Flotation Circuit Optimisation

Application

Optimising individual flotation cells improves performance locally, but recovery is ultimately determined by how the rougher, scavenger and cleaner stages work together. A change in one stage propagates through the circuit, and challenging or variable mineralogy — low-grade and altered ore in particular — can erode recovery in ways that no single cell controller can fully arrest.

Minealytics Flotation Circuit Optimisation is a whole-of-circuit intelligent system. It integrates all flotation stages and their variables into one dynamic, predictive model and provides real-time control recommendations for operating variables across the entire circuit, so the stages are coordinated toward a common objective rather than tuned in isolation.

Solution

The system models the circuit as a connected whole and uses that model to recommend and adjust operating parameters in real time:

• Dynamic simulation: a predictive model of the full circuit simulates how the stages respond to changes before they are made.
• Automatic parameter adjustment: operating variables such as airflow, reagent dosage and cell levels are adjusted automatically to keep the circuit at its optimum.
• Arresting recovery loss: by anticipating the circuit’s response, the system helps arrest recovery loss under challenging, variable mineralogy, including low-grade and altered ore.

Because the model spans rougher, scavenger and cleaner stages together, recommendations account for the downstream consequences of upstream changes rather than optimising each stage against the others.

Integration

Flotation Circuit Optimisation is designed to work with the infrastructure an operation already has. It integrates with existing DCS and PLC systems and with expert systems, and it incorporates real-time sensor data and on-stream analyser measurements to ground its recommendations in current circuit conditions.

The system is built to be operator-friendly, presenting recommendations in a form that supports rather than supplants operator judgement, and it is scalable and replicable across operations so a proven configuration can be rolled out to other circuits.