Studies in the computation of compressible and viscous flow
Abstract
The theme has been adaptive solution refinement. A novel approach called Truncation Error Injection (TEI) was introduced during the course of research. The idea behind TEI is very simple, i.e., the exact nodal value of the solution to a differential equation could be obtained on any grid and from solving a difference equation that models the differential equation if the truncation error were known. Although the TE is not known in general, it can be approximated on a local grid patch. This approach of approximating the local error due to discretization in effect decouples a problem of multiple disparate length scales into problems of single length scale so that they can be solved more efficiently on a computer than the original problem. Three types of applications have been demonstrated. In addition to solution refinement by TEI, we have shown that the decoupling of the unsteady computation from the steady one by TEI could significantly reduce the computing time and storage for flutter prediction, and that viscous effects can be computed separately and injected into the solution of an inviscid solver for viscous flow computation. Some of the advantages of this approach are: it requires very little modification to the base solver; no compatibility problems in using different grids and different solvers; readily suited for multi processors.
 Publication:

Arizona Univ., Tucson Report
 Pub Date:
 October 1988
 Bibcode:
 1988uat..reptT....F
 Keywords:

 Compressible Flow;
 Computation;
 Computerized Simulation;
 Differential Equations;
 Prediction Analysis Techniques;
 Truncation Errors;
 Viscous Flow;
 Adaptation;
 Approximation;
 Compatibility;
 Computational Grids;
 Decoupling;
 Flutter;
 Injection;
 Mathematical Models;
 Problem Solving;
 Refining;
 Revisions;
 Fluid Mechanics and Heat Transfer