How to Diagnose Pump Cavitation in the Field

Back to Blog

Confirm the Cavitation Signature First

Start with your ears and the gauges, because cavitation has a signature that is hard to mistake once you have heard it. The noise is a sharp crackling or a sound like the pump is passing gravel, caused by vapor bubbles collapsing violently against the impeller as the fluid re-pressurizes. It rises and falls with load rather than staying at one steady pitch, which distinguishes it from the constant whine of a failing bearing or the rhythmic knock of a bad coupling. Put a hand or a screwdriver to the bearing housing and the casing; cavitation vibration is felt broadly across the casing, not localized at one bearing.

The instruments back up the ear. A cavitating pump shows a discharge pressure that swings and hunts rather than holding a steady value, because the vapor pockets are disrupting the flow the impeller can deliver moment to moment. Flow falls off and the head the pump makes drops below its curve at that speed. Motor amps often wander with the pressure. If all three signs move together, the noise and the erratic gauges and the lost flow, you are looking at cavitation and not a purely mechanical fault.

Rule out the close mimic before going further. Air binding produces a similar loss of flow but a different sound, more of a smooth surging than a sharp crackle, because the pump is trying to move a compressible gas pocket rather than collapsing vapor. The distinction matters for the fix, so the note on what pump air binding is is worth a look if the noise is soft rather than gravelly. The definition page on what pump cavitation is covers the physics the field signs are pointing at.

Prove It With a Suction-Side Head Check

Ears confirm cavitation is likely; a head check proves it. Cavitation happens when the pressure at the impeller eye drops to the vapor pressure of the liquid, so the test is whether the net positive suction head available (NPSHa) has fallen below the net positive suction head the pump requires (NPSHr). Read the suction pressure at the pump inlet and convert it to head, add the velocity head, and subtract the vapor pressure of the liquid at its actual temperature. That is your available head. Compare it to the NPSHr curve on the pump datasheet at the current flow.

The arithmetic is the diagnosis. If NPSHa sits comfortably above NPSHr with the usual margin, the suction side is healthy and the noise is coming from somewhere else, so redirect the investigation to mechanical causes. If NPSHa has collapsed to at or below NPSHr, the physics of cavitation are satisfied and you have your confirmation. The concept of how much cushion you need is treated in the note on what NPSH margin is, and the underlying quantity in what net positive suction head is. A trend of suction pressure over the last shift often shows exactly when the margin ran out.

Watch the temperature term, because it is the one people forget. Vapor pressure climbs steeply with temperature, so a suction condition that was fine on a cool morning can cavitate on a hot afternoon or when the process liquid heats up, even though nothing about the piping changed. If the pump only cavitates when the liquid is hot, the vapor-pressure term is your answer and the fix is about lowering liquid temperature or raising suction pressure, not repiping.

Trace the Cavitation to Its Cause

With cavitation confirmed, the causes fall into a short list ordered by how often they turn up. The most common is a restricted suction: a plugged suction strainer or filter, a partly closed suction valve, or a collapsed suction hose, all of which throttle the flow into the pump and drop the pressure at the eye. Walk the suction line and check the strainer differential and every valve position. This is the cheapest cause to find and fix, so it comes first.

The next tier is a suction condition that is genuinely marginal by design or by change. A dropping source level raises the lift and cuts NPSHa, a rising liquid temperature raises vapor pressure, or the pump is simply running too far out on its curve where NPSHr is high. Running out toward runout is its own trap because NPSHr rises with flow, so a pump pushed to high flow can cavitate even with a clean suction, which the note on what pump runout flow is explains. Throttling the discharge slightly to pull the operating point back left often stops the noise, which is itself a confirming test.

The rarest but most damaging cause is a suction-side design that never had enough margin, where the fix is engineering rather than operations: raising the source tank, shortening or enlarging the suction line, cooling the liquid, or lowering the pump. Before recommending that, be sure you have exhausted the cheap causes, because a plugged strainer that gets blamed on plant design wastes a shutdown. The value of continuous suction-pressure trending on a platform such as Merobix is that it timestamps when the margin started eroding, which usually points straight at whichever term changed.

When to Escalate

Escalate to a mechanical inspection when the suction-side head checks out healthy but the noise and vibration persist, because you are then likely looking at impeller damage from past cavitation, a worn wear ring, or a bearing fault that mimics the sound. Cavitation that ran for a while leaves eroded, pitted impeller vanes that keep the pump noisy and inefficient even after the suction cause is corrected, and that is a teardown finding, not a field adjustment.

Escalate to engineering when the cause is a genuine NPSH shortfall built into the installation and no operational change restores the margin. Repiping the suction, relocating the pump, adding a booster pump, or cooling the process are design changes that need proper hydraulic review and sign-off. Field diagnosis has done its job once it has named the term that is short; closing that gap safely is a design decision for qualified personnel.

Frequently Asked Questions