TRW REPORT NO.
14103-6004-RO-OO
LASER HOLOGRAPHY STUDY
OF
OIL-FIRED BURNER COMBUSTION
ENVIRONMENTAL PROTECTION AGENCY
-------
This report was furnished to the Environmental Protection Agency by
TRW Systems Group in fulfillment of Contract CPA 70-4, Modification 7.
The contents of this report are reproduced herein as received from TRW
Systems Group. The opinions, findings and conclusions expressed are
those of the authors and not necessarily those of the Environmental Pro-
tection Agency. Mention of company or product names does not constitute
endorsement by the Environmental Protection Agency.
-------
LASER HOLOGRAPHY STUDY
OF
OIL-FIRED BURNER COMBUSTION
Prepared by
A. B. Witte and B. J. Matthews
TRW Systems Group
One Space Park
Redondo Beach, California 90278
Contract CPA 70-4
ENVIRONMENTAL PROTECTION AGENCY
Technical Center
Research Triangle Park, North Carolina
November 1971
-------
FOREWORD
This report summarizes the holographic interferometry studies of
oil fired burner flames accomplished under Environmental Protection
Agency Contract CPA 70-4, Modification No.7, dated 20 April 1971. This
report is submitted by TRW Systems Group, TRW, Inc., in accordance with
provisions of the modified contract.
Work on this contract was accomplished by personnel of the TRW
Fluid Mechanics Laboratory. The Principal Investigator was Dr. Arvel
B. Witte. Dr. Ralph F. Wuerker of TRW's Systems Group Research Staff
served as a consultant in the areas of laser physics and holocamera
design. The TRW Project Manager was Mr. Birch J. Matthews. Technical
direction and administration of the contract for the Environmental
Protection Agency was the responsibility of Messrs. William B.
Kuykendal and Robert M. Statnick.
The authors wish to thank Messrs. Frank Gomes and Bruce Winston
of the TRW Fluid Mechanics Laboratory for their assistance in con-
ducting the holography experiments. In addition, the assistance and
cooperation of Mr. Blair Martin of the Environmental Protection Agency
was greatly appreciated.
iii
-------
CONTENTS
l.
2.
INTRODUCTION. . . . . . . . . . . .. . . . . . . .. . .. . . . .. . . . . . .
EXPERIMENTAL APPARATUS .......... .......... .....
3.
4.
2.1 FUEL OIL BURNERS ..........................
2.2 LASER HOLOGRAPH DEVELOPMENT ...............
2.3 THERMOCOUPLE PROBES .......................
TEST PROCEDURE.................................
DATA REDUCTION PROCEDURE .......................
5.
RESUL TS ........................................
6.
7.
5.1 SEQUENCE OF PROGRESS AND PROBLEMS SOLVED ..
5.2 HOLOGRAMS.... . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3 HOLOGRAPHIC INTERFEROGRAMS ................
5.4 INTERFEROGRAM DATA REDUCTION ..............
5.5 THERMOCOUPLE MEASUREMENTS .................
CONCLUSIONS. . .. . . . . . . .. . .. . . . .. . . . . .. . . . . . . . . . .
RECOMMENDATIONS. . . . . . . .. . . . .. . . . . . . .. . .. . . . . . . .
8.
REFERENCES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
APPENDIX A, THERMOCOUPLE RADIATION
CORRECTION, CONDUCTION LOSSES AND
FREQUENCY RESPONSE .............................
v
Page
3
3
3
11
15
17
21
21
21
23
35
40
44
45
46
-------
ILLUSTRATIONS
1.
2.
ABC Standard Burner ...........................
ABC Mi te Burner...............................
3.
4.
Beckett Bantam Burner .........................
5.
Burner Tes t Ri g ...............................
Schematic of Wide-Angle Two-Beam Transmission
Holocamera ....................................
6.
Schematic of Holocamera Scene Volume, Maximum
View Angel, and Fog-Reducing Shutter Position..
Chromel-Alumel Flame Thermocouples
7.
8.
........... .
Hologram of Coordinate Systems and Background
G ri d ..........................................
9.
Burner Spray Pattern Burner Off. Spray on for
'2 Seconds.....................................
10.
Burner Spray Pattern Burner Off. Spray on for
2 Seconds.....................................
11.
Reconstruction photographs of a ruby laser inter-
ferogram of open flame combustion from the ABC
Mite oil burner. The photos differ by the focus of
the copy camera. The direction of flow is from
ri gh t to 1 eft. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.
Burner Interferogram.
Burner on for 5 Seconds...
,':)
!-..J.
Burner Interferogram. (Reconstructed from
Bleached Plate). Burner on for 5 Seconds
...... .
14.
Burner Interferogram (Reconstructed from
Bleached Plate). Burner C~ for 1 Second ........
15.
Infinite Fringe Interferogram of ABC Mite
Burner Centered at 10 inches from Head ..........
16.
17.
Interferogram of ABC Mite Burner ................
Infinite Fringe Interferogram Combustion Starting
Transient in ABC Mite Showing Intermittent
Comb us t ion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
vii
Page
3
4
5
6
7
12
14
22
24
24
26
28
28
29
29
30
32
-------
18.
19.
ILLUSTRATIONS (Continued)
Infinite Fringe Interferogram Combustion Starting
Transient in ABC Mite Burner Showing Two Burning
Clouds of Gas........................................
Reconstruction photograph of the Mite burner
approximately 1/2 second after ignition. This
infinite fringe interferogram illustrates the
effect of placing a thermocouple probe in the
combustion zone.......................................
20.
Reconstruction photograph of a finite fringe inter-
ferogram of the Beckett oil burner recorded approxi-
mately 2 seconds after ignition. Note the signif-
icant reduction in the level of turbulence with this
burner in comparison with the Mite illustrated in
F; gu re 1 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
21.
Reconstruction photograph of a finite fringe inter-
ferogram of the Beckett oil burner. The recording
shows the fringe shift due to vaporizing fuel spray
prior to full ignition and combustion. Note the
discontinuity at the top of the burner suggesting
a non-uniform spray distribution and/or the onset
of combustion in this local region....................
Infinite Fringe Interferogram of Bernzomatic Flame ...
22.
23.
Finite Fringe Interferogram of Bernzomatic Flame
Infinite Fringe Interferogram of Hot Air Gun Jet .....
24.
25.
Finite Fringe Interferogram of Hot Air Gun Jet........
Infinite Fringe Interferogram of Bernzomatic Burner
Oriented Vertically..................................
26.
27.
Temperature Profile at the Exit Plane of the Beckett
Burne r ...............................................
28.
29.
Density Profile for Bernzomatic Burner at Exit Plane ..
Density Profile for Hot Air Gun at Plane 0.5 Head
Diameters from Exit Plane ..........................,..
Temperature measurement of the Mite Burner flame
using a 0.051 inch diameter bare wire chromel-
a 1 ume 1 th e rmo co up 1 e ...................................
30.
viii
Page
32
33
33
34
36
36
37
37
38
39
41
42
43
-------
1.
INTRODUCTION
Furnace emission characterizations and fuel demand data for oil burners
indicate that during winter months space heating is a significant source
of SOx, NOx and particulate emissions. A recent evaluation of flame
retention devices on small space heater, high pressure at~mizing gun burners
showed that these emissions can be substantially reduced. However, it is
not obvious why some of these retention devices are more effective than
others, although the microstructure of the resulting flame is thought to
influence emission levels.
The primary objective of this project was to design instrumentation and
evaluate its feasibility for characterizing flame turbulence and temperature
gradient for several burner configurations. In this way, it may be possible
to relate basic flame characteristics to the generation of emissions, parti-
cularly NO which is formed and fixed in the primary combustion zone.
