AIAA AVIATION Forum 10.2514/6.

2022-3869
June 27-July 1, 2022, Chicago, IL & Virtual
AIAA AVIATION 2022 Forum

Non-Intrusive Polynomial Chaos Approach

for Nonlinear Aeroelastic Uncertainty Quantification

Jeffrey P. Thomas∗ and Earl H. Dowell†

Duke University, Durham, NC 27708-0300

Presented is a point collocation non-intrusive polynomial chaos approach for nonlinear

aeroelastic uncertainty quantification analysis. A unique aspect of our approach is that we
are using a nonlinear frequency-domain harmonic balance aeroelastic solution methodology,
which allows us to solve directly for a limit cycle oscillation response amplitude versus reduced
velocity response curve. Other researchers have typically framed their aeroelastic uncertainty
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

quantification analysis methods around time-domain aeroelastic solvers, where an indirect

measure of aeroelastic stability, such as aeroelastic damping, is the dependent variable that
one performs an uncertainty quantification analysis for. Our method allows one to determine
directly, uncertainty quantification for an aeroelastic limit cycle oscillation response curve, as
well as unstable branches of an aeroelastic limit cycle oscillation response curve. We demonstrate
the technique for uncertainty in freestream Mach number and airfoil static unbalance for a
two-dimensional model of the NASA Benchmark Models Program B0012 NACA 0012 airfoil
section aeroelastic wing configuration.

Nomenclature

𝑏 = airfoil semi-chord length, 𝑐/2

𝑐 = airfoil chord length
𝑐𝑙 = airfoil sectional coefficient of lift
𝑐𝑚 = airfoil sectional moment coefficient about elastic axis
𝑐¯𝑚0 = zeroth harmonic or mean moment coefficient about elastic axis
𝑐¯𝑙1 = first harmonic unsteady coefficient of lift
𝑐¯𝑚1 = first harmonic unsteady moment coefficient about elastic axis
𝐷ℎ = airfoil structural plunge coordinate damping
𝐷𝛼 = airfoil structural torsional damping
𝑒 = nondimensional location of airfoil elastic axis, measured positive aft of airfoil mid-chord, (see Fig. 1)
ℎ = airfoil plunge coordinate degree of freedom
𝛼 = airfoil pitch coordinate degree of freedom
𝛼𝑒0 = angle-of-attack for which there is zero moment about the elastic axis of the airfoil.
𝛼¯ 0 = zeroth harmonic or mean airfoil pitch amplitude
ℎ¯ 1 = first harmonic unsteady airfoil plunge amplitude
𝛼¯ 1 = first harmonic unsteady airfoil pitch amplitude
𝐼𝛼 = airfoil second moment of inertia about elastic axis
𝐾ℎ = airfoil plunge stiffness
𝐾𝛼 = torsional stiffness about elastic axis
L = vector of aeroelastic solution variables
𝑀∞ = freestream Mach number
𝑚 = structural mass
𝜇 = mass ratio, 𝜇 = 𝑚/𝜋𝜌∞ 𝑏 2 𝑠
𝜇 𝑀∞ = mean freestream Mach number
𝜇 𝑥𝛼 = mean airfoil static unbalance
𝑁𝑆 = number of samples

∗
Research Assistant Professor, Department of Mechanical Engineering and Materials Science, Senior Member, AIAA.
†
William Holland Hall Professor, Department of Mechanical Engineering and Materials Science, Honorary Fellow, AIAA.

Copyright © 2022 by Jeffrey P. Thomas and Earl H. Dowell. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.
 𝑝 = order of polynomial chaos expansion
Q = vector representing discrete harmonic balance flow solution
Re = real or in phase part of complex valued quantity
Im = imaginary or out of phase part of complex valued quantity
R = vector representing residual of aeroelastic governing equations
𝑅𝑒 ∞ = freestream Reynolds number
𝑟𝛼 = radius of gyration of airfoil about elastic axis, 𝑟 2𝛼 = 𝐼 𝛼 /𝑚𝑏 2
𝜌∞ = freestream density, respectively
𝑆𝛼 = first moment of inertia of airfoil about elastic axis
𝑠 = airfoil span
𝜎𝑀∞ = freestream Mach number standard deviation
𝜎𝑥 𝛼 = static unbalance standard deviation
𝜃 = random variable
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

𝑉 = reduced velocity, 𝑉 = 𝑈∞ /𝜔 𝛼 𝑏
𝜔 = dimensional frequency (radians/second)
𝜔¯ = reduced frequency based on airfoil chord 𝜔¯ = 𝜔𝑐/𝑈∞ , respectively √︁
𝜔ℎ = plunge coordinate natural frequency based√︁on wing mass, 𝜔 ℎ = 𝐾 ℎ /𝑚
𝜔𝛼 = pitch coordinate natural frequency, 𝜔 𝛼 = 𝐾 𝛼 /𝐼 𝛼
𝑥𝛼 = airfoil static unbalance, 𝑥 𝛼 = 𝑆 𝛼 /𝑚𝑏
𝜁ℎ = plunge coordinate damping coefficient, 𝜁 ℎ = 𝐷 ℎ /2𝑚𝜔 ℎ
𝜁𝛼 = pitch coordinate damping coefficient, 𝜁 𝛼 = 𝐷 𝛼 /2𝐼 𝛼 𝜔 𝛼
𝑈∞ = freestream velocity

