AIAA SciTech Forum 10.2514/6.

2018-0029
8–12 January 2018, Kissimmee, Florida
2018 AIAA Aerospace Sciences Meeting

Standard Test Cases for Transition Model Verification and

Validation in Computational Fluid Dynamics

James G. Coder1
University of Tennessee, Knoxville, TN 37996

Laminar-turbulent transition modeling in RANS-based computational fluid dynamics is

becoming increasingly prevalent in simulations. Accordingly, there is a need for standardized
test cases for transition model verification and validation to support both implementation of
existing models into flow solvers and development of new models. Six test cases are put forth
in this paper for this purpose, and encompass both two- and three-dimensional flows with a
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

variety of transition mechanisms. The two-dimensional cases are the zero-pressure-gradient

flat plate, a wind-turbine airfoil, and a general-aviation airfoil. The three-dimensional cases
are a prolate spheroid, a sickle-shaped wing, and a high-speed elliptic cone. Experimental data
are available for all cases. Flow conditions for the cases are provided along with suggested
gridding guidelines, with grids being made publicly available for some of the cases.

I. Introduction

L AMINAR-TURBULENT transition modeling in computational fluid dynamics (CFD) is, at present, an immature
discipline and a topic of much ongoing research. This is due in large part to the current paradigm of CFD being
dominated by Reynolds-averaged Navier-Stokes (RANS) methods for simulating flow about practical geometries (e.g.
commercial aircraft in transonic cruise). Inviscid CFD methods, such as potential flow or Euler, may include the
effects of viscosity through coupling with external boundary-layer solvers. In such cases, transition can be readily
predicted using any of a wide variety of non-CFD-based modeling approaches. On the other end of the simulation
fidelity spectrum are highly resolved, large-eddy simulations (LES) and direct numerical simulations (DNS). Here,
the physics of the transition process may be simulated from first principles without modeling; however, this generally
requires a prohibitive amount of computational effort for practical usage.
Outside of RANS CFD, the science of laminar-turbulent transition has seen rich development in both fundamental
theory and modeling approaches. A common trait of prevailing non-CFD transition prediction models is that they are
inherently non-local. This may entail the use of integral boundary-layer properties, such as for local correlation models
[1-3] or database methods for linear stability theory [4-7], or solving the global flow stability characteristics [8].
However, this runs counter to the typical structure of RANS solvers, where the numerical region of influence is
restricted to a local stencil and parallelizable sparse-matrix operations dominate. Even fluid dynamic turbulence,
which itself has non-local characteristics, is modeled locally to fit within this computational framework. Nevertheless,
various approaches have been proposed through the years to couple RANS codes with external transition-prediction
modules, such as was done in INS2D [9] and, more recently, in the DLR TAU code [10].
Newer developments in CFD-based transition modeling have focused heavily on PDE-based methods. These may
be solved with the same numerical algorithms as the mean-flow equations and commonly used turbulence models,
which also adds compatibility with unstructured methods and amenability to massive parallelization. These models
are generally either phenomenological in nature, in which surrogate modeling variables are solved, or are physics
based with governing equations built up directly from first-principles arguments. Prominent phenomenological models
are the local-correlation models (cf. Langtry and Menter [11]) and the amplification factor transport model (cf. Coder
and Maughmer [12]), while the most prominent physics-based model is the laminar kinetic energy of approach of
Walters et al. [13]. Of these, the Langtry-Menter model has seen widespread adoption in a number of CFD codes,
including commercial, academic, and government.
In response to the potential paradigm shift from fully turbulent to transitional simulation capabilities, the CFD
Transition Modeling Discussion Group was formed within the AIAA Applied Aerodynamics Technical Committee.
The specific purpose of this group is to: 1) Bring together modelers, code developers, CFD practitioners, and
fundamental researchers; 2) Identify community needs, expectations, and restrictions for CFD transition model

1
Assistant Professor, Dept. of Mechanical, Aerospace and Biomedical Engineering, Senior Member, AIAA.

1
American Institute of Aeronautics and Astronautics

Copyright © 2018 by James G.

