Prediction of Helicopter Blade- Vortex Interaction Noise using Motion Data from Experiment

Size: px

Start display at page:

Download "Prediction of Helicopter Blade- Vortex Interaction Noise using Motion Data from Experiment"

Bethany Gibson
5 years ago
Views:

1 Prediction of Helicopter Blade- Vortex Interaction Noise using Motion Data from Experiment Yoshinobu Inada, Choongmo Yang, and Takashi Aoyama Computational Science Research Group (CSRG) Institute of Aerospace Technology (IAT) Japan Aerospace Exploration Agency (JAXA) Computational Science Research Group 27/12/21

2 Outline 1. Introduction Background and Objectives 2. Methods and Results Single Blade Calculation Full CFD Calculation (Overlapped grid method) Acoustic Analysis 3. Conclusion

Helicopter BVI Noise Background Sound Pressure (Pa) 5 #2 #1 #3 6 12 18 ψ (deg) Large impulsive

3 Helicopter BVI Noise Background Sound Pressure (Pa) 5 #2 #1 # ψ (deg) Large impulsive noise caused by the blade and vortex interaction, inhibiting a flexible operation of helicopter

capturing/coupling + Acoustics Rotor Blade Euler (direct vortex capturing) + Acoustics Grid Type Grid points

4 Background CFD analysis for helicopter aerodynamics Single Grid ModelM MENTOR Multidisciplinary Euler/Navier-Stokes Tool for Rotorcrafts Full CFD Model (Overlapped( Grid Model) Euler/NS code + Wake capturing/coupling + Acoustics Rotor Blade Euler (direct vortex capturing) + Acoustics Grid Type Grid points Computing time Accuracy Purpose Single grid.5~5 millions 1~1 hr/case moderate Design/Industry Overlapped grid 2~6 millions 1~5 hr/case excellent Academic

5 Background Acoustic analysis for noise prediction Using CFD Results Direct Calculation Kirchhoff Method FW-H Method Computational Aeroacoustics (CAA) HSI Noise BVI Noise Blade Shock Wave HSI Noise Subsonic Region Supersonic Region

6 Background Questions to be solved for correct prediction of BVI noise flap Q2 twist lead-lag Aero-elastic elastic analysis (CFD( CFD/CSD) Q1 Q3 Q1. Can accurate blade motion be simulated by using CFD/CSD coupling method? Q2. If the blade motion is correctly simulated, can CFD capture the tip vortex trace and simulate the blade vortex interaction (BVI) accurately? Q3. Can acoustic code predict BVI noise correctly from obtained pressure data?

7 Objectives BVI noise prediction using blade motion data from experiment Q2 Q3 As the middle step for full coupling aero-elastic analysis, primary focus on the second and third questions using elastic blade motion data captured from the experiment (HART II experiment).

Vibration by NASA Langley, US Army, DLR,

elastic deformation Pressure distribution

8 HART II Experiment International co-operative operative project on higher harmonic blade motion control BL : Baseline MN: Minimum Noise MV: Minimum Vibration by NASA Langley, US Army, DLR, DNW, ONERA Blade air-loading Blade elastic deformation Pressure distribution on blades Noise Wake etc. Available for research use Noise contour PIV

HART II Calculation Flap, Lead-lag (R).6.4.2 -.2 HARTII Exp.

9 HART II Calculation Flap, Lead-lag (R) HARTII Exp. Data flap torsion lead-lag Torsion (deg) Blade motion with elastic deformation ψ (deg) Full CFD Aerodynamic/ Acoustic Analysis Single blade 3D CFD + Beddoes model

10 Single Grid Model Blade grid (chord normal span) blade ( ) 1 = 919,8 Rotor Blade Numerical Scheme for rotor (Curvilinear Grid): Governing Equation: 3D unsteady Euler equation Space: Beam-Warming scheme + TVD scheme - 2nd Order Accuracy - MUSCL approach using minmod limiter Time: Euler Backward Implicit Time Integration Newton iterative method in unsteady calculation Solution algorithm: Upwind Line Gauss-Seidal Relaxation Method

11 Wake Model Flow Beddoes Generalized Wake Model Swirl velocity 2 Γ r vθ () r = 2 π r r 2 2 ( r ) a + Translation by blade elastic deformation Y Position x = cosψ v r v v + μ Δψ x v + X e ψ b ψ v y = v r v sinψ v ψ v ( R ) dψ zv = μzδψ v + / Ω ψ b + Y e + Z e Z (x v, y v, z v ) X downwash (v)( downwash inside the disk 8E 3 v = v μ x y Ex E y 15π outside the disk 8E 3 v = 2v + 1 2μ x y E y 15π (van der Wall, 2)

12 Acoustic Analysis P CFD calculation blade observer pressure distribution Ffowcs Williams & Hawkings (FW-H) equation p 1 ρvn 1 pb cosθ pb cosθ x, t) = ( ) 4 dσ + dσ + dσ π t rλ c t rλ r Λ ( 2 Sound pressure Thickness noise Load noise (far) Load noise (near) Λ = 2 1/ 2 ( 1+ M n 2M n cosθ ) Σ : Influential surface

13 Condition Twist angle (/R) Blade motion parameter Rotor radius, R (m) Blade chord length, c (m) Blade motion data Collective pitch angle, θ 3.2 Lateral cyclic pitch angle, θ 1C Longitudinal cyclic pitch angle, θ 1S 2. HHC lateral cyclic pitch angle, θ 3C HHC long. cyclic pitch angle, θ 3S Precone angle, b 2.5 Thrust Coefficient, C T.44 Tip Mach number, M tip.6387 Inflow ratio, μ.15 Angle of tip path plane, α TPP 4.5 or 5.3 BL -1.1

(xv, yv, zv) X Z HART II Experiment Cal. with blade elasticity μ Y Cal.

14 Results (Single Grid Cal.) Leading edge differential pressure distribution 3% chord ψb ψ ψv Y (xv, yv, zv) X Z HART II Experiment Cal. with blade elasticity μ Y Cal. without blade elasticity 1.5 μ P Blade vortex interactions (BVI) are occurring X

15 Results (Single Grid Cal.) Blade air-loading and acoustic results Blade air-loading.2 without elasticity.2 with elasticity CnM2.1.5 CnM2.1.5 experiment no elasticity experiment with elasticity Acoustics Sound Pressure Level (Pa) PSI experiment no-elasticity Sound Pressure Level (Pa) PSI experiment with-elasticity PSI (deg) PSI (deg)

Full CFD Model Full CFD Model (Overlapped grid system) Specifications of grid system Grid type rotor only (X Y Z) = number of grid points with fuselage coarse medium fine medium Inner background grid

16 Full CFD Model Full CFD Model (Overlapped grid system) Specifications of grid system Grid type rotor only (X Y Z) = number of grid points with fuselage coarse medium fine medium Inner background grid = 3,335, = 14,4, = 6,9, = 38,4, Outer background grid = 321, = 56, Blade grid (chord normal span) blade ( ) 4 = 1,899,5 Fuselage grid = 123,753 Total ~5,56, points ~16,6, points ~61,4, points ~41,, points Inner background spacing.17c (=.15R).99c (=.6R).66c (=.4R).99c (=.6R)

17 Full CFD Model Numerical Scheme for rotor and fuselage (Curvilinear Grid): Governing Equation: 3D unsteady Euler equation Space: Beam-Warming scheme + TVD scheme - 2nd Order Accuracy - MUSCL approach using minmod limiter Time: Euler Backward Implicit Time Integration Newton iterative method in unsteady calculation Solution algorithm: Upwind Line Gauss-Seidal Relaxation Method Numerical Scheme for background grid (Cartesian Grid): Governing Equation: 3D unsteady Euler equation Space: Compact TVD scheme (4th Order Accuracy) Simple High-resolution Upwind Scheme (SHUS) - Advection Upstream Splitting Method (AUSM) Time: Explicit Time integration Four stage Runge-Kutta method (4th Order)

18 Results (Full CFD Cal.) Effect of blade elasticity consideration Blade air-loading Measured and calculated coefficient of blade loading at 87% spanwise position with/without elastic deformation for BL case

coarse 29 23 5 = 3,335,.17c (=.15R) Results (Full CFD Cal.) Effect of grid size medium 45 4 8 = 14,4, Effect of Grid Dependency.99c (=.

19 coarse = 3,335,.17c (=.15R) Results (Full CFD Cal.) Effect of grid size medium = 14,4, Effect of Grid Dependency.99c (=.6R) fine = 6,9,.66c (=.4R) Measured and calculated coefficient of blade loading at 87% spanwise position with/without elastic deformation for BL case

20 Results (Full CFD Cal.) Effect of fuselage Measured and calculated coefficient of blade loading at 87% spanwise position with/without elastic deformation for BL case

21 Results (Full CFD Cal.) Acoustic results blade observer Measure and calculated sound pressure for BL case using high pass filter at 4/rev.

22 Conclusions For both single grid and full CFD calculation Consideration of blade elastic deformation remarkably enhanced the calculation quality, thus indicating the necessity of CFD/CSD coupling in a future prediction system. For single grid calculation Beddoes generalized wake model successfully simulated the tip vortex trace in the wake, realizing the reasonable quality of calculation, but overestimation of air-loading and the BVI noise still remain. For full CFD calculation The calculation quality is dependent on the inner-background grid size, proper grid size should be used to enhance the precision. Fairing effect also should be considered for high quality of calculation.

23 Thank you for attention

Rotorcraft Noise Prediction with Multi-disciplinary Coupling Methods. Yi Liu NIA CFD Seminar, April 10, 2012

Rotorcraft Noise Prediction with Multi-disciplinary Coupling Methods Yi Liu NIA CFD Seminar, April 10, 2012 Outline Introduction and Background Multi-disciplinary Analysis Approaches Computational Fluid