Electrode-Parameter-Based Fault Diagnosis and Capacity Estimation for Lithium-Ion Batteries in Electric Vehicles
Yiming Xu, Xiaohua Ge, Ruohan Guo, Cungang Hu, Weixiang Shen
Abstract
In this study, we present an open circuit voltage (OCV) reconstruction method to extract electrode parameters of electric vehicle lithium-ion batteries for short-circuit (SC) fault detection and capacity estimation. More specifically, a set-valued observer is first employed to identify OCVs in real time. Then, data screening is developed to check current polarity to calculate the mean OCVs at the same SOC during discharging. These mean OCVs are further used as inputs for the particle swarm optimization algorithm to determine electrode parameters. After that, the obtained electrode parameters are utilized to detect SC fault occurrence with the help of a K-nearest neighbor algorithm and accurately estimate battery capacity. The feasibility and applicability of the proposed method are demonstrated through extensive experiments with various types of batteries under different SC resistances, current profiles, and battery aging levels.