Ice-structure interaction in shallow water - A study based on laboratory-scale experiments and discrete element simulations

The global warming and retreating sea ice cover open up new transportation routes and offer new opportunities for exploratory activities in the Arctic regions. The increasing levels of marine activities in these regions constantly increase the demand for offshore structures in ice-covered sea areas. Optimizing the design of offshore structures still requires new engineering insight on ice-induced loads and the mechanics of ice-structure interaction. During an ice-structure interaction process, ice is drifting against a structure and failing into ice blocks, which form an ice rubble pile. This process may subject the structure to high ice loads. Many offshore structures operate in shallow water, meaning that the rubble pile may ground, which affects the further ice loading process. This thesis studies the ice-structure interaction against a wide, sloping structure in shallow water. The work consists of model-scale experiments and full-scale numerical simulations. The model-scale experiments were performed in the Aalto Ice Tank by pushing an ice sheet against a ten-meter-wide, inclined, structure in shallow water. The structure consisted of ten identical segments and the loads measured on the structure were analysed in different resolutions. The experiments were conducted in three tests series with varying ice strength. The novelty within the experiments was in the very wide structure and the segmentation of the structure. The ice loading process in the experiments showed two distinct phases: the load on the structure (1) first increased linearly with a rate of increase depending on the ice mass above waterline, after which (2) a steady-state phase with an approximately constant load level was reached. The experiments also showed that the magnitude of ice loads was not directly proportional to the ice strength, as the weakest ice yielded higher loads than the ice having twice its strength. The loads on the individual segments correlated strongly on the level of the entire interaction process, suggesting that the assumption of non-simultaneous ice load does not apply. In the numerical modelling, a two-dimensional finite-discrete element method was used to simulate the ice-structure interaction process against a shallow water structure in full-scale. The simulations allowed for investigations on the ice loading process in full-scale and the focus was on the influence of water depth and the ice thickness on the ice loads. The simulations showed that the peak ice loads increase with decreasing water depth. The simulations also indicated that scaling the rubble pile dimensions with the square root of the ice thickness correctly account for the influence of ice thickness on them. This suggests that the water depth to ice thickness ratio, often used in ice action-related problems in shallow water, is dubious when considering the ice-structure interaction process. Both the simulations and the experiments indicated that the loads applied on the bottom, simulating the seabed, are small compared to the loads applied on the structure.