Status of Turbulence Modeling For Hypersonic Propu PDF

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/262948688
Status of Turbulence Modeling for Hypersonic Propulsion Flowpaths
Article in Theoretical and Computational Fluid Dynamics · December 2013

DOI: 10.1007/s00162-013-0316-z
CITATIONS READS
45 196
4 authors, including:
Nicholas J. Georgiadis William A. Engblom

NASA Embry-Riddle Aeronautical University
83 PUBLICATIONS 1,421 CITATIONS 52 PUBLICATIONS 616 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Jet CFD View project
Turbulent Heat Flux View project
All content following this page was uploaded by Nicholas J. Georgiadis on 31 August 2015.
The user has requested enhancement of the downloaded file.

NASA/TM—2012-217277 AIAA–2011–5917
Status of Turbulence Modeling for Hypersonic

Propulsion Flowpaths
Nicholas J. Georgiadis, Dennis A. Yoder, and Manan A. Vyas
Glenn Research Center, Cleveland, Ohio
William A. Engblom
Embry Riddle Aeronautical University, Dayton Beach, Florida
April 2012
NASA STI Program . . . in Profile
Since its founding, NASA has been dedicated to the • CONFERENCE PUBLICATION. Collected
advancement of aeronautics and space science. The papers from scientific and technical
NASA Scientific and Technical Information (STI) conferences, symposia, seminars, or other
program plays a key part in helping NASA maintain meetings sponsored or cosponsored by NASA.
this important role.
• SPECIAL PUBLICATION. Scientific,
The NASA STI Program operates under the auspices technical, or historical information from
of the Agency Chief Information Officer. It collects, NASA programs, projects, and missions, often
organizes, provides for archiving, and disseminates concerned with subjects having substantial
NASA’s STI. The NASA STI program provides access public interest.
to the NASA Aeronautics and Space Database and
its public interface, the NASA Technical Reports • TECHNICAL TRANSLATION. English-
Server, thus providing one of the largest collections language translations of foreign scientific and
of aeronautical and space science STI in the world. technical material pertinent to NASA’s mission.
Results are published in both non-NASA channels
and by NASA in the NASA STI Report Series, which Specialized services also include creating custom
includes the following report types: thesauri, building customized databases, organizing
and publishing research results.
• TECHNICAL PUBLICATION. Reports of
completed research or a major significant phase For more information about the NASA STI
of research that present the results of NASA program, see the following:
programs and include extensive data or theoretical
analysis. Includes compilations of significant • Access the NASA STI program home page at
scientific and technical data and information http://www.sti.nasa.gov
deemed to be of continuing reference value.
NASA counterpart of peer-reviewed formal • E-mail your question via the Internet to help@
professional papers but has less stringent sti.nasa.gov
limitations on manuscript length and extent of
graphic presentations. • Fax your question to the NASA STI Help Desk
at 443–757–5803
• TECHNICAL MEMORANDUM. Scientific
and technical findings that are preliminary or • Telephone the NASA STI Help Desk at
of specialized interest, e.g., quick release 443–757–5802
reports, working papers, and bibliographies that
contain minimal annotation. Does not contain • Write to:
extensive analysis. NASA Center for AeroSpace Information (CASI)
7115 Standard Drive
• CONTRACTOR REPORT. Scientific and Hanover, MD 21076–1320
technical findings by NASA-sponsored
contractors and grantees.
NASA/TM—2012-217277 AIAA–2011–5917

Glenn Research Center, Cleveland, Ohio
William A. Engblom
Embry Riddle Aeronautical University, Dayton Beach, Florida
Prepared for the

47th Joint Propulsion Conference and Exhibit
cosponsored by the AIAA, ASME, SAE, and ASEE
San Diego, California, July 31–August 3, 2011
National Aeronautics and

Space Administration
Glenn Research Center

Cleveland, Ohio 44135
April 2012
Acknowledgments
The authors would like to thank the Test Resource Management Center (TRMC) and the Test and Evaluation/Science and
Technology (T&E/S&T) Program for their support. This work was sponsored by the TRMC T&E/S&T Program through the High
Speed Systems Test (HSST) area (formerly APTT) and the NASA Fundamental Aeronautics Program.
Trade names and trademarks are used in this report for identification
only. Their usage does not constitute an official endorsement,
either expressed or implied, by the National Aeronautics and
Space Administration.
This work was sponsored by the Fundamental Aeronautics Program

at the NASA Glenn Research Center.
Level of Review: This material has been technically reviewed by technical management.
Available from
NASA Center for Aerospace Information National Technical Information Service
7115 Standard Drive 5301 Shawnee Road
Hanover, MD 21076–1320 Alexandria, VA 22312
Available electronically at http://www.sti.nasa.gov


National Aeronautics and Space Administration
Glenn Research Center
Cleveland, Ohio 44135
William A. Engblom
Embry Riddle Aeronautical University
Dayton Beach, Florida 32114
This report provides an assessment of current turbulent flow calculation methods for
hypersonic propulsion flowpaths, particularly the scramjet engine. Emphasis is placed on
Reynolds-averaged Navier-Stokes (RANS) methods, but some discussion of newer meth-
ods such as Large Eddy Simulation (LES) is also provided. The report is organized by
considering technical issues throughout the scramjet-powered vehicle flowpath including
laminar-to-turbulent boundary layer transition, shock wave / turbulent boundary layer
interactions, scalar transport modeling (specifically the significance of turbulent Prandtl
and Schmidt numbers) and compressible mixing. Unit problems are primarily used to
conduct the assessment. In the combustor, results from calculations of a direct connect
supersonic combustion experiment are also used to address the effects of turbulence model
selection and in particular settings for the turbulent Prandtl and Schmidt numbers. It is
concluded that RANS turbulence modeling shortfalls are still a major limitation to the
accuracy of hypersonic propulsion simulations, whether considering individual components
or an overall system. Newer methods such as LES-based techniques may be promising, but
are not yet at a maturity to be used routinely by the hypersonic propulsion community.
The need for fundamental experiments to provide data for turbulence model development
and validation is discussed.
Nomenclature
Cf skin friction coefficient
Cp specific heat at constant pressure
Dt turbulent species diffusivity
h static enthalpy
H fuel injector ramp height for scramjet test case
k turbulent kinetic energy
kt turbulent thermal conductivity
M Mach number
Mt turbulent Mach number
P static pressure
Pk production of turbulent kinetic energy
q˙w rate of wall heat flux
P rt turbulent Prandtl number √
Ry turbulent Reynolds number based on wall distance = ρy k/µ
Rt turbulent Reynolds number = ρk/µω
NASA/TM—2012-217277 1
Rex plate Reynolds number
Reν vorticity-based Reynolds number = ρy 2 Ω/µ
s streamline coordinate
Sij rate of strain tensor
Sct turbulent Schmidt number
St Stanton number
t time
T static temperature
Tr temperature ratio
Tt stagnation temperature
Tw wall static temperature
Uj velocity vector
U∞ freestream velocity
Wij vorticity tensor
x, y, z Cartesian coordinates
y+ wall normal coordinate
turbulent dissipation rate
s solenoidal turbulent dissipation rate
µ dynamic viscosity
µt dynamic eddy viscosity
ω specific turbulent dissipation rate = /k
Ω vorticity magnitude
φ fuel equivalence ratio
ρ density
ρ∞ freestream density
T
τij turbulent stress tensor
Introduction
Flowpaths within hypersonic propulsion systems, in particular the scramjet engine, are characterized by
several aerodynamic and thermodynamic features which are very difficult for currently available computa-
tional fluid dynamics (CFD) methods to calculate accurately. At the same time, experiments cannot provide
all of the details of scramjet flows because of the difficulties in simulating actual operating conditions in
ground experiments and making appropriate measurements in flight tests. As a result, the need for com-
petent CFD methods to be used in conjunction with experiments is high, in order to successfully carry out
hypersonic research and development programs, as outlined in detail by Drummond et al1 and Tishkoff et
al.2 One CFD modeling area that has remained extremely challenging, yet very critical to the success of
hypersonic propulsion system simulations is turbulence modeling, which is the focus of this paper.
Scramjet powered hypersonic vehicles are among the most integrated vehicle / propulsion system config-
urations of all aircraft. Nevertheless, we may identify flow features specific to the primary propulsion system
components that provide the greatest turbulence modeling challenges, as identified in Fig. 1. For many
hypersonic vehicles, the entire forebody beginning with the leading edge forms part of the engine inlet. Here
boundary layer transition has a large effect on the thickness of the boundary layers that enter the enclosed
portion of the scramjet engine and also on the the thermal properties of the vehicle forebody. Knowledge
of the boundary layer state is important enough that programs such as HyperX / X-43 have implemented
forced transition locations via boundary layer trips.3, 4
Within the inlet and isolator, shockwave/boundary layer interactions severely complicate the flow, and
provide many challenges to turbulence modeling efforts. Within the combustor, the complexities are nu-
merous. These include three-dimensional turbulent mixing in the presence of a chemically reacting flow,
recirculating flow at the locations of fuel injection, heat transfer at the combustor walls, compressibility ef-
fects, and continued shockwave / boundary layer interactions. Further downstream in the expansion system,
the flow may still be reacting and the complexities are similar to those at the beginning of the combustor
region.
The practical state-of-the-art in CFD for hypersonic propulsion flowpath analyses during the past two
NASA/TM—2012-217277 2
decades has been Reynolds-averaged Navier-Stokes (RANS) methods, typically using two-equation eddy
viscosity closures such as k-ω or k- models. In terms of more advanced RANS techniques, some limited
attempts have been made to use algebraic Reynolds stress models, but substantial improvement over the
linear two-equation eddy viscosity approaches has not been demonstrated. There have not been many
attempts to use full Reynolds-stress closures for scramjet analyses. The flow features described in the
preceding paragraph all provide major challenges to the most commonly used two-equation eddy viscosity
RANS techniques. Transition models coupled to RANS are highly empirical at best, and have only been
shown to have potential applicability at high freestream turbulence levels, which is not the case in actual
flight simulations. For hypersonic vehicle inlets, RANS methods have been shown to be quite deficient
for calculating the details of shockwave/boundary layer interactions. Isolator flowfields frequently contain
several shock boundary layer interactions. In ramjet mode, this separated flow enters the combustor creating
a complex flowfield that no RANS models can accurately compute.
The complexities occurring farther downstream in the combustor and expansion system, namely high
temperatures, compressibility effects, recirculating flow in the regions of fuel injection, and other three-
dimensional flow features, further overwhelm currently practical CFD capabilities. Baurle5 provides a com-
prehensive overview of turbulent reacting methods for supersonic combustion while Ebrahami6 addresses
validation challenges for high speed combustion problems from scramjets to rockets. Turbulent chemistry
interactions are obviously crucial in these regions, yet no RANS-based techniques have been shown to accu-
rately represent the actual flow state. Constant values for the turbulent Prandlt and Schmidt number are
typically used to extend the RANS-based eddy viscosity used in the momentum equations to the turbulent
heat transport and species transport in the energy and individual species transport equations, respectively.
This is known to be a significant limitation,5, 7 and while some efforts have been underway in recent years to
develop more sophisticated RANS-based models for such scalar transport, not enough testing and calibration
has been performed to date to determine the potential of these techniques.
Beyond RANS-based methods, large eddy simulation (LES) techniques and hybrid RANS-LES techniques
have received significant attention over the past decade for application to aerodynamic flow analyses. While
progress has been made in developing and using such techniques, especially for flows away from walls,
LES-based techniques cannot be considered a mature science, free of modeling parameters that need to be
set by the analyst. Effects of grid sensitivity are fundamentally different from RANS-based techniques. In
RANS, it is the expectation that as one refines a computational grid, a grid-independent solution is achieved.
In contrast for LES, as the grid is refined, typically more small turbulent structures are resolved instead
of modeled, and the solution can indeed be expected to change. In the subsonic combustor community,
LES techniques have been shown to have some promise,8 in particular for flows where wall effects can be
considered small, and where turbulent velocity scales are on the order of the mean flow velocity. In a scramjet
combustor, however, the situation is different, as the core flow velocity is still of appreciable magnitude, and
wall boundary layer effects are important. References 9 and 10 have demonstrated some of the earliest
application of LES-based methods to scramjet combustors. While LES-based methods have promise for
calculating the largest scales of turbulence, chemical reactions occur at the microscale, so the subgrid scale
modeling used in LES for scramjet simulations (or any combustor for that matter) is very important. Some
research into developing more accurate subgrid scale closures to handle turbulence /chemistry interactions
is underway,11 but such methods are far from being considered mature enough to be applicable to practical
hypersonic propulsion system analyses.
In this paper, the chief modeling challenges briefly outlined in this introduction are addressed. The
current practical state-of-the-art in RANS-based techniques is emphasized, utilizing some of our own recent
experiences. Our efforts utilized the Wind-US code12–14 which is a production RANS solver with several
available turbulence models. Discussion of more computationally intensive techniques such as LES, including
hybrid RANS-LES approaches, will be provided. Suggestions for the required research to result in improve-
ments to these methods will be described. A discussion of experimental efforts that would aid modeling
efforts will also be provided.
RANS Turbulence Modeling

