances for accurate readings, though.

The CEM sampling line (see Figure 3-1) consists of a

1/2"-inch diameter stainless steel probe, a 11-cm glass fiber

filter, a water-cooled condenser, a water collection bowl, a

diaphragm pump, 60 feet of 3/8"-outer diameter polyethylene

tubing, a 500-ml drying flask filled with silica gel, and a

5-port sampling manifold. The filter, condenser, bowl, and

pump sit up on the sampling platform while the manifold and

drying flask are installed inside the instrumentation

trailer. The filter is held in a glass filter holder and is

backed by a stainless steel frit. The filter holder sits in

a box near the stack. The pump sends about 10 standard

ft /hr of stack gas sample into the trailer. A vent line is

installed before the drying flask so that the pump can draw

more stack gas than the instruments require. This allows

then instruments to respond faster to changes in the stack

gas composition.

Four continuous emissions monitors are installed inside

the instrumentation trailer. A Beckman 864 CO2 monitor, a

Beckman 866 CO monitor with zero/span accessories, a Servomex

777 Combustion Analyzer that monitors 02, and a Beckman 952A

NOx analyzer are all connected to the manifold. The CO2 and

CO monitors are nondispersive infrared (NDIR) analyzers. The

Servomex uses a paramagnetic cell for measuring 02 content.

The NOx analyzer uses the chemiluminescent reaction between

NO and 03 to measure both NO and NO2. The NO2 in the gas

sample is reduced to NO in a catalytic converter inside the

meter. The meter reads both NO and NO2 (NOx) when the NO2 is

reduced and reads just NO when the NO2 is not reduced. The

meter can read just NO2 by rapid switching back and forth

between reading NOx and reading NO and subtracting the two


The 14.3% CO2, 87.9 ppm CO, and 156 ppm NOx span gas

tanks are installed inside the trailer. The 02 analyzer uses

ambient air drawn from outside the trailer as a 20.9% 02

calibration gas. The 02 analyzer has an accuracy of 0.1%

02. The CO2 monitor's accuracy is 0.1% CO2. The accuracy of

the CO monitor's is 1 ppm CO. A 7:1 dilution system is

installed before the NOx analyzer. The dilution is necessary

since the highest range of the NOx meter is only 0-25 ppm and

the span tank has 156 ppm NOx The accuracy of the NOx meter

at its highest range is 1 ppm NO More details on the use,

calibration, and wiring hook-ups of the continuous emissions

monitor are in Appendix B.

A new HC1 continuous monitoring system was set up in

July 1991 (see Figure 3-1). It consists of a glass sampling

probe, a glass dilution chamber, a McNeil HC-3 HC1 sensor

cell, a membrane tube for drying the dilution air, two

rotameters, a 24-volt power supply, a load resistor, and a

diaphragm pump. The glass probe has a bend at the end.

When it is inserted into the stack, the bent end points

downstream. The stack gas is therefore drawn into the probe

in a direction opposite stack gas flow so that particulate

entrainment is minimized.

The analog outputs of the CO2, CO, and NOx monitors are

connected to the 0- to 5-volt analog input of the QuaTech A/D

module. The analog outputs of the 02, HC1 monitors are

connected to the FEM.

Particulate concentration in the stack is measured with

a standard EPA Method 5 [95] sampling train. This

concentration is an integrated average (as opposed to

continuous) measurement over an 1-hr sampling time. Hydrogen

chloride in the stack is measured by EPA Method 26 [96].

Sampling times of 30 to 60 minutes produce an integrated

average concentration. Semi-volatile organic compounds

(boiling points greater than 100 OC or vapor pressures

between 10-7 mm Hg and 10-1 mm Hg) and volatile organic

compounds (b.p. less than 1300C or vapor pressures greater

than 10-1 mm Hg) in the stack gas are measured by EPA

Modified Method 5 [97] and VOST [98-101] sampling trains,

respectively. Modified Method 5 measurements are 1- to 4-

hour integrated averages while VOST measurements are 20-

minute averages. These four sampling systems use a pump and

meter to draw a sample of stack gas through a filter,

adsorbent trap, and/or impinger solution to collect the

desired pollutantss. An extended VOST method, which

includes sampling for some of the lighter semi-volatile

compounds, is being developed by the CCTL group [102-103].

The ash is weighed after each burn and samples of the ash are

occasionally analyzed for composition, metal content [104],

alkalinity (pH) [105], and leachability. Leachability is

determined by using the EPA Extraction Procedure (EP)

Toxicity [106-107] or Toxic Characteristic Leaching Procedure

(TCLP) [108] tests.

The data acquisition program records all temperature,

control setting, air flow rate, and stack gas readings once

every 11 seconds onto a pair of floppy diskettes. The CO2

and CO readings recorded are actually the average of 10

successive readings, which helps to suppress noise, while the

other readings recorded are just single readings. The

program automatically can record up to 1800 readings during

up to 5.75 hours on each pair of floppy diskettes. Gas meter

readings, weather, settings of the underfire and biomass

blowers, weights of garbage bags fed, quality of garbage,

types of solid fuel(s), biomass continuous feed rate, and

sampling times are manually recorded on data sheets.


The CO2, CO, NOx, and HC1 monitors are turned on first

and allowed to warm up for at least one hour while the

sampling line is set up. At least 15 minutes before garbage

is to be fed the data logger, computer, and amplifier are

turned on and the data acquisition program is started. The

02 analyzer is turned on and allowed to warm up. The

sampling pumps are started, the CO2, CO, and NOx monitors are

then calibrated (zeroed and spanned), and the 02 analyzer is

calibrated. After all garbage is fed, the calibrations of

the CO2, CO, NOx, and 02 monitors are checked, and the

monitors are shut off. The data acquisition program is

stopped, and the amplifier, computer, and data logger are

shut off. The floppy diskettes, containing the continuously

monitored data, and any other data recorded by hand (garbage

bag weights, HC1 sampling time, natural gas meter readings,

etc.) are taken back to the office for later analysis.

A typical example of the protocol used for a typical

experimental burn at CCTL is shown in Table 3-1. All

monitoring, sampling, and analysis procedures closely adhered

to quality assurance/quality control (QA/QC) procedures [109]

developed for CCTL's Florida Department of Environmental

Regulation contract, "Measurement and Minimization of Toxic

Incineration Products".

Table 3-1. Typical Protocol for Experimental Burns.


Goals: 1.

VOST sampling while burning straight good garbage.
HC1 sampling while burning straight good garbage.
M2 volumetric flow rate measurement.
HC1 (continuous) emissions monitoring.
NOx (continuous) emissions monitoring.

Get VOST/Carbosorb stuff from Tacachale.
Purge VOST and/or Carbosorb traps.

Inform Tacachale of burn on Saturday.

Bring VOST/Carbosorb stuff out to Tacachale.




Open gate for garbage delivery.



Start CEM monitors.

Set up HC1 sampling train with grey meter box at far right port. 09:00
Set up VOST/Carbosorb sampling train with blue meter box at M5 port.
Hook up HC1 meter box thermocouple to digital temperature reader.

Set up stack CEM train.
Set up HC1 CEM train.
Set up cooling water system.

Warm up incinerator on natural gas.

Start computer.
Calibrate CEM meters.

Start feeding garbage.
Turn on the big and biomass blowers.
Set ram feed timer to 5.75 minutes so
Set stoker timer to 5 minutes so that

Run 1st BC1 (A22A) for 45 minutes.

Run 1st VOST (A22A) for 20 minutes.

Run 2nd VOST (A22C) for 20 minutes.
Run 2nd HC1 (A22B) for 45 minutes.

Run 3rd VOST (A22B) for 20 minutes.


that it feeds every 7.0 minutes.
it stokes every 7 minutes.







Set up MM5 probe with blue meter box at M5 port.
Connect probe thermocouple to digital temperature reader.
Run M2 (A22D) stack gas volumetric flow rate measurement.

Feed rest of garbage.

Do burndown and cleanup, with stoker timer set to 5 minutes.


The goal of this review is to seek relationships between

the operating conditions and the emissions, with emphasis on

determining the effects of fuel rates, temperatures, and

oxygen levels on carbon monoxide emissions and volatile

organic compound emissions. The relationships found should

be useful for control purposes. Rather than generating new

emissions data, this review will examine the three years of

accumulated emissions data from the CCTL incinerator

Simple relations exist between input rates of garbage,

gas, and air and output temperature, output rates of stack

gas, and concentrations or 02, C02, H20, HC1. These

relations, such as increasing gas input rate raises

temperature levels until the oxygen is exhausted, will serve

as a basis for more complex models needed for various

pollutant emissions. These models can then be analyzed to

determine the best operating conditions for lowest emissions

of certain pollutants at various input conditions.

High carbon monoxide emission levels have been used to

indicate poor combustion conditions in an incinerator [9].

Though CO levels may not correlate well with

chlorohydrocarbon (CHC) and principal organic hazardous

constituents (POHCs) destruction efficiencies [9], avoidance

of high levels of CO can be a conservative bound on high

levels of organic emissions. Benzene (C6H6), another product

of incomplete combustion (PIC), can serve as a surrogate for

other PIC and polynuclear aromatic hydrocarbon (PAH)

emissions [9].

Particulate emissions is another measure of an

incinerator's combustion performance. Semi-volatile organic

compounds in the stack often condense on the particulate

matter. Free chlorine in the stack gas can chlorinate

organic compounds found on particulates and in the stack gas.

Metal catalysts, especially cuprous chloride (CuC12), present

on particulates can help produce free chlorine from HC1 via

the Deacon reaction [14-15]:

2 HC + 02 --> Cl2 + H20 (4-1)

CCTL Trial Burn Data

ICAAS has conducted trial burns at the CCTL incinerator

for more than three years. During this time, incinerator

operating conditions, temperature and oxygen levels, solid

fuel feed rates, and natural gas rates have been monitored

for most burns. Input power levels are calculated by

multiplying the feed rate of a fuel by its higher heating

value (HHV). The higher heating value represents the maximum

energy available from the combustion of a fuel. Thermal

efficiencies are always based on a fuel's higher heating

value. The as-fired HHVs for gas, waste, and biomass are

typically 1050 BTU/cuft, 4500 BTU/lb, and 7000 BTU/lb,

respectively. The HHVs for gas and biomass are averages of

measured values. The HHV for waste is estimated from the

waste's paper, plastic, food stuff, and water content.

Continuous emissions data for CO2 and CO have been

collected for one and one-half years. Particulate emissions

were measured using EPA Method 5 [95] over a one year period.

Some semi-volatile organic compounds have been detected in

the stack gas using EPA Method 0010 [97] but have not been

quantitatively measured. Volatile organic compounds (VOCs)

in the stack gas have been measured during the last year

using VOST [98-101]. Total bottom ash from most burns has

been weighed. For several burns the ash has been broken down

into categories or into two or three sizes. Fine bottom ash

has been periodically analyzed for metal content, alkalinity,

and leachability.

Over 1,000,000 temperature, 100,000 oxygen (02), and

20,000 carbon dioxide (CO2) and carbon monoxide (CO) data

points have been recorded at the CCTL using its continuous

emissions monitors. These data were measured on a dry volume

basis and have been reduced to single integrated values for

each sampling run. Data calculations to determine integrated

temperatures and emissions levels are given in Appendix C.

The data collected cover a wide range of fuels, fuel loading,

and operating conditions. Thirty-nine sampling runs occurred

when just non-hazardous institutional waste was burned with

natural gas assist. Thirty-three of these runs (see Table 4-

1) include data for continuous emissions monitoring of 02,
















X C0

.v 00

0 fE




v 0.E



- g. 0

p* A



ur &
Sj ;

u a

X 0

^ ^^.h


* h*
*j ja
*^ E



.*0. .. ....... *.a. .

. .. . .. a 0 0 ...9 00 .0a. ....... ...

S -L-S5SSS-- aaaoa"uat.^ .>.Lo.LL > L.ELL
0 00 a

a 0OO000OC0000000000000 0OENEnNENEN

a M 0.cc o >>o 0aOC o o o> yy0

a. ooo oo :.. g.
C00000000000000000000 44444~~O
CCo CCO OOOOOOOOOCoooo 000000000 000

CO------C~~d~dd~oeln~oo 0E...--....
0 O0-.-.-.000000000000---1000000000

1; ~ <<< 0 ~ ~ < << UU< <<<< - o-----B--000o----------------
- - - --.........CCCCCCCCCCC

lEE g ll l l g l E E lE --f l l l E 00 C l EEEE' ^ ll

EE i-^^ ^ o Egg E tEE

* *4 S S
0 EENENCo000EN0 *NEIEN N.... 0 i~~^ otlE.CE oi.E
o g EN0 o 0 ENCoN 0 0EN CE- o50 EN C C.E oCEN..EnC...r.


- - - -no -- - - - --oc~r~n

toEN ..CC 00,E50N0 000 E...cC..E IOnCEnC 0.
*n I wnnnin0ior r-frnnoo^-

. . O. . ..OOCn C COC EN .N EN ENEN.c C

-- ^-r-- -- -- ---ic~o co,^-,,- --
0E~n nCCEN0nn0EEN00 0E 0E 0 CON CEN000CENC



*~ 55


- - - - - - --,^^ ,^^ -^ - -^

^-^ i*^^ nn~orn r.nrr ^-^.i ^

ENENEE OCOEN..N..EN 0 o0C1<0 C00EN.N0..O o

- - - - --000nC000 0 0N..E OCENNECCC 00
CCENC E .'gtElowoc- - --- --rno

9-EN.N go r Oo CO EN nn.CNENNENn 0COEn oCN

0 5



0! 0

g 0)



C >


s c


0 44
0 0

S 4

' C

8 (0

I e,
H Q)


t cn!



a 0)

o o


S t-i


; dC


a S
s +-
( 1-







j i

K *;


u 6





0 .

oo oo










0 0 00

* C N 0 C 0 C
J r-


E .N .N . .

&>* rn^OO r
RO c 0^ ^^

!0 SSS^S^

Li. n r' *r'

jj nocn c
: aOO

pg r" > e<-t

pr a r4n aw

i a w

10 0 -* 0 0

CO2, and CO on a dry basis. Most runs also have HC1 and/or

VOC sampling data. The other 6 runs (see Table 4-2) of the

39 include data on particulate and 02 emissions. The CCTL

data is analyzed here for relations between incinerator

operating parameters and emissions of CO, particulates, and

VOC. The VOC data includes benzene and some CHC and aromatic

hydrocarbon emissions data. The VOC emissions will be

studied here as a group, rather than as individual compounds.

Individual chlorinated VOCs will be studied in Chapter 6.

Carbon Monoxide Data

Of the 33 sampling/CEM runs, 27 had full data for

temperatures, 02, and CO. Both the big (underfire) and bio

(overfire) blowers were on for 24 of these runs. The CO

emissions for these 24 runs were compared to various

operating parameters using linear regression analysis. A

dimensionless statistical quantity, the coefficient of

determination (R2), will be used here as a measure of the

"goodness of fit" of a relation between a dependent variable

(an output such as CO and other emissions data) and

"independent" variables (inputs such as temperature or other

operating parameters). Higher values of R2 indicate better

fits, with 1.0 indicating a perfect fit. A more in depth

explanation will be given in the Review of Statistical

Procedures section in Chapter 6. A computer spreadsheet

program was used to generate the linear regression results.

