NASA Technical Reports Server (NTRS) 20020025991: Quantitative Surface Emissivity and Temperature Measurements of a Burning Solid Fuel Accompanied by Soot Formation

6 Pages·0.43 MB·English

by  NASA Technical Reports Server (NTRS)

AIAA-2001-0627
QUANTITATIVE SURFACE EMISSIVITY AND TEMPERATURE MEASUREMENTS OF A
BURNING SOLID FUEL ACCOMPANIED BY SOOT FORMATION
Nancy D. Piltch, NASA Glenn Research Center, Cleveland, Ohio,
Richard D. Pettegrew, NCMR, Member AIAA, and
Paul Ferkul, NCMR, Senior Member AIAA
Abstract Introduction
Surface radiometry is an established technique for non- Infrared (1R) thermography as a technique for non-
contact temperature measurement of solids. We adapt contact temperature field measurements of solid sur-
this technique to the study of solid surface combustion faces requires knowledge of the emissive properties of
where the solid fuel undergoes physical and chemical the surface for accurate interpretation of the IR signal.
changes as pyrolysis proceeds, and additionally may In general, the surface emissivity and other radiative
produce soot. The physical and chemical changes alter parameters are afunction of temperature, wavelength,
the fuel surface emissivity, and soot contributes to the and viewing angle, as well as surface composition and
infrared signature in the same spectral band as the sig- finish _'2'3.While the radiative properties of various
nal of interest. We have developed a measurement that materials have been examined under limited condi-
isolates the fuel's surface emissions in the presence of tions 4'5'6'7's,materials undergoing the physical and
soot, and determine the surface emissivity as a function chemical changes of combustion introduce additional
of temperature. A commercially available infrared complexity. In particular, emissions from the flame
camera images the two-dimensional surface of ashless must be isolated, for example by optical filtering, from
filter paper burning in concurrent flow. The camera is the broadband surface radiation. Soot, if present, fur-
sensitive in the 2 to 5gm band, but spectrally filtered to ther complicates the interpretation of the signal, as it is
reduce the interference from hot gas phase combustion a broadband emitter, and therefore cannot be removed
products. Results show a strong functional dependence by filtering. Finally, in the mid-IR region where our
of emissivity on temperature, attributed to the combined detector operates, there is only a small spectral window
effects of thermal and oxidative processes. Using the that is free of gas phase emission, so that two-
measured emissivity, radiance measurements from wavelength pyrometry is not feasible.
several burning samples were corrected for the presence
of soot and for changes in emissivity, to yield quantita- This paper describes a technique for the experimental
tive surface temperature measurements. Ultimately the determination of the temperature dependence of the
results will be used to develop a full-field, non-contact emissivity of a burning thin solid fuel in the presence of
temperature measurement that will be used in space- soot. Part of this technique is a procedure to reduce the
based combustion investigations. effect of soot on the temperature field measurements for
moderately sooty flames. As two-dimensional tempera-
ture field measurements are ultimately desired, all
measurements were made in an imaging configuration.
Formulation
The measurement scenario consists of imaging a burn-
Copyright 82001 by the American Institute of Aero- ing sample surface with a mid-IR camera, while simul-
nautics and Astronautics, Inc. No copyright is asserted taneously acquiring an independent point temperature
in the United States under Title 17, U.S. Code. The measurement on the sample surface, using a thermo-
U.S. government has a royalty-free license to exercise couple. The image is filtered to block gas-phase emis-
all rights under the copyright claimed herein for Gov- sions (from CO, COz, H20 and vapor phase hydrocar-
ernmental purposes. All other rights are reserved by bons). A small hole in the fuel sample, near the ther-
the copyright owner. mocouple, allows the imaging of soot (if present) from
1
American Institute of Aerort_.uIi_s and Astrqnautics .
nss_sa prepnnt orreprimofa paper intended for presentation at a
conference. Because changes maybe made before formal
publication, thisismade available withtheunderstanding that itwill
notbecited orreproduced without thepermission oftheauthor.
AIAA-2001-0627
both sides of the sample. The signal from the soot can (flame reject) filter transmission and the detector sensi-
be subtracted from the detector signal, leaving only the tivity are spectraily uniform within the filter bandpass,
IR emissions of the solid surface. The corrected detec- (ii) lateral temperature variations (perpendicular to the
tor signal and the thermocouple measurement together flame spread direction) are small, (iii) scattering and
provide the surface emissivity. The correction reduces absorption of radiation by soot can be neglected, (iv)
to the familiar ratio of actual radiance to blackbody the detector response is approximately linear and, (v)
radiance when soot is absent. the imaging depth of field is sufficiently large to in-
clude the flames on both sides of the burning sample.
