![BMW Group
Dr. Norbert Gr端n
Lecture Series on
Road Vehicle
Aerodynamics
VKI, Brussels
May 30-June 3, 2005
Page 10
Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics.
Macroscopic Quantities
Macroscopic quantities, such as density, pressure, velocity, etc.
are computed by statistical methods from the state vectors
DENSITY
MOMENTUM
ENERGY
Higher order moments (Energy Flux, Stress Tensor)
are also available locally (do not involve derivatives)
=
j
j txntx ),(),(
rr
=
j
jj txnctxu ),(),(
rrrr
[ ] =
j
jjj txnccmtxE ),(),( 2
1
rrrr
0=u
r
0u
r
Vector length
denotes number
of particles moving
in that direction
m
T
kVRMS 3= 1000 m/s for oxygen at 20属 C
Particle velocities can be much higher
than the resulting macroscopic velocity](https://image.slidesharecdn.com/31db77c4-7a63-43c0-8d4b-231b981fffc9-150507192334-lva1-app6891/85/VKI-RVAD-2005-BMW-Presentation-10-320.jpg)
![BMW Group
Dr. Norbert Gr端n
Lecture Series on
Road Vehicle
Aerodynamics
VKI, Brussels
May 30-June 3, 2005
Page 24
Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics.
CFD Process: Simulation Timestep
Simulations are always run in transient mode
The physical time per timestep is determined by resolution and test conditions
[ ]epsec/timest
V
x
Maa
V
x
Vt LatticeLattice
=
=
Strictly there is no room left for the user to control the timestep
Artificially elevating the Mach number increases the time step
Example:
mmx
epsec/timesttMaCT
smV
2
10515.020
/50
6
=
==属=
=
That means 1 second of physical time requires 200,000 timesteps
Using Ma=0.30 instead of Ma=0.15 cuts the run time in half !](https://image.slidesharecdn.com/31db77c4-7a63-43c0-8d4b-231b981fffc9-150507192334-lva1-app6891/85/VKI-RVAD-2005-BMW-Presentation-24-320.jpg)
![BMW Group
Dr. Norbert Gr端n
Lecture Series on
Road Vehicle
Aerodynamics
VKI, Brussels
May 30-June 3, 2005
Page 40
Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics.
Applications: NASTRAN Interface (Structure)
Select parts per PID
Match NASTRAN parts
with the PowerFLOW model
PLOADs [N/mm2]
on the PowerFLOW model
Map area loads onto
the NASTRAN model
Forces on
individual parts](https://image.slidesharecdn.com/31db77c4-7a63-43c0-8d4b-231b981fffc9-150507192334-lva1-app6891/85/VKI-RVAD-2005-BMW-Presentation-40-320.jpg)
VKI-RVAD-2005-BMW-Presentation
- 1. Application of a Lattice-Boltzmann Code in Automobile and Motorcycle Aerodynamics. Dr.-Ing. Norbert Gr端n Aerodynamics Simulation Lecture Series on Road Vehicle Aerodynamics von Karman Institute for Fluid Dynamics, Brussels May 30 June 03, 2005
- 2. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 2 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Outline Aerodynamic Development Process Physics Overview Simulation Process Validation Examples Various Applications Conclusion
- 3. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 3 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. CFD in the Aerodynamic Development Process Simultaneous Usage of Experimental & Virtual Tools Serial Development PhaseConcept Phase Prototypes 100% Windtunnel Model CFD-Models A C D F C C F A B C D E F Styling- Freeze StylingCompetition A C D F C C F A B C D E F 40% Windtunnel Models CFD-Models Proportion-Studies CFD-Models StylingProcessAeroAnalysisTools
- 4. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 4 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Requirements on CFD as a Productive Tool Accuracy (CD <賊0.005, CL <賊0.010), at least for trends Geometry input preparation minimized Ability to handle complex geometries (underhood & underbody) Deliver results in a reasonable timeframe (over night) Easy to use (by non-numerics specialists)
- 5. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 5 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. CFD Resources 285 222 95 48 248 253 0 50 100 150 200 250 300 1997 1998 2000 2001 2002 2004 2005 TotalNumberofProcessors SUNSUN SGI 2 x SGI (95+127) 1 x SGI (127) 2 x HP (je 63) SUN 1 x SGI (160) 2 x HP (je 63) decicated PowerFLOW servers NumberofProcessors Speed-U p Efficiency on Parallel Computers
- 6. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 6 New Goal : Construct simplified microscopic description (mesoscopic) that still contains the essential micro-physics to achieve desired macroscopic behaviour. Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Motivation for Lattice-Boltzmann Methods (LBM) Microscopic ______________ Mesoscopic ______________ Macroscopic Microscopic ______________ Mesoscopic ______________ Macroscopic Kinetic Theory Lattice Methods Navier-Stokes Kinetic Theory Lattice Methods Navier-Stokes Simulate fluid at microscopic level since the physics is simpler and more general than macroscopic, continuum (PDE) approach. However, complete reproduction of molecular dynamics is much too expensive (today and also in the near future).
