LAUSR

Abstract The tectonic fault poses a great threat to buried pipe segments. Buckling or rupture of pipes will ensue due to large strain induced by continuously increasing fault movements. In this study, a reliability-based methodology is conceptually proposed for designing new buried pipes or assessing existing pipes installed in fault areas. Uncertainties associated with the influencing factors of pipe structure and seismic loading (fault displacement) generated by earthquakes are considered. For pipe structural reliability, a developed back-propagation neural network (BPNN)-based surrogate model is utilized for the strain demand term in the limit state function for reliability calculations. Given statistical data for the basic variables including diameter, wall thickness, internal pressure, intersection angle or dip angle, and soil spring resistances, the conditional probability of failure can be obtained at a particular fault displacement by Monte Carlo simulation. Uncertainty of the fault displacement is estimated based on probabilistic fault displacement hazard analysis (PFDHA) having results given as probability of exceeding a certain fault displacement level for a specific site, then the occurrence probability of a specific fault displacement at a specific time can be indirectly obtained. The actual probability of failure (PoF) is calculated as an integral of fault displacement based on the total probability which is the product of the conditional PoF of pipe structure and occurrence probability of the specific fault displacement. This developed methodology is applied to a case study about a pipe project from China. Results show that the probability of failure increases with the elapsed time but stay within the safety margin for the next 30 years. Generally, the result of PFDHA is an essential input parameter in reliability analysis for pipes buried in fault regions. The multidisciplinary combination of structural reliability analysis and fault displacement hazard analysis is of practical significance for the reliability-based study of pipes in fault regions.