A case-based system for process design of injection moulding

Abstract

Process design of injection moulding mainly involves the selection of moulding machine, mould design, cost estimation, and the determination of injection moulding parameters, which traditionally is performed by experienced engineers. Some researchers have attempted to automate the process design by using the simulation, process windows, expert systems, and the artificial neural network approach. In this paper, an artificial intelligence technique, case-based reasoning (CBR), is adopted to develop a case-based system for process design (CBSPD) of injection moulding, which aims to derive a process solution for injection moulding quickly and easily without relying on the experienced moulding personnel. In the system, experience of the process design is represented in cases which are stored in the case library in a structural manner. After the input of the part, production and the quality information, the system searches for the proper cluster of cases and the closest case is then retrieved based on the pre-defined indexes and the two stages of similarity analysis. Two types of adaptation, substitution and transformation, have been introduced to adapt the closest case for the new problem. This approach will not allow the fragile knowledge of the process design for injection moulding to be represented easily, but will facilitate a self-learning capability in the CBSPD.