\

AEDC-TR-91-1 ,

Liquid Film Cooling in Rocket Engines

William M. Grisson
Physics Department
Morehouse College
Atlanta, Georgia 30314

Match 1991

Final Report for Period June 1987to December 1989

I Approved foe public release; distribution I$ unlimited. I

PROPERTYOF U.S. AIR FO~C~

TECHNICAL REPORTS TECHNICALLIBRARY

FILE C OP_Y.
.,~,,., . , u ~ ~,,~,. '.,'m~d

ARNOLD ENGINEERING DEVELOPMENT CENTER

ARNOLD AIR FORCE BASE, TENNESSEE
AIR FORCE SYSTEMS COMMAND
UNITED STATES AIR FORCE
 NOTICES

When U. S. Government drawings, specifications, or other data are used for any purpose
other than a def'mitely related Government procurement operation, the Government thereby
incurs no responsibility nor any obligation whatsoever, and the fact that the Government
may have formulated, furnished, or in any way supplied the said drawings, specifications,
or other data, is not to be refJh"ded by implication or otherwise, or in any manner licensing
the holder or any other person or corporation, or conveying any rights or permission to
manufacture, use, or sell any patented invention that may in any way be related thereto.

Qualified users may obtain copies of this report from the Defense Technical Information
Center.

References to named commercial products in this report are not to be considered in any
sense as an endorsement of the product by the United States Air Force or the Government.

This report has been reviewed by the Office of Public Affairs (PA) and is releasable to
the National Technical Information Service (NTIS). At NTIS, it will be available to the
general public, including foreign nations.

APPROVAL STATEMENT

report h u been reviewed and approved.

PAUL LACASSE, Capt, CF

Facility Technology Division
Directorate of Technology
Deputy for Operations

Approved for publication:

FOR THE COMMANDER

KEITH L. KUSHMAN
Technical Director
Directorate of Technology
Deputy for Operations
 REPORT DOCUMENTATION PAGE Form Approved
!
OMB No. 0704-0188
Pubh¢ reporting burden/or this collection of Information Isestlmlm.,d to average ! hour per response, Indndlng the time for reviewing instructions, mrchlng existln(j data sources,
gilt hermg and maintaining the data needed, and completing and mv,ewlng t he c011e~lonof Information Send comments rega~llng this burden estimate or any othor aspect of this
collection of Information. mdudln9 suggestions for rnducln9 this burden, to Weshlngt(m Headquarters Sorvlces,Directorate for InfOrm,tim1 Operationsand Reports, 121SJefferson
Davis Hlqhwayr Sune 1204,Adlrviton. VA ;t2;10;R.4302,and to the Office of Management and Budget. Papemror, Reduction Project ~0704-01U).Washlnqton. DC 20503
I AGENCYUSEONLY(Leave blank) I 2. REPORTDATE [ 3 REPORTTYPEAND DATESCOVERED
March 1991 Final, June 1987- March 1989
4. TITLEAND SUBTITLE S. FUNDINGNUMBERS
Liquid FilmCooling in Rocket Engines
F49620-85-C-0013
6. AUTHOR(S)
Grissom, William M., Morehouse College

7 PERFORMING
ORGANIZATIONNAME(S)ANDADDRESS(ES) 8. PERFORMINGORGANIZATION
REPORTNUMBER
William M. Grissom
Physics Department AEDC-TR-gl-1
Morehouse College
Atlanta, GA 30314
9. SPONSORING/MONITORINGAGENCYNAMES(S)AND ADDRESS(ES) 10. SPONSORING/MONITORING
AGENCYREPORTNUMBER
Arnold Engineering Development Center/DO
Air Force Systems Command
Arnold AFB, TN 37389-5000

11 SUPPLE~IENTARY
NOTES
Available in Defense Technical Information Center (DTIC).

12a DISTRIBUTION/AVAILABILITYSTATEMENT 12b. DISTRI6UTIONCODE

Approved for public release; distribution is unlimited.

13. ABSTRACT(Maximum 200 words)

A one-dimensional analytical model of I.iquid film cooling in rocket engine, combustion

chambers is developed and compared with existing data. Tile vapor generated at the li.quid
interface greatly decreases the convective heat flux and is treatea as a "transp.iration m
process. The radiant he.at i.s absorbed at the walls and transmitted to the liqui~l film by
ooiling, which can.mad to burnout of the film. Downstream of the liquid film, the vapor
provioes continued thermal protection to the wall and is treated as a gaseous film coofi.ng
process. A standard correration is expressed in differential form to allow for t.he
nonuniform free-stream flow in the nozzle. Liquid film lengths are well predicted by t.he
model. Downstream of the dry-out I)oint, wall temperatures are well predicted up to .t.he
start of convergence. Downstream of the converging turn, a correction term, correlated by
a "centrifugal paramgte.r," is required to accoun~ for increased mixing at the converging
turn: Downstream of the throat, wall temReratures decrease very quickly because o7
acceleration of the boundary.layer gases, an effect not included in the model.

14. SUBJECTTERMS 15. NUMBEROF PAGES

liquid-propellant rocket motors transpiration 74
liquid film cooling heat bansfer 16 PRICECODE
gaseous film tooting
17 SECURITYCLASSIFICATION J 18 SECURITYCLASSIFICATION J 19 SECURITYCLASSIFICATION 20 LIMITATIONOF ABSTRACT
OF REPORT I OF THISPAGE I OF ABSTRACT SAME AS REPORT
UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED
COMPUTERGE~,ERATED Standard Form 7.98(Rev, 2-89)
Prescrl~-sdby ANSI S~l Z)9.18
296-102
 AEDC-TR-91-1

PREFACE

The research reported herein was sponsored by the Air Force Office of Scientific
Research/AFSC under Contract F49620-85-C-0013/SB5851-0361 as a follow-on to work
done in the summer of 1987 on the Universal Energy Systems Summer Faculty Research
Program at Arnold Engineering Development Center, Arnold AFB, Tennessee. The author
returned to AEDC during the summer of 1988 to continue work on the project. Apprecia-
tion is extended to C. C. Limbaugh and W. K. McGregor for suggesting the topic, to M.
A. Simmons for running the NASA equilibrium code, and to Mike Powell of the Air Force
Astronautics Laboratory for providing numerous references to rocket test reports. M. K.
Kingery supervised the Summer Faculty program at AEDC.

The reproducibles used in this report were supplied by the author.

 AEDC-TR-91 - 1

CONTENTS

Pag_.__~e

1.0 INTRODUCTION .................................................. 5

2.0 LIQUID FILM EVAPORATION ...................................... 6
2.1 Convective Heat Transfer ....................................... 6
2.2 Radiative Heat Transfer ........................................ 11
2.3 Comparison With Other Analyses ................................ 16
2.4 Summary of Liquid Film Evaporation Analysis . . . . . . . . . . . . . . . . . . . . . 18
3.0 GASEOUS FILM COOLING ......................................... 21
3.1 Free Stream Turbulence ........................................ 23
3.2 Foreign Gas Injection .......................................... 24
3.3 Thermal Radiation ............................................. 25
3.4 Non-uniform Free Stream Flow .................................. 25
3.5 Turning Effects ................................................ 28
3.6 Chamber Contour ............................................. 29
3.7 Summary of Gaseous Film Cooling Analysis . . . . . . . . . . . . . . . . . . . . . . . . 30
4.0 COMPARISON WITH EXPERIMENTS ................................. 32
4.1 Liquid Film Evaporation ........................................ 32
4.2 Gaseous Film Cooling .......................................... 43
5.0 CONCLUSIONS ................................................... 49
REFERENCES .................................................... 50

ILLUSTRATIONS

Figure Pag___ee

1. Mechanics of Liquid Film Cooling ................................... 53

2. Transition to Fully-Developed Flow .................................. 54
3. Comparison of Transpiration Relations ............................... 55
4. Total Emittances of H20 and CO2 at I atm . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5. Gaseous Film Cooling Analysis ...................................... 57
6. Chamber Contour Parameters ...................................... 58
7. Comparison with Data of G. R. Kinney, et al . . . . . . . . . . . . . . . . . . . . . . . . . . 59
8. Comparison with Data of G. Morrell ................................. 60
9. Comparison with Data of R. C. Kesselring, et al . . . . . . . . . . . . . . . . . . . . . . . 61
I0. Comparison with Data of Ewen and Rousar . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
II. Correlation of Turn Turbulence with Centrifugal Parameter . . . . . . . . . . . . . 68
12. Comparison with Kinney's Figure 4 .................................. 69

NOMENCLATURE ................................................ 70
 AEDC-TR-91-1

1.0 INTRODUCTION

Many liquid rocket engines employ a film of liquid fuel as thermal protection
for the combustion chamber walls. This process was experimentally studied in the
1950's and 1960's, however no general analysis was developed. Since then a
n u m b e r of fundamental heat transfer studies have been performed which allow the
liquid film cooling problem to be analyzed as a collection of several fundamental
processes.

The mechanisms involved in liquid film cooling are depicted in Figure 1.

Heat is transferred from the hot free stream gas to the liquid film by both radiation
and convection. This energy is absorbed in vaporizing the liquid in the protective
film on the wall. The vapor generated flows outward from the liquid film,
decreasing the normally expected convective heat flux by the well-known
transpiration cooling process. Downstream of the liquid film the vapor mixes with
the free stream gas entrained in the boundary layer, lowering the wall temperature
through the well-known "gaseous film cooling" process. This provides thermal
protection downstream of the dryout point.

The existing heat transfer correlations are for the flow of a low turbulence
gas at constant velocity over a fiat plate. These were extended to the case of
developing, accelerating, and turbulent free stream flows. Since the goal was to
develop a simple model, suitable for design and test analysis, only a simple
one-dimensional model was considered.
AEDC-TR-91-1

2.0 LIQUID FILM EVAPORATION

2.1 CONVECTIVE HEAT TRANSFER

The flow over the combustion c h a m b e r walls is almost like that over a flat
plate in a uniform free stream flow, the difference being that it arises from an
effective stagnation point rather than an a b r u p t leading edge. However, for a
turbulent b o u n d a r y layer the exact starting conditions quickly become unimportant.

In the absence of the liquid film, the n o r m a l convective heat flux can be
calculated using o n e o f the many flat-plate data correlations (Reference 1). Most
agree within 5%. One simple correlation with an analytical basis, t e r m e d "Colburn's
Equation", is based u p o n a 1/7th p o w e r law velocity profile (Reference 2).
Expressed in terms of the friction factor, Cfi and the Reynold's n u m b e r based u p o n
the distance x from the leading edge:

Cfo = 0.0592 Rex'° .2 (Equation 2.1)

Using Reynold's m o m e n t u m - h e a t analogy for t u r b u l e n t convection, a

non-dimensional heat transfer coefficient, the Stanton number, is expressed:

Sto = Z Cfo Pr "° .s (Equation 2.2)

Colburn's Equation is applicable for Rex< 1.10 ~ . In most rocket engines this
condition is satisfied.

In a rocket engine there is a large t e m p e r a t u r e difference between the

c o m b u s t i o n gases a n d the liquid film, so that the t e m p e r a t u r e at w h i c h the
properties are evaluated is important. It is c o m m o n practice in b o u n d a r y layer
correlations to evaluate properties at the free stream gas temperature. A correction
factor for the free stream to wall temperature ratio is then multiplied, the exponent
being 0.26, 0.4, o r 0.5 d e p e n d i n g u p o n the a u t h o r (References 3, 4, or 5,
respectively). However, evaluating all of the properties at the mean temperature of
the free stream and wall eliminates such correction factors, and is assumed here.

As the b o u n d a r y layer grows, the cylindrical geometry becomes important.

The most extensive measurements in fully-developed pipe flow at high temperatures
were by H u m b l e (Reference 6). He obtained a correlation which can be written
(using Nu = StRePr):
"Sto = 0.023 ReD"° .2 pr-0 .s (Equation 2.3)

6
 AEDC-TR-91-1

where the Reynold's n u m b e r is based on the pipe diameter. Humble found that if
the property values were evaluated at the mean temperature, then no temperature
correction factor was necessary.

Bartz (Reference 7) found that Equation 2.3 predicted the heat flux at the
throat of an RFNA/N2 F4 rocket if the constant was 0.026. In Bartz' tests the flow
was n o t completely developed, which explains the slightly larger constant. This
so-called "Bartz' Equation" has been c o m m o n l y used to calculate convective heat
fluxes in rocket engines. Not surprisingly, it usually underpredicts the convective
heat flux upstream of the throat where the assumption of fully-developed flow is n o t
justified.

For x>3.53 D the flat plate correlation (Equation 2.2) predicts a lower heat
flux than that for fully-developed flow (Equation 2.3). Certainly the heat flux cannot
fall below that for fully-developed flow. To patch these two limiting cases together,
Equation 2.2 can be used, but with x replaced by an effective x e (Reference 8):

Xe = 3"53 D Ii + 13.53x ]-m ]-i/m

The exponent m = 1.£ gives the best comparison with the data of Barbin and Jones
(Reference 9), as shown in Figure 2. Their data was taken at Re D = 388,000 which
is in the range of the data analyzed in Section 4.1. The o p t i m u m e x p o n e n t is
probably a function of the Re]mold's number.

The convective heat flux for the flow of free stream gas over the dry walls of
the combustion chamber is then calculated as:

Qconv = ho ~T

where h o = GmeanCpgSto . Gmean is the gas mass flow per area, evaluated at the
m e a n temperature between the gas and liquid film. It is scaled from the chamber
value (Gcl 0 as:
Gmean = Gch(Tg/Tmean)[0Jg-Ul)/Ug ]

The second factor accounts for the velocity of the free stream relative to the liquid
surface. The liquid surface velocity (UI) is given in Section 2.2.2.

7
AEDC-TR-91 -1

2.1.1 Free-stream Turbulence

The convective heat flux increases by a factor Kt = 1 + 4.0 e t in the presence

of an rms turbulence fraction, e t , in the free-stream flow (Reference I0).

Two studies measured the turbulence intensities in rocket engines. Hersch

(Reference 11) f o u n d e t = 0 . I 0 to 0.05 at distances o f 2 to 8 inches from the
injector in a LOx/GH2 engine, while Talmor (Reference 12) found e t = 0.20 to 0.15
at distances of 6 to 23 inches from the injector in an N204/AzS0 engine.

2.1.2 Transpiration Corrections

The total heat flux d u e to both convection and radiation, Qtot, is absorbed in
the liquid film, causing an initial temperature rise: d T l i q / d x = Q t o t / ( U C p ) , where I"
is the local liquid mass flow rate per circumference. After the liquid film reaches
the saturation temperature, Tv, the evaporation rate per area is:

;v : Qtot/A : (Qconv+Qrad)/A (Equation 2.4)

This vapor flows away from the liquid film, similar to gaseous transpiration
through a p o r o u s wall. Transpiration decreases the wall shear stress and convective
heat flux by a factor d e p e n d a n t u p o n the '~lowing parameter': P , ~ v / G .

The simplest transpiration analysis assumes that the transpiring vapor does
not change the thickness of the viscous sublayer in the b o u n d a r y layer. Normally
the velocity profile in the viscous sublayer is linear. With b l o w i n ~ it b e c o m e s an
exponential function. The ratio of the wall shear stress in the blown (Tw) to
unblown case (~wo) is:

Lw - z where: z = 2F
~wo ez - 1 Cfo

This analysis is t e r m e d the "Couette flow model" (References 5 and 13) or

alternatively "film theory" (Reference 14). Since StfCf/2, the convective heat flux is
decreased by the same factor. It is convenient to rearrange the expression to a
logarithmic form, where the argument contains the actual Stanton n u m b e r (St),
rather than the unblown value (Sto). Substituting Qconv = St G Cpg AT, gives a
simple result:

h__ = I n C l + H ) where: H = F = C p g r ~ T + Orad]

ho H St A L h J
 AEOC-TR-91 1
-

Without radiation, the convective heat flux decrease depends only upon the
non-dimensional factor Cpg~T/~, termed the "Spalding transfer number" in droplet
evaporation research. With radiation, h must be determined implicitly.

Results of more sophisticated transpiration analyses are compared with the

Couette flow model in Figure 3. Since the existing experimental data shows
considerable scatter, there is no advantage in using the more complicated analyses.

To account for an injectant different than the free stream gas, the parameter
H must be multiplied by a correction factor, KM, being either the vapor to gas
specific heat ratio to the power 0.6 (References 5 and 15) or the gas to vapor
molecular weight ratio to the same power (Reference 16), when Mc<Mg. These
two forms are identical for ideal gases with the same 7. For M c >Mg, Rubesin
(Reference 16) gives the exponent as 0.35. The data of References 5 and 15 was for
Mc < Mg only and thus does not contradict this second exponent. These corrections
are confirmed by Landis' turbulent boundary layer model (Reference 3).

Since the specific heats were considered in the transpiration model, they
should not require a separate accounting. However, a molecular weight difference
would ~equire a correction term, because a lower density injectant would displace a
greater volume in the boundary layer, decreasing the heat flux. The effect should
be less p r o n o u n c e d when the injectant has a greater density than the free stream
gas, as given by Rubesin's second exponent. For this reason the correction term
based upon the molecular weight ratio, as suggested by Rubesin, is used in the
model as:

= IM_S a where: a= 0.60 for M c <M g

KM Mc 0.35 for Mc>Mg

2.1.3 Liquid Film Disturbances

Two types of disturbance to the surface of the liquid film have been
o b s e r v e d . T h e s e have b e e n studied by Kinney (Reference 17), Knuth
(Reference 18), and Gater (Reference 19).

The first type of disturbance is the appearance of random, small-scale

structures, with a pebbled appearance. These disturbances are always present and
cause an effective surface r o u g h n e s s which may increase the heat flux.
Photographic studies show that the length scales of these disturbances decrease at
AEDC-TR-91-1

h i g h e r velocities, although no general correlations were determined

(References 17, 18).

The other type of disturbance, realized at higher coolant flow rates, is the
appearance of large waves traveling in the flow direction. These waves exist only at
the upstream positions where the liquid film is thick e n o u g h to sustain them. The
mass loss rate in the regions where these waves do exist is 2 to 4 times the normal
evaporation rate, independent of the film thickness. Shearing of droplets from the
crests of the waves, without evaporation, is considered the primary mechanism of
increased mass loss.

The point of onset of the large waves is well defined. Knuth was able to
correlate both his data and that of Kinney for the transition point. Assuming that the
mixture above the liquid film is mostly vapor and that the vapor to liquid viscosity
ratio is greater than 0.05, Knuth's correlation for the critical liquid mass flow p e r
circumference is:
[~cr = 1.01 x 10 s ~v2//J£

Gater found quite different results. In photographic studies he noticed large

waves only for water coolant. Mass was lost without evaporation for all of his test
conditions. Furthermore, the mass loss rate was d e p e n d a n t u p o n the local film
thickness, proportionate to the local flow rate.

T h e s e differences might be e x p l a i n e d by the different e x p e r i m e n t a l

arrangements. Kinney and Knuth both m e a s u r e d the length of the liquid film for
different liquid flow rates by determining the dryout point, using axially spaced
thermocouples. Gater used a fixed film length and captured the excess liquid in a
downstream slot. Obviously the potential existed for some of the liquid to flow past
the capture slot. At high injection rates it is also difficult to insure that all of the
liquid is placed on the wall. Knuth considered this p r o b l e m in detail. Either of
these problems might explain the different results of Oater. In favorably comparing
his p r o p o s e d correlation with Kinney's data, Gater employed an adjustable constant
and failed to properly account for quantities on a unit area basis.

Liquid film waves have also been studied with application to chemical
processes. Hanrany and Hersham (Reference 20) give a correlation for the free
stream velocity at the transition to large waves in terms of the gravitational constant.
However, they were unable to explain the results of Kinney a n d Knuth in a
horizontal tube. Apparently their correlation is useful only in vertical tubes at low
gas velocities. Woodmansee and Hanratty (Reference 21) measured:

10
, AEDC-TR-91 - I

l'cr = 0.0641 kg/m-s for water, which is about 3 times lower than predicted by
Knuth's correlation. Of related interest, Tatterson and Dallman (Reference 22) give
a correlation for the m e a n diameter of the droplets sheared off of the liquid film.

The present model is valid only w h e n the liquid flow rate is below the critical
value given by Knuth's correlation. The small-scale disturbances are not assumed to
increase the convective heat flux.

2.2 RADIATIVE HEAT TRANSFER

In most rocket combustion chambers the radiant heat flux is negligible in

comparison with the convective heat flux (References 23 and 24). However, the
transpiration of vapor from the liquid film decreases the n o r m a l l y expected
convective flux to the extent that the radiant flux may b e c o m e d o m i n a n t in
determining the liquid evaporation rate.

The radiation from combustion products is difficult to calculate from first

principles d u e to the complicated molecular spectra of the many species present.
Molecular band models attempt to replace the actual spectral lines present with a
statistical distribution of lines, having the same gross properties. Such modeling is a
major research effort and was not warranted in the present study. However, with
access to spectral codes, it is preferred to the simple approach described here.

Fortunately, in a high temperature, high pressure environment many of the

gas spectral properties can be m o r e simply treated. High temperature causes each
spectral line to be "doppler b r o a d e n e d " d u e to kinetic motion. Even m o r e
significant is "collision broadening", due to high pressure. This broadening tends to
smear the spectrum lines together into continuous bands. With a large n u m b e r of
different species, there would be few gaps in the spectrum. Soot radiation aids in
filling the gaps. In the limit, a very d e n s e gas mixture in t h e r m o d y n a m i c
equilibrium radiates as a perfect blackbody.

2.2.1 Total Emittance of Gas

The total emittance of a gas is the ratio of the radiant intensity of the gas to
that of a blackbody at the same temperature, averaged over the entire spectrum.
This quantity is m u c h easier to measure than the detailed spectral data and has been
available for many years. Unfortunately data is available only for H 2 0 and CO2,
however, this covers many cases of combustion interest since symmetric diatomic

11
AEDC-TR-91 -1

molecules, such as Ns, do not radiate significandy.

The m o s t recent emittance data at I a~rn pressure is compiled by Siegel and

Howell (Reference 25). This data is replotted in Figure 4 as a function of the optical
density, Pop~ which is the gas partial pressure times the path length t h r o u g h the
gas. Note that at very high optical densities the emittances reach limiting values, ~ i~
of 0.825 for H s O and 0.251 for COs. These curves may be analytically fit to a
function:
, = , [ ~[1 f+ optl-n ] -1/n

with coefficients:

H20 COs
T('K~ ctatm, m~ n T ~°K) c(atm,m) n
i000 0.165 0.45 i000 0.05 0.6
2000 0.90 0.65 1500 0.075 0.6
3000 2.05 0.61 2000 0.15 0.6

A three point interpolation is used for temperatures between these.

A correction is required for pressures other than 1 arm by multiplying a

correction factor, Kp, to each emittance. The curves for these pressure corrections
are fit by the functions:

H20: Kp = 1 + C l { 1 - e x p [ ( 1 - P [ l + N w ] ) / C 2 ] )

where: CI = 0.26 + 0.74 exp(-2.5 PH20)

Cs = 0.75 + 0.31 exp(-10 PH20) atm

COs: log10 Kp = 0.036 Pco2-°-4SS[l+(2 lOgl0P)-m]-l/m

where: m = 100 PCOs

An additional correction is necessary to account for overlaps in the two

spectra. This correction (for T> 1200K) is fit by:

AE = 0 . 0 5 5 1 Kx [ 1 - e x p ( - 4 P o p t ) ] [1-exp(-12.5 Popt)]

where: Kx = 1 - I N----X-w
Nw+Nc2 - 1 In

n = 5.5 [1+(1.09 Popt)-a'as]-I/8"a8

12
 AEDC-TR-91-1

In the equations above:

PH20 , PCO= ' Popt <> atm•m

P <> atm
Nw = mole fraction o f w a t e r in m i x t u r e
N c = mole fraction o f CO2 in m i x t u r e

The total gas emittance is then:

~ g = e H 2 o + ~CO2"Ae

To evaluate these terms it is necessary to input the optical path length

through the gas. Since this generally varies over the sight angle, the most direct
method would be to calculate the emittance at a n u m b e r of sight angles and to
average them, weighted by the projected area at each angle. However, a simpler
approach is to use an overall effective length. A very simple expression is quite
accurate in determining the effective length (Reference 25): Leff=0.95(4V/A) ,
where V is the chamber volume and A is the surrounding surface area.

In applying this formula, the downstream section o f the chamber can be

assumed to be an infinite cylinder, for which L etf = 0.95 D. If the cloud of droplets
coming out of the injector is assumed to be perfectly reflective, then the upstream
direction can also be considered an infinite cylinder. Since the droplets have
n u m e r o u s partially reflecting surfaces this is probably the best assumption.

Alternatively, if the droplets are assumed to be perfectly absorbing then the

effective upstream length is: L etf = 0.95D[4x/(D+4x)]. The upstream and
downstream emittances are then averaged to obtain the total emittance. With this
assumption the radiant heat flux must be calculated at every axial position.

It is also necessary to consider the possibility of reflective walls. Reflective

walls increase the effective sight paths. The simplest correction, d u e to Egleti
(Reference 26), is to multiply the effective length above by a factor Aw to the power
-0.85 ; where Aw is the wall absorptivity.

With the gas emittance determined, the radiant heat flux is calculated as:

Qrad = o Aw ~ g [Tg' - Tv 4 ]

where: o = 5.67 10"s W/m 2 • • K'

13
AEDC-TR-91-1

A final concern is that the above relations are for the radiation from hot gases
in chemical equilibrium. However, there is evidence that the radiation from
transient species can he important. Ziebland (Reference 26) m e a s u r e d an emittance
of 0.22 in a small O2/H2 rocket operating at 10 arm. The radiation peak, which
occurred at a position 8 cm from the injector face, was almost three times the
expected value. Some 20 cm downstream, the emittance settled to the expected
value. This initially large radiation was attributed to radiation by transient OH.
While the radiation from OH in aunospheric flames contributes only a faint blue
glow o f n o significant intensity, at high p r e s s u r e s it b e c o m e s significant
(Reference 27). Since OH concentrations are difficult to calculate, radiation by OH
was not considered in the present model.

2.2.2 Liquid Film Burnout

The radiant heat can penetrate the liquid film and be absorbed directly at the
combustion chamber walls. It is then conducted into the liquid film by boiling heat
transfer. With a high e n o u g h heat flux the liquid film can "burnout", as in normal
pool boiling.

The same geometry of a thin liquid film flowing across a heated surface has
been studied by Monde and Katto (Reference 28), Katto and Ishii (Reference 29),
and Mudawwar, et. al. (Reference 30). They all correlated the b u r n o u t heat flux as:

n2

Pv AU Pv 2

where Pv and p~ are the densities of the vapor and liquid, o is the surface tension,
is the latent heat of vaporization, L is the length of the heated surface, and U is the
average velocity of the liquid film.

The constants fit to the data by each g r o u p and the parameter ranges are
given below:

Group C nl n2 U(m/s~ L(cm~

Monde and Katto 0.0591 0.725 0.333 3 t o 26 0.56, 1
Katto and Ishii 0.0164 0.867 0.333 1.5 t o 15 1 - 2
Mudawwar, et.al. 0.0881 0.867 0.432 0.4 t o 2 6.4, 1 2 . 7

Monde and Katto studied a circular jet of water and Freon®113 impinging
d o w n w a r d o n a heated plate. Katto and Ishii s t u d i e d a plane jet of water,

14
 AEDC-TR-91-1

Freon®ll3, and l,l,l-trichloroethane flowing across a heated plate. Mudawwar, et.

al. studied a film o f F l u o r i n e r # liquid (FC-72) flowing d o w n w a r d along a heated
surface. Since they used a single liquid, they a s s u m e d the same density ratio
exponent, n t , as Katto and Ishii. Since Katto and Ishii's test conditions appear to be
closer to those for liquid film cooling in rocket engines, their correlation is used.

All of the researchers observed that w h e n the critical heat flux was exceeded,
the liquid film separated from the heated surface. Katto and Ishii observed that with
water, the separation occurred u p o n first contact with the heated surface, whereas
the organics m a i n t a i n e d contact with the surface for s o m e distance before
separating. Mudawwar, et. al. observed separation u p o n first contact with the
surface for their Fluorinerto tests. They also observed that the film reattached to
the surface w h e n the surface temperature was decreased.

