Thermal flow analysis and electrical safety control strategies for electric vehicle power batteries

Abstract

Electric vehicles (EVs) have become an essential component of the global drive towards sustainable development due to their low carbon emissions and high efficiency, increasing demands for reliable thermal-electrical safety in lithium-ion Power Battery Packs. However, traditional single-point thermistor sensing cannot capture the three-dimensional nonuniform thermal field of battery packs, while infrared thermal imaging and distributed optical fiber temperature sensing provide feasible solutions for high-precision, high-spatiotemporal-resolution full-scale thermal monitoring. A combined thermal-flow-analysis-and-electrical-safety-control strategy for the operation of a 24 cell, prismatic lithium iron phosphate (LFP) battery-pack at 2C-discharge condition was investigated in this paper. A volume-averaged heat-generation model based on the Bernardi formula and anisotropic extended a three-dimensional energy-conservation equation are used to describe the distribution of temperatures in a liquid-cooled pack. This study also reserves hardware interfaces and data fusion modules for infrared thermal imaging and distributed optical fiber temperature sensing, enabling real-time calibration between simulated thermal fields and measured optical thermal fields. A peak in cell temperature at 42.3°C was identified through the application of a finite-element model; Inter-cell differences were approximately 7.8°C. To determine which cell is most thermally vulnerable by setting thermal runaway threshold values and risk scores. Based on this research results, a hierarchical, temperature adaptive battery management system (BMS), which includes Coulomb counting-based SOC estimate, thermal drift-compensation and multi-level protection logic to prevent over voltage - overtemperature-overcurrent situations. The integrated framework can reduce the peak temperature by 23.1% relative to passive cooling, while achieving an state-of-charge (SOC) estimation root-mean-squared-error of only 1.87% over a stable working range of 10-45°C. When connected to optical thermal diagnosis data, this framework can further extend the thermal runaway early warning time by more than 2.3 times compared with traditional single-point sensing. The results offer Quantitative Design References for Thermally-Informed Electrical Safety Strategies of Next-Generation EVs' Battery Systems.