轨道结构隐蔽性病害多源智能检测装备研发与融合辨识技术数据集

形成了多源检测数据集，具体涵盖以下类型数据：在轨道结构隐蔽病害多源波场响应特性方面，采用有限元方法计算了CRTSII型无砟轨道结构弹性波传播和超声波传播数据，为隐蔽病害波场响应特性分析提供了数据基础。在隐蔽病害多源融合识别方面，制作了实体模型，采用加速传感器、SIRIUS采集仪和MIRA1040超声检测仪采集了不同类型和尺寸隐蔽病害的弹性回波和超声回波数据，根据识别结果和缺陷预设参数，构建了不同病害的检测置信度矩阵，为融合识别方法的估计结果关联度矩阵提供数据基础，为多源融合识别方法提供检测数据；在轨道结构隐蔽病害多源检测装备方面，在CRTSII板式无砟轨道足尺模型采用轨道结构多源隐蔽病害检测装备开展现场检测，包括了检测区域的全部弹性回波数据和13个预设隐蔽缺陷区域的超声阵列回波数据，预设隐蔽缺陷类型涵盖了四种典型的隐蔽缺陷，包括不同尺寸的功能层脱空、功能层上离缝、功能层下离缝和底座板空洞（质量劣化），为研发的多源检测装备的检测指标评定提供检测数据基础。
数据类型：csv/txt/xlsx。
采集方案：通过查阅国内外弹性波/超声检测无砟轨道隐蔽病害的文献和标准，设计了弹性波/超声多源阵列检测方案，征求了国内行业领域内专家意见并通过了专家组评审。以CRTSII型板无砟轨道结构为对象建立了精细有限元数值模型，采用Python自编程序和ABAQUS/ANSYS有限元软件对不同工况下轨道结构多源波场传播开展计算，采用ABAQUS/ANSYS软件进行波场数据采集。以石家庄铁道大学无砟轨道结构足尺模型为检测对象，以弹性波阵列引导的所有疑似存在隐蔽病害的区域为检测区域，采用自主开发的超声阵列检测装置（包括超声阵列全矩阵采集仪和16个干点耦合式超声波传感器组成的检测阵列）对检测区域开展检测阵列超声波检测，激励源主频为50KHz，采样频率为1MHz，在检测区域从左至右依次采集轨道结构内部的超声回波数据。
采集地点：石家庄
采集时间：2022年5月－2024年5月
设备情况：高性能计算机/加速度传感器/SIRIUS采集仪/MIRA1040超声检测仪/自主开发的轨道结构隐蔽病害多源检测装备。

A multi-source detection dataset is developed, which specifically covers the following types of data:
Regarding the multi-source wavefield response characteristics of hidden defects in track structures, the finite element method is used to calculate the elastic wave propagation and ultrasonic wave propagation data of CRTS II slab ballastless track structures, providing a data basis for the analysis of wavefield response characteristics of hidden defects.
For multi-source fusion recognition of hidden defects, a physical model is fabricated. Accelerometers, SIRIUS data acquisition instruments and MIRA1040 ultrasonic detectors are used to collect elastic echo and ultrasonic echo data of hidden defects with different types and sizes. Based on the recognition results and preset defect parameters, a detection confidence matrix for different defects is constructed, which provides a data basis for the correlation matrix of estimation results of fusion recognition methods, and provides detection data for multi-source fusion recognition methods;
Regarding multi-source detection equipment for hidden defects in track structures, on-site detection is carried out on the full-scale CRTS II slab ballastless track model using the self-developed multi-source hidden defect detection equipment for track structures. The data includes all elastic echo data of the detection area and ultrasonic array echo data from 13 preset hidden defect areas. The preset hidden defect types cover four typical hidden defects: delamination of functional layers with different sizes, upper gaps of functional layers, lower gaps of functional layers, and voids in base slabs (quality degradation). This provides a data basis for the evaluation of detection indicators of the developed multi-source detection equipment.

Data types: csv/txt/xlsx.
Collection plan: By reviewing domestic and foreign literature and standards on elastic wave/ultrasonic testing of hidden defects in ballastless tracks, a multi-source array detection scheme for elastic wave/ultrasonic testing is designed. Opinions from experts in the domestic industry are solicited and the scheme passes the review of the expert group. A refined finite element numerical model is established based on the CRTS II slab ballastless track structure. Self-developed Python programs and finite element software ABAQUS/ANSYS are used to calculate the multi-source wavefield propagation of the track structure under different working conditions, and ABAQUS/ANSYS software is used to collect wavefield data. The full-scale ballastless track structure model of Shijiazhuang Tiedao University is taken as the detection object, and all areas suspected of having hidden defects guided by the elastic wave array are taken as the detection areas. The self-developed ultrasonic array detection device (including an ultrasonic array full matrix acquisition instrument and a detection array composed of 16 dry-point coupled ultrasonic sensors) is used to carry out ultrasonic array detection in the detection areas. The dominant frequency of the excitation source is 50kHz, the sampling frequency is 1MHz, and the ultrasonic echo data inside the track structure is collected sequentially from left to right in the detection areas.
Collection location: Shijiazhuang
Collection time: May 2022 – May 2024
Equipment: High-performance computer/accelerometer/SIRIUS data acquisition instrument/MIRA1040 ultrasonic detector/self-developed multi-source detection equipment for hidden defects in track structures.

搜集汇总

背景与挑战

背景概述

该数据集由石家庄铁道大学发布，专注于铁路轨道结构隐蔽性病害的智能检测与识别。它包含多源检测数据，如弹性波和超声波传播数据，以及现场采集的回波数据，用于支持病害响应特性分析、多源融合识别方法研究和检测装备性能评估。数据采集于2022年至2024年，涵盖多种格式，总数据量约290.87MB，源自国家重点研发计划项目。