Thomas A. Zang. M. Yousuff Hussaini. Contract. No. NAS1- ..... to acknowledge helpful discussions with Thomas. Corke, Thorwald. Herbert and Y. Kachanov. 5 ...
NASA ICASE
Contractor Report
Report No.
181948
89-78
ICASE MULTIPLE PATHS TO SUBHARMONIC LAMINAR BREAKDOWN IN A BOUNDARY LAYER
Thomas A. Zang M. Yousuff Hussaini
Contract
No.
October
1989
Institute NASA Hampton, Operated
NAS1-18605
for Computer Langley
Applications
Research
Virginia
in Science
and Engineering
Center
23665-5225
by the Universities
Space
Research
Association N90-13722
(NAeA-CR-181948) MULTIPLE PATHS TO SU6HARMONIC LAMINAR B_EAKOOWN IN A BOUN_ARY LAYER Final Report (ICASF) 15 p CSCL 2OD
G3/34
National Aeronaulics and Space Adminislration Langley Research Center Hampton, Virginia 23665-5225
unclas 0251708
MULTIPLE
PATHS
BREAKDOWN
TO
SUBHARMONIC
IN
A BOUNDARY
Thomas
LAMINAR LAYER
A. Zang
Computational
Methods
Branch
NASA
Research
Center
Langley Hampton,
VA 23665 and
M. Yousuff
Hussaini
_
ICASE NASA
Langley Hampton,
Research
Center
VA 23665
ABSTRACT
Numerical duced
by the
dimensional shear
simulations
layer
secondary subharmonic
demonstrate instability disturbances
that
laminar
of two-dimensional need
not
take
breakdown
in a boundary
Tollmien-Schlichting the
conventional
layer
waves lambda
in-
to three-
vortex/high-
path.
1This research was supported by the National Aeronautics and Space Administration under NASA Contract No. NASl-18605 while the second author was in residence at the Institute for Computer Applications in Science and Engineering (ICASE), NASA Langley Research Center, Hampton, VA 23665. J_
The classicalboundary-layer transitionexperiments of the early 1960%
[I],[2],[3]led
to the perception that the laminar-turbulent transitionprocess in a B1asius boundary layer may
be
called
viewed
as a sequence
the
primary
understood.
The
and
the
of instabilities.
secondary
primary
The
instabilities,
instability
first
two
are linear
in an incompressible
instabilities at onset
flat-plate
in the
and
sequence,
are now very
boundary
layer
well
is known
as a Tollmien-Schlichting (TS) wave. The TS wave, as it equilibrates nonlinearly, becomes susceptible to what is known as the secondary instability. This is truly three-dimensional in character
and
essentially
of three
instability, the
is thoroughly
and
secondary
on the
vortices
fundamental tertiary
vortex
the
much
secondary
wave
tices
which
primary
later
stabilities
The
detuned waves
To date,
the
shorter
longer
wavelengths
Numerical shear
layers
instability which
and
address
This
instability
of the
stage
suffers
peak
is possibly
imminent.
classical
a
in the
this stage
what
becomes
secondary
so-called
with
through
formation
system
the
Also associated appear
and
evolution
layers.
all of the
that
layers
are
and
This
experiments,
streamwise
the resulting
involves
to yield
their associated detuned
with
spikes
the
of the
of lambda
combination
vor-
modes,
the
wavenumber
have
not been
observed
Instead,
these
of the spectrum,
formation
exper-
wavelength
breakdowns.
broadening
observed
pattern
so-called
combine
by a rapid
associated
the
was
of "spikes",
but
of the
not
in-
just
to
also to even
[7].
of the
of subharmonic details
and
or the
of boundary-layer
spikes
case,
wave, instability
subharmonic
wavelengths
the
vortex
and an associated
dominated
wavenumbers
to breakdown
simulations
stage
primary
in which
nonlinear
are
secondary
the
a streamwise
The
spikes
both
ensues
a "spike".
subharmonic
secondary whose
the
to lead
streamwise
and
layer
spot
to have
of the primary
high-shear
for either appear
called
turbulent
case,
high-shear
multiple
a subharmonic
In the
that
of secondary
wave.
and
from
[5]-[6].
is double
experimentally
usually
is now known
arising
is staggered.
i.e., a pair
to detached
Thereafter,
the subharmonic
splitting") direction.
instabilities
breakdown.
breakdown
imentally
leads
secondary
first
wavelength,
streamwise
oscillogram,
K-type
In the
by a kink in the high-shear
of instabilities,
scenario,
is called The
instability
elements.
sequence
streamwise
known
instability,
instability.
(a "peak-valley
in the
manifested
perturbation
particular
same wave
aligned
secondary
hairpin
a rapid
the
[4]. The
secondary
secondary
primary
are
instability
streamwise are
have
of the
by Herbert
- the fundamental
detuned
wave
scale
lambda
kinds
the
reviewed
K-type
transition
of the
the characteristic
[s],[g],[10],[ii] and the
breakdown
transition
have produced
[12]. In this paper
laminar
breakdown
of boundary-layer
transition
we report
induced
by
early
high-
secondary
numerical
simulations
subharmonic
secondary
instability. Numerical
simulations
and
Hussaini
[9]. In the
the
evolution
is tracked
point not
z0 and necessarily
present in time
the approximation in t),
the
base
simulations rather is that flow
have
the parallel
than
in space.
the displacement is strictly
in the
been
reviewed
flow approximation The
focus
thickness streamwise
is on the is constant direction
recently
by Zang
is employed vicinity
and
of some
in x (although and
it is given
by ub(y,t) = (us(y%/'i'_), 0, 0), where us(r/) is the Blasius velocity profile which follows from the similar boundary-layer equations, and X = zo + eft; c! is the speed of a moving computational
frame.
Lengths
are
scaled
by
the
displacement
thickness,
6_, at z0,
and
velocitiesby the free-streamvelocity, U_.
In these
Although
to be parallel,
the
boundary
layer
[12], with the speed of the The numerical simulations the
algorithm
described
is assumed
units,
X/zo
it is permitted
computational frame, c1, equal reported here were performed in [13], with
the
nonlinear
= 1 + R_--;cft, where to thicken
d = 1.72. in time
to the phase speed of the 2D wave. with the boundary-layer version of
terms
treated
in skew-symmetric
form
[14]. If a and rections, L, and
fl denote
the
respectively,
= 27r/a
and
representations
the
L_ = 2_r/19.
s_ are integers
direction. al numbers
fundamental
then which
(In the cases k, = _,/s_ with
wavenumbers
fundamental
The
specify
imposed the
in the
wavelengths periodicity
number
streamwise
in these lengths
of subharmonics
are
and
directions s_L,
that
are
and
spanwise
di-
are
given
by
where
s_
s,L_,
permitted
in each
presented in this paper s, = 1 and s, = 1, 2, or 9.) The rationand /c, = _/s, label the Fourier wavenumbers in the numerical
respect
to the
fundamental
wavenumbers
a and
t9.
Case KKVF KKVS
Re 1100
wave TS
a 0.250
/3 0.000
1100
TS
0.250
1100
TS
0.250
0.209 0.000
.08624
+ .00333i
0.018
0°
2
- .01946i
0.008 36 °
2
36 °
2
36 °
9
0.209 0.000
KLSL
730 730
SQ TS
0.127 0.254
0.241 0.000
KLD
730 730
SQ TS
0.127 0.254
0.241 0.000
730
SQ
0.113
0.241
Table
.03385
.09292 + .00160i .03925 - .02254i .09292
where
U2D(y)
Sommerfeld
simulations
normalized
so that
the
velocity.
oblique
the
least
The
parameters
in mind.
numerical
recommended
of Kovasznay, corresponding amplitudes shows wave cases has
the
two
modes
of the
a and
amplitudes
the
points
both normal
Zang
phase
ft.
relevant These
are 1, and,
shift
and actual
types shear,
undergone
whether
between
equation
when
the
(Orr-
eigenfunctions used
are
in the form
e3D measure
relative
the to the
two-dimensional
and
the
with
(However,
initial
wave
with
for
resolution
to meet
required
grids
on the
symmetry
particular
component
w, £s included
sufficient spanwise
were chosen
the
order
was
is a TS
reference.