Figure 1shows a schematic layout of the fuel surface.
Surface TC Measure ,_Sampte
Sootradiation, reflectedofffud (Rp,) _._
Sootradiation,trznsmittedthroughfed (Rz_)
Feelsurface emission (SrQ 4
Sootradiation, direct (12)
/ Surface IR Measurermmt'
/ (spotD
Soot IR Measuremtmt
(Spot 2)
Figure 2
Radiance Components in the Observed Image
Figure 1
Figure 2 shows a schematic of the radiant contributions
Fuel Surface Arrangement and to the detector signal. The total radiance arriving at the
Measurement Locations detector from spot 1in figure 1,RTOT,is:
A Type K thermocouple mounted to the fuel surface RTOT = 0sR+ TsR+ esSB + R (1)
with a bit of epoxy provided an independent tempera-
ture measurement, needed for the determination of sur- Where the soot radiance from one side of the flame is
face emissivity. Once the emissivity was determined, R, the blackbody radiance from the surface is SB,and
the thermocouple was no longer required and the meas- the radiative properties of the fuel are us,%, and Ps for
urement scheme becomes truly non-contact. the emissivity, transmissivity and reflectivity.
The contribution of soot to the total radiance was iso- The time-encoded data were the total front view radi-
lated by simultaneously imaging through a small hole in ance coming from the fuel surface and soot (Eq. 1), the
the fuel. Imaged radiance there includes only emissions radiance of the soot alone, measured through the hole in
from soot in flames on both sides of the burning fuel. the fuel surface, and the thermocouple output. All radi-
The hole diameter was chosen to be much larger than ance signals were collected in detector counts, propor-
the spatial resolution of the camera, so that a local soot tional to photons received. The thermocouple signal
radiation measurement could be made without also col- was converted to equivalent counts by calibration
lecting radiation from the nearby sample surface. The against a blackbody source. Thermocouple data were
hole was located along the centerline of the fuel, and not corrected for radiative losses. The detector signal
the thermocouple was mounted on the surface nearby for spot 1on the fuel surface is St, and represents RTot;
along a horizontal line the signal from spot 2 is Sv Since the flame is symmet-
ric front and back, $2is twice the value of the single-
The measurement requires the following assumptions: side soot radiance R, where scattering and self-
(i) the fuel surface and soot emissions, the bandpass absorption are neglected. No assumption of soot emis-
American Institute of Aeronautics and Astronautics
AIAA-2001-0627
sivity is required. Apart from a proportionality constant Experimental Results
representing collection geometry, camera throughput,
and A/D conversion factors, Experimental results showing the surface emissivity (in
the 3.7-3.9 _m band) as a function of temperature are
plotted inFigure 3. This figure shows eight tests that
S_= RTOT (2) followed a similar qualitative trend, although with sig-
$2= 2R (3)
Emissivity vs. TC Temperature (K)
Conservation of energy gives:
1=_,+p,+ '_, (4) • 10.15 Ashless filter paper
• 10,20 4.0 ps|a, 21% 02
1.0 7 • 10,26a
Substituting and rearranging gives: 0.8 • 10.27a Upward Burning
10.28a
• 10.28b
_-_0.6 + 10.28(: ++++_1_
_,= ($1 - $2) /(SB- $2/2) (5)
- 10.29 4-
"_ 0.4
++ Lk + "" &12_
Note that we treat p, + "csas a single variable and do not 0.2 • T&+ + ..,, q_
determine the two quantities independently.
