An experimental study on the progressive shear behavior of rock joints with tooth-shaped asperities

Abstract

Rock mechanics have commonly tended to focus on static stability where the initial peak shear strength is the principal concern. Recently, a need has arisen to assess dynamic stability because of an increased concern for earthquake, rock burst and explosion safety. Many computer models of rock masses [1–3] are usually based on the superposition or extrapolation of static experimental results. As numerical analysis capabilities increase, a means is required for obtaining a complete shear stress–displacement relationship for rock joints from pre-peak to their residual state. Essentially, the shear stress–displacement curves of rock joints generally fall into two classifications. Some shears reach peak strength, drop off rapidly and finally taper off in strength to a residual value. Others behave more or less plastically with little or no distinct peak. Actually, this type of shear stress–displacement curve is greatly influenced by the joint surface roughness. These changes can be related to the selective destruction of the asperities present on both joint faces. Because this information is not usually compiled, the transition from peak to residual values is often assumed to decrease either linearly or exponentially when the joints exhibit strain softening behavior [4–6]. Understanding the progressive failure process after the first peak strength is a meaningful matter. Joint roughness has an essential influence on the development, and as a consequence, the shear strength of joints. To date, joint roughness has only been considered as a dilation parameter that effectively increases the friction angle above the basic friction angle by some angle. Generally, the dilation term is not constant, but gradually decreases due to the asperities being destroyed with the increase in shear displacement. Patton [7] recognized that the asperity of rough joints occurs on many scales. He first categorized asperity into first-order (waviness) and second-order (unevenness) categories. The behavior of rock joints is controlled primarily by the second-order asperity during small displacements and the first-order asperity governs the shearing behavior for large displacements. Barton [8] and Hoek and Bray [9] also stated that at low normal stress levels the second-order asperity (with highestangle) controls the shearing process. As the normal stress increases, the second-order asperity is sheared off and the first-order asperity (with a longer base length and lower-angle) takes over as the controlling factor. In numerical modeling, Cundall et al. [10] intended to simulate the intrinsic mechanism of progressive joint damage under shear illustrated in Fig. 1. This is based on the hypothesis that the asperity roughness is not a constant attribute and account must be taken for the effects of changes in roughness, due to asperities being worn down and destroyed. He assumed that a complete shear stress-displacement curve of a rock joint is made up of many sizes of asperities, such that the resultant curve is more or less a summation of many basic curves. In his report, two basic curves for a given tooth-shaped asperity were proposed. These curves depend upon the normal stress level and asperity angle. At lower normal stress (see Fig. 1), shear stress initially causes elastic deformation of the asperity up to the yield deformation. The sliding of one asperity up over another (dilation of the joint) then follows. At a certain shear displacement, the resistance of the asperity is exceeded and shearing through the asperity occurs. Further shearing leads to residual shearing resistance along the newly created surface. At high normal stress, no dilation occurs and the asperities are immediately sheared off at the base. In order to enhance the applicability of this constitutive *Corresponding author. Tel.: +886-2623-4215; fax: +886-26209747. E-mail addresses: yang@freebsd.ce.tku.edu.tw (Z.Y. Yang).