The
criterion
of 192 × 96 ×
exploited
to halve
direction.)
were
performed
Vasudeva laminar
and
detached
symmetry
and after 4_ periods layer is present and (The
figure
KKVS
of the K-type simulations with
1 demonstrates
strong
spanwise
conditions
respectively,
Figure exhibit
in the
the
KKVF
breakdowns,
of breakdown
two roll-ups.
under
[2]. The
experiment.
Ou/Oy,
1. They
frequency,
[15]; this typically stage.
subharmonic
in Table
temporal performed
in that
simulations
to the
indicates
for the KKVF simulation, a strong detached high-shear already
e_[(o,ls®)_-_z] },
linear
are given
The were
and
Komoda, and
table
breakdown
of grid
first
fundamental
1
at x = 0 and z = 0. Thus, e2D and of the 2-D and oblique disturbances
8 measures
wave.
simulations
laminar
number The
The
(SQ)
by Krist
256 by the the
form
wavenumbers
streamwise
of five simulations
or a Squire
direct
1.78 x 10 -5
waves.
experiments wave
angle
- .02082i
(1)
stable
prescribed
maximum
The
0.00622
Simulations
+
given in Eq. (1), these maxima occur maxima of the streamwise fluctuations freestream
4.5 x 10 -s
+
for the
their
- .02254i + .00160i
are of the
.+
U_D(y ) are
and
or Squire)
0.00622
.03630
i: Nonlinear
0.00622 1.78 x 10 -5
+ .00160i
.03925 .09292
u(x,0) = 1
1
0.018 0.008
0.125 0.254
for the
8 x
+ .00333i - .01263i
SQ TS
conditions
0
.08624 .09396
ii00 730
initial
C
0°
KLSH
The
W
that high-shear
plane
after
experiment addressed
the former
the
simulation
at these
high
initial
layers.
The
figure
3_ periods
of the
2-D
for the KKVS simulation. In both by the time of the illustrations each
corresponds
to the two-spike
stage.)
Three-
dimensionalflow-field visualizations of both casesreveal that the transition is associated with the formation of a large-scalelambda vortex/detached high-shearlayer for which the breakdownoccurs first in the symmetry plane and then spreadsoutwards in the spanwise direction. (The vortex lines for the KKVS caseshownthe left frame of Figure 2, for example, indicate the presenceof a strong large-scalelambda vortex.) The subharmonicinstability evolvesmore slowly in this case,as is typical of secondary instabilities at large amplitudes of the primary wave. Nevertheless,the KKVS simulation doesfurnish unmistakeableevidencethat strong detachedhigh-shearlayers are not exclusively the province of K-type breakdowns. Continuation of these two simulations to later times showsthat the strong similarity of the breakdownsremains,except,of course,that in the former casethe patterns are aligned in the streamwisedirection, whereasin the latter casethey are staggeredwith twice the streamwisewavelength. We shall refer to this type of breakdown,which in this caseis commonto both the fundamental and subharmonicinstabilities, as a primary-scalebreakdown in deferenceto the scaleof the vortical patterns (relative to the scaleof the primary wave)which ultimately roll-up. The next two simulations were under the conditions of the Kachanov and Levchenko [6] experiments, which focusedon subharmonic instabilities. The initial amplitudes here, particularly for the three-dimensionalcomponents,were much smaller than those for the other
experiment.
until
after
KKVS KLSL
3-D
the
cases
Indeed, primary
the
cases
the amplitudes
wave
had
breakdowns
differed
only
passed
occurred
in the
Figure 3 illustrates the components for-t}le-_e
were
so low that
Branch
II of the
well before
amplitude
time
immediate
down
in these
down
in the
contains
only vortex
smaller
scale
breakdown
but
detached it originates
plane familiar
or "peak"
plane
and
spreads
places,
scale
the
roll-up,
such
as the
breakdown,
than
KLSH
at many
the
primary
simulation KLSL although
and
KLSH
and
in the
4.
