Enhanced Battery Power Control at Low Energy Level for Electrified Powertrains via Load Prediction

Battery power control at low energy level is crucial to appropriate operation of electrified powertrains. This paper presents a Model Predictive Control (MPC)-based method to tackle this issue in a computationally efficient manner. The method takes future load disturbances into consideration to more accurately manage battery voltage constraint while performing control optimization. A forward Dynamic Programming (DP) algorithm is implemented to derive the optimal control policy in the MPC framework. To resolve the well-known computational issue in DP, L "curse of dimensionality", model order is reduced by reasonably approximating the two-dimensional feasible state space with an array of linear lines. The advantage of the proposed method in battery over-discharge prevention is verified by applying it to an electric vehicle driving on a typical city road.