The difference between the computed solutions and the exact solutions results from the errors due to the finite precision arithmetic and the truncation error. Both of these errors are the functions of the spacing between discrete points. To minimize these errors, it is often necessary to adapt a given computational mesh for a specific numerical problem. A new solution adaptive criterion is developed in the present study for the solution of computing compressible flows on unstructured meshes. The characteristics and performances of the new criterion are compared with a modified Mitty's adaptive criterion. The Euler equations are discretised by three kinds of third-order TVD schemes for the convection term, which are explicitly solved by a five-stage Runge-Kutta scheme. The focus of present study is to develop a general form of solution adaptive criterion and the refinement and de-refinement procedure for coarse initial meshes to obtain accurate solution. Results are presented for inviscid two-dimensional circular cylinder flow, compression ramp flow, supersonic bump channel flow and transonic bump channel flow, as well as three-dimensional intake flow.
|Number of pages||1|
|Publication status||Published - Jan 1 1997|
|Event||Proceedings of the 1997 ASME Fluids Engineering Division Summer Meeting, FEDSM'97. Part 24 (of 24) - Vancouver, Can|
Duration: Jun 22 1997 → Jun 26 1997
|Other||Proceedings of the 1997 ASME Fluids Engineering Division Summer Meeting, FEDSM'97. Part 24 (of 24)|
|Period||6/22/97 → 6/26/97|
ASJC Scopus subject areas