z

Available online at http://www.journalcra.com

INTERNATIONAL JOURNAL
OF CURRENT RESEARCH
International Journal of Current Research
Vol. 8, Issue, 08, pp.35973-35982, August, 2016

ISSN: 0975-833X
RESEARCH ARTICLE
FLUTTER ANALYSIS OF AIRFOIL USING FLUID STRUCTURE COUPLING BASED
ON EULER EQUATIONS
*,1Eng. Magdy Saeed Hussin, 2Prof. Dr. Mohamed A. El-Samanoudy and 2Dr. Ashraf Ghorab
1Arab Organization for Industrialization, Aircraft factory, Cairo, Egypt
2Mechanical Power Engineering Dept., Ain-Shams
Ain Shams University, Cairo, Egypt

ARTICLE INFO ABSTRACT

Article History: Aero-elasticity

elasticity is the subject that describes the interaction between the deformation of an elastic
Received 14th May, 2016
structure in airstream and resulting aerodynamic force especially on aircrafts. This work studies the
Received in revised form numerical analysis of the airfoil flutter. To do this work, Euler equations are integrated in time to get
21st June, 2016 aerodynamic forces, then, at each time step, these aerodynamic forces are coupled with airfoil
Accepted 18th July, 2016 equations of motion to simulate its flutter. Two numerical techniques are use used in this work ;
Published online 20th August, 2016 McCormack’s technique to evaluate aerodynamic forces, and Newmark’s technique to evaluate airfoil
dynamic response. MATLAB program was developed for implementation of these techniques. For
Key words: solution verification and validation, the work was implemented on NACA 0012 airfoil because of
availability of its section parameters and aero-elastic
aero elastic characteristics. Most previous researches on
Aero-elasticity, flutter didn’t focus on two important flutter points. The first point is the numerical error produced in
Airfoil flutter, flutter
er velocity when using low coupling frequency (CFD-CSD).
(CFD CSD). The second point is flutter velocity
McCormack’s technique, obtained by classical method is so close to that obtained by using sophisticated codes like Euler
Newmark’s technique, equations. As a result, the two cases have been investigated in the next pages and results presented :
Euler equations. (I) Substantial error at using long coupling step (with time step=0.002s flutter velocity=174m/s, with
time step=0.01s flutter velocity=166.8m/s). (II) Classical method can be used as first estimation of
flutter because it give tolerable results (flutter velocity based classical=175m/s, flutter base Euler
equations=174m/s).

Copyright©2016, Eng. Magdy Saeed Hussin et al. This is an open access article distributed under the Creative Commons Att
Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Citation: Eng. Magdy Saeed Hussin, Prof. Dr. Mohamed A. El-Samanoudy

El and Dr. Ashraf Ghorab, 2016. “Flutter
Flutter analysis of airfoil using fluid
structure coupling based on Euler equations”, International Journal of Current Research,
Research 8, (08), 35973-35982.

INTRODUCTION aerodynamic surfaces under dynamic motion was calculated by

Wagner’s and Theodorsen’s
heodorsen’s Functions (Fung, 1955; Jan R.
Subsonic airfoil flutter is the topic of this research. This is Right and Jonathan E.Cooper, 2007). Aircraft flutter studies
owing to the fact that flutter commonly leads to failure in based on classical methods don't give accurate results in
aircrafts. So, it is a critical factor in the design of some transonic regimes because the non non-linear in the transonic
airframes and aircraft engines. As the problem of aircraft regime results in large variations of aerodynamic forces with
flutter involves
lves interaction between Computational Fluid small changes of the aerodynamic shape. Thirdly, time
Dynamic (CFD) with Computational Structure Dynamic marching analysis of flutter problems has been used, this
(CSD), it is important to couple their codes. Historically, method uses coupling between fluid models and structure
flutter problems at first were solved based on the direct models. Time marching analysis is accurate and adequate for
eigenvalue approach; where lift gradient iss prescribed by the nonlinear fluid/structure models. This work assumes quasi
classical, incompressible, non-viscous
viscous aerodynamic theory. steady aerodynamic flow during simulation and fluid domain
This method does not include structure damping and assumes mesh is attached to the airfoil surface and moves with it
linear aerodynamic and structural models. Secondly, damping accordingly where pitch angles are assigned to the boundary
from viscous structure and unsteady aerodynamics are conditions. In the past, the aerodynamic models which used in
incorporated
porated in the classical model, where the behaviour of aero elasticity were two-dimensional
dimensional strstrip theory or three-
dimensional panel methods, however, these models didn’t
*Corresponding author: Eng.. Magdy Saeed Hussin, predict the occurrence of shock waves in the transonic flight
Arab Organization for Industrialization, Aircraft factory, Cairo, Egypt. regime (Jan R. Right and Jonathan E.Cooper, 2007). Right
now, there is an increasing interest in the effect of aerodyna
aerodynamic
35974 Eng. Magdy Saeed Hussin et al. Flutter analysis of airfoil using fluid structure coupling based on Euler equations

