ASSI F I E D

C LAC

AD 274 109

ARMED SERVICES TECHNICAL INFORTON AGENCY

ARIJNfN HALL SATM
ALINGTON 12, VIRCJNIA

UNCLASSIFIED
NOTIC3: Mhen goverment or other draings, speci-
fications or other data are used for any purpose
other than in connection with a definitely related
goverment procurement operation, the U. S.
Goverment thereby incurs no responsibility, nor any
obligation vhatsoever; and the fact that the Govern-
ment my have forulated., fuarnished., or in any way
supplied the said dravings, specifications, or other
data is not to be regarded by implication or other-
vise as in any manner licensing the holder or any
other person or corporation, or conveying any rights
or permission to minufacture, use or sell any
patented invention that may in any way be related
thereto.
 ARL 189
tog7.,0

SURVEY ON HEAT TRANSFER AT HIGH SPEEDS

jST R. G. ECKERT

I~ UNIVERSITY OF MINNESOTA
MINNEAPOLIS, MINNESOTA

*_ 5 ,-3 r

"DECEMBER IldI

AERONAUTICAL RESEARCH LABORATORY

OFFCE OF AEROSPACE RESEARCH
UNITED STATES AIR FORCE

'k-
ARL 189

SURVEY ON HEAT TRANSFR AT NIGH

ERNST R. G. ECKERT

UNIVERSITY OP MINNESOTA
MINNEAPOLIS, MINNESOTA

DECEMBER 1961

CONTRACT AF 33(616)-214
PROJECr 7064
TASK 70138

AERONAUTICAL RESEARCH LABORATORY

OFFICE OF AEROSPACE RESEARCH
UNITED STATES AIR FORCE
WRIGHT-PATTERSON AIR FORCE BASE, OHIO

200 - April 1962 - 27-1139 G 1140

 FORWORD

This technical report has been prepared by Dr. 'T. R. G. Eokert, Professor of
Mechanical 1 ngineering at the Univrsity of Minnesota under Contract AF 33(616)-
2214 for the Aeronautical Research Laboratory, Office of Aerospace Research, Wright-
Patterson Air Force Base, Ohio. The work reported hprein was accomplished on Task
70138, "Investigation of Heat Transfer" of Project 7064, "Research on Aerodynamic
Flow Fields". Mr. Trich Soehngen, ARL, was the project scipntist.

This technical report supersedes WADC TR 54.70 dated April 1954.

Appreciation is extended to Miss Carolyn Byers for typing the manuscript.

i
iii
 this report In a relse at VW U Si.-70 ublob Wresemted a ovey of tb.
3itera6' beat tPImOfe at "S in@lurmi voloeltles pub~lahd prodo.1mst3y
Gam
la Soe psme 3953 to 3951. !be IMP MOW Of PeM' made It nmory to
resklet owwenueeemlaw~ to mfae..s along skoh the wonue" is easmtat
MM 6001WO t too zestfiotive sloGs swidamme ... 3 be presek thIat
I me4
for deale a-Q1 a. gpza32 ume Sn= bsp ew airoraft the Snftme of a
pom varietis OR beet, tsemote Is m31. V) rq~ela6tions bays been o3Jsocted
or developed ddah pewdft the elotm of friotie fto roewowy fwtetov
sod beet tminda eaffioemt for 32mnr bowdary 3aar flow eLoq serfosee of
oomtat teqeratee'.a with an aourso better the 3 per cent. A greaer momr-
leftV I&the pveIsetiom at the abov ar mebu is caused by the poor I2efte
af ?raelul or values for air at high tqwuSe. A decided aiwmm t the
rebram pveaute Sn this rertu 3 h fact that they appear to ho34 for
iS n
ow reammsema Phou mdtlw ymriatiem.
heiti fw frIOUGS, rMOMy md bat tuMefer have also bass MOOMSIOed
for tindmb bomfy lepers. A inthod he. bam deuorihed by dusk the offet of
as aritmd3 inp
myteqw~ s- -aogte
ee inaem be establidod. go
Salme afthOw mamWlso skekms a ofdr mmiil at f2Jht at hM* alte.e
adl of oun at b* fliat veloltie has beem. Gemmsed. Ifo
ooeuate
this dSun e eSfjmeaU as wall as from the muvqp binr3 Via rqpns
boom Upaum-%I dulb ft resema s at qpesal ocineu'Sm Sepormo.
 LIST OF COrTrMS
Section Page
I INTRODUCTION ... . . . . . . . . 1
II COIMRSION OF FRICTION AND HTAT TRAWSF'R RnLATIONS FOR WrDGr.S,
COMS AND CYLIDARS INTO RELATIONS FOR FLAT PLATES . . . . . 4
III HEAT TRAWSF R TO A FLAT PLATE WITH CONSTANT SURFACE TI§4P"AATURE . 7
LAMINAR BOUNDARY LAYER FLOW. . . . . . . . . . . . . .*. . . . . 8
Constant Property Solutions ....... . . ... 9
Solutions for umkmT4 . .. . .. . . .. ........ . 10
Solutions for Air with Actual T-mperature Variation of all
Properties . .. .. .. .. . . . . . ...... .... 13
Reoreed Calculation Procedure ..... ............ 21
TURBUL T BOURDARY LAYER FLOW. ....... ................. 24
IV INFLUENCE OF PRESSURE AND TRMPRATURr VARIATION AO'G THE SURFACE
ON HvAT TRAMSM * * . . . .. . . . . ... . . . . ... . . 29
VARIABLE PRESSURE. .e * * o * * . *.
.. . . . . . ... . .. 29

VARIABL, WALL T!24PRATURE. . . . . . . . . . . . . . . . . . . . 30

V FRICTION AND HrAT TRAFSFTR AT LOW DM.SITIES .. .. . .... ... 36
VI FRICTIOq A"!HAT TRANSFER UNDER HIGH T-- PWRATURE CONDITIONS. . . 41
VII TRANSITION FROM LAINAR TO TURBULFNT FLOW. . . . . . . . . . . . .
AMDt'DUM.. . . . . . . . . . . . . . .. 45

Appendix

I R'F'RE!'TCE TFMPERATUR. ...................... 46

II HEAT BALANCE EQUATION. . . . .. . ... . . ......... 48

III LIMIT FOR ACCELRATION FOR WHICH BODUNDRY LAYIR-FLOW CAN BE

RIOARDRD AS QUASI-STTADY .. .. .. .. .. .. .. .. .. . .. 50

IV EVALUATION OF INTGRALS (Cb- 8) AND (Cb -9) NEAR S"z. . . . . 51

V REFTRENCS . . . . . . . . . . . ...... 53

I
 IBM or FZI3MS

1 Piofiles af Static Yepaew T an Total Tuesb T+ In

Ug Vm Doi~mday Lqer e0000
o .ee..e...ee ee . * 9e* a 73
2 Galoulatm Lamr Flat Plate Reovw7' Factor as FvUat± of
P11a e .e 0 0 0 0 00 0a 0 0 00 0 0 0 0 0 0 * 0 0 0 0 0 0 * 714

3 Pv,'p t Aofr a nointmofqwn . . .. ..00 ..... 7

14 CaIslatei M~ FrIctim CoeftiUlemt for Luinr 3cmavy Laque
anllFlat tU0 * * 0 00 0 0 a0e0 0 00 00 * 0 0 0 76

5 Teqieabw. Pz'ofilse ad~ Wummen Tepraue. In M& 9ps

Bounary Lqw.o. * so **0000000000
* *e*oooa 77
6 OaiiateI Lmdr lecov=7 Fetoi e" astm at Oask Disko. . .. 7
7 Osiamite Imaw hthaw aeMU7 rum"os hi FI#A t ~t. .* .*
a Ratio at Frictm Factr to hBAbsW Sbatm Iemb for Lo
lie a Folat late. *0000e*9 e*e*....e..e.... e... 0
9 ftato Aetual Omnt PgWtgw Tuft.imt Friation Factor for

10 0alcuate TbumtIamaotme forliA Mt . .*** .. SI

11 Nash wer~au'S tAIM Or3Jw . . 0 00 .
0 . . . . . . .*
0 83
12 Lt DIet, Transfer m Flat Plate viti Varible 1.31 Twpnrtmr. 61
13 OaIouiation Wabut bmof~r hAUN Fiat MWat edh TiJabIe Wal
fte ku**
00000 0 0000000000000000

35 bom ftbla llaest 0000000000000000000000ftes. ..

28 Awmae bAbo fo
Bbw for Ent ond asu'IOr lIn M£ Ar g
Fiet mijAt 00 00 0 00 00 0 00 00 0 00

20 Dal00atiu of~ Dfrne

D !MWU Ofo
' JAW
mr lieu AlUK
Fla PlMUat *****************......AUK
FI PMUe 9* 9e 99 e* *9 e 9
Figure Page
22 Calculation of Reference Temperature for Laminar Flow Along

LIST OF TABLE-S
Table Page
I& Wall at Recovery Temperature .......... 15

Ib Cole
0
Wall0. . . . . . . . . . . ........ 9
16

IV
 SCTION I

IO.TL'.,

The purpose of the present report is to rxesent a survey and suraary of

presently available information on heat transfer at hi6h supersonic speeds. It is
hoped that the paper is prepared in such a way that it is useful for the desi6ner
who, in order to calculate temperatures in aircraft at high speeds, has to have
information on the convective heat transfer between the aircraft skin and the
athosphere. It is well known that by the aerodynic heating effect aircraft which
fly with supersonic speeds, are heated up to such an extent that the skin terperature
and the temperature inside the aircraft require the special attention of the aesi6ner.
Corresponding to the gro,.in6 ipjortance 01 cooliiL p.obleias in acroi,.utics, tne
nutber of papers which deal with high velocity flow has steadily increaseu durini
recent years. This is evidenced by the fact that to-usy, for example, each issue
of the "Journal of tae %eronutical iciences* contains at least one contribution
to the heat transfer problem.. In this situation a swurb:ery has to be restricted in
its scope. In tis report only convective neat transfer will be discussed ana
this only for surfaces along wnich the pressure is constant. The literature
collected in the appendix is thought to be complcte in this field. such a restric-
tion is per.Assible since aircraft flying through the atroisphere at high speeas will
always have very slender sapes so that the pressure variation alon& the surface is
not large and its effect on heat transfer is of minor importance. Some re.,arks on
the influence of pressure variation are however included. Lxcluded also from tue
discussion will be surfaces which are cooled by a process known as transpiration
cooling. In this process the skin is manufactured froam a porous material and a
cooling ,jAediwd is blown through the skin into the atmosphere and builds up U cool
film which protects the surface from the influence of the hot boundary l~yer.
Information on this cooling process and its influence on heat transfer under high
velocity conditions is still meager (Ref. 1, 2, 3, 4). The discussion is also
restricted to heat transfer under steady state conditions. It has been shown that
for rates of change in spced as they are usually encountered in aircraft the
convective neat transfer may be regarded as quasi stationary (See Appendix III.).
This means the heat transfer at each instant may be calculated with formulae valid
for steady state into which velocity and property values are introduced as they are
encountered at this specific instant (Ref. .5,6).

The essential features of the temperature distribution as it is encountereu

near a wall moving with high speed may be discussed with the help of f'ibure 1. In
this figure the temperatures are plotted over the distance from the wall. In any
high speed gas flow two different temperatures have to be distinguished: the
static temperature T as it would be measured by an instrument which is at rest
relative to the flow at each location, and the total Temperature Tt measured by an
instrument which slows the flow adiabatically down to the velocity zero, measured
relative to the wall. The full lines represent the temperature aistribution for
the case that no heat is conducted from the surface into the interior : the wall
or transferred to the enviromiient by radiation. The profiles of both temperatures
are shown in the figure.
TIe static temperature increases throughout the boundary layer in the
direction toward the wall because of the heat generated by internal friction.
Under the conditions as specified no heat flow from the wall into the fluio occurs
and therefore the temperature gradient at the wall is zero. However, tne wall

1
temperature is hi~her than the static temperature in the stream outside the boundary
layer because of the effect of "aerodynamic heating'.
It is interesting to observe the behavior of the total temperature throughout
the boundary layer. This temperature measures the total energy content of the
fluid, internal and kinetic, and this energy content toes not change in a steady
flow from which energy is not extracted, either as heat or as mechanical enerar
both conditions are fulfilled by the boundary layer as a whole. Therefore, the
avera e value of the total temperature has to be the same within the boundary l.yer
as in the outside flow. dhen a defect of the total temperature in the layers near
the wall is encountered for fluids with a Prandtl iamber smaller than one, then it
has to be compensated by excess of this temperature in the outer region of the
boundary layer. For fluids with Prandtl number larger than one, the conditions are
reversed, the total temperature being larger near the wall and smaller in the outer
region of the boundary layer than in free stream.
The difference between the total and static temperatures at any point is
determined by the local velocity V. For a fluid with constant specific heat cp
this difference is given by the expression shown in the diagram. 4or variable
specific heat the difference betwegn the corresponding enthalpies it - i is equal
to the kinetic energy of the air V-/2. For gases under the condition of no heat
transfer, this wall temperature which will be called in this report the "recovery
temperature" is almost equal to the total temperature in the stream outside the
boundary layer. This leads at supersonic velocities to very high wall temperatures.
Fortunately, conditions are usually much more favorable on the skin of aircraft
flyinL at high speed because heat is radiated away from the surface of the aircraft
or, for short flight duration, heat is being stored in the skin or other structural
elements. The te.perature profile found in the boundary layer under such circum-
stances is shown by dashed lines in iure 1. It is interesting to note the
peculiar shapes wich the total temperature profile may assume. This profile is
the one which can be comparatively easily measured by total temperature probes,
whereas today no satisfactory direct measurement of local static temperatures is
possible.
From the discussion of Figure 1 it is apparent that heat transfer in high
speed flow is essentially influenced by two parameters which are usually not
present in other heut transfer problems: a marked effect of frictional heating
and a large terperature variation which occurrs in the neighborhood of surfaces in
a high velocity flow field, even when the surface itself is cooled down to the
outside stream temperature.

It will be shown later, that for the case when the wall-temperature is cooled
to the stream static temperature a maximum temperature within the boundary layer
arises which for a .LAach number 10 is five time the stream temperature or, for a
-a.ch number 20 is equal to 18 ti.es the stream temperature. These temperature
variations are unusually large and cause the property values to change extensively
throughout the boundary layer. The temperature within the boundary layer may
become so large that it causes dissociation of the air.
Frictional heatink as determined by the velocity of the air stream, large
temperature variation, and high temperatures are the parameters which will have to
be discussed extensively in this report with regard to their influence on heat
transfer,

2
 ,Lnthcr conition whi. is con.ected with ,Lircraft a 4l1icatiox is tie low
.ressure rin cor:-eu3oiinly tie lo;i density )f t.ae air, encountrca by ttie eircrcft
at hih fZlijit altitude. The air .;t niii altitudes is so rarifie.. t.aUt the leniths
of tue 1,Lths Wi;ich t.a si1Lnle Lir ;,olecules travel bet-cen collisions bcuiae
considerable nd of the swrie order f i.Lsnitu.e as the thickness of the boundary
l ,cr. Ihis c.*-ncs nuct transfer conditions from whxat is usully experienced in
air flow under norzr.al ipessure. This factor will also be disoussed in its
influence on heut transfer in one of the following chapters.

I.
 1-TO .&U. iTI;.-
. N Ti. ,"T ,i ,

..ectijn III w.ill 6CU. LXtLf~ivClJ .Iit:i t.llfriction and a t tr-nsfr

canLit,.ln F3 t.-e, tist is lam.inar or turbulcnt boundury layers on surfaces in
t.:o-cL.ensionul flo: for ,-igh tee jrcsurc does not ciuzx±~e in flow Uirectiun.
Juc.a conUiti ns ur rcalized on a flat plate parallel to the stream direotion s
long as the boundary layers are very thin so that the displacement of the flow bY
this boundary layer can be neglected.
Constant pressure is in supersonic flow also encountered along the surface of
wedges and cones when the shock waves by which the oncoming flow is turned into a
direction parallel to the surface are attached to the vertex of the object. tor
this case siaple conversion formilas exist bY which friction and heat transfer data
for cones and wedges can be obtaired from information available for flat plates.
These will be discussed in this chapter. Such conversion rules are Important
because parts of missiles or aircraft have conical shapes and also because maw
teat results are available on cones and wedges and can in this way be compaed with
flat plate data.
friction and heat transfer on the surface of ai cylinder in a flow Marallel to
Its axis are identical with the ones on a flat plate as long as long as the mallest
radius of the curvature of the cylinder cross section is large compared with the
boundary layer thickness. Some investigations as to the limiting conditions for
this analogy will also be discussed.

Jor a wedge in a supersonic flow which is directed perpendicular to its edg4

and which has attached shock waves the flow conditions are constant in the field
between the shooks and the surface of the wedge. Constant I.ach number, constant
total and static temperatures and constant density prevail in this field. dince
the flow on such a wedge Is two-dimensional and since the pressure along the
surface of the wedge is constant, the relationship for skin friction and heat
transfer, valid for flat plates, may be imediutely applied to the surfaces of such
a wedge. Conditions between the shock and the surface correspond to the fl.w
conditions at the flat plate outside the boundary layer.

