Battery fault diagnosis and thermal runaway warning based on the Feature-Exponential-Function and Dynamic Time Warping method

With the development of new energy technologies, electric vehicles are becoming more and more widely available. However, faults within the battery pack such as thermal runaway and internal short circuit are still a serious problem for electric vehicles, which is a challenge for the existing fault diagnosis methods. In this paper, firstly, Gaussian Smoothing Filter (GSF) is used for noise reduction of voltage data to get higher quality voltage data. Secondly, a novel Feature-Exponential-Function (FEF) method is proposed for extracting the voltage features between battery pack. This FEF feature value can effectively distinguish the faulty batteries from other normal batteries, which is easy to detect the faulty batteries, and the calculation method is straightforward. Finally, the improved Dynamic Time Warping (DTW) algorithm is proposed to accomplish the automatic detection and localization of faulty batteries online. Validation on operational vehicles with internal short circuits and thermal runaway in cloud data shows that the method can effectively detect faulty batteries and enable early warning of thermal runaway of batteries.