EPA-650/2-74-031-a
April 1974
Environmental Protection Technology Series
-------
EPA-650/2-74-031-0
APPLICATION OF HOLOGRAPHIC
METHODS TO THE MEASUREMENT
OF FLAMES AND PARTICIPATE,
VOLUME I
by
A .B . Witte and D .E. Haflinger
TRW Systems Group
One Space Park
Redondo Beach, California 90278
Contract No. 68-02-0603
ROAP No. 21ADG-51
Program Element No. 1AB014
Project Officer: William B. Kuykendal
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, N. C. 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
April 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
TRW REPORT NO. 23523-6001-TU-00
APPLICATION OF HOLOGRAPHIC METHODS
TO THE
MEASUREMENT OF FLAMES AND PARTICIPATE
VOLUME I
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL PROTECTION AGENCY
Washington, D.C. 20460
111
-------
ABSTRACT
The report gives results of the application of a pulsed ruby laser
holographic interferometer to the study of flames, in hopes of extracting
temperature profile data. The principle involved is to record holograph-
ically the interferogram which presents a three-dimensional record of the
interference fringe pattern. The density profile and hence the temperature
profile can be calculated from the fringe shift information. The report
presents data for a methane-air burner operating both as a diffusion flame
and as a premixed flame. The large number of fringe shifts recorded on an
interferogram complicated the reduction of the methane-air data, but it was
possible to correlate the interferometrically derived temperature data with
thermocouple measurements. Application of the technique to a 0.2 gal/hr oil
burner was unsuccessful because the highly turbulent flame caused an inter-
ference pattern that could not be deciphered. This report was submitted in
fulfillment of TRW Project No. 23523 and Contract No. 68-02-0603 by TRW
Systems Group under the sponsorship of the Environmental Protection Agency.
Work was completed as of November 1973.
IV
-------
CONTENTS
Page
1. INTRODUCTION 1
2. EXPERIMENTAL APPARATUS 3
2.1 Hauck Burner 3
2.2 Methane-Air Burner 4
2.3 Holocamera 5
3. TEST PROCEDURE. 7
4. DATA REDUCTION PROCEDURE 9
_ i
4.1 Introduction . -; 9
4.2 Abel Inversion Technique ; 9
4.3 Comparison Model to Account for Scene Difference ... 10
4.4 Premixed Flame Fringe Shift Equation 11
4.5 Diffusion Flame Fringe Shift Equation 12
4.6 Density Calculation Procedure 15
4.7 Sample Calculations . 16
4.8 Excess Air Effect 17
4.9 Velocity Ratio and Reynolds Number 18
5. RESULTS. 21
5.1 Sequence of Progress and Problems Solved 21
5.2 Hauck Burner 22
5.3 Flame Interferometry 22
5.4 Premixed Methane-Air Flame 22
5.5 Diffusion Methane-Air FUme 35
6. CONCLUSIONS AND RECOMMENDATIONS 56
6.1 Conclusions 56
6.1.1 Interferometric Data Reduction Procedures ... 56
6.1.2 Diffusion Flame 56
6.1.3 Premixed Flame. 57
6.1.4 Oil Burners 57
6.2 Recommendations 57
REFERENCES 58
-------
ILLUSTRATIONS
Page
2-1 Sectional View of Hauck Proportioning Oil Burner
2-2
2-3
2-4
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
5-5
5-6
5-7
5-8
5-9
5-10
5-11
and Magnified View of Nozzle Section
Methane-Air Burner
Air-Methane Premix Fitting
Schematic of Holocamera
Diagram of Light Beam
Molecular Weight Versus Excess Air for Methane-Air
Combustion . . . .
Glades tone- Dale Constant Versus Excess Air for
Methane-Air Combustion
Methane-Air Burner with Annul us Insert
Infinite Fringe Interferogram of Methane-Air Burner
With Chimney
Finite Fringe Interferogram of Methane-Air Burner
With Chimney
Infinite Fringe Interferogram of Methane-Air Burner
Without Chimney
Finite Fringe Interferogram of Methane-Air Burner
Without Chimney
Interferogram of Pre-Mixed Flame With Condition
Shown in Table 5-1
Interferogram of Pre-Mixed Flame With Condition
Shown in Table 5-1
Comparison Between Equations (4-5) and (4-6) for
Interferogram Data EPA No. 5-5
Fringe Shift for Interferogram No. 5-5. .
Density Solution for Interferogram EPA No. 5-6
Fringe Shift for Interferogram EPA No. 5-6
Effect of Excess Air on Density Profiles for
Interferogram No. 5-6
3
4
5
6
14
18
19
20
23
23
24
24
25
25
27
28
29
30
31
vi
-------
ILLUSTRATIONS (Continued)
5-12 Test Interferogram for Rectangular Box with
Pyrex Windows 32
5-13 Effect of Wall Temperature on Density Profile for
Interferogram No. 5-6 33
5-14 Interferogram with Thermocouples Attached 34
5-15 Temperature of Methane Air Premixed Flame Versus Percent
Excess Air Assumed 36
5-16 Partial Pressures of Combustion Products 37
5-17 Density Profile for Interferogram No. 5-14 With
14 Percent Excess Air. 40
5-18 Schematic of Typical Fringe Pattern Divided Into Its
Three Compositional Regions 42
5-19 Reconstructed Interferogram 42
5-20 Reconstructed Interferogram 43
5-21 Reconstructed Interferogram 43
5-22 Reconstructed Interferogram 43
5-23 Reconstructed Interferogram 43
5-24 Reconstructed Interferogram 43
5-25 Reconstructed Interferogram 44
5-26 Reconstructed Interferogram 44
J
5-27 Reconstructed Interferogram 44
5-28 Reconstructed Interferogram , 44
5-29 Reconstructed Interferogram , . . 44
5-30 Reconstructed Interferogram 46
5-31 Reconstructed Interferogram 46
5-»32 Reconstructed Interferogram 46
5-33 Reconstructed Interferogram 46
vii
-------
ILLUSTRATIONS (Continued)
Page
5-34 Reconstructed Interferogram 46
5-35 Reconstructed Interferogram 47
5-36 Reconstructed Interferogram 47
5-37 Reconstructed Interferogram 47
5-38 Reconstructed Interferogram 47
5-39 Reconstructed Interferogram 47
5-40 Plot of Peak Temperatures of Reduced Interferograms
and Thermocouple Data 49
5-41 Temperatures Profiles With r 0.825 inch, Varying
Reynolds Number, Constant UA/UF 50
5-42 Temperature Profiles with r = 0.825 inch, Medium
Reynolds Number, Varying UA/UF 51
5-43 Temperature Profiles with r = 0.825 inch, High
Reynolds Number, Varying UA/UF 52
5-44 Temperature Profiles with r = 0.4 inch, Medium Reynolds
Number, Varying UA/UF 53
5-45 Temperature Profiles with r = 0.4 inch, High Reynolds
Number, Varying UA/UF 54
TABLES
5-1 Test Condition for Interferograms 26
5-2 Test Conditions for Diffusion Flame With Radius of
ftnnuTus Equal to 0.825 Inch 39
5-3 Test Conditions for Diffusion Flame With Inner Radius
of Annulus Equal to 0.4 Inch 39
viii
-------
ACKNOWLEDGMENTS
Work on this contract was accomplished by personnel of the TRW Fluid
Mechanics Laboratory. The Principal Investigator was Dr. Arvel B. Witte;
the Project Engineer, Donald E. Haflinger. The TRW Project Manager was
Birch J. Matthews. Technical direction and administration of the contract
for the Environmental Protection Agency was the responsibility of Mr. Wil-
liam B. Kuykendal.
The authors wish to thank Messrs. Frank Gomes and Robert Briones of
the TRW Fluid Mechanics Laboratory for their assistance in conducting the
holography experiments. In addition, the assistance and cooperation of Mr.
Blair Martin of the Environmental Protection Agency was greatly appreciated.
ix
-------
Metric System Conversion Table
SYMBOL
in.
ft
yd
Ib
°F
psi
WHEN YOU KNOU
inches
feet
yards
pounds
Fahrenheit
temperature
pounds per
square inch
MULTIPLY BY
25.4
0.3048
0.9144
0.453592
5/9 (after
subtracting 32)
51.71
TO FIND
millimeters
meters
meters
kilograms
Celsius
temperature
torr
SYMBOL
mm
m
m
kg
°C
torr
mm
m
m
kg
°C
torr
millimeters
meters
meters
kilograms
Celsius
temperature
torr
0.03937
3.28084
1.09361
2.20462
9/5 (then
add 32)
0.01933
inches
feet
yards
pounds
Fahrenheit
temperature
pounds per
square inch
in.
ft
yd
Ib
°F
psi
-------
1. INTRODUCTION
In situ probes may be used to measure the properties of chemically re-
acting flows encountered in premixed and diffusion flames. Generally they
will disturb the flow to produce conditions not normally existent at that
point. They also may require extensive calibration of the probe to make
accurate measurements possible. In the current study holographic interfer-
ometry was used to "remotely sense" the index of refraction, density and
temperature fields associated with these flames. No flow field disturbances
were caused in this way.
In order to characterize a flow field uniquely, species compositicn,
two state variables, and one dynamic variable must be known. An interfer-
ometer measurement yields the index of refraction field, n(x,y,z). n in
turn is a function of species mass fraction, C., and density. In equilib-
rium, C. can be expressed as a function of two state variables, for example,
pressure and temperature, when the initial species fractions are known. For
the premixed flame in equilibrium, the state variables and composition can
be determined uniquely for low speed flames (constant pressure). Data re-
duction of the diffusion flame is likely to be more complicated because, al-
though local equilibrium may apply, the species iray diffuse as well, leaving
the initially unreacted composition at a point unknown. In the present
study, techniques are developed using holographic interferometry to charac-
terize flame structure and to obtain density and temperature profiles. With
the current technique, averaged measurements are made along a line of sight.
Data which exhibit axial symmetry can be reduced to localized quantities
(radial profiles) by means of inverting the integral equation for fringe
shift, a technique established by Abel and described in Reference 1. For
the "not so axially symmetric data" either statistical averaging techniques
have to be used or the use of wide-angle holographic interferorretry and a
more sophisticated 3-D data reduction scheme must be employed. Use of a 3-D
reduction technique was first applied to holographic interferorretry by
Witte2 and later applied extensively by Matulka3. The extent to which pro-
gress toward data reduction of these flames can be made is discussed in some
detail in this study.
-------
Two burner flames were investigated. The first, produced by an oil
fuel fired burner by Hauck, yielded far too many fringes — a condition
resulting from the fact that the burner could not be operated satisfactorily
at the low range of fuel-air rates. The large number of fringes observed
was analogous to that measured in burners of somewhat larger size and
reported during an earlier EPA program. This type of burner is described
in the next section; however, the flame was not studied interferometrically
because of the fringe resolution problem mentioned above.
The second burner operated on methane and air over a wide range of
excess air and as a premixed and diffusion flame. Holographic interferome-
tric data reduction techniques are developed for reducing the interfero-
grams recorded of this burner flame.
-------
2. EXPERIMENTAL APPARATUS
2.1 HAUCK BURNER
A schematic of the Hauck Burner* is shown in Figure 2-1. This burner
has a maximum air intake rate of 28 cfm with a maximum fuel intake of 1.2
gph and 0.2 gph minimum. To operate the burner, oil must be supplied at ap-
proximately 20 psi to the burner pressure regulator. Oil pressure is reg-
ulated to the burner from 2 to 10 psi. A separate air blower is required
AIR INLET
OIL INLET ,
MICRO-METERING
OIL VALVE
QUICK-DISCONNECT
OIL VALVE LEVER
A.
B.
C.
D.
E.
F.
Single Control Lever
Inner Air Nozzle
Tangential Inlet Openings
Oil Orifices
Stationary Outer Air Nozzle
Outer Air Nozzle Opening
Figure 2-1. Sectional View of Hauck Proportioning Oil Burner
and Magnified View of Nozzle Section
product of Hauck Manufacturing Co., Lebanon, Pennsylvania.
-------
to supply air at 0.25 psig. The fuel oil was No. 2 distillate obtained
*
from Gulf crude. The oil had an API gravity equal to 36. The fuel oil
was fed to the burner from a small pressurized tank.
2.2 METHANE-AIR BURNER
A schematic of the methane-air burner is shown in Figure 2-2. The de-
sign concept for this laboratory methane-air burner was provided by EPA.
Additional information was obtained from Prof A.F. Sarofim, Massachusetts
Institute of Technology. The burner consists of two concentric cavities in-
to which the air and gaseous fuel are introduced. Both cavities are filled
with glass wool to dampen feed system fluctuations and obtain uniform fuel
distribution. The air and methane next pass through 12 gauge stainless
steel tube bundle to straighten the
flow and break up any large eddies.
A screen separates the tube bundle
and concentric cavities are housed
in a 3.60-inch-diameter tube. Gas
and air are injected through a base
plate. A retainer ring and four re-
tainer bolts complete the assembly.
A glass chimney was initially used
to preclude secondary air entrain-
ment into the burner flame. Later
the glass chimney was replaced with
a thinner quartz chimney for optical
reasons. To accomplish a premix
flame configuration, methane is in-
troduced into a 1/4-inch-diameter
tube which has twenty-five 0.059-
inch-diameter holes drilled in it.
The tube slips inside a 1/2-inch-
diameter bulkhead tee fitting carry-
ing the air. The mixer is shown in
Figure 2-3.
Figure 2-2. Methane-Air Burner
\
BOTTOM PLATE
ASSEMBLY
*An API gravity of 36 is equivalent to a specific gravity of 0.845.
4
-------
Figure 2-3. Air-Methane Premix Fitting
2.3 HOLOCAMERA
The laser holograph is comprised of a pulsed ruby laser and a holocam-
era. The pulsed ruby is used in this investigation because sufficient en-
ergy can be produced to penetrate the burner flame during a pulse width
small enough to stop the action of the combustion phenomena. The pulsed ru-
by laser has been described in detail in an earlier TRW report to EPA and
will not be discussed further here.
The holocamera design is a modified version of the greater scene depth
holocamera used in later series of tests described in Ref. (4). The basic
modifications were as follows:
1
a) The entire holocamera was rotated 90 degrees to permit in-
sertion of the burner such that burning took place verti-
cally. To have used the.previous holocamera configuration
would have resulted in contamination of the reference beam
by the burner flame and resulting by-products. The flame
produced by vertically oriented burner is more nearly sym-
metric than those at an angle to the gravity field.
b) A new finite fringe rotator to rotate the hologram verti-
cally, was designed and built to provide horizontal fringes.
-------
c) The outer pair of lenses from the quadratic lens set,
referred to in Ref. (4), was appropriately spaced to
produce a focal length of 21 inches, and apertures down
to 11 inches in diameter to correct for abbreviations.
d) Hologram and prism plate were reoriented at 52.5 degrees
and 75 degrees as shown on Figure 2-4 in order to reduce
reflections.
INCOMING
LASER
BEAM
REFERENCE BEAM
MIRROR
•~\
\ ^
O^
\ K\/
\ I \\ \S? 52.!
- i / ^-\-
-j»M««35!
HOLOGRAM
FOCUSING
LENS SET
VOLUME
Figure 2-4. Schematic of Holocamera
-------
3. TEST PROCEDURE
A test procedure was established and followed for recording holograms
and interferograms of the burner flame. Starting with the burner in place,
the following steps were taken to make the recordings:
1) Ignite burner
2) Adjust regulator pressures and flow rate
3) Turn off lights
4) Uncover film plate
5) Record the test scene holographically
6) Shut off methane input
7) Record the comparison scene holographically
8) Cover film plate
9) Lights on
10) Develop plate.
Ideally, the comparison scene should be recorded before the test scene;
however, with this burner, it was necessary to reverse the order because the
quartz chimney had to be removed in order to ignite the burner. Removing
the chimney could leave movement fringes in the resulting interferogram.
In the premixed flame configuration, thermocouple readings were record-
ed during the time the test scene was holographically recorded. Some ther-
mocouple readings were also taken at a later time period. Comparison of
thermocouple readings showed that good test-to-test repeatibility could be
achieved without any loss in accuracy.
In order to save time and also to avoid any obscuration, wall thermo-
couple readings for the diffusion flame configuration were also recorded at
u
a different time period than the interferogram recordings. The annulus (see
Section 2.3) was changed to vary the mixing region thickness as well as
overall stoichrometry. This modification simply requires lifting out the
center core of the burner.
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
""Product of Agfa-Gevaert, Antwerp, Belgium.
-------
HRP developer was typically 4 minutes for an interferogram. The acid fixer
used was Nacco fix.
Interferograms were reconstructed with a Model 124A Spectra Physics
helium neon gas laser (X. = 0.6943|i) illuminator. The recordings were photo-
graphed using Polaroid Type 52 film, having an ASA rating of 400. When de-
sired, negatives of the reconstructed scene were made using Polaroid Type
55 P/N film.
8
-------
4. DATA REDUCTION PROCEDURE
4.1 INTRODUCTION
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 microdensitometer
trace; and computing density by means of the standard equation for fringe
shift,
S(x,y) -J- / P(x, y, z) p dz (4-1)
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 reconstruc-
tion, a 15 mw continuous wave He-Ne laser (K= 0.6943 micron) was used to
illuminate the holographic interferogram. A copy camera was positioned be-
hind 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 nomi-
nal 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 neg-
atives was then recorded on a Joyce precision microdensitometer having a
slit size of about 0.010 inch by 0.010 inch.
4.2 ABEL INVERSION TECHNIQUE
Fringe number S was measured relative to the undisturbed (light fringe)
background gas, i.e., relative to S^ = 0. The first dark fringe occurs
where the change in optical path isK/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
*When many fringes are available, use of the microdensitometer may be
unnecessary.
-------
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 axi symmetric. With the assumption
of axi symmetric flow, the equation can be inverted as the Abel integral to
obtain
f* —2 ^
*>"-•-* /2 i
y
\
When this equation is rewritten as
X 9 f ds
" " ^ 2 7~2
J o dr
(4-3)
it may be cast into the following finite difference form (the Schardinvan
Voorhis approximation) for which the density is assumed constant in each of
N thin angular rings of thickness A:
M 1 / o o / 9 9
p(y, _ P = -2A. v (s s ) ^k + i)2-i2- ^2-i2
iy' *° TTAK 2, k k+1 2k + 1
k=T (4.4)
where r. =Ak, y = Ai and r =AN has been used and the indice i takes on
t\ S
the following values: i = 0, 1, 2, ..., N-l. K is the Gladestone-Dale con-
stant, e.g., for Pffl K = 2.926 x 10"4 for air when P^ = 1 atm.
4.3 COMPARISON MODEL TO ACCOUNT FOR SCENE DIFFERENCE
Accounts must be made for different chemical species between the double
exposures creating the interferograms because the comparison scene is record-
ed with air and the test scene recorded with combustion products.
10
-------
Therefore, the Abel Inversion technique was modified to accept (3 and
(3^ by using the Gladestone-Dale equation for a dilute gas. The original
program was used to solve the fringe shift equation
(4-5)
which does not account for the difference in gas composition between the
test and comparison beam recordings. By using the Gladestone-Dale relation-
ship, n 1 = P P/PS, and the basic fringe shift equation, S = •]•• j(n-nj dz,
one gets
B p r/A pco^ . \
dx (4-6)
which accounts for the difference in composition. The Gladestone-Dale con-
stant for the reaction products of methane-air combustion, S , is shown to
-4 T
be 3.208 x 10 for stoichiometric conditions.
4.4 PREMIXED FLAME FRINGE SHIFT EQUATION
For the premixed flame configuration, further modification of the
fringe shift equation had to be carried out due to the occluded region which
appears on the edges of the quartz chimney, an observation which will be
discussed later.
This occluded portion of the interferogram at the wall of the quartz
chimney had caused concern because initial calculation had indicated a cen-
terline temperature T/Tm <2 for the premixed configuration, which appeared
to be far too low. It was concluded,that the fringe shift readings within
the occluded region could not accurately be determined. One possible method
of alleviating this source of error is to measure the waTl temperature of
the quartz chimney and to include this additional information in the fringe
shift equation. By rearranging the fringe equation,
S - 4- / (n n ) dz (4-7)
CO
11
-------
into two separate integrals and considering the wall temperature, then
Now, again, using the Gladestone-Dale relationship, it can be shown that the
fringe shift equation is:
- 2
6 p
roo rco
X PC
'COQ
(4-9)
4.5 DIFFUSION FLAME FRINGE SHIFT EQUATION
In the diffusion flame configuration, the same occluded region appears
on the recorded interferograms and, therefore, the fringe shift equation has
to be modified in a similar manner as presented above. However, there is an
additional consideration that must be taken into account. In the case of
the diffusion flame, the Gladestone-Dale constant is no longer uniform across
the chimney region and, therefore, must be varied in the evaluation of the
fringe shift equation. In the outer portion of the chimney, there is air
with no unburned fuel nor combustion products. Moving towards the center of
the chimney region, combustion products are encountered, and, finally, at
the center of the quartz chimney, the gases are probably a mixture of un-
burned fuel with some combustion products. To evaluate the fringe shift
equation, the following modification (which is shown in greater detail than
previously) must be made for the different species configuration. The gen-
eral fringe number equation is
S = 4" / (n - n )
X J
Now, separating this equation into two separate integrals as before we
get
S =
J_
X.
njdz
/ ("» - ".
dz
(4-10)
12
-------
Using the Gladestone-Dale relationship, n 1 P P/PS> such that
0T
(species condition)
(wall condition)
and
n =P
00 CO
(reference condition)
and substituting these relationships into the fringe number equation, one
gets
W
TT
dz
+ 1 -
W V + P»f W
T~ )dz + T\^
CO
uo
'00
'dz
(4-11)
After some algebra
_ ^^£w/& ^^ _p .A dz + ^/\ A r
- * P.pJ P0 pw P« / ^ P«\p« / -/
dz
(4-12)
The integral / dz can be transformed into terms of r/rQ as can be seen by
referring to Figure 4-1.
1/2
dr
(4-13)
13
-------
LIGHT RAY
Figure 4-1. Diagram of Light Beam
and therefore
. B, i. i M £ Jl. ,Vu zS.ift !
Pco
rco
where
1/2-
(4-14)
S fringe shift
X = irradiation wavelength
P = Gladestone-Date constant
P = density
z = distance along light ray
subscripts: ® = comparison scene condition of air burner is turned off
T = local test scene condition (which varies for diffusion
flame configuration and is constant having combustion
product species for premixed flame configuration)
o = standard conditions at which (3's are evaluated
w = wall conditions
14
-------
4.6 DENSITY CALCULATION PROCEDURE
The procedure for both the premixed flame configuration and the diffu-
sion flame is about the same. First, correct the fringe numbers by adding
the function X . -( - r2) where,
3 P /3 «
X 2 — — I — -
-g- l) (4-15-1) (premixed)
9 °o
x.
(4-15-2) (diffusion)
Now, calculate the quantity
(4-16-1) (premixed)
w
(4-16-2) (diffusion)
Then dividing by the appropriate Y quantity one gets the density ratio,
which is the flame density normalized by the density at the conditions of
the comparison scene recording (room temperature and pressure).
The value of P is calculated as follows from the Gladestone-Dale equa-
tion for species i - 1, 2 ... N. For any dilute gas,
n - 1
Sp.
— J-
poi
(4-17)
P.
(4-18)
15
-------
where P. is evaluated for species i and C- is the mass fraction. However,
by definition
n - 1 = PT
(4-19)
Or
Therefore, combining these two equations yields
C..P,
(4-20)
n
Or
roi
P - can be shown to be 0.00279 m. Ib_/ft at atmospheric pressure and 0°C
where m. is the molecular weight of species i.
4.7 SAMPLE CALCULATIONS
As an example, consider the chemical reaction of premixed methane-air
combustion at stoichiometric conditions
CH,
2"02 + 2'(3.76)N2-C02 + 2'HgO + 2'(3.76)N2 (4-21)
From this reaction, the following is calculated:
Reactions
Products
co2
H20
N2
Number of
Moles (N)
1
2
7.34
Molecular
Weight (m)
44
18
28
N x m
44
36
205.52
285.52
Mass fraction
(c,)
0.1541
0.1261
0.7198
1 .0000
Reactions
Products
C0£
N,
ci
0.1541
0.1261
0.7198
P,
4.51 x TO"4
2.54 x 10"4
2.97 x 10~4
P°i
0.12276
0.05022
0.07812
Vp0l
3.674 x TO'3
5.058 x 10"3
3.802 x 10"3
c. P
1 °i
.0189
.00633
.05623
"iV.,
5.662 x IQ'4
6.378 x TO"4*
2.737 x 10"3
16
-------
The following useful parameters can also be calculated:
Parameter
Equation
Value
CO
1/m
in
p /P
0}
-------
28.5
28.0
27.5
CH4 * 2 • 0+X) • (3.76) • N2
- CO2 + 2-H20 + 2-(HX)-(3.76)N2
10
20
30
40
50
60
70
80
90
100
Figure 4-2.
AMOUNT OF EXCESS AIR, X(%)
Molecular Weight Versus Excess Air for
Methane-Air Combustion
where X is the amount of excess air. Excess air is usually defined in terms
of percent of theoretical air. such that:
AIR
actual
AIR
theoretical
AIR
x 100%
(4-23)
theoretical
The amount of excess air also affects the Gladestone-Dale constant, p ,
for the resultant reaction product which is shown in Figure 4-3.
4.9 VELOCITY RATIO AND REYNOLDS NUMBER
For diffusion flames, two parameters should be defined; (1) the veloc-
ity ratio, and (2) Reynolds number.
The velocity ratio is defined as
U Q A,.
a a f _
(4-24)
Qf *
18
-------
3.70
3.18
3.16
3.14
3.12
3.10
3.08
10
20
30
40
SO
70
80
90
100
AMOUNT OF EXCESS AIR, X(°il
Figure 4-3,
Gladestone-Dale Constant Versus Excess Air
for Methane-Air Combustion
where r is the annulus radius, R is the burner radius (see Figure 4-4). Q,
a
is the air flow rate and Qf is the methane flow rate.
The Reynolds number is defined as
(4-25)
where the density and viscosity are defined at room temperature and pressure,
u, is the average velocity of air and methane (u = [uft + uf]/2), and L is
the distance above the burner surface at which the data are being reduced.
19
-------
t
TTT
AIR METHANE AIR
FUSED
. QUARTZ
CHIMNEY
BURNER
SURFACE
TUBES
GLASS WOOL
• ANNULUS
ANNULUS
Figure 4-4. Methane-Air Burner
With Annul us Insert
20
-------
5. RESULTS
5.1 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 indica-
ted:
• Methane burner was fabricated for use in creating both premixed
and diffusion flames.
• Finite fringe and infinite fringe interferograms were
recorded in diffusion flame configuration.
• Occluded region was greatly reduced.
• Laminar premixed flame was easy to produce.
t Finite fringe interferograms were recorded of premixed
f1ame.
t Modification of data reduction procedure to eliminate
occluded region problem was accomplished.
• Center!ine and wall thermocouple temperature measurements
were made of the premixed flame.
• Data reduction of the interferograms for the premixed
flame was accomplished.
• Finite fringe interferograms were recorded of diffusion
f1ame.
• Film fogging problem (during high flow rates in diffusion
flame tests) solved with filters.
\
• Wall thermocouple temperature measurements were made of the
diffusion flame.
• Data reduction of the interferogram for diffusion flame was
accomplished.
• Both laminar and turbulent flames were observed in the dif-
fusion flame configuration.
21
-------
5.2 HAUCK BURNER
A small size (as low as 0.2 gph) fuel oil-fired burner was selected for
study because it represents a compromise between the large (1 gph) burners
tested during a previous EPA study and the methane-air burner which is an
approximation to the oil-fired burner for the purpose of instrumentation
development. During testing of this burner, it was found that the flow rate,
turbulence, and illumination were much larger than expected. It was then
decided that the concentration of efforts should be placed on the methane-
air burner, because no new information could be obtained with this Hauck
burner.
5.3 FLAME INTERFEROMETRY
Holographic interferograms of methane-air combustion were recorded to
ascertain what problems might be encountered in obtaining good quality re-
construction. Infinite and finite fringe recordings were made both with and
without the glass (Pyrex) chimney. Figures 5-1 through 5-4 show the results
of these initial interferograms of a diffusion flame. These figures indi-
cate that finite fringe interferograms would be best suited for this study.
Also, the requirement to control the ratio of fuel to oxidizer dictated the
need to use the chimney even though the occluded region caused additional
problems in reducing the fringe pattern. Note condensation in Figure 5-1.
5.4 PREMIXED METHANE-AIR FLAME
In the premixed configuration, initial tests were conducted under
stable flame conditions. A stable flame is defined as a steady blue -violet
flame which is attached at the burner surface. It was found that when the
*
fuel mixture was too rich , the flame would rise up the chimney and go out.
In the opposite condition, when the fuel mixture was too lean, the flame
had the tendency to be "blown" out. Over the range of allowed operating
conditions a laminar-like flame was observed.
