Inconsistency Evaluation of Lithium-ion Batteries Based on Ledoit-Wolf Covariance Estimation and Mahalanobis Distance

Inconsistencies among lithium-ion batteries can significantly affect the performance and lifespan of battery packs. Variations in manufacturing processes and operating conditions may lead to differences across various parameters, causing overall performance degradation and increasing the risks of overcharging, over-discharging, and thermal runaway. To address these inconsistencies, this study introduces an analytical method based on Mahalanobis distance, enhanced with Ledoit-Wolf covariance estimation. By utilizing an adaptive sliding window in combination with these techniques, the method effectively evaluates variations in internal resistance and identifies inconsistent cells, thereby improving the safety and reliability of lithium-ion batteries.