PARALLEL ADAPTIVE MESH REFINEMENT TECHNIQUES FOR PLASTICITY PROBLEMS

Accurately modeling the nonlinear properties of materials can be compu tationally expensive. Parallel computing offers an attractive way for solving such problems. However, the efficient use of these systems req uires the vertical integration of a number of very different software components. To investigate the practicality of solving large-scale, no nlinear problems on parallel computers, we explore the solution of two - and three-dimensional, small-strain plasticity problems. We consider a finite-element formulation of the problem with adaptive refinement of an unstructured mesh to accurately model plastic transition zones. We present a framework for the parallel implementation of such complex algorithms. This framework using Libraries from the SUMAA3d project, allows a user to build a parallel finite-element application without w riting any parallel code. To demonstrate the effectiveness of this app roach on widely varying parallel architectures, we present experimenta l results from an IBM SP parallel computer and an ATM-connected networ k of Sun UltraSparc workstations. The results detail the parallel perf ormance of the computational phases of the application during the proc ess while the material is incrementally loaded. (C) 1998 Elsevier Scie nce Ltd. All rights reserved.