This report describes a series of model tests and laboratory experiments carried out in the ice tank at IMD, between September 1986 and February 1987, for Melville Shipping Limited, and sponsored by the Transportation Development Centre. The purpose of the tests was to study the effect of friction between the ship model and the ice on the resistance measurements in level ice. The model test results show the effect of surface finish, characterized by friction coefficient or surface roughness, on resistance, within the range of parameters covered.

Results of propulsion tests in level ice on a model of the WTGB 140-ft. Great Lakes icebreaker are presented and compared to available full-scale data. In spite of the difficulties in exactly modeling full-scale conditions, the predictions based on the model test results of the ship performance compared reasonably well to those measured during full-scale trials. Several possible sources of errors are identified. In particular, duplication at the model scale of the ship hull's ice friction coefficient is considered to be critical in determining the ice resistance and the corresponding propulsion characteristics, namely propeller speed, thrust and torque. Keywords: Coast Guard ships; Ship models; Icebreakers; and Model tests.

Tests in level ice on an idealized icebreaker bow in the shape of a simple wedge were conducted in the test basin. The horizontal and vertical forces on the wedge were measured, and floe size distribution in the wake of the wedge was observed. From the force measurements, the ice wedge/hull friction factor was calculated and found to be in general agreement with the friction factor measured in separate friction tests. The ice floe length and ice floe area measured in the current study were found to follow log-normal probability distributions defined by the length average L and area average A and corresponding standard deviations S sub L and S sub A. The results of these tests and other tests conducted at another facility showed that the ratios A/h squared and L/h (A = average floe area, L = average floe length, h = ice thickness) were, for the same type of model ice, directly proportional to the parameter sigma/gamma h (sigma = ice bending strength, gamma = specific weight of water) and sigma/gamma h, respectively, and independent of the velocity and ice strain modulus or ice characteristic length. However, the coefficients of proportionality appear to depend upon the type of model ice used in the tests. The ratios S-sub-L/A were independent of sigma/gamma h but varied with the bow shape and the type of ice. The available field data are not sufficient for meaningful comparison with the laboratory results. (Author).