Integrated battery power capability prediction and driving torque regulation for electric vehicles: A reduced order MPC approach

Abstract

To comply with battery power constraint in the operation of electric vehicles, traditional methods usually estimate battery’s power capability first and then regulate vehicle’s driving torque based on the estimation. These methods have fundamental drawbacks due to their separated nature. A better way is to integrate the estimation and regulation together, which increases the model complexity of the control problem though. To tackle this issue, a reduced-order model predictive control (MPC)-based approach is proposed in this paper, where the dimension of the control model is reduced from two to one by exploiting a quasi-linear relationship between the two state variables. Rigorous mathematical justification proves that sufficient accuracy is retained as long as MPC’s prediction horizon is determined according to the initial states and battery’s polarization dynamics. The superiority of the proposed method is validated by model-in-the-loop tests, which demonstrate that the proposed method reduces the possibility of over-discharge and can make a flexible trade-off between power constraint handling and target vehicle speed tracking.