The unsteady-state, one-dimensional convective mixing equation was solved with finite end boundary conditions for a pulse of tracer admitted into a steady flow, assuming a constant dimensionless mixing factor D *. Radiotracer tests were performed on two continuous wet overflow ball mills, a laboratory mill of 0.3 m i.d. and a pilot-scale mill of 0.91 m i.d. Counting through the mill case enabled the tracer concentration to be measured at L/3, 2L/3 and L for the small mill, and L/2 and L for the pilot-scale mill. The results were consistent with the mixing model with D * = 0.5 and 0.3, respectively, with D * constant along the mill. The mean residence time of water was about 0.85 that of the solid, showing the slurry density in a mill to be higher than that of the feed and product streams.

We carried out wet milling batch tests in a laboratory scale ball mill on a minus 600 μm UG2 ore as feed. In this paper, we aim to investigate the nature of the UG2 ore using a simple method that relies only on the breakage kinetic data and knowledge of basic process modeling skills. Our results show that a two component model more accurately predicts the breakage behavior of the UG2 ore when compared to the homogeneous or true first order model. This paper takes the view that due to the effectiveness of the two component model in describing the milling behavior, it may be hypothesized that the ore might best be described as composing material of two different hardness; termed the ‘soft’ and ‘hard’ components. Our results also show that for all solid concentrations investigated, the rates of breakage of the ‘soft’ component are much higher than for the hard component. We also found that at 20% solid content the difference between the rates of breakage of the soft component (0.15 per second) and that of the hard component (0.0159 per second) is 0.1341 per second, while at 50% solids the difference is 0.111.