Rotorcraft Noise Prediction with Multi-disciplinary Coupling - - PowerPoint PPT Presentation

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 Dynamics – Rotor Wake Modeling Methods – Computational Structural Dynamics – CFD/CSD Coupling Procedure – Acoustic Analysis

• Results for High speed impulsive noise prediction
• Results for Blade vortex interaction noise prediction
• Concluding remarks

Introduction and Background

Leishman, ‘Principles of Helicopter Aerodynamics’

• Complex interactional aerodynamics

Dynamic Loads and Structure Dynamics Engine/Drive Train Dynamics Acoustic Noise Aeroelastic Response Flight Controlling System

Rotor Source Noise

Lighthill’s Formulation

Ffowcs Williams-Hawkings Formulation

• Numerical Solution to FW

Numerical Solution to FW Numerical Solution to FW Numerical Solution to FW-

• H equation

H equation H equation H equation

• − =
• + "#

$# +

• $#$%

('#,%) )

• Three source terms:

– = +,- + + .- − ,- thickness source (monopole)

• Requires rotor blade geometry and kinematics

– "# = /#,%-% + +.# .- − ,- loading source (dipole)

• Requires rotor blade geometry, kinematics and surface loading

– '#,% = +.#.% + /#,% − + − +0 #,% quadrupole source

• Requires flow field around the rotor blade (volume integration)

Noise prediction method

• The noise standard became ever more stringent, and the

cost of flight testing and wind tunnel experiments was increasing

• Computational aero-acoustics gets more attention

– Direct Numerical Method – Hybrid Numerical Method

• High-speed impulsive (HSI) noise represents one of the

most intense and annoying forms of noise generated by helicopter rotors in high-speed forward flight.

• Blade Vortex Interaction (BVI) noise represents another

type of intensive noise generated by helicopter rotor in low-speed decent flight close to ground.

Approaches

High-fidelity Wake Modeling (Particle-VTM) Computational Structure Dynamics (CSD) High-order Computational Fluid Dynamics (CFD)

Shock Wave Boundary Layer

Noise Propagation (WopWop, RNM)

CAMRAD II, DYMORE Multi-body, Nonlinear FUN3D, OVERFLOW NASA RANS Flow Solver

Blade-Vortex Interaction

A systematic coupling approach among multiple disciplines to predict rotor noise

Computational Fluid Dynamics

TURNS (Transonic Unsteady Rotor Navier-Stokes by Prof. Baeder)

– Compressible unsteady Reynolds Averaging Navier-Stokes (RANS) solver – Inviscid terms are computed using 3rd MUSCL, and 5th order WENO scheme – Viscous terms are computed using 2nd central differencing – Second order time accuracy with Newton-type sub-iteration – Baldwin-Lomax and Spalart-Allmaras turbulence models – The low dispersion and dissipation Total Variation Diminishing (STVD) scheme developed by Helen Yee is implemented

Rotorcraft Wake Modeling Methods

• Free-wake module

– Wake geometry is decided by potential flow based method – CFD calculates the wake effects with the inputted wake geometry – Heavily depend on empirical input

• Overset grid methodology to capture wake directly

– Physics based, high resolution wake capturing method – Grid dependency – Numerical dissipation diffuses the tip vortex too rapidly

• Particle Vortex Transport Method (PVTM) (Dr. Phuriwat Anusonti-Inthra)

– Solves the incompressible vortex transport equation using a Lagrangian (vortex particle) approach – Fully coupled with CFD

• Uses CFD in near body to capture vortex generation
• Uses PVTM in other domains for calculating vortex

evolution

Computational Structural Dynamics

• CAMRAD II (Dr. Wayne Johnson)

– Comprehensive Finite Element Analysis with Multi-body Dynamics – Forced periodic solutions for steady and level flight to get trim solutions – The structural dynamics response is very important for the rotor blade simulations in forward flight conditions

CFD/CSD Coupling Procedure

TURNS / CFD Aero

CAMRAD II Lifting line aero with uniform inflow + “delta” force CAMRAD II trim solutions Forces and motions at quarter chord Blade Motions CFD Surface Aero- loading Aero-loading Difference ∆F/Mn+1 = (F/Mcfd – F/MCII) + ∆F/Mn Check convergence CAMRAD II Aero-loading

• A loose coupling

strategy based on a trimmed periodic rotor solution

• The comprehensive

airloads are replaced with CFD airloads with a ‘delta’ force method

• Use the lifting line

aerodynamics to trim and CSD to account for blade deformation in the comprehensive analysis

Acoustic Analysis

• PSU-WOPWOP (Prof. Brentner)

– Solves Farassat’s retarded-time formulation 1A of the Ffowcs Williams-Hawkings (FW-H) equation – Propagate and compute the tone noises at any given observer locations – Impermeable surface method

• Noise source based on the blade loadings or surface pressure

predicted by CFD or comprehensive analysis

• Thickness noise and loading noise

– Permeable surface method

• Noise source based on the flow field solutions provided by CFD on

a specified surface around the rotor blade

• Thickness noise, loading noise, high-speed impulsive noise

High-order STVD scheme for Permeable Surface method

Permeable Surface surrounding the blade

• r

Acoustic Data Surface Physical Blade Surface

High-order CFD scheme provides sufficient spatial and temporal accuracy to propagate acoustic characteristic waves to the acoustic data surface

• Implemented the low

dispersion and low dissipation Symmetric Total Variation Diminishing (STVD) scheme introduced by Helen Yee into our current CFD solver , which replaces the

• riginal 2nd order ROE

scheme inside the code

High Speed Impulsive Noise Prediction with CFD+overset grid method

• The DNW acoustic wind tunnel test is conducted in the

large low-speed facility (LLF) at the Duits Nederlands Windtunnel (DNW) for a scaled model of the UH-60a helicopter main rotor

• High speed forward flight case with advance ratio 0.3010

and the rotor tip Mach number of 0.8737

• CFD + overset background grid method coupled with

CSD, with 10 million total grid points.

