R-Adaptive Mesh Quality Improvement via the Target-Matrix Paradigm

P. Knupp, U. Hetmaniuk, L. Diachin

Objective:

Use applied and computational math to research & develop state-of-the-art mesh improvement methods, ultimately for implementation in the Mesquite library.

Current mesh improvement methods tend to be problem-specific, ad-hoc, limited in scope, and have weak formulations.

Approach:

Node-movement strategies based on properly formulated quality metrics, objective functions, and numerical optimization techniques.

Target-matrix paradigm provides a general framework for many different mesh improvement goals.

New Metric for Improving both shape & size of elements:

Significance:

Applications not only need meshes, but good meshes that improve accuracy and efficiency in simulations.

Research Areas: R-adaptivity, ALE rezoning, Guaranteed Invertibility, Surface Mesh Optimization, High-order mesh smoothing, Efficient Numerical Solvers, Parallel mesh optimization, and High-aspect ratio elements.

Accomplishment:

(Return to Applied Math program list)