Recent research on longitudinal pipe-soil interaction shows that traditional analysis models are inadequate and too conservative, especially when cohesive soils are involved. The practical implication for SNAM, whose network extends over the entire Italian territory where slow ground movements inducing longitudinal soil-pipe interaction are frequent, is that the management of the gas pipeline has to rely mainly on field measurements.
The correct assessment of the interaction forces was therefore included as an important part of a wider research program, whose aim is to perform pipe risk analysis by which the structural vulnerability for some SNAM typical scenarios can be quantified as a function of such parameters as the pipe section geometry, the type of soil, the burial depth, the length of pipeline section involved and the magnitude of the soil imposed displacements.
Experimental activities specifically regarding longitudinal, static, interaction problems were carried out; in particular pull-out tests were performed on two out of use pipelines, having two different diameters: 8” and 24”. For each site, four different test conditions were investigated where type and compaction state of the material surrounding the pipe varies. The behaviour of the pipe embedded in the original clayey backfill was compared to that observed after such fill was excavated and replaced around the pipe, to simulate conditions after standard stress relieving works. Different fills where then used, made of either granular soil or granulite; this latter was used in order to ascertain the possible benefits of using light artificial materials to mitigate soil-pipe interaction phenomena. Site testing was accompanied by a careful geotechnical investigation both in the field and in laboratory that included direct shear tests of interfaces using coated steel pipe specimens.
A comparison between the results obtained and the existing state of the art is presented. This comparison allowed to verify the effectiveness of some interpretative models, and in particular the convenience of adopting effective stress based models rather than pure cohesive ones (i.e. total stress), even in the case of clayey soils.