Coder. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.
 capabilities and implementation; and 3) Promote improvements to the state of the art of CFD transition modeling.
Pursuant to these goals, the discussion group is connected with and leverages the efforts of other groups within AIAA,
such as the FDTC Transition Discussion Group and the Turbulence Model Benchmarking Working Group (TMBWG),
who are responsible for the invaluable NASA Turbulence Modeling Resource (TMR) [14].
Building from the activities of TMBWG and the NASA TMR, it was recognized that a suite of standard test cases
is essential for model verification and validation exercises. For existing models, consistency in implementation and
results from code to code is essential. Performing a method of manufactured solutions study would be ideal, but is not
always practical. Instead, grid convergence studies can be used to assess the continuum value predicted by the
implemented model. For the same model variant, all CFD codes should agree with each other. Any discrepancies in
continuum results as compared to vetted implementations should then be investigated. A particular challenge of this
endeavor for transition models is that they often treat the various transition mechanisms in a segregated fashion. That
is, not all model terms and functional behaviors are guaranteed to be active for every test case. Consequently, a variety
of test cases is necessary to verify a model’s implementation. It is also preferable for the test cases to be simple enough
so as to not burden one’s computational resources. For the development of new models, it is always beneficial to have
test cases with high-quality experimental data available for comparison, including both localized quantities (e.g. skin-
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

friction distributions, surface pressures, transition locations, etc.) as well as integrated quantities, such as forces and
moments, if possible.
As an outgrowth of the discussion group, a special conference session on CFD Transition Modeling and Predictive
Capabilities has been organized for the AIAA SciTech 2018 conference held in January 2018 in Kissimmee, FL. The
sessions are intended to be a forum for discussing best practices in transitional flow simulations, including grid
resolution studies and flow-solver strategies. For this special session, a set of two- and three-dimensional test cases
were defined, and participants were requested to perform at least one two-dimensional and one three-dimensional
case. All types of transition prediction methods are of interest as long as they can be coupled to a CFD code, and the
goal is for all final flow solutions to be RANS in nature.
This paper describes these proposed test cases for laminar-turbulent transition modeling: 1) a zero-pressure-
gradient flat plate [15]; 2) the S809 wind-turbine airfoil [16]; 3) the NASA NLF(1)-0416 general aviation airfoil [17];
4) an inclined, 6:1 prolate spheroid [18]; 5) the TU Braunschweig sickle wing [19]; and 6) the HIFiRE-5b high-speed
elliptic cone [20,21]. Key aspects of their underlying motivation and development are included, and references to
relevant experimental data are provided where available. Due to the inherent dependency on the underlying turbulence
model in transition flow simulations, it is intended that these cases be used in addition to the TMR turbulence model
verification cases [14].

II. Summary of Test Cases

Case 1 – Zero-Pressure-Gradient Flat Plate

The first two-dimensional test case that was defined is the one that is the most geometrically simple: the zero-
pressure-gradient flat plate. A grid family for this case is available through the NASA TMR that is intended for fully
turbulent simulations. The grids are suitable for transition model verification and validation for low free-stream
turbulence intensities (FSTI), such as for the conditions of the Schubauer and Klebanoff experiment [22]; however,
they are ill-suited for high FSTI cases where bypass transition dominates, such as the ERCOFTAC T3 series [15]. For
one, the streamwise resolution of the grids is insufficient to adequately capture the transition region at the design
Reynolds number. While it would be feasible to lower the Reynolds number, the wall-normal spacing then becomes
excessively fine, which hinders the ability to infer meaningful best practices based on the results.
Accordingly, a new grid family has been generated specifically for bypass transition model testing, and inlet
conditions have been defined for the T3A and T3B cases. The recommended boundary conditions are a nozzle inlet
condition with specified total conditions corresponding to a free-stream Mach number of 0.2, a Riemann characteristic
upper boundary, a constant-pressure outlet (P/P¥ = 1.0), symmetry on the lower boundary upstream of the plate, and
viscous adiabatic walls on the plate itself. The medium grid is pictured in Fig. 1 and labeled with the boundary
conditions. The sizes for all meshes in the new family are listed in Table 1, and the conditions for the two cases (T3A
and T3B) are described in Table 2. The inlet turbulence conditions were determined based on measured free-stream
turbulence decay, the analytic behavior of the Menter SST model [23] (which is the basis for the widely used Langtry-
Menter transition model), and the size of the computational domain. A comparison of the measured and predicted
intensities using the Langtry-Menter model in the OVERFLOW 2.2n solver on the finest grid are plotted in Fig. 2 for
both cases.

