State Estimation for SAG Mills utilizing a simplified model with an alternative measurement

摘要

The semi-autogenous grinding (SAG) mill is one of the main devices utilized for grinding solid materials in mining activities. SAG mills exhibit a non-linear behaviour, whilst the variables of interest (states) are not readily available (not subject to measurements). Tipically, the estimation of SAG mill states is carried out utilizing the Extended Kalman Filter, albeit with system ouputs that are not available in practice. In this paper, we propose to estimate the states utilizing the Unscented Kalman filter with a simplified model of the mill. We also propose the utilization of a system output that is available in practice, but not tipically used. We illustrate the benefits of our proposal with numericall simulations.