CFD Analysis of Flow Fields around a Wing-Body Airplane Model with an Unstructured Grid Method

 

Fluid Mechanics Laboratory, Department of Aeronautics and Astronautics,

Kyushu University, Fukuoka, 819-0395, Japan

 

Introduction

Recently, owing to a rapid development of computational performance and advancement of analysis method, Computational Fluid Dynamics (CFD) makes a rapid progress. In the aerospace field, CFD also plays an important role that is sometimes equal to that in an experiment or a test flight. In our laboratory, several researches have been conducted to establish advanced Reynolds Averaged Navie Stokes (RANS) models and LES-RANS hybrid models and to improve their prediction accuracy. So far, reasonable results have been achieved for flow fields around a 2-D aerofoil or some 3-D models with simple shapes. However, it is still difficult to apply these models to engineering flow fields that consist of 3-D complex shapes.

The purpose of this research is to assess a prediction accuracy of existing turbulence models and to obtain valuable knowledge for improving them, leading to the development of further advanced turbulence models applicable to various complex engineering flow fields. As a part of this objective, the present study aims to analyze flow fields around an airplane model with an unstructured grid method.

 

Numerical method and target of analysis

In this study, all the calculations are performed by the flow-simulation code ¡°FrontFlow/red [1]¡±, which is an open source program developed in the project ¡°Revolutionary Simulation Software 21 (RSS21) [2]¡±. In this study, large eddy simulation (LES) with the Smagorinsky model is adopted. A discretization method is based on a node-centered finite volume method, and the advection term is discretized by the 3rd¨Corder upwind differential scheme. In this study, Euler implicit method is used for the time integration.

DLR-F6 [3] model is selected as the present target of analysis. It was used in the 2nd AIAA CFD Drag Prediction Workshop [3] in June 2003. In this study, a Wing-Body (WB) model, which is a simpler configuration in the workshop, is chosen. The Reynolds number is Re=based on the mean aerodynamic code length and the inflow velocity M=0.75. Figure 1 and Table 1 show the wind channel model and the wind channel data, respectively.

 

 

Table 1 Summary of wind channel data.

Mach number

0.75

Re (based on MAC)

3.0¡Á106

Reference temperature

305K

Mean aerodynamic chord

141.2 mm

Half model reference area

72700.0 mm2

 

Grid system and computational conditions

In this study, prism cells are used near the airplane and far fields for the purpose of resolving a boundary layer and avoiding rapid change of volume of cells. It is noted that hexahedral cells are adopted in some particular regions near the wing and the wing-body junction. Tetrahedral cells fill up rest of the computational space. Figure 2 shows the computational grid used in this study and Table 2 shows computational grid data. To reduce the number of grid nodes, the model is cut in half and the symmetry boundary condition is specified at the cut section (Fig. 3).

 

 

Table 2 Computational grid data.

Nodes

927656

Elements

Tetrahedral

1338418

Hexahedral

505746

Pyramid

29342

Prism

324590

Total

2198096

 

 

 

Sample Results

Representative sample results are shown below. Figure 4 shows streamlines and color-contour lines of Mach number. On the other hand, pressure (Cp) distributions around airplane are illustrated in Fig. 5.

 

 

Fig. 4 Streamlines and color-contour lines of Mach number.

 

 

Fig. 5 Pressure (Cp) distributions around airplane.

 

 

References

[1] Toshio Kobayashi: Numerical computation using Advance/FrontFlow/red, Advance Soft

[2] http://www.rss21.iis.u-tokyo.ac.jp/index.html

[3] http://aaac.larc.nasa.gov/tsab/cfdlarc/aiaa-dpw/Workshop2/workshop2.html

 

 

All Rights Reserved, Copyright(C) 2008, Fluid Mechanics Laboratory,

Department of Aeronautics and Astronautics, Kyushu University.