This study compares the energy efficiency of a real battery pack with that of a simulated battery pack using a hardware battery simulator, through experimental testing on a dedicated inertia dynamometer designed for light quadricycles. The real battery pack under investigation consists of lithium-ion cells with a nominal voltage of 48 V, a nominal capacity of 100 Ah, and a total energy content of 4.8 kWh. The pack is equipped with a Battery Management System (BMS) that monitors voltage, current, temperature, and State of Charge (SOC). The battery simulator was configured to replicate the same electrical specifications, including the OCV–SOC curve, internal resistance, and current limits, enabling a direct comparison between theoretical and experimental behavior.
Experimental tests were conducted on a custom inertia test bench developed to overcome the limitations of conventional chassis dynamometers when testing light quadricycles, characterized by low mass, limited power output, and independent four-wheel drive systems. Adjustable inertial masses were employed to simulate vehicle masses between 500 and 600 kg. Multiple driving cycles were repeated under controlled conditions. Data acquisition was performed via CAN communication using an IXAAT interface and a dedicated LabVIEW application, while a PLC-based closed-loop control system regulated vehicle speed to accurately reproduce the target driving cycles.
Energy consumption was calculated by accounting for electrical inefficiencies and mechanical losses, applying corrective factors for rolling resistance and aerodynamic drag. The comparison between the real battery’s residual SOC, measured by the BMS, and the theoretical SOC provided by the simulator revealed increasing deviations over repeated cycles. These discrepancies are mainly attributable to internal resistance effects, conversion inefficiencies, and thermal dynamics not fully captured by the simulation model. Although the simulator demonstrated good predictive capability in estimating overall energy consumption, it slightly overestimated the residual SOC under dynamic load conditions, highlighting the need for advanced electro-thermal model calibration for real-world applications.