The railway signal intelligent operation and maintenance system with cloud computing at the center faces the problems of excessive load, high energy consumption, and insufficient real-time performance. Therefore, a solution for railway signal intelligent operation and maintenance platforms based on edge-cloud collaboration was proposed. After elaborating on the system architecture based on edge-cloud collaboration, this paper designed an intelligent operation and maintenance platform composed of data collaboration layer, edge computing layer, and cloud computing center layer to realize real-time scheduling and decision-making of operation and maintenance tasks. By designing the computing offload model based on reinforcement learning, some computing tasks of operation and maintenance data were offloaded to edge nodes. This can alleviate the computing pressure of the cloud computing center, reduce system delay and energy consumption, and improve the energy efficiency of the system. Simulation experiments show that the optimal computing offloading strategy obtained through reinforcement learning features lower time delay and energy consumption than deploying computing tasks on the cloud or edge alone. This study is of significance to improve the railway signal intelligent operation and maintenance and enhance energy efficiency in this field.

针对传感器水质监测所获取的大量数据,在处理设施工况稳定、外部影响因素及发生影响概率基本不变的条件下,通过筛选代表性、可信度高的数据作为样本,建立BP神经网络模型获取水质指标数据的拟合曲线,可用于传感器自动监测数据质量的智能辨识,对非故障原因导致的测量偏差、原水水质和设施状况异常提供预警.该方法可以保证水质自动监测数据质量,实现无人值守水处理设施精准维护,降低现场工作人力成本.

为预防铁路安全事件的发生,设计ZPW-2000轨道电路故障诊断系统的总体结构,明确4大组成部分,包括数据预处理、数据分析、数据服务及数据应用.首先,通过改进的旋转门转换(SDT)算法对轨道电路电气特性数据进行数据压缩;其次,分段线性拟合轨道电路不同状态下的模拟量数据,计算得到特征值;最后,通过密度聚类的特征提取方法诊断轨道电路故障,并通过系统验证9种常见轨道电路故障.结果表明:改进的SDT算法可以有效对轨道电路电气特性数据进行数据压缩,压缩后的数据经过分段拟合后可以有效提取特征值,应用密度聚类算法可以生成诊断模型,提升轨道电路故障诊断的准确性,从而提高信号设备维护效率和维护水平.

列控系统态势感知是指对可能引起列控系统信息安全态势发生变化的态势要素进行获取、理解、评价以及预测的过程.提出一种基于SVD Entropy和SVM联合算法的列控系统态势感知技术,以识别信息安全风险,避免列控系统安全事故发生.该联合算法能够处理列控系统多源异构数据集,通过设定奇异值熵阈值,完成多源异构数据的降维与特征提取;再结合支持向量机多分类算法,实现数据的分类与融合.仿真实验表明,当奇异值熵阈值设定为0.85时,在保持分类精度基本不变的前提下,SVD Entropy和SVM联合算法能有效压缩系统运算时间,具有较高的计算实时性.该联合算法为后续实时在线处理数据,完成态势评价、预测提供了理论支持.

以数据为驱动的高速铁路信号智能运维技术,是保证列车安全正点运行的重要保障.在阐述数据源、数据汇集、数据存储、数据分析、数据应用、数据展示等系统架构设计的基础上,结合数据融合、道岔智能诊断、轨道电压曲线智能分析和故障预测与健康管理等关键技术,提出设备综合监测、全生命周期管理、智能故障诊断、综合运维分析、生产作业管理、应急调度指挥、设备PHM、车地一体化分析等8种系统功能设计,可以为提升高速铁路信号运维水平,转变电务运维方式提供参考.

我国铁路信号监测技术三十多年的发展历程,大体上可以分为初步探索、逐步规范和深化应用 3个阶段,先后研制和应用了TJWX、TJWX-2000、TJWX-2006和CSM型等 4代产品.本文对每个发展阶段及每一代产品的技术特点进行了介绍,对信号监测未来的发展进行了展望;提出了信号监测将向着集中化、智能化、自动化和一体化方向发展;指出在智能高铁快速发展的今天,信号监测将发挥更大的作用.

城市轨道交通信号系统制式和国铁有很大的不同,为实现信号集中监测系统在城市轨道交通中的应用,对其系统结构、软件架构、监测技术等进行了探讨.

针对电务实操考核中经常出现的器材损坏、培训效率低、培训记录不全等问题,提出了一种电务实操智能考核系统设计方案.通过与电务段实训基地的配套使用,能够实现智能化考核、智能化演练,并能对考核和演练的全过程进行记录.该系统已应用在长春电务段实训基地,取得了很好的效果.

The reliability of data transmission plays significant role in reflecting the real-time train running status,monitoring operated situation of trackside equipment.It is vulnerable to electric traction,noise in terference and signal degradation and other factors.The reliability of data transmission needs to be improved.In this paper,we propose a novel method based on the strong error correcting ability of Low Density Parity Check (LDPC) code to increase the reliability of the data transmission of the trackside monito ring equipment.LDPC code is defined by parity check matrix,the error correcting ability is determined by the degree distribution which means the row (column) weight of the parity check matrix.A degree distribution optimization algorithm for regular LDPC codes based on extrinsic information transform chart is presented,and some good results are introduced.According to the characteristics and requirements of two different kinds of message types transmitted in the trackside monitoring,three kinds of LDPC codes are designed.The error correcting capability of the designed LDPC codes are investigated via simulations.Simulation results show that the bit error rate of the system decreases from 10-2 to 10-6,the reliability of the trackside monitoring system can be significantly improved as LDPC codes are employed.

传统冷链物流与仓储监控过程不够透明,货主和第三方物流服务商彼此信任度低.本文结合冷链物流理论,设计一个完整的低温食品冷链运输、仓储、配送流程,使用RFID温度标签和托盘共用系统,来实现冷链物流的高效运转,以保障食品在整个供应链中的安全性.

根据信号集中监测网现有的网络带宽和网络需求,分析在搭建电务段集中监测中心过程中,利用负载均衡技术(以F5-BIG-IP为例)建立网络和服务器平台,以解决在组建大规模集中监测网及重大故障应急处理过程中,面对大量用户并发访问监测平台可能引起的网络堵塞,服务中断等数据安全问题.

在区域联锁主控站,车务人员无法实时观察被控站道岔动作情况,这给行车带来安全隐患.为此设计了一套电流表远程监测系统以解决此问题,本文对系统的结构、系统功能及采用的关键技术进行了论述.

在设计分析信号集中监测系统时,主要对噪声的产生及如何降低噪声进行了研究.引入了系统设计的效率函数及相关系统的功率与传输信息位的函数关系,进而阐述了功率、热量、电磁干扰之间的内在联系.深入分析了信号集中监测系统底层设计的相关技术及注意事项,使监测系统底层硬件设计达到相对优化的程度.

随着物联网技术在铁路运输中应用深度和广度的不断增加和扩展,逐步促进了铁路运营模式的转变.在阐述物联网与铁路物流关系的基础上,从技术层面对物联网在铁路运输信息网络领域的应用进行研究,分析物联网技术在铁路货运集散中心的应用,并提出铁路运输物联网应用的建议.

To improve the accuracy of Trace Hotbox Detection System(THDS),the comprehensive forecast of hotbox is discussed in this paper,on the basis of forecast of 5T System implemented on Golmud-Lhasa section of Qinghai-Tibet Railway,related information of TFDS.The related solutions of TFDS information are also given in this paper,and the application effect is good.

物联网理论和技术为转变铁路经济增长方式提供了支撑。从网络经济和网络化物理系统入手,构建包含传感网络、运输网络、信息网络三个基本网络层次的铁路物联网,提出以集散中心和信息中心为网络节点、面向广大用户的物联网运作模式、组建铁路物联网运输联盟方式的实施策略。

The authors introduced the system structures,monitoring objectives and system functionality of the centralized monitoring system of Xinfengzhen marshalling yard signals.Meanwhile,the system features and new techniques were explained.

According to the maintenance requirements of signal equipment after the large-scale speed-up campaign,a set of intelligent analysis and fault diagnosis system were designed based on microcomputer monitoring.The design principle,key technologies,and realized functions of the system were given.

According to the characteristics of Qinghai-Tibet Railway and the way of its maintenance,the MMT system was developed to maintain and monitor ITCS system and also to record and display the signal monitor information with stations.Its design principle,key technologies applied in the system,and system function were given.

Through combining currently popular computer technology, a new cab signal inspection data analysis system based on B/S mode was developed and its design principle and implementation method were described.