A Methodology to Develop and Validate a 75-kWh Battery Pack Model with Its Cooling System under a Real Driving Cycle

A major issue of battery electric vehicles (BEV) is optimizing driving range and energy consumption. Under actual driving, transient thermal and electrical performance changes could deteriorate the battery cells and pack. These performances can be investigated and controlled efficiently with a thermal management system (TMS) via model-based development. A complete battery pack contains multiple cells, bricks, and modules with numerous coolant pipes and flow channels. However, such an early modeling stage requires detailed cell geometry and specifications to estimate the thermal and electrochemical energies of the cell, module, and pack. To capture the dynamic performance changes of the LIB pack under real driving cycles, the thermal energy flow between the pack and its TMS must be well predicted. This study presents a BTMS model development and validation method for a 75-kWh battery pack used in mass-production, mid-size battery SUV under WLTC. Eighty thermocouples, pressure, and coolant flow sensors are installed on the different battery cells, bricks, and modules to capture the time-series thermal and electrical performance changes. The dual e-motor vehicle is tested on a chassis dynamometer to measure transient pressure drop, inlet-outlet coolant temperature, battery pack and brick temperature, and voltage. These data are used to validate the integrated battery pack and its TMS cooling circuit. The pack model consists of 4416 cells based on 2RC branches equivalent-circuit model (ECM) calibrated using an electrochemical Pseudo 2D approach. The integrated BTMS model includes all inlet and outlet cooling flow channels to validate its thermal and electrical performance under steady-state and WLTC tests. The total model can capture dynamic performance changes, such as pack and brick temperature, pressure drop, and voltage, with over 90% accuracy. The model can be used in complete vehicle model simulation with low computation time and high accuracy for future VTMS studies.