MODELLING SHIP PERFORMANCE IN ICE USING BAYESIAN NETWORKS

摘要

Navigation in ice has received substantial attention over recent decades. This increased attention has led to the development of numerical and semi-empirical models that characterize ship performance in ice. These models consider numerous parameters such as; level ice thickness, ridged ice thickness and ice concentration in additive fashion. However, they fail to account for the joint effect of the above ice features on ship speed. Moreover, the effect of ice compression on ship performance is usually omitted. This paper introduces probabilistic models, based on field observations, that predict a ship's speed and the situations where a ship is probable to get stuck in ice based on the joint effect of selected ice features, such as the thickness and concentration of level ice, ice ridges, rafted ice, and ice compression. To develop the models a Bayesian Belief Network is used. The case study presented in this paper considers a single and unassisted trip of an ice-strengthened bulk carrier between two Finnish ports in the presence of challenging ice conditions and the obtained results show very good prediction power of the models.