A Low-power Battery Cooling using an Adaptive Liquid-cooled Battery Thermal Management System

Abstract

Electric vehicles (EVs) are gaining popularity due to their environmental and economic benefits. However, their performance and safety depend on the thermal management of their power battery packs, which generate heat during operation. Excessive heat can degrade the battery capacity and lifespan, and even cause fire hazards. Therefore, this research aims to design a low-power battery cooling system using an adaptive liquid-cooled battery thermal management system. The main contributions of this research are: 1) developing a thermal model for the lithium-ion battery pack based on heat generation and dissipation mechanisms; 2) proposing an adaptive control strategy based on deep neural network (DNN) and dwarf mongoose’s coati optimization algorithm (DMCOA) to regulate the battery temperature and minimize the energy consumption of the cooling system; and 3) validating the proposed model and strategy through simulations using AMESim and Matlab.

The proposed thermal model consists of two parts: a heat generation model based on the Bernardi heat generation rate model, which considers the internal resistance, entropy change, and reaction heat of the battery; and a heat dissipation model based on Newton’s law of cooling, which considers the liquid flow rate, heat transfer coefficient, and fluid temperature in the cooling channels. The thermal model is applied to a single cell and then extended to the entire battery pack using the principle of energy conservation. The proposed adaptive control strategy uses the mass flow rate of the water pump as a control parameter and the battery temperature as a state variable. It employs a DNN to predict the vehicle speed fluctuations based on the driving cycle, and then uses a DMCOA to optimize the temperature cost function, which balances the battery temperature and energy consumption objectives.

To evaluate the performance of the proposed thermal model and control strategy, a simulation model of the battery’s cooling system was built using AMESim, which was coupled with Matlab for data processing and optimization. The simulation results showed that the proposed thermal model could accurately capture the temperature dynamics of the battery pack under different operating conditions. The simulation results also showed that the proposed adaptive control strategy could effectively regulate the battery temperature within a safe range, while reducing the energy consumption of the cooling system by an average of 15% and 27% compared to conventional PID and MPC methods under predefined NEDC cycle conditions.

-