Predictive Energy Management for Dual-Motor BEVs Considering Temperature-Dependent Traction Inverter Loss

In this article, a predictive energy management system (EMS) for dual-motor battery electric vehicles (BEVs) is proposed, considering temperature-dependent traction inverter loss. First of all, we establish a high-fidelity BEV powertrain under a hardware-in-the-loop (HIL) testbed. The high-frequency switching of power electronics and electrothermal traction inverters is considered. Subsequently, because transient current and voltage-based electrothermal models in the literature are unsuitable for vehicle-level EMS design, we propose an innovative control-oriented inverter loss model and introduce approximate junction temperature dynamics in the predictive EMS. Then, based on Pontryagin's minimum principle, we propose a fast solution algorithm, making it possible to validate the EMS in a real-time HIL testbed. To the best of our knowledge, temperature-dependent traction inverter loss has not yet been studied in EMSs. The traction inverter loss model has been experimentally validated and used in the proposed predictive EMS to provide more comprehensive validation. Results have shown that the proposed predictive EMS can reduce the power loss by 5%-9% compared to the widely used instantaneous optimization-based controller in academia.