Aerodynamic Simulations with AcuSolve

Innovation Intelligence®

Aerodynamic Simulations with AcuSolve

Dr. Marc Ratzel

April 2013

Copyright © 2012 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.

Agenda

1. Motivation

• Applications of aerodynamics

2. AcuSolve for external aerodynamics

• Quick overview of AcuSolve

3. Aerodynamic examples

• Automotive aerodynamics

• Fluid-Structure-Interaction for a 100m wind turbine

4. Summary


1. Motivation

• Aerospace

• No ground effects

• A380 900 km/h (Mach 0.75)

• Interest: lift/drag ratio, stall point, capacity,…

• Automotive

• Proximity to the ground

• My car 150 km/h (Mach 0.13)

• Interest: drag, lift, noise,…

• Buildings

• Several objects in close proximity

• Avg. wind speed 20 km/h (Mach 0.02)

• Interest: pressure loads, max. speed,…


Characteristics of external flows

• Numerical model

• Large models (+50Mio volume elements)

• Complex geometry (e.g. engine, underbody)

• Very fine mesh near walls (Boundary layer, BL)

• Physics

• Different scales time/length scales

(e.g. sounds travels much faster than airflow,

high/low frequency, small/large eddies)

• Turbulence

• Transient phenomena (e.g. airborne noise CAA, wake)

• Fluid-Structure-Interaction

(e.g. structural vibration of hood noise, eigenmodes)

• Discontinuities (e.g. shock wave aerospace)

• Moving boundaries (e.g. train, wheels of a car)

Boundary layer mesh

Turbulent wake, wind turbine

Shock wave


2. AcuSolve (CFD solver)

• General

• General purpose, 3-dimensional, unstructured solver

• Based on Finite Element method (GLS)

• Originated at Stanford University (T. Hughes et al.)

• Integration with other HyperWorks tools (HyperMesh, HyperView, RADIOSS,..)

• Numerics

• 2nd order accuracy in time and space for all flow variables

• Designed from day one for large scale problems

• Advanced turbulence models (SST, k-w, SA, DES, DDES, …)

• Comprehensive physics (fan/radiator component, sliding mesh, radiation,…)

• Robust fast volume mesher included (e.g. 90Mio car, engine, underbody, 80min)

• Very low requirement for element quality (e.g. tetras in boundary layer, skew>0.999)

Excellent fit for external aerodynamics


AcuSolve (Fluid-Structure-Interaction, FSI)

• Rigid body dynamics

• 6 DOF rigid body solver

• No structural displacement

• Practical FSI (P-FSI)

• No run-time coupling

• Structural displacement computed in AcuSolve based on eigenmodes

• Limited to linear structural displacement

• Directly coupled FSI (DC-FSI)

• Codes communicate during run-time

• Loads/displacement exchange

• Non-linear structural behaviour

• Supported for RADIOSS, Abaqus, MD-Nastran


3.1 Automotive aerodynamics

• Drag force

• Accounts for 75% of the car’s resistance (100km/h)

• Computation:

• Drag reduction by minimizing CD shape optimization

• Drag coefficients CD

drag side

lift

Car shape Frontal area

Modern car: ~ 0.29 Eiffel Tower: 1.8 – 2.0 Man (upright): 1.0 – 1.3


Asmo model (Daimler & Volvo)

Amed body (S.R. Ahmed)

C_p, underbody C_p, rear C_p, roof

Drag coeff.: 0.162 (exp) / 0.164 (AcuSolve)

Drag coeff.: 0.23 (exp) / 0.23 (AcuSolve)

Classical external aero benchmarks

AcuSolve

Exp. (Volvo, Daimler)


Fluid-Structure-interaction for rear wing

• P-FSI analysis

• Generic model of an automotive rear wing

• Soft plastic material, thickness of 2mm

• 20 & 100 eigenmodes computed with OptiStruct

fixed

monitor point

Displacement of monitor point

(20 & 100 eigenmodes) Down force for rigid and elastic

4%


Fluid-Structure-interaction for rear wing

• P-FSI analysis

• Generic model of an automotive rear wing

• Soft plastic material, thickness of 2mm

• 20 & 100 eigenmodes computed with OptiStruct

fixed

monitor point

Structural deformation (vel. contour) Structural deformation (vel. contour)


Altair’s Virtual Wind Tunnel (VWT)

• External automotive CFD analysis

• Advanced physics (rot. wheels, radiator, FSI,…)

• Efficient case setup


Altair’s Virtual Wind Tunnel (VWT)

• External automotive CFD analysis

• Advanced physics (rot. wheels, radiator, FSI,…)

• Efficient case setup

• AcuSolve as CFD solver

• Automatic post-processing


3.2 Wind turbine (Fluid-Structure-Interaction analysis)

• Facts

• Cooperation with Sandia National Laboratories, CA, USA

Blade length 100m

Weight 1.1t

Max. chord 7.6m

Material Fiberglass, Resin, Foam,…

Max. operation speed 7.44 RPM (tip speed 80m/s)

Power output ~ 13MW

human scale 1.8m


Numerical models

• Structural model

• OptiStruct used as solver (eigenmode analysis)

• Composite model

• 100 eigenmodels

• CFD model

• Steady state, Multiple-Reference-Frame (MRF)

• Spalart-Allmaras RANS turbulence model

• ~ 50Mio elements

inflow periodic

outflow farfield

blade

Not true to scale

Eigenmode

CFD mesh


Results (varying inflow speed & rotor RPM)

• Surface streamlines

• Power / Thrust

4.0 m/s Wind Speed 17.0 m/s Wind Speed

Larger separation bubble

Discrepancy with FAST results Some discrepancy

(due to separation bubble)

Remark: FAST and WT_Perf are commonly used tools in the wind turbine domain, containing simplifications for the CFD part


4. Summary

• Challenges for external aero

• Large complex models (+50Mio cells)

• Advanced physics (e.g. diff. scales, turbulence)

• Transient (e.g. FSI, noise)

• AcuSolve

• Accurate, scalable, robust CFD solver

• Fluid-Structure-Interaction (FSI) capabilities

• Altair’s Virtual Wind Tunnel for external automotive aero

• Aerodynamic examples

• Automotive: Classical benchmarks (ASMO, Ahmed)

& FSI rear wing

• Wind turbine: FSI of turbine blade

• Both cases very good match with exp. data

Parking turbine blade

Virtual Wind Tunnel

Aerodynamic Simulations with AcuSolve

Technology

rights reserved

rear wing

boundary layer

2mm 20

100 eigenmodes computed

noise

underbody

acusolve