This total separation of the film from the surface in the experiments may
have been d u e to the constant heat flux condition imposed. In a rocket engine the
liquid film w o u l d be less susceptible to b u r n o u t than in these heat transfer
experiments b e m u s e separation of the liquid film from the surface would generate
a region of droplets a n d bubbles which would cause scattering, decreasing the
radiative transmission through the film.

Use of these b u r n o u t correlations is questionable since they are expressed in

terms of the overall heated length. Ideally, the b u r n o u t point should be expressed
in terms of local conditions, such as the local film thickness. They are mainly useful
as an o r d e r of magnitude calculation. With this in mind, the b u r n o u t heat flux for
several rocket fuels is calculated below for the case: L = 5 cm and U = 1.5 m/s, at a
pressure of 100 psia (data Reference 31):

Fuel Tsat A P~ Pv o~ Qbo

(°K) (10SJ/kg) (kg/m s) (kg/m s) (10-SN/m) (kW/m =)
MMH 433 663 720 9.79 20.2 419
AZ50 413 870 778 9.50 18. 548
H2 29 335 70 5.75 0.6 17.6

In m o s t rocket engines the radiant flux slightly exceeds these critical values.
However, some o f the radiant heat is absorbed by the liquid film. Since hydrogen
has such a low predicted b u r n o u t heat flux and does not absorb infrared radiation,
it appears unsuitable for liquid film cooling.

Katto et. al. suggest a correction factor w h e n the liquid is injected at a

temperature below the saturation temperature, however it was not correlated over a

15
AEDC-TR-91-1

large range and is significandy larger than such correction factors for pool boiling.
If the b u r n o u t point is controlled only by local conditions, and the liquid is heated
to the saturation temperature at the b u r n o u t point, then no sub-cooling correction
should be necessary, therefore n o n e is used in the present model.

The fraction of radiation transmitted t h r o u g h the liquid film is: exp(-~t) ;

w h e r e a is the absorptivity of the liquid, averaged over the s p e c t r u m of the
radiation. The liquid film thickness, t, is calculated from the wall shear stress.
Assuming laminar flow:

where I" is the local coolant mass flow per circumference. The surface shear stress,
~vo is calculated by Equation 2.1, with the transpiration correction of Section 2.1.2.
For simplicity h / h o is used in place of ~w/~-wo.

The average liquid film velocity, needed for the correlation, is: U = I'/(pt).
The surface velocity of the liquid, n e e d e d in Section 2.1, is twice this average film
velocity.

Using these values, the b u r n o u t heat flux is calculated with Katto and Ishii's
correlation at each axial position in the rocket chamber a n d c o m p a r e d with the
radiant heat flux transmitted t h r o u g h the liquid film. A warning results w h e n the
radiant flux exceeds the b u r n o u t heat flux. However, as discussed in Section 4.0,
this b u r n o u t warning was n o t found to give a reasonable c o m p a r i s o n with any of
the data analyzed.

2.3 COMPARISON WITH OTHER ANALYSES

2.3.1 Purdue University Jet Propulsion Center (JPC):

A n u m b e r of liquid film cooling analyses have been presented by researchers

from the Jet Propulsion Center at Purdue. In early studies Zucrow a n d Graham
(Reference 32) attributed the increased mass loss rate after transition not to iargc
surface waves, but to variations in the termination point of the liquid film a r o u n d
the circumference. Zucrow and Sellers (Reference 33) attempted to model the mass
transfer with no real success.

16
 AEDC-TR-91-1

In a later analysis Warner and E m m o n s (Reference 34) gave a result which

was restated by Ziebland (Reference 26). For the no blowing case it reduccs to an
unusual result:
st = (PrUg)'

It was not possible to r e p r o d u c e their calculations with the suggested constants,

even after correcting for an apparent sign error in their Equation 5, so their analysis
was not considered further.

Gater, et. al. (Reference 35) a t t e m p t e d to relate all o f the p r e v i o u s

expressions from the .[PC. In a later report (Reference 19), Gater mentions the
transpiration analysis of Section 2.1.2, but discounts it due to the anomalies in his
data m e n t i o n e d in Section 2.1.5.

2_~-2 Shembharkar and Pai (Reference36)

These authors use a Prandfl mixing length turbulence model for the gas
convection, c o u p l e d to a viscous flow m o d e l for the liquid film. Landis
(Reference 3) and Economis (Reference 37) reported similar turbulent b o u n d a r y
layer m o d e l s which gave no real i m p r o v e m e n t over the simple Couette flow
transpiration model, so the need for this complexity is questionable.

In their d a t u m case, the starting b o u n d a r y layer thickness is 12.5 ram,

corresponding to a downstream starting position of x= 0.812 m, by the b o u n d a r y
layer growth relations of Section 3 . 0 . Since the boundary layer is well developed
u p o n contacting the relatively short liquid film, the evaporation rate s h o u l d be
a l m o s t constant. Instead, they calculate an initially high evaporation rate of
0.2 kg/s-m 2, d r o p p i n g exponentially to 0.124 kg/s-m 2 , possibly d u e to numerical
problems. The simpler analysis of the present r e p o r t predicts an almost constant
evaporation rate o f 0.106 kg/s-m 2 . The difference between this and their final
evaporation rate is exactly accounted for by the blowing correction factor (h/ho). It
is not apparent that the m o m e n t u m of the transpiring vapor was properly included
as a b o u n d a r y condition in their model. This vapor m o m e n t u m is what causes the
convective heat flux decrease in transpiration cooling.

17
AEDC TR 91 1

2.4 SUMMARY OF LIQUID FILM EVAPORATION ANALYSIS

This section summarizes the equations and calculation order used in the
liquid film evaporation model. A simple algebraic formulation is given first, suitable
when radiation is negligible and the flow is either fully-developed or purely
boundary layer.

2.4.1 Simple Expression

Without radiation, the liquid film length, Lc , is determined from the coolant
flow (per circumference), F, and the convective evaporation rate, mconw as:

L c = F/mconv

Substituting Equation 2.4, with the transpiration correction,

LC ffi g
GCpgdT St o (h/ho)

For fully developed flow (Lc> 5D), Sto is constant and is calculated with
Equation 2.3, allowing Lc to be calculated above.

For pure boundary layer flow (Lc<2D), with liquid injection at x=O, the
average Stanton number is 1.25 times higher than that at x f L c (found by integrating
Equation 2.2 from xffiO to x=Lc). Thus:

Stolavg : 1 . 2 5 [ 0 . 0 2 9 6 ( G L c / ~ g ) - O - 2 P r -°-6]

Substituting above:

LC = 61.62 u [ ), F ]1.28 pr O . 7 8
G Cpg AT ~g(h/ho)

Without radiation, the transpiration correction of Section 2.1.2 reduces to a

simple form:

h/ho ffi ln(l+HJ where: H = Cpa

~. AT K M
H

18
 AEDC-TR-91-1

2A.2 Complete Fomulation

The e q u a t i o n s u s e d in t h e liquid film e v a p o r a t i o n FORTRAN p r o g r a m

"liquid.for" are s u m m a r i z e d below.

The r a d i a n t emittance o f the hot c o m b u s t i o n p r o d u c t s is calculated first in

subroutine "Emittance" as:

Eg = Emittance(T,Lefl~Tg,Nw,Nc)

where: L etf = 0.95 D Aw'° .a 8

giving: Qrad ffi o Aw ~ g [Tg' - Tv 4 ]

Subroutine "Emittance" is a direct coding o f the curve fits o f Section 2.2.1.

Parabolic interpolation is u s e d between the three t e m p e r a t u r e curves for CO 2
and H 2 0 in function "Fit". Since a constant L etf is a s s u m e d (i.e. reflective
droplets), the radiative heat flux is constant, i n d e p e n d e n t o f axial position.

Axial steps:
Beginning at the liquid film injection point, the local convective coefficient
without transpiration is calculated as:

h o = KtG C p g S t o

where: K t = l+4e t "turbulence correction factor"

St ° = %CfPr-O.,
Cf = 0. 0 5 9 2 R e x -° • 2

Re x = G X e / # g
G = Gch(Tg/Tm) [ (Ug-UI)/Ug ]
xe = 3.53D(1 + [x/(3.53D)]-1"2} °'a88s

Tm = ~(Tg+Tv)

The convective heat flux with transpiration is calculated implicitly as:

h ffi h o l n ( l + H ) / H

I
where: H = Cpg KM mva p /h

= a ; if <

a=0.35 if M v > M g

19
AEDC-TR-91-1

mva p is the liquid evaporation rate per surface area, calculated below.
The wall shear stress is decreased by this same factor, so that:

~w ffi ~wo(h/ho)

where: ~w ffi %CfG(Ug-U1)

The convec~ve heat flux is then: Qconv = h AT.

The initial temperature rise of the liquid film is:

dTliq/dx = Qtot/(FCp) for Tliq<Tv

where: Qtot=Qconv+Qrad

Once Tliq=Tv, the liquid evaporates at a rate (per area):

mva p = Qtot/~

which decreases the liquid flow p e r circumference at a rate dI"/dx = - mva p.

The position at which F = 0 is the "film-cooled length", Lc.

The liquid film thickness and average velocity are calculated using the simple
laminar "Couer~ flow" result:

t = sqrt[ r/(pTw)]
Uli q ffi I ' / ( p 0

The surface velocity of the liquid is twice this average velocity: UI = 2Uli q

The possibility of liquid film b u r n o u t is tested. The b u r n o u t heat flux limit is

calculated as:

Qbo = 0 - 0 1 6 4 ~ p ~ ° * S s 4 p v ° ' I s s ( U l i q ° / x l i q ) ° ' s s s

This is c o m p a r e d with the radiant heat transmitted to the wall:

Qw = Qrad exp(-~t)

When Qw > Q b o , a warning is printed in the output file. However, this warning
has been found unreliable.

20
 AEDC-TR-91-1

3.0 GASEOUS FILM COOLING

After the liquid film has vaporized it continues to provide thermal protection
to the wall by calorimetric mixing with the hot free stream gas entrained in the
b o u n d a r y layer. N u m e r o u s correlations have been presented for this process,
termed "gaseous film cooling". They are normally expressed in terms of a "cooling
effectiveness":
m Ta - Taw
Tg - Tc

where Tg is the free stream gas temperature, T c is the initial coolant temperature,
and Taw is the "adiabatic wall" temperature in the boundary layer which results
from calorimetric mixing of the two gas flows. The cooling effectiveness is a
function of a non-dimensional distance CA')downstream of the injection point:

X=Kx

where: K - G pg°'2s Mc-1"2s

Mc = coolant mass flow per circumference

G ~ pqUq = free stream mass flow
x d6w~stream distance

Most authors arrange the constant K in terms of a coolant Reynolds number,

giving the false impression that the injection velocity is important. As long as the
coolant is injected with low relative velocity so that there are no jet effects, the exact
geometry of the injection point is unimportant.

The most successful analysis is that of Kutateladze, et al (Reference 38) and

StoUery, et al (Reference 39, Equation 27) for gaseous film cooling on a flat plate in
turbulent flow, with coolant injection at the leading edge. To specify the conditions
of the boundary layer downstream of the injection point, they conceptually replace
the injected coolant with an equivalent mass of free stream gas. To provide the
proper boundary layer growth rate, they identify an effective leading edge a distance
x o upstream of the true leading edge (at x= 0). Their analysis can be extended to
the case of injection at a point x i downstream of the true leading edge as follows,
with the nomenclature defined in Figure 5.

Prom the solution to the boundary layer integral equations for a 1/7th power
law velocity profile (Reference 2), the boundary layer thickness and mass flow (per
width) a distance x' from the effective leading edge is:

21
AEDC-TR-91-1

6 = 0.371x'Rex '-0-2 where: x'E x+x o

Mbl = 7 / 8 G~ = 0 . 3 2 5 M c [ X + X o ) ° ' ' (Equation 3.1)

The fictitious upstream point Xo is found from a mass balance at the

injection point:

Mbzl ~=~i = Mc + ~up

o.3,s c(xi+Xo)O.. = + o.3,5 xio.. (Equation 3.2)
Thus,
Xo = (3.08+Xi0.8)*.~8- Xi

where: Xi = K xi

The total mass of free stream gas entrained up to position x is:

(Equation 3.3)
From a calorimetric heat balance:

cpc cpCt~c JJ
giving,
=
[ i + Cpc(0.325[X+Xo]
o., - 11
]-, (Equation 3.4)

Other correlations based u p o n similar analyses have been p r o p o s e d

(Reference 39, 40, 41). However, most of these are less careful about accounting
for the effects of the injectant upon the boundary layer growth downstream of the
injection point. At large downstream distances most reduce to the same form. The
correlation above gives excellent c o m p a r i s o n with most experimental data
(Reference 59, 42).

This analysis can be extended to the case of distributed injection, as exists in

the liquid film cooling process, with the vapor injected continuously along the
liquid film length. However the resulting expression is unwieldy. In any event, the
exact injection point is of m i n o r concern. Librizzi and Cresci's analysis
(Reference 41), which gives predictions very close to Equation 3.4, does not even
require the position of coolant injection.

This analysis contains two assumptions which are known to be wrong. These
are that all of the gaseous coolant remains in the boundary layer and that all of the
gases in the b o u n d a r y layer are at the same temperature, Taw. In fact,
measurements of the concentration and temperature profiles show an S-shaped

22
 AEDC-TR-91 -I

profile for both (Reference 43). However, these profiles do maintain a similar
shape, growing away from the wall at the same rate as the b o u n d a r y layer thickness.
This suggests that the above analysis can be reinterpreted.

The S-shaped temperature profile can be conceptually divided into a two

d o m a i n region, an outer layer at the free stream temperature and an inner layer at
Taw. Since this inner layer thickness is a constant fraction of the total b o u n d a r y
layer thickness, the same mass ratio of hot gases will be entrained as in the original
analysis, giving the s a m e cooling effectiveness. However, in this case the
interpretation is that Taw is the temperature o n l y in the region very close to the
wail.

Even with this re-interpretation a difficulty remains, in that the analysis

predicts no free stream gas entrainment in fully-developed flow, which is certainly
unrealistic for t u r b u l e n t flow. Indeed, Equation 3.4 is u s e d in Section 4.2 for
fully-developed flows, with satisfactory results. That the analysis leading to
Equation 3.4 is apparently flawed should not detract from the fortuitous result that
it correlates all existing data well.

