Coupling of wear sensor measurements with numerical modelling for grinding mill charge dynamics prediction

Abstract

Accurate knowledge of charge dynamics in grinding mills is essential for optimising throughput, energy efficiency, and liner life. However, progressive liner wear continuously alters the internal geometry, making realtime charge prediction challenging. This article presents a novel methodology that couples in-situ wireless wear sensors with numerical modelling to predict mill charge dynamics throughout the liner lifecycle. Sparse point-based thickness measurements from embedded sensors are combined with discrete element method (DEM)-derived wear intensity distributions to reconstruct progressive global liner profiles via a topological evolution algorithm. The reconstructed worn geometry drives two complementary modelling streams: a continuum power draw model for total charge level estimation, and a pre-computed DEM database for shoulder and toe angle prediction. The methodology was validated on a 36ft SAG mill over a full 188-day liner campaign. Predicted transient total charge and toe angles showed strong agreement with an independent MillSense instrumentation. The framework enables condition-based reline scheduling and adaptive control of mill operation, representing a significant advance towards digital tools enabled grinding circuit optimisation.