Due to its complex phenomenon, ice-induced load process could not be physically possible to be modeled by a specific well established probabilistic model. Therefore, a non parametric approximate probabilistic approach should be performed based on the available data. This paper describes the procedure and also the extrapolation to get the short term and long term extreme values. Comparison with the classical approach, where the initial distribution of ice-induced load is assumed to be one of the well established probabilistic models, was made. The comparisons discussed in this paper were explored by using the same available data of full scale measurement on board a coastal guard vessel KV Svalbard during the winter 2007. There was a tendency that the non parametric approach produces more conservative results.
展开▼