A n u m b e r of corrections to Equation 3.4 are necessary to account for effects

not considered in the standard fiat-plate boundary layer analysis. Corrections for
free stream turbulence, foreign gas injection, thermal radiation, and changing free
stream conditions are considered below.

3.1 FREE STREAM TURBULENCE

The flat-plate boundary layer relations used in deriving Equation 3.4 assume
no free stream turbulence. Two experimental studies m e a s u r e d the effects of
t u r b u l e n c e u p o n gaseous film cooling (References 44 and 45). A possible
correction to the above analysis is to multiply the non-dimensional distance X by a
factor Kt . 1 + C t. This allows the increased mixing caused by the free strcam
turbulence to be accounted for as an effectively, larger downstream distance. When
applied to the data of Marek and Tacina (Reference 44), Ct is found to be constant
with X and varies with the free stream turbulence rms fraction, et~ as: Cot = 8.67 e t ,
for e t = 0.07, 0.14, and 0 . 2 3 . For e t = 0 . 3 5 , the data follows this relation for
X< 1.5, thereafter Ct increases as X ° .s

Unlike Marek and Tacina, Carlson and Taimor (Reference 45) did not
directly measure the turbulence level, but inferred it based on screens placed in the
flow. The length scales of their turbulence may also have differed considerably

23
AEDC-TR-91 -1

from Marek and Tacina's. Analyzing their data gives: Cot = 11.7 e t • In their data Ct
varied slightly with X at all turbulence levels. The expression above is an average
over X. Since their greatest turbulence level was 22%, the increased mixing found
by Marek and Tacina at 35% turbulence was not confirmed. For the present model
the two results were averaged, giving a correction term:

Kt ffi 1 + 10.2 e t

3.2 FOREIGN GAS INJECTION

Goldstein, et. al. (Reference 43) found that the effectiveness values were
about 30% higher than expected when helium replaced air as the coolant injected
into a heated airstream. These results were confirmed by Burns and Stollery
(Reference 46). However, Carlson and Talmor (Reference 45) questioned whether
the effect should be attributed to the coolant properties, since the helium injection
velocities w e r e lower than the air injection velocities at the same effectiveness
values.

In the present model the effect is assumed to be real and is attributed to the
molecular weight difference, for the same reasons stated in Section 2.1.2. Goldstein
fit his data with an empirical equation (Reference 43, Equation 12). Apparently he
failed to recognize that this empirical equation could be p u t into the form of
Equation 3.4 by multiplying a correction factor, KM, of 0.76 to the (Cpg/Cpc) ratio.
Assuming a p o w e r function d e p e n d a n c e of this correction factor u p o n the
molecular weight ratio, and generalizing from this specific case of a coolant to gas
molecular weight ratio of 0.138, the correction factor can be written in general as:

x M -- (Mc/Mg)o.x,

With these corrections for turbulence and foreign gas injection, Equation 3.4
becomes:

32 , tx xo,0 ' - i l l -1 ,E uation

24
 AEDC-TR-91 -1

3.3 THERMAL RADIATION

The radiant heat flux is transmitted through the b o u n d a r y layer and

absorbed at the wall. This heat is then conducted back into the boundary layer
gases, r e q u i r i n g a wall temperature, T w, in excess of the b o u n d a r y layer
temperature, Taw:

T w = Taw + Qrad/h

This form is suggested for gaseous film cooling with a non-adiabatic wall
(Reference 47). The heat transfer coefficient is calculated by Equation 2.2,
evaluated at x'= x + x 0 .

The radiant heat also enters into the heat balance of the boundary layer
gases. Assuming that it is spread evenly over the entire boundary layer mass, the
rate of temperature increase with distance is:

/ ~ a w ffi O r a d
~x NblCpg

3.4 NON-UNIFORM FREE STREAM FLOW

Equation 3.5 was derived for a constant free stream gas flow rate and
temperature. If these quantities change as the flow proceeds downstream, then
Equation 3.5 cannot be directly applied. To demonstrate this, consider a sudden
decrease in the flow rate per area, G. Applying Equation 3.5 locally would imply a
sudden decrease in the non-dimensional distance, X, and a step increase in the
effectiveness, ~. This would suggest that the gases spontaneously unmix, violating
the second law of thermodynamics.

Due to the parabolic nature of the b o u n d a r y layer equations, the local

conditions determine only the rate of change, in the downstream direction, of the
boundary layer properties. Equation 3.5 is an integral result for the special case of
constant flow conditions. However, it can easily be rearranged in the form of a
differential equation.

The rate of change of ~ with position x is determined by differentiating

Equation 3.5 • All parameters are held constant in differentiating, since such was
assumed in deriving the integral expression. Expressing the result in terms of the
local value o f , :

25
AEDC-TR-91-1

KMCpg L.

II is found by' numerically integrating this expression. The initial condition

on ,1 at the injection point c o u l d be found from P.quation 5.5. However, this
predicts rl < 1 at the injection point w h e n injection is downstream o f the leading
edge, which is unsupported. A m o r e realistic assumption that ,7 = 1 at the injection
point is used as the initial condition in the present study. The above approach is
t e r m e d the "differential rl" f o r m u l a t i o n to distinguish it f r o m the "integral
correlation" of Bquation 5.5 •

The free stream mass flow rate per area, G, varies inversely with the local
cross-sectional area as:
G = Och(Ach/A)

Increases in G, as the nozzle converges, increases the rate o f free stream gas
entrainment through the coefficient K.

The other effect of free stream acceleration is a d r o p in the static temperature

of the free stream gas. Upon entrainment into the b o u n d a r y layer the original
stagnation temperature is recovered as the kinetic energy is converted back into
thermal energy. However, some of this heat is then lost by conduction back into the
cooler free stream gas. The result is that the free stream gas attains a temperature
in the b o u n d a r y layer lower than the stagnation temperature. This is t e r m e d the
"recovery temperature", T r, and is normally a constant fraction of the difference
between the static and stagnation temperatures:

T r = T o - ( 1 - r ) (To-T s)

where T O and T s are the static and stagnation temperatures of free stream gas and r
is the "recovery factor".

This recovery temperature should replace Tg in the equations above. For a

fiat-plate b o u n d a r y layer without gaseous film cooling, r = Pr 1/a (Reference 48). It
is assumed that this relation holds for the gaseous film cooling case, although this
assumption is questionable.

From the isentropic relations for compressible flow:

T s = To/[1+

26
 AEDC-TR-91-1

The local Mach number, M, is found implicitly from one of two forms:

for M<I
M = 1At [7-~ + M~r]Tr/2
or
M = [Tr(MAt)2/Tr - ~-~_i]% for M>I

where: A t m (A/Athroat)
7r_=~ +l
7-i

3.4.1 Differential Entrainment Formulation

When the recovery temperature changes, the "differential ~" formulation

becomes invalid, for the subtle reason that the local value of ,1 no longer
determines the local state of the boundary layer. This problem is remedied by
recognizing that the true local state of the boundary layer is determined by the local
mass flow rate in the boundary layer, Mbl. From Equation 3.1, the local
entrainment rate is:

dl~bl]dxe - - 0 . 1 9 6 3 G [~bl]- ° ' ' s (Equation 3.6)

The calorimetric heat balance in differential form is:

dMblCpg (Tr-Taw) = (l~gCpg+l~cCpc) dTaw

The specific heat on the left side should be evaluated at the mean temperature
of T r and Tavo while those on the right should be evaluated at Taw. Substituting
Equation 3.3, and dividing both sides by dx:

d-'
= d~aT
xw d{Ibl
+ ', de (Tr-Taw)
x [~4bl [Cpg[Cp--q _ l] l~Ic] -1 (Equation 3.7)

To apply these relations, Equations 3.6 and 3.7 are integrated to determine
Mbl and Taw at each x position. As stated, Taw is equated to T c (or ,7 = 1) at x = 0.
The initial condition on Mbl is found from Equation 3.2:

Mbl[xi ffi M c [ 1 + ° - n s x i ° "' ]

In applying this to the curved walls of a rocket chamber, x is measured along the
contour of the nozzle.

27
AEDC-TR-91 -1

This "differential entrainment!' formulation is the most general. Although the

derivation is m o r e involved, it requires no m o r e computation than the previous
"differential n" formulation. For the case of constant free stream gas temperature,
w h e r e the local value o f r/ is directly related to Mbi, b o t h f o r m u l a t i o n s are
equivalent. And, of course, for constant free stream conditions, both reduce to the
integral correlation (Equation 3.5).

To include the effects of free s t r e a m turbulence and foreign gas injection, the
coefficient in Equation 3.6 is multiplied by the factor I(t , and the (Cpc/Cpg) ratio in
Equation 3.7 is multiplied by 1/K M. Both corrections are consistent with the
placement of these factors in the differential r/formulation.

3.4.2 Circumferential Change:

In a rocket engine the contraction of the nozzle decreases the circumference,

causing an increase in the boundary layer mass flow per circumference:

Io ;-
Since this increase is not d u e to free stream entrainment, it does not enter
into the energy balance. However, it is important in that it affects the total b o u n d a r y
layer flow rate, w h i c h is f o u n d by a d d i n g the changes d u e to free stream
entrainment and this circumferential change. This total b o u n d a r y layer flow rate
determines the local rate of free stream gas entrainment, as given by Equation 3.6.
In adding this term to the analysis, the coolant flow rate per circumference should
also be calculated locally, by scaling the chamber value.

3.5 TURNING EFFECTS

Carlson and Talmor (Reference 45) measured the effectiveness downstream

of a sharp turn in a rectangular duct. Their results can be fit as:

.=[l+cxn] ''
with coefficients:

turnina anule C n
30" 0.40 2.2
45" 0.75 2.1
60" 1.40 2.0

28
 AEDC-TR-91 - 1

To reduce approximately to Equation 3.4 the exponent, n, should approach

0.8 as the turning angle approaches 0". Since the e x p o n e n t is increasing in the
opposite direction their results are somewhat anomalous. For u n t u r n e d flov~ in a
circular duct their results followed Equation 3.4. Possibly these results were d u e to
disturbance of the boundary layer by the sharp turn in their wind tunnel.

Ewen a n d Rousar (References 49-51) studied g a s e o u s film cooling in

converging nozzles. Their data is analyzed in Section 4.0 and leads to a correlation
for an increased entrainment in terms of a "centrifugal parameter".

3.6CHAMBERCONTOUR

In general the shape of most rocket combustion chambers can be expressed

in terms of a few parameters: 11 , r : , 0 c, D t, r 2 , and 0 d, as shown in Figure 6. Given
these, the limits o f each segment are expressed:

12 = I, + r, sin0 C
D 2 = Dch - 2 r, ( 1 - c O S 0 c )
Ds = Dt + 2 r 2(1-Cos8C)
Is = i~ + (D2-Ds)/(2tanSc)
It = i s + r 2 sin8 c
i s = I t + r 2 sin8 d
Ds = D t + 2 r 2(1-cos0d)

The chamber diameter is then calculated at any axial position, I, by:

For: i, < 1 < 12 D = Dch - 2(r,-sqrt[rl~-(l-l,)2])

12 < 1 < is D = D 2 - 2(l-12)tan0 c
ls < 1 < 15 D -- D t + 2(r2-sqrt[r22-(l-15)2])
16 < 1 D = D6 + 2 (i-is)tan0 d

The downstream position along the wall, x, is related to the axial position, I,
as:

( ~ x ) 2 = C~l) 2 + ( ~ D / 2 ) 2

This expression is integrated as axial steps •I are made, to determine the contour
length.

29
AEDC-TR-91-1

3.7 SUMMARY OF GASEOUS FILM COOLING ANALYSIS

This section summarizes the equations and calculation o r d e r used in the

gaseous film cooling FORTRAN computer program "gas.for". A simple formulation
is given first, suitable when the free-stream mass flow and temperature are constant
and radiation is negligible.

Simple Expression:

With constant free stream conditions, the wall t e m p e r a t u r e is easily

calculated from Equation 3.5:

Taw = T r - . (Tr-T c)

where:
i]]-i
X=Kx
K = G pgo.~s Mc-*'26

Complete Formulation:

Starting at the end of the liquid film, steps dx along the wall contour are made. The
increase in boundary layer flow due to free stream entrainment is:

* = 0.1963KtG(.g/Mbl [ i)-o. a, dx

where the turbulence correction I ~ l + 10.2e t . The new boundary layer flow is
projected as: Mbl[ i+ 1 = l~IblJ i+dMt" Following the 2nd o r d e r Runge-Kutta
method, the change is recalculated (dM2) using l~Ibl[ i + 1 • The average of the two
is used:

The increase in boundary layer flow (per circumference) due to chamber

contraction is:

dMc -- - ~4bl ( l / D ) (dD/dx) dx

where dD/dx is the rate of change of diameter with contour length x, as expressed
in Section 3.6.

30
 AEDC-TR-9.1-1

The boundary layer flow at the next axial step is then:

The increase in b o u n d a r y layer t e m p e r a t u r e d u e to the free stream mass

entrainment is:

The increase in boundary layer temperature due to thermal radiation is:

dTrad = [Qrad/(CpgMbl) ]dx

The boundary layer temperature at the next step is then:

dTaw[ i+l = dTaw[ i + (dTe+dTrad)

The wall temperature is higher than the boundary layer gas as:

T w = Taw+Qrad/h

where h is the convective heat transfer coefficient, as determined in Section 2.1.

The recovery temperature (Tr) is determined by the local area ratio, as

discussed in Section 3.4 • The local area ratio is d e t e r m i n e d by the c o n t o u r
equations of Section 3.6.

31
AEDC-TR-91-1

4.0 COMPARISON WITH EXPERIMENTS

4.1 LIQUID FILM EVAPORATION

4.1.1 G.R. Kinney et al air-water experiment (Reference 17)

These heat transfer measurements were performed at fairly low temperatures

in a tube with almost fully-developed flow. This made radiation negligible and the
evaporation rates almost constant over the liquid film length. For these reasons the
evaporation rate could be calculated by hand using the formulation of Section 2.4.1.
For completeness the computer program was used to include the slight effects of
radiation, interface motion, and boundary layer development.

The static pressure varied from 1.4 to 2.47 arm over the runs. An average
saturation temperature of 240 F is stated by Kinney, corresponding to a saturation
pressure of 1.7 arm. However, a liquid interface temperature of 200 F is given in
Kinney's Figure 4. This difference might be attributed to mass transfer to the
unsaturated gas. A pressure of 1.7 arm and an interface temperature of 200 F was
assumed in all of the calculations, since no itemization was given for each run.

The free stream gas consisted of the products of gasoline/air combustion.

For property values, p u r e air was assumed. For radiation calculations 10% CO 2
and 1096 H 2 0 were assumed for all runs. In any event, radiation was insignificant
at the low t e m p e r a t u r e s of the tests. Since the free stream t u r b u l e n c e was
unknown, a value of 0% was used. The property values c o m m o n to all runs were:

Gas: Coolant:
air water
M= 29. M= 18. p= 3.03 10 - 4 k g / m - s
P= 1.7 a t m T v = 366 K pl = 962 k g / m 3
T o = 300 K p v = 1.01 k g / m 3
Cpl= 4 2 1 0 J / k g - K o = 60.3 1 0 - S N / m
A = 2.27 1 0 e J / k g ~ = 0.

32
 AEDC-TR-91 -1

The property values o f air are listed b e l o w (Reference 54) for each free
stream gas temperature used in Kinney's tests.

Tg (F) 800 900 1200 1400 1600

Tg (K) 700 756 922 1033 1144
pg (kg/m3) 0.857 0.794 0.651 0.582 0.525
pg (10-Skg/N-s) 3.349 3.515 3.976 4.264 4.550

a t T m e a n (K) ffi 533 561 644 700 755

C p g (J/kg-K) 1036 1042 1062 1075 1088
/Ja (10-Skg/m-s) 2.80 2.90 3.17 3.35 3.51
P~ 0.698 0.698 0.701 0.702 0.703

Kinney stated only the free stream Reynold's numbers. Using the viscosity
values above, the gas mass flux (Gel0 for each test was back-calculated from the
stated Re)mold's n u m b e r . The liquid film length (Lc) at an average coolant flow
[" =0.08 k g / s - m was d e t e r m i n e d for each test and the average evaporation rate
calculated as: m v = I'/L c

Similarly, the experimental evaporation rates for both normal evaporation

(~Vexpl) and after transition to the large waves (~Vexp2) were calculated from the
slopes o f the data plots in Kinney's Figure 6, giving:

TEST Re D Gch Lc mvcalc mvexpl mvexp2

(lO 5 ) ( k g / s - m 2) (om) [ kg/s-m 2 ]
2insmooth:
800F 4.4 290.1 86.6 0.0924 0.141 ---
800F 5.6 369.2 71.3 0.112 0.200 ---
1200F 3.5 273.9 50.8 0.157 0.236 0.450
1600F 2.5 223.9 38.5 0.208 --- 0.611

4 in s m o o ~ :
900F 6.0 207.6 106.7 0.0750 0.0794 0.326
900F 8.2 283.7 83.1 0.0963 0.109 0.226
900F 9.9 342.5 71.5 0.112 0.153 0.366
1400F 5.3 222.4 54.7 0.146 --- 0.440
1600F 4.7 210.5 48.0 0.167 --- 0.629

4 in rough:
800F 6.8 224.1 119.9 0.0668 0.131 ---
800F 11.0 362.6 81.5 0.0982 0.226 0.458
1400F 5.4 226.6 53.9 0.149 0.326 0.930
1400F 6.7 281.2 45.4 0.176 --- 0.896

33
AEDC-TR-91 -1

C o m p a r i s o n between the calculated and experimental (mVexpl) values is

given in Figure 7. While the calculations d o not compare favorably with all of the
m e a s u r e m e n t s in an absolute sense, they do correlate the data well. The calculated
evaporation rates are a factor 3596 lower than the data for the t w o "smooth" tubes.
Assuming a free stream t u r b u l e n c e intensity of 8.7596 w o u l d account for this
discrepancy, as discussed in Section 2.1.1.

Several other factors could contribute to the discrepancy. Most significant is

that the actual propellant flow rates were not stated, so that the gas mass flux (Gch)
had to be back-calculated without knowing the value of gas viscosity Kinney used in
calculating the stated Reynold's n u m b e r s . Also, the specific heat of the actual
combustion products is slightly higher than that for the p u r e air assumed. Finally,
the "pebbling" of the liquid surface observed by Kinney would give an effect similar
to pipe roughness, increasing the convective heat flux.

The model is not able to predict the evaporation rates w h e n large waves are
present, given by the mVexp2 values. When the large waves are present, the rates of
mass loss are increased by a factor of 2.56, on average.

34
 AEDC-TR-91 -1

4.1.2 E.L. Knuth air-water tests (Reference 18)

These tests were p e r f o r m e d shortly after Kinney in a 2.9 inch diameter tube
with fully-developed flow. The gas properties are again taken as those of p u r e air.
10% CO2 and 10% H 2 0 in the products is assumed for the radiation calculations.
The liquid interface temperature is taken as 559 K, based u p o n Knuth's Figure I I .
This disagrees with the stated static p r e s s u r e of 1 arm. The discrepancy is
attributed to mass transfer to the unsaturated gas. The c o m m o n properties u s e d
were:

Gas: Coolant:
air water
M= 29. M = 18. g = 4.26 1 0 - 4 k g / m - s
P= 1 a t m T v = 339 K pl = 980 k g / m 3
T c = 300 K pv = 0.1664 kg/m3
C p l = 4188 J / k g - K o = 65.1 1 0 - S N / m
A = 2.34 1 0 S j / k g a = 0.

The gas properties used in each calculation were (Reference 54):

TEST 20-30: 45-54: 89-99:

T g (K) 613 901 1230
p g (kg/m3) 0.576 0.392 0.287

a t T m e a n (K) = 476 620 785

C p g (J/kg-K) 1025 1056 1095
~ g (E-5kg/M-s) 2.59 3.10 3.60
Pr 0.699 0.700 0.703

With these property values, the liquid film lengths and evaporation rates for
each test were determined at a coolant flow of 0.08 kg/s-m. Unlike Kinney, I(nuth
gave the mass flow o f the freestream gas directly, eliminating one unknown.
o

TEST G Lc mvcalc mvexp

(ka/s-m 2 ) (cm~ (ka/s-m 2 ) (ka/s-m2)
20-30 181.0 165.9 0.0482 0.0586
45-54 91.6 125.3 0.0638 0.0586
89-99 69.2 92.3 0.0867 0.0633

The calculated values agree fairly well with the m e a s u r e d values, however
there is not e n o u g h data to generalize.

35
AEOC-TR-91-1

4.1.3 G. Morrell LOx/NH s rocket (Reference 52)

These test were p e r f o r m e d in a 4 inch diameter rocket, with liquid oxygen

and a m m o n i a propellants. The liquid coolant was injected 2.8 inches downstream
of the p r o p e l l a n t injector. Water, ethanol and a m m o n i a w e r e tested as liquid
coolants. The a m m o n i a was at super-critical conditions and no sharp evaporation
points were determined, hence it is omitted here. Since the free stream turbulence
was not measured, a value of zero was used in the calculations.

The pressure varied from 16.9 to 17.7 arm in the water tests. For simplicity
the following saturated coolant properties at 17.4 arm were used in all calculations:

coolant TV(K) I1 AsJO

~-~] [Col
I~ ~V ~I a 91 9V Fcr
~., lkq-Kl - I
water 480 1.91 4530 15.9 129 36.2 857 9.09 0.198
ethanol 448 0.574 5126 13.1 121 7.91 637 27.6 0.143

The properties for water are from Reference 54. Most of the ethanol properties are
from Reid (Reference 55): Tv is from Appendix A, ~ from Equation 7-9.4, Pl from
Equation 3-11.9,/Jv from Table 9-2, and tJI from Table 9-8. The remaining ethanol
p r o p e r t i e s (Cpl, o, and Pv) are from various tables in Perry's H a n d b o o k
(Reference 56).

The free stream gas differed in composition for each run, with the O / F mass
ratio given by Morrell. The water coolant tests were all lean, with a reaction:
x I N H s + x~O 2 -, 1.SxxH20 + ~xxN 2 + (x2-0.75xl)O 2

The water vapor fraction, XH2 O, was related to the O / F mass ratio by:
XH20 ffi 4 8 / [ 4 0 + 1 7 ( O / F ) ]

The molecular weight of the products was calculated as:

M g = 544 r 1 + ( O / F ) 1
4 0 + 1 7 (O/F)

The following properties were used for each p r o d u c t species in the gas, at a
m e a n temperature of 1720 I~

C ~g Pr
(J/k~-K~ (10 -6 ka/m-s)
H20: 2684 5.85 0.9
N s : 1266 5.61 0.708
02 : 1167 6.78 0.75

36
 AEOC-TR-91-1

The specific heats were calculated from Table 3-2 of Holman (Reference 57)
for low pressure. From Holman's Figure 7.8, the corrections for chamber pressure
are negligible. The viscosities were calculated from the Chapman-Enskog
correlation, using the values in Appendix B of Reid (Reference 55). The Prandfl
numbers were extrapolated from the charts in Appendix A of Kays (Reference 5).

The viscosity of the products was calculated at an O/F mixture of 1.61, using
the mixture weighting of Wilke (Reference 55, Eq. 9.5-13), giving: #g = 5.86 l0 "s
kg/m-s. The specific heats were weighted by the mass fractions of the product
species and the Prandti numbers by the mole fractions for each run. With these
property values the liquid film lengths were calculated for each test:

Test O/F Tg P G
jCP g Pr F Llmea s Llcal c
[k_k_q_l [ a--I Fk-g-I
(K) ( a t m } Is-m21Ika-KI Is-ml (om~ (om)
water:
1 1.61 2950 17.4 226 2120 0.847 0.269 21.2 24.9
2 1.56 2960 17.8 234 2141 0.848 0.258 19.4 23.0
3 1.64 2945 17.0 225 2110 0.846 0.213 14.2 19.6
4 1.42 2978 17.4 233 2196 0.852 0.219 14.6 18.9
5 1.50 2963 17.0 207 2164 0.850 0.209 16.2 19.8
6 1.66 2940 17.1 229 2104 0.846 0.130 7.01 11.45
7 1.72 2935 17.5 220 2082 0.844 0.128 9.86 11.65
8 1.41 2978 17.1 209 2199 0.852 0.262 18.8 24.6
9 1.51 2963 16.9 214 2159 0.849 0.262 20.7 24.8
i0 1.75 2930 17.7 225 2072 0.844 0.290 20.4 27.7
II 1.67 2942 17.7 224 2100 0.845 0.296 21.7 28.0

ethanol:
12 I. 18 " 17.8 223 " " 0.809 14.2 50.0
13 I. i0 " 17.4 222 " " 0.869 16.2 54.4
14 i. 17 ,i 17.8 225 " , O. 8 3 4 17.4 51.4
15 1.12 " 17.4 211 " , 0.859 19.4 55.2
17 1.25 " 18.0 209 n , 0.872 17.7 56.1
18 1.17 " 17.7 228 " , 0.869 18.0 53.4
19 1.13 i. 17.6 224 " , 0.716 16.2 43.8
20 1.25 P 18.0 224 " - 0.713 16.2 43.5
22 1.41 2978 18.1 236 2199 0.852 0.570 14.6 33.1
23 1.37 " 17.9 225 " " 0.598 15.8 35.9
25 1.48 " 18.1 230 " , 0.535 ii.0 31.4
26 1.31 w. 18.2 237 " - 0.525 10.4 30.2
28 1.54 " 18.0 238 . - 0.341 10.3 19.0
29 1.53 " 17.8 225 " , 0.369 9.55 21.4
30 1.59 " 18.2 231 " , 0.318 10.2 17.9

37
AEDC-TR-91-1

A c o m p a r i s o n of the calculations with the experimental data for water

coolant is given in Figure 8. The calculated liquid film lengths are an average of
25% longer than m e a s u r e d . Since convection a c c o u n t e d for 62% of the total
evaporation rate, multiplying the convective heat flux by a factor of 1.4 w o u l d bring
the calculations into best agreement with the measurements. This corresponds to a
free stream turbulence of 9%, which is a reasonable assumption.

In all of the ethanol tests the liquid coolant flow rate i f ) greatly exceeded
Knuth's critical value for the formation of large waves ffcr), given in the coolant
table above. Such large waves w o u l d explain why the m e a s u r e d liquid film lengths
were an average of 2.75 times shorter than calculated, which compares favorably
with Kinney's results for the mass loss rate with large waves present. For simplicity,
the gas p r o p e ~ e s at an O / F of 1.41 were used in all of the ethanol calculations.

Katto's correlation for radiant b u r n o u t is tested in the program and predicted

liquid film lengths 10% shorter than listed above with water coolant. However, with
ethanol coolant the b u r n o u t condition predicted liquid film lengths of less than
I cm, which is unreasonable. Therefore the b u r n o u t prediction was n o t d e e m e d
trustworthy.

38
 AEDC-TR-91-1

4.1A Warner and Emmons H21air rocket (Reference 34)

These tests w e r e m a d e in a 79.8 m m diameter r o c k e t c h a m b e r b u r n i n g air

and H~. The liquid film cooling test section was 276 m m d o w n s t r e a m o f the
injector. The data given in their Figure 5 was analyzed. The gas was a s s u m e d d u e
to the stoichiometric reaction o f air and H 2 , giving 0.653 N 2 and 0.347 H~ O m o l a r
fractions, and a molecular weight o f 24.5. A t e m p e r a t u r e o f 2222 K and a p r e s s u r e
o f 34 arm w e r e reported.

Since the mass flow o f gas (Och) was not given, it was back-calculated from
the stated Reynold's n u m b e r o f 68,000. At the free stream t e m p e r a t u r e the species
viscosities w e r e 6.626 10 "s kg/m-s for N 2 and 7.186 10 "s k g / m - s for H 2 0 . Using
WUke's m e t h o d , the free stream gas viscosity was 6.806 10 "s kg/m-s. Using this
value: Och = 58.03 kg/s-m 2 .

The gas properties, evaluated at a m e a n t e m p e r a t u r e o f 1300 K, are listed

below:
C Pr
{J/k~-K) (10-ska/m-s]
N= 1226 0.705 4.680
H20 2494 0.90 4.584
mixture 1549 0.773 4.670

The p r o p e r t y values w e r e d e t e r m i n e d in the same m a n n e r as for Morrell's data.

