Multi-response Robust Design Based on Principal Component and Grey Relational Analysis

Robust design is an effective approach to improve product quality and reduce manufacturing cost. Taguchi method, which is widely used, is a traditional method of robust design, but it can´t solve the multi-response problem which often exists in engineering practice, even some responses are correlated. This paper uses principal component analysis (PCA) to eliminate the correlation among multiple responses and reduce the dimensionality of design problems, and then transform the multi-response problem into a single-response problem via grey relational analysis (GRA) so as to obtain the optimal combination of controllable factors. This method has a high practical value because it doesn´t involve complicated process of mathematics or statistics. Finally, the effectiveness of the proposed method is illustrated via an engineering example.