For a mining company to start the mining stage of an ore, several previous stages need to be performed, such as the mineral research stage, drilling, sample preparation and analysis, along with QA/QC, creation of the geological model, estimation of resources and block model generation, resource classification, pit optimization and sequencing. All mining planning and sequencing steps are based on the Geological Model, which is generated through modeling and estimates using traditional and/or geostatistical methods. Since the model is created from an estimate, it is necessary to validate it. During block model creation, several important validations are carried out, but the main ratification of the model is obtained with the reconciliations, with the mine already in operation. Reconciliations are the comparison between a measurement and an estimate, in this case, between the production/mining data performed (volume/tonnage/quality) and the estimate made in the block model. As a result, reconciliation becomes an instrument used to validate the geological model, to detect failures in operations and provide opportunities for improvement. In the present work, it was detected during the monthly reconciliations that one of the ore quality parameters (grit) was not in harmony in the comparison between the real data and the data from the geological model. First, it was assumed that the model was coherent and an audit was carried out in the field to verify how the samples that represented the real quality of the ore were collected. In the field, it was seen that the samples were collected inappropriately in the ore piles, the sampling was non-probabilistic, the collection frequency was inadequate and the preparation was carried out in an inappropriate place and with inappropriate tools. Based on these considerations, a work team was formed and after several analyzes and debates, it was decided to interrupt the sampling of the piles and change it to a sampling where the product of the blunger feed (in the form of pulp) is presente, since was assumed that it would be a representative sampling and with the appropriate frequency. This type of sampling was already performed, but more samples throughout the day began to be collected and more analyzes were also carried out, allowing both production, using these data instantly to control the quality of the blend, and Geology, using in reconciliations and in helping to control the ore mine's quality, can benefit from the new process. After the start of the new sampling, it was possible to verify in the reconciliations that the quality data between production and the Geological Model became adherent, confirming that the discrepancy existed due to a problem in an operational phase (sampling) and not due to a failure in the Geological Model.
|
