Automation of Tenanted Arches NDT Inspections using Robotics and Machine Learning

这一创新项目旨在自动化检测包层租户拱门，而无需移除包层，并检测拱门中的次表面缺陷，这是目前使用现有方法无法实现的。该项目将侧重于使用机器人设备结合无损检测（NDT）技术，如探地雷达（GPR）和X射线后向散射（XBS），以建立一个准确的三维图片的整体健康的租户拱门。这一建议是一个延续的成功活动之间的MTC和网络铁路，租户拱门项目流，其中，GPR和XBS技术被用于手动扫描大伦敦内的包覆租户拱门，快速且无需移除或损坏任何包覆层。在项目的最后阶段，可以手动收集所需数据，然后快速生成有关覆层材料后面所有相关缺陷的大小和位置的人类可读信息。到目前为止，所有的检查都是由操作员手动完成的，因此，本提案将集中于完全自动化的过程，以便更快和更安全。拟议项目包括四个主要活动：为机器人设备定义自动化平台，以在不去除覆层的情况下检查租户拱门，增强数据处理技术以提高NDT扫描中的图像质量，在3D空间中部署用于自动缺陷检测和可视化的机器学习算法，并进行概念验证演示，展示系统和传感器的集成。通过自动化检查，增强数据分析和利用先进技术，该项目旨在简化租户拱门检查，提高收集信息的质量和数量，并改善预测性维护计划。最终，该项目将为铁路基础设施中的包层租户拱提供更安全、更快速、更可靠的检测解决方案。

This innovative project that aims to automate the inspection of cladded tenanted arches without the need to remove the cladding, and to also detect sub-surface defects in the arches, which is not currently possible with current methods. The project will focus on the use of robotics devices coupled with Non-Destructive Testing (NDT) techniques such as Ground Penetrating Radar (GPR) and X-ray Backscatter (XBS) to build an accurate and 3D picture of the overall health of the tenanted arches.This proposal is a continuation of the successful activities between the MTC and Network Rail, the Tenanted Arches Project stream, in which GPR and XBS technologies were utilised to manually scan cladded tenanted arches within Greater London, quickly and without the need to remove or damage any cladding. In the latest stages of the project, it was possible to manually collect the required data and then quickly produce human readable information about the size and location of all relevant defects behind the cladding material. All the inspection thus far has been completed manually by operators, so this proposal will focus on fully automating the process such that it is faster and safer.The proposed project involves four main activities: defining an automation platform for robotics devices to inspect tenanted arches without cladding removal, enhancing data processing techniques to improve image quality in the NDT scans, deploying machine learning algorithms for automated defect detection and visualisation in 3D space, and conducting a proof-of-concept demonstration to showcase the integration of systems and sensors. By automating inspections, enhancing data analysis, and leveraging advanced technologies, the project aims to streamline tenanted arch inspections, increase the quality and quantity of collected information, and improve predictive maintenance scheduling. Ultimately, the project will provide a safer, faster, and more reliable inspection solution for cladded tenanted arches in railway infrastructure.