Investigation of Airfoil Aero Acoustics due to Different Stall - - PowerPoint PPT Presentation

Slide 1

Investigation of Airfoil Aero Acoustics due to Different Stall Mechanism using Large Eddy Simulation (LES)

Voo Keng Soon Vince Lim Nee Sheng Winson Tan Chun Hern DSO National Laboratories 05 Nov 2012

Slide 2

Brief Review of Available Experimental Data

Slide 3

Literature Containing Experimental Data

Brooks, T., Pope, D. and Marcolini, M., “Airfoil Self-Noise and Prediction”, NASA Reference Publication 1218, 1989

• Identification of 5 airfoil self-noise mechanisms
• Contains a series of acoustic tests of 2D and 3D NACA0012 airfoils
• 6 2D blade sections of chord length from 2.54cm to 30.48cm (1in to 12in)
• 5 3D blade sections of chord length from 5.08cm to 30.48cm (2in to 12in)
• wind tunnel speeds up to Mach 0.21 (Re based on chord up to 1.3x106)

Slide 4

NACA0012 Blade Sections[1]

2D models span: 45.72cm chord: varies trailing edge: sharp, < 0.05mm thick 3D models span: 30.48cm chord: varies trailing edge: sharp, < 0.05mm thick

[1] Brooks, T., Pope, D. and Marcolini, M., “Airfoil Self-Noise and Prediction”, NASA Reference Publication 1218, 1989

Slide 5

Acoustic Test Setup[1]

Free jet nozzle M1 M2 M5 M4 M8 M7

1.22m 30° [1] Brooks, T., Pope, D. and Marcolini, M., “Airfoil Self-Noise and Prediction”, NASA Reference Publication 1218, 1989

Slide 6

Brief Review of Different Airfoil Self-Noise Mechanisms

Slide 7

Airfoil Self-Noise Mechanisms[1][2]

[1] Brooks, T., Pope, D. and Marcolini, M., “Airfoil Self-Noise and Prediction”, NASA Reference Publication 1218, 1989 [2] Brooks,T. And Schlinker, R., “Progress in Rotor Broadband Noise Research”, Vertica, vol. 7, no. 4, 1983, pp. 287-307

Slide 8

Experimental Data[1]

[1] Brooks, T., Pope, D. and Marcolini, M., “Airfoil Self-Noise and Prediction”, NASA Reference Publication 1218, 1989

50 55 60 65 70 75 80 85 90 500 10000

Sound Pressure Level (dB) Frequency (Hz)

Self-noise Spectra for NACA0012

5.08cm chord, aoa 08.4° 5.08cm chord, aoa 15.4°

Slide 9

CFD Simulation of NACA0012 at AOA of 8.4° to show Turbulent Boundary Layer Trailing Edge Noise and Boundary Layer Separation Noise

Slide 10

Meshing and Testing Conditions

• Trimmer mesh of ~11 million cells
• cells aligned to air flow
• RANS, LES
• Mach 0.208 (71.3m/s)
• 0.0508m chord, 0.1286m span
• Reynolds Number studied: 230,000
• Periodic boundary conditions
• Time-Step: 1.0e-5s
• Simulated time: 0.1s
• Aeroacoustics module: Ffowcs Williams - Hawkings

Slide 11

Mid-Span Velocity Distribution (LES) and Streamwise Reynolds Stress

0.5 chord after TE 2.0 chord after TE 3.0 chord after TE

• 1
• 0.5

0.5 1 0.000 0.005 0.010

Z / chord <U'U'> normalized by ρV2

• 1
• 0.5

0.5 1 0.000 0.005 0.010

Z / chord <U'U'> normalized by ρV2

• 1
• 0.5

0.5 1 0.000 0.005 0.0

Z / chord <U'U'> normalized by ρV2

generates turbulent boundary layer trailing edge noise and boundary layer separation noise

Slide 12

Acoustic Pressure

Local pressure deviation from the ambient atmospheric pressure caused by sound wave Plot shows characteristics of a dipole

Slide 13

Sound Pressure Level vs. 3rd Octave Bands

Simulation Peak SPL: 65.65 dB at Center Frequency: 3171 Hz Experiment Peak SPL: 65.61 dB at Center Frequency: 3171 Hz

40 50 60 70 500 10000

Sound Pressure Level (dB) 3rd Octave Frequency Bands (Hz)

Sound Pressure Level vs. 3rd Octave Bands

Experimental Data Observer Point (Corrected for Span)

Reasonably close fit between experimental data and simulation

Slide 14

Far-Field Sound Directivity Pattern

10 20 30 40 50 60 90 75 60 45 30 15 345 330 315 300 285 270 255 240 225 210 195 180 165 150 135 120 105

Microphones are placed at a distance of 70x chord around airfoil SPL vs. Receiver Positions

