Abstract： Determining optimal process parameter settings critically influences productivity, quality, and cost of production in the plastic injection molding industry. Selecting the proper process conditions for the injection molding process is treated as a multi-objective optimization problem, where different objectives, such as minimizing product weight, volumetric shrinkage, or flash present trade-off behaviors. As such, various optima may exist in the objective space. This paper presents the development of an experiment-based optimization system for the process parameter optimization of multiple-input multiple-output plastic injection molding process. The development integrates Taguchi’s parameter design method, neural networks based on PSO (PSONN model), multi-objective particle swarm optimization algorithm, engineering optimization concepts, and automatically search for the Pareto-optimal solutions for different objectives. According to the illustrative applications, the research results indicate that the proposed approach can effectively help engineers identify optimal process conditions and achieve competitive advantages of product quality and costs.6211
Keywords Plastic injection molding . Back-propagation neural networks . Particle swarm algorithm . Multi-objective . Optimization
1 Introduction
  Optimizing process parameter problems are routinely performed in the manufacturing industry, particularly in setting final optimal process parameters. Final optimal process parameter setting is recognized as one of the most important steps in injection molding for improving the quality of molded products. Traditionally, the process conditions are often determined by experienced engineers or based on reference handbooks and later improved and fine-tuned by trial and error and Taguchi’s parameter design method on the shop floor. This method depends greatly on the experience of molding operators and could potentially be costly and time consuming, especially with new resins or new applications, thus it is not suitable for complex manufacturing processes. Hsu [1] argued that when using a trial-and-error process, it is impossible to verify the actual optimal process parameter settings. Moreover, Taguchi’s parameter design method can only find the best specified process parameter level combination which includes the discrete setting values of process parameters. Application of the conventional Taguchi parameter design method is unsuitable when one of the process parameter variables is continuous, and it cannot help engineers obtain optimal process parameter setting results [2]. Furthermore, when engineers deal with a multiresponse process parameter design problem, the conventional Taguchi parameter design method runs into difficulties [3]. Advanced methods are highly demanded to model and optimize the injection molding process with the purpose of manufacturing high-quality plastic parts.
  The quality characteristics of plastic injection-molded products can be roughly pided into three kinds: (1) the dimensional properties, (2) the surface properties, and (3) the mechanical or optical properties [4]. Previously, researchers showed that product weight is a critical quality characteristic and a good indication of the stability of the manufacturing process in plastic injection molding. Yang et al. [4] revealed that product weight is an important attribute for plastic injection-molded products because the product weight has a close relation to other quality properties (e.g.,surface and mechanical properties), particularly other dimensional properties (e.g., thickness). Kamal et al. [5] showed that controlling the product weight is of great commercial interest and can produce great value for production management. One of the most common defects in injection molding is flash. Flash occurs when excess plastic material is extruded from the edges of a mold. Flash may be caused by changes in the processing conditions such as injection pressure, melt temperature, clamp pressure, improper feeding materials, or mold damage [6]. In many cases, the excess material must be trimmed manually, which lowers process efficiency. To reduce inefficiency caused by flash, in-process flash detection techniques are essential. Zhu and Chen [7] described the development of a fuzzy neural network-based in-process mixed material caused flash prediction system for injection molding processes. During the plastic injection molding process, one of the biggest challenges is shrinkage which deteriorates the quality of produced parts [8]. Therefore, product weight, flash, and volumetric shrinkage are feasible quality characteristics which can be used as important responses in the process parameter optimization of plastic injection molding. 塑料注塑工艺英文文献和翻译:http://www.youerw.com/fanyi/lunwen_3625.html