I. Introduction
Aeroelastic Uncertainty Quantification (UQ) is currently an active research area. A recent survey paper by
Chassaing et al. [1] provides an excellent overview of the many research efforts currently underway. In the following,
we present a Non-Intrusive Polynomial Chaos (NIPC) approach for nonlinear aeroelastic UQ analysis. A unique aspect
of our UQ approach is that it is formulated around a nonlinear frequency-domain Harmonic Balance (HB) aeroelastic
solution technique.
To demonstrate our aeroelastic UQ approach, we consider a two-dimensional two degree-of-freedom (plunge and
pitch motions) model of the NASA Benchmark Models Program B0012 configuration, which consists of the NACA 0012
airfoil section (See e.g. Refs. [2–6]). The governing frequency-domain aeroelastic equations for the two-dimensional
NASA B0012 model can be written as (See Fig. 1 for geometric details)

R(Q, L) = 0 (1)

where


 𝛼¯ 0  

 𝐿     𝑅1 


 1 




 𝑉










𝐿   
  
 
𝑅  

 2 
 𝜔¯ 
 
 2
 
  
  
  
 

L = 𝐿3 = ¯ℎ1
 , R = 𝑅3 ,

 
  Re
 
 
 

 𝐿4   𝛼¯¯1 𝑏   𝑅4 

     
  
  


 


 𝐿5  
 ℎ 1

 
 𝑅5 

    Im 𝛼¯ 𝑏 
   

 1 

2
 2𝑉 2
𝑅1 = 𝛼¯ 0 − 𝛼𝑒0 − 𝑐¯𝑚0
𝜋𝑟 2𝛼
 2 
ℎ¯ 1 ℎ¯ 1 ℎ¯ 1
         
Re( 𝑐¯𝑙1 ) 1 2 2 √ 𝜔ℎ 𝜔ℎ
𝑅2 = − 𝜇𝜔¯ Re + 𝑥 𝛼 𝑉 − 𝜇𝜔𝜁 ¯ ℎ Im 𝑉+ Re
𝜋 𝛼¯ 1 4 𝛼¯ 1 𝑏 𝜔𝛼 𝛼¯ 1 𝑏 𝜔𝛼 𝛼¯ 1 𝑏
ℎ¯ 1
    
Re( 𝑐¯𝑚1 ) 1 2
𝑅3 =− 2 + 𝜇𝜔¯ 𝑥 𝛼 Re + 𝑟 2𝛼 𝑉 2 + 𝑟 2𝛼
𝜋 𝛼¯ 1 4 𝛼¯ 1 𝑏
 2 
ℎ¯ 1 ℎ¯ 1 ℎ¯ 1
       
Im( 𝑐¯𝑙1 ) 1 2 2 √ 𝜔ℎ 𝜔ℎ
𝑅4 = − 𝜇𝜔¯ Im 𝑉 − 𝜇𝜔𝜁 ¯ ℎ Re 𝑉+ Im
𝜋 𝛼¯ 1 4 𝛼¯ 1 𝑏 𝜔𝛼 𝛼¯ 1 𝑏 𝜔𝛼 𝛼¯ 1 𝑏
ℎ¯ 1
  
Im( 𝑐¯𝑚1 ) 1 2 √
𝑅5 =− 2 + 𝜇𝜔¯ 𝑥 𝛼 Im 𝑉 2 + 𝜇𝜔𝑉 ¯ 𝜁 𝛼 𝑟 2𝛼 .
𝜋 𝛼¯ 1 4 𝛼¯ 1 𝑏

For the two-dimensional two degree-of-freedom model of the NASA Benchmark Models Program B0012 NACA 0012
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

configuration, the structural parameters are:

𝜇 = 1000, 𝑥 𝛼 = 0.0, 𝑟 2𝛼 = 1.024, (𝜔 ℎ /𝜔 𝛼 ) = 0.646, 𝜁 ℎ = 0.0, 𝜁 𝛼 = 0.0,

𝑒 = 0.0, and 𝛼𝑒0 = 0.0◦ .

c
b b
eb xαb

Mean PositionMid Chord Mass Center

+h Elastic Axis

+αe Kα

Kh

Fig. 1 Geometry for "Typical" (Plunge/Pitch) Two Degree-of-Freedom Airfoil Section Aeroelastic Model.

We are using the nonlinear frequency-domain HB Computational Fluid Dynamic (CFD) solution method of
Hall et al. [7], as well as the nonlinear frequency-domain HB aeroelastic solution method of Thomas et al. [8, 9]. We
consider a nominal case freestream Mach number of 𝑀∞ = 0.75, freestream Reynolds number of 𝑅𝑒 ∞ = 1, 000, 000,
and a mean angle-of-attack of 𝛼¯ 0 = 0.0◦ . Figure 2 shows the 193x49 mesh points o-grid computational grid used for
the analysis along with computed Reynolds Averaged Navier-Stokes (RANS) CFD model steady flow Mach number
contours for the nominal 𝑀∞ = 0.75 flow condition. For the results presented here, the CFD method is a variant of
standard Lax–Wendroff scheme (See Refs.[10] and [11]) in conjunction with the one-equation turbulence model of
Spalart and Allmaras [12].
Next, Fig. 3 shows computed steady flow surface pressure coefficient distributions for freestream Mach numbers
ranging from 𝑀∞ = 0.73 to 𝑀∞ = 0.77, along with the computed transonic Mach number flutter onset reduced velocity

3
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

(a) Computational Grid. (b) Steady Flow Mach Number Contours.