right
The
frame
KLSL
in the
does
in the which
near-wall
such
this
much
with
some
in the
sym-
subharmonic
This
in detached
break
that
commences
region.
eventually
break
visualizations,
direction.
places
not
plane
develop
which
spanwise
in some
does
break-
symmetry
2, reveal
5. Thus,
of a breakdown
plane,
case
are associated
shear
in Figure
laminar
The
of Figure
These
strong
locations,
ultimate
previously.
plane.
as shown
the vortices
of the
wave
breakdown shear
We shall
down
in the layers refer
to
are on a much
wave.
exhibits
case.
KKVF
The
energy [9] in the 2-D and II is reached at t = 675,
three-dimensional
symmetry
outward
symmetry
since
and
paradigm
spanwise
kinetic Branch
discussed
layer,
wall,
(In the
perturbation.
character
Eventually,
the
classical
occurring
the
layers.
near the
diffuse,
than
shear
follow
in other
The
from
not
it as a small-scale smaller
shear
case shown
away
but does
symmetry,
KLSL
strong
3-D
in Figure
two simulations
detached
form
is the
is documented
as the
weak
vortices
plane,
is more
a very
initial
did not occur
primary wave. However, the primary from the subharmonic waves [16].
for this paper
This
manner
lines for the
moderately metry
two cases. same
as the
interest
curve.
II was reached.)
evolution of the normalized simulationsl On tl_is scale
two
breakdown
neutral
Branch
of the
as evidenced by the subsequent decay of the eventually resumes growth because of feedback Of more
laminar
The
smaller
a stronger shear
layer
scale
features
detached of the form
high-shear
latter off the
case
does
symmetry
layer
in the
eventually plane
symmetry exhibit
as well.
This
the is
an intermediate casebetweenthe distinct primary-scalebreakdownof the KKVS simulation and the small-scalebreakdownof the KLSL run. The principal factor determining the type of laminar breakdownwhich occursappearsto be the amplitudeof the primary waveat the stageof incipient three-dimensionality, i.e., when the
three-dimensional
component
reaches
a level
comparable
to that
the primary amplitude is large at incipient three-dimensionality, occurs; when it is small, a small-scale breakdown ensues; and level, the additional
breakdown numerical
boundary-layer
thickness
a primary-scale amplitude chosen
has a mixture simulations.
one.
held Here
fixed
the
for a primary-scale for the
breakdown
KLSH
To date
there
have
been
breakdowns.
as they
in the
were
the
wave
and
actual
shear
experiment
focusing
observations
behavior
When
then
reaches larger
simulation
of the
primary
KKVF
to have
amplitude is associated TS
wave
KKVS
simulation. done
past
layers
in the
been
is
primary-scale
high-shear
to identify
to the
appear
a suffÉcient
a clean
of detached
by two with the
is manifestly
KLSL
be as easy
[2] corresponding
on subharmonic
and
if an even
stability
would
breakdown
unstable
of the
eventual
layers
the
Moreover,
breakdown
the
then
e4 = 1.78 x 10 -4,
no experimental Such
value,
to occur.
primary.
This speculation is supported simulation is repeated, but
is always
with
small-scale
amplitudes
subharmonic no experiments
primary
for example,
Thus
initial
at its initial
breakdown
case,
develops.
with its small Branch II.
of both features. When the KLSL
of the
a primary-scale breakdown when it is at an intermediate
for case
However, at such
large
amplitudes. At lower ing to even KLD, wise
longer
allowed
scales
mode rapid
Note
This
type
is quickly
The breakdown although
reveal
that
however,
that
in a boundary features
of both
The authors are pleased Herbert and Y. Kachanov.
both
final simulation, for 9 stream-
The can
to illustrate basic
study
two
that types
in a given
helpful
5
has of the
for the
detuned there
are
primary
is indicative
the
(4/9,1)
initial
and
conditions,
together.
to a primary-scale which
amplitudes
be present
to acknowledge
leads those
of the
of the
allow not
Notice
also
initial
been
Flow
breakdown.
