Optimization of structures under stochastic loads

Abstract

The objective of this study is to develop a method for design under multiple stochastic loads based on optimization. The objective function, such as cost, is minimized under the constraint that the probabilities of various limit states being reached are within allowable limits. The loads are treated as random processes and their combined effects evaluated by the load coincidence method. The limit states considered are either yielding or first plastic hinge at the member level or plastic collapse at the system level. A penalty function approach and a Sequential Unconstrained Minimization method are used. Sensitivity studies with respect to the load parameters are carried out. The dependence of the resultant optimal design on the constraints used, e.g. at the member or system level, is investigated. The proposed approach to the problem is shown to be a viable method and the design using bi-level (member and system) reliability constraints is shown to yield more risk-consistent results.