Optimisation of an industrial scale ball mill using an online pulp and ball load sensor
Master Thesis
2013
Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
University of Cape Town
Department
License
Series
Abstract
The secondary milling circuit at Waterval UG2 Concentrator had undergone a circuit change with the commissioning of the IsaMill, a horizontally stirred mill, in parallel with the secondary ball mill. The operation treats the PGM bearing UG2 ore type and produces a final concentrate enriched with PGM's. The concept was to treat the finer silicate rich fraction in the IsaMill and the coarser chromite rich fraction through the ball mill. This circuit is typical of a UG2 plant in which maximum silicate with minimal chromite breakage is targeted. As a result of the circuit change an opportunity for optimisation around the industrial scale ball mill was considered for this study. Of concern in this study were new operating conditions for the mill in the changed circuit at which improved performance could be obtained. Another objective was to investigate if a difference in breakage response for the silicate and chromite fractions could be identified for different operating conditions in the ball mill. The secondary mill at Waterval UG2 Concentrator was already fitted with an online ball and pulp load sensor, the Sensomag. The information obtained from the sensor is in the form of shoulder and toe positions for the ball and pulp filling in the rotating ball mill. The mill was surveyed at various ball filling and mill % solids conditions and information from the online sensor was used to understand the mill performance, particularly with regards to mill load behaviour. Hence a final objective was to demonstrate that the information obtained from the online sensor could be related to mill operating conditions. The sensor output was envisioned to eventually form part of the mill control philosophy. Samples were taken of the mill feed and discharge streams at the different operating conditions and analysed for grind as well as PGM and Cr₂O₃ content. The majority of the PGM's in the UG2 ore are in the silicates and thus the PGM distribution results would indicate the amount of breakage in the silicate fraction. Cr₂O₃ is used as an indicator of the chromite content in UG2 ore. In order to identify optimum mill performance the results were analysed using different measures which include general grind, particle and species distributions, reduction ratios, sieve efficiencies and specific energy. By comparing the results the differences and limitations of certain techniques were identified. It was found that the mill performance varied at different operating conditions. The optimum ball filling was found to be around 30%, which is similar to site operational target. The optimal % solids for this mill however seems to be higher than what the mill is typically operated at. No peak in % solids for mill performance was obtained. Scope exists to determine how far from the investigated maximum of 75% solids (by mass) does the optimum in-mill density lie for this mill. Thus new optimum conditions in terms of % solids do exist for the mill in the modified circuit. Results also showed that the size reduction of the silicates increased with an increase in mill % solids and ball filling degree. For chromite, the mill % solids did not appear to have any effect at low ball fillings, but a slight shift was observed at the higher ball fillings tested. The trend suggests that the size reduction of both silicates and chromite increased with an increase in ball filling, albeit at different rates. Finally, the test work has demonstrated that the online sensor outputs can be related to mill performance. Differences in shoulder and toe positions for the ball and pulp loads were distinct between operating conditions. Improved grind performance was observed at conditions that resulted in lower free pulp angles. Thus the sensor could be used as a control tool to identify and maintain optimum mill operational conditions. The Sensomag should be incorporated into a mill controller that looks at more than just mill ball filling. Conditions that result in optimum mill efficiency can be identified and the mill may be controlled using the sensor data. It is recommended that the mill continue to be run at 30% ball filling and at higher mill % solids than the maximum reached in this work.
Description
Keywords
Reference:
Keshav, P. 2013. Optimisation of an industrial scale ball mill using an online pulp and ball load sensor. University of Cape Town.