[1A1] Transforming railway structure inspections using robotic-coupled non-destructive testing and automated analysis
H Murtaza, T Hovell, N Pietrow, A Birch and M Begg
The Manufacturing Technology Centre, UK
Comprehensive assessment of structural defects within railway infrastructure, such as tenanted railway viaduct arches, plays a crucial role in maintaining the safety, reliability and efficiency of transportation systems. These structures are currently examined using rudimentary manual techniques to examine for surface and subsurface defects, which are slow, subjective and expose engineers to hazardous conditions for extended periods.
This study presents an automated non-destructive testing (NDT) platform that integrates a bespoke wall-crawling robot with X-ray backscatter (XBS) and ground-penetrating radar (GPR) to inspect arches without removing corrugated cladding. A data processing framework was developed, incorporating a machine learning model for XBS data and computer vision models for GPR datasets, to enable automated defect detection, classification and reporting. Additionally, a 3D mapping framework was created to link surface and subsurface defects, such as mortar degradation, spalling, voids, fractures, calcite deposition, missing bricks and voiding, to arch geometry for virtual visualisation and maintenance planning.
On-site trials in tenanted arches demonstrated the capability of the system to improve inspection speed, reliability, digital traceability and examiner safety. This work advances NDT by enabling novel non-invasive inspection of cladded infrastructure, offering practical insights for railway asset management. These findings enhance safety and longevity of railway arches through scalable, data-driven maintenance strategies.
This study presents an automated non-destructive testing (NDT) platform that integrates a bespoke wall-crawling robot with X-ray backscatter (XBS) and ground-penetrating radar (GPR) to inspect arches without removing corrugated cladding. A data processing framework was developed, incorporating a machine learning model for XBS data and computer vision models for GPR datasets, to enable automated defect detection, classification and reporting. Additionally, a 3D mapping framework was created to link surface and subsurface defects, such as mortar degradation, spalling, voids, fractures, calcite deposition, missing bricks and voiding, to arch geometry for virtual visualisation and maintenance planning.
On-site trials in tenanted arches demonstrated the capability of the system to improve inspection speed, reliability, digital traceability and examiner safety. This work advances NDT by enabling novel non-invasive inspection of cladded infrastructure, offering practical insights for railway asset management. These findings enhance safety and longevity of railway arches through scalable, data-driven maintenance strategies.
