Electronic Theses and Dissertations (Masters)

Permanent URI for this collection

https://hdl.handle.net/10539/37975

Browse

Browsing Electronic Theses and Dissertations (Masters) by Keyword "Bencat failure"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Slope Failure Prediction at Husab Open Pit Mine in Namibia
    (University of the Witwatersrand, Johannesburg, 2023-12) Thikusho, Christine Runguro; Watson, Bryan P.
    The study is focused on Domain D at Husab Mine in Namibia. The purpose of the study was to improve prediction of pending slope failures for planar and wedge configurations. Planar and wedge failures are similar in that little strain is required to initiate failure. Slope monitoring systems such as ground based radars, interferometric synthetic aperture radar and prisms were reviewed from the available literature. The data from the mine’s satellite monitoring data and the ground-based radar instruments was analysed. Slope prediction methods were used to back-analyse the failures, to determine if failure prediction times were possible. A case study was incorporated from the neighbouring Rössing Uranium mine, to supplement the data. The data utilised for the study was downloaded from the slope monitoring instruments on the mine i.e., the interferometric synthetic aperture radar, ground-based radar and tension crack data. The following slope failure predictive tools were investigated; the strain deformation approach; the inverse velocity method; the slope gradient method; the acceleration and velocity approach; and Displacement/Time plots. The back-analysis work done proves that the following slope failure predictive methods were able to predict failure at least 3 days before failure: velocity, cumulative displacement and inverse velocity. It appears that the Husab mine failure mechanism is not as brittle as previously assumed and failures are not necessarily instantaneous. Therefore, failures should be identified early, and the necessary risk mitigation measures implemented proactively. The ability of back analysing large volumes of stored data is important in the study of failure prediction.