SOH estimation method of lithium-ion battery based on multilayer feedforward neural network

摘要

It is difficult to estimate the state of health (SOH) of lithium-ion battery based on electrochemical mechanism model in real time application, since the mechanism model of the battery is too complex. In this paper, a SOH estimation method based on feedforward neural network is proposed. Only the operation data of the battery is required to train the neural network, which will be used to estimate the SOH of battery. Hence, the mechanism model is unnecessary for SOH estimation. Two schemes for estimating the SOH of battery are designed, which are based on internal resistance and capacity, respectively. Simulation results show that the estimating errors of SOH are below 5%.