System-Level Energy Management Optimization Based on External Information for Power-Split Hybrid Electric Buses
Xiaodong Sun, Ziyin Dong, Zhijia Jin, Xiang Tian
Abstract
In recent years, the improvement of fuel economy in power-split hybrid electric buses (PS-HEBs) has emerged as a prominent subject of discussion. This article introduces a system-level optimization (SLO) that relies on the collection of road information. The control strategy is categorized into online and offline parts. First, the speed information throughout the entire driving process is obtained by integrating the global positioning system with real-time vehicle map data. Subsequently, the K-means clustering method is employed to classify distinct kinematic segments based on several characteristic parameters. Based on this point, the back propagation neural network prediction model is constructed by monitoring the variations in speed sequence, enabling the prediction of future speed curves using historical speed. Besides, a nested architecture is utilized for optimizing the motor to attain collaborative optimization across multiple disciplines. Finally, the proposed strategy achieves adaptive energy management for different road segments by incorporating the predicted vehicle speed and pertinent information into a rule-based strategy. Through simulation and hardware-in-the-loop experiments, the superiority of the proposed strategy in dealing with road segment changes is demonstrated, leading to a substantial reduction in fuel consumption.