Improved nonlinear solution techniques in the VAST finite element program


  1. Jiang, L.
  2. Chernuka, M.W.
Corporate Authors
Defence Research Establishment Atlantic, Dartmouth NS (CAN);MARTEC Ltd, Halifax NS (CAN)
This report presents theoretical and implementation details related to recent improvements in various linear and nonlinear solution techniques in the VAST finite element program. In order to improve the efficiency of the VAST solver, the preconditioned conjugate gradient method was implemented to provide an alternative means for solving systems of linear algebraic equations. Seven pre-conditioners, derived from the stationary iterative methods and incomplete factorization algorithms, were implemented and benchmarked. Numerical results suggest that for well conditioned problems, iterative solves can be considerably more efficient than the direct solver. However, for poorly-conditioned problems, such as those resulted from structural analysis of thin shell structures, the performance of the preconditioned conjugate gradient method was not satisfactory. In order to further improve the efficiency of the iterative solver, pre-conditioners based on the multi-level or multi-grid methods have been recommended. In order to provide an accurate and reliable method for generating residual stresses for load-shortening curve computations of stiffened panels, a linear thermal stress capability was developed in the VAST program. A stress scaling factor was employed to ensure that the maximum residual stress along the plate-web junction equals the yeild stress. TRUNCATED
Iterative solvers;Direct solvers;Pre-conditioners;Scaling factors;Linear thermal stress
Report Number
DREA-CR-2000-101 — Contractor Report
Date of publication
01 Feb 2000
Number of Pages
Hardcopy;Document Image stored on Optical Disk

Permanent link

Document 1 of 1

Date modified: