Thermo-structural optimization of air-cooling battery thermoregulation system using surrogate modelling and NSGA-II

The present work investigates an efficient air-cooling system based on simultaneous thermo-structural optimization using numerical and surrogate-based evolutionary modelling techniques. The electro-thermal model of a lithium titanate oxide battery has been coupled with the dynamics model of an electric vehicle subjected to a realistic drive cycle. Multi-objective optimization of thermo-structural parameters has been carried out using NSGA-II considering average battery temperature, temperature gradient, pressure loss, and packaging efficiency as the performance parameters. Natural convection cooling proved inadequate for battery thermal management as the average battery temperature exceeded 45 degrees C at the end of the simulated drive cycle. Polynomial-based regression equations were formulated using response surface methodology with an R2 value exceeding 0.966 for reliable prediction of the performance parameters. The optimized air-cooling system achieved an average battery temperature of 38.37 degrees C, a temperature gradient of 4.77 degrees C, a pressure loss of 12.08 Pa, and a packaging efficiency of 62.41 %. The proposed air-cooling system resulted in a reduction of about 16.4 % in the average battery temperature and 14.57% in the maximum battery temperature relative to the natural convection cooling of the battery pack. The present study proposes coupling surrogate and genetic evolutionary models to optimize the thermo-structural parameters of the battery thermoregulation system while considering the dynamic power requirements of electric vehicles.