A cone.may now be considerea which is located in a supersonic flow directed

parallel to the cone ais. The onstream conditions may be such that the shook wave
Is attached to the apex of the cone. The shook has in this ase a conical shape
with axis and apex coincident with the solid cone. Strem conditions in this case
are not constant In the field between this shook and the cone surface, however,
they are constant along iaW conical surface with the asme axis and apex as long as
the displacemnt of the flow bY the boundary layer Is neglegible. Ab thin boundary
layers, therefore, this condition will also hold along a conical envelope very near
to the cone surface and just outside the boundary layer. dhen friction or heat
transfer data from flat plates "re to be converted to that on the surfaces of cones
then the flow conditions, existing in this envelope which in pratically O ocident
with the cone surface, correspond to the flow conditions for the plate. agai the
pressure is constant along the surface in flow direction.
 The flow, however, is not two-diensional and therefore conversion forbiulas
have to be usdC in order to obtain fropi flat plate data relationships valid for the
surface of cones. Hantzsche and "lendt (Ref. 12) show by a transformation of the
laminar boundary layer equations for flow along the cone surface to the flat plate
case that the temperature recovery factor remains unchanged by this transformtion
and has the same value for cones as for flat plate. Local friction coefficients
and heat transfer coefficients, valid for the flat plate, have to be multiplied Iv
4Y to obtain friction and heat transfer coefficients on the surface of the cone at
the same distance from the leading edge and under identical stream conditions.
This fact may also be stated in a different way, namely, that local friction and
heat transfer coefficients are id ,ntical for cones and flat plates at locations for
which the cone Reynolds number i: times the value of the flat plate Reynolds
number. At such corresponding locations witiiin the boundary layer arc coiletely
identical and therefore not only the heat transfer coefficient and friction
factors but also the stability conditions within the boundary laer and therefore
the conditions for transition to turbulence are the same.
An analogous transformation on the turbulent boundary layer equations written
in time mean values has been made by Gazley (Ref. 11), and in a veiy general form
taking into account the variation of the property values by Van Driest (We. 9).
The result of this transformation is that again conditions within the boundary
layer are identical on such locations at the flat plate and on the cone surface for
which the cone Reynolds number is equal to twice the flat plate Reynolds number.
IV the s&iLe rules stated, friction and heat transfer can be calculated from
formulas which are valid for flat plates.
For sme purposes it is Important to know also how much the local and the
average friction and heat transfer values differ on the cone surface and the flat
plate at locations with the same R"nolds number. For laminar flow it is alrma4
stated that the local hat transfer coefficient and friction factor on the cone are
by the factor VTlarger than on the flat plate. The avoerage heat transf coeffi-
cient and friction factor for the cone are larcer by thp factor 2/ 'nil It
may be seen that tho average friction and heat transfer values do not diffor much
for cones and flat plates when they are compared at the same Reynolds number. For
the turbulent friction coefficient, using Blasius equation and Reynolis analogy,
the following rolationships for a turbulent boundary layr can be.derived: The
local friction and heat transfer cofficipnts are by the factor 2 * 1 15 larger
than for a flat plate. For th, avorai.e values h e factor is 1.022. Therefore,
the difference between average friction and heat traesfer coefficients on cones
and flat plates considering locations with the same Reynolds number is only 2 per
cent for turbulent flow. It may be added here that a transformation of the boundary
layer equations for arbitrary bodims of revolution in rotationally symetric flow
to the two-dimensional case has beon presented In Reference 14.
For a cylinder with circular cross section in a flow-subsonic or ajuemmio--.
which is parallel to the cylinder axis conditions within the boundary layer are
principally different frot., the conditions on a flat plate uince cylindrical areas
within the boundary layer I)Lrollel to the surface inercase in area 1roportional to
the distance froA. the cylinder axis. The influence of this condition on heat
transfer and friction in a turbulent boundary loyer hi-.s been determined analytically
by Jakob and Dow (Rof. 13), by it. U. Lckert (Ref. 10), and by Saeban (Ref. 17). no.
latter Investigator found thct thin condition does nut influence the recovery factor
so that flat plate recovery factors can bc tizaedtitely apl:lied to cylinders in axal
flow.
 Aith rerara to the averu6e skin friction the different predictions vary
widely. Al experij..cntal investieation in reference 65, z:,ude on cylinder-cone
coijbinations. found that at a Reynolds nunber around 10 an increase of the average
friction factor of .5 percent was connected with a change of the length L to diameter
d ratio from 8 to 23. The calcul.tion by Takob and Zw gives for a ohange of L/d
from 8 to 23 an increase of 1,5 .ercent. H. U. Takort's method results for the same
conditions in 1.5 percent increase. The evidnee presented doas not give a conclu-
sive quantitative infonation on how much friction and heat transfer on a cylinier
are different from tho one on a flat plate. It can b- concluded ho!,,vovr, "t-atup
to a ratio of boundary layer displacemant thickaess to cylinder radius of 1:100 to
1:50, the difforenoe botw.en cylinder and flat plate data should be nogligible.

For cylinders in a flow normal to their axis the condition for heat transfer
are rather involved. A bouriary layer builds up around the upstream part of the
surface. The Urge presaure variation along the surface makes the development of
this boUtdary layer quite different from the conditions considered before in this
chapter. The effect of a pressure variation on the boundary laer will be considered
In samwhat more detail in a later chapter. Heat transfer to the downstream part of
the cylinder where the flow separates Trom the surface is still poorly understood.
Extremely low recovery temperatures (partially below static upstream - temperature)
have been o served in this region at subsonic upstream speeds. (Pef. 178. 179)

6
 QiTION III
BL^T THANfaR TO A ?Lf.T PL.Th JITH CONSTANT &iEFAUl TWMPRATUM

This section deals with what is usually oalled the flat plate case. bbre
exactly, it deals with heat transfer in two-dimensional flow, the pressure of which
is constant along the surface. lor thin boundary lovers this type of flow Is found
on a flat plate. There are, however, special conditions like very low pressure,
very high velocity where boundary layers become quite thick. The flow outside the
boundary layer Is then displaced and curved along a flat plate, according to the
Increase of the displacement thickness of the boundary layer in downtrem direction.
As a consequence of this, the pressure decreases in flow direction. Constant
pressure in such a came would be rather obtained tV a slightly ooncave surface. In
supersonic flow, constant pressure along a surface is also found on welges adascon
with attached shocks. The simple rules with which flat plate formal"a for heat
transfer and friction can be transferred to the conditions on wedges and cones have
been discussed in the previous chapter.

A constant temperature along the surface is also specified for this chapter.
Generally, on missiles and aircraft the temperature will vary along the surface
and it is known today that such a temperature variation has a considerable effect
on heat transfer. The plate with constant temperature, however, serves as a
standard condition to which comparatively simple formlas apply. A variation of
the surfaces temperature has to be dealt with specifically IV calculations in seek
individual case. This procedure will be discussed in a later chapter.

Conditions on the wall surface of the plate will be denoted by a aubscript w,

conditions in the stream outside the boundary layer bV a subscript a. Jkoe It Is
assumed that flow conditions in the field outside the boundry lever (voloeIV V*,
temperature T. 9density 1P.) are counant.- The symbol T denotes absolute e rtm s
(in OR.). In the flow ? mans alwAWs statie temperature. The total teerat~ w
(as measured ty a total temperature probe at rost relative to the plate) Is
Indicated by a subscript t. The following dimensionless flow paramters ere used,

ae Lve (B- 1)

- s(B -2)

The Reynolds number Re is based on the distance A from the leding s4 of the
plate. The property values V and A in the AWnolds number will be introduced at
different locations (wall or strew.), whereas the ach namer bla is alwwa based on
the sound velocity a. at stream temperature.

7
 The local shearing stress #T w which the flow exerts on the plate surface Is
expressed by a dimensionless friction factor defined in the following way

- c s 0-
( 3)
The local convective heat flow per unit time end area at the plate surface is
described by a heat transfer coefficient h which is defined in the folloiing way
bI'-0)
U (B -4)
Tis the actual wall temperature and Tr to the temperature which the specific
location on the surface assumea when the convective heat transfer is zero. 2his
temperature will be cal..ed 'recovery temperature' since it is conventionally
described tbr the temperature recovery factor Sq. (Db-4). Te heat transfer
coefficient is In turn expressed non-dimiensionally tV the Nusselt umber
hk (B-5)
or TV fte Stanton number

To (B - 6)

UANIUR DUDARY LAY"R FLW

The differential equations which describe the flow end enagf conditions for a
lInasr boundary layer are well eetablished under normal pressure conditions In tge
contimaim flow regime., Muarous solutions of these differential equations have bees,
wod out and ane today available to give infozuation in a wide reap of sonditioas.
extending for Instance to very high bc umber and lsrge temperature diffexeaces.
These results generally cheik satisfactorily with experimental invasti.gatioms, when
the experiments were nod at the boundary condition. which correapond to the bouAi
Iqev solutions. Therefore * in the lmnar rap, the most reliable Information oam
be obtained frm calculations. The main difficulty In solving the differeatial,
equations is connected with the fact that the property values (viscosity,, beat
oadustivity, density, and specific heat) very for all gass and flids with
temperature and pressure. It is quite difficult to tabs ito proper acount the
actual varilation of the property values In the solution of the differential
equations. U flat plate came Is in this respect simpler than flow and heat
tresefer for general conditions. At the flat plate, as mentioned above, the
pressare Is constant along taie surface.* In the thin boundary layer the pressure
also doe not chowg In a direction perpendicular to the siwfaW so that it is then
oestent over the whole field In which the flow and emeWg transfer ocow. in
consequence of this fact only the variation of density with tmetu has to be
accounted for In solutions of the boundary layer equatioas for teflat plate.- in
this connection It might be pointed out that the terminoloff whic h aa becomee
customary, nosely to call a boundary layer on a flat plate which develops under
hila flow velocities a 4compressible boundary layer' Is sot too well ahosen. So
established meaning of the term In VkW&ics Is a changp of volume In onsequense of a
chanme In pressure and such a chap does not occur In a boundary layer an a flas
plate since the pressure In the whole field of interest Is constant. Ibis expresses
 itself in the fact that the cheracter at the boundary layer equations is exactly
the same for mbsonic as in the supersonic range.

ismt rondg Solutions

The first solutions of the laminar boundary layer equations were obtained with
the assumption that the property values in these equations are constant (independent
of pressure and tet.prature). Under this condition the local skin friction factor
was found to be

O- 0.66 /(B

The local uasselts number was determined as

,U - 0.332)S
(Ba - 2)

when the heat transfer coefficient h on which the Nsselt number is based is
defined by the following equation

.W
- (Tr - 'w
(Ba - 3)
in which Tr is the temperature which the plate assume when it is only in convective
heat exchangs with the high speed flow but otherwise not cooled or heated (by
* radiation, or by heat conduction into the solid wall interior). This wall tempera-
ture is in high speed flow larger than the static temperature in the sti*= outside
the boundary layer because of the erodyna c heating effect within the buad
layer and its value has to be know if the heat flow on a surface unde hi speed
conditions Is to be calculated. The calculations which were mrne under the
assumptions of constant property values showed that the following temperature rasios

which is called temperature recovery factor is a function of only the Frandtl

namber of the fluid. 2his function is shown over a wide range of Prandtl uamber
in rieure 2. (Bit. 3 ). It can be approximated bV:

r.pW (,- 5)

in the ran. of Prandtl numbers between 0.5 and,5. The wall temperature Tr which
is deoribed to the temperature recovery factor will in this report be calfed
areovery temperature'.

9
 The heat transfer coefficient is often expressed in a dimensionless form by a
pakameter called Jtanton number:

t = h/ .Op.-T (Be - 6)

It can esily be seen that the following connection exists between the Nusselt
number and the Stanton Number:

KU/ .h(Be
Iu - 7)

A comparison of Equation Be - 1 for the skin friction factor and Equation Be - 2

for the local ausselt number results in the folloving:

St -0.5 of PV" (DR - 8)

This relation between skin friction and heat transfer is useful because it still
holds under certain conditions when the property values vary with temperature. This
will be discussed in the next paragraph.

leynold's analogy between skin friction and heat transfer can be expressed by
the following equation:

2 (8.- 9)

The beat transfer calculated from BeynoldIs analogy the tfore differs from the
correct relationship for laminar flow by the factor P1d3. Jbr games the Frandtl
umber is not too different frm 1 and therefore Reynolds analo gives a reasonable
first appmcimation to the real heat transfer value. The relationships which have
been deterained for constant property values can be expected to hold satisfactorily
for real fluids or gases as long as the temperature variation throuhout the
bouaday layer Is comparatively small. It has been found that this is the case as
long as the ratio of wall to static free stream temperature is not too lreg (sa
between 0.5 and 1.5). Property data for determining the Reaynolds and Prandtl
nambers nm be formd with temperatures Tw orT The advantag of the coast=&
property solutions is that they depend on a minL;m of parameters and that they
apply generally to any fluid.

In the supersonic range and for lari temperature differences the fact that
the property values actually depend on temperature has to be accounted for in a
solution of the laminar boundary laer equations. Ligure 3 shows that the viscosity
and heat conductivity vary considerably faster with temperature then the specific
heat and the Prandtl number. Consequently the next step was to obtain boundary
liyer solutions for variable viscosity and conductivity but constant Frandtl umber
and specific heat. The equation

O(BA-1)

10
 olo:t.. t .'; const:.at I-'.n~tl xnw,.bcr andc i-cii'ic hcLt, the heut conductivity
-'L.ric3 ruorotion~li; to thc. yie~di. 7acre'.ore the lutw of tecipTcrntux-e Le;,en( ence
:XLS to b. -c~ulibe
.1.-Ali onc roicrty on1j'. The follo.winL luw was usswaiec to
E.Prczi; Lhc .j(rtr ecjcenr'cncy o: t(vic. 3ity'

T are
Iand law the viscosity and temperature at some reference condition. The uF, Of
triis brings two advantages, first it simplifies the calculation r;rocedure a-i
secondily it smalces the result dependent on a 4haracteristic temperature ratio only,
for instance, on the ratio of wall tezaperatureto free stream static or total
teniperoture. This temperature ratio as well as the exponent 40 have to be added to
the paramters mentioned above for the constant property value solutions.
Lquation Be - 11 expresses the actual temperature variation of the vicosity
for gases end specifically for air reasonably veil in a temperature range which is
not too laree. However, it Is found that the nmrical value of the exponent 0b
depends on the temperature level. For very low temperatures 01is approximately 1
and it decreases with Increasing temperature asymptotically towards the value 0.5.
Consctuently, in different calculations different numerical values for this
rI-,rameter were ued. The following general results were obtained: (For a suAmary
or tiaese calculations see Reference I5). It was found that the expression Ba - 5
for the recovery factor still holds uider the assumptions of this paragraph. In
the ae way the relationship Be - 8 between skin friction and heat transfer is
still valid. On the other hand It was noted that the skin friction and the heat
transfer vary now with Each number, with the characteristic tesiperature ratio, with
the value of the parmter 4 , and the tenmperature level. Only the variation of the
friction coefficient has to be discussed because of the validity of equation Be -8
Specifically the variation of the skin friction coefficient with the mentioned
parar~sters depends on the wVe In which they are defined, from equatiojns B - 1 and
B - 3 which define the skin friction coefficient and the Beynolds nwaber It can be
seen that for variable property values a specific temiperature noet be chosen at
which these values are introduced into the right hand side of the equations.
Usually the freestresa static temperature is chosen as such a temperature. In
this case the skin friction coefficient Is found to decrease considerably with
increasing i-ach zimber, with increasing temperature ratio T,1/f 1 , and with decreasing
value of the paramter *6. Figure 4 taken from reference 1413 show the dependeacie
of the friction factor on Ruch number. The dashed curves and the calculations %W
Van Driest, and Young and Tnsan will be discussed later on. It Is Interesting to
note that Q.specific value I makes the dependence of friction factor on As and
T /T vanish. The relations Be - 1 to Be - 9 hold. therefore, not only for constant
proprties but also when viscosity and heat conductivity vary proportionally to the
absolute temperature. This fact adds considerably to the Importance of the constant
property solutions.
 If the property values in quations B - 1 and B - 3 for the skin friction
factor for 6A different from one are introduced at the wall temperature, then the
trend of the variation of the skin friction factor with Mch number is opposite to
the one mentioned above.

It may therefore be expected that property values introduced at a proper

reference temperature situated somewhere between the temperature extreme encountered
within dhe boundary layer will cause the variation of the friction factor with Nis
and T /T to vanish. Actually, Flubesin and Johnson (Ref. 49) have shown that it is
possitle'to find such a reference temperature T* and that it can be expressed with
good accuracy by the following equation taken from Rubesin's and Johnson's japers

I Ot2(l )") . 5g
(Ba - 12)

This expression for a reference temperature can also be transformed into the
following:

To a 0.5 (TW- To) * o.19 (Tr - ?9)

(Be - 13)
In this form the relation for the reference temperature appears especially
instructive, as can be seen for same special oases sketched in figure 51

Fbr low flow velocities the recovery well temperature T is practically equal
to the stream temperature To. The temperature profile in thd case has the shape
indicated in the upler diagram of the figure and quation Be - 13 shows that in
this case the reference temperature T* Is located approximiately hilfway between
wall temperature T. and stream temperature To. This agrees with the results of
other investigations on heat transfer at low velocities.

Flr high velocities but a small rate of heat transfer the wall temperature
Tw is appm iately equal to the recovery temperature Tr . The temperature profile
is for this case indicated in the center portion of the figure and 1gsation Ba-13
showns that the difference between reference temperature T* and stream temperature
Tsto now 77 of the difference between wall temperature * and stream temperature
T . Since the fluid layers near the wall will Influence the heat transfer to the
wall most strongly and since these l yers have now, In an averae, a higher
temperature than in the first case considered, the value for the reference
temperature is very credible.
As a third example let us sasum that the wall is cooled to the stream
temperature as indicated in the lower diagram of the figure. it will be seen
later on that in this case the difference between the maximum temperature T
within the boundary layer and the stream temmjrature is approximately 25X J the
difference between recovery temperature and stream temprature. Fbr the difference
between reference temperature and stream temperature the corresponding value is
19%. which Is again possible. It should be pointed out that Figure 4 presents
friction factors for the case where the wall is at recovery temperature. Jor a
cooled wall the friction factors are higher and f or a heated wall lower than the
values in the figure. Calculations of the friction factor with the reference
temperature as eiven y Rqsmtions Be - 12 and Ba - 13 properly accounts for these
variations as will be shown In the following chapter.

12
Solutions for Air with Actual Temperature Variation of All Properties

For very high velocities as are expected in future aeronautical applications

the teiq:erature variation thoroughout the boundary layer becomes so large, even in
the case where the well is cooled down to practically the free-stre.i static
temperature, that it is doubtful how valid the results of calculations are which are
based on the assumptions mentioned in the prccieding pararaph. Acco'dinlly, a
number of recently published palers consider the specific heat and the Prandtl
nun.ber as well as the heat conductivity and the viscosity to vary with tewperuture.
id1 of the calculations have been made for air and r.ore accurate relattionships
than a power law for the variation of the vicosity and the heat conlductivity with
temperature have been ussumed.

Stherland' s relationship:

(Ba - 14)
with the constant C =32P R. for air, or tabulated values are used to express the
viscosity. Tabulated values are used for the specific heat and lrandtl number.
The conductivity is again fixed by equation Be - 10. The main uifference in the
various sets of calculations is caused by the fact that different laws for the
Frandtl number dependence on temperature have been used. Klunker and . acLean
(Ref. 41) as well as Young and Janssen (Ref. .59) use Pandtl namber data as contained
in the Gas Tables und in a paper by frius and Boelter (Mef. 177). Van i2iest
(Ref. 36) on the other hand bases his calculations on Prandtl number values which
have been recaea~inded by the National Bureau of Standard (Ref. 156). Aare 3
indicates the Prandtl numbers as used in both sets of data. It .a be recognized
how lrg the differences in both sets of 1trandtl number data are. All of the
calculations had to extruolAte keandtl number values to larger temperatures.
This is indicated by the dashed part of the lines. The large differences in both
sets of data reflect the leak of accurate knowledge of heat conductivity and
Prandtl number for air at hijier temperatures and is caused by the fact that it is
extrertly difficult to rmasure heat conductivity values for gases at larger
temperatures.

The calculations mentioned above reflect the real conditionr as encountered by

an aircraft in bhe atmosphere better than the ones discussed in the previous
paragraph since they are based on mavre accurate property value datal however, they
have one disadvantage, numely, that flow ana heat transfer now appear to depend on
one additional parameter namply , on the temperature level in the air stre m. The
governing parameters ae now Reynold's number, I.sch nmber, krandtl number, a
churacteristic temperature ratio and a characteristic temperature, for instance wall
to stream temperature ratio, and stream temperature. The number of parmeters is
considerable and correspondingly the calculations become rather extended.

13
 A~ll the calculations show that the friction factor varies proportional to the
square root of the Reynold's number In the ownse w4 as in 3quation Be - 1. Instead
of the constant 0.64 in this equation, however, a function of Mah number, tempera-
ture ratio and temperature level appears.* The most extensive calculations of this
nature have been done ty Young and Janssen (Ref. .59). The authors calculated
friction factors form wide range of 1Jsich numbers (up to 30) for a large range of
wall to streari temperature ratios, (1.5 to 22) and for three values of the static
stream temperature (100,400,8000 B.). They also determined a reference temperature
such tht introduction of the property values at that temperature nukces the value
#Wequal to the incompressible value 0.664. Young and Janssen reccamended
quation Be - 12 for ik up to 5 and give the following relationships which
approximate the calculated reference temperatures well in a Nkah number range from
.5 to 10 S

fort mr t T* s o.0556 (1s?rt) 6 (Be - 15)

fort Tv Tr: ?*As - o.70. * j *) 0.9@ (Too't)

Since It Is inconvenient to use different equations for different Mch number

ranges and for the wall at recovery temperature and at a different value, it was
Invest igated whether the reference temperatures an be expressed In the whole Mach
umber ranLe by one relationship. The calculations presented in A&ppendxd~ I
resulted In the following equations for the reference temperature

g0o. -T
(To - 4) * 0 22 r -?o)( a s 16)
The equation ay also be written as follows:

(za 16a)

with r Indicating the recovery factor and V the ratio of specific heats.
Table Ia contains In the fowrt coum the friction parameters Of 4i taken
from, Rf. 59. In the reference the property values woz based on the streem
temperature. The fifth colum contains the se friction, parmmtor as calculated
with the reference temperature given In the last equation In the following ways
Xquation Be - 1 was assumed to hold when the properties are based on the
reference temperature

(o .6M.4
 1 2 3 4 5 6 7 8 9 10 11 12
Riun TOMR I. (Cie'). (cfre). To, TI, x ~ (p
Nr.
3c 99 4.23 0.690 0.685 -0.7 400 405 1.7 0.845 0.864 -2.2
4J 101 042 0.6.51 0.650 -0.2 800 793 -1.0 0.839 0.834 0.6
2d 99 8.17 0,620 0.619 -0.2 1182 11_5 -2.5 0.828 0.820 1.
5a 100 9.57 0.593 0.592 -0.2 1559 1506 -3.0 0.807 0.810 -U.4
id 10% 1.67 055.5 0.555 -0.0 2246 2134 -5.3 0.807 0.803 0,.5
6a 102 15.15 0.501 0.505 o.8 3565 3322 -7.0 0.798 0.797 0.1
7a 90 20.45 0.462 0.470 1.7 .5135 4870 -5.1 0.781 0.793 -1.5
30 )Ioo 0.22 0.674 0.664 -1.5 400 403
41 402 2.43 0.629 0.630 0.2 800 791 -2.2 0.840 0.827 1.6
2e 400 3.147 o.6o1 o.600 -0.2 1182 1l59 -3.0 0.828 0.815 1.6
5b 401 1.28 0.575 0.576 0.2 1556 1516 -3.5 0.813 0.809 0.6
i0 405 5.52 0.5.41 0542 0,2 2246 215 -5.1 0.806 0.803 0.4
64 401 7.49 0.499 0.49b -0.8 3565 3366 -6.3 0.796 0.797 -0.1
7b 388 9.5, o.458 0.56 -0.4 5135 4898 -4.8 0.780 0.792 -.5
80 398 12.44 0.419 0.414 -0.7 8160 7868 - .7 0.786
96 396 35.95 0.383 0.375 -2.1 126IL5
1.e 794 0.26 0.671 0.664 -1.0 800
2f 800 1.74 0.638 0.635 -0.5 1182 190 2.1 0.813 0.812 0.1
50 808 2.45 0.612 0.610 -0.3 1550 1558 1.1 0.816 0.806 1.2
if 797 3.52 0.576 0.574 -0.3 2246 2235 -1.0 0.788 0.803 -1.9
6f 810 4,95 0.533 0.527 -1.1 356.5 3510 -1.6 0.806 0.797 1.1
7f 792 6.45 0.490 0.490 0.0 35 5162 0.8 0.794 0.792 0.0
8. 805 8.54 0.48 0.443 -. 1 816 86135 -0.3 0.786
9f 765 11.33 0.1405 0.3" -1.5 126&5
Jima runa were not fully evaluated since the property values were not known
for them.

Table Ia. WVal at re*overy temperature.

 1 2 3 4 5 6 7 , 9 10 11 12

Run T T * f
,Tr. ORt ;x Ma of fp), (Cf ine)*a '2St~ (Pr*)~

14d 97 400 8.63 0.667 0.661 -0.9 0.911 0.795 -14 0.934 -17
15d 101 400 11.84 0.629 0.64 -0.8 0.862 0.774 -11 0.924 -19
20f 94 400 19.19 0.584 0.583 -0.2 0.804 0.754 -6 0.878 -16
lf 101 1182 8.93 0.613 0.614 -0.2 0.873 0.766 -14 0.916 -19
11f 102 1182 12.37 0.594 0.589 -0.8 0.814 0.755 -7 0.876 -16
16f 105 1182 18.74 0.553 0.552 -0.2 0.776 0.746 -4 0.835 -12
13f 96 2246 13.39 0.553 0.547 -1.1 0.853 0.742 -14 0.898 -20
18f 80 2246 22.40 -0.524 0.526 0. 0.799 0.742 - 8 0.842 -14

1o 402 1182 4.55 0.591 0.590 -0.2 0.722 0.758 5 0.758 0

lie 405 1182 6.31 0.572 0.572 0.0 0.704 0.753 8 0.751 1
166' 401 1182 9.70 0.538 0.534 -0.7 0.691 0.750 9 0.734 2
17g 404 1182 14.83 0.481 0.478 -0.6 0.686 0.733 7
22s 410 1182 18./ 0.450 0.446 -0.9 0.692 0.731 6
13. 398 2246 6,68 0.531 0.532 0.2 0.7" 0.750 6 0.7"4 1
18e 394 2246 10.20 0.503 0.504 0.2 0.688 0.740 7 0.720 2.6
19e 394 2246 15.50 0.461 0.458 -0.6 0.692
23c 416 2246 18.70 0.432 0.437 -1.1 0.695 0.730 6
10d 796 1182 3.29 0.621 0.617 -0.6 0.726 0.758 4 0.755 0
lid 805 1182 4.55 0.602 0.604 0.3 0.718 0.750 4 0.756 -1
16d 792 1182 6.99 0.568 0.558 -1.8 0.708 0.750 6 0.745 1
17h 812 1182 10.57 0.513 0.510 -0.6 0.702 0.735 5
22d 784 1182 13.43 0.476 0.477 0.2 0.677 0.731 8
ld 801 2246 4.77 0.574 0.559 -2.6 0.725 0.750 3 0.755 -1
18d S0 2246 7.22 0,536 0.52 -2.2 0.672
19C 800 2246 10.96 0.488 0.48 0.0 0.694 0.732 5
23. 792 2246 13.63 0.462 0.464 0.4 0.685 0.730 6
So.e runs have not be.m fully evaluat-d, sinep the property values were not
known for thom.

Tablo lb. Cooled wall.

16
The friction parameter based on stream conditions is then

(Ba - 18)
dith the known values of the stream temperature and the reference temperature
Xquation Ba-18 can be used to determine the friction parameter as based on stream
conditions. This parameter is compared in Tables Is and b with the result of the
calculations by Young and Janssen, and the error in per cent is entered into the
sixth or seventh colwm. It is seen that the maximum difference between the friction
factors calculated by both procedures exceeds only in three cases slightly the value
2%. Generally the error is probably within the accuracy of the boundary lear
solutions. Figure 4 contains as dashed lines also friction parameters (C
which have been calculated with incompressible relationship Ba - 1 and the reference
temperature as given by Equation Be - 16 with the same procedure as used in
connection with Table 1. It can be recognized that the agreement between the exact
calculation and the simple calculation procedure using the reference temperature is
excellent regardless of the specific law for the viscosity variation.

The papers which consider the temperature variation of all property values
report on recovery factors which can not be expressed by the simple Ruation Ba - .5,
References 41 and 59 use the following definition for the temperature recovery facton

(Be - 19)
ligure 6 shows these recovery factors and indicates a considerable variation with
1ach namber and stream temperature. This large variation of the recovery factor is
surprising and it may be supposed to be caused by the fact that according to the
definition Equation Ba - 19 the difference between recovery temperature T and
static temperature T Is compared with the difference between total tempJature and
stream temyerature wiich a fluid would have if it had a constant specific heat equal
to its real value at stream condition. Actually, the specific beat varies through-
out the boundary lsyer and an average value should be used to deteraine the
difference between total and static temperature. It ma be calculated how well this
avere4p value for the specific heat is approximated by the specific heat which
corresponds to the reference temperature given in Squation Sa - 16. Therefore the
recovery values in igAre 3 were recalculated (Aef. 23) using the following
definition for the recovery factor

17 ( -20)

17
 Ta
6 4000I.

TV ma (of OR (of t) 0$ ($!h

no 4.55 0.91 O.9" o.311 -o.SS
6.31 0.572 0.738 .319 -0.525
9.70 o.530 o.684 0.325 -o.417
14.83 o.4& o.752 0.313 -o.321
2246 6.68 o.531 1.460 o.312 -1.365
1.20 0.90 1.065 00319 -o.660
15.50 0.461 0.967 o.305 -o.4g0

174
Owe*e recovery factors are also plotted in Figure 6 and It oan be seen that most of
the variation of the recovery factor with Moch umber and streom temperature has
disappeared. The remaining variation is well expressed by lquation Be - .5 when the
Prandtl number is introduced at the reference temperature described by Nquation
Do - 16. This can be recognized from table Is In which the recovery temeo tr as
calculated by Young and Jansen is contained in colum 7 and the recovery temperature
determined with Equations Do - .5 and Do - 16 In colum 8. The difference in per cent
Is entered in column 9. It exceeds only in a few cases 5 per cent.
Crocco already pointed out (Ref. 34) that the relationships obtained for a gas
with constant specific heat remain materially unchanged for a gas with variable
specific heat when the product c dT is replaced by the differential of the enthalyw
i. In newer papers (Ref. W4 anl 36) this fact is utilized.
Van Driest bases his recovery factors on enthalpy instead of temperature and
uses the following definitions for the enthalwV recovery factor

(Re - 21)
It can be observed that this definition is very similar to the ohe given in the
above Iquation Us - 20. ,iguvr 7 contains as solid lines the enthalIV recovery
factors as calculated in Reference 36 for two different conditions in the stream
outside the boundary layer. The upper graph of the figure presents the recovery
factor for a stream with a static temperature of 40003. the lower diapm gives
the recovery factor for a total temperature In the strem equal 1000,. A static
temperature of 4000R. was chosen to approximate conditions as encountered by an
aircraft in flight through the atmosphere whereas a total temperature of l000F. Is
supposed to represent conditions as found in wind tunnels. A comparison of the
recovery factors in this figure taken from Ref. .59 with the enthlV recovery
factors In Table la disclos a marked difference especially at highe ftsh numbers.
This Is due to the different temperature variation of haendtl umber on which the
calculation by Young and Uanseen on one side, and by Von Driest on the other side Is
based. It was pointed out before. that the lower Prandtl amber values were used
in Reaf. .59. whereas Van Wriest adopted the set of K4ie values in Figure 3. Also
Inserted In the figure as dashed lines are recovery factors which have been
calculated using Equation Do - 5 into which the Prandtl uamber was introduced at a
reference temperature as given by Equation Re - 16. It ma be obeerved that the
agreemnt between the exact calculation and the calculation from Equation Re - 5 is
very satisfactory. The maximm deviation between both sets of calculations Is loes
than 2 per cent. From the data published in Reerence Do - 28 the onthalff recovery
factors have been calculate~d Inserted in Table Is as colum 10. They cean be
compared With the values MPr 1 in column 11. 2Me difference In per cent is entered
Into column 12. It may be obeerved that the difference is only in one case larger
than 2 per cent. It is smaller than the error in colam,6 for the temperature
recovery factors. Use of the enthellwjecovery factor Is therefore, preferable. It
is also remarkable that the value OW1iw appboxismtes the actual recovery factor so
well regardless of the Praodtl umber variation on which calculations have been
baned.

1s
 Hoat transfer cooeffioints have been calculated by Young and Uansen for the
sam rang. of prameters as the friotion factors. One result of this as of all
other calculation procoedures is that the Maslt number as a dimnsionless parsmter
for the heat transfer coefiloent increases proportional to the square root of the
Reynolds number. AccpQM y, Youn and Jansuen (Rbf. 59) presented their data as
the parameter W qW In i/ which all the ]roperty values have been Introduced
atom strom temperatures assuming uhat the fbaselt number is also proportional to

Since It has been shown that the friction factor can be expresed very
accurately by a simple fozrula, It appears advantapous to calculate the heat
transfer coefficient with the help of the friction factor. 4Ierefore the ratio
af/2 St which according to Nquatlon Ba - 8 has the value PrI for a fluid with
constant property values, was detezined fzL Young and Janssen's data. With
property values based on strem temperature tae following equation holds

~28A A--.I1
(Pa'j
(Ra - 22)

The paramters on the right side of the equation are tabulated in Deference 59.
Now It will be checked how well Iuation Be - 8 approximates the values calculated
ty Young and JZeasen when all properties in this equation we introduced at the
reference temperature described by lquation Ra - 16. With Uquations B - 3 and
B - 6 the ratio friction factor to Stanton mmber cas be written

%( - 23)

The only property value in this equation is the specific heat c . Therefore the
corresponding ratio based on properties at reference temperaturg is

(1a - 24)
This pwmster as calculatod from Young and Jmassea's data Is cotainsed In colum 8
of Table lb. Tho values PrES3 we coatainhd in column 9 and the pwroentao
dirferense in column 10. It can be observed that the egeement for the heat
transfer parmter is not as good as for the friction factors and thestuerature
recovery factors. Only for the two strem tempraturea 4000R. and :o. is it
within 10 per cent. Calculations 1W Xlumr and INedan (f. I1 and 42) are less
extensive theam th ose In Reference 59 but are Ssnerall in good eoomat with the
latter. Therefore a comparison with the imcompressible relatioship was not made
for thes calculations.

