To determine the effects of dynamic loading rate on the fracture normal deformation and normal stiffness, a series of dynamic compression tests on artificial fractures in mortar specimens were carried out at different loading rates ranging from 10–1 MPa/s (quasi-static) to 10 4 MPa/s (highly dynamic). The fractures are of different numbers of asperities and consequently of different contact areas, which simulate the natural rock joints. The normal stress-closure response curves of fractures are measured under different dynamic loading rates. Accordingly, a nonlinear (hyperbolic) model of dynamic fracture normal behavior is established, which is modified from the existing static BB model by considering the effects of loading rate. Finally, the relationship between the dynamic stiffness and static stiffness is presented.
Over the past thirty years, a number of models of fracture normal behaviours have been proposed for quasi-static monotonic or cyclic loading conditions. Among them, two types of empirical models are more commonly used. One is the hyperbolic models proposed by Goodman (1968), Kulhaway (1975), Bandis et al. (1983), Another is the logarithmic models proposed by Swan (198 I) and Brown and Scholz (1985 and 1986). Goodman (1976) made laboratory measurements of fracture closure as a function of normal stress on artificially induced tensile fractures. The hyperbolic model of the fracture normal behavior proposed by Bandis et al. (1983) is commonly used in rock mechanics and engineering, and is usually termed as static BB model. Under quasi-static and cyclic loading/unloading conditions, the BB model describes that the initial load and unload cycles may cause a hysteresis between them. Successive load/unload cycles can continue to stiffen the fractures, and the BB model eventually tends to a hyperbolic elastic model without the hysteresis between the load and unload cycles. The above models of fracture normal behavior are obtained under quasi-static loading conditions, and the effects of loading rate are not taken into consideration. In the present experimental study, the laboratory tests on the fracture normal behavior are carried out under the dynamic monotonic loading condition. The loading rate applied normal to fractures covers a wide range from 10-¹ to 10–4 MPa/s, so that the effects of loading rate can be examined. The tests at very low loading rates between 10-¹ to 10–0 MPa/s are treated as quasi-static tests, which are equivalent to the standard for quasi-static rock deformability tests specified in the ISRM Suggested Methods for Rock Characterization, Testing and Monitoring (198I).
Fracture-inclusive specimens are composed of two rectangular blocks of intact mortar each with a dimension of 70x70x70 mm. Table I shows the basic mechanical properties of the mortar material. Between the two mortar blocks is there a fracture perpendicular to the longitudinal axis of the rectangular specimen with an overall dimension of 70x70xl40 mm. The fracture has a saw-tooth surface with a partial contact area through generating saw-tooth asperities stripes on one of the fracture walls when being precasted, whereas another fracture wall is made flat and smooth.
