A Model for Evaluating Remaining Energy of Retired Battery Based on Improved PSO-BP Neural Network

Abstract

AbstractAiming at the insufficiency of the remaining energy detection methods for decommissioned batteries specified in the current national standards, it is proposed to use neural networks to build a model for evaluating the remaining energy of decommissioned batteries to provide a basis for the subsequent recycling and reuse of decommissioned batteries. In view of the problems of traditional BP neural network being easy to fall into local optimal solutions, it is proposed to increase the momentum term and the learning rate adjustment function at the same time; for the problems of slow convergence of the PSO algorithm, the linear decreasing weight method is used as the weight update strategy, and the learning factor is adjusted at the same time. The improved PSO algorithm is combined with the improved BP neural network to construct a PSO-BP neural network model to evaluate the residual energy of decommissioned batteries. It has been verified that the PSO-BP model has a faster convergence speed and a higher accuracy during the training process, and the accuracy rate of the residual energy evaluation result of the retired battery is 98%, which has high practical application value.KeywordsImproved BP algorithmPSO algorithmPSO-BP neural network