Finite Element Modeling Approaches for Predicting Strain Demand in Buried Pipelines Subjected to Landslides

This study investigates how pipeline element type and pipe-soil interaction modeling affect longitudinal strain predictions for inelastic steel pipelines subjected to symmetric lateral landslide-induced ground displacements. Three finite element soil-spring approaches are evaluated using beam and shell pipeline models under two extreme soil conditions: cohesionless silt sand and cohesive stiff clay. Results show that modeling choices significantly affect predicted strain demands. Strain development depends strongly on soil stiffness and ground displacement magnitude. For silt sand, both beam-based models ( Pipe-PSI and Pipe-Spring ) predict tensile strains below 3% and compressive strains below 2% for displacements up to 2.5 m, indicating their differences can be neglected for screening-level assessments in cohesionless soils. For stiff clay, both models predict strains exceeding common screening limits. The Pipe-Spring model reaches the 3% tensile limit at approximately 0.32 m displacement versus approximately 0.48 m for Pipe-PSI , and the 2% compressive limit at approximately 0.25 m versus approximately 0.45 m for Pipe-PSI . Comparison with the Shell-Spring model reveals that beam-based soil-spring models underestimate peak strains once local deformations develop. While beam and shell models agree well at relatively small ground displacements (approximately 0.5 m in tension and 0.25 m in compression), the Shell-Spring model predicts substantially higher tensile and compressive strains at larger displacements due to its ability to capture local buckling and post-buckling behavior. Overall, the findings clarify the applicability and limitations of common soil-spring approaches, providing a rational basis for selecting modeling strategies to support more reliable strain demand estimation and pipeline integrity management.