2
American Institute of Aeronautics and Astronautics
 Table 1. Zero-pressure-gradient flat plate grid dimensions

Upstream
Grid Level Dimensions Symmetry Length Number of Points
Tiny 45 x 25 13 pts 1,125
Coarse 89 x 49 25 pts 4,361
Medium 177 x 97 49 pts 17,169
Fine 353 x 193 97 pts 68,129
Extra-Fine 705 x 385 193 pts 271,425

Table 2. Zero-pressure-gradient flat plate T3A and T3B case description.

Inlet Mach number 0.2

Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Reynolds number 200,000 per grid unit

Reference temperature 540 R
Reference pressure 14.7 psi
Inlet turbulence conditions Case 1A (T3A):
• Turb. intensity = 5.855%
• µT/µ = 11.9
Case 1B (T3B)
• Turb. intensity = 7.216%
• µT/µ = 99

Figure 1. Schematic of zero-pressure-gradient flat plate grid (medium-resolution grid shown).

3
American Institute of Aeronautics and Astronautics
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Figure 2. Comparison of measured and predicted free-stream turbulence decay for the T3A and T3B cases.

Case 2 – S809 Wind-Turbine Airfoil

The second two-dimensional test case is the S809 airfoil, which was designed by Somers in the 1980s for stall-
regulated wind turbines and subsequently tested in the TU Delft Low-Speed Wind Tunnel [16]. Due to the availability
of high-quality wind-tunnel measurements and its wind-turbine applications, it has become a favorite for transition
model development [9,11,24]. Whether or not it was the conscious intent of these researchers, the S809 has flow
features that make it especially challenging for transitional CFD simulations. For one, the airfoil features a sharp upper
corner of the low-drag lift coefficient range (the so-called “drag bucket”), which is indicative of a rapid change in
transition location over a relatively small angle-of-attack range. This feature can be seen in the profile-drag polar
plotted in Fig. 3. Additionally, the airfoil features a short but intense and interactive laminar separation bubble on each
the upper and lower surface throughout the drag bucket at a Reynolds number of 2 million. This is evident in the
surface pressure distribution for a 1° angle of attack plotted in Fig. 4, which is originally from Ref. [16]. For typical
grid generation approaches, and without a priori knowledge of the flow behavior, inadequate resolution of the bubble
may result. That is, the separation bubble may be resolved across a single cell in the streamwise direction, which leads
to inaccurate predictions and/or non-physical solution unsteadiness.
For this case, a grid-refinement study with at least three resolutions is prescribed for angles of attack of 1°, 6°, and
9° for the conditions described in Table 3. The 1° angle of attack was chosen as being in the middle of the drag bucket,
which tests the ability of models to capture the laminar separation bubble. The 6° angle of attack was chosen as it is
just beyond the upper corner of the drag bucket where the transition location is rapidly moving. This is expected to
amplify the effects of grid refinement on the predictions. Lastly, the 9° angle of attack was chosen as being close to
stall as observed in the experiment.
Recommended grid spacings for a medium-resolution mesh are also provided in Table 3 for this case, but a family
of structured, two-dimensional, C-type grids has been generated and made available. To generate this grid sequence,
a hyperfine mesh (not actually included in the sequence) was first generated using a conformal mapping method. Then,
coarser meshes were generated using every 2nd, 3rd, 4th, 6th, 8th, and 12th point of the hyperfine mesh. With the inherent
elliptical nature of the conformal mapping, all coarser levels retain the smoothness and orthogonality of the original.
These grids are pictured in Fig. 5, and their dimensions are listed in Table 4.

