| dc.description.abstract | The subject of ice-induced vibrations on bottom-founded vertically sided offshore structures has gained renewed interest with the growth of the offshore wind industry, stimulated by many nations committing to ambitious climate targets. A phenomenological numerical model (Hendrikse 2017) has been developed to predict the occurrence and severity of dynamic interaction events, and Aalto University is involved in the further development of the numerical model, focusing of validation through model scale testing. The numerical model requires calibration based on empirical reference measurements for different types of ice. The work in this thesis includes implementing the numerical model in MATLAB, developing new strategies for increased calibration accuracy, and model scale testing to gather calibration data. This thesis aims to gather insight on the subject of ice crushing in model scale, to arrive at a set of numerical model input parameters that reproduce the behavior of the model ice at Aalto Ice tank, as well as to explore new methods for determining the creep-related input parameters to the numerical model.
A new method for determining creep parameters shows promise, and the results suggest that the cube-root relationship between force and creep velocity previously assumed in the phenomenological model does not apply to model ice, and that Newtonian creep is more accurate. The results also indicate that the model parameters can not be set to mimic delayed-elastic response of ice as measured in a limited indentation test, as it does not yield accurate statistical loads at constant indentation rate tests where ice crushing occurs. The overall calibration effort was unsuccessful due to an unexpected trend between ice velocity and mean global ice load in the continuous brittle failure regime. The trend can not be explained by the phenomenological model, and likely stems from the FG-ED model ice not behaving like natural ice in crushing failure. | |