and structural nonlinearities and the effect of that on the ' I ' , is section mass moment of inertia about support point,
aeroelastic behaviour; CFD based Euler/ Navier–Stokes
coupled with structural models are used now .This work S  mxcg , is the mass unbalance
presents a procedure for solving fluid-airfoil interaction in two-
dimensional subsonic flow. The solution of fluid flow is based The potential energy is simply the energy stored in the two
on explicit solution of unsteady Euler equations by springs, and given by:
McCormack’s techniques (Anderson, Jr.1995; Klaus A.
Hoffmann and Steve T. Chiang. 2000) using MATLAB finite
1 1
differences code. The CSD is based on the direct time V  K h h 2  K 2 ………………..(2)
integration of airfoil equation of motion by Newmark’s 2 2
method, where implicitly the solutions of the flow field are
coupled in time by structural equation of motion (Michel Section dissipative energy '' given by:
Géradin and Daniel J. Rixen, 2015). Results of Newmark’s
1 1
method have been investigated at different CSD-CFD coupling  C h h 2  C  2 ………………...(3)
frequencies to investigate the numerical error. We also 2 2
calculated flutter boundary for this problem by classical
methods and the results compared with that of numerical The equations of motion can be obtained by using Lagrange’s
method equation:

Airfoil aeroelastic model

d  Tk  T
  k  (  )  V  L …………….. (4)
Figure1. Shows sketch of the airfoil which is considered as a dt   h.

 h
 h h
rigid section, supported by translational and rotational springs,
so the model features are only two degrees of freedom,
d  Tk  T
  k  (  )  V  M
pitching ' ' and heave 'h' .Chord length is 'c' , ' xcg ' is ……………..(5)
dt   .    
distance between center of mass and center of support, ' ' x cp' is
 

position of center of pressure, ' ' x0' is position of support, This leads to the airfoil equation in matrix form:
' L' ,' M' are resultant aerodynamic lift and moment ..
 m - S   h  C h 0  h   K h 0  h   L  …(6)
respectively (positive direction as on the sketch), ' Kh ,' Kα'  S       
  I    ..  0 C    0 K     M 

 
are section stiffness properties in heave and pitch respectively,
' Ch' ,' Cα' are corresponding damping properties, ' m' is the ..
, F  [L M]T
Μ U   [C]U  K U  F, U  h   …...(7)
T

section mass.  

Flutter investigation methods

Time marching method based on CFD

In this approach airfoil equation of motion (7) has been solved

numerically in time domain. Here aerodynamic forces were
calculated by CFD based on unsteady Euler equations. In this
approach quasi-steady aerodynamic flow employed to get the
aerodynamic forces.

