Combustion Science and Technology

ISSN: 0010-2202 (Print) 1563-521X (Online) Journal homepage: http://www.tandfonline.com/loi/gcst20

COMPUTATIONAL AND EXPERIMENTAL STUDY OF A

TWO-SECTION POROUS BURNER

MATTHEW T. SMUCKER & JANET L. ELLZEY

To cite this article: MATTHEW T. SMUCKER & JANET L. ELLZEY (2004) COMPUTATIONAL
AND EXPERIMENTAL STUDY OF A TWO-SECTION POROUS BURNER, Combustion Science
and Technology, 176:8, 1171-1189, DOI: 10.1080/00102200490457385

To link to this article: https://doi.org/10.1080/00102200490457385

Published online: 11 Aug 2010.

Submit your article to this journal

Article views: 232

View related articles

Citing articles: 51 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=gcst20
Combust. Sci. andTech., 176: 1171^1189, 2004
Copyright # Taylor & Francis Inc.
ISSN: 0010-2203 print/1563-521X online
DOI: 10.1080/00102200490457385

COMPUTATIONAL AND EXPERIMENTAL STUDY

OF A TWO-SECTION POROUS BURNER

MATTHEW T. SMUCKER AND JANET L. ELLZEY*

Department of Mechanical Engineering and the Applied
Research Laboratories, University of Texas at Austin,
Austin,Texas, USA

In this study, experiments and computations were conducted on a two-section

porous burner operated on propane=air and methane=air mixtures. The
burner consisted of an upstream section of reticulated yttria-stabilized
zirconia with 23.6 pores per centimeter (ppc) and a downstream section of
3.9 ppc. Measurements of axial and radial temperatures, pressure drop, and
emissions levels were recorded. The predictions from a one-dimensional
transient mathematical model with full chemistry were compared to experi-
mental results. Both computations and experiments showed that the stable
operating range increases with equivalence ratio. The predicted upper limit
agrees well with experiments but the lower limit is somewhat overpredicted.
The average temperature in the exhaust stream increases with both inlet
velocity and equivalence ratio and is relatively uniform across the burner.
Pressure drop is much greater for reacting flows than cold flow and generally
increases with inlet velocity. Measured levels of unburned hydrocarbons,
oxides of nitrogen, and carbon monoxide are low.

Keywords: porous burner, emissions, premixed flame, excess enthalpy

Received 30 May 2003; accepted 22 December 2003.

This work was supported by the Office of Naval Research through Dr. Gabriel Roy
and also by the Independent Research and Development Program at the Applied Research
Laboratories, University of Texas at Austin. The authors would also like to acknowledge
Porvair Corporation for supplying the porous ceramics for these experiments. They also
thank Mr. John Witton of Cranfield University for his helpful discussions.
*Address correspondence to [email protected]

1171
1172 M. T. SMUCKER AND J. L. ELLZEY

INTRODUCTION
Weinberg (1971) was one of the first researchers to suggest that heat
recirculation without dilution in a burner was the way to both increase
efficiency and reduce emissions. Such a burner made use of ‘‘borrowed’’
heat to increase the temperature and, thus, reaction rates in the flame.
Although no experiments were completed at that time, a design incor-
porating a nonmixing counterflow heat exchanger was suggested as a
possibility to achieve the heat recirculation.
The idea of inserting a porous solid into the flame zone of a burner was
proposed by Takeno and Sato (1979a). The porous matrix was modeled as
a semi-infinite solid through which the flame was allowed to propagate as
fuel and air mass flow were increased. They evaluated the idea using a
mathematical model that included one-step chemistry and convection and
conduction heat transfer, but neglected radiation. This work was followed
soon after by a second study (Takeno and Sato, 1979b), which extended the
computational model to account for heat transfer to the surroundings and
included an experimental study. In the experimental study, the porous solid
inserted into the flow was a sintered iron block with many small, straight
channels through which the gas flowed. Premixed methane and air were
preheated by this block and resulted in flame stabilization just downstream
of the block exit. This technique extended the lean flammability limit
slightly, but its effects were limited due to oxide buildup on the iron block.
More experimental work followed this early attempt. Echigo et al.
(1986) constructed a burner out of three separate layers of porous media
made from stainless steel mesh. The first and third layers were thin and
more densely packed, while the middle layer was much thicker and more
porous. The porosity of the middle layer was varied during the study and
was found to affect the stable range of the burner.
The potential applications for porous burners as radiant heat sources
in industrial applications was recognized and dictated the direction of
research for a time. Reticulated foams were used in the next generation of
burners and the open-cell structure increased radial mixing and heat
transfer. Sathe et al. (1990) made use of lithium aluminum silicate
ceramic foam for a methane-fueled one-section porous radiant burner. It
was discovered that radiant output and flame speeds were highest for a
flame located in the center of their burner.
Hsu et al. (1993) studied a burner consisting of two sections of
partially stabilized zirconia (PSZ) reticulated ceramic. The upstream
TWO-SECTION POROUS BURNER 1173

