Water leakage mapping in concrete railway tunnels using LiDAR generated point clouds

Abstract

Light detection and ranging (LiDAR) is a key non-destructive testing (NDT) method used in modern civil engineering inspections and commonly known for its ability to generate high-density coordinated point clouds of scanned environments. In addition to the coordinates of each point an intensity value, highly dependent on the backscattered energy of the laser beam, is recorded. This value has proven to vary largely for different material properties and surfaces. In this study properties such as surface colour, roughness and state of saturation are reviewed. Different coloured and concrete planar targets were scanned using a mobile LiDAR scanning system to investigate the effect distance, incidence angle and ambient lighting have on targets of differing properties. The study comprised controlled laboratory scans and field surveying of operational concrete railway tunnels. The aim of field tests was to automatically extract water leakage areas, visible on tunnel walls, based on the intensity information of points. Laboratory results showed that darker coloured targets resulted in a lower recorded intensity value and larger standard deviation of range. Black targets recorded the lowest intensities (0 - 4 units) with 50% higher standard deviations of range, on average, compared to all other coloured targets which recorded standard deviations of around 12 mm. The roughness of each coloured target showed to largely influence the recorded intensity, with smooth surfaces recording higher standard deviations of measurements. Concrete targets proved that a difference in roughness and saturation was detectable from intensity data. The biggest change was seen with saturated targets where a 70 to 80 % lower intensity value was recorded, on average, when compared to the same targets in their dry state. The difference in target roughness showed to have no effect on intensity when saturated. The laboratory data provided an important reference for the interpretation and filtering of field point clouds. Ambient lighting had no significant effect on all measurements for both the coloured and concrete targets. Field tests conducted on an operational concrete railway tunnel confirmed and demonstrated the ability to rapidly identify, extract and record areas of water leakage based on the intensity and spatial information of point cloud data. This is particularly useful as water ingress is known to degrade concrete, resulting in the earlier onset of corrosion, spalling and loss of strength. The mobile LiDAR scanning system used here proved capable of reducing survey time, which would allow for shorter interval revisits, while providing more quantitative information of the leakage areas. Long-term continuous monitoring of the internal structure of a tunnel will reduce the life cycle costs by removing the need for personnel to enter the tunnels for visual assessments and enable remedial work to be better planned by analysing a virtual 3D point cloud of the tunnel before stepping foot onto site.