SOC Estimation of Lithium-Ion Batteries With AEKF and Wavelet Transform Matrix

Due to harsh electromagnetic environment in electric vehicle (EV), the measured current and voltage signals can be seriously polluted, which results in an estimation error of state of charge (SOC). The proposed denoising approach based on wavelet transform matrix (WTM) can analyze and denoise the nonstationary current and voltage signals effectively. This approach reduces the computation burden and is convenient to be programed in microcontroller unit, which is suitable for EV real-time application. The steps of the approach are as follows: 1) decomposition of the current and voltage signals based on WTM; 2) denoising of the wavelet coefficients under the thresholding rule; and 3) reconstruction of the denoised current and voltage signals based on inverse WTM. A battery-management system prototype was built to verify the approach on a Li(NiCoMn)O 2 battery module with nominal capacity of 200 Ah and rated voltage of 3.6 V. SOC estimation error with the proposed denoising approach is limited within 1%. Compared to the maximum error of 2.5% using an adaptive extended Kalman filter without denoising, an estimation error reduction of 1.5% is achieved.