Three different coolants were tested. The properties o f each at the c h a m b e r

p r e s s u r e are given below:

coolant Tv A CDI ~v ~I o Pl Pv Fcr

(K) rio.L1 f I flO-6k _1 flo-s 1 r rkal
[ kql I k a - K I I m-sl I ml I ma I Is-m]
water 514 1.76 4779 17.1 Iii.0 28.4 813 17.0 0.266
ethanol 480 0.440 5126 14.0 83.0 4.18 564 50.3 0.239
ammonia 345 0.937 4103 11.9 79.4 10.25 523 25.2 0.180

The coolant properties w e r e d e t e r m i n e d as in Morrell's tests, with the exceptions

that ~, Cpl, Pl, and Pv for a m m o n i a were scaled from the values at 525 K in Table
3-209 o f Perry (Reference 56).

With these p r o p e r t y values, the liquid film length and average evaporation
rates for each test w e r e calculated at a coolant flow 0") o f 0.08 kg/s-m, giving:

39
AEDC-TR-91 -1

Coolant Lo mvcal c mvexp

(cm) (ka/s-m 2 ) (ka/s-m 2 )
water 29.05 0.275 0.439
NH3 14.25 0.561 1.159
ethanol 16.85 0.475 0.886

Both the ammonia and ethanol data are somewhat questionable since they
extrapolate to a positive coolant flow at zero film cooled length, indicating an initial
mass loss at injection. The ammonia data also showed significant scatter. The flow
rates in the ammonia tests were centered about Knuth's critical value, which may
explain the scatter.

The calculated evaporation rate for water coolant is 5796 lower than
measured. Radiation accounted for 62.8% of the calculated evaporation rate,
making the calculations very sensitive to the assumed gas temperature. Given the
uncertainties in the test conditions, it is futile to speculate at length as to the
discrepancy.

40
 AEDC-TR-91 -1

4.1.5 R.C. Kesselring, et ai OFa/BaHe rocket test 0reference 53)

These tests were performed in a rocket engine using diborane fuel, which is
u n c o m m o n and very expensive. The combustion chamber was a thin n i c h e shell of
9.45 cm diameter, used for short duration tests. The free stream gas conditions
were falrlF constant for all runs, so that the average properties below were used in
analyzing all runs:
Tg= 3900 K, P=6.8 arm, Gch = 150.7 k g ] s . m a .

The NASA/Lewis Chemical Equilibrium code gave the following flee stream mole
fractions and properties:
0.537 HF, 0.225 BOF, 0.063 Ha O, 0.036 Ha, 0.031 OH, Mg=25.6

and a~ Tmean= 2200 K:

Cpg=3665J/kg-K, ~g=8.37 10"s kg/m-s, Pr=0.80

The properties of saturated BaHe coolant at 6.94 atm from Kit and Evered
(Reference 58) are:

Tv= 243.8 K, Mc = 27.69, X = 3.72 106 J/kg, Cpl = 3235 J/kg-K

#1 = 3.6 10"8 kg]m.s, al = 341 kg]m 8, pv = 29.4 kg]mS, o = 7.16 lO'SN/m

Kit and Evered give X = 5.173 106 J/kg at Tb= 180.04 K at 1 arm, which compares
well with Perry's h a n d b o o k (Reference 56). Scaling X as suggested in Reid
(Reference 55) to 243.8 K, using Tcrit = 289.7 K, gives the value above. The vapor
viscosity was not found. However, due to the very low liquid viscosity, the critical
flow rate for forming large waves, Fcr, is estimated to exceed 100 k g ] m . s, making
such waves unlikely.

Thermocouples were located on the outside of the thin chamber. A steady

temperature rise during a run was indicative of the absence of liquid film on the
wall at that axial position. Where the liquid film did exist the wall temperature
actually decreased during a firing, due to the low saturation temperature of
diborane. Unfortunately, the thermocouples were spaced 2 inches apart, making it
difficult to determine the liquid film lengths accurately. The best approximations
are given b e l o ~

41
AEDC-TR-91 -1

TEST I" (ku/m-s~ L c finl

1 0.583 3< Ix: <5
3 0.445 I< I,o <3
4 0.451 I< I ~ <3
6 0.191 i< Ix: <3
9 0.306 i< L c <3
I0 0.532 = 3
II 0.435 I< L c <3

Surprisingly, the authors assume that the liquid is immediately evaporated,

based u p o n calculations of the normally expected heat flux without transpiration.
They interpret the t h c r m o c o u p l e data as meaning that this vapor refrains from
mixing with the hot gases until s o m e arbitrary distance downstream.

No information o n the emittance of the combustion products was available,

so the emittance was treated as an adjustable parameter. Assuming 8% free stream
turbulence, the best comparison between the data and calculations was obtaincd
with an assumed emittance of 17%, as shown in Figure 9. Radiation accounted for
90% o f the heat flux in this case. The calculated liquid film length is s o m e w h a t
insensitive to the assumed emittance. For example, halving the emittance to 8.5%
increases the calculated liquid film length by 58%, whereas an 82% increase might
be expected. The reason is that the vapor liberated by radiation greatly decreases
the convective heat flux t h r o u g h transpiration. If the radiative flux decreases,
convection increases, keeping the evaporation rate somewhat constant.

Given the n u m e r o u s species in the products, o n e might expect a larger

emittance, however the optical density of 64 atm-cm was fairly low. To critically
judge the analysis for this test requires an i n d e p e n d e n t determination of the gas
emittance.

42
 AEDC-TR-91-1

4.2 GASEOUS FILM COOLING

4.2.1 Ewen and Rousar (References 49-51)

The goal in all three test programs was to study gaseous film cooling in a
converging-diverging nozzle, for application to rocket engines. The a u t h o r s
analyzed their data in terms of an empirical model. Although their model is similar
to the differential entrainment formulation of Section 3.4.1, it is n o t based u p o n
standard correlations. For this reason only their raw data was used.

In the first two reports (1972 and 1973) testing was d o n e on a flow bench in
a laboratory, using heated nitrogen for the free stream flow and either hydrogen or
nitrogen coolant. The free stream turbulence was measured to be 4.1% for all tests.
The free stream flow was fully-developed u p o n contacting the coolant injector,
leading to a difficulty in specifying the injection position from the "leading edge", a
p r o b l e m discussed later. Only those tests with a coolant injection velocity less than
the free stream velocity were analyzed (to avoid jet effects), although all of the data
was qualitatively similar. The test conditions are given below:

Test chamber Gch T coolant Mc Tc Aet

(kq/s.m2~ (3) ~kq/som~ (K) ~%)
1972:
3 30°-IR 630 763 H2 0.189 294 38
5A " " 629 764 H2 0.185 285 28
8 " " 636 756 H2 0.122 Iii 6
IIA conical 631 761 H2 0.185 292 0

1973:
101L 30°-IR 581 820 Ns 0.355 341 0
101H " " 577 822 N2 0.528 320 0
102 30°-2R 528 880 Hs 0.170 270 16
103A 15°-2R 591 816 H2 0.173 275 18
104AL ,, i, 599 815 N2 0.360 345 0
104AH " " 596 817 N2 0.677 327 0
105A con. m o d 597 817 H2 0.157 270 0

For test 104AL, the free stream temperature was stated as 915°K in Rousar
and Ewen's report, apparently as a misprint. Also, for tests 102, I03A, and 105A,
the stated coolant injection temperatures, T c, were adjusted to be consistent with
the axial t h e r m o c o u p l e readings.

In the last report (1977), measurements were made in an 0 2 (g)/H2 rocket

engine using hydrogen, nitrogen, and helium coolants. All tests used the same
chamber (P/N 1182134). The coolant injector was 178 m m d o w n s t r e a m of the

43
AEDC-TR-91 -1

propellant injector. The test conditions are given below:

Test O/F Gch Ta coolant Mc Tc Ae t

(ka/som 2 ) fg) (ko/s.m) (K) (%)
102 3.89 242.6 2670 H2 0.587 290 0.
103 3.81 228.5 2640 H2 0.504 290 0.
105 5.75 241.3 2990 H2 0.478 290 0.
II0 7.71 263.1 3080 H2 0.414 290 0.
IIi 7.99 260.6 3080 H2 0.362 290 0.
114 6.27 237.5 3030 H2 0.345 290 0.
115 5.93 234.9 3005 H2 0.408 290 0.
116 7.60 246.4 3075 H2 0.309 290 0.
117 4.00 219.5 2720 H2 0.391 290 0.
119 7.76 283.7 3080 He 0.515 310 0.
120 7.46 270.8 3075 He 0.469 310 0.
122 7.75 299.1 3080 He 0.595 310 0.

The coolant flow rates (Wc) were not given directly and were back-calculated
from the stated total mass flow (WT), the coolant to fuel mass flow ratio (Wc/WF),
and the O/F ratios as:

wc = (Wo/WF) WT/(I+O/F)
The Mc values tabulated above are the coolant mass flows (Wc) per circumference.

In all of the tests the products contained excess hydrogen. From the
reaction stoichiometry, the mass fraction of H 2 0 in the products is given by:

mH20 = 9 / 8 (O/F)/[I+(O/F)]

the remainder being H 2. The flame temperatures, Tg, listed are from a plot of
somewhat conflicting values from a n u m b e r of literature sources, mostly at 1 aim
pressure.

The specific heats of the various gas species were calculated as (Reference 57):

Cp[ N 2 = 1 4 1 5 - 2.88 10SK/T + 55.48 IOeK2/T 2 J / k g . K

Cp[ H2 = 12,047 + 2.176K-x T + 31,180I~/, /'I '~ J/kg. K
C p [ H 2 0 = 4615- 1.03 10sI~A/T~ + 9.68 10SK/T J / k g , K
Cp[ He = 5238 J/kg. K

These specific heats w e r e weighted by the species mass fractions to

determine the product specific heats.

44
 AEDC-TR-91 -!

The viscosity o f each species was calculated from the Chapman-Enskog

constants. The mixture weighting m e t h o d o f Wilke was used to d e t e r m i n e the
productMscosity (Reference 55):

t~l N 2 = 1.23 10 "e J T k g / m . s

#]Ha =6.7110"JT kg]m.s
# [ H 2 0 = 8.22 10-" J T k g / m . s

The parameters describing the chamber contours, discussed in Section 3.6,

are tabulated below:

chamber Dch 11 rI 0c Dt r2 0d
~mm~ (mm~ (mm~ (deo) (mm) (mm~ (deo)
30°-IR 31. 88.9 15.5 30. 15.5 7.8 15.
30°-2R 31. 88.9 31. 30. 15.5 15.5 15.
150-2R 31. 88.9 31. 15. 15.5 15.5 15.
conical 31. 7.9 12.7 4.4 15.5 7.8 15.
(modified) 31. 0. 12.7 4.4 15.5 7.8 15.
1182134 99.6 106.7 51.8 30. 52.1 28.7 15.

The results of each calculation are c o m p a r e d with the data in Figure 10. The
simplest geometry was test 11A, with a gradually converging conical chamber. For
comparison, the calculations are shown both with the differential e n t r a i n m e n t
formulation (Section 3.4.1) and with the simpler integral correlation (Section 3.0)
(based u p o n the starting conditions). The injection point was assumed at Xin = 0 for
both. The differential formulation gives the best c o m p a r i s o n u p to the throat.
D o w n s t r e a m o f the t h r o a t the wall t e m p e r a t u r e decreases m u c h faster than
predicted by the model, a difficulty with all of the calculations which is discussed
later.

The comparison is not as g o o d for the cylindrical chamber data, such as for
test 3. The wall temperature increases very rapidly downstream of the converging
turn (at 88.9 mm). The calculated value might be expected to rise d u e to the
increase in the free stream mass flow as the chamber area decreases. However, as
shown by the solid curve, this is not the case. The reason is that the contraction
also increases the b o u n d a r y layer mass flow per circumference, as discussed in
Section 3.4.2, and the two effects tend to cancel. Disabling this circumferential
c o r r e c t i o n t e r m yields a better c o m p a r i s o n , as s h o w n by the d a s h e d curve.
However, such fortuitous comparison is not a suitable reason for discounting a
physical effect, so the circumferential term is left in the model.

45
AEDC-TR-91-1

As in m o s t modeling efforts the comparison can be i m p r o v e d by judicious

selection of input parameters, as shown in the second figure. To force comparison
in the straight chamber section, a downstream injection point (Xin) of 20 m m was
used. While arbitrary, this could be d e f e n d e d by arguing that in fully-developed
flow the injection point is considered an infinite distance from the "leading edge".
However, with the exception of tests 3 and 5A, Xin=0 was always the o p t i m u m
choice. This is s u p p o r t e d by heat transfer data taken by Ewen a n d Rousar which
s h o w the initially high convective heat transfer rates o f a leading edge at the
injection point. Apparently the coolant injection disrupts the flow such that a n e w
b o u n d a r y layer grows from the injection point.

To force comparison downstream of the converging turn it was necessary to

assume an increase in the free stream turbulence d u e to the turn, Ae~ of 38% for
test 3. Similar c o r r e c t i o n factors are necessary with all t u r b u l e n c e m o d e l s
(Reference 59). The values required for each test are tabulated above.

Such a correction is defensible only if it is not arbitrary. This extra "turning

turbulence" can be correlated by a "centrifugal parameter":

Fc= (pg-Pc)v--~2rt
= rtpg
tG211-p~gg]

w h e r e r I is the radius of the converging turn, Pc and pg are the coolant and gas
densities at the turn. Note that Fc has dimensions. For all of the 1977 tests in the
rocket engine, the correction was negligible. Given the low densities in most rocket
chambers, this is generally true.

As shown in Figure 11, the correlation between Aet and F c can be fit by a
function:

~e t = 5.26 {[1+(0.029 Fc)S]t/s - 1} for Pg > Pc

= O. for Pg < Pc
where: FO <> kN/mS
The aberrant point is d u e to test 8 of 1972, for which it was impossible to obtain
any satisfactory comparisons.

It was not possible to obtain a reasonable comparison with the data of rests
119-122 which used helium coolant. The rate of temperature rise with non-reactive
h e l i u m coolant was overpredicted by a factor of 1.7, raising the question that
combustion effects are important and that the favorable comparisons with hydrogen
coolant in the rocket were merely fortuitous.