For nearly two decades, the practical state-of-the-art in RANS turbulence modeling has employed two-
equation eddy viscosity turbulence models. In linear eddy viscosity models, the Boussinesq approximation is
used to relate the turbulent Reynolds stress to the mean rate of strain tensor through a turbulent (or eddy)
NASA/TM—2012-217277 3
viscosity:

T
2 ∂uk 2
τij = µt 2Sij − δij − ρkδij (1)
3 ∂xk 3
The rate of strain is tensor, Sij , is given by:

1 ∂uj ∂ui
Sij = + (2)
2 ∂xi ∂xj
It is beyond the scope of this paper to describe all two-equation models that are in use today. A detailed
list of several RANS turbulence models that have been used in hypersonic simulations is presented in Ref.
15. It should be noted that the vast majority of the literature concerning validation and comparison of
RANS turbulence models has not been in the hypersonic regime, but at much slower speeds, particularly
subsonic and transonic. A majority of the two-equation turbulence models in use are either k − or k − ω
formulations, which we focus on here. While not discussed here, a model that has received attention for high
speed flows in recent years in the k − ζ (enstrophy) model as described in Refs. 16–18. Here we describe
in some detail one turbulence model that is widely used, and has been utilized for the majority of our work
in recent years. This model is the Shear-Stress Transport (SST) two-equation turbulence formulation of
Menter.19 The SST model has been shown to be robust and relatively accurate for a broad range of flows,
including wall boundary layers and free shear layer regions. The SST model is a two-layer model which
employs the k − ω model of Wilcox20 in the inner region of boundary layers and switches to a k − model in
the outer region of boundary layers and in mixing regions. The outer k − model is transformed to provide a
second set of k − ω equations with a blending function used to transition between the two sets of equations.

∂ρk ∂ρUj k ∗ ∂ ∂k
+ = Pk − β ρωk + (µ + σk µt ) (3)
∂t ∂xj ∂xj ∂xj

∂ρω ∂ρUj ω α ∂ ∂ω 1 ∂k ∂ω
+ = Pk − βρω 2 + (µ + σω µt ) + (1 − F1 )2ρσω2 (4)
∂t ∂xj νt ∂xj ∂xj ω ∂xj ∂xj

2
Pk = min 2µt Sij Sij − (µt Smm − ρk)Snn ; µt Ω2 (5)
3
F1∗ = tanh(arg41 ) (6)
= exp −(Ry /120)8

F4 (7)
F1 = max (F1∗ , F4 ) (8)
√ ! !
k 500ν 4ρσω2 k
arg1 = min max ; ; (9)
β ∗ ωy ωy 2 CDkω y 2

1 ∂k ∂ω −20
CDkω = max 2ρσω2 ; 1.0 × 10 (10)
ω ∂xj ∂xj
and from these the turbulent viscosity is given as

ρk a1 ρk
µt = min α∗ ; (11)
ω ΩF2
F2 = tanh(arg22 ) (12)
√ !
k 500ν
arg2 = min 2 ∗ ; (13)
β ωy ωy 2
The constants (φ1 ) associated with these equations for the inner, k − ω, model are:
4
5/18 + (Rt /8) .025 + Rt /6
σk1 = 0.85, β1∗ = 0.09 · 4 , α1∗ = ,
1 + (Rt /8) 1 + Rt /6
5 0.1 + Rt /2.7
σω1 = 0.5, β1 = 0.075, α1 = · , a1 = 0.31
9 1 + Rt /2.7
NASA/TM—2012-217277 4
and the constants (φ2 ) for the transformed k − model are:
σk2 = 1.0, β2∗ = 0.09, α2∗ = 1,

σω2 = 0.856, β2 = 0.0828, α2 = 0.4403, a1 = 0.31
These constants are blended using the same switching function, F1 that is found in the model equations such
that φ = F1 φ1 + (1 − F1 )φ2 for any of the given parameters.
The SST turbulence model has become popular for its ability to handle separated flows and complex
geometry in the near wall region due to the strengths of the k−ω model, while maintaining the characteristics
of the k − model to be more accurate in free shear layers.
In eddy viscosity models such as the SST formulation, the ultimate quantity that is used in the momentum
equations to model the turbulent stresses is the eddy viscosity, µt . In most RANS codes, a constant turbulent
Prandtl number is used to extend the eddy viscosity to model the turbulent thermal transport via the relation:
P rt = µt Cp /kt (14)
The turbulent species diffusivity is extended from the eddy viscosity in a similar manner via the turbulent
Schmidt number:
Sct = µt Cp /Dt (15)

In most applications at low Mach numbers and without reactions or significant heat transfer effects, the
air is treated as a calorically perfect gas; a constant turbulent Prandtl number on the order of unity is
typically used and no species transport equations are solved. However, the turbulent Prandtl and Schmidt
numbers have significant effects for supersonic combustion problems as will be discussed in a later section.
Beyond two-equation models, full Reynolds-stress closures, or second moment closures as they are some-
times referred to, offer a more complete representation of the three-dimensional turbulent stress field. How-
ever, the improvements in accuracy over linear two-equation models have not been shown to be sufficiently
high enough to justify the additional computational cost.21 As a result, the international CFD community
has not broadly accepted the usage of full Reynolds stress closures, even with the known limitations of linear
eddy-viscosity models. As a compromise beyond two-equation viscosity models, there has been some work
to develop and investigate non-linear explicit algebraic stress models (EASMs). Unlike linear two-equation
models, EASM formulations are sensitive to turbulent stress anisotropies and have a direct relation to the
full Reynolds stress model.22 As a result, EASM models have the capability to include more relevant flow
physics than the linear models. However, they are also solved using a two-equation approach and as a result
are not significantly more computationally expensive than linear two-equation models.
The turbulent stress tensor of the EASM (for both k − and k − ω) is:
T
τij = 2µt {Sij − 31 Skk δij + [a2 a4 (Sik Wkj − Wik Skj )
(16)
−2a3 a4 Sik Skj − 31 Skl Slk δij ]} − 23 ρkδij

where Sij was defined previously in Eq. 2 and Wij is:

1 ∂ui ∂uj
Wij = − (17)
2 ∂xj ∂xi
The eddy viscosity, µt is:

µt = Cµ∗ ρkτ = −ρkα1 (18)
where the turbulent time scale is τ = 1/ω = k/. The quantity α1 /τ is equivalent to −Cµ∗ and is obtained
from the solution to a cubic equation at every point in the flow field. The closure constants, a2 , a3 , a4 ,
and entire solution procedure are described in Refs. 23 and 24. The form of the k − ω model used as the
underlying two-equation model for the EASM described here is provided in Ref. 25.
In the following sections, we consider the performance of RANS turbulence modeling for component
or subset problems that exhibit the turbulence modeling-dominated features found within the hypersonic
propulsion system flowpath as described in the introduction and shown schematically in Fig. 1. We begin
with a discussion of laminar-to-turbulent boundary transition. Consideration of shock wave / turbulent
NASA/TM—2012-217277 5
boundary layer interactions, which dominate the flow within the scramjet isolator but are pervasive to the
entire scramjet propulsion flowpath, are considered next. Moving toward the rear of the propulsion flowpath,
consideration of turbulence modeling factors in the combustor and exhaust system are presented in the last
section.
Forebody/Inlet Transition Modeling

Hypersonic flight vehicles frequently have boundary layers with significant laminar regions on vehicle
forebodies and inlet surfaces because of the low freestream disturbances and low densities characteristic of
the altitudes that such vehicles fly within. While flight tests can reproduce these atmospheric conditions,
scaling issues due to Reynolds number affect the ability to reproduce the effects of boundary layer state,
and in particular the point of transition. Ground test facilities frequently have freestream turbulence levels
that are high enough to affect the transition point. Whereas the mechanism of transition in free flight is
usually modal growth, the high freestream turbulence experienced by ground test articles frequently leads
to bypass transition, in which the transition to turbulent flow is dictated by large freestream disturbances.
In this section, we consider a RANS-based bypass transition model. It is generally agreed that RANS-based
techniques are not applicable to modal growth situations.
Details of the formulation presented here are provided in Ref. 26. The model is based on the SST turbu-
lence model and was built starting from a previous SST-based transition model.27–29 Several modifications
were made to enable: (1) consistent solutions regardless of flow field initialization procedure and (2) fully
turbulent flow beyond the transition region. In the following, we highlight only the key aspects of the imple-
mentation of the bypass transition model. Building upon the SST turbulence model, the key modification
is made to the production term in the turbulent kinetic energy equation:

∂ρk ∂ρUj k ∗ ∂ ∂k
+ = P T M · Pk − β ρωk + (µ + σk µt ) (19)
∂t ∂xj ∂xj ∂xj
where P T M is termed the “production term multiplier.” The ω equation is not modified.
The final form of the transition model that is recommended based on the work of Ref. 26 is:
P T M = 1 − 0.94(P T M 1 + P T M 2) F3 tanh (y + /17)2

(20)
Rt 2 1
F3 = e−( 3 ) (1 − P (Rt )) + P (Rt ) (21)
2
2.5 −(Rt −3)2
P (Rt ) = √ e 2 (22)
2π

(3.28 × 10−4 )Re − (3.94 × 10−7 )Re2 + (1.43 × 10−10 )Re3 ; Re < 1000
v v v v
P T M 1 = 1 − CP T M 1 (23)
 0.12 + (1.00 × 10−5 Rev ; Rev > 1000

− |K|0.4 Rev ; K < 0
PTM2 = 80 (24)
0; K>0
1.0 ≤ CP T M 1 ≤ 2.0 (Recommended range) (25)
where Rev is the vorticity-based Reynolds number, Rt is the turbulent Reynolds number, and the pressure
gradient parameter, K, is given by
µ dp
K = − 2 3 1 − M2 (26)
ρ U ds
In Refs. 28 and 29, P T M 2 was formulated for flows with significant internal flow pressure gradients,
and specifically for flows within low pressure turbine stages. In all cases examined in this work and that
of Ref. 26, where flows with significant pressure gradients were not examined, P T M 2 was not found to be
significant.
This model is intended for flows where bypass transition is the key transition mechanism as opposed to
transition dictated by modal growth. Validation of the new transition model was performed for flows ranging
from incompressible to hypersonic conditions.
NASA/TM—2012-217277 6
Figure 2 shows a comparison of simulations for a widely used data set for incompressible transition flow
over a flat plate(referred to as the T3A test case30 ). The fully turbulent SST solution shows the rapid
transition (much closer to an entirely turbulent flow) that provided the motivation for the present work. It
does a poor job of accurately capturing the early laminar behavior and the transition location which would
lead to incorrect results for the drag on this flat plate. For the transition model solutions, the effect of the
key parameter, CP T M 1 , is to control the transition onset location. As with nearly all RANS-based transition
prediction schemes, the width of the calculated transition region is shorter than indicated by experimental
data. The solution with CP T M 1 = 2.0 captures the transition onset location best while decreasing CP T M 1
to 1.0 (the default value from original formulation) captures the end of the transition zone (i.e. where fully
turbulent flow is realized) best.
Hypersonic transitional flows present unique challenges in both modeling and experimentation. To base-
line the present model’s ability to predict transitional behavior in hypersonic conditions, it was validated
against transition data taken on sharp nose cones in the AEDC tunnel B at Mach 7.93.31 The simulation is
performed on a 7◦ half-angle cone. As in the experiment, the wall temperature was set at 0.42Tt , where Tt
is the freestream stagnation temperature. Several unit Reynolds numbers (Re/m) were evaluated in Ref. 31
to provide a complete scan of the transition region. These data show good agreement for transition location
across a range of Re/m, allowing the simulation to use only one, ≈ 6.8E6 Re/m, corresponding closely with
the center of the experimental range. Inflow conditions are calculated to match those of Ref. 31. Insight
from Ref. 32 indicated that an inlet freestream turbulence intensity of 1.25% was appropriate for AEDC
tunnel B.
Figure 3 shows the Stanton number (St), as defined by Ref. 31 and as shown in Eq. 27, versus Reynolds
number for the SST model alone, the SST transition model using CP T M 1 = 1.0 and CP T M 1 = 2.0, and the
experimental data.
St = q˙w /(ρ∞ U∞ (h(Tt ) − h(Tw ))) (27)
As in the incompressible case, the predicted behavior is overly abrupt. However, the model accurately
captures the minimum heat transfer value, corresponding to transition onset for CP T M 1 = 2.0 and the
location where the flow becomes fully turbulent for CP T M 1 = 1.0.
The transition model solutions are a marked improvement over the SST model alone, which vastly over-
predicts the heat transfer by indicating a fully turbulent state from the leading edge of the test article. For a
model in a high-speed wind tunnel that has a significant laminar region, the error due to transition onset can
be quite large. The total heat transfer integrated over the length of the cone differs by 38.7% between the
standard (fully turbulent) SST model and the SST-based transition model with CP T M 1 = 1.0. Accurately
capturing this behavior is especially important when trying to evaluate thermal properties and heat transfer
behavior near the tip of a hypersonic vehicle. The large discrepancy in heat transfer rates indicated by the
two models would substantially alter the vehicle’s predicted temperature profile.
While the transition model presented here demonstrates some promise for bypass transition, it is empha-
sized that a general RANS-based predictive technique for the modal growth situation that is more important
in free flight has not been demonstrated nor is likely feasible. LES is likely not a candidate either, because
the transition process is dominated by amplification of initially very small disturbances, which is in contrast
to the LES philosophy of calculating the largest turbulent scales while modeling the smallest scales. In
the long run, direct numerical simulation (DNS) probably offers the best prospects for such a predictive
capability. In the interim, techniques such as eN methods, in which stability theory is used to predict the
location of transition,33–35 will remain the realistic option.
Shock Wave Turbulent Boundary Layer Interactions

The shock wave / turbulent boundary layer interaction (SWTBLI) is a very common phenomena that
has significant effects throughout the hypersonic propulsion system flowpath. The problem has been studied
extensively over the past few decades in laboratory experiments and in computational fluid dynamics (CFD)
efforts. While gains in understanding of the underlying fluid dynamics of SWTBLI have been made, research
continues into understanding the complex interaction of a shock wave with a turbulent boundary layer. In
a typical SWTBLI, the adverse pressure gradient induced by the shock system causes a flow separation that
frequently is unsteady and three-dimensional. Without question, control of SWTBLI has yet to be mastered.
While there have been numerous CFD studies, there is not a single approach that has been identified as
optimal for calculation of SWTBLI.
NASA/TM—2012-217277 7
For practical aerodynamic analyses that contain one or more SWTBLIs as is the case for a scramjet
isolator, RANS is still the only feasible approach. More recently, large eddy simulation (LES) and hybrid
RANS-LES techniques have been investigated for application to the SWTBLI problem, but typically have
been restricted to small unit problems, as applying an LES-based technique to a more complex system (i.e.
aircraft inlet) involving SWTBLIs is still largely prohibitive due to the range of scales involved. Knight and
Degrez36 and Knight et al37 provide comprehensive overviews of a broad range in CFD methods as applied
to SWTBLIs, in particular those investigated under AGARD and RTO working groups. Consideration of
both RANS and LES methods is made. Reference 38 focuses on an assessment of RANS-based methods
while a survey of LES-based approaches as applied to SWTBLIs is presented in Edwards.39 The overall con-
clusions of these survey papers is that RANS methods are inherently unable to calculate some of the crucial
features of SWTBLI, in particular the unsteadiness of the shock system and separated flow. In addition, the
three dimensional features are also troublesome for RANS-based techniques. LES-based techniques may be
promising, but neither sufficient maturity of these techniques nor experience using them has been realized.
In an inlet or isolator the shocks closest to the leading edge of the vehicle have the highest approach Mach
number. Further downstream, the shocks may occur at lower approaching Mach number, but the boundary
layers are typically thicker and may have already been subject to effects of upstream SWTBLIs. Such is the
case in scramjet isolators.
We applied several two-equation models including eddy-viscosity models and EASMs to a Mach 5 SWBLI
problem in the current work. Mach number contours for an SST Wind-US solution are provided in Fig. 4.
Comparisons of several solutions using linear two-equation turbulence models and EASMs with experimental
data from Ref. 40 are provided in Fig. 5. It may be observed that the linear k − model solution provides
poor agreement with experimental data. While the k − based EASMs, with and without compressibility
corrections, marginally improve upon the linear k − solution, the largest improvement is obtained by using
the k − ω formulations, including the Menter SST model. The k − ω EASMs provide improvement just past
the shock impingement point. The solution obtained with the SST model also provides reasonable agreement
in the post-shock region, with somewhat lower calculated skin friction than the experimental data and the
other k − ω models (linear and EASM).
A recent workshop considering CFD calculations for a set of SWTBLI cases was held in conjunction
with the 48th American Institute of Aeronautics and Astronautics (AIAA) Aerospace Sciences Meeting. An
overview of the workshop objectives is presented in Benek41 with a summary of overall findings presented in
Benek.42 Several investigators contributed solutions to this workshop. Contributions included RANS and
LES-based computations. DeBonis et al43 provides a comprehensive assessment of the CFD calculations,
including uncertainty analysis of the submitted CFD results and experimental data obtained for the same
configurations. Hirsch44 examined some of the CFD trends, in particular effects of turbulence modeling,
RANS versus LES, and numerical schemes submitted by several investigators and includes the results from
our efforts described here.
One of two SWTBLI configurations considered in the workshop was a test case obtained at the Institut
Universitaire des Systemes Thermiques Industriels (IUSTI) in Marseille, France.45 The experimental data
was utilized in the European Union SWTBLI research project referred to as UFAST.46 The UFAST experi-
ments utilized an 8 degree shock generator which spanned the entire width of the tunnel with an approaching
Mach 2.25 flow. The supply stagnation pressure was 50.5 kPa and the stagnation temperature was 293 K. A
schematic of the experimental configuration representing that used in the UFAST experiments is shown in
Fig. 6 as taken from DeBonis et al.43 Particle Image Velocimetry (PIV) was used in all cases to characterize
the interaction region, with both mean flow velocities and turbulent statistics obtained.
For this test case, we utilized the Menter SST k-ω and k-ω based EASM two-equation models and the
one-equation model due to Spalart and Allmaras.47, 48 Axial velocity contours obtained for the UFAST test
case are shown in Fig. 7. The oblique wave originating from the sharp leading edge of the shock generator
may be observed, along with the SWTBLI region centered about x = 320mm. The complex interaction
on the top of the tunnel actually results in a flow separation that is much larger than that in the focused
SWTBLI region on the bottom.
The experimental data was collected within the region focused around the bottom wall interaction. Axial
velocity contours for the UFAST test case using the three turbulence modeling approaches are compared to
the experimental PIV measurements in Fig. 8. The CFD solution slices represent the same physical domain,
with the same contour levels as the experimental data. The CFD solutions were interpolated to the same
physical locations where the PIV data were taken, as was required by the organizers of the AIAA workshop,
NASA/TM—2012-217277 8
for purposes of computing differences between solutions and experimental data.43 Examining these contours,
the extent of the adverse pressure gradient effects indicated by the Menter SST solution are larger than the
other solutions and experimental data. Further comparisons of the CFD solutions and experimental data
are made for the axial velocity profiles at four axial locations in Fig. 9. One may observe that the CFD
solutions plotted do not go all of the way to the wall, and again this is due to interpolation of the CFD
results to the PIV measurement locations.
The comparisons in Fig. 9 show the same trends as indicated by the contour plots of Fig. 8. In
particular, the size of the flow separation indicated by the SST solution is greater than that indicated by the
experimental data or the other turbulence models. It appears that the SA and ASM models do not predict
as large a SWTBLI region as the experimental data. These results are very similar to those reported for SA
and SST by Bhagwandin and DeSpirito.49 They also considered a k- model which produced solutions that
highly underestimated the effect of the SWBLI. This is generally the expected performance of k- models in
separated or adverse pressure gradient flows.21
Combustor / Exhaust System Modeling

