The majority of exploration, development and production offshore Newfoundland has occurred in shallow water. Currently there are two floating production, storage and offloading (FPSO) vessels and one gravity based structure producing oil on the Grand Banks. In recent years, there has been a move to deeper water offshore Newfoundland. There have been significant discoveries at Bay du Nord and Mizzen, and new licence blocks are opening up towards the north and northwest of the existing producing fields. With the move towards deeper water, new challenges arise. Supply facilities, search and rescue equipment and other required infrastructure is much further away. A GBS structure is no longer an option due to the water depth. A disconnectable concrete Spar is being considered as a concept for deep water locations off Canada’s east coast. The platform is designed to withstand iceberg impacts, or to be disconnected and moved off location to avoid impacts from very large ice features. Global design loads (ice crushing forces on the platform) and mooring systems global loads were estimated using a two-step approach. First the concrete Spar was assumed to be fixed (i.e. not able to move during an iceberg impact) and quasi-static global design loads were estimated using the Iceberg Load Software (ILS). The ILS was developed to model iceberg impacts with a fixed platform such as a gravity based structure (GBS). Assuming the concrete Spar to be fixed is a very conservative assumption for floating platforms which are free to move upon impact, albeit limited by a mooring system. Second, the inertial properties of the platform and the mooring compliance were approximated using a one-dimensional timestep model. Since the time domain model is not as efficient computationally, a subset of the simulated impacts from the quasi-static analysis was used as input into the time domain model, and dynamic design loads were estimated. The resulting design loads were used by designers to ensure that the structure meets the requirements of ISO 19906:2010. The end product is a more effective design for the platform, while not compromising the safety of the personnel onboard or the integrity of the structure, mooring system or risers.

In the calculation of ice loads, reliance is generally placed on analysis of empirical data. The main reason for this is that the complex behaviour of ice under stress makes it difficult to use calculations based on mechanics directly in design. Ice is extremely brittle and prone to fracture, and develops areas of very high compression and confinement during compressive failure, which result in substantial changes in the microstructure and behaviour of the material. These are termed highpressure zones. Progress in the analysis of ice failure using advanced methods of mechanics is reported. This has included studies aimed at understanding the behaviour of high-pressure zones. All work includes analysis of time-dependent effects, essential for a proper analysis of ice failure. The paper includes a discussion of the role of fracture and the development of high-pressure zones. Fracture processes are understood but the randomness of the flaw structure of ice in the field suggests that there is a strong variation in magnitude and position of high-pressure zones. This is supported by observations in the field on vessels operating in ice. The analysis of high-pressure zones is outlined; these produce the high local pressures of great importance in design. As noted, very high pressures occur at the centre of the zones on small areas, of the order of 70 to 100 MPa, accompanied by very high confinement and shear. The response of the ice is profoundly altered under these conditions, as compared to virgin ice. Methods of damage mechanics have been used successfully to analyze the failure of high-pressure zones, and a distinct failure load has been obtained. The methods of mechanics require a non-classical approach that recognizes changes of material behaviour with time. A scaling rule related to the mechanics allows modelling on a smaller scale in the laboratory. Fracture events make the process of interaction highly variable in time. In the view of the writers of the present work, full time domain simulation of the process of interaction is not possible. A reasonable way to move forward is to combine the progress in high-pressure zone analysis with empirical analyses of high-pressure zone areal density and intensity.