Abstract
Important short-term planning decisions such as direction of mining, mineable ore inventory, and plant feed ore material characteristics are made based on a blast-hole production model. A blast-hole production model is a numerical characterization of the blasted areas that are ready to be shoveled in the pit. The short-term plan has a direct impact on the weekly plant feed and concentrate production. The proper mining of the deposit, in the short range, depends on the accuracy and detail of the blast-hole production model.
The geologic attributes of the blast-hole production model are estimated pre-blast, using blast-hole data and a geologic interpretation of the area. The post-blast configuration is calculated based on pre- and post-blast topographic surveys as well as blast pattern parameters. Then, industry standard Machine Learning algorithms are implemented in the form of in-house applications to calculate productivity scenarios based on dig-boundaries, ore material blending, and concentrate production. These algorithms are continuously monitoring different production databases, including drill and blast, fleet management, and concentrator plant to update their parameters. The generation of these scenarios are used to assist in the design of realistic short-term plans.