Research On Energy Saving Techniques Based On Parameter Optimization In Injection Molding Process

Energy saving and emission reduction has become a key national policy in China. Catering forthe urgent demand on the low cost and high efficiency energy saving techniques in injection moldingindustry, this thesis focuses on the parameter optimization based energy saving methods for bothsingle-product and multiple-product cases. The proposed methods can reduce energy use and assureproduct quality at the same time. They are purely data-based techniques, which require no re-purchaseor re-build expenses for upgrading the existing low-end injection molding machines. The keycontributions are summarized as below.(1) For a single-product injection molding process, a dual-objective parameter optimizationframework is developed to reduce energy use and assure product quality simultaneously. First,Taguchi’s method is used to design experiments for generating initial process data for process analysis,modeling and optimization. Then, analysis of variance (ANOVA) is used to determine the key factorsfor both product quality and energy consumption. Following that, artificial neural network (ANN) isused to map the nonlinear relationships between the control factors and the response variables, and todevelop the prediction model. Finally, genetic algorithm (GA) based lexicographic method is adoptedto search the optimal process parameter settings for the formulated dual-objective optimizationproblem. An updating procedure is also included in the proposed framework.(2) For a multiple-product injection molding process, a single-machine-scheduling strategy isproposed for minimizing the global energy consumption. First, the total energy consumption in amultiple-product process is studied, which contains three parts: switchover energy, transitional energyand stable operation energy. Based on the transitional process data, a partial least square (PLS) basedtransitional energy consumption model is developed. Then, a signed di-graph (SDG) model isdeveloped to describe the global energy consumption, where the nodes represent the stable productionmodes and the arcs represent the transitional modes. The scheduling problem is finally formulated as atypical travelling salesman (TSP) problem, for which, a GA-based lexicographic optimization methodis proposed to search the optimal production sequence and the corresponding optimal parametersettings for the stable production modes.(3) Experimental research has been done in the center for polymer processing and systems(CPPS) of Hong Kong university of Science and Technology (HKUST). Results demonstrated that theproposed two energy saving techniques are effective, which can reduce energy consumption significantly, and meanwhile, the obtained process parameter settings can contribute to the productionof satisfying product quality.