RELIABILITY-BASED OPTIMIZATION OF ROBUST FEASIBILITY IN MECHANICAL DESIGN USING PVP TECHNIQUE AND FUZZY FORMULATION

摘要

This paper presents a three-stage integrated optimum design procedure for obtaining the highest design level of robust feasibility and optimum design performance with random variables. The consideration of structural system reliability was merged in the formulation to construct reliability-based robust feasibility optimum design that extends the usability of the proposed method. The parameter variation pattern (PVP) has been adopted to show the correlations among design variables that can be expressed and manipulated. The strategy of multiobjective fuzzy formulation was applied for simultaneously dealing with the maximization of feasibility degree, the optimization of design performance, and the maximization of system reliability. A simple algorithm to compute the amount of modification for each design variable to promote the robustness and design goal performance is introduced in the paper. In general, the proposed robust feasible design is capable of increasing feasibility up to more than 99.9%. Three mechanical design examples illustrate the integrated design methodology, which clearly show the proposed method yielding designs best balanced between the design goal performance and robust feasibility under original design constraints.