Вы находитесь на странице: 1из 9

National Conference on

Contemporary Approaches in Design and Manufacturing


ACCET, Karaikudi-4.

PARAMETRIC STUDY OF AIRFOIL ANALYSIS USING CFD


V.GIRIDHARAN *, S.RAJAGOPAL **, A.LURDHARAJ ***, &
Dr.S.ELANGOVAN ****.

*V.Giridharan, M.E. student, aeronautical, M.I.T, Anna University.


giriaero_mit@yahoo.co.in

**S.Rajagopal, Scientist ‘D’, ADE, DRDO.


Raja_gopal_2001@yahoo.com

***A.Lurdharaj, Scientist ‘F’, ADE, DRDO.


**** Dr.S.Elangovan, Assistant Professor, MIT, Anna University.
subelango@yahoo.co.in

ABSTRACT
Aeronautical Development Establishment (ADE), DRDO is involved in the design and development of Unmanned
Aerial Vehicles (UAV). It has successfully designed, developed and delivered UAVs to the Indian Services, which
are aerial target system and tactical UAV. Currently it is involved in design and development of a medium
altitude long endurance (MALE) UAV. During the process of conceptual and preliminary design of the MALE
UAV, ADE has designed an indigenous airfoil that is required to have high aerodynamic efficiency in turn
resulting in long endurance of the UAV. The theoretically designed airfoil has to be analyzed computationally and
experimentally in order to ensure that it gives the desired performance. In order to carry out the computational
analysis using any CFD codes, proper choice of the turbulence model has to be done. Therefore, parametric study
on different airfoils to check the effect of various turbulence models like one-equation model (spalart-alarmas
model) and two equations model (k-ε and k-ω model), was carried out. For this study, three airfoils (NACA
641612, NLF-0416, and GAW) were considered, which had experimental test values. Airfoils were modeled in
GAMBIT and solved in FLUENT with different turbulence models. The results obtained with different turbulence
models have been compared with experimental values and also with XFOIL (another airfoil analysis code, which
is used for design of airfoils) values. Obtained values agreed well with the experimental values. This study
concludes that for lower angle of attack one-equation model predicts well for higher angle of attack k-ε two-
equation model is well suited. With this conclusion, the ADE designed airfoil is also analyzed and the results are
presented.
Keywords: Stalling angle, Turbulence model, Lift curve slope, Endurance.
1. INTRODUCTION

In the recent years, there has been lot of research going on in the design and development of
Unmanned Aerial Vehicles (UAVs). It is because of their high potential usage for both civil and
military applications. They are in the forefront of current aeronautical research as they are able to
perform variety of missions, including reconnaissance, surveillance, targeting and so on. ADE has been
involved in the design and development of UAVs for the past 25 years. It has successfully designed,
developed and delivered UAVs to the Indian Services, which are aerial target system and tactical UAV.
Currently it is involved in design and development of a medium altitude long endurance (MALE)
UAV.
In order to satisfy the long endurance requirement of MALE UAV, the aerodynamic
efficiency of the wing has to be high and therefore the airfoil has to be designed in such a way that it
produces a minimum operating Cl of 1.3 to 1.4 (so that the endurance parameter is high thereby giving
maximum endurance for the aircraft), which in turn decides the maximum Cl of the airfoil to be around
2.1. The airfoil also comes up other requirements of operating at adverse weather conditions (ice
formation) and low Reynolds number (the aspect ratio of the wing is large with a taper ratio of 0.4,
hence chord variation is much thereby making the wing to operate at various Reynolds number). The
catalogue airfoils don’t have these high operating Cl and don’t satisfy the other requirements also.
Hence ADE has designed an indigenous airfoil that satisfies the all the requirements.
The theoretically designed airfoil has to be analyzed computationally and experimentally in
order to ensure that it gives the desired performance. The best way to carry out the computational
analysis is to use well-validated commercially available CFD software. Whenever the CFD
computations with commercially available software are used for the analysis of airfoils, proper choice
of the turbulence model have to be made. The standard one equation and two equation turbulence
models available in the commercial software have been well validated over the last few years and have
been extensively used for attached and mild separated flows. The objective of this investigation is to
identify the best turbulence model for the purpose of low speed and low Reynolds number airfoils,
which are being used for the endurance UAVs. In order to identify the best turbulence model, a
parametric study of effect of various models on the airfoil parameters have been carried out and
presented in this paper. Three low speed airfoils were used for this parametric study and the
methodology of analysis are briefly describes in the following sections.