19
 Van Dr'iest* has also calculated beat transfer coefficients with the Pramdtl
amber variation as proposed by the National huam of Standards. 1h used In the
pr'esentation of his results a Stanton number which Is based on enthalpies I instead
of temperatures analogouslyr as in the definit ion of the recovery factor. Instead
of the custmary definition of the Stanton number

(Ba - 2)
he defines It in the follwing way

46
S61
2i (o1O- 1W)
(Ba - 25a)
ths Paper contains plots of the parametor cf/ 2 St I whit fo ns1tant tIMA
number has according to IXauation Be - 8 the value (0) . te 8 shos thi
parmter plotted over the lJah umber for two diffeeaft ecaditims In this free
atrem. Soe upper diagram holds for a fee sftrea static temperature5 Wf OO3. mad
the lover diepen holds ficr a total free strain te-eatr or 10001. a 54W3.
Compred with these boundary lWer solaaticas we *aan the "Its of as rOPPEnte
proedue
Aobset th pem i- 2 8 equal to Prh mis~romese the frmtl
amboer at Soi etsa t PWUUe C si;6 IF kte ln -16 D uq be
obserwed that the epmemt betwoo bot sets of ee2.euLatoms Is pAs smallest.
sodeviation is In the whole ramp Investigated Iss them I per est.- 3 viow ot
tugs goo ap'emnt the prmter eg*2 at1 was als o eleulst figs Tomsaid
Jmsemi's deta. It Is contained Is ;;IQ ' ot able lb. OoUM &GLVe spain t"e
per Gent diftremoe between the actual Talamo In oo0in 7 ind go aPONS0USe 1
2patioa Ba - 8 as gonaad In sb= 3. SO em be Ame1WeO theaO is
poer tbr a owow -feae'- 2 MW". " 80002. A Itee--- urh Of 10003.
OW eew In wind teint Seatst is MMv emOued is Noe ==w. JI
Iseiga ealeulaticas. therefore, the ase of the Stanton Waer boei an eatha1W Is
peterable. ftis Stenton iaber em he detesuized fras *&"s ]k - S with mn
error Uoss then 2.6 par seat. Whe ealculations of seewewV testote or hoat
transfer oceffiestsu we based as eaftalm. it is prbiv a smieft to
deteulm the retisreae ecedition at shish the Newly valm -N46" s ifve~m
a ratarSFOe eat6hMAl inWSteatof thefSGSe t M era- - (1010ie 28 - 16).-&
this reese the ealeuLaat pmma as 4 seribe& la fteaUxi I wa epeate en
the basis ot enthalpiss inted of temperature. It e 9*m that wit Nis
sousay obtainablis the equation

(200~
vhich has the am numerical constants as lquation B - 16, deteruinss the referense
conditions which, when used for the p opertles in the equation for the friction
factor, gives best areement with exact boundary layer solutions. he reference
enthalp may also be exressed In a different way as

1* . .5(is #i..+ .22 h Lt1 (h)2is

(m - 2 6 a)
A comparison ot asured temperature recovery factors and heat transfer
coefficients vith calculated values has been mde In a recent paper tq G. Aer
(Obt. 1.55). Zber shows that masured temperature recovery factors In a law
boundary layer on cones vary around the value 0.85 with a mauims deviation of
I per cent. Within this limit the values age also with the calculated results
which have been presented In this paper.
In the FIgure 4 of his paper, Sher coampred mmure reoovery factors with
values which have been calculated bV Slunker end 16Lasn and points out that for
larger Moh mmbers a considerable difference exists. It has to be kept In mind,
however, that the calculated data b1 l Lunhmr and blean have beem obtained for
tree-strem conditions which were different frm the ones for which the mam e-
mrnts have been msde. 2he maured recovey tastoers have to be oompsred with the
values in Fisare 7 and exibit then en apemat which sain Is within I par sot.
The peculiar fast that temperature recovery factors measured on a flat plate wer
consistently bi4er than the owe deterumed tram come md seamnd values mund
0.88 has been explained t' Srobk:n (Asf. 43) as sawed 1 heat oommetisa In the
solid plate aterial. Heat transfer ooefficle nts ould omly be wared with an
accuracy of atrcinmately 10 per cent and within this liit &Smia apremat
between valos masured on cones and calculated heat transfer coefficiets have
been observed.

Plecoemded Calculation hvoedare

On the basis of the investigation eds In this chapter the olloving pWeeem
is recomended for the calculation of the friction and beat transfer of flat plaste
in supersonic lamnar flows
he wall shearing stress is calculated from the following equation

2 (Jk - 27)
with the friction factor appearing In this equation given t the relationship

f a 0.66./V/ 71 a 0.6 ix
A (Ba - 28)

21
Tbe Woperty values are introduced into both of the above equations at a reference

temperature which my be tentatively calculated from the following relationship

T= =T,+o.5 (Tw - T) +o.22 (Tr - Ts ) (Be - 29)

This pWocedure was shown to give agreement with exact boundary larer solutions
within :t 2 per cent.

Heat transfer coefficients are defined kV the following equations&

%. h (Tr - Tw ) (Ba - 30)

for constant specific beat, ors

q W a hi (I "-i W)
a -3 )

for variable specific heat.

The recovery temprature Is detezuined tw the relation

Tr T +r--?A-
2 cp
(h - 32)

for constant specific heat and bV

ir Is + r "

2 Ok - 33)

for variable specific beat.

le recovery factor appearing in both of the above equations is oealouLaed

from the equation

r (20,- 30
into which the Prandtl mber has again to be Lntrodue4d at the refesaee earwatuf
as given ir Iquation ha - 24. Alternatively the Pramdtl umber to be Introduced in"

22
Equation Bh - 3& can also be determined from the reference enthalW

*a o.5 (i.+ iw) +~o.22 ri tI (Ma) is

2
(BA- 35)
iBscovory factors celculatc A in this way agree within t 2 per cent vith the results
of exact boundary layer solutions.
Mhe beat transfer coefficient appearing In Equation So - 30 and Be - 31
calculated from the Stanton number value

St z
9 OPVsO (- 36)
for constant specific heat or from

h
9 O(h- 37)
for variable specific heat. The Stanton numbers are obtainable from the relaticashil

St = 0~ (f*)%

Into which the Frandtl numibers have again to be Introduced at the reference
temperature given by Equation Be - 28 or the reference enthallW Equation Ba - 35.
Heat transfer coefficients calculated In this way apree within 2.6 per cent with
the results of exact boundary layer solutions, for free stra em peaue above
30001t.
2Me main unertainty In the calculation of recovMr factors and heat mtraser
coefficients Is caused by our insufficient present da kawmled.e of Preadtti
mauhers for air at hider temperatures. Prandtl umber va&Ues as given In the
Oas Tbles and reoumaded by the National Aares of fandwds differ for isetaos
at a ioeaperature of 20000R. %y12 per cent ad accordingly the recovery factors
calculated with thee Prandtl nubers differ by 6 per cent sai heat transfer
coefficients by 8 per cent.

23
Ttl~BMIZWT BDOU'ARY LAYIM FLOW
In the preceeding chapter it has been demonstrated that the calculation of
friction and heat transfer in lmdnar boundary layer flow has been refined to an
accuracy which Is very difficult to parallel in experiental investigations. In
turbulent boundary layers the situation io coouletely different. The present day
understanding of turbulent exchange of nmentum and enerff especially under
conditions of widely varying temperature In restricted to such an extent that exact
solutions of the boundary layer equations can not be obtained. In rvV calculation
procedure assumptions have to be Introduced on the exchange mechanims and these
assumptions have to be continuously checked with experiments. In obtaining friction
and heat transfer data we have, therefore, to rely heavily on masurements.
Celculat ions are necessary and useful in order? to extend the reape of parameters
outside the experimentally investigated one. However, it will be necsseary to
select from the different calculations procedures published in the literature those
which egree well with experimental data In the Investigated rapg. Comparatively
extensive experimental Investigations are available for the friction factor of
plates whose temperatures are equal to the recovery temiperatures.
In Figure 9 (Hof. 65) the information on the Influence of b~ch umber on this
friction factor is collected. The measurements, the results of which are presented
In the figures* have been made on flat plates or cylinders. Plotted Is the ratio of
the friction factor In compressible flow to the friction factor In incomressible
(meame constant property valaae) flow. Thebfriction factor is %pinbased onk the
property values taken at strei temperature. It may be oberve that It decreases
considerably with Increasing Mach aumber. The Influence of Reynolds number on the
friction factor has been found to be the mem for supersonic as for low velocity
flow. Therefore Figure 9 applies to aay value of the Fleynolds umber. Friction
factors which have been msred on come V Bradfield (3sof. 18) hare values which
we appoximately 20 per cent hghr then the results of the meammeiments collected
In Viwce 9. The reason for this discrepans is not explained until now.
'Leav 10 (3sf. 65) presents the predictions which different published theories
make on the change of friction factors with Nock amber. It ma be observe that
the varnous preditions differ to a very hI* degree. Soe best ageement to the
experimeintal results presented In Figure 9 give calculations mate IF ilson,
3bOM, GOPe, AMMUkl.oiahel, and Tooker. ht*eve with ths emwlse It io
diffieult to amae predictions on the friction factor at hier Ifs ,erI becase
the different theoris deviate considerably at Mobh imumbers above five.
Faced with this situation it Is mpo~prate to look for a simple calculation
preewe which em with the exparimental results In the investigate rmpg.
YOU"n at Jamieson (sf. 59) Pointed out that the I" of constat rOPeWV relaticas
with the refaves temperature established for lon.aar boundary layers gives Mood
agroet for turbulent boundary lovers as well. Jbr ths Ni'oeduoe the relaim-.
ship for inda friction in a coastat; properly fld flow mst be establiahed at
first.* Severl equations are offered in the literature for this purpose.
TheAolus equation

Lfz o.0296 R7
2
(5b - 1)
gives good ageement with masured values up to Pl. a 107 . At larger erynolds
numbere Equation Bb - I gives friction factors which we too low.

A good representation of the measured average frict ion factors in the whole
rang up to He a 109 is obtained by the Karan-Schoenherr equation

r .;- logi ° (R. )

rhe local friction factor msy be calculated from the averag value o the
relation

of = 0.557 af
o.557 + 2 r

(ab - 3)
Simpler to apply Is a forula bV Sahls-Or~us for the local friction factor

cf o.370
(lOglo ae) (B -

and a relation by Prandtl-Schli hting for the avere. friction factor

- o.455
Of go Re) °M
Clo10

(lb - 5)
&oth of these relations represent masrements up to 11e - 10 9 very well.

SIT the asusmennt. plotted In Figure 9 have bon ad at ftnolds rumbea

below lo. th3 Masius Equation (lb - 1) will be used to establish the eoh mmber
dependence of the friction factor with the use of the reference tem~ ature . as
friction factors In Figure 9 we based on property values at strems teerature

(Mb - 6)

25
and aoared with the friction factor ofi of an incompressible (constant property)
fluid at stream temperature

f = o.0296
2
(. V8 x/ F),
(B - 7)
Assuming that for a variable property (compressible) fluid the shearing atream is
given Wy the constant property value relationship when the property vlues ae
introduced at the reference temperature T* remlts in the equation

(* V2
(g
VX /)O 2 (Bb - 8)
A omarison between the Iquationa lb - 6, Bb - 7, and Bb - 8 gives the following
forula by vhich the ratio of the friction factor in supersonic flow to the friction
factor for a fluid with constant property value can be calculated.

C: (j,,.o. 2 (9,)..

(B - 9)

%he remlts of such a calculation in which it was assumed that the viscosity varies
proprotional to the power of 0.76 of the absolute temperature Is indloated %V a
daubod line in Fiegre 9. It mir be observed that this dashd line aveps the
maWssud values verY vell. It ale arees at h4er Mash numbers with the oalmela-
tionm IV Vil=on, Mbn*ghm, and Tucker as presented in FiCure 10. It can therefore
be reocomended for a plamible extrapolation of the test results into the hg1bar
Mach number range. A pap-r (by . H. Abbot - Ref. 181) presented to the fourth
meting of AGMiD wind tunnel and ardi testing panel contains the remalts of test
on thin-valled tubes in a firing ramp. Two series of toot points vere obtained,
one at Msh number 3.9 and for the vail cooled to strem static temerature, the
ootheM
r at9. t oh numbr 7.25 and 1.8 T . rPresents
dah ad ine in this foi
bJt uoutheofratio
tosto'ao
poatsasweealoeuated
saova ia
with Foations lb - 9 and Do - 16. te ageement is especial3 -tor the meier
Mmh=ro for wlh according to Abbott the tets r produoed mat$a1astietiasl3,

Som friction ocefficlents on a cooled vawl we also reportod In Reference 180.

2Wewoe ieasmed at Mah numbur beWN 5 and 7.7. tis values o/Of for theme
tests we all lower them even the ouwve fb the val at rooem tepmpeure shwn
in PFge 9. It is believed that this is caused 1W th feet that the meuarmmts
vose m oa a vind tumol vell along which the preaure wes not content but
decreased in flow dirotion.

26
 Reference 73 contains the results of a very extensive investigation vhich has
been ade at the NAA to study friction factors and heat transfer coefficients for
air flowing tuulLently through a round tube with larp temperature differences and
subsonic velocities. Me essential result of these tests wag that the normal constant
property value relationship described the reults of these tests satisfactorily hean
the property values vere introdused at a reference temperature which was situated
half wr between the wall teoerature aad the air tsperature. It may be observed
that equation Ba - 16 contains the same relation for the reference temperature for
the case that the air velocity is mall (recovery temperature equal to wall
temperature). Therefore, It uq be justified to reoammad use of constant property
value relationshps together vith equation Be - 16 m reference temperature for
turbulent flow along a flat plate.
I plausible deductions Ackerasnn (Ref. 62) and Squire (Ref. 85) determined
the following relationship for the recovery factor In a turbulent bouadary lWer

(Mb - 10)
Test results on flat plate. as well as on come agree very von with this
prediction so that it Is toda in common usage to describe the temperature recovery
in a turbulent boundaz7 lqer on a flat plate. The test result& showed no conclusive
variations of the recovery factor with Reynolds amber or with Mach amber. A
slight decrease of the recovery factor with Mach number which has been predicted kV
an extension of Squire's calculation into the compessible flow rangs (Rsf. 87) has
tip to now not bee confirmad by 'xo-riments, probably, bocause this offect is too
small in thp rangf of Mach numbers investigated till now. It has to be rompmbered
that all experimental information has been obtained at a total tomporature of the
free stream around 1000F.
the simplest vq to calculate heat transfer in turbulent flow is again to
derive it from the friction factor with an equation

St (b-u
atS a f (Bb- 11)
analogous to relationship Be - 8 in which the function S has to be deterained
experimet ally.
Ncr amel temperature differences, and low velocities Golburn (Rof. 66) has
found that thSs factor S can be exWessed for turbulent flow as well as for laminar
flow as 1/fr V3 . Accurate test results for which this expression could be
checked In high velocity flow vith large teperature d fferences are not known to
the author.

27
 hubesin (Ref.- 82) obtained In a calculation In which he had again to make
plausible assumptions on the turbulent exchazge Mcham the result that the
factor S agrees for high velocity high t.5lpratur. difference flow for a Reynolds
uamber below i66 with the expression 1ffrw and that It decrease& slightly with
Reynolds amber so that for a Reynolds number of 107 it "as for air the numerical
value 1.20 and for a Reynolds umber equal 109 the value 1.18. The Influence of
Mach number and of the vall to stream temperature ratio was in this calculation
found to be extremely small.
Van Driest (Ref. 36) makes In his well known calculation of heat transfer in
the turbulent high velocity bounday layer the eaumption that the Prandtl number
and the ratio of turbulent diffuslvities for mofntm and heat are both equal to
one. for such a fluid Reynolds analog Is exactly fulfilled. This means that for
3k M 4 the friction factor of a wall vith the tempeature equal to the recovery
temperature as calculated ty Von Driest Is approximately 20 per cent larger then the
cm meaued experimentally. It Is therefore undlerstandable that the heat transfer
cefficient calculated by Van Diest smes well with experiments and vith the
predictions of Rubesin.
ALl the discussion In this chapter assumed that the plate surface Is smoth.
It In known that roudmnes increases friction and heat transfer In turbulent flow
when the roughness exceeds certain limits. ktensive systemtic Information on
rough surfaces exists only for the friction coefficient in low velocity flow. A
good survey over this field is for Instance contained in Reference 27.
Stuasing our discussion on turbulent boundary laers the following
recamendatlon is made for a calculation procedures Oonotant property relationships
should be used to calculate the friction factor, the recovery factor, and the heat
transfer coefficient, namely Squatione lb - 1 to Bb - 5 for the friction factor,
EquatIon Bb - 10 for the recovery factor, and 3quation Bb - 11 for the heat transfer
coefficient. The factor S in the latter lquatiog varies for air between 1.25 and
1.18 In the Reynolds number range from 10.7 to IVw. Property values should be
Introduced Into these equations at the reference temperature given by equation
as - 16 and the recovery factor as well as the Stanton number describing the heat
transfer should be based on enthalpies Instead of temperatures vhen the temperature
range Is such that the specific heat varies considerably within the bounday layer.
The recomendations have to be considered as tentative ad have to be ahecked. ty
more experiments at large Ikch numbers and especially at large temperature
differences and high temperatures.

28
 SETION IV

INFlENCE OF PRESSURE AND TEMPEAAIE VARIATION ALDW DM

SURFACE ON HEAT TRANSIFE

VARIABL PR%!SSTRF,

Up to now the pressure was assumed to be constant along the surface of the body
to which heat is transferred. Actually, the pressure usually varies along the body
an veil as along the airfoils of aircraft and it Is important to know the influence
of such a pressure variation on heat transfer.

Exact solutions of the boundary laer equations for laminar flaw and for
fluids with constant properties exist for certain types of velocity variation
especially for a velocity outside the boundary lqer which increases proportional
to some power of the distance from the leading edge.

Approximate procedures which use the integrated momentum and energy equation
have also been worked out for the calul ation of heat transfer on surfaces along
which the pressure varies in some arbitrary way. Information obtainable from
calculations is much more restricted for high velocity flaw in fluids with variable
properties. Only in a very few cases solutions of the boundary lqer equations have
been obtained and even procedures which use the integrated boundary iaer equations
become rather tedious.

Generally it is known that a pressure variation influences heat transfer

coefficients to a lower degree than the skin friction parmeters. This Is caued
by the fact that the pessuare gradient enters the engff equation only in am
indirect way. In addition to this fact, the shapes, used for aircraft bodies and
airfoils designed to fly at high velocities, have to be mne very streamlined in
order to reduce the drag. On such configurations which for supersonic flow have
also sharp noses or leading edges the pressure variation Is comparatively maUll and
therefore the effect on heat transfer is of minor im oc. This finding is
subetantiated tV experiments contained in Reference 6I in which avere beat transfer
coefficients were measured on a cylindrical body with a conical nose. Mw pressure
along the cylinder could be varied by an adjustment of the wind tunne walls. In
addition to a constant pressure along the surface, pressure variations ware
investigated corresponding to Nwh number variations aong the surface as shon in
Figure 11. It Is mntiowd in the referenced report that uah a variatio
corresponds apoximtely to the pressure variation as found arowd a 3 per cent
thickness becouvex ar-foil or to the pres su variation around a peaolie
body of revolution with a length to diamter ratio equal to U. It was found that
for pressure variations a indicated n Fipue 11 the avers.. heat transfer
coefficient varied only to a degree which was oaller then the aoosW with which
heat transfer coefficients could be measued.

fte emn conclusion was reached in Reference 26 where local heat transter
values and recovery factors have been masured on cones with straight ad parabolic
generatrices. It can therefore be expected that for bodies with Mal angle noses
or loading edges and for flow under mall angle of attack th constant pressre
relations give heat transfer data of uff icient accuracy for desigan puposes. With

29
respect to cooling a rounded nose would be advantageous for missiles since in this
vy the large heat transfer coefficients near leading edges are avoided. Heat
transfer on such a rounded nose vil decidedly be different from that on a flat
plate. An extensive treatment of the Influence of pressure variation on heat
transfer to bodies of arbitrary shape would exceed the scope of this paper. Jor
this reason no further discussion on the Influence of pressure variations vil be
made and the reader is referred to the literature presented in the appendix.