Flow governing equations

The aerodynamic model written in conservation forms as:

Fig.1. Sketch of NACA 0012 geometry and parameters
Q E  G
  …………..…. (8)
Derivation of motion equations as follows (Fung, 1955; Jan R. t x  y
Right and Jonathan E.Cooper, 2007):
   u   v 
Section kinetic energy ' Tk ' given by:    2   
 u   u     uv  …..(9)
Q ,E    ,G   2 
v 
1 1    uv   v   
Tk  mh 2  hS  I 2 ……………...(1)  E 
 t
 ( E   )u 
 t 
 ( E   )v 
 t 
2 2
35975 International Journal of Current Research, Vol. 08, Issue, 08, pp.35973-35982, August, 2016

1 t
  (   1 ) ( E t  ( u 2  v 2 )) (10) t  t  Q 
t
2 Q  Q   t (20)
 t i , j
Here ' Q' , ' E' , ' G' are variable vectors and flux vectors. '  '
Step 2: Corrector step:
is density, ' u ' , ' v ' are Cartesian velocity components , ' P' is
static pressure, ' E't is the total energy per unit volume, '  ' is  Q 
t  t
 Eti,j t  Eit1,tj G i,t j t  G i,t j-1t 
 
ratio of specific heat. In this work, equations transformed from  t     


 (21)
Cartesian coordinates ( x , y ) to general curvilinear coordinates  i , j  
(  , ) for ease of assign airfoil boundary conditions. The
transformed equations from Cartesian coordinates to Coupled Fluid-Structural Interaction Procedure
computational coordinates are expressed as (Klaus A.
Hoffmann and Steve T. Chiang, 2000): In this work Newmark's technique has been used to solve
airfoil equation (7) in time domain through external coupling
 Q  E G with CFD. This technique presents single-step integration
  0 (11) method to get structural displacements and velocities at next
t  
time step as (Michel Géradin and Daniel J. Rixen, 2015;
Q Manjuprasad et al., 2009):
Q (12)
J
..
1
E  (  xE   yG ) (13) Μ U  [C]U  K U  F (7)
J  
1 givens are mass matrix ' M' , damping matrix ' C' , stiffness
G  ( xE   yG ) (14)
J matrix ' K' , initial displacements ' U0' ,  .
and velocities U 0
1 Specify Newmark’s parameters '  ' ,'  ' . Calculate
J (15)
x y  x y integration constants ' a' ,' b' and calculate effective stiffness
matrix K̂ .
' J' is Jacobian, and  x  Jy ,  y   Jx ,  x   Jy , y  Jx
are metrics of transformation. The superscript“__” which used  1
to designate generalized coordinates will be dropped for the aM C (22)
 t 
remainder of analysis.
1 
bM  C(  1 )t (23)
Numerical model of Euler equations 2 2

In this application McCormack’s technique was used because ˆ  K C  M 1

K
it is much simpler in its application. It is an explicit finite-  t  t 2
difference technique which is a second-order-accurate in both (24)
space and time.
Iteratively for each time step calculate:
Q E G
  0 (16)
t   ˆ  F  F  aU   bU

F t  t t t t (25)
t  t  Q 
Q  Qt    t (17) U  ΔF
t  t
ˆ /K
ˆ U
t (26)
 t  average
 1 1  1  
  U t  t  2
( U t  t - U t ) - Ut - Ut  Ut (27)
  t t 2
t  t
   
t  t  U t  1   t U t  t U t  t
t
 Q  1   Q   Q   (18) U (28)
        
 
 t  average 2   t i , j  t i , j 
    
 from predictor step from corrector step
NB: at   1 / 2 ,   1 / 4 the solution is unconditional stable
 (19) (21)  with asymptotically the highest accuracy
Step 1: Predictor step:
Direct eigenvalue method (classical)
t t t t t
 Q  E  E G G 
i 1, j i, j i, j1 i, j In this method, we use Eigenvalues solutions to obtain the
      (19) natural frequencies and damping ratios of the system at
 t i , j     different airspeeds, the lift gradient with respect to angle of
35976 Eng. Magdy Saeed Hussin et al. Flutter analysis of airfoil using fluid structure coupling based on Euler equations

attack is constant and obtained by thin airfoil theory. Resultant The unsteady aerodynamic forces and effective angle of attack
lift/moment about airfoil support point '0' given by (Houghton are given by (Fung, 1955):
and Carpenter, 2003):
1  1 3 
C  is the unsteady pitch angle  eff    h   c  x0  (36)
L  q  cC L  q  c L  , (29)
V V  4 

