Pump Cavitation: Causes, Symptoms, and Prevention Across Pump Types

Updated July 2026

What Is Pump Cavitation?

Pump cavitation is the formation and collapse of vapor bubbles inside a pump, a rapid phase change that begins when local pressure drops below the vapor pressure of the liquid and then rebounds around the impeller. This cycle repeats thousands of times a second, with each implosion striking the pump impeller and casing in a localized shock wave.

In basic fluid dynamics terms, an experiment measured a single collapsing bubble’s shock wave at approximately 2.46 MPa (about 357 psi) near atmospheric pressure, with the imploding liquid jet striking a surface at roughly 103 m/s — a high velocity by any pump standard — as water flows through the collapsing-bubble zone at high flow. [1] Cavitating pumps generate tens of thousands of these micro-implosions every second, each one landing inside the pump instead of dissipating harmlessly outside it.

Suction cavitation and discharge cavitation are the two root types, and discharge cavitation occurs less often than its suction-side counterpart but does just as much damage to the pump housing and impeller. Suction cavitation occurs when the pressure at the pump inlet is insufficient — there isn’t enough Net Positive Suction Head Available (NPSHA) to overcome the pump’s required level (NPSHR). Discharge cavitation happens when there’s not enough room for the fluid that entered the pump to exit — high discharge pressure and a sudden pressure increase against a restriction or blocked outlet together create a recirculating eddy near the impeller outlet, and the resulting drop in pressure there mirrors what a low-NPSH suction problem does at the inlet. (Entrained/dissolved gas cavitation, often called gaseous cavitation, causes similar effects but not destructive implosions — this is the difference that matters for diagnostics, covered in H2-3.)

What Causes Pump Cavitation, Suction Side vs Discharge Side

Fully 80% of pump failures trace back to the pump’s internal operating conditions on the suction side, states 30-year Ingersoll Rand and Union Pump engineer and six-patent holder Terry Henshaw in Pumps & Systems. [4] Knowing which side to troubleshoot first can save hours — the common causes below sort into a checklist by side of the pump.

The Discharge-Side Blind Spot: every SERP Guide we reviewed discussed the suction-side in detail and only gave discharge cavitation a throwaway sentence, if any. That balance is backwards for a meaningful share of real failures. A well-documented case of a problem with a discharge-side root: Three fuel oil pumps operated for years with partially closed suction valves to compensate for underpowered motors — an intentional workaround that should have addressed suction issues. Unfortunately, this resulted in mechanical seal failure every four to six months (the description said catastrophic failure upon dissection), although vibration and temperature remained within expected ranges and classic impeller pitting wasn’t present. [5] The workaround was papering over a design-and-discharge issue, not solving it.

One thing to be up front about: The advice “never throttle a suction valve” is widely circulated but isn’t universally true. A closed experiment comparing throttling a suction valve to pulling a vacuum on a supply tank of a recirculating loop was published in Pumps & Systems. Throttling the suction valve actually pushed the pump operating point away from cavitation (it actually added head, thus decreasing the flow rate and thus decreasing the NPSHR needed), whereas pulling a vacuum on the supply tank caused visible cavitation.[6] The prevailing advice — control the flow from the discharge side or by a VFD, not a throttled suction valve — still holds for the vast majority of open systems, but it depends on where you are on the system curve, not a blanket rule.

How to Tell If a Pump Is Cavitating, Symptoms and Diagnostic Signals

The Rocks-in-a-Can Diagnostic Sequence: The rattling noise of a gravel-choked washboard is perhaps the most iconic signature of cavitation, but it isn’t unique to it, and it isn’t always present (dissolved gas in the fluid may cushion the collapse shock.[7]) So consider the noise, but not as a sole diagnosis; instead use it as one input for a short sequence of tests.

Beyond a simple discharge-throttle test, also be on the lookout for: pump head falls or flow drops below the rated curve; abnormal vibration, often focused at the bearing housing; and increasingly, premature detection by spectrum-analyzing vibration monitors. Research is moving in this direction: a 2026 bibliometric study of 645 papers shows the field’s interest in multi-signal fusion (vibration, noise, pressure pulsation), and in using machine-learning algorithms for classifying events. One study found a deep-learning model could achieve more than 98% classification accuracy under test conditions.[8] (However, that study is frank about one limitation: Most current systems can reliably detect that cavitation is happening but have trouble finding low-level, early-stage cavitation before significant damage has been done.[8])

What Cavitation Does to a Pump, Damage Mechanics and Failure Forensics

Bubble collapse near a metal surface does more than make noise. Each implosion forms inside the pump rather than harmlessly outside it, generating localized heat beyond the melting point of the surrounding metal and damaging pump components well beyond the impeller alone. That failure pattern is diagnostic — it tells a technician how severe and how long the cavitation has been running.