- 7. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 7 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. LBM vs. Traditional CFD Methods Real Fluid Free molecules in continous space Kinetic Theory Microscopic particles (Boltzmann Equation) Real Fluid Free molecules in continous space Kinetic Theory Microscopic particles (Boltzmann Equation) Traditional CFD Methods ___________________________ Chapman-Enskog Expansion Statistical Method applied to real gases Navier-Stokes Equations Conservation of Mass, Momentum and Energy Numerical Methods Discrete Approximation of Partial Differential Equations Traditional CFD Methods ___________________________ Chapman-Enskog Expansion Statistical Method applied to real gases Navier-Stokes Equations Conservation of Mass, Momentum and Energy Numerical Methods Discrete Approximation of Partial Differential Equations Lattice-Boltzmann _________________________________ Simulation of Particle Dynamics No integration of partial differential eqn. Movement & collisions conserve mass, momentum and energy No numerical instabilities Lattice-Boltzmann _________________________________ Simulation of Particle Dynamics No integration of partial differential eqn. Movement & collisions conserve mass, momentum and energy No numerical instabilities Results Fluid dynamic quantities at discrete points in space and time Results Fluid dynamic quantities at discrete points in space and time
- 8. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 8 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Basics of Kinetic Theory Boltzmann Equation ),,(),,(),,(),,( tcxCtcxfctcxf t tcxf dt d rrrrrrrrr =+ = Describes the rate of change of the velocity distribution function due to nonequilibrium Velocity Distribution Function ),,( tcxf rr Gives the number of particles at time t per unit volume in phase space around x and c Collision Term C satisfies the necessary conservation laws = 0)()( cdcCc rrr 両 Mass Momentum Energy 1)( =c r 両 cc rr =)(両 2 2 1 )( cc rr =両 Describes fluid behaviour using the interactions of air molecules = cdtcxftx rrrr ),,(),(Density = cdctcxftxutx rrrrrr ),,(),(),(Momentum = cductcxftxE rrrrrr 2 ))(,,(),(Energy
- 9. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 9 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Basics of Lattice Methods Replace the continuous velocity distribution function by a discrete set of particle velocities defined on a lattice of equal shaped cubic cells Vtxiftxintcxf ≠ ),(),(),,( rrrr },...,1;{ miicc = rr Particle dynamics is now described by the Lattice Boltzmann Equation ),(),(),( txiCtxintticxin rrr +=++ The collision operator C determines if a lattice system produces a physically meaningfull fluid behaviour During an elementary time interval particles can only hop from one center of a cell to one of the m near neighbouring cells according to their velocity )1(=t x r ticx + rr ic r
- 10. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 10 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Macroscopic Quantities Macroscopic quantities, such as density, pressure, velocity, etc. are computed by statistical methods from the state vectors DENSITY MOMENTUM ENERGY Higher order moments (Energy Flux, Stress Tensor) are also available locally (do not involve derivatives) = j j txntx ),(),( rr = j jj txnctxu ),(),( rrrr [ ] = j jjj txnccmtxE ),(),( 2 1 rrrr 0=u r 0u r Vector length denotes number of particles moving in that direction m T kVRMS 3= 1000 m/s for oxygen at 20属 C Particle velocities can be much higher than the resulting macroscopic velocity
- 11. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 11 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Transport Coefficients in Lattice Methods Kinetic Theory allows to compute viscosity and thermal conductivity from the velocity distribution function ! T D D a MFPMFP 2+ == 了了僚 The molecular viscosity depends on the mean free path between collisions and the speed of sound (temperature). Viscosity is set by adjusting the relaxation parameter of the collision operator { } ( ) ( ) jc eq jc eq jjcj jjjjj nn nntxn txnCtxntcxn += = +=++ 1 ),( ),(),()1,( r rrrr Lattice-Boltzmann Equation Viscosity is reduced by reducing the mean free path or equivalently the time between collisons Collision frequency for2<c 0>Lattc僚 c/1 錚 錚 錚 錚 錚 錚 錚 錚 = 2 11 cT 僚 Chapman-Enskog Expansion 錚 錚 錚 錚 錚 錚 錚 錚 + = 2 11 2 2 c D T 了 Viscosity Thermal Conductivity
