Beruflich Dokumente
Kultur Dokumente
WILLIAM S. SARIC
Mechanical and Aerospace Engineering
Arizona State Umvers~ty
Tempe AZ 85287.6106 USA
ELI RESHOTKO
Mechanical and Aerospace Engineering
Case-Western Reserve University
Cleveland OH 44106 USA
DANIEL ARNAL
Department Atrothermodynamique
ONERA / CERT
31055 Toulouse. FRANCE
At times, the initial instability can be so strong that the growth The implication of this pamlle/jlow approximation is that the
of linear disturbances is by-passed (Morkovin. 1969) and stability of the flow at a particular station (x, z) is determined
turbulent spots or secondary instabilities occur and the flow by the local conditions at that station independently of all
quickly becomes turbulent. This phenomenon is not well others.
understood but has been documented in cases of roughness
This leads to a system of homogeneous, ordinary differential
and high freestream turbulence (Reshotko, 1986). In this case,
equations for the amplitude functions r(y). The homogeneous
transition prediction schemes based on linear theory fail com-
bamdary conditions (the disturbances must vanish in the
pletely.
freestream and the velocity fluctuations are zero at the wall),
In the flight environment without large surface roughness, the establish an eigenvalue problem: when the mean flow is speci-
most successful of the transition estimation techniques are tied, nontrivial solutions exist only for certain combinations of
those based on the results of linear stability calculations. the parameters a,b,o and R, where R is the Reynolds mm-
Presently we have only developing knowledge of the nonlin- her. This constitutes the dispersion relation.
ear phenomena and breakdowns at supersonic and hypersonic
In this paper, the discussion will be restricted to the spatial
speeds. Thus, much of what is discussed herein relies on linear
theory, which is more relevant than the temporal theory for
theory.
boundary-layer flows. Thus, w is real, a and /? are complex:
This paper has benefited from other recent reviews of the a = a, + ia; and fl= 8, + ifl, Then r is expressedby:
important issues, principally those of Morkovin (1987, 1988).
Morkovin and Reshotko (1990). Reed and Saric (1989). Re- r= r(Y) p(-air-/$r) expb(a,x+P,z -ox)] (2)
shotko (1994). and Reed et al. (1996. 1997). In particular, the
work of Reed et al. (1997) directly complements this report It is possible to define a wavenumber vector i = (a,,&)and
au amplification vector .i = (-ai,-bi)with angles c and @
1 GENERAL BACKGROUND with respect to the x direction:
with
F.,,, =exp
Mach Number M,
marching procedure in the x directton. This method makes it
posstble to treat nonlinear mode interactions with a numerical Figure 1. Wave numbers 2nt/P corresponding to
effort which is much less than that required by DNS. inviscid compressible instabilities in insulated flat-plate
boundary layers. (2D Waves). Solid lines are subsonic
3 STATE-OF-THE-ART CONSIDERATIONS waves. Dashed lines are supersonic waves. (Mack
1969).
3.1 Streamwise Instability
Supersonic disturbances (c, < I -I/M,) were long considered
Theoretical studies have concentrated principally on deter- irrelevant since supersonic neutral disturbances have non-
mining the character and properties of the growing normal vanishing amplitudes at infinite distances from the wall. Mack
modes. In their pioneering study. Lees and Lin (1946) intro- (1969) has shown. however. that amplifying outgoing ruper-
duced the generalized inflection criterion (maximum of the sonic disturbance amplitudes do decay exponentially for large
local angular momentum @l/l& ), and the classification of y , albeit slowly, and are therefore acceptable nomxal modes.
wnvenumbers (frequencies) notlccably larger ,M/lack 1199Oa)
has also calculated cases of h,ghly-cooled boundary layers at
M = 3 and M = 4 5. again m the inwscid hmlt. and found
second-mode superswuc disturbances that xc dcstabilized by
sucrion. Although the growth raLes of these latter disturbances
approach the levels associated with subsonic disturbances, the
disturbance wavenumbers and corresponding frequencies arc
very large. The significance of these disturbances to the tran-
Won process is likely to be negligible since Mack (1990b)
further indicates that they are strongly damped at fimte Rey-
nolds number.
where A is the wave amplitude and the index 0 refers to the lies in the pressure disturbances radiated by the turbulent
streamwise position where the wave becomes unstable. The boundary layers developmg along the nozzle walls. This leads
envelope of the total amplification curves is: to low transition Reynolds numbers, i.e., small values of the
N-factor at transition: typical values are in the range between
1 and 4.
Compressible flows. When compressibility begins to play a laminar boundary layer is less noisy than a turbulent one. This
role, the problem becomes more complex, because the most was done in the quier tunnel built at NASA Langley with a
unstable waves are often oblique waves. As a consequence, a freestream Mach number M, = 3.5 (see description in Beck-
new parameter enters the dispersion relation: the angle with 1983) Notable features are the use of boundary-layer
9, between the streamwise direction and the wavenumber bleed slots upstream of the throat, a careful pohshing and a
vector. It is still assumed that the amplification takes place in careful design of the nozzle walls contour in order to mini-
the x-direction only. i.e., fi = 0 or ij = 0 , but fl, (or q) mix the development of GOtiler vortices. With a lammar
needs to be specified boundary layer on the nozzle walls, the measured pressure
fluctuations can be one or two orders of magnitude below
For second-mode disrurbances, the problem is rather simple,
those measured in conventional facilities.
because the most unstable direction is q = 0 i.e., fi, = 0. As
,B, is also zero, it is sufficient to drop bz in the exponential Several transition experiments were carried out in the quiet
term of equation (I). tunnel On a flat plate. transition Reynolds numbers RXT as
high as 12~10~ were measured; this result corresponds to N-
For first-mode disturbances, the most unstable direction is no
facrors around IO (Chen and Malik 1988). The flow on super-
longer the x-direction; in addition, this direction varies when
sonic sharp cones constitutes a second case where the frees-
the computation proceeds downstream. Two strategies are
tram Mach number is constant in the streamwse direction.
essentially used for two-dimensional flows:
Measurements performed in the quiet tunnel give values of
. At each streamwise location, one finds the direction, RX7 close to 7 or 8~10~. The predicted transition Reynolds
q~, , of the wave which gives the maximum value -ai, number computed for N = IO is 8~10~ , m good agreement
of a, The N-facror is then defined by: with experimental data (Chat and Malik 1988).
dtrect comparisons of the disturbance enwmnment we dlffi- The eN method was also used by Malik et al. ii990b) for the
cult to perform. indirect comparisons are made by looking at rather complex reentry-f experiments. The reentry-F vehtcle
the value of the N-factor at transition. In this respect, the fltght (Johnson et al. 1972) cowsted of a 5 half angle cone with an
experiments on the so-called AEDC cone (Fisher and Dough- imtial nose radius of 2.54 mm. Computatmns were performed
erty 1982) provided us with some interestmg information. at an altitude corresponding to a freestream Mach number
Malik (1984) computed the N-factor for four flight-test points close to 20. The basic flow was calculated by equilibrium gas
correspondmg to M, = 1.2, 1.35. 1.6 and I95 He found Navier-Stokes and PNS codes. At the measured transition
transitton Reynolds numbers at the onset of transition that location, the N-facmr was around 7.5. Roughness effects
were correlated by N-factors between 9 and I I. probably affected the transttion mechanisms. so that the value
of N would be somewhat larger for a perfectly smooth surface.
Correlations in the vertical direction were obtained using two on the sharp-tipped model at small angles of attack are by
probes separated verrically by 1.47 mm and traversing them Doggett et al. (1997). while those on a slightly different sharp-
vertically through the boundary layer With appropriate signal tipped model with wall cooling are by Blanchard (1995). AIt
analysis (see Kimmel and Poggle 1997). the second-mode three sets of experiments are summarized by Wilkinson
waves are shown to have a flattened strocture (Fig. 7) that (1997). In all cases, the cone portions of the models and a
agrees qualitatively with shadowgraphs of the rope waves portion of the flare region were free from any tunnel noise.
for the same configuration These rope waves which appear Only the aft portions of the flare were subject to low-level
before boundary-layer breakdown, seem to be a nonlinear radiated noise from the part of the nozzle wall boundary layers
manifestation of the second mode With an increase in Rey- that were undergoing some mild Gdrtler instability growth.
