Adaptive Mesh Generation within a 2D CFD Environment Using Optimisation Techniques
A method for generating optimal solution adaptive computational grids in terms of smoothness, local orthogonality and a volume weighted functional has been decribed. This method is 2 dimensional but it extends to 3 dimensions. The first operation in this grid adaption is to locate all regions which require grid refinement or coarsening. A procedure has been developed which detects where the grid needs adapting. It returns an indicator which has been termed the ’adaptive sensor’. This indicator is scaled and filtered and then used when adapting the grid. The second operation in this grid adaption is to generate the new grid which involves the development of a method to transfer information from the old grid to the new grid. This method is divided into two separate parts. The first part is a search to decide where the new grid point is positioned with respect to the old grid and the second part interpolates the information from the old onto the new grid. Both the interpolation scheme and the search method are discussed.
Using the adaptive grid generation method developed, varying degrees of coupling have been considered. Tests have been carried out on flow over a 10% bump in a channel, ranging from extremely loose coupling (producing adaptive grids externally to the flow code execution) to extremely close coupling (producing adaptive grids following each flow code cycle). The resulting solutions and execution times have been compared and a recommendation made as to the level of coupling to be used for the most efficient use of the grid adaption procedure.
KeywordsMaster Cell Grid Point Mach Number Flow Solver Adaptive Grid
Unable to display preview. Download preview PDF.
- Bristeau, M. O. and Periaux, J. ‘Finite element methods for the calulation of compressible viscous flows using self-adaptive mesh refinement’, VKI Lecture Notes on CFD, March 1986.Google Scholar
- Brackbill, J. U. and Saltzman, J. S. ‘Adaptive Zoning for Singular Problems in Two Dimensions’, J.Comput. Physics 46 (1982).Google Scholar
- Carcaillet, R. ‘Generation and Optimisation of flow-adaptive computational grids’, MSc Thesis University of Texas at Austin, 1986.Google Scholar
- Pierre, D. A, ‘Optimization Theory with Applications’, ISBN 0–486–65205–X, 1986.Google Scholar