An Accurate State of Health Estimation for Retired Lithium-ion Batteries Based on Electrochemical Impedance Spectroscopy

Abstract

With the energy crisis and environmental pollution intensifying, lithium-ion batteries (LIBS) are widely used in new energy industries such as energy storage and electric vehicles. There is a consensus in these industries that the retirement of lithium batteries will usher in a peak in the next few years. Therefore, the capacity estimation and reutilization of retired LIBS has become a hot issue of social concern. In this paper, an accurate estimation model for state of health (SOH) estimation of retired LIBS is established, which is based on electrochemical impedance spectroscopy (EIS) and back propagation (BP) neural network. After comparing the EIS curves under different SOH, we select the maximum impedance of the imaginary part and the impedance amplitude at 0.01Hz and O.IHz in the EIS as the inputs of BP neural network, and the actual SOH is used as the output. The mean absolute error (MAE) and root mean square error (RMSE) of samples for verification are 0.59% and 1.38%, so the SOH estimation model has high accuracy and adaptability for retired lithium-ion batteries.