- 12. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 12 Simple 2D Model with 4 directions and 3 speeds Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Concept of Particle Models The fluid is composed of a very large number of particles (not molecules, this is a mesoscale model) Particles are only allowed to move in certain directions on the lattice with limits on how far they can get in a single time step (their speed) The state of the fluid is represented by the number nji of particles moving with speed (energy) j in direction i 1 Possible Directions 2 3 4 Particle with speed 1 in direction 4 Particle with speed 1 in direction 4 Particle with speed 2 in direction 3 Particle with speed 2 in direction 3 A model allowing 3 speeds (0,1,2) and 4 directions re- presents the particle popu- lation by 9 state vectors nji n0 ( = n01 = n02 = n03 = n04 ) n11 , n12 , n13 , n14 n21 , n22 , n23 , n24 State vectors are integers ! Particle with speed 0 Particle with speed 0 The maximum number of particles per state depends on the number of bits for state vectors !
- 13. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 13 Repetition evolves time (t -> t+1) and forms an inherently transient solver The process of evolving (solving) the update equation is inherently parallel (computationally efficient) and stable (computationally robust) is the collision operator that exactly conserves local mass, momentum and energy jC Also drives local distribution to equilibrium (entropy maximized)n eq j Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Fluid Fluid Interaction Dynamics in the fluid consists of two steps : MOVE & COLLIDE Update equation { }),(),()1,( txnCtxntcxn jjjjj rrrr +=++ Time t Time t+1 n1 n2 n2 n1 Example : Mass Conservation n'1 + n'2 = n1 + n2
- 14. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 14 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Fluid Surface Interaction Facets Solid Body Voxels Surfels Automatic discretization The intersection of voxels with the facets representing solid bodies creates surfels which define the computational surface resolution. In each timestep surfels gather and scatter particles, altering their momentum according to the boundary conditions Surface forces depend on the momentum exchange between fluid and wall Vin Vout Specular Reflection Vtin Vnin Vtout Vnout Vin Vout Bounce Back Reflection Slip Condition Normal component inverted Tangential component unchanged Momentum balance normal force only No Slip Condition Normal component inverted Tangential component inverted Momentum balance normal & tangential force
- 15. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 15 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Reynolds-Number Regimes Regime Reynolds Number PowerFLOW __________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________ Low Re < 10,000 Direct Simulation Transitional 10,000 < Re < 100,000 currently not applicable High Re > 100,000 Boundary Layer Simulation approximate values, actual values problem dependent Regime Reynolds Number PowerFLOW __________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________ Low Re < 10,000 Direct Simulation Transitional 10,000 < Re < 100,000 currently not applicable High Re > 100,000 Boundary Layer Simulation approximate values, actual values problem dependent Solid Wall Solid Wall Modeled Flow Direct Simulation Using a large number of voxels the boundary layer is resolved down to the wall with zero velocity at the wall. Particles are bounced back from the wall exactly canceling their momentum. Boundary Layer Simulation The presence of the wall is modeled by a shear stress at the slip surface. Particles loose momentum at the slip surface according to the (modified) law of the wall.
