Artificial Neural Network Based Dual-functional System for Thermal and Fire Safety Management for Lithium-ion Batteries

Description

In the development of global industry, the far-flung use of traditional fuels brings tons of greenhouse gas emissions which harms the environment and accelerates global warming. Therefore, it is an urgent need to develop the secondary energy source for transportation like electric vehicles (EVs). Among the energy storage devices, lithium-ion batteries (LIBs) with high energy density, long lifespan and high efficiency, have been regarded a popular alternative and booming in the market. However, their characteristics of high energy density and less thermally stability bring in key challenges for thermal safety issue. Given the LIBs are sensitive to the working temperature, battery thermal management systems (BTMSs) are developed by researchers to dissipate heat and control battery temperature in the battery pack.However, the current BTMSs are mostly designed or compared based on limited parameters settings and can hardly focused on multiple parameters simultaneously. This research gap may lead to an omission of the potential optimal solution due to the blank study in some unknown or un-designed parameters. The implementation of machine learning techniques, or more specifically, artificial neural networks (ANN), can be a potential method to conduct the system optimization. Although some previous studies proved that the efficiency of using ANN to solve battery thermal problems, to the best of our knowledge, the application of ANN to battery research is emerging and there is still a large room to implement ANN to battery system design and optimization. For example, the detailed temperature distribution of LIB and battery pack have not been fully investigated.Furthermore, a safer BTMS should also have the capability of mitigating battery thermal runaway (TR) since the battery fires happen now and then. There is, however, scarce available information regarding the intelligent protection system design in the BTMS. To the best of our knowledge, there exists no relevant ANN-based BTMS which investigated dual functions (i.e. controlling temperature in normal-use and predicting TR propagation in extreme-condition) simultaneously.This project aims to develop an innovative ANN-based dual functional BTMS for LIBs, with which the thermal and fire safety management could be improved efficiently. The ANN model will be established and trained based on comprehensive data collections in this study including experimental data and CFD simulation results with mathematical model analysis. The outcomes of project can be a pioneering study to design and optimize an advanced ANN-based dual-functional BTMS, which may provide new references and guidance for future BTMS design.