Cyclones in Iron Ore Processing: More Than Meets the Eye

When a cyclone underperforms in an iron ore plant, the instinct is to blame the cyclone itself. But what if the real culprit is hiding upstream — or downstream — in plain sight? In this blog, we unpack insights from Multotec's webinar Cyclones: More Than Meets the Eye presented by Ernst Bekker, which challenges conventional thinking and reveals how seemingly minor issues can quietly drain millions from the bottom line.

The webinar explores cyclone performance in two critical iron ore processing environments: dense media separation (DMS) for hematite ores and hydrocyclone classification in magnetite milling circuits. The central message is clear: even perfectly sized and selected cyclones can fail to deliver optimal results if surrounding process conditions — media quality, sampling accuracy, feed preparation, mechanical alignment, or circuit design — aren't properly optimised. Through real-world case studies, the presentation demonstrates how small deviations can cost operations tens of millions of dollars annually, and why experienced suppliers look beyond the cyclone to diagnose performance issues.

As higher-grade direct-shipping ores become scarcer and environmental pressures mount, iron ore processing is becoming increasingly critical. Processing upgrades are now essential to deliver higher Fe content, reduce impurities like silica, alumina, sulphur, and phosphorus, and lower the carbon footprint of steelmaking.

How does a dense media separation plant operate?

At its core, DMS exploits density differences between valuable ore and waste rock. The principle is elegantly simple: introduce a mixture of particles into a liquid medium with an intermediate density, and light material floats while heavy material sinks. Adjust the liquid density, and the split is controlled.

In an industrial plant, a DMS system has three main sections. Feed preparation handles screening, sizing, and controlling slimes and fines — the first line of defence against poor performance. The separation section, typically a dense media cyclone, is where gravity separation happens at high throughput. Finally, the media recirculation system is critical because the separation medium, usually ferrosilicon, is expensive and must be recovered efficiently.

Cyclone selection involves careful consideration of input values like yield targets, mass balance, and design parameters, along with constraints such as top size, breakaway size, number of cyclones, and spigot requirements. If the cyclone is selected correctly with accurate information, it should perform well. But that's only the beginning of the story.

What constraints are faced in dense media separation?

In a lab test, waiting indefinitely for a perfect density split is possible, a clean vertical line separating floats from sinks. In a DMS cyclone, residence time is short, so that sharp separation becomes a slope, and particles near the cut density get misplaced. This is measured by the EP value (Probable Error), which represents the spread of the partition curve. A lower EP means better separation efficiency and less misplacement of material.

The webinar shared an evaluation of two ore bodies — one with low near-density material, which is easy to separate, and one with high near-density material, which is difficult to separate. Near-density material is everything sitting close to the cut density. For example, if cutting at 3.5, all material between 3.4 and 3.6 is considered near-density. Poor separation efficiency had modest impact on low near-density ore but massive impact on high near-density ore, where much more material sits in that vulnerable zone. The economic difference between good and poor separation efficiency can reach millions of dollars.

But even with good separation efficiency, there's another hidden constraint: cut density drift. Optimal DMS performance requires both a low EP and operating at the correct cut density, the point where a particle has a 50/50 chance of reporting to overflow (floats) or underflow (sinks). SCADA systems might appear to keep operations on target, but one case study revealed the SCADA setpoint in one band while tracer tests showed the actual cut density had drifted elsewhere. The culprit was media instability. New ferrosilicon behaved differently than "in-process" media, causing settling rate changes and density control problems.

The impact is staggering. A cut density shift of just 0.03 (from 3.315 to 3.345) resulted in yield drops of 2-3%, Fe quality changes of 0.25-0.4%, and economic losses of approximately US$ 13-25 million depending on ore characteristics. The solution lies in treating media quality and stability as a first-class control variable. Operators need to monitor media characteristics including stability, rheology, and settling behaviour, and control slimes (−45 µm (micron)) tightly, especially critical when moving toward ultra-high density DMS, where operating densities of 3.4 to 4.0 create a much narrower operating window.

There's also the challenge of measurement accuracy. If sampling and lab systems can't reliably detect a 0.4% Fe quality shift, money may be lost without anyone knowing. The webinar emphasised that 80% of sampling error comes from the sampling activity itself, not the analysis. Robust sampling systems and protocols are essential before blaming the cyclone.

Finally, mechanical and operational issues like misalignment, internal steps, turbulence, inconsistent spigot discharge, and installation errors can all destabilise performance. The webinar shows images of two cyclone spigot discharges on the same distributor behaving completely differently, while operations continued. When cyclones on the same module behave differently, it indicates that the separation process is not optimised.

How to achieve optimal energy efficiency in magnetite grinding processes?

The webinar presented a case study where Multotec supplied cyclones for a magnetite circuit with primary, secondary, and tertiary milling stages. The target was to process approximately 400 t/h and upgrade from 34.5% Fe to 66% Fe. The plant couldn't hit 66% Fe, so operators tried pushing finer grind by increasing −71 µm material to improve liberation. They added water to the hydrocyclones to drive a finer cut, which did increase −71 µm content, but only to a plateau.

Re-examining the data revealed something critical: the −71 µm fraction increased when feed rate decreased, yet grade didn't respond. This pointed to an energy limitation. The mill data showed a ceiling at around 96% −71 µm. The cyclone cannot produce a finer overflow than the mill can produce, so no amount of water addition or cyclone tweaking would overcome this fundamental constraint.

Comparing the old and new circuit designs revealed the new design had omitted "slimers", equipment used to remove silica from the circuit. When the magnetic separator concentrate was examined, silica entrapment was found. The material was ground fine enough, but silica couldn't escape the circuit. Without the slimers, grinding as fine as possible wouldn't improve grade.

The value of looking beyond the cyclone

A single 610 mm diameter cyclone treating 134 t/h at 60% yield can generate over US$ 9 000 per hour or US$66 million per year. Yet when performance drops, losses accumulate fast. Experienced suppliers understand that cyclones sit at the centre of a complex process with multiple interdependent flow streams. When something goes wrong, the cyclone often gets blamed, but the real issue may lie elsewhere.

Watch the full webinar here to explore these case studies in detail and discover how small process changes can unlock millions in value.