4
American Institute of Aeronautics and Astronautics
 Table 3. S809 wind-turbine airfoil case description.

Mach number 0.1

Angle of attack 1°, 6°, and 9°
Reynolds number 2 million (based on chord length)
Reference temperature 540 R
Reference pressure 14.7 psi
Freestream turb. Intensity 0.07% (or Ncrit = 9)

Gridding guidelines Recommended medium-grid spacings:

LE spacing: 0.0010 chords
TE spacing: 0.0005 chords
Dy = 6.5 x 10-6 chords

Growth rates < 1.2 normal to wall

Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Far-field boundaries at least 1000 chords

from surface
Comments Participants should run all three angles of
attack for at least three grid levels (coarse,
medium, fine) generated based on
medium-grid guidelines.

Quantities of interest include section lift,

drag, and pitching-moment coefficients,
surface pressure distributions, and
transition locations (if predicted).

Free-stream eddy-viscosity ratio may

vary depending on modeling approach
and best practices

Table 4. S809 structured-grid family dimensions.

Grid Level Dimensions Wake-Cut Length Number of Points

Tiny 353 x 49 49 pts 17,297
Coarse 529 x 73 73 pts 38,617
Medium 705 x 97 97 pts 68,385
Fine 1057 x 145 145 pts 153,265
Extra-Fine 1409 x 193 193 pts 271,937
Ultra-Fine 2113 x 289 289 pts 610,657

5
American Institute of Aeronautics and Astronautics
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Figure 3. S809 experimental drag polar for Re = 2 x 106 (from Ref. [16]).

Figure 4. S809 experimental and theoretical surface pressure distribution for Re = 2 x 106 and 1° angle of attack (from
Ref. [16]).

6
American Institute of Aeronautics and Astronautics
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Figure 5. S809 structured-grid family.

Case 3 – NASA NLF(1)-0416 General Aviation Airfoil

The third two-dimensional test case is the NASA NLF(1)-0416, which was designed by Somers as a general
aviation airfoil and tested in the NASA Langley Low-Turbulence Pressure Tunnel (LTPT) [17]. This test case is
intended to serve as a higher-Reynolds number complement the S809 case, as experimental data are available for
Reynolds numbers up to 9 million. For these Reynolds number, upper surface transition is typically natural (i.e. not
through a separation bubble), while transition on the lower surface is through a small bubble in the cove region. The
presence of natural transition in the low-drag lift-coefficient range is in contrast to the S809, and requires transition
models to be applicable at moderate-to-high Reynolds numbers (note that laminar separation bubbles are the “free
lunch” of CFD-based transition models).
Details of the prescribed conditions for the test case are provided in Table 5. This includes a grid resolution study
for two angles of attack, and an alpha-sweep using the medium-resolution grid, all at a Reynolds number of 4 million.
Recommended grid spacings are also included in Table 5, and like the S809, a family of structured, C-type grids were
generated and made available. This grid family was also generated using a conformal mapping method, and are
pictured in Fig. 6 with the dimensions listed in Table 6 (note that these are identical to those for the S809).

7
American Institute of Aeronautics and Astronautics
 Table 5. NLF(1)-0416 general aviation airfoil case description.

Mach number 0.1

Angle of attack Case 3A (grid resolution study): 0° and 5°
Case 3B (alpha sweep): [-4°,8°] in 2°
increments
Reynolds number 4 million (based on chord length)
Reference temperature 540 R
Reference pressure 14.7 psi
Freestream turb. Intensity 0.15% (or Ncrit = 7.2)
Gridding guidelines Recommended medium-grid spacings:
LE spacing: 0.0010 chords
TE spacing: 0.0005 chords
Dy = 3.5 x 10-6 chords
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Growth rates < 1.2 normal to wall

Far-field boundaries at least 1000 chords
from surface
Comments Participants should run Case 3A on at
least three grid levels (coarse, medium,
fine) generated based on medium-grid
guidelines and Case 3B on at least the
medium grid.

Quantities of interest include section lift,

drag, and pitching-moment coefficients,
surface pressure distributions, and
transition locations (if predicted).

Free-stream eddy-viscosity ratio may

vary depending on modeling approach
and best practices

Table 6. NLF(1)-0416 structured-grid family dimensions.

Grid Level Dimensions Wake-Cut Length Number of Points

Tiny 353 x 49 49 pts 17,297
Coarse 529 x 73 73 pts 38,617
Medium 705 x 97 97 pts 68,385
Fine 1057 x 145 145 pts 153,265
Extra-Fine 1409 x 193 193 pts 271,937
Ultra-Fine 2113 x 289 289 pts 610,657

8
American Institute of Aeronautics and Astronautics
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Figure 6. NLF(1)-0416 structured-grid family.

