Modelling battery packs of real-world electric vehicles from data sheet information

Lithium-ion batteries have emerged as the leading enabling technology in developing Electric Vehicles (EVs), But, large-scale publicly available EV data are extremely difficult to find. So it becomes difficult to research and disseminate new methods for monitoring the battery pack of an EV. In this work, we propose a Simulink-based approach to define a virtual-EV model that simulates EV battery pack signals starting from input driving sessions. The battery pack module within the virtual-EV has been fine-tuned using data gathered from real-world EV data sheets. Moreover, the battery pack module includes thermal and aging models, impacting on the output signals, considering the temperature of the surrounding environment and the initial State of Health (SOH) of the battery pack. The virtual-EV generates time series of vehicle's speed, and battery pack's current, State of Charge (SOC), voltage, and average internal temperature according to the input driving cycle. We defined two Simulink EV models emulating two distinct real-world-EVs. Then, we assessed the performances of the simulators comparing the simulated data and real EV data signals collected by the same real-world-EV models, and we obtain, for both simulated EV models, R2 values higher than 0.70 and an RMSE of at most 7V and 8% for the voltage and SOC of the battery pack, respectively.