Advances in Theoretical and Applied Mechanics, Vol. 11, 2018, no.

1, 49 - 59
HIKARI Ltd, www.m-hikari.com
https://doi.org/10.12988/atam.2018.8104

Probabilistic Study of the Aerodynamic

Around a 3D Wing
Soufiane Elouardi, Rabii El Maani and Bouchaib Radi

LIMII, FST Settat, Route de Casablanca, Settat, Morocco

LSMI, ENSAM Meknès, Marjane 2, Meknès, Morocco

Copyright c 2018 Soufiane Elouardi, Rabii El Maani and Bouchaib Radi. This article
is distributed under the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original work is properly
cited.

Abstract
Aerodynamics can be defined as the science that deals with the ma-
nipulation of interactions between air and a structure. During this ma-
nipulation and control, numerical simulation is used as a tool for pre-
dicting the phenomena encountered in these interactions. Although dis-
persion and uncertainty occur naturally in all aspects of an analysis and
deterministic analyses do not take it into account. Probabilistic charac-
terization quantifies the reliability or quality of the product by means
of a statistical analysis. In this paper, we present the validation of a
numerical simulation of a 3D transonic flow on the ONERA M6 wing
for which the numerical results, realized with ANSYS/FLUENT , c will
be compared with the experimental data and the NASA CFD results.
Then statistical distribution functions are evaluated to describe the un-
certain parameters relative to the fluid around the wing and determine
the influence of the choosen random quantities on the results of the
aerodynamic analysis.

Keywords: Aerodynamics, CFD, probabilistic analysis, uncertainties

1 Introduction
To begin, aerodynamics can be defined as the science that treats the air flow
on different mobiles, in particular, airplane wings that are a very interesting
application of aerodynamic systems [1, 4]. The number of searches in this area
50 Soufiane Elouardi, Rabii El Maani and Bouchaib Radi

is increasing rapidly due to new technologies and robust design and simula-
tion methods. Since the use of aerodynamics through mathematical analysis,
empirical approximations, wind tunnel experimentation and computer simu-
lations have provided a rational basis for flight development [6, 16]. Recent
research work has been developed on different aspects such as turbulence, opti-
mization, and boundary layers [17]. These studies can cover the entire velocity
range of low velocity, transonic, supersonic, and hypersonic flows as well as
turbulence and flow control. System properties such as mass flow rates and
pressure drops, as well as dynamic fluid forces such as lift, drag and pitch
moment, can be easily calculated in addition to wake effects [2, 9, 8, 14]. For
wing design, the requirement for new tools capable of accurately predicting
aerodynamic behavior is realized.
Numerical simulation of computational fluid dynamics can be applied for
the early detection of adverse effects on stability and control behavior [18, 12].
At the same time, uncertainty is an inevitable problem in the field of research.
Because planes have a complicated operating environment and the sophisti-
cated mechanical structure itself. Material property is one of the uncertainties
in the aircraft system. As well as in the practical design of the aircraft wing,
there are uncertainties that cause fluctuations in the performance of the air-
craft, so it is necessary to anticipate these uncertainties in the design and
therefore, it is essential to quantify the uncertainties in simulation and design
to analyze the risk of problems of uncertainties related to manufacturing, in-
frastructure, and design [11]. Deterministic modeling is not suitable to take
into account stochastic parameters, but with the development of computer
technology, CFD (Computational Fluid Dynamic) simulation has been widely
used to solve the problems of aerodynamic flows [3, 7].
In a different context, the uncertainties of the input and model param-
eters have different characteristics. The most popular methods in research
include probability theory, evidence theory, possibility theory, interval analy-
sis, and convex modeling. Among them, the theory of probabilities is a theory
more widespread or better known to engineers. Its relative advantages are
due to a solid theoretical basis and a long development time [19]. In this
paper, we present the validation of a numerical simulation of a 3D transonic
flow on the ONERA M6 wing for which the numerical results, realized with
ANSYS/FLUENT , c will be compared with the experimental data and the
results of the NASA CFD for different turbulence models [13]. The lift (Cl ) and
drag (Cd ) cœfficients along the surfaces of the upper and lower wings are given
in the deterministic case. Then, statistical distribution functions (Gaussian
distribution, uniform distribution, etc.) are evaluated to describe the uncer-
tain parameters relating to the fluid around the wing and to determine the
influence of the chosen random quantities on the results of the aerodynamic
analysis.
Probabilistic study of the aerodynamic around a 3D wing 51

2 Computational Fluid Dynamics (CFD)

With the development of computer technology, CFD technology has been
widely used to solve the problems of aerodynamic mechanics. Traditional CFD
simulation is deterministic. However, a variety of uncertainties are inevitable
in CFD simulation with the increasing complexity of the fluid problem leading
to the mismatch between CFD simulation results and real results. For numer-
ical simulation, the fluid flow problem is defined by the conservation laws of
mass, momentum, and energy. These laws are expressed in terms of partial
differential equations that are discretized by a finite element technique [10].

2.1 Equation of continuity

The equation for conservation of mass, or continuity equation, can be written
as follows:
∂ρ
+ ∇ · (ρ→
−v ) = Sm (1)
∂t

Where ρ is the density, v is the velocity and the source Sm is the mass added
to the continuous phase from the dispersed second phase and any user-defined
sources.

2.2 Moment Equation

Conservation of momentum in an inertial (non-accelerating) reference frame is
described by:

∂ → →
−
(ρ−
v ) + ∇ · (ρ→
−
v→−
v ) = −∇p + ∇ · (τ ) + ρ→
−
g +F (2)
∂t

Where p is the static pressure, τ is the stress tensor (described below) and ρg
→
−
and F are the gravitational body force and external body forces, respectively.

2.3 Spalart-Allmaras turbulence model

The Spalart-Allmaras model is an equation model that solves a modeled trans-
port equation for kinematic eddy viscosity [15]. The Spalart-Allmaras model
was designed specifically for aerodynamic flows and aerospace applications
involving wall-bounded flows and has been shown to work well for boundary
layers with adverse pressure gradients. The transported variable in the Spalart-
Allmaras model ν̃ is identical to the turbulent kinematic viscosity except in
52 Soufiane Elouardi, Rabii El Maani and Bouchaib Radi

the near-wall region. The transport equation for ν̃ is:

 ( ) !2 
∂ ∂ 1 ∂ ∂ ν̃ ∂ ν̃  −Yν +Sν̃
(ρν̃)+ (ρν̃ui ) = Gν +  (µ + ρν̃) + Cb2 ρ
∂t ∂xi σev ∂xi ∂xi ∂xi
(3)
Where Gv is the production of turbulent viscosity, Yv is the destruction of
turbulent viscosity that occurs in the near-wall region due to wall blocking
and viscous damping, σν̃ and Cb2 are constants, ν̃ is the molecular kinematic
viscosity and Sν̃ is a user-defined source term. The production term Gv is
modeled as:
Gν = Cb1 ρS̃ ν̃ (4)
 6
1/6
1+Cω ν̃
Where fω = g 6
g +Cω6
3
, g = r + Cω2 (r6 − r), and r ≡ S̃κ2 d2
3
The model constants have following values:
Cb1 = 0.1355, Cb2 = 0.622, σν̃ = 32 , Cv1 = 7.1,

Cω1 = 3.239, Cω2 = 0.3, Cω3 = 2.0, κ = 0.4187

3 Formulation of the probabilistic study

Although dispersion and uncertainty occur in all aspects of an analysis, deter-
ministic analyzes do not take it into account. Then, probabilistic studies are
essential analyzes so that aerodynamic manufacturers can answer various aero-
nautical problems. Probabilistic characterization quantifies the reliability or
quality of the product through statistical analysis. It combines deterministic
characterization with statistical analysis tools to address the effect of statistical
variability and uncertainty on product behavior.

3.1 Six Sigma analysis

Figure 1: The curve of the function of a Gaussian distribution

Probabilistic study of the aerodynamic around a 3D wing 53

The purpose of a Six Sigma analysis is to understand the impact of the un-
certainties associated with the input parameter of your design. A Six Sigma
analysis allows you to determine how the uncertain model inputs affect the
results of an analysis using statistical distribution functions (such as Gaus-
sian distribution, etc.) to describe uncertain output parameters. Six Sigma
Analysis allows you to determine whether your product satisfies Six Sigma
quality criteria. An output parameter that characterizes product performance
is typically used to determine whether a product’s performance is satisfactory.
The parameter must fall within the interval bounded by the lower specification
limit (LSL) and the upper specification limit (USL).
The Mean is a measure of average for a set of observations. The mean of a
set of n observations is defined as follows:

n
1X
µb = yi (5)
n i=1

The standard deviation of a set of n observations is defined as follows:

v
n
1
u
b 2
u X
σb = t (yi − µ) (6)
(n − 1) i=1

Sigma level is calculated as the inverse cumulative distribution function of

a standard Gaussian distribution at a given percentile. Sigma level is used
in conjunction with standard deviation to measure data dispersion from the
mean. For example, for a pair of quantile X and sigma level nα , it means that
the value of X is about nα standard deviations away from the sample mean.
The cumulative distribution function Fi associated with xi :

x − xi
P (X ≤ x) = Fi + (Fi+1 − Fi ) (7)
xi+1 − xi

4 Numerical simulation
4.1 Presentation of the validation
We compare the results of our numerical simulation precisely with experimental
results on a generic configuration such as the ONERA M6 wing which presents
a database for these tests documented by Volker Schmitt and François Charpin
in the AGARD report [13].
54 Soufiane Elouardi, Rabii El Maani and Bouchaib Radi