Case 4 – Inclined, 6:1 Prolate Spheroid

The first three-dimensional test case is the 6:1 prolate spheroid, which is a defined as

1
x 2 + 36y 2 + 36z 2 =
4 (1)

where x spans [-0.5,0.5]. This shape serves as a geometrically simple surrogate for axisymmetric (or nearly so)
aerodynamic bodies, such as fuselages. At non-zero angles of attack, the flow field is strongly three-dimensional. The
three-dimensionality also extends to the boundary layer; however, the details of the behavior are quite different from
those on a wing [25]. It has been studied experimentally by DVFLR (now DLR) [18], and the transition behavior has
been studied numerically by various researchers [25-27]. The strong influence of crossflow on transition for this case
places a strain on the predictive capability of CFD based transition models, leading to pronounced discrepancies
between prediction and measurement as typified by the results shown in Fig. 7.
Flow conditions for this test case are listed in Table 7, and consist of three angles of attack at constant free-stream
Reynolds and Mach number. The free-stream turbulence intensity is specified; however, it may be desirable to impose
streamwise and crossflow critical amplification factors based on the experiment. Gridding guidelines have also been
provided in Table 7. A single-block structured grid with dimensions 257x129x129 has been made available, and is
pictured in Fig. 8. This grid represents only the right half of the cylinder, thus requiring a symmetry-plane boundary
condition on four computational boundaries. A family of unstructured, fully three-dimensional grids have been made
available by DLR [28]. Grid sizes for members of this family are defined in Table 8, noting that the nominal grid size
is defined as “prism layers X surface nodes”. While a grid-refinement study is not explicitly specified for this case, it
is an overall good exercise to undertake whenever possible. The provided unstructured grids are pictured in Fig. 9 for
all refinement levels.

9
American Institute of Aeronautics and Astronautics
 Table 4. Inclined, 6:1 prolate spheroid case description.

Mach number 0.13

Angle of attack 5°, 10°, and 15°
Reynolds number 6.5 million (based on spheroid length)
Reference temperature 540 R
Reference pressure 14.7 psi
Freestream turb. Intensity 0.15% (or, critical Tollmien-Schlichting
and crossflow amplification factors
calibrated based on wind-tunnel data)

Gridding guidelines Recommended spacings:

LE/TE spacing: 0.1% of spheroid length
Dy = 2.3 x 10-6 spheroid lengths
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Growth rates < 1.2 normal to wall

Far-field boundaries at least 100 lengths
from surface
Comments Quantities of interest include surface skin
friction distributions, surface pressure
distributions, and transition locations (if
predicted).

Structured-overset and unstructured grid

systems to be made available, though
participants are encouraged to generate
their own.

Free-stream eddy-viscosity ratio may

vary depending on modeling approach
and best practices

Table 8. Prolate spheroid unstructured grid dimensions.

Grid Level Prism Layers Surface Nodes Total Nodes

032x001250 32 1,253 53,938
048x002813 48 2,814 159,936
064x005000 64 5,002 363,532
096x011250 96 11,259 1,164,171
128x020000 128 20,026 2,715,698
192x045000 192 45,001 8,949,137
256x080000 256 80,062 21,071,815

10
American Institute of Aeronautics and Astronautics
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Figure 7. Example of discrepancies between experimental and computational skin friction based on choice of transition
model (from Ref. [25]).

Figure 8. Medium-resolution single-block structured mesh for 6:1 prolate spheroid.

11
American Institute of Aeronautics and Astronautics
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Figure 9. Unstructured, 064x005000 prolate spheroid grid.

Case 5 – TU Braunschweig Sickle Wing

The second three-dimensional test case is the TU Braunschweig Sickle Wing, which was design and
experimentally tested by Petzold and Radespiel [19]. This geometry, picture schematically in Fig. 10, features
progressively increasing sweep angles and was design as a transition-prediction validation case for spanwise-varying
crossflow. Both surface roughness and tunnel turbulence intensity were documented as driving factors in the
boundary-layer receptivity. Transition locations were measured using infrared thermography, and the model skin was
specifically designed to enhance the contrast for these measurements. Details of the geometry and a variety of test
data have been made available through the AIAA CFD Transition Modeling Discussion Group.
Four flow conditions, including transition prediction guidelines based on linear stability theory, are specified for
this test case and are described in Table 9. For this grid, only one solution resolution is required, and the gridding
guidelines are also included in Table 9.

