Study of the rubble ice action in scale-model ice ridge impact on seabed structures
MetadataVis full innførsel
The action of first-year ice ridges on offshore structures was studied using model basin tests, small-scale laboratory investigations and finite element modelling (FEM). Scalemodel experiments were performed at Hamburgische Schiffbau-Versuchsanstalt (Hamburg Ship Model Basin HSVA) in Germany. Seabed structures were impacted by modelled first-year ice ridges. Ice ridge keel geometry, ice thickness and driftspeed were all modelled at 1:20 scale, corresponding to a full-scale ridge with a keel 10 m deep. Four ice ridges were built with keel depths between 0.45 and 0.55 m. The main test variables were the temperature and rubble submersion time (5 h with Tair = -0.5°C and 18 h with Tair = 4°C) and the interaction speed (0.045 and 0.22 m/s). The structures included two underwater cubes with 0.7 m side lengths and one cone with a 60° slope and a waterline diameter of 0.6 m. Additional tests were performed to characterise the mechanical properties of the model ice keels, including punch tests, retaining wall tests, oedometer tests and piling tests. Small-scale laboratory work was performed in the cold laboratory at NTNU to investigate freeze-bond effects on the resistance of first-year ice ridges and their deformation process. The rubble was composed of saline ice blocks at - 7°C, and freeze-bonds of three different strengths were created between the blocks by submerging them for 0 min, 10 min or 20 h. Fourteen shear box tests with three different vertical confinements were analysed. The punch tests and interaction tests were modelled in three dimensions using Abaqus V8.2 with the nonlinear Eulerian finite element method. The rubble's behaviour was described with the Drucker-Prager material model. The following points were highlighted in the analysis of the experiments: • The horizontal loads on the subsea structures reached a steady state of progressive local failure correlated with the keel profiles. • The time series from the retaining wall tests showed that the plate load was correlated with the penetration speed, and a wedge extrusion was observed. • The oedometer test showed behaviour typical of a loose soil. The Young modulus was measured at 0.9 MPa, and the hydrostatic compressive yield curve was determined. • The finite element analysis of the punch tests showed that:A cohesive softening occurred in the rubble. It was assumed that the cohesion decreased linearly from its initial (peak) value to 0 over the first 2% of plastic strain.For the short-submersion-time ridges, it was necessary to use a vertical distribution of the cohesion representing the vertical distribution of the freeze-bond strength. The cohesion was distributed linearly in the keel, from 0 Pa to 1.2 kPa at the keel top.In the long-submersion-time tests, the cohesion was uniformly distributed in the keel and its value did not exceed 0.5 kPa.The ice density and the keel depth were the main parameters governing the frictional resistance of the keel. Because of the low confinement at the rubble failure planes, the importance of the friction angle was reduced; therefore, this test is not the most appropriate for determining this angle.• In the pile tests, the critical angle was higher than the repose angle, indicating that cohesion can develop in a short period of time (less than 5 min). The friction angle was estimated to be between 30º and 45°. • The shear box experiments at NTNU showed that rubble failure was initiated by the breaking of freeze-bonds. The analyses revealed that increased freeze-bond strength contributed to: o Higher rubble shear resistance in the primary failure mode. o Higher strain localisation and dilatation under low confinement deformations. Finite element simulations of the model ice ridge keel interactions with the structures were performed using the material model and mechanical parameters determined from mechanical tests: • The numerical models were able to estimate the keel load and its deformed shape. • A progressive failure of the keel occurred, which agreed with observations reported from small- and medium-scale physical tests. The action of the model ice ridge keels on the cone was computed by analytical and numerical methods. The analytical method involved keel load calculations using the recommendations given in ISO/FDIS 19906 with minor modifications to account for the geometry of the structure. The following conclusions were drawn: • ISO recommendations are insufficient and potentially underconservative with respect to the calculation of loads from wide ridge keels on simple conical structures (at model scale). • It is necessary to consider the surcharge due to rubble accumulation. • Numerical methods can be used to better understand the behaviour of unconsolidated ridge keels and to calibrate analytical models.