Wear prediction is important in the development and optimisation of liner design, and for helping to manage reline strategies. DEM is able to predict wear and profile evolution of liners over the life cycle for specific sets of conditions. However, real operation conditions often vary - usually in ways that are not known in advance, which limits the ability to use such model predictions in managing associated wear. In this study, we use a specific case of a tumbling mill to explore issues around the variability of wear caused by different operating conditions within the applicable range for the mill, and consider how these can be included in a modelling framework in order to assist with mill liner management. Crown Copyright (C) 2014 Published by Elsevier Ltd. All rights reserved.