Initial holographic interferograms recorded are shown in Figures 5-5
and 5-6, with test conditions shown in Table 5-1. These interferograms are
used to show the effects of the various modifications of the data reduction
procedure results.
*
"Rich" in this case may be a misleading term since all flames were
operated with some excess air. This burner would not operate under true
rich fuel conditions.
22
-------
Figure 5-1. Infinite Fringe Interferogram
of Methane-Air Burner With
Chimney
Figure 5-2. Finite Fringe Interferogram
of Methane-Air Burner With
Chimney
23
-------
Figure 5-3. Infinite Fringe Interferogram
of Methane-Air Burner Without
Chimney
Figure 5-4. Finite Fringe Interferogram
of Methane-Air Burner Without
Chimney
24
-------
Figure 5-5. Interferogram of Pre-mixed
Flame With Condition
Shown in Table 5-1
Figure 5-6. Interferogram of Pre-mixed
Flame With Condition
Shown in Table 5-1
25
-------
Table 5-1. Test Condition for Interferograms
Hologram
Number
5-5
5-6
Air
Regulator
Pressure
30 pslg
30 psig
Flow
Meter
16800
cc/
min
16800
cc/
min
Methane
Regulator
Pressure
30 psig
30 psig
Flow
Meter
1200 +
cc/
min
93Q
cc/
min
Conditions
S.S. screen
quartz
stack
No screen
stack on
Remarks
Air flowing
for no fire
condition
Low flame
no popping
A comparison between Equations (4-5) and (4-6), as shown in Section 4,
is made in Figure 5-7. The solid curve is the density profile for inter-
ferogram 5-5 using Equation (4-5) for which hot and cold combustion gases
occur in the test and comparison scenes, respectively. The dashed curve
represents Equation (4-6) in which account is made of the change in index
between scene recordings. The fringe shift profile used for this com-
parison is shown in Figure 5-8. As can be seen in Figure 5-7. this first
modification tends to slightly lower the resulting density profile as com-
pared with the original fringe shift equation.
Note the discontinuity of the density profile at the centerline.
This discontinuity (r/r 0 of Figure 5-7) can be explained by examining
the S-curve of Figure 5-8. The maximum in the S-curve is to the left of
center, such that ds/dr , _ Q * 0. The inverted fringe shift equation
for density °
p(y) B
i -
00
TTKD
CO
i:
dr£
(r2 - y2)
V2
shows that with a negative ds/drl. _ n the density will be initially
r/rQ - u
larger than if the ds/dr^r _ Q is positive. Thus, the density is dis-
continuous as r approaches r = 0, a condition which is physically impossible
Interferograms will be identified by their figure number.
26
-------
P
_P_
Poo
Figure 5-7. Comparison Between Equations (4-5) and (4-6) for
Interferogram Data EPA No. 5-5
-------
r
ro
Figure 5-8. Fringe Shift for Interferogram No. 5-5
for this flow problem. Figure 5-9 shows a similar phenomenon for inter-
ferogram 5-6. In this figure, the solid curves indicate the computed
density across the quartz chimney with the same discontinuity shown at
r/rQ = 0. For axisymmetric flow, ds/dr is zero. Therefore, to eliminate
the density discontinuity, the left and right sides of the S-curve were
averaged. The dashed curve in Figure 5-9 is the result of S-curve aver-
aging. Figure 5-10 shows the actual fringe shift and the average fringe shift
for interferogram 5-6.
The temperature ratio T /T is just the inverse density ratio P/p
co reo
since the molecular weights are nearly the same. The centerline tempera-
ture ratios for these two interferograms appear far too low, T/T^ 2. This
problem was discussed earlier in Section 4 and will be given further
treatment later on in this section.
For the density calculations up to this point it was assumed that
stoichiometric conditions exist. (See Section 4.8 and Figures 4-1 and
4-2). To show what effect the amount of excess air has on the calculation
28
-------
ro
AVERAGED
SOLUTION
1.0
.6
.4
.2
0
r
.2
.4
.6
.8
1.0
Figure 5-9. Density Solution for Interferogram EPA No. 5-6
-------
—2
1.0 .8
.6
.4
.2
.2 .4
.6
..8
Figure 5-10. Fringe Shift for Interferogram EPA No. 5-6
of the density profile, a sample case was done using interferogram 5-6,
The results are shown in Figure 5-11. The solid curve is the density
profile as defined by stoichiometric conditions. The dashed curve is the
density profile with an assumed 100 percent excess air (twice the amount
of air needed for stoichiometric conditions). Allowing for the amount
of excess air will lower the density for any one fringe shift profile.
In Figures 5-1 and 5-2, about 50 percent scene within the Pyrex
chimney region is occuluded and, therefore, it is impossible to make any
determinable conclusion about the fringe shift profile in the occluded
portion of the chimney. The Pyrex chimney has a wall thickness of 2.5
millimeter. By replacing this chimney with a fused quartz chimney with
a wall thickness of 1.5 millimeter, the occluded regions were sufficiently
decreased even though some waviness in the wall is observed. The result of
this change decreased the occlusion to about 10 to 20 percent.
Since the quartz chimney also produced some occlusion, it was decided
to use a rectangular box with Pyrex windows, used on an earlier EPA pro-
gram, to circumvent the occluded areas produced by the edge of the round
quartz chimney. An adapter was designed and fabricated to permit the
30
-------
1.0
STOICHIOMETRIC CONDITIONS
100% EXCESS AIR CONDITIONS
P_
P.
Figure 5-11. Effect of Excess Air on Density Profiles for Interferogram No. 5-6
-------
mounting of the glass sealed box around the methane-air burner. Figure
5-12 shows an interferogram recorded with this configuration. The chief
problem encountered with this approach is that the flame has a tendency
to "wander" about inside the box, a characteristic which is not conducive in
providing axisymmetric phenomona. For this reason, the rectangular box
approach was abandoned.
It was noted earlier that data reduction procedures up to this point
produce center!ine temperatures which were too low. Referring to the
modification in Section 4.4, a sample calculation was made using the data
recorded in interferogram 5-6 to illustrate the feasibility of this
*
procedure. It was assumed that the wall temperature was T /T^ = 2 and
that stoichiometric combustion occurred.
A comparison between previous methods and this approach is shown in
Figure 5-13. The centerline temperature is now calculated to be
which is in a more probable temperature range.
Figure 5-12. Test Interferogram for Rectangular Box
with Pyrex Windows
*
Thermocouple data support this value.
32
-------
oo
Co
.8 -
P
pob
.6
.5
.4
.3
.2
CD
CD
.1 .2 .3 .4
.5 .6
.8
.9 1.0
Figure 5-13. Effect of Wall Temperature on Density Profile for Interferogram No. 5-6
-------
To apply the data reduction method above, chimney wall and gas center-
line thermocouples were installed. The two wall thermocouples were of
copper-constantan type. The thermocouples were bonded with sauereisen
to each side of the fused quartz chimney at the scene centerline. The
wall temperature was measured to be 456°F. The fused quartz chimney
centerline thermocouple was of a chromel-alumel type. The gas centerline
temperature was measured to be 1670°F. The air and methane were measured
as 49870 cc/min and 4578 cc/min, respectively.
Two interferograms were taken at the same test conditions as were
reported above. Also, additional thermocouple readings were recorded to
insure repeatability of the initial test. These readings were well within
1 percent of the original readings. Figure 5-14 shows one of the inter-
ferograms recorded and used for data reduction purposes. All three thermo-
couples can be seen in this interferogram.
