Using artificial intelligence to improve manganese output and reduce waste at GEMCO

Through Innovate32, our approach to better enable innovation at South32, our Australia Manganese operation has developed a new artificial intelligence advisory system. The Advanced Process Control (APC) advisory system helps to improve overall recovery and output from each ton of mined manganese ore processed at the operation. 

Manganese recovery in the concentrator - the area of the processing plant that recovers manganese from mined ore - is primarily driven by two variables: the quality of the feed and density control. Historically, density control has been achieved by infrequent interventions from the control room operators on the advice of geologists. The APC advisory system uses analytics and machine learning to make near real-time density setpoint recommendations to the operator, who can implement them to improve the yield in the concentrator. The algorithm used in the advisory system is based on sophisticated statistical relationships that have been derived from the control variables, blend characteristics and product quality.   

Since deployment of the advisory system in FY24, we have seen an increase of 19,000 tonnes in the production of saleable product and a 12 –15 per cent reduction of manganese content in waste.  

Over time we will work on refining the advisory system, which will become ‘smarter’ by capturing learnings and operator feedback.  

Using the learnings from this initiative, we are actively assessing and progressing similar value-adding opportunities for artificial intelligence across our business.