For linear, 2-parameter (1 independent variable and a

constant) fits (see Table 4-3), the best R s were for CO

versus CO2 with an R2 of 0.134; versus PCC temperature

(Tpcc), 0.140; versus SCC temperature (Tscc), 0.148; and

versus hot stack temperature (Thstk), 0.138. All other R2s

were less than 0.07, including CO versus 02, though for the

most part CO remains low (less than 24 ppmv dry basic or 100

ppmv dry basis corrected to 7% 02) as long as the dry 02

level is above 17.5%. The high level of oxygen here is due

to dilution air being dumped into the CCTL stack. Without

the dilution the oxygen level is about 12.4%. An R2 of 0.118

or more here indicates that, statistically, there is only a

10% probability the data had this relation by random chance

(see Review of Statistical Procedures section in Chapter 6).

For quadratic (1 independent variable, its square, and a

constant) fits, the R2s for CO versus Tpcc, Tscc, and Thstk

were 0.248, 0.198, and 0.196, with temperatures for minimum

CO of 1730 OF (see Figure 4-1), 1805 OF, and 1560 OF, respec-

tively. For an R2 of 0.248 the equivalent correlation

coefficient, r, is 0.498. These quadratic fits have much

improved R s over the linear fits. Bilinear (2 independent

variables and a constant) fits for CO versus Tpcc and CO2 and

versus Tpcc and 02 yielded R s of 0.228 and 0.148, respec-

tively. Adding a Tpcc2 term to these last two fits increased

the R2s to 0.260 and 0.257, with PCC temperatures for minimum

CO emission in the range of 1700 F to 1760 F, depending on

CO2 or 02 concentration (see Figures 4-2 and 4-3).

Table 4-3. Linear Regression Results from CO Data Review.

Functional Relation

CO = 32.34373 0.04878 (waste rate)

CO =-14.47260 +



7% 02 =





14.14515 (gas power)

+ 2.643393 (input power)

- 17.3666 02

1154.305 61.5348 0.

+ 39.79183

+ 0.151814

+ 0.102442

+ 0.103479





CO =

CO =

CO @

CO =

CO =

CO =

CO =

CO =

CO =

CO =

CO =

CO =

CO =

CO =










4725 5.524 Tpcc + 0.001598 Tpcc2 + 18.50

7930 8.472 Tpcc + 0.002433 Tpcc2 + 2.418



(gas power) = 2.382266 0.16545 (waste power)

Tpcc = 1801.244 0.03626 (waste rate)

Tpcc = 1775.353 + 3.604221 (input power)

Thstk = 1903.200 64.8644 (input power)

Thstk = -162.634 + 0.952666 Tscc























-315.245 + 0.134705 Tpcc +

-97.6802 + 0.134535 Tpcc -

6742.107 7.755420 Tpcc +

1851.408 2.044650 Tscc +

1542.362 1.97369 Thstk +


--------- ---------------- ------ -

.. .- -- -- ----.

0 1800 1850 1900

Figure 4-1.


Figure 4-2.

PCC Temperature (F)
Quadratic fit of CCTL carbon monoxide data
to primary combustion chamber temperature.


rr J.-o nLd-i _on F







PCC Temperature (F)
Carbon monoxide data fitted with a quadratic
function of PCC temperature and a linear
function of oxygen concentration.










PCC Temperature (F)
Figure 4-3. Carbon monoxide data fitted with a quadratic
function of PCC temperature and a linear
function of carbon dioxide concentration.





I I l


Particulate Data

Particulate emission results were compared to various

operating parameters using linear regression analysis. The

best 2-parameter fits (see Table 4-4) were for particulate

emission (part) versus stack oxygen content with an R of

0.656 and versus natural gas input power, with an R of

0.765. Higher order fits were not attempted due to the low

number of data points. These analyses show that to minimize

particulate emissions, the oxygen in the stack should be as

low as possible (see Figure 4-4), while the natural gas input

should be as high as possible (see Figure 4-5). The result

with low oxygen is expected since lower excess air leads to

less carry over of particles from the PCC. The result with

natural gas is also expected since the combustion of natural

gas produces virtual no particulate emissions. However, high

natural gas flow rates correspond to low waste feed rates, so

there must be some trade-off between waste feed rate and

particulate emissions.

Table 4-4. Linear Regression Results from Particulate Data

Functional Relation R2

part = 0.166319 0.02840 (input power) 0.426045
part = 0.174602 0.05804 (gas power) 0.764566
part = 0.088428 0.00354 (waste power) 0.005993
part = 0.362099 0.00015 Tpcc 0.173282
part = 0.192295 0.00005 Tscc 0.065743
part = 0.243911 0.00009 Thstk 0.237539
part = -0.864920 + 0.052694 02 0.655576







17.80 18.00 18.20

02 (% dry)

O data fit

Figure 4-4. Particulate emissions compared to oxygen






--------------- ------ ... ...-------------

7] El

. I . I I I I I I I I I I I I I I





Gas input power (MMBTU/hr)

O data fit

Figure 4-5. Particulate emissions data compared to natural
gas input.


Volatile Organic Compound Data

Eleven of the 33 runs in Table 4-1 occurred during VOST

sampling. One of these runs occurred during two VOST

samplings. A detailed list of the emissions of 43 volatile

organic compounds sampled using VOST for these 12 samples is

shown in Appendix D. To facilitate analysis of this data,

the results from the twelve samples were ranked from 1 to 12,

with the run with rank 1 having the highest number of

compounds for which its emissions were the lowest or second

lowest and the least number of compounds for which its

emissions were the highest or second highest. The ranking

scheme used emission values from four different ways of

formatting the data: non-blank corrected ng/m3, blank

corrected ng/m non-blank corrected pg/kg, and blank

corrected pg/kg (see Appendix D). Blank correction involves

subtracting from the emission value for a compound the am-

of that compound that appeared on the field blanks taken

along with the samples. The amount appearing on the field

blanks is considered a baseline contaminant level. Usually

the baseline contaminant level is much less than the emission


The final rankings appear in the last column of Table

4-1. For the run with two samplings, the average of the two

rankings was used. The rankings were used in linear

regression analysis against various combustion conditions.

The R2s for all analyses are shown in Table 4-5. The best 2-

parameter fits were for VOST ranking (vr) versus waste feed

rate (WR) and versus PCC temperature (Tpcc), with R2s of

0.255 and 0.140, respectively. VOC emissions fell with

increasing PCC temperature and rose with increasing feed

rate. For 3-parameter fits, best results were for VOST

ranking versus WR and WR2 (see Figure 4-6) at an R2 of 0.329,

and for VOST ranking versus WR and Tpcc (see Figure 4-7) at

an R2 of 0.646. The curve for the WR-WR2 case yields lowest

VOC emissions at a waste feed rate of 315 lb/hr. The signs

of the coefficients for the WR-Tpcc case,

vr = 59.39418 + 0.35314 WR 0.03774 Tpcc, (4-2)

are as expected; the positive sign on the WR term indicates

that increasing the waste feed rate should increase emissions

while the negative sign on the Tpcc term indicates that

increasing the temperature should decrease emissions.

Equation 2 can be solved for PCC temperature for a VOST

ranking in the lower half (vr 5 6.5):

Tpcc a 1402 F + 0.936 b/r WR, (4-3)

where PCC temperature is in F and waste feed rate is in

Ib/hr. For example, with a 400-lb/hr waste feed rate, the

PCC temperature should be above 1776 F.

The highest ranking VOST data point in Figure 4-7 is far

away from the rest of the data points. With this point

removed the regression of VOST ranking on WR and Tpcc results

in R2=0.773 and Equation 4-3 becomes

Table 4-5. Linear Regression Results from VOST Data Review.

Functional Relation

vr = 12.69634

vr = -2.53150

vr = -1.41614

vr = 43.32779

vr = 15.75597

vr = 14.67748

vr = 81.83877

vr = 6.771868

vr = 7.194926

vr = 7.329590

vr =

vr =



vr=20.92161 -

vr=33.10936 -

vr = 129.6102

vr = -2.52417

- 2.80065 (gas power)

+ 0.023015 (waste rate)

+ 2.020129 (input power)

- 0.02075 Tpcc

- 0.00485 Tscc

- 0.00496 Thstk

- 4.22762 02

- 0.26058 CO2

- 0.07187 CO

- 0.01956 CO @ 7% 02

+ 0.299665 CO (w/o highest CO)

+ 0.070243 CO7% (w/o high CO)

0.10194 (waste rate) + 0.000161 wr2

15.1771 (input power) + 2.131065 ip2

- 9.57080 02 + 0.149382 022

+ 6.497911 CO2 1.21779 CO22

vr=59.39418 + 0.035314 (waste rate) 0.03774 Tpcc

vr (w/o high)=59.40090 + 0.030509 wr 0.03691 Tpcc




















Tpcc 1401.54 + 0.93572 (waste rate)

Tpcc 2 1433.24 + 0.82658 (waste rate) (w/o highest vr)

2 1 1111 1111 1 R II a I 1 1 I @1 I
250.0 300.0 350.0 400.0 450.0 500.0 550.0

Waste rate (lb/hr)

o data fit

Figure 4-6.

Vost emissions compared to waste feed rate.



1700 1750 1800 1850


PCC temperature (OF)

0 data fit

Figure 4-7. Vost emissions compared to PCC temperature
and waste feed rate and plotted against

Tpcc 2 1433 OF + 0.827 lb/hr WR, (4-3)

For this equation, a 400-lb/hr waste feed rate yields a

minimum PCC temperature of 1764 oF.

Summary of Significant Relations

The review of CO emissions data from CCTL shows that to

operate the incinerator with low emissions, the diluted

oxygen level should be under 17.5% (12.4% undiluted) the

PCC temperature should be about 1730 OF, and the SCC

temperature should be about 1805 OF. Particulate data show

using low levels of oxygen for low particle emissions. The

VOST data show for low VOC emissions to use a 315-lb/hr feed

rate, or use a PCC temperature of 1776 OF with a 400-lb/hr

feed rate.


Kinetic modelling of the combustion process can show

fundamental bases for relationships between inputs to the

incinerator, operating conditions, and emissions, whereas an

energy analysis code, such as that of Kodres [110-113], will

only calculate the temperature and major combustion products

for given inputs of waste, air, and support fuel.

The concentration of CO and other minor species can be

estimated from a thermodynamic equilibrium analysis. For CO

the appropriate equilibrium reaction is

CO <--> CO + 02 (5-1)

[114] with an equilibrium constant of

Kp= PCO PO1/2 PCO1 (5-2)

The partial pressures of 02 and CO2 at the CCTL incinerator

in the stack before the dilution "T" (see Figure 2-1 in

Chapter 2) are estimated at 0.097 and 0.068, respectively.

These estimates are based on 02 and CO2 measurements after

the dilution "T" and a dilution factor of 1.5:1 based of mass

flow determinations. For an average temperature of 1800 F

(1256 K), K=10 278. Solving Equation 5-2 with these

values for pCO yields pO=1.15xl0-8 or 0.0115 ppm. Actual
values for pC yields pC0=.15xlO or 0.0115 ppm. Actual

values of CO at the same location average about 20 ppm.

Therefore, equilibrium analysis is not realistic enough to

describe the emissions from an incinerator.

Kinetic modelling also can show how minimizing toxic

formation and maximizing toxic destruction are related.

These relationships can be used to create or fine-tune

empirical models (like those derived from reviewing test

data) useful for controlling incinerator emissions.

A typical gaseous chemical reaction is of the form

EG + J --> E + GJ. (5-3)

The kinetic reaction rate, k, is of the Arrhenius form,

k = A Tn exp (-Ea/(Ru T)), (5-4)

where A and n are constants for the reaction, T is the

absolute temperature, Ea is the activation energy of the

reaction, Ru is the universal gas law constant. The constant

n is zero for most reactions. The ratio E /R is often
a u
listed in tables of chemical reaction rate parameters instead

of just E The change of concentration of the species in

the above reaction is given by

-d[EG]/dt = -d[J]/dt = d[E]/dt = d[GJ]/dt = k[EG][J]

= [EG][J] A Tn exp (-Ea/(Ru T)) (5-5)

For a more complex example,

c C + d D --> e E + f F (5-6)

equation 5-5 would become,

1 [C 1 dD] 11 [E d[F k B
c dt d dt e dt f dt AB
= [A][B] A Tn exp (-E /(Ru T)) (5-7)

The units of the constant A depend on the number of reactant

species. For a system of reactions, the solution for the

time-dependent concentrations of the species involved

requires the solution of a coupled set of nonlinear

differential equations.

A chemical pyrolysis/kinetics code for modelling coal-

water-gas combustion [115-116] was modified for studying

emissions from a modular incinerator. The code now uses the

LSODE routine [117] to numerically solve the coupled kinetic

differential equations. The LSODE routine is a Gear-type

solver for stiff equations, such as those used in chemical

kinetics. The reactions and their rate parameters are now in

a data file that is read by the code. Inputs of species to

the combustion system are now in lb/hr. The species' amounts

are now calculated, displayed, and printed in kmol/hr. The

concentration of the third body M is calculated as the total

concentration of all species at each time step. The third

body is used for energy transfer in reactions where a

molecule is split into two or where two molecules recombine.

The code is outlined in Figure 5-1 and listed in Appendix E.

The code is currently run on the DEC VAX cluster at CIRCA at

the University of Florida. It has also been run on SUN SPARC

workstations. A typical run takes 2-3 minutes on the VAX as



For each reaction:
Add to list of
reaction data
Add to list
of species
Add to list
of reactants for
each reaction
Add to list of
reactions destroying
each species
Add to list of
reactions producing
each species

SFor each amount:
Add to list
of species
Add to amount
for this species

If first time thru\
write species names
to output file /

Read list
of time steps



Figure 5-1. Kinetics code



Write initial time,
temps, and amounts
to output file

For each time step:
Call integrator
(it calls cdot)
Write time, temps,
and amounts
to output file





)t subroutine
late temperature


program flowchart.

Calculate total amount
Calculate total volume
Calculate each specie's
Calculate third body
For each specie calculate time
rate of change of concentra-
tion using Arrehnius equation
For each specie calculate
time rate of change of amount

opposed to an estimated 6-7 hours on an 4.77-MHz 8088-powered


Most of the reaction rates in the code input file (see

Appendix F) were compiled from various sources [118-120].

Where reaction rates were not known estimates were generated

from thermal destruction [121-124] and from similar reaction

mechanisms. To get to gas-phase reactions the code uses a

simplistic yet reasonable pyrolysis process. Semi-global

reaction rates for the formation of chlorinated and non-

chlorinated benzene, phenol, furan, and dioxin were developed

in modelling data from the Pittsfield-Vicon incinerator study

[1,7,125]. This modelling is still going on by the CCTL.

The kinetics code was written by CCTL to give quick

results when studying combustion and pyrolysis processes. A

commercial code could have been used, but using a code that

one has written oneself has many advantages: the code can be

modified at will to adapt to the problem at hand; the code

can be made to run on existing hardware; the input to and

output from the code can be made into any convenient form;

and writing one's own code forces one to learn the

fundamentals of the process. There are currently no

reasonably-sized (mainframe computer-sized) kinetic models

for chlorohydrocarbon combustion [126].

Rather than keeping track of heat of reactions, heat

transfer, and combustion gas temperature, the code requires a

temperature-time profile be imposed. This avoids modelling

the heat transfer characteristics of the combustion system

Temperature (OF)
2000 -i








0 1 2 3 4 5

Time (sec)

Figure 5-2.

Temperature-time history for CCTL incinerator.

and greatly simplifies the computer code. A typical

temperature-time profile for the CCTL incinerator is shown in

Figure 5-2. This profile is based on measurements of flow

rate and temperature at various places along the path of the

combustion gases through the CCTL incinerator (see Appendix

G). The code starts after the heating of a slug of pre-mixed

input materials to 1000 OF. The code then follows the slug

of gases as it travels through the incinerator along the

temperature-time profile. The rapid rise in temperatures in

the PCC and SCC as shown in Figure 5-2 is due to the heat

released from the natural gas burners and waste combustion in

the PCC. The end of the profile at 5.7 seconds is at the end

of the horizontal stack. The sampling ports for M5, VOST,

and HC1 are at 5.1 seconds. The sudden drop off after the

SCC is due to the dilution "T" (see Figure 2-1 in Chapter 2)

where ambient air is drawn in, diluting and cooling the stack

gas. Since this code uses no fluid dynamics, the gaseous

species are assumed to be instantly mixed and cooled when

more air is added at the dilution "T" and instantly mixed

when natural gas and air is added at input to the SCC. The

code stops generating output when the slug reaches the 5.1-

second sampling ports.

Inputs to the Code

The code requires an input file divided into four

sections (see Appendix F). The first section lists, for each

reaction, the chemical formula, the Arrhenius reaction rate

constants (A, n, E /R ) (kmol/m -K based), and, if known, the
a u
enthalpy of reaction (kcal/kmol), though it is not used by

the code. By keeping the reaction and their rates in a data

file instead of the main program, the program is made generic

and does not have to be recompiled for a different set of

reactions. The second section lists the input rates (in

lb/hr) of chemical species into the incinerator. The third

section lists the time steps for which the code will produce

output. The fourth section lists the temperature-time

profile parameters. For simulating processes such as the

CCTL incinerator with consecutive, different temperature-time

profiles and multiple input locations, the second, third, and

fourth sections may be repeated with the new inputs as

needed. The output file of the code lists the time,

temperature, and effluent rate (in kmol/hr) of the various

species at each time step. For the CCTL, the final time step

ends where the combustion gases reach the sampling ports 260

inches downstream of the "T" at 5.1 seconds.

Waste Composition

The waste in the code is composed of five parts: water,

polystyrene, PVC, ash, and a garbage "molecule". The

pyrolysis/kinetics code was previously developed for a

macromolecular model for the structure of coal with unit

molecule such as C35H2603A [115-116], which is compatible

with the mass fractions of carbon, hydrogen, oxygen, and ash

(with an "atomic weight" of 50) for coal. The institutional

waste at Tacachale contains about 25% by weight water, and 7%

by weight ash, with the bulk mainly made up of plastics,

paper, and some food stuffs (see Table 5-1). The coal

pyrolysis model is used here since garbage, like coal, can be

modelled to be made of large molecules due to its paper,

plastic, and food stuff components.

To be compatible with typical measurements of 02, C02,

H20, and HC1 in the stack at CCTL, 400 lb of waste was

modelled to consist of 100 lb of water, 20 lb of polystyrene,

2 lb of PVC, and 278 lb of garbage "molecules". In the code,

the PVC accounts for all of the chlorine in the waste. The

0- I rn

0 I '








N 0O


( ,1










I m


0 4

10 (



, I


o 0
D 0

D o

o o

0 0

n 0

o o
0 LA
o a

o o
o o

o 0
o o

o o

o o

O m

oA A

L 0

o C0
& &

i Im
SI m
o\ le

278 lb of garbage molecules includes 28 lb of ash. The unit

molecular formula of the garbage is taken as C21H42021A, with

a molecular weight of 701, where the ash "atom" A has an

atomic weight of 71. This molecular weight is comparable to

those used for coal molecules in the old pyrolysis/kinetics

code, and used a set of C, H, and 0 empirical coefficients

that were closest to whole numbers.

Pyrolysis Steps

The breakdown of the garbage molecule occurs here in

three pyrolysis steps, modelled after the coal pyrolysis

process used in the pyrolysis/kinetics code:

C21H42021A --> C16H20A + 5 CO2 + 11 H20, (5-8)

C6H20A --> C6A + 10 CH2, (5-9)

C6A --> 3 C2 + A. (5-10)

The coal pyrolysis model is used here since the pyrolysis of

garbage, like coal, can be modelled as the breakdown of

large molecules. In the first step, all oxygen is driven off

as loosely bound CO2 and H20, leaving a tar. In the second

step, the remaining hydrogen is driven off as CH2, leaving a

char. In the third step, the char breaks down into C2 and

ash. The CH2 and C2 enter into the combustion process as

they are evolved. In the kinetics code the ash is ignored

since it is never considered a gaseous component. The same

reaction or decay rate is used for each of these steps:

T 7 -1
k = 1000 (3000 K) sec (5-11)

This reaction rate was developed from modelling coal mass

loss versus temperature and time data from Massachusetts

Institute of Technology (MIT) [115-116,127]. Carbon monoxide

levels increase by about 10% if the 1000 constant in Equation

5-11 is doubled to speed up the pyrolysis process.

The polystyrene first breaks down into its monomer

styrene, which then breaks down into benzene and acetylene

(see Figure 5-3). The PVC breaks down directly into HC1 and

acetylene. These breakdowns were modelled from thermal

destruction of polymers [121-124]. The aromatic sequence is

based on work by Green et al. [4,7,125].

Results from the Kinetics Code

Typical inputs of natural gas to the CCTL are 30 and 37

lb/hr of CH4 to the PCC and SCC, respectively. Typical waste

input is 400 lb/hr (100 lb/hr of water, 20 lb/hr of

polystyrene, 2 lb/hr of PVC, 250 lb/hr of ash-free garbage

"molecules", and 28 lb/hr of ash). The ash is not actually

included in the input file. Typical air inputs are 2679,

2679, and 9911 lb/hr to the PCC, SCC, and dilution "T",

respectively. The air consists of 76.08% by weight N2,

22.98% 02, and 0.94% H20. The water content (1.5% by volume)

in the air is due to the high humidity typically found in

Florida. These inputs, with peak PCC and SCC temperatures of

1800 OF and 1975 F, are considered the base case (Run 30)

here. The output for this base case is listed in Appendix H


0 HC]

H20 +



+ C2H2




OH C6H6 + C2H2

H5C1 C6H50H+H 3C2H2

I40HC1 1 \\3C2H2+O

H20+C12H80 C12 C12H802+H2

C12H7OCl+HC1 C12H702C1+HC1

Figure 5-3. Kinetic aromatic and chlorination sequences.


and graphed in order decreasing emission rate in Figure 5-4.

For all cases examined, excluding N2, 02, and H20, the

species' amount change little after the dilution "T" at 4.2

seconds. The sudden drop in temperature appears to halt most


The first kinetic code runs (Runs 1-19) used various

waste pyrolysis rates to get the garbage, tar, and char

burnouts in the appropriate range and do not yield results

comparable to the later ones examined here.

Initial kinetic model studying (Runs 20-29) examined the

effects on carbon monoxide and benzene due to varying the

temperatures in the PCC and SCC, while keeping waste, air,

and gas rates constant. For minimal carbon monoxide

emission, the optimal peak PCC and SCC temperatures are 1875

F and 2050 F, respectively (see Table 5-2). This would

correspond to measured temperatures of about 1850 OF for the

PCC at 1 second in the temperature-time profile (see Figure

5-2) and about 1925 F for the SCC at 3.5 seconds. These

times are where the thermocouples are located in the

temperature-time profile. For minimal benzene emission,

optimal peak PCC and SCC temperature are 1900 F and 2025 F,

respectively (see Table 5-2). This would correspond to

measured temperatures of about 1875 OF for the PCC and about

1900 F for the SCC. The residence time for each chamber was

kept constant for these cases since the residence time

decreases only 1% for each 25 F rise in temperature (due to


1E+03 N2
I 02
1E+02 B20

-- H20(IQ)
1E+01 .......- H20(LIQ)

1E+00 - - -- - - - CO

1E-01 H2


1E-03 . 'i- POLYSTYRN
I ... C6H6
1E-04 -----

S-- char
.c 1E-05
1E-06 HCl

1E-07 -- ----
S" C2
....... I .......- O
0 1E-08 0

1E-09 -I --- CBH20

1E-10 ... ............... ....O

I -' C12
1E-11 --------- - ------ -
~ CH2
1E-12 CBO


1E-15 --------------- -------- C6H5C1

1E-16 ..
o 1 2 3 4 5 6 CH
Time (sec) ClDX

Figure 5-4. Base case (Run 30) kinetics code output.

