Read Microsoft PowerPoint - HP_CAE_Symposium_Mueller.ppt text version

19th Annual HP CAE Symposium

THE CHALLENGES OF FLUID STRUCTURE INTERACTION

Alan Mueller CD-adapco Seattle

ACM1

Take the FSI challenge

2

Slide 2 ACM1 Ask an engineer what he thinks FSI is and you will get a 100 different answers - all of them right. The truth of the matter - FSI is actually a very broad subject with a range of challaenges from the relatively simple to the extremly difficult. Often an engineer looking from outside may not appreciate what goes on "under the hood" and it may be difficult to asses the level of difficulty of a problem The movie depicts a VOF simulartion of wave loading on a relatively realistic off-shore platform. It gives a glimmer of what is to come in the area of FSI. So join me in this FSI challenge, it may not be quiite as simple as the "Pepsi Challenge" but I hope that it will be ever bit a satisfying.

Alan Mueller, 4/1/2008

Outline

· · · ·

FSI Instances and Classification Mapping and Data Exchange Mesh Motion Demands Robustness and Stability

3

Instances of Fluid-Structure Interaction (FSI)

· ·

Pressure-actuated valves, pumps Flow induced vibration of structures

­ ­

cables, risers, towers, etc. flutter sloshing in fuel tanks, wave loading, etc.

·

Free surface applications

­

· · ·

Solid/Fluid thermal interaction Tire hydroplaning Biomedical applications

An understanding of both solid and fluid domains is essential for an accurate simulation.

4

Fluid-Structure Interaction Classifications

·

1 way or weakly 2 way coupling

­ ­ ­ ­

Deformation has little effect on fluid motion Mechanical, thermal loads: fluid structure Surface temperatures: structure fluid Thermal or hydrodynamic induced loads » Steady state » Transient (multiple transfers)

·

2 way coupling

­ ­ ­ ­

Deformation or rigid body motions significantly impact the fluid motion Mechanical, thermal loads: fluid structure Displacement & temperatures: structure fluid Motions require CFD moving mesh capability

2 Way coupled FSI places an extreme burden on the CFD solver to produce good, body conforming grids

5

Avenues for FSI Coupling in STAR-CD/STAR-CCM+

·

1-way or weakly 2-way coupling

­ ­ ­

Prep/Post: STAR FEA MpCCI, DCI: STAR FEA STAR fluid STAR solid MpCCI, DCI: STAR FEA STAR fluid STAR solid

·

2-way coupling

­ ­

6

Outline

· · · ·

FSI Instances and Classification Mapping and Data Exchange Mesh Motion Demands Robustness and Stability

7

STAR Integration Pre/Post FEA support

· ·

·

·

Import FEA Mesh (ABAQUS, ANSYS) Non-conformal mapping STAR-CCM+ FEA ­ Surface to Surface ­ Volume to Volume st ­ Conserved & 1 order Export FEA Loads ­ Pressure, Nodal Forces ­ Heat flux, Nodal Heat ­ Heat Transfer Coefficient, Ambient Temperatures Import FEA Results import from/export to the native FEA format ­ Displacement Java scripting for dynamic exchange ­ Temperature

8

Steady-State Bidirectional Heat Transfer Analysis

Fluid Polyhedral Grid

Heat Transfer Coefficient

Fluid Temperature

Solid Temperature (ABAQUS)

9

ABAQUS and STAR Direct Coupling Interface (DCI)

ABAQUS MPI socket process 0 process 1

k

All face matching and interpolation on STAR faces local to each processor

n

process 2 ... process n STAR

i m

n

e

10

Aquaplanning Simulations

Important Considerations ­ Tread surface contacts pavement ­ Hydrodynamic loads significantly deform tire ­ No circumferential symmetry ­ Periodic frequency as tread patterns move in/out of contact ­ Tire is able to slip (motion defined by contact friction)

·

STAR-CD

­ ­ ­ ­

Moving Deforming Mesh Arbitrary Sliding Interface VOF for fluid/air interface Cavitation Steel Belted Tires Lateral Treads Vehicle weight, Fluid loads Static or Dynamic Pass fluid loads Pass surface deformation11

·

ABAQUS

­ ­ ­ ­

·

DCI

­ ­

STAR Fluid

·

STAR Rigid Body

Structure Fluid

Stucture Motion defined as Rigid Body (6DOF)

­ ­

Fluid Forces and Moments Body Surface position

· ·

Fluid mesh translates/rotates with rigid Structure Implicit Coupling between Fluids/Structure Canister Example