One remark mny be added, namely that for flow with pressure Uadients the
correlation of the actual friction and heat transfer parameters with the constant
property relations by specification of a reference temperature is not possible. This
can be seon from Rfer-nce 99 where calculations are made under the assumption
that the product of density and viscosity is independent of temperature. Friction
factor and "Tusselt number art, found to vary with thp ratio wall to stream tenp-ra-
ture. wo such variation woull bp in-icatel by t!- constant property relation for
any choic- of the ref-re-cg tqrp-ratir.

VARIABL' WALL TIP URATW

That a temperature variation along the surface of a heated object is a

parameter of Importance for the heat trasfer coefficient has been only recently
generally recognized. Sage and Falimer (Ref. 115) presented Por lasnar flow an
approximate treatment of this problem am early as in 1931 but It was only after
150 that intensive work was started to investigate the dependence of heat transfer
on this parmter. All calculations on beat transfer at variable vail temperature
are based on the fact that for constant property values the energ equation vhich
describes the heat fat thro a boundary layer into a vail Is linear and that
therefore general solutions of this equation can be obtained br a superposition of
individual special solutions. 7hi linearity of the eneg equation is in laminar
flow essentially still preserved when the property values vary in such a war that
the product of density times viscosity Is constant. Under this condition and for a
specific type or temperature variation along the suface

TV Tr ' MC i Cl(Ob -1)

Chapman and lbesin (rf. 113) obtained exact solutions of the l-minr boundary
layer equations describing the flow and heat transfer to a flat plate. Fleare 12
contains heat transfer coefficients based on the local temerature difterenoe
- , as calculated in thi refereana. The parmter M/41 is plotted over the
-- -L which determines the temperature veriation along the mrfase. It MW be
observed that with increasing value of 0 a considerable increase in the local heat
transfer coefficient as determined tv the pearstor 114140 occurs. Men the
temperature difference between the wall aM recovery temperature, vhich determines
the heat transfer. increases linearly with distance from the leading edge the heat
transfer coefficient will be bw 65 per cent larger than the heat transfer coeffieient
under corresponding conditions bet for a constant wall temperature. For a tempeatWwe
diftrence which increases proportional to the square of the distance from the
loading edge, the heat transfer coefficient is larger by a factor of 2 thea the hest
transfer coefficient for constant wall teMaperature.

)0
 Conditions similar to the ones just discussed occur near the nose or loading
edge of a body or airfoil of an aircraft In high speed flight when the aircraft
suface is cooled either bW radiation or by heat storage In the solid material. In
this ase the vail surface temperature is determined by a balance between the beat
transferred from the air to the surface br convection and the heat removed trami the
sutface tr radiation or b' heat storae in the Interior. It io knon that the heat
transfer coefficient on the leading edge of objects is very large (theoretically on
the flat plate It Is equal to Infinity) and that It decreases with increasing
distance from the leading edge Corresponding]y, the temperature of the object near
the leading edg is practically equal to the recovery temperature and It drope with
Increasing distance becuse the heat removal through radiation becomes mre and
moy effective an comp ad with the convective heat transfer. Temperature distri-
butions in wings of dimond shaped as calculated bv KW* (jbf. 116) show this
characteristic behavior.* It has to be expected, tht the heat transfer coefficients
In a close region downtrem of the loading edge of such objects will be conciderably
larger than the heat transfer coefficients calculated under the assumption of a
constant vall temperature. Correspondingly the surface temperatures will be higher.
It is possible to determine heat transfer for lemner flow conditions and. for
an arbitrary variation of the wall tmperature ty superposition of individual
solutions which have been obtained for r temperature variation according to Xquation,
Ob - 1 bw expressing the temperature variation along the surface as a series:

T. - Tr . AP
(Cb - 2)

Teis method which has been use" kW different athors has the disadvantage that the
partial solutions are knon only for lanar flow conditions and for a limited
uniber of terms In fquation Gb - 2. It Is often not possible to express a tempera-
ture variation occuring on an object bF a power series with this limited umber of
term satisfactorily. For this reason anther method will be described here which
can be applied to aW' temperature variation and to turbulent as well as lamiar flow.
It Is mainly based on work by Bad (Mat. 110) and Wool&n (Rat. M2). ~Ibs method
starts from a solution of the boundary 1lqer equations for a step-ise varying wall
temperature and replaces an arbitrary wall temperature variation IV an ifite
aber of infinitely smasl steps. An Integration poodare of the known solution
tor one step then gives the answer for the arbitrary varying wall temperature. this
derivation obtains the following equation:

q(x) sfh IX) ATd (S) dt + h(S:

1 x) [Tw ) - W P

(Gb - 3)
for 4he heat flow at the distance X from the leading edg Into a sufac. In this
equation h is the heat transfer coefficient which Is found at the location X whem
the temperature over the surface Is 9t first equal to the recovey temeratur'e e
then ump sumddenly at the location I to a certain value which afterwards is again
constant. Tis the temperature variation actually found along the surface. The
intepastion V. to be performed disregarding discontimaities In the wall temperature

31
Tw . The sumation expressed bt the second term on the right hand side of the
equation has to be considered when the w I temperature Tw varies discontinously
along the surface. In this case Tw ( S ) is the vall temperature Immediately
downstre, of a sudden temperature Jump and Tw ( AT ) is thecorresponding
temperature -i-edlately upstrem of the Jump. If the tenperature of the surface
differs at the loading edge from the recovery temperatuo then this summation has
also to be applied to the loeading edge vwith a tmperature T ( d ) oqual to the
reoovery temperature. The procedure will be explained once more for a specifio
temperature variation later on.

The following expressions for the heat transfer coeffiioents vhich are needed
in the above equation have been found by approximate calculations for the conditions
on a flat plate. For a laminar boundary layer (Ref. 121)

h( ~
x2) 2 o, 33k4RVp-pL - ctY
X
(Gb - 4)
for a turbulent boundary layer (Rof. 122)

h(",x) a o.0289 k Re*,* pr [ i

-1

X xx
(Ob - 5)
These equations have been verified by experiments. Since the calculation procedure
described In this chapter Is baed on more simplifying aumptions than the results
discussed for a constant val temeorature, it seems dvant ous to deterine bF the
mthod of thAs chapter only a oarrection factor vhich ha to be aplied to the
constant wall temperature heat transfer coefficients in order to take Into account
the influence of a vail temperature variation. This calculation vil be done in
the folloving lines.

Neglecting the Influence of the variable wall temrature on heat trasfor bV

using the heat transfer coefficient for a constant wall temperature the heat flow
at the location z is described by the folloving equation

q0(x) z h(Ox) [T x) - T] (Cb - 6)

rm this and the previous Vqmtion Ob - 3 ye obtain the relation

q(x) [h(d) - TW(i)

q(x) h(O 'x)IjTW W - Tr)
(Cb - 7)

32
for the ratio of the actual heat flow q(z) to the heat flow %,(z)calculated under
the simplifying assumption. For a lamnar boundury layer 03 a flat plate , this
equation can be transformed by use of 3quation, Cb -iinto

~ I[tcf"
dTx +Ei~' IT () T
d(xT (X)- r (Ob - 8)

and tor a turbulent boundary layer on a flat plate into the relationship

q(x) a 1 .1.ALj TrIf TAI

(Ob - 9)
Bach of the above two equations can be solved as soon as fte wall temperature T1 as
a function of the distance from the leading ig is ;reesribed, for Instance by
Leiwial or graphical integration of the Integrals apearing in the equations.
let us assaw, for Instance,* that the wall temperature as indicated 1n the
upper diagrm of Figure 13 Is prescribed. this temperatture variftion has two
discontimities, one at the leading edge ( Iis 0 ) an ons atl § i, Te local
heat flow at the location x for lamnar flow ma have to be deIund n order
tution
to find the Integral int build the differential quotient dbk/d5%
mltipl01A
ths7ale0 and Plot the Produce over S as indicated
In the bmwa diagram of E.3. l3J'e area under this cume represents the
Integral. fte rumation haa to beextended over I w@0 ad 11 an,~

j [i~
TW(Vt )-T (N-)) a [T,)-r + 1 Tw(%,)-T ('I-)]

Cob - 10)
Te calculation is however quite tedious becase the iatepatin has to he pert~no
for aW location z along the surface torwkwich the loca heat flow is to be Gstemied.
Additionall'I the term uner the insewa siOs ten" aaus twer"s Infinity Vhs
the variable under the Inteal sign approaches the Uinit z aw own in Flomz 13.
Therefore the contribution of the region naer z to the Integral has to be evaliastd
by som special procedure, as miggusted In Appendix IT.

Me denominator In both Equations Ob - 8 and Ob - 9 can be written also In te

following formg

- WV j
fdT,

(Ob - 12)

33
It is in the example equal to the area under the dashed curve in Fgaro 13 Flus the
sum of the two sudden temperature changes at the leading edge and at | 15L.

The procedure described assumed that the wall temperature distribution is

known. The Inverse problem to find the wall temperature for a prescribed convective
heat flow has been solved In Reference 117. Actually, the wall temperature has
usually to be determined from a heat balance between the convective heat transfer
and the radiative heat exchange of the surface with the surrounding and the heat
conducted away into the solid interior. Generally, this problem can be solved only
by a tedious trial and error procedure.

In this situation it is important to note that by treatment of some simplified

case it has been determined that the ratio of the differences between wall and
stream temperatures, calculated by the procedure described in this ohapter)to the
corresponding temperature difference based on a simplified procedure which neglects
the influtaoe of the wall temperature variation on the heat transfer coefficient Is
considerably smaller than the ratio of the heat transfer coefficients determined tV
both procedures.

Lihthill (Ref. 119) based his calculations on the assumption that a thin-
walled flat plate receives heat by laminar convection from a gas stream and
discharges heat by radiation Into a surrounding with a temperature of 00 Rankine.
Heat conduction within the wall and aW other heat removal has been neglected.
Using the expressions given bv Squations Cb - 3 and Gb - 4 for the convective heat
flow into the surface and balancing this heat flow ty the radiative heat transfer
an Integral equation Is obtained which has been solved. Mhe result of this
calculation which is contained in Figure 2 of the referenced report can be used to
make a comparison with a calculation based on heat transfer coefficiente for
Isothermal surfaces. It has been found that the wall temperature calculated by
Lighthill differs up to 10 per cant from the absolute wall temperature obtained by
the simplified calculations.

Another calculation has been made V fryson and Edwards (lbf. 112) for the
following situation: A flat plate is Initially at a constant temperature and is
suddenly s1bjected to the Influence of a high velocity gas flow of constant or
exponentially decreasing velocity. Under the Influence of the convective heat
transfer the temeature adjusts itself gradually to the temerature of the stream.
Radiation has been neglected and the convective heat flow Is balanced with the heat
stored within the val (no heat conduction within the vall paallel to the sufae
and constant temperature throughout the wall normal to the surase we assumed).
Te use o Squation Ob - 3 for the heat flow results In an integro-ditterential
equation which has been solved apd ocmp ed with the relts of the oppoximate
pooewes which neglects the influnoe of a tengerature variation on the beat
tresah ooeieients. It was found that the difference between wall tes eature
and statio etrom temeature as obtained by the exact calculation Is up to 20 per
cent lugpr than the difference obtati.d br the approximate prooodar. A diff remne
of 10 per sent in th wall tem rature as found I Ltghthil. or of 20 per oet in
the temgerature difren a determined by kyson ocrrespo-ft as a wall tempera-
ture of 15000 Rnkine and a static strem temperature of 100°R. to a temperatae
differene of aroued 2W00f. and such a diffrence m for Instmo , be doolding in
9ke delsion whether a certain material can be used or not for the skin of an
aircraft. It mq also be that the temperature diffwenoe on special locations like
the lading edg and the trailing ed4e of & solid wing WLioh is besin heated up by
aerodyanio heating (as calculated in efenrone 116) mw be larger than in the above

34
examples. Near the leading edge the temperatures wil be increased and near the
trailing edge decreased by the effect discussed In this chapter. In design
calculations it will therefore have to be decided from the situation whothr the
Influence of a variable vall temperature has to be included into the calculations
or vhether the mh simpler procedure which neglects this influenoe is considered
as sufficiently accurate.

A check on the accuracy of the method discussed in this chapter has been made
by Lighthill who compared it with Chamn and Jabesin' a solution for launer flow
and a wall temperature variation according to Squation Cb - 1. geemnt vithin
3 per cent was obtained.

35
 SECTION V

PRICTION A) HEAT TRAMM AT IDW DWITIES

.hesolutions for friction and heat transfer which have been presented up to
nov are subject to a amber of limitations. One of these limitations will be
discussed In the present chapter. It Is connected with the fact that air coLahsts
of single molecules. In deriving the differential equations which are the basis
for the relationships developed, it has been saumed that the properties of the
fluid vary steadily throughout the field investigated. Actually energ and mentum
is carried without change by the single molecules from one to the next collision.
As long as the men free path length of the molecules is very small compared vith
the length dimnsions with which we are concerned, we can assume the variation to
occur steadily.

The man molecular path length increases inversely proportional to the

density and therefore with decreasing density it has to be expected that a limit is
reached where the free molecular path length is of the sae order of magnitude as
the characteristic dimension. The ratio of free molecular path length to the
characteristic length is called Knudsen number. In boundary layer flow as it has
been considered up to now the rate of change of the properties Is largest in a
direction normal to the surface throughout a length equal to the boundary layer
thickness. For this type of flow the Knudsen number is defined as the ratio ct
man free molecular path length to boundary layer thickness. The Knudsen number
In my be expressed by the flow parameters which have been used before, namely the
Reynolds number and the Pbch number. From gas kinetic considerations the following
relationship is derived

0.79? e
(D - 1)

in which 7 ia the ratio of specific heats at constant pressure and volume.

For boundary layer flow the Reynolds umber Is based on the bounday layer
thickness as characteristic length. Changing to the Reynolds number bead on the
distance along the surface moasured from the leading edge results in the following
relationshipe for lmnar flow along a flat plate

(D - 2)
and for turbulent flow

36-3
 According to the above discussion the molecular structure makes :itslf felt
when the Knudsen number exceeds a certain limit. This limit can be veil illustrated
in a diagram which uses the Reynolds number and the Nkch number as abscissa and
ordinate respectively. Such a diaram is presented in Figure 14. Inserted in the
diagram is the approximate limit between turbulent and laminar flow. This limit is
determined by a large number of factors which will be discussed in more detail In a
later chapter. The line given in Figure 14 should be taken only an a very approxi-
mate indication. The line with the break on this boundary between laminar and tur-
bulent flow indicates the conditions for which the Knudsen number for laminar flow
as an expression fbr the ratio of free molecular path length to boundary layer
thickness has the value 0.003. In the field on the right hand side of this line the
free molecular path length is less than 1 per cent of the boundary layer thickness
and the relationships developed in the preceding chapters apply.

With increasing distance from the limiting line towards the left hand side the
free molecular path length increases more and more and it han to be expected that
the molecular structure of the air causes larger and larger deviations from the
conditions encountered in a continuum. Near the limiting line these deviations
occur only within the imnediate neighborhood of the solid surface. Their min
effect is to create finite flow velocities right at the wall surface and a sudden
variation of the temperature (temperature jump). AcoordingLy the regime just to
the left of the limiting line is called slip flow regim.

With decreasing density finally a state is reached where now inversely the
free molecular path length is much larger than any dimension of the body. In mch
a case the molecules impinge on the body surface practically with the omntum and
energy which prevail in large distance from the body. The field in Figure 14 in
which this condition applies Is termed free molecular flow regime. The limit for
this regime is usually fixed by the condition that the value ft/fB which is not
mach different from the Knudsen number has the value 10 (Hof. 140)

1a/Re -
(D - 4)
No boundary lawer exists under this condition. Therefore Re is based on a character-
istic body dimension. The line described by equation D - 4 is also shown on the
left hand side of figure 14.

Another limit to the relationship for contininaum flow presented in the

previous chapters is connected with the fact that with decrealsg Reynolds amber
and increasing laoh number the boundary lover thickness increases. Mhe boundaz
layer equations on which the solutions are based become invalid when the boundary
layer thickness exceeds a certain limit. The displacement of the flow by the
boundary laer becomes then so large that it comes the presswe to vary notably
along a flat plate. This effect is especially large whe" the flow field Is limited
by a shock wave and when this shock wave is comparatively neae to the auface.
Different authors have made estimates on the limit beyond which this bouandry layer
shock wave Interaction becomes important. The results indicate gonarally that the
limit is situated near to the limit between the slip flow and contiaum flow region
somewhat displaced toward the left for low Nah numbers and towrd the right at high
Moch nmbers. Correspondingly, for high Maeh nml trs the slip effect will be the
first one which modifies the contiuum flow conditions when the Reynolds number
decreases, whereas at low Maob numbers the mentioned boundary laer shock wave

37
interaction will cause the first modification.

In general the friction and heat transfer conditions outside the rang. of
continuum flow are extremely complicated and therefore little investigated analyti-
cally as well as experimentally. A comparatively good knowledge exists only for the
regime far to the left in Figure 14 which has been termed "free molecular flow'
regime. Very mach work remains to be done in the Islip flow$ regime and especially
in the 'mixed flow" regime.

In this situation it is important to visualize the regions in the Reynolds-

Hich number field which will be of special importance in future aeronautical
applications. In this paper we are concerned with cooling problems and from this
point of view we can obtain the desired information in the following wow: Cooling
problems vili awqs become critical when the skin of the aircraft exceeds certain
temperature limits. The surface temperature of the aircraft is determined by the
combined effects of convective heat transfer from the atmosphere to the surface, of
emission of radiation from the surface, of irradiation from the sun, and finally of
cooling either by storage of heat in the solid material of the aircraft or by some
cooling process. All of these heat exchange processes depend on the flight altitude
and the flight velocity, both of which can be expressed br the Reynolds number and
ch =umber at which the flight occurs. Two curves calculated on the basis of the
mentioned considerations In hppendix II have been inserted In Figie 14. Along
these curves a wall temperature of 15000 R is obtained by the surface of the aircraft
when the surface radiates heat like a black bodr and when cooling by heat storage or
by mom cooling process does not occur. One of the curves holds for a location on
the aircraft which has a distance of one foot from the leading " and the second
curve for a distance of 5 ft. The break in the curve at Be w 2.100 is connected with
the fact that cooling conditions are more favorable in a laminar boundary layer than
in a turbulent one. Depending on the specific conditions the location of the
transition point may shift to the right or left from the position indicated in the
floe.

Reynolds number and Mch number together with the other data which have been
mentioned, fix the flight altitude along the two curves end the corresponding values
we also indicated. The left hand curve holds also for the condition that the
distance L Is equal to 5 ft. but the emisivity of the surface for beat radiation is
equal to 0.2. he two curves therefore indicate the Reynold and IMh number
conditions for which an uncooled surface of an aircraft ssums a temperature of
1500R in stea fliht. If cooling Is applied to the surface either by heat
storage which is possible for short flight duration or by sm cooling process then
the flight velocity and correspondingly the Mach number can be higher for a certain
Reynolds umber without increasing the surface temperature above the value of
150R. For this condition therefore the flight would occur at a point which is
located in Figare 14 above and to the right of the curv. Aircraft or missile
during their flight may encounter conditions as given by their Reinolds end Mich
number which belqng to the range located below and to the left of the two curves in
Figure 14. This then indicates that the wall temperature even for steady flight
will be lover than the assumed limit of 15000R and generally that such flight oondi-
tions are not critical from a cooling standpoint. The conclusion therefore Is that
the range near the two curves in Figure 14 is the one for which Information on heat
transfer is most important. With this In view we will now review the infozltion
on friction and heat transfer which is available for the regimes outside the
continuum flow region.

38
 Comparatively good information is available for the conditions in the free
molecular flow regime. The kinetic theory of gases gives the possibility to
calculate friction and heat transfer to objects in a flow with uniform velocity and
uniform temperature. Figure 15 and 16 show such data. The figures have been taken
from Reference 135. Figure 15 gives Stanton numbers for a monatomic and a diatomic
gas. The information on a monatomic gas is important because the air molecules
dissociate at very high altitudes. The value A tV which the Stanton numbers are
divided in the figure is the accommodation coefficient. Its value for air is not
too well known. Usually it is assumed to be equal to 0.9. An interesting fact may
be observed in Figure 16 which presents the recovery factors for two-molecular and
one-molecular gases as a function of the Nach number. It can be seen that quite a
aumber of configurations among them the flat plate in a flow parallel to its
surface have recovery factors which are larger than one. We have therefore to
expect that with decreasing density the recovery factor which in continuum flow of
air has a value of 0.85 or 0.9 increases and assumes values above 1 in the free
molecular range. This trend has actually been verified experimentally.

Information on heat transfer in the mixed flow regime is very limited. This
is an unfortunate fact since we have seen that this region is of special Importance
for future aeronautical applications. Calculations which have been made rest on a
comparatively unecure basis since it it todmW even not known exactly what the
differential equations are which describe the flow and energ transfer process in
this region. Experimental information has been obtained at the Ames Laboratory of
the NACA and at the University of California in lerkeley. Such Information Is very
valuable in view of the difficulties of obtaining analytic information. However,
these teats are restricted to Mach numbers below 4 and from Figure 14 it can be
son that the very high NIwh number condition is especially important in this range.
It is not at all sure whether information obtained at small Mach numbers can be used
for the high 4oh number regime as well. One additional factor may be observed in
Fip r. 14. Any information obtained from analysis and all the experimental
information obtained so far deal with lainar flow region. The curves indicating
the flight conditions under which the surface of an aircraft assume a temperature
of 500"R however, indicate that at least part of the flight of aircraft and
missiles has to be expected to be In the turbulent region, at least ewar the slip
flow regime and then exists practically no knowledge on heat transfer under
turbalent flow conditions outside of the continuum regime.

In the rest of this chapter information will be discussedwhich has resAlted

from investigations in the slip flow and the mixed flow regime for linar flow
conditions. Near the contimam regime the normal boundary layer equaLons still
hold to a good approximation. Only the boundary conditions have to be changed to
take into account the finite velocities and the sudden change in Invsrature
occuring on a solid surface (slip flow end temperature jump) and the Interaction
between shock waves and boundary layer. Phalen (Ref. 132) has made a calculation
on this basis aich is assent ally a perturbation of the normal contnum solutioL.
This calculation determines to a first approximation the changes which occur in the
norml continuum solutions when the flow enters the slip flow regim. boe results
of the calculations are expressed ty the following equations for the skin friction
coefficient

(D - 5)

'9
and for the heat transfer coefficient

(D - 6)
In both equations the subscript zero indicates the normal continuum solution, for
instance, as discussed in the previous chapters of this paper. The subscript 1
indicates the chang which this continuum solution undergoes to a first approximation
and the parameter 8 is essentially the Knudsen number:

(D - 7)

The calculation procedure is restricted to tae assumption of a constant specific

heat for the air but Is applicable to any variation of viscosity and Prandtl number
with temperature. Table II gives the results of a numarioal evaluation. The
friction factors and heat transfer coefficients as calculated by Young and Janssen
(R1f. 59) are used as the continuum solution and the changes occuring to a firet
approximation have been calculated by the procedure described in Reference 132. Irom
bU II and from the 3quations D - 5 to D - 7 It mar be observed that the shearing
stress increasee above the value in continuum flow when the slip flow region is
entered. Tha Increase Is cased by the boundary layer shock wave Interaction. The
heat transfer parameter decreases below the value In continuum flow as a consequence
of the temperature Jump occuring at the surface under slip flow conditions. Both
chanes we of the order of mnJtude of 1 per cent at the limit which is indicated
In 11Pws 14 between continuum flow and slip flow regime. The difference between
the actual and the continuuma solution IncreaseAjith decreasing Maudson number and
Is about 10 per cent when the parameter ft/ 4( assumes the value 0.1 Indicating
that the free mleculer path length has increased to approximately 1/10 of the
boundary laver thickness.

An indication of how friction and heat transfer change throughout the whole
region between continuum and free mlecular flow may be obtained from Fha-ee 17
and 18 which are taken from a paper by Shaaf (nf. 138) mamrizing the work of
the Berkeley roup. The figures contain as dashed lines the solutions obtained
from the boundry laer equations and valid in the continuum flow regime as well as
the solutions obtained for the free molecular flow reome. The full line gives the
recommnded correlation. Some experimental points have also been included. The
Reynolds and Mach numbers appearing in FigAte 18 are based on the conditions behind
the shock wave. The value * is half the apex angle of the cone.

40
 SETION VI

ERICTION AND HAT WANSF UND HIG TDIPA TURE C0NMTIONS

In the flight of aircraft through the atmosphere with extreme Mach number very
high temperatures are encountered within the boundary layer. Thes temperatures
cause the air to dissociate and finally to Ionize (at temperature above 20,00°R).
In this chapter the limitations will be discussed which are imposed on the boundary
lsyer solutions in the previous chapters under high temperature conditions.

Figure 19 has been prepared to indicate what temperatures have to be expected

within the boundary layer at high Mach numbers. The figure holds for a stream
static temperature of 4000 R. When the surface assumes the recovery temperature
then this temperature will also be the highest static temperature rhich is encount-
ered within the boundary layer. This recovery temperature of a flat plate is
indicated in the figure for laminar and turbulent flow conditions. The full line
has been calculated assuming constant specific heat and a recovery factor of 0.85
for laminar flow. The dashed line is calculated for a recovery factor of 0.9
corresponding to turbulent flow and agan assuming constant specific heat. A
third curve has been added from the results of Reference 134, which considered the
specific heat to vary, in order to indicate what variation in the recovery tempbra-
ture is caused by the variation of the specific heat.

In aircraft applications the surface temperature will usually be cooled by

radiation or by some cooling process. Under such conditions the ini4man tempature
occurs somewhere within the boundary laver (See also Fig. 5). Curves have been
added in the figure which indicate this maxiamm temperature within the boundary
layer for the case that the wall temperature is cooled down to the stream static
temperature and for the case that the abeolute wall temperature is four time the
free static temperature (at a temperature of 16000R). For a gas with a Prandtl
number equal to 1 it can be easily derived that the difference between the maxium
temperature within the boundary layer and the stream static temperature is equal
to one fourth of the difference between recovery temperature and static temperature
in the stream for laminar as well as for turbulent flow unde the condition that the
wall temperature is equal to the static stream temperature. For the calculation of
the curves in Figure 19 it has been assumed that this condition still holds
approximately for air with a Prandtl number of 0.72.

It can be determined from Reference 165 that dissociation starts at an altitude

of .50 miles in the air with a temperature around 3,000oR. and at sea level with a
temperature around 6,O00OR. The Figure 19 in which these limits are also inserted
indicates that no dissociation within the boundary layer is expected up to bch
numbers of 6 or 10 for the case where the wall temperature is kept at the recovery
temperature and that dissociation will occur only at lMach numbers above 12 to 18
when the wall temperature is cooled to values below 16000R. The latter case is the
one which is more likely to occur on aircraft.

Two papers (Bef. 142. and 143) have been published recently which Investigate
what effect dissociation will have on friction and heat transfer in laminar flow
along a flat plate. SAch calculations are connected with considerable uncertainties.
They are for instance based on the assumption that the air adjusts itself at each
location within the boundary layer to the equilibrium dissociation belonging to the
local temperature. It is actually quite doubtful whether the fluid particles stay

41
for a sufficient time in the high temperature zone to assume equilibrium dissocia-
tion. Also diffusion of the different gases into which the air dissociates into
the cooler regions may occur. All these facts re neglected in the calculations.
i nother uncertainty is based on our insufficient knowledge of property values of a
Las at extremely high temperature and under the condition of dissociation. In this
point the assumptions made In both sets of calculations differ considerably. Fo
instance Voore (Hef. 143) assumes that the Prandtl number in dissociated air
increases up to values around 6 whereas Crown (Ref. 142) reasons that even with
dissociation the iPandtl number can never assume values larger than 1. Nevertheless
both calculations come in some Important points to the same conclusion. It is found
that dissociation under the assumed conditions has a negligible effect on the skin
friction and on the heat flow entering or leaving the marfacoe per unit area as
long as the a urface temperature is kept at values which are below the temperature at
which the dissociation in the air starts. This situation holds even though very
rach higher temperatures may be encountered within the boundary layer. The condition
that the wall temperature is lower than the dissociation temperature--lower than
say 3,OOOR--will be fulfilled in almost all aeronautical applications.

Crown finds additionally in his calculations which are based on the asaneption
that the Pmandtl number is constant and that the product of viscosity and density
is a linear function of the enthalW that the recovery factor can be represented
within a few per cent by the well-known relationship

r.
(E - 1)

He also finds in fluation 58 of his report an expression which can be transformed

into the following relation fur the ratio of Stanton numbers to friction factor

2 1 - ft '2/3.
O apTr- ic-j Tw

(Z - 2)

This equation Is identical with Equation Ba - 8 and indicates that the Stanton
number is practically independent of dissociation when based on *nthaW.

Ibore performed his calculation considering that the Frandtl number for air
varies with temperature. BnthalW recovery factors calculated from the recovery
temperatures contained In his report have been ompared with enthalW recovery
factors calculated from equation

(3 - 3)
with the Frandtl number introduced at the reference enthaljW as given by AqIation
b - 26. The areement is not too good for this set of data. lbzrula (Z - 3)
gives recovery temperatures whi.h are up to 20 per cent too high. For design
calculatcns the actual recovery temperature under conditions of dissociation is of
little interest since it is so high that no presently known material can withstand
it. Surfaces of aircraft flying at such Ma.ch numbers vwill have to be cooled
considerably. Into calculations of the convective heat transfer under such
conditions the recovery temperature enters only as a means to express the beat flov
conveniently (3quation e - 3) and for this purpose the recovery temperature
determined by Iquation E - 3 gives better results than the actual recovery
temperature.

Crown presented In his paper formulas for a calculation of friction and heat
transfer which are valid for an, temperature dependence of the properties in the
sam way as the equations presented in the preoceeding chapters. The method
discussed in this paper is believed to have the advantage that it Is in accord with
the procedures with which in heat transfer calculations a variation of properties
In accounted for.

The findings which have been discussed are very useful information for an
estimate of heat transfer coefficients under high temperature conditions. Of
course, It has to be kept In mind that they we based on assumptions which still
have to be backed up by experimnts and that they hold directly only for laminar
boundary layers. Experimental investigations under high temperature conditions are
needed for more accurate knowledge of friction and heat transfer conditions. They
are however, extremely difficult to perform. In wind tunnels for Instance, It is
almost hopeless to increase the temperatures to such values that on surfaces, which
are c)oled to conditions similar to the ones expected on high velocity aircraft the
inflence of dissociation can be studied. Prom Figure 19 for exmple, it can
easily be determined that for a statio temerature of 4000B. in an air strem and a
pressure correspooding to an altitude of 50 =les the total temperature or stanation
temperature has to be raised to values above 8,000 to 10,000OR. before aW
dissociation will occur In boundary layers which develop on surfaces cooled to a
temperature of l16000R. It Is to be expected that even considerably higher total
temperatures we necessary to obtain a marked effect of dissociation on friction
and heat transfer at the surface.

43
 SWTION VII

1RANITION MON WOW TO ZMW NfLW FLOW

In order to apply the information presented in the previous chapters it is

necessary to have additional knovledge as to which portion of a surtace is covered
tV turbulent boundary layers or, in other words where the transition from laminar
to turbulent flow occurs. This point is the weakest link in the whole procedure
to predict wall temperatures and heat transfer coefficients. The present day
knowledge has been recently sumaried by Galey In Reference 147. The reader
therefore is referred to this reference. Only one figure may be added here which
presents the results of a stability calculation by Van Driest (Ref. 145) since It
eztends to considerably higher Hach numbers than the information contained in
Reference 147. Figure 20 shows as a function of Mach number and of wall to stream
static temperature the Reynolds number at which boundary layers on a flat plate
become unstable for small fluctuations of certain wave length.

The full lines in the figure indicate the critical Reynolds numbers for air of
394.R. static stream temperature assuming a rondtl number 0.72 and a viscosity
which follows Suthrland's relation. Limited by the curve ie a 00 Is a field In
which the flow fluctuations always damp out. The border of this field was also
calculated for a gas w ith a viscositr varying proportional to temperature ( Jt a
constant) and for two Pkandtl numbers. The dashed line holds for these conditions.
It may be recognized that both parmters have considerable Influence on stability.

It iS known especially from experiments at a lover Mach amber that a certain

time or length Is necessary for these fluctuations to increase to such an amount
that they transform the flow to turbulence. Accordingly the critical Raynolds
number for transition Is by a factor up to 100 larger than the Reynolds number
predicted from Flgure 20. However, qualitatively the trends redicted by stability
calculations have been verified generally by experiments. The figwe can therefore
be used to mkn estimates an the Influence of Nah number and wall to stream
temperature ratio on the critical hernolds number at which transition occurs. Bven
in this respect however, one has to be careful, since Gamely presents experimental
information which in m ases does not conform with predictions of the stability
theory. Figure 20 together with the Information in teferencea 247 has been used to
estimate the critical Reynolds number for a wall to stream temper tue ratio equal
to 3 and this estiMted ourve hs been entered into Figure 14. %r caeWim this
curve with the two lins which describe the flight conditions for an aircraft with
a surface temperature of 1500o0. It mw be concluded that probably up to an altitude
around 30 or 40 miles the boundary layers will be turbulent over the major portion
of the surface. At higher altitudes, when the aircraftt enters the mixed flow
region and oves towards the free aeleeulm flow relms, the turbulence has to die
down sm here beause certainly no turbulence can be imagned in the free molecular
flow regime.
 ADEN

E. R. Van Driest recently extended his calculations of friction and beat

transfer in a lainna boundway laer on a flat plate to cover a very vide rane of
Hach numbers (1 0 to 16) of total stream temperatures (T. 0 to 20000F) and of the
ratiot enthalw at surface to enthalW in stream (I./t o I to 6) (nOf. 182)
Additionally, results for 400OR stream temperature are presented.

The method to calculate friction factors, recovery factors and heat transfer
parameters 2 St/c with the reference temperature proposed In this report has been
checked against tie results of Van Dlet's calculations. Agreemnt of the
frictin factors was within 4 per cent, of the enthalW recovery factors within
2 1/2 per cent and of the heat transfer paramter within 1 per cent.

'5
 APPDNX I
REFEECE T~EERAZME

In this appendix it vil be described how the Equation Be - 16 has been

obtained from the data in Reference 59. In this paper the reference temperature T'
is tabulated. Thais temperature is determined in such a way that the friction
factor, calculated from tle boundary lqer solutions and based on property values
at the Temprature T' agrees with the friction factor for a constant property value
field.

The temperature T' will now be approximated by a temperature To as gives by an

equation of the form:

'N 0 Ts+ A(T - a

Ts) B(T-r T.)
(I - 1)

Tables 2, 3, and 4 in Reference 59 represent the temperature T' for the condition
that the wall temperature is equal to the recovery temperature. In this case
Equation I - 1 changps to:

T*/V. - I* (A*B) (Tr/T. - 1)

(I - 2)

This equation is represented by a straight line in Figure 21. In this diagram the
temperature ratio T'/T takon from Reference 59 is plotted over T./T. . It can be
observed that the points representing the exact values of the reference teraature
can be well approximated by a straight line. The tangent of the line yieldss
A + B= 0.72.

Tables 5, 6, 7 in Reference 59 contain the reference temperature T' for

conditions where the wall temperature Is different from the recovery temperature.
In order to obtain the ocstants A and B individualy foam thes data, Equation
I - 1 Is brought into the form:

T- ToaA(Tw - T.) * (o.72 -A) (Tr- T.) (1- 3)

Or$

o.7 4 .o.26 *A(I'aife)

(U - 4)

This relation Is again represented by a straight line in Figus 22. In this

diam we plotted two valus 0.72 (T -T')/T over (T -T )/T . These data were
calculated fram the data in Reference 39. Itais sen fha this valuas can alo be

46
vell approzinate4d t a straight line. Th tanent of th Insaerted lne gives
A = 0.50. which yields: B = 0.22. In this way Njuation Ba - 16 is established.