Figure 10. Diagram of the TU Braunschweig Sickle Wing wind-tunnel geometry.

12
American Institute of Aeronautics and Astronautics
 Table 9. TU Braunschweig sickle wing case description.

Case A B C D
Mach number 0.156 0.259 0.259 0.158
Angle of attack -2.6° -2.6° -0.3° +6.0°
Reynolds number 2.75 x 106 4.45 x 106 4.43 x 106 2.75 x 106
Reference temperature (°C) 22.62 28.38 28.98 25.98
Reference pressure (Pa) 98919. 98974. 103614. 100801.
Freestream turb. Intensity Tux=0.05% @ 50m/s
critical N-Factors for linear stability & eN method
(specific for sickle wing model)
NTScrit=11.60 ± 0.80
NCFcrit=8.4 ± 0.75

Gridding guidelines Typical surface grid resolution (recommended)

Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

400 pts chordwise (200 pts per side)

400 pts spanwise
Clustering near kinks advised.
Boundary layer resolution:
Depending on transition prediction method.
For linear stability solver, a minimum of 40pts
inside BL is recommended.

Interference effects with WT side walls are low for

cases A, B, C – but not for case D. Thus, WT walls
should be included in CFD simulations.
(see sample grid provided)

Comments Quantities of interest include transition locations

over span, transition mechanism near kinks and
surface pressure distributions.

CAD definition and sample unstructured grid

available.
IR images and extracted transition lines available.
Provided cP distributions come without any wall
corrections.

Remark: Transition over inboard section (φ=30°)

is affected by increasing disturbance level towards
WT floor.

13
American Institute of Aeronautics and Astronautics
 Case 6 – HIFiRE-5b
A common trait of the previous test cases is that they were for low-Mach number flows; however, laminar-
turbulent transition is very much relevant in the high-speed, hypersonic flow regime [20,21,29]. It is nevertheless
desirable to be able to seamlessly predict transition in the CFD environment for hypersonic flows. Therefore, the third
three-dimensional test case was selected to be the HIFiRE-5b, which is a high-speed, elliptic cone as pictured in Fig.
11. Such a configuration may feature a wide variety of transition mechanisms, including first- and second-mode
instability, crossflow instabilities, and attachment-line instabilities, as well as interactions between the mechanisms.
The geometry of the cone is analytically defined, as found in Ref. [20], and a Pointwise database file is available upon
request. Experimental transition data obtained during flight testing are described in Ref. [21].
Details of the flow condition of interest are described in Table 10 along with specified gridding guidelines. Of
these, there are some very important details that cannot be overlooked. First, a constant-temperature wall boundary
condition is specified for the cone. Consequently, the gridding requirements are more stringent than for adiabatic walls
due to the strong temperature gradients near the surface. Second, it is regarded that only the payload geometry needs
to be defined and that the rest of the stack may be neglected. This is motivated by the fact that the flow is hypersonic,
and the only upstream propagation of information is in the boundary layer below the sonic line. If the computational
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

outflow coincides with the end of the cone, it is furthermore advised that highly stretched “sponge layers” be used
near the boundary to mitigate deleterious boundary influences. Last, the high free-stream Mach number may place a
strain on typical numerical schemes. It is advised that either shock-fitting or a specially tailored numerical flux be
employed for this case.

Figure 11. HIFiRE-5b elliptic cone diagram and coordinate system.

14
American Institute of Aeronautics and Astronautics
 Table 10. HIFiRE-5b, high-speed, elliptic cone case description.

Mach number 7.80

Angle of attack 1.2°
Reynolds number 8.99 x 106 per meter
Reference temperature 388 R
Reference pressure 0.5 psi
Wall temperature 671.67 R
Freestream turb. Intensity 0.07% (or, critical First Mode, Second
Mode, and/or Centerline amplification
factors calibrated based on existing
databases)

Gridding guidelines Recommended spacings:

Dy = 2.54 x 10-7 meters
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Growth rates < 1.2 normal to wall

Truncate the domain at the aft end of the

cone with highly stretched cells near the
outflow boundary.
Comments Quantities of interest include surface skin
friction distributions, surface pressure
distributions, and transition locations (if
predicted).

