Lithium-ion battery state of health estimation using the incremental capacity and wavelet neural networks with genetic algorithm

Accurate state of health (SOH) is a crucial factor for the regular operation of the electric vehicle. Compared with the equivalent circuit methods, the data-driven methods do not rely on the battery model and do not need to measure the open-circuit voltage. This paper proposes an on-line method based on the fusion of incremental capacity (IC) and wavelet neural networks with genetic algorithm (GA-WNN) to estimate SOH under current discharge. Firstly, IC curves are acquired, and the important health feature variables are extracted from IC curves using Pearson correlation coefficient method. Second, The GA is used to optimize the initial connection weights, translation factor and scaling factor of WNN; then, the GA-WNN model is applied to estimate battery's SOH. Third, the established model is verified by battery data. Finally, the experiment results show that the SOH estimation error of this method is less than 3%.

Automated summary

This paper proposes an online method based on incremental capacity and wavelet neural networks, which can estimate the health status of the battery under current discharge. By extracting important variables from IC curves and optimizing the WNN model parameters using genetic algorithm, the SOH of the battery is successfully estimated with an error of less than 3%.