Previous work at TRW has shown .laser holography to be2ao effective
research tool in the study of high temperature phenomenon. ,J Laser hologra-
phy offers an opportunity to permanently record for later detailed study
high resolution flame detail in three dimensions. In addition, the result-
ing high energy, coherent illumination is sufficient to penetrate radiant
flames while producing a high signal to noise ratio.
In all studies of this type, the main problem encountered is film
fogging by a luminous subject. Temperature data have been calculated
from hoJographic interferograms in former studies involving turbulent
wakes.4,5 The primary requirement imposed by the data reduction procedure
upon the data is that fringes can be resolved, numbered, and that only
small statistical excursions from an otherwise axisymmetric mean flow are
encountered. This latter restriction can be removed when holocameras are
developed with 1800 viewing as is now believed to be possible.6
The following sectiors of this report describe the experiments and
discuss the experimental technique (Sections 2 and 3, respectively). Data
from the experimental program is analyzed in Section 4. The results are
summarized in Section 5.
-------
2.
EXPERIMENTAL APPARATUS
2.1
FUEL OIL BURNERS
A schematic of each high-pressure atomizing-gun burner is shown in
Figures 1, 2, and 3. The fuel oil was No.2 distillate obtained from
Gulf crude. The oil had an API gravity equal to 36.* The fuel oil was
gravity fed to the burner which was mounted in a test rig. A burner box
9 inches square by 15 inches in length with Pyrex No. 7740 glass sides
was used to approximate actual operation while permitting optical access
for holography. The flue gases were vented to the roof of the laboratory
through a 20 foot stove pipe with the aid of a small exhaust fan. A
schematic of the test rig is shown in Figure 4. Fuel comsumption of the
ABC Mite was measured to be 1 m1/sec or about 0.9 gal/hr.
2.2 LASER HOLOGRAPH DEVELOPMENT
A significant fraction of the project effort was devoted to developing
a laser ho10camera suitable for producing high quality interferograms.
This camera and the procedure for its application are described below.
Basic Laser Holograph - Short Scene Depth
The laser holograph is comprised of a pulsed ruby laser and a holo-
camera. The pulsed ruby is used in this investigation because sufficient
energy can be produced to penetrate the burner flame during a pulse width
small enough to stop the action of the combustion phenomena. The pulsed
ruby laser has been described in detail in an earlier TRW report to EPA
and will not be discussed further here.7 The holocamera design is a
modified version of the apparatus shown in Figure 5 developed on another
contractS by Dr. Ralph Wuerker of TRW. Modifications to the holocamera
shown in Figure 5 for use with this contract were as follows: (1) the
angle described by mirror No.5, the hologram and the quadruple lens set
is 90° not 75°; (2) the scene beam-hologram angle is 45° not 52-1/2°;
and, (3) a plate rotator was installed to permit finite fringe operation.
The purpose of recording finite fringes is to avoid fringe ambiquity
problems.
The advantages of this holocamera are that it has a very wide viewing
angle (measured to be 40°)9 and good resolution (about 20 microns). 0
The scene depth, namely the distance between the hologram and the objec-
tive.1ens is 17 inches. The side wall of the 9 x 9 inch square cross-
sectlon burner box was located about two inches from the quadratic lens
set. This position permitted the largest viewing aperture of the burner
box; however, the aperture was not large enough to permit viewing of both
top ~n~ bottom walls of the burner, a requirement imposed so that boundary
condltlons are known for data reduction purposes. The scene depth was
subse~ue~t1y expanded so that the burner side wall could be observed. A
descrlptlon of the larger scene depth ho10camera follows.
* An API gravity of 36 is equivalent to a specific gravity of 0.845.
2
-------
w
ELECTRODES
AIR TUBE
FUEL
PIPE
ABC Standard Burner.
Figure 1.
-------
+=-
ELECTRODES
AIR TUBE
Figure 2. ABC MITE Burner.
-------
AIR TUBE
FUEL
PIPE
(J'1
ELECTRODES
COMBINATION END CONE
AND FLAME RETENTION SHIELD
Figure 3.
Beckett Bantam Burner.
-------
BURNER BOX
9X9X 15 INCHES
. GAS THERMOCOUPLE
FUEL TANK
r~
BURNER
m
Figure 4.
PYREX 7740 GLASS
SI DE WALLS
Schematic of Burner Test Rig.
FUE L GAS
STACK
EX HAUST
FAN
FLOW DIRECTION
-------
......
SCENE BEAM
Figure 5.
Schematic of Wide-Angle Two-Beam Transmission Holocamera.
-------
Modified Laser Holograph - Large Scene Depth
One of the convenient features of the original 11 inch focal length
objective lens set used in this holocamera is that the lens set was
quadratic; namely, it had two matched pairs of glass lenses. (The quad-
ratic set weighed 85 pounds).
Two iterations on holocamera design were implemented to extend the
scene depth by about a factor of two. The first design utilized the two
outer lenses of the quadruple lens set. This set comprised a concave-
convex pair with a resultant focal length of 18 inches. Holograms
recorded with this setup showed severe distortion. The inner lens set
(focal length 20.5 inches), also a concave-convex pair, was next installed
and proved to be superior in performance to the first set. Good quality
holograms of a coordinate system in the scene volume as well as on the
objective lens were recorded in the process of path matching the system.
These modifications required that the holocamera be lengthened by a
factor of two. Modifications were accomplished by the addition of sections
to each end of the "breadboard" holocamera. The new system allows
observation of the top and bottom of the box.
Helium Cell Fringe Compensator Design
The purpose of the compensator is to reduce the overall fringe shift
so that it will remain within the picture (hologram). This apparatus
will not reduce the sensitivity of the index of refraction changes and
gradients (density, temperature gradients). The compensator is com-
prised of a plexiglass box 6 inches deep with a 13 inch square cross-
section. During the comparison beam exposure, the compensator (box) is
filled with helium at atmospheric pressure. During scene beam exposure,
the box is filled with air at 1 atmosphere. The compensator was bolted
to the quadratic lens set mount on the side opposite the hologram plate.
If the box filled with air comprises the comparison scene, denoted
subscript 2, and the box filled with helium comprises the test scene,
denoted subscript 1, then the fringe shift relation is developed.
Exposure 1:
Exposure 2:
Helium
(1 Atm)
Air
(1 Atm)
...
Lc
~
Lc
~
8
-------
For the case of planar geometry the fringe number becomes:
1
5 = - (n - n ) L
A 1 2 c
where the index of refraction
p.
n. = 1 + 13. -2-
1 1 Psi
After some algebra, one obtains
5=113 ~(~.~~-1\ L
A 2 Ps2 P2 T2 132 I) c
where
5 = fringe shift
13. = G1adestone-Da1e constant for species
1
and 0.36 x 10-4 for helium)
P = gas density
P = gas pressure
T = temperature
A = radiation wavelength
L = cell width (15cm)
c
and subscript
s = standard conditions of temperature and pressure at which both
~IS are evaluated.
i (2.92 x 10-4 for air
Taking P1 = P2' Tl = T2' the result becomes
5 = -3.32 Lc
= -50 fringes
Thus, relative to air, helium causes a fringe shift of -50 fringes.
9
-------
If the burner box were completely filled with 2000°F air, and
s = 4.2 (~ s - 1) LB'
then S ~ -74 fringes (relative to air) where the burner box width LB = 22.5 cm.
Because the gas in the box is cooler at the walls, the absolute fringe shifts
would be somewhat less. For the simplified case cited above, the fringe
shift of the flame relative to helium would be about -24 fringes. Thus,
the compensator so designed would appear to be adequate.
