Prediction and optimization of preheating temperature for CGFRPP composite laminates based on BP neural network

The integrated overmolding of fiber-reinforced thermoplastic composite laminates has emerged as an innovative manufacturing process. The laminate temperature uniformity is the main factor affecting product quality. In this paper, the temperature of continuous glass fiber-reinforced polypropylene (CGFRPP) laminate was measured using a thermal imaging camera. A finite element model was established according to the heat transfer equation to analyze the temperature distribution of the laminate. The Grey wolf optimization (GWO) algorithm and chaos mapping were used to optimize the BP neural network. The CIGWOBP temperature prediction model was established. By independently controlling the heating power of the heating tubes, the temperature field of the laminate became uniform. The maximum temperature gradient was 9.4°C and the optimized efficiency was 51%. The laminate structure was optimized with a preheating time of 30 s. These results are expected to guide future efforts in achieving temperature uniformity in heated composite laminates during the integrated over-molding process in the automotive industry.