A 2024 study documented melting directly on the surface of copper (melting point 1,083°C) and stainless steel (melting point 1,440°C) test specimens subjected to continuous cavitation; failure progressed from plastic deformation along the pit’s edges, through creep, to fracture as neighboring pits coalesced into a single, visible one.[9] Ductile metals tend to fail this way, whereas brittle metals such as tungsten fail along grain boundaries — even the fracture pattern provides information about the material as well as the intensity of the cavitation.[9]

How Specific the Forensic Evidence Can Be: A documented case of damage on a 330-MW-class, eight-stage boiler feed pump illustrates just how detailed the postmortem analysis can be. Only one eye (the coupling side) of the double-suction first-stage impeller was damaged — the outboard eye stayed intact. Analysis revealed that the cause wasn’t material defect, but a combination of suction recirculation and a split in flow distribution between the two impeller eyes, necessitating a flow-balance redesign, not a metallurgical remedy.[10] And the asymmetrical damage to the two eyes of a double-suction impeller actually points to the geometry of the upstream suction piping, not the pump casing itself.

This cost issue is far from academic. Trade-press reporting puts downtime costs for large industrial pumps at more than $200,000 a day for operating expenses alone, before counting replacement parts,[11] and that severe, sustained cavitation can take out pumps in minutes rather than years.[11] Cavitation bubble-implosion damage is particularly severe on high-energy pumps, generally defined as those exceeding roughly 850 ft of head or 600 HP per stage. For the mining industry specifically, unplanned pump-related downtime can average around $180,000 per event, and it isn’t uncommon for mine operations to budget 35-50% of their annual expenses on maintenance because their pumps are run until they fail rather than being proactively serviced.[12]

Cavitation Across Pump Types, What Changes by Architecture

Here’s the honest truth about all three top cavitation guides we came across when doing the research for this article: the highest three-scoring results all treat cavitation as a generic, pump-type-agnostic phenomenon. One guide flatly states, in the very first paragraph, that the cavitation principles it discusses are universally applicable to “all pump types” and “are very similar,” and it’s not until later that it zeroes in on centrifugal pumps alone — without ever returning to any other pump type in the text. The other two follow a similar, albeit unstated, trajectory. Generic advice found in the vast majority of guides will be, at best, incomplete if your pump is anything other than a standard centrifugal pump.

The 9-Type Cavitation Behavior Profile

Drawing from manufacturer engineering literature, NFPA fire-pump test standards, and forum-documented field cases, the following table is far from a cookie-cutter guide.

The fire-pump row warrants a second glance due to the exceptional specificity of the governing standard. NFPA 25 requires churn (no-flow) testing weekly for diesel fire pumps (minimum 30 min run) and monthly for most electric fire pumps (minimum 10 min run), with investigation required whenever discharge/suction pressure differential exceeds 95% of the rated pump pressure.[16] Annex A.14.2.6 specifically requires the engagement sequence for the jockey and main fire pumps to step by step (jockey engages on 10 psi drop, main engages on a 5 psi drop after the jockey),[13] to both avoid nuisance short-cycling and pressure spikes resulting from excessively wide activation gaps. The minimum separation distance of 10 pipe diameters from the suction flange to any downstream elbow or tee on a horizontal split-case fire pump also applies, although horizontal split-case fire pumps don’t require that minimum separation distance if they have an integrated, inline impeller, which positions the impeller beneath the motor rather than in a separate suction casing, at the cost of increased maintenance access complexity.[14]

NPSH Margin, The Core Prevention Math

Net Positive Suction Head Available (NPSHA) is a system characteristic — the absolute suction head of liquid at the impeller datum, minus the liquid’s absolute vapor pressure. Net Positive Suction Head Required (NPSHR) is a pump characteristic the manufacturer publishes — the minimum suction head the pump needs. NPSHA must exceed NPSHR, and the difference is the margin — maintain a positive margin of NPSHA over NPSHR at every operating point, not just the design point.[17] The governing standard was recently updated in ANSI/HI 9.6.1:2024 (replacing the 2017 version), changing from a standard 10-foot margin over the lab-derived NPSH3 value (which is not published by all manufacturers) to a variable 10-segment market-specific guideline for petroleum, chemical process, power, water and wastewater, pulp and paper, slurry, oil and gas, irrigation, general service, and building services.[17]

