Pipeline walking is a complex phenomenon, governed by the axial pipe-soil interaction which is not yet fully understood, and which has been gaining increasing attention over the last few years. Axial friction (as any other aspect of soil mechanics) is usually bounded by drained and undrained responses, while in real life it will often be somewhere in the middle. For pipeline expansion it could be critical as while the ends might move a couple meters in the few hours it takes to heat up, somewhere close to the middle it will move a few millimeters only (so the level of drainage is totally different along the same line, during the same loading).
Geotechnical frameworks to address the different aspects governing the soil response to pipeline axial movement have recently been published. However, the current practice has been to lump all the time dependent effects back into a single equivalent friction factor, based on a representative pipeline velocity.
This paper presents the results of ‘true’ rate dependent pipeline walking analyses, and compares them to those obtained using constant equivalent frictions. While good agreement is observed for particular cases, differences up to 30% of the walking rate are observed. Examples show that the results of rate dependent pipeline walking analyses are significantly influenced by how the temperature changes over time along the pipeline length, thus realistic modeling of the heat up and cool down processes is crucial.
The rate dependent model employed describes the axial resistance as a hyperbolic function of the pipe velocity. Additional aspects which are expected to influence the soil response (e.g. consolidation time between sweeps, progressive compression and consolidation hardening) have been neglected, and shall be considered in further studies.
