Uncertainty quantification of rock quality designation at the Gazestan Phosphate deposit

Several Geomechanical variables reflect the characteristics of the rock masses and in recent years have been considered as the Geometallurgical responses in mining industry. These variables are highly effective in the energy consumption and in further stages of mineral processing, impact the grinding circuit for throughput, metal recovery and reagent consumption in mineral processing system. For instance, rock quality designation not only (RQD) is a very popular parameter that is able to define the domains with excellent and poor qualities of rocks, but also can be used to predict relative throughput rates in SAG milling circuit. Geostatistical modelling of this regionalized variable by linear interpolation methods such as kriging and Inverse Distance Weighting (IDW) has a longstanding problem with smoothing effect and non-additivity constraint that leads to a systematic bias in the produced maps. On the contrary, non-linear geostatistical methodology can be employed instead. In this study, the multi-Gaussian kriging approach as a non-linear geostatistical technique is adopted for probabilistic domaining of RQD at unsampled locations through a phosphate deposit located in Iran. The resulting estimates are checked thoroughly and can be applicable for defining the probable area dominated by soft, moderate and hard rocks applicable for mine design and mineral processing plant optimization.