Study on the prediction model of bulletproof performance of glass fiber reinforced composites based on improved neural network

DOI:	10.11809/bqzbgcxb2023.07.022
Keywords:	BPNN; glass fiber reinforced composites; bulletproof performance; Fireworks Algorithm
Abstract:	Glass fiber reinforced composite is a kind of material commonly used in the field of weapon protection, and its bulletproof performance is one of the most important indicators to evaluate material properties. The traditional evaluation method of bulletproof performance of materials is the manual experimental comparison data, which has many defects, such as heavy workload, complex experimental methods, high cost and low detection efficiency, and the evaluation results are prone to large deviations. Back propagation neural network (BPNN) can reduce output errors by training autonomous learning rules. Aiming at the problems of traditional material performance evaluation methods, this paper proposes to predict the bulletproof performance of glass fiber reinforced composites through BPNN, which can reduce the workload and improve the efficiency of material bulletproof performance evaluation. Aiming at the defects of BPNN, Fireworks Algorithm (FWA) is used to optimize its weight and threshold. Finally, based on the improved BPNN model, a bulletproof performance prediction model is built. The experimental results show that the prediction accuracy of FWA BPNN model exceeds 99.5%, while the prediction accuracy of PSO BPNN model, GA BPNN model and BPNN model is 99.00%, 98.50% and 98.06% respectively. The experimental results show that the FWA BPNN model proposed in the study can efficiently and accurately predict the bulletproof performance of glass fiber reinforced composites, which is of positive significance to the development of the field of weapon equipment protection in China. It can reduce the time of material performance evaluation, improve the accuracy of material performance evaluation, and provide data support for the selection of weapon equipment materials.
Published:	2023-07-28
PDF HTML