Uplift of shallowly buried pipe sections in saturated very loose sand

Abstract

Contribution by R. PeekThe paper by Byrne et al. (2013) together with the companion paper by Williams et al. (2013) present perhaps the most comprehensive and systematic set of upheaval resistance tests in saturated clean sand. The results also appear very consistent with the theories and explanations provided, which further attests to the quality of the experimental work, and the power of the relatively simple model for development and dissipation of excess pore pressure, and the vertical slip model for uplift resistance. STABILITY UNDER FORCE-CONTROLLED LOADING?The paper shows that for loose backfill, to the point where it contracts when sheared, significant and predictable uplift resistance is developed for slow displacement-controlled loading of the pipe, ensuring essentially drained conditions, but that rapid displacement-controlled loading essentially liquefies the soil, resulting in near zero or even negative uplift resistance. But what would happen under force-controlled loading, which is more representative of the loading the pipe exerts on the soil?The authors conclude that 'It would therefore [i.e. in view of the test results] be prudent to avoid this behaviour [i.e. contractive behaviour of the soil leading to a rise in pore water pressure] by ensuring that soil backfill is sufficiently densified after placement, or sufficiently free draining, that loss of strength does not occur.' This discussion endorses that conclusion even when the rate of heating of the pipe is slow compared to the time needed to dissipate excess pore water pressures. The reason for this is that the upheaval loads in a heated pipeline are force-controlled rather than displacementcontrolled as in the tests. Indeed, as the yielding of the soil due to the uplift force increases the out-of-straightness, this also increases the uplift force, so the type of loading on the soil is even more severe than force-controlled. (At least this applies for small amounts of uplift, typical of those that develop up until upheaval buckling instability.) The implications of such an increasing-force loading condition are not clear from the results of the displacement-controlled tests reported.Based on the results in the paper it takes about 50 s for the excess pore pressure to dissipate for a burial depth of H ¼ 0 . 25 m to the centre line of the pipe. For typical flowline the burial depth might be 1 m, so that, for the same soil, the pore pressure dissipation time is 50 s 3 (1 . 0/0 . 25) 2 ¼ 13 . 3 min. The heating of a flowline, from ambient to near the maximum operating temperature, could happen that fast, for example when dual flowlines are used and fully developed production is switched from one flowline to the other, but typically heating takes longer. The question is whether under such slower loading conditions one can rely on the significant uplift resistance for very loose soils under drained conditions established in this paper.In theory if a test under displacement-controlled loading produces an uplift resistance F ¼ F(t ), then applying ...

Uplift mobilisation resistance of subsea pipelines in loose sand

“…2 it is also clear that the experiments of Trautmann et al (1986) and Matyas & Davis (1983) undertaken in dry sand conditions are similar to those undertaken in saturated conditions. However, the studies undertaken more recently in dry conditions (Martindill, 2005;Thusyanthan et al, 2010;Byrne et al, 2013) indicate higher mobilisation displacements when compared with those undertaken in earlier studies.…”

“…Figure 6, on the other hand, proposes an extension to the DNV zone where only data for very loose sand are presented. Here, the extension zone mainly captures the data reported by Byrne et al (2013), currently the most comprehensive study in terms of very loose sand. The zone suggested goes beyond the recommended DNV zone for 3?5 , H/D , 7?5 to include these data and also gives an extension for shallow depths.…”

“…A few tests on very loose sand with RD , 10% were undertaken by Schaminee et al (1990). More recently, studies reported by Byrne et al (2013) and Williams et al (2013) have provided a comprehensive dataset of approximately 60 tests performed on very loose sand (0% , RD , 17%).…”

“…Byrne et al (2013). It also contains test results commissioned directly by Technip, called Technip A and Technip B (Technip OED, 2003).…”