Thesis
Spanish
ID: <
http://hdl.handle.net/10251/114943>
Abstract
In a metallurgical project, a problem to solve is the quantification of the recovery, the degree and the kinetic speed of the valuable mineral that is expected to be obtained from the material that will be processed. One of the methods to perform the quantification is to carry out the metallurgical flotation process, which consists in the separation of the valuable material from the non-valuable one in a metallurgical laboratory. Because the modeling of the complex flotation process is complicated, it is essential to be able to identify variables with greater explanatory power to obtain a model as parsimonious as possible. In the present Master-thesis project, multivariate analysis methodologies and machine learning techniques have been used to predict metallurgical results, which could later be used in the prediction of the entire geological deposit. Satisfactory results have been achieved in the prediction of the degree of copper in the flotation concentrate, demonstrating the robustness of the techniques applied as predictive tools. TFGM