In order to validate the temperature calculated by the fringe shift
recorded on the interferograms, comparisons were made to the gas temperature
readings that were recorded from the thermocouple measurements and those
computed from a theoretical adiabatic flame temperature calculation for
a premixed methane-air flame. The method used included the dissociation
of the combustion products over a range in percentage of excess air. This
Figure 5-14. Interferogram with Thermocouples Attached
34
-------
method is described in a book entitled "Flames" by A. 6. Gaydon and H. G.
Wolfhard. The results of this computer program are shown in Figures 5-15
and 5-16. In Figure 5-15, the temperature decreases as the percentage
of excess air is increased from stoichiometric conditions (zero percent
excess air) to 300 percent excess air. In Figure 5-16, the various com-
bustion products are shown in terms of their individual partial pressures
and how they are related to the percentage of excess air. Both the tempera-
ture and partial pressure of the combustion products have been verified
at stoichiometric conditions. In Figure 5-17, the density profile for
interferogram 5-14 is shown. The results of the measured fringe shift
from this interferogram indicates a centerline temperature of 1777°F, which
is much lower than the theoretical temperature at the indicated value of
14 percent excess air for this case. However, the fringe shift calculation
of temperature is higher by 6 percent than the recorded thermocouple read-
ings which are expected to be low because neither radiation nor heat con-
dition.loss corrections were made. The differences between the theoretical
temperature and those of the fringe shift calculations and the thermocouple
temperature reading may be caused by heat loss to the surroundings since
the measurements were taken 3 inches above the actual burning surfaces.
5.5 DIFFUSION METHANE-AIR FLAME
Since a turbulent flame could not be produced in the premixed con-
figuration, diffusion flames were studied to: (1) discover if a turbulent
flame could be produced and a data reduction scheme provided for it, and
(2) determine what effect, if any, fuel injection radius and air annulus
radii would have on the temperature profiles.
i
In the study of effect of annulus radius, two annuli of different
radii were used. The choice of annulus radii was somewhat arbitrary.
However, to make a comparison of the effect that the annulus radius has on
temperature profiles a significant difference in radius should be used.
Thus, the inner annulus radii were selected to be 0.825 and 0.4 inch.
Interferograms were taken at five different velocity ratios obtained by
varying the amount of excess air at a given fuel rate. The range of excess
air provided by the larger annulus radius is from 0 to 300 percent. Also,
it was attempted to vary the Reynolds number by two orders of magnitude.
35
-------
3000
T<«F)
2000
Ok
1000
100
200
300
%EXCESS AIR
Figure 5-15.
Temperature of Methane Air Premixed Flame Versus Percent Excess Air
Assumed: Dissociation of combustion products and no losses due to
radiation, conduction nor convection)
-------
50 100 150 200
Percent of Excess Air
250
300
Figure 5-16. Partial Pressures of Combustion Products
37
-------
Then, using the smaller annul us, a comparison series of tests was conducted.
The flow rates for this series of tests were taken such that the velocity
ratios remained the same, and the velocities of the individual gases were
also the same between the two test series. However, the range of excess
air was altered as dictated by the velocity requirements stated above.
In order to maintain the same velocities between these series of
tests, the volumetric flow rates for each gas were adjusted according to
the following:
where subscripts 1 and 2 refer to test series 1 and 2, respectively, and
where Q is the flow rate and A is the area. By expressing this relation-
ship in terms of the respective radii, one obtains for each gas,
n(R2 - r 2)
Q2 - Q,—5 2 (A1r)
2 ] rr(R2 r/)
and
'(Methane)
Therefore, when using the smaller inner radius (r« = 0.4) for the second
test series, the air flow rate was increased by a factor of 1.39, and the
methane flow rate was decreased by a factor of 4.25. This considerably
increased the range of excess air from 489 to 2260 percent for the second
series of tests. The test conditions are shown in Tables 5-2 and 5-3.
These tables also include wall thermocouple temperature results and some
visual flame descriptions.
38
-------
Table 5-2. Test Conditions for Diffusion Flame with Radius
of Annul us Equal to 0.825 Inch
Reynolds No.
Factor
1
2
3
*,*
•; Excess
fir
0
50
100
ZO'J
300
0
50
100
200
300
0
50
100
200
300
Flow Rate (cc/min)
Air
945
1,120
1.8JO
2,340
3,780
9,450
14,200
18,900
28,400
37,800
04,500
142,900
189.JOO
284 ,000
373,000
1 'ethane
99.1
991.0
9910.0
Velocity
Patio
u la.
a T
3.15
4.72
G.29
9.45
12.58
3.15
4.72
6.29
9.45
12.58
3.15
4.72
6.29
9.45
12.53
He
T
158
483
662
G45
1,210
1 ,580
1,330
6.S20
0,450
12, ua
15,800
Ti
T Very turbulent flame
39
-------
.6
.5
.4
-p- .3
roo
.2
.2
.4 .5
.6
.7
.8
.9 1.0
Figure 5-17.
Density Profile for Interferogram No. 5-14 With
14 Percent Excess Air
-------
For the interferogram data reduction procedure of the diffusion flame,
the chimney cross section was divided into three regions (as shown in
Figure 5-18):
1) The outer most region of the annulus was considered pure air.
2) The next region (moving towards the centerline) was identified
as the stoichiometric combustion zone.
3) The centerline region was considered as pure unburned methane.
Using this description of zones in initial evaluation of temperature
(results not shown), it was found that when the third region was assumed
to be pure methane, very low temperatures at the centerline were computed
(T/Tg, < 1). Also, the temperature profile would markedly differ in the
third region if the thickness of this region was varied. To explain
further, the temperature would "drop" as the temperature profile passed
through the second region to the third and, therefore, showed a dependency
L '
on the selected location of the boundary between these regions. This
problem does not occur between the first and second region and was tested
by shifting the location of their boundary.
Since the location of the third region boundary greatly influenced
the temperature in the third region, it was not possible to define this
region clearly. However, the location of the third region does not affect
the calculation of the peak temperature in the temperature profile because
the composition at the peak temperature is dominated by air and combustion
products whose index and Gladestone-Dale relationships do not differ
significantly. This problem cannot be solved using interferometry alone,
but requires a separate measurement of gas composition.
Figures 5-19 through 5-29 are interferograms of the diffusion flame
with r = 0.825 inch and with excess air, x, and velocity ratio, ug/uf as
shown. The peak temperature, taken from the reduced temperature profile
47
-------
TYPICAL
FRINGE
PATTERN
Figure 5-18.
Schematic of Typical Fringe Pattern Divided
Into Its Three Compositional Regions
Figure 5-19
r = 0.825
u = 0.0267
^ = 12.58
x = 300%
I = 3.13
Figure 5-19. Reconstructed Interferegram
-------
00
Figure 5-21
r =
U =
i i ~
uf
x =
I =
0.825
0.0817
3.15
0%
2.89
r =
u =
=
«*f
X =
0.825
0.112
4.72
50%
4.0
Figure 5-22
r = 0.825
u = 0.143
$ = 6.29
x = 100%
I- = 3.03
Figure
r =
u
"a _
uf
x =
I =
5-23
0.825
0.205
9.45
200%
3.3
Figure
r =
u =
uf
x =
I =
5-24
0.825
0.267
12.58
300%
4.0
Figures 5-20 through 5-24. Reconstructed Interferograms
-------
Figure
r =
U =
ill =
uf
x =
I-
5-25
0.825
0.817
3.15
0%
4.30
Figure
r =
u =
"a _
uf
x =
t =
5-26
0.825
0.817
4.72
50%
3.65
Figure
r =
u =
ua _
uf
x =
1 =
T.O
5-27
0.825
1.12
6.29
100%
3.97
Figure 5-28
r = 0.825
u
= 1.43
= 9.45
"f
x = 200*
I = 3.0
Figure 5-29
r = 0.825
u = 2.05
^ - 12.58
"f
x = 300%
4.41
1 =
Figures 5-25 through 5-29. Reconstructed Interferograms
-------
data, is also shown in the legend of each figure and is plotted later on.
