Much of the available triaxial compressive strength data for columnar saline ice has been generated from conventional triaxial tests. The advent of hydraulic "true" triaxial testing systems has allowed researchers to determine the compressive strength of ice using loading paths that were not accessible with conventional triaxial tests systems. The question that is considered here is whether data obtained from different loading paths can be directly compared. The deformation of laboratory-grown S2 columnar saline ice (LGSI) under triaxial compressive loading was investigated at a temperature of -10°C at a constant strain rate of 3.9 x 10~(-5) s~(-1) applied in the direction of greatest load, using an MTS multiaxial servo-hydraulic testing system. All of the tests were carried out on 160 mm cubic samples. In one set of tests the samples were proportionally loaded with confinement ratios of R_(21)= σ_(22)/σ_(11)=0.25 and 0.50, with R_(31)=σ_(33)/σ_(11) variable; σ_(11) and σ_(22) are the normal stresses applied in two orthogonal directions across the columns and σ_(33) is the normal stress applied along the columns. A second set of tests was performed to simulate conventional triaxial loading. In this case the samples were loaded hydrostatically to an initial pressure σ_o and then further loaded along x_1 at a constant strain rate of 3.9 x 10~(-5) s~(-1) while σ_(22) and σ_(33) were held constant at σ_o. In this scheme x_1 and x_2 are perpendicular to the long axis of the columns; x_3 is parallel to the column axis. Comparison of the yield stresses from the proportional loading tests and simulated conventional loading tests did not reveal any significant difference in the yield strengths from the two loading paths, within experimental scatter. Further analysis has shown that there is statistically no significant difference in the flow stresses taken at other points on the stress-strain curve (σ_(11) vs. ε_(11)), up to 2% inelastic deformation.
展开▼