Fig. 2 NASA Benchmark Models Program B0012 NACA 0012 Configuration Computational Grid and Computed
Mach Number Contours for 𝑴∞ = 0.75, 𝑹𝒆 ∞ = 1, 000, 000, and 𝜶¯ 0 = 0.0◦ .

trend for freestream Mach numbers ranging from 𝑀∞ = 0.70 to 𝑀∞ = 0.83. Shock formation near the quarter chord
of the airfoil section can be seen to be starting at roughly a Mach number of 𝑀∞ = 0.75 from the surface pressure
coefficient plot, and the well known transonic Mach number "flutter speed dip trend" (Dowell [13]) can be seen from the
flutter onset reduced velocity versus Mach number plot.
For the nonlinear aeroelastic Limit Cycle Oscillation (LCO) analysis, we are modeling the LCO response amplitude
of the first harmonic unsteady pitch amplitude (𝛼¯ 1 ) as a function of the reduced velocity (𝑉). A unique aspect of
the nonlinear frequency-domain HB aeroelastic solution approach of Thomas et al. [8, 9] is that the unsteady pitch
amplitude is the independent variable in the analysis, and the reduced velocity is the dependent variable. With
time-domain aeroelastic solution approaches, the situation is the opposite, i.e., reduced velocity is the independent
variable, and unsteady pitch amplitude is the dependent variable. One benefit of having the unsteady pitch amplitude as
the independent variable is that one can compute unstable branches of an LCO response curve.
Figure. 4 shows the computed deterministic LCO unsteady pitch versus reduced velocity response curve for the
two-dimensional two degree-of-freedom NASA Benchmark Models Program B0012 NACA 0012 configuration for the
𝑀∞ = 0.75, 𝑅𝑒 ∞ = 1, 000, 000, and 𝛼¯ 0 = 0.0◦ flow condition. As can be seen, there is a slight region of unstable
aeroelastic LCO response from zero degrees of unsteady pitch amplitude, up to roughly 1.0 degree of unsteady pitch
amplitude.

II. Uncertainty Quantification Analysis

As a demonstration of our nonlinear frequency-domain aeroelastic solution methodology UQ analysis approach, we
consider there to be uncertainty in the freestream Mach number. Specifically,

𝑀∞ = 𝜇 𝑀∞ + 𝜃𝜎𝑀∞ (2)

where the mean freestream Mach number is 𝜇 𝑀∞ = 0.75, the standard deviation of the freestream Mach number is
𝜎𝑀∞ = 0.005, and 𝜃 is a random variable with zero mean and unit variance, i.e., 𝜃 ∼ N (0, 1). Next, following the NIPC
UQ approaches presented in Hosder et al. [14] and Stanford and Massey [15], and Stanford et al. [16], a fourth-order
chaos expansion using Hermite polynomials is chosen for the point-collocation NIPC uncertainty propagation for the

4
 26
0.9
Airfoil Surface Pressure Coefficient, -cp=-(p-p∞)/q∞

0.8 25

Flutter Onset Reduced Velocity, V=U∞/ωαb

M∞=0.73
0.7 24
M∞=0.74
0.6 M∞=0.75
23
M∞=0.76
0.5
M∞=0.77 22
0.4

0.3 21

0.2 20
0.1
19
0.0
18
-0.1
17
-0.2

-0.3 16
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.70 0.71 0.72 0.73 0.74 0.75 0.76 0.77 0.78 0.79 0.80 0.81 0.82 0.83
Airfoil Surface Location, x/c Freestream Mach Number, M∞

(a) Surface Pressure Coefficient. (b) Transonic Flutter Onset Reduced Velocity Trend.

Fig. 3 NASA Benchmark Models Program B0012 NACA 0012 Configuration Surface Pressure Coefficient
Distributions and Transonic Mach Number Flutter Onset Reduced Velocity Trend, 𝑹𝒆 ∞ = 1, 000, 000 and
𝜶¯ 0 = 0.0◦ .

5.5

5.0

4.5
LCO Pitch Amplitude, α1 (deg)

4.0

3.5

3.0

2.5

2.0
Deterministic Harmonic Balance CFD Aeroelastic Solution
1.5

1.0
Hopf Bifurcation Linear Flutter Onset Reduced Velocity, V=17.982
0.5

0.0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb
"Hard" / "Effective" Flutter Onset Reduced Velocity, V=17.912

Fig. 4 Computed Unsteady Pitch Limit Cycle Oscillation Response Amplitude Versus Reduced Velocity Trend,
𝑴∞ = 0.75, 𝑹𝒆 ∞ = 1, 000, 000, and 𝜶¯ 0 = 0.0◦ .