The sequence of photographs are in order of increasing Reynolds number.
Fringe shift near the wall is minimal, indicating small changes in tempera-
ture. Proceeding inward towards the center!ine, the abrupt fringe shift
results from the lower density, higher temperature "flame sheet" occurring
at the air-fuel interface. The trend with increasing Reynolds number which
is directly proportional to
is that the flame width decreases, a result consistent with a laminar mixing-
reaction zone. The interferometric data for r = 0.4 inch are shown in
Figures 5-30 through 5-39 and exhibit the same trends as the data at
r = 0.825 inch.
During some preliminary work conducted before actual interferometry on
diffusion flames, thermocouple measurements were taken. The thermocouple
was located about 0.25 inch above the burner surface and emersed at the edge
of the visible flame front. This location was selected after some probing
was done to find where the maximum temperature occurs. (It should be noted
that when reducing the interferogram data, it was not always possible to
select a fringe which corresponds to the thermocouple location.) The
results of the thermocouple measurements indicate only a slight increase in
temperature with increase of excess air. The phenomenon differs greatly
with the findings of premixed flame configuration which indicated a sub-
stantial temperature dependence on the percentage of excess air.
45
-------
CTl
Figure 5-30
r = 0.4
u = 0.0817
Hi = 3.15
uf
x = 489%
T
T_
5.5
Figure 5-31
r = 0.4
u = 0.112
Hi = 4.72
uf
x = 786%
I = 3.95
Figure 5-32
r = 0.4
U = 0.143
Hi = 6.29
uf
x = 1080%
T = 4.17
Figure 5-33
r = 0.4
0 = 0.205
Hi
uf
= 3.33
Hi = 9.45
x = 1680%
T
Figure 5-34
r = 0.4
u = 0.267
Hi = 12.58
x = 2260%
I - 3.33
Figures 5-30 through 5-34. Reconstructed Interferograms
-------
Figure 5-35
r = 0.4
u = 0.817
Si- - 3 IS
uf J>lt>
= 489%
= 3.16
Figure 5-36
r = 0.4
1.12
4.72
u
Hi
uf
x = 786%
± =2.30
Figure 5-37
r = 0.4
u = 1.43
Sf • «•»
x = 1080%
2.40
Figure 5-38
1 =
L
r
U
"f
x
T
TL
0.4
2.05
9.45
1680%
3.10
Figure
r =
u =
uf
x =
1
t
5-39
0.4
2.67
12.58
2260%
4.68
Figures 5-35 through 5-39. Reconstructed Interferograms
-------
Figure 5-40 shows the peak temperature as a function of averaged
streams' velocities,"!?, which provides a Reynolds number per unit length range
of two orders of magnitude, varying from~160 to 16,000 per foot. The mean
temperature appears to be about I/Ten ~ 3.5 or T = 300 x 3.5 = 1050°K
(1430°F). Disallowing the one high data point near "u" = 0.08 ft/sec, vari-
ations of the data about the mean temperature increase with u or Re. This
observation is consistent with the idea of an unstable laminar flame sheet
in the transition region. Recall that at the highest Reynolds numbers the
flame looks turbulent. See especially the highest Reynolds number interfero-
gram data at r 0.4.
The solid curve in the intermediate range of IT represents flame thermo-
couple measurements. Good agreement is obtained with the interferometric
data. No corrections were made to the thermocouple measurements for
radiation losses. By using the radiation correction model developed on the
earlier EPA contract, a correction to increase the measured temperature by
about 20 percent is indicated. This correction would place the measurements
very near the upper range of the interferometric data rather than at the
average.
Local temperature profile data are shown in Figures 5-41, 5-42 and
5-43 at r = 0.825 inch and Figures 5-44 and 5-45 at r = 0.4 inch. Shown in
the legend are also the values of excess air, x, velocity ratio u,/u^, and
or
average velocity, u. Examination of these data reveal the following
trends:
a) Strong TT (Reynolds number) dependence on the peak tempera-
ture location is observed for constant ua/uf and x for a
102 range in Reynolds number. As Re increases, the flame
sheet moves towards the air side. This result is Indicative
of what might be expected of a turbulent flame or a
"transitional" flame (Figure 5-41).
b) For each r value of the annulus (compare Figures 5-42 and
5-43 and 5-44 and 5-45), this same effect is observed inde-
pendent of velocity ratio of excess air.
48
-------
6 r
D
5 ~
oo
D
o
o
o
o
D
o
8
vo
O r = 0.825
D r = 0.4
THERMOCOUPLE DATA
10
-2
10
-1
J( FT/SEC
Figure 5-40. Plot of Peak Temperatures of Reduced Interferograms
and Thermocouple Data
-------
X UA/UF U
1 300 12.58 .0267
2 300 12.58 .267
3 300 12.58 2.67
Figure 5-41. Temperature Profiles With r = 0.825 inch, Varying Reynolds Number,
Constant UA/UF
-------
K.O
O.I'...
1.2 0.3
RxRCO)
0.
0.5
0.6
0.7
0.8
0.9
1*0
Figure 5-42. Temperature Profiles With r = 0.825 inch, Medium Reynolds Number,
Varying UA/UF
-------
ro
X UVUF U
0 3.15 0.825 B
4.72 1.12
1.43
2.05
2 50
3 100 6.29
4 200 9.45
5 300 12.58
IllJi!! ii
:» ,:: .: :. : :!••:.::;:;:::::::::::::::::::::::::::::::::::;::::::
<' : i: : r:;\: "::::::::: :i
IX i i _.- ..i.. ,.••• . ,
*« " • 5 • «•!*"••!••••••••••••• • : ... . •
." • " - t - «••••••••!•••••••••••••••!•••••••••••••••••••!•••••••
• " •••••••••••••••••••••••••••••••••••••••••••••••!••••••«
• • • • I •I•••••«••••••••••••••••••••••••••P«*••••!•••••••••••••
• I • • • •••••••••••••••••••••••••••••••••••••••••••••••!•••••••
: : :: : i:::::::::::::::::::::::::::::::::::::::::::::::::
: : : : : ::::::::::::::::::::::::::::::::i:::::::::::::::::::::::
0.0 O.J
0.2 0.3 Q.* 0.5
RsRlO)
= 0.825 INCH
0.6
Figure 5-43. Temperature Profiles With r = 0.825 inch, High Reynolds Number
Varying UA/UF
-------
U1
CO
X U/VUF
| 1. 489 3.15 .0817
•2. 786 4.72 .112
3.1080 6.29 .143
]4. 1680 9.45 .205
J5. 2260 12.58 .267
r=0.4 INCH
0.2 0.3
RxFUO)
0.5
0.6
0.7
o.a
0.9
1.0
Figure 5-44. Temperature Profiles With r
Varying UA/UF
= 0.4 inch, Medium Reynolds Number,
-------
01
-pi
X UA/UF U
1 489 3.15 .817
2 786 4.72 1.12
3 1080 6.29
4 1690 9.45
5 2260 12.58
:o.o o.i
0.8
0.9
1.0
Figure 5-45. Temperature Profiles With r = 0.4 inch, High Reynolds Number,
Varying UA/UF
-------
c) For a given Re, considerable variation in position of peak
temperature signal is present. This is believed to be
caused by the fluctuating or transitional character of the
flame zone.
d) No clear difference exists between the r = 0.4 and 0.825 inch
cases.
