Flow-Induced Noise Technical Group

Transcription

1 Flow-Induced Noise Technical Group Center for Acoustics and Vibration Spring Workshop April 30, 2014 Presented by: Dean E. Capone, Group Leader 1

2 Overview The mission of the Flow-Induced Noise Group of the Center for Acoustics and Vibration is the understanding and control of acoustic noise and structural vibration induced by fluid flow. Topical Research Area Presentations Dr. Ken Brentner: PSU-WOPWOP noise prediction code Dr. Robert Campbell: Fluid Structure Interaction (FSI) Research, including A Wind Turbine Project 2

3 Ongoing Projects Ongoing projects Project Topic: Fluid-Structure Interaction (FSI) of a Flexible Strut with Strong Turbulent Upstream Vortices Student: Abe H. Lee (Ph.D in Acoustics) Advisors: Dr. S.A. Hambric, Dr. R.L. Cambpell Project Topic: Modelling and Measurement of Turbulent Boundary Layer Unsteady Shear Stress in Elastomer Layers Student: Cory Smith (Ph.D in Acoustics) Advisors: Dr. D. E. Capone and T. A. Brungart Project name: High Cycle Fatigue Simulations and Measurements Sponsor: Pratt & Whitney PI(s): Philip Morris Students/degree levels: Michael Lurie, PhD Project name: Adjoint Design for Low Noise Sponsor: Office of Naval Research PI(s): Philip Morris Students/degree levels: Nidhi Sikarwar, PhD 3

4 Ongoing Projects Ongoing projects Project name: Simulation of Jet Noise Reduction Devices Sponsor: Office of Naval Research PI(s): Philip Morris Students/degree levels: Matthew Kapusta, MS Project name: Rotorcraft Broadband Noise Predictions Sponsor: Bell Helicopter Textron Inc. PI(s): Kenneth Brentner, Philip Morris Students/degree levels: Abhishek Jain, MS Project name: Nonlinear Sound propagation from Distributed Sources Sponsor: Unsponsored PI(s): Philip Morris Students/degree levels: Donald Hyatt, MS 4

5 PENNSTATE An Overview of PSU-WOPWOP: A General Purpose Ffowcs Williams Hawkings Solver Kenneth S. Brentner Department of Aerospace Engineering CAV Workshop April 30, 2014 Kenneth S. Brentner, Dept. of Aerospace Engineering 1

6 Historical Background Ffowcs Williams Hawkings Equation J. E. Ffowcs Williams and D. L. Hawkings (1969) Philosophical Transactions of the Royal Society of London, Series A, 264, Sound generation by turbulence and surfaces in arbitrary motion. Laid the framework for treating sound field of a surface moving at high speed Surfaces are replaced by discontinuities in the flow-field Generalized conservation equations valid everywhere in space The FW-H equation is the most general form of Lighthill's acoustic analogy PENNSTATE Prof. J.E. Ffowcs Williams Kenneth S. Brentner, Dept. of Aerospace Engineering 2 The Impact of Farassat s Formulations Nearly 40 Years On

7 Key Concepts: Ffowcs Williams Hawkings Equation PENNSTATE Rearrangement of Navier-Stokes equations into an inhomogeneous wave equation 2 p ( x, t) t x 2 x x Q ( f ) F ( f ) T H( f ) i i i j ij f = 0 describes the integration surface M > 1 Thickness displacement of fluid generates sound Loading accelerating force distribution generates sound (includes BVI noise) Quadrupole All volume sources, non-linear effects nonuniform sound speed Kenneth S. Brentner, Dept. of Aerospace Engineering 3

8 PENNSTATE Rotor Noise Theory Aeroacoustics is governed by the conservation equations of fluid dynamics Navier-Stokes equations wave equation Three main computational approaches: acoustic analogy treats real flow effects by fictitious sources; exact in principle Ffowcs Williams Hawkings equation (1969) appropriate when solid bodies are present (Lighthill analogy) Kirchhoff approach based upon wave equation actual sources replaced by their influence on a surface direct computation (CFD and CAA) high spatial and temporal accuracy needed Kenneth S. Brentner, Dept. of Aerospace Engineering 4

