Near-optimal prediction of geometrical requirements of injection moulded parts using Mamdani-based fuzzy logic controller

Abstract

Injection moulding process is popularly used to fabricate complex and intricate parts with thermo plastic and composite materials. In this paper, Mamdani-based fuzzy logic (FL) controller has been developed to predict the quality of the parts produced using plastic injection moulding machine. It is to be noted that the quality of the parts produced depends on various geometrical requirements, such as global warpage, lower edge surface planarity and hole circularity of the manufactured part. However, these geometrical requirements depend on various input process parameters, namely melting temperature, mould temperature, packing time and packing pressure. As the input-output relationship of the injection moulding process is highly non-linear, FL technique is considered to model the process. It is important to note that the performance of the FL system depends on its knowledge base (that is, rule base and database) developed by the human expertise. In the present paper, genetic algorithm (GA) is used to optimise the optimal knowledge base of the FL system. Further, the prediction accuracy of the developed models has been tested with the help of 20 test cases and found reasonably good accuracy.