SOC Estimation for Aging Lithium-Ion Batteries Using an Enhanced DTW-AEKF Algorithm

To ensure the stability and safety of lithium-ion batteries, it is essential to accurately estimate their State of Charge (SOC). This paper presents a novel SOC estimation technique that combines the Adaptive Extended Kalman Filter (AEKF) with an enhanced Dynamic Time Warping (DTW) algorithm to address the decline in SOC accuracy caused by battery aging. The approach leverages the charging voltage curve characteristics, optimizing the weight coefficients and search direction of the DTW algorithm. The enhanced DTW method is then used to evaluate the similarity of charging voltage profiles, indirectly determine the battery's aging state, and estimate the battery's maximum usable capacity. Using this data, a new SOC calculation matrix is created by integrating it with a second-order Thevenin equivalent circuit model. The AEKF is then employed to estimate the SOC using this matrix. Experimental results show that the suggested DTW-AEKF mode achieves a root mean square error (RMSE) of 0.359% and an average SOC estimation error of 0.141% over multiple battery cycles. These outcomes demonstrate the exceptional accuracy of the improved DTW-AEKF algorithm across the battery's life cycle.