r-Refinement grid adaptation algorithms and issues Original Research Article

Issues concerning r-refinement adaptation are discussed in terms of the research underway at North Carolina State University. A brief justification for adaptation is offered and adaptation criteria are stated. The NCSU adaptation algorithms DSAGA and SIERRA are described and results obtained with these algorithms are used to illustrate the adaptation issues. Grid quality, truncation errors added and reduced by adaptation, and current research areas are discussed. Qualitative analysis of grids produced by the adaptive algorithms reveals that conventional standards of grid quality are violated frequently in regions where solution improvement due to adaptation is evident. Alignment of the grid with solution features is shown to be a more appropriate indication of mesh improvement than cell shape. Benefits and weaknesses of r-refinement adaptation are summarized and areas for further research are noted.