Hydrocyclone Design and Operation: How Geometry Sets the Cut Size

A hydrocyclone does one job with almost nothing inside it: it turns pressure into rotation and lets density decide where each particle ends up. There are no moving parts, no chemicals and no power feed beyond the pump that pushes slurry through the tangential inlet. That simplicity is why the device has survived nearly a century of competition from screens, centrifuges and lamella settlers.

A Dutch idea from the late 1930s that never left the plant floor

The hydrocyclone first appeared in the Netherlands in the late 1930s as a swirling-flow classifier for mineral slurries. The same geometry later proved useful for thickening, dewatering and even sorting, and it has been a fixture in mineral processing and in wastewater pretreatment ever since.

Construction is deliberately plain. A hollow cylindrical section sits on a conical section; the cylinder diameter sets the size class of the unit. Available sizes run from 50 mm to 1,000 mm, and 125 to 500 mm covers most plant installations. A vortex finder drops into the centre of the cylinder from the top plate, the feed inlet enters tangentially, and an apex or spigot opening is left at the bottom of the cone.

Two vortices, one decision

Slurry enters under pressure and is immediately forced into a circular path by the cylindrical wall. Coarse particles carry high inertial centrifugal force, are thrown to the wall, migrate downwards and leave through the bottom as underflow. Fine particles approach the wall more slowly and are swept inward and upward, exiting through the vortex finder as overflow. Gravity contributes, but centrifugal acceleration is what keeps the unit compact.

The feed is delivered tangentially by a sand pump or by static head, at 0.5 to 2.5 kg/cm 2 and a velocity of about 5 to 12 m/s. Higher velocity buys sharper classification up to a point, then turns into wear on the apex and the inlet, which is the trade every operator eventually makes.

Four flow features explain most of the difference between a well-running unit and a poor one:

• Outer and inner vortex. Both rotate the same way but travel in opposite axial directions. The outer vortex carries coarse heavy solids to the apex; the inner vortex carries fine light solids to the overflow.
• Short-circuit flow. Part of the feed is retarded by wall friction, moves up, then inward under the top plate, then down along the outside of the vortex finder and straight into the overflow. It never gets classified, so it drags efficiency down.
• Circulation flow. Fluid migrating from the outer to the inner vortex cannot all leave through the overflow, so the surplus circulates up and down in the annulus between finder and wall.
• Air core. As the spiral radius shrinks, tangential velocity rises and pressure falls below ambient. Air comes out of solution and outside air is drawn in through the discharge openings.

The zero-velocity envelope marks the locus of zero axial velocity and separates the two vortices; for a given geometry its size and shape barely change. The maximum tangential velocity locus marks where particle tangential velocity peaks during inward migration.

Geometry decisions that set the cut

Diameter governs both capacity and cut size, and both grow with it. The feed inlet diameter affects capacity, cut size and efficiency, with a larger inlet giving a coarser cut; it is normally scaled to the unit as D i = (0.2 to 0.26) D. Cone angle trades fineness against capacity: fine classification uses 10 to 15 degrees, coarse classification and thickening use 20 to 45 degrees, and internal resistance climbs with the angle so throughput falls at constant inlet pressure.

The vortex finder is the single most influential dimension because it touches every process indicator. A larger finder raises overflow volume, coarsens the overflow, strips fines from the underflow and lifts underflow concentration; it is sized at D o = (0.2 to 0.4) D. Insertion depth sets the cut as well, shallower being finer and deeper coarser, with h = (0.3 to 0.7) D typical. A thicker finder wall can lift efficiency slightly, cut internal energy loss and add capacity.

The apex is the most wear-prone part of the body. A larger apex gives a finer cut and a smaller one a coarser cut, at d = (0.15 to 0.25) D, and the apex is how operators tune the cone ratio, defined as apex diameter over finder diameter and usually held between 0.35 and 0.65. Feed viscosity coarsens the cut in proportion to its square root and also raises capacity and split ratio.

Operating levers and their limits

Feed pressure is the lever everyone reaches for first, and it is overrated. Raising pressure lifts slurry velocity and centrifugal force, improving classification efficiency and underflow concentration, but the effect on cut size is marginal while energy consumption climbs sharply and apex wear becomes severe. Coarse duties run at low pressure (0.05 to 0.1 MPa); fine and clayey feeds need 0.1 to 0.3 MPa. Feed rate moves the cut in the opposite direction to intuition: a higher rate coarsens it, a lower rate fines it.

Feed concentration matters as much as pressure. High concentration raises viscous drag, coarsens the cut and lowers efficiency. For a cut size of 0.074 mm, a feed concentration between 10% and 20% is appropriate. Particle size distribution around the cut is equally decisive: too little near-cut material leaves coarse particles in the overflow, while too much of it pushes fines into the underflow, and both cases cost performance.

Why it still earns its place in a treatment train

For wastewater duty the argument is footprint. A hydrocyclone removes grit and heavy solids ahead of flotation, filtration or biological units with high capacity per unit volume, classification efficiency of 80% to 90% at fine cuts, low capital cost and low material consumption. Sand and grit removal is the classic application; the same body also dewaters slurry and thickens tailings before they go underground as backfill, and in gravity separation plants it doubles as a desliming device.

The drawbacks are honest ones. Power consumption is relatively high, and wear at high pressure is severe, which is why apex and inlet parts are often supplied in cemented carbide or silicon carbide. In closed-circuit grinding the small working volume leaves no buffer against swings in ore throughput, so control is less forgiving than with a mechanical classifier. Neither problem is fatal. Both simply have to be designed for rather than discovered later, which is the difference between a cyclone that runs for years and one that is rebuilt every season.