Utilization of Gamma & Weibull Distribution Functions in Validating Experimental-based Result of Ore Amenability for Coarse Gangue Rejection

Abstract The mining industry faces two-pronged challenges, that is the increasing demand for commodities and declining ore quality, resulting in higher energy consumption and operational costs but poor productivity. Multiple methods have been proposed to address this issue, including early coarse gangue rejection which requires further amenability testing by either experiments and statistical analysis. Experimental approaches, such as Gangue Rejection Amenability Tests (GRAT), have demonstrated efficacy but required substantial time and resources. Statistical methodologies utilizing lognormal distributions offered an alternative, but unable to validate statistical outcomes against experimental results. This study investigated the potential of alternative statistical methods (i.e., gamma and Weibull distributions) to validate GRAT. Through the analysis of grade and density parameters data drawn from eight ore fractions, this research constructed an accurate statistical model for predicting ore separation results. The model’s accuracy was evaluated based on coefficient of determination, heterogeneity of data, and comparison of statistical predictions of cumulative mass and metal yield with laboratory results. The results of employing two distributions did not yield accurate predictions of GRAT results. The statistical values tended to be higher than the experimental results for both parameters. We suggested that alternative functions be tested for more representative and suitable use in coarse gangue rejection applications.