上海轨道交通通信信号系统电子设备多元维修策略

Multi-faceted Maintenance Strategy for Electronic Devices in Shanghai Rail Transit Communication and Signaling System

[目的]上海轨道交通电子设备故障率逐渐上升,传统的自主维修速度慢、返厂维修价格昂贵、周期长,需要大量的库存备件,造成了维护资源的严重浪费。对此,提出一种由快速检测、深度维修、预测维修所组成的电子设备多元维修策略。[方法]通过搭建综合性电源模块检测平台,快速检测电源模块,实现故障的精准定位,从而提高维修效率。针对老旧的显示单元进行深度维修,用LED(发光二极管)灯条进行改造,优化功耗性能,并进一步延长使用寿命。利用深度学习技术,基于电子设备的基础信息、日常检修数据、运行监测数据、历史故障等信息搭建预测剩余使用寿命的模型,进行预测维修,科学、合理地动态调整检修维护周期,解决传统预防维修模式中采用固定检修周期的问题,避免了过度维修或欠维修情况的发生。[结果及结论]在现场实际应用中,自主设计的综合性电源模块检测平台能够快速精准定位故障点,维修效率大幅提升了83%;LED灯条改造后,TOD(司机驾驶显示单元)运行功耗下降了30%,显示效果提升显著,并延长了使用寿命。在枪机摄像头的剩余使用寿命预测准确率达到了96.4%。预测维修能够精准调控检修周期,使运维成本下降了53%,使故障率下降了64%,成效显著。

[Objective]As the failure rate of electronic devices in Shanghai rail transit system keeps on increasing,traditional maintenance approaches—characterized by slow speed of self-repair,high costs of factory repair,long repair cycle,and substantial inventory requirements—result in significant waste of maintenance resources.To address these issues,an electronic device based multi-faceted maintenance strategy consisting of rapid detection,in-depth repair,and predictive maintenance is proposed.[Method]A comprehensive power module testing platform is established to enable rapid detection on power modules and precise fault localization,thereby enhancing repair efficiency.For outdated display units,in-depth repairs are performed,with LED new technology for replacement and upgrade to optimize power consumption and extend service life.Predictive maintenance is carried out using deep learning techniques to build models predicting the remaining service life based on electronic device fundamental information,daily maintenance data,operation monitoring data,and historical fault.Through conducting scientific and reasonable dynamic adjustments of maintenance cycles,the issues of adopting fixed repair cycle in conventional preventive maintenance mode are addressed,avoiding excessive repair or maintenance insufficiency.[Result&Conclusion]In practical applications,the self-designed comprehensive power module testing platform significantly improves fault localization accuracy and increases repair efficiency by 83%.After retrofitting the TOD(train operator display)units with LED light strips,the power consumption is decreased by 30%,significantly improving the display quality and extending the service life.In the remaining service life prediction using bullet camera,the accuracy rate reaches 96.4%.The predicative maintenance enables precise adjustment of maintenance cycles,resulting in a 53%reduction in operation-maintenance costs and a 64%decrease in fault rates,demonstrating substantial effectiveness.