9 PENNSTATE Comparison with Other Approaches Acoustic analogy: FW-H equation (surface & volume integrals required) 2 2 p ( x, t ) Q ( f ) Fi ( f ) T t x i x i x j practical approximation (only surface integrals) 2 p ( x, t) t H ( f ) x Q ( f ) F ( f ) Kirchhoff: physical sources represented by mathematical sources on surface (only surface integrals) 2 p (,) x t Q ( f ) F i ( f ) t x i i i Common use inappropriate for most rotor noise prediction Direct computation: limited to relatively near field (full domain discretized) provides input to other approaches ij Kenneth S. Brentner, Dept. of Aerospace Engineering 5

10 PENNSTATE PSU-WOPWOP Numerical solution to Farassat s Formulation 1A of the FW-H equation Requires geometric and loading or flow data as input Arbitrary source and observer motion can be specified including deformable surfaces Many sources can be specified at one time (e.g., multiple rotors) Observer parallel Accepts Variety of Input Data Types Compact patch (Line) Regular surfaces Permeable acoustic data surfaces (ADS) Used in varied applications Rotorcraft noise Discrete frequency and broadband Maneuver Civil noise certification Near real-time Permeable Surface applications Wind turbine noise Open rotor noise (commercial aircraft) Jet noise Landing gear and airframe noise Acoustic scattering PSU-WOPWOP can compute the acoustic pressure gradient Kenneth S. Brentner, Dept. of Aerospace Engineering 6 The Impact of Farassat s Formulations Nearly 40 Years On

11 Rotor Noise Prediction during a Complex Maneuver PENNSTATE Right Turn (a) Height Above Ground (m) c 1000 Turn Bank 22 sec North-SouthPosition(m) Level Flight 57 sec a Level Flight 14 sec b 0 Climb1sec Flight Ends 80 sec East - West Position (m) Very complicated source motion Aircraft motion is complex Each blade is moving in an independent, aperiodic trajectory Significant blade-vortex interaction during aggressive right turn - Brès, G.A., Brentner, K.S., Perez, G., and Jones, H.E.: Maneuvering Rotorcraft Noise Prediction. Journal of Sound and Vibration, Vol. 275, No.3-5, August 2003, pp Chen, H.N., Brentner, K.S., Lopes, L.V., and Horn, J.F.: An Initial Analysis of Transient Noise in Rotorcraft Maneuvering Flight. International Journal of Aeroacoustics, Vol. 5, No. 2, April 2006, pp Chen, H.N., Brentner, K.S., Lopes, L.V., and Horn, J.F.: An Initial Analysis of Transient Noise in Rotorcraft Maneuvering Flight. International Journal of Aeroacoustics, Vol. 5, No. 2, April 2006, pp Kenneth S. Brentner, Dept. of Aerospace Engineering 7

12 Supersonic Jet Noise and a Synthetic Phased Array Highly detailed CFD prediction provided input data for PSU-WOPWOP (ITAC and PSU) PSU-WOPWOP predicted the noise on two virtual arrays Acoustic array data used with advanced phased array processing (OptiNav) PENNSTATE Nelson, C.C., Cain, A.B., Dougherty, R., Brentner, K.S., Morris, P.J., "Application of Synthetic Phased Array Techniques to Hot Supersonic Jet Noise," AIAA Kenneth S. Brentner, Dept. of Aerospace Engineering 8

13 PENNSTATE Wind Turbine Noise Prediction Unsteady Navier-Stokes computation (OVERFLOW) of wind turbine flow field PSU-WOPWOP noise prediction Nelson, C., Cain, A., Raman, G., Brentner, K. S., and Morris, P. J., "Studies of a Wind Turbine Using OVERFLOW and PSU-WOPWOP," AIAA , Kenneth S. Brentner, Dept. of Aerospace Engineering 9

14 Acoustic Scattering for a Notional Quad Tiltrotor PSU-WOPWOP computes acoustic pressure gradient NASA FAST scattering code predicts scattered field Sound Pressure Level 10 m below the front rotor plane (6 times rotor BPF : Hz, L=-0.5) Incident field PENNSTATE Total field Total field Kenneth S. Brentner, Dept. of Aerospace Engineering 10

15 Questions? PENNSTATE Kenneth S. Brentner, Dept. of Aerospace Engineering 11

16 ARL Fluid-Structure Interaction (FSI) Efforts at the Applied Research Laboratory Presented by: Robert Campbell Noise Control and Hydroacoustics Division Applied Research Laboratory The Pennsylvania State University Presented at: CAV Workshop 30 April 2014

17 ARL Fluid-Structure Interaction Fluid Force on Structure Fluid Motion Structure Motion No Fluid Slip/ Flux at Surface 2

18 ARL Outline FSI Modeling Overview ARL Partitioned FSI Solver Overset Grid Technology and FSI Simulations Application Examples Unsteady Hydrofoil (in Cylinder Wake) 3D Flag Benchmark Wind Turbine Student Projects 3

19 ARL FSI Modeling Overview: Monolithic vs. Partitioned Fluid structure interaction (FSI): Fully-coupled motion of a deformable solid and a surrounding and/or contained fluid Two approaches to FSI modeling Monolithic Governing equations for both the fluid and solid cast in terms of the same primitive variables (velocity and pressure) Single discretization scheme applied to entire domain Fluid Motion Fluid Force on Structure No Fluid Slip/ Flux at Surface Solve Continuum Mechanics Equations (Flow, Acoustics, and Structure) Structure Motion t n t n+1 Partitioned Fluid and solid domains modeled separately Separate discretizations of each domain Stress and displacement communication across the domain interface Solve Flow t n t n+1 Solve Structure 4

20 ARL ARL Partitioned Solver Overview FSI solver based on a partitioned approach with body-fitted meshes: Independent flow and structural solvers that communicate at the fluid/solid interface OpenFOAM for the flow solver Custom FEANL structure class for the structural solver fsiinterface class to interface the solvers OpenFOAM fsiinterface FEANL Fluid Mesh Motion 5

21 ARL Partitioned FSI Solver Subiterations each time step to ensure fluid and structure interaction is converged Under-relaxed structural displacements to improve convergence Variable under-relaxation coefficient (w) determined using Aitken s method START OpenFOAM fsiinterface Fluid Mesh Motion FEANL Move Fluid Mesh Solve Fluid No Solve Solid t>t end Yes STOP i++ No Yes t=t+dt Fixed-Point Iteration 6

22 ARL Overset Mesh Methods An overset mesh assembly is comprised of a set of overlapping grids Interpolation is performed across the overlapping boundaries in order to create a continuous domain Overset grids enable the meshing of complex geometries Component grids can be built independently Components can be easily modified, added, or removed 7

23 ARL Overset Mesh Advantages: Translating Body Continuous Re-Meshing 8

24 ARL Overset Mesh Advantages: Translating Body Continuous Re-Meshing Morphing Mesh 9

25 ARL Overset Mesh Advantages: Translating Body Continuous Re-Meshing Morphing Mesh Overset Mesh 10

26 ARL Overset Mesh Motion Traditional Mesh Motion: Structural Displacements Apply as BC to Full Mesh Displacement Linear Solver for Mesh Displacement Move Mesh Overset Mesh Motion: Structural Displacements Apply as BC to Full Mesh Displacement Linear Solver for Mesh Displacement Move Mesh UpdateDCI (Suggar++) Subset Overset Mesh Motion: Structural Displacements Apply as BC only to Overset Interface Mesh Displacement Linear Solver only for Overset Interface Mesh Displacement Move only the Overset Mesh UpdateDCI (Suggar++) Significant savings in mesh motion calculations 11

27 AMPLITUDE (Ay/D) TE Tip Deflection, m ARL Partitioned Solver Validation FSI with Overset Meshes: Turek Hron Benchmark 6 x TE Tip Deflection Test 47, BaselineAOA Test 48, BaselineAOA Test 44, -0.25º Simulation, AOA 0 Simulation, AOA Time, s Campbell, R. L. and Paterson, E. G. Fluid structure interaction analysis of flexible turbomachinery, Journal of Fluids and Structures, 27(8): , DES of Self-Excited Cylinder in Cross-Flow 2D Sim 3D Sim [a] DAMPING 2 m (2 ) (2 ) 2 S 2 D Lee, A. H., Campbell, R. L., and Hambric, S. A. Coupled delayeddetached-eddy simulation and structural vibration of a self-oscillating cylinder due to vortex-shedding, Journal of Fluids and Structures, DES FSI of Upstream Cylinder Forcing of a Hydrofoil 12

28 ARL Three-dimensional Flag Benchmark Images provided by Cooper Elsworth/Grant Dowell Three dimensional Turek experiment mounted in 12 Tunnel Goal is to publish 3D turbulent validation data for FSI simulations 13

29 ARL Three-dimensional Flag Benchmark Experiment (Dowell) Y X Tip displacement is measured from the center of the flag trailing edge in the y-direction Y = 0 is determined by the mean of the location values 14

30 ARL Three-dimensional Flag Benchmark Model (Elsworth) Objective Perform validation and verification of the overset FSI solver Simulate Grant s 3D experiment 15

31 DOE Program: Cyber Wind Facility *PLATFORM-WAVE HYDRODYNAMICS and 6-DOF MOTIONS (Hybrid URANS/LES +VOF) MESOSCALE, WEATHER (URANS/WRF) *ATMOSPHERIC BOUNDARY LAYER TURBULENT WINDS (4-D LES) *BLADE AERODYNAMICS, SPACE-TIME LOADINGS (Hybrid URANS/LES) WAKE TURBULENCE BLADE-WAKE-ATMOSPHERE (Actuator Vortex Body Embedding within LES) WAKE- TURBINE INTERACTIONS (wind plant) *BLADE AND *TOWER ELASTIC DEFORMATION (FEM, Modal model + FSI) *Shaft Torque, *Drivetrain Loadings CYBER WIND FACILITY highly resolved 4-D cyber data coupled atmospheric turbulence-blade loadings-shaft torque data coupled wave structure platform motion turbine loadings data experiment design, test-bed, turbine design, controls concepts and testing advanced correlations for ALM and other design tools using look-up tables *sensors, controllers, diagnostics 16

32 ARL DOE-Sponsored Cyber Wind Facility daytime large-eddy simulation multi-scale structured + unstructured grid design Moderately Convective Boundary Layer: Turbulence Structure Size similar to Rotor Diameter Turbulence inflow at the microscale Unsteady flow simulation over turbine blades 17

33 The Cyber Wind Facility Team Jim Brasseur (PI), Eric Paterson, Sven Schmitz, Rob Campbell, Sue Haupt Principal Investigators Balaji Jayaraman, Ph.D., DOE Research Associate, Mechanical Engrg Sven Rob Sue Pankaj Amir Mehdizadeh Javier Alex Brent Jim Eric Balaji Tarak Ganesh Adam Brent Craven, Ph.D., DOE ARL Research Associate Tarak Nandi, DOE PhD student, Mechanical Engineering Alex Dunbar, DOE PhD student, Mechanical Engineering Javier Motta-Mena, DOE PhD student, Mechanical Engineering Ganesh Vijayakumar, NSF PhD student, Mechanical Engineering Adam W. Lavely, NSF PhD student, Aerospace Engineering Pankaj K. Jha, DOE Graduate Student, Aerospace Engineering Amir Mehdizadeh, DFG Postdoctoral Fellow Industry Partner: GE GR 18

34 ARL Cyber Wind Facility: Single Rotating Blade in Atmosphere 19

35 Flap-wise Tip Deflection, m ARL Cyber Wind Facility Next Step: Introduce Blade Flexibility Initial Spacing of 5.2 m (on a 63 m blade!) Parked Rotor Simulation with Uniform 10 m/s Inflow Time, s 20

36 ARL Student Projects Abe Lee: Propeller Crashback Cooper Elsworth: Partitioned FSI Mesh Convergence Metrics Javier Motta-Mena: Wind Turbines with ABL Turbulence Kenneth Aycock: Blood Vessel and IVC Filter FSI Jason Sheldon: Flow-induced Loads by Unsteady Wake Erica Lieberknecht: FSI Monolithic Solver Development Grant Dowell: FSI 3D Flag Benchmark Experiment Nick LaBarbera: Towed Array Dynamics Jason Halwick: FSI for Cavitation Erosion Byron Gaskin: Cancer Cell Migration Michael McPhail: Human Voice Production 21

37 ARL External Perspective We are/have organized FSI mini-symposia on FSI: Fluid-Structure Interaction Algorithms and Applications, 11 th World Congress on Computational Mechanics, Barcelona, Spain, July 20-25, 2014 Fluid-Structure Interaction Algorithms and Applications, 12 th US National Congress on Computational Mechanics, Raleigh, NC, July 22-25, 2013 Fluid-Structure Interaction Algorithms and Applications, 10 th World Congress on Computational Mechanics, Sao Paulo, Brazil, July 8-13, 2012 Fluid-Structure Interaction Algorithms and Applications, 11 th US National Congress on Computational Mechanics, Minneapolis, MN, July 25-29,