Shutters and Filters
In order to minimize film (plate) fogging it was necessary to
experiment with a number of mechanical shutters to "shield" the film
from the intense burner flame. Three shutters were evaluated.
Focal Plane Roller Blind Shutter --
This shutter had a 4 x 5 inch rectan9ular format and was of Graflex
origin. In order to keep the dark (cloth) blind from burning, a large
solenoid driven leaf type outer shutter, made by Harvard, was used to
protect the blind. The fastest shutter speeds ranged between 80 and
100 milliseconds. Adverse fogging occurred at these speeds. The shutter
was located at plane A-A in Figure 6.
Leaf Shutter 2-5/16 inch Diameter --
A spring-loaded shutter taken from a Fairchild camera was operated
at 6 milliseconds, which removed the fogging problem but reduced viewing
area. Solenoid actuation was added to the shutter. The shutter was
located at position B-A in Figure 6.
Leaf Shutter 3-1/2 inch Diameter --
A spring-loaded shutter taken from a Kodak Air Force mapping camera
was operated at 9. 18, and 40 milliseconds. At a shutter speed of 9
milliseconds, holograms and interferograms were essentially free of fog.
In addition, this shutter has 3 times the area of the former shutter and
thus allows substantially larger viewing angle of the flame. The shutter
was located at position B-A in Figure 6.
Fi lters
Filters were used in order to reduce the amount of fogging during
hologram plate.exposure and thus increase the signal-to-noise ratio.
The burner radlates strongly in the visible and infrared. A Kodak
Wratten No: 79 filter wa~ used to reduce the fogging by visible radiation.
The transmlsslon propertles of the Wratten 70 filter are as follows:
10
-------
------- Scene Volume Boundary
------ Max. View Angle Construction
A-A Focal Plane Shutter Location
B-A Forward Shutter Location
PERIPHERAL
_1:-;/-
...----...-.-*' ./
-
........-"- ./
-- /'
B ---- /
---- /
--/- /~
............. (3, VIEW ANGLE
A-- -- ----- J.
""""".................... ~
""""".................... "
--,--~-~
..................~
-
PE RI PHE RAL .....
RAY
REFERENCE
BEAM
,
OPTICS
CENTERLINE
--'
--'
PLATE ANGLE, a
...SCENE
BEAM
4" X 5"
HOLOGRAM
PLATE
(AGFA GEVAERT
8E75 ANTIHALATION
BACKED)
SCENE CENTER
MAX. VIEW ANGLE
(MODEL CENTERLINE)
QUADRUPLE OBJECTIVE
LENS SET 13" DIA.
Figure 6. Schematic of Holocamera Scene Volume, Maximum View Angel, and Fog-Reducing Shutter Position
-------
Wavelength
in Microns (p)
Percent
Transmittance
.65
.66
.67
.68
.69
.70
.63
10.5
35.0
55.2
70.3
79.0
Good transmission properties are observed for wavelengths above 0.66p
with a transmittance of almost 79 percent at the ruby wavelength, A =
0.6943p. There are no good filters in the range of 0.7 and beyond
(infrared) which do not exclude the ruby wavelength; however, the
spectral response of Agfa 8E75 emulsion beyond 0.76p is negligible.
From the fog levels encountered early in the test program with the
relatively slow speed focal plane shutter, one may conclude.that these
oil-fired burners radiate strongly in the 0.65 to 0.75p reglon.
In addition to the Wratten No. 70 gelatin filter, Corning Glass
Works filters number CS-2-58 and CS-2-64 were evaluated. These filters
essentially duplicate the Wratten No. 70 in the visible, but also
exclude radiation for A>2.75p. No difference in fog level was noted
when the Corning glass filters were used in series with the Wratten No.
70 filter.
2.3 THERMOCOUPLE PROBES
Probe Design
Both shielded and unshielded chromel-alumel thermocouple probes
were built during the program. It was desired to measure both mean
temperature and temperature fluctuation. Installation of the probes
in the burner box was accomplished using a swivel-seal joint located
near the head end of the burner box. The swivel joint permitted temp-
erature traverses to be made during burner operation.
The mean temperature measurement thermocouples were packed in a
refractory oxide insulation su~rounded by an inconel sheath. The junc-
tion was formed by stripping back the inconel sheath. Probe construction
is shown in Figure 7. The unshielded junction mean temperature probe
(Figure 7a) used 0.051 inch diameter wires to demonstrate that chromel-
alumel thermocouples could be used in the burner combustion environment.
12
-------
--'
w
0.0625
3.5"
0.35"
BAKED
SAUEREISEN
JOINT
0.011"
WIRE
7-b
0.25
0.050 t
-.l 0.75
1 0.051" ~
WIRE
7-0
0.011"
WIRE
0.0625
t
0.5"
---1
7-c
Figure 7.
Chromel-Alumel Flame Thermocouples
-------
Radiation losses occur when unshielded thermocouples are employed
in a high temperature environment. Recognizing that a radiation shield
would be needed to minimize these losses, the probe shown in Figure 7b
was designed and built. The thermocouple junction now exchanges radiation
with the hot refractory shield rather than the IIcoldll walls of the burner
box. Holes were drilled in the shield to allow the hot gases to aspirate
over the thermocouple junction. In this design, smaller wires (0.011
inch diameter) were chosen to minimize the conduction losses to the
support.
The thermocouple probe designed to measure temperature fluctuations
is shown in Figure 7c. Use of a radiation shield would seriously impair
the time response of this type of probe. Thus, a correction must be
applied to the measurement for radiation losses. Because of the large
LID of the wire, conduction losses to the support are negligible.
An estimate of unshielded thermocouple error due to radiation losses
and an analysis of thermocouple frequency response to the turbulence of
the flame is presented in Appendix A. In addition, conduction losses
through the thermocouple support are also considered in the appendix.
14
-------
3. TEST PROCEDURE
The test procedure was established and followed for recording holo-
grams and interferograms of burner combustion. Starting with the burner
in place, the following steps were taken to make the recordings:
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Shutter cocked and energized
Room lights out
Uncover film plate
Record the comparison scene holographically
Recock shutter
Turn burner on
Record the test scene holographically
Turn burner off
Cover film plate
Lights on
Develop plate
Remove and clean windows on burner box.
In one instance, it was determined that the pyrex windows could
only withstand about 90 seconds of uncoo1ed operation before cracking. *
Not all of the flame interferometry recordings were made using the
windowed burner box. Some reGordings were taken with the box removed
and the flame unconfined (open flame tests). For these tests, a piece of
1/4 inch thick P1exig1ass plate was used to protect the quadruple lens
set of the ho10camera from flame impingement.
Some preliminary work was done with the helium cell; however, on
the first formal test with this device, a slight overpressure resulted
causing one of the Plexiglas wall bonds to fail. The cell was not used
again during this program due to a reordering of priorities.**
* Use of a nitrogen gas purge over the windows would increase the oper-
ational life of the Pyrex windows. (In the future, quartz windows may
have to be used if true adiabatic wall operating conditions seem
des i rab 1 e.
** Damage to the cell is minor and may be easily repaired such that the
concept may be tried again in the future.
15
-------
All of the interferograms and holograms were recorded on Agfa 8E75
anti-halation backed 4 x 5 inch glass plates.* Development time in
Eastman Kodak HRP developer was typically 4 minutes for an interfero-
gram. The acid fixer used was Nacco fix.
Interferograms were reconstructed with a Model l24A Spectra Physics
helium neon gas laser (A = 0.6328~) illuminator. The recordings were
photographed using Polaroid Type 52 film having an ASA rating = 400.
