Performance analysis on liquid-cooled battery thermal management for electric vehicles based on machine learning

In this paper, the coupling system of liquid-cooled battery thermal management system (BTMS) and heat pump air conditioning system (HPACS) for battery electric vehicles (BEV) is designed and analyzed. The performances of liquid-cooled BTMS are concerned and analyzed from the perspective of air conditioning based experimental data. Besides, an automatic calibration model of the liquid-cooled BTMS based HPACS is established to predict cooling capacity and system coefficient of performance (COP) of the BTMS by support vector regression (SVR). To better obtain three hyper parameters (the penalty coefficient C, the RBF kernel function parameter gamma, and the insensitive loss coefficient epsilon) of SVR model, particle swarm optimization (PSO) algorithm is introduced to optimize above three parameters. It is found that compared to SVR model, the correlation coefficient (R) of cooling capacity and system COP for the proposed PSO-SVR model in this paper is improved 2.1% and 2.8% respectively, the mean squared error (MSE) of and cooling capacity and system COP is reduced 87.8% and 82.9% respectively, which indicated that PSO-SVR model can be used as a new method to fit the complex nonlinear relationship among the system COP, cooling capacity and other influencing factors of the liquid-cooled BTMS based HPACS.