M  q cC L ( x0  xcp )  q cC L ( lc ) C L
(30) L  q  cC L  q  c  eff (37)
Cl 
 q c 2  l  q c 2 C m
 C  c2 
1 M  L( x0  xcp )  q c L   (38)
q    V2 is the free stream dynamic pressure   16V 
2
After substituting (36), (37), (38) into equation (7) we get:
Substituting these lift, moment expressions in airfoil equation
(7) and neglect structure damping:
h h  h  0 
[K]  [A]   [M]   [C]  [C A ]     (39)
..      0 
 m - S  h   K h 0  h  C L   (31)
 S   ..        q c  
  I     0 K      lc  Where [A], [C A ] are aerodynamic stiffness and damping

 
matrix respectively;
h  ho  t
Assume      exp and substituting this
  
 o C l 0 1 
displacement vector in equation (31) we get: [A]  q c 0 - (x  x )
  0 cp 

 m - S  2  K h 0  Cl 0 1   h0  0 (33) (32)

     q c 0 - (x  x )      
  S I  0 K    0   0  0
cp  
1 3 / 4c  x0 
   (40)
Cl V V

In matrix form the above equation can be written as: [CA ]  q c 2
 (x0 - xcp ) (x0  xcp )( 3 / 4c  x0 ) c 
 -  
0 1  V V 16V 
[K]  [A]   [M]h   0,
2 0
[A]  q c
Cl

0 - (x  x

)
(34)
 0   0 cp
h
And the non trivial solution is: Assume x1  U    ,
 

det [K] [A] λ 2[M]  0,  (35) h 
 
h   x1   
The above equation is an eigenvalue problem which when x 2   U   ,x        this lead to

solved we get flutter velocity for the undamped system. The    x2  h 
C L  
parameters '   ' ,' V ' ,' c' ,' C L ' ,' C m ' , , q are free

stream density, free stream velocity, chord length, lift
coefficient, moment coefficient, lift gradient, and free stream
x 1  x 2   U 
dynamic pressure respectively. The previous analysis has been
presented in (Jan R. Right and Jonathan E.Cooper, 2007; After substituting the above expressions into equation (39) we
Manjuprasad, 2009). get state space form:

State space method (classical) 0 0 1 0  , or x   [S]x (41)

x   0 0 0 1

x
In this method the aeroelastic equation (7) can be solved based - [M]-1[K  A] - [M]-1[C  C ] 
 A 
on matrix approach by transforming it from second order to
first order (state space) (Jan R. Right and Jonathan E.Cooper,
2007; Sdmenath Mukherjee et al., 2008) Flutter conditions can be obtained from equation (41) either by
time marching or by system’s eigenvalues. The eigenvalues
.. derived as follow:
ΜU  [C]U  K U  F (7)
 
35977 International Journal of Current Research, Vol. 08, Issue, 08, pp.35973-35982, August, 2016


Assume, x    x  [S] - [I]  0 
where [I] is with increasing angle of attack. Figure 2a shows that lift
identity matrix. This assumption transforms equation (41) to coefficient increase linearly with angle of attack like thin
airfoil theory.
eigenvalue problem as:

