During the design phase of the Confederation Bridge in Canada, the Canadian government Public Works and Government Services Canada asked the National Research Council (NRC) of Canada to provide information on ice loads from large ridges using a wide range of predictive technologies. The NRC put together a team that looked at loads from several sources including analytical models, physical model tests, finite element models, discrete particle models, and full-scale data. The ice loading scenario was an extreme first-year ridge loading one of the bridge piers. A large number of analytical models were used and the load components were separated into those from the consolidated layer, the sail, and the keel. An upper bound prediction from this approach gave a value of 16 MN on a pier, but the assumptions that were used to arrive at this value did not match observed behavior in the physical and numerical studies of the program. Physical model tests indicated that the loads could be 10.5 MN with a load of 7.3 MN from the keel and 3.2 MN from the consolidated layer and sail. A finite element analysis indicted a range of predicted values of 10 MN to 12 MN depending upon the assumptions used. A discrete particle analysis predicted load values from 2.2 MN to 9.5 MN depending upon the assumptions used in describing the stiffness of the ridge. A review of full-scale measurements on lighthouses and ships suggested that the loads could range from 7.3 MN to 10.4 MN. These predicted values compare to the highest load measured on the Confederation Bridge over a twenty-year span of just over 8 MN. This paper outlines the approaches used for this prediction study and their resulting predictions. It shows the value of using multiple approaches for load predictions for offshore structures in ice-covered waters.
Read full abstract