Figure 2: Geometric layout of the ONERA M6 wing and fluid mesh

For this, we carried out an aerodynamic analysis by ANSYS/Fluent soft-

ware with a Mach number of 0.8395, an angle of attack of 3.06 and a Reynolds
number until 11.72E6. We validate our aerodynamic simulation results found
by ANSYS/Fluent with the database for the ONERA M6 wing consisting on
the pressure cœfficient Cp , located in two different sections y/b = 0.2, 0.44
and the lift (Cl ) and drag (Cd ) cœfficients for different turbulence models.
A hybrid-unstructured mesh was generated with 375 263 cells and 102 432
volume nodes points for CFD computations.

Table 1: Comparison of Cl and Cd for different turbulence models

NASA CFD S-A k- Standard k- RNG k-ω Standard k-ω SST
Cl 0.1410 0.1310 0.1250 0.1299 0.1102 0.1288
Cd 0.0088 0.0097 0.0145 0.0103 0.0236 0.0096

Figure 3: Cp distribution for different turbulence models at y/b=0.2 & 0.44

Probabilistic study of the aerodynamic around a 3D wing 55

Figure 3 shows our numerical simulation results of the pressure cœfficient

using the five models of turbulence (S-A, k −  Standard, k −  RNG, k − ω
Standard, and k − ω SST) compared to the experimental results. According
to the table comparing the different turbulence models, we observe that the
SA model is the most compatible and closest for this configuration with the
results of the NASA-CFD.

4.2 Probabilistic results

This analysis was performed on the random value of the aerodynamic cœffi-
cients forces exerted along the surfaces of the upper and lower wing to quantify
input parameter values such as viscosity, Mach number and angle of attack.

Figure 4: Distribution function of the input parameters: Angle of attack;

Viscosity; Mach number

Table 2: Input parameters properties.

Lower Bound Upper Bound Distribution type Mean Standard Deviation
Viscosity 1,5129E-05 2,0658E-05 Normal 1,7894E-05 8,9470E-07
Mach Number 0,5072 0,6927 Normal 0,6000 0,0300
Angle of attack 0,0451 0,0616 Normal 0,0534 0,0026
56 Soufiane Elouardi, Rabii El Maani and Bouchaib Radi

The design of the input parameter experiments (Viscosity, Mach number,

and attack angle), as shown in Figure 4, is a technique used to determine
the location of the sampling points; it is part of the six sigma analyzes. We
represent in this plane of experience the distribution cumulative function and
the density of probability function. Table 2 presents the distribution properties
of each input parameter, the upper and lower limits, the mean, the standard
deviation and the type of distribution associated with the input parameter
used to generate the experimental design.

Figure 5: The cumulative distribution function and the Probability density

function of the drag and lift cœfficients.

Table 3: Output parameters properties.

Lower Bound Upper Bound Distribution type Mean Standard Deviation
Cd 0,0047 0,0137 Normal 0,0082 0,0011
Cl 0,0652 0,1497 Normal 0,1018 0,0114

After releasing the six-sigma analysis on the aerodynamic cœfficients based

on a sampling set up as a normal distribution of the uncertainties of the differ-
ent types of parameters on the aerodynamic wing ONERA M6, we obtained
distributions of the two aerodynamic cœfficients (Cl and Cd ), as illustrated in
Figures 5 with properties of each parameter presented in Table 3. We note
that the area included under the curve at the standard deviation value in the
two output parameters, we can say that 99.99 of individuals are found in a
range of plus or minus six standard deviation.

5 Conclusions
Six Sigma analysis is a probabilistic analysis considered as the main tool for
evaluating the performance of a process in order to detect statistical variability.
Probabilistic study of the aerodynamic around a 3D wing 57

One of the basic principles of Six Sigma is the reduction of uncertainty. Indeed,
a customer’s dissatisfaction always results from a gap between an expected
situation and a real situation. This difference is due in large part to the
variability of the processes (variability on materials, variability in geometry
and variability on the conditions in which the process evolves).
However, we have shown easily that, even though the process is perfectly
under control, it is not possible to detect small displacement of the process.
The smallest detectable shift is 1.5 Sigma. It is, therefore, an estimate of a
percentage of nonconformities in a given situation. It is, therefore, an estimate
of a percentage of non-compliant in a given situation. A single simulation
gives an indication of the operation of a design under ideal conditions, but
a six-sigma analysis investigates the performance range based on the design
and operating parameters. It allows designers and engineers to have a robust
design due to the uncertainty of model formulation and input parameters.

References
[1] A. Abbas, J. DE Vicente and E. Valero, Aerodynamic Technologies to
Improve Aircraft Performance, Aerosp. Sci. Technol., 28 (2013), no. 1,
100 - 132. https://doi.org/10.1016/j.ast.2012.10.008

[2] M. H. Dickinson, K. G. Gotz, Unsteady Aerodynamic Performance of

Model Wings at Low Reynolds Numbers, J. Exp. Biol., 174 (1993), no.
1, 45 - 64.

[3] A. El Hami, B. Radi, Fluid-Structure Interactions and Uncertainties: An-

sys and Fluent Tool, John Wiley and Sons, 2017.
https://doi.org/10.1002/9781119388937

[4] R. El Maani, B. Radi and A. El Hami, Validation Numérique d’un Modèle

Aérodynamique: Application à Une Aile D’avion, Incertitudes et Fiabilité
des Systèmes Multiphysiques, iSTE OpenScience, 17 (2017), no. 1.
https://doi.org/10.21494/iste.op.2017.0118

[5] R. El Maani, Étude Basée Sur l’optimisation Fiabiliste En Aérody-

namique, INSA de Rouen, 2016.

[6] B. J. Holmes, C. J. Obara and L. P. Yip, Natural Laminar Flow Experi-

ments on Modem Airplane Surfaces, NASA Tech. Pap., 1984, 2256.

[7] J. B. Ingram-Cotton, M. J. Hecht, R. J. Duphily, M. Zambrana, T. Hi-

ramoto, C. O’connor, Reliability Program Requirements for Space Sys-
tems, Aerosp. Rep. NO. TOR-2007 (8583)-6889, US Aerosp. Corp., 2007.
58 Soufiane Elouardi, Rabii El Maani and Bouchaib Radi

[8] R. H. Liebeck, Design of the Blended Wing Body Subsonic Transport, J.

Aircr., 41 (2004), no. 1, 10 - 25. https://doi.org/10.2514/1.9084

[9] R. M. Hicks, P. A. Henne, Wing Design by Numerical Optimization, Jour-

nal of Aircraft, 15 (1978), no. 7, 407 - 412.
https://doi.org/10.2514/3.58379

[10] C. H. U. Liu, E. S. Cursi, C. Zheng, E. I. D. Mohamed, E. L. Abdelkhalak,

Airfoil Shape Multi-Point Optimization Based on Sensitivity Analysis.

[11] C. L. pettit, Uncertainty Quantification in Aeroelasticity: Recent Results

and Research Challenges, J. Aircr., 41 (2004), no. 5, 1217 - 1229.
https://doi.org/10.2514/1.3961

[12] A. Rizzi, Modeling and Simulating Aircraft Stability and Control—the

SimSAC Project, Prog. Aerosp. Sci., 47 (2011), no. 8, 573 - 588.
https://doi.org/10.1016/j.paerosci.2011.08.004

[13] V. Schmitt, Pressure Distributions on the ONERA M6-Wing at Transonic

Mach Numbers, Experimental Data Base for Computer Program Assess-
ment, AGARD AR-138, 1979.

[14] J. Sobieszczanski-Sobieski, R. T. Haftka, Multidisciplinary Aerospace De-

sign Optimization: Survey of Recent Developments, Struct. Optim., 14
(1997), no. 1, 1 - 23. https://doi.org/10.1007/bf01197554

[15] P. Spalart, S. Allmaras, A One-Equation Turbulence Model for Aerody-

namic Flows, In 30th Aerospace Sciences Meeting and Exhibit, (1992),
439. https://doi.org/10.2514/6.1992-439

[16] H. Tijdeman, Investigations of the Transonic Flow around Oscillating Air-

foils, Nationaal Lucht-en Ruimtevaartlaboratorium, 1977.

[17] R. Vinuesa, S. M. Hosseini, A. Hanifi, D.S. Henningson, P. Schlat-

ter, Pressure-Gradient Turbulent Boundary Layers Developing Around
a Wing Section, Flow, Turbul. Combust, 99 (2017), no. 3-4, 613 - 641.
https://doi.org/10.1007/s10494-017-9840-z

[18] R. Von Kaenel, A. Rizzi, J. Oppelstrup, T. Goetzendorf-Grabowski, M.

Ghoreyshi, L. Cavagna, A. Berard, CEASIOM: Simulating Stability and
Control with CFD/CSM in Aircraft Conceptual Design, In 26th Inter-
national Congress of the Aeronautical Sciences, ICAS, Vol. 1, 2008, 14 -
19.
Probabilistic study of the aerodynamic around a 3D wing 59

[19] W. Yao, X. Chen, W. Luo, M. Tooren, J. Guo, Review of Uncertainty-

Based Multidisciplinary Design Optimization Methods for Aerospace Ve-
hicles, Prog. Aerosp. Sci., 47 (2011), no. 6, 450 - 479.
https://doi.org/10.1016/j.paerosci.2011.05.001

Received: November 12, 2018; Published: December 12, 2018