ESCL_에너지저장변환 실험실

Energy Storage Conversion Lab.

Publication

International

Multi-Variate Discrete Wavelet Technique for Advanced State-of-Charge Estimation of a Lithium-Ion Cell
Content type 2017
Title of paper Multi-Variate Discrete Wavelet Technique for Advanced State-of-Charge Estimation of a Lithium-Ion Cell
Author Woonki Na, Sesha Kovvuri, and Jonghoon Kim
Publications 105
Status of publication published
Vol Energy Procedia
Link 관련링크 https://www.sciencedirect.com/science/article/pii/S1876610217310767 46회 연결

This paper introduces an innovative approach for replacing the active and passive equalization circuits used for

balancing the voltage and state-of-charge imbalances among Lithium-Ion cells. In this research, design and

implementation of multi-variate discrete wavelet technique (MDWT) for de-noising experimental discharging

charging signals are considered. WMULDEN function in MATLAB is used for MDWT in a battery application. The

experimental charging discharging voltage signal (ECDVS) is applied as source data for the WMULDEN based

analysis. The WMULDEN is a strategy based on wavelet decomposition. It has an ability to extract useful

information from multiple non-stationary signals simultaneously by analysing in the decomposed data through

principal component analysis (PCA). By using WMULDEN analysis, the de-noised experimental discharging

charging voltage signals from Lithium-Ion cells can be derived. The signal to noise ratio (SNR) and the execution

time comparisons have been performed between two methods, MDWT and univariate Discrete Wavelet Transform

(DWT). The comparison results show that MDWT has a better performance than DWT in terms of Signal to noise

ratio and the execution time.