Numerical stability of dynamic relaxation analysis of non‐linear structures

Abstract
The estimation of the parameters (‘fictitious densities’) which control the convergence and numerical stability of a non‐linear Dynamic Relaxation solution is described. The optimal values of these parameters vary during the iterative solution and they are predicted from the Gerschgörin bounds, that is rowsums of the stiffness matrix, which are divided into constant and variable parts for computational convenience. The procedure is illustrated by reference to the analysis of an axially loaded beam on a non‐uniform elastic foundation.

This publication has 8 references indexed in Scilit: