Abstract:
Reinforced concrete (RC) structures are prone to spalling damage. Successful detection and quantification of spalling is critical for monitoring the structural safety condition. Point cloud generated by emerging surveying technologies such as photogrammetry, laser scanning and Light Detection and Ranging (LiDAR) have been growingly used for rapid structural inspections and data-driven asset management. The integration of Building Information Modelling (BIM) and point cloud lays a new pathway for visualisation and management of geometric and semantic information of the structure. This solution will be particularly useful for post-earthquake impact assessment and building condition assessment when the timing is a major constraint. However, it is still a global challenge to automatically detect and quantify the spalling damage in RC structures from point cloud data.
Furthermore, little research can explain how to integrate the spalling damage into as-damaged BIM in the process of 3D BIM reconstruction from the point cloud. Focusing on damaged RC columns, this research presents a computational framework to detect the spalling from the point cloud, quantify concrete loss and exposed rebars, and to regenerate the 3D geometry of as-damaged BIM. Specifically, an algorithm was firstly designed to identify the spalling damage from the captured point cloud. After re-aligning the coordination system, the point cloud model was sliced into layers for further detection of the spalling damage and quantification of exposed reinforcement from concrete, the exposed surface area and lost concrete volume. Finally, geometry information was used to reconstruct the as-damaged BIM models. The proposed framework was tested on full-scale RC columns, and the results showed that meaningful spalling damage information to civil engineers can be extracted automatically, and that the reconstructed 3D BIM is accurate in representing the geometry of the physical structure.
