Dry Running

What it is. The mechanical seal depends on a ~1 μm fluid film between the faces to seal AND lubricate at the same time — without that film, it's pure metal-to-metal friction between rotating surfaces. Dry running is exactly that: the seal chamber runs short of fluid (emptying, a suction vortex, gas ingress, or a piping plan that stopped working), the film disappears, and the faces generate heat through direct friction. Since the faces are designed for a micrometer-thick film, not direct contact, temperature rises extremely fast — minutes, not hours — until thermal shock, radial cracks in the ceramic faces, and total destruction.

How to recognize it in the field. It's one of the few seal failure modes with a clear, simple early signal: seal chamber temperature rising fast and out of pattern. It doesn't require sophisticated analysis — a well-calibrated temperature sensor with an alarm detects the event in minutes, enough time to act before total destruction.

Why it matters. It is typically the component's fastest and most expensive failure mode — from normal operation to a destroyed seal in a matter of minutes, often leading to process leakage (with safety/environmental risks depending on the fluid). The root cause is almost always outside the seal: cavitation at the pump suction, a low tank level, or a sealing-plan block valve mistakenly closed.

Film-loss pathways

1. Empty chamber — low suction level, lost prime, or the pump operating without enough fluid circulating through the chamber.
2. Vortexing/gas ingress — an obstructed or undersized recirculation (Plan 11) lets dissolved gas or air ingress replace the liquid in the chamber.
3. Inoperative piping plan — a valve mistakenly closed, a clogged Plan 11 orifice, a stopped Plan 23 circulation pump, or lost barrier pressure on Plans 53.
4. Cavitation at the pump suction — severe cavitation reduces the effective fluid density in the chamber, an indirect but real path to partial dry running.

Framework A — diagnosis

Random, β ≈ 1: the event is triggered by a one-off operating condition (a closed valve, a low level), not by cycle accumulation — a new seal and an old seal are equally vulnerable if exposed to the same film loss. There is no characteristic failure age for this mode.

Framework B — prescription

Continuous CBM via temperature, the component's shortest and most critical P-F — the difference between detection in seconds/minutes and late detection is the difference between "seal saved" and "destroyed seal + process leak." The second layer (level/pressure and interlock checks) attacks the cause, preventing the dry-running condition from ever setting in.

Why destruction is so fast — the PV factor physics without a film

Without a lubricating film, the face friction coefficient jumps from very low values (hydrodynamic regime) to typical dry carbon-ceramic contact values — an order of magnitude higher. Since frictional heat generation scales with the product of normal force × speed × friction coefficient, the same PV condition that was safe in the lubricated regime greatly exceeds the faces' thermal limit in the dry regime — hence the timescale of minutes, not hours: the "PV factor" discussed under Mechanical Seal assumes a film is present; without it, the physics changes category.

Relationship to the upstream process

Dry running is rarely a defect of the seal itself — it is almost always a symptom of a process or pump condition that the seal merely reveals first (being the component with the smallest thermal margin). This makes it an excellent "canary": a recurring dry-running history at the same pumping point should trigger investigation of tank level, suction sizing or cavitation before any repeat seal replacement — replacing the seal without treating the cause just guarantees the next failure.