Abstract—Internal Short Circuit (ISCr) is main cause of dangerous incidents such as thermal runaway in a lithium-ion battery. However, if constant currents are applied to the battery as a load current, existing model-based methods have difficulty in estimating parameters in an equivalent circuit model of the battery accurately, resulting in problem of detection of the ISCr. In this paper, we propose a method for detecting the ISCr in the lithium-ion battery using the Convolutional Neural Networks (CNN). Data pre-processing is conducted to enlarge the effect of ISCr in terminal voltages of the battery, and then the CNN algorithm is used to classify the degree of the ISCr faults. Dataset for the CNN is obtained from a MATLAB/Simulink battery model. The proposed method shows classification result with high accuracy of 96.0% and consequently contributes to detecting the ISCr in the battery early. 
 

Index Terms—internal short circuit, early detection, convolutional neural networks, data pre-processing, lithium-ion battery

Cite: Minhwan Seo, Teadong Goh, Minjun Park, and Sang Woo Kim, "Detection of Internal Short Circuit for Lithium-ion Battery Using Convolutional Neural Networks with Data Pre-processing," International Journal of Electronics and Electrical Engineering, Vol. 7, No. 1, pp. 6-11, March 2019. doi: 10.18178/ijeee.7.1.6-11