For several years, the University of Virginia (UVA) has conducted research on a benchmark scramjet
combustor, shown schematically in Fig. 10, to investigate dual-mode ramjet/scramjet engine behavior.50–52
This configuration consists of a two-dimensional convergent-divergent nozzle providing a Mach 2 flow to a
rectangular isolator. An electric heater is used to provide clean air at approximately 1200 K to the test article.
To examine test media effects, water vapor and carbon dioxide delivery systems may be used to replicate
the vitiated air levels typical of combustion heated facilities simulating Mach 5 flight. For reference, the
key stations through the UVA configuration are listed in table 1. Downstream of the isolator, an unswept
ramp contains a single round fuel injector which provides unheated hydrogen to a combustor section. The
burned flow then enters a divergent nozzle and exhausts to the freestream. The fuel injector geometry is a
convergent-divergent nozzle. We have applied RANS based calculations to this configuration,12, 53, 54 with
much of the focus on determining the capability of CFD to replicate differences between clean air and vitiated
air on engine performance as a function of fuel equivalence ratio. Gupte et al55 used finite element analysis to
examine the thermal-structural response of the UVA test article and found substantial effects of the thermal
deformation on the predicted flowfield when comparing CFD solutions from the baseline to the thermally
deformed configuration.
Station x/H
Isolator Entrance -45
Fuel Injection Plane 0
Nozzle Exit 57
Table 1. Key positions for UVA scramjet configuration
Most recently, the results of Vyas et al54 demonstrated the ability to qualitatively reproduce the equiv-
alence ratios where the combustor operation switched from ramjet to scramjet mode in the experiments of
Rockwell et al.52 In all of these experiments, the equivalence ratios were relatively low, on the order of 0.5
or less, in comparison to the typical equivalence ratios on the order of unity that a scramjet intended for
actual flight will experience. Variants of the baseline UVA scramjet combustor flow have been developed, as
discussed in Goyne et al,56 to enable higher equivalence ratios in anticipation of planned flight tests.
A major difficulty in drawing solid conclusions of CFD-to-experiment comparisons of scramjet predictions
is that frequently the only quantity available for comparison is surface pressure measurements as a function
of axial position in the scramjet flowpath. In CFD, several factors can contribute to significant changes in
the pressure distribution, i.e. turbulence model selection, settings for the turbulent Prandtl and Schmidt
numbers, chemical kinetics choice, wall temperature and/or heat transfer modeling. Typically an analyst will
make a selection for all of these model settings, and then vary one of the parameters, such as the turbulent
Schmidt number, to optimize the CFD generated pressure distribution until it passes most closely through
the experimental pressure distribution. While this is the current “state-of-practice,” optimizing such settings
for one scramjet configuration at one operating point does not guarantee the same settings will be optimal
NASA/TM—2012-217277 9
at other operating points, say with different freestream inflow conditions and fuel equivalence ratios. Over
a period of several years, the efforts discussed in Refs. 12, 53, 54 considered effects of turbulence model
selection, settings for the turbulent Prandtl and Schmidt numbers, and chemical kinetics modeling. We also
worked closely with the experimentalists, who made significant efforts to carefully measure heat transfer
through the scramjet walls, to model the wall thermal conditions as closely as possible.
The discussion here focuses on findings specific to turbulence modeling. Figure 11 shows Mach number
contours in the UVA scramjet flowpath for the fuel off (φ = 0.0) case using the Chien k − model. The
top plot shows contours from the facility nozzle entrance through isolator, combustor and exhaust nozzle.
The middle plot focuses on the region near the end of the isolator and combustor, while the bottom plot
extends from the combustor to the exhaust nozzle exit plane. Supersonic flow exits from the facility nozzle,
through the combustor, and into the expansion nozzle with no pressure rise in the combustor, which would
occur for fueled cases. In contrast to a flight vehicle nozzle, where by design the flow would continue to
expand through the exhaust nozzle, the flow in the UVA experiments is overexpanded due to limitations in
the overall system nozzle pressure ratio, and a shock system with flow separation may be observed near the
exhaust nozzle exit.
Figure 12 shows a comparison of RANS solutions obtained using the Menter SST model and the Chien
k − model. Through most of the flowpath, the pressure distributions are very similar, but in the last half of
the exhaust nozzle, the SST pressure distribution deviates from the experimental data and the k − solution.
This is the result of the SST solution separating much farther upstream in the nozzle. As discussed in the
SWTBLI section, typically k − models are not considered a good choice for flows with separations, and the
SST model has developed a broad reputation for performing better in such situations. However, as shown in
the SWTBLI results, it appears that SST overreacts in moderate to highly separated flows situations. This
may be in part due to the actual shear stress limiter term active within the Menter SST model, but the
performance is generally similar to other k − ω models.
Our initial set of calculations simulating an entire sweep in fuel equivalence ratios, as reported in Bhag-
wandin et al,53 used the SST model and 7-species, 8-step H2 -air kinetics due to Evans and Shexnayder57
modified to include third body efficiencies other than unity, as taken from the model of Jachimowski.58
After obtaining the result shown in Fig. 12, which was not expected, we completed an additional sweep in
equivalence ratios for both clean and vitiated air as reported in Vyas et al,54 using 9-species, 13-step H2 -air
kinetics due to Peters and Rogg59 and the Chien k- model. The Peters and Rogg model was selected for
the subsequent studies because the complete kinetics set was developed to investigate vitiated air. These
latter calculations were able to reproduce the experimentally observed “mode transition” or point at which
the core flow entering the combustor region changed from subsonic (ramjet) to supersonic (scramjet). Figure
13 provides a comparison of Mach number contours for two cases, both at the same fuel equivalence ratio,
φ = 0.26, but in one case the supply flow was clean air and the other has vitiated air with 10 percent water
vapor per mole. It may be observed that the clean air case is in ramjet mode while the vitiated air case is in
scramjet mode. These results are largely due to chemical kinetics effects, although to obtain this qualitative
agreement with experimental results on the equivalence ratio where mode transition occurred required close
calibration of the choices in turbulence model settings, including turbulent Prandtl and Schmidt numbers.
Figure 14 shows the strong sensitivity of the solutions to turbulent Schmidt number.
We emphasize that these choices may not be optimal for all scramjet configuration simulations. In
particular, the performance of the k − model may be considered somewhat fortuitous, and in general is not
considered an optimal choice for flows with separations. The conclusion here is that the “state-of-practice”
in RANS as applied to complete scramjet systems requires calibration of the modeling parameters for one
case, i.e. selecting fixed values for the turbulent Prandtl and Schmidt numbers and then trusting that the
trends of the solutions will remain qualitatively accurate at other operating points. As mentioned previously,
it is difficult to draw definitive conclusions on the validity of a particular CFD modeling choice for a complex
system such a scramjet when only pressure distributions as a function of axial position are available. For
this reason, usually unit problems are considered to investigate the behavior of individual model settings.
Scalar Transport Modeling

