Abstract
This paper presents a novel real-time energy management strategy (EMS) for plug-in hybrid electric vehicles (PHEVs), which combines the adaptive neuro-fuzzy inference system (ANFIS) and the model predictive control (MPC). A two-objective EMS with two state variables is defined by integrating the battery aging and fuel economy in the objective function. First, the dynamic programming (DP) approach is applied offline to obtain the globally optimal solutions. Then a real-time predictive EMS is proposed, in which DP carries out a moving-horizon optimization. Contrary to the charge-sustaining HEVs, the optimal trajectory of the battery state-of-charge (SOC) in PHEVs does not fluctuate around a constant level. Thus, determining the desired value of SOC for the real-time moving-horizon optimization is a challenging issue. Unlike the EMSs with a pre-determined reference for SOC, a trained ANFIS model constructs the real-time sub-optimal SOC trajectory in advance. Finally, the effectiveness of the proposed approach is shown through simulation. The proposed EMS is examined over multiple real-time driving cycles, and the results indicate that the total cost is increased compared to those unaware of battery aging. The real-time EMS is then compared to different approaches. While suboptimal, the proposed EMS is real-time implementable, and the results are found to be close enough to those of optimal controller, compared to the two other tested approaches.
A Mixed Logical Dynamical-Model Predictive Control (MLD-MPC) Energy Management Control Strategy for Plug-in Hybrid Electric Vehicles (PHEVs)
