汽车工程 ›› 2022, Vol. 44 ›› Issue (4): 514-524.doi: 10.19562/j.chinasae.qcgc.2022.04.007

所属专题: 新能源汽车技术-电驱动&能量管理2022年

• • 上一篇    下一篇

融合工况识别的增程式电动汽车模糊能量管理策略研究

陈勇1,魏长银1,李晓宇1(),李彦林1,刘彩霞1,林霄喆2   

  1. 1.河北工业大学机械工程学院,天津市新能源汽车动力传动与安全技术重点实验室,天津  300130
    2.吉利汽车动力总成研究院,宁波  471002
  • 收稿日期:2021-11-04 修回日期:2021-11-29 出版日期:2022-04-25 发布日期:2022-04-22
  • 通讯作者: 李晓宇 E-mail:lixiaoyu@hebut.edu.cn
  • 基金资助:
    宁波市科技计划项目(2019B10111);国家重点研发计划(2018YFB0106403)

Research on Fuzzy Energy Management Strategy for Extended-Range Electric Vehicles with Driving Condition Identification

Yong Chen1,Changyin Wei1,Xiaoyu Li1(),Yanlin Li1,Caixia Liu1,Xiaozhe Lin2   

  1. 1.School of Mechanical Engineering,Hebei University of Technology,Tianjin Key Laboratory of New Energy Vehicle Power Transmission and Safety Technology,Tianjin  300130
    2.Geely Powertrain Research Institute,Ningbo  471002
  • Received:2021-11-04 Revised:2021-11-29 Online:2022-04-25 Published:2022-04-22
  • Contact: Xiaoyu Li E-mail:lixiaoyu@hebut.edu.cn

摘要:

针对模糊能量管理策略设计仅依赖专家经验很难适应复杂工况的问题,本研究提出了一种基于神经网络工况识别的增程式电动汽车模糊能量管理策略。首先,基于中国货车行驶工况(CHTC-HT)数据,利用改进遗传算法优化的BP神经网络构建工况识别模型;其次,根据所识别的工况类型,融合电池SOC及整车需求功率参数,设计了自适应模糊能量管理策略,通过实时获取发动机功率输出实现能量优化分配;最后,通过硬件在环测试验证了所提出的方法。结果表明自适应模糊策略油耗相比规则策略降低9.67%,比模糊策略降低7.84%,有效提高了整车经济性。

关键词: 增程式电动汽车, 模糊能量管理策略, 神经网络算法, 改进遗传算法, 硬件在环

Abstract:

For the problem that the design of the fuzzy energy management strategy is difficult to adapt to complex driving cycles owing to only dependent on expert experience, a fuzzy energy management strategy is proposed for extended-range electric vehicles with working condition identification based on neural network. Firstly, the working condition identification model is designed based on the data of Chinese high truck driving cycle (CHTC-HT) with back propagation neural network optimized by improved genetic algorithm. Subsequently, combining the identified working condition with battery state of charge and vehicle power demand, an adaptive fuzzy energy management strategy is developed to implement optimized energy distribution by real time acquisition of engine output power. Finally, the proposed method has been verified by hardware in the loop test. The results show that the proposed adaptive fuzzy energy management strategy can reduce fuel consumption by 9.67% compared with CD-CS strategy and by 7.84% compared with fuzzy energy management strategy, which effectively improves the fuel economy for extended-range electric vehicles.

Key words: extended-range electric vehicles, fuzzy energy management strategy, neural network algorithm, improved genetic algorithm, hardware in the loop