Slope stability analyses of rock slopes and a comparison of limit equilibrium and continuum numerical methods.

Master of Science in Environmental Science. University of KwaZulu-Natal, Durban 2016.

Open pit mines are amongst the world’s largest geotechnical structures accounting for a large proportion of the world’s metals and minerals. The pressure of future supply of these resources meeting the demands of a growing population leads the mining industry into precarious environments in attempts to access deeper set resources. Increasing depths associated with surface mining operations directly relate to slope angles where steeper slopes are economically appealing due to a reduction in waste rock removal, but at the expense of an increasing risk of slope failure. Rock slope instability in such environments represents a significant hazard and is the cause of serious injuries and fatalities as well as major financial losses. It is therefore essential to rigorously manage this hazard through regular rock slope stability analyses. There are numerous types of slope stability methods of analysis, all of which aim to safely uphold natural or anthropogenic structures. The Limit Equilibrium Method is particularly useful in hard rock masses with distinctive discontinuities. It allows for the prediction of stable or unstable conditions by calculating the stresses acting on a block and subsequently comparing them to the shear resistance provided by the discontinuity separating the block from the rock mass. This method however requires the user to define a potential failure surface in advance thus indicating its usefulness to a limited range of failure modes. Open pit mines are however dynamic environments which give rise to more complicated responses by the rock mass. The availability of computers and the continued advancements made in computational power has encouraged the development of sophisticated computer codes which are capable of modelling increasingly complex problems. More complex rock slope movements can be determined by numerical modelling techniques such as the Finite Element Method and Finite Difference Method. These methods can analyse the stability of slopes directly through the use of the Shear Strength Reduction technique where strength properties are reduced until failure takes place. This study explores the aforementioned techniques applied to three critical profiles (A, B, C) selected based on areas of known concern in an open pit mine. Structurally controlled (i.e. planar, wedge, toppling) and non-structurally controlled (i.e. circular) failure mechanisms were assessed via LEMs and compared with numerical models represented as pseudo-discontinuum media. Pseudo-discontinuum media were based on a range of joint network models incorporating kinematically feasible joints which vary in terms of orientation, length, spacing and persistence. These included the Parallel Deterministic network of infinite length and of finite length, the Ubiquitous Joint network model and the Veneziano Joint network model. The geotechnical model was based on data obtained from four different consultants over a period of 15 years which spans conceptual to design levels. Materials were modelled based on the Generalized Hoek-Brown and Equivalent Mohr-Coulomb failure criteria and results were reported in terms of safety factors, probability of failure and shear strain. Rock masses represented as continuum media in non-structurally controlled Limit Equilibrium Methods and numerical methods determined stable conditions with good agreement in safety factors between all methods and both strength criteria as well as shear strain accumulation zones between the numerical models. Structurally controlled Limit Equilibrium Method results revealed that profile AA’ is the most critical slope with a significant probability of planar and wedge failure at stack angle level. Safety factors for large scale planar failure of profile BB’, although stable, remains below the acceptance criteria for the overall slope angle. Profile CC’ produced acceptable safety factors and was deemed stable. These results correlated with that of the Ubiquitous Joint network model in the Finite Difference Method and the Parallel Deterministic network of infinite length in the Finite Element Method. Shear strain accumulation of predicted failure modes is however better modelled by the former. The introduction of rock bridges along discontinuity planes in the Parallel Deterministic network of finite and Veneziano Joint network model significantly contributed to stability, reaching minimum safety factors of 3.77 and depicting approximately circular failure surfaces. A rockfall analysis showed the significance of fall body shape on rockfall trajectories along topographies experiencing crest loss to varying degrees. The three-dimensional analysis revealed that good slope/cut slope topography performs well in terms of catching falling bodies at upper benches as opposed to the moderate and poor slope/cut slope topography which permits movement to the bottom of the slope. The two-dimensional analysis however overestimated results where a minimum of 80% of falling rocks reached the pit floor. Approximately angular shaped rocks achieved greatest velocities in the three-dimensional analysis as opposed to the circular bodies doing so in the two-dimensional analysis.