One detail competitors covering this topic consistently miss is that NPSHR isn’t a fixed number for a given pump. NASA testing on two different pumps operating on water found NPSHR falling sharply as liquid temperature rose: on one test pump, measured NPSHR dropped from about 12 ft at roughly 70°F to about 2 ft at roughly 410°F, an 83%-plus reduction on that specific unit and fluid.^[18] This is known as the thermodynamic suppression effect, and NASA’s own data notes the magnitude varies by fluid properties, pump design, speed, and flow coefficient — it is not a number you can copy onto a different pump. The practical takeaway is narrower and more useful: a flat NPSH margin figure applied without regard to operating temperature can be needlessly conservative for hot or volatile services, and pretending the margin is temperature-independent is the more common mistake. Suction specific speed (Nss) is the other variable worth checking alongside NPSH margin — typical values fall between 7,000 and 15,000, and pump users have historically set lower maximum Nss limits after problems traced back to oversized impeller eyes.^[4]

How to Prevent and Fix Pump Cavitation, Step-by-Step

There is no single way to prevent cavitation that works for every installation, since several factors often combine to cause cavitation rather than one isolated culprit. The best way to prevent pump cavitation is to work through this checklist in order to confirm it’s not an installation problem before declaring the pump itself defective:

• Check filters and strainers on the suction side and clean or replace anything clogged — a blocked strainer is the single most common cause suction problem we see
• Verify suction pipe diameter and keep 10 pipe diameters of straight pipe upstream of any bend or fitting to keep flow velocities steady, as a standard rule of thumb^[13]
• Minimize suction lift height by installing the pump below or near the fluid source
• In open systems, regulate flow using a valve or VFD on the discharge side instead of throttling the suction, since every extra bend and valve added upstream of the suction inlet raises the pressure required as the liquid accelerates into the impeller eye
• Re-calculate NPSHA based on the maximum operating temperature of the fluid rather than the design ambient temperature
• Check that the pump is operating within approximately 70% to 120% of its Best Efficiency Point (BEP) under normal conditions

Two preventive methods worth knowing, even if not ready to be implemented: Some modern VFDs now include anti-cavitation controls that monitor the shaft torque in real-time and automatically slow the pump when cavitation begins and speed up again after it subsides.^[19] For services prone to air ingestion rather than vapor cavitation, certain designs such as double-volute impellers and high-specific-speed (over 3,000 rpm) impellers can tolerate air better than conventional designs.^[20]

“We size suction piping and NPSH margin for the hottest, lowest-tank-level condition a system will actually see — not the nameplate condition. A pump that never cavitates on a cold commissioning test can still cavitate every summer afternoon once the process fluid runs hot and the supply tank runs low. Sizing for the average case is the single most common design mistake we see in RFQs.”

BBP Engineering Team, Pump Applications

When a Piping Fix Isn’t Enough, Choosing a Cavitation-Resistant Pump

The Piping-Fix-or-Pump-Swap Verdict: if the pump still cavitates during normal operation after correct piping, discharge control, and adequate NPSH margin are in place, the problem now lives inside the pump itself. Fixing it shifts focus from piping changes to pump-level choices: lower-NPSHR impeller geometry, a dual-suction design, or more cavitation-resistant materials.

Choosing the right pump for the application in the first place is cheaper than any of these after-the-fact fixes — which is why the checklist above asks you to assess the entire pump system before blaming the equipment. Entire pump system design, from tank to discharge, drives how fast cavitation increases from a minor nuisance to a warranty claim.

Choosing the correct pump material to resist damage to components, tested per ASTM G32 vibratory cavitation-erosion methodology, is a documented, quantitative decision, not a guess — good piping design and material selection work together, since even the best alloy cannot compensate for fluid flow that arrives at the impeller eye already turbulent. A paper from the Texas A&M/Flowserve turbomachinery symposium rates various impeller alloys using ASTM G32 vibratory cavitation-erosion tests; one chrome-manganese austenitic stainless casting alloy (tradename X-Cavalloy) displayed 9x times the life in lab tests over the baseline cast carbon steel and field experience since 1994 supports a conservatively documented 3x-or-greater life improvement in real-world pump service applications over impellers which were prematurely failing due to cavitation. [21] Even more detailed, a specific 1986 power plant feed water pump case involved the combined redesign of an impeller and the replacement of CA15 stainless with CA6NM stainless material to achieve at least a 10x reduction in the rate of cavitation erosion versus the baseline design. [21]

It’s the geometry changes that also have a direct and specific payoff: a patented impeller change of 8 to 24 outlet blades brought the needed NPSHR down from 5.4 m to 1.1 m on a low-specific-speed pump running at roughly 6,870 rpm[22] in the test lab. On pumps where replacement can’t be justified, a sacrificial insert retrofit which is bolted to the volute closest to the impeller and which is field proven as a repair strategy on semi-open impeller cavitation damage, can avoid replacing the casing, and is covered by patent.[23]

FAQ, Common Pump Cavitation Questions