欢迎访问《汽车安全与节能学报》,

汽车安全与节能学报 ›› 2023, Vol. 14 ›› Issue (3): 299-309.DOI: 10.3969/j.issn.1674-8484.2023.03.005

• 汽车安全 • 上一篇    下一篇

重型载货汽车长下坡制动工况辨识

史培龙1(), 高艺鹏2, 张子豪1, 赵轩1, 余强1   

  1. 1.长安大学 汽车学院,西安 710064,中国
    2.北京理工大学 机械与车辆学院,北京 100081,中国
  • 收稿日期:2022-09-27 修回日期:2023-04-24 出版日期:2023-06-30 发布日期:2023-06-30
  • 作者简介:史培龙 (1984—),男(汉),陕西,副教授。E-mail:peilongshi@chd.edu.cn
  • 基金资助:
    国家自然科学基金面上项目(52172361);榆林市科技计划(CXY-2020-021);中央高校基本科研业务费专项资金(300102222201)

Identification of braking condition for heavy truck on long downhill

SHI Peilong1(), GAO Yipeng2, ZHANG Zihao1, ZHAO Xuan1, YU Qiang1   

  1. 1. School of Automobile, Chang’an University, Xi’an 710064, China
    2. School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081, China
  • Received:2022-09-27 Revised:2023-04-24 Online:2023-06-30 Published:2023-06-30

摘要:

传统持续制动系统过度依靠驾驶人主观判断实现开启和关闭,易出现不当操作引起整车制动性能热衰退问题,为此该文提出了重型载货汽车长下坡制动工况构建及辨识方法,为持续制动系统介入或退出主动控制提供依据。以重型载货汽车长下坡试验数据为基础,包含制动踏板动作、车速及GPS数据,利用短行程划分、K聚类和编码技术,基于Markov- Monte Carlo方法构建了长下坡制动工况,总时长1 194 s,路程21.18 km;基于滚动时间窗原理建立反向传播(BP)神经网络工况辨识模型并进行离线训练和识别验证,结果表明:一般制动和强制动工况识别准确率达到89.30%,显示出提出的重型载货汽车长下坡制动工况构建及辨识方法能够有效识别车辆制动状态。

关键词: 汽车工程, 重型载货汽车, 制动工况, Markov-Monte Carlo方法, 工况辨识, 滚动时间窗

Abstract:

The traditional continuous braking system excessively relies on the subjective judgment of the driver to open and close, which is prone to the problem of thermal decline of the whole vehicle’s braking performance caused by improper operation. Therefore, this paper proposed a method to construct and identify the braking condition of heavy truck on long downhill, which provided a basis for the continuous braking system to intervene or withdraw from active control. Based on the long downhill test data of heavy trucks, including brake pedal action, vehicle speed and GPS data, the long downhill braking conditions was constructed using short stroke division, K-clustering, coding techniques and the Markov-Monte Carlo method, where the total time was 1 194 s and the total distance was 21.18 km. Based on the principle of rolling time window, the Back Propagation(BP) neural network working condition identification model was established, and offline training and identification verification was carried out. The results show that the recognition accuracy of general braking and forced braking working conditions is 89.30%, proving that the proposed method can effectively identify the braking status of heavy trucks on long downhill.

Key words: automotive engineering, heavy truck, braking condition, Markov-Monte Carlo method, condition identification, rolling time window

中图分类号: