The joint surface roughness is one of the important parameters which influences the mechanical and hydro-mechanical behavior of rock joints. In most cases, the joint roughness coefficient (JRC) was used to quantitatively express the roughness degree. In this study, we applied the mean Z2 values of all profiles on the joint surface, rather than picking out one or several typical profiles, as estimated parameter to predict the JRC values. The calculated Z2 values and JRC values diminish with the increase of sampling intervals. Through analyzing the measured JRC values under three sampling intervals, the 1.0 mm interval is pointed out the most appropriate one to evaluate the joint roughness.
In addition, we evaluated and compared the JRC values of joints within granite specimens that have different mechanical properties and weathering state by two different methods (Z2 method and backward analytical method). From obtained results, The Z2 method could relatively accurately predict the JRC values of unweathered material at 1.0 sampling interval, while overestimate the JRC values in weathering state. It may be attributed to that the weather process may weaken the mechanical properties of JCS values and basic friction angle which are the controlling mechanical factors of definitional JRC values. Moreover, the Z2 roughness metric does not take into consideration the mechanical proprieties and topography characteristics of weathering rock mass. In total, Z2 method should be improved by considering additional parameters related to mechanical properties of rock joint (i.e. distribution of contact area within joint surface).
It has long been recognized that the rock joint is one of the crucial components of a rock mass. The rock joint surface plays a significant role in controlling the mechanical and hydro-mechanical behavior. In previous study, a parameter of joint roughness coefficient (JRC) is proposed to quantitatively express the roughness degree (Barton and Choubey, 1973). Nevertheless, measuring definitional JRC values still needs a tilt test or direct shear test to make a reliable estimation. To overcome this limitation of experiments, ten typical profiling lines are defined and visual comparisons are made to estimate roughness degree (Barton, 1977). However, this method is strongly dependent on the subjectivity and experience level of the researcher. To remove this bias, the root mean square first derivative values (Z2), a statistical numerical parameter, is revealed, which is also easily to be determined.