Structural Optimization Using the Design of Experiments and Mathematical Programming.

Abstract

The authors proposed a new practical optimum design method that consists of two steps: the design of experiments and mathematical programming. The design of experiments is used to generate approximate evaluation functions for the controlling behavior depending on the changes in design variables of an object structure, by a series of finite element analyses (FEA). Based upon an orthogonal array of a combination of design variables, effects of the design variables can be calculated by a relatively small number of FEA, and then the approximate evaluation functions are generated by those effects based upon analysis of variance. The evaluation functions can also be used as direct tools for estimating the behavior of design structure. Finally, a successive quadratic programming (SQP) method is employed to solve the optimization problem of the approximate evaluation functions. It is confirmed that the proposed method can be used for almost all kinds of the nonlinear problems including the impact behavior of structures, and that it can be carried out in much smaller number of FEA than the other existing methods. As an example, the present method was used to solve an optimum design problem of an automobile seat frame subjected to impact loading. It was found that the present method is a very effective and powerful tool for the optimum design of various practical design problems.