nolds number approaching transition, the waves become more The experm~entsincluded a detailed study of the mean flow
erect approachmg the angles of turbulent structures. and the stationary freestream disturbance patterns in the Mach
6 nozzle. The experiments documented the growth of the
Cone exoeriments at M = 5 in the Caltech T5 Hvoervelocity
dominant second-mode disturbances with adverse pressure
Shock Tunnel. Boundary-layer transition experiments were
gradient, small nose bluntness, wall cooling and small angle
performed on a Y half-angle cone at zero angle-of-attack in
of attack,
air, nitrogen and carbon dioxide (Germain et al. 1993: Adam
and Hornung 1997). With air and nitrogen as test gases. reser- Under adiabatic wall conditions, the recovery temperature
voir enthalpies varied from 4 MI/kg to 27 Ml/kg at stagnation distributions on the sharp-tipped model showed transition
pressures from IO MPa to 85 MPa. For carbon dioxide, reser- occurring only at the very end of the flared section. Hot-wire
VOXenthalpies varied from 3 MJikg to IO Ml/kg and reservoir spectra for this case taken at each streamwise measuring
pressures from 40 MPa to 95 MPa. Dissociative effects are station showed a dominant second-mode peak at about 226
clearly important. The wall temperature is not specified. kHz corresponding to a linear stability prediction of 230 kHz
However, assuming it to be about 300 K gives a wall enthalpy (Balakumar and Malik 1994). There are however also har-
of just under 0.6 Ml/kg for air and nitrogen and just under 0.5 monic peaks around 449 kHz and 670 kHz indicating some
MJ/kg for carbon dioxide. So these tests are conducted under nonlinear behavior. Very small nose bluntness ehminated
very cold-wall conditions where under linear stability argu- transition but still showed some growth of the second mode at
ments. the second mode should dominate. around 230 kHz. With slightly larger nose bluntness there was
neither transitron nor evidence of second-mode growth. At
However, while the disturbance environment is nor quantita-
two degrees angle of attack, the boundary layer on the wind-
tively described, it is generally regarded as noisy. Therefore
ward side is stabilized and the dominant second-mode fre-
the relationship between the obtained results and linear stab&
quency increased, while on the leeward side the boundary
ity argoments is not clear. Adam and Homung (1997) state
layer is destabilized and the dominant frequency sharply
that no clear relationship is found to exist between the transi-
decreased. Also on the leeward side, localized peaks over a
tmn Reynolds number based on the boundary layer edge
range of low frequencies are observed upstream of the second-
conditions and the resewox enthalpy. However, when refer-
mode growth. These may be associated with crossflow insra-
ence temperature conditions are used instead, the different test
bility.
gases give distinguishable results which are ordered according
to their dissociation energies. These authors also compare Under cold wall conditions (sharp-tipped model at zero angle
their T-5 results with Reentry-F flight data reported by Wright of attack), the second-mode frequency is at slightly higher
and Zoby (1977). The flight data are in a Mach number range frequency as expected theoretically (Balakumu and Malik
from about 8 to 20 and are taken in a relatively quiescent 1994). the first harmonic (at just under 600 kHz) is compara-
environment. Nevertheless, Adam and Homung (1997) show ble in amplitude to the fundamental and transition is advanced
that the transition Reynolds numbers under reference tem- to mid-flare.
perature conditions are between 600,OCOand 2,COO,C0Ofor
Transition studies on curved comoression surfaces in the
both the T-5 and the Reentry-F flight data. This collapse of
Calsoan 48.inch Hvoersonic Shock Tunnel. Pressure and heat
data is interesting and unexpected. More data are needed to
transfer measurements were made on a curved compression
show that this result is other than fortuitous.
ramp whose geometry is typical of an inlet to a scramjet en-
c gine (Holden and Chadwick 1995). The measurements were
Lawlev Mach 6 Owet Tunnel. Experiments were conducted made at Mach numbers 10, I I and 12 for a range of Reynolds
on 5 half-angle cones followed by large-radius circle-arc numbers at each Mach number to place the beginning of
flares. The experiments at adiabatic wall conditions with and transition at various stations along the ramp. Whereas the thin-
wthout small bluntness are by Lachowtr et al. (1996). Those film heat flux gauges were capable of obtaining transient data.
and these were used to more closely identify transition loca- clear that the equilibrium and nonequilibrium solutions can
tions, only mean heat transfer data are presented. In the ah- differ sigmficantly depending on the rates of the reactions
sence of spectral mformation. it is not possible to identify the relative to the time scales of convection and dtffusion For
dommnnt modes of the pro-transitional growth. example. some of the equilibrium modes were determined to
he superson~ modes, each of which was a superposition of
The transition results are not particularly sensitive to the small
incoming and outgomg amplified solutions in the inviscid
differences in Mach number in the M = 10 - I2 range. At the
region of the shock layer. (No similar solutions were found for
highest unit Reynolds numbers, transition occurs on the 5
the nonequilibrium shock layer.) The magnitudes of these
wedge preceding the curved portion of the ramp. As unit
modes oscillated with y in the inviscid region of the shock
Reynolds number IS lowered, transmon moves downstream as
layer This behavior is possible only because the shock layer
expected until at the lowest unit Reynolds numbers tested, it
has a finite thickness. They are also unlike Ma&s higher
occurs around the end of the curved ramp. Although transition
modes (except for the second) in that the disturbance-pressure
Reynolds numbers are not computed, it appears that the high-
phase for all of these supersomc modes changed most BUOI
est transition Reynolds numbers are obtained at the highest
the inwscid region of the shock layer (The dlsturhance-
unit Reynolds numbers. Some of the reduction in transition
pressure phase change for Macks higher modes occurs across
Reynolds number with unit Reynolds number is due to the
the VISCOUF region of the flow, i.e the boundary layer ) In fact.
influence of the adverse pressure gradient on the compression
the disturbance-pressure phase change for all of these super-
ramp, and in fact, the length of the transition zone is reduced
sonic modes through the boundary layer is comparable to that
when the transition occurs on the curved ramp. In these ex-
of Macks second mode.
periments the maximum end-of-transition Reynolds number is
about 3 x IO, below what one expects in flight at these Mach Another effect of the chemical reactions is to increase the size
numbers The facility may not he quiet enough for flight- of the region of relative supersonic flow primarily by reducing
quality data for the reported tests the temperature in the boundary layer through endothermic
reactions, mcreasing the density, and hence decreasing the
3.1.4 Effects of Chemisrry and Bow Shock (Reed et al 1997)
speed of sound. This reduces the frequency of the higher
Linear stability solutions for hypersonic flows are comphcated modes; in particular. the most unstable one. the second mode.
for some of the following reasons. I) At hypersonic speeds, The higher modes in the reacting-gas cases are also more
the gas often cannot he mod&d as perfect because the mo- unstable relative to the corresponding perfect-gas modes The
lecular species begin to dissociate due to aerodynamic heating. first modes are. however, more stable.
In fact, sometimes there are not enough intermolecular colli-
Finally, the finite thickness of the shock layer has n slgmficant
sions to support local chemical equilibrium and a nonequilih-
effect on the first-mode solutions of all of the families. The
Rhum-chemistry model must be used. 2) The bow shock is
effect on higher-mode, higher-frequency solutions does not
close to the edge of the boundary layer and must be included
seem to he as large as long as they are suhsomc. This is per-
in studies of transition.
haps what one would intuitively expect because the shock is
Mahk (1987, 1989, 1990) investigated the stability of an likely stiff and hence difficult to perturb with smaller-
equilibrium-air boundary layer on an adiabatic flat plate. wavelength. larger-wavenumber, higher-frequency distur-
Malik et al. (1990) used the eN method for the reentry-F bances However, the nonparallel effects are known to he
experiments; the haslc state was calculated by equilibrium-gas large for first-mode solutions, and so B complete quantitative
Navier-Stokes and PNS. Gasperas (1990) studied stability for description of the effects of the finite shock-layer thickness
an imperfect gas. Stuckert and Reed (1994) analyzed the awaits either a PSE solution or a DNS analysis
stability of a shock layer in chemical nonequilibrium and
The inclusion of the bow shock IS especially critical to studies
compared results with the flow assuming 1) local chemical
of leadmg-edge receptivity as demonstrated by Zhong (1997).
equilibrium and 2) a perfect gas.
His DNS results over a blunt wedge show that the instability
Stuckert and Reeds coordinate system for both the basic-state waves developed behind the bow shock consist of both first
and stability analysis fit the body and bow shock as coordinate and second modes, His results also mdicate that external
lines. This makes it easier to apply the linearized shock-jump disturbances. especially entropy and vorticlty disturbances.
conditions as the disturbance boundary conditions (e.g. Stuck- enter the boundary layer to generate instnhlhty waves mainly
en 1991). At the surface of the cone, for the nonequilibrium in the leading-edge region.
calculations, the species mass fluxes were set to zero (non-
catalytic wall). whereas for the equilibrium calculations the
disturbances were assumed to he in chemical equilibrium. It is
tolott! (1991) wth the growth rare computed at the maximum
of the mass flux. The nonparallel effects become apprecmble
It is now recognized that the growth rate of two-dimensional
81 M, =45only
waves is only weakly affected by nonparallel effects. This was
demonstrated for the Blasius flow by the computations of This conclusion, however. IS no longer valid for oblique
Gaster (1974) who used the method of multiple scales This wwes. as it can be seen in Fig. 8 (Bertolotti 1991) which
result was confirmed by PSE and DNS computations. PSE displays the evolution of the total growth rate for the same
results were published by Bertolott~ (1991) and by Chang frequency, but with v = 55. The mean flow nonparallelism
et al. (1991) for supersonic Mach numbers. A good agreement exerts B strong destabilizing effect, and this effect increases
was achieved with the multiple scale analysis of El-Hady with increasing Mach number. The importance of this differ-
(1991) at M, =1.6. ence on the N-factor could be significant for supersonic flows
4 since the oblique W~XS are the most unstable ones and for
hypersonic Mach numbers when first mode disturbances are
dominant.
