عنوان مقاله [English]
Simulation methods give the most probable models that could be used to obtain the best and worst case of a resource. Therefore, average of these models is similar to the estimation model but does not contain the estimation models problems. In the introduced new method after determining the rock type probability models the grade distribution is evaluated using geostatistical simulations. This method in addition to containing the probability models uses simulated grades that may be most authentic. Performance of the presented method is evaluated using a case study of a porphyry copper mine in Chile.
Geological modeling is one of essential processes in ore reserves and resources evaluation that is prior to the grade estimation. It containins the deterministic division of deposit into smaller subdomains that are called geological units. Although this categorization leads to better behavior than spatial grade variability, but it is not possible to present the uncertainty models in the boundaries of geological units that are needed to classify the ore reserves and resources according to international standards such as JORC. This article presents an approach that uses geological uncertainty that could be incorporated in the evaluation of mineral resources. Introduced method is applied on the rihoblanco deposit case study in chile.
Methodology and Approaches
At first to obtain probabilistic descriptions of three different lithological units the geostatistical simulation have been used in this deposit (granodiorit, tourmaline breccia and another breccias). Then to determine the grade uncertainty of the deposit, the conditional simulations have been applied. Finally, mineral resource model is presented using weighting the copper grade realizations by occurrence probability of different geological units. Presented model makes the production of a grade model possible. The grade model contains the geologic models uncertainty with avoidance of artificial models that caused by deterministic boundary hypotheses between high grade and low grade areas of ore in near the geologic boundaries.
Results and Conclusions
Introduced method presented better results compared to the kriging method. Results showed that evaluation of the deposit geological uncertainty prevented production of unrealistic models created in models like kriging method. Reliability of the presented model was evaluated using less standard deviation of weighted realizations conditional variance. It showed the suitable reliability of the model. Furthermore, coefficient of variation in boundary domains was low which showed the method is suitable and uncertainty in boundaries is low. In addition to that, the presented new method compared with general simulation of grade distribution showed that the new method predicts higher tonnage percentage in grade-tonnage curves.