An investigation into wear of media in stirred media milling

Abstract

The beneficiation of low grade and finely disseminated ores has increased the adoption of stirred media mills because of their superior energy efficiency compared to tumbling mills. However, the highly dynamic and energy intensive milling environment accelerates grinding media wear, affecting both process efficiency and operating costs. This study investigated the wear behaviour of ceramic grinding media in stirred media mills through combined experimental and numerical (Discrete Element Method, DEM) approaches. The work aimed to identify operating conditions and media interactions that promote wear and to evaluate the capability of DEM to analyse energy transfer mechanisms associated with media wear. Direct wear quantification was not possible due to limitations such as uncharacterised media, the absence of the slurry phase, and the inability to resolve separate normal and tangential forces required to compute shear impact energies. Experimental tests were conducted using a laboratory-scale stirred mill with ceramic beads as grinding media. Specific energy input was varied at 5, 10, and 20 kWh/t, while feed sizes of -1180 + 850 µm and -650 + 425 µm were used to evaluate their influence on grinding performance and wear. Results showed that increasing specific energy input produced finer products, confirming the relationship between energy input and grinding efficiency. Coarser feed exhibited higher power draw and torque due to increased resistance and viscous losses, consuming more energy for the same milling duration. Wear tests performed on the coarse feed for 2 h and 4 h showed measurable bead degradation, with a linear wear-rate trend under the tested conditions. DEM simulations revealed predominantly low-energy collisions through the collision energy spectra, consistent with abrasion dominated milling environments. Predicted power draw values were 13% to 34% lower than experimental measurements, primarily due to the exclusion of the slurry phase, which typically accounts for up to 34% of total power losses in stirred mills. Despite these constraints, DEM reproduced realistic energy transfer trends, complementing the experimentally observed trends. The study confirms that integrating experimental data with DEM analysis offers valuable insights into the mechanisms governing energy dissipation and media wear in stirred milling environments. Further work should employ DEM software capable of resolving stress fields for integration into energy based wear models, include detailed characterisation of ceramic media properties, and incorporate breakage models to establish a more comprehensive representation of the grinding process.