An integrated predictive energy management for light-duty range-extended plug-in fuel cell electric vehicle

A reliable power distribution strategy is of great significance towards the performance enhancement of fuel cell electric vehicles. In this work, a novel model predictive control-based energy management is developed for a fuel cell based light-duty range-extended hybrid electric vehicle. To fulfill the model predictive control framework, a cooperative speed forecasting method based on Markov Chain and fuzzy C-means clustering technique is proposed, which contains multiple predictive sub-models for handling different driving patterns. The final prediction results are obtained by synthesizing the forecasted speed profiles from all sub-models with the quantified fuzzy membership degrees. Besides, an adaptive battery State-of-Charge reference generator is built, which can regulate the SoC depleting rates against various power requirements. Combined with the forecasted speed and SoC reference, the desirable control actions are derived through minimizing the performance index over each finite time horizon. As a result, under the realistic urban-based postal-delivery mission profiles, the proposed strategy can achieve over 3.79% equivalent hydrogen consumption saving and over 40.04% fuel cell power dynamics decrement against benchmark strategy. Moreover, the presented predictive energy management is robust to certain level of trip duration estimation errors, further indicating its suitability for real applications.