Hydrocyclone Classification Roping Case Study

This case study follows a realistic mineral processing upset: a grinding circuit begins sending coarse material to flotation because several hydrocyclones move from spray discharge to roping discharge. The symptoms first appear as lower recovery and unstable cyclone pressure. The root cause is not a single bad sample. It is a coupled operating problem involving slurry density, water addition, cyclone pressure, apex condition, circulating load, and delayed operator evidence.

The case teaches how a processing engineer connects plant measurements to physical classification behavior. A cyclone cluster can be mechanically intact and still make the wrong separation if feed density, pressure, water balance, and geometry move outside the operating envelope.

Case Summary

Item	Engineering relevance
System	Ball-mill grinding circuit with hydrocyclone classification before flotation.
Normal objective	Send adequately ground overflow to flotation while returning coarse underflow to the mill.
Event	Several cyclones show roping underflow and coarse overflow.
Main indicators	Lower cyclone pressure, high feed density, high circulating load, coarser overflow P80, lower recovery.
Hidden weakness	Density meter bias and process-water valve restriction masked the water-balance change.
Corrective action	Restore process water, reduce fresh feed temporarily, isolate suspect cyclones, inspect apexes, recalibrate density measurement, and reset operating limits.

The central engineering question is:

Did the cyclone cluster still provide the classification cut required by the downstream flotation circuit?

During the event, the answer was no. The overflow became too coarse, circulating load rose, and recovery fell.

Initial Operating Envelope

The normal operating envelope for the circuit is:

Variable	Normal target
Fresh ore feed	300\ \text{t/h} dry solids
Cyclone underflow return to mill	600\ \text{t/h} dry solids
Circulating load ratio	200\%
Cyclone feed solids concentration	62\% by mass
Cyclone feed pressure	95\ \text{kPa} plus or minus 10\ \text{kPa}
Overflow particle size	P80=150\ \mu\text{m}
Flotation recovery at target grind	88\%
Normal underflow pattern	open spray, not a solid rope

The circuit can tolerate ore hardness variation if density, pressure, and water addition remain controlled. It cannot tolerate a sustained shift that sends coarse, poorly liberated particles to flotation.

Event Timeline

1. A maintenance shift replaces a section of process-water piping and reopens the circuit.
2. The control room sees cyclone pressure drifting between 65 and 85\ \text{kPa} instead of the normal 95\ \text{kPa} target.
3. The density meter reports acceptable density, but manual samples later show higher solids concentration.
4. Operators notice roping discharge from several cyclone underflows rather than a stable open spray.
5. Overflow samples show P80 rising from 150\ \mu\text{m} to about 230\ \mu\text{m}.
6. Flotation recovery falls from 88\% to 82\% on the same ore blend.
7. Mill power rises and the circuit becomes harder to stabilize because circulating load increases.

The early alarm looked like ordinary ore variability. The later evidence showed a classification failure.

Symptom 1: Circulating Load Increase

Circulating load ratio is:

Under normal conditions:

During the event, the underflow return rises to:

so:

Total mill solids feed changes from:

to:

The mill is now trying to process one-third more solids internally without an equivalent increase in useful throughput. High circulating load can be acceptable in some designs, but here it coincides with coarse overflow, pressure instability, and roping. That combination indicates poor classification, not deliberate high-recycle operation.

Symptom 2: Slurry Water Deficit

The cyclone feed during the event has about:

The target solids concentration is:

At target density basis, total slurry mass should be:

Target water flow is:

Manual samples during the event show about:

Then total slurry mass is:

Event water flow is:

The estimated water deficit is:

That is not a small instrument detail. A water deficit of this size changes slurry viscosity, cyclone pressure drop, apex discharge behavior, and classification sharpness. It can push cyclones toward roping.

Symptom 3: Coarse Overflow and Recovery Loss

The flotation circuit expects overflow near:

During the event:

The relative increase is:

Coarse overflow means some valuable mineral remains locked in gangue. If feed grade is stable, a recovery drop from 88\% to 82\% can be estimated as lost recovered metal:

For a copper feed of:

lost recovered copper is:

The exact value depends on assays, sampling, and time delay, but the calculation shows why classification faults become economic faults quickly. The plant is not merely producing a different particle size; it is losing recoverable metal.

Failure Mode Analysis

The investigation identifies several interacting failure modes.

