Reliability-Based Structural Optimization Under Stochastic Time Varying Loads

Abstract

A methodology for reliability-based structural optimization under time varying random loadings with emphasis on realistic modeling of the loadings and the effect of uncertainties of loadings on the design is developed. It includes studies on optimization problem formulation, loading modeling, efficient reliability calculation as well as mathematical nonlinear programming. As the most difficult problems in such optimization are the accurate evaluation of the reliability under combined loads and the large amount of computation required in the reliability analysis, the primary attention is given to these problems.

A tri-modal point estimate for the system reliability is developed, which gives more accuracy and continuity of the reliability function without extensive computational effort. A general methodology for large system designs is developed, by which the computational work can be reduced drastically.

The load combination rules in the current codes are examined through the use of the proposed load models and the load combination method. Based on the results, some improvements over the current load combination rules are suggested.