A typical challenge in heap leaching operations is dealing with high fines contamination coming from the crushing section. The high fine fraction in stacked modules results in poor leaching efficiency, localized flooding due to poor permeability, and high acid consumption.
The IntelliSense.io solution to address the root cause of high fine fraction in the content of the stacked module is the Two-week inferential fine fraction prediction Model.
It is a classification model used to predict the probability of having the event of a high fine (above a certain threshold), two weeks in advance. based on a material transport model that uses different data sources such as from Data mine or mine plan and/or lab reports.
The model explains material influences that might be influencing the model output(Probability of high fines event occurring).
This Virtual Sensor can be visualized using a brains.app dashboard. This is a reference dashboard showing the high fines probability virtual sensor with the key variables:
Navigate to access the Two-week inferential Fines Prediction dashboard in your brains.app.
Select historic data and choose the period (normally two weeks ahead of the current mine plan).
Observe the Line graph to see how the prediction of high fines probability is trending against the threshold(red line).
See the major material property influences contributing to the prediction i.e. Basically, when the blue line is above the red line, this indicates that there is a high probability of producing fine material from crushing that exceeds the threshold.
Potential action to take(Blending ore), in order to mitigate this predicted event
This allows you to:
A reduction of localized flooding events on leaching pads
Potential OPEX savings on acid consumption
A higher metal recovery due to improved leaching kinetics by changing mine planning or ore blending.