When desired, negatives of the reconstructed scene were made using
Polaroid Type 55 PIN film.
* Product of Agfa-Gevaert, Antwerp, Belgium.
16
-------
4. DATA REDUCTION PROCEDURE
Data reduction is comprised of three parts: photographing the
desired view of the holographic interferogram and recording the fringe
pattern using a microdensitometer; interpreting fringe number from the
mincrodensitometer trace; and computing density by means of the standard
equation for fringe shift,
z
S(x,y) = f f G(x, y. z) - PooJ dz
o
The holographic interferogram is mounted in a reconstruction jig
whose orientation is the same with respect to the reference beam as it
was during the original recording of the holographic interferogram.
During reconstruction, a 15 mw continuous wave He-Ne laser (A =
0.6328 microns) was used to illuminate the holographic interferogram.
A copy camera was positioned behind the holographic interferogram. The
camera could be focused on the event which was a known distance behind
the holographic interferogram and bounded by the viewing cone of the
hologram.
The interferograms were photographed with either Polaroid Type 55
PN film which provides a positive and negative copy of the event using a
nominal 1 to 10 second exposure at f/5.6 or Polaroid Type 52 which is
about 8 times faster. The photographic density of the fringe pattern
on these negatives was then recorded on a Joyce precision micro-densi-
tometer having a slit size of about 0.010 inch by 0.010 inch.
Fringe number S was measured relative to the undisturbed (light
fringe) background gas, i.e., relative to S = O. The first dark fringe
occurs where the change in optical path is ooA/2. Interpretation of the
change in optical path in wavelengths of light, i.e., fringe number S,
is aided by keeping in mind the standard equation for fringe shift.
Starting at the edge of the flame the fringe number begins to decrease,
the first dark fringe yielding a value of S = 1/2 fringe.
The mean density profile calculations carried out here were made
under the assumption that the flow field was axisymmetric. With the
assumption of axisymmetric flow, the equation can be inverted as the
Abel integral to obtain12
P (y) - P = - L-
00 7TK
2
rs ds dr2
f dr2
(/ - y2) 1/2
y2
17
-------
When this equation is rewritten as
r2
s
.:\ f ds ( 2 2) 1 /2
p(y) - Poo = - 'irK 2 dr2 d r - y
y2
it may be cast into the following finite difference form (the Schardin-
van Voorhis approximation13) for which the density is assumed constant
in each of N thin angular rings of thickness Do:
P (y) - P =
00
~k2_i21
2:\ N-l
'lrllK L
k=i
~(k + 1)2 + i2'
(Sk - Sk+l) 2k + 1
where rk = 6.k, Y = 6. i and r s = 6. N has been used and the i ndi ce i takes on
the following values: i = 0,1,2, ..., N-l. K is the Gladestone-Oale
-4
constant, e.g., for p K = 2.98 x 10 for air when p = 1 atm.
00 00
Because the,interferometer comparison scene is recorded in air and
the test scene recorded with combustion products, account must. be made
for different chemical species. The fringe number equation is,
S = .!. f ( n- n ) dx
:\ 00
Using the Gladestone-Oale relationship, n-l = S ~, one gets
Ps
S = Soo Poo f(~ Pooo .t?.- -1)' dx
:\ Po 00 SOO PTO Poo
18
-------
where
S = fringe shift
A = irradiation wavelength
a = Gladestone-Dale constant
P = dens ity
x = distance along light ray
subscripts
~ = comparison scene conditions of air before burner is ignited
T = test scene conditions having combustion product species
o = standard conditions at which ~IS are evaluated
The procedure then is one of calculating the quantity
~ P~o L
a~ PTO P~
by the standard Abel technique just described and then dividing by the
a P
quantity ~ ~ to get the density ratio, L, which is the flame
I-'~ PTO P~
density normalized by the density of air at the conditions of the com-
parison scene recording (room temperature and pressure).
The value of a is calculated as follows from the Gladestone-Dale
equation for species i = 1,2 ... N. For any dilute gas,
n - 1 = Ea.
1
p.
--1
Poi
a.
1
= pE - C.
Poi 1
where a. is evaluated for species i and Ci is the mass fraction.
by defi~;tion
However,
n - 1 = aT ~
OT
Therefore, combining these two equations yields
C.a.
- E--L!.
aT - POT Poi
19
-------
Poi can be shown to be 0.00279 mi lbm/ft3 at atmospheric pressure and O°C
where m. is the molecular weight of species i.
1
As an example, consider the case of burning butane whose chemical
reaction at stoichiometric conditions is
CH3 (CH2)2 CH3 + 6.502 + 6.5 (3.76) N2 ~
4 C02 + 6.5 (3.76) N2 + 5 H20
From this reaction the following is calculated:
Reaction C. s. Poi SiP oi Ci Poi m. C. S. P .
Products 1 1 1 1 1 01
C02 O. 186 4.51 x 10-4 O. 1275 3.53 x 10-3 0.0237 44 0.656 x 10-3
N2 0.723 2.97 x 10-4 0.0812 3.66 x 10-3 0.0587 28 0.265 x 10-3
H20 0.095 2.54 x 10-4 0.0522 4.88 x 10-3 0.00496 18 0.464 x 10-3
P = z:C.p .
o 1 01
= 0.0874
z:C. (S./p .) = 3.770 x 10-3
1 1 01
Therefore,
S = (0.0874) (3.77 x 10-3)
T
= 3.29 x 10-4 (NOTE: for air S
00
=
2.92 x 10-4)
The average molecular weight is calculated as,
C.
= z: -1.
m.
1
-
m
After some algebra
m = 28.3 (for air m = 28.7)
S T P 000
From these results, one can show that s- --- = 1.14.
00 PTO
20
-------
5. 1
5.
RESULTS
SEQUENCE OF PROGRESS AND PROBLEMS SOLVED
Before proceeding with a discussion of the results, a chronological
perspective of the progress and problems leading to those results will
be given. The following areas were investigated and progress made as
indicated.
.
.
Film fogging problem solved with filters and shutters
Laser light absorption shown not to be a problem
Holograms of the fuel spray and holographic interferograms
of the flame were recorded with 17 inch scene depth ho10camera
Holographic interferograms were recorded of the flame with a
41 inch scene depth holocamera
Finite fringe interferograms were recorded of flame
Both infinite fringe and finite fringe interferograms were
recorded of a small propane burner and of a hot air source
Preliminary data reduction of the interferograms was
accomplished
Preliminary thermocouple temperature measurements were made
of the burner flame.
.
.
.
.
.
.
5.2 HOLOGRAMS
A ho10gram* of a 1 inch square grid coordinate system placed in the
scene at 45° to the beam axis is shown in Figure 8. A faintly observed
1 cm square grid located at the objective lens is also shown. The copy
camera was focused on the coordinate system in the foreground and thus
the grid in the background is out-of-focus.
The purpose of recording holograms, as distinguished from holo-
graphic interferograms, on this program was to observe the fuel spray.
The effort was meant to be on a quick look basis rather than an in depth
study.
* These figures are called holograms and interferograms although they
are more properly identified as photographs of one view of the recon-
structed holographic scene, whether a hologram or holographic inter-
ferogram.
21
-------
11
Fi gure 8.
Hologram of coordinate systems and background grid.
22
-------
In Figure 9 is shown a burner spray pattern on the windows of the
burner box. The burner was off but the spray was allowed to run for 2
seconds before the hologram was recorded. Without burning, the spray
quickly mars the transmission through the box as can be seen from the
irregular visibility of the background centimeter grid. The wavelike
pattern on the walls is reproducible qualitatively from shot to shot.
Agglomerated droplets are visible on the windows besides the large wave-
line films observed.
In Figure 10 is shown a burner spray pattern which was recorded by
placing an opaque checkerboard grid at the plane of the centimeter grid
(i.e., at the objective lens). The burner box windows were removed for
this test. Recording the spray hologram in this way allows dark field
illumination viewing of the spray pattern against the opaque strips of
the checkerboard. There was no preponderance of evidence that spray
droplets could be resolved in this holo~ram with the unaided eye, although
individual scattering centers (droplets) were observed on some of the
other holograms near the periphery of the lens. The droplet size is
believed to be ~25 microns.* The single droplets could be recorded via
low angle forward scattering holography using a two- or three-beam holo-
camera arrangement as demonstrated and described in References 13 and 14.
These holograms were recorded in the small holocamera because reso-
lution of individual droplets is expected to drop off as scene depth is
increased. No further work or analysis was done on either the spray
holograms or on optimizing the resolution of the holocamera in this
"quick-look" effort.
5.3 HOLOGRAPHIC INTERFEROGRAMS
Short Scene Holocamera Results
A two inch diameter, leaf-type mechanical shutter was located and
installed before the film plate holder of the holocamera apparatus. The
shutter speed was measured photoelectrically and determined to be 6 msec;
approximately 10 times faster than the focal plane shutter used originally.
The faster shutter together with a Wratten No. 70 filter resulted in a
successful holographic interferograms of open flame combustion from the
ABC MITE Model S oil burner. The recording was free of any objectionable
fog.
* The resolution of this diffuse
was previously measured using Air
and found to be ~25 at a distance
plate. This work is described in
illumination transmission holocamera
Force 1951 resolving power targets
of 12 inches from the holographic
Reference 8.
23
-------
24
Figure 9.
Burner Spray Pattern Burner Off.
Spray on for 2 seconds.
Figure 10.
Burner Spray Pattern Burner Off.
Spray on 2 seconds.
-------
The recording was made on Agfa 8E75, 4 x 5 inch plate; however,
since the shutter diameter was only two inches, one had the feeling of
looking through a "knothole" when viewing the virtual image reconstruction.
This is seen in the accompanying reconstruction photographs (Figure 11).
The reconstruction photos were made using a bellows-type view camera
with a 127 mm focal length lens at an fill aperture. All reconstruction
photographs were recorded on Polaroid Type 52 film. In the first photo-
graph, the view camera was focused on a centimeter grid inscribed on the
1/4-inch acrylic sheet covering the quadruple lens set of the holocamera.
In the original scene, this grid was located approximately 15 inches from
the center of the holographic film plate. The succeeding shots were made
by extending the bellows from the rear (film holder) side of the camera
in one inch increments. In addition to changing the focus~ the recon-
structed image was simultaneously magnified.
The purpose of this test was to determine the effect of increased
shutter speed on the fog level of the Agfa plate. Since the flame was
essentially unconfined, burner operating conditions were not monitored
and systematic reconstruction procedures were not used as quantitative
data would not be reduced from this interferogram. From a qualitative
point of view, the test was very successful in that a fog free inter-
ferogram of burner combustion was achieved. The series of four recon-
struction photos demonstrate the three dimensionality of the holographic
interferogram and show that the combustion scene is not opaque to the
ruby scene light.
Careful examination of the four reconstruction photographs of
Figure 11 reveal subtle changes in the fringe patterns at different focal
positions of the camera. The quantity and pattern of the fringes indicate
a high degree of turbulence in the flame. When viewing the reconstruction
from extreme angles against a dark field background, it is possible to
detect individual oil droplets by scattered laser scene light. This is
not seen in the accompanying reconstruction photographs. The droplets
may be too small to resolve in the diffuse scene light. This observation
does suggest, however, that forward scattered light holography techniques
could be employed to assess droplet life and number density distribution
in the combustion volume.
A larger 3.5 inch diameter shutter, mentioned in Section 2.2~ was
installed and used successfully to record holographic interferograms of
the burner flame. This shutter has three times the area of the former
shutter and hence represents an improvement in holographic coverage of
the flame. The sp~eds Jf the mechanical leaf-type shutter are 9, 18,
and 40 milliseconds. To date only the 9 millisecond exposure has been
used. The Wratten No. 70 filter was also used.
25
-------
~
en
Figure 11.
Reconstruction photographs of a ruby laser interferogram of
open flame combustion from the ABC Mite oil burner. The
photos differ by the focus of the copy camera. The direc-
tion of flow is from right to left.
~
-------
A description of some of the burner infinite fringe holographic
interferograms recorded with the short focal length holocamera is given
next. Figures 12 and 13 show two reconstructions of the same holographic
interferogram. The burner head is just visible in the right side of the
picture. A centimeter grid is shown out of focus in the background.
The view camera used in reconstruction had a focal length of 127 mm.
Figure 12 was copied at f/32 for 6 minutes. Figure 13 was copied at
f/32 for 6 seconds. The difference between the two reconstructions is
that the plate photographed in Figure 13 had been bleached. This process
returns the silver to silver halide and allows a brighter picture to be
recorded. The right central portion of the photographs, where the tur-
bulent flame predominates, shows close fringe spacing and large gradients.
Outside this central region, cooler gases exist and larger fringe spacing
is observed.
The most important feature observed in these two figures is that the
burner flame is very turbulent. The eddies appear to range in size by
at least a factor of 30.
Shown in Figure 14 is a reconstruction of a burner interferogram in
which the burner had been turned off for 1 second. Residual burning lasts
for about 2 seconds. The fringe spacing is much larger in Figure 14
than for Figures 12 and 13 indicating that the density gradients are
lesser for the former case than the latter two cases. Different regions
of the hot gas are observed here just like in the former case. These
regions are delineated by changes in the fringe spacing. Observed im-
mediately to the left of the burner head is a small contour fringe. This
fringe is believed to be related to evaporation of the spray in this
region. In the left hand portion of the picture the oil spray has im-
pinged on the window producing a poorer quality gas interference pattern
there.
In Figure 15 an infinite fringe interferogram of the ABC Mite burner
flame is shown at a distance of 10 inches from the burner head. Irregular
large scale fringes are in the upper left diagonal half of the picture
indicating small gradients and low temperatures. Smaller fringes are in
the lower half indicating large gradients and high temperatures. It ap-
pears that the luminous part of the flame is in the lower half region.
Strong turbulence confuses the fringe structure and numbering in that
region.
Long Scene Depth Holocamera Results
In order to see the flame boundaries, the holocamera was increased
in scene depth from 17 to 41 inches (see Section 2.2). An infinite.
fringe interferogram of the ABC Mite flame confined by the burner b?x 1S
shown in Figure 16. The fringes are better resolved at the boundar1es,
but it is not possible to follow one fringe completely because of the
extreme turbulence.
27
-------
~,'"
Figure 12.
Burner Interferogram.
Burner on For 5 Seconds.
l'l--C~
. ,Yf1fi!.- ~~~ ~~i
:0 ~'T(l.-~f,'IiIi~~. -'. ,',1. r~'&." .". '~
~ -'" - ' ., ~. .{, ~j: ;:', ',j ~':E'" ,~~.:
- :;-', - 'rO-~".,). '~", ~;~..;-
~;E~~~,--:}: i2 ~/~i, "c'~~;.i~.:-~.:~~~_~I-t~~0F,(;
. -'.','~ 7-..;;" (c'}'. ~', . ;:--:;th"'", ,.~-...-...\;~~". ~, ...
