Online Extraction of Electrochemical Impedance Spectroscopy Pattern Based on EV Load Profile and Short Time Fourier Transform for Diagnosis of Lithium-ion Battery Safety

Batteries which convert chemical energy into electrical energy and play an important role in our daily lives and across various industries. The chemical reaction that occurs within a battery during operation of an electric vehicle (EV) depends on environmental conditions such as temperature, load, and humidity, and these variables greatly affect battery performance and life. However, there is challenging to analyze these electrochemical phenomena with only simple electric signals, such as direct current (DC) signal. For compensation, passive electrochemical impedance spectroscopy (PEIS) is derived by applying a signal processing algorithm that converts time series signals into frequency-domain signals in this study. Based on this, we propose to learn a siamese convolutional neural network (SCNN) algorithm on the converted PEIS pattern to the image and to effectively diagnose the state of the battery through similarity evaluation.