A Study On Robust Design For Dynamic Multi-response System

With the change of customers'requirements, the products extends from single quality characteristic to multiple quality characteristic, and transfer from static quality characteristic to dynamic quality characteristic. Accordingly, robust design also extends from traditional static single response system to dynamic multi-response system gradually. Compared with the researches on robust design for static system, the study robust design for dynamic system is relatively little. So this paper studies the robust design for dynamic multi-response system, and proposes a new robust design method that is simple and feasible.This paper introduces the research situations of dynamic robust design, and integrates the existing robust design methods, proposes a new robust design method on account of the main problems existing in dynamic multi-response robust design. According to the idea of dual-response surface methodology, this paper uses the desirability function to measure both system sensitivity and system variation of each response. In view of the problem that correlations exist between each response, we use principal component analysis to eliminate the correlations. Aiming at the information incompletion of orthogonal design, this paper uses grey relational analysis to gain the grey relational grade (GRD) of system sensitivity and variation principal component sequence relative to their respective ideal solutions, and transfers them into integrated GRD by means of TOPSIS. Then we can determine the optimal control factor level combination through factor effect analysis in light of two-step approach, and predict the signal-noise ratio of each response under the optimal factor combination. At the end of this thesis, we conduct case study on two dynamic multi-response system cases with the proposed method, and compare the results with the references'results, which confirms the validity and superiority of the proposed method.