2. TURBULENCE MODELS:

The Spalart-Allmaras model6 is a relatively simple one-equation model that solves a modeled
transport equation for the kinematic eddy (turbulent) viscosity. This embodies a relatively new class of
one-equation models in which it is not necessary to calculate a length scale related to the local shear
layer thickness. The Spalart-Allmaras model was designed specifically for aerospace applications
involving wall-bounded flows and has been shown to give good results for boundary layers subjected
to adverse pressure gradients. It is also gaining popularity for turbo machinery applications. In its
original form, the Spalart-Allmaras model is effectively a low-Reynolds-number model, requiring the
viscous-affected region of the boundary layer to be properly resolved. This model makes it the best
choice for relatively crude simulations on coarse meshes where accurate turbulent flow computations
are not critical. Furthermore, the near-wall gradients of the transported variable in the model are much
smaller than the gradients of the transported variables in the - or - models.
1
The standard - model is a semi-empirical model based on model transport equations for
the turbulence kinetic energy ( ) and its dissipation rate ( ). The model transport equation for is
derived from the exact equation, while the model transport equation for was obtained using physical
reasoning and bears little resemblance to its mathematically exact counterpart. In the derivation of the
- model, it was assumed that the flow is fully turbulent, and the effects of molecular viscosity are
negligible. The standard - model is therefore valid only for fully turbulent flows. The shear-stress
transport (SST) - model was developed by Menter to effectively blend the robust and accurate
formulation of the - model in the near-wall region with the free-stream independence of the -
model in the far field. To achieve this, the - model is converted into a - formulation. The
SST - model is similar to the standard - model, but includes the following refinements:
The SST model incorporates a damped cross-diffusion derivative term in the equation.
The definition of the turbulent viscosity is modified to account for the transport of the turbulent shear
stress.
The modeling constants are different.
These features make the SST - model more accurate and reliable for a wider class of flows (e.g.,
adverse pressure gradient flows, airfoils, transonic shock waves) than the standard - model.

3. COMPUTATIONAL GRID AND BOUNDARY CONDITIONS:

The geometries and the grids are generated with Gambit (version 2.1.), a preprocessor of Fluent The
domain of the airfoil has a C-shape in order to get structured grid. Element used is quadrilateral1.The
computational domain has a fine grid at airfoil and coarsens toward outer boundary. The grid of airfoil
surface is refined at the leading edge where the largest pressure gradients, due to leading edge suction.
Also fig1 give an idea of the grid around an airfoil with 8340 cells. For all the airfoils same procedure
has been attempted for grid generation. For the case of boundary conditions airfoil is considered as
wall, where as the domain is considered as far-field. The conditions of far-field flow are pressure-
101328 pa; Reynolds number- 2million; turbulent intensity -.01%. The direction of flow is specified in
Cartesian coordinates and depends on the angle of attack. More details on number of nodes, edges,
faces and elements are given below

Number of nodes : 8602


Number of faces : 17808
Number of cells : 8430
Element used : quadrilateral
Mesh scheme : Mapped mesh.

4. FLOW SOLVER AND METHODOLOGY:

2,6
Fluent version 6.1 is used for simulations. The segregated solver (i.e. each equation solved
individually) is used with an implicit formulation and second orders scheme for pressure terms,
QUICK (Quadratic Upwind Interpolation Convective for Kinematics) scheme for convective derivative
and a second order upwind scheme for other flow variables. The constant is chosen for density. The
Spalart-Allmaras model and the standard - model were used as turbulence model. The proper
choice of turbulence must be done and is discussed later. No transition model is implemented in fluent
because of the lack of a well-documented and validated transitional model for airfoils at this moment.
Before simulation of E2 airfoil, a series of simulations on all other three airfoils are done. Experimental
data is available for comparisons. These simulations are necessary to address a few important
questions, such as which turbulence model should be used to predict lift of the profile accurately and
the drag as best as possible? At what angle flow gets stall? Finally, a total of 120 two-dimensional
computations are executed on three different turbulence models for 8 angles of attack. Every
computation consists of 1500 iterations. Convergence reached after above 500 iterations for lower
angles of attack, but Convergence reached after above 1500 iterations for higher angles of attack.

5. RESULTS AND DISCUSSION:


The results obtained with different turbulence (S-A model and K-E model) have been
compared with experimental values and also with Xfoil values. From these values it is observed which
turbulence model well agreed with experimental values for airfoil analysis and the same model is
considered for wing analysis also. The results of comparison are shown in table-1. Lift curve slopes (CL
vs. α) for all the airfoils are plotted with different comparisons are shown in fig2 to fig6. In fig2 the lift
curve slope obtained from fluent for four different airfoils are plotted. E2 airfoil gives maximum lift
coefficient compared to all other airfoils as expected.
In fig3 lift curve slope for NACA 641612 is plotted with comparisons of experimental values
and fluent K-e values. Similarly fig4 and fig5 gives the lift curve slopes for NLF-0416and GAW-1
airfoils respectively. From the figures it is clear that for lower angles of attack S-A model matches well
with experimental data, for attached flow K-e model matches well. The uncertainty on the
experimental data is less than 2% for lift coefficient for attached flows and (2-10)% for separated
flows. Fig7 shows the velocity vectors in the order of S-A model, K-e model for NLF-0416 .The ADE-
LS-E2 airfoil, was also analyzed using different turbulence (S-A model and K-E model) and the results
are shown in fig6. The velocity vector for ADE-LS-E2 airfoil is shown in fig8.

TABLE-1: COMPARISION OF CLMAX FOR DIFFERENT AIRFOILS


AIRFOIL FLUENT VALUES EXPERIMENTAL XFOIL
VALUES VALUES

S-A model K-ε model

NACA641612 1.3022 1.4007 1.397 1.6


Re = 1.5 mil

NLF-0416 1.5806 1.5406 1.58 1.7


Re = 2 mil

GAW 1.6409 1.6256 1.6044 1.85


Re = 2 mil

ADE-LS-E2 1.6299 1.7 - 2.1


Re = 2 mil

Fig1: airfoil mesh details (same for all the airfoils)


Full view

Magnified view
Fig2: Lift curve for different airfoils

Fig3: FOR NACA 641612

Fig4: FOR NLF-O416


Fig5: FOR GAW

Fig6: FOR ADE-LS-E2 AIRFOIL

Fig7: VELOCITY VECTOR FOR E2: At CLMAX


7.a S-A model
7. b K-ε model

FIG8: VELOCITY VECTOR FOR E2: At CLMAX


8. a S-A model

8. b K-ε model
6. CONCLUSIONS:

This study concludes that for lower angle of attack one-equation model predicts well, for
higher angles (i.e. upto stalling angle) k-ε two-equation model is well suited for the analysis of
indigenous designed ADE-LS-E2 airfoil, the k-ε two-equation model available in the commercial
software has been used. Anyhow the design data would be frozen only after evaluating the performance
of the airfoil in the experimental wind tunnel test, which is scheduled to be held shortly.

7. REFERENCES.
1. F.R.Menter and J.A.Ekaterinaris “Computation of Oscillating Airfoil with One-and Two-Equation
Turbulence Models”, AIAA Paper, vol.32, no12, december1994

2. David C.Wilcox, “Simulation of Transition with Two-Equation Turbulence model”, AIAA


Journal, vol.32, no2, Feb1994.

3. Patrik Rautaheimo and Timo Siikonen, “Improved Solid-Wall Boundary Treatment in Low
Reynolds-Number Turbulence Models”AIAA Journal, vol39, no5, May 2001.

4. Taodu and Zi-Ziuwu, “Mixed Analytical/Numerical method for Low Reynolds Number
K-E Turbulence Models”, AIAA Journal, vol.42, no6, June 2004.

5. J.R.Edwards “Development of a One –Equation Transition/Turbulence model”, AIAA Journal,


vol.39, no9, September 2001.

6. Fluent users guide: section11.2.4, 11.2.5 and 11.2.9.

8. ACKNOWLEDGEMENTS:

The authors would like to acknowledge the support rendered by all the members of aerodynamics
division in carrying out this work and Director, ADE for allowing presenting this work.

Вам также может понравиться