Gas cyclones: what wear at the inlet scroll means that cone wear does not

A cyclone inspection that reports only a minimum remaining thickness has thrown away the diagnosis. Two vessels can show the same thinnest reading for entirely different reasons: one because the gas enters too fast against a duct bend, the other because separated dust is being held in the cone and re-entrained. The number is identical; the corrective action is not.

Divide the inspection into the inlet and upper barrel, the mid barrel, the cone, and the dust outlet, and report a finding for each. Four observations cost little more than one and carry incomparably more information.

Loss concentrated at the inlet, on the opposite wall or in the first turn of the scroll usually reports the duct rather than the cyclone. A bend too close to the entry, a damper that has been left partly closed, or an unequal split between parallel units all deliver the dust to one part of the wall instead of distributing it around the barrel.

These faults are correctable upstream, generally at lower cost and with a benefit to separation efficiency as well as to wear. Lining the affected patch harder treats the bruise and leaves the cause running, and the next inspection will find the loss again in the same place.

Wear low in the cone and around the dust outlet is rarely caused by the incoming gas — the material there has already been separated and slowed. It is caused by that material failing to leave. A restricted outlet, a build-up bridging the cone, or an airlock passing air upward keeps separated dust circulating in the narrowest part of the vessel, where it grinds continuously against the wall.

The same fault also destroys separation efficiency, because re-entrained dust leaves through the outlet it was collected to avoid. Wear and performance decline together here, which is why cone wear should trigger an examination of the discharge before any lining decision is taken.

Gas cyclones frequently run hot, and the surface temperature governs the retention method long before it troubles the ceramic. Just as important is the lower bound: a surface that falls below the dew point during a start-up, a shutdown or a low-load period will condense, and a damp deposit behaves nothing like a dry one. It sticks, it bridges the cone, and in some gas streams it turns mildly acidic.

Record the full temperature history including the transitions, not just the steady-state figure. A lining specified for the normal operating condition and installed with a bonding system chosen for the same figure can be defeated entirely by what happens during two hours of start-up.

Where cyclones are arranged in parallel, differing wear between nominally identical units is one of the most reliable signals available, and it is almost always about flow split rather than about the units themselves. A manifold that favours one branch delivers more gas and more dust there, and that unit wears first and separates worst.

Rotating spare units through positions is tempting as a way to even out replacement, but it also erases the evidence. Keep a fixed position map with dated observations instead, and correct the split once rather than absorbing it forever in spares.

The order matters more than the choice. Correct the inlet geometry and the flow split first, restore free discharge second, and specify the lining last — against the duty that remains once the first two are fixed. Doing it in the reverse order produces a well-lined vessel that still wears out on schedule and still separates poorly.

Only after those corrections is a lining comparison meaningful, because only then does the remaining wear describe the duty rather than the defect. Bring the four-region findings and the position map to that discussion, and the specification will follow from them.