HIGH-PERFORMANCE COMPUTING ISSUES FOR MODEL REDUCTION EXPANSION/

Several methods are investigated for finite element model expansion, w ith particular attention paid to the role of high performance computin g for this problem in the area of structural dynamics. Model expansion is used in model updating, which concerns itself with fine tuning fin ite element models so that they agree with experimentally obtained mod al data. Three popular methods are explored: static expansion, dynamic expansion and Improved Reduced System. Beam and plate examples are us ed to investigate the performance of these methods on an Intel Paragon when faced with large problems and very few measurement data. (C) 199 8 Elsevier Science Limited. All rights reserved.