天津地铁10号线弓网在线监测分析系统数据集

一、数据采集流程：接触网悬挂状态检测装置、几何参数测量仪、燃弧、硬点检测装置等设备通过连续采集多个接触网运行关键监测点数据，包含不局限于定位点状态、弓网燃弧、导高拉出值、螺栓松动、异物等超100万原始数据、报警数据。将关键数据以时间序列统计、分析，研究不同缺陷类型与时间、位置之间的相关性。通过数据积累、交叉验证、报警检测等识别数值变化规律，提前给出设备故障预警，以便提前采取应对措施，减少故障发生率，快速定位故障点，延长设备使用寿命。
为保障数据模型能够实现自动运算和预测预警功能，依据天津津铁供电公司维护经验制定了数据阀值，以阙值控制数据状态和有效性，经检测设备、巡检模型逻辑运算后，将运算结果和报警信息等存储至数据库中。部分数据阙值规则设置详建数据知识产权登记申请书

1. Data Collection Process:
Overhead contact line (OCL) suspension state detection devices, geometric parameter measuring instruments, arc discharge detection devices, hard spot detection devices and other equipment continuously collect data from multiple key monitoring points during OCL operation. The collected data includes over 1 million pieces of original and alarm data covering but not limited to positioning point status, pantograph-catenary arc discharge, catenary height and stagger value, bolt looseness, foreign objects and other items. Key data is statistically analyzed in time series to study the correlation between different defect types, time and location. Through data accumulation, cross-validation, alarm detection and other methods, the variation rules of numerical values are identified, and early warnings of equipment faults are issued in advance, so as to take preventive measures ahead of schedule, reduce the fault incidence rate, quickly locate fault points, and extend the service life of equipment.

To ensure that the data model can realize automatic operation and predictive early warning functions, data thresholds are formulated based on the maintenance experience of Tianjin Jintie Power Supply Company, which are used to control the status and validity of collected data. After logical operations via detection equipment and inspection models, the operation results, alarm information and other related data are stored in the database. For the setting rules of some data thresholds, please refer to the Data Intellectual Property Registration Application Document for details.

搜集汇总

背景与挑战

背景概述

该数据集由天津津铁供电有限公司提供，包含31592条记录，每日更新，用于监测天津地铁10号线接触网设备的运行状态。数据集涵盖缺陷等级、告警类型等21个字段，通过在线监测和算法预警，解决人工巡检效率低和故障定位问题，提升供电设备维护的准确性和效率。