A simulated annealing algorithm for discrete minimal weight design of shallow trusses with stability constraints

Csébfalvi, Anikó Borbála (2002) A simulated annealing algorithm for discrete minimal weight design of shallow trusses with stability constraints. In: Long papers of the Fifth World Congress on Computational Mechanics, (WCCM V). Technische Universitat Wien, Vienna, pp. 1-10. ISBN 3-9501554-0-6

Preview

Text 1269609.pdf Download (46kB) | Preview

Abstract

A new simulated annealing algorithm is proposed for discrete minimal weight design of shallow space trusses. The process allows us to consider the global and local stability problems as well. Because of the highly nonlinear phenomena, it is essential to use appropriate and safe models that detect these effects as accurately as possible. In this paper, a non-linear path-following method is applied for the analysis of non-linear stability problems. The design variables are the cross-sectional areas selected from a given set of cross-sections. In the simulated annealing procedure, a truss structure is characterized by its cross-sectional areas, its weight, and its maximal constraint violation free load intensity factor. A random neighbor is defined to be any truss that can be created by replacing the cross-sectional area of one element group in the current truss by its immediate predecessor or successor area. To demonstrate the efficiency of the simulated annealing method proposed, two numerical examples, a 24-member, and a 30-member dome structure are presented where structural instability constraints and member buckling are considered as well. Computational results indicate that the simulated annealing procedure is a viable approach for a very difficult structure optimization problem.

Actions (login required)

Edit Item