- 16. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 16 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Turbulent Wall Model ( ) Byu += ++ ln 1 虜 Assumption: Universal velocity profile of a turbulent 2D boundary layer with dp/dx=0 ++ = yu :505for も + y :5for + y 0.5 4.0 B 虜 u u u =+ u yy =+ w u = PowerFLOW Extension: include the effect of a longitudinal pressure gradient 錚 錚 錚 錚 錚 錚 += ++++ x p fAAyUyU 1mit)/()( The wall model provides the wall shear stress to alter the momentum of scattered particles. w
- 17. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 17 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Approaches to Turbulence Modeling Dissipation dl LLength 僚/2 dl UL /Time Turbulent Scales 4/3 Re/ dlLRange ( ) ( ) 2/12 Re/// 僚dlULRange RANS = Reynolds Averaging All scales of motion are described by statistical methods (time averaged) LES = Large Eddy Simulation Alle Skalen werden berechnetmodeled computed via modified unsteady Navier-Stokes equations Filter Width (Grid Size) DNS = Direct Simulation All scales of motion in space and time are computed VLES = Very Large Eddy Simulation modeled computed unsteady Coherent anisotropic eddiesUniversal eddies
- 18. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 18 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. CFD Simulation Process CAD/CASModel CATIA/ALIAS CAD/CASModel CATIA/ALIAS Clay Model POLYWORKS Clay Model POLYWORKS Simulation Model (Surface Facetization) ANSA, QUICKMESH, PowerWRAP, ... 1-5Days Simulation Model (Surface Facetization) ANSA, QUICKMESH, PowerWRAP, ... 1-5Days Simulation PowerFLOW 1Day Simulation PowerFLOW 1Day Postprocessing PowerVIZ Postprocessing PowerVIZ ResultResult Shape Modification of CAD/CAS Data Shape Modification of CAD/CAS Data Morphing of the Surface Mesh (PowerCLAY) Morphing of the Surface Mesh (PowerCLAY) Turnaround 2-6 Days Turnaround 2-6 Days
- 19. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 19 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. CFD Process: Geometry Preparation (Wrapping) Complete STL Data (imperfect facetization) Gaps & holes Overlaps & intersections Interior details Wrap Wrapped Surface Facetization (ready for simulation) Water-tight single solid Controlled granularity Interior details removed Preparation timereduced from daysto hours! Complete Set of CAD Data Export or facetize without cleanup or de- featuring
- 20. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 20 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. CFD Process: Geometry Morphing Modification of the surface facetization instead of changing the CAD data which would require a re-facetization.
- 21. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 21 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. CFD Process: Geometry Input The surface facetization represents the geometry only. It does not set the resolution for the simulation. Depending on the level of detail up to 2-3 million facets are used.
- 22. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 22 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. CFD Process: Modular Assembly The complete configuration may be composed of any number of components. Components may be arranged in an arbitrary fashion and also intersect each other.
- 23. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 23 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. CFD Process: Automatic Discretization Voxels (Fluid Cells) Solid Body Facets (Geometry) Surfels (Surface Elements) Typical voxel counts for external aerodynamic cases range from 20-100 milion cells. Geometry representation embedded in a lattice of cubic cells (with different levels of resolution).
- 24. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 24 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. CFD Process: Simulation Timestep Simulations are always run in transient mode The physical time per timestep is determined by resolution and test conditions [ ]epsec/timest V x Maa V x Vt LatticeLattice = = Strictly there is no room left for the user to control the timestep Artificially elevating the Mach number increases the time step Example: mmx epsec/timesttMaCT smV 2 10515.020 /50 6 = ==属= = That means 1 second of physical time requires 200,000 timesteps Using Ma=0.30 instead of Ma=0.15 cuts the run time in half !
