Gas Pipelines are exposed to many threats and ASME B31.8S groups the threats into ‘Time Dependent’, ‘Resident’ and ‘Random or Time Independent’ threats. The time dependent threats include internal and external corrosion and stress corrosion cracking (SCC). Resident threats include manufacturing, construction and equipment defects and malfunctions. Random or time independent threats include third party (interference) damage, incorrect operational procedure, weather and outside forces (including geotechnical).

Pipeline failure probability from third party threats, manufacturing, construction and equipment threats are usually covered by generic data based on statistics from years of pipeline operations. However, for outside forces, earth movement threats from geohazards are dependent upon site specific data. In lieu of site specific data, estimates of pipeline failure probability can be made using the damage potential for different geohazard categories and the associated incident frequencies. The damage potential depends on the soil pipeline interaction for the geohazard, the section of the pipeline exposed to displacement and the pipe and girth weld strain capacity.

DNV GL has developed a methodology for Structural Reliability Assessment (SRA) of pipeline girth welds applying a probabilistic machine learning (ML) approach coupled with Finite Element Analysis (FEA) to evaluate the probability of fracture for pipelines subjected to large deformations. The methodology is combined with incident rates following the principles of the approach used in the UK, which is summarized in IGEM/TD/2. The approach detailed in IGEM/TD/2 is applicable to any pipeline where terrain type can be zoned and combined with an estimate of the geohazard incident rate for each zone. The SRA is used to derive the failure frequencies at different locations and finally combined with the predicted consequences of a pipeline failure in a Quantitative Risk Assessment (QRA). This paper describes the methodology and presents a case study to illustrate the application of the methodology.