Assessment of Rockfall Deformation and Related Geometric Characteristics via Terrestrial Laser Scanning
MetadataShow full item record
Due to the required numerous and extensive rock cuts through steep topography and adjacent steep natural slopes, rockfalls are a common risk to which the Canadian National Railway (CN) and Canadian Pacific Railway (CP) rail corridors in Western Canada are exposed, which can result in economic or human losses in the event of a track blockage or derailment. Some rockfall blocks demonstrate pre-failure deformation prior to their release from the slope which can be detected via analysis of 3D point clouds obtained via repeated terrestrial laser scanning (TLS). Understanding what factors affect the likelihood and magnitude of a future rockfall’s pre-failure deformation, and subsequent establishment of expected deformation thresholds, may prove useful for hazard management. The research presented herein has several objectives: (1) Develop a workflow for the semi-automated assessment of rockfall pre-failure deformation via roto-translation, (2) Validate the accuracy of the roto-translation algorithm, (3) Develop algorithms to objectively assess rockfall block and post-failure scar geometric properties from 3-dimensional point cloud data, and (4) Populate a database containing deformation and geometric characteristics of past rockfalls and investigate correlations between these sets of properties. This study utilizes TLS data collected on average every three months from May 2013 to March 2019, at an active rock slope above a rail line in the Thompson River valley in British Columbia, Canada. A database of 55 rockfalls was populated, including both pre-failure deformation and geometric properties. Eighty-seven percent of these events demonstrated detectable pre-failure deformation. When considering all rockfall events in the database simultaneously, the strength of correlations between geometric and deformation parameters was limited. The only correlation strong enough to support identification of potential parametric deformation thresholds was between pre-failure toppling magnitude and rockfall volume. Rockfalls at this site with shallowly dipping sliding planes and high volume will in general be expected to have comparatively lower levels of detectable pre-failure deformation, especially toppling if the sliding plane is shallow, and will leave behind a relatively smooth sliding plane if significant pre-failure toppling occurs. The high frequency of rockfalls demonstrating detectable pre-failure deformation is promising for the feasibility of future application of monthly-scale TLS and iterative-closest-point roto-translation techniques for the monitoring of rockfall pre-failure deformation. However, if time-to failure is to be reliably estimated, there must be a greater understanding of what environmental factors and/or geometric characteristics are correlated with the onset and duration of the tertiary acceleration phase of creep.