Neural Networks and Support Vector Machine Models Applied to Energy Consumption Optimization in Semiautogeneous Grinding

摘要

Semiautogenous (SAG) mills for ore grinding are large energy consumptionrnequipments. The SAG energy consumption is strongly related to the fill level of thernmill. Hence, on-line information of the mill fill level is a relevant state variable tornmonitor and drive in SAG operations. Unfortunately, due to the prevailing conditions inrna SAG mill, it is difficult to measure and represent from first principle model the staternof the mill fill level.rnAlternative approaches to tackle this problem consist in designing appropriate datadrivenrnmodels, such as Neural Networks (NN) and Support Vector Machine (SVM). Inrnthis paper, NN and a SVM (specifically a Least Square-SVM) are used as Nonlinearrnautoregressive with exogenous inputs (NARX) and Nonlinear autoregressive movingrnaverage with exogenous inputs (NARMAX) models for on-line estimation of the fillingrnlevel of a SAG mill. Good performances of the developed models could allowrnimplementation in SAG operation/control hence optimizing its energy consumption.