Dynamic energy management for a novel hybrid electric system based on driving pattern recognition

• Published in: Applied Mathematical Modelling Vol. 45; pp. 940 - 954
• Main Authors: Lei, Zhenzhen, Qin, Datong, Liu, Yonggang, Peng, Zhiyuan, Lu, Lilai
• Format: Journal Article
• Language: English
• Published: New York Elsevier BV 01.05.2017
• Subjects: Adaptive control; Batteries; Battery cycles; Cluster analysis; Clutches; Computer simulation; Driving conditions; Electric vehicles; Energy consumption; Energy management; Fuel consumption; Gear trains; Hybrid electric vehicles; Hybrid systems; Hybrid vehicles; Knowledge bases (artificial intelligence); Particle swarm optimization; Pattern recognition; Power consumption; Power efficiency; Switching
• ISSN: 0307-904X; 1088-8691; 0307-904X
• DOI: 10.1016/j.apm.2017.01.036

This paper focuses on the dynamic energy management for Hybrid Electric Vehicles (HEV) based on driving pattern recognition. The hybrid electric system studied in this paper includes a one-way clutch, a multi-plate clutch and a planetary gear unit as the power coupling device in the architecture. The powertrain efficiency model is established by integrating the component level models for the engine, the battery and the Integrated Starter/Generator (ISG). The powertrain system efficiency has been analyzed at each operation mode, including electric driving mode, driving and charging mode, engine driving mode and hybrid driving mode. The mode switching schedule of HEV system has been designed based on static system efficiency. Adaptive control for hybrid electric vehicles under random driving cycles with battery life and fuel consumption as the main considerations has been optimized by particle swarm optimization algorithm (PSO). Furthermore, driving pattern recognition based on twenty typical reference cycles has been implemented using cluster analysis. Finally, the dynamic energy management strategy for the hybrid electric vehicle has been proposed based on driving pattern recognition. The simulation model of the HEV powertrain system has been established on Matlab/Simulink platform. Two energy management strategies under random driving condition have both been implemented in the study, one is knowledge-based and the other is based on driving pattern recognition. The model simulation results have validated the control strategy for the hybrid electric vehicle in this study in terms of drive pattern recognition and energy management optimization.