11'(,-;,.),.'.. ,,~t . "I,,' "',. ~~~~ J'A--.;.t~.r., ",' i
"',~,;Jl£\l-~~\i:'''''','",' ;,*,~.Y'\:"i.~.t.:~" ",: .~,~i;,'''',k4
<:£~;;.:-.' ""'..:._.}~ ~~"'~"~3:\I.-..tr,.~..~. ..~ 'i
~. . ~,-~.~ '" ",...,.,-<,.", ",,;('i$!'..,.,. '/ ''''''' j
':~~~~'~~:-~~:~~~,r. ..:..~.~};,~;!b~'(:
,- ";.- if :..:.. "'-,.
~.' .~... ii.
. :~.?".""'::r:.
- . ~ ,-
.,
Fi gure 13.
Burner Interferogram. (Reconstructed from Bleached Plate).
Burner on for 5 Seconds.
28
-------
Figure 14.
Figure 15.
Burner Interferogram (Reconstructed from Bleached Plate).
Burner Off For 1 Second.
Infinite Fringe Interferogram of ABC Mite Burner
Centered at 10 Inches From Head.
29
-------
Figure 16.
Interferogram of ABC MITE Burner.
30
-------
During our work with the ABC Mite burner, it was noticed that the
starting transient of the burner was not a smooth process. Intermit-
tent, repetitive starts were common. Two interferograms showing this
phenomena are given in Figure 17 and 18. In Figure 17 combustion
extinguished itself completely after an initial burning cloud of fuel
oil ignited. In Figure 18 a second burning cloud appears to be
emerging from the burner. It is anticipated that poor combustion
and high emissions are present at this time.
A finite fringe interferogram of the ABC Mite after 1/2 second
combustion is shown in Figure 19 with one of the thermocouple probes.
An early time recording was made to avoid the large thermal free con-
vection currents which initiate subsequently and add to the obscuration
of the fringe pattern. The probe was included to show what distur-
bance, if any, was introduced by the probe. It is believed that the non-
symmetrical fringe pattern is caused by combustion turbulence and not
the probe. The fringe resolution is particularly good in this photo-
graph, accentuating the turbulence present in the flame. The flame
appears to be asymmetric with respect to the spray axis which is along
the burner head centerline. A preliminary interpretation of this fringe
pattern suggests that combustion may not have been fully established
and that the flame had lifted off the face of the burner. As noted,
this recording was made very early after ignition.
Next it was decided to see whether the Beckett and ABC Standard
burners were as turbulent as the ABC Mite. An interferogram of the
Beckett is shown in Figure 20 at a point 1 second after ignition and
presumably past the combustion transient discussed earlier. The
characteristic feature of this interferogram contrasted to the ABC Mite
is that the flame of the Beckett appears more concentrated along the
burner centerline. Possibly this occurs because of poorer spray
distribution. It may be a contributing factor to the Beckett emitting
twice as much NO as the Mite, a result determined in Reference 1. Some
attempts at data analysis were made on this interferogram. This is
discussed in Section 5.4.
The reconstruction photograph of Figure 21 is a finite fringe
interferogram recorded using the Beckett burner. The recording was
made prior to ignition of the oil droplet spray. The fringe shift
apparent in this illustration is due to changes in the optical path
from vaporizing fuel droplets. The pattern appears symmetrical in the
shape of a cone with the exception that some fringe shift appears at
the top of the burner head near the face. This suggests that the spray
pattern produced by this burner is not always entirely ~ni!o~m. .The .
recording may have caught the very onset of fuel spray 19n1tlon ln thlS
one local region.
31
-------
~
Figure 17.
Infinite Fringe Interferogram Combustion Starting Transient
in ABC Mite Showing Intermittent Combustion.
Figure 18.
Infinite Fringe Interferogram Combustion Starting Transient
in ABC Mite Burner Showing Two Burning Clouds of Gas.
32
-------
Figure 19.
Reconstruction photograph of the Mite burner approximately
1/2 second after ignition. This finite fringe interfero-
gram illustrates the effect of placing a thermocouple probe
in the combustion zone.
Figure 20.
Reconstruction photograph of a finite fringe interferogram
of the Beckett oil burner recorded approximately 2 seconds
after ignition. Note the significant reduction in the level
of turbulence with this burner in comparison with the Mite
illustrated in Figure 19.
33
-------
.~
Figure 21.
Reconstruction photograph of a finite fringe interferogram
of the Beckett oil burner. The recording shows the fringe
shift due to vaporizing fuel spray prior to full ignition
and combustion. Note the discontinuity at the top of the
burner suggesting a non-uniform spray distribution and/or
the onset of combustion in this local region.
34
-------
Because of the large scene filled with a hot, turbulent burner
flame, the fringes are irregular, closely spaced and difficult to
interpret. It was suggested that the understanding of burner combustion
especially interferometrically, would proceed faster if smaller burners'
or a small air heater might be studied in a somewhat more controlled
manner. To this end, both a Bernzomatic propane flame and a laboratory
hot air gun were recorded. These are illustrated in Figures 22, 23, 24,
and 25. Both infinite fringe and finite fringe interferograms were
recorded. Bouyancy effects are particularly noticeable. As the hot gas
forward velocity decreases due to entrainment, bouyancy forces cause the
gas to rise. The important features to observe of the small scale burners
are the following:
1.
2.
Fringe spacing is large enough to resolve individual fringes.
Lateral boundaries are wholly contained within the pictures
so that fringe numbering can be accomplished.
Turbulence is not so intense that the interpretation is
obscurred.
3.
An improvement to the study of small burners would be to orient the
flame axis vertically so that the bouyancy forces would be symmetrical
to the normal symmetries of the flame such as that shown in Figure 26
of the Bernzomatic recorded during an independent TRW study. Data
reduction can be best achieved for symmetrical fringe patterns or
patterns perturbed only slightly from a mean symmetrical pattern. Large
asymmetries can be treated if a wide angle camera is used to record wide
angle data. Some data reduced from the Bernzomatic will be discussed in
Section 5.4. The hot air gun interferograms shown in Figures 24 and 25
have lesser fringe shifts than the Bernzomatic because of lower temp-
erature. Turbulence is in evidence on a much broader scale range than
for the Bernzomatic, which has larger scales represented largely because
of the induced motion caused by bouyancy forces. The turbulence in
evidence in Figures 24 and 25 was caused by a grid and small blower,
part of the air gun. Preliminary data were also reduced for the gun and
is given in Section 5.4.
5.4 INTERFEROGRAM DATA REDUCTION
The Abel inversion integral (Section 4) was applied to 3 cases to
obtain some preliminary feasibility of reducing the data on the Beckett
(Figure 20), the Bernzomatic (Figure 26) and the hot air gun (Figure 24).
Because the fringes could not be resolved near the centerline of
the Beckett, data could only be reduced from .5 burner head radii out to
the outer boundary as shown in Figure 27. The main feature is the very
large temperature gradient starting at about R/RB = 0.6. The gradient
there is
6T/T - 6T
6R/R - 7.5 or 6R = 800°F/inch.
35
-------
.,
Figure 22.
Infinite fringe interferogram of Bernzomatic flame.
Figure 23.
Finite fringe interferogram of Bernzomatic flame.
36
-------
Fi gure 24.
Infinite fringe interferogram of hot air gun jet.
Figure 25.
Finite fringe interferogram of hot air gun jet.
37
~
-------
,""
Figure 26.
Infinite fringe interferogram of Bernzomatic
burner oriented vertically.
38
-------
8 2.0
I-
"-
I--
o
}-
~
W
\0
w
0<:
:J
I-
~
~ 1.6
:E
w
I-
2.6
TEMPERATURE CALCULATIONS
- STOPPED IN REGION NEAR
BURNER CENTERLINE WHERE
FRINGES COULD NOT BE
RESOLVED. RESOLVED.
