Abstract
M.Phil. (Civil Engineering)
As open-pit mines become deeper, the stability aspects of large slopes can be at the limit or beyond precedent practice. There are new design demands, scale related challenges, continuous blasting effect to the rock mass, various complex interactions between rock stress, fractures, and water flow, over long periods. Thus, there is a greater probability of slope failure associated with deeper and eventually steeper slope development. Interactions between the factors affecting slope stability in open-pit mines are therefore more complex and often difficult to define. Determination of the non-linear behaviour of such a multivariate dynamic system is often challenging and demanding. In addition, the problem is often dictated by non-linear equations. Therefore, there is a need for the proposal of an intelligent slope stability index. A new objective holistic approach is proposed in order to address the status of stability of slopes in open-pit mines. Through this approach the primary role of rock mass structure, in situ stress, waterflow, and construction effect have been extended into 18 system parameters describing the overall environment, intact rock quality, rock mass properties, in situ rock stress, hydraulic conditions, discontinuities properties, geometry, construction, and history of instabilities. Soft computing methods such as Back Propagation, Self-Organising Maps, and Artificial Neuro Fuzzy Inference System algorithms are used in conjunction with the Rock Engineering Systems approach to assess the stability status of 141 open-pit slopes in a complex rock engineering system. Implementation of the Rock Engineering System allows for the complex interaction investigation between the influential parameters using the generic interaction matrix. The central reasoning behind employing ANN to slope stability in open-pit mines is its ability to identify interesting and useful patterns in data. Hence, new knowledge is gained with regards to the open-pit engineering environment in an attempt to reduce the associate uncertainty. An Intelligent Open-pit Mine Slope Stability Index is proposed to assess the potential status regime from an objective holistic point of view. The reliability of the predictive capability is computed as the Mean Squared Error, and further validated through a Receiver Operating Characteristic curve. Together with the Mean Squared Error, and Receiver Operating Characteristic curve of 92%, 97% and 98% for Back Propagation, Self-Organising Map, and Artificial Neuro Fuzzy Inference System respectively, the application illustrates that the prediction of slope stability through Artificial Neural Networks produces fast convergence giving reliable predictions, and thus being a useful tool at the preliminary stage of open-pit slope stability project.