高铁基础设施隐蔽病害数据库及知识图谱数据集

高铁作为现代交通运输体系的重要组成部分，其安全稳定运行关乎国计民生。然而，高铁基础设施在长期服役过程中，受环境、荷载等多因素耦合作用，易产生各类隐蔽缺陷，严重威胁运营安全。为深入揭示高铁基础设施隐蔽病害的孕育机理与演化规律，并最大化病害数据的应用价值，本数据集采用多源协同采集体系进行构建，集成轨道检测车、高清摄像设备、GPS/北斗定位系统及便携式检测仪等专业设备，结合铁科院与各路局既有病害台账系统，通过自动检测、人工巡检与系统接口对接相结合的方式，实现病害数据的高效、全面获取。为保证数据真实有效，所有采集设备均按国家计量标准定期校准，严格执行标准化现场采集流程与数据校验机制，并通过第三方测试与专家评审对数据质量进行验证，确保数据来源可靠、格式规范、内容准确。在数据处理与分析过程中，参考行业已有研究及规范标准，对多源异构数据进行清洗、整合与结构化处理，并利用统计分析和数据挖掘方法进行多次交叉核验，排除异常值与偶然误差，确保数据集的科学性、一致性与可用性。在此基础上，进一步构建了高铁基础设施隐蔽病害知识图谱，共包含6374个节点、13283条关系，数据量为21.6MB。

As an important component of the modern transportation system, high-speed railway (HSR) plays a critical role in national economy and people's livelihood. During long-term service, however, HSR infrastructure is prone to various hidden defects under the coupled effects of multiple factors such as environment and loads, which seriously threatens operational safety. To deeply reveal the incubation mechanism and evolution law of hidden defects in HSR infrastructure and maximize the application value of defect data, this dataset is constructed using a multi-source collaborative acquisition system. It integrates professional equipment including track inspection vehicles, high-definition imaging devices, GPS/BeiDou positioning systems and portable detectors, combines the existing defect ledger systems of China Academy of Railway Sciences (CARS) and various railway bureaus, and realizes efficient and comprehensive acquisition of defect data through the combination of automatic detection, manual inspection and system interface docking. To ensure the authenticity and validity of the data, all collection equipment is regularly calibrated in accordance with national metrological standards, standardized on-site collection procedures and data verification mechanisms are strictly implemented, and data quality is validated via third-party testing and expert review, ensuring reliable data sources, standardized formats and accurate content. During data processing and analysis, referring to existing industry researches and normative standards, multi-source heterogeneous data is cleaned, integrated and structured, and multiple cross-verifications are conducted through statistical analysis and data mining methods to eliminate outliers and accidental errors, ensuring the scientificity, consistency and availability of the dataset. On this basis, a knowledge graph of hidden defects for HSR infrastructure is further constructed, which contains a total of 6374 nodes and 13283 relationships, with a data volume of 21.6 MB.