714 Shorter Communications

AckrloM~lrdg~m~,lt-A portion of this material is based upon 6. T. F. Zien, Approximate analysis of heat transfer in
work supported by the National Science Foundation under transpired boundary layers with effects of Prandtl
Grant No. ENG-77-1X26X and the author thanks them for number, Inl. J. Heut Mtrss TrumfGr 19, 5 I3 521 (1976).
their support. 7. T. F. Zien. Approximate calculation of transient heat
conduction, A/AA JI 14(j), 4044406 (1076).
x. W. M. Kays. Corwectioe Hw cold Mass Trmsfer. p.
222. McGraw-Hill, New York (I 966).
REFERENCES Y. H. Schlichting, Boundary La~cr Thcor~, 4th Edn
1. V. N. Volkov, A refinement of the Karman- (Translated by J. Kestin), p. 214. McGraw-Hill, New
Pohlhausen integral method in boundary layer theory, J. York (lY60).
Engng Phys. 9(5), 371-374 (1965). IO. C. R. Illingworth, The laminar boundary layer as-
7. V. N. Volkov and V. K. Li-Orlov, A refinement of the sociated with the retarded flow of a compressible fluid,
integral method in solving the heat conduction equa- ARC R & M 2590 (lY46).
tion, Heat Transfer--Soviet Res. 2(2), 41-47 (1970). I I. J. Sucec, Approximate solution to a class of transient
3. L. A. Bromley, Effect of heat capacity of condensate, forced convection problems. J. Engag Pw 99(4), Ser. A,
I/Id. Emyq Chrm. 44( I’), 2Y66-296X (I 052). 567~ 574 (lY77).
4. T. F. Zien, A new integral calculation of skin friction on I2 B. T. Chao and L. S. Cheema, Unsteady heat transfer in
a porous plate, AlA/1 J/9(7), 1423-1425 (1971). laminar boundary layer over a Rat plate, Int. J. Hrut
5. T. F. Zien, Skin friction on porous surfaces calculated Mass Trrrmfer 11. I31 I 1324 (lY6X).
by a simple integral method. AIAA J/ 10(1(j), 13. H. T. Chao and D. R. Jeng. Unsteady stagnation point
l267-- 126X(I Y72). heat transfer, J. Heat Trtrmfrr 87. 221 -220 (1065).

Inr. J. Hear Mass Trunsfrr. Vol. 22. pp 774-176 0017 9310’79’0501-0774 $0200;0
@ Pergamon Press Ltd. 1979. Printed in Great Britilm

ON PREDICTING BOILING BURNOUT FOR HEATERS

COOLED BY LIQUID JETS*

J. H. LIENHARD and R. EICHHORN

Boiling and Phase Change Laboratory, Mechanical Engineering Dept.,
University of Kentucky, Lexington, KY 40506, U.S.A.

(Rewicrd I2 Jurlc lY7X trod ill revised fiwm IO October 197X)

NO%lENCLATURE where 4 is the vapor escape velocity at the peak heat Hux,
A, area ; r,,,“,, divided by the jet velocity. u,. Thus $ = q,,,/p,h~gu,.
0, exponent expressing We, dependence of 6; The ratio of the saturated liquid and vapor densities is
D, disc diameter; designated as r and We, G p,rufDjs. where D is the
d, jet diameter; diameter of the heater.
i?,k.E., rate of energy ; rate of kinetic energy; More recently Katto and Ishii [3] achieved comparable
latent heat of vaporization ; success in correlating data for plane jets impinging at a 15”
‘I&
4mw peak (or “burnout”) boiling heat llux ; angle to a square heater. This time they based WrJ on the
r. P rlP, ; length of the plate and found:
[If, liquid velocity ; f$ = 0.01 64r”-“h7/We~“. (2)
11’vapor velocity;
UP“cl,” maximum vapor velocity;
YmrJPa~l ,y ;
We,, “liquid” Weber number, p,u:D/rr.

Greek symbols
3. fraction of liquid How directed into spray;
P. D/h ;
d. surface area average droplet diameter;
P/.P,, saturated liquid and vapor densities ;
fl, surface tension ;
4. by,,,!~~y)
= Ym.JPs~frs[lr.
INTRODUCTION

MONDE and Katto [I] recently provided an extremely

successful correlation of peak heat Hux (q,,,..) data for
saturated liquid jets impinging on discs. Their configuration
is shown in Fig. I, They correlated 937; of I50 original and
previous [2] qmrx data for water and Freon I 13, within
+3Sy,, of the best line through them. The correlation can
be written as:
4 = 0.0745r0 ‘*s/We:3 (I) we,= -
p,du:
c

*We are pleased to acknowledge support for this work

under NSF Cirant ENG77-2502’). FIG. I. Jet and heater configuration with nomenclature.
 Shorter ~ommuni~tions 175

In 1076, we addressed a different external How-boiling splashed out than there was with water. Since water has a
burnout problem, namely that of predicting the peak heat value of r almost an order of magnitude greater than the
Hux during the cross-How of saturated liquid over a other two fluids, we shall presume that z = a(r). Equation
horizontal cylindrical heater [4]. To predict q,,,,, in that (5) then takes the form:
case we offered the following mechanical energy stability
criterion:
The vapor removal wake must become unstable and
collapse when the net mechanical energy added to it by
THE EVALUATION OF k/:6)’ ’
the heater becomes positive.
The term (d/6)“- in equation (6) presents a serious
In that’tlow configuration, vapor kinetic energy is added difficulty; however we can obtain some guidance in its
and newly created surface energy in the escaping vapor evaluation both from dimensional analysis and from the
sheet consumes mechanical energy. When too much vapor literature. The functional dependence of d should be:
kinetic energy is added to the sheet it becomes unstable and
must widen abruptly to lower ug. In so doing the sheet 6 = 6(p,, ps, rg, 0, liquid film tl~ickness)
detaches from the leading edge of the cylinder, insulates it where the liquid velocity is probably unimportant and
and brings about burnout. where the film thickness (see Fig I) depends on the local
Our present aim is to try to apply this criterion to the radius of the disc. Since most of the disc area lies close to a
boiling configurations studied by Katto et al.. and in so radius of D/2, we can replace the thickness with d( I -cc)/fi.
doing to discover more about the physical mechanisms There are six variables in length. force, and time, so we look
involved in jet-cooled heater burnout. for three dimensionless groups:

We shall formulate a burnout prediction for the disc

configuration shown in Fig. I and when it is complete we
shall note any differences that might arise in the plane jet If the groups (with the exception of pr/p,) obey a power-
configuration. At burnout, the rate of mechanical energy law relation. then this becomes
supplied to the “wake”configuration equals that absorbed by d 1.1
it: _ 8’ “(Wr;@‘/j-“)f;(r). (7)
0 6,
K.&r + ~~~.“~~,~~~“~~~~~
= &err;lee+ K:E.,i,,ti braring. (3)
In 193X. Nukiyama and Tanasawa [S] provided the
The vapor generated next to the surface blows through classical expression for the Sauter mean diameter of
the sheet, tearing holes in it. Its primary elfect is to divide droplets formed in spray nozzles. It consisted of two terms,
the liquid into two portions. One portion (with mass How the second of which is a viscous term that is quite negligible
fraction, U) is a spray of droplets of average diameter,? S, in our range of interest. The first term is of the form:
which absorbs a great deal of surface energy and departs at
a slightly upward angle from the plane of the heater. The
other portion (with mass How fraction. 1-a), consists of This expression is missing a lel~gth-presL~mabiy an
the shards of the original sheet that remain attached to the appropriate jet diameter. d. With d-“Z multiplied on the
heater and depart horizontally. We also presume that the left of the expression above, it can be arranged as
velocity of both portions remains close to or. Thus the
extreme left and right hand terms in equation (3) balance
one another. The remainder of the equation then takes the
form:
This result would give a as I/6 in equation (7). and no
dependence on r. Any handbook or survey article on sprays
will offer competing formulas for droplet size in different
spray-making configurations, none of which are the same as
which reduces to the one under consideration. They generally yield values of
(I on the order of magnitude of I/h or less and minimal, if
(5) any, dependence on r.
Returning to equation (5) with equation (7) we find
where the quantity @ = D/d = (A~rrn/4tn)“Z is one which L ~~(~~~~~~~~~,~-u
V2”
Katto et oi. did not include explicitly in their correlation.
Or
The actual appearance of burnout under the postulated
@,j p3 (1 - .%li),f3- 60,
mechanism is as follows: As burnout is approached the
(9)
lower (horizontal) liquid How remains intact, and the spray E * i-1WC,
above it is quite fine. Then the lower How suddenly stops,
the spray becomes more coarse, and the angle between the The Monde and Katto correlation for the axisymmetric
plate and the spray increases. The liquid ceases to absorb jet arrangement leads to fs(r) = r’.“s while Katto and
kinetic energy and to protect the plate, which is now I&ii’s result for the plane jet geometry gives f;(r) = r”.s6’,
insulated around its periphery. This model is verified both In the latter case p would be (AII(.~,~:/A~,)~~’instead of D/d.
by Monde and Katto’s description and by our own Both cases have (I = 0 so that the /? dependence is absent in
observation of the process in a barbotage analogue which the equations. Geometry effects evidently remain, but more
we built in our laboratory. importantly, the assumptions of power law dependences
The work of Katto and I&ii casts some light on the throughout (that we have also made, in part) might not
nature of r in our equation (5). They observed that with jets adequately reflect the true mechanisms. As we will show
of Freon 1 I3 and trichloroethane there was far less spray below, a slight dependence of & on p is present, and it is
only approximately true that n is equal to more over the
whole range of we,.
tWe are being a little casual about averages here. 6 is the
diameter of the droplet with auernye ,sur/bce eriergy. Later COMPARISON OF CORRELATION WITH DATA

we relate IS to a Sauter mean diameter. Since the various We have plotted the group /@/r’.“’ against We,r!fiJ in
averages differ by a constant factor which need not be Fig. 2 using the saturated boiling data that were correlated
evaluated, no harm will be done. in [I]. Some error was doubtless incurred by scaling
776 Shorter Communications

FIG. 2. The data of Mondc and Katto under the present scheme of correlation.

numbers from the graph in [I]_ but this should be small in CONCLUSION
comparison with the inherent scatter of the points. Figure Z
The rate-of-energy stability criterion advanced in [3]
reveals:
appears to be applicable in the present problem. Its use
(I) About 94’5, of the data correlate within &Xzi of the
results in a predicted correlation which works slightly
best line through their midst. This is a substantially better
better than-although it is similar to-the empirical
correlation even than given by equation (I ).
expression given by Monde and Katto. The present analysis
(2) The slope of the curve varies from 3/X on the left suggests two foci for future experimental work. (I) The
through I,‘3 in the middle to l/4 on the right. These slopes
spray mass Row fraction largely dictates &(r). Thus Y
give values of (4 in the droplet diameter relation. equation
should be observed as a function of I’ and of the
(7). equal to - 114, 0 and + l/h (Nukiyama’s case),
configuration. (2) Spray droplet sizes should be measured
respectively. They are reasonable values in the light of what
in these boiling configurations to fix “a” a priori.
droplet diameter expressions look like in other cases. And it
appears that Monde and Katto’s choice of Weft3 was the
best single exponent they could have used.
The data of Katto and Ishii for planar jets are less
abundant than those of Monde and Katto and they show a HEFEREYCES

little more inherent scatter. Our correlation of those results M. Monde and Y. Katto, Burnout in high heat-Hux
gives a = 0. This collapses our correlation, equation (9), boiling system with an impinging jet. Iat. J. Neat Mass
exactly into equation (?)_ and we get the same numbers as T~arlsf~r 21, 29s-305 (I 97X).
Katto and fshii did.* Y. Katto and M. Monde. Study of mechanism of
In the present correlations we have taken n to be a burnout in high heat&x boiling system with an
constant. although nothing in our dimensional analysis impinging jet, in Proceedings of the 5th I~~termdmzl
prevents it from varying with r. We have aiso accepted Heat Trarrsf~r Cm&rerrce. Tokyo, Vol. 4, B6.2. Japan
Katto et al.‘s power laws for f;(r). Unfortunately there are (1974).
only 2 values of r, in each case. with which to fix these Y. Katto and K. lshii, Burnout in a high heat fIux
functions 12 of 3 values in the Katto-Ishii paper are almost boiling system with a forced supply of liquid through a
the same.) Professor Katto has informed us of ongoing high plane jet, in Procredirrgs of tlw 6th Irfternatiorml Weat
pressure experiments which will greatly extend the range of Truasf& CorQererrce, Toronto ( 197X).
r. When data from these tests are available, the r-indepen- J. H. Lienhard and R. Eichhorn. Peak boiling heat flux
dence of a and the function f;(r) can be questioned further. on cylinders in a cross flow, fur. J. Heat Mass Tram&-r
19, 1135-l I42 (1976).
*We are grateful to Mr. Md. Z. Hasan for correlating the S. Nukiyama and Y. Tanasawa, Trans. Sot. Mech. Enyrs
Katto-lshii data for us. (Japan) 4,86 (1938).