Mesh CFD
Mesh CFD
Mesh CFD
Dimitri J. Mavriplis ICASE NASA Langley Research Center Hampton, VA 23681 USA
11th International Meshing Roundtable September 15-18, 2002 Ithaca New York, USA
Overview
History and current state of unstructured grid technology of CFD
Influence of grid generation technology Influence of solver technology
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Overlapping grid system on space shuttle (Slotnick, Kandula and Buning 1994)
Advantages
Complex geometries Adaptivity Parallelizability
Enabling factors
Maturing grid generation technology Better Discretizations and solvers
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Mature technology
Numerous available commercial packages Remaining issues
Grid quality Robustness Links to CAD
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Agglomeration Multigrid
Edge-data structure graph of grid Agglomeration Multigrid levels = graphs Excellent load balancing up to 1000s of processors
Homogeneous data-structures (Versus multi-block / overlapping structured grids)
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Practical Examples
VGRIDns tetrahedral grid generator NSU3D Multigrid flow solver
Large scale massively parallel case Fast turnaround medium size problem
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
3.1 million vertices, 18.2 million tets, 115,489 surface pts Normal spacing: 1.35E-06 chords, growth factor=1.3
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Current approach
Generate coarse (O(10**6) vertices on workstation Refine on supercomputer
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Edge data-structure Line solver in BL regions near walls Agglomeration Multigrid acceleration Newton Krylov (GMRES) acceleration option Spalart-Allmaras 1 equation turbulence model
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Parallel Implementation
Domain decomposition with OpenMP/MPI communication
OpenMP on shared memory architectures MPI on distributed memory architectures Hybrid capability for clusters of SMPs
Weighted graph partitioning (Metis) (Chaco) Coarse and fine MG levels partitioned independently
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Mesh independent property of Multigrid GMRES effective but requires extra memory
Parallel Scalability
Cases Run
Baseline grid: 1.6 million points
Full drag polars for Mach=0.5,0.6,0.7,0.75,0.76,0.77,0.78,0.8 Total = 72 cases
120 Cases (excluding finest grid) About 1 week to compute all cases
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Adaptive Meshing
Potential for large savings through optimized mesh resolution
Well suited for problems with large range of scales Possibility of error estimation / control Requires tight CAD coupling (surface pts)
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Mapping techniques
Hessian based Grid quality
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Refinement Criteria
Weakest link of adaptive meshing methods
Obvious for strong features Difficult for non-local (ie. Convective) features
eg. Wakes
Surface grid motion Interior grid motion Grid sensitivities Automation / Parallelization
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Sensitivity Analysis
objective function (e.g., Stress, CD)
Manual differentiation Automatic differentiation tools (e.g., ADIFOR and ADIC) Complex variables Finite-difference approximations
Grid Sensitivities
Grid v
f
Grid Grid
f s
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
Potential for large efficiency gains Spectral element methods Discontinuous Galerkin (DG) Streamwise Upwind Petrov Galerkin (SUPG)
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA
High-Order Discretizations
Require more complete surface definition Curved surface elements
Additional element points Surface definition (for high p)
Conclusions
Unstructured mesh CFD has come of age
Combined advances in grid and solver technology Inviscid flow analysis (isotropic grids) mature Viscous flow analysis competitive
Complex geometry handling facilitated Adaptive meshing potential not fully exploited Additional considerations in future
Design methodologies New discretizations New solution techniques H-P Refinement
11th International Meshing Roundtable September 15-18,2002 Ithaca New York, USA