AbstractsEngineering

The creation of underground mine excavation disturbs the original state of the rock mass surrounding the excavation which often leads to instability of the excavation. This poses a threat to the safety of personnel and equipment in and around such excavations. Therefore the stability of underground mine excavations has always been a major concern to geotechnical engineers. The stability of the excavations depends on physical and mechanical properties of the rock masses as well as the in situ stress condition. For stability analyses, these parameters are determined either by in situ investigations or laboratory tests. Because of the inherent uncertainties associated with natural materials such as rock masses, the precise values of the properties are never known. The sources of these uncertainties could be inherent variability caused by random process (aleatory uncertainty) or it could be a knowledge-based uncertainty (epistemic uncertainty) such as measurement error or model transformation uncertainty. Traditional deterministic methods, which use mean or single characteristics value of the input parameter for stability analyses, do not consider the inherent uncertainties in the rock mass properties hence the underlying effects of the uncertainties become obscure. Depending on the distributive character of these uncertainties, the deterministic methods may lead to results which are not representative of real behaviour. Therefore, for a realistic stability analysis of an underground excavation the uncertainties must be adequately considered in the analysis using probabilistic methods. With probabilistic methods, the likelihood of occurrence of unsatisfactory performance or failure of the underground excavation can be estimated with respect to a predefined tolerable limit. Even though probabilistic methods have been increasingly used in slope stability analyses, the application of probabilistic methods has not received considerable attention in the stability analyses of underground mine excavations for instance. Therefore, there is need for an increased understanding of the effect of uncertainty in the rock mass properties on the stability of underground mine excavations so as to facilitate the application of the probabilistic methods in the stability analysis of the underground excavation. This thesis is therefore aiming in that direction and thus the objective. To achieve this aim, the uncertainties in rock mass properties were quantified so as to determine the statistical parameters and the probability density functions for the random rock mass parameters. Different probabilistic methods, each with different sampling techniques and assumptions, were incorporated with commercial finite difference numerical code to analyse the stability of underground mine excavations for different case studies. The probabilistic methods considered were Point Estimate Methods (PEM), Artificial Neural Network (ANN), Response Surface Method (RSM), Monte Carlos Simulation (MCS) and Strength Classification Method (SCM). The results of these…