UDEC and RFPA2D simulations on the influence of the geometry of partially-spanning joints on rock mechanical behaviour

Abstract

The Authors used Universal Distinct Element Code (UDEC) and two-dimensional Rock Failure Process Analysis (RFPA2D) programs, which use the Discrete Element Method (DEM) approach and the Finite Element Method (FEM) approach, respectively, to simulate the mechanical behaviour of rock-like materials with partially-spanning joints in different geometries. The results from the two programs were compared with the results of an experimental study, conducted on cement-mortar specimens with partially-spanning joints in different geometries, in order to evaluate the feasibility of the two programs in simulating the mechanical behaviour of rock-like materials with partially-spanning joints. Three different partially-spanning joint geometrical properties, i.e. joint location, orientation and trace length, were considered in numerical simulations using both UDEC and RFPA2D and the experimental study. For partially-spanning joint location, both numerical programs produced reasonably consistent results with experimental results for the variation of Uniaxial Compressive Strength (UCS) against joint location, especially for higher values of joint location. However, considering the overall variation of UCS against joint location we proposed that the joint location is of negligible influence on the UCS of the rock. Variations of UCS against partially-spanning joint orientation for the experimental work and UDEC simulations were observed to match very closely, whereas RFPA2Dresults have underestimated the UCS for all joint orientations. The selection of continuously yielding joint constitutive model for the joints in UDEC simulation, which is more representative of the joints used in the experimental study, can perhaps be the reason for the more accurate replication of experimental results using UDEC. Moreover, both numerical simulations verified the result observed in the experimental study in which the UCS of test specimens is minimal when the partially-spanning joint is orientated at an angle of 45°. The strain distribution characteristics obtained from both numerical programs generally agreed. The fact that when the joint is oriented in 45° angle, the influence from the joint on failure of rock is maximum and with increasing and decreasing joint orientations from 45°, the contribution from the joint for the rock failure is less. The UCS of test specimens was observed to decrease linearly with increasing joint trace length from the results of the experimental study, and the results of numerical simulations from both numerical programs showed a reasonably good agreement with the experimental results. According to the strain distribution characteristics of the experimental and numerical simulation results from both programs, relatively longer partially-spanning joints can significantly influence the failure of test specimens, whereas samples with relatively shorter partially-spanning joints fail with considerable rupture in the intact material. Finally, we can conclude that, while both numerical approaches are capable of satisfactorily simulating mechanical behaviour of rock-like materials with partially-spanning joints, UDEC (DEM approach) with its more versatile features can provide more promising results.