5
 aeroelastic reduced velocity dependent variable, i.e.,
𝑝=4
∑︁
𝑉 ( 𝛼¯ 1 , 𝜇 𝑀∞ , 𝜎𝑀∞ , 𝜃) ≈ 𝑉˜ ( 𝛼¯ 1 , 𝜇 𝑀∞ , 𝜎𝑀∞ ) 𝑗 Ψ 𝑗 (𝜃) (3)
𝑗=0

where 𝑝 = 4 for our demonstration test case. Five deterministic solutions for 𝜃 0 = 0.0, 𝜃 1 = −2.0, 𝜃 2 = 2.0, 𝜃 3 = −4.0,
and 𝜃 4 = 4.0, (i.e., 𝑀∞0 = 0.75, 𝑀∞1 = 0.74, 𝑀∞2 = 0.76, 𝑀∞3 = 0.73 and 𝑀∞4 = 0.77) are computed for each of the
12 unsteady pitch amplitudes shown in Fig. 4 (i.e., 𝛼¯ 1 = 0.001◦ , 0.2◦ , 0.5◦ , 1.0◦ , 1.5◦ , 2.0◦ , 2.5◦ , 3.0◦ , 3.5◦ , 4.0◦ , 4.5◦ ,
and 5.0◦ ), which allows us to then compute the five NIPC expansion coefficients 𝑉˜ ( 𝛼¯ 1 , 𝜇 𝑀∞ , 𝜎𝑀∞ ) 𝑗 , 𝑗 = 0, 1, ..., 𝑝 = 4
in Eq. 3, for each of the 12 unsteady pitch amplitudes.
To next analyze the statistics of the uncertainty in the freestream Mach number, for each unsteady pitch amplitude,
we consider 𝑁 𝑆 = 100, 000 samples of the normally distributed random variable 𝜃 (i.e., 𝜃 𝑖 , 𝑖 = 0, 1, ..., 𝑁 𝑆 = 100, 000),
and we compute 𝑁 𝑆 = 100, 000 reduced velocities (𝑉 ( 𝛼¯ 1 , 𝜇 𝑀∞ , 𝜎𝑀∞ , 𝜃 𝑖 ), 𝑖 = 0, 1, ..., 𝑁 𝑆 = 100, 000) corresponding to
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

each 𝜃 𝑖 using Eq. 3. We can then construct probability density distribution histograms for each of the 12 unsteady pitch
amplitudes as shown in Fig. 5 and Fig. 6. For the histograms, 100 equally spaced bins between reduced velocities of
𝑉 = 17 and 𝑉 = 21 are established, so that the bin width is Δ𝑉 = 0.04, and 𝑛 𝜃𝑖bin is the number of 𝑉 ( 𝛼¯ 1 , 𝜇 𝑀∞ , 𝜎𝑀∞ , 𝜃 𝑖 )
within each bin. Additionally, we can plot the cumulative probability distributions as shown in Fig. 7. The red shaded
regions in Fig. 7 indicate the 95% probability range for reduced velocity for each unsteady pitch amplitude.

17
16 o
α1 = 0.001
Probability Density Distribution, (nθ /Ns)/∆Vbin

15 o
α1 = 0.2
14 o
α1 = 0.5
13 o
α1 = 1.0
ibin

12 o
α1 = 1.5
11 α1 = 2.0
o

10 α1 = 2.5
o

9 o
α1 = 3.0
o
8 α1 = 3.5
o
7 α1 = 4.0
o
6 α1 = 4.5
o
5 α1 = 5.0
4
3
2
1
0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb

Fig. 5 Probability Density Distributions for Each Unsteady Pitch Limit Cycle Oscillation Response Amplitude,
𝑴∞ = 𝝁 𝑴∞ + 𝜽𝝈𝑴∞ , 𝝁 𝑴∞ = 0.75, 𝝈𝑴∞ = 0.005, 𝜽 ∼ N (0, 1), 𝑹𝒆 ∞ = 1, 000, 000, and 𝜶¯ 0 = 0.0◦ .

From the 95% probability range of reduced velocity for each unsteady pitch amplitude shown in Fig. 7, one can
similarly add the 95% probability range for reduced velocity as error bars to each unsteady pitch amplitude of the
aeroelastic LCO response curve as shown in Fig. 8, We believe the "narrowing" of the 95% probability range for reduced
velocity around an LCO pitch amplitude of 𝛼¯ 1 = 2.0◦ is due to the fact that there is a "crossing" of the deterministic

6
 17 17 17
16 16 16
Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

15 15 15
14 14 14
13 13 13
ibin

ibin

ibin
12 12 12
11 11 11
10 10 10
9 9 9
8 8 8
7 7 7
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(a) 𝜶¯ 1 = 0.001◦ . (b) 𝜶¯ 1 = 0.2◦ . (c) 𝜶¯ 1 = 0.5◦ .

17 17 17
16 16 16
Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

15 15 15
14 14 14
13 13 13
ibin

ibin

ibin
12 12 12
11 11 11
10 10 10
9 9 9
8 8 8
7 7 7
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(d) 𝜶¯ 1 = 1.0◦ . (e) 𝜶¯ 1 = 1.5◦ . (f) 𝜶¯ 1 = 2.0◦ .

17 17 17
16 16 16
Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

15 15 15
14 14 14
13 13 13
ibin

ibin

ibin

12 12 12
11 11 11
10 10 10
9 9 9
8 8 8
7 7 7
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(g) 𝜶¯ 1 = 2.5◦ . (h) 𝜶¯ 1 = 3.0◦ . (i) 𝜶¯ 1 = 3.5◦ .

