Battery Safety: Data-Driven Prediction of Failure

Abstract

Accurate prediction of battery failure, both online and offline, facilitates design of safer battery systems through informed-engineering and on-line adaption to unfavorable scenarios. With the wide range of batteries available and frequently evolving pack designs, accurate prediction of cell behavior under different conditions is very challenging and extremely time consuming. In this issue of Joule, Li et al. 1 used data from a previously reported finite-element model to train machine learning algorithms to predict whether a cell will undergo an internal short circuit when exposed to a selection of mechanical abuse conditions. The presented approach aims to alleviate, and yet is still limited by, a common challenge facing data-driven prediction methods: access to robust, plentiful, high-quality, and relevant experimental data.

Data-Based Interpretable Modeling for Property Forecasting and Sensitivity Analysis of Li-ion Battery Electrode

“…With the rapid development of machine learning and intelligent computing techniques, data-based modeling solutions have become one of many popular tools for the final battery production management [5,6]. A great deal of research has been conducted for internal battery states estimation [7,8], battery aging prognostics [9,10], battery fault diagnostics [11], cell equalization management [12], charging control [13,14], and energy management [15,16].…”

Data-model alliance network for the online multi-step thermal warning of energy storage system based on surface temperature diffusion

“…At present, the fault diagnosis methods of battery energy storage systems are mainly divided into battery model methods and data-driven methods. 7 , 8 The method based on battery model achieves fault diagnosis by comparing the predicted value of the model with the actual measured value of the battery. The premise is to establish a reliable and accurate battery model.…”

Automated feature extraction and selection for data-driven models of rapid battery capacity fade and end of life

“…Beyond the knee point, a cell can be considered to have lost its value for a given application. Forecasting that point is therefore crucial for understanding the lifetime value of lithium-ion batteries [21], [22]. Previous attempts at datadriven health prediction have tried to estimate the timing of the knee point [7], [23] or the cycle life to 80% capacity [19] respectively.…”