Crack growth and faulting in cylindrical specimens of chelmsford granite

Abstract This study focuses on processes of fracture propagation and faulting in specimens of Chelmsford granite subjected to various end-boundary conditions and confining pressures. Chelmsford granite is a brittle material and contains numerous small defects which are preferentially oriented along three mutually perpendicular planes. The defects significantly affect the physical properties of the granite and the way the granite fails. Theoretical solutions for stresses within cylindrical elastic bodies indicate that stresses within a test specimen are markedly nonuniform. The stresses are not simply axial and equal to the axial load divided by the cross-sectional area of the specimen, as is assumed by many rock mechanicists. Regardless of end-boundary conditions of specimens, cracks in Chelmsford granite propagate parallel to the direction of axial loading. Cracks in a specimen of granite, with thin neoprene inserts or teflon inserts placed between its ends and the loading platens, grow most intensely near the longitudinal axis, at the ends of the specimen. A theoretical solution for stresses in specimens subjected to these end conditions indicates that compressive axial stresses and tensile radial stresses are maximum in those places. Thus, a combination of Griffith's theory of crack propagation and the theoretical solution for stresses seems to account for the observed crack pattern. Uniform loading was achieved for Chelmsford granite by inserting specially made steel discs between a specimen and the platens of the loading machine. Cracks in specimens subjected to uniform loading propagate relatively uniformly throughout the specimens, as predicted by theory. Specimens subjected to end-boundary conditions of uniform loading or of teflon inserts or of neoprene inserts, fail by longitudinal splitting, parallel to the direction of axial loading, and then by buckling of the split slabs. Specimens that were placed in direct contact with the relatively rigid platens of the loading machine also are stressed highly nonuniformly. The areas of most intense crack growth are two bands, extending from the corners of a specimen to a distance of about one-quarter of the specimen height from the specimen ends, as predicted theoretically. The specimens of granite subjected to direct contact end conditions or to triaxial loading, regardless of end condition, fail by faulting. The type of faulting depends on frictional properties of the platens. The fault surfaces consist of small steps, arranged as in a staircase. The surfaces of the steps are roughly perpendicular to the loading axis of a specimen and the vertical risers are surfaces of cracks that have propagated axially. Thus, immediately before faulting occurs, a specimen appears as though it consists of closely fitting, tiny beams, bounded by axial cracks. We have developed a theory of faulting of specimens of Chelmsford granite, based on beam-buckling theory and frictional contact among beams. The theory seems to describe relations between average ultimate compressive strengths and confining pressures, ranging from zero to 10,000 psi, for samples of Chelmsford granite. Neither the Coulomb criterion nor the modified Griffith criterion fit the experimental data.