Implementation of EKF combined with discrete wavelet transform-based MRA for improved SOC estimation for a Li-Ion cell

This work gives insight to the design and implementation of an improved extended Kalman filter (EKF) algorithm combined with the discrete wavelet transform (DWT) for a Li-Ion cell. The discharging/charging voltage signal (DCVS) is applied as source data in the DWT-based analysis due to its ability to extract information from the non-stationary and transient phenomena simultaneously in both the time and frequency domain. Through using the multi-resolution analysis (MRA) implementing the wavelet decomposition and reconstruction, two DWT techniques in processing the EKF-based state-of-charge (SOC) estimation have been newly introduced to improve the battery management (BMS) performance. Firstly, irrespective of the measurement noise model and data rejection for compensating the model errors in the existing method, this work requires less computation and satisfies an improved SOC estimation by the use of a low frequency component An of the DCVS, as terminal voltage of a cell. Secondly, the denoising technique in processing the noise-riding DCVS dealing with a high frequency component Dn enables us to estimate SOC without any modification of the algorithm. Experimental results indicate the robustness of the proposed approach for the reliable SOC estimation.