最新刊期

      DENG Qian, YU Hangyi

      摘要：To address the challenge of health state identification for lithium titanate (LTO) traction batteries in rail transit under complex operating conditions, significant seasonal variations, and scarcity of fault samples, this paper proposes an unsupervised health state identification method based on operating point-cloud density modeling. Leveraging the structural characteristics of LTO battery modules and cells, the method maps onboard operational monitoring data into operating point clouds within a feature space, thereby constructing a healthy operating region through neighborhood density characterization of normal operational states. Low-density samples deviating from this healthy region are used to represent potential anomalous states. To address the parameter sensitivity issue inherent in single-scale density models, a multi-scale neighborhood density consistency analysis is introduced, enabling continuous quantification of battery anomaly severity. Experiments are conducted using actual LTO traction battery operational data from rail transit vehicles, wherein a data pipeline comprising an operational data store (ODS), a data warehouse detail (DWD) layer, and a data warehouse service (DWS) layer is constructed and deployed onboard. Experimental results demonstrate that, without requiring fault sample annotation, the proposed method can reliably identify the health state of traction batteries, reduce the risk of false anomaly detection caused by seasonal variations, and exhibit strong engineering feasibility under constrained onboard computational resources, thereby providing a scalable technical approach for traction battery health management in rail transit applications.  
      关键词：rail transit;lithium titanate power battery;health state identification;operating point cloud;density-based clustering;unsupervised learning;onboard deployment   
      0
      |
      1
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 157654467 false
      更新时间：2026-06-03

      HOU Yuxuan, HUANG Zhihui, CHEN Xuejing, YAN Bohan, Tian Mingjie

      摘要：To improve the lateral stability performance of 25T passenger coaches during operation, a simulation analysis was conducted based on the Linear Quadratic Regulator (LQR) strategy of secondary lateral dampers. The study examined the impact of LQR control on the vehicle's lateral stability performance at different speeds and the influence of time delays on LQR control effectiveness. Additionally, adaptive compensation control and Smith predictor compensation control were employed to mitigate system time delays, with a comparative analysis of their compensation effects. The simulation results demonstrated that LQR control effectively enhances the lateral stability performance of the vehicle while having minimal impact on its operational stability and curve negotiation capability. Both Smith predictor compensation control and adaptive compensation control can effectively reduce the influence of time delays on LQR control. However, compared to Smith predictor compensation, adaptive compensation control does not require precise knowledge of the time delay magnitude, offering nearly equivalent performance with greater stability. Consequently, it exhibits superior overall practicality over Smith predictor compensation control.  
      关键词：semi-active control;dynamic performance;time delay;time delay compensation;estimated compensation   
      6
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 157467162 false
      更新时间：2026-05-29

      SHI Yixuan, MAO Qingzhou, DAI Huanyun, WANG Qunsheng, SHI Xiangyu

      DOI：10.13890/j.issn.1000-128X.2024.01.236
      摘要：The safety and reliability of traction inverters are critical to the stable operation of high-speed EMU vehicles. To judge the working state of the inverter, combined with the vehicle system dynamics, a fault diagnosis method of traction inverter based on motor vibration signal analysis was proposed. In this method, the motor vibration signal is taken as the monitoring object, and the characteristic signal of abnormal motor vibration caused by current distortion when the inverter is faulty is used to judge. Taking the traction transmission system of CRH2 EMU as an example, the main circuit simulation model of a three-level traction inverter based on the SVPWM strategy was established. Taking the power switch elements as the object, the structural fault modes of the inverter are simulated. The results indicate that inverter faults will significantly affect the harmonic content of the output current on the AC side, especially with a significant increase in the 5th, 7th, 11th, and 13th harmonic content; After these harmonics are input to the traction motor, they are converted into 6P(1-s) and 12P(1-s) frequency doubling pulsating torque that acts on the motor, causing its output vibration to contain corresponding harmonic components of the frequency. By analyzing the vibration signals of the traction motor during the actual operation of high-speed multiple units, it was found that under normal circumstances, there is no obvious presence of such harmonic frequency components; When there is an inverter fault, it will cause the frequency content to be very obvious. Therefore, the vibration signal analysis method proposed in this article can effectively monitor and diagnose the working status of traction inverters and has a certain engineering application value.  
      关键词：Electric multiple units;traction inverter;motor vibration;current harmonics;fault diagnosis   
      469
      |
      1068
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 56193450 false
      更新时间：2024-05-08
    0