The Use of Acoustics to Improve Load Estimation in the Cannington AG Mill

The estimation of the grinding mill charge weight is vital to ensure its efficient performance. Mill underloading or overloading can cause losses through decreased production or unplanned down time. Metso Minerals Process Technology Asia Pacific and South America (MMPT) has developed two systems for measuring mill load. The first is an acoustic system (SmartEarÖ) and the second is an online model-based charge estimation system. The aim of both systems is to facilitate better operation of grinding mills. SmartEarÖ performs analysis on an audio signal from the mill and allows each different frequency component to be viewed, analysed and used in process control for load estimation. By isolating different frequencies, SmartEarÖ can give an estimate of mill load. The online dynamic mill model uses measurable mill parameters (such as power) to predict the current total load. Kalman filtering can also be utilised to further increase the accuracy of the model. Both of MetsoÆs systems have been installed at BHP BillitonÆs Cannington mine to measure the load within their autogenous (AG) mill. In this paper, the methods of mill load estimation are reviewed and compared to the bearing pressure reading, which is the current method used by Cannington to control mill load. The redundant estimates of mill load provide increased confidence and consequently more robust grinding operation under a wide range of conditions.