det [S] - [I]  0, (42) Table 1. Airfoil properties and initial conditions of motion
Modeling unsteady aerodynamic forces by Theodorsen’s
function Geometry Profile:NACA0012 airfoil(symmetric)
Inertia
The quasi-steady assumption is not sufficiently accurate for m=51.5kg, I  2.275kgm 2 , x0  0.4m
flutter calculations because it doesn’t predict the dependency , x cg  0.0429
of aerodynamic forces on the airfoil oscillation. The key tools
to analyse these effects are Wagner’s and Theodorsen’s Stiffness K h  50828.463N/m ,
functions (Jan R. Right and Jonathan E.Cooper, 2007). K   35923 .241Nm/rad
Theodorsen’s complex function C(k) is used here to model the Damping Ch
phase difference between the aerodynamic loading and the h   0.01 , Ch  32.358Ns / m
response. 2 Khm
0.165 0.335  
C
 0.01 ,
C  5.71Nms/rad
C(k)  1 -  , k  0.5 2 K h I
0.045 0.30
1- i 1 i Initial displacements
k k h0  0.01m , 0  0 rad
0.165 0.335 (43) Initial dh  d 
C(k)  1 -  , k  0.5 velocities   0.001m / s ,   0.01 rad/s
0.041 0.32 dt  0 dt  0
1- i 1 i
k k
c Time marching flutter simulation
k , k is reduced frequency,  is frequency obtained
2V
Numerical results of flutter analysis using CSD-CFD with
from imaginary part of eigenvalue. The modified aerodynamic
simulation time step equal 0.002s are shown in Figures 3-5.
matrices are as follow:
From these figures, we can investigate that the airfoil motion
0 1  and the corresponding aerodynamic lift. These figures show
Cl that at free velocity 174.5m/s heave and twist oscillate
[A]  C(k)* q c 0 - (x  x ) (44)
  0 cp  unboundedly and airfoil displacements increase exponentially
with time and system is unstable due to the aerodynamic forces
1 3 / 4c  x0  overcome the restoring forces due to structural stiffness. As
   (45) free velocity reaches 174m/s the system reaches the stable
Cl V V

[C A ]  C(k)* q c 2 point where both heave and twist motions are simple harmonic
  (x0 - xcp ) (x0  xcp )( 3 / 4c  x0 ) c 
 -   motion and their amplitudes remain constant with time and this
 V V 16V 
velocity is the flutter velocity of the airfoil. At free velocity
173.5m/s heave and pitch motions oscillate and their
Numerical results and discussions
amplitudes decay with time. Figure 5 show that lifting force
Results are presented here for symmetric NACA 0012 airfoil follows the pitch motions.
(Figure 1) of unit chord, and unit span and the fluid-structure
interaction is numerically simulated with small initial Stability and numerical errors of simulation algorithm
conditions, all calculations executed by using MATLAB
codes. The stiffness and damping properties of springs are The dynamic coupling in this work is by using Newmark’s
chosen so that flutter instability occurs in the subsonic regime. technique with average acceleration. This method is better to
Here we assume infinite rectangular wing, so airfoil properties solve aero-elastic system, because the algorithm is
remain the same at any location along the span. The airfoil unconditionally stable and don't have numerical error in the
structural properties and initial conditions are listed in Table 1. amplitude (amplitudes of oscillation don't depend on the length
of simulation time step).
NACA 0012 airfoil performance at steady conditions
Unfortunately length of simulation time step affects on the
Steady pressure distribution along airfoil surface and the period of oscillation because the period depends on the
corresponding coefficient of lift were calculated for different simulation time step; as step length increases as oscillation
angle of attack at Mach=0.5 as shown in Figure 2. These period elongates (Michel Géradin and Daniel J. Rixen, 2015).
results are obtained to validate the present CFD code with the Flutter simulation using longer time step (0.01s) are shown in
available results. From figures at '   0 0 ' the net area Figures 6-8.From these Figures we can see that flutter velocity
enclosed by Cp curve is zero, so no aerodynamic force acting is 166.8 m/s. which is less than flutter velocity estimated in
0 0 previous section when the time step was (0.002s), So length of
on airfoil. At '  2 ,   6 ' the stagnation point creeps simulation time step has strong impact on the solution
towards the lower surface and the area enclosed by Cp curve accuracy.
increases with angle of attack so aerodynamic forces increase
35978 Eng. Magdy Saeed Hussin et al. Flutter analysis of airfoil using fluid structure coupling based on Euler equations

(a) (b)

(c) (d)

Fig. 2. Airfoil performance at Mach=0.5.(a) ' CL ' vs. ' ' .(b) ' C p ' distributions at   0 0 . (c) ' C p ' distributions at   2 0 ,(d)
' C p ' distributions at   6 0

Fig. 3. Simulation of pitch motions with free velocities173.5m/s.-174m/s-174.5m/s

35979 International Journal of Current Research, Vol. 08, Issue, 08, pp.35973-35982, August, 2016

Fig. 4. Simulation of heave motions with free velocities173.5m/s.-174m/s-174.5m/s

Fig. 5. Simulation of lift with free stream velocities173.5m/s.-174m/s-174.5m/s

Fig. 6. Simulation of heave motions with free velocities166.7m/s.-166.8m/s-167m/s

35980 Eng. Magdy Saeed Hussin et al. Flutter analysis of airfoil using fluid structure coupling based on Euler equations

Fig. 7. Simulation of pitch motions with free velocities166.7m/s.-166.8m/s-167m/s

Fig. 8. Simulation of lift with free stream velocities166.7m/s.-166.8m/s-167m/s

(a) (b)

Fig. 9. Trend of motion frequency/amplitude vs. free stream velocity for undamped system: (a) Trend of motion amplitude vs. free
velocity. (b) Trend of motion frequency vs. free velocity
35981 International Journal of Current Research, Vol. 08, Issue, 08, pp.35973-35982, August, 2016

(a) (b)
Fig. 10. Trend of motion frequency/amplitude vs. free stream velocity for damped system: (a) Trend of motion amplitude vs. free
velocity (b) Trend of motion frequency vs. free velocity

Table 2. Flutter velocity of NACA0012 airfoil using different methods

Method Flutter velocity

Classical direct eigenvalue approach, without any damping 175m/s.(Fig.9)
Classical method with aerodynamic and structural damping and Theodorsen’s function C(k) About 189m/s. (Fig.10)
Time domain simulation using FDM based Euler equations for generating aerodynamic forces in quasi steady flow (simulation 174m/s.
step=0.002s) (Figs.3-5)
Time domain simulation using FDM based Euler equations for generating aerodynamic forces in quasi steady flow (simulation 166.8m/s.
step=0.01s.) (Figs.6-8)
Time domain simulation using FEM based Euler equations for generation of aerodynamic forces in quasi steady flow 174.2m/s.
(Manjuprasad et al., 2009)
Time domain simulation using FEM based N-S solver (viscous laminar) for generation of aerodynamic forces in quasi-steady 184.55m/s.
flow (Amit Kumar Onkar et al, 2011)
Time domain simulation using ANSYS FLOTRAN CFD solver (viscous turbulent) for generation of aerodynamic forces in quasi 192.45m/s.
steady flow(Davinder Rana et al, 2009)
Time domain simulation using Panel code for generation of aerodynamic forces (Sdmenath Mukherjee et al., 2008) 161.3m/s.

Classical flutter analysis using direct eigenvalue method becomes positive and the other negative. Hence the system
becomes unstable, which is the flutter condition.
Figure 9 present graphically the results obtained by solving
eigenvalue problem defined by equation (35). From the Figure Comparison between flutter velocities obtained by
we can see that for this system as the air speed increases, the different methods
two frequencies (heave and pitch) move closer to each other
and the damping of both modes remains at zero. The two The flutter velocity obtained by present time marching code
frequencies coalescence at velocity 175 m/s, and at this point has been compared with those obtained by classical and other
one real part crossing velocity axis towards positive area, so methods as shown in Table 2, the results of comparison are as
one of the damping ratios becomes positive and the other follow:
negative. Hence the system becomes unstable, which is the  The present flutter velocity based Euler equations is low
flutter condition. compared to the velocities obtained using other time
domain methods. This due to the fact that the present
Flutter analysis using State space method for the damped approach doesn't include viscous effect of flow(no
case friction drag) which leads to high generation of
aerodynamic forces for airfoil compared to laminar
Figure 10 present graphically the results obtained by solving viscous and turbulent viscous flow.
equation (42) using state space method. In Figure10 the  From the Table we see that at the same conditions
imaginary parts of the roots indicate the circular frequencies different simulation time steps result in different flutter
(radians/sec) of the two modes (heave/pitch) while the real velocities (CSD numerical error).To get reasonable
parts indicate decay/increase of amplitudes (damping ratio) results simulation step should be  0.1Tn , where Tn the
with time. Figure 10 shows that the two frequency modes
period of higher mode frequency (Michel Géradin and
move closer to each other near instability condition and at
Daniel J. Rixen, 2015).
flutter condition (189m/s) one real part crossing velocity axis
towards positive direction. In this case it is clear that presence  From the Table we also see that flutter velocity
of structural/aerodynamic damping delay flutter occurrence obtained by classical method (175m/s) is so close to
from flow velocity 175m/s up to about 189m/s. For flow that obtained by using sophisticated codes based Euler
regimes beyond this velocity, one of the damping ratios equations (174m/s), so classical method give tolerable
results and can be used for flutter first estimation.
35982 Eng. Magdy Saeed Hussin et al. Flutter analysis of airfoil using fluid structure coupling based on Euler equations