55
-------
6. CONCLUSIONS AND RECOMMENDATIONS
6.1 CONCLUSIONS
6.1.1 Interferometrlc Data Reduction Procedures
a) The basic equation for fringe shift was inverted to provide
unique determination of the density field for the case of
axisymmetric phenomena, which turned out to be a fairly
good approximation in most cases. Modifications were
included for differences in molecular weight and refractive
index between the reference and test scenes of the burner.
b) A data reduction model was developed which allowed the
reduction of interferometric data when the chimney wall
became hot. Basically, the procedure allowed the incorpora-
tion of a thermocouple wall temperature measurement which
obviated the high fringe density and internal reflection
present at the chimney wall for the premixed flame.
c) It was shown by calculation that errors in knowledge of the
composition for the premixed flame produced small errors in
the final computed density and temperature field because
the index, index-density, and density-temperature relation-
ships were dominated by those properties of air, at least
for those conditions investigated here for stoichiometric
and excess air combustion.
d) For the data reduction model developed for the diffusion
flame, it was found that the air side of the flame could
be computed somewhat independently of knowing what the
actual species composition was in the diffusion layer
whereas the features of the fuel side of the sheet were
strongly dependent on the composition in the diffusion
layer between the fuel and the flame sheet. Thus, the
overall variation of index, index-density, and density-
temperature across the flame is too large to expect an
interferometric measurement alone to provide temperature,
at least for the fuel side of the flame sheet.
6.1.2 Diffusion Flame
a) The diffusion flame was operated over a Reynolds number
range ojMOO. Reynolds number per unit length,
Re/L = u/y, was varied by increasing both air and fuel flow
rates and by increasing the air flow rate (increasing
percent excess air) at fixed fuel rate.
b) Variations about the peak temperatures calculated from
these data increased as a function of Reynolds number.
This phenomenon is believed to be caused by the increas-
ing frequency of unstable spots (turbulent spots) in the
flame sheet.
56
-------
c) Thermocouple data provided excellent agreement with mean
of the interferometric data.
d) The large range in Reynolds number investigated provided
data for the diffusion (laminar) flame and at the highest
two Reynolds number, data for the turbulent flame.
6-1.3 Premixed Flame
Largely because of the extensive work done to provide a suitable data
reduction procedure for the case of the premixed flame, i.e., to account for
the occluded region at the chimney wall, data were recorded and reduced at
only one Reynolds number and one value of excess air. Good agreement was
obtained between the thermocouple inserted in the flame and the interfero-
metric data albeit the measurements were much lower than the adiabatic flame
temperature.
6.1.4 Oil Burners
At present, holographic interferometry does not appear applicable to
highly turbulent-complex-real flames (i.e., oil burners). However, labora-
tory flames that are in the transitional or slightly turbulent regime can
be evaluated by this technique.
6.2 RECOMMENDATIONS
a) Future studies should be conducted with a wrap-around
180-degree viewing-angle holographic interferometer to
provide adequate three-dimensional coverage so as to allow
accurate measurement of strongly asymmetric flames caused
by either or both turbulence or asymmetric burner
configurations.
b) Diffusion flame cannot be analyzed quantitatively by
interferometry alone. A mass composition probe is neces-
sary to provide data on the fuel-side diffusion zone of
the flame sheet. It does not appear that a complicated
non-equilibrium sampling procedure is justified since the
index of the non-equilibrium products properly averaged
may be equal to the equilibrium value. Thus, a simple
aspirating probe may be adequate.
c) It is strongly recommended that the feasibility of chemi-
luminescent spectroscopy be examined to measure local
chemical species concentrations as well as temperature
profiles in flames of this type. This technique has re-
cently been developed and used successfully on several
chemical laser programs at TRW. Its chief advantage lies
in being able to make the measurement remotely without
disturbing the flow.
57
-------
REFERENCES
1. F. 0. Weyl, "Analytical Methods in Optical Examination of Supersonic
Flow," Navord Report 211-45, Navy Department Bureau of Ordnance,
Washington, D.C., 11 December 1945.
2. A. B. Witte, "Three Dimensional Flow Field Analysis by a Holographic
Interferometry," Final Technical Report, TRW Report No. 12414-6005-RO-OO,
15 February 1971.
3. R. D. Matulka and D. J. Collins, "Determination of Three-Dimensional
Density Fields from Holographic Interferograms," Journal of Applied
Physics, 42, 1109, March 1971.
4. A. B. Witte and B. J. Matthews, "Laser Holography Study of Oil-Fired
Burner Combustion," EPA Contract CPA 70-4, TRW Report No. 14103-604-RO-OO.
5. A. G. Gaydon and H. G. Wolfhard, Flames. Their Structure, Radiation
and Temperature, 3rd edition, Chapman and Hall Ltd., London, pp. 288-304,
1970.
6. R. Friedman and J. B. Levy, Combustion and Flames. 7, 195, 1963.
58
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
\, REPORT NO.
\, . .
EPA-650/2-74-031-a
3. RECIPIENT'S ACCESSION>NO.
4. TITLE AND SUBTITLE
Application of Holographic Methods to the Measurement
of Flames and Particulate, Volume I
5. REPORT DATE
April 1974
6. PERFORMING ORGANIZATION CODE
11982
7. AUTHOH(S)
A..B. WitteandD.E. Haflinger
8. PERFORMING ORGANIZATION REPORT NO.
23523-6001-TU-OO
9. PERFORMING OROANIZATION NAME AND ADDRESS
TRW Systems Group
One Space Park
Redondo Beach, CA 90278
10. PROGRAM ELEMENT NO.
1AB014; ROAP 21ADG-51
11. CONTRACT/GRANT NO.
68-02-0603
12. SPONSORING AGENCY NAME AND ADDRESS
SPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
The report gives results of the application of a pulsed ruby laser holo-
graphic interferometer to the study of flames, in hopes of extracting temperature
jrofile data. The principle involved is to record holographically the interferogram
which presents a three-dimensional record of the interference fringe pattern. The
density profile and hence the temperature profile can be calculated from the fringe
shift information. The report presents data for a methane-air burner operating both
as a diffusion flame and as a premised flame. The large number of fringe shifts
recorded on an interferogram complicated the reduction of the methane-air data, but
it was possible to correlate the interferometrically derived temperature data with
thermocouple measurements. Application of the technique to a 0.2 gal. /hr oil burner
was unsuccessful because the highly turbulent flame caused an interference pattern
:hat could not be deciphered.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Pollution
holography
interferometers
Femperature Measurement
jlame Photometry
Combustion
Diffusion Flames
Air Pollution Control
Holographic Ihterfero-
metry
Flame Measurement
Combustion Research
Premised Flames
13B
14B
20F
21B
9.-DISTRIBUTION STATEMENT
19. SECURITY CLASS (Tliis Report)
Unclassified
21. NO. OF PAGES
69
Unlimited
20. SECURITY CLASS (Tills page)
Unclassified
EPA Form 2220-1 (9-73)
59
------- |