No abstract

This study applies nonlinear model predictive control (NMPC) to the torque-vectoring and front-to-total anti-roll moment distribution control of a four-wheel-drive electric vehicle with in-wheel-motors, a brake-by-wire system, and active suspension actuators. The NMPC cost function formulation is based on energy efficiency criteria, and strives to minimize the power losses caused by the longitudinal and lateral tire slips, friction brakes, and electric powertrains, while enhancing the vehicle cornering response in steady-state and transient conditions. The controller is assessed through simulations using an experimentally validated high-fidelity vehicle model, along ramp steer and multiple step steer maneuvers, including and excluding the direct yaw moment and active anti-roll moment distribution actuations. The results show: i) the substantial enhancement of energy saving and vehicle stabilization performance brought by the integration of the active suspension contribution and torquevectoring; ii) the significance of the power loss terms of the NMPC formulation on the results; and iii) the effectiveness of the NMPC with respect to the benchmarking feedback and rule based controllers.

Abstract-Fully electric vehicles are rapidly gaining user and market interest worldwide, due to their zero direct emissions, appealing driving experience and fashionable perception. Unfortunately, cost, range and reliability have not reached the desired targets yet. Since consumers are prone to spend money to have a more reliable system, Design-for-Reliability will be a useful tool for the Design of tomorrow's EVs, justifying part of the increased cost for these products. In this work, a vertical model-based approach to design-for-Reliability of power converters for EVs is presented, paying special attention to thermally-induced aging. The design starts from various driving cycles, properly assembled to describe the vehicle mission, then load profiles for the converters are found and the resulting thermal stress is quantified. The converter lifetime can be estimated, taking into account also parameter dispersion, and requirements for the active thermal control of the parts modeled achieved, thus giving practical information to the system designers.

No abstract