Electrical and Electrochemical Fusion Model Design and Safety Diagnosis of Lithium-ion Batteries Based on Probability Density Function to Reflect Optimal Dynamic Characteristics

This paper presents an optimal voltage model that incorporates both electrical and electrochemical characteristics using a weight-based fusion model. Moreover, it proposes a battery safety diagnosis method by applying an Unscented Kalman Filter based on this model. The proposed algorithm uses deterioration data in which the swelling phenomenon of a nickel-manganese-cobalt battery occurred. In addition, a weight-based fusion model was designed to select the optimal terminal voltage value by fusing direct current, alternating current, and open circuit voltage models. Using this model, an Unscented Kalman Filter algorithm was applied to diagnose battery safety through the SOC estimation error (residue), and the charging state was estimated based on the deterioration profile in which the swelling phenomenon occurred. Verification of the proposed method confirmed that it exceeded the threshold, which is the failure determination criterion set after the occurrence of battery swelling. A comparative verification of the proposed method was also performed using normal deterioration data.