On the Scalability of Model-Scale Ice Experiments

Abstract

Ice model-scale tests are a frequently used mean to assess and predict the performance of ships and structures in ice. However, ice model-scale tests may not be treated as a black-box where any full-scale scenario can be tested and a Froude-scalable result is obtained. Prior to scaling a thorough analysis of the physical processes is required and whether they can be transferred to full-scale. Model-scale ice is an empirically developed compound-material, consisting of frozen water, voids of air and other artificial dopants. The model ice manufacturing process and dopant amounts have been adjusted to achieve Froude-scalability for the ice thickness and certain force response levels, i.e. ice resistance tests of ships breaking ice in the bending mode. However, not much is known about the internal mechanical processes of model-scale ice and how the scaled force levels are reached. This may add uncertainty to ice model tests and their application on new fields. Recent research indicated that the internal mechanics of model-scale ice and natural sea ice are different, which is also challenging some of the existing scaling approaches. Mechanical specimen tests in full-scale and model-scale are usually compared by stresses, i.e. relating the failure load to the cross-sectional properties. However, depending on the tests different stress combinations might lead to failure, such as different geometries and dimensions may cause qualitatively different stress distribution, which ultimately limits the comparability of the tests. Subsequently, this paper presents a qualitative assessment on selected topics to assess the differences of model-scale ice and natural ice and the influence of the specimen geometry. Furthermore, existing scaling approaches are discussed in context with recent research findings.