Experimental studies on shear failure of freeze-bonds in saline ice: Part I. Set-up, failure mode and freeze-bond strength

Abstract

Abstract The strength of freeze-bonds in thin saline ice has been investigated through two series (in 2008 and 2009) of experiments in the Hamburg Ship Model Basin (HSVA) as a function of the normal confinement ( σ ), the submersion time (Δ t ) and the initial ice temperature ( T i ). The freeze-bonds were mostly formed in a submerged state, but some were also formed in air. The experimental set-up was improved in the 2009 experiments. In 2008 a ductile-like failure mode dominated (78%), whereas in 2009 the brittle-like dominated (93%). We suggest that this is a combined ice and test set-up effect. The 2009 experimental procedures allowed for careful sample handling giving higher strength and it was softer. Both these things should provoke a more brittle-like force–time response. The average freeze-bond strength in brittle-like samples was around 9 kPa while in ductile-like samples was around 2 kPa. The maximum freeze-bonds strength were measured for short submersion times, from 1 to 20 min, and reached a maximum value of 30 kPa. A Mohr–Coulomb like failure model was found appropriate to represent the freeze-bond shear strength as function of the normal confinement. Saline freeze-bonds in saline water had cohesion/friction angle around 4 and 1.4 kPa/25° for the brittle- and ductile-like samples respectively, which fitted well with previously published data. A bell-shape dependence for τ c vs. Δ t was found, which agreed with the predictions by Shafrova and Hoyland (2007). We suggest that this is essentially a freeze-bond porosity effect and propose three phases in time with subsequent cooling, heating and equilibrium to account for this trend. Qualitative experiments showed that the submersion time and the initial ice temperature were strongly coupled. To account for the connection between contact time, block dimensions and ice properties and the freeze-bond strength, dimensionless number were used. Fourier scaling was more appropriate than Froude scaling to scale freeze-bonds. The freeze-bonding made in air developed fast (in less than 30 s) when the ice was cold and dry, but no freeze-bonding occurred for the same contact times when the ice was warm and wet.