ROBUST DESIGN METHOD FOR DIVERSE CONDITIONS

摘要

This study demonstrates a robust design method for diverse conditions that are diverse design object, user features, and the environments that the object will be used in. In diverse conditions, two design problems are conceivable. One of the two problems occurs when an objective characteristic is in a non-normal distribution due to strong non-linear, discontinuous or stochastically fluctuant objective functions. We focus on the probability density function of the objective characteristic and propose an index to estimate the feasibility of an objective characteristic within a tolerance. The other problem is when a control factor has an adjustable range that enables users to adjust to meet their needs. To derive the optimum adjustable range of the control factor, we propose an index to estimate the ratio of the sum of sets of the probability of the objective characteristic within the tolerance. The proposed method using these indices is applied to an example where the appropriate cushion angle of a public vehicle seat is determined. As a result, the solutions using the indices are more robust than the conventional one.