基于探地雷达的高铁无碴轨道多层成像研究

Ballastless track is used as the main track type in the construction of the high-speed railway in China. Despite greatly enhancing the smoothness ,stability and safety of the line, ballastless track also has a higher demand on the health and stability of the structure. As the time of the large-scale use of ballastless track in China is still short, the detection of the disease is also lack of experience. Ground penetrating radar as a commonly used non-destructive testing (NDT) equipment, has a great potential in the ballastless track disease investigation. This project is aimed at the complexity of ballastless track and the lack of internal reinforcement, inter-layer gaps and mortar cavities. The three-layer dielectric model of the ballastless track, including concrete track slab, CA mortar layer and concrete support plate is established..A microwave imaging method based on ground penetrating radar (GPR) and full-waveform inversion (FWI) is proposed for the three-layer structure of ballastless track, in order to accurately conduct quantitative analysis on the interface and the target body (e.g. inter-layer gap) in real-time. .The project involves electromagnetic field theory, physical/mathematical modeling and machine learning algorithms and other interdisciplinary techniques. The following studies will be conducted:.1) Conduct theoretical analysis of the propagation of electromagnetic waves in multi-layer plate media. Build the multi-layer plane wave forward model of the track with dielectric constants and interlayer thicknesses..2) Study the full-waveform inversion optimization algorithm with the help of support vector machine (SVM);.3) Establish the migration imaging algorithm for the three-layer ballastless track with the dielectric constant model obtained by full-waveform inversion.

无碴轨道是我国高速铁路建设使用的主要轨道类型。由于我国高铁无碴轨道大规模投入使用的时间尚短，在病害的检测上也还缺乏经验。探地雷达作为一种常用的无损检测设备，在无碴轨道的病害检测上具有非常大的潜在使用价值。本项目针对无碴轨道构成的复杂性和内部容易出现的钢筋缺失、层间离缝和砂浆层空洞等问题建立无碴轨道的混凝土轨道板、CA砂浆层和混凝土基座三层结构的介电模型。提出基于探地雷达和全波反演法的高铁无碴轨道三层结构的微波成像，用于准确、实时地进行分层界面和目标体（如层间空隙）的定位定量分析。项目涉及电磁场理论、物理数学建模以及机器学习算法等多类跨学科技术，将主要进行以下研究：1）对电磁波在多层平板介质的传播进行理论分析，建立关于介电常数和层间厚度的路基多层平面波正演模型；2）结合支持向量机，研究全波反演优化算法；3）基于全波反演获取的介电常数模型建立适用于无碴轨道三层结构的偏移成像算法。

无碴轨道是我国高速铁路建设使用的主要轨道类型。由于我国高铁无碴轨道大规模投入使用的时间尚短，在病害的检测上也还缺乏经验。探地雷达作为一种常用的无损检测设备，在无碴轨道的病害检测上具有非常大的潜在使用价值。本项目针对无碴轨道构成的复杂性和内部容易出现的钢筋缺失、层间离缝和砂浆层空洞等问题建立无碴轨道的混凝土轨道板、CA砂浆层和混凝土基座三层结构的介电模型。提出基于探地雷达和全波反演法的高铁无碴轨道三层结构的微波成像，用于准确、实时地进行分层界面和目标体（如层间空隙）的定位定量分析。具体研究内容包括：1）对电磁波在多层平板介质的传播进行理论分析，建立关于介电常数和层间厚度的路基多层平面波正演模型；2）基于全波反演获取的介电常数模型，结合SAR（合成孔径雷达）技术，建立适用于无碴轨道扫描图像的聚焦成像算法；3）利用主成分分析法，研究缺陷扫描图像的特征提取和分析方法。本项目的研究成果为无碴轨道乃至混凝土建筑的缺陷检测提供了一种直观而高效的手段。