中国机械工程

• • 上一篇    下一篇

[智能感知]基于粒子群优化支持向量机算法的行驶工况识别及应用

石琴;仇多洋;吴冰;李一鸣;刘炳姣   

  1. 合肥工业大学汽车与交通工程学院,合肥,230009
  • 出版日期:2018-08-10 发布日期:2018-08-06
  • 基金资助:
    国家自然科学基金资助重点项目(71431003);
    安徽省高校自然科学研究资助项目(KJ2013B250)

DCR and Applications Based on PSO-SVM Algorithm

SHI Qin;QIU Duoyang;WU Bing;LI Yiming;LIU Bingjiao   

  1. School of Automobile and Transportation Engineering,Hefei University of Technology,Hefei,230009
  • Online:2018-08-10 Published:2018-08-06

摘要:

实车采集4种典型行驶工况数据,采用随机数法提取并扩充行驶工况识别训练及测试样本,利用多元统计理论对数据进行处理,基于粒子群优化的支持向量机(PSO-SVM)算法来进行行驶工况识别,分析了识别周期及更新周期对行驶工况在线识别精度的影响。将行驶工况识别技术应用在插电式混合动力汽车的能量管理策略中。仿真结果表明,相对于未采用行驶工况识别技术以及采用传统SVM算法进行工况识别的能量管理策略,基于PSO-SVM算法工况识别的能量管理策略使整车燃油经济性分别提高9.836%和4.348%,并且电池荷电状态(SOC)变化相对平稳,有利于提高系统效率和延长电池寿命。

关键词: 行驶工况识别, 粒子群优化, 支持向量机(SVM), 插电式混合动力汽车

Abstract: The random number method was used to extract the DCR training and text samples based on collecting 4 typical driving cycle data in real vehicle tests. Multivariate statistical theory was used to process the data and the PSO-SVM algorithm was proposed to conduct DCR. The influences on the on-line identification accuracy of recognition period and update period were analyzed. Finally, DCR was applied in plug-in hybrid electric vehicle energy management strategy. The simulation results show that comparing with the energy management strategy without DCR and the energy management strategy with DCR based on traditional SVM algorithm, the fuel economy of the energy management strategy with DCR based on PSO-SVM algorithm increases by 9.836% and 4.348% respectively, and the battery state of charge(SOC) changes more steady, it is beneficial to improve the system efficiency and prolong the battery lifes.

Key words: driving cycle recognition(DCR) , particle swarm optimization(PSO), support vector machine(SVM), plug-in hybrid electric vehicle

中图分类号: