Experimental investigation of spudcan penetration in multi-layer clays with interbedded sand layers

Highly layered soils are prevalent in several areas of current offshore activity and in particular in emerging provinces and fields. However, no investigation has been carried out for spudcan penetration in stratified sediments with more than two layers. This paper reports centrifuge modelling of spudcan foundations penetrating through 3 similar to 6 clay layers with interbedded stronger clay, silica sand and carbonate sand layers. The aim was to provide insight into the behaviour of spudcan foundations in stratified sediments comprising layers of different drainage conditions and mineralogy. The motivation for this paper emanated directly from the future needs identified by the recent revised version of ISO 19905-1. It was found that the downward soil deformation was extended into 3 similar to 4 underlying layers of sand and clay with thickness ratios (t/D) less than 0.6. The limiting squeezing depth was 0.17 similar to 0.18D when approaching stiff clay-moderate clay layers and 0.28D when approaching dense sand-stiff clay layers. These depths were enhanced (to 0.3 similar to 0.35D) by the soil plug trapped beneath the advancing spudcan. The values obtained using the ISO criteria either overestimated or underestimated these depths. Punch-through failure at a shallow penetration depth occurred with a load-spreading gradient of 1 (horizontal): 3 (vertical) in silica sand and with a much smaller gradient of 1:7 in carbonate sand. In both sands, the spreading angle reduced with increasing depth of punch-through failure, and diminished quickly immediately after a failure. In contrast to ISO guidelines, the plug base, where the end bearing was mobilised, was located in one of the lower layers moved down with the advancing spudcan. For all cases, a soil plug developed at the base of the advancing spudcan, with the height of the plug accumulating through trapped clay and sand, regardless of their strength, from layers above and below the advancing spudcan. The height of a soil plug was reduced partly by soil backflow and squeezing out from contact with a strong layer. The observed soil failure mechanisms and corresponding penetration resistance profiles highlight uncertainties in ISO 19905-1. Guidelines are provided to improve the suggested punch-through and squeezing methods in ISO 19905-1. Implications for the bottom-up design approach proposed by ISO, and also a top-down design approach used by some practitioners, are discussed.