Optimization for Liquid Cooling Cylindrical Battery Thermal Management System Based on Gaussian Process Model

Abstract

Abstract The power of electric vehicles (EVs) comes from lithium-ion batteries (LIBs). LIBs are sensitive to temperature. Too high and too low temperatures will affect the performance and safety of EVs. Therefore, a stable and efficient battery thermal management system (BTMS) is essential for an EV. This article has conducted a comprehensive study on liquid-cooled BTMS. Two cooling schemes are designed: the serpentine channel and the U-shaped channel. The results show that the cooling effect of two schemes is roughly the same, but the U-shaped channel can significantly decrease the pressure drop (PD) loss. The U-shaped channel is parameterized and modeled. A machine learning method called the Gaussian process (GP) model has been used to express the outputs such as temperature difference, temperature standard deviation, and pressure drop. A multi-objective optimization model is established using GP models, and the NSGA-II method is employed to drive the optimization process. The optimized scheme is compared with the initial design. The main findings are summarized as follows: the velocity of cooling water v decreases from 0.3 m/s to 0.22 m/s by 26.67%. Pressure drop decreases from 431.40 Pa to 327.11 Pa by 24.18%. The optimized solution has a significant reduction in pressure drop and helps to reduce parasitic power. The proposed method can provide a useful guideline for the liquid cooling design of large-scale battery packs.