- 25. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 25 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. CFD Process: Transient Simulation No explicit convergence criterion, user monitors key quantities to decide when to stop the simulation. 100,000 Timesteps (1 Timestep = 4.7 10-6 sec.) Averaging Window
- 26. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 26 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Validation Models (Scale 1:2.5). 5series touring Open Convertible 5series Limousine with/without Mirror Calibration Motorcycle
- 27. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 27 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Validation: Aerodynamic Forces CZ2 0.114 0.105 CZ1 0.067 0.070 PowerFLOW 3.4: 0.252 CX BMW Windtunnel: 0.252 CZ1 CZ2 -0.038 0.009 -0.027 0.006 PowerFLOW 3.4: 0.276 CX BMW Windtunnel: 0.292
- 28. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 28 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Validation: Surface Pressure Distribution Top Centerline (Geometry not to scale) PowerFLOW Experiment Bottom Centerline (Geometry not to scale)
- 29. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 29 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Validation: Near Surface Flow Topology
- 30. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 30 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Validation: Reynolds Effect
- 31. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 31 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Validation: Motorcycles (Windshield Variations) 0,300 0,320 0,340 0,360 0,380 0,400 0,420 0,440 Serie LT Sport Cx*A Windkanal (Aschheim) PowerFLOW Hot Wire Measurement PowerFLOW Serie LT Sport
- 32. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 32 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Analysis of Drag Generation -0,06 -0,04 -0,02 0,00 0,02 0,04 0,06 0,08 0,10 0,0 0,1 0,3 0,4 0,5 0,6 0,7 0,9 1,0 0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35 0,40 0,0 Cx(x)Verteilung Cx(x)Integral PowerFLOW 3.4: 0.371 CX BMW Windtunnel: 0.382
- 33. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 33 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Analysis of Lift Generation -0,03 -0,02 -0,01 0,00 0,01 0,02 0,03 0,0 0,1 0,3 0,4 0,5 0,6 0,7 0,9 1,0 -0,40 -0,30 -0,20 -0,10 0,00 0,10 0,20 0,0 Cz(x)Verteilung Cz(x)Integral CZ1 CZ2 0.011 0.013 0.143 0.123 PowerFLOW 3.4 BMW Windtunnel
- 34. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 34 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Visualization : Transient Surface Pressure
- 35. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 35 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Visualization : Near Wall Streamlines
- 36. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 36 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Visualization : 3D Streamlines Colors represent Near Surface Velocity Distribution
- 37. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 37 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Visualization: Transient Isosurface VX=0 Reverse flow (Vx<0) inside the isosurface
- 38. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 38 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Visualization : Transient Flow Field Slices
- 39. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 39 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Visualization : Transient Isosurface Cpt=0 For Cpt=0 the total pressure loss is equivalent to the dynamic free stream pressure
- 40. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 40 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Applications: NASTRAN Interface (Structure) Select parts per PID Match NASTRAN parts with the PowerFLOW model PLOADs [N/mm2] on the PowerFLOW model Map area loads onto the NASTRAN model Forces on individual parts
- 41. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 41 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Applications: ABAQUS Interface (Heat Transfer) Mapping heat transfer coefficients from a PowerFLOW simulation onto an ABAQUS structure model
- 42. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 42 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Applications: Detail Optimization (Mirror)
- 43. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 43 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Applications: Underhood Flow
- 44. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 44 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Applications: Motorcycles - Transient Flow Field Near Surface Velocity
- 45. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 45 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Applications: Motorcycles - Transient Flow Field Reverse Flow (Vx<0)
- 46. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 46 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Applications: Motorcycles - Transient Flow Field Different Windshields TouringTouring StandardStandard SportSport Helmkraft StandardStandard TouringTouring SportSport
- 47. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 47 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Applications: Motorcycle Acoustics Punkt 92 0 20 40 60 80 100 120 140 10 100 1000 10000 Hz dB(A) Experiment Berechnung_grob Punkt 101 0 20 40 60 80 100 120 140 10 100 1000 10000 Hz dB(A) Experiment Berechnung_grob Punkt 110 0 20 40 60 80 100 120 10 100 1000 10000 Hz dB(A) Experiment Berechnung_grob Small timesteps enable high sampling rates for pressure and velocity fluctuations. The upper frequency limit is determined by the background noise of the Lattice-Boltzmann method. The lower frequency limit depends on the physical time (number of periods) covered by the simulation.
- 48. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 48 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Applications: Electronics Cooling Goals: qualitatively Heat Transfer Distribution quantitatively Surface Temperatures Heat Transfer Coefficient Velocity Magnitude
- 49. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 49 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Applications: Windtunnel Design
- 50. BMW Group Dr. Norbert Gr端n Lecture Series on Road Vehicle Aerodynamics VKI, Brussels May 30-June 3, 2005 Page 50 Application of a Lattice-Boltzmann Code in Road Vehicle Aerodynamics. Conclusion + Maturity level sufficient for external aerodynamics + Short preprocessing phase due to automatic meshing. + Capability to handle complex geometries (underhood/underbody). + Steep learning curve due to ease-of-use. + Does not require a numerics expert. - Optimization loops still slower than the wind tunnel. - Hardware requirements high for rapid turnaround. - Thermal management capabilities still under development.