1 INTRODUCTION
Joint shear behavior is rediscussed on the basis of a 2D statistical characterization and a 3D modelling of a joint roughness morphology. The shear processes and progressive degradation phenomena on the joint surfaces are analyzed in relation with normal stress and shear displacement. These discussion and analysis are based on the results of an experimental study on the shear behavior of unfilled natural rock joint replicas.
The behavior of single, irregular joints submitted to direct shear conditions, under either constant normal stress or constant normal stiffness, is very complex. Scholtz (1990) stated that there is no constitutive law for friction quantitatively built upon micro-mechanical framework because of the complexity of shear contacts, the topography of contacting surfaces and the evolving surfaces topography during sliding. Even if a numberless research works were done in the last three decades, to solve various problems in relation with joint shear behavior, on which the ISRM Commission on Rock Joints organized three symposium (Stephansson 1985; Barton & Stephansson 1990; Goodman & Myer 1992), a recent review of the literature (Stephansson & Jinq 1995) pointed out that there is still a large number of problems to solve before an overall understanding of the phenomenon. Over this period, a large number of models were formulated to predict rock joint shear strength (Patton 1966; Ladanyi & Archambault 1969; Barton 1973; Swan & Zonggi 1985; Haberfield & Johnston 1994; to cite a few), several of which were compared on the basis of appropriate physical constraints for empirical relations and their ability to satisfy the principles of mechanics (Gerrard 1986). Other models were proposed for the deformability of rock joints (Goodman 1976; Bandis et al. 1981; Gentier 1986; Saeb & Amadei 1992). While these models of joints have substantially improved our understanding of rock joint behavior, their limitations must also been recognized. The mechanical properties of rock joints depend essentially on the degree of matching between the joint walls; on the surface of contact between them; on their morphological characteristics; on the magnitude of normal stress (or load) acting on them as well as the rock material properties of the joint walls and the presence of infilling materials. The dominant factor, influencing practically all these aspects of the mechanical behavior or rock joints, is the roughness morphology and the difficulty is its characterization and modelling (Stephansson & Jing 1995). Joint roughness profiles provide an incomplete and biased characterization of the surface morphology (Riss & Gentier 1990). Very few works account for joint walls asperity angularity distribution in modelling the mechanics of rock joint shear behavior and particularly none has done an attempt to integrate a 3D representation of the surfaces morphology. These observations and the basic concepts for modelling joint shear process on irregular surfaces are discussed on the basis of 3D statistical description of joint surfaces morphology in relation with direct shear test results from natural joint replicas submitted to constant normal stress loading conditions for given shear displacements.
