A Lithium-Ion Batteries Fault Diagnosis Method for Accurate Coulomb Counting State-of-Charge Estimation

Real-time and accurate estimating state-of-charge (SOC) of a lithium-ion battery is a critical but technically challenging task for battery management systems. Coulomb counting algorithm is an effective real-time SOC estimation algorithm but suffers from three typical faults: initial SOC fault, battery capacity fault, and biased load current measurement fault, making its estimation accuracy challenging in practice. To solve the above-mentioned problem, this paper proposes a model-based fault diagnosis algorithm for the Coulomb counting algorithm. The proposed algorithm effectively diagnoses the faults, where the diagnosis requires merely the load current and the terminal voltage of the battery without extra measurements or prior knowledge of the battery. Also, this algorithm is performed alongside the Coulomb counting algorithm either intermittently or remotely in the cloud to ensure the real-time SOC estimation feature of the Coulomb counting algorithm. To showcase the performance of the proposed algorithm, two experiments: a battery discharging experiment using a standard electric vehicle driving profile and a Monte Carlos experiment were performed. Both experiments well-demonstrate the effectiveness of the proposed battery electric circuit model-based SOC estimation algorithm in diagnosing the three typical faults of the Coulomb counting algorithm with 100% true-positive rates.