Centrifuge tests to assess capacity performance of suction caissons without lids

A programme of centrifuge tests has been performed at the National Geotechnical Centrifuge Facility (NGCF) at the University of Western Australia (UWA) to provide insight on the monotonic capacity performance of suction caissons without lids. Two equivalent sized model caissons were considered in the tests – one featuring a built-in lid (the ‘control’ caisson) and a second ‘open-top’ model with no lid.
The tests were conducted at an acceleration of 35g in two different soil samples. The first sample was a fine silica sand, saturated with a viscous pore fluid, while the second sample was a slightly overconsolidated kaolin clay, saturated with water. The overall test programme comprised vertical loading tests (six in sand and two in clay) and catenary loading tests (eight in sand and five in clay) on the model caissons at different pulling velocities.
Before each loading test, the caisson model was jacked vertically into the soil at the testing acceleration. The depth profiles of installation resistance in each soil were in good agreement, with resistances at full embedment in the range 2.6 to 3.3 kN for sand and 95 to 105 N for clay.
Vertical loading tests in sand were completed at velocities of 0.0002 mm/s and 3 mm/s to explore a range of drainage responses. For the ‘control’ caisson, the loading velocity had a significant effect on the vertical capacity and mobilisation displacement, with a maximum capacity of approximately 1.1 kN for fast loading – up to approximately five times that of the slow loading. In contrast, the vertical capacity for the ‘open-top’ caisson was similar for both loading velocities and in good agreement with that measured in tests using the ‘control’ caisson when loaded at the lower velocity. These findings indicate that the ‘open-top’ caisson is likely to provide similar vertical capacity in sand as a conventional caisson (i.e. a caisson with a lid) when drained conditions are generated, but that a conventional caisson is likely to generate higher capacity under partially drained or undrained conditions.
Vertical loading tests in clay were completed at a constant velocity of 0.5 mm/s. Tests using the ‘control’ caisson gave peak vertical capacities of approximately 350 N with a reverse-end bearing mechanism mobilised, indicated by the presence of a full soil plug in the caisson after the test. This capacity was three times the peak vertical capacity of the ‘open-top’ caisson – with the ‘open-top’ caisson only offering a slightly higher peak capacity than the peak installation resistance.
Catenary loading tests in sand were completed at velocities of 0.001 mm/s and 0.1 mm/s. The early stiffness response for both model caissons under both velocities were similar. Higher catenary loads of up to approximately 3 kN (eventually) developed for tests at the higher velocity – these peak loads were approximately 1.4 and 1.5 times the peak loads measured in tests at the lower velocity using the ‘control’ and ‘open-top’ caissons respectively. The slower rate tests using the ‘control’ caisson gave peak loads that were approximately 20% higher than those with the ‘open-top’ caisson. However, once the submerged weights of the model caissons were accounted for, this difference reduced to approximately 10%, indicating that the open-top caisson is likely to achieve similar capacity in sand under catenary loading conditions.
Catenary loading tests in clay were completed at a constant velocity of 0.5 mm/s. Early results from these tests indicated that the capacity of the ‘open-top’ and ‘control’ caissons were similar under catenary loading conditions.
The self-weights of the model caissons were greater than the opposing soil resistance forces in the clay sample. This meant that (after the caissons were released from the installation apparatus) some additional settlement occurred before reaching a state of force equilibrium and before commencing the catenary loading event. This settlement was less for the ‘control’ caisson due to the bearing resistance that developed at the underside of the caisson lid. To normalise for this, additional tests were performed on caissons that were modified to have the same weight and installed to the same depth. Under these ‘normalised’ conditions a similar initial peak capacity was mobilised under catenary conditions in clay for both caisson types. However, with continued pulling, the ‘open-top’ caisson response was characterised by another gradual increase towards a second (much greater) peak load – almost twice the initial peak. No such load increase was observed for the ‘control’ caisson, which may be due to the geometry of the ‘open-top’ caisson that allows for continued embedment into deeper, stronger soil during catenary loading.