The performance of a grinding mill can be greatly influenced by the slurry hold-up inside. As more and more autogenous (AG) and semi-autogenous (SAG) grinding mills are operated in closed circuit with cyclones, the limitations that can be imposed by this are becoming increasingly apparent. A literature review was performed at the Julius Kruttschnitt Minerals Research Centre (JKMRC) during the early 90's. This revealed that relevant literature on the subject was almost non-existent. Hence a research program was initiated at the JKMRC to study the variables which affect the transport and hold-up of slurry inside a mill.
A literature review conducted at the beginning of this project revealed that previous investigators neglected to study the effects of mill aspect ratio on the relationship between flowrate and hold-up. Another aspect that has been neglected in the past is the effect of slurry rheology on the previously mentioned relationship.
This research project was conducted with the aim of exploring the effect of mill aspect ratio and slurry rheology on the relationship between flowrate and hold-up. The investigations were conducted at the laboratory and pilot scales and experiments were performed using water and slurries. As well as conducting dynamic experiments, tests were performed using a static bath. This was designed to measure the axial water profiles and how these are affected by design and operating conditions.
Experiments were conducted at the laboratory scale using a variable length mill, thus allowing experiments to be conducted at three aspect ratio (D/L) configurations: high, square and low. The experimental conditions for each aspect ratio configurations included five levels of flowrate, three charge levels and three mill speeds. These were randomised to avoid the effects of unwanted but significant variables, such as, time dependency. Also two repeat tests were performed for each experimental condition.
Experiments at the pilot scale were also conducted using three aspect ratio configurations, with the same number of levels for each variable. Due to time constraints the experimental program was designed using a Central Composite Rotational Design (CCRD). The use of this technique reduced the numbers of experimental conditions significantly, whilst still producing a statistically valid set results.
The analysis of the experimental results revealed that aspect ratio did have an effect on the flowrate - hold-up relationship under both dynamic and static conditions. It was found that mean hold-up tended to decrease with mill aspect ratio. Static bath experiments later revealed that the decrease in hold-up was caused by a decrease in the mean level of water inside the mill as the mill length was increased. However the maximum level of water increased with increasing mill length.
Experiments were also conducted using slurries. The solids in the slurry comprised a mixture of minerals containing roughly 60% by weight of a nickel concentrate and 40% by weight of cyclone underflow from the Mt. Isa Copper Concentrator. This slurry showed pseudoplastic with yield behaviour. The results showed that the fractional hold-up appeared to increase linearly with the apparent viscosity of the slurry. It was also shown that hold-up increased with charge level in a non-linear fashion.
The flowrate - hold-up relationship developed by Latchireddi (2002) was used to determine if this model was capable of predicting the results obtained from the experiments. This exercise showed the Latchireddi model was only able to predict the results for a high aspect ratio configuration. It over-predicted the hold-up for the square and low aspect ratio configurations.
The Latchireddi model was modified to take into consideration the effects of mill aspect ratio. Non-linear least square techniques were used to incorporate an aspect ratio term (AR). The new model parameters g1 to g7, were fitted using a total of 633 data points collected during the experimental stages of this project.
Two further models were constructed, the Maximum Height model and the Maximum Fractional Hold-Up model. Their aim was to correlate the maximum height of the water level and the fractional maximum hold-up corresponding to the maximum height of the profile, to the mean hold-up. These models were constructed from data collected at the laboratory scale only and therefore, experiments should be conducted at pilot and industrial scales in order to validate them.