=,';PT t)tlr-
0.0
I ]
Experimental Setup 300 500 700
TC Temperature (K)
The fuel used in this work was Whatman #4 Ashless
Filter Paper, a thin, cellulosic material similar to other
fuels used in combustion studies 9'1°. The 5 x 10cm Figure 3
sample was taped to a thin metal holder and placed in a
Surface Emissivity in the 3.7-3.91Ltm Band_
26-liter chamber. A 0.27 mm diameter (29 gauge) re-
as a Function of Temperature
sistively heated Kanthal wire, woven in a sawtooth pat-
tern on the bottom edge of the fuel, provided the igni-
nificant scatter. These tests indicate a value of Es(T) of
tion source resulting in upward flame spread. A Type
approximately 0.3 +/- 0.1 at the lowest recorded tem-
K thermocouple (3-mil wire diameter) mounted about 5
peratures of about 400 K. The emissivity then drops to
cm up from the ignited edge, just to the side of the cen-
a minimum of about 0.1 near 530 K, climbs to a maxi-
terline, using a dot of epoxy for good thermal contact
mum of about 0.6 by 680 K, and finally drops sharply
provided the independent surface temperature meas-
to slightly below 0.4 by 750 K, at which point the tests
urement. At the centerline of the sample (next to the
terminated as the fuel was consumed.
thermocouple) a 5-mm hole allowed imaging of the
soot field alone. The spatial resolution of the camera
These eight tests are averaged to determine a relation
(at 30 cm working distance using a 20-degree forelens
between emissivity and temperature for the fuel used
with close-up attachment) is approximately 0.3 mm, so
that a 5 mm diameter hole allowed such a measurement. here. Using this relation, we can calculate the fuel sur-
face temperature using only the two IR signal meas-
Images were recorded in the mid-IR using acommer-
urements (S] and $2, described earlier).
cially available camera (single element detector, scan-
ning type) with a 3.8 +/-0.05 _tm bandpass filter which
Figure 4 shows the surface temperature (at apoint) cal-
effectively eliminated contributions from CO, CO2,
culated in this manner for two tests, along with the ac-
H20 and vapor phase hydrocarbons. The detector dy-
tual thermocouple traces. The calculated curves do not
namic range permitted imaging over a temperature
replicate the thermocouple traces exactly, due to the
range of approximately 400 K to 750 K, which corre-
averaging of the emissivity over multiple tests. These
sponds to preheat and pyrolysis of the fuel surface.
differences (between the calculated IR temperature and
Images were recorded at 30 Hz video frame rate and
the recorded thermocouple temperature) are an indica-
matched up with nearly simultaneous thermocouple
tion of the limitation of this technique, which depends
data. All tests were conducted in normal gravity, with
on how well the emissivity-temperature curve is known.
21% Oz at a pressure of 28 kPa.
3
American Institute of Aeronautics and Astronautics
AIAA-2001-0627
The non-monotonic thermal behavior of the emissivity
Calculated IR Temperature vs Time may be somewhat surprising. Though not yet con-
800 Ashless filter paper firmed, this behavior is believed to be aresult of the
4.0 psia, 21% O= combined effects of both pyrolysis and oxidation lead-
ing to structural and chemical changes in the fuel.
_700
Upward Burni_
There appears to be a qualitatively similar result re-
Test 10-_ ported for cellulose undergoing pyrolysis without com-
bustion 13.Emissivity for this case was measured by
600 /,,_/ I_ erature
FTIR reflectance assuming zero transmittance. Our
'//Y" I -'- TC Temperature
5oo /_,r Test10-291-- IRTemperature values differ numerically, due to the narrow bandpass
,_ __---TC Temperature of our measurement and our inclusion of transmittance.
400
1 2 3 4 5 Absorba nee of KimWipes vs.Wa'velength
Time (seconds) 2.0 Spc'Ctr+l ReJIton Ilallgcd
1.9' I'_,inK Ramc F/ll¢r
1.8'
Figure 4
1.7"
Calculated IR Temperature and Thermocouple avJl 1.6-
.Temperature vs. Time (including soot correction) 12C
1.4-
< 1.3!
Discussion
1.2'
1.l
These tests show a low overall emissivity and astrong I.rl p
temperature dependence of the emissivity. Therefore,
0.9
constant emissivity values reported in various refer-
iL 6 s I0 L2 14 16 IS 20 22 24
ences [_,_zmay not be appropriate for burning samples. _J_kngllt (micro.s)
The reported measurements may have been made at a Figure 5
different wavelength, different temperature, or in non-
S.pectrally Dependent Absorbance of KimWipes at
pyrolyzing environments. It isnot intuitive that the
Room Temperature_ Using an FTIR
value of the emissivity should be as low as that deter-
mined here; the literature TM12states that the emissivity
of paper should be in the range of 0.8 to 1.0. A plausi-
An area of concern is the validity of neglecting scatter
ble explanation for this discrepancy can be seen by re-
and self-absorption of radiation by soot. This isequiva-
calling that the current measurements were made in a
lent to the standard simplified models of gas emission.
narrow spectral bandpass (3.7 - 3.9 ftm). Figure 5
The simplified model is expected to capture most of the
shows the mid-infrared absorption spectrum of Kim-
observed phenomena with greater accuracy than com-
wipes (trade name) a cellulosic laboratory wipe fre-
plete neglect of soot effects. With improved knowledge
quently used in combustion studies at room tempera-
of soot distribution in these flames, a more complete
ture. While this spectrum changes with temperature,
approach may be developed if the measurement preci-
and may differ in detail from that for ashless filter pa-
sion warrants it.
per, it demonstrates that the radiative properties can be
very different in a narrow bandpass compared to the
The influence of soot on the observed radiance is
overall spectrally averaged quantity. The spectral re-
shown in Figure 6. This figure shows the measured
gion that is free of gas phase emission is a region of
emissivity of the fuel as a function of temperature, and
relatively low absorbance and therefore low emissivity.
was one of the test points used in this study. The upper
The same spectral region will be used for the non-
trace is calculated from the raw data without correcting
contact field thermometry, which is the ultimate goal of
for the presence of soot, while the lower trace includes
the project and therefore is the region in which these
our correction. The greatest difference is at low tem-
calibrations must be performed.
peratures, corresponding to the preheat region. In this
American Institute of Aeronautics and Astronautics
AIAA-2001-0627
ity results. Additional scatter may arise from spot-to-
Emissivity vs Temperature spot variation in the surface condition.
Effect of Soot Correction
The theoretical limit to which the emissivity can be
,.
determined (using the equipment inthis study, based on
the manufacturers' stated detector and thermocouple
accuracies) isestimated to be + 12% (relative error).
:_ • • Emiss w/o soot corr I _ This translates to a maximum uncertainty in surface
temperature of + 15K. Clearly we have not yet reached
this limit, but the technique described here has the po-
" * t t it* tential for improvement when the independent surface
I
0.0 I , , temperature measurement (thermocouple) is refined.
400 500 600 700 800 We will attempt this in future work.
TC Temperature (K)
Conclusions
Figure 6
Tests of the technique were conducted using thin paper
Surface Emissivity (3.7-3.9pm Band) as a Function fuel (ashless filter paper) in upward burning normal
of Temperature, with and without Soot Correction gravity tests. These tests show a strong dependence of
emissivity on temperature, with values everywhere
region, the downstream tip of the flame, which has the much lower than those commonly published. The low
highest soot volume fraction, appears in the field of values are attributed to the low spectral absorption by
view of the camera and contributes substantial radiance cellulose in the narrow band used to eliminate interfer-
to the observed signal. As the flame propagates down- ence by gas phase emissions. These effects must be
stream, the fuel surface temperature rises and its radi- accounted for inany experimental effort that uses IR
ance increases. Simultaneously, the camera views a thermography for quantitative temperature measure-
less sooty portion of the flame, so that soot contributes ment.
proportionately less radiance. The two traces are essen-
tially identical within experimental error for tempera- Examination of the data from these tests shows a fair
tures higher than about 550 K. amount of scatter; much of this is thought to be due to
inconsistencies with the mounting technique of the sur-
The high amount of test-to-test scatter in the emissivity face thermocouples. The uncertainties caused by this
measurements requires some comment. One of the effect are thought to account for the variations observed
factors that contributes to the experimental scatter is the from test to test.
mounting technique of the thermocouples. The junc-
tions were fixed to the sample surface using epoxy; The measurements show asignificant influence of soot
however, the exact amount of epoxy used varied radiance on the determination of- surface emissivity,
slightly from test to test. The size of the thermocouple particularly in the downstream (preheat) region. The
beads in the various tests (along with the epoxy used to effect of the soot-signal compensation algorithm was
mount them) was approximately 0.22 to 1.5 mm. The studied by calculating the surface emissivity while first
corresponding thermal response times reach a maxi- including, then neglecting soot. The test data indicate
mum of about 0.5 msec, indicating that the thermal in- that for the conditions and configuration employed,
ertia of the bead is not a concern. failure to account for soot can cause over prediction of
the emissivity in this region. For the particular case
The thermal length scale of the flame (cdVb,oy_,t) is studied, the over prediction was about a factor of two.
approximately 1ram. Hence, a temperature gradient
may occur across the bead, introducing an uncertainty In summary, a technique was developed for determin-
in the temperature measurement. Since the determina- ing the temperature-dependent emissivity of a thin
tion of the emissivity is based on correlation between burning sample. This method provides a simple way
the thermocouple data and IR data, the uncertainty in to isolate emissions from the soot field in the gas phase
the thermocouple data will propagate into the emissiv- flame and implicitly includes the effects of physical and
chemical changes occurring in the fuel surface during
American Institute of Aeronautics and Astronautics
AIAA-2001-0627
combustion. The technique is applicable to a wide
range of materials and up to moderate soot loading. No
assumption of optical opacity of the fuel is required. 7Pettegrew, R., Piltch, N., and Ferkul, P.:
Ultimately it may permit non-contact thermometry Emissivity Measurement of a Thin, Pyrolyzing Solid,
without the need for a surface-mounted thermocouple 37thAIAA Aerospace Sciences Conference, AIAA 99-
and without the reliance on tabulated emissivity values 0700 (1999)
generated under different conditions.
8Serio, M.A.., Pines, D.S., Bonano, A.S., Solomon,
P.R.: An Instrument for Characterization of the Ther-
Acknowledgements
mal and Optical Properties of Charring Polymeric Ma-
terials, pp. 1447-1453, 25thSymposium (International)
The authors wish to express their gratitude to the
on Combustion, The Combustion Institute, 1994
NASA Aeronautics and Space Administrations' Office
of Biological and Physical Research who supported this
work. 9Bhattacharjee, S., Altenkirch, R.A., Olson,
S.L., Sotos, R.G.: Heat Transfer to a Thin Solid Com-
Trade names or manufacturers' names are used in this bustible in Flame Spreading at Microgravity, Journal of
report for identification only. This usage does not con- Heat Transfer-Transactions of the ASME, 113:(3) 670-
stitute an official endorsement, either expressed or im- 676, August, 1991
plied, by the National Aeronautics and Space
Administration. 10Bhattacharjee, S., Altenkirch, R.A,
Sacksteder, K.: Effect of Ambient Pressure on Flame
Spread Over a Thin Cellulosic Fuel in aQuiescent, Mi-
crogravity Environment, Journal of Heat Transfer-
iSiegel, R., and Howell, J.R.: Thermal Radia-
Transactions of the ASME, 118:(1) 181-190, February,
tion Heat Transfer, Hemisphere Publishing Corpora-
1996
tion, Washington (1981).
2 txGrober, H., Erk, S., and Grigull, U.: Funda-
Arakawa, A., Saito, K., and Gruver, W.A.: mentals of Heat Transfer, 3raEdition, McGraw-Hill
Automated Infrared Imaging Temperature Measure- Series in Mechanical Engineering, (1961)
ment with Application to Upward Flame Spread Stud-
ies, Part I, Combustion and Flame, Vol. 92, pp. 222-230 12Incropera, F.P., and DeWitt, D.P.: Funda-
(1993). mentals of Heat and Mass Transfer, 3rd Edition, John
Wiley& Sons, (1981).
3Staaf, O., Ribbing, C.G., and Andersson,
S.K.: Temperature Dependence of the Band Emittance t3Suuberg, E.M., Milosavljevic, I., and Lilly,
for Non-gray Bodies, Applied Optics, Vol. 35, No. 31, W.D.: Behavior of Charring Materials in Simulated Fire
pp. 6120-6125 (1996). Environments, NIST-GCR-94-645 (1994).
4Haugh, M.J.: Infrared Thermometry for Low
Emissivity Metals, Instrumentation Society of America
Transactions, Vol. 22, No. 3(1983).
5Olstad, S.J., Tanaka, F., and DeWitt, D.P.:
Evaluation of a Method for Measuring Spectral Emis-
sivity at Moderate Temperatures, AIAA 20thThermo-
physics Conference, AIAA-85-0991 (1985).
6Zhang, Y.W., Zhang, and C.G., Klemas, V.:
Quantitative Measurements of Ambient Radiation,
Emissivity, and Truth Temperature of a Greybody:
Methods and Experimental Results, Applied Optics,
Vol. 25, No. 20, pp.3683-3689 (1986).
American Institute of Aeronautics and Astronautics