The geometric definition may be found in

AIAA Paper 2010-4985, and a Pointwise
database definition is available on
request. Structured-overset and
unstructured grid systems to be made
available, though participants are
encouraged to generate their own.

Free-stream eddy-viscosity ratio may

vary depending on modeling approach
and best practices

III. Conclusion
Six test cases have been put forth for verification and validation of laminar-turbulent transition models in the CFD
environment. These include a zero-pressure-gradient flat plate, two airfoils, a prolate spheroid, a sickle-shaped wing,
and a high-speed cone. This assortment of cases encompasses a wide variety of transition mechanisms, including
Tollmien-Schlichting instabilities, laminar separation bubbles, bypass transition, crossflow instabilities, and high-
speed first and second mode instabilities. By no means are these cases intended to be truly complete with respect to
the full behavior of individual transition mechanisms, and these cases will likely evolve to meet ever-changing
demands. However, it is hoped that they will provide a benefit to the transition modeling community by improving
model implementation and fostering new model developments.

Acknowledgments
The author would like to express sincere gratitude to those who helped and consulted in the development of these
test cases, including but not limited to Ryan Glasby and Doug Stefanski of the UT/ORNL Joint Institute for
Computational Sciences, Matthew Tufts and Roger Kimmel of the Air Force Research Laboratory, Stefan Wernz of
Raytheon, Andreas Krumbein of DLR, and Rolf Radespiel of TU Braunschweig.

15
American Institute of Aeronautics and Astronautics
 References
[1] Michel, R., “Etude de la Transition et Calcul de la Trainee des Profiles d’aile en Incompressible,” ONERA Publ. 58, 1951.
[2] Eppler, R. and Somers, D. M., “A Computer Program for the Design and Analysis of Low-Speed Airfoils,” NASA TM-80210,
1980.
[3] Abu-Ghannam, B. J. and Shaw, R., “Natural Transition of Boundary Layers: The Effects of Turbulence, Pressure Gradient,
and Flow History,” Journal of Mechanical Engineering Science, Vol. 22, No. 5, 1980, pp. 213-228.
[4] Smith, A. M. O. and Gamberoni, N., “Transition, Pressure Gradient, and Stability Theory,” Douglas Aircraft Rept. ES-26388,
Long Beach, CA, 1956.
[5] van Ingen, J. L., “Suggested Semi-Empirical Method for the Calculation of the Boundary Layer Transition Region,” Dept. of
Aerospace Engineering, Delft. Univ. of Technology Rept. VTH-74, Delft, The Netherlands, 1956.
[6] Stock, H. W. and Degenhart, E., “A simplified en method for transition prediction in two-dimensional, incompressible
boundary layers,” Z. Flugwiss. Weltraumforsch., Vol. 13, 1989, pp. 16-30.
[7] Arnal, D., “Transition Prediction in Transonic Flow,” IUTAM Symposium Transsonicum III DFVLR-AVA, Springer-Verlag,
Berlin, 1988, pp. 253-262.
[8] Theofilis, V., “Global Linear Instability,” Annual Review of Fluid Mechanics, Vol. 39, 2011, pp. 319-352.
[9] Brodeur, R. R. and van Dam, C. P., “Transition Prediction for a Two Dimensional Navier-Stokes Solver Applied to Wind-
Downloaded by AUBURN UNIVERSITY on January 10, 2018 | http://arc.aiaa.org | DOI: 10.2514/6.2018-0029

Turbine Airfoils,” 2000 ASME Wind Energy Symposium / AIAA Aerospace Sciences Meeting, AIAA Paper 2000-0047,
Reno, NV, Jan. 2000.
[10] Krumbein, A., Krimmelbein, N., and Schrauf, G., “Automatic Transition Prediction in Hybrid Flow Solver, Part I:
Methodology and Sensitivities,” Journal of Aircraft, Vol. 46, No. 4, 2009, pp. 1176-1190.
[11] Langtry, R. B. and Menter, F. R., “Correlation-Based Transition Modeling for Unstructured Parallelized Computational Fluid
Dynamics Codes,” AIAA Journal, Vol. 47, No. 12, 2009, pp. 2894-2906.
[12] Coder, J. G. and Maughmer, M. D., “Computational Fluid Dynamics Compatible Transition Modeling Using an Amplification
Factor Transport Equation,” AIAA Journal, Vol. 52, No. 11, 2014, pp. 2506-2512.
[13] Walters, D. K. and Leylek, J. H., “A New Model for Boundary Layer Transition Using a Single-Point RANS Approach,”
Journal of Turbomachinery, Vol. 126, No. 1, 2004, pp. 193-202.
[14] Rumsey, C. L., Smith, B. R., and Huang, G. P., “Description of a Website Resource for Turbulence Modeling Verification
and Validation,” 40th AIAA Fluid Dynamics Conference, AIAA Paper 2010-4742, Chicago, IL, June-July 2010.
[15] Savill, A. M., “Some Recent Progress in the Turbulence Modeling of By-Pass Transition,” Near-Wall Turbulent Flows, ed.
R. M. C. So, C. G. Speziale, and B. E. Launder, Elsevier, New York, 1993, p. 829.
[16] Somers, D. M., “Design and Experimental Results for the S809 Airfoil,” NREL/SR-440-6918, 1997.
[17] Somers, D. M., “Design and Experimental Results for a Natural-Laminar-Flow Airfoil for General Aviation Applications,”
NASA TP-1861, 1981.
[18] Kreplin, H.-P., Vollmers, H., and Meier, H. U., “Wall shear stress measurements on an inclined prolate spheroid in the DFVLR
3 m x 3 m low spee wind tunnel,” Gottingen, DFVLR-AVA, report IB 22-84 A 33, 1985.
[19] Petzold, R. and Radespiel, R., “Transition on a Wing with Spanwise Varying Crossflow and Linear Stability Analysis,” AIAA
Journal, Vol. 53, No. 2, 2015, pp. 321-335.
[20] Kimmel, R. L., Adamczak, D., Berger, K., and Choudhari, M., “HIFiRE-5 Flight Vehicle Design,” 40th AIAA Fluid Dynamics
Conference, AIAA Paper 2010-4985, Chicago, IL, June-July 2010.
[21] Kimmel, R. L., et al., “HIFiRE-5b Flight Overview,” 47th AIAA Fluid Dynamics Conference, AIAA Paper 2017-3131,
Denver, CO, June 2017.
[22] Schubauer, G. B. and Klebanoff, P. S., “Contributions on the Mechanisms of Boundary-Layer Transition,” NACA TN-3489,
1955.
[23] Menter, F. R., “Two-Equation Eddy-Viscosity Turbulence Models of Engineering Applications,” AIAA Journal, Vol. 32, No.
3, 1994, p. 1598-1605.
[24] Medida, S. and Baeder, J. D., “Application of the Correlation-based g-Reqt Transition Model to the Spalart-Allmaras
Turbulence Model,” 20th AIAA Computational Fluid Dynamics Conference, AIAA Paper 2011-3979, Honolulu, HI, June
2011.
[25] Grabe, C., Shengyang, N., and Krumbein, A., “Transition Transport Modeling for the Prediction of Crossflow Transition,”
34th AIAA Applied Aerodynamics Conference, AIAA Paper 2016-3572, Washington, D.C., June 2015.
[26] Krimmelbein, N. and Krumbein, A., “Automatic Transition Prediction for Three-Dimensional Configurations with Focus on
Industrial Applications,” Journal of Aircraft, Vol. 48, No. 3, 2011, pp. 1878-1887.
[27] Langtry, R. B., Sengupta, K., Yeh, D. T., and Dorgan, A. J., “Extending the g-Reqt Local Correlation Based Transition Model
for Crossflow Effects,” 45th AIAA Fluid Dynamics Conference, AIAA Paper 2015-2474, Dallas, TX, June 2015.
[28] Krumbein, A., Private Communication, 2017.

16
American Institute of Aeronautics and Astronautics