关键词: 逼近理想解排序, 多目标优化, 回归分析, 稳健设计, 信噪比, 注塑工艺参数

Abstract: Present approaches for the robust optimization of injection molding process parameters cannot thoroughly analyze the influence of uncontrollable factors due to the lack of outer experimental runs. And they are often single-objective optimization with no comprehensive evaluation on the characteristics of molded parts. To overcome these shortcomings, a novel multi-objective robust optimization method is proposed on the basis of parameter design using inner-outer array, analysis of signal to noise ratio (SNR) and the technique for order preference by similarity to ideal solution (TOPSIS). The principal process parameters that have the most significant influence on the part quality are firstly screened out by executing two-level orthogonal experiments and analysis of variance (ANOVA), and then both inner and outer orthogonal arrays can be designed by taking the errors of principal process parameters as noise factors. The quadratic polynomial models for predicting the part quality of given process parameter scheme are formulated on the basis of uniform experimental design and stepwise regression analyses, which are then utilized to efficiently compute the values of quality indices corresponding to all parameter schemes arranged according to the inner-outer array. Thus the SNR of every chosen quality index corresponding to all the parameter schemes of the inner array can be efficiently obtained, based on which the preference of all inner schemes are sorted and the most robust one can be determined. The feasibility of the proposed method is demonstrated by the robust optimization of the process parameters for the injection of a mobile phone panel.

Key words: Multi-objective optimization, Injection molding process parameter, Regression analysis, Robust design, Signal to noise ratio (SNR), Technique for order preference by similarity to ideal solution (TOPSIS)