主办：陕西省汽车工程学会
ISSN 1671-7988  CN 61-1394/TH
创刊：1976年
汽车实用技术

汽车实用技术 ›› 2026, Vol. 51 ›› Issue (7): 13-19.DOI: 10.16638/j.cnki.1671-7988.2026.007.003

• 新能源汽车能量自适应管理 • 上一篇    

基于工况识别的纯电动汽车制动能量回收 控制研究

丁磊 1，董志辉 1，吴颂 2，韩立星 1，蔡建军 3   

  1. 1.柳州城市职业学院；2.东风柳州汽车有限公司； 3.重庆大学机械传动国家重点实验室
  • 发布日期:2026-04-08
  • 通讯作者: 丁磊
  • 作者简介:丁磊（1986－），男，硕士，高级工程师，研究方向为新能源动力系统优化与控制
  • 基金资助:
    广西高校中青年教师科研基础能力提升项目（2025KY1940）

Brake Energy Recovery Optimization Method for Pure Electric Vehicles Based on Operating Condition Recognition

DING Lei1 , DONG Zhihui1 , WU Song2 , HAN Lixing1 , CAI Jianjun3   

  1. 1.Liuzhou City College of Vocation and Technology； 2.Dongfeng Liuzhou Motor Company Limited; 3.State Key Laboratory of Mechanical Transmissions, Chongqing University
  • Published:2026-04-08
  • Contact: DING Lei

PDF

摘要： 纯电动汽车面临续航里程焦虑问题，如何提升电池能量使用效率成为研究热点。现有 纯电动汽车制动能量回收策略在实际运行时受到行驶工况的影响，难以获得最大的回收效率， 研究提出一种基于工况识别的纯电动汽车制动能量回收控制策略，优化车辆在实际道路行驶 环境下的能量回收效果。该方法利用 K-means 聚类算法，对实车制动能量回收时的行驶状态 进行分类，利用机器学习算法对能量回收不足的行驶状态进行识别训练，建立车辆运行工况 识别模型。针对回收效果不佳的行驶工况，设计改进控制策略，并建立在线执行模块，对车 辆控制策略实时切换。结果表明，所提策略可以准确识别制动效果较差工况，通过切换优化 策略提高制动能量回收效率，对比原制动能量回收方案续航里程提升 3%～4%。

关键词: 纯电动汽车；制动能量回收；工况识别；机器学习

Abstract: Pure electric vehicles face the problem of range anxiety, and how to improve battery energy efficiency has become a research hotspot. In response to the shortcomings of existing pure electric vehicle braking energy recovery strategies in actual operation, this study proposes a pure electric vehicle braking energy recovery control strategy based on condition recognition to optimize the energy recovery effect of pure electric vehicles in actual road driving environments. This method uses K-means clustering algorithm to classify the driving state of the actual vehicle during braking energy recovery, and uses machine learning algorithm to identify and train the driving state with insufficient energy recovery, establishing a vehicle operating condition recognition model. Design improved control strategies for driving conditions with poor recycling efficiency, establish an online execution module, and switch vehicle control strategies in real-time. The results show that this method can accurately identify working conditions with poor braking performance, improve the efficiency of braking energy recovery by switching optimization strategies, and increase the range by 3%~4% compared to the original braking energy recovery scheme.

Key words: pure electric vehicles; braking energy recovery; driving condition recognition; machine learning