Experimental and numerical investigations to assess the behaviour of a buried pipeline in areas with high geological instability

Ground displacements such as landslides, fault movements, soil liquefaction which may be caused by seismic activity are one of the most dangerous phenomena that can involve buried pipelines, e.g. for oil and gas transportation, or water and sewage. This aspect is currently an important part of research and a challenge for lifelines owners that are interested in prevent or limit pipeline damages. Within the framework of GIPIPE1 research program (SAFETY OF BURIED STEEL PIPELINES UNDER GROUND-INDUCED DEFORMATIONS) new full-scale facilities have been developed and adopted in order to investigate pipe-soil interaction mechanism (in particular sand and 8” 5/8 X65 steel pipes). The new experimental facilities have been designed to perform two groups of tests: simple interaction tests (axial pullout and transversal pullout test) and complex interaction tests (reproducing a pipeline crossing landslide). A system of steel containers (stationary and fixed) in which pipe samples are buried within the sand, have been assembled. Numerical analyses have been performed using strength parameters of sand and steel obtained from laboratory testing and subsequently validated by means full-scale experimental results. The outcomes of the experimental activity showed some differences in soil reaction on pipe by increasing the relative density of soil filling and using a smoother coating. Moreover peak soil resistances estimated with equations suggested by ASCE guidelines [4] cannot predict satisfactory measured axial and lateral soil reactions. This is a confirmation of previous studies in which was evidenced the effect of soil dilation in the annular soil zone around the pipe during axial relative movement between pipe and soil causes an increase of the normal stress at pipe soil interface, in particular the horizontal direction is significantly constrained by the surrounding soil mass leading to an higher increase in lateral soil stress in this direction respect to the vertical direction. Therefore this phenomenon leads to a lateral earth pressure coefficient K which is greater than K0 (coefficient of pressure at rest) as suggested in the ASCE guidelines [4], therefore for a better estimation of soil response using that equation it is suggested to measure the ratio between horizontal stress and the vertical stress during a full-scale axial pullout test. Pipes submitted to lateral soil displacement with a constrained uplifting show as expected a greater soil reaction than that estimated by ASCE [4] and PRCI [20]. As far as the landslide/fault test are concerned, the maximum soil relative density (Dr) achieved during experimental tests performed in this study was around 40%. This level of density led to a low stiffness of soil mass hence a limited global deformation of a 24 m embedded pipe during landslide/fault tests in which one caisson was moved up to 4 m respect to the initial position. These experimental findings confirm that sand with a low value of maximum achievable density may prevent from high loads developing on pipelines, in contrast to native soil which can apply higher loads. Numerical analyses and their validation gave us a suitable instrument to estimate the pipe soil response for large ground displacements phenomena.