46
 AEDC-TR-91-1

For instance, the free stream temperature close to the wall might have been
considerably less than that used in the calculations above. Since many injectors provide
a fuel rich zone near the wall, this is plausible. Such an assumption would make the He
coolant calculations compare more favorably with the data. Normally this assumption
would also lead to a significant under-prediction of the wail temperatures with H2 coolant.
However, considering combustion of the H2 coolant in the boundary layer could restore
the wail temperatures to the measured values. Data with other coolants is needed to ascer-
tain whether combustion effects could explain the discrepancy.

In all of the calculations it was not possible to account for the rapid decrease
of wall t e m p e r a t u r e d o w n s t r e a m of the throat. Since the wall t e m p e r a t u r e
decreased even faster than the free stream gas static temperature, increased mixing
with the free stream gases could not account for this rapid decrease. Instead, it
might be attributed to the static temperature change of the b o u n d a r y layer gases as
they accelerate. Since the present formulation lumps together all of the boundary
layer gases, it w o u l d be difficult to incorporate such an effect into the model. A
turbulent b o u n d a r y layer model, which considers the velocity profile across the
b o u n d a r y layer, w o u l d be better suited to this task.

All of the calculated results assumed a recovery factor of zero, meaning that
the enU'ained free stream gas is at the static temperature. However, upstream of the
nozzle t h r o a t t h e r e is little difference b e t w e e n the static a n d stagnation
temperatures, so that the choice of recovery factor is not significant. Downstream
of the throat the present model is unable to accurately predict the wall temperature,
regardless of the choice of recovery factor.

47
AEOC-TR-91 -1

4.2.2 G.IL Kinney, et al (Reference 17)

The data in this report was reviewed in Section 4.1.1 pertaining to calculation
of liquid film lengths, which was the focus of their study. The authors do, however,
give one plot of temperatures downstream of the dryout point in their Figure 4.
This data is useful in determining how to apply the gaseous film cooling analysis to
the special case of liquid film cooling. The property values are from Section 4.1.1.

The start of the liquid film was 40 inches from the tube entrance, and the
liquid film was 20 inches long. As argued in the previous analysis of Ewen and
Rousar's data, the gaseous coolant injection tends to disturb the boundary layer
such that the injection point acts effectively as a "leading edge". The simplest
assumption is that all of the vapor is injected at the termination of the liquid film
(Xin=O). The calculated wall temperatures compare favorably with the data, as
shown in Figure 12. This is compared with an alternate assumption that the
effective leading edge is in the center of the liquid film (Xin= 254 mm), which gives
very p o o r comparison with the data. Therefore, as was found with Ewen and
Rousar's data, taking Xin = 0 generally gives the best comparisons with existing
data for all cases.

48
 AEDC-TR-S 1-I

5.0 CONCLUSIONS

A"simple one-dimensional model gives satisfactory comparison with existing

data for liquid film lengths in rocket engines. Convective and radiant heat transfer
m u s t both be considered. At high coolant flow rates large waves o n the film can
decrease the coolant efficiency. Knuth's correlation is useful in predicting this
transition. Radiant b u r n o u t of the liquid film is possible, however the use of
existing b u r n o u t correlations is questionable.

Downstream of the liquid film a standard gaseous film cooling correlation,

m o d i f i e d to a differential form, satisfactorily predicts the wall t e m p e r a t u r e s
upstream of the throat as the vapor mixes with the free stream gas. A correction
factor, correlated in terms of a "centrifugal parameter", is n e e d e d to account for
increased mixing due to the converging turn. However, in m o s t rocket engines it is
not significant. Downstream of the throat the wall temperature drops very rapidly
d u e to acceleration of the b o u n d a r y layer gases, which is n o t predicted by the
present model.

49
AEDC-TR-91 - 1

REFERENCES

I. Cebeei,T. and Bradshaw, P. Physical and Computational aspects of Convective

Heat Transfer. Springer-Verlag, 1984.
2. Schlichting, H. Boundary Laver Theory. McGraw-Hill, 1981.
3. Landis, R.B. "Numerical Solution of Variable Property Turbulent Boundary Layers
with Foreign Gas Injection". FhD Thesis, UCIA, 1971.
4. Reynolds, W.C., et. al. NASAMemo 12-I-58W, 1958.
5. Kays, W.M. Convective Heat and Mass Transfer. McGraw-Hill, 1966.
6. Humble, L.V., et. al. NACAReport 1020, 1951.
7. Bartz, D.R. "A Simple Equation for Rapid Estimation of Rocket No~le Convective
Heat Transfer Coefficients', Jet Propulsion - ARS Journal, Jan. 1957, p. 49.
8. Churchill, S.W. and Usagi, R. "A General Expression for the Correlation of Rates
of Transfer and other Phenomena", AIChE Journal. Vol. 18, No. 6, (Nov. 1972),
pp 1121-1128.
9. Barbin, A~R. and Jones, J.B. "Turbulent Flow in the Inlet Region of a Smooth
Pipe', Trans. ASME - |ournal of Basic Engineering. March 1963, p. 29.
I0. Pletcher, R.H. "Progress in Turbulent Forced Convection", Trans. ASME - Journal
of Heat Transfer. Vol. 110, Nov. 1988, p. 1129.
11. Hersch, M. ARS Journal. Vol. 31, (1961), p. 39 •
12. Talmor, E. A.I.Ch.E. Journal, Vol. 12, (1966), p. 1092.
13. Hartnett, J.P. et. al. in Heat and Mass Transfer in Boundary Layers. Vol.l N.
Afgan, editor, Pergamon, 1972.
14. Bird, R.B., Stewart, W.E., and Hghtfoot, E.N. Transport Phenomena. Wiley
(i 0)
15. Brunner, M. J. A S M E Paper no. 64-WA]HT-50.
16. Rubesin, M.W. et al. Handbook of Heat Transfer. W.M. Rosenhow, editor.
McGraw-Hill, Ch. 8, p. 171.
17. Kinney, G.R. et al. "Internal Liquid Film Cooling Experiments with ..."N A C A
Report 1087, 1952.
18. Knuth, E.L. "The Mechanics of Film Cooling". JPL Memo 20-85, September 1953.
19. Gater, R.A. et al. Jet Propulsion Center TM-69-1, Purdue Univ., contract Nonr
1100(21), 1969.
20. Hanratty, T.J. and Hersham, A. "Initiation of Roll Waves", AIChE Journal, vol. 7,
no.3, Sept. 1961, p. 488.
21. Woodmansee, D.E. and Hanratty, T.J. "Base Film over which Roll Waves
Propogate", AIChEJournal, vol. 15, no. 5, Sept. 1969, p. 712.
22. Tatterson, D.F., et. al. "Drop Sizes in Annular Gas-Liquid Flows", AIChE Journal,
vol. 23, no. 1, Jan. 1977, p. 68.

50
 AEDC-TR-91-1

23. Research Summary No. 36-4, Volume II. Jet Propulsion Laboratory, Pasadena,
California, September 1, 1960. pp 82.84.
24. Sutton, G.P. et al. "Advanced Cooling Techniques for Rocket Engines",
Astronautics and Aeronautics, January 1966, pp 60-71.
25. Siegel, R. and Howell, J.R. Thermal Radiation Heat Transfer, second edition,
Hemisphere, (1981), Section 17.5 - 17.6.
26. Ziebland, H. and Parkinson, R.C. "Heat Transfer in Rocket Engines",
AGARD-AG-148-71, Sept 1971, Ch.6.
27. Burrows, M.C. NASATND - 2541, 1964.
28. Monde, M. and Katto, Y. "Burnout in a High Heat Flux Boiling System with an
Impinging Je¢', Int. |ournal of Heat and Mass Transfer, Vol.21, 1978, p.295.
29. Katto, Y. and Ishii, K. "Burnout in a High Heat Flux Boiling System with a forced
supply of liquid through a Plane Jet", 7th Int. Heat Transfer Conf., Vol.4, Toronto,
1978, p.435.
30. Mudawwar, I.A., Incropera, TA. and F.P. "Critical Heat Flux in Falling Liquid
Films". Particulate Phenomena and Multiphase Transport, T.N. Veziroglu, ed.,
Volume 2, Hemisphere, 1988.
31. CPIA Liquid Propellants Manual, Chemical Propulsion Information Agency,
Laurel, Md.
32. Zucrow, M.J. and Graham, A.R. "Some Considerations of Film Cooling for
Rocket Motors", Jet Propulsion - ARS Journal, June 1957, p. 650.
33. Zucrow, M.J. and Sellers, J.p. "Experimental Investigation of Rocket Motor Film
Cooling", ARS Journal, May 1961, p. 668.
34. Warner, C.F. and Emmons, D.L. "Effects of Selected Gas Stream Parameters and
Coolant Properties on Liquid Film Cooling", Trans. ASME |ournal of Heat
Transfer, May 1964, p. 271.
35. Gater, RA. et al. Jet Propulsion Center TM-65-6, Purdue Univ., Contract Nonr
1100(21), October 1965.
36. Shembharkar, T. R. and Pal, B.R. "Prediction of Film Cooling with a Liquid
Coolanf', Int. Journal of Heat and Mass Transfer, Voi. 29, No. 6, 1986, p. 899.
37. Economos, C. "The Compressible Turbulent Boundary Layer with Mass
Transfer". Phd thesis, Polytechhnic Inst. of Brooklyn, 1968.
38. Kutateladze,S.S. et al. High Temperature (Soviet),Vol.1, Sept-Oct 1963.
39. Stollery, J.L. and EI-Ehwany, A.&M. "A note on the use of a Boundary Layer
Model for correlating Film Cooling Data", Int. Journal of Heat and Mass Transfer,
Vol.8, 1965, p.55.
40. Hatch, J.E. and Papell, S.S, NASATND-150, 1959.
41. Librizzi, J. and Cresci, R.J. "Transpiration Cooling of a Turbulent Boundary
Layer in an Axisymmetric Nozzle", AIAA[ournal, Vol.2, No.4, April 1964.

51
AEDC-TR-91-1

42. Goldstein, R.J. et al. "Film Cooling Effectiveness with Injection through a Porous
Section', Trans. ASMEIournal of Heat Transfer, August 1965, p.353 •
43. Goldstein, R.J. et al. "Film Cooling with Helium Injection into an Incompressible
Air Flow', Int. I. of Heat and Mass Transfer, Vol. 9, 1966, p.1341.
44. Marek, C.J. Taeina" R.R. "Effect of Free-stream Turbulence on Film Cooling",
NASATN D-7958, June 1975.
45. Carlson, L.W. and Talmor, E. "Gaseous Film Cooling at Various Degrees of Hot
Gas Acceleration and Turbulence Levels', Int. Iournal of Heat and Mass Transfer,
Vol.ll, 1968, p. 1695.
46. Burns, W.IC and J.L. Stollery, "The Influence of Foreign Gas Injection and Slot
Geometry on Film Cooling Effectiveness', Int. [. of Heat and Mass Transfer,
Vol. 12 1969, p. 935.
47. Hartnett, J.P. et al in Handbook of Heat Transfer. W.M. Rosenhow, editor,
McGraw-Hill, Ch. 17, Sec. B.2, 1973.
48. Eckert, E.R.G. and Drake, R.M. Analysis of Heat and Mass Transfer. McGraw-Hill,
1972.
49. Ewen, R.L. "Hydrogen Film/Conductive Cooling", NASA CR-120926, November
1972 (Aeroje0.
50. Rousar, D.C. and Ewen, R.L. "Hydrogen Film Cooling Investigation",
NASACR-121235, August 1973 (Aerojet).
51. Rousar, D.C. and Ewen, R.L. "Combustion Effects on Film Cooling",
NASACR-135052, February 1977 (Aerojet).
52. Morrell,G. "Investigation of Internal Film Cooling of 1000 Ib Thrust Liquid
Ammonia, Liquid Oxygen Rocket Engine Combustion Chamber'', NACA RM
E51E04, July 1951.
53. Kesselring,R.C. et al. "Boundary Cooled Rocket Engines for Space Storable
Propellants". NAS7-767, NASA-CR-129260,June 1972, (Rocketdyne).
54. Brodkey, R.S. and Hershey, H.C., Transport Phenomenon - a unified approach.
Tables A.1 and A.2 McGraw-Hill, 1988.
55. Reid, R.C. et. al. The Properties of Gases and Liquids. 4th ed. McGraw-Hill
(1987)
56. Perry, ed. Chemical Engineer's Handbook. McGraw-Hill.
57. Holman, J.P. Thermodynamics. 2nd ed., McGraw-Hill (1974)
58. Kit, B. and Evered, D.S. Rocket Propellant Handbook. MacMillan (1960)
59. Gibson, M.M. and Younis, B.A. "Modeling the Curved Turbulent Wall Jef',
AIAAlournal, 20, no. 12, December 1982.

52
 _J

O O

O
O o

O o

o
o FREE STREAA
O
GAS
O
O o
vAPoR

L I Q U I D F]L/~ ~-- I - ' ~ GASEOUS F I L/~ COOL I NG

COOL ! NG
Figure 1. Mechanics of Hquid film cooling.

31,
I11
O
0,,4
,3
I0
 m
O

:p
1.3 1-1
¢o
..s

Dafa of Ref. 9

1.2

1.1 " ~
m =[ .2
•

1.0 O

boundary layer
0.9

0 2 4 6 8 10 12 14 16
x/D
F'~quure 2. T r a n s i t i o n to fully-developed flow.
 1.0

0.9

0.8

Landis
0.7

0.6
St
Sto Rubesin
0.5

Brunner
0.4

0.3

Flow
0.2

0.1

I I I I ,J I !
0
0 1. 2. 3. 4. S. 6. 7.
F/St
Figure 3. Comparison of transpiration relaUons. m

co
d
 >
m
0
(3

t
100 l
82.5%
80 l
3000°K m m m m m m m m m m ~ q m m q~
,.t

H20 2000°K
60 1000°K

40

23 1%
2O
t~4J
4J
b,

¢I.
d 10
W
= 8
•~ 8
E
h.i
1000°K
1500°K t COz
4 2000°K

I
0.1 1.0 10.

Popt -- Pp" L (atm-m)

Fil~'e 4. Total emittances of H=O and (:O2 at 1 a m .
 @

Tg Am
i " N

/ / / ~ 1 ----~~s] Tow
,MI

'~ x xi P "'/~c mc
o b
X
S. Caseous film cooling ~ s i s .

