This study investigates the energy recovery potential of regenerative suspension systems in battery-electric vehicles and their direct impact on ride comfort under diverse urban driving conditions. Moving beyond simplified estimation methods, a comprehensive electromechanical modeling framework was developed in MATLAB/Simulink. This framework integrates a validated 2-DOF quarter-car model with an explicit generator dynamics and a longitudinal BEV model, ensuring that the regenerative damping force is directly coupled with the vehicle’s motion. The longitudinal model was validated against NEDC data with less than 1% deviation, and the suspension’s electromechanical feedback was found to be consistent with established literature. Simulations were conducted over a speed range of 2–50 km/h on semi-circular and trapezoidal speed bumps, as well as on stochastic ISO 8608 Class A and Class B road profiles. The results reveal that energy recovery is highly sensitive to both vehicle speed and road roughness. For discrete road inputs, maximum energy values of 0.406 Wh and 0.659 Wh were recovered for semicircular and trapezoidal bumps, respectively. Under continuous excitation on ISO Class B roads, the system demonstrated a significant recovery potential of up to 70.4 Wh per kilometer at low speeds, representing a meaningful supplementary energy source. A critical inverse relationship was observed between vehicle speed and harvested energy; as speed increased to 50 km/h, the reduced interaction time between the suspension and road irregularities led to a sharp decline in recovery efficiency. Furthermore, the ride comfort analysis, evaluated per ISO 2631 criteria, highlighted a distinct trade-off: while lower speeds favored both superior energy recovery and passenger comfort, higher speeds required a balance between energy gains and increased vertical accelerations. Overall, this research demonstrates that integrating physics-based electromechanical coupling provides a more accurate and promising outlook for regenerative suspension systems in realistic driving scenarios.
Demir et al. (2026) studied this question.