The same procedure was carried out replacing in all of the above equations
the temperature k" the enthalW corresponding to the respective temperature. 7he
points Indicated tv the white arabols in Figuwe 21 and 22 were obtained and it =W
be observed that the sme straight line appoximates the enthaiW ratios as wll as
the teWature ratios.

47
 AIPPrTDIX II

HAT BALANCE TQUATION

The temperature which the skin of an aircraft assumes is to be calculated for

the following conditions: The aircraft flies horizontally with constant speed for
a time sufficient for the skin to assume its steady state temperature. 'o cooling
process is used to remove heat from the skin and thn temperature is determined by a
balance between the heat exchanged with the universe by radiation. The skin surface
may be considered "black" with respect to radiation.

The h-at balance per unit area is written:

hi (4- 1,,) . -- %.u.

_ 6,T4

(II - 1)
For turbulent flow the 'quations Ba-38, Ba-7, and Ba-8 are used to express the
heat transfer coefficient by the Reynolds number. The rpcovery enthalpy is
Pxorssed by ?quation Ba-21, Ba-47 and the relation:

2s 2 i
(II - 2)

this results in:

o.0296 (as)O (iW .8(a 2 rX6 ~

(II - 3)

This equation gives for the prescrlbed wall temp-rature of Tw = 1500OR and stream
t-mp-rature the necessary relation bptween Reynolds - and Mach numbers. The stream
temmnrature, however, is connected with thn flight altitude and this value is a
Punc'ion of 1i Fi-d Ma t.rourh t!;- follon~nr r-!htion for the density of air:

quA- I
x asM
(II - 4)
In addition, Pr ani A d~o~n1 on the raf' r-nc- n"halpy ani th-r-for- on stream
t-morature (tomperature of atmosphere in altitude) and on the Mach number. The
problem to ietermine the Rgynolds number which corresponds to a prescribed Mach
number may therefore be solved by a trial and error procedure: Assume a stream
tomperature, calculate from ? cation (II -3) the Reynolds number, from rjuation
(II - 4) the air density and 2heck in tables for proprties of the atmosphere
whther the assumed stream temperature corresponds to the altitude as determined by
the air density. The result of such calculation for x-l and xa5 ft is inserted in
Figure 14.

48
 Laminar flow conditions occur mainly in the mixnd flow regime and the heat
transfer coefficients for this region wouli have to be taken from Figure 18. This,
howov-r, makes the calculation v-ry inconvenient. Therefore the calculations were
made for laminar continuum flow, -xpressing the heat transfer coefficient in
7quLtion (II- 1) by Pquations (Ba-37) and (Ba-28) and were extended only as far
as the continuum relations are axpacted to hold still with a reasonable accuracy.
The curves of constant wall temperature in the upper right hand corner wore
calculated with the heat transfer coefficient for free molecular flow.

49
 L114T FOR ACCELARAION FOR dHCH B0UNDAEW LAYER-
IIDW CAN ES REGARDD S QUASI-SMUZ

F. 1-bore shows in Ref. 5 that boundary layer flow along a flat plate may be
treated as quasi-steady when the di~snsionless parameters:

xVs 2V V(n)

are small. In these parameters V1 in the first time derivative of the velocity,
V the second etc. The influence of the first one of these parameters on the
friction factor and recovery factor of an insulated flat plate with laminar boundary
layer has been evaluated. It was found that the shearing stress is larer by
2.5 per cent and the recovery factor smaller by 2.8 per cent than the steady state
values for flow with constant acceleration when the condition:

t' 10 - 2

(III - 2)

is fulfilled. This means, for instance, that the acceleration of a plate moving
with 1000 ft per sec. speed has to be larger than 10,000 ft/sec2 or larger than
300 tiwas the gravitational acceleration before deviations larger than 2.5 per cent
from steady state conditions are encountered at a location of 1 ft distance from
the leading edge. For a distance of 10 ft the necessary acceleration is 1000
ft/sec 2 or 30 g.

50
 APPNODL IV

EVAiUATION OF INTEGRLS (Ob - 8) and (Gb - 9) NLARi%x

The integrals in Equation (Cb - 8) and (Gb - 9) are of the farm:

J J ()a] b d -
Ml (x

The integral may have been evaluated to such a value 1. near x that for the
remaining interral from SI to x it can be replaced tr:

.7 .A [ (1 ),L]b dy
t~o

(IV - 2)

with the average temperature gradient (dTw/d )m in this interval determinable

with sufficient accuracy !V an estimate.

DI a change of the variable 5 to: y I-

x 3 and introduction ofr a a x-S2 we
obtain:

,~ -
J;- -c>]"
(b>, d,
(IV - 3)
Expanding (l-(y/x)) a
into a series and neglecting terms higher than first order
gives:

(.Z)4. 1 -
x x

(IV - 4)
Introduction into(3N - 3) and integration finally results in:

(Iv -
.5)

51
The location at le can always be chosen such that the approximation (IV - 4) is
fulfilled with sufficient accuracy.
Then CIV - 5) gives the contribut ion of the interval between S.and x to the
integrals (Cb - 8) end (Cb - 9).

52
 APPDIU V

R~nMENC

Mit this ammedlx llteraturo on the sub"ec of beat tim-

for at high uelocittse to cllected, whih hes boom published
@law Jaal7 1950* Literature pWUblihd In proviom 7OUW JA
oalr tmcluded irnfwr m It hes been rofowonoe In this vopwt

IMNODUCTIM

1. Broadwe~Jl,. 3. and Shermsn Pe Suval of Pbrom Nell

hbat-TaMfow Litoratue. EMAMW ]:-t-
~Sao iiV'Orsty or smonigin Projmat H M9-7, huwl
1953

Voi U , No
EsO9 9 fl-335* 1952
3. Bomrt, l. R. 0. Beat tamfow i imi Arm ion

k. lluiuw, Z. Be and Ivy Reome. An g sorR ni

5. Mmor, Pradate 9. vmt*WbrW= FlI1ow.

6. stevaren, K. on the twoivNtuce or a list lat. in

apT"Afimsmas, Vol. k9 Part 2,. pp. U82-19 IM5

=UM OF FRICT2I MWNT UAU M3

1ATM
FM =M, 00n, AM CYLZNDEM 70 M1
3AlIOU FOR TO3 IA1' LAI

7. Draltild, W. S. Sam Nots owThe _feUatt

5,
 8. BradMSaA 1. S. The bmentm itou Ltgu for Can-
fUsut
OJmiFlw
~saibi Tm!W a lo
tonou nlow, ValYSIUly or mlmscta, Iii Nmii IT;

9.Van Driest, 3. R. Tuitulamt Do~dr A naCwI

310. ickert,, Urn U. 8 tfJ.d Treatment of the Turbulent Bourn-

Of 1 Awonntl 3000j OL,. J N. Is pp. 23-28, 38,

11. 0.51.7, C. Thorstio Ivalustion of the Turbulent ada Fric-

12. Mmtohs W. and Wendts .TeLrBonay ro

Ciroular Cons at Zenisf~ iSaz e stes
Rep. ad Trawul. No. 70, Dri~sh R A Pq Mg, IM~
13. Jekobq M. md Dow, W. N. Heat Trafer from ty11drJ Cal Sur-
fes to Air In I~' 'zm!n0w VUslutfsunimaG
Hi~at- ectin. ranactioin L5 .N., Vol.3 -5,pp

14. Mo~r, ftW.ll

Wr.11 !oust= !o6= onBoieCiof lw

16. Moore, Franklin K. Laminar ftouda i on aCnsinca]

17. Sebang R. A. aid Bod,, R. 81d-Friction en Hat-framfer

Xclesu, Vol. 1v5, no. 10, pp. 671-675, 1951

54
 HUT? IIAJBP TO A FLAT RA&TE WITH CONSTANIT SURFACE TZUMTURZ

18. Bradfield, I. S., DeCouruin, D. 0., and Buer,, C. B. Zk-

ki t~gtic oftm Chacterlot~.-s orthe
wrieata
.1. 1 A6 PpiU1a, lY53

19. Bradfield,, W. S., DeCoursin, D. 0., and Blumer, C. B. @iar-

mummoTta Iiiut of Tecbnolog, *Dept. of Aeronautical

Eagirmering,, Rosemount Research Labortories, 1952
20. dn adi, B. and Sternberg, J. Reader's Porms On Bourn-

TVA ~roaMI&L
CI~OWVol ly No y.pp. 6145, 6 4i6 p
1952
21. Cole., Donald Reader's Fiorum, Direct Messuremunt of 3gM-
sonic Skin FrictIon. Journal of the AeronmotIcal Sclenoes,
VoU. lys 9o.T10, p. 717, 1952
22. 0mbr, 0. R. Recent Investigation ofrature Recove7

of the Aeronautical Sciences, Vol. ly, Wo.=1, pp. 1-6,

1149 3952
23. Eckert, E. R. 0. Beat Transfer S iu. g6tCn

214. Fischer, W. W. and Morrin, R. Ho Suiarnie Convective Eeat-

25. Hilton, W.1F. Wind-Tupl Tests for Tempwatuas R4c99"W7

Factors atCIM
SuesncVlcourml or ORe
Asronatioa acienoes, V4l. 1,No. 2, Ipp. 97-100, 1951
26. Soberr Richard 29mison of Theoretical and IerlnA1
met-framofer aawacterlsatics of mallie or Revolution at
Sup-onl 3poft.U.B., I.A.C.A.t Report No. IM5,

27. Schlichtizig, H. and Dram, 0. Groanscbihttbeorle. Karlaruim,,

1951

95
28. Sboulburgs Re BH., Hilla J. A. F.,9 and Ravaes HBA. A
Wimeumtal Detw.aut of Flat late Reovey
nwwnzw vj m 3.4. .I.T.
aI DUPMzOM.O LalowToa7 WIN Tinl NsPoW X9 36q
1952
29. Slack, Ulla 0. lAwet m ofatRt Trmefer

30. Stalder, Jadon R.9 Rdmsuim Mums Wes and 1dlanM,

fbarval A DeteiSAlm~ at the Ludaw-. Trqltioul-,

33. 911agom, NU.1 . -MWtodo

-t1. moa T ofWte
mom~~~~
Vee emorua
MaCA. WMaN. 75P19YJ

IS Impr35e6-068,
5 19538

34. v L.
kiaost M. emlaw -k nGomssp

36. Von Ahlast. 3. Re On the aa lt

56
37. Bakerts 3. R. 0. md Drowits, 0.De wbqr- f

38. Ntmrlsi Cwlo VolooLty ad owtmDtiui

pp. 3947s 65a 1*9

39. Orootemulas P. ~ae' For I Ogmat. on want Trmfen
1w
UW. Aeo l aallnas Tl . , n. 9,pp. 5950

of a on. imOt W
I~~tiU4 md Phbia1 Saimms vol. 199, No. 1o~9,
PP. 53348, VWk
1.1. Sludsrt 3. B., and M~ami, F. B3vsd Lmlmr Mwotion and
Buat frmfw 86McN-I-m tim 1- to 10g U.S.,
N.L.C4.9 Toimloa oIMe lb. heWp jy~r.

1.2. Kludas 3. B. d MiAmi, 1. B~md Xffe.t of !anval Nror

PP. 572-581, 1953

46. KwsqwsbodkS, 3. Ze mA~m'. PaMI 3=0 ROIftk m "A

bpMq, o.Me zu eT. p1369 1953

45-. LUrti, 0. TtA LmMArWPllenw

Inc., Now York, 19%1
1.6. van Le, guvemt ftrmuntia st~eting an

57
47?. Lqy a. and Sobs., R. A. Reader'. Farms Roe= Factors

pp. 35!5-357, 1952

he8. Maidw9 3AnUl C.,0 and PAppw , Cotitine C. I M
Inve aaImn of the Local aid &=W Skinu1Ictici
homw!20 "Morn aYA s at s
urn OK Z-4- U.S., N.A.C*Aot 704=10a NMe RD* iW,
1952
1.9. lubeel, V. 1.9 and Johzmong H. A. A Critical Review of
LIat(4n
Pr md Net-freanfer- Slt a of im ri
Noe2~ 4p 19 *33K Vol.3W,

50, Schuv er, Ridmrd and Ooveg Forrest 3. O mno ho

51. Scbllobt1ui, L6 Sur De"w lt der Lmlm'm.e is dmdmbt

52. Sibulkin, M. and Ebe, 0. R. On the Variation of Ne9t frazufer

209NO.rgmpp. 286-2961, 1953

53. StswulorteKom 3 eae j2=~bead O ~

30riMAp
R andPyical Sciencee, Vol. 200,
No. 1060, pp. 84a-100, 11

55. Tifford, Arthur N. Reader's Forms On the Sef.. Iffocts

58
56. TiLfford, Artwr 3. adM@ap Shng Reader's Forwas On

199 3-11 vPP. 787-789P 1952

57. Tifford, Arthur N. Reader's Forme fteifled Cowm-
f TOL o tW .£5, p358,93599

58. o mg . A. D.o lwiotion In the Lamar Boundary Layer

3A U=
0 swOwo onau QRmer17
TROP ET~i"Ifeist,
Wo Lndon To. 1, Part II,
Pp. 137-164t 1%9
59. Young, Gorage 3.1.e and Jmu.., Earl TeOmvi
law Journal of therm emefCee,
VO*IF# goo 4t pp. 229-236# 288, 1952
60. Young, A. Do Skin Friction Lu the Ladma !2gm11-

61. Young, A. D. Bomi &onz and Skin Frictio n u hee"

&!L
s. mr...,"Vol. 55,

TMW.3I UDI r Lim now

62. Ackrmn 0. Plat Therommer Lu ft Velosit. new with

64.. Brini &, Pal F. and Diasoma vicK a.Do

21e0o end Skim Priotum at Mac.p

65. Cqim, Dem R. and Restart Robert S. MY -at@ Of Tur-

59
66. Co1bump A. P. A detbod of 00 at F~oed Coetiom

Smat-brm Bts, Pams

ap man3, !a nm qun

nflm.
up.mint5i.oa6
Aww 1ee9Etp
1.111U
O*27

1933 9p.555k,15
71. Dow1)4., 3.Skint Friction im t forli

St~rdUnvamt ft pp. 1-59 195

753 ISLlap 336,15

X--Tv
PP. W-60l15
76. eel loUtM for
Mrtins John Joseph Reader's Forms. Ors60

Toe zo 8
Me ZqPP.1k-10# 1953
77. Moadafm, A.J.s!00 Istwess !Eriintal vessurwenta

zou Mwaren
onnou u~rr a qPsi No. 401 955,95
78. Wn~faghan, Be J. and Woka, J. Re The Mossnts of Boat

79. Johiwon, Je 2. sed Monahans R. J. Abauapnt of Meet Trans-

MpreNo.59, 1951, 194.9

80. Monaghan R. J. Neft's Forums Further Cossteats on OTu-

sotVOOZ, IW. .9,pl. 2 1953

81. Pappas,9 COintantiM C. and Rubsin, Morris W. Heat Tranfer

82. hub"1 Mn
orris W. A Mod'f led I~ Ana= Sr ths Cam-

83. Rubesin, Morris, We, Mqdsu, Randall C., and Varga, Steven A.
Af jOel and E=Imntal Investigation of the SkIn

84.. Shang S. F. Invest atim of Turbulent DOu =hmoa

85. Squire, H. 5. Beat Trafer Calculation for Aerofoils. R.aM*

61
 86. Stine, Howard A. and Sohemrr, Richard ixrimantal Inveeti-

87. Tucker, Maurice, and Mauolen, 4wpa..- R. Turbulent Boundr

88o Wilson, Robert I* Turbulent BoundaryLayr Characteristics

or teAronauti Sc lCe, Vol. 1,N. 9, 195D

89. roung, A. D. Thejtiozw of Motion and Energ and the

INFLUENCE OF PRSURS AND 1TI2ATUR3 VARIATION ALONG~ THE MUMAC OF RUT TRANSER

VARIMDI F3UU
90. Cooper, ICM and Tilin, M. P. The Lrninar Flow about Ver

91. Dimemeam We Reader's Forms Calculation of the Thermal.

pp. 64,659 1951

92. Drake, Jr... Robert M. Calculation Method for Three-Dinemuional

93. Eckert, 1. R. 0. and Livingood, Johun N. B. Metbod. for Cal-

cutlation of Beat frainfer In Leminar Region or Air

62
9)4o Ickert, Has U. tin Mthod for 2!j sible

Air Deveomnt Center, 1953

95. Ferrari, Carlo StW' of the Boga - Deo p t in

N'. p pp 33-57v 1950

96. Ferrari, CarloTh wuetBudrb rina2 m
sibi. FlumithPstv rsn*Gain.dunio
the Aronatical Scienes .Vl1. Vo-U o-79 o ,pp.J0-7,
-4047
1951

4~749 194a6 U.S., L.A.C.A.g Technical Mmrazdia No.