Acoustic Experiments Set-up

• Mic. 1

3R 3R 30o

• Mic. 1
• Mic. 7

Wind

Top View

• Mic. 7

3R R Side View

The acoustic predictions and measured sound pressure at microphone 1 and microphone 7 are compared

Acoustic Experimental Measurements (Microphone 1)

• 100
• 50

50 100 0.25 0.5 0.75 1

Normalized Time Sound Pressure (Pa)

EXP-Microphone 1

HSI Noise

Noise due to vortex

Acoustic Experimental Measurements (Microphone 7)

• 100
• 50

50 100 0.25 0.5 0.75 1

Normalized Time Sound Pressure (Pa)

EXP-Microphone 7

HSI Noise Noise due to vortex

Impermeable Surface Method (Microphone 1)

• 70
• 60
• 50
• 40
• 30
• 20
• 10

10 20 30 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

No-dimension Time Sound Pressure-Total

Impermeable Surface Method EXP - Microphone 1

Impermeable Surface Method (Microphone 7)

• 100
• 80
• 60
• 40
• 20

20 40 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

No-dimension Time Sound Pressure-Total

Impermeable Surface Method EXP - Microphone 7

Permeable Surface Method: Acoustic Data Surface

Red – Blade; Blue – Baseline Grid; Orange – New Grid

Acoustic Predictions for Different Acoustic Data Surfaces

• 70
• 60
• 50
• 40
• 30
• 20
• 10

10 20 30 0.01 0.02 0.03 0.04

Time Sound Pressure-Total

Baseline Grid Grid No. 1 Grid No. 2 Grid No. 3 EXP

Permeable Surface Method (Microphone 1)

• 70
• 60
• 50
• 40
• 30
• 20
• 10

10 20 30 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

No-dimension Time Sound Pressure-Total

Impermeable Surface Method Permeable Surface Method EXP - Microphone 1

Permeable Surface Method (Microphone 7)

• 100
• 80
• 60
• 40
• 20

20 40 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

No-dimension Time Sound Pressure-Total

Impermeable Surface Method Permeable Surface Method EXP - Microphone 7

Blade Vortex Interaction Noise Prediction with CFD+PVTM method

Particle Vortex Transport Method: Coupled with CFD, where vortex source term come from RAN solution

• First-principle method
• Conserve vorticity
• Gridless wake (No grid adaptation)

Paper: Anusonti-Inthra, P. “Validations of Coupled CSD/CFD and Particle Vortex Transport Method for Rotorcraft Applications: Hover, Transition, and High Speed Flights” Proc. 66th AHS Forum, Phoenix, AZ, 2010

L

S v u dt d + ∇ + ∇ ⋅ = ω ω ω

2

HART II – Baseline PVTM Set up

• CAMRAD II

– 5 beam elements

• CFD grid

– 1.5M cells × 4 blades – Boundaries

• 0.5c: downstream
• 1.0c: other directions
• PVTM resolution

– Fine: 0.5c (5R×1.25R) – Medium: 1c(10R ×3.75R) – Coarse: 2c (15R ×6.25R)

(a) HART II CFD grid

• Higher Harmonic Control

Aeroacoustic Rotor Test (HARTII) performed in October 2001 in the Large Low-Speed Facility (LLF) of the DNW Wind Tunnel

• Flight Conditions:

– Low speed decent flight – Advance ratio, µ = 0.15

HART II Baseline Loading Predictions

r/R = 0.8700

0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 90 180 270 360

Azimuth M2CN

Exp 0 Iteration 1st Iteration 2nd Iteration 3rd Iteration 4th Iteration 5th Iteration 6th Iteration

CFD with Overset Grid 4.82 million near-body grid points 5.69 million back-ground grid points PVTM

Tunnel Wind

4 m

4 m

2.7 m 2.7 m

Rotor Disk Plane (Radius = 2m) Microphone Plane 2.215 m below the rotor disk

HART II Baseline Experimental Noise Map

HART II Predicted BVI Noise Contour

CFD + overset grid Method CFD + PVTM Method

Concluding Remarks

• A multi-disciplinary analysis method with coupled

algorithms among CFD/Wake/CSD/Acoustics to predict the rotor source noise

• The rotor blade loadings and acoustic signatures

predictions are compared with DNW high-speed forward flight case, and HART II low-speed decent case

• For High-speed forward flight case with transonic

impulsive noise, the acoustic signature predictions are more dependent on the acoustic and CFD coupling procedure, where the permeable surface method is giving much better results than the impermeable surface method

Concluding Remarks (cond.)

• For Low-speed decent flight case with blade vortex

interaction noise, the predictions are more dependent on the rotor tip wake modeling/capturing, where the PVTM method gives better results than the CFD overset grid method with relative coarse grid.

• Overall, for this kind of hybrid CAA method for rotorcraft,

the accuracy of the noise prediction is greatly dependent

• n the accuracy of the blade loading and rotor wake

simulation results from CFD. Thus, any improvement of the accuracy of CFD simulations for rotorcraft complex flow phenomena will also great improve the noise prediction