A broadly used experiment for validation of supersonic reacting flows is that of Burrows and Kurkov.60
A schematic of their experimental setup is shown in Fig. 15. Despite the fact that these experiments
were performed a number of years ago, the data obtained for this hydrogen-vitiated air test case remains
widely used today. Using the 8-step kinetics of Evans and Shexnayder discussed previously, we present
NASA/TM—2012-217277 10
results here that exhibit the sensitivity of the solutions to settings for the turbulent Prandtl and Schmidt
numbers. Figures 16 and 17 provide temperature contours in the mixing section which illustrate the effects
of variation in these scalar transport settings on the ignition point. The ignition point in the experiment
was approximately x = 22cm. Exit plane contours are shown in Fig. 18. While both quantities have
considerable effects, the turbulent Schmidt number effects are more pronounced. The Schmidt number
controls the turbulent species transport, and considering the Arrhenius rate form of the chemical kinetics
used in most RANS combustion simulations as well as the actual physical process requiring sufficient reactants
to be present for combustion to occur, it is not surprising that the turbulent Schmidt number has such an
important effect on the combustion behavior.
Unfortunately the optimal turbulent Schmidt number and turbulent Prandtl number for one case are not
likely optimal for another. Further, the current common practice of setting constant values to P rt and Sct for
modeling scalar transport is very likely inadequate for the complex flows occurring in scramjet systems. This
finding has been been reported elsewhere, such as in Ref. 7. Work has been underway to investigate variable
Prandtl/Schmidt number formulations, or “scalar variance methods” to be used for improving the modeling
of turbulent heat and mass transport in scramjets and high speed reacting jets such as those described in
Refs. 61–65. The treatment of turbulent-chemistry interactions, such as those involving probability density
function (PDF) formulations66, 67 is beyond the scope of this paper. However, one study that demonstrated
a variable Schmidt number model incorporated with a PDF approach is described in Ref. 68. While these
efforts are promising, the methods are not mature enough yet to be in broad use by the scramjet community.
Sufficient experimental data to thoroughly develop and validate these methods is also not available, yet very
much needed.
Compressible Mixing
In the scramjet combustor and into the exhaust nozzle, turbulent mixing is highly three dimensional and
subject to compressibility effects. In this section we briefly discuss the status of modeling compressible
mixing. A number of experiments may be found in the literature concerning planar mixing layers and effects of
compressibility on mixing layer growth rates. In recent years, most computational and experimental research
efforts has focused on jet flows, driven by aircraft noise considerations. For the high speed compressible jet
problem, one of the most widely used data sets is from the benchmark experiments performed at the NASA
Langley Research Center (LaRC) Jet Noise Laboratory by Seiner et al.69 These experiments investigated an
axisymmetric water-cooled Mach 2 nozzle. The nozzle operated at fully expanded conditions for this series of
cases. In our studies, the turbulence models examined were the Menter SST and Chien k- formulations. In
addition, these two models were used with and without the Sarkar compressibility correction,70, 71 as shown
in Eq. 28.
= s (1 + αMt2 ) (28)
and the turbulent Mach number is defined as:
√
Mt = 2k/a (29)
where s is the solenoidal dissipation rate solved in the dissipation equation for the k- model where
ω = /k. The coefficient, α, is set to the default value of 1.0 here for use with both turbulence models.
The Sarkar compressibility correction increases the rate of turbulent kinetic energy dissipation and then
lowers the resultant turbulent viscosity. Compressibility corrections such as the Sarkar formulation, are
most commonly employed for free shear layers or jets. For wall bounded flows, compressibility effects are
not considered to be important for Mach numbers less than about 5, as discussed by Wilcox.21 Rumsey
examined compressibility corrections applied to k − ω models for hypersonic boundary layers in Ref. 72.
In Dembowski and Georgiadis,73 a comparison of computations to experiments were made for all of the
operating points, with variations in jet stagnation temperature that were measured in the experiments of Ref.
69. Results are presented for the highest temperature case with nozzle plenum stagnation temperature set
to 1370 K in Fig. 19. For this and other operating points, the Chien k- solution and SST solutions with no
compressibility corrections provide similar solutions with faster jet mixing than indicated by the experimental
data. The Chien and SST model results obtained with the Sarkar compressibility correction indicate much
longer potential cores and slowest initial mixing of any of the solutions for all of the temperatures investigated.
In all of the cases shown in Ref. 73 where several jet operating temperatures were investigated, solutions
NASA/TM—2012-217277 11
obtained with the compressibility correction overpredict the potential core length, as shown for the highest
temperature case considered here. In results for supersonic jet calculations shown by Gross et al74, 75 the
performance of SST with and without the Sarkar correction is analogous to the results obtained here. As
shown in Ref. 76, the potential core length behavior observed here for the uncorrected models is not the
same for lower Mach number jets, where two-equation turbulence model results generally indicated longer
potential core lengths, and implied slower initial mixing rates, than found experimentally. As a result, while
compressibility corrections could be tailored, for example via the Sarkar coefficient, α, for a particular case,
there are not enough physics in such models to accurately capture compressibility effects. The conclusion
here is that compressible mixing is yet another area in which RANS turbulence modeling improvements are
still necessary.
LES-based Methods for Combustor and Exhaust System

With all of these limitations of RANS-based techniques for the scramjet combustor and exhaust nozzle re-
gions, there is hope that LES and hybrid RANS-LES may offer better prospects for improving the ability to
calculate regions dominated by turbulent mixing. There have been significant efforts underway in develop-
ing and applying LES for mixing problems, especially turbulent jets issuing from aircraft exhaust nozzles.
However there are a number of unresolved issues in LES that significantly affect the accuracy of turbulent
jet flows and other mixing dominated problems. There have been some limited attempts to use LES for
scramjet combustors, as in Refs. 9 and 10. In the combustor, the LES problem of accurately capturing
turbulent-chemistry interactions is analogous to the RANS problem. Some research into techniques such as
filtered density functions, the LES analogy to the RANS usage of PDFs, are being examined.11, 77
Considering the scramjet system as a whole, the need to accurately treat both the wall bounded regions
and the mixing layers away from the walls provides difficult challenges. At first impression, hybrid RANS-
LES methods seem to be a promising solution to this problem, and indeed they should be and are being
explored.78–80 However the interface region where RANS switches to LES, is one of the most significant chal-
lenges. In aerodynamic flows, hybrid RANS-LES has found success where the RANS regions are demarcated
from the LES regions and some physical feature is responsible for the instabilities that generate and sustain
turbulence in the LES regions.81 In the scramjet, frequently a flameholder with a rearward facing step can
serve as the physical mechanism behind which large scale eddies may be generated, to drive the necessary
unsteadiness of the LES region. Countering this, however, are the large values of eddy viscosity convected
from the thick isolator boundary layers calculated with RANS, which serve to damp the unsteadiness in the
computation.
Experimental Validation Data

The need for fundamental experiments to provide data for method development and validation is very
high. As discussed previously, while full scramjet simulations are informative, it is very difficult to perform
validation of specific CFD modeling features for complete scramjet systems. One reason is that historically
only static pressure instrumentation is employed in scramjet experiments. Further, it is nearly impossible
to parametrically isolate the benefits of one particular modeling feature, whether it be turbulence modeling,
chemical kinetics modeling, or heat transfer, when the effects of the modeling limitations in the others is
unknown. For these reasons, unit problems are more useful in model development and validation. While there
are a number of such unit problem experiments reported in the literature, most of the data available consists of
mean flow measurements. Improvements in turbulence modeling for hypersonic propulsion requires turbulent
statistical data from simultaneous measurements of velocities, temperatures, and species concentrations.
One current effort addressing this need is the US National Center for Hypersonic Combined Cycle Propul-
sion.82 This center is developing and using advanced experiments and diagnostics such as Coherent Anti-
Stokes Raman Scattering (CARS), PIV, Rayleigh techniques, and Planar Induced Flourescence (PLIF) to
obtain more detailed data including turbulent statistics and the simultaneous measurements of velocities,
temperatures, and species concentrations. This center is also developing advanced turbulent calculation
methods targeted at the supersonic combustion problem, including RANS, LES, and hybrids. NASA’s
Fundamental Aeronautics Program, the United States Air Force Office of Scientific Research and the Test
Resource Management Center’s Test and Evaluation/Science and Technology (T&E/S&T) Program are also
sponsoring experimental research efforts from geometrically simple unit flow problems to fully integrated
systems and also sponsoring computational method development and validation efforts. With hypersonic
NASA/TM—2012-217277 12
funding historically very cyclical in the United States, it is hoped that a sustained effort will be possible, as
the research needs are great.
Some recommendations for specific experimental data needs that are not currently being broadly ad-
dressed are as follows: In the area of SWTBLIs, advanced measurement techniques such as PIV have been
recently applied to fundamental experiments, such as those described earlier in this paper. However, nearly
all of these experiments have been conducted in facilities with small tunnel cross sectional areas, where
sidewall effects frequently contaminate the focus SWTBLI region. It would be highly desirable to utilize an
experimental rig where such sidewall effects could be eliminated. PIV has become a powerful and broadly
used experimental technique. Obtaining high accuracy in regions such as the beginning of a jet shear layer,
where the turbulent flow state is far from equilibrium, is quite a challenge. However, it is this region where
both RANS turbulence models and direct calculation methods (i.e. DNS and LES) need data for method
development and validation.
Thermal boundary layers obviously have significant effects on hypersonic propulsion flowpath aerother-
modynamics. While heat transfer and surface temperature measurements are being made in some select
configurations, this is the exception rather than the rule. Such measurements are very difficult to make, but
crucial in attempts to accurately represent the thermal boundary layer states as well as identifying potential
thermal deformation which in turn can alter the flowpath cross sectional area and aerodynamics. For the
combustor, detailed mappings of the turbulent flow state for velocities, temperatures, and species concentra-
tions are necessary. While some fundamental experiments are attempting to obtain these quantities, perhaps
revisiting an experiment such as the Burrows-Kurkov test case (which only provided mean flow quantities)
with these more modern experimental techniques (i.e. PIV, CARS, and PLIF) would greatly assist model-
ers in determining if a particular modeling enhancement was actually improving the fidelity of the correct
turbulent flow physics.
Conclusions
This paper presents one group’s perspective on the current state of turbulent flow calculation methods for
hypersonic propulsion flowpaths. It is not intended to serve as a comprehensive review of all computational
techniques in use today, but does discuss key issues throughout the vehicle propulsion flowpath. These
include laminar-to-turbulent boundary layer transition, shock wave / turbulent boundary layer interactions,
and modeling of the combustor and exhaust system. Emphasis was placed on Reynolds-averaged Navier-
Stokes (RANS) methods, although prospects of newer methods such as Large Eddy Simulation (LES) based
techniques were also provided. It is concluded that RANS turbulence modeling shortfalls are still a major
limitation to the accuracy of hypersonic propulsion simulations, whether considering individual components
or an overall system. Newer methods such as LES-based techniques may be promising, but are not yet at a
maturity to be used routinely by the hypersonic propulsion community.
RANS turbulence model development in the past decade has been minimal for all aerodynamic regimes.
Most of the programmatic support for turbulent calculation techniques has been toward LES-based tech-
niques. While the broader CFD community will likely be using LES more in the future as computing power
continues to improve, there is still a need in the foreseeable future for better RANS-based techniques. The
subsonic aircraft community, while also examining hybrid RANS-LES techniques, is currently beginning
to revisit pure RANS models because of the importance to aerodynamic predictions of quantities such as
vehicle drag. Support for improving some of the modeling shortfalls identified in this paper for hypersonic
propulsion flowpaths is also warranted. Finally, more detailed experimental data to help guide modeling
improvements is also needed.
NASA/TM—2012-217277 13
References
1 Drummond, J. P., Cockrell, C. E., Pellett, G. L., Diskin, G. S., Auslender, A. H., Exton, R. J., Guy, R. W., Hoppe, J. C.,
Puster, R. L., Rogers, R. C., Trexler, C. A., and Voland, R. T., “Hypersonic Air Breathing Propulsion - An Aerodynamics,
Aerothermodynamics, and Acoustics Competency White Paper,” NASA TM 2002-211951, Nov. 2002.
2 Tishkoff, J. M., Drummond, J. P., Edwards, T., and Nejad, A. S., “Future Direction of Supersonic Combustion Research:
Air Force/NASA Workshop on Supersonic Combustion,” AIAA Paper 97-1017, Jan. 1997.
3 Berry, S. A., DiFulvio, M., and Kowalkowski, M. K., “Forced Boundary-Layer Transition on X-43 (Hyper-X) in NASA
LaRC 31-Inch Mach 10 Air Tunnel,” NASA TM 2000-210315, 2002.

4 Berry, S. A., Auslender, A. H., Dilley, A. D., and Calleja, J. F., “A Hypersonic Boundary-Layer Trip Development for
Hyper-X,” Journal of Spacecraft and Rockets, Vol. 38, No. 6, Nov. 2001, pp. 853–864.
5 Baurle, R. A., “Modeling of High Speed Reacting Flows: Established Practices and Future Challenges,” AIAA Paper
2004-267, Jan. 2004.

6 Ebrahami, H. B., “An Overview of Computational Fluid Dynamics for Application to Advanced Propulsion Systems,”
AIAA Paper 2004-2370, June 2004.

7 Baurle, R. and Eklund, D., “Analysis of Dual-Mode Hydrocarbon Scramjet Operation at Mach 4-6.5,” Journal of Propul-
sion and Power , Vol. 18, No. 5, Sept. 2002, pp. 990–1002.
8 Pitsch, H., Desjardins, O., Balarac, G., and Ihme, M., “Large Eddy Simulation of Turbulent Reacting Flow,” Progress
in Aerospace Sciences, Vol. 44, No. 6, Aug. 2008, pp. 466–478.

9 Fureby, C., “LES Modeling of Combustion for Propulsion Applications,” Proceedings of the Royal Society A, Vol. 367,
July 2009, pp. 2957–2969.

10 Berglund, M., Fureby, C., Sabel’nikov, V., and Tegner, J., “On the Influence of Finite Rate Chemistry in LES of Self-
Ignition in Hot Confined Supersonic Airflow,” European Space Agency Special Report ESA SP-659, 2008.
11 Givi, P., “Filtered Density Function for Subgrid Scale Modeling of Turbulent Combustion,” AIAA Journal, Vol. 44,
No. 1, Jan. 2006, pp. 16–23.

12 Georgiadis, N. J., Yoder, D. A., Towne, C. E., Engblom, W. A., Bhagwandin, V., Power, G. D., Lankford, D. W., and
Nelson, C. C., “Wind-US Physical Modeling Improvements to Complement Hypersonic Testing and Evaluation,” AIAA Paper
2009-193, Jan. 2009.
13 Nelson, C. C. and Power, G. D., “CHSSI Project CFD-7: The NPARC Alliance Flow Simulation System,” AIAA Paper
2001-0594, Jan. 2001.

14 Nelson, C., “An Overview of the NPARC Alliance’s Wind-US Flow Solver,” AIAA Paper 2010-27, Jan. 2010.
15 Roy, C. J. and Blottner, F. G., “Review and assessment of turbulence models for hypersonic flows,” Progress in Aerospace
Sciences, Vol. 42, 2006, pp. 469–530.

16 Robinson, D. F., Harris, J. E., and Hassan, H. A., “Unified Turbulence Closure Model for Axisymmetric and Planar Free
Shear Layer Flows,” AIAA Journal, Vol. 33, No. 12, Dec. 1995, pp. 2324–2331.
17 Robinson, D. F. and Hassan, H. A., “Modeling Turbulence Without Damping Functions Using k-ζ Model,” AIAA Paper
97-2312, June 1997.

18 Alexopoulos, G. A. and Hassan, H. A., “A k-ζ (Enstrophy) Compressible Turbulence Model for Mixing Layers and Wall
Bounded Flows,” AIAA Paper 96-2039, June 1996.

19 Menter, F. R., “Zonal Two Equation k − ω Turbulence Models for Aerodynamic Flows,” AIAA Journal, Vol. 32, No. 8,
Aug. 1994, pp. 1598–1605.

20 Wilcox, D. C., “Reassessment of the Scale-Determining Equation for Advanced Turbulence Models,” AIAA Journal,
Vol. 26, No. 11, Feb. 1988, pp. 1299–1310.

21 Wilcox, D. C., Turbulence Modeling for CFD, DCW Industries, 2nd ed., 1998.
22 Rumsey, C. L. and Gatski, T. B., “Summary of EASM Turbulence Models in CFL3D with Validation Test Cases,” NASA
TM 2003-212431, June 2003.

23 Rumsey, C. L. and Gatski, T. B., “Recent Turbulence Model Advances Applied to Multielement Airfoil Computations,”
Journal of Aircraft, Vol. 38, No. 5, Sept. 2001, pp. 904–910.

24 Rumsey, C. L., Gatski, T. B., and Morrison, J. H., “Turbulence Model Predictions of Strongly Curved Flow in a U-Duct,”
AIAA Journal, Vol. 38, No. 8, Aug. 2000, pp. 1394–1402.

25 Yoder, D. A., “Initial Evaluation of an Algebraic Reynolds Stress Model for Compressible Turbulent Shear Flows,” AIAA
Paper 2003-0548, Jan. 2003.

26 Denissen, N. A., Yoder, D. A., and Georgiadis, N. J., “Implementation and Validation of a Laminar-to-Turbulent Tran-
sition Model in the Wind-US Code,” NASA TM 2008-215451, Sept. 2008.

27 Langtry, R., A Correlation-Based Transition Model using Local Variables for Unstructured Parallelized CFD Codes,
Ph.D. thesis, University of Stuttgart, 2006.

28 Langtry, R. B. and Sjolander, S. A., “Prediction of Transition for Attached and Separated Shear Layers in Turbomachin-
ery,” AIAA Paper 2002-3641, July 2002.

29 Menter, F., Ferreira, J. C., Esch, T., and Konno, B., “The SST Turbulence Model with Improved Wall Treatment for
Heat Transfer Predictions in Gas Turbines,” Proceedings of the International Gas Turbine Congress - IGTC2003-TS-059 , Nov.
2003.
30 Savill, A. M., “Some Recent Progress in the Turbulence Modeling of By-pass Transition,” Near-Wall Turbulent Flows,
edited by C. S. R.M.C. So and B. Launder, 1993, pp. 829–848.

31 Kimmel, R. L., “The Effect of Pressure Gradients on Transition Zone Length in Hypersonic Boundary Layers,” Journal
of Fluids Engineering, Vol. 119, Mar. 1997, pp. 36–41.

32 McDaniel, R. D. and Hassan, H. A., “Role of Bypass Transition in Conventional Hypersonic Facilities,” AIAA Paper
2001-0209, Jan. 2001.
NASA/TM—2012-217277 14
33 Reshotko, E., “Transition Issues at Hypersonic Speeds,” AIAA Paper 2006-707, Jan. 2006.
34 Reshotko, E., “Is Reθ / Mθ a Meaningful Transition Criteria?” AIAA Paper 2007-0943, Jan. 2007.
35 van Ingen, J. L., “The eN method for transition prediction. Historical review of work at TU Delft,” AIAA Paper 2008-
3830, June 2008.

36 Knight, D. and Degrez, G., “Shock Wave / Boundary Layer Interactions in High Mach Number Flows - A Critical Survey
of Current CFD Prediction Capability,” AGARD AR-319 , Vol. 2, 1998.

37 Knight, D., Yan, H., Panaras, A. G., and Zheltovodov, A., “Advances in CFD Prediction of Shock Wave Turbulent
Boundary Layer Interactions,” Progress in Aerospace Sciences, Vol. 39, No. 2, Feb. 2003, pp. 121–184.
38 Zheltovodov, A. A., “Some Advances in Research of Shock Wave Turbulent Boundary Layer Interactions,” AIAA Paper
2006-0496, Jan. 2006.

39 Edwards, J. R., “Numerical Simulations of Shock/Boundary Layer Interactions Using Time-Dependent Modeling Tech-
niques: A Survey of Recent Results,” Progress in Aerospace Sciences, Vol. 44, No. 6, Aug. 2008, pp. 447–465.
40 Schulein, E., “Optical Skin Friction Measurements in Short-Duration Facilities,” AIAA Paper 2004-2115, June 2004.
41 Benek, J., “Overview of the 2010 AIAA Shock Boundary Layer Interaction Workshop,” AIAA Paper 2010-4821, Jan.
2010.
42 Benek, J. and Babinsky, H., “Lessons Learned from the 2010 AIAA Shock Boundary Layer Interaction Workshop,” AIAA
Paper 2010-4825, Jan. 2010.

43 DeBonis, J. R., Oberkampf, W. L., Wolf, R. T., Orkwis, P. D., Turner, M. G., and Babinsky, H., “Assessment of CFD
Models for Shock Boundary-Layer Interaction,” AIAA Paper 2009-4823, Jan. 2010.
44 Hirsch, C., “SBLI Lessons Learned - CFD Simulations of Two Test Cases,” AIAA Paper 2010-4824, Jan. 2010.
45 Dupont, P., Piponniau, S., Sidorenko, A., and Debieve, J., “Investigation by Particle Image Velocimetry Measurements
of Oblique Shock Reflection with Separation,” AIAA Journal, Vol. 46, No. 6, June 2008, pp. 1365–1370.
46 Unsteady Effects in Shock Wave Induced Separation, edited by P. Doerffer, C. Hirsch, J.-P. Dussauge, H. Babinsky, and
G. Barakos, Springer, 2010.

47 Spalart, P. R. and Allmaras, S. R., “A One-Equation Turbulence Model for Aerodynamic Flows,” AIAA Paper 92-0439,
Jan. 1992.
48 Spalart, P. R. and Allmaras, S. R., “A One-Equation Turbulence Model for Aerodynamic Flows,” La Recherche Aerospa-
tiale, , No. 1, 1994, pp. 5–21.