­ ­ ­

·

CFD of both air and water VOF (Volume of Fluid) free surface capability Rigid canister of given mass, mass moments

12

STAR Fluid

· · · ·

STAR Solid Stress

·

· ·

Build one mesh in one GUI environment Polyhedral meshing advantage Immediate picture of impact of fluid on solid stresses Finite Volume, implicit iterative solver requires significantly less memory than FEA Implicit fluid/solid coupling via memory using similar data structures and iterative procedures Tighter fluid/solid exchange at subiteration level Targeted Applications:

­ ­ ­

Conjugate heat transfer and thermal stresses Fluid-Structure Interaction Casting and Solidification (fluids cells become solid cells)

13

Flow, Heat, and Thermal Stress - Manifold

Gas Temperature Gas Velocity

Manifold Temperature

Manifold effective stress

14

Turbine Blade Analysis

· ·

·

Fluid Flow and Conjugate Heat Transfer Static load: Elastic solid with centrifugal and thermal loads Creep : Viscoplastic solid with stress and temperature dependent creep rates temperature

effective stress fluid flow

15

Examples of Finite Volume Stress Analysis

Temperature (2M cells)

Stresses (6M DOF)

Solid (12.6M DOF), Fluid (1 M cells) 16

Scalability: Engine Block and Head (12.6 M DOF)

· · ·

Thermoelastic stress solves on Cray in 14 CPU min (28 Processors) Convergence rate independent of # processors Good scalability to about 20K cells per processor

M achine : Cray node AM opteron 250, 2.4Ghz, D 4Gb RAM per node, 2 processors per node, Rapid Array M PI Problem : 4.2M cells, 12.6M DOF, 9.8Gb RAM

32 28 24 speed-up 20 16 12 8 4 0 0 4 8 12 16 20 24 28 32

1 3 speed up 7

HP ProliantScalability of Solid Stress Solver Opteron 2 16 node(AMD Turbo with 700414 cells, on White cluster dual core/node,Infiniband,8Gb/node, 1GHz

10 9 8

6

5

4

2

CPU processors

0 0 4 8 12 16 20 24 28 32 36 number of processors

17

Outline

· · · ·

FSI Instances and Classification Mapping and Data Exchange Mesh Motion Demands Robustness and Stability

18

Defining the Mesh at all times

·

·

·

Fluid mesh must conform to boundaries and maintain cell quality Mesh motion is difficult not simply because a structure moves Mesh motion is difficult when a structure moves close to other structures (rigid or not)

The goal : Define the mesh motion with as little user Think Contact! intervention as possible

19

Mesh Evolution Strategies in STAR-CD/STAR-CCM+

· · · ·

Mesh translation/rotation Mesh morphing (constant topology) Sliding Interfaces Topology changes

­ ­ ­

Cell Insertion/Deletion Re-meshing/Interpolation Parallel Meshing

· ·

Immersed Boundary Method (Low Re) Overlapping Meshes

Many "preprocessing" meshing features available 20 to STAR-CD/STAR-CCM+ during solve

Mesh Translates/Rotates with Rigid Body

· ·

Mesh conforms to motion of the body External fluid BC must be preserved during motion

Even rigid body motion adds to the mesh motion complexity

21

Advanced Morphing Polyhedral Meshes

·

· · ·

Small number of control points define motion on moving boundary region Morpher preserves quality of grid, boundary layer User control on other boundary regions Morphing in Parallel

22

Outline

· · · ·

FSI Instances and Classification Mapping and Data Exchange Mesh Motion Demands Robustness and Stability

23

Explicit vs Implicit Coupling Explicit coupling is the most widely used coupling between FEA and CFD codes, but ... · A tighter implicit coupling requires a deeper and more intimate integration of the solvers Stable increasingly STAR fluids/solids are implicitly coupled · The stiff, and massive structure · CD-adapco is working with our FEA partners to Unstable develop fully implicit coupling.

·

increasingly incompressible, and massive fluid

24

Conclusions

·

Fluid/Structure Interaction is a broad field with many distinct challenges and classes

­

STAR offers many strategies by which to couple the fluid to the structure appropriate for the particular physics

·

Mesh motion places extraordinary demands on the CFD to preserve grid quality

­

STAR offers several strategies by which to evolve the mesh at solver time

25

Enjoy the Conference

26

Information

Microsoft PowerPoint - HP_CAE_Symposium_Mueller.ppt

27 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

1213056