| dc.description.abstract | The present work focus on oil spreading in broken ice, and the content of this thesis falls into three categories:
• The physical and numerical modelling of oil spreading in ice.
• Ice models, and parameters describing the ice cover.
• Experiments on oil spreading in broken ice.
A background study was carried out to investigate existing models for simulating oil in broken ice. Most of them describe motion of oil simply as a function of the ice motion, and do not take advantage of the possibilities that recent ice models provide. We decided to choose another direction, starting from scratch with equations describing the flow of oil on top of a water surface. The equations were implemented numerically, including proper boundary conditions to account for the presence of physical restrictions in the form of ice floes in the simulation area. The implementation was designed to be able to apply data on ice motion calculated by an existing dynamic ice model. A first validation of the model was carried out using existing experimental data. As those data were obtained in a different setting, the recorded parameters and setup of the experiment were not ideal for our purpose. However, we were able to conclude that our model behaviour was reasonable.
We have carried out statistical analyses on meteorological data of wind speeds, temperatures, floe sizes and ice thickness to obtain probability distributions describing the parameters. Those data had been collected in the Pechora Sea. Wind and temperature had been recorded for a period of 30-40 years. For this region we also had available Argos satellite data from four buoys drifting in the ice in April-June, 1998. The Argos data were carefully analysed to suggest probability distributions and return periods for certain speeds.
Indoor basin tests were carried out to obtain data on spreading of oil in broken ice. A set of 20 tests was conducted, each with different type of oil, ice concentration, slush concentration or ice motion. In short, oil was introduced in the center of a 6x4.5 m section of the basin which contained polygonal ice floes. The ice cover was brought into motion, and the test was surveyed by overhead cameras. The videos were analysed to obtain the spatial distribution of oil as well as the paths of all the floes.
Finally, the same scenarios were simulated with the oil spreading model, and the data from simulations and tests were compared. It turned out that the model did not perform well in the initial test phase, probably due to the presence of some slush, which is not yet included in the model description. However, the model performed better in the later stages of the test, which is probably related to the fact that the slush tended to move toward the edges during the tests. | nb_NO |