Conclusion REFERENCES
In the present work a number of different approaches have Amit Kumar Onkar, A. Arun Kumar et al. 2011. "Flutter
been employed for modeling and analysis of linear/nonlinear prediction of an airfoil using fluid structure interaction in
flutter of symmetrical NACA 0012. To solve airfoil equation time domain through finite element method based Navier-
of motion in time domain we externally coupled CFD model Stokes solver", NAL TechnicalMemorandum, TM-ST-11 06.
(McCormack’s technique) and CSD model (Newmark’s Anderson, Jr. D. 1995. "Computational fluid dynamics: the
technique) by using MATLAB code. In time marching basics with applications", McGraw-Hill.
approach quasi-steady aerodynamic flow employed to get the Davinder Rana, Sandeep Patel et al. 2009. "Time domain
aerodynamic forces from the unsteady aerodynamic forces. simulation of airfoil flutter using fluid structure coupling
The airfoil is not moved and the pitch angle of airfoil is through FEM based CFD solver", National Aerospace
assigned to inlet free velocity. There are four important Laboratories-CSIR, Bangalore, India.
observations on the results of this work. The first observation Fung Y.C. 1955. "An introduction to the theory of aero-
is the numerical error produced in flutter velocity when using elasticity", John Wiley and Sons Inc.
low coupling frequency (CFD-CSD), to avoid this error use Houghton E.L. and P.W. Carpenter, 2003. "Aerodynamic for
reasonable time step in coupling (with time step=0.002s flutter Engineering Students", Fifth edition, Butterworth-
velocity=174m/s, with time step=0.01s flutter Heinemann.
velocity=166.8m/s). The second observation is flutter velocity Jan R. Right, Jonathan E. Cooper, 2007. "Introduction to
obtained by classical method is so close to that obtained by aircraft aero-elasticity and loads", Wiley, John &Sons.
using sophisticated codes like Euler equations, so classical Klaus A. Hoffmann, Steve T. Chiang. 2000. "Computational
method can be used as first estimation of flutter because it give fluid dynamics", fourth edition, engineering education
tolerable results(flutter velocity based classical=175m/s, flutter system USA.
based Euler equations=174m/s). The third observation is the Manjuprasad, M., Neeraj Kumar Sharma, Davinder Rana et al.
present flutter velocity is low compared to those obtained by 2009. "Time domain simulation of airfoil flutter using fluid
viscous laminar based Navier Stokes and viscous turbulent structure coupling through FEM based Navier-Stokes
based Average Navier Stokes. This is due to the present work solver", NAL Technical Memorandum,TM-ST-09-03.
neglects viscous effect of flow which lead to increasing in Michel Géradin, Daniel J. Rixen, 2015, "Mechanical vibrations
generation of aerodynamic forces (flutter velocity based theory and application to structural dynamics", John Wiley
Euler=174m/s, based N-S=184.55m/s, based turbulent= & Sons.
192.45m/s). The fourth observation is that the analysis model Sdmenath Mukherjee, M. Manjuprasad, Amit Kumar Onkar
in this work can be adjusted directly to predict flutter in et al. 2008, "Time domain simulation of airfoil in the
transonic and supersonic flow. subsonic regime using fluid structure coupling through
panel method", NAL structural technologies division.

*******