Health-Aware Speed-Prediction-Based Energy Management Strategy for Fuel Cell Logistic Vehicles

This paper develops a two-layer energy management strategy for fuel cell hybrid trucks, incorporating considerations of operational costs. In the upper layer, an enhanced speed predictor is proposed, combining fuzzy C-means clustering with long-and-short-term memory neural network. It can identify driving states in real time and choose the appropriate sub-models for speed forecasting. In the lower layer, a multi-objective function including fuel consumption cost and power-source aging cost is established. And sequential quadratic programming is used to derive the optimal control action inside each receding horizon. Additionally, the performance differences induced by speed prediction techniques and the length of prediction horizon are analyzed. Results show that vehicular total operating cost can be reduced by at least 6.76<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> and fuel cell degradation cost by at least 11.52<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> compared with benchmark solutions.