Mesh deformation is a required component for many computational analyses of transient dynamic systems. Elliptic Smoothing has proven to be a promising mesh manipulation technique. Likewise, mesh deformation algorithms based on radial basis function volume interpolation methods have also proven to be effective. Regardless of the underlying algorithm, unstructured mesh smoothing in three dimensions is a computationally intensive process that requires execution on multiple processors to be effective for most problems of practical interest. Additionally, for large problems, the entire computational mesh does not typically exist on a single processor, which requires the smoothing to be distributed along with the mesh. The focus of this paper is to explore methods for increasing scalability to the mesh smoothing process by making the process distributive across multiple computer processors.
展开▼