Failure mode	Evidence	Consequence
Process-water restriction after maintenance	Valve travel check and flow trend show reduced water addition.	Feed density rises and cyclone pressure falls.
Density meter bias	Manual samples disagree with online density during air entrainment.	Operators underestimate solids concentration.
Roping underflow	Visual underflow pattern and high underflow density.	Poor separation and unstable return load.
Apex wear or partial blockage on two cyclones	Inspection finds worn and scaled apex liners.	Individual cyclones behave differently from cluster average.
Delayed particle-size feedback	Lab P80 result arrives after the pressure/density upset.	Coarse overflow persists long enough to affect recovery.
Control response based on pressure alone	Pressure controller cannot distinguish water deficit from ore hardness or cyclone geometry.	Wrong corrective action is possible.

No single measurement proves the cause. The case becomes clear only when water balance, density samples, pressure trend, underflow pattern, P80, and recovery are interpreted together.

Corrective Engineering Decision

The operating team takes a staged response:

1. Reduce fresh feed from 300\ \text{t/h} to 250\ \text{t/h} to protect the mill and flotation circuit.
2. Restore process-water flow and verify valve position after maintenance.
3. Isolate cyclones with abnormal underflow pattern.
4. Inspect and replace damaged apex liners.
5. Recalibrate density measurement against manual timed samples.
6. Re-establish cyclone pressure at 95\ \text{kPa} plus or minus 10\ \text{kPa}.
7. Hold the circuit in stable operation until overflow P80, underflow pattern, density, and recovery confirm recovery.

The decision avoids a common mistake: increasing mill feed to recover tonnage while classification is already failing. More feed would have increased circulating load and pushed more coarse material downstream.

Validation After Correction

After water flow is restored and two apex liners are replaced:

Variable	Event condition	Corrected condition	Target
Cyclone feed solids	70\%	63\%	about 62\%
Cyclone pressure	65 to 85\ \text{kPa}	92 to 100\ \text{kPa}	95\pm10\ \text{kPa}
Circulating load	300\%	210\%	near 200\%
Overflow P80	230\ \mu\text{m}	155\ \mu\text{m}	150\ \mu\text{m}
Underflow pattern	rope on several cyclones	open spray	open spray
Flotation recovery	82\%	87.5\%	about 88\%

The corrected condition is acceptable for restart to normal feed because the classification indicators move together. One good value would not be enough. Pressure, density, P80, underflow pattern, and recovery all need to support the same conclusion.

Transferable Lessons

Hydrocyclone classification is a system behavior. It depends on feed solids, water balance, pressure, geometry, apex condition, vortex finder condition, ore particle-size distribution, viscosity, and control response.

The main lessons are:

• Roping is a classification alarm, not only a visual oddity.
• Cyclone pressure alone is not enough; pressure must be interpreted with density and underflow pattern.
• Density instruments need manual validation when slurry aeration, wear, or calibration drift is credible.
• Circulating load should be reviewed with overflow size and downstream recovery, not treated as a standalone KPI.
• Maintenance work on water circuits can create metallurgical losses if return-to-service checks are weak.
• Corrective action should protect downstream recovery before chasing lost throughput.

Release Criteria

Before returning to full feed rate, the circuit should meet:

1. Stable cyclone pressure inside the approved range for a defined period.
2. Cyclone feed density confirmed by online and manual samples.
3. Underflow discharge pattern visually checked or instrumented for abnormal roping.
4. Overflow P80 within the flotation feed requirement.
5. Circulating load inside the operating envelope.
6. No known blocked, worn, or mismatched apex liners in the active cluster.
7. Recovery, concentrate grade, and tailings grade reconciled against the corrected condition.
8. Shift handover notes explaining the water valve, density calibration, isolated cyclones, and residual monitoring actions.

Engineering Closeout

A defensible closeout statement is:

The grinding-classification circuit upset was caused by a water-balance and cyclone-condition fault that drove several cyclones into roping. Manual density samples, pressure trends, underflow observations, increased circulating load, coarse overflow P80, and recovery loss support the diagnosis. Restoring water flow, replacing damaged apex liners, recalibrating density measurement, and reducing feed during recovery returned the circuit to the approved operating envelope.

This is the useful engineering conclusion: the plant did not have a generic recovery problem. It had a classification stability problem with measurable hydraulic, metallurgical, and operational evidence.