| Summary: | Experimental testing in model ice is still the state-of-the-art method for the performance prediction of ships and structures in ice. For frequently produced ice thicknesses large experience is available to estimate the growth rate and the final thickness. However, the prediction of growth rate for model ice thicknesses outside this range come with major uncertainties. An accurate growth prediction of the ice requires a physical growth model that can account for the complex boundary conditions. Such a model is also useful when boundary conditions change, for example the renewing of the cooling system. In this work, we propose a growth prediction model for model ice in The Hamburg Ship Model Basin (HSVA). In the HSVA, the ice is artificially seeded during the beginning of ice growth to modify the mechanical properties of the ice. We used a traditional thermodynamic ice growth model and modified it so that it takes the seeding procedure into account. We used measurements of air temperature and ice thickness to tune the tank specific model parameter. For the best fitting parameters, the mean squared difference between our model and measurements is 2.17 mm. We speculate that most of this difference is due to a strong post-growth that is not correctly represented in the model. Nevertheless, the model provides a significant improvement and help for the modelling of thin ice. By modelling the post-growth process individually, the model could be further improved. This, however, requires more measurements of the ice thickness during post-growth.
|
|---|