Improvement of Rubble Ice Generation in Numerical Simulation of Ice Ridge and Structure Interaction

Abstract

Vessels and offshore structures operating in the Arctic Ocean are expected to face ice ridges during operation. Ice ridges are formed when sea ice layers collide and break, creating a pile of rubble ice along a line. Experimental model tests in ice tanks are performed to assess loads on offshore structures interacting with ice ridges and the ability of a ship to break through an ice ridge. Computational simulation of the ice ridge and structure interaction at the early design stage would allow the prediction of the loads and behaviour of the structure before testing in model ice basin. The numerical simulation under development at the Hamburg Ship Model Basin (HSVA) is capable of creating ice ridge and simulating the interaction with ships and offshore structures. In the pre-existing version of the numerical simulation, the rubble ice particles were represented by three-dimensional rectangles. In order to improve the rubble ice geometry, and consequently, the interaction between ice-ice and ice-structure, this master thesis offers improvements on the geometry modeling, where they are represented by three-dimensional polygons. The dimensions of the polygons in the simulation were obtained from samples of model scale rubble ice generated from ice ridges constructed at HSVA ice basin. The axis-aligned bounding box dimension and the number of edges of the rubble ice were collected. The data was fit in beta probabilistic distribution function (PDF). Thereafter, a PDF random variate of the axis-aligned bounding box and number of edges are the input for the creation of polygons in the simulation. In addition, the algorithm for the calculation of the volume, moment of inertia, superficial area and mass of the polygon shaped rubble ice was implemented and validated. Subsequently, punch test simulation and ship simulation were carried out to compare the effects of the implementation with the pre-existing code. The software was validated with experimental model test data from HSVA.