section was a small-pore PSZ foam and various sizes of larger-pore PSZ
were tested as the downstream section. In this design, the upstream
section was defined as a preheating region and the downstream section as
the stable burning region. In this configuration, the flame stabilized in the
downstream section.
An important reason for the low emissions produced by porous
burners is that their extended stable operating region allows them to
operate at lean equivalence ratios. Khanna et al. (1994) conducted a
study on a two-section methane porous burner with 25.6 pores per
centimeter (ppc) PSZ ceramic matrix in the upstream section and 3.9 ppc
PSZ in the downstream section. Levels of carbon monoxide (CO) and
oxides of nitrogen (NOx) were measured at the exit of the burner and the
results showed that the levels for both species decreased with decreasing
equivalence ratio. Also, radiant output was measured and found to
increase with inlet flow velocity.
Similar results were documented by Trimis and Durst (1996). They
conducted experiments on a porous burner that consisted of two sections
of packed spheres with differing diameters. This research showed that
combustion was stable for a wide range of air=fuel ratios, emissions were
low, and the burner had a high-power dynamic range of 20:1.
Mital et al. (1997) studied a much thinner two-section porous burner
made of cordierite with a silicon carbide coating. The results from this
study determined that CO, unburned hydrocarbon (UHC), and NOx
levels decreased with decreasing equivalence ratio, but CO and UHC
levels decreased with increasing firing rate while NOx levels increased.
Porous media material properties impose limitations on burner
operation. Because of this, a number of porous materials have been tested
for use in porous burners. Many early burners were made from PSZ.
Later in a study by Mathis and Ellzey (2003), zirconia-toughened mullite
(ZTM) and yttria-stabilized zirconia (YZA) were evaluated for use in a
two-section porous burner in an effort to improve on the performance of
PSZ burners. Of these materials, YZA was found to have the material
properties that resulted in the largest burner operating range; thus, fur-
ther work was conducted with this material. Also in this work, the
upstream (small-pore) section of the burner was shortened with only a
small effect on the stable operating range. Emissions measurements were
made during this study and levels were low for all conditions tested.
Models of the combustion within porous burners have steadily
become more sophisticated as computational capabilities have improved.
1174 M. T. SMUCKER AND J. L. ELLZEY

Sathe et al. (1990) used a one-dimensional steady-state model with a

one-step chemical mechanism and found that they had good agreement in
the trends of flame speed and radiative output shown from experimental
data. Rumminger et al. (1996) modeled a methane-fueled porous burner
with a multistep mechanism using a one-dimensional approach. They
concluded that almost all of the methane is consumed within the porous
medium and that NOx compounds are formed near the flame front. Pre-
dicted temperatures were higher than those measured by their experiments.
Barra et al. (2003b) completed a computational study of a
two-section porous burner using a one-dimensional transient approach
that included complete chemistry for methane combustion. Results
showed that heat recirculation efficiency decreased with increasing
equivalence ratio. Also, their analysis showed that the heat recirculation
to the incoming fuel=air mixture was dominated by either solid conduc-
tion or radiation, depending on the equivalence ratio and inlet flow
velocity. In later work, Barra (2003) computationally investigated the
effects of heat loss on the combustion stability range and found that
when heat losses were included, the stable velocity limits decreased as well
as the range between them.
Reviews of the state of research in porous burner technology (Howell
et al., 1996; Viskanta and Gore, 2000) summarized recent findings and
pointed out some areas where continued research was necessary. These
areas included fluid mechanics within porous media, porous media
property data, and continued experimentation and comparison to porous
burner numerical models.
The present research used a mathematical model with full propane
chemistry to predict the stable operating range, and the size of the
experimental burner was increased in an effort to reduce heat losses per
unit area. For comparison, experimental and computational results were
obtained for operation on methane. In building on the experimental work
done by Mathis and Ellzey (2003), important additional measurements
were made, including pressure drop across the burner for cold and
reacting flow conditions and exhaust temperature measurements.

EXPERIMENTAL APPARATUS
The experimental apparatus is shown Figure 1. The burner in this study
consisted of two cylindrical sections of YZA ceramic foam (each 10.16 cm
in diameter and 5.08 cm in height) with different pore diameters. YZA
TWO-SECTION POROUS BURNER 1175

Figure 1. Diagram of the experimental apparatus.

ceramic foam was used in this research because it has been shown to have
the most favorable material properties of the materials available for
experimentation (Mathis and Ellzey, 2003). The upstream section had
23.6 ppc and the downstream section had 3.9 ppc. These sections were
inserted into the bottom half of tubular rigid alumina insulation, leaving
the upper half empty. In previous experimentation, similar two-section
burner designs have been shown to stabilize the flame over a wide range
of conditions (Khanna et al., 1994; Mathis and Ellzey, 2003).
The flow system in the experimental apparatus used independently
controlled metering systems for the fuel and air to supply a premixed
mixture for a wide range of equivalence ratios and flow velocities. The
fuel was CP grade propane or methane and the air was pressurized
building air filtered for oil, water, and particulates.
Pressure drop across the porous media was measured with a U-tube
manometer connected between a pressure tap just upstream of the PM
and the room atmosphere. The manometer had the capacity to measure a
pressure difference of 20 inches of water.
Temperature measurements during experimentation were taken using a
combination of K- and B-type thermocouples. For the axial temperature
measurement, the thermocouples were inserted into ports (ceramic tubes) in
the insulation until they reached the outer circumference of the porous
1176 M. T. SMUCKER AND J. L. ELLZEY

media. Nine of these thermocouple ports were spaced every 1.27 cm along
the axial direction of the burner at a fixed radial location, starting at the
bottom (upstream end) and finishing at the exit surface of the porous media.
For measurement of the exhaust gas temperature, five K-type
thermocouples were attached to a stainless steel ‘‘comb’’ with five tines,
each spaced 2.54 cm apart. This comb was then secured to a stand and
lowered into the exit of the burner, allowing temperature measurements
to be taken along a diameter of the burner at various locations in the
exhaust stream.
Emissions measurements in this research were taken using a set of three
analyzers from Rosemount Analytical: one each for UHC, NOx, and CO.
Exhaust samples from the burner were drawn out of the exhaust stream
through a quartz probe, dried, and supplied to the analyzers through a
network of stainless steel tubing. The total error in these measurements was
calculated as the sum of the calibration gas error, the round-off error, and
errors inherent to each analyzer, such as zero or span drift.
To take data at each test condition, the air and fuel flow rates were
adjusted to the appropriate settings and the flame was allowed to reach
steady state, as determined by the axial temperature profiles. This often took
as long as an hour. When steady state was reached, the temperature profile
was recorded and the pressure drop and the emissions levels were measured.

COMPUTATIONAL METHOD
The computer code in this study is similar to that used by Barra et al.
(2003a), but with different geometry, material properties, and fuel
composition. It uses a one-dimensional transient approach that includes
the effects of solid- and gas-phase conduction, solid radiation, solid-to-
gas heat transfer, species diffusion, and full chemistry.
The following conservation equations for mass, gas energy, solid
energy, and gas species, are solved:

@ðrg eÞ @ðrg euÞ

þ ¼0 ð1Þ
@t @x

@Tg @Tg X @Tg X

rg Cg e þ rg Cg eu þ reYi Vi Cgi þe o_ i hi Wi
@t  @x  @x
@
 @T
g  
¼e kg þ rCp Ddjj
hv Tg
Ts ð2Þ
@x @x
TWO-SECTION POROUS BURNER 1177

@Ts @ 2 Ts   dqr
rs Cs ¼ ks 2
þ hv Tg
Ts
ð3Þ
@t @x dx

@Yi @Yi @
rg e þ rg eu þ ðreYi Vi Þ
eo_ i Wi ¼ 0 ð4Þ
@t @x @x
where e is porosity; r is gas density; u is gas velocity; t is time; Cg is
specific heat of the gas; Tg is temperature of the gas; x is distance; Yi, Vi,
Cgi, oi, hi, and Wi are the mass fraction, diffusion velocity, specific heat,
molar rate of production, molar enthalpy, and molecular weight of the
ith species, respectively; kg is gas thermal conductivity; hv is the volu-
metric heat transfer coefficient between the Porous Media and the gas; Ts,
Cs, and ks are the temperature, specific heat, and effective thermal con-
ductivity of the porous medium, respectively; and qr is radiant heat flux in
the x direction (Barra, 2003).
The chemical reaction mechanism for propane includes 70 species and
463 elementary reactions (Qin et al., 2000). This mechanism was built
upon the GRI 3.0 chemical kinetics mechanism (Smith et al., 1999),
adding elementary reactions for hydrocarbon compounds that contained
up to three carbon atoms. This work was done with the intention
of modeling propane (C3H8) chemistry. The mechanism was optimized by
comparison to published ignition delay and flame speed data and predicts
the laminar flame speed of propane within 5% (Qin et al., 2000).
There is a limited amount of information about the material prop-
erties of YZA ceramics available in the literature. Values were found for
the downstream section (Fu et al., 1998; Mital et al., 1996), but none were
available for the smaller-pore material of the upstream section. Because
these property values were not available, values for PSZ ( Younis and
Viskanta, 1993) and alumina (Hsu and Howell, 1992) of appropriate pore
size were used.
The material properties used in the computations are summarized in
Table 1. In the table, C and m are constants used in a correlation to
determine the volumetric Nusselt number for convective heat transfer
within the porous media:

Nuv ¼ C  Rem ð5Þ

where Re is the Reynolds number.
The initial input conditions were conditions thought to be near the
middle of the operating range for each equivalence ratio. After the
1178 M. T. SMUCKER AND J. L. ELLZEY

Table 1. Material property data used in computations

Upstream Downstream

Porous media type 26.5 ppc PSZ 3.9 ppc YZA

Pore diameter 0.029 cm 0.26 cm
Porosity 0.835 0.85
Extinction coefficient 17.07 cm71 1.05 cm71
Conductivity 0.2 W=mK 0.3 W=mK
Albedo 0.8 0.78
C 0.638 0.252
m 0.42 1.08

solution for these conditions was obtained, it was used as a restart for
higher- and lower-velocity conditions. In this way, solutions for lower
and higher inputs were obtained until unstable solutions were encoun-
tered. Stepping down in velocity, the computations eventually reached a
point where flashback occurred. At this point, the flame front propagated
farther upstream with each time step and exited the computational
domain. Stepping up in velocity, the velocity was eventually high enough
to result in blowoff. In this case, the flame propagated downstream with
each time step. As unstable conditions were encountered, the incremental
change in input velocity was decreased so that it was possible to specify
upper and lower limits within 1 cm=s.

RESULTS AND DISCUSSION

The stable operating range of a burner is defined as the range of
equivalence ratios and inlet velocities for which a flame stabilizes within
the downstream section of the burner. Outside the limits of the stable
operating range, the flame experiences blowoff, flashback, or quenching.
In theory, these are simple concepts. Blowoff occurs beyond the upper
limit of the burner. In this case, the inlet flow velocity of the unburned
mixture is greater than the effective flame speed of the mixture in
the porous media and the flame propagates downstream and out of the
burner. Flashback is the opposite of this situation. For flashback, the
fuel=air mixture enters the burner at a velocity lower than the effective
flame speed, and the flame propagates upstream. Quenching also occurs
at the lower stability limit and was observed in a number of instances. In
this case, when the inlet fuel=air velocity is lowered, the flame attempts to
TWO-SECTION POROUS BURNER 1179

flashback but then encounters the small-pore upstream section and can
no longer sustain itself due to rapid heat loss.
Although these concepts were simple in theory, they acquired new
complexity in practice. For example, blowoff during the experiments did
not occur as a distinct event but was a gradual process. This was because
the flame did not blow off of the entire burner surface at once. First, the
flame started to blow off at one location, causing a visible dark spot on
the burner, but the flame was still able to stabilize within the burner at
other locations. The spot appeared dark because the porous matrix in
that location was not radiating strongly, indicating that the temperature
was significantly lower there than in the nearby radiating matrix. As inlet
velocity increased further, the flame continued to stabilize for successive
conditions and the cool spot increased in size. Eventually, the spot
encompassed the location of the thermocouples, and this is where blowoff
was said to occur. The flame blowoff was repeatable, starting at the same
location for each test run, and so it was possible to compare the data
from different runs.
This nonuniformity in blowoff was a major concern when it was first
observed and served as an impetus to investigate flow at the exit of the
burner. A laser doppler velocimetry study was undertaken by Wharton
(2003) in which velocity measurements were taken for both hot and cold
flow conditions at the exit of the burner. In addition, the cold flow
operation of each section of porous media was tested without the presence
of the other section to determine the independent effects on flow velocity.
Flashback was difficult to define experimentally. The upstream
section of porous media was meant to serve as a barrier to the flame and
prevent such an occurrence. For this reason, it was not desirable to
operate the burner in a condition where the flame was stabilized within
the upstream section of porous media. Because of this, flashback was
defined in this study as any condition for which the flame was located
within the upstream section.
The response of the flame to changes in inlet velocity at an equiva-
lence ratio of 0.60 is shown in Figure 2. As velocity increases, the flame
first moves upstream slightly, then reverses direction and proceeds
downstream until blowoff. Also, the highest peak temperature occurs at
an inlet velocity somewhere in the middle of the stable inlet velocity
range. At the low-inlet-velocity end, the peak temperature increases with
inlet flow velocity, but only up to a point. Somewhere near the middle of
the range, the peak temperature starts to decrease again until blowoff.
1180 M. T. SMUCKER AND J. L. ELLZEY

Figure 2. Axial temperature profiles for propane with an equivalence ratio of 0.60.

Examination of similar data for various equivalence ratios showed

the response of the axial temperature profiles and stable operating range
as equivalence ratio increased. First, the peak temperatures increased
with each increasing equivalence ratio, from a peak temperature of
1444 K for an equivalence ratio of 0.55 to a peak temperature of 1869 K
for an equivalence ratio of 0.65, resulting in a difference of 425 K. This
trend was expected, because that is the result for any conventional
premixed flame; the flame temperature increases with equivalence ratio to
a peak value at an equivalence ratio of about 1.1.
To see if fuel composition had an effect on burner performance and
emissions, experiments were conducted with methane. Burner operation
for methane and propane was similar. The flame exhibited the same
movement downstream to blowoff and the same variation in peak
temperature with increasing inlet velocity. The main difference was in the
stable operating range.
Stable operating limits were determined experimentally for five dif-
ferent equivalence ratios from 0.55 to 0.65. Experiments were not
conducted for richer mixtures because of the temperature limitations of
the porous media. Equivalence ratios richer than 0.65 would have
TWO-SECTION POROUS BURNER 1181

produced temperatures in excess of the maximum operating temperature

of YZA (1873 K), so such experiments could not be conducted safely.
Computations of the stable operating range, however, were
completed for a wider range of equivalence ratios for both propane and
methane. Figure 3 is a graphical representation of all of the experimental
and computational stable operating ranges for the burner. For
comparison, the laminar flame speed of propane is also shown.
The predicted stable operating limits in terms of firing rate (Figure 3)
follow similar patterns for both propane and methane. In each case, both
the upper and lower stability limits increase with increasing equivalence
ratio and the upper limit increases more sharply than the lower limit. This
has the effect of widening and shifting the stable operating range with
increasing equivalence ratio. The main difference is that the predicted
limits for methane are lower than those for propane. In particular, the
upper limit decreases more than the lower limit, indicating that the upper
limit is more sensitive to fuel composition than the lower limit. This effect
is also seen in the experimental data and was probably due to the dif-
ference in laminar flame speeds of methane and propane.
Firing rate is a function of both equivalence ratio and inlet velocity.
The experimental firing rates for operation on propane range from

Figure 3. Comparison of predicted and experimental stability ranges.

1182 M. T. SMUCKER AND J. L. ELLZEY

475 kW=m2 at an equivalence ratio of 0.625 and a velocity of 22 cm=s to

2650 kW=m2 at an equivalence ratio of 0.65 and a velocity of 118 cm=s.
The predicted upper stability limit agrees well with experimental
results for most of the tested equivalence ratios. In general, the predicted
and measured values are within a few centimeters per second. The only
point at which this comparison starts to break down is at the leanest
equivalence ratio. For the equivalence ratio of 0.55, the experimental
upper limit differs significantly from the computational value. In this
region, a significantly higher value was predicted than measured. This
had the effect of extending the computed stable operating range of
the burner to much leaner equivalence ratios than those seen in the
experiments.
There was significant discrepancy between the predicted and mea-
sured lower limits. As mentioned earlier, the predicted lower limit
increased slightly with increasing equivalence ratio and while the com-
puted and experimental values for propane agreed at the lowest equiva-
lence ratios tested, their values diverged significantly as equivalence ratio
increased. No trend could be determined for the experimental lower
stability limit for methane, because only one data point was determined,
but its value was lower than the predicted value.
There were several differences between the experiments and the
computational model. First, in the computations, data for PSZ rather
than YZA was used for the upstream section with 26.3 ppc. Barra et al.
(2003a) showed that the lower limit is primarily affected by the properties
of the upstream section and so it is reasonable that the computed lower
limits differ from experimental values. In addition, extinction rather than
flashback was observed at the lower limit in the experiments. Another
source of uncertainty in the computational model comes from the
material property approximations at the interface between the large- and
small-pore halves of the burner. To eliminate the discontinuity at the
interface between the two sections, the properties are smoothed over a
few millimeters. Therefore, the complexities of the fluid mechanics at the
interface are not accurately represented in the model.
To investigate the effect of inlet velocity on the radial exhaust
temperature profiles, the equivalence ratio was held constant at 0.60
while velocity was varied (Figure 4). At the lower end of the stable range,
an increase in velocity has a larger effect on average temperature than the
same increase at the high end of the stable range. With the increase from
30 to 50 cm=s, the temperature profile increases by about 110 K on average
TWO-SECTION POROUS BURNER 1183

Figure 4. Effect of changing inlet velocity on exhaust temperature profile.

and the profile becomes more uniform, the exception being the thermo-
couple at the far right. With a further increase in inlet velocity from 50 to
70 cm=s, the average increase in temperature is much less, only about
30 K. Also, the temperature difference within the profile decreased dra-
matically, from 66 K at 50 cm=s to 20 K at 70 cm=s.
The effect of varying equivalence ratio on radial temperature profiles
was also studied (Figure 5). In this case, the inlet flow velocity is constant
at 35 cm=s and the equivalence ratio increases from 0.55 to 0.65. Varying
the equivalence ratio with the velocity held constant had less effect on the
shape of the temperature profile but did serve to increase the average
temperature of the exhaust stream.
As shown in Figure 6, pressure drop generally increases with
flow speed and it is significantly higher for a reacting flow than for cold
flow conditions. This is due in part to the temperature. As the gas flows
through the reaction zone, its density decreases and hence its velocity
increases, resulting in greater frictional losses and a larger pressure drop.
For some of the conditions, pressure drop increased with velocity up
to a certain point after which it leveled off. This behavior can be
explained by examining the peak temperature variation in the axial
temperature profiles. For this burner the peak temperatures in the axial
1184 M. T. SMUCKER AND J. L. ELLZEY

Figure 5. Effect of changing equivalence ratio on exhaust temperature profile.

Figure 6. Pressure drop across the burner for various equivalence ratios.
TWO-SECTION POROUS BURNER 1185

Figure 7. Comparison of emissions levels for propane and methane operation.

temperature profiles increased with increasing velocity up to a point, and

then decreased again until blowoff. This temperature variation resulted in
an increase and then a decrease in pressure drop. When this effect is
added to the base pressure drop of cold flow, the trend is reasonable. In
the 0.65 equivalence ratio case, the highest peak temperature occurs for a
flow speed of 80 cm=s, which coincides with the peak in the pressure
drop data.
With the advent of more stringent regulations, emissions of UHCs,
NOx, and CO have become a very important consideration for burners.
Emissions of CO are extremely sensitive to the equivalence ratio. By
burning lean, these emissions may be maintained at very low levels
provided combustion is complete. Also, NOx emissions are controlled by
burning at low-temperatures. The porous burner operates effectively in
the lean, low- temperature regime due to heat recirculation to the
unburned mixture.
For the majority of the emissions data, the variation at a fixed
equivalence ratio is less than the calculated error in the measurements.
Hence, it is difficult to identify trends in the results. However, the pre-
cision of these measurements is much better than the calculated errors.
That is, the measured levels do not fluctuate from point to point to the
1186 M. T. SMUCKER AND J. L. ELLZEY

extent indicated by the error analysis, giving an indication that the trends
in this data are valid.
Emissions measurements were taken using methane with an equiva-
lence ratio of 0.65. Data from this experiment are displayed along with
the corresponding propane data for comparison in Figure 7. The propane
and methane data follow the same trends of increasing NOx and CO with
increasing velocity. Also, the NOx and CO levels are higher for propane
compared to methane. UHC levels for both fuels were fairly constant and
low. For equivalence ratios lower than 0.65, NOx and CO levels were
lower than those shown in the figure. All emissions for operation on
methane were less than 10 ppm, and for operation on propane only the
CO emissions rose to levels significantly higher than 10 ppm.

CONCLUSIONS
In this study, a two-section porous burner was investigated through
experiments and computations. In the experimental study, the stable
operating range with propane was identified for a burner with a 23.6-ppc
upstream section and a 3.9-ppc downstream section, each made from
YZA ceramic foam. This same burner was then operated with methane
for comparison. The computational study of the burner used a
one-dimensional approach with a complete chemical mechanism for
propane.
The experimental stable operating range for the burner with the
23.6-ppc upstream section was small at an equivalence ratio of 0.55 but
widened dramatically as equivalence ratio was increased. The upper limit
agreed well with computations. Previous work (Barra, 2003) on a smaller
burner showed some discrepancies that were thought to be caused by heat
losses, but because the burner in this study was larger, those heat losses
became less important relative to the heat release. Although non-
uniformities were observed near blowoff, these did not appear to
significantly affect burner performance.
Although the predicted upper limit agreed well with experiments, the
lower limit was overpredicted. The measured lower limit remained fairly
constant while the predicted limit increased with equivalence ratio. The
lack of agreement on the lower limit may be explained by a number of
factors. The material properties of the upstream section used in compu-
tations were not that of YZA because these values were not available. In
TWO-SECTION POROUS BURNER 1187

the experiment, quenching was observed whereas the computations

predicted flashback.
Both the experimental and computational methane limits were lower
than the corresponding propane limits for the equivalence ratios tested,
illustrating the effects of the lower laminar flame speed of methane.
Again, the experimental and computational values of the upper limit
agreed, whereas the experimental lower limit was lower than predicted.
The radial temperature profiles were measured just downstream of
the exit face of the burner. They showed good uniformity, with variation
on the order of 50 K. With increasing equivalence ratio the temperature
increased as expected. With increasing inlet velocity, the temperatures
increased and became more uniform.
The pressure drop was measured for cold flow as well as for reacting
flows. Pressure drops for reacting flows were always higher than the
corresponding cold-flow values. This was because the increased tem-
peratures in the reacting flow had the effect of lowering density, thus
increasing velocity and frictional effects and resulting in a higher pressure
drop across the burner. For several conditions, the pressure drop
increased initially and then leveled off with increasing inlet flow velocity.
This trend followed that observed in the peak temperature.
Most of the variation in emissions of UHC, NOx, and CO was within
the uncertainty of the measurements, but the precision in the measure-
ments suggests that trends are valid. The emissions data indicate that
both NOx and CO increased with firing rate whereas UHC remained
fairly constant. Overall the emissions were low, with most measured NOx
and UHC levels below 10 ppm, and most CO levels below 20 ppm.

REFERENCES
Barra, A.J. (2003) Computational Study of a Two-Section Porous Burner.
M.S. Thesis, The University of Texas at Austin, Austin, TX.
Barra, A.J., Diepvens, G., Ellzey, J.L., and Henneke, M.R. (2003a) Numerical
study of the effects of material properties on flame stabilization in a porous
burner. Combust. Flame, 134, 369.
Barra, A.J., Ellzey, J.L., and Henneke, M.R. (2003b) A Computational Study of
Heat Recirculation in Porous Burners. Central States=The Combustion
Institute 2003 Spring Meeting, Paper B22, Chicago, IL, March 2003.
Echigo, R., Yoshizawa, Y., Hanamura, K., and Tomimura, T. (1986) Analytical
and Experimental Studies on Radiative Propagation in Porous Media with
1188 M. T. SMUCKER AND J. L. ELLZEY

Internal Heat Generation. Proceedings of the Eighth International Heat

Transfer Conference. Edited by C.L. Tien, V.P. Carey, J.K. Ferrell.
Fu, X., Viskanta, R., and Gore, J.P. (1998) Measurement and correlation of
volumetric heat transfer coefficients of cellular ceramics. Exp. Thermal Fluid
Sci., 17, 285.
Howell, J.R., Hall, M.J., and Ellzey, J.L. (1996) Combustion of hydrocarbon
fuels within porous inert media. Prog. Energy Combust. Sci., 22, 121.
Hsu, P. and Howell, J.R. (1992) Measurements of thermal conductivity and
optical properties of porous partially stabilized zirconia. Exp. Heat Transf.,
5, 293.
Hsu, P., Evans, W.D., and Howell, J.R. (1993) Experimental and numerical study
of premixed combustion within nonhomogeneous porous ceramics. Combust.
Sci. Technol., 90, 149.
Khanna, V., Goel, R., and Ellzey, J.L. (1994) Measurements of emissions and
radiation for methane combustion within a porous medium burner. Combust.
Sci. Technol., 99, 133.
Mathis, W.M. Jr. and Ellzey, J.L. (2003) Flame stabilization, operating range,
and emissions for a methane=air porous burner. Combust. Sci. Technol., 175,
825.
Mital, R., Gore, J.P., and Viskanta, R. (1996) Measurements of radiative prop-
erties of cellular ceramics at high temperatures. J. Thermophys. Heat Transf.,
10, 34.
Mital, R., Gore, J.P., and Viskanta, R. (1997) A study of the structure of
submerged reaction zone in porous ceramic radiant burners. Combust. Flame,
111, 175.
Qin, Z., Lissianski, V.V., Yang, H., Gardiner, W.C., Davis, S.G., and Wang, H.
(2000) Combustion Chemistry of Propane: A Case Study of Detailed Reac-
tion Mechanism Optimization, Proc. Combust. Instit., 28, 1663.
Rumminger, M.D., Dibble, R.W., Heberle, N.H., and Crosley, D.R. (1996) Gas
Temperature Above a Porous Radiant Burner: Comparison of Measure-
ments and Model Predictions, Proc. Combust. Instit., 26, 1755.
Sathe, S.B., Kulkarni, M.R., Peck, R.E., and Tong, T.W. (1990) An Experimental
and Theoretical Study of Porous Radiant Burner Performance. Proc.
Combust. Instit., 23, 1011.
Smith, G.P., Golden, D.M., Frenklach, M., Eiteneer, B., Goldenberg, M.,
Bowman, C.T., Hanson, R.K., Song, S., Gardiner, W.C., Jr., Lissianski,
V.V., and Qin, Z. (1999) http:==www.me.berkeley.edu=gri_mech= (accessed
in July 2002.).
Takeno, T. and Sato, K. (1979a) An excess enthalpy flame theory. Combust. Sci.
Technol., 20, 73.
Takeno, T. and Sato, K. (1979b) A Theoretical and Experimental Study on an
Excess Enthalpy Flame. 7th ICOGER, Gottingen, Germany, 20–24 Aug.
TWO-SECTION POROUS BURNER 1189

Trimis, D. and Durst, F. (1996) Combustion in a porous medium — Advances

and applications. Combust. Sci. Technol., 121, 153.
Viskanta, R. and Gore, J.P. (2000) Overview of cellular ceramics based porous
radiant burners for supporting combustion. Environ. Combust. Technol., 1,
167.
Weinberg, F.J. Combustion temperatures: The future? Nature, 233, 239.
Wharton, J. (2003) Experimental Study of Exit Flow from a Porous Burner. M.S.
Thesis, The University of Texas at Austin, Austin, TX.
Younis, L.B. and Viskanta, R. (1993) Experimental determination of the volu-
metric heat transfer coefficient between stream of air and ceramic foam. Int.
J. Heat Mass Transf., 36, 1425.