摘要
为车辆管理人员及时制定维修策略提供依据,该文设计了基于迭代学习的地铁车辆速度传感器故障识别方法。依据霍尔速度传感器在地铁车辆制动系统运行时的输出波形,构建故障状态时速度传感器的线性时不变系统。利用迭代学习算法的迭代过程,依据线性时不变系统的理论输出值与速度传感器的实时采集值,利用跟踪器迭代运算传感器残差值,获取传感器故障信号;将传感器故障信号输入学习过程的决策逻辑识别模型,决策逻辑识别模型利用加权KNN分类器,输出地铁车辆速度传感器故障类型,实现地铁车辆速度传感器故障识别。实验结果表明,该方法可以精准识别地铁车辆速度传感器的虚接故障、断线故障等不同类型故障,故障识别准确率高。
The fault identification method of subway vehicle speed sensor based on iterative learning is studied to provide a basis for vehicle managers to formulate maintenance strategies in time.According to the output waveform of the Hall speed sensor when the subway vehicle braking system is running,a linear time-invariant system of the speed sensor is constructed in the fault state.Using the iterative process of the iterative learning algorithm,according to the theoretical output value of the linear time-invariant system and the real-time acquisition value of the speed sensor,use the tracker to iteratively calculate the sensor residual value to obtain the sensor fault signal.Input the sensor fault signal into the decision-making process of the learning process.The logic recognition model and decision logic recognition model use the weighted KNN classifier to output the fault type of the subway vehicle speed sensor,so as to realize the fault recognition of the subway vehicle speed sensor.The experimental results show that the method can accurately identify different types of faults such as virtual connection faults and disconnection faults of subway vehicle speed sensors,and the fault identification accuracy is high.
作者
刘尧
LIU Yao(The First Operation Company Branch of Beijing Subway Operation Limited Company,Beijing 102200,China)
出处
《自动化与仪表》
2023年第4期107-111,共5页
Automation & Instrumentation
基金
北京市重点科技项目(G20211381)。
关键词
迭代学习
地铁车辆
速度传感器
故障识别
跟踪器
残差值
iterative learning
subway vehicle
speed sensor
fault identification
tracker
residual value