1229, 1931
9oKorobkin, I. Local Flow Conditioval RecOVWr Factors, and
98. Nat-Trawfor Coefficients on Us Woe* or a sipre
%~I' der at a Mach Nmber of 218. iIAVORD Rejportn2665,r

99. Lev S. and Saban, Re. A. Skin Friction auid Beat Transfer
for Laminar Bovind!jry-r Flow with variabl Mr
orteels andM rable Fise-Btrem Velocity Paper No. 53-
BA-39 Bertelay, Clifornia.
100. Schuh LreBo4 of3 ri ur. Reports. and

101. Schuhg He A New Method for Calculating Leinar Heat Transfer

Roal Institut of Qoas-sectin or Isonesoiies
Tecbitor
Techno Note NTH Aero TN 33p 1953
102. Saban, Re. A. and Drake, Re. be, Jr. Local Hsat-fronsfer
in Kir-pee A rerm. fper N. 5 -A , Berkely,

63
103. Seban, Rt. A. and Drake,, R. U., Jr. Local Beat Transfer

iii, Issue No. 6, 1952

104g. Tifford, A. N. and Chu, S. T. Heat frazisfer In Lunn

Dogg Layrs Subject to Sface Pressure and T mr-
Mach., Ohio State Un- Preas 363-377, 1952
105. Tifford, Arthur N. and Cam, Shang To Beat fraufer an

Center, 1953
106. fruckeinbrodt. Z. A A Ee te Method for the Calculation,

Fow*.ora of te Arona c cences, Vol. ,y

V7,pp. 428-1429, 1952
107e Tuaker, Marice Appoximate Calculation of Turbulent Boem-

108. Tucker,, -Mrie AN 4 te Turbulent opayaw evl

opmnt In pln !M W Fo IITeal
Insulated surfaces with A lcation to :!2e=. 2mi-Tunzial
0;;nFo;r Vorrectics. U.3., NLAaC.A*, Technical Note No.
20415, 1950

109. Wimbrowp WiLli Rt. and ScIbrrer, Richard Lainar-Boumm

U.S., N.A.C.A.0 Technical No10No 214, 1750

VARIABLE WALL TEMUATURE

110e Bod,Re Heat Transfer to ala&Lnw Bue=Lyrwt

reqaraur. ints or Engineering Rlesearch Univ. or

m~Yireipereey, Calif., 1950

Surface T0---MP0 ur.APidNt.p 7,-5

64
112s Brybon, A. Z. and Edwards, Rt.N. The Effect of Nonuniform
Surf ace Temprature on the Trans lent AeldyMie WaeTing
of Thin-skinned Bodies.* Journal or tin Aeronautical
Sciences. Vol. 1,1W. 7, pp. 471-475, 1952
113. Chapmn, Rt. and Rubesin, Morris W. IMeature and Velocity

of the Aeronautical Sciences,, Vol. 16, NO* 9, pp. 54-

565, 19149

11chr
I!!tsaraur.Zeluachrixt fuer angew. Mat.- und

115. Fage, A. and Falkner, V. M. Relation between Heat Trasfer and

Surface Friction for LMMnnw Flow* ARC RAM No. 1400, 1M31
116. Kaye, JosephThe Transient TeMprature Distribution in a
Win Fli-n-Ia
suerv a pees.journal of the naeu-
TM sc~iines o 17, No.I10PP. 787-807, 816, 1950

117. Klein, J. and Tribus, X. Forced Convection from Ion-Isothermal

Surfaces. University of Michigan, Engr. Res. lImt., Proj.
W92W7 1952
118. Levy, Solomon fsat-frmwfer to Comat-oet Laminar Bairn-
9!.yr Flow th Power-Ainctiou Free-atream Velocit-
cences, To1 1, NO 5, pp.3~1-34a8, 1952
119. Lighthill, N. J. Contributions to the Theor of Beat Transfer
Proceedings seisA ath. adPhysical Sciences, Vol.
202,, No. 1070, PP. 359-377, 1950
120. Rubee in,Morris I. The Effect of Arbitrary Surfaaa-Temprature
4A!o a Flat Piats on the Convective Bat Transfr in an
e 7rbuent
Boplay N -w.U.S., I.A.C.A.,

121. Rubeein, Abrrie W. AnA4 lli Investiga&tion of the Beat Tranm

for Between a ?FuindFiti~ats Parallel to the
Diecto ofFo E a *ti~
Dson i L&
Temeraure 3Twhes, umvftity.ef caIU'ceuia, E 0e . 1914

65
122. Scesa, S. and Sauer, F. M. An E!primental Invetiation of
Convective Heat Transfer to Air from a Flat Plate RWi
alte5se Discontinuous Surface Temperature. Trans*
AWE74j 79PP. Lz>1 12552 1952

123. Schuh,Heat Reader'safForums

H. Trnse 1r OnWlls
As totic Solutions
eoratu i a for
T the

o. 146,
147,,t 1953ronautc
o Scences, V 0, , pp.

124. Schuh, H. The Solution of the Laminar-B d -L r uation

for the Plate for Velocity and eperature Hoe
for Variable 4an Mroeries and for the D Oruwin
Fleld at H!gh Concentration. (Z W B/F B/1"BO, Augist 18,
19WJ U.S., N.A.C.A,, Technical Memorandum Mo. 1275, 1950

125. Tifford, Arthur N. Reader's Formes On the Theory of Heat

Transfer Throug a Lln arDoudary Laer. Journal of
the Aeronautical Sciences, Vol. 15, INo. 1, pp. 283, 284a
1952

FRICTION AND HEAT TASFE AT LOW DENSITIES

126. Bertram, Mitchel H. An Approximate Method for Deteu±!.ng the

Di.placement Effects and Viscous Drag or LN a HUd
Iwsr in Two-d ion nic Flow. U.S.,
Technical Note 2773, 1952
127. DeLauer, R. D. xp.rimental Investigation of Heat Transfer
at Hypersonic Hach Number. Bureau of Aercnautics, Navy
Dept. and agamatsu, H. T. California institute of Tech-
nology, 1953

128. Drake, R. M. and Backer, G. H. Heat Transfer from Spheres in

Supersonic Flow to a Rarefied Gas. AS Preprint no.
51-A55s, 1951

129. Drake, R. M. amd Maslach, G. J. Heat Transfer from Righ*

Circular Cones to a Rarefied Gal in Supersonic Fow. UC8
Report HS-150-91, 1952
130. Leo., Loster Reader's Forums on the BoV!, -Luations

pp. 1143-14.5t 1953

131. Lees, L. and Probsein, R. F. HyesncVsosnwoe
a Flat Plato. Princeton Uhvers y ron. En. L. Fp.
go. 19% 52
132. Msln, Stepben H. Second Approuimation to Laninar Co. r-

133. Mushlm, H. Zetimate of SJiU Effect on 22eauaibl. Laminar

Bo~&Tj~ SknFito.US, N.. TA,--cifcal

1314. Monweloer, T. The Two-Dimnional Laminar Bo2Md L r: at

.Hyersni-NO40Col. o A rcnautics, G onied

135. Oppenhein., A. K. Generalized Theory of Convective Heat Trans-

fer in a Free-Molecule ~ournhl
Jlw of the Aeronau-
tical Sciences,, Vol. 20, N. 1, p. 4.9, 1953
136. Probstein, Ronald F. and Loss, Lester, On the Recovery Factor
for Hyprsonic Flow with a Self-IndUced Pressure Caradiont.
Princeton University, Aeronautical Engineering Lab. Re-_
port No. 217, 1953
137. Sauer, F. M. and Drake, R. if., Jr. Forced Convection Heat
Tranfer from Horizontal !ylin~ in a Rarefied Gas.
Journal of the Aeronautical sciences, Vol. 20, 1~3j,
p. 175, 1953
138. Scaaf, S. A. goat Transfer Sympoium. Theoretical Consid-

139. Stalder, Jackson R.,, Goodwin, Glen, and Creager, Marcum 0.

Heat Transfer to Bodies in a Hi gh-Sped Rarefied-Gam
Stream. U.S., N.A.C.A., Technical Note no.. 243a5, 1951
140# Tsien, H. S. Su reer q!ics, Mechanics of Rarefied Games.
Jouna
ofth Aronautica Sciences, Vo. 13 No. 1,
pp. 653-6144, 19146

67
1U-i Wegenier, Peter P., linkler, Eva M., and Sibulkin, If. A

AsrontMa Sciences, Vol. M-7o0 3, pp. 210-222,

IRIC'IIN AND HEAT 1'RAMSIU UNDER HiCIH 1=PSRATuRz CONDITIoJs

14i2. Crown, Jo C. The L-Inna Bo~l Layer at Hprsonic Spes

14.3. moore,0 L. L. A Solution of the Linar Bud= L rIa

Sciences, Vol 19, 9. 8, pp. 5050518, 1952

TRANSITION MRM LAMIMAR TO TURBULENT FLOW

1441. Braslow, Albert L. and Visconti, Fioravante Investigat ion of

Bound= r Realde Nusber for Transitin on an NG

Note No. 17049 1945~

14.5. Van Driest. 3. R. Calculation of the Stability of the Laminar

wit Bea mr er. Journa or th Aronantic S-noes

Vol. 19, No. 129 p. 801, 1952
14.6. Dryden,, Hugh L. Review of Published Data on the Effect of
Nou aym
on Transition from Lmninar to TurbulentFo.

14.7. Gasley, Carl Jr.,, on~yL2 tbl n Trasition i

Subeonic and 6Wroi lw ora rtnAron-- a
Rcinee, Vol. 20, 19. , p. 19, 1953

68
14~8. Higgins, Robert W. andl Pappas, Constantine C. An erimental

1499 Lees, Loster Stabilit of the Sy~roonic Laminar Layer with

a Pressure Gradient. Princeton University, Aeronautical
gIQMerkngULa.,Report No. 167, 1950
150. Scherrer, Richard Bo M Trasition on a Cooled 200

151. Sdaerrer, Richard, Wimbrow, William R., and Gowen, Forrest S.

Hat-Transfer and Bomimr-Lar Transition on a Neated
2 0 0 Maen at a Each mnaer or 1.2,3 U,So, I.AsCoAspRe
search MmmoriM=7 moo £5125, 191j
152. Sternberg, Joeeph A fenjtIvsigto ftePsi

gn xMr
i. ra of teAronautia Sciences, o.

GENERAL LITERATUE

153. Bolter, L. I. K., Popexuliek, Ko, Young, 0., and Andamben, J.R.
An Investigation of Aircraft Heaters rnX-Nocturnal fIrr-
dIation as a Function of Alte and Its Use in Dee-
aination of goat Re-fremnts of Aircraft. U.S., N.A.C.A*,

15l4. Brun, S. et Plan, Me La Covctlkon Foe de la Chaleur Avg

du C.N.R.Sts No. 18, 1952

155. Dber, Go Rt. Some Aspects of Supersonic Heat Tranufer. Pro-
ceedings bfte 3. Midwetern Conftenoe on TlIM mechanics,,
1953
156. Hilsenrath, J. The NS-ACA Tables of Therml Properties of
Owses. Table Z.U4, Nati. bureau of Stewdards, 1950

157. Hilton, W. F. 4wrtur fets in Aeronautice., The Aero-

Plane, P. D#I5

69
158. Huston, Wilber B., WaVrf ield, Calvin N., and Stone, Anne Z,

159. Johmion,, H. A. at al. A Deeia Manual for Datemii h

ThrndAl. Chr ca onm

Subs iMS BatS rONicr
Piwt 1.Jbion H A.Old Slack1, A. W. iA l andr
o
o "h e AroM ar Moheno
of t. A
arun Ae. oynmn H. A. at PSsonic L.d 3diation.
Part . Johwon H A. Slack 3. . Phys~ial and

Aircraft Heat Balmose

UnIweify of Caiorniao Dept. of Mach. Ing. Research
Projects. U.S. Army Air Forces Technical Report No. 5632.

160. [ays# J., Keees Jo Bog Elingememith, X. lag Ketchus Go U.,,

and Toog, To r. Measurement of Recovery Factors and
Friction (oefficiente for Supersonic Flow o Ar n
Tas.j=N ~a~ue Data anReut Hase on a
Umt-DiiDNiniona FIOw Mod6l. Ann. meeting Alnw-c
ioE
ftv. 1,954 Pip-or 5-3~a,20 pp. a J1. apple Hoch.
19, 1, PP. 77-96, 195D
161.. [aye, Jo, Ienzan, J. H., mnd Shoulbergo Re. H. Measuement of
Resvery Factors and Friction CofficiEVRTMi7.8-0-ic
Fe 1air in a Tube, awarne Discuss Heet Trenefrr
Mon Ne. T1-13, 1951, Sect. II. London ast.
OnCu.,
Mach. Sagre.; Now Tork, Awrican Soc. Mach. breo 1951.
162. [aye, Joseph, Leoan. J. H., and McAd, W. H. 29prt of Pro-
on Meroemnts of Friction Coefficients Reer Fce
tr and Heat-Trmfe Coefficients ?or Superonlc Jiow
-VA-In a KIMe. Arim Society of Hoch. Nage.
vol. LxXIII, go. III, p. 267.
163. [aye J.o Toon, T. To, and Shoulberg, Re. I Mssuremt of

sinlFlow U, - IlofoR
atrince Rawo. AM. Meti
hAtlantic CitY, aper 51--29(b), 1951

70
164,. Keeimn,, J. H. and Ket Je Gas Tables. J. Wley & So=$, New
York,, 1948
165. Krieger, F. 4. and White, W. B. The 2I itomanMw ?hezuo-

RM Mh Rim (apdrat ion,, I - 149, l1ii9

166. Lini, T. C. and Street, R. 1. Effect of Variable Viscoeity

167o Lo, Hbu Determination of Transient Skin Temperature of

coicl odesIWI SFP M M I E V te
U.S., I.A.C.A.9 Techncal Rota no* l7Z52 191g5

16o-176, 1951
169. Mn rIler, T. Surimee Conditione of the Beat Trasferred

170. Ryaii, Lly .1M~nso ~r!n!q Col% Zurich,

Aerodynamlk, Mitteilungeui Hr. 18, PP. 7-52,p 1951
171. Soherer, Richard The Iffecte of Aerdyamc Heating

173.S BuliI eat Trufera ew thfae!Faur ea

o to bli
Re ~ o Evolutio
aet M. Jun
R 0% U.3. N.AC.e
SiiTVIF7 7 ,p.70 1 1o.95219
172~. SiiJoh . ft of G t ation fae the eatur

1951
 175. Stepbones Z. D. TAtoenbre--k Rview. Techft. Note WU

176. Thomeng Rt. V. and UhIpple, F. L. Reader's Forams Astro-

ballitic heat Trut~re -journal of the Aero mtCal
Hclnce,Vl. 5,no. 9, pp. 636, 637s 1951
177. b-ibu IL ad Doolter, L. M. K. An Investitation of Aircraft
Beaters 11-Properties of Uspoi. U.S., X.A.C.1.0 A.R.R
"t9 174E
178. ZckrtI.G and WisesW. Meqg e opeauvreln
suf der Oberfiagobwscol
h1aauower unbhjjor
Korir Forsohg.QebqIng. -svol.13.1942.pp.246
179. Ra,.eEprmnso !AzM g M*Mitt.Inst.f.
A~r~VIIIIO
B~dft*OCMKOM@h~lZUrich,1951,
So.18. pp.7-50.
180. Loeb,Ks WiuaklerpZ.M. and Perab,J,.kperigiontal Investiggtion of
TurbulentDoundar &more in Hypersoni a-Flow.
InseAeroSci. 1954. Preprint No.452.
181. Abbot,1.1. Sam Factors Contributing to Scale Rffectsat Super-.
soi ooos Paper presented at 4 th Meeting of AGAUt
Wiadtun and Model Testing Pawsl, London, Dn,1sand.
Sept.3-U1 ,1953.
182. Van Driest,E.R. The LaaixorDoundaw Iyez With Variable fluid
ftOASD North American Aviation Inc. Report Nr.
AL 1866194.

72
 Tt

Tr

/v
/2 a

TwV

Figure 1. Profiles of Static Temperature T and Total Temperature Tt

in High Spnel' Boundary Layer.

73
 ol
0

C4

_ _ _ _ _

744

__ _ _ _ _ _ _ _ _ 0 0
 140

5w4

cYl
a u,

75z
 +3

r P.4

I
I. 0

I 5.

00

__ -_ _ - _ _ _ _ _ .-4.
3k-

Ia--a

CM,

-v r4

76.
 T -

TV"-

Tw

Tprnpprature in High
Figure 5. Temporature Profiles and Rqf~'r-%!cm
Speed Bouniary Layers.

77
 0

0~ N0

*L

L-'--
a $

00

/C,

* 0o

0(

0
OD,

784
 0.87

0.8

o od

0.81

10, 0 100 CONE

a 200
OB4, x600

0 2 4 68 10
M2

Figure 7. Calculated Laiminiar Fnthalpy Recovery Factors for Flat Plate

79
 T3 4P0*R

H Ma

4- 1 * F~

80
 I4-
*-r i

C-D
4.3

*00

o I 0

4 45

I I .0

V 22

so 400x 0 '

-- -- 0

oOJ~0 - 000

CD CY 0 NO
4
81
 wj
zW

4c IW 4cI 49

aa.w - P'
Jo

.H
a-
4

°rl
,0

-i 0

1U

040

r.

82
1~14

4)

US qxbD C-,

0
4 ~4
,0

4UX

0 -
oU

0.S
IU

83
 1.0

0.8e000 0

0.6
Nu
R0. 4

R- zO. 72
ATuo xUS
0.2

0 _ _ _ _

0 2 4 6 8 10

Figure 12. Laminar Hm'it Tranisfer on Flat Plate

with Variable~ Wall Tpm-prature.

84
 WT
Q-.Tr

XX

lf-

-d
7 -T .r

Figire 13. Calculation of Heat Transfer Along Flat

Plate with Variable Wall Tempprature

85

I
 tF--
E
2

2_

0
e-
80 04D 0 40V

CD 40)

86
 (0

0,0 000 -O0

10 C,

Iit
04

C-0

00
00x

0 0 I0Y 0
N0 0 0
-i'7o 0
 a_ o
0

1 2 0

CID)

0y SL.

14U
d. H- -v 9 -0-T-

ON 0
w A
o0 0QC
 A44
/

41-3

4-)

amU

0 0..

.4.

00

89-
 0

0N ~U
TI

0 * C

w 14

___ ___" 1____ 5.

r.

f!--t
5.4
-

di d

* 00

ddc; o~
-oo0 0 0

90
 0 ~
,n a
-o -) (D N- - 4

I--

IL

0D

0 pI~

91
IL

*~ I,

0
r4

IT
\jP-1

00

I4.)

Wr

92-
 +3

•4.

4
-- -

43

93
 44

c4-,

- - coo
" III
G 0

I I

-4:

@ 41 I

I,- 0.i

94

94-
4 )

d)O~ edU
0. 'd 10 10 -Tj 4,l
v5 44
kS5 Q 0 0 to 1.

54 .02 c~-

4k be ) GI -04 - )t~

a~ o'4 :u,0
0d : d 0- 0

5 4 1.4

ot (d,, 0 s
~
(a,

o 0 N4 .0 ' a . v 0 . v4) *. 3
44 -' 0" a k4O .~ - 4) - 4 . 0 .

.5~0 rno 4)0

* 4 4'N
4 4
4~ -0 )~4) 0 4) (d

Ln (A r0. r o 0 o'
0 0 u.

a - o. z-