17 17 17
16 16 16
Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

15 15 15
14 14 14
13 13 13
ibin

ibin

ibin

12 12 12
11 11 11
10 10 10
9 9 9
8 8 8
7 7 7
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(j) 𝜶¯ 1 = 4.0◦ . (k) 𝜶¯ 1 = 4.5◦ . (l) 𝜶¯ 1 = 5.0◦ .

Fig. 6 Probability Density Distributions for Each Unsteady Pitch Limit Cycle Oscillation Response Amplitude,
𝑴∞ = 𝝁 𝑴∞ + 𝜽𝝈𝑴∞ , 𝝁 𝑴∞ = 0.75, 𝝈𝑴∞ = 0.005, 𝜽 ∼ N (0, 1), 𝑹𝒆 ∞ = 1, 000, 000, and 𝜶¯ 0 = 0.0◦ .

7
 1.0 1.0 1.0

0.9 0.9 0.9

0.8 0.8 0.8

Cumulative Probability, F(V)

Cumulative Probability, F(V)

Cumulative Probability, F(V)

0.7 0.7 0.7

0.6 0.6 0.6

0.5 0.5 0.5

0.4 0.4 0.4

0.3 0.3 0.3

0.2 0.2 0.2

0.1 0.1 0.1

0.0 0.0 0.0

17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(a) 𝜶¯ 1 = 0.001◦ . (b) 𝜶¯ 1 = 0.2◦ . (c) 𝜶¯ 1 = 0.5◦ .

1.0 1.0 1.0

0.9 0.9 0.9

0.8 0.8 0.8

Cumulative Probability, F(V)

Cumulative Probability, F(V)

Cumulative Probability, F(V)

0.7 0.7 0.7

0.6 0.6 0.6

0.5 0.5 0.5

0.4 0.4 0.4

0.3 0.3 0.3

0.2 0.2 0.2

0.1 0.1 0.1

0.0 0.0 0.0

17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(d) 𝜶¯ 1 = 1.0◦ . (e) 𝜶¯ 1 = 1.5◦ . (f) 𝜶¯ 1 = 2.0◦ .

1.0 1.0 1.0

0.9 0.9 0.9

0.8 0.8 0.8

Cumulative Probability, F(V)

Cumulative Probability, F(V)

Cumulative Probability, F(V)

0.7 0.7 0.7

0.6 0.6 0.6

0.5 0.5 0.5

0.4 0.4 0.4

0.3 0.3 0.3

0.2 0.2 0.2

0.1 0.1 0.1

0.0 0.0 0.0

17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(g) 𝜶¯ 1 = 2.5◦ . (h) 𝜶¯ 1 = 3.0◦ . (i) 𝜶¯ 1 = 3.5◦ .

1.0 1.0 1.0

0.9 0.9 0.9

0.8 0.8 0.8

Cumulative Probability, F(V)

Cumulative Probability, F(V)

Cumulative Probability, F(V)

0.7 0.7 0.7

0.6 0.6 0.6

0.5 0.5 0.5

0.4 0.4 0.4

0.3 0.3 0.3

0.2 0.2 0.2

0.1 0.1 0.1

0.0 0.0 0.0

17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(j) 𝜶¯ 1 = 4.0◦ . (k) 𝜶¯ 1 = 4.5◦ . (l) 𝜶¯ 1 = 5.0◦ .

Fig. 7 Cumulative Probability Distributions for Each Unsteady Pitch Limit Cycle Oscillation Response
Amplitude, 𝑴∞ = 𝝁 𝑴∞ + 𝜽𝝈𝑴∞ , 𝝁 𝑴∞ = 0.75, 𝝈𝑴∞ = 0.005, 𝜽 ∼ N (0, 1), 𝑹𝒆 ∞ = 1, 000, 000, and 𝜶¯ 0 = 0.0◦ .

8
 5.5

5.0

4.5
LCO Pitch Amplitude, α1 (deg)
4.0

3.5

3.0

2.5

2.0
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

Deterministic Harmonic Balance CFD Aeroelastic Solution

1.5 Mean and 95% Probability Range for Reduced Velocity

1.0

0.5

0.0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb

Fig. 8 Unsteady Pitch Limit Cycle Oscillation Response Amplitude as a Function of Reduced Velocity, Including
95% Probability Range for Reduced Velocity for Each Unsteady Pitch Amplitude, for Uncertainty in Freestream
Mach Number, 𝑴∞ = 𝝁 𝑴∞ + 𝜽𝝈𝑴∞ , 𝝁 𝑴∞ = 0.75, 𝝈𝑴∞ = 0.005, 𝜽 ∼ N (0, 1), 𝑹𝒆 ∞ = 1, 000, 000, and 𝜶¯ 0 = 0.0◦ .

𝑀∞ = 0.74 and 𝑀∞ = 0.76 LCO response curves also around an LCO pitch amplitude of 𝛼¯ 1 = 2.0◦ as can be seen in
Fig. 9.

