Discrete Wavelet Transform-Based Feature Extraction of Experimental Voltage Signal for Li-Ion Cell Consistency

The difference in electrochemical characteristics among Li-ion cells in the battery pack inevitably result in voltage and state-of-charge (SOC) imbalances caused by cell-to-cell variation. Therefore, in this approach, with lower requirements of active and passive balancing circuits, a novel approach based on the discrete wavelet transform (DWT) that are well suitable for analyzing and evaluating an experimental charging/discharging voltage signal (ECDVS) is newly introduced to minimize the aforementioned problem. The ECDVS can be applied as source data in the DWT-based analysis because of its great ability to extract variable information of electrochemical characteristics from the nonstationary and transient phenomena simultaneously in both the time and frequency domains. By using the wavelet decomposition implementing the multiresolution analysis (MRA), it is possible to discriminate Li-ion cells that have similar electrochemical characteristics corresponding to information extracted from the ECDVS over wide frequency ranges. Consequently, experimental results showed the clearness of the proposed DWT-based approach for cell discrimination very well.