-I
 m

]
I 2
I

Dch _ ,DL
e (:::::

Figure 6. Clamber contour parameter.

 0.40
D
2" smoolh
II
4" smoolh
E ~a
4" rough
C~ 0.30- ..........................................................................................................................................................................................................................
"4
v

[--I
t-
O
0.20-

0
O.
(0 i -.
>
(])
"(3
0.10-
:3
(/)

(I)
E

0.0¢ I
0.00 0.05 0.10 0.15 0.20

calculated evaporation rate (kg/s-m2)

Figure 7. Comparison with data of 6. R. Kinney, et al. 0"4
 0

(D
30"'" I i ' i

o 1 I I
0 5 10 15 2O 25 30

Calculated Liquid Film Length (cm)

Figure 8. C o m p a r i s o n w i t h d a t a o f G. Morreil.
 1=1
Eg = 17%

E
v
0

E)')
r-

E
oN.,..,,
~

"10

ET

0.0 0.1 0.2 0.3 0.4 0.5 0.6

coolant flow per circumference (kg/s-m)
Figure 9. Comparison with data of R. C. Kesselring, et al. 0

-I
,=
¢D
 rn
O

-- Test 3 (1972) -- D "~ ' ' -4

Test 3 (1972) ,=

- - xln = 0. - i ~ ':~- - i ~ "~ - - xln ~- 20 mm - -

e I j~..~-~t _e"

D o w m , ~ u m p o r t i o n (mini Domm poemon (mini

a. T e s t 3 ( 1 9 7 2 ) , ~ e t = O, x i . = O b. Test 3 (1972), Aet= 38%, xtn = 20 ,.in

" i i! I ~ " i, 1I li 1 t1
----Test 5A i ,~-- TEST8 ........ ! ~ .... I

I i i~-r"
Ii l E~

fi i /'
* .... * ! I
Downetream poeitlon (mm) Downmnmm posmon (mini

c. Test SA, A e t ~- 2 8 % , ] d n = 3 2 mum d. T e s t 8, A e t = 6 %

F i g u r e lO. C o m p a r i s o n with data of Ewen and Rousar.
 500

480 --
TEST11A,(1972)
i
i
i
i t~
L t..
800
750 -
TES'r 1OIL, (1973)

460' --
x''-°----!-- j• i 700
,I i
! <'
[]

I I ~ !
•, 440'

420

400
----t-----~

]
,
i
:
........... ~ - ~ i - . - ~ ~ l

i=
'--.~J~.-
I I

i
,
'

'
, ~
,= t 6O0
l

~
"--
,1- ~ t
,-,il

i
!
---

i L
i ~ I l , 550 / l
380
i 3i .
I i I 500 ! i
Ji
. . . . . . +.__..~ ~ ..................... . . . . + .............. + ............. @ ............. I
36O I ! m .

,/, ~ , _..........T--,~ ~_'. . . . - - - ~--1

~........... t i ,~o
340

/i
! I

320
I ! I ! I
4.00 I
I i i i i !
300 . . ~_~___~, ..... _~ l ~ - - ........ l 350
" i ! i z
! t/

280
t
i i
: i i i 1
i /
I 3000 2|0 , , i I
20 40 60 80 100 120 140 40 60 80 100 120 140
Dowl~mampoeitlon(mm) Downitrimmp~ill~(mm)

e. Test 11A (1972), xin = 0 f. Test 101L (1973)

700 r "I ~ 500 7" I' r T r .l
/ TEST 101H, (1973) ~ I / TEST 1021 (1973) =_ ~
8So-~[ ~---T-----r--iI 1/'~- ]' -i! - ]" -I ,150-I Act = 16% --i
6oo ....... ~ / , I J" ! I

i I ,~ i i i

i,oo/
J, o/ ---7 i -:. '
i ~ ' ' ' i i

'~°~-~~-
300
0
i
!
20
i
,.
40
•
60
Downstream
! I
80
I ~
100
pollitlolrl (mm)
i iI,
120 14.0
iI
~ 2 5 0 --I
0
i
20
,
40
I
60
Downstream
80
l
,

pmdtion
.,
~
100
(mm)
I
120 140 >
m
c~

9
-".1
g. Test 101H (1973) h. Test 102 (1973), ~et = 16%
F i s u r e 10. Continued. .3
 500 800
! i i 1 I I I
TEST 103A, (1973) ! f m

r-1 750' -- TEST 104A1., (1973) -

.-I
____~et=lS% _4 1 _ / I ,:p
450. 700'
=1,

~650
E ', | i J r , i/"
t"
t
/ ' i. i
l 400 800

Z. 550 !
E
.~ 350 500
m J/4i ii Ii m t
i 1,/-! I i 450
I
,. I]
I
500 /, '
t ,,i,
I
!I I
400
' ! 1 i ! !I
1
I
350,
! I J
i i ~ I , I
250 i 300
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
Oowniirellm poeliion (mm) Downstream position (mm)

1. Test 1O3A (1973), ~ e t = 18°7o J. Test 104AL (1973)

7°°1 TEST' 104AH,

' '
(1973) J , J~" I .60
-- TEST 105A, (1973)
i
I
400 J
e 5~o I
|i 5oo ~
i /~|
r/i,---1---~;
i ~, ,
i
,; ,i',
!
,'I
i i e
|,,o iI
- ~ J
I "/ ! ! ! I t I 35o
J
i

400 / r
' ~ ~
.I
~
I
i r
'
I I
-'"° ,
320
/4t
i
I
!I
350
]
I t
.---~-.--,---=-~
I I
I
~
!
~
I
E
300 !
/ [ Ji

500
~
i
i . I ! ~ ,i
i
! i
1
i =
260
i

i
0 20 40 60 80 100 120 140 0 20 40 60 80 00
Downstream po/tion (ram) Downstream pmelUon (mm)

k. Test 104AH (1973) L Test 105A (1973)

Figure 10. Continued.
 90O ' ' S ' 900
x
. . . . . T E S T 102, (1977) i ! . . . . . .
T ^ 800-
800
;
2" 700 _ _ ~ _ _ _ ~
i. . . . . ~ 1_ _ ~ _I _ _ 700
, Sxxt: X

i 600
i 600 . . . . . . 2--
i
~
i÷ ÷÷
-+-+-.-*--.--4
I i
. . . . . .
' i÷ ÷ ÷ ~ ' T
I 500 500
m
400 . . . . . . . . . . . .

400 I
300 - 300- z i i
i
200 i 250
200
50 100 150 200 250
0 50 100 150 200
Downstream poMtion (ram) DownMrum poMtlon (mm)

m. Test 102 (1977) n. T e s t 1 0 3 ( 1 9 7 7 )

O~ 900 1000
' I +r~vXA
800" - - T E S T 105, (1977) 900 - - T E S T 1,1 O, (1977) --.

• , 700- !

i j/"÷
x
T
1

÷ " I" ÷
xx ~. 800'

S 600- i ..... ~ ~ ÷ _ I
3 700"
i , 4- i ÷
600-
500- . ~h""
. . . . . . . . . . .
i i E
=. $ 500- - - - - i ..... i. . . .

400 ...................... i ! ,
i I
400

300 30O i

200 200
50 100 150 200 5O 100 150 200 ]>
Downstream IXmHIon (mm) Downstream pooitlon (ram) m
o
-.4
~u
o. T e s t 1 0 5 ( 1 9 7 7 ) p. T e s t I I 0 ( 1 9 7 7 )
¢o
F i g u r e 10. C o n U n u e d .
 1100 lO0O /
j >
I ]
~
i
) A
m
1000 -- TEST 111, (1977) i 900- mTEST 114, (1977) J
'-4
,:p
900
i l

/
(¢1
...... I ~
~. 800 / ,< ~ ^ .A
I

G 800 700
I 700
I 6o6
J 600 i ÷ ...........
J 500
~¢ 500

400
400

300
/" 1
300 i
2OO
I
200 0 50 100 150 200
50 100 150 200
Downstream posnlon (ram) Downstream IXmWon (mm)

q. T e s t 1 ! ! 0 9 7 7 ) r. Test 114 (1977}

1000
O's i I , i
900' -- TEST 115, (1977) ooo- -- TEST 11'6, (1977) / x ~

i
800

700 '
',

] ~
I
4-
]
T ¥
--- soo

700
f
600 i..
I
.-~ i
,f
]l i 6o0
E
500 / ~ iJ ~ i
r ................. ............... - 500 j / ,
400
500
.... ~ 400
/,
/ " J 300 . I
200 ; 200
50 100 150 200 50 100 150 200
Dowrmtream p o ~ l i o n (ram) Downstream p o t i o n (ram)

s. T e s t 115 0 9 7 7 ) t. Test 116 (1977)

Figure 10. C o n t i n u e d .
 IOOOi i ~ "x ' Iooo /

900 TEST117,(1977) .......... ~ •

t~o 9ool ,/ ~,
J x x
x I x

l '~ i ~ x,~ x l ~ 8ooi ~ t - ~ .... I ~

-) s00 ................ h ............ ~'. . . . . . I - / / . . . _ j ~ , , ' _,__

.........÷.~...TEST 1191 (1977)
................
I/ i i .e Coo,..*

200 1 [ ] ] / ,oo, , , , ,i , * ,
0 50 100 150 200 0 50 100 150 200
Downstrmmm IXmition (mm) Downstream p o ~ l l o n (mm)

u. Test 117 (1977) v. Test 119 (1977)

1100] i / i ~ ilxl 1°°°1 '/ : i

l OOO~........................... ~ - - - / ~ ..................... ~ . . . . ~ - ~ . . . . . . . . 9oo~----'./-. - ~ .............

9OOl-- ................/ - - ...........~ - ~ - + ~ - - ~ ~ ,oo ..... ~ ~ ~---.:-----~-

x , i -= 700 .... -~ T ...........................~ ....

; ~ ÷ = ÷ I-- / i - ~ * ~ ÷

i sooJ' [ ~J~" .................... ~ ' TEST 1 2 0', (1977) "ook-/--i,

J i .....
- soo I ~ - - i i TEST 122, (1977)
soo ...............~..._.i................... i- He Coolant ~ I/ "! He Coolant
400 • ! i j

30O 3oo V ~ '= J

0 50 100 1S0 200 0 S0 100 150 200
Downstream position (ram) Downmream position (mm) m

-I
w. Test 120 (1977) x. Test 122 (1977)
Figure 10. Concluded.
 m

n
-4
,=
CD

A
t
V

r- i
iO

(n

(1,1
t. ;
O
t-- i r-i

O0 E}
ii n
O
Q i
i

,,,Q 1

I--1

-100 0 50 100 150 200 250 3OO

Fc (kN/m3}
Figure 11. Correlation of turn turbulence with centrifugal parameter.
 x i n = 0. D

f O
r-I
/ D

r-I

xin = 2 5 4 m m

300
0 100 200 300 400 500 600 700
Downstream position (ram)
Figure 12. Comparison with Kinney's Figure 4.
AEDC-TR-91 -!

NOMENCLATURE

A = local cross-sec'donal area o f c h a m b e r

Aw = absorp~vity o f c h a m b e r walls
Cf = skin friction factor -- 2 ~ w / o g U g ~
C p -- specific heat per mass
D = diameter of combustion chamber at position x
e t = free stream turbulence intensity fraction
aet= increase in turbulence due to converging turn
F = blowing parameter = m v / G
G c h = free stream gas mass flow per area = p U g
G m e a n = mass flow at T m e a n relative to liquid surface = p m e a n C U g - U l)
h = convective heat transfer coeff.
rI = F / S t .
K = Gpg°'2S/ict'2s
Kg = t h e r m a l conductivity o f free stream gas
KM = c o r r e c t i o n factor f or m o l e c u l a r weight
Kt = c o r r e c t i o n factor f or t u r b u l e n c e
L = optical path length
L etf = average optical path length
m v = total liquid e v a p o r a t i o n rate p e r surface area
M = m o l e c u l a r weight o r
Mbl = b o u n d a r y layer mass flow rate p e r c i r c u m f e r e n c e
M c = g a s e o u s c o o l a n t mass flow rate p e r c i r c u m f e r e n c e
n = no. o f m o l e s
N c = m o l e fraction o f CO2
N w = m o l e fraction o f w a t e r v a p o r
Pr = Prandfl n o o f gas = / J g C p g / K g
P = absolute p r e s s u r e
Q = h e a t flow p e r a r e a
r = r a d i u s o f c o n v e r g e n c e arc in nozzle
Re D = R ey nol d's n o b a s e d o n d i a m e t e r = GD//~g
S t = 5tanton n o - h / ( G Cpg)
T = absolute t e m p e r a t u r e
T c = t e m p e r a t u r e o f liquid o r ga s eous cool ant
T v = saturation t e m p e r a t u r e o f c o o l a n t
AT = T Tv
U = axial velocity

"t0
 AEDC-TR-91-1

x = d i s t a n c e f r o m injection p o i n t ( a l o n g c o n t o u r )
xe = x corrected for developing pipe flow
X ffi d i m e n s i o n l e s s d i s t a n c e = K x
y ffi d i s t a n c e f r o m wall

Greek Symbols-

-t = C p / C v = specific h e a t ratio
6 = b o u n d a r y layer thickness
e = gas emissivity
= film c o o l i n g effectiveness
F = liquid c o o l a n t m a s s flow rate p e r c i r c u m f e r e n c e
X = latent h e a t o f v a p o r i z a t i o n o f c o o l a n t
* = X + CplCrv-Tc)
# = d y n a m i c viscosity
p = mass density
P o p t = optical d e n s i t y -- P L
o = s u r f a c e t e n s i o n o f c o o l a n t o r S t e p h a n ' s const.

subscripts:

a w - adiabatic wall
bl - b o u n d a r y l a y e r
bo - burnout point
c - c o o l a n t , Olquid o r g a s e o u s )
ch - v a l u e s in cylindrical c h a m b e r s e c t i o n
c o n y - convective
c r - at t r a n s i t i o n to "large waves"
g - free s u ' e a m gas
I - liquid coolant
m e a n - m e a n t e m p e r a t u r e b e t w e e n free s t r e a m gas a n d liquid
o - f o r "dry-wall" c o n d i t i o n s , w i t h o u t t r a n s p i r a t i o n
tad - due m radiation
up - upstream
v - vapor
w - e v a l u a t e d at wall

71