Fine grinding is becoming an integral part of mineral processing plants to liberate valuable minerals from fine-grained low-grade ore bodies. Stirred mills are becoming recognized by the industry as a more efficient technology selection compared to the ball mill for fine grinding (< 100 µm). The increasing number of stirred mill installations in mineral processing concentrators has necessitated a process model development that enables full circuit modelling and simulation. Many researchers have developed mathematical models to assist in the design and optimization of the stirred mill operation including gravity induced stirred mills such as the Metso Vertimill and the Nippon Eirich Tower Mill.
Most of the developed models lack sufficient responses to changes in process conditions. Furthermore, there is an internal classification effect that is previously mentioned in several publications but was not quantified through experiments. Therefore, a research program was designed and carried out to describe and model different sub-processes related to the operation of the gravity induced stirred mill. The sub-process models (i.e. grinding and internal classification) were combined into a single model structure to represent the operation in a gravity induced stirred mill.
During this research, comprehensive test work was carried out in a batch gravity induced stirred mill using Cu-Au ore and limestone to evaluate the effect of operating conditions (i.e. specific energy consumption, slurry density, grinding media size and stirrer tip speed) on particle size and fines generation (-75μm). The test data showed a finer product size when the mill operates at higher specific energy, lower slurry density, smaller grinding media size and higher stirrer tip speed. The test work also identified that attrition breakage mechanism is predominant in the gravity induced mill. Sub-process models that relate the selection function to the process operating conditions were developed. The internal classification effect on mill operation was also measured through test work by using industrial grade silica to identify its effect on the mill operation. The result showed that particle classification takes place inside the gravity induced mill in certain favorable conditions. Models were developed to link the classification parameters and the mill operating conditions to be included in the combined model structure.
A process model (using a time-based population balance technique) was developed, integrating individual sub-processes models such as breakage, selection and classification functions. The model was validated with industry survey data in different process conditions. The validation result showed that the developed model was capable to predict the mill product size distribution when mill power, feed rate, slurry solids concentration and grinding media size are varied.
The thesis has successfully developed a process model of the gravity induced stirred mill with predictive capability. Moreover, it has developed an in-depth understanding of the laboratory scale gravity induced mill behaviour and its breakage mechanism. The inclusion of the classification with the breakage process into a single model structure is novel in comminution process modelling.