When sea or lake ice interacts with concrete offshore structures in Arctic regions, the frictional forces between the ice and the structure cause abrasion of the concrete surface of the structure. This may endanger the structural integrity when the steel reinforcement gets exposed and experiences corrosion, and must therefore be taken into account in the design process. For the design of concrete offshore structures in Arctic conditions, an accurate description and prediction of ice-structure interaction is required. The interaction between moving ice and concrete surfaces is mainly governed by friction and the so-called stick-slip phenomenon. This phenomenon has been observed during laboratory and field testing and, although the physics of this phenomenon are believed to be well understood, the corresponding static and kinetic friction coefficients reported in literature have a widespread range and are inconclusive. This thesis aims at a more accurate identification of the ice-concrete friction coefficients.

For this graduation project, an experimental set-up was designed and stick-slip tests were carried out at Memorial University of Newfoundland, Canada. Additionally, a numerical model describing stick-slip behavior between ice and concrete was created. For the experimental set-up, a cylindrical fresh water columnar ice sample with a 50 mm radius and 50 mm height was attached to four springs with the same stiffness. The springs were attached to a support structure and throughout the test campaign, the stiffness of these springs was varied between 20 - 70 N/m per spring. To simulate one-dimensional ice-concrete interaction, the ice sample was placed near the edge of a rotating concrete slab. The normal load on the ice sample was varied from 0.7 to 2 kg by adding weight. In addition, the concrete velocity as experienced by the ice was varied between 0.15 and 0.82 m/s by increasing the rotational rate of the concrete slab. The static and kinetic friction coefficients were obtained from the experimental data and their dependence on normal load, velocity and spring stiffness was analyzed as well. The static friction coefficients found over the whole range of tests varied from 0.1 to 0.5. The analysis showed that the static friction coefficient decreases with increasing normal load and with increasing velocity. The kinetic friction coefficient was found to be in the range of 0.08 to 0.4 and may on average be obtained as 0.7 times the static friction coefficient. The kinetic friction coefficient, too, decreases with an increase in normal load and velocity. The influence of the spring stiffness was not clearly identified.

The friction coefficients that were calculated using the experimental data were provided as input to the numerical stick-slip model. An analysis was performed to verify that the model displays similar regression with the varied mass, velocity and spring stiffness, compared to what was observed in the experiment. The output was compared to the experimental data, and it was found that the model describes the stick-slip behavior as seen during the experiment with an accuracy between 84 and 99%. Although some further improvements to the model can be implemented, in general it is concluded that under the made assumptions, the model is valid for the prediction of stick-slip behavior as observed during the experiment.