Slide 15

Comparison of Coefficients

At 8.4°AOA Drag Coefficient, CD Lift Coefficient, CL Published Data[3] for NACA0012 at Re 0.36x106 0.01748 0.81798 RANS 0.02481 0.83163 LES 0.02047 0.93481

Reference values for simulations:

• reference area = 6.5068E-3 m²
• reference velocity = 71.3 m/s
• reference density = 1.17683 kg/m³

[3] Robert E. Sheldahl, Paul C. Klimes, “Aerodynamic Characteristics of Seven Symmetrical Airfoil Sections Through 180-Degree Angle of Attack for Use in Aerodynamic Analysis of Vertical Axis Wind Turbines”, Sandia National Laboratories, SAND80-2114, 1981

LES reported slightly higher CL. Mesh is coarse for LES. To consider modelling eddies in turbulent boundary layer (TBL) using RANS and the outer-flow with LES => Detached Eddy Simulation (DES)

Slide 16

CFD Simulation of NACA0012 at AOA of 15.4° to show Large Scale Separation (Deep Stall) Noise

Slide 17

Meshing and Testing Conditions

• Trimmer mesh of ~11 million cells
• cells aligned to air flow
• URANS, LES
• Mach 0.208 (71.3m/s)
• 0.0508m chord, 0.1286m span
• Reynolds Number studied: 230,000
• Periodic boundary conditions
• Time-Step: 2.0e-5s
• Simulated time: 0.2s
• Aeroacoustics module: Ffowcs Williams - Hawkings

Slide 18

Mid-Span Velocity Distribution (URANS) and Streamwise Reynolds Stress

0.5 chord after TE 2.0 chord after TE 3.0 chord after TE

• 1
• 0.5

0.5 1 0.000 0.050 0.100

Z / chord <U'U'> normalized by ρV2

• 1
• 0.5

0.5 1 0.000 0.050 0.100

Z / chord <U'U'> normalized by ρV2

• 1
• 0.5

0.5 1 0.000 0.050

Z / chord <U'U'> normaliz

large scale separation stall generating noise

Slide 19

0.000 0.050

• 1
• 0.5

0.5 1

<U'U'> normaliz Z / chord

0.000 0.050 0.100

• 1
• 0.5

0.5 1

<U'U'> normalized by ρV2 Z / chord

0.000 0.050 0.100

• 1
• 0.5

0.5 1

<U'U'> normalized by ρV2 Z / chord

Mid-Span Velocity Distribution (LES) and Streamwise Reynolds Stress

0.5 chord after TE 2.0 chord after TE 3.0 chord after TE

large scale separation stall generating noise

Slide 20

Acoustic Pressure

Local pressure deviation from the ambient atmospheric pressure caused by sound wave Plot shows characteristics of a dipole

Slide 21

Sound Pressure Level vs. 3rd Octave Bands

Simulation Peak SPL: 84.6 dB at Center Frequency: 1010 Hz Experiment Peak SPL: 88.9 dB at Center Frequency: 1010 Hz

60 70 80 90 500 10000

Sound Pressure Level (dB) 3rd Octave Frequency Bands (Hz)

Sound Pressure Level vs. 3rd Octave Bands

Experimental Data Observer Point (Corrected for Span)

Similar trend between experimental data and simulation

Slide 22

10 20 30 40 50 60 70 90 75 60 45 30 15 345 330 315 300 285 270 255 240 225 210 195 180 165 150 135 120 105

Far-Field Sound Directivity Pattern

Microphones are placed at a distance of 70x chord around airfoil SPL vs. Receiver Positions

Slide 23

Comparison of Coefficients

At 15.4°AOA Drag Coefficient, CD Lift Coefficient, CL Published Data[3] for NACA0012 at Re 0.36x106 0.21114 0.64258 URANS 0.31031 0.92369 LES 0.17406 1.15898

Reference values for simulations:

• reference area = 6.5068E-3 m²
• reference velocity = 71.3 m/s
• reference density = 1.17683 kg/m³

[3] Robert E. Sheldahl, Paul C. Klimes, “Aerodynamic Characteristics of Seven Symmetrical Airfoil Sections Through 180-Degree Angle of Attack for Use in Aerodynamic Analysis of Vertical Axis Wind Turbines”, Sandia National Laboratories, SAND80-2114, 1981

LES reported higher CL. Mesh is coarse for LES. To consider DES.

Slide 24

Results and Conclusions

• For the case involving 8.4° AOA,
• a reasonably close fit between the experimental data and simulation
• similar peak SPL and center frequency
• LES reported a slightly higher CL.
• For the case involving 15.4° AOA,
• comparable trend between the experimental data and simulation albeit loose fit
• a difference of ~5dB in the peak SPL, similar center frequency
• LES reported a much higher CL.
• Mesh for a fully resolved LES has to be sufficiently fine to resolve the small eddies
• Considerations for future work:
• Refining the existing mesh
• Switching to Detached Eddy Simulation for additional comparison