~
RB = BECKETT BURNERHEAD RADIUS.
R = RADIUS POINT OF MEASUREMENT.
T = FLAME TEMPERATURE AT RADIUS R.
T = AMBIENT ROOM TEMPERATURE.
"-
~
2.4
2.2
1.8
1.4
1.2
1.00
1.0
RADIUS RATIO, R/RB
Figure 27.
Temperature profile at the exit plane of the Beckett Burner.
2.0
-------
A small correction for the fact that the gas in the comparison scene is
air and in the test scene is combustion products has not been made but
is discussed in Section 4.
The Bernzomatic density profile is shown in Figure 28. The approx-
imation has been made that the indicies of refraction are unchanged be-
tween the comparison and test scenes. If account is taken of this effect,
as shown in Section 4, the density shown in Figure 28 should be about
15% less. Because the molecular weight is about the same between test
and comparison scenes, the temperature ratio is just the inverse of the
density ratio. The density drops rapidly at the edge of the burner to
a minimum (temperature maximum). This rises to a relative maximum and
then to a relative minimum on the centerline. The non-monotonic behavior
is believed to be caused by passage of the cool gases into the reaction
zone. The temperature gradient at the edge of the Bernzomatic is about
a factor of 10 larger than for the Mite. Considerable profile detail
can be achieved by this method. Burner turbulence may playa strong
role in producing these largely different temperature gradients. For
instance, mixing at one edge of a turbulent flame is far more rapid than
for a laminar one. This tends to make a more uniform density/temperature
jet like that shown for the hot air gun in Figure 29. The temperature
of the jet on the centerline is about 100°F above ambient temperature.
This result comes directly from the density profile shown and the perfect
gas law.
Use of small burners holds greater promise for holographic data
reduction in a controlled experiment for the following two reasons:
1.
Flame size is smaller, producing fewer fringes which can
be resolved and interpreted, and;
Flame turbulence level can be controlled by using screens,
boundary layer trips, and other forms of turbulence control
and generation.
2.
5.5 THERMOCOUPLE MEASUREMENTS
~f the th~ee the~mocouple designs fabricated, shown in Figure 7,
and ~l~c~ssed 1n Sect10n 2.3, data have been obtained only with the
fe~slb111t~ mo~el, a 0.051 inch diameter barewire chromel-alumel probe
(~lgure 7a). Ihe data for the Mite in the burner box are shown in
F1gur7 30. Tem~erat~re, 5 ;nc~es from the burner head, is shown as a
fun~t10n 0: rad1al d1stance Wh1Ch has been normalized by burner head
rad1u~. ~lth a !ew exceptions, the temperatures are within 100°F for
a~l a1r.m1x se~t1ngs of the Mite. Using a radiation correction estab-
11shed 1n Sect10n 2.3, the peak temperature should be about 26000R
Based upon d = 0.051", t = 0.75", .
H = BT~ ' k = 1.12 BTU
g ft hroF hr ftOF
(chromel), the conduction er~or was shown to be negligible. The peak
temperature seems low. Poss1bly, fuel spray is impinging on the ther-
mocou~17'. More work needs to be done on this type of measurement in
the v1C1n1ty of the fuel spray.
40
-------
PIPIfO .6
~
--'
1.0
.
-
XI D = 0 .
.
.
.
. Ii
. - .
.
. .
. .
.
.
4 .. . .
.
. ..
.
. .
. 18... . 4
. . .
.9
.8
.7
.5
.4
.3
.20
.4
.2
.3
.1
.5
.6
.7
.8
.9
1.0
1.1
r/rb
Figure 28.
Density profile for Bernzomatic burner at exit plane.
-------
.9
. -
41
.
.
- .
It -
4. . . . . . . . .
. . . .
X/D = .5
D = 4.8
). = . 6943E-4
a = .935E-1
# 1 .292E-3
. 1
.2
.3
.4
.5
.6
.7
.8
.9
1.0
1.1
1.0
.8
.7
.6
.po
N
PIPoo
.5
.4
.3
.20
r/rb
Figure 29.
Density profile for hot air gun at Plane 0.5 head diameters from exit plane.
-------
1400
u..
~
I-
1300
~
w
1200
Fi gure 30.
1800
1000
t
.
....
.
.
.
....
.
. .
....
AIR MIX SETTING
. 1.6 (FUEL RICH) .
. 2.6 ....
.... 5.0 (FUEL LEAN)
.
....
1700
1600
1500
1100
o
1.0
2.0
3.0
4.0
r/rb
Temperature measurement of the Mite Burner flame using a 0.051 inch diameter barewire
chrome1-a1ume1 thermocouple.
-------
6.
CONCLUSIONS
1.
The feasibility of recording fog-free holographic interferagrams
of residential fuel fired oil burner flames was demonstrated with
two holocameras designed to observe the near and far field of the
burner.
2.
The most important feature observed interferometrically was that
the flame structure is extremely turbulent. This turbulence made
it impossible to resolve the fringe pattern on the centerline of
the burner.
3.
Preliminary data reduction of the interferograms showed that on
the outside edge of the burner, the temperature gradient, ~T, was
about 800°Fjinch. ~R
4.
Barewire thermocouple measurements made of the flame were corrected
for radiation, and are believed to be lower than the true flame
temperature because of fuel droplet impingement and evaporation.
Excellent quality holographic interferograms were recorded of a
small burner (Bernzomatic burner). Fringes could be resolved,
read and interpreted. Temperature gradients calculated at the
edge of this laminar flame were about a factor of 10 higher than
for the Beckett fuel oil burner turbulent flame supplied by EPA.
5.
6.
Recording a holographic interferogram of a hot air gun showed the
feasibility of controlling the turbulence level of a flame-like
hot gas. Preliminary data were also reduced in this case.
44
-------
7.
RECOMMENDATIONS
1.
Study the full-scale burner using holographic interferometry by
sampling only a small part of the burner flame. This concept
can be implemented by using a "cookie cutter" manifolded laterally
and vented apart from the test sectio~
Study a subsca1e burner which produces a resolvable number of
fringes, provides more flexibility and control of the pertinent
variables, especially the mixing and turbulence factors. Measure
mean flow and turbulence properties of the flame.
2.
3.
Study both burners cited in (1) and (2) above using cooled probe
anemometry. Fast response .006 inch diameter cooled quartz tube
probes (with platinum film sensing element) are available commer-
cially and would provide such measurements as turbulence intensity,
power spectral density, auto correlation and eddy integral scale.
Ion probes may also be considered for this task. The thermocouple
probes designed and fabricated under the present contract should
be tested and evaluated. These measurements are necessary to
supplement and validate the interferometry measurements of mean
and fluctuating properties. Temperature measurements in the
presence of liquid droplets may require design of a special
temperature/enthalpy probe.
Study combustion of liquid droplets holographically to determine
a model for liquid droplet combustion in a turbulent flow field.
Determine the relationship between droplet size and turbulent
eddy size for most efficient combustion.
4.
45
-------
8.
REFERENCES
1.
Howekamp, D. P., IIF1ame Retention - Effects on Air Pollutionll,
Paper submitted at the Ninth Annual Convention, National Oil
Fuel Institute, Atlantic City, N.J. 9-11 June 1971.
Matthews, B.J., and Wuerker, R.F., liThe Investigation of Liquid
Rocket Combustion Using Pulsed Laser Ho1ographyll, AIAA 5th
Propulsion Joint Specialist Conference, United States Air Force
Academy, Colorado, Springs, Colo., 9-13 June 1969.
2.
3.
"Studies of the Erosion of Aluminum Cartridge Casesll, TRW Report
No. 13774-6002-RO-00 Army Contract DAAA25-70-C-0152, Oct. 1970.
4.
Witte, A.B., Fox, J., and Runga1aier, H., IILoca1ized Measurements
of Wake Density Fluctuations Using Pulsed Laser Holographic
Interferometry, II Presented at AIAA Reacti ng Turbu1 ent F1 ows
Conference, San Diego, Calif., 17-18 June 1970.
Witte, A.B., Fox, J., and Runga1dier, H., IIHo1ographic Interfero-
metry Measurements of Mean and Localized Fluctuating Wake Density
Field for Cones Fired at Mach 6 in a Ballistic Rangell, AIAA Paper
71-564 Presented at AIAA 4th Fluid & Plasma Dynamics Conference,
Palo Alto, Calif., 21-23 June 1971.
5.
6.
Witte, A. B., Wuerker, R. F., and Hef1 i nger, L. 0., IIWi de Ang1 e
Holographic Interferometry and Data Reduction to Three Dimensional
Density Fie1dll, Technical Brief, TRW Systems, June 1970.
Matthews, B. J. and Kemp, R. F., IIHo1ographic Determination of
Injected Limestone Distribution in Unit 10 of the Shawnee Power
P1ant,1I TRW Final Report No. 14103-6001-RO-00, EPA Contract
CPA 70-4, June 1970.
7.
8.
Matthews, B.J., Wuerker, R.F., and Harrje, D.T., IISmall Droplet
Measuring Technique,1I Final Report, AFRPL-TR-156, Air Force
Rocket Propulsion Laboratory, Edwards, California, July 1968.
Witte, A.B., IIThree Dimensional Flow Field Analysis by Holographic
Interferometry,1I Final Tech. Rept. 15 Feb. 1971, TRW Report No.
12414-6005-RO-00.
9.
10.
Personal communication from W. Kuyhenda1 to Birch Matthews and
Ralph Wuerker, TRW Systems Group.
Caldwell, F.R., IIThermocoup1e Materia1sll, NBS Monograph No. 40
1 March 1962.
11.
12.
Wey~, F.J.,~ "Ana1ytica1 Methods in Optical Examination of Super-
sonlC Flow , N~vord Report 211-45, Navy Department Bureau of
Ordnance, Washlngton, D.C., 11 December 1945.
46
-------
APPENDIX A
Thermocouple Radiation Correction
A simple estimate of the thermocouple radiation error was calcu-
lated by assuming that an energy balance exists between the convection
heat flux to the wire and the radiation flux away from the wire, neglec-
ting the background radiation back to the wire. The model develops as
one considers a section of wire:
qradiation
qconvection
In the steady state, qconvection = qgradiation and,
hg (Tg ~ Tw) = £ a Tw4
Where re-radiation back to the thermocouple has been neglected so as to
yield a conservative estimate.
(1 - ~w) = E~ ~~
g g g
and
Tw -
Tg -
3" -1
[1+ E~:W ]
A-l
-------
where
T = temperature
h = connective heat
g = gas side
w = wall
transfer coefficient
E = wire emissivity
cr = Stefan-Boltzman constant
For oxidized chromel-alumel thermocouples E ~ 0.87.11 For a 0.011
inch diameter wire at 22000R and h = 72 BTU/ft2 hroF, the correction
becomes g
TW
T = 0.84
g
Thus, the gas temperature would be 2600oR. Because the precise value of
E is unknown and may vary as conditions change between stoichiometric
and excess air operating conditions, it will be necessary to use a
radiation-shielded aspirating thermocouple probe of the type discussed
in Section 2.3.
Thermocouple Conduction Losses to Support
The support losses are computed by equating the conductive flux in
the wire to the convective and radiative flux from the gas in the steady
state:
~ 4~xr-
br~ qXifix )
d / '\ hg(Tg-Tw)~dLl.x
qrad7rd~X
A-2
-------
The energy balance in steady state is,
7Td2
(qx+~x - qx) ~ - hg(Tg - Tw) 7Td ~x + qr 7Td ~x = 0
Taking the limit as ~x""O the differential equiation becomes:
d2T
~+
d 2
x
4 ~ (T g - T w) - 4~~ = 0
Tg-TW - x
Neglecting radiation and letting e = - x = - one obtains,
T g- Two' t'
d2e
- Ke = 0
d-2 -
x
The boundary conditions are:
at x = 0,
~= 0
dx
x = 0,
e = 1
(corresponds to the thermocouple
junction location)
(corresponds to the support
location)
The solution is:
Tg-TW - cosh KX
Tg-TwO - cosh K
-K
K» 1 .. 2e .
cosh KX
=
2e-K at x = 0 (the junction of the thermocouple)
where
Tg = gas temperature
T = wire temperature
KW = 4:9 d (~)2
A-3
-------
x = x/t normalized distance from junction
h = gas side heat transfer coefficient
d = wire diameter
t = wire length
k = wire thermal conductivity
Subscripts
g = gas
w = wire
o = supports
For the conditions of these measurements with the .011 inch diameter
wire, K = 38, the wire junction temperature reads the gas temperature.
Thermocouple Frequency Response
The following discussion is an analysis of the frequency response
of a thermocouple wire to the turbulence of a flame. For this analysis,
a physical model was assumed where: (1) the thermocouple wire alter-
nately heats and cools via convection; (2) the radiation losses are
neglected; and (3) the wire end losses are neglected. In the analysis to
follow, the nomenclature listed below is used.
Physical Model:
.
.
wire alternately heats
neglect radiation
neglect end losses
and cools via convection
.
d = wire diameter
p = wi re dens ity
A = wire cross sectional area
T = wire temperature
t = time
Cv = specific heat at constant volume
h = gas side heat transfer coefficient
w = frequency of fluxuation, rad/sec
f = W/27T (Hertz)
A-4
-------
k = thermal conductivity
Nu = ~d Nusselt number
j = imaginary number
a = ~ thermal diffusivity
pCv
Subscri pts
g = gas.
o = amplitude
w = wire
The energy balance becomes
dT
pA cv d~ = h (Tg - Tw) TId
dT w hd k 4
at = K . - 2" (Tg - Tw)
cd'
p v
---- --
Nu a
To solve, one assumes a sinusoidal variation in gas temperature;
Let
T = T e -jwt
g go
T = T e-jwt
w wo
- 4
(3 - Nu a -
d2
A-5
-------
Then
-Two jw = B T - B T
go wo
Two (l-jW)=l
T go B
Two 1 + jw/ B
TgO=1+~)2
To get the magnitude Two Two
of T' T
go gO
/I w/ B , Two 1
+~y T-
go ~l ~J
II
~
+ *
desire that
Because we
T
wo -1
Tgo
this implies that
Therefore,
T 1
TWO = 1 - 2"
go
~r
= 1 - E:
where
E: « 1
Then
and
,2:
f - 4 Nu a~
- 21Td2
< 1
~=
B
A-6
(*f
+ ....
« 1
-------
For the conditions encountered here
d=O.l1,
Nu = 1.5 based on
- O. 24 ft2,
a - hr
Reynolds number of 5
and specifying that
1 >
TWO
T ~ .9.
go
the frequency thatone can measure becomes
f < 34 Hz
This f corresponds to our being able to measure an eddy size A, of
A = r = 1 ~~/~~c = .029 ft (.348 in)
where u is the convection velocity. Observation of the existing flame
interferograms show that the predominant eddy size is this size or
larger so that recording should be possible.
A-7
------- |