III. Randomness in Two Input Variables

We next consider the case where there is randomness in two input variables. In the case, freestream Mach number
and static unbalance. Specifically,

𝑀∞ = 𝜇 𝑀∞ + 𝜃 𝑀∞ 𝜎𝑀∞ 𝑥 𝛼 = 𝜇 𝑥 𝛼 + 𝜃 𝑥 𝛼 𝜎𝑥 𝛼 (4)

where the mean freestream Mach number is again 𝜇 𝑀∞ = 0.75, the standard deviation of the freestream Mach number
is again 𝜎𝑀∞ = 0.005, 𝜃 𝑀∞ is a random variable with zero mean and unit variance, i.e., 𝜃 𝑀∞ ∼ N (0, 1), the mean
static unbalance is 𝜇 𝑥 𝛼 = 0., the standard deviation of the static unbalance is 𝜎𝑥 𝛼 = 0.0025, and 𝜃 𝑥 𝛼 is a random
variable with zero mean and unit variance, i.e., 𝜃 𝑥 𝛼 ∼ N (0, 1). For each of the the 12 LCO response amplitudes,
we compute 25 solution samples, corresponding to random variable values of 𝜃 𝑀∞ = −4.0, −2.0, 0.0, 2.0, 4.0 and
𝜃 𝑥 𝛼 = −4.0, −2.0, 0.0, 2.0, 4.0. A two-random-variable fourth-order "total order" polynomial chaos expansion consisting
of 15 terms
𝑃=14
∑︁
𝑉 ( 𝛼¯ 1 , 𝜃 𝑀∞ , 𝜃 𝑥 𝛼 ) ≈ 𝑉˜ ( 𝛼¯ 1 ) 𝑗 Ψ 𝑗 (𝜃 𝑀∞ , 𝜃 𝑥 𝛼 ) (5)
𝑗=0

is used for the NIPC expansion (See e.g. Eldred [17]). The least squares method is then used to compute the 15 NIPC
expansion coefficients 𝑉˜ ( 𝛼¯ 1 ) 𝑗 based on the 25 solution samples, for each unsteady pitch amplitude.
Similar to Figs. 5 and 6, Figs. 10 and 11 show the computed probability density distribution histograms for each
of the 12 unsteady pitch amplitudes when using 100,000 normally distributed samples for both 𝜃 𝑀∞ and 𝜃 𝑥 𝛼 for the
fourth-order "total-order" expansion (Eq. 5). Figure 12 shows the 95% probability range for the limit cycle oscillation
response curve when considering uncertainty in freestream Mach Number and static unbalance. As can seen, when
compared to Fig. 8, the predicted uncertainty in reduced velocity is more uniform over the entire range of LCO
amplitudes.

9
 5.5

5.0

4.5
LCO Pitch Amplitude, α1 (deg)
4.0

3.5

3.0

2.5

2.0 M∞=0.74 Deterministic Harmonic Balance CFD Aeroelastic Solution

Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

M∞=0.75 Deterministic Harmonic Balance CFD Aeroelastic Solution

1.5 M∞=0.76 Deterministic Harmonic Balance CFD Aeroelastic Solution

M∞=0.75 Stocastic Case Mean and 95% Probability Range for Reduced Velocity

1.0

0.5

0.0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb

Fig. 9 Unsteady Pitch Limit Cycle Oscillation Response Amplitude as a Function of Reduced Velocity, Including
95% Probability Range for Reduced Velocity for Each Unsteady Pitch Amplitude, for Uncertainty in Freestream
Mach Number, 𝑴∞ = 𝝁 𝑴∞ + 𝜽𝝈𝑴∞ , 𝝁 𝑴∞ = 0.75, 𝝈𝑴∞ = 0.005, 𝜽 ∼ N (0, 1), 𝑹𝒆 ∞ = 1, 000, 000, and 𝜶¯ 0 = 0.0◦ .

IV. Conclusions
As an advantage to time-domain methods, the nonlinear frequency-domain harmonic balance aeroelastic solution
approach allows one to directly compute aeroelastic limit cycle oscillation response trends. The approach also enables
one to compute sub-critical branches of limit cycle oscillation response curves, for configurations that exhibit sub-critical
aeroelastic limit cycle oscillation response. The non-intrusive polynomial chaos technique for uncertainty quantification
is demonstrated for predicting uncertainty quantification for the aeroelastic limit cycle oscillation response trend for a
transonic aeroelastic airfoil configuration. We demonstrate using the method when there is uncertainty in Mach number,
as well as Mach number and static unbalance.
For future studies, we intend to perform Monte Carlo simulations with the order of 1,000 or 10,000 simulations
using the full computational model in order to have a baseline of uncertainty quantification statistics, to which we can
then compare the non-intrusive polynomial chaos uncertainty quantification statistics to. Again, when compared to time-
domain approaches, the nonlinear frequency-domain harmonic balance aeroelastic solution approach is advantageous in
that it allows one to directly compute aeroelastic limit cycle oscillation response trends. We will consider uncertainty in
additional variables, e.g., flow parameters such as mean angle-of-attack and Reynolds number, as well as structural
properties such as mass ratio and structural mass, stiffness, and damping. We will also consider different ordered
non-intrusive polynomial chaos expansions, as well as using gradient data computed from discrete adjoint sensitivity
analysis to help reduce the number of samples required to formulate the non-intrusive polynomial chaos expansions.

V. Acknowledgments
The authors would like to thank Jared Cooper and Michael Devore of Barron Associates, Inc. for bringing to our
attention the various methodologies that are suitable for aeroelastic uncertainty quantification.

References
[1] Chassaing, J., Nitschke, C., Vincenti, A., Cinnella, P., and Lucor, D., “Advances in Parametric and Model-Form Uncertainty
Quantification in Canonical Aeroelastic Systems,” Aerospace Lab, , No. 14, 2018, pp. 1–19. https://doi.org/10.12762/2018.AL14-

10
 17
16
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

o
α1 = 0.001
Probability Density Distribution, (nθ /Ns)/∆Vbin

15 o
α1 = 0.2
14 o
α1 = 0.5
13 o
α1 = 1.0
ibin

12 o
α1 = 1.5
11 α1 = 2.0
o

10 α1 = 2.5
o

9 o
α1 = 3.0
o
8 α1 = 3.5
o
7 α1 = 4.0
o
6 α1 = 4.5
o
5 α1 = 5.0
4
3
2
1
0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb

Fig. 10 Probability Density Distributions for Each Unsteady Pitch Limit Cycle Oscillation Response Amplitude
for Uncertainty in Freestream Mach Number and Static unbalance, 𝑴∞ = 𝝁 𝑴∞ + 𝜽 𝑴∞ 𝝈𝑴∞ , 𝝁 𝑴∞ = 0.75,
𝝈𝑴∞ = 0.005, 𝜽 ∼ N (0, 1), 𝒙𝜶 = 𝝁 𝒙𝜶 + 𝜽 𝒙𝜶 𝝈𝒙𝜶 , 𝝁 𝒙𝜶 = 0.0, 𝝈𝒙𝜶 = 0.0025, 𝑹𝒆 ∞ = 1, 000, 000, and 𝜶¯ 0 = 0.0◦ .

11
 17 17 17
16 16 16
Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

15 15 15
14 14 14
13 13 13
ibin

ibin

ibin
12 12 12
11 11 11
10 10 10
9 9 9
8 8 8
7 7 7
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

(a) 𝜶¯ 1 = 0.001◦ . (b) 𝜶¯ 1 = 0.2◦ . (c) 𝜶¯ 1 = 0.5◦ .

17 17 17
16 16 16
Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

15 15 15
14 14 14
13 13 13
ibin

ibin

ibin
12 12 12
11 11 11
10 10 10
9 9 9
8 8 8
7 7 7
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(d) 𝜶¯ 1 = 1.0◦ . (e) 𝜶¯ 1 = 1.5◦ . (f) 𝜶¯ 1 = 2.0◦ .

17 17 17
16 16 16
Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

15 15 Probability Density Distribution, (nθ /Ns)/∆Vbin 15

14 14 14
13 13 13
ibin

ibin

ibin

12 12 12
11 11 11
10 10 10
9 9 9
8 8 8
7 7 7
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(g) 𝜶¯ 1 = 2.5◦ . (h) 𝜶¯ 1 = 3.0◦ . (i) 𝜶¯ 1 = 3.5◦ .

17 17 17
16 16 16
Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

Probability Density Distribution, (nθ /Ns)/∆Vbin

15 15 15
14 14 14
13 13 13
ibin

ibin

ibin

12 12 12
11 11 11
10 10 10
9 9 9
8 8 8
7 7 7
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb Reduced Velocity, V=U∞/ωαb

(j) 𝜶¯ 1 = 4.0◦ . (k) 𝜶¯ 1 = 4.5◦ . s (l) 𝜶¯ 1 = 5.0◦ .

Fig. 11 Probability Density Distributions for Each Unsteady Pitch Limit Cycle Oscillation Response Amplitude
for Uncertainty in Freestream Mach Number and Static unbalance, 𝑴∞ = 𝝁 𝑴∞ + 𝜽 𝑴∞ 𝝈𝑴∞ , 𝝁 𝑴∞ = 0.75,
𝝈𝑴∞ = 0.005, 𝜽 ∼ N (0, 1), 𝒙𝜶 = 𝝁 𝒙𝜶 + 𝜽 𝒙𝜶 𝝈𝒙𝜶 , 𝝁 𝒙𝜶 = 0.0, 𝝈𝒙𝜶 = 0.0025, 𝑹𝒆 ∞ = 1, 000, 000, and 𝜶¯ 0 = 0.0◦ .

12
 5.5

5.0

4.5
LCO Pitch Amplitude, α1 (deg)
4.0

3.5

3.0

2.5

2.0
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

Deterministic Harmonic Balance CFD Aeroelastic Solution

1.5 Mean and 95% Probability Range for Reduced Velocity

1.0

0.5

0.0
17.4 17.6 17.8 18.0 18.2 18.4 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2
Reduced Velocity, V=U∞/ωαb

Fig. 12 Unsteady Pitch Limit Cycle Oscillation Response Amplitude as a Function of Reduced Velocity,
Including 95% Probability Range for Reduced Velocity for Each Unsteady Pitch Amplitude, for Uncertainty in
Freestream Mach Number and Static unbalance, 𝑴∞ = 𝝁 𝑴∞ + 𝜽𝝈𝑴∞ , 𝝁 𝑴∞ = 0.75, 𝝈𝑴∞ = 0.005, 𝜽 ∼ N (0, 1),
𝒙𝜶 = 𝝁 𝒙𝜶 + 𝜽 𝒙𝜶 𝝈𝒙𝜶 , 𝝁 𝒙𝜶 = 0.0, 𝝈𝒙𝜶 = 0.0025, 𝑹𝒆 ∞ = 1, 000, 000, and 𝜶¯ 0 = 0.0◦ .

