HIIMRIIIII
PB95-232864
NTIS
hfiforffMtfmi is our 'lnHrtyiiWi*
FUNDAMENTAL STUDIES ON THE CHARACTERIZATION
AND FAILURE MODES OF INCINERATOR
AFTERBURNERS
(U.S.) ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC
1995
U.S. DEPARTMENT OF COMMERCE
National Technic at Information Service
-------
BIBLIOGRAPHIC INFORMATION
PB95-232864
Report Nos: EPA/600/A-95/089
Title: Fundamental Studies on the Characterization and Failure Modes of Incinerator
ATterourners.
Date: 1995
Authors: C. Bass. R. Barat. G. Sacchl. and P. M. Lemieux.
Performing Organization: Env1ronmental Protection Agency, Research Triangle Park, NC.
Air ana tnerqy tngmeering Research Lab.**New Jersey Inst, of Tech.. Newark. Dept. of
Chemical Engineering. Chemistry, and Environmental Science.^Massachusetts Inst, of
Tech.. Cambridge. Dept. of Chemical Engineering.
Type of Report and Period Covered: Rept. for Mar-Dec 94.
~ " Presented at the International Incineration Conference. Seattle.
NTIS Field/Group Codes: 68A (A1r Pollution & Control). 81A (Combustion & Ignition).
m (Industrla I Oiertlstry & Chemical Process Engineering)
Price: PC A03/MF A01
Availability: Available from the National Technical Information Service. Springfield.
VA. iffilbl
Number of Pages: 15p
Keywords: *Incinerators. *A1r pollution sampling. *Combustion efficiency. *Failure
moaes. neprints. Afterburners. Air pollution control equipment. Hydrocarbons, Chemical
analysis. Gas analysis. Gas chromatography. Combustion products. combustion kinetics.
Products of Incomplete combustion.
Abstract: The paper describes a combined approach of bench and pilot scale experiments
wat is revealing new insights into the complex interactions between chemical reaction
and mixing which affect product of incomplete combustion (PIC) emissions and secondary
combustion chamber (SCO performance. The bench scale reactor allows well defined
experiments which are relatively easy to control and inexpensive. The pilot scale
reactor allows experiments under near-commercial scale SCC conditions. Using Its
volatile organlcs sampling train (VOST)-Hke sampling and gas chromatography (GC)
systems. combustion products are analyzed with high sensitivity, but at low costs and
snort turnaround times.
-------
EPA/600/A-S5/089
FUNDAMENTAL STUDIES ON THE CHARACTERIZATION
AND FAILURE MODES OF INCINERATOR AFTERBURNERS
Charles Bass and Robert Barat
Dept. of Chemical Engineering, Chemistry, and Environmental Science
New Jersey Institute of Technology
Newark, NI 07102
Guido Sacchi
Department of Chemical Engineering
Massachusetts Institute of Technology
Cambridge. MA 02139
Paul M. Lemieux
U.S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
Combustion Research Branch
Research Triangle Park, NC 27711
1995 International Incineration Conference
May 8-12, 1995 Seattle, WA
Abstract
The combined approach of bench and pilot scale experiments is revealing new
insights info the complex interactions between chemical reaction and mixing which
affect product of incomplete combustion (PIC) emissions and secondary combustion
chamber (SCC) performance. The bench scale reactor allows well defined
experiments which are relatively easy to control and are inexpensive. The pilot scale
reactur allows experiments under near-commercial scale SCC conditions. Using its
Volatile Organic* Sampling Train (VOST)-llke sampling and gas chromatography (GC)
systems, combustion products are analyzed with high sensitivity, but at low costs and
short turn-around times. A set of diagnostics that associates specific PICs with
selected failure modes is being developed and tested with the bench scale reactor.
These diagnostics identify the failure modes of poor alomization and poor mixing
energy. The combination of diagnostics and modeling will then be applied to the
pilot scale reactor when run under poor atoinization and low mixing energy
conditions. Linking diagnostics to failures modes has the potential to improve the
design and control of commercial SCCs.
-------
INTRODUCTION
The prevention of products of incomplete combustion (PICs) from
incinerator afterburners ia a major challenge, the solution of which
will enhance public acceptance of incineration. The formation of PICs
is highly dependent upon the local ratio of fuel and oxidant. Ifieir
amount and composition are sensitive to both turbulent mixing and
chemical kinetic constraints.
Incinerators that fail to adequately destroy the principal organic
hazardous component (POHC) or produce pollutants above regulatory levels
do so for a number of reasons. These include kinetic and thermodynamic
failures such as Improper stoichiomatry (too little or too much excess
oxygen), inadequate temperature, insufficient time to react, or mixing
failures. Mixing failures include incomplete mixing of the POHC with
surrounding oxidants. This can be caused by insufficient turbulent
mixing energy or the presence of large (rogue) droplets that move
through a large percentage of the reactor before combusting.
Another failure mode, identified by Trenholm et al. (1), is the reduced
destruction removal efficiency (DRE) as the feed POHC concentration
decreased. Wendt (2) suggested that these data might imply a
fundamental limit on our ability to reduce organic stack emissions. In
an attempt to explain these data, Bass et al. (3) performed computer
simulations with detailed chemical mechanisms (4) of afterburner
incineration of chlorinated hydrocarbons. Tbe results suggested that,
while macromixing is important, chemical kinetics in a premixed
simulation are insufficient to explain the low POHC failure mode (2) ,
Micromixing must also be included.
While such understanding is needed for environmentally safe incineration
of mixed chemical wastes, laboratory and modeling studies necessarily
must be limited to selected model wastes. An important class of wastes
which is the focus of our research is chlorinated hydrocarbons (ClHCs).
The multiple kinetic effects of ClHCs on hydrocarbon combustion are due
largely to the major role of the Cl atom in pyrolysis and combustion
(5,6). The rapid H atom abstractions by Cl accelerate the formation of
carbon-based radicals, enhancing the potential for molecular weight
growth and PIC formation (e.g., chlorobenzene) . The resulting hydrogen
chloride (HC1) competes with the burnout of carbon monoxide (CO) by
hydroxyl (OH) radical, yielding less heat — thus destabilizing flames
and increasing PIC potential.
Insufficient mixing energy can result in PIC formation which would not
occur in premixed or sufficiently mixed systems. Formation of aromatics
has been observed in a fuel-lean hot flue gas into which chloromethane
(CH3CI) had been injected under poor micromixing conditions (7) .
We have undertaken a research effort to characterize and model PIC
formation during various incinerator failure nodes. This effort involves
a three-way collaboration that'utilizes facilities at NJIT, MIT, and
EPA. Experiments include studies on both bench scale and pilot
incinerator simulators. A parallel modeling effort applies fundamental
chemical kinetic mechanisms to these combustor geometries using a
variety of mixing schemes. Our objective is to develop a set of
diagnostics to identify various failure modes and explain these failure
2
-------
modes in terms of kinetic and nixing limitations. Such a fundamental
understanding will lead to PIC control strategies.
STUDXff* ON THI BBNCH ICALI
Experiments with bench scale reactors offer the opportunity to study
selected phenomena under conditions much better controlled and defined
than with large scale devices. The two-stage turbulent flow bench
combustor at MIT has been used to study the effects of mixing
limitations on PIC formation.
A cross section of the two-stage atmospheric pressure combustor is shown
in Figure 1. The primary tone is a torus of 250 ml volume and 3.2 cm
minor diameter. Premixed fuel gas and air enter the torus from an outer
stainless steel circular manifold through a series of 32 jets. The jet
tubes (1 Bin 2D) are angled 20 degrees off radius. The high velocity
turbulent flow of each jet into the torus entrains surrounding bulk gas
to create a highly backmixed zone. Residence times are 5-10 ms in this
primary zone, which is well characterized as a perfectly stirred reactor
(PSR) under most conditions (8). A FSR is characterized by homogeneity
of composition and temperature.
The hot reactive gases exit the torus, passing over a flow straightener
into the secondary (linear) stage. Additional gases or vapors are
injected at this point. Second stage residence times are on the order of
15-20 ms in this 30 cm long, 5 cm ID secondary reactor. Further details
about the combustor are provided elsewhere (9) . The first stage ideally
simulates the flame/mixing chamber of a secondary combustion chamber
(SCC), while the second stage simulates the downstream SCO burnout zone.
Prevaporized material injected into the bulk flow enters in radial
cross-flow through eight jets at a single axial location (7). The gas
jet momentum is set to allow the injected material to penetrate through
the perpendicular main baseline cross flow to the reactor tube wall.
This scheme produces a macroscopically homogeneous fluid across the tube
diameter without significant micromixing (7) . The experiments performed
included injection of CH3CI into a hot baseline first stage exit flow
derived from fuel-lean or fuel-rich ethylene (C2H4)/air combustion. An
adjustable probe is inserted into the second stage for extraction of- gas
samples downstream of the injection point. These samples are then
analyzed by gas chromatography (GC) . Further details appear elsewhere
(10). This setup allows for studies of PIC formation under mixing-
limited conditions.
The research has suggested that mixing constraints decrease the extent
of reaction by delaying molecular mixing, as compared to a premixed
environment. Mixing constraints also cause a distribution of local
equivalence ratios + (+ - (fuel/air)actu«i/(fuel/air)¦toichioMtric) ««
compared to the theoretical value assuming perfect mixing. Some of the
high t values result in pyrolytic pockets which lead to the formation of
species which would not otherwise be formed in a premixed or perfectly
mixed case. In addition, for those species which have a strong non-
linear dependence on +, these high + values can lead to differences in
concentrations as compared to the completely mixed flow field.
Several parameters have been Identified as diagnostics of failure modes
for practical incinerators which typically run fuel-lean: ICO - C0oJ.
ICO - COoJ/COq, F/ICO - C0o), C/[CO - C0o], F/(C*F] where CO - measured
3
-------
CO concentration, COQ « calculated CO concentration under perfect
mixing, F ¦ unburned fuel, and C = byproduct hydrocarbon* (not including
F) . The matrix below illustrates how these quantities serve as failure
mode indicators.
[C0-C0o]/C0o
F/[CO-COo]
C/[CO-COo]
F/[C+F]
Indication
Low
High
Low
High
Poor atomization
Moderate
Low
High
Low
Moderately poor
mixing — no benzene
High
Moderate
High
Low
Extremely poor
mixing - benzene
For example, high values of F/ [C+FJ and F/(CO - C0Q] indicate that
unburned fuel is present, suggesting poor atoipization at the fuel
injector.
Results have shown that, for an overall fuel-lean system, the detection
of benzene is a marked indicator of poor mixing. The critical $ for
benzene formation in a fuel-rich, premixed environment is 2.2. Hence, $
fluctuations due to poor mixing can result in localized $ values in
excess of 2.2. High values of the ratios [CO - C0o)/C0o and C/ [CO -
C0ol, with or without the measured presence of benzene, will further
indicate the relative degree of poor mixing.
STUDIES ON TK8 LARGE PILOT SCALE
A critical component of our research effort to characterize PIC
formation and understand failure modes has been experimental and
modeling work with a large pilot scale rotary kiIn incinerator
afterburner. Efforts have centered on a reactor engineering model
characterization of the device, and the creation of a PIC and continuous
emission monitor database from C1HC combustion.
Experimental Apparatus
The EPA/AEERL's Rotary Kiln Incinerator Simulator (RKIS) is located
in the EPA Environmental Research Center hazardous waste incineration
research laboratory in Research Triangle Park, NC. Shown in Figure 2,
this facility has been described in detail elsewhere (11) . The
locations of sampling points (numbers) and dopant injection points
(letters) are noted.
The 73 kW prototype rotary kiln section contains a 61 cm long, 76.3 cm
diameter recess which contains the solid waste during incineration. For
this study, the kiln was operated with a natural gas/air feed only. From
the kiln section, the combustion gases enter the transition section.
The gases then flow into the experimental 73 kW SCC, which is the focus
of our studies. The SCC consists of three regions: the 61 cm diameter x
71 cm long mixing chamber, the 61 cm diameter x 178 cm long linear flow
section, and the stack transition section. A replaceable choke
separates the mixing chamber from the linear flow section. Currently,
the installed choke has a 15.24 cm diameter. A conical refractory
insert has been installed into the first plug flow section to provide a
4
-------
gradual divergence from the choke diameter to the linear flow nection
diameter and minimize recirculation zones downstream of the choke. The
afterburner provides heat and flame to this SCC. It is fired by natural
gas, and is capable of operating on air or oxygen-enriched air. For
these tests, air was the only oxidizer used.
Nitrogen-pressurized liquid surrogate waste is injected at point B
through an air atomizing nozzle. The two liquid surrogate wastes in this
study were Carbon Tetrachloride (CCI4) and Dichloromethane (CH2CI2)•
The RXIS is equipped with gas analysis and data acquisition systems.
Trace volatile organics are measured by on-line gas chromatography (GC).
Oxygen (O2), carbon monoxide (CO), carbon dioxide (CO2) nitric oxide
(NO), and total hydrocarbons are determined by continuous emissions
monitors (CEMs) . Sampling probes are located at ports 1 and 5. This
setup allows for on-line, simultaneous monitoring of gas-phase species
at the exits of both the kiln and the SCC.
Based on GC analysis of preliminary runs, and GC/mass spectrometry (MS)
analysis of several samples collected using the EPA-standard Volatile
Organics Sampling Train (VOST) method, a list of target analytes h
-------
column effluent is split to deliver sample to both detectors
simultaneously. Electronic analog integrators relate peak areas to
sample concentration. Sample concentrations were calculated based on
response factors derived from multipoint calibration standards
measurements.
The experiments were designed to examine differences in SCC
effectiveness under system failure mode conditions at three sample
locations (residence times) within the SCC and at three operating
conditions of the afterburner, as defined by its This particular set
of experiments concentrated on the potential failure mode of poor
atomization by varying the pressure of the atomizing air in the liquid
waste feeder. The test matrix is shown in Table 2.
TABLE 2: TEST MATRIX
Sample Location
3 4 5
Atomization Pressure and Dopant
High
Carbon Tetrachloride
S
R, L
R, L, S
Dichloromethane
S
R, L
R, L
Carbon Tetrachloride
s
R, L
R, L, S
s
R, L
R,L,S
S - afterburner operated stoichiometric, $ • 1
R - afterburner operated fuel-rich, $ > 1
L - afterburner operated fuel-lean, $ < 1
The mean droplet size of the atomized spray was altered to examine the
effects of rogue droplets passing through the SCC. Two extremes of the
atomization air were used: high <274 kPa) and low pressure (170 kPa) .
Throughout all of the tests, the primary natural gas burner was operated
at a constant 73 kW firing rate at near stoichiometric conditions. The
afterburner was operated at a constant 51 kW resulting in temperatures
of approximately 1000 °C throughout the system. It was desired that the
kiln be operated on low excess air so that significant combustion would
not be occurring in the transition duct due to excess oxygen from the
kiln effluent. The gases leaving the kiln had an oxygen concentration
of approximately 1% by volume. The combustion conditions in the SCC
were changed by varying the combustion air going into the afterburner.
In the tests, the dopant was injected at point B. Samples were
collected from the SCC at point 3, 4, or 5.
Results from C1HC Doping
The concentration data were put through a linear regression using the
JMP® Stepwise algorithm, considering $, the residence time, their
squares, their linear cross terms, and the dopant. The Stepwise
algorithm was then used to determine which predictors are statistically
significant. This methodology can determine which of the experimental
variables can be used to account for the variances in the data. Table 3
lists the significant predictors that resulted from the regression
analysis, plus the value.
6
-------
There are several Interesting observations to note about the regression
analysis. First, benzene emissions (see Figure 3) were accounted for
with correlation coefficient of 0.919S by considering only the
afterburner $, independent of what dopant was being fired at the time.
Toluene, on the other hand, gave a very high R2 value of 0.9362, but all
predictors were significant. Trichloroethylene (TCE) provided a fairly
Table .3.1 Regression Analysis of SCC Data
Compound
Ir
Most Significant Pfffdlctpn
Benzene 0.9195
Carbon Tetrachloride 0.6781
Chlorobenzene 0.5356
Chloroform 0.5482
1.2 Dichlorobenzene 0.4213
1.3 Dichlorobenzene 0.2429
1.4 Dichlorobenzene 0.4288
1,2 cis-Dichloroethane 0.3310
1,2trans-Dichloroethane 0.7989
Dichloromethane 0.7376
Ethyl Benzene 0.8203
Methyl Ethyl Ketone 0.8221
Perchloroethylene 0.6598
Styrene 0.2226
Toluene 0.9362
1,1,1 Trichloroethane 0.0769
Trichloroethylene 0.6277
o-Xylene 0.8637
m,p-Xylenes 0 .6691
RT, dopant**
all****
dopant*, RT
all****
none
RT, dopant****
~
all****
all
~ , RT
all****
all
none
all
none
all***
all****
RT
t • equivalence ratio, RT » residence tine
• No significant differences were seen between CCI4 and CH2CI2 as dopants, but the
presence/absence of chlorinated dopant was significant.
** Significance of dopant variable was between runs with CCI4 as dopant and runs
without" CCI4 as dopant
*** TCE was a low level contaminant in all samples
**** Significance of dopant variable was between runs with CHjClj and without CH2CI2
good correlation and all predictors were significant. Unfortunately,
TCE was found as a trace contaminant in all samples; hence, the TCE data
are of questionable quality. Hie investigators are endeavoring to
eliminate this contaminant so that a blank correction would not be
necessary, chlorobenzene, 1,2 dichlorobenzene, 1,2 dichloroethane,
methyl ethyl ketone, and o-xylene, were found to be significantly higher
only during runs where CH2CI2 was the dopant, indicating that CH2CI2 has
a higher propensity than CCI4 to form intermediates that promote
molecular growth.
Characterization of the SCC
The modeling of PIC formation in the SCC requires use of a fundamental
chemical mechanism with hundreds of elementary reactions and many dozens
of species. Ideal reactors provide a convenient mixing model for
simulating the combustor while using detailed mechanisms.
In the plug flow reactor (PFR), which assumes no axial mixing,
composition varies axially, but is constant across the flow path at any
7
-------
given axial location. The PFR is solved by Integrating a set of
ordinary differential equations from an Initial condition. As stated
earlier, the PSR is characterized by homogeneous compositions and
temperature throughout the entire zone. The PSR 1® solved by finding
the solution of a set of non-linear algebraic equations.
Software packages such as CHEMKIN II (12) add to the convenience by
allowing formulation of these equations at run time. More complicated
reactive mixing models using finite difference techniques are
computationally expensive and become impractical to solve for chemical
mechanisms beyond a few dozen species even on modern supercomputers
(10) .
Unfortunately, a single ideal reactor cannot reflect the complexities of
turbulent mixing in reactive flows such as combustion. Instead, a
series of ideal reactors can be used, where each ideal reactor describes
a macromixing zone. Beer and Lee (13} suggested the use of a PSR
followed by a PFR to describe a simple combustor with*a swirling flame.
The PSR represents the strong backmixing in a swirling flame and the PFR
represents the downstream burn-off region. Beer and Lee proved the
validity of this model using helium tracers to determine the residence
time distribution (RTD) of their combustor. The PSR-PFR model has"been
successfully applied in modeling a two-stage combustor by Mao et al.
(14). Swithenbank et al. (15) used a more complex multiparameter ideal
reactor model to describe combustion in a gas turbine.
Determining reasonable model parameters is the main drawback to
increasing the complexity of an ideal reactor network. The result is
often a number of undetermined, adjustable parameters which reduce the
value of the model. Experimental tracer-based RTD data can be used to
resolve the model parameters. This process becomes increasingly
difficult with increasing complexity of the model.
The RTD represents the scalar transfer function of a linear system.
Thus it cannot even distinguish the order of ideal reactors -- a
distinction that is important once kinetics are applied. Further, the
RTD-based ideal reactor model represents a linear approach to non-linear
turbulent mixing. Despite these drawbacks, we chose this as the best
approach for incorporating detailed chemical kinetic mechanisms in _
modeling a practical system because of its relative computational ease
and successful use in the past.
Approach
Development of an ideal reactor SCC model began with experimental tracer
studies to determine the RTD. Chemically inert sulfur dioxide (SO2) gas
(15% SO2, balance N2) was injected in both stepwise and pulse modes into
the SCC at locations A, B, and C. Gas samples were drawn from locations
1 through S for analysis by an on-line ultraviolet SO2 analyzer as
functions of elapsed time.
Figure 4 highlights the mixing zones of the SCC. Kiln gas enters the
mixing chamber tangentially. A swirled natural gas/air afterburner
fires on the axis of the chamber. The combustion products pass through
a choke and expand through a funnel to the linear flow section. In the
mixing chamber two distinct mixing zones are apparents the backmixing
region of the swirl burner and the tangential mixing region of the
entering kiln gas. The distinction between the two zones was physically
8
-------
apparent by the radial temperature profile (taken horizontally at raid
chamber) and (rem the data f'jich showed no tracer sampled from the
tangential region when injs^ted into the SCC burner. Further analysis
of the RTD data suggested the need of adding two additional mixing
regions: an entrainmcnt zone around the flame and a dissipation zone
immediately beyond the choke.
M
-------
ACRMOHLBDOEMBNTi This work is supported fay the Northeast Hazardous
Substances Research Center and by the U.S. EPA, ABERL.
1. Trenholm, A., Gorman, P., Jungclaus, O., U. S. Environmental
Protection Agency Report EPA-600/2-84-181a (NTIS PB85-129500) (1984).
2. Wendt, J. O. L.,Twenty Fifth Synpoaium (Int.) on Combustion, The
Combustion Institute, Pittsburgh, PA (1994).
3. Bass, C., Barat, R. B.» and Lemieux, P., Twenty Fifth Synpoaium
(Int.) on Combustion, Combustion Institute, Pittsburgh, PA (1994).
4. Chiang, H., Park, B., und Bozselli, J. *»., Twenty Fifth Symposium
(Int.) on CombustIon, Combustion Institute, Pittsburgh, PA (1994).
5. Ho, W. P., Barat, R B., and Bozzelli, J. W., Combustion and Flame,
vol. 88, p. 26S (1992) .
C. Wang, L., Jalvy, P., and Barat, R. B., Combustion Science and
Technology, vol. 97, no. 1-3, p. 13 (1994).
7. Brouwer, J., Sacchi, G., Longwell, J. P., and Sarofim, A. F.,
Combustion and Flame, vol. 99, p. 231 (1994).
8. Barat, R. B., Combustion Science and Technology, vol. 84, no. 1-6,
p. 187 (1992).
9. Nenniger, J. E., Chomiak, J., Kridiotis, A., Longwell, J. P., and
Sarofim, A. F., Twentieth Symposium (Int.) on Combustion, p. 473",
The Combustion Institute, Pittsburgh, PA (1984).
10. Brouwer, J., Ph.D. Thesis, Department of Mechanical Engineering,
Massachusetts Institute of Technology, Cambridge, MA (1993).
11. Linak, W., Kilgroe, J., McSorley, J., Wendt, J., and Dunn, J., J. of
the Air Pollution Control Association, vol.37, p.54 (1987).
12. Kee, R. J., Rupley, F. M., and Miller, J. A., Sandia National
Laboratory Report SAND89-8009B UC-706 (1993).
13. Beer, J. M. and Lee, K. B., Tenth Symposium (International) on
Combustion, Combustion Institute, Pittsburgh, PA - poster (1965).
14. Mao, F., Kretkowski, D., and Barat, R. B., Combustion Science and
Technology•, vol. 102, no. 1-6, p. 145 (1994).
15. Swithenbank, J., Poll, I., and Vincent, M. W., Fourteenth Symposium
(Int.J on Combustion, Combustion Institute, Pittsburgh, PA (1974).
PROBE
VEH»
Number! - Sampling Points; Letters > Dopant Injection Pclnu
Secondary CemburtlMi Ctambw
End View
WATER
SPRAY
Dud from Win
FLOW
reED set men
JETS
INJECT OH
Vta* Port
S/./X // // //
n
00 o 11
i '""I
^
/
Kin Section TramMon S»dlon
Figure 1: Cross section of two-stage
bench scale combustor
Figure 2: Rotary kiln incinerator
simulator with SCC
10
-------
•
*
O
m
0 CCI4
m
~ CH2CI2
~
O Blank
1K>
K *
0.80 0.90
1.00
4>
1.10 1.20
Fig. 3: SCO Benzene emissions
11
-------
Zone t; Swirling Flame Zone 3; Pntrainment
\ ||_ _ fifi r al / Choke
3 3 31.0 cm
Zone 2: Tangential Mixing
Zone 4; Dissipation
Fig, 4: Detail of SCC mixing zones
Mixing Chamber
PFR 2 (Stack.
lwg« HmM Loss]
PFR t (Small Hut loss)
CcV
rw iii\ ,
8yp» Way
t-—< C<"
Bypass Fraction 0
Fig. 5: SCC ideal reactor model
12
-------
bypass 1
1
bypass 2
0,4 OS
Bypass Fraction
3.0E-08
2.5E-0®
bypass 1
2.0E-06
1.SE-06
1.0E-06
0.5E-06
bypass 2
0.00025
Oj 0.00020
0.00015
% 0.00010
0.00005
0.4 0.6
Bypass Fraction
bypass 1
bypass 2
0.2 0.4 0.8
Bypass Fraction
0.8
Fig. 6: Selected simulated PIC emissions from
SCC ideal reactor model [Indicated
bypass fraction varied while other
held at 0.5; f = 1.24; POHC is CH2CI2;
CH4/air afterburner flame]
13
-------
AEERL-P-1267 /iNmr S!mm E^nwvftrt
i.RtTOHTNO. 2.
EPA/900/A-95/0S9
3.
4. TITLE AND SUBTITLE
Fundamental Studies on the Characterization and
Failure Modes of Incinerator Afterburners
6. REPORT DATE
6. PERFORMING ORGANIZATION COOS
». authors c.Bass and R. Barat (NJIT), G. Sacchi (MIT),
and P. M. Lemieux (EPA)
(, PERFORMING ORGANIZATION REPORT NO.
•. PERFORMINO ORGANIZATION NAME ANO AOORESS
New Jersey Institute of Massachusetts Institute of
Technology Technology
Dept. of Chem. Engrg Dept. of Chem. Engrg.
Newark. NJ 07102 Cambridge, MA 02139
10. program KlIMUnT N6.
1 i. CONYRAGT/dftANT WSl
NA (Inhouse)
12. SPONSORING AGENCY NAM! AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research TrianglePark, ifc 27711
IS. TYPE OP REPORT AND PERIOO COVERED
Published paper; 3-12/94
14. SPONSORING AGENCY COOE
EPA/600/13
iMUFfUMiNTARVNCTii AEERL project officer is Paul M. Lemieux. Mail Drop 65. 910/
541-0962. For 1995 International Incineration Conference, Seattle, WA, 5/8-12 '95.
is. abstract jhe paper describes a combined approach of bench and pilot scale experi-
ments that is revealing new insights into the complex interactions between chemical
reaction and mixing which affect product of incomplete combustion (PIC) emissions
and secondary combustion chamber (SCC) performance. The bench scale reactor al-
lows well defined experiments which are relatively easy to control and inexpensive.
The pilot scale reactor allows experiments under near- commercial scale SCC con-
ditions. Using its volatile organics sampling train (VOST)-like sampling and gas
chromatography (GC) systems, combustion products are analyzed with high sensi-
tivity. but at low costs and short turnaround times. A set of diagnostics that asso-
ciates specific PICs with selected failure modes is being developed and tested with
the bench scale reactor. These diagnostics identify the failure modes of poor atomi-
zation and poor mixing energy. The combination of diagnostics and modeling will
then be applied to the pilot scale reactor when run under poor atomization and low
mixing energy conditions. Linking diagnostics to failures modes has the potential to
improve the design and control of commercial SCCs.
IT. KEY WORDS AND DOCUMENT ANALYSIS
». DESCRIPTORS
b.lOENTIPIERS/OPEN ENDED TERMS
c. cosati FkM/Growp
Pollution Sampling
Incinerators Gas Chromatography
Combustion Atomizing
Emission
Organic Compounds
Volatility
Pollution Control
Stationary Sources
Afterburners
13 B 14B
07D
21B 13H, 07A
14G
07C
20M
18. DISTRIBUTION STATEMENT
Release to Public
1®. SECURITY CLASS (Thlt Htport)
Unclassified
21. NO. OF PAGES
13
20. SECURITY CLASS (Thtipaf )
Unclassified
22. PRICE
CPA Worm 2220-1 <•-»«
------- |