Prediction of Remaining Capacity of Lithium Battery Based on QPSO-LSTM Model
To overcome the problem of low accuracy in predicting lithium-ion battery capacity,a quan-tum particle swarm optimization improved long short-term memory(QPSO-LSTM)neural network bat-tery capacity prediction technology is proposed.Firstly,the basic principles of QPSO and LSTM algorithms are analyzed.Then,to solve the problems of poor prediction accuracy and difficulty in determining hyperpa-rameters for long-term data,the QPSO algorithm is used to optimize the main hyperparameters of the LSTM model,such as the number of neurons and learning rate.The QPSO-LSTM model is constructed.Finally,taking NASA batteries as the analysis object,four prediction models,QPSO-LSTM,PSO-LSTM,LSTM,and GA-BP are used to predict the remaining capacity of two different types of batteries.The prediction results show that the QPSO-LSTM model had high prediction accuracy,with an error within 1.5％,providing an effective method for predicting the remaining capacity of batteries.
万方数据