07, URL https://hal.archives-ouvertes.fr/hal-01935239.

[2] Bennett, R. M., “Test Cases for Flutter of the Benchmark Models Rectangular Wings on the Pitch and Plunge Apparatus,”
Verification and Validation Data for Computational Unsteady Aerodynamics, edited by L. P. Ruiz-Calavera, RTO-TR-26
AC/323(AVT)TP/19, North Atlantic Treaty Organization Research and Technology Organization, BP 25, 7 rue Ancelle, F-92201
Neuilly-sur-Seine Cedex, France, 2000, pp. 173–200. https://doi.org/10.14339/RTO-TR-026.

[3] Rivera, J. A., Dansberry, B. E., Bennett, R. M., Durham, M. H., and Silva, W. A., “NACA 0012 Benchmark Model Experimental
Flutter Results with Unsteady Pressure Distributions,” AIAA Paper 1992-2396, 1992. https://doi.org/10.2514/6.1992-2396.

[4] Schwarz, J. B., Dowell, E. H., Thomas, J. P., Hall, K. C., Rausch, R. D., and Bartels, R. E., “Improved Flutter Boundary
Prediction for an Isolated Two-Degree-of-Freedom Airfoil,” Journal of Aircraft, Vol. 46, No. 6, 2009, pp. 2069–2076.
https://doi.org/10.2514/1.30703.

[5] Levin, D., “Convolution and Volterra Series Approach to Reduced Order Modelling of Unsteady Aerodynamic Loads and
Improving Piezoelectric Energy Harvesting of an Aeroelastic System,” Ph.D. thesis, Duke University, May 2020.

[6] Stanford, B. K., Jacobson, K. E., and Massey, S. J., “Transonic Aeroelastic Modeling of the NACA 0012 Benchmark Wing,”
AIAA Paper 2020-2716, 2020. https://doi.org/10.2514/6.2020-2716.

[7] Hall, K. C., Thomas, J. P., and Clark, W. S., “Computation of Unsteady Nonlinear Flows in Cascades Using a Harmonic Balance
Technique,” AIAA Journal, Vol. 40, No. 5, 2002, pp. 879–886. https://doi.org/10.2514/2.1754.

[8] Thomas, J. P., Dowell, E. H., and Hall, K. C., “Nonlinear Inviscid Aerodynamic Effects on Transonic Divergence, Flutter, and
Limit-Cycle Oscillations,” AIAA Journal, Vol. 40, No. 4, 2002, pp. 638–646. https://doi.org/10.2514/2.1720.

[9] Thomas, J. P., Dowell, E. H., and Hall, K. C., “Modeling Viscous Transonic Limit Cycle Oscillation Behavior Using a Harmonic
Balance Approach,” Journal of Aircraft, Vol. 41, No. 6, 2004, pp. 1266–1274. https://doi.org/10.2514/1.9839.

[10] Ni, R.-H., “A Multiple-Grid Scheme for Solving the Euler Equations,” AIAA Journal, Vol. 20, No. 11, 1982, pp. 1565–1571.
https://doi.org/10.2514/3.51220.

13
 [11] Saxor, A. P., “A Numerical Analysis of 3-D Inviscid Stator/Rotor Interactions Using Non-Reflecting Boundary Conditions,”
Gas Turbine Laboratory Report 209, Massachusetts Institute of Technology, March 1992.

[12] Spalart, P. R., and Allmaras, S. R., “A One-Equation Turbulence Model for Aerodynamic Flows,” AIAA Paper 1992-0439,
January 1992. https://doi.org/10.2514/6.1992-439.

[13] Dowell, E. H., A Modern Course in Aeroelasticity, 5th ed., Springer, Berlin, New York, 2015.

[14] Hosder, S., Walters, R. W., and Balch, M., “Point-Collocation Nonintrusive Polynomial Chaos Method for Stochastic
Computational Fluid Dynamics,” AIAA Journal, Vol. 48, No. 12, 2010, pp. 2721–2730. https://doi.org/10.2514/1.39389.

[15] Stanford, B. K., and Massey, S. J., “Uncertainty Quantification of the FUN3D-Predicted Flutter Boundary on the NASA CRM,”
AIAA Paper 2017-1816, January 2017. https://doi.org/10.2514/6.2017-1816.

[16] Stanford, B., Sauer, A., Jacobson, K., and Warner, J., “Gradient-Enhanced Reliability Analysis of Transonic Aeroelastic Flutter,”
AIAA Paper 2022-0632, January 2022. https://doi.org/10.2514/6.2022-0632.
Downloaded by Northwestern Polytechnic University on April 29, 2024 | http://arc.aiaa.org | DOI: 10.2514/6.2022-3869

[17] Eldred, M., “Recent Advances in Non-Intrusive Polynomial Chaos and Stochastic Collocation Methods for Uncertainty Analysis
and Design,” AIAA Paper 2009-2274, May 2009. https://doi.org/10.2514/6.2009-2274.

14