,...
,.f
simdar with those described
Mach I 6 and 4.5 flat-plate
before
flows.
were reported
In general, nonparallel
for
1
/,..;
/ J\ I I .5
I, 4.5
.J,<\
Y 1-l effects appear to be less stgniticant for oblique waves near the
lower branch of the neutral curve but become more important
at higher Reynolds numbers near the upper branch.
noted as Lin the figure) and on the windward ray (denoted BS wth experiments over a range of unit Reynolds numbers and
W) Results at zero angle of attack are given for comparison. angles of attack.
The most striking feature is that N increases from windward to Thl*netransition process on a delta wing was investigated at
leeward ray for first mode disturbances. whereas it decreases Imperial College gun tunnel within ESATRPTran et 81.
for second-mode disturbances. As a consequence, predicted (1995). A large number of parameters (angle of attack, angle
transition occurs earlier on the leeward ray than on the wind- of sweep. unit Reynolds number, leading edge bluntness) WBS
ward ray (in agreement with the experiments) if one assumes investigated. Linear. local stability computations have then
that it is triggered by first mode disturbances. But transitmn been performed for a few experimental contiguratmns (Amal
would appear earlier on the windward ray if it were induced eta1 1996) With sharp leading edges, the boundary layer
by second-mode disturbances. This could be an indication that development was nearly two-dimensional. and experimental
second-mode. high frequency disturbances do not play any
transition locations were correlated with N-factors between 1 larger for three-dimensional flows than for two-dimensional
and 2 As soon as the leading edge radius was increased, the flows, resulting in a larger scatter m the (linear) N-J&or at
boundary layer flow became highly three-dimensional, and transItlo.
transition was induced by crossflow instability at much larger
In spite of these deficiencies, the eN method can help to
values of the N-factor From this (rather limited) series of
understand the physics of the transition phenomena. As for
computations, it appears that the N-factors at transition are
two-dimensional flows, it is useful for parametric studies, as
likely to be very different depending on the type of dommant
soon as the N-factor for the different types of disturbances has
instability. Reed and Haynes Reed and Haynes (1994) investi-
been fixed from some reference case
gated the local instability properties of a supersomc rotating
cone at zero angle of attack, which was used as a model of a 3.3 Giirtler Instability
swept wing. The results were then used to develop a simple
Boundary layers over concave surfaces are suh,ect to a cen-
criterion which was applied to several experiments on cones at
trifugal instability that appears m the form of stationary
angle of attack.
streamww-oriented. counter-rotating vortices usually called
Shortcomings of the N-factor Method. As the eN method 1s GBrtler vortices. A description of the physical mechanisms
based on linear stability only. receptwity and nonlinear and a recent revtew IS given by Saric (1994h).
mechanisms are not taken into account In addition the
A graphic example of this mechanism can be observed as
nonparallel effects are neglected in the local procedure. and rt
streamwise surface striations on reentry vehicle heat shaelds.
has been shown that they could he important for oblique
The differential ablation between the nose cone and the heat
waves.
shield of reentry vehicles produces a concave surface which
Several particular problems arise for three-dimensional flows. induces a Gdrtler instability. The vortices then cause spanwise
The first one is to choose the best strategy of integration of variations m the heat transfer.
the N-facror. The envelope method is widely used. but its
3.3. I Swrpl- Wing Flows and Sragnation Flows
physical meaning is not clear in many cases. A classical
problem occurs for swept wings, especially in transonic con- Hall (1985) has shown that beyond a very small sweep angle.
ditions: in the vicinity of the point of minimum pressure, rapid the prmcipal instability on a swept wmg with concave curve-
variations in f+~, can be observed when the dominant instabil- ture i9 a crossflow instability and not a Gbnler lnstahility
ity suddenly changes from the crossflow to the streamwise This analysis reduces the concern about Gdrtler problems in
type The other strategies are often used in order to avoid swept-wing flows. Bassom and Hall (1991) extend these ideas
From B practical point of wew, the most important issue is the circulation criterion appears to be satisfied near the stagnation
value of the N-factor at the onset of transition. Concerning region of bluff bodies and a Gdrtler mstahlhty has been a
two-dimensional flows, it is has been shown (paragraph 3.1.2) suspect regarding the angin of streamwise vomcity in a
that the eN method with N q lOcan be applied to predtct boundary layer such as observed by Klehanoff et al. (1962).
transition for a wde range of flows in a low background Stuart (1984) has shown conclusively that an instability does
large scatter I the values of N at the unset of transition A Induced by GLinler ~ortxes in a compressible flow IS in the
possible reason is that crossflow and streamwise instabilities design of quiet supersonic wind tunnels. The mayor source
are not initiated by the same type of forced disturbances. For of freestream noise is radiation from the turbulent boundary
instance, the crossflow disturbances are very sensitive to layers on the nozzle. Since the accelerating pressure gradient
micron-sized roughness elements which have no effect on generally stabilizes Tollmien-Schlichting instabilities. transi-
streamwse disturbances, see Radeetsky et al. (1993). Another tion is induced by Gdrtler vortices on the concave region of
reason is that there exists a multiplicity of nonlinear processes the nozzle (Beckwith et al. 1985. Chen et al 1985). Efforts to
before the breakdown to turbulence, depending on the relative calculate the growth of the instability with compressibility are
part of stationary and travelling unstable modes. In particular, given by Hbmmerlin (1961), Kobayashi and Kohama (1977).
it is now clear that the extent of the nonlinear regime IS much El-Hady and erma (1983a.h). Tumin and Chernov (1988).
2-14
u- x~exp(oz)expb(az+~--wr)] (17)
+
periments on a swept cylinder (M.. = 7.1. M, = 5 13). The
critical Reynolds number for adiabatic conditions is SOOJ
e
??;, E 172. This surprismg result means that natural transi- ,
tion is hkely to occur at a Reynolds number which is smaller
than that corresponding to leading-edge contaminatmn 400 8
(R,;. 250 see next paragraph) For a wall temperature rntm
-
T&w around 0.4, R,, increnses up to 13-500. I e.
$:? 4500. 300
I
PCrirena Bushnell and Huff-
man (1967) observed that laminar flow on a smooth attach- 200
Me
ment line could extst up to at least R-,, =8x10, where 7
R,,, is the Reynolds number based on freestream conditions
/I ,>,I I,,
and leading-edge diameter. This criterion is valid for Mach
d. 2.5 5.0
numbers M, between 2 and 8. and for sweep angles larger
Figure 13. Transition Reynolds numbers as a function of
than 40. Another criterion proposed by Poll is R 5 660 at M, on a smooth leading edge without end plates
transitmn. It will be discussed later on (TJR,, a 0 4) From (Galliard 1993).
Exoerimental Results. There are only a few experimental Recent experiments were performed by Holden et al. (1994) in
results available for the problem of natural transition along the Calspan 48.inch shock tunnel on a highly swept cylinder,
the attachment line of swept models at high-speed condttions. at freestream Mach numbers from 10 to 12. Attachment line
Careful experiments-were conducted by Creel et al. (1987) I measuwnents with a smooth leading edge, with end plates
the quiet tunnel at NASA Langley; under adiabatic condi and with roughness strips were reported. With a smooth lead-
tions, transition was found to occur for i? close to 700, a ing edge. tmnsttion was observed for values of R from
value which IS nearly twice the theoretical value computed by 600 to 800. The corresponding freestream Reynolds numbers
Malik and Beckwith (1988). However, it is interestmg to note based on cylmder diameter were above 8x10 However.
that the experimental value IS close IO that of the cnwal turbulent bursts were observed at R z 550, which could be
Reynolds number iF computed for incompressible flow. This attributed to a misalignment of the nosetip wth the
observation led Poll to the conclusion that R 9 660 could be freestream.
a universal value for transition on smooth leading edges.
Experiments were conducred by Murakami et al. (1995), !n
The most interesting finding of Creels experiments is that the
the hypersonic Ludwieg-tube wind tunnel of DLR at Mach
wind tunnel noise has no effect on the natural transttmn
numbers M, =5 and 6.9 (spanwise Mach numbers
Reynolds number (see Rg. 16) This can be explained by a
M, = 1 9 and 2.9). Transition was detected on swept cylin-
theory on the receptivity of supersonic laminar boundary layer
ders by liqu&crystal techmque. In the absence of gross dis-
to acoustic disturbances; calculations by Gaponov demon-
turbance sources, the Reynolds numbers 2 for natural
strated that external noise cannot generate unstable waves
transition were in the range 680.750, corresponding to values
when the boundary-layer thickness IS constant, as it is the case
of R+ I the range 0.9-1.2~10~. These values agree wth
along the attachment line.
Creels data.
The measurements by Gaillard (1993). shown in Fig 13, have
3.4.2 Artachenr-Lme Conrnmrnarion by End Plares
been performed on circular cylinders as well as on cylmders
having a nearly flat surface around the attachment line (in Emwincal Criteria. Leading edge contammation is likely to
order to reach large values of ii ). The experimental results occur when a swept body in attached to B solid surface (fuse-
for TWIT, - 0.4 are plotted in Fig. I3 It can be seen that the lage. wind-tunnel wall. etc ). This problem has been widely
transition Reynolds number decreases when the spanwise studied for low-speed flows (see reviews in Poll I978 and
Mach number increases. Gaillard also studied the influence of Amal 1992 for instance). and a simple criterion. based on the
the wall temperature. He found rhar cooling has a small stabi- value of F, was developed by Pfenninger (1965). If iT is
hang effect for spanwise Mach numbers up to about 5. and lowec than 250. the bursts of turbulence convected along the
2-17
wall arc ddmped and vnmsh ns they travel along the attach- ment wtth the findings of hlurakaml et aI ilY95) and supports
ment l,ne However. for ii > 250, these bursts an: self- their cxplanatmn for the increase of R: m the exe of short
wstalmng They grow. overlap and the leading-edge region end plates A more surprlsing observaoon was made in the
hecomes turbulent. case of aery shon end plate (L/D = 0.53); in Ih,s case. the
Reynolds number at attachment-hne contaminatmn was re-
Several series of experiments were devoted LO the study of
duced to about 200 The explanation proposed by Gaillard is
leading-edge contamination at high speeds. In most of the
that. for long and short end plates, [he man source of
cases, end plates were used as sources of gross disturbances.
gross disturbances IS the boundary layer (ather turbulent or
The first xtempl to quanrlfy attachment-hoe confaminaoon in
laminar) at the end platelcylmder junction For very shorr
these conditions is due to Bushnell and Huffman (1967). who
end plates, the shock generated at the plate leadmg edge be-
observed that the boundary-layer flow on the leading edge
gins to play the dominant role, leading to a decrease m F:
became turbulent If R_, >2x10S(M_ between 2 and 8.
sweep angle larger than 40).
Region I for 0 < d/q < 0.7 the woe has no effect, and
transition IS triggered by Imear mechamsms; the main pa-
~arne~eris the freestream disturbance level.
Region II: for 0.7 <d/T < 1.5, the wire begins to control
transition. The location of the first turbulent spots moves
closer to the wire when p is increased.
Region III: for 1.5 <d/q < 1.9, the flow is exher fully
Figure 14. Leading edge contamination criterion at high
turbulent or fully laminar behind the wire At a fixed
speed. From (Poll 1985)
value of d/q this change in the boundary layer stfllcnxe
Exoerimental Results. In the experiments of Murakami et al.
occurs for a very small van&loo of ii. Gaster (1967)
(1995). it was found that disturbances generated by end plates
and Cumpsty and Head (1967) observed this phenome-
parallel fo the Incoming flow Induced leading edge confami-
non wtnch Poll called flashmg
nation at free stream Reynolds numbers R,,D of approxi-
mately 0 15 to 0 5x10. in reasonable agreement with Bush- Region IV: for d/q > 1.9, turbulent bursts always appear
nells criterion. When expressed in term of ji , leading-edge immediately behind the wire. But. if ?? is lower ihan
contamination occurred for values which are significantly about 250, they decay more or less rapldly as they are
larger than those due to Pfenningers criterion and other ex- convected along the attachment line If ii is greater than
perimental data (mean value around 300-350). The explana- this critical value, the size of the burst increases and
tion could be that the disturbances generated at the end-
leading-edge contamination occurs It is clear that there
plate/cylinder junction were small because the boundary layer
is a strong slmilaity with the leading-edge contamination
on the end plate ~8s laminar
induced by a wmg-wall JUnCtiOn. One can deduce that
The influence of the end-plate length was lovesligated III there exists a minimum Reynolds number cR 5 -50
) be-
detail by Costa (1940) and Gaillard (1993) at CEAT Poitiers. yond which every turbulent structure generated by a
For long end plates (i.e. L/D > 1 .where L is the plate gross disrurbance source becomes self-susttimng. devel-
length), leading-edge contamination was observed for
ops and makes the leadmg edge turbulent.
R I 250. For L/D = 0.79, the attachment-bne boundary
layer became turbulent for ii z 330. This result IS in agree-
boundary layer is disturbed locally by the roughness ele
fllet.
Usmg a multiple-scales analysis, Reshotko and Khan (1980) stability calculations for their assumed night conditions how-
showed that for a blunted flat plate. upstream of the location ever found no growing disturbances in this portmn of the
2-m
flowfield. Clearly. the effects of nose bhmtmg on stability and TS. swept attachment-lme and crossflow instability and tran-
transition are not resolved. sition provided that It is ascertained that the results are not
affected by tunnel noise. Reed et al (1997) contain additional
3.5 Receptivity and roughness
discussions on this topic.
3.5.1 Use of Ground Based Facilaies
3.5.2 What is Known About Supersonic Receprivity
Transition experiments done into the 1960s and the transition
The disturbance environment affecting boundary layers on
Reynolds number correlations developed from them mcluded
vehicles in flight can come from two sources: a) atmospheric
among the factors affecting transition a so-called unit Rey-
disturbances or particulate effects as modified by passage
nolds number effect, which usually showed itself as an in-
through the inevitable shock waves, and b) mechanical and
crease m transition Reynolds number with increase in wind-
acoustic disturbances coming from other parts of the vehicle.
tunnel total pressure Reshotko (1969) pointed out that this is
Bushnell (1990) supplies background, references and specula-
in fact due to the spectral character of the disturbance em-
tions on the triggering mechanisms associated with such
ronment. To that point, the disturbance environment had not
disturbances.
been specifically considered among the experimental factors
affecting transition. Morkovin (1969) enlarged upon this by Receptivity phenomena for supersonic and hypersonic bound-
pomting out that one must determine the means by which the ary layers are just now beginning to receive serious attention.
disturbance environment generates growing disturbances in a Whatever the receptivity mechanisms, the receptivity consid-
boundary layer, a process that he called receptivity Also in erations are altered considerably by the presence of attached
this time frame, Pate and Schueler (1969) showed convinc- or bow shock waves and the fact that while vorticity and
ingly that transition behavior in conventional supersonic entropy disturbances are convected along streamlines. acoustic
tunnels was dominated by the sound radiated onto wind-tunnel disturbances propagate along (relative) Mach lines (Morkovin
models by the turbulent boundary layers on the tunnel 1987) Any freestream disturbance field is altered in passage
walls an artifact of wind-tunnel testing with no flight coun- through a shock wave. When entropy or vorticity disturbances
terpart This was confirmed by the important flight experi- on streamlines remote from the boundary layer pass through
ments of Dougherty and Fisher (1980) Their transition cone the shock wave, they are modified and are pamally converted
had already been run in many transonic and supersonic wind to pressure waves that impinge on the boundary layer along
tunnels throughout the world In wind tunnels for M > 2 they (relative) Mach lines (Mack 1975). Sound waves change
consistently obtained the familiar transition bucket with a amplitude and refract through shock waves and give rise to
Re,, minimum between Mach numbers 4 and 5. However, in vorticity and entropy disturbances. Furthermore, irregularities
fhght experiments with the cone model mounted ahead of the on the vehicle surface generate pressure waves that are mcdi-
fuselage of an F-15 aircraft, the transition Reynolds number fied by the shock and reflected back toward the body. Once
corrected to adiabatic wall conditions showed consistent the disturbances are intemalized in the boundary layer. the
increase with Mach number over the entire flight speed range problem of matching wavenumbers and frequencies of ampli-
whereas the values for Re,, obtained m conventional wind fied normal modes remains conceptually similar to that at low
tunnels started falling at M > I .6. speeds. Very recently, Zhong (1997, 1998) calculated by DNS
methods the receptivity of two-dimensional and axisymmetric
The Pate and Schueler (1969) work led directly to the devel-
hypersomc (M = 15) blunted configurations to freestream
opment of quiet supersonic wind tunnels. A conventional
acoustic disturbances (upstream of the bow shock wave). He
quiet supersonic tunnel is one where laminar boundary layers
was able to show the processing of the incoming disturbances
are maintained over as much of the nozzle and test sectron
by the bow shock wave leading to vonicity, entropy as well as
walls as possible. The character, limitations and accomplish-
acoustic disturbances downstream of the shock wave, and the
ments of the NASA-Langley M = 3.5 pilot quiet tunnel are
subsequent development of first- and second-mode distur-
described by Beckwith et al. (1990). Quiet supersonic tunnels
bances in the boundary layer.
have been developed further and shown to give flight-quality
transition data (Wilkinson 1997). At M> 8, even conventional 3.13 Roughness
tunnels with turbulent tunnel-wall boundary layers display an
While surface roughness can have a profound effect on transi-
essentially quiet behavior as the primary Instability frequency
tion, the mechanisms associated with single roughness ele-
bands tend to be above the tunnel noise spectrum. However if
ments are only partially understood while those responsible
the path to transition involves subharmonic frequencies, then
for transition with distributed roughness are not yet known.
those tunnels may not be sufficiently quiet. Conventional
This has led to a large body of empirical information in the
supersonic and hypersonic tunnels may still be used to study
bterature that is not fully consistent. These correlations are
generally based on 2D parameters such as Re, , k/6 k/6 layers on the tunnel walls. Conventional qwet tnnc,s do
whereas the distribured roughness IS inherently 3D. The three this by maintaining laminnr boundary layers on the tunnel
dimensionality IS introduced info the correlations by providing walls (see Beckwth and Chen 1989). At M > 8. even conven-
separate curves for each 3D shape and distnbutwn. Neverthe- tional tunnels with turbulent rldewalls dnplay some quiet
less, these correlations are still the operative database for behawor as the instability bands tend to be separatedfrom the
dealing with distributed roughness.The Re#/M, correlations tunnel noise spectrum (Fig. 18). However, while the primary
that are so prevalent m reentry vehicle transition work are now instability may be noise free. subharmonics may not be. and
understood to be primarily a reflection of surface roughness so the transition process beyond primary instablhty can be
effects. affected by the tunnel nose. This issue merits further study.
Conventional supersonic tunnels may stdl be used for TS.
The bulk of applicable roughness data for hypersonic vehicles
swept attachment-line and crossflow mstability studies, pro-
comes from wind-tunnel and flight tests of the Shuttle orbiter
vided that it is ascertained that the results are not affected by
vehicle. The roughness on this vehicle is principally from the
tunnel noise. Additional material on this topic is contaIned in
gaps and grooves between the tiles of the thermal protection
Reed et al. (1997).
system (TPS). The prime applicable correlations are due to
Benin et al. (1982). These correlations have been recently
reviewed and revised (Bouslog et al. 1997) The latest ver-
sions show good collapse of the wind-tunnel data The flight
data follow the trends of the wind-tunnel data but are about 1
60% higher. It is suggested that this is due to the reduced
noise level in flight.
How dces one define a quiet tunnel? Seemingly It should be process to the flight environment might be determined.
one where the frequency bands of unstable modes are not
excaed by acoustic radiation from the turbulent boundary
4 Whenever possible, tests should involve more than one downstream boundary condition should he non-reflecting in
facility. Tests should have ranges of overlapping pa- order that spurious feedback phenomena be absent. Herbert
rameters, and whenever possible. experiments should (1994), Reed (1994). and Reed et al. (1998) present critical
have redundancy I trans,tm measurements. discussions of the relevant issues.
Morkovin has suggested strongly that Guideline Number Four At present. these reqwrements are marginally attainable only
be applied also to computational studies. for small portions of simple flows. The calculation of a full
vehicle is still a number of computer generatmns ahead of us.
The above guidelines apply as well to flight experiments Parallelization of the codes and running them in parallel on
Their implementation for flight experiments requires special multiprocessors can enable the implementation of such calcu-
attention to a number of factors: lations. But until the overall capability is available, the esti-
The measurement of disturbance environment must he matron of transition locations will probably have to depend on
incorporated into the model design and in fact must be linear stability properhes of the calculated laminar flows and
part of the model. their synthesis into an eN procedure that properly accounts
for the three-dimensionality of the flow and the disturbances
Attention has to be given to the maintenance and moni- as well as for surface CrYatle effects.
toring of test conditions such as Mach number. Reynolds
4.4 Design considerations for hypersonic flight
number, angle of attack, yaw angle and surface tem-
perature for the duration of the measurementperiod Vehicle design often starts by choosing a configuration based
on inviscid reasoning and then fine tuning that configuration
Attention has to be given to the mamtenance and moni-
to optimize its features. For fbght up to M = 3, this generally
toring of model surface condltmns for each flight. This means reducing form drag, wave drag. and friction drag as
includes protection of the model surface before launch, well as improvmg the low-speed high-hft properties of the
and recovery of the vehicle for inspection and recondi- configuration. Above M = 3, as flight Mach numbers increase,
tioning of the surface prior to the succeeding flight reducing aerodynamic heating loads becomes the primary
consideration Every attempt IS made to minimize the need for
Because stability phenomena at supersonic and hyper-
active cooling. This means giving very detimte attention to
sonic speeds occur at frequencies of hundreds of kHz
delaying transition and taking advantage of passive cooling
and even to MHz levels. there is a need for very high
through radiation from the aerodynamically heated surfaces to
data sampling rates, especially when monitoring multiple
the surroundings. Rarely does this alter the original general
channels. This poses special problems in data acquisition
shapeof the configuration.
and data reduction. Reliable digital telemetering of data
An exception is the recent experience with NASP the U.S.
from the vehicle may also be necessary m order to mini-
National AeroSpace Plane (Eiswirth and Lau 1997). This is a
mize weight and volume of the data acquisition equip-
case where the baseline configuration was axlsymmetric. The
Itlent.
forebody served as the compression surface of the mlet and
4.3 Transition prediction or estimation was subject to adverse pressure gradient, Giirtler instability
and crossflow instability when the body was at angle of at-
One can dream about an eventual capability of computing the
tack. Because of the axial symmetry. the entropy layer became
full flowfield about a vehicle or vehicle component including
successively thinner with distance downstream so that the
the transition process Since transition is the result of unsteady
boundary layer edge conditions were beyond swallowing and
phenomena as discussed m this article, the computation
subject to second and higher mode instabilities. All of these
through transition requires time-accurate solutions of the
factors promote earlier transition. In thn situation, the TS as
compressible, three-dimensional, unsteady Navier-Stokes
well as the GBnler and crossflow instabilities are not amelio-
equations in a domain large enough to contam the complete
rated by surface cooling. By the time of the configuration
phenomenon, and with a grid that is tine enough over the
development, it had already been estabhshed in the NASA-
extended boundary layer to capture the detailed linear eigen-
Langley Quiet Pilot Tunnel that flat plates (2D) had higher
functions and the subsequentnonlmearities. In addition, initial
transition Reynolds numbers than cones at the same
and boundary conditions must be properly specified and en-
freestream conditions (Chen and Mahk 1988). This had also
forced. The initial and upstream boundary conditions should
been verified by eNcalculations (Mack 1987, Elias and
represent properly the physical disturbance environment to
Eiswirth 1990).
which the flow is subjected. Truncation error and round-off
Introduce disturbances whose spectrum is non-physical. The
2-23
This prompted a reconsideration of the geometry of the vehi- Bassom, A.P., Hall, P. 1991 Vortex instabilities in
cle. The forebody was replaced by a wedge-like configuration three-dimensional boundary layers: the relationship be-
tween G6Rler and crossflow vortices. J. Fluid Mech. 232
with rounded comers at its spanwise edges. The lower surface
pp. 647-80.
was again a compression surface but the upper surface had a
slightly favorable pressure gradient. The leading edge of Beckwith, 1.E 1975 Development of a high-Reynolds-number
quiet tunnel for transition research. AIAA J. 13(3) pp. 300.
course had to be blunted for its own protection. The blunting
6.
was sufficient so that the boundary layers on both the lower
and upper surfaces are within the entropy layer where the edge Beckwith, I E. Chen. F.J., Malik. M.R. 1990 Transition re-
search in low-disturbance high-speed wind tunnels, Lomr-
Mach numbers are low enough that first mode TS instabilities,
nor-Turbulenr Transirion III, pp. 227-38. Eds. D. Amal,
which can be controlled by surface cooling, are dominant. The R. Michel. Springer-Verlag.
upper surface could be radiation cooled over the flight range
Beckwith, I.E., Creel. T.R., Jr., Chen, F.I., Kendall. 1.M. 1983
while the lower surface required active cooling through heat
Freestream noise and transition measurements on a cone
exchange with the cryofuel. Crossflow instabilities were in a Mach 3.5 pilot low-disturbance tunnel. NASA TP
confined to the rounded edge regions and did not add greatly 2180.
to the cooling requirements. The transition behavior was
Beckwith, I.E.. Malik, M.R., Chen, F.I., Bushnell. D.M. 1985
verified by eN type calculations using a 3D code known as Effects of nozzle design parameters on the extent of quiet
the eMalk code (Malik 1989, Malik and Balakumar 1992, test flow at Mach 3.5. In Lmninar-Turbulenr Transirion
Schwcerke 1993). Ed. V.V. Kozlov. New York: Springer-Verlag pp. 589.
6W.
This case is instructive for future configuration development
Benin, J.J., Hayden, T.E., Goodrich, W.D. 1982 Shuttle
in that it shows that transition considerations could constmc-
Boundary Layer Transition due to Distributed Roughness
lively alter the basic configurational shapeof a vehicle.
and Surface Cooling. J. Spacecrafl and Rockers. /9(5) pp.
389.96.
ACKNOWLEDGEMENT
Bertolotti, F.P 1991 Compressible boundary layer stability
The authors would like to acknowledge the discussions and analyzed with the PSE equations. AIAA Paper 91.1637.
helpful suggestions of Professor Helen Reed on much of the Blanchard,A.E. 1995 An invesrigation of wall-cooling @em
content of this report. on hypersonic boundary-layer srabiliry in n quier wrnd
runnel. PhD Dissertation, Old Dominion University, Dept.
REFERENCES of Mechanical Engineering, Norfolk VA.
Adam, P.H., Homung, H.G. 1997 Enthalpy effects on hyper- Bouslog, S.A., Benin. J.J., Berry, S.A., Caram, J.M. 1997
velocity boundary layer transition: experiments and free Isolated Roughness Induced Boundary Layer Transition-
flight data. AIAA Paper 97.0764. Shuttle Orbiter Ground Tests and Flight Experience.
AMA Paper 97.0274.
Amal, D. 1992 Boundary-layer transition: Prediction, appli-
cation to drag reduction. AGARD Rep 786 (Special course Bushnell, D.M 1990 Notes on Initial Disturbance Fields for
on skin friction drag reduction). the Transition Problem. In lnsrabiliry and Trans~~ton
Vol I Eds. Hussaini. Voight., Springer-Verlag.
Amal, D. 1994 Predictions based on linear theory. In Progress pp. 217-32.
in Tronsirion Modelling AGARD Rep. 793.
Bushnell, DM., Huffman, J.K. 1967 Investigation of heat
Amal, D., Kufner, E.. Oye, I., Tran. P. 1996 PROGRAMME transfer to leading edge of a 76 swept fin with and with-
TRP TRANSITION Computational results for transition
out chordwise slots and correlations of swept-leading-
prediction. Study Note 7. edge transition data for Mach 2 to 8. NASA-TM-X-1475.
Amal, D.. Vignau, F., and Laburthe, F. 1991 Recent super-
Cattafesta. III, L.N., Iyer, V., Masad, J., King. R., Dagenhart.
sonic transition studies with emphasis on the swept cyl- I. 1995 Three-dimensional boundary-layer transition on a
inde case. Conf: on Boundmy Layer Transition and Con- swept wing at math 3.5. AIAA J. 33 (I I)
trol, Cambridge.
Chang, C.-L., Malik. M., Erlehacher. G., Hussaini, M. 1991
Balakumar. P.. M&k, M.R. 1994 Effect of adverse pressure Compressible stability of growing boundary layers using
gradient and wall cooling on instability of hypersonic parabolized stability equations. AIAA Paper 91-1636.
boundary layers. High Technology Corporation, Hampton,
VA Rep. HTC-9404. Chang. C.-L., Malik, M.R. 1993 Non-parallel stability of
compressible boundary layers. AIAA Paper 93.2912
Chapman, G.T. 1961 Some effects of leading edge sweep on
boundary-layer transition at supersonic speeds NASA-TN-
D-1075
Chen, F. and Malik, M. 1988 Comparison of boundary layer El-Hady, N.M., Verma, A.K. 1983b. G&tler instability of
transition on a cone and flat plate at Mach 3.5. AIAA Pa- compressible boundary layers. AIAA J. 22 pp. 1354-55.
per88.0411.
Elias, T.I., Eiswinh, E.A. 1990 Stability Studies of Planar
Chew F.J., Malik. MR., Beckwith, I.E. 1985 Instabtlittes and Transition in Supersonic Flows. AIAA Paper 90.5233.
transition in the wall boundary layers of low-disturbance
Fisher, D. and Dougherty, N. 1982 Inflight transition meas-
supersonic nozzles. AIAA Paper 85.1573.
urements on a 10 deg. Cone at Mach numbers from 0.5 to
Chen, F-J , Malik, M.R., BeckwIth. I.E. 1989 Comparison of 2. NASA TP 1971.
Boundary Layer Transition on a Cone and Flat Plate at
Fu, Y.B., Hall, P. 1991a Nonlinear development and secon-
Mach 3.5. AIAA J. 27(6) pp. 687.93
dary instability of Giirtler vortices in hypersonic flows.
Creel, T.R , Beckwith 1.E and Chew F.J. 1987 Transition on ICASE Rep. 91-39.
swept leading edges at Mach 3 5. J. Aircraft 25(10)
Fu, Y.B., Hall. P. 1991b Effect of Gi%tler vortices, wall cool-
Cumpsty, N. and Head, M. 1967. The calculation of three- ing and gas dissociation of the Rayleigh Instability in a
dimensional turbulent boundary layers Pan II: attachment hypersonic boundary layer. ICASE Rep. 91.87.
line flow on an infinite swept wing. Avon Qunrr., XVIII,
Fu, Y.B., Hall, P. 1994 Crossflow effects on the growth rate
Part 2.
of inviscid G6irtl.z vortices in a hypersonic bamdaty
Da Costa, I.D. 1990 Contribution d le%de de In rranririon de layer. J. Fluid Mech. 276 343.67.
bord daarraque par contamination en e!coulement hyper-
Fu, Y.B., Hall, P. Blackaby, N. 1993 On the GBrtler instabil-
sonique. Masters thesis, Poitiers University, France.
ity in hypersonic flows: Sutherland law fluids and real gas
Dagenhart, J.R. 1981 Amplified cmssflow disturbances in the effects. Phil. Trans. R. SK. Lo& A 342 325-77.
laminar boundary layer on swept wings with suction.
Gaillard, L. 1993 Elude de lo rransrrion de bord dottnque sur
NASA TP 1902.
un cylindre en j7lche en koulemenr hypersonique. Mas-
DalImann, U., Hein, S., Koch, W., Bertolotti, F.. Simen. M , ters thesis, Poitiers University, France.
Stoke. A., Gardner, A., and Nies, 1. 1996 Status of the
Gasperas,G. 1990 Stability of the laminar boundary layer for
theoretical work within DLRs non empirical transition
an imperfect gas. Lombtar-Turbulent Transition II. Eds.
prediction project 2 European forum on Lominor Flow
D. Amal, R. Michel. Berlin: Springer-Verlag. pp. 291.
Technology, Bordeaux.
302.
Dando, A. 1992 The inviscid compressible GBrtler problem in
Gaster, M. 1967 On the flow along leading edges. Aeron.
three-dimensional boundary layers. Theo&. Comput.
Qum, XVIII Part 2.
Fluid Dyn. 3 pp. 253-65.
Gaster, M. 1974 On the effects of boundary-layer growth on
Dando. A., Seddougui, S.O. 1991 The inviscid compressible
flow stability. J. Fluid Me& 66 pp. 465-80.
Gdrtler problem. ICASE Rep. 91.54.
Germain, P., Cummings, E. and Homung, H. 1993 Transition
de Luca, L.. Cardone, D. Aymer de la Chevaletie, D. Fon-
on a sharp cone at high enthalpy; new measurements in
teneau, A. 1993 GOttIer instability of a hypersonic bound-
the Shock Tunnel T5 at GALCIT. ALU Paper 93-0343.
ary layer. Exp. Fluids 6 10.6.
GBriler, H. 1955 Dreidimensionale InstahilitiIt der ebenen
Doggett, G.P., Chokani, N., Wilkinson, S.P. 1997 Effect of
Staupunktstriimung gegeniiber wirbelartigen StBrungen. In
angle of attack on hypersonic boundary-layer stability.
Fifty Years of Boundary layer Research. Eds. GBrtler and
AIAA J. 3X3) pp. 464-70.
Tollmien. Vieweg and Sohn. Braunschweig.
Dougherty, N.S , Fisher, D.F. 1980 Boundary layer transition
Hall, P. 1985 The GBrtler vortex instability mechanism in
on a IO-degree cone: Wind tunnel/flight data correlation.
three-dimensional layers. Proc. R. Sac. London Ser. A 399
AMA Paper 80-0154.
pp. 135-52.
Eiswirth, E., La, K 1997 Boundary Layer Transition on
Hall, P., Fu. Y.B. 1989 On the Gbrtler vortex instability
Hypersonic Configurations-Lessons Learned from NASP.
mechanism at hypersonic speeds. Theoret. Compur. Fluid
Privnre communication to E. Reshotko.
Dyn. I pp. 125.34.
Eiswirth. E., Lau, K 1997 Boundary Layer Transition on
Hall, P., Malik, M.R. 1989 The growth of G(irtler vortices in
Hypersonic Configurations -Lessons Learned from NASP.
compressible boundary layers. J. Eng. Math. 23
AIAA Paper 97.1909.
pp. 239.45.
El-Hady. N.M. 1991 Nonparallel instability of supersomc and
Hall, P., Malik, M.R., Poll, DIA. 1984 On the stability of an
hypersonic boundary layers. Phys. Fluids A 3(9) pp. 2164.
Infinite swept attachment-line boundary layer. Proc. Roy.
El-Hady. N.M., Verma, A.K. 1983a Gworth of Giirtler Sot. Land. A 395 pp 229.45.
vortices in compressible boundary layers along curved
surfaces. J. Eng. Appl. Sri 2 pp. 213.38.
2-25
Htimmerlin, G. 1955 Zur lnstahilitlstheorie der ebenen Stau- Kendall, I.M. 1975 Wind tunnel experiments relating to so-
punktstr0mung. In Fij?y Years of Boundary Layer Re- personic and hypersonic boundary-layer transition. AIAA
search. Eds. G&tler and Tollmien. Vieweg and Soho, J. 13(3) pp. 290-9.
Brauschweig.
Kimmel R.L.. Demetriades, A., Donaldson. 1. 1996 Space-
Htimmerlin. G. 1961 Uber die StabilitU einer kompressiblen time correlation measurementsin a hypersonic transitional
StOrmung kings einer kookaven Wand bei verschiedenen boundary layer. AIAA J. 34( 12) pp. 2484-9.
Wand-temperaturvehsltnissen. Deutsche Versucbsanstalt
Kimmel, R.L., Poggie, J. 1997 Disturbance evolution and
fur Luftfahrr, Bericht 175.
breakdown to turbulence in a hypersomc boundary layer,
Haniti, A. 1995 Nonlocal stability analysis of the compressi- ensemble-averagedstructure. AIAA Paper 97.0555.
ble boundary layer on a yawed cone. Ph.D. diss., KTH,
King, R.A. 1991 Mach 3 5 boundary layer transitloo on a cone
Stockholm.
at angle of attack. AIAA Pope, 91.1804.
Hanifi,A., Henningson. D., Hein, S., Bertolotti, F , and Simen,
Klebanoff, P.S., Tidstrom. K.D.. Sargent, L.M. 1962
M. 1994 Linear nonlocal instability analysis the linear
Three-dimensional nature of boundary layer instability. J
NOLOT code. FFA Rep. 1994-54.
Flurd Mech. 12 pp. l-34.
Herbert, Th. 1988 Secondary instability of boundary layers.
Kohayashl, R., Kohama, Y. 1977 Taylor-Gtirtler instability of
Ann. Rev. Fluid Mech. 20 pp. 487.526.
compressible boundary layers. AlAA J. I5 pp. 1723.7.
Herbert, Th. 1994 Paraholized stability equations. In Progress
Krogmann, P. 1977 An experimental study of boundary layer
in Transition Modelling. AGARD-FDP VKI Course. Rep.
transition on a slender cone at Mach 5 AGARD-CP-224.
793. Ed. W. S. Saric
Krogmann, P. 1996 Privore communication to W.S Soric, 26
Herbert. Th, Esfahanian, V. I992 Stabdity of hypersonic flow
Nov. 1996.
over a blunt body. AGARD CP 514 Theorelical and Ex-
penmental Methods in Hypersonic Flows. Lachowitz, IT.. Chokani, N.. Wilkinson, S.P. 1996 Bound-
ary-layer stability measurements in a hypersonic quiet
Holden, M.S., Bower, D.R., Chadwick, K.M. 1995 Measure-
tunnel. AIAA J. 34(12) pp. 2496-500
ments of boundary layer transition on cones at angle of
attack for Mach numbers from I I to 13. AIAA Paper No. Lees, L., Lin, C.C. 1946 Investigation of the stability of the
95.2294. laminar boundary layer in a compressible fluid. NACA TN
111s.
Holden, MS., Chadwick, K.M. 1995 Studies of laminar,
transitional, and turbulent hypersonic flows over curved Mack, L M. l965a Computation of the stability of the laminar
compression surfaces.AIAA Paper 95.0093. compressible boundary layer. In Merhods in Camp. Phys.
Vol. 4 pp. 247.99. Ed. B. Alder. Academic Press. New
Holden, M.S., Kelly. I.M. 1995 Attachment-line transition
York.
studies on swept cylindrical leading edges at Mach mm-
hers from 10.12. AlAA Paper 95.2279. Mack. L M. 1965b The stability of the compressible laminar
boundary layer according to a direct numerical solution
Holden, M.S., Kelly. J.M., Bower, D. 1994 Attachment-line
AGARDograph 97, pt. I pp. 329-62
transition studies on swept cylindrical leading edges at
Mach numbers from 10.12. CALSPAN Tech. Rep. 26106. Mack, L.M. 1969 Boundary layer stability theory JPL Rep.
900.277. Rev. A.
Howarth. L. 1948 Concerning the effect of compressibility on
laminar boundary layers and their separation. Proceedings Mack, L.M 1975 Linear stability theory and the problem of
of the Royal Society of London A. Vol. 194. supersonic boundary-layer transition. AIAA J. 3 pp.
278.89
Illingworth, CR. Steady flow in the laminar boundary layer of
a gas Proc. R. Sot. London Ser. A Vo. 199. Mack, L.M 1984 Boundary-layer linear stability theory.
AGARD Rep. 709 (Special coarse on stability and transi-
lallade, S.. Amal. D., Ha Mirth, H 1990 Theoretical study of
tioo of laminar flows).
GBrtler vortices; linear stability approach. In Laminor-
Turbulent Tronrition. Eds. D Amal, R. Michel. Berlin: Mack. L.M. 1987 Stability of Axisymmetric Boundary Layers
Springer-Verlag. pp. 563-72. on sharp Cones at Hypersonic Mach Numbers. AIAA Pa-
per 87.1413.
Johnson, C.. Stainback, P., Wiker, K., and Bony, L. 1972
Boundary layer edge conditions and transition Reynolds Mack, L.M 1990a On the inviscid acoustic-mode instability
number data for a flight test at Mach 20 (Reentry-F). of supersonic shear flows, Part I: two-dimensional waves.
NASA TM X2584. J. Theor. and Camp Fluid Dyn. 2 pp. 97-123.
Johnson, H.B., Candler, G V., Hudson. M L. 1997 Numerical Mack, L.M. 1990b Private communicarron to E. Reshotko.
Study of Hypersonic Boundary Layer Transition on a
Malik, M. 1990a Numerical methods for hypersonic boundary
Blunt Body. AIAA Paper 97-0554.
layer stability J. Camp. Phys. 86 pp.376.423.
Malik, M., Balakumar. P. 1992 Instability and transition in Pfenninger. W. 1965 Some results from the X-21 program.
threedimensional supersonic boundary layers. AlAA Paper Part 1. Flow phenomenon at the leading edge of swept
92.5049 wings. AGARDograph No. 97.
Malik, M.R. 1984 Instability and transition in supersonic Pfenninger, W. 1977 Laminar Flow control Laminanzatmn..
boundary layers. In Laminar Turbulent Boundary Layers. AGARD Rep. R-654. (Special course: concepts ,n drag re-
Energy Sources Technology Conference, New Orleans. duction).
Malik, M.R. 1988 Stability theory for NASP transition pre- Pfenninger, W and Bacon, Jr, J.W. 1969. Amplified laminar
diction. F$fh Notional Aero-Space Plane Symposium, Pa- boundary layer oscillations and transition at the front at-
per 57. tachment line of a 45 deg. swept flat-nosed wing with and
without boundary-layer suction. In Viscous Drag Reduc-
Malik, MR. l989a Prediction and control of transition in
tion. Ed. C. Wells. Plenum Press.
supersonic and hypersonic boundary layers. AIAA J.
27(1 I) pp. 1487-93. Poggie. I., Kimmel. R.L. 1997 Disturbance evolution and
breakdown tu turbulence in a hypersonic boundary layer:
Malik, M.R. l989b el : A New Spatial Stability Analysis
instantaneous structure. ALU Paper 97.0556.
Program for Transition Prediction Using the e Method.
High Technology Rep. HTC-8902 Poll, D. 1978 Some aspects of the flow near a swept attach-
ment line with particular reference to boundary layer tran-
Malik, M.R., Beckwith, I.E. 1988 Stability of a supersonic
sition CranJield College of Aeronautics Tech. Rep. 7805.
boundary layer along a swept leading edge. AGARD
CP 438. Poll, D. 1985 Boundary layer transtion on the windward face
of space shuttle during reentry. AIAA Paper 85.0899
Malik, MR., Spall, R.E., Chang, C.L. 1990a Effect of use
bluntness on boundary layer stability and transition. AIAA Poll. D.I.A. 1984 Transition Description and Prediction in
Paper 90-0112. Three-Dimensional Flows, AGARD Rep. 709. (Special
course: stability and transition of laminar flows).
Malik, M.R., Zang, T.. Bushnell, D.M. l990b Boundary layer
transitmn in hypersonic flows AIAA Paper 90.5232. Poll, D.I.A. 1985 Some observations of the transition process
on the windward face of a long yawed cylinder. J Fluid
Morkovin, M.V. 1969 Critical evaluation of transition from
Mech. 150.
laminar to turbulent shear layers with emphasis on hyper-
sonically traveling bodies. Air Force Nigh: Dynamics Poll, DSA. 1986 A new hypothesis for transition on the
Laborarory Rep. AFFDL-TR-68-149. windward face of spaceshuttle. J Space. Rockets, 23(6).
