A predictive model that allows mill fill level and the lifter height to be determined from measurements of the head, bulk and impact toe positions for a generic SAG mill has been developed using data from a series of three dimension DEM simulations. A Design of Experiment (DOE) approach is used to define a series simulation conditions with the two factors being the lifter height and the fill level. The charge structure is characterised by the locations of the head, shoulder, bulk toe and impact toe. These are determined by visual analysis of the particle trajectories. The spatial variation of the liner pressure distribution can provide valuable information on the structure of the charge. This was found to vary systematically with lifter height and fill level. The relationships between the visual key locations and the key features of the pressure distributions were identified and found to be well described by linear equations. Finally, parametric models are fitted to the DEM data for both the lifter height and fill level allowing them to be estimated from measurements of the key charge locations. These could be used as part of a control strategy for the mill and/or to assist with re-line scheduling. Crown Copyright (C) 2015 Published by Elsevier Ltd. All rights reserved.