Modeling and optimization of micro heat pipe cooling battery thermal management system via deep learning and multi-objective genetic algorithms

Battery thermal management and electrochemical performance are critical for efficient and safe operation of battery pack. In this research, a multi-physics model considering the battery aging effect is developed for micro heat pipe battery thermal management system (MHP-BTMS). A novel multi-variables global optimization framework combining multi-physics modeling, deep learning and multi-objective optimization algorithms is established for optimizing the structural parameters of MHP-BTMS to improve battery thermal management and electrochemical performance simultaneously. It is found that MHP-BTMS fails to control the temperature of aged battery pack due to the higher heat generation caused by solid electrolyte interphase formation. After 1000 cycles, the maximum temperature and maximum temperature difference were increased by 3.32 K, 2.49 K, 2.04 K and 1.78 K, 1.46 K, 1.26 K, respectively. It is also found that the battery electrochemical performance during the cycling is highly related to battery thermal behaviors. MHP-BTMS with 0.004/s inlet velocity achieved the best performance in preventing SEI formation and battery aging effect, which was lower by 7.01 nm (SEI) and 1.65% (aging), 2.31 nm and 0.58% as compared to 0.002 and 0.003 m/s cases. Besides, MHP-BTMS with optimized inlet velocity, MHP arrangement and cold plate can improve cooling performance and electrochemical performance. Multi-variables global optimization can provide the optimal structure parameters of MHP-BTMS under the different combinations of weighted coefficients and optimization strategies to achieve the trade-off between battery thermal issues and electrochemical performance. In addition, it is demonstrated that the weighted coefficients and optimization strategies in this novel framework can be changed according to the actual needs in engineering applications.