Morkovin, M.V 1969 On the many faces of transition. Potter, J.L. 1974 The unit Reynolds number effecr on bound-
Viscous Drag Reduction. Ed. C.S. Wells. Plenum. ary layer transition. Ph.D. Dissert. Vanderbilt Univ.
Morkovin, M.V. 1987 Transition at hypersonic speeds.ICASE Potter, I.L. 1980 Review of the mfluence of cooled walls an
Inkrim Report I (NASA Contractor Rep. 178315). boundary-layer transition. AIAA .I. 18(g) pp. 1010-Z.
Morkovin, M.V. 1988 Recent insights into instability and Radeztsky, R.H. Jr., Reibert, M.S., Saric, W.S., Takagi, S.
transition to turbulence in open-flow systems. AIM Paper 1993 Effect of micron-sized roughness on transition in
88.3675. Also ICASE Report No. 88.44, (NASA Conrrac- swept-wing flows. AIAA Paper 93-0076.
mr Report 181693).
Reed, H.L. 1994 Direct Numerical Simulation of Transition:
Morkovin, M.V., Reshotko, E. 1990 Dialogue on progress and The Spatial Approach. In Progress in Trmsirion Model-
issues in stability and transition research. Lamimzr- ling. AGARD-FDP VKI Course. Rep. 793. Ed. W.S. Sark.
Turbulent Transition 111.Eds. D. Amal. R. Michel. pp.
Reed, H.L., Haynes, T.S. I994 Transition correlations in
3-30. Berlin: Springer-Verlag.
three-dimensional boundary layers. AL44 J. 32(S) pp.
Momsette, E. 1976 Roughness induced transition criteria for 923.9.
space shuttle-type vehicles. J. Aircrofl, 13(2).
Reed, H.L., Haynes, T.S., Saric. W.S. 1998 Computational
Murakami, A., Stanewsky, E., Krogmann, P. 1995 Boundary Fluid Dynamics Validation Issues in Transition Mcdeling.
layer transition on swept cyhnders at hypersonic speeds. AIAA J. 36(5) pp. 742-51.
AMA Paper 95.2276.
Reed, H.L., Kimmel, R.L.. Schneider. S.P., Amal. D. 1997
Pate, S.R. 1978 Dominance of radiated aerodynamic noise on Drag prediction and transition in hypersonic flow. AIAA
boundary-layer transition in supersonic-hypersonic wmd Paper 97.1818 (also AGARD-CP-600, Vol. 3 pp. Cl5-l-
tunnels: theory and application AEDC-TR-77-107. 17).
Pate, S.R., Schueler, C.J. 1969 Radiated aerodynamics noise Reed, H.L., Saric W.S. 1989 Stability of three-dimensional
effects on boundary-layer transition in supersonic and hy- boundary layers. Ann. Rev. Fluid Mech 21 pp. 235-84.
personic wind tunnels. AIAA J. 7 pp. 450-7.
2-27
Reed, H.L.. Saric, W S , Amal, D. 1996 Linear slablbty the- Stetson, K F, KImmel, R.L., Thompson, E R , Donaldson,
ory applied to boundary layers Annu. Rev, F&id Mech 28 J.C.. Siler. LG., 1991 A Comparison of planar and coot-
pp. 389.428. cal boundary layer stability and transition at a Mach num-
ber of 8 AIAA Paper 91-1639.
Reshotko, E, Saric, W.S., Nagib, t1.M. 1997 Flow Quality
Issues for Large Wind Tunnels. AIAA Paper 97.0225. Stetson. K.F., Thompson, E.R., Donaldson. J.C., Siler, L.C.
1984 Laminar boundary layer stability experiments on a
Reshotko, E. 1969 Stability Theory as a Guide to the Evalua-
cone at Mach 8, Part 2: blunt cone. AIAA Paper 84XXIC.5
tion of Transition Data. A/AA J 7(6) pp. 1086.91,
Stewartson, K. Correlated Compressible and Incompressible
Reshotko, E 1986 Stability and Transition. what do we
boundary Layers. Proc R. Sot. Lmzdon Ser. A Vol 2OiJ
know? Proc. U.S. Nar;l. Congress Appl. Mech. Austin,
pp. 84-100.
TX. ASME, 421.
Stuart, J.T. 1984 Instability of laminar flows, nonlinear
Reshotko, E. 1994 Boundary layer instability, transition, and
growth of fluctuations and transition to turbulence. In
control. ALU Paper. 94.0001
Turbulence and Chaotic Phenomena in F1uid.r. Ed. T.
Reshotko, E. Khan, M.M.S. 1980 Stability of the laminar Tatsumi. Amsterdam: North-Holland pp. 17.26
boundary layer on a blunted plate in supersonic flow. In
Stuckert, G K. 1991 Linear stnbihry theory of hypersonic,
Laminnr Turbulent Transition I, Springer-Verlag.
chemrcally reacting wcous flows. Ph.D. Dissefl.. Ariz.
Reshotko, E , Beckwith. I. 1958 Compressible laminar bound- State Univ., Tempe.
ary layer over a yawed infinite cylinder with heat transfer
Stuckert. G K , Reed, H.L. 1994 Linear disturbances in hyper-
and arbdrary Prandtl number. NACA Tech. Rep. 1379.
sonic. chemically reacting shock layers. AIAA J. 32(7)
Saric. W.S. 1994a Physical description of boundary-layer pp 1384.93.
transition: experimental evidence. lo Progress in Transr-
Tran. P., Seraudie, A., Wendt, V.. and Poll, D I.A. 1995
fmn Modelling. AGARD-FDP VKI Course. Rep. 793. Ed.
PROGRAMME TRP TRANSITION Experimental results
W.S. Sari,
for transition prediction. Study Note 6
Sark, W.S. l994b GBrtler vortices. Ann. Rev. Fluid Mech. 26
Tumin. A.M , Chernov, Yu. P. 1988 Asymptotic analysis of
pp 379.409.
flow instability in a compressible boundary layer on
Schneider, S.P.. Haven. C.E. 1995 Quiet flow Ludwieg tube curved surface. Ah. Prikl Mekh Tekh. Fiz. 3 pp 84-9.
for high-speed transition research. AIAA J 33(4) pp. 688.
Van Ingen, J.L. 1956 A suggested semi-empirical method for
93.
the calculation of the boundary-layer transition region.
Schwoerke, S. 1993 Quiet Tunnel Results and Analysis for Rep. No VTH 71 and 74, Dept. Aero. Eng , Tech. Univ
Various Forebody Shapes at Mach 3.5. Narronal Aero- Delft, Neth.
Space Plane Technology Review, Paper No. 158.
Wendt, V , Kreplin, H., Hiihler, G., Grosche, F., Krogmann.
Simen, M., Dallmann, U. 1992 On the instability of hyper- P., Simen. M. 1993 Planar and conical boundary layer
sonic flow past a pomted cone Comparison of theoretical stability experiments at Mach 5. A/AA Paper 93.5112.
and experimental results at Mach 8. DLR-FB 92.02.
Wendt, V Simen. M.. Hanifi, A. 1995 An experimental and
AGARD CP 514 Theorerrcal and E.rperimenral Methods
theoretical mvestigation of instabilities in hypersonic flat
m Hypm-sonic Flows.
plate boundary layer flow, Pbys. Fluids 7(4) pp. X77-87.
Skuratov, A, Fedorov, A. 1991 The lammar-turbulent transi-
Wilkmson. S. 1997 A Review of Hypersonic Boundary Layer
tion past roughness at the attachment line of a swept cyl-
Stability Experiments in a Quiet Mach 6 Wind Tunnel.
inder m supersonic flow. lo AN SSSR, MewI. Zhidk I
AIAA Paper 97-1819.
Gum (tmnslated in englirh). (6) pp. 2835.
Wright. R.L., Zoby, E.V. 1977 Flight boundary layer tram-
Smith, AM.O., Gamberoni, N. 1956 Transition, pressure
don measurements on a slender cone at Mach 20 AIAA
gradient and stability theory. DAC Report No. ES 26388.
Paper 77.719.
Spalart, P.R. 1989 Direct numerical study of leading-edge
Zhong. X. 1997 Direct Numerical Simulation of Hypersonic
contamination. AGARD-CP438.
Boundary-Layer Transition over Blunt Leading Edges
Spall. R.E., Malik. M R. 1989 Gdrtler vortices in supersonic Part II: Receptivity to Sound. AIAA Paper 97-0756.
and hypersonic boundary layers. Phys. Fluids A I
Zhong. X. 1998 Direct Numerical Simulation of 3-D Hyper-
pp. 1822-35.
sonic boundary Layer Receptivity to Freestream Distur-
Stetson, K.F. 1982 Mach 6 experjments of transition on a cone bances AIAA Paper 98.0533.
at angle of attack. J. Spacecraft Rockets 19 (5) pp. 397.
403