Today
Recovery and grade are observed after the fact
- Outcomes are read from lab assays once the tonne has moved.
- Setpoints are adjusted on judgement and recent trends.
- A shortfall is found once it is already in the concentrate.
By the time the assay returns,the feed has already changed.
Recovery and concentrate grade move with head grade, particle size and solids flowrate, but reagent dosing and other setpoints are usually adjusted periodically against lagging assay results. INSUS predicts circuit condition from feed data the plant already records and puts setpoint recommendations in front of the metallurgical team to review.
01
Feed data as a leading indicator, not only an after-the-fact assay.
02
Read-only on your SCADA and LIMS. No writes to the circuit.
03
Every recommendation reviewed by the metallurgical team.
The problem
Feed changes continuously.Setpoints wait for the next assay.
Recovery and grade respond to feed characteristics that shift continuously: head grade, particle size, solids flowrate and hardness. In differential circuits treating complex sulfide ores, each stage concentrates a different mineral against its own targets. Reagent dosing, air rate and levels are still adjusted periodically on judgement and recent assay trends, so a shortfall is often identified once it is already in the concentrate.
01Head grade and size keep shifting
02Setpoints adjusted periodically
03Recovery confirmed later, from assays
- 01Head grade and size keep shifting
- 02Setpoints adjusted periodically
- 03Recovery confirmed later, from assays
The shift
From periodic adjustmentto predictive setpoint guidance.
The metallurgical team still makes the call. What changes is that feed data becomes a leading indicator of circuit condition instead of something reconciled only after the assay.
With INSUS
Circuit condition is predicted from the feed
- Condition is predicted from feed data on this circuit's own history.
- Setpoint recommendations are generated for the team to review.
- Shortfall risk is flagged before it is fully in the concentrate.
How it works
From feed datato a setpoint recommendation.
This fits OptiStudio as quality and process decision support beside existing flotation controls. Integration is read-only. It has no authority to change setpoints.
Connect
Link feed, circuit control and assay data
Connect feed head grades, particle size and solids flowrates; flotation SCADA and DCS setpoints including reagent dosing, air rate and level control; the process historian; and LIMS recovery and grade results. All read-only.
Understand
Learn how this circuit responds to its feed
From historical data, learn how feed characteristics and setpoints have translated into recovery and grade for this plant's ore and circuit, stage by stage.
Predict
Predict circuit condition from current feed
Predict the expected recovery and grade trajectory from current feed data, and flag a likely shortfall before it is fully reflected in concentrate output.
Advise
Search setpoints for better recovery and grade
Search combinations of reagent dosing and related setpoints such as air rate and level, using the predictive model to estimate their effect. The metallurgical and operations team decide whether to apply the recommendation.
Verify
Check every prediction against the assay
Log predictions and recommendations against actual recovery and grade results. Track prediction accuracy openly and refine the model against this plant's ore and circuit over time.
Getting started
Start with one circuitand the assays you already run.
For minerals plants that already record flotation feed, setpoints and LIMS results. Delivered through OptiStudio as recovery and grade decision support. Expert Services can help with data assessment and validation. Day one is advisory and read-only. Each model is calibrated to one plant's stage, ore and equipment.
- What we need. Historical head grades, feed particle size, solids flowrate, reagent dosing and other setpoints, and matching recovery and grade assays across a representative range of feed conditions.
- What you see first. One circuit or differential stage running predictions and recommendations alongside existing operating practice, with a validation report comparing predicted and actual recovery and grade.
- Where the limits are. Setpoint changes stay a human decision. Where head grade or particle size comes from periodic sampling rather than online measurement, predictions are only as current as those inputs.
Related for minerals
Other mineralsstarting points
Flotation sits beside grinding energy and Strategic Project documentation on the same minerals operation.
- 01Comminution Energy & Stability OptimizationUpstream of the bank. Stabilising the grinding circuit changes the feed particle size flotation depends on.
- 02CRMA Strategic Project ComplianceOrganise Strategic Project documentation against Critical Raw Materials Act criteria rather than optimising a circuit.
- 03Minerals industryThe wider picture for minerals processing: comminution energy, flotation recovery and grade, and the metallurgical record.
Questions metallurgical teams ask us
Feed characteristics shift continuously while setpoints are adjusted periodically against lagging assays. This use case predicts circuit condition from feed data and recommends setpoint changes for metallurgical review.
Historical feed head grades, particle size, solids flowrate, setpoint data and matching recovery and grade assays. No new hardware. Integration is read-only with SCADA, DCS, historian and LIMS.
It predicts the expected recovery and grade trajectory from current feed data and can recommend reagent dosing and related setpoints such as air rate and level control.
Not in this use case as scoped. Setpoint changes remain a human decision. The metallurgical team reviews every recommendation before anything is applied.
Against the pilot circuit's own historical recovery and grade baseline. Predictions are logged and compared with actual assay results. External study figures are not treated as promised results for your plant.
Through OptiStudio as process and quality decision support for the flotation circuit. Expert Services can support data assessment, integration and validation with your metallurgical team.
Models are calibrated per circuit, stage and ore, so they do not transfer between plants unchanged. Predictions are only as current as the feed inputs behind them, especially where sampling is periodic rather than online.
Flotation circuit optimisation
Read the feed before the concentrate tells you.
Start with one circuit and the feed, process and assay data you already record. Test whether predicting circuit condition early would have changed the recovery and grade decisions your team made last quarter.