Prediction of extreme loads and fatigue damage for a ship hull due to ice action

Abstract

Knowledge about statistical behavior of local ice loads on ship hulls is valuable for the purpose of safe voyage in ice-covered waters, both for short and long term time horizons. Based on the loads that are measured, e.g. by an ice load monitoring system installed on board of the vessel, short term duration statistics can be established and utilized in order to predict the extreme ice loads into the near future for subsequent parts of the voyage. As a consequence, voluntary speed reduction and/or route alteration can be decided upon. Here, the short term statistics are obtained for an interval which is characterized by stationarity conditions with respect to the ice thickness, vessel speed, and propulsion power.

Certainly, the conditions in term of ice thickness, vessel speed, and propulsion power vary significantly throughout the service life of the vessel. This emphasizes the need to establish long term statistics. The extreme value predicted by long term statistics is useful for scantling and design in relation e.g. to ultimate limit state criteria. On the other hand, if the vessel has already been designed based on a certain ice class, such a long term prediction is useful for checking and quantifying the risk when the vessel is going to be deployed for operation in a certain ice-covered route. Moreover, by applying the long term conditional statistics, the prediction of accumulated fatigue damage of hull details can also be performed.

It was then realized that a systematic and proper procedure was in demand in order to handle the very important issues mentioned above. Accordingly the main objectives of this dissertation were to establish a probabilistic approach for prediction of short and long term extreme ice-induced loads and for prediction of accumulated fatigue damage in the local components of a ship hull.

Eventually, this dissertation outlines a new generalized statistical model for representation of ice-induced loads on a ship hull, i.e. the three-parameter exponential model, which well suited for fitting the whole range of the distribution. A new systematic procedure for prediction of the long term extreme ice-induced load on a ship hull is also presented. This procedure is referred to as the m-nautical mile maximum approach. Furthermore, a new systematic procedure is described which is based on the Palmgren- Miner’s rule for prediction of fatigue damage for a local component of the ship hull. This procedure can be applied even for the case when a limited data base is available. Formulations for accumulated fatigue damage for various models of ice-induced stresses are developed. This includes formulations for the "classical" one-slope SN-curve, for the case when a lower stress threshold is introduced,for the case that a stress concentration factor is given, and for combinations of these three cases. The proposed methods/procedures are applied to several calculation examples, which utilize full scale data from the expedition of KV Svalbard during the winter of 2007 in the Northwestern part of the Barents Sea.