scientific article; zbMATH DE number 3907941

From MaRDI portal
Publication:3685360





[For the entire collection see Zbl 0566.00025.] Dynamic relaxation (DR) forms of the governing static equations in nonlinear structural analysis may be used to derive incremental/iterative solution processes in a dynamical context. Explicit implementations of force-level DR equations have enjoyed success in treating problems with strong, sharp nonlinearities. Stiffness-level DR forms, on the other hand, give rise to Newton-type methods suitable for smooth nonlinearities. Application of the latter forms has been hampered by the lack of simple stepsize control strategies, which require an estimate of the directional derivative of the stiffness matrix. In the present paper it is shown how quasi-Newton (QN) updata formulas can be used to derive economical estimates. An analogy between QN and two-step-integrated DR methods is also examined and shown to offer potential insight on parameter selection for the Broyden QN update family.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3685360)