Optimization of injection-molded light guide plate with microstructures by using reciprocal comparisons

Injection molding is considered as an effective way to manufacture light guide plate (LGP) with microstructures. However, the determination of processing parameters has always been difficult pertaining to shrinkage and warpage, bringing about variations in quality of the injection molded product. This study proposed a procedure for solving the optimization problem utilizing reciprocal comparisons in data envelopment analysis. This method attempts to improve the comparisons of efficiency between different systems (CEBDS) which has been done in other works. The objectives are to make the depth and angle of the V-cut microstructures as close to the target values as possible, and make the residual stresses of light guide plate minimum. First, Taguchi method with orthogonal array L18\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hbox {L}_{18}$$\end{document} was applied to reduce the number of experiments. Then, the significant factors which have profound effect to the quality characteristics were confirmed using the ANOVA and main effect analysis. Next, CEBDS and reciprocal comparisons were conducted to optimize the multi-parameter combination. The results were compared and investigated both theoretically and experimentally. The reproducibility of the experiment was verified by confirming a confidence interval of 95 %. It is inferred that the reciprocal comparisons approach is far superior to the CEBDS approach. The results also demonstrated improvements of LGP qualities compared to the best achieved in Taguchi orthogonal array experiment.