Open pit slope design requires sufficient geotechnical data to establish a representative geotechnical model for slope stability analyses. Geotechnical data uncertainty can result in a poor performance of the pit slopes with potential adverse consequences (loss of life, increased costs, operational delays, etc.). It is generally accepted that different confidence levels in the collected geotechnical data are targeted per project stage, with an increased confidence level from the early stage to the more mature stages of a mine. The absence of quantitative guidelines is an important issue due to the risk associated with data uncertainty that can be quite high.

This thesis proposes a methodology to quantify the confidence level in the employed geotechnical data for open pit slope design. It furthermore provides quantitative guidelines for the minimum number of boreholes, laboratory tests and joint orientation data required to reach targeted levels of confidence in the collected geotechnical data.

Three South African mine sites were selected to construct comprehensive case studies. The analyses of the collected geotechnical data allowed to quantify the impact of collecting additional data in the slope design process for different project stages in the life of a mine. The minimum number of boreholes to minimize the variation of the interpreted geological model from the real rock mass was evaluated by progressively increasing the number of boreholes simulated in 3D geological models. Statistical analyses, using the small-sampling theory, were performed to determine the minimum number of specimens required to reach targeted levels of confidence in the normally distributed laboratory testing results. A new methodology was developed to accommodate for lognormally distributed data. Structural data were analysed using discrete fracture networks (DFN) by simulating sampling boreholes in the models. The minimum number of joint orientation data required to reach targeted levels of confidence was evaluated by progressively increasing the simulated drilling density.

The principal contribution of this thesis is an innovative methodology for planning geotechnical data collection campaigns to insure to reach targeted levels of confidence in the collected data. These quantitative measures can eventually replace subjective measures in subsequent risk analyses.