Developing an abrasion characterisation test for measuring superficial breakage in comminution

Abstract

Three motivations were presented for conducting this research. Firstly, comminution is primarily concerned with breaking of rocks therefore it is imperative to fully understand the breakage mechanisms involved in size reduction processes to be able to develop reliable and predictive models and consequently facilitate process optimisation. Impact breakage has been investigated extensively in literature as well as at the Julius Kruttschnitt Mineral Research Centre (JKMRC). However, surface breakage is poorly understood despite the fact it contributes significantly to new surface area generation in grinding. Another motivation for conducting this research was a contribution to the Unified Comminution Model (UCM). The UCM is a mechanistic model framework, and thus requires, as fundamental inputs, the response of a rock particle to the common modes of breakage found in comminution, independent of each other. Hence, this research aimed to isolate the abrasion mechanism with an appropriate device and to develop an abrasion characterisation test for measuring superficial breakage in comminution. This was because most of the prior comminution experiments investigating abrasion were conducted in tumbling mills which neither isolated the mechanism nor eliminated secondary breakage. Lastly, Leung (1987) reported that low specific energy impact events can produce similar product size distributions to that of abrasion. But, the abrasion experiments were conducted in a tumbling mill. Hence, the possibility of substituting low energy impact breakage for shear abrasion breakage was investigated as well. Abrasion tests and impact tests were conducted over a range of energies with the same ore type and the results were compared. Surface breakage experiments were conducted with single ore particles and particle beds. The single particle tests included a novel use of the Steel Wheel Abrasion Tester (SWAT) device as well as single impact tests with the JK Rotary Breakage Tester (JKRBT). Batch (bed) experiments were conducted on the bench scale with a planetary mill and the results compared with the single particle outcomes. A novel methodology was followed to produce the results which included the application of an insert to minimise secondary breakage. The results showed that despite the fact that energy directly contributed to the production of product mass during abrasion, the primary driver of mass loss in ore particles was the applied load. It was found that the mass loss rate (g/kJ) was directly proportional to the applied load during the steady state phase of the rock’s response to the abrasion mechanism. This introduced the possibility of establishing an abrasion index for rock particles. The findings also revealed that, at face value, neither low nor high energy, single point, single impact breakage produced appearance functions similar to that of abrasion. Therefore, single impact breakage mechanisms cannot be used as a proxy for abrasion breakage mechanism, at least not for the energy range in which the experiments were conducted (0.005 ‒ 3 kWh/t). The primary outcome of this thesis was the IMLAT (Incremental Mass Loss Abrasion Test). The IMLAT provides the methodology and outputs necessary to characterise a rock’s response to the steady state abrasion mechanism in comminution. However, it was never the aim of this project to develop the models or generate the abrasion index of all ores, but merely to demonstrate how one would go about achieving it. In other words, this thesis paves the road to an abrasion index of ores and a mechanistic abrasion model. Both of which could be meaningful in the comminution context. Results from the planetary mill experiments revealed that a bed of particles can produce appearance functions similar to the IMLAT (single particles). The IMLAT was laborious and produced significant noise pollution. Moreover, the statistical significance of the results was questionable due to the small number of samples in each test. Therefore, it was recommended that future research investigate the possibility of conducting batch tests in a planetary mill as a proxy for the IMLAT. These tests would be fast and simple and therefore easily repeated to improve the statistical significance of the results. Moreover, it would be the first truly batch abrasion characterisation test.