49 Bhagwandin, V. A. and DeSpirito, J., “Numerical Prediction of Supersonic Shock Boundary-Layer Interaction,” AIAA
Paper 2011-859, Jan. 2011.

50 Goyne, C. P., McDaniel, J. C., Quagliaroli, T. M., Krauss, R. H., and Day, S. W., “Dual-Mode Combustion of Hydrogen
in a Mach 5 Continuous-Flow Facility,” Journal of Propulsion and Power , Vol. 17, No. 6, Nov. 2001, pp. 1313–1318.
51 Goyne, C. P., Rodriguez, C. G., McDaniel, J. C., Krauss, R. H., and McClinton, C. R., “Experimental and Numerical
Study of a Dual-Mode Scramjet Combustor,” Journal of Propulsion and Power , Vol. 22, No. 3, May 2006, pp. 481–489.
52 Rockwell, R. D., Goyne, C. P., Haw, W., Krauss, R. H., McDaniel, J. C., , and Trefny, C. J., “Experimental Study of
Test Medium Vitiation Effects on Dual-Mode Scramjet Mode Transition,” AIAA Paper 2010-1126, Jan. 2010.
53 Bhagwandin, V., Engblom, W. A., and Georgiadis, N. J., “Numerical Simulation of Hydrogen-Fueled Dual-Mode Scramjet
Engine Using Wind-US,” AIAA Paper 2009-5382, Jan. 2009.

54 Vyas, M. A., Engblom, W. A., Georgiadis, N. J., Trefny, C. J., and Bhagwandin, V., “Numerical Simulation of Vitiation
Effects on a Hydrogen-Fueled Dual-Mode Scramjet,” AIAA Paper 2010-1127, Jan. 2010.

55 Gupte, A. A., Engblom, W. A., Goyne, C. P., and Rockwell, R. D., “Effect of Thermally Induced Deformation in UVa
Supersonic Combustion Facility,” AIAA Paper 2011-503, Jan. 2011.

56 Goyne, C. P., Cresci, D., and Fetterhoff, T. P., “Short Duration Propulsion Test and Evaluation (HyV) Program,” AIAA
Paper 2009-7926, Oct. 2010.

57 Evans, J. and Shexnayder, C., “Influence of Chemical Kinetics and Unmixedness on Burning in Supersonic Hydrogen
Flames,” AIAA Journal, Vol. 18, No. 2, Feb. 1979, pp. 805–811.
58 Jachimowski, C., “An Analysis of Combustion Studies in Shock Expansion Tunnels and Reflected Shock Tunnels,” NASA
TP 3224, July 1992.

59 “Reduced Kinetic Mechanisms for Applications in Combustion Systems,” Lecture Notes in Physics, edited by N. Peters
and B. Rogg, Springer Verlag, 1993.

60 Burrows, M. and Kurkov, A., “Analytical and Experimental Study of Supersonic Combustion of Hydrogen in a Vitiated
Airstream,” NASA TM-X 2828, Sept. 1973.

61 Brinckman, K. W., Calhoon, W. H., and Dash, S. M., “Scalar Fluctuation Modeling for High-Speed Aeropropulsive
Flows,” AIAA Journal, Vol. 45, No. 5, May 2007, pp. 1036–1046.
62 Brinckman, K. W., Calhoon, W. H., Mattick, S. J., Tomes, J., and Dash, S. M., “Scalar Variance Model Validation for
High-Speed Variable Composition Flows,” AIAA Paper 2006-715, Jan. 2006.

63 Kenzakowski, D. C., Papp, J., and Dash, S. M., “Evaluation of Advanced Turbulence Models and Variable
Prandtl/Schmidt Number Methodology for Propulsive Flows,” AIAA Paper 2000-0885, Jan. 2000.
64 Chidambaram, N., Dash, S. M., and Kenzakowski, D. C., “Scalar Variance Transport in the Turbulence Modeling of
Propulsive Jets,” AIAA Paper 99-0235, Jan. 1999.

65 Xiao, X., Hassan, H. A., and Baurle, R. A., “Modeling Scramjet Flows with Variable Turbulent Prandtl and Schmidt
Numbers,” AIAA Journal, Vol. 45, No. 6, June 2007, pp. 1415–1423.
66 Hsu, A. T., Tsai, Y. L. P., and Raju, M. S., “Probability Density Function Approach for Compressible Turbulent Reacting
Flows,” AIAA Journal, Vol. 32, No. 7, July 1994, pp. 1407–1415.
67 Baurle, R. A., Hsu, A. T., and Hassan, H. A., “Assumed and Evolution Probability Density Functions in Supersonic
Turbulent Combustion Calculations,” Journal of Propulsion and Power , Vol. 11, No. 6, Nov. 1995, pp. 1132–1138.
NASA/TM—2012-217277 15
68 Keistler, P. G., Gaffney, R. L., Xiao, X., and Hassan, H. A., “Turbulence Modeling for Scramjet Applications,” AIAA
Paper 2005-5382, June 2005.

69 Seiner, J. M., Ponton, M. K., Jansen, B. J., and Lagen, N. T., “The Effect of Temperature on Jet Noise Emission,” AIAA
Paper 92-02-046, 1992.

70 Sarkar, S., Erlebacher, G., Hussaini, M. Y., and Kreiss, H. O., “The Analysis and Modeling of Dilatational Terms in
Compressible Turbulence,” Journal of Fluid Mechanics, Vol. 227, 1991, pp. 473–493.
71 Sarkar, S. and Lakshmanan, B., “Application of a Reynolds Stress Turbulence Model to the Compressible Shear Layer,”
AIAA Journal, Vol. 29, No. 5, May 1991, pp. 743–749.

72 Rumsey, C. L., “Compressibility Considerations for k-ω Turbulence Models in Hypersonic Boundary-Layer Applications,”
Journal of Spacecraft and Rockets, Vol. 47, No. 1, Jan. 2010, pp. 11–20.
73 Dembowski, M. A. and Georgiadis, N. J., “An Evaluation of Parameters Influencing Jet Mixing Using the WIND Navier-
Stokes Code,” NASA TM 2002-211727, Aug. 2002.

74 Gross, N., Blaisdell, G. A., and Lyrintzis, A. S., “Evaluation of Turbulence Model Corrections for Supersonic Jets using
the OVERFLOW Code,” AIAA Paper 2010-4604, June 2010.

75 Gross, N., Blaisdell, G. A., and Lyrintzis, A. S., “Analysis of Modified Compressibility Corrections for Turbulence
Models,” AIAA Paper 2011-279, Jan. 2011.

76 Georgiadis, N. J. and DeBonis, J. R., “Navier-Stokes Analysis Methods for Turbulent Jet Flows with Application to
Aircraft Exhaust Nozzles,” Progress in Aerospace Sciences, Vol. 42, No. 5, July 2006, pp. 377–418.
77 Madnia, C. K., Jaberi, F. A., and Givi, P., “Large Eddy Simulations of Heat and Mass Transport in Turbulent Flows,”
Handbook of Numerical Heat Transfer , edited by W. J. Minkowycz, E. M. Sparrow, and J. Y. Murthy, John Wiley and Sons,
2009.
78 Edwards, J. R., Boles, J. A., and Baurle, R. A., “LES/RANS Simulation of a Supersonic Reacting Wall Jet,” AIAA
Paper 2010-370, Jan. 2010.

79 Boles, J. A., Edwards, J. R., and Baurle, R. A., “Large-Eddy/Reynolds-Averaged Navier-Stokes Simulations of Sonic
Injection into Mach 2 Crossflow,” AIAA Journal, Vol. 48, No. 7, July 2010, pp. 1444–1456.
80 Fulton, J. A., Edwards, J. R., Goyne, C. P., McDaniel, J. C., and Rockwell, R., “Numerical Simulation of Flow in a
Dual-Mode Scramjet Combustor,” AIAA Paper 2011-3714, June 2011.

81 Georgiadis, N. J., Rizzetta, D. P., and Fureby, C., “Large-Eddy Simulation: Current Capabilities, Recommended Prac-
tices, and Future Research,” AIAA Journal, Vol. 48, No. 8, Aug. 2010, pp. 1772–1784.
82 McDaniel, J. C., Chelliah, H., Goyne, C. P., Edwards, J. R., Givi, P., and Cutler, A. D., “US National Center for
Hypersonic Combined Cycle Propulsion: An Overview,” AIAA Paper 2009-7280, Oct. 2009.
Figure 1. Schematic of scramjet-powered hypersonic vehicle with key turbulent flow physics.
NASA/TM—2012-217277 16
T3A Data
SST - fully turbulent
SST - transition, CPTM1=1.0
0.006 SST - transition, CPTM1=1.5
SST - transition, CPTM1=2.0
Laminar Correlation
Turbulent Correlation
0.004
Cf
0.002
0
0 200000 400000 600000
Re x
Figure 2. Incompressible transition comparison.
0.01
Re/m = 3.3E6
Re/m = 3.9E6
Re/m = 6.6E6
Re/m = 8.2E6
SST - fully turbulent
SST - transition, CPTM1 = 1.0
SST - transition, CPTM1 = 2.0
St
0.001
1E+06 1E+07
Rex
Figure 3. Hypersonic transition comparison.
NASA/TM—2012-217277 17
M: 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
0 100 200 300 400 500

x (mm)
Figure 4. Mach number contours for shock-wave boundary-layer interaction problem.
0.012
Schulein, et al. (1996)
Schulein (2004) [optical]
SST
0.010 Wilcox k-ω (1988 model)
Chien k-ε
ASM k-ε
ASM k-ω
0.008
Skin Friction, Cf
0.006
0.004
0.002
0.000
-0.002
300 350 400 450 500
x (mm)
Figure 5. Skin friction along bottom wall for shock-wave boundary-layer interaction problem.
NASA/TM—2012-217277 18
y
θ 
x
M∞ 
PIV Plane 
Figure 6. Schematic of typical SWTBLI experimental configuration (courtesy of DeBonis et al43 ).
Figure 7. Velocity contours for UFAST test case using SST.
NASA/TM—2012-217277 19
(a) Experiment
(b) Menter SST
(c) Spalart-Allmaras
(d) k − ωASM
Figure 8. Axial velocity contours for UFAST test case in vicinity of SWTBLI.
NASA/TM—2012-217277 20
15 15
Expt Expt
SST SST
10 SA 10 SA
k-t ASM k-t ASM
y (mm)
y (mm)
5 5
0 0
0 200 400 600 0 200 400 600
u (m/s) u (m/s)
(a) x = 280 mm (b) x = 300 mm
15 15
Expt Expt
SST SST
10 SA 10 SA
k-t ASM k-t ASM
y (mm)
y (mm)
5 5
0 0
0 200 400 600 0 200 400 600
u (m/s) u (m/s)
(c) x = 320 mm (d) x = 340 mm
Figure 9. Axial velocity profiles for UFAST test case in vicinity of SWTBLI.
NASA/TM—2012-217277 21
Figure 10. Schematic of UVA direct connect dual-mode combustor (courtesy of Rockwell et al52 ).
Figure 11. Mach number contours for scramjet flow case, fuel off.
NASA/TM—2012-217277 22
3
Expt., Clean Air
Wind-US, SST
Wind-US, Chien k- ¡
2.5
2
P / Pref
1.5
0.5
0
-40 -20 0 20 40 60
x/H
Figure 12. Pressure distributions for scramjet flow case, fuel off.
NASA/TM—2012-217277 23
(a) Clean air
(b) Vitiated air
Figure 13. Mach number contours for scramjet case, φ = 0.26.
NASA/TM—2012-217277 24
4
Expt., Clean Air

