Development of Brownfield Exploration Algorithms for El Teniente Exploration Management Division CODELCO Headquarters.

This project was an initiative of the exploration department of CODELCO, which main goal was getting a predictive model to identify potencial exploration targets in the neighborhood of a current active operations (aka Brownfield Exploration) called “El Teniente” in the O’Higgins Region of Chile. To do this the client provided us with several data for different sources, most of it in 2D such as geochemistry, geophysics studies, geological maps, elevation maps, lithology and satellite multispectral images. They also provided us with 3D geophysics studies and some drillholes.

The problem we solve in this project is a current unsolve problem because it is a kind of new. Almost all the exposed ore deposits were discovered, so they need a new approach to find new potential deposits, this means they need to know if there are a deposit underground. The idea of this project was give them some hints about where could be placed a new ore depósit using easy-to-collect and cheap data, avoiding random drillhole campaign. We noticed this problem will be experienced by all the industry in some moment, so we decided to go further and propose a new methodology from this work.

Our new methodology, in few words, combine the different sources of data, 2D and 3D, using machine learnign, geostatistic algorithms and the geologist knowledge to identify the footprint left by the porfiric process which create the deposit. The information in these key zones are used to generate a map of probability to find an economic interesting location.

As stated, this methodology was applied in the neighborhood of “El Teniente”, an operation property of CODELCO, also, given the promising ideas, in the final stages of the project, they asked us to apply it at another operation called “Gabriela Mistral”, also property of CODELCO, both with a good reception by the client.