Performance of suction caissons with a small aspect ratio

Abstract

Suction caissons with a smaller aspect (length to diameter) ratio are increasingly used for supporting offshore structures, such as wind turbines and oil and gas production facilities. The design of these stubbier foundations is usually governed by lateral loads from wind, waves, or currents. It is desired to have more physical understanding of the behavior of less slender suction caissons under cyclic lateral loading condition and to have robust design tools for analyzing these laterally loaded caissons. In this study, one-g model tests with 1:25 and 1:50 suction can foundation scale models with an aspect ratio of one are conducted in five different soil profiles: normally consolidated clay, overconsolidated clay, loose siliceous sand, cemented siliceous sand, and cemented calcareous sand. This test program involves monitoring settlements, lateral displacements (walking), tilt, lateral load and pore water pressures in the suction can during two-way cyclic lateral loading at one, three and five degrees of rotation. The model foundations are monitored during installation, axial load tests, and pullout tests. In one and two-degree (±0.5 and ±1 degree) rotation tests, the suction can does not have significant walking or settlement in all the five soil profiles after 1000 load cycles. However, more significant walking or settlement may occur at extreme conditions such as the 5-degree (±2.5 degrees) rotation tests. Gaps between the foundation wall and the soil may also form in these extreme conditions in overconsolidated clay, cemented siliceous sand, and cemented calcareous sand. Plastic limit analysis, finite element analysis, and finite difference analysis are used to evaluate the laterally loaded suction can in clay. The plastic limit analysis originally developed for more slender suction caissons appears to predict a lateral capacity close to the measured short-term static capacity of the caisson with an aspect ratio of one when undisturbed undrained shear strength of soil is used. However, this plastic limit model underestimates the long-term cyclic lateral load capacity of the caisson when the remolded undrained shear strength was used. The finite element model developed in this study can simulate the development and effect of a gap between the foundation and surrounding soil as observed in the experiments in overconsolidated clay. The lateral load-displacement response predicted by this finite element model matches well with the experimental data. Finally, finite difference analysis for a rigid caisson with lateral and rotational springs was developed by fitting the lateral load-displacement response of the suction can in clay. The calibrated p-y curves for rigid caisson are significantly stiffer and have higher ultimate resistance than the p-y curves recommended by API which is consistent with other studies. This finite difference model provides an efficient approach to analyze a laterally loaded caisson with a small aspect ratio in clay.