轨道交通车辆轮轴固死故障预测模型研究
洪旭陈美霞滑瑾
Research on Rail Transit Vehicle Wheel Axle Locking Fault Prediction Model
HONG XuCHEN MeixiaHUA Jin
-
作者信息:中车南京浦镇车辆有限公司， 210031， 南京
-
Affiliation:CRRC Nanjing Puzhen Co., Ltd., 210031, Nanjing, China
-
关键词:
-
Key words:
-
DOI:10.16037/j.1007-869x.2024.05.035
-
中图分类号/CLCN:U260.33
-
栏目/Col:研究报告
摘要:
［目的］由于轮轴组装不当、检修时操作过激使轮轴出现凹凸面或作业时电腐蚀造成擦伤缺损等情况，均会加剧轨道交通车辆轮轴热轴、轮轴固死等异常故障的发生。如不及时处理，可能造成严重的行车事故。因此，有必要研究轨道交通车辆轮轴固死故障预测方法及预防措施。［方法］以南京地铁S7号线列车出现轮轴固死故障为例，通过对列车监测数据的分析和挖掘，探究其潜在的相关性和周期性，为列车故障诊断提供依据；基于牵引模式，结合实时牵引电流、各架轴速、参考速度等信号进行大数据分析，建立轮轴固死故障预测模型；对轴承卡滞、轴内阻力增大及轴系老化导致性能下降等原因引起的轮轴固死现象进行判断，辨识其健康状态，实现对轮轴内部可能出现的一些老化、卡滞现象进行故障预测。［结果及结论］建立了轨道交通车辆轮轴固死故障预测模型；对列车实际运行数据进行分析，验证了该预测模型的可行性。但由于轮轴固死的样本数据很少，后续还需采用历史值结合新增值的方法进行在线学习，来提高模型预测的准确性。
Abstracts:
［Objective］ Improper assembly of wheel axles, excessive operation during maintenance leading to axle surface irregularities, or abrasions and defects caused by electrical corrosion during operation exacerbate abnormal faults such as wheel axle overheating and locking in rail transit vehicles. If left unattended, these faults may result in serious accidents. Therefore, it is necessary to study predictive methods and preventive measures for wheel axle locking faults in rail transit vehicles. ［Method］ Taking the occurrence of wheel axle locking faults in Nanjing Metro S7 Line trains as an example, analysis and mining of train monitoring data are conducted to explore their potential correlations and periodicities, providing a basis for train fault diagnosis. Based on traction mode, big data analysis is performed using real-time traction current, axle speeds, reference speed signals to establish a prediction model for wheel axle locking faults. Assessment of wheel axle locking phenomena caused by bearing seizure, increased internal resistance, and aging of the axle system is carried out to identify their health status and predict potential aging or seizure within the wheel axle. ［Result & Conclusion］ A prediction model for wheel axle locking faults in rail transit vehicles is established and validated through analysis of actual train operation data. However, due to limited sample data on wheel axle locking, subsequent online learning methods combining historical values with new data are required to improve the accuracy of the prediction model.
- 上一篇: 地铁车辆可清洁式均匀送风道数值仿真分析
- 下一篇: 城市轨道交通ATO与TCMS功能融合设计
过刊浏览
资料下载
友情链接