U X m m r- m ch m a, m %o r- m 0 h i o w O m ui r- ch vr- r-.r N u- r- r- &n -)Oc OU r r- o w nm C4


%0 t~ M Ln c,4 m c,4 y -4 o c ,4u cj cj c,4 cu c w r- w r- w %o w w a N N %~o L n -4 r, O > 'D M 14 Ln r-, M k
%0 m -0 % c % 4 -- 4 NO I N 0 M I I I I I

c'l>, o o e4 n 4

N -r-I NC NN N N N N N .~ N N~ N N N N N -4 t rn~t r-4 t n N 0 N n C; N NNN

lA* p~~- 0N0000~ ~ No'~~o~~oP~N~-I0O
'C 0( 0 0 0 0 0 0 0 0 0 0. 0 0 00 0 0 0 0 0 000
-o $ r-4 -1 -4 r-4 o ch -i ch ch ch c c o i co m r-i r- v C-i o n r- m r- ch c,4 c,4 ci -t C4 % C %0 a, -4 r-


o ~u
>, " N C-4 C-4 C- N C-4 N N N N N N 0 0 V M %0 V 0 04 r-4 r- r- V r tnN 0 tn tnN N M N rn 0 M -.7 -

I (
U 4 w -4 o .r % r4 %o .r co m w co co m O m h t co Ln %o %o o wi wl w m v C r4Ch c4w oc4 U1 '

U4 0 ~000000~~~N ~ ~ OO~N~0~

o ~
u 5

r. % h.)r 0r o m NN N N N r- m -rr- (-4 v 0a,-4'0 M VOc~ -4 M.7r'J'0 -4'O


u ~0000000000000000000010~00JOOO00000 0000000

UX 00O1000000OOO0OO(~NN0O~a~ON0,gN00000000000000
C; C; C; C; C; C; C; C; 4 C Y*% %o P P 4 ch r-* k,3 k, ) o o . . . . . . . .
c,4 o c-i co c4 m co -o o ki ea w w woo~

VX ~~~000~000000000000000~~000000000000000000
0 m M m M M M r4 m N r4 P O OO OO OO -I o Ch fn C4 N Ch O O n %D %DPO

IE O~~~~~~~~~~~00~ ~ ~~~~0000000i~~ulu0~c~000O0000
o 00 %0000ovoooooooooo IN N %ooorooooooo-ANlCoir-
U'-4 1-
r-40 i % nLnwi i o wi wi wi %oo % o %o -i h o m o 4 ch m o Ln %n vi o wi wi o %o %o wi wi o 0 o -.7 4

o4 h h ov000000000000000000000%00%00 Mr-00 Ma00lAOlA0lA0
1*-4 ~
-; 44 4; 4444 C; C;4 0 ,44C ; ;C 0,0, C

0 0 0 0 0 0 0 0 0 0 Ch Ch ON ON ON ON ON ON 7 Ch h Ch0 0 C 0 Ch o C Ch I 4 C
L) m ~ ~~Omom~oooo~ooooooooooh~vrvrooooooooo0000

gas's higher velocity), and the accuracy of the residence

time calculation is at best 5%.

Non-chlorinated benzene emissions were calculated by the

code at levels of 100-3000 pg/kg (output benzene over input

waste), which compares well with typical emission measure-

ments at CCTL (see Appendix D). For chlorinated benzenes,

represented by monochlorobenzene in the code, the code's

prediction of 100-300 ng/kg is about an order of magnitude

low, though the reaction rates in the code were designed

[4,7,125] so that the code matched the results of the

Pittsfield-Vicon data [1].

Varying the ratio of PCC air to total air from 60% to

40% (Runs 30-34) had little effect on the CO, C6H6, or C6H5C1

concentrations. For cases where the waste, gas, and air

rates were varied together so as not to change the

temperature-time profile (Runs 30, 35-42), there was little

effect on carbon monoxide, with possibly a minimum occurring

with a 350 lb/hr waste feed rate. Benzene emissions were

negligibly affected. Chlorobenzene emissions were lower for

lower feed rates.

When the waste, gas, and air rates are lowered to 90% of

their base case values, with no change in the temperature-

time history (Run 43), CO decreased from 247 ppm to 230 ppm,

while C6H6 decreased negligibly and CH 5Cl remained the same.

When the waste rate is increased to 450 lb/hr, with the rest

of the inputs at their base case value and with peak PCC and

SCC temperatures raised to 2004 OF and 2087 OF to account for

the extra power input from the waste (Run 44), CO decreased

to 59 ppm. With these higher temperatures, if the waste rate

is dropped back to 400 lb/hr and the gas rate to the PCC is

raised to compensate (Run 45), CO increases almost negligibly

to 61 ppm. Benzene emissions for the last two runs were all

but eliminated, while C6 HC1 emissions were substantially


Decreasing the air flow rate into the PCC by 5% and

increasing the air flow rate into the SCC by 5% from the base

case, while raising the peak PCC temperature to 1877 oF to

compensate (Run 48), increases CO emission from 247 ppm to

300 ppm (see Table 5-2). Increasing the air flow rate into

the PCC by 5% and decreasing the air flow rate into the SCC

by 5% from the base case, while lowering the peak PCC

temperature to 1730 F to compensate (Run 53), also increases

CO emission, to 336 ppm. So, for the same total air flow

rate, the base case is near optimal for CO emission.

Decreasing either air rate solely by 5% (Runs 46 and 49) cuts

CO emission in half, while increasing either air rate solely

(Runs 50 and 52) by 5% doubles CO emission. However, the

lowest PCC air rate cases had the lowest benzene emission,

mostly likely due to the increased PCC temperature. Benzene

emission trends followed those of CO. Chlorobenzene

emissions were lower for lower PCC air rates, but were hardly

affected by the SCC air rate.

Lowering the PCC air rate substantially while raising

the PCC and SCC temperatures to compensate (Runs 30, 56-61)

decreases the CO and C6H6 output, with most of the decrease

due to increases in the PCC and SCC temperatures. Benzene

and C6H5Cl emissions trends followed those of CO here as


Conclusions from the Kinetics Code

Carbon monoxide emissions predicted by the code are very

high compared to actual measurements. The input amounts and

reactions to the code need to be improved to reduce the CO

emissions. The temperature-time profile could be fattened to

allow more time at higher temperatures, which should reduce

the CO emissions. The air brought in at the dilution "T"

could be gradually mixed. This also would keep the

temperature of the combustion gas higher until the two

streams are fully mixed.

The fast quench of reactions after the dilution "T" may

minimize formation of certain organic compounds that would

otherwise form due to slow cooling.

The initial kinetics code modelling (Runs 20-29) shows

minimal CO emissions with the PCC and SCC at 1850 OF and 1925

OF, respectively, while minimal benzene emissions with the

PCC and SCC at 1875 OF and 1900 OF, respectively. Later

kinetics code runs show lower CO with decreased air rates,

mainly through increased temperatures. The PCC air rate can

only be lowered so far before the incinerator becomes


Results from runs 20-29 are questionable since the

temperature was varied without changing the inputs levels of

waste, air, or natural gas. Results from runs 31-34 are also

questionable since air flow rates were varied without

changing the temperature-time profile.

The significant conclusions relevant for minimization of

emissions are to burn hotter, avoid chlorine in the input

(the code has shown increased chlorinated organic emissions

with increased PVC input in other studies [4,7,125]), and

burn longer (have a greater residence time).


Conventional wisdom on the subject of the effect of

chlorinated input, particularly in the form of PVC, on the

levels of chlorinated organic emissions from incineration can

be summarized by the following statements:

There is no evidence that the amount of PVC in the
waste affects the levels of PCDDs/PCDFs at the
boiler outlet, tertiary chamber, or the stack.
[1:10-4] There is no statistically significant
relationships between the levels of PCDDs/PCDFs and
NO SO2, THC [total hydrocarbons], or HC1. [1:10-

There are multiple sources of chlorine in MSW
[Municipal Solid Waste]. Hence attempting to
reduce PCDD/PCDF emissions based on a strategy of
lowering the chlorine content of MSW by separating
chlorinated plastics in unjustified.... No rela-
tionship exists between PCDD/PCDF emissions and the
amount of PVC in the waste (or the concentration of
HC1 in the flue gas).... Incinerator operating
temperature significantly affects the levels of
PCDDs/PCDFs produced, and the levels of PCDDs/PCDFs
and carbon monoxide are related. [2]

These statements refer to the results of the Pittsfield-Vicon

incinerator study [1]. This study attempted to determine the

relation between operating temperature and PVC input on PCDD

and PCDF emissions. However, only 4 pairs of runs studied

PVC input, while the data analysis for PCDD/PCDF versus

PVC/HC1 relations used all 19 runs, which had many variations

in temperature, CO, and input PCDD/PCDF levels that could

have obscured such relations. Also, only linear regression

analysis was used to examine the data.

The above statements contradict a logical conclusion

that reducing chlorinated input reduces chlorinated organic

emissions. Obviously, if there is no chlorine in the input

waste stream, there can be no chlorinated organic emissions.

In MSW 50% of the chlorine comes from PVC in the waste.

However, in Medical Waste Incineration (MWI) PVC accounts for

most of the chlorine in the waste. Green et al. [3-8] has

reanalyzed the Vicon data and found relationships between

chlorinated hydrocarbon (C1HC) emissions and HC1 levels.

Emissions data from the CCTL incinerator has been compared to

the results of emissions testing on several California

medical waste incinerators [3-5,7]. Emissions of HCl and

ClHCs from the CCTL incinerator were both about one-tenth the

level as from the California incinerators, which indicates

probable relations between chlorine levels and C1HC emissions


Correlation of emissions with operating parameters is a

crude field since stack sampling is a one decimal point

science at best and stack emission data usually are very

noisy. Many sampling runs are needed at identical and

different operating conditions to form a sufficient set of

data for statistical analysis. To evaluate the validity of

the conventional wisdom, and to bring order to this crude

field, data on chlorinated emissions from the CCTL

incinerator will be analyzed to determine any relations

between chlorine input and HC1 output, and between HC1 output

and chlorinated organic compound emissions. An attempt is

made to systematically analyze the CCTL data using linear,

multivariate, and nonlinear approaches.

PVC, HC1, and VOST Data from CCTL

Since the chlorine level in the waste is not known for

each burn, it is useful to have a surrogate to indicate the

level of chlorine in the waste and to compare chlorinated

organic emissions with. The best surrogate is the major

product of chlorine combustion, HC1. The CCTL has made over

30 measurements of HC1 emissions with different levels of

chlorine in the input [3] (see Table 6-1). The PVC pipe

added during the earlier experiments was assumed to contain

40% PVC and 60% filler material. PVC itself contains 56.8%

by weight chlorine.

Data from CCTL sampling runs where both HC1 and VOST

(volatile organic sampling train) measurements were made will

be analyzed to determine if any relationships exist between

chlorinated volatile organic compound (VOC) emissions and

chlorine input. The VOST measurement procedure examines 43

compounds (see Appendix D). Twenty-seven of these compounds

are chlorinated. However, only a few compounds show up in

sufficient amounts and often enough to be considered

indicators of overall volatile organic emissions. Benzene,

toluene, (mono)chlorobenzene, the dichlorobenzenes, and

chloroform fit this criteria. Other compounds such as

ethylbenzene, and the xylenes may also fit this criteria.

Eighteen VOST runs occurred while sampling for HC1 and

burning just NHW or NHW and PVC. Six runs had PVC resin

added into the incinerator while NHW was burning. Data for

feed rates,

Table 6-1.

temperatures, CO, HC1, and many VOC emissions for

CCTL PVC and HC1 Test Data.






Cl HC1
input output
lb/hr lb/hr

12.50 p 2.84
6.50 p 1.48
4.65 p 1.06
3.02 p 0.69
10.00 p 2.27
18.00 p 4.09
9.00 r 5.11
9.00 r 5.11
18.00 r 10.22
9.17 r 5.21
5.25 r 2.98
3.00 r 1.70


400-lb/hr feed
rate equivalent
added Cl HC1
input output
lb/hr lb/hr



p=PVC pipe (22.7% Cl), r=PVC resin (56.8% Cl)



r-I *

0 0n

) 0
1o C



r. .




r- 1
0 0
E c


0 en

, (N o on
Z o Z

U -W m
U) U
S(N Zn


N a

E-, rj





U 4n



.. ZL

m m
U a

~o ~

a 0
C ~ -



o .a

-. o
So v
c a

~o o
Ed- + U
N) 0 e

0. -

m o a
U o

1 -1

z oz
0l I< M


n o

CN z
c^ O

0 0 % 0
o o m

v co
o a, 00

L -

U, m

n (N

z z

o (N in o0
C; C ; C

o o 0 0

D co 4 coi
V( I M
1 N0 m
oT r~~

0 C4
o N

o 1

o o,
o *' ,
o 0 O

OD m
*o r-
(N Or


o O CN 0 0 (N
- O 0' 0 0 -
%0 T 0 0 OT
z 4C;C

- 0 40
O0 0 ~

U ,
0 r-

(N r-

r- oo co
rf- r- In

Un 07 O7
Ue) t- ('n

< 0,


in m

^ oo



o O
o 4

o ru
o o-


O % q
(,o~ U%
O~ ro

0 00
o m
o (N

- P

Mi e)

%0 0
.N N.

(N (N

(N (N
,O C!

Uo ,

0 o
S (N


r- 0

o o
oo o
N U,

M (o

(N (N

fN C
(N N
(N N

o N

r- CN

- V N C- M "
w Ln r oo V r a,
r^ r^ r^ oo o >o

(N oa Un -
*O 0 0 T
- r- c-
r-1 -I < F-

*o c) r- Ue -
o o >o %0 LO
C rD r r r -
- = -

L) .r ) a a c 0 0 I O
~~ m O~~~, N C', ~


-- LO, 0 0 0 v vl 0 k 0 0
- N - (N m 0 (
co a, i 0
s S ^s s s ^ S s3 ^

o 0

o *'
>-< r-t

these 18 runs is shown in Table 6-2. The CO, HC1, and VOC

emissions were normalized to the waste (including any added

PVC) feed rate. Carbon monoxide emission is listed in (mg/hr

CO)/(kg/hr input waste), or mg/kg for short; HC1 is listed in

g/kg; and the VOCs are listed in pg/kg. These unit result in

data ranges of 1 to 1000 so that very small numbers are


The analysis of this data will use standard statistical

procedures for linear and multivariate regression. Also some

nonlinear analysis of the data will be performed based on

earlier analysis of Vicon data by Green et al. [3-8].

Review of Statistical Procedures

In statistical analysis data is usually modelled using a

simple linear equation:

Y = a + B X, (6-1)

where Y is the fit to the dependent variable Y, a is the

intercept, B is the slope, and X is the independent variable.

The intercept and slope are determined by a least

squares method. The least squares method determines a and B

that minimize the sum of squared errors (SSE). The SSE is

the sum of the squares of the differences for each data point

between the model's prediction for the dependent variable and

the actual value of dependent variable:
SSE = (Y -Yi)2. (6-2)

The slope and intercept are calculated by

B X a-Y B X, (6-3a,b)

where V is the average value of V for V=X or V=Y and
SSVW = (V-Vi)(W-W), (6-4)

with V=X and W=Y for SSXY and with V=X and W=X for SSXX. A

measure of how well the model fits the data is the

correlation coefficient, r [128]:

r = B SSY' (6-5)

The SSYY (defined by Equation 6-4 with V=Y and W=Y) is the

sum of the squares of the differences between the average

value of the dependent variable and the actual value of

dependent variable for each data point. When the model fits

the data (dependent variable) well, r is near 1 for positive

B and near -1 for negative B. If the model does not fit

data well, r is near zero.

For multivariate regression a typical model is

Y = B0 + B1 X1 + B2 X2 + ... + Bm Xm, (6-6)

where Y is the fit to the dependent variable Y, B0 is the

intercept, Bj are the coefficients or parameters, and X. are

independent variables or are made up of functions of the

independent variables, e.g., X2, X1*X, ln(X2), etc. The B.s

are also determined by the same least square method used in

linear regression. Equation 6-6 also can be written as

Y = 0 X + B1 X + + B X, (6-7)

where X0 is a dummy variable always equal to 1. The

inclusion of X0 is necessary when using matrix methods to

solve for the least-square coefficients. The B0 is therefore

also a coefficient, albeit the coefficient of an independent

variable whose value is always 1.

The most frequently used statistical quantity for

determining how well a multivariate regression model fits a

set of data is the coefficient of determination, R2 [128-


R =1 SST' (6-8)

2 2 2
where SST=SSYY. For linear regression (m=l), R =r An R
value of 0.500 is equivalent to an r value of 0.707, so R is

a more conservative measure of fitness. The correlation for

a model is statistically significant at the 5% level [130]


F (n m 1) R
Fm R= > F(m,n-m-1,5%) (6-9)
m (1 R2

R2 > F(m,n-m-1,5%) (6-10)
m F(m,n-m-1,5%) + (n-m-l)'

where m is the numbers of coefficients in the model excluding

the intercept, n is the number of data points, and F is a

statistical quantity, the F-value or the inverted beta (F)

distribution, and F(m,n-m-1,P%) is the critical F-value for a

P% significance level [129-130]. A P% significance level

indicates that there is only a P% likelihood that the data

ended up along the lines of the model by random chance.

Critical F-values and their associated minimum R2 are listed
2 2
in Table 6-3. When the R2 exceeds the minimum R for a 5%

significance level, the correlation is well-fit.

The R2 can be negative if the intercept (constant) term

is not included. Negative R s indicate that the fit is worse

than fitting data to the mean of the dependent variable.

Most statistical books never consider models where there is

no intercept term. Some statistical qualities are no longer

valid when the intercept is excluded. The intercept is often

not counted among the coefficients of (or degrees of freedom

of) a multivariate regression model, as is the case for m in

Equations 6-9 and 6-10. The definition of R2 is based on

SST, which equals SSE when the model only contains the

intercept parameter. The R2 is therefore zero when the model

contains only the intercept parameter.

Table 6-3. Critical Values of t, F, and R2 at the 5%
Significance Level [128].

t(5%/2) F(5%) R2(5%)
DOF m=l m=2 m=3 m=4 m=5 m=l m=2 m=3 m=4 m=5
8 2.306 5.32 4.46 4.07 3.84 3.69 0.399 0.527 0.604 0.658 0.698
9 2.262 5.12 4.26 3.86 3.63 3.48 0.363 0.486 0.563 0.617 0.659
10 2.228 4.96 4.10 3.71 3.48 3.33 0.332 0.451 0.527 0.582 0.625
11 2.201 4.84 3.98 3.59 3.36 3.20 0.306 0.420 0.495 0.550 0.593
12 2.179 4.75 3.89 3.49 3.26 3.11 0.284 0.393 0.466 0.521 0.564
13 2.160 4.67 3.81 3.41 3.18 3.03 0.264 0.370 0.440 0.495 0.538
14 2.145 4.60 3.74 3.34 3.11 2.96 0.247 0.348 0.417 0.470 0.514
15 2.131 4.54 3.68 3.29 3.06 2.90 0.232 0.329 0.397 0.449 0.492
16 2.120 4.49 3.63 3.24 3.01 2.85 0.219 0.312 0.378 0.429 0.471
17 2.110 4.45 3.59 3.20 2.96 2.81 0.207 0.297 0.361 0.411 0.452

A coefficient is statistically significant [128] at the

5% level if the range defined by the coefficient (its t-

value times its standard error) does not include zero, i.e.,

there is a 95% or greater probability that the coefficient is

not zero. The t-value is from the Student's t test for 2.5%

(half of the 5%, indicating a two tailed () test) and the

number of degrees of freedom (DOFs) (number of data points

minus number of coefficients (including the intercept)). The

t-values for 5% significance level are listed in Table 6-3.

The t-values average around 2.2 for 8 to 17 DOFs.

If the t-ratio for a coefficient (the absolute value of

the coefficient divided by its standard error) is greater

than its t-value, then the range defined by the coefficient

(its t-value times its standard error) does not include zero,

and the coefficient is statistically significant at the 5%

level. The standard error of a coefficient is a measure of

how well the dependent variable is fit by all of the

independent variables compared to how well the independent

variable whose coefficient is being tested is fit by the

other independent variables. In other words, the standard

error partially answers the question, "does this independent

variable add any new information to the fit?". Statistically

significant coefficients are labelled well-defined.

Two strategies are useful in multivariate regression

analysis. One strategy is to start with a linear model and

add independent variables, one at a time, hoping to increase
2. The other strategy to to start with many independent
R The other strategy to to start with many independent

variables and throw out, one at a time, those whose

coefficients are least defined, until only well-defined

coefficients remain. Both strategies are used here in the

analysis of the CCTL ClHC data.

A computer program in BASIC was written to facilitate

linear regression analysis of the CCTL data. The program

reads a file containing the data shown in Table 6-2, with

each column headed by the compound's or temperature's name.

The user specifies a dependent variable and the independent

variables to which to fit the dependent variable. These

variables can be made up of any function of the compounds

and/or temperatures. The program determined the number of

valid data points (N), the number of degrees of freedom

(DOF), the coefficients by least-square calculation, the

coefficient of determination (R2), the standard error of the

dependent variable (S =ISSE/DOF), the standard error of the

coefficients, and the t-ratios of the coefficients. The user

can also specify that data points for a compound's highest or

zero values not be used. The program is interactive, and all

output is written to a file for later editing and printout.

The program can also plot the data and the fit to the

computer screen.

Effect of Chlorine Input on HC1 Emissions

Figure 6-1 shows a graph of HC1 output plotted against

added Cl input (from Table 6-1) where both the HC1 output and

added Cl input are in lb/hr and are normalized to a 400-lb/hr

HC1 out



C 0n



An A

put (lb/hr)


0.00 2.00 4.00 6.00 8.00 10.00 12.00

Cl added (Ib/hr)

Figure 6-1. HC1 emissions compared to added chlorine input
with both HC1 and added Cl normalized to a 400-
lb/hr waste feed rate.

waste feed rate. The normalization is necessary to present

the data in a 2-dimensional plot. The linear regression fit

shown on the graph is

HCl = 1.114 lb/hr + 0.7478 Cl, (6-11)

with a coefficient of determination, R, of 0.735 (or a

correlation coefficient, r, of 0.857). The 1.114 lb/hr

constant indicates a 0.2785% chlorine content in the waste.

The 0.7478 factor implies that not all of the chlorine ends

up as HC1, or at least not immediately. Chlorine does tend

to be a flame inhibitor [10-12], which would explain a slow

burn out. When the unnormalized data is subject to a

regression analysis with HC1 output as the dependent variable

Ej J^gmD


0 i i i i i I I i i i

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs