United States
Environmental Protection
Agency
Industrial Environmental Research?
Laboratory
Research Triangle Park NC 2771
Research and Development
EPA-600/S7-81-110 Aug. 1981
Project Summary
Mixing Phenomena in
Industrial Fume
Afterburner Systems
A. A. Putnam and H. A. Arbib
This report reviews the physical
mixing phenomena involved in the
reactions that occur in afterburners or
fume incinerators. Mixing the after-
burners is considered from three points
of view. First, the typical designs of
afterburner components that are in-
volved in the mixing phenomena are
covered. With the paucity of informa-
tion available on performance, there is
no clear-cut indication of the superi-
ority of any particular design. Second,
consideration is given to the possible
application of mathematical modeling
principles developed for studies of
conventional furnaces to afterburner
design. Although the problem for the
afterburner is basically simpler, prac-
tical application of mathematical
modeling still seems some time off.
Third, empirical relations available in
the literature for describing the per-
formance of jet flow systems similar
to various afterburner components are
presented. Use of these relations
permits an estimate of time-tempera-
ture histories of various flow paths
through different afterburner designs.
Overall, the design of adequate after-
burners from the mixing point of view
seems wellwithin the current state of
the art.
This Project Summary was devel-
opad by EPA's Industrial Environmen-
tal Rasaarch Laboratory, Research
Triangle Park, NC. to announce key
findings of the research project that Is
fully documented in a separate report
of the same title, (sea Project Report
ordering Information at back).
Introduction
The aim in an industrial afterburner is
to consume a pollutant down to a safe
lower level while balancing the eco-
nomics of auxiliary fuel requirements,
afterburner size and complexity, operat-
ing and maintenance costs, and meeting
any other specific requirement, such as
turndown capabilities. While the kinetics
of the specific reactions that are desired
will control minimum time-temperature-
concentration requirements, the mixing
patterns in the afterburners will be
critical in providing these conditions.
Furthermore, the complexities of provid-
ing mixing patterns of a desired size will
control the size of the afterburner, the
amount of auxiliary fuel actually re-
quired, and the pressure drop through
the afterburner. It is therefore clear that
a detailed consideration of the various
mixing phenomena that occur in indus-
trial afterburners could result in sig-
nificant advances in their design.
One might designate three character-
istic times in an afterburner system,
namely:
1. Chemical time,rc. For a first-order
reaction, rc is proportional to l/k
where k is the rate constant. For
other orders, the rate constant is
multiplied by an appropriate aver-
age concentration.
2. Mixing time, rd. The mixing time is
proportional to L2D, where L is a
characteristic linear dimension
and Dv is the effective diffusion
coefficient.
-------
3. Residence time, rr. The residence
time is proportional to L/V where
V is an average gas velocity.
One notes that the ratio of the mixing
time to residence time is proportional to
the Reynolds number, which is expected
to be high in the usual afterburner. This
means that mixing across the combustor
must be examined critically. While the
procedure for computing the flow and
mixing in nonreacting systems can be
quite satisfactory, where reaction sys-
tems are considered, the problem of
"unmixedness" enters. Unmixedness
was first analyzed in some depth by
using a Gaussian distribution as a basis
for computational procedure. A turbu-
lent diffusion flame is made of pockets
of unreacted fuel, air, unreacted mix-
ture, reacting mixture, and products of
combustion. A similar situation occurs
for a turbulent premixed flame. As a
result, a sampling at one point may
show both reaction products, and fuel,
oxygen, and nitrogen. In dealing through
mathematical modeling with the physi-
cal (as contrasted to the chemical)
aspects of combustion in a combustion
system, the phenomenon of unmixed-
ness must be added to the other flow
considerations. However, it is well
known that the jets and vortex flow sys-
tems that make up the usual combustion
system do not show a Gaussian-type
distribution on the scale of interest.
Quite common in the literature are
pictures and sketches of the rolling-up
vortices at the region of highest velocity
gradient, which normally is in the
highest reaction rate region.
The ratio of the chemical time to the
residence time is V/kL, the reciprocal
first Damkohler group. If the rate con-
stant is very high, or V/kL is low, the
reaction can be inhibited by lack of
proper mixing but not the chemistry.
This is the usual case in combustion
systems. However, in fume incinerators,
it is possible that the rate constant for
the fumes [not the fuel for the gas
burner(s)] is low and the performance
will be limited by both the reaction time
and the mixing time. As a result, an
optimum fume incinerator must be
designed for a particular class of fumes.
To complicate these chemical aspects
further, the presence of a flame is
important for contaminant removal.
Evidence indicates that, when using
electric heat energy, much higher tem-
peratures are required—1088 to 1255
K—to obtain the same efficiency
achieved with a direct-flame system at
810 to 1033 K. If satisfactory incinera-
tion is to be achieved at the lowest
possible temperature, the type of flame
and the design of the combustor are also
important factors to be considered. In
1980, studies of the self-ignition tem-
perature of kerosene sprays downstream
of a gas turbine combustor can were
reported in the literature. The tempera-
tures varied from about 500 K to 1300 K,
with an inverse correlation of the self-
ignition temperature with a measure of
the nonequilibrium ionization of the
igniting gas from the turbine. This
would indicate that the more intimately
and quickly the primary products of
combustion in a fume incinerator mixed
with the polluted secondary air, the
lower the temperature or the shorter the
time could be for complete combustion
of the fumes.
Estimates of the chemical reaction
time for various fumes are covered in
Reference 1, a parallel study to the
present one. In the present study, the
factors entering into the estimate of the
mixing time are reviewed. The review
covers three aspects of the mixing
problem, namely, a consideration of the
various direct flame afterburner systems
that are in present use, a consideration
of the possible mathematical modeling
of fume incineration systems, and a
review of the performance parameters
associated with the various components
of an afterburner.
Review of Direct Flame
Afterburner Systems
An effective direct flame afterburner
provides (a) contact between the con-
taminants in the air and the burner
flame, (b) time for the combustion
process to be completed, (c) sufficiently
high temperature for the complete oxi-
dation of the combustibles, and (d) flow
patterns that ensure adequate mixing
while preventing excessive quenching.
(It is common for legislation or regula-
tions to prescribe a pair of fixed values,
such as 0.3 sec for "b" and 1300°F for
"c," without regard to type of pollutant.
The analysis in Reference 1 shows this
is not an adequate approach.) To do this,
the contaminated gases are delivered to
the afterburner where they are mixed
with the burner flame or flames in the
upstream part of the unit, normally a
refractory-lines chamber. They then
pass through the remainder of the
chamber, where the combustion process
is completed prior to discharge to the
atmosphere. A typical fume afterburner
uses a mixing plate or grid burner. An
array of line burners, multijet burner, a
nozzle-mix burner, or a premix burner
could be used in the same general con-
figuration. Variations on the recuperator
from no recuperator to more extensive
ones can also be made. A catalytic
afterburner would have the catalyst
section inserted at the downstream end
of the combustion chamber, and the
combustion chamber would not run as
hot, thus saving on the required amount
of supplemental fuel.
Mathematical Modeling of
Turbulent Combustion
At present there are three main types
of mathematical models available for
predicting the performance of furnaces.
These can be classified according to the
dimensionality of the model. The sim-
plest model is the zero-dimensional,
well-stirred single-zone model. This
model has the great advantages of
simplicity and computer economy, and
is capable of predicting the overall
performance of a conventional furnace
surprisingly accurately. Thus, despite
various shortcomings, the well-stirred
model with well-established empirical
correction factors is used as the basis
for routine design of radiant sections of
boilers and fired heaters. Limitations of
the well-stirred model lie in the fact that
it does not predict spatial variations in
temperature and heat flux along a
furnace and it cannot predict the effect
of changes in flow pattern and flame
length. These limitations are critical to
any use in fume reactor design.
The second class of models are one-
dimensional, such as the long furnace
model in which the gas velocity and
temperature transverse to the flow
direction are assumed to be uniform and
axial radiation is neglected. This model
is economical and is capable of predict-
ing axial variations of temperature and
heat flux. It can deal with changes in
flame length but not recirculation of
gases within the furnace. When the
recirculation zone is of limited extent, a
combination of well-stirred and plug
flow models can be used successfully.
This modeling approach is applicable to
certain fume reactor designs.
The third class of models are multi-
dimensional. Zone and multiflux meth-
ods are available for solving that part of
the calculation concerned with radiative
heat transfer. However, these must be
coupled with some method for calculat-
ing the flow and heat release. In prin-
ciple, this can be done by solving the
governing equations for conservation of
mass, momentum, and energy using
-------
(say) a finite-difference solution proce-
dure. However, these methods require
considerable expertise and computer
time, and often produce unreliable
predictions. The unreliable results
probably arise from the need for various
turbulent mixing coefficients which
must be evaluated by trial and error
comparison with data on a similar
system. For these reasons they are used
rarely (if at all) for design of industrial
furnaces.
In time (about 1985) these methods
are expected to become more reliable
and gain wider acceptance in furnace
design. Because of some simplifying
factors, practical application of this
approach may be even sooner in the
fume incinerator field.
Jetlike Components of
Afterburners
Examination of afterburner compo-
nents reveals that many of them can be
classified as resulting in Jetlike action in
the flow field. Looking at the compo-
nents as regions in the flow field, one
could consider the regions as jet flow
regions. For instance, discrete source
)urners can be treated as hot gas jets
interacting with surrounding fume
laden air. While the simple jet in an
infinite environment of identical prop-
erties to the jet has been discussed in
many reports, some consideration of the
jetlike flames in an afterburner shows
that many variations in the jet and
environmental flow patterns can lead to
significant differences in results. On the
other hand, the smaller variations in
temperature and reaction rate in the
combustion chamber section of an
afterburner, as contrasted to normally
expected combustor rates in the com-
bustion region of a furnace, permit the
use of the simpler nonreacting jet
results in analysis. In this report,
emphasis is placed on "cold jet" results
that are applicable to afterburner design.
For orientation. Table 1 lists several
Bets of type variations that one might
encounter in a combustion system. For
a given jet, a type must be specified from
each set; this leads to a large number of
variations in jets even without consider-
ation of details. For the present study,
several examples in the literature that
are pertinent to afterburners are cited.
Conclusions
This report on afterburner mixing
jhenomena is a companion to an earlier
report. Reference 1, which discusMd
Table 1. Jet-Type Variations Classified in Sets
a. Geometry of jet exit
1. Axially symmetric
2. Two-dimensional
3. Rectangular
4. Wall
b. Environmental velocity
1. Zero
2. Coaxial positive
3. Coaxial negative
c. Environmental properties
1. Temperature and composition same as in jet
2. Temperature different
3. Composition different
4. Temperature and composition different
d. Number of jets
1. Single jet
2. Multiple parallel jets in line
3. Multiple parallel jets in pattern
4. Multiple nonparallel jets
e. Relation of jet to wall
1. Nonimpinging jet
2. Jet impinging at 90°
3. Jet impinging at other than 90°
f. Environmental geometry
1. Jet in open
2. Jet in enclosure without recirculation
3. Jet in enclosure with recirculation
g. Swirl condition
1. Zero swirl
2. Low swirl
3. High swirl
h. Buoyancy
1. No buoyancy effect
2. Positive buoyancy parallel to jet direction
3. Negative buoyancy parallel to jet direction
4. Nonparallel buoyancy effects
5. Radial
6. Annular
7. Other
4. 90° cross flow
5. Non 90° cross flow
chemical aspects of afterburner sys-
tems. The report on chemical kinetics
brought together information for esti-
mating time-temperature-composition
relations necessary to carry out the
oxidation of various organic fumes. In
the present report, sufficient informa-
tion is given to permit the estimation of
comparative values for various fluid
dynamic systems that can be used as
afterburners. These values can then be
compared to the requirements based on
Reference 1 to determine if a design is
acceptable. The report is divided into
three phases, namely,.a review of direct
flame afterburner systems, a considera-
tion of the possibilities of using mathe-
matical modeling in designing such
systems, and a review of the empirical
relations that can be used in the de-
scription of the mixing associated with
various jet-like components of an after-
burner.
Direct flame afterburner systems may
be considered as composed of a fume
polluted air source, a recuperator sec-
tion (usually), an approach section to the
burner section, a burner section, a com-
bustion chamber section, possibly a
catalytic section, the other side of the
recuperator section (usually), and an
exhaust. In this study, the sections
involved in the mixing process related to
the incineration of the fumes, namely,
the approach section, the burner sec-
tion, and the combustion chamber sec-
tion, are considered. A large variety of
designs in these three sections are
found, with no clear-cut preference for
any particular combination. This is
ft US. GOVERNMENT PRINTING OFFICE 1«61 -757-012/7296
-------
because of conflicting requirements
relative to capital cost, maintenance
cost, size, complexity, fuel consumption,
needed flexibility, and pollutant type, as
well as variations in code restrictions
and contract specifications. However,
well-documented field studies of actual
performance would undoubtedly elimi-
nate many of the available fume after-
burner systems as viable products.
Mathematical modeling of combus-
tion systems is receiving great attention
in the current literature. Therefore, the
pertinence of these studies to the
design of practical fume afterburners is
considered in some detail. It is pointed
out that mathematical modeling is only
one of many possible design approaches.
In the proper formulation of the pertinent
equations, the character of the turbu-
lence must be understood before suitable
approximation can be made in mathe-
matically specifying the turbulence at
any point in a combustion system. Even
then, the mathematical models contain
arbitrary coefficients based on the fit of
analytical results to experimental re-
sults for a particular experimental flow
system. For best results, then, the
particular system on which the coef-
ficients are based should resemble the
afterburner design of interest. Compar-
ing the complexity of the typical after-
burner with the simplicity of most ex-
perimental systems used in connection
with specifying the parameters of a
mathematical model, some pitfalls of
this design approach become apparent.
On the other hand, because the flow
and reaction region of interest in the
primary fume incineration process does
not involve the high fuel concentration,
high reaction rate, and high tempera-
tures of the more usual types of com-
bustion systems, simplifications are
possible in the application of mathemat-
ical modeling to fume incinerator design.
It is concluded that the practical use of
mathematical modeling for fume after-
burner designs is at least 5 years away.
For the same reason that there ap-
pears to be a possible simplification in
the use of mathematical modeling,
there also appears to be a possible
simplification in using literature on
various forms of jets to estimate the
time-temperature-composition profiles
as the hot combustion gases are mixed
into the fume-laden secondary air flow.
From such information, one can check
whether the time-temperature-compo-
sition requirements predicted for a
specific fume are met. Because of the
low reaction rate, "cold jet" data maybe
used rather than having to invoke the
complications of intense combustion
process taking place in the jet flow. This
opens up a wide range of applicable dat
in the literature, with rather simpl
relationships available for predictin
the mixing features of a jet and it
surroundings. Specifically, inthisstudi
the following types of turbulent jets ar
reviewed—circular free jet, plane fre
jet, free compound jet, enclosed com
pound jet, circular jet with swirl, jets i
cross flow, and impinging jets. Th
information in this literature is believe
to be sufficient to permit the calculatio
of adequate time-temperature-compc
sition curves for most reasonable fum
afterburner designs.
Reference
1. Barnes, R.H., M.J. Saxton, R.E
Barrett, and A. Levy. Chemica
Aspects of Afterburner System:
EPA-600/7-79-096 (NTIS PB 298465
April 1979.
A. A. Putnam and H. A. Arbibare with Battelle-Columbus Laboratories, 505 King
Avenue, Columbus, OH 43201.
John H. Wasser is the EPA Project Officer (see below).
The complete report, entitled "Mixing Phenomena in Industrial Fume After-
burner Systems," (Order No. PB 81 -222 259; Cost: $11.00, subject to change)
will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
FS 0000329
U S E^VIR PROTbCi'IUN
HEGIOM 5 LIBRARY
230 S DEAPBQRN
CHICAGO II. 60604
------- |