Real-time energy management of fuel cell hybrid electric vehicle based on variable horizon velocity prediction considering power source durability

Effective energy management strategy (EMS) is essential to ensure the safe and efficient operation of fuel cell hybrid electric vehicle (FCHEV). To improve the economy and reliability of FCHEV, a real-time EMS using variable horizon velocity prediction-based Pontryagin's minimal principle (PMP) considering both fuel economy and power source durability is proposed in this paper. The adaptive PMP is achieved by online updating of the co-state using velocity prediction. To improve the accuracy of velocity prediction, a variable horizon velocity prediction method based on fuzzy C-means clustering (FCM) and radial basis function neural network (RBF-NN) is constructed. Then, to improve the power source durability, a fuel cell (FC) power variation limiting factor with a weight coefficient is incorporated into the Hamiltonian function. And the average daily operating cost is introduced to determine the weight coefficient by evaluating the trade-off between the hydrogen consumption, the FC durability and the battery durability. The simulation results show that the proposed method is able to accurately predict driving behavior and update the co-state in real time. Compared with rule-based EMS, the proposed adaptive PMP achieves a 8.9% reduction in hydrogen consumption and maintains a relatively low FC power change rate, improving fuel economy and power source durability.