Wind-US, Prt=.9, Sct=.5
3.5 Wind-US, Prt=.9, Sct=.7
Wind-US, Prt=.9, Sct=.9
2.5
P / Pref
1.5
0.5
-40 -20 0 20 40 60
x/H
Figure 14. Turbulent Schmidt number effect on clean air scramjet case, φ = 0.26.
Figure 15. Schematic of Burrows-Kurkov experiment (taken from Burrows and Kurkov60 ).
NASA/TM—2012-217277 25
(a) Prt = 0.5, Sct = 0.7
(b) Prt = 0.7, Sct = 0.7
(c) Prt = 0.9, Sct = 0.7
Figure 16. Temperature contours for Burrows-Kurkov test case illustrating effect of P rt on ignition point.
NASA/TM—2012-217277 26
(a) Prt = 0.7, Sct = 0.5
(b) Prt = 0.7, Sct = 0.7
(c) Prt = 0.7, Sct = 0.9
Figure 17. Temperature contours for Burrows-Kurkov test case illustrating effect of Sct on ignition point.
4 4
3.5 3.5
Experiment Experiment
Sct = 0.7, Prt = 0.5 Sct = 0.5, Prt = 0.7
3 Sct = 0.7, Prt = 0.7 3 Sct = 0.7, Prt = 0.7
Sct = 0.7, Prt = 0.9 Sct = 0.9, Prt = 0.7
2.5 2.5
y (cm)
y (cm)
2 2
1.5 1.5
1 1
0.5 0.5
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
H2O Mole Fraction H2O Mole Fraction
(a) Turbulent Prandtl number variations (b) Turbulent Schmidt number variations
Figure 18. Exit profiles for Burrows-Kurkov test case.
NASA/TM—2012-217277 27
1.2 2.5
Seiner Data Seiner Data
1.1 SST SST
SST with Sarkar SST with Sarkar
1 Chien k-¡ 2 Chien k-¡
Chien k-¡ with Sarkar Chien k-¡ with Sarkar
0.9
Mach No.
0.8 1.5
u / Ujet
0.7
0.6 1
0.5
0.4 0.5
0.3
0.2 0
5 10 15 20 5 10 15 20
x/D x/D
(a) centerline axial velocity (b) centerline Mach number
1.2
Seiner Data
1.1 SST
SST with Sarkar
1 Chien k-¡
Chien k-¡ with Sarkar
0.9
0.8
Tt / TT-jet
0.7
0.6
0.5
0.4
0.3
0.2
5 10 15 20
x/D
(c) Centerline stagnation temperature
Figure 19. Comparison of solutions for heated Mach 2 jet.
NASA/TM—2012-217277 28
Form Approved
REPORT DOCUMENTATION PAGE
OMB No. 0704-0188
The public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the
data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this
burden, to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302.
Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not display a currently valid OMB
control number.
PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS.
1. REPORT DATE (DD-MM-YYYY) 2. REPORT TYPE 3. DATES COVERED (From - To)
01-04-2012 Technical Memorandum
4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER
Status of Turbulence Modeling for Hypersonic Propulsion Flowpaths
5b. GRANT NUMBER
5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S) 5d. PROJECT NUMBER

Georgiadis, Nicholas, J.; Yoder, Dennis, A.; Vyas, Manan, A.; Engblom, William, A.
5e. TASK NUMBER
5f. WORK UNIT NUMBER

WBS 031102.02.03.0829.11
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION
National Aeronautics and Space Administration REPORT NUMBER
John H. Glenn Research Center at Lewis Field E-18032
Cleveland, Ohio 44135-3191
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITOR'S

National Aeronautics and Space Administration ACRONYM(S)
Washington, DC 20546-0001 NASA
11. SPONSORING/MONITORING
REPORT NUMBER
NASA/TM-2012-217277
12. DISTRIBUTION/AVAILABILITY STATEMENT
Unclassified-Unlimited
Subject Categories: 01, 02, 07, and 34
Available electronically at http://www.sti.nasa.gov
This publication is available from the NASA Center for AeroSpace Information, 443-757-5802
13. SUPPLEMENTARY NOTES
14. ABSTRACT
This report provides an assessment of current turbulent flow calculation methods for hypersonic propulsion flowpaths, particularly the
scramjet engine. Emphasis is placed on Reynolds-averaged Navier-Stokes (RANS) methods, but some discussion of newer methods such
as Large Eddy Simulation (LES) is also provided. The report is organized by considering technical issues throughout the scramjet-powered
vehicle flowpath including laminar-to-turbulent boundary layer transition, shock wave / turbulent boundary layer interactions, scalar
transport modeling (specifically the significance of turbulent Prandtl and Schmidt numbers) and compressible mixing. Unit problems are
primarily used to conduct the assessment. In the combustor, results from calculations of a direct connect supersonic combustion experiment
are also used to address the effects of turbulence model selection and in particular settings for the turbulent Prandtl and Schmidt numbers. It
is concluded that RANS turbulence modeling shortfalls are still a major limitation to the accuracy of hypersonic propulsion simulations,
whether considering individual components or an overall system. Newer methods such as LES-based techniques may be promising, but are
not yet at a maturity to be used routinely by the hypersonic propulsion community. The need for fundamental experiments to provide data
for turbulence model development and validation is discussed.
15. SUBJECT TERMS
Hypersonics; Turbulence; Transition; Heat transfer; Scramjet; Large Eddy Simulation (LES)
16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF 18. NUMBER 19a. NAME OF RESPONSIBLE PERSON
ABSTRACT OF STI Help Desk (email:help@sti.nasa.gov)
a. REPORT b. ABSTRACT c. THIS PAGES 19b. TELEPHONE NUMBER (include area code)
U U PAGE UU 34 443-757-5802
U
Standard Form 298 (Rev. 8-98)
Prescribed by ANSI Std. Z39-18
View publication stats

Status of Turbulence Modeling For Hypersonic Propu PDF

Uploaded by

Copyright:

Available Formats

Status of Turbulence Modeling For Hypersonic Propu PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Status of Turbulence Modeling For Hypersonic Propu PDF

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Status of Turbulence Modeling for Hypersonic Propulsion Flowpaths

Article in Theoretical and Computational Fluid Dynamics · December 2013

Nicholas J. Georgiadis William A. Engblom

SEE PROFILE SEE PROFILE

Jet CFD View project

Turbulent Heat Flux View project

The user has requested enhancement of the downloaded file.

Status of Turbulence Modeling for Hypersonic

Status of Turbulence Modeling for Hypersonic

Prepared for the

National Aeronautics and

Glenn Research Center

This work was sponsored by the Fundamental Aeronautics Program

Available electronically at http://www.sti.nasa.gov

Nicholas J. Georgiadis, Dennis A. Yoder, and Manan A. Vyas

RANS Turbulence Modeling

and from these the turbulent viscosity is given as

σk2 = 1.0, β2∗ = 0.09, α2∗ = 1,

Sct = µt Cp /Dt (15)

where Sij was defined previously in Eq. 2 and Wij is:

The eddy viscosity, µt is:

Forebody/Inlet Transition Modeling

P T M = 1 − 0.94(P T M 1 + P T M 2) F3 tanh (y + /17)2

Shock Wave Turbulent Boundary Layer Interactions

Combustor / Exhaust System Modeling

Table 1. Key positions for UVA scramjet configuration

Scalar Transport Modeling

LES-based Methods for Combustor and Exhaust System

Experimental Validation Data

LaRC 31-Inch Mach 10 Air Tunnel,” NASA TM 2000-210315, 2002.

2004-267, Jan. 2004.

AIAA Paper 2004-2370, June 2004.

in Aerospace Sciences, Vol. 44, No. 6, Aug. 2008, pp. 466–478.

July 2009, pp. 2957–2969.

No. 1, Jan. 2006, pp. 16–23.

2001-0594, Jan. 2001.

Sciences, Vol. 42, 2006, pp. 469–530.

97-2312, June 1997.

Bounded Flows,” AIAA Paper 96-2039, June 1996.

Aug. 1994, pp. 1598–1605.

Vol. 26, No. 11, Feb. 1988, pp. 1299–1310.

TM 2003-212431, June 2003.

Journal of Aircraft, Vol. 38, No. 5, Sept. 2001, pp. 904–910.

AIAA Journal, Vol. 38, No. 8, Aug. 2000, pp. 1394–1402.

Paper 2003-0548, Jan. 2003.

sition Model in the Wind-US Code,” NASA TM 2008-215451, Sept. 2008.

Ph.D. thesis, University of Stuttgart, 2006.

ery,” AIAA Paper 2002-3641, July 2002.

edited by C. S. R.M.C. So and B. Launder, 1993, pp. 829–848.

of Fluids Engineering, Vol. 119, Mar. 1997, pp. 36–41.

2001-0209, Jan. 2001.

3830, June 2008.

of Current CFD Prediction Capability,” AGARD AR-319 , Vol. 2, 1998.

2006-0496, Jan. 2006.

Paper 2010-4825, Jan. 2010.

G. Barakos, Springer, 2010.

tiale, , No. 1, 1994, pp. 5–21.

Paper 2011-859, Jan. 2011.