for only one yet
of the the
3 _, s "7, 13 andS. as _,
harmonics
parameter low
both
evolve
broaden-
wavenumber
in the
eventually
than
is also possible aimed
is included
simulation
complete
layer. types
The domain
streamwise
involves
difference
expensive
at sufficiently
here
the
which
modes
and
particular
more
breakdown
reported
interactions.
computational
setting
the former
sum
so a more
a small-scale simulations
and
this
by
instability
only
other
is much
harmonics,
suspect,
component
generated
in the
by spectral
wavenumber to be 4 the the streamwise 6f various harmonics. This simulation
although
of such
and
breakdown
difference
room
wave
secondary
that
of simulation
streamwise We
primary
[6], [7] suggest
sum and
by leaving
component the evolution
development
field visualizations
reports
by various
or combination,
1) modes.
latter
and
effects
of the
three-dimensional Figure 6 shows
a de-tuned, (5/9,
experimental
for both
subharmonics
initial wave.
the
amplitudes
or two
performed.
three-dimensional
instability.
are several
primary-scale
paths and
to laminar small-scale,
case.
discussions
with Thomas
Corke,
Thorwald
References [1] P.S. Klebanoff, boundary-layer
K.D. Tidstrom, and L.M. instability. J. Fluid Mech.,
Sargent. 12:1-34,
The 1962.
three-dimensional
nature
[2] L.S.C. Kovasznay, H. Komoda, and _ B.R. Vasudeva. Detailed flow field In Proc. 1962 Heat Transfer and Fluid Mechanics Institute, 1-26. Stanford Palo Alto, 1962. [3] F. 1%.Hama and J. Nutant. Detailed flow-field a thick boundary layer. In Proc. 1963 Heat 77-93. Stanford Univ. Press, Paio Alto, 1963. [4] T. Herbert. 526, 1988.
Secondary
instability
observations in the Transfer and Fluid
of boundary
layers.
Ann.
[5] Yu. S. Kachanov, V. V. Koslov, and V. Ya. Levchenko. wave in a boundary layer. Izv. Akad. Nauk SSSR. Mekh. Russian). (See also Fluid Dyn. 12:383 (1978).) [6] Yu. S. Kachanov laminar-turbulent
and V. Ya. Levchenko. transition in a boundary
Rev]
in transition. Univ. Press,
transition: Mechanics
Fluid
of
process in Institute,
Mech.,
20:487-
Nonlinear development of :a Zhidk. I Gaza, 3:49, 1977 (in
The resonant interaction layer. J. Fluid Mech.,
of disturbances at 138:209-247, 1984.
[7] T. C. Corke. Measurements of resonant phase locking in unstable axisymmetric lets and boundary layers. In Nonlinear Wave Interactions in Fluids, (1%.W. Miksad, T.1%. Akylas and T. Herbert, eds.), 37-65. ASME, New York, 1987. [8] A. Wray and M. Y. Hussaini. Numerical Roy. Soc. Loudon, Series A, 392:373-389,
experiments 1984.
in boundary-layer
stability.
Proc.
[9] T. A. Zang and M. Y. Hussaini. Numerical simulation of nonlinear interactions in channel and boundary-layer transition. In Nonlinear Wave Interactions in Fluids, (R.W. Miksad, T.R.. Akylas and T. Herbert, eds.), 131-145. ASME, New York, 1987. 5_
[10]
E. Laurien and L. Kieiser. Numerical simulation Of boundary-layer sition control. J. Fluid Mech., 109: 403-_40, !980.
[11]
K. S. Yang, P. 1%. Spalart:, and J. H. FerzigerTransition in Boundary Layers. 1%eport No. Stanford University, 1988.
_
_
=
[12] P. 1%. Spalart and K. S. Yang. Numerical flow. J: Fluid Mech., 178:345-365, 1987. [13]
T. A. Zang and S. E. Krist. Numerical channel flow. Theoret. Comput. Fluid
[14]
T. A. Zang. tions. Appl.
[15]
S. E. Krist and TP 2667, 1987.
On the rotation Numer. Math., T. A. Zang.
r
tran-
Numericai Studies:on Laminar'turbulent TF-37. Dept. of Mechanical Engineering,
\
!
study "
experiments Dynamics,
Simulation
[16] J. D. Crouch. Analysis of three-dlmenslonal of NASA/ICASE Workshop on Instability Voigt, eds.) Springer, New York, 1989.
and
.....
and skew-symmetric in press. Numerical
transition
of ribbon-induced
transition
on stability and 1:41-64, 1989, forms
transition
for incompressible
of Channel
Flow
in Blasius
in plane
flow simula-
Trans{tion.
NASA
structures in boundary layers. In Proc. and Transition, (M. Y. Hussaini and R. G.
II
._o E t,,,Q
1 ii,-i
(/) 0
0
l-I
-_ II m
m
t-
,._
E
_ 0 .o
"13 C. I.I. o_,q
r
0 _0 60
_ o_ "_
7
rf_
_o m
C_ J_ mmm m
C_
E _0
0
o_
H
0
_I, ml
c_ mE
01FI_
L
m
_
l_
mo o1_q
1.4
O0
giim
lu
D,,
Q_
,E
a_ a_ _o
_m4
q_
r
r_
in
(:3 0 0 Ol
i
I
C) 1=='
X
0 C3 (4)
(D II
\
'
!
0 0
1--4
II
E ¢)
0 1.4 NT
_)
==
U
o_ ..I..1
¢)
X
o
m
I::I 0
|
| | 0
_ m
0 _.I
(I) C::
e6_
c_
H i!
O
1
_r_ t E
m
r
_E K
l z h i__ L..
!
f
11 j C_
NE_
w_
_O r,O
l0
ef,,q
V
O
0
r_
f
o(
._%,, m
C_
0
1l
_
o J,,-i
o
_
opqW
o
• _,,t
0 I
I
I
I
I
_
IL
o
--_
0
_
_
m cJ
_
o
0
,._
-
U
\
I _
0
I 0
"::2"0
o
=
|
I
o
i-4
12
_
_lOr_l _L_Or_ulCS SP_a¢e Aom_niSIIal_O
Report
an(_
I. Report No, NASA CR-181948 ICASE Report 4. Title and Subtitle
Documentation
Page
2. Government AccessionNo. No.
MULTIPLE PATHS IN A BOUNDARY
3. Recipient'sCatalog No.
89-78 5. ReportDate
TO SUBHARMONIC LAYER
LAMINAR
BREAKDOWN
October
1989
6. PerformingOrganizationCode
7. Author(s)
8. Performing OrganizationReportNo.
Thomas A. M. Yousuff
89-78
Zang Hussaini
10. Work Unit No. 505-90-21-01
Performing Organization Name and Address 11. Contract or Grant No.
Institute for Computer Applications in Science and Engineering Mail Stop 132C, NASA Langley Research Center Hampton, VA 23665-5225 12. SponsoringAgency Name and Address National
Aeronautics
and
Space
NASI-18605 13. Type of Reportand Period Covered Contractor
Report
14. Sponsoring Agency Code
Administration
Langley Research Center Hampton, VA 23665-5225 15. Supplementa_ Notes Submitted Langley Richard
Physics
Technical Monitor: W. Barnwell
to Review
Letters
Final Report 16. Abstract
layer waves tional
Numerical simulations demonstrate that laminar breakdown induced by the secondary instability of two-dimensional to three-dimensional subharmonic disturbances need not lambda
vortex/high-shear
layer
transition,
path.
18. DistributionStatement
17. Key Words (Suggestedby Author(s)) boundary-layer mechanics
fluid
34 01
- Fluid Mechanics and Heat - Aeronautics (General)
Unclassified 19. SecuriW Classif.(of this report) Unclassified
in a boundary Tollmein-Schlichting take the conven-
_. SecuriWClassif. (of thispage) Unclassified
Transfer
- Unlimited 21. No. of pa_s 14
_. Price A03
NASA FORM 1626OCT86 NASA-Langley,
1989
i
i J
t
|
Z
|
-