CFD modelling of aerodynamic vortex generators on an airfoil with - - PowerPoint PPT Presentation

CFD modelling of aerodynamic vortex generators on an airfoil with OpenFOAM

Lorena Fernández, Rubén Gutiérrez & Beatriz Méndez

lfernandez@cener.com, rgutierrez@cener.com

Santiago de Compostela- 2nd OpenFOAM iberian 29th May 2018

2

Outline of the presentation

• 1. Introduction
• 2. Geometry
• 3. Mesh
• 4. CFD simulations
• 5. Conclusions & future work

Ø 1 7 1 . 2 m 196.2 m

3

• 1. INTRODUCTION

Low drag Vortex Generators

Vortex generators delay separation in the airfoil to higher angles of attack and increase the maximum lift. Vortex generators create longitudinal vortices that mix high momentum air form the outer flow down to the boundary layer near the surface making the flow more resistant to separation. Drawback: drag penalty

4

Vortex generators implementation on different blade regions

• 2. GEOMETRY

~ 5 mm Leading Edge Protection System Chord = 1m ~ 3 m m

A pseudo-2d approach is done. A segment of the blade is simulated with the following assumptions:

• Constant twist and chord.
• 2D flow without rotational effects.
• Fixed inflow velocity. (Fixed AoA).

Simulated region

• Boundary located at 25 chords.
• The Reynolds number is 6 millions
• cells of the mesh 10 millions

Case background Geometry

5

• 3. MESH

Efficient meshing: Mesh in 3 parts (merged and use of AMIs)

6

• 3. MESH

Cf-Mesh ICEM-CFD Salome-Meca BL limitation

Equal BL parameters (VG & Airfoil surfaces) Solution with more elements in BL region

Meshing tools:

7

Avoid extension of boundary layer thickness, causing large aspect ratio cells.

Propagation nodes problem O-grid Solution

• 3. MESH

8

• 4. CFD SIMULATION

Steady-state solver SimpleFOAM SIMPLEC Algorithm fvSolution Consistent true

Schemes Solver

Grad → cellMDLimited Gauss linear 1 Div → bounded Gauss linearUpwind Laplacian → Gauss Linear Limited 0.5 Interpolation → linear snGrad → limited 0.5 OpenFOAM version 5

9

Slip AMI Walls Far field Slip AMI

Boundary conditions

• 5. CFD SIMULATION

Name Condition Regions FalField freestream With the velocity indicated with respect to AoA Far field at 25m from the airfoil Slip slip (or empty without the VG, a 2D case) Sides of the domain AMI cyclicAMI Interfaces between 3 mesh pieces Walls wall Airfoil profile and the VGs

10

Clean situation: Transitional model K-KL Dirty situation (Fully-Turbulent): Fully-turbulent model KW-SST (Chosen model)

• 1. Upstream the LEPS 2. At LEPS 3.Downstream the LEPS

1 2 3

Turbulence model

• 4. CFD SIMULATION

11

K-Epsilon Spallart Allmaras K-Omega

Slice (Normal to x) Slice (Normal to x) Slice (Normal to x) Good flow but the “vortex” generated by the VG are not well captured Non realistic flow Good flow and “vortex” but a wrong flow is downstream

Turbulence model

• 4. CFD SIMULATION

12

U field from use potentialFoam K-Omega K-Epsilon as a good initialization turbulence model K-Epsilon (first) K-Omega

Initialization

• 4. CFD SIMULATION

13

• 5. CONCLUSIONS & FUTURE WORK
• Optimize the CFD study of a VG and an cross-section of an airfoil
• Study difgerent VG and airfoil shapes
• Meshing strategy optimized
• Initialization was solved
• CFD 3D for all the blade
• K-KL turbulence model without LEPS
• T

ransient solver

• Semi-empirical models (Wendt,Bay…) avoiding fully-mesh computations
• Simulate only the VG’s infmuence area with airfoil profjles’s boundary layer as an input

Conclusions Future work

¡THANK YOU!

Thanks for your atuentjon. Questjons?