How to optimise your plant for different feed types?

Typically, a minerals processing plant (e.g. a concentrator or coal wash plant) is fed by materials that can change from one time to the next based on where the material is coming from in the mine. And different materials have different recoveries (or yields) based on their physical and chemical properties. Traditionally, a metallurgical or process engineer will work with the plant operators to set the operational tactics (the parameter set points) to optimise recovery for a particular feed type. This is done using fundamental metallurgical approaches and experience. It depends largely on the skill set and experience level of the plant personnel and this leads to variability in performance.

Imagine if you could take the knowledge and experience at your site and amplify it 10x? And then have that amplified knowledge available 24/7, no matter who is on shift nor which part of the mine is delivering feed. What if you could provide your site team with an on-demand decision making tool that helps them maximise the value delivered by the plant, no matter what the circumstances?

These were the questions we asked ourselves at Interlate and this led us to develop a tool that determines all the plant set points within seconds to maximise the recovery for a particular feed type. The tool looks at different feed types, considers the plant’s past performance and matches plant set points that will enable optimum performance. This provides a benchmark based on the best performance the plant has ever achieved and provides a foundation to drive even greater performance.

The tools use historical data as a starting point and then applies a hyper-variant machine learning capability to optimise the set points going forward. The advantage that comes from using a data driven approach means the site team can make decisions with extremely high levels of statistical confidence – typically 95% confidence. And as the data collected increases over time, more and more fine tuning is possible until the plant is fully optimised. It is also responsive to real-time data which means that it can adapt to changing circumstances protecting value at all time.

Brett Harries,
Technical Lead, Interlate