Load distribution along fully grouted cable bolts based on constitutive models obtained from modified Hoek cells

This thesis investigates the mechanism of bond failure of both conventional and modified geometry cable bolts (Garford bulb and nutcase) and introduces models that can address the responses of these under different conditions. To accomplish this a series of pull out tests was performed under constant radial confining pressure in a pressure vessel and load-displacement behaviour as well as radial dilation due to the axial pull were determined.; For the conventional cable, such data was already available through the laboratory tests performed earlier by MacSporran (1993) but to accomplish the same thing for the modified geometry cables a new modified Hoek cell was designed and manufactured. Tests were performed at a range of different constant confining pressures.; Compared to the conventional cable results, higher bond capacities and stiffnesses were obtained in the modified geometry cables and larger radial dilations were recorded which correspond to the presence of the "bulge" in the samples. The axial load response of the 25 mm Garford bulb and 21mm nutcase cables was practically the same whereas the obtained radial dilation was less in the latter.; A frictional model, similar in concept to that proposed by Amadei and Saeb (1990) for the rock joints, was introduced to simulate the mechanics of the bond failure in the conventional cable. For the modified geometry cables, however, the bond failure mechanics was different since the presence of the deformed structure created a considerable area of the grout which needed to be sheared during axial pull and required consideration of the cement shear behaviour in the model. Lacking enough data about the properties of the cement annulus in direct shear, a test programme was arranged to investigate that using cylindrical grout samples. A model, similar to the so called "continuous yielding model" proposed by Cundall and Hart (1984), was developed which enabled the prediction of the observed responses in the laboratory.; Load distribution along long bolts was investigated both analytically and numerically. Knowing the behaviour of conventional and modified geometry cables for a short length, a finite difference numerical scheme was developed to determine the distribution of axial load along a long length of cable. Both continuous and discontinuous rock mass displacements (cracks opening) were modeled and the approach was validated against closed form solutions. All these were implemented into a software program called CABLE (Computer Aided Bolt Load Estimation) which can be implemented in numerical stress analysis packages (such as FLAC, UDEC, MAP3D (non-linear), PHASE 2 etc ...) to account for the rock-reinforcement interaction. Using the code, many questions about observed cable failures can be answered and it can also be used to back analyze some of the earlier reinforcement practices and help to achieve an optimized design.