Fracture mechanics and tribology of rock and rock joints

Abstract: In the first part of the thesis fracture toughness tests of rocks are described. Specimens of Stripa Granite have been tested in three point bending, compact tension and short rod tension. For testing in three point bending two test procedures were used: Specimens without precracking were loaded to complete failure, then approximate fracture toughness, Km , and fracture toughness from secant method, Ksec' were evaluated; Specimens were precracked in a single loading cycle and later loaded and unloaded several times to obtain the apparent fracture tougnesses, K Q Km-data from testing with different notch depth were statistically analyzed in order to determine the relationship between Km and the crack length, a. The main conclusion drawn from the tests are the following. Km-values obtained from specimens without precracking are inevitably lower than KQ-values. This is due to neglect of microcrack growth preceeding crack extension. Acceptable KQ-values can be obtained from the three point bending tests on a common core specimen with 40 mm in diameter and a notch depth ratio of 0.3. Fracture tougness, Ksc' from short rod tests are found to be size independent, and the data is consistent with KQ-values obtained from compact tension tests and three point bending tests. Finally, short rod tests on filled rock joints are presented and the results are shown to be scattered. In the second part of the thesis direct shear tests on joints of granite and slate are presented. A new device to measure vertical and horizontal displacements during shearing is described. A compliance approach is proposed instead of a stiffness approach in evaluating joint behaviour. The joint compliances (normal, shear and off-diagonal) are stress, and or displacement dependent. The normal and off-diagonal compliances obey a power law while the secant shear compliance has a linear relationship with displacement. A study of the complete shear curves shows that not only the shear stress, but also the resultant of the normal stress and shear stress should be considered in evaluation of the shear data. For rough rock surfaces two factors are found to contribute to the shear stress before sliding occurs. The first is the adhesion which occurs at the region of real contact, and the second is a deformation along the fracture surface. A new model for the shear mechanism is proposed. The joint surface parameters of specimens were studied by means of profile measurements. The change of surface parameters is small before and after shear tests. Based on the Hertzian contact theory, new asperity models for closure and shear of rock joints are proposed. The models allow for prediction of normal and shear compliances of rock joints. Good agreement was obtained between experimental data and predicted values according to the theories. For slate the predicted ultimate coefficient of friction is 0.32 - 0.327 and the experimental value is pu = 0.324, and the predicted secant shear compliance is 0.01 m/Gpa, the experimental value is 0.012 m/GPa. Furthermore, the shear model gives a theoretical explanation to some of the phenomenon observed in shear testing.