Research on Parameter Identification in Electrochemical Models of Lithium-Ion Batteries Based on Identifiability Analysis and Hierarchical Strategies

The electrochemical model of lithium-ion batteries features a host of microscopic and electrochemically relevant parameters. Hence, it bears substantial challenges to concurrently identify all the unknown parameters of the model solely with the measured current and voltage data, let alone yield accurate results. Since the identifiability of the parameters varies under different work conditions, it is necessary to perform a parameter identifiability analysis to identify the corresponding parameters utilizing highly identifiable work conditions. In this paper, we commence the establishment of a high-precision and low-complexity reduced-order simplified model for the full-order Pseudo-two-dimensions (P2D) model, culminating in a parameter identifiability analysis strategy leveraging the Fisher Information Matrix (FIM). Via analyzing the parameter identifiability of the reduced-order simplified model under different work conditions, a parameter identification strategy for the electrochemical model is proposed in conjunction with the open circuit voltage (OCV) test, work condition test, capacity test, and direct current internal resistance test. The results demonstrate that the proposed parameter identification strategy boasts high precision, where the relative error of the identified parameters falls within 4%, and the mean absolute error (MAE) of the voltage is below 2mV in the simulation test analysis.