| Summary: | Thesis deals with finite element (FE) simulation of borehole jack (BHJ) tests in ice. The purpose of the thesis was to simulate elastic-plastic range shown in BHJ tests, compare measured ice stiffnesses with obtained from simulations, fit plastic range, investigate plastic properties and stress descending in surrounding ice. Eight BHJ tests carried out at different depths of different boreholes in the first-year ice ridge have been taken for simulation. Tests were performed in ice blocks of sail and in consolidated layer. Two-dimensional (2-D) and three-dimensional (3-D) models were developed in FE program Abaqus 6.8-2. Plane strain situation was assumed in 2-D model. In reality this can correspond to the BHJ test at the depth of sufficient ice confinement. 3-D model considered thicknesses of ice blocks. Elastic moduli obtained in uniaxial tests of samples taken in the vicinity of the BHJ tests were used as a material input. The ice in simulation was assumed isotropic, because of severe distortion, and independent of strain rate, since it is not known at the beginning of the test. Simulations were carried out up to the measured peak stresses, softening and fracture of the ice was not covered. Stiffnesses obtained in 2-D simulations were lower than the measured, average error was 27.1%, while 3-D model gave stiffnesses in general higher and average error was 55.8%, however high value of the error is governed mainly by two results. Stiffnesses simulated with 2-D and 3-D models showed the same pattern as elastic moduli. It was shown that the thinner ice confinement of BHJ leads to the lower stiffnesses, and 2-D model could not take this fact into account. Fitting of elastic-plastic range was carried out with 3-D model, therefore lower values of elastic moduli had to be taken for fitting of elastic range. Some facts, which could lead to lower measured values of in situ stiffness, are discussed: interpretation of BHJ output pressure-displacement diagram, action of BHJ on both sides of the borehole. Introducing of plastic hardening was required for fitting of measured curves up the peak stresses, thinner confinement required higher degree of hardening. Fitting of measured diagrams required introducing of input materials with peak stresses in average 1.8 times higher than the ones of corresponding brittle uniaxial specimens. It was shown that to some extend plastic hardening in BHJ tests is due to confinement of surrounding ice. And also concluded that testing conditions could increase the brittleness of uniaxial samples, therefore plastic range in uniaxial samples was not shown. Stress descending in 2-D and 3-D models showed profile similar to the theoretical solution of stress descending in the circular plate with a concentric hole with internal pressure. 3-D model showed closer profile. Obtained results had good correspondence with the measured value. In experiment of Barrault and Strub-Klein (2009), at the distance of 4 borehole diameters stress descended down to 2% of stress measured by BHJ, while it was 3.6% in 2-D, 0.4% in 3-D model, theoretical approach gave 1.2%. Therefore it is possible to conclude that theoretical approach can be used for preliminary estimations of stress descending quite safely. Many uncertainties arose due to the fact that input used for BHJ tests simulation was taken from the BHJ tests performed in the sea ice ridge. Simulation of BHJ tests carried out in the level ice of known thickness, with input material parameters obtained from the field testing will help to validate numerical model. Introduction of fracture mechanics in the model is also required for further improvements. CoMEM - Coastal and Marine Engineering and Management Hydraulic Engineering Civil Engineering and Geosciences
|
|---|