Abstract
Risk assessment and reliability models try to predict the probability of landslide-induced pipeline failures based on detailed, site-specific studies. Because these models are mainly designed to be used on a site-by-site basis, applying them over long pipelines is a challenge due to the challenges of collecting vast amounts of data for those long distances.
A new GIS-based method has been developed to produce an order of magnitude approach to estimating the annual probability of landslide-caused failures (POFs) of pipelines over entire transmission systems using historical data that can range from loss of containment, loss of serviceability, and significant deformations caused by landslides.
This new method uses high-resolution light detection and ranging (LiDAR) mapping to detect and delineate terrain anomalies interpreted to be the geomorphic response to ground deformations caused by landslides. The possible landslides are inventoried to record their activity, relative relationship with the pipeline, proximity to the centerline, length of intersection with a pipeline, and the angle of incidence between the perceived direction of movement of the potential landslide and the pipeline. This method also integrates regional landslide susceptibility maps depicting the relative likelihood of soil units to landslide occurrence along the pipeline corridor and its surrounding areas.
In the presented study case, the developed method is applied to an approximately 19,312-kilometer (12,000-mile) pipeline system located in the United States and Canada. The application of the model yielded results that significantly help the operator to prioritize and optimize the allocation of resources for landslide management. The model can be replicated over multiple pipeline systems and customized to the particular needs of the end-users.
