Artificial intelligence applied to the control of the underground mining dilution
Luis Montiel Petro, Promine; Kilian Bao, DT Solutions Services; Yvan Dionne, Promine
Canada has a large mining potential, in where some operating mines exploit or are planning to exploit deep horizons and some other new deep mines will start operating in the coming years. The majority of underground mines follow the longhole stopping methodology. This strategy is characterized by the blasting of large stopes, a process that presents challenges pertaining to the management of both overbreak and underbreak. This project consists of the development of a program that forecasts the occurrence of both overbreak and underbreak through the use of artificial intelligence algorithms. The software analyzes data pertaining to the drilling, blasting, rock quality, support, dilution and operational loses of existing stopes in various underground mines to find patterns that are matched with data for planned stopes to predict the occurrence of both overbreak and underbreak. When analyzing a planned stope, the software will use these predictions to recommend changes to the planned drilling and blasting strategies to reduce dilution and operational loses. The software is intended to provide engineers with better information to improve their decision-making process by taking advantage of the best of two worlds, the human-machine interaction. Studies in risk analyses have demonstrated that this interaction provides better results. This project has benefited from a multidisciplinary team. Special thanks go to Laval University, Agnico Eagle, IAMGOLD, Promine and DT Solutions Services.
Dilution; Operational loses; Artificial intelligence; Longhole stopping