| dc.description.abstract | Scientific and engineering efforts in mine planning theory are focused on improving the speed and size capacity of existing algorithms. They look for changing from minutes to seconds, and hundreds of thousands to million blocks of an ore body representation. However, mining practices are so full of manual work and personal decisions, that algorithmic solutions are changed considerably by the mine planner, lasting weeks in this process to achieve operative final results.
This thesis proposes a different point of view, joining some of such hand work decisions, like design and ore body modeling.
The developed work concentrates on parametric representations of mine design and an ore body model, through volumes and morphological tools, optimized by simulated annealing. It is shown that it is possible to model and optimize a final open pit, with road, benches and switch-backs design and to fit an ore body with parametric volumes, which include geological knowledge.
The principal applications of such results are: mine design could be obtained in minutes instead of weeks, the project value will not change because of handmade decisions, mine operation and geological units will have a common language through parametric volumes, geostatistical predictions will depend on geological knowledge and fitted data, geological uncertainty would be modeled from parameter stochasticity, so stochastic optimization could be implemented from simulations. In fact, mine planning algorithm inputs would no longer be a block model, but directly drill hole data.
Despite that some numeral examples were developed, real cases were not the scope of this thesis work. The value of this work concentrates on proposing ideas and a new field of investigation in mine planning, focused on more realistic mining needs and bringing different tools, as those that until today were the paradigm, and trying to join professional areas that work separately. | |
