The first number that matters is the pressure margin above the liquid’s vapor pressure at the first-stage impeller eye. A hot, preflashed crude stream can cross that limit abruptly after a small change in suction pressure, temperature, flow, or gas content. The reported pump is a two-stage, top-top unit whose first impeller has double suction. It handles crude oil at 145–160 °C, with 165 °C identified as the maximum design temperature, specific gravity 0.7888, and a reported vapor-pressure range of 3.28–3.84 kg/cm²a.
Loss of suction, lower flow, lower discharge pressure, gravel-like noise, and rising vibration can result from true vapor cavitation, gas breakout, external air ingress, or a suction-side vapor pocket. These mechanisms need different corrections even when the operating symptoms sound alike.
Pressure, temperature, and timing limits
True cavitation begins when local absolute pressure falls below the crude oil’s vapor pressure. Vapor cavities form near the impeller inlet and then collapse after entering a higher-pressure region. The collapse produces pressure pulses, noise, vibration, and potentially localized surface damage. Suction pressure does not have to be negative gauge pressure; the controlling comparison is local absolute pressure versus vapor pressure.
Calculate NPSHA from conditions at the pump suction flange:
NPSHA = P_suction,absolute/(ρg) + v²/(2g) − P_vapor/(ρg)
Compare that value with the pump’s NPSHR at the actual flow and speed, not only at the design point. Convert a gauge-pressure reading to absolute pressure before using it, and obtain vapor pressure for the measured crude composition and temperature. The reported 3.28–3.84 kg/cm²a range is absolute, but its temperature and composition correlation must match the fluid present during the event.
| Quantity | Reported value or source | Why it matters |
|---|---|---|
| Operating temperature | 145–160 °C |
Higher temperature raises vapor pressure and reduces suction pressure margin. |
| Maximum design temperature | 165 °C |
A design limit is not proof of adequate cavitation margin. |
| Specific gravity | 0.7888 |
Used to convert pressure into liquid head; use density at operating temperature if available. |
| Vapor pressure | 3.28–3.84 kg/cm²a |
Must be compared with absolute pressure at the impeller inlet. |
| Failure timing | Abrupt loss; recovery after flushing for hours | Points toward a changing gas inventory, vapor pocket, temperature condition, or restored prime rather than simple progressive wear. |
Competing hydraulic and gas mechanisms
True cavitation, liberated process gas, and external air ingress can all reduce developed head. Vapor cavities collapse as pressure rises. Permanent gas bubbles primarily compress, expand, and redistribute; they do not behave like condensable vapor. Gas accumulation can still block part of the impeller eye, disturb double-suction loading, reduce flow, and create vibration without producing the same erosion mechanism as vapor collapse.
| Observed behavior | Likely mechanism | Deciding check |
|---|---|---|
| Noise and vibration decrease when discharge flow is reduced | Insufficient cavitation margin at the original flow | Trend suction absolute pressure, temperature, flow, and noise during a controlled throttle test. |
| Noise persists despite a substantial controlled flow reduction | Entrained gas, air ingress, trapped vapor, or mechanical contact | Vent a safe high point, inspect leak paths, and compare vibration spectra with hydraulic load. |
| Normal service returns after hours of flushing | Cooling, gas displacement, vapor-pocket removal, restored wetting, or restored prime | Record casing temperature, vent discharge, suction pressure, and restart conditions before and after flushing. |
| Abrupt loss of both flow and discharge pressure | Gas binding, suction blockage, flashing, loss of prime, or impeller torque-transfer problem | Correlate suction pressure and motor load with flow and discharge pressure at event onset. |
| Vibration rises only after hydraulic performance collapses | Hydraulic instability is driving vibration | Confirm that vibration follows the pressure and flow change rather than preceding it. |
Diagnostic approaches compared
A controlled discharge-throttling test is the fastest discriminator when the system permits it. Lower flow generally reduces inlet velocity losses and the pump’s required suction head. If gravel-like noise, vibration, and unstable head recede together, true cavitation is the leading mechanism. Treat this as a directional test rather than a standalone diagnosis because throttling can also alter gas transport.
Gas-removal testing addresses a different mechanism. Venting the casing or another approved high point can reveal trapped gas, while simultaneous pressure and temperature trending shows whether the liquid was near flashing. At 145–160 °C, use closed, pressure-rated handling routed to an approved recovery system; direct manual venting of hot crude creates burn, fire, and exposure hazards.
Piping inspection finds conditions that calculations based only on vessel level and line size may miss. Review the suction route for high points, pockets, downward entry into the top suction connection, an elbow immediately before the pump, asymmetric flow into the double-suction impeller, and transient pressure loss across the strainer. Check the inlet arrangement against ANSI/HI 9.8 using the applicable edition and pump geometry.
Recommended isolation procedure
- Capture the healthy baseline. Record suction pressure at the pump flange, discharge pressure, flow, crude temperature, motor current or power, speed, vibration, and control-valve position. Use absolute suction pressure for vapor-margin calculations.
- Capture the transition. Trend the same signals fast enough to establish which quantity changes first. A falling suction pressure before head loss points upstream; stable suction pressure followed by loss of motor load and flow directs attention to gas binding or internal torque transfer.
-
Calculate the operating margin. Use the measured suction pressure, velocity, density, and vapor pressure at the event temperature. Read
NPSHRfrom the pump curve at the measured flow and speed. - Reduce flow under an approved operating procedure. Throttle only on the discharge side while monitoring minimum-flow restrictions, temperature, vibration, and motor load. A prompt reduction in noise and vibration supports vapor cavitation.
- Check for accumulated gas. Use the installed closed vent or an engineered degassing path. Record whether gas or liquid exits and whether head returns immediately after gas removal.
- Measure suction losses. Obtain pressure on both sides of the strainer if connections exist. A visually clean strainer can still develop a meaningful pressure drop from fouling, flashing, or trapped gas.
- Inspect the piping profile. Compare elevations and slopes against an as-built sketch. Mark every high point, reducer, valve, elbow, branch, and diameter transition between the source and pump.
- Repeat under matched conditions. Reproduce the comparison at similar temperature, flow, source pressure, and fluid composition. A correction is verified only when the symptom stays absent at the former failing operating point.
Suction piping and gas-management corrections
Correct trapped-gas geometry first. Regrade suction piping where practical, remove unintended high points, and provide engineered venting at locations where gas collects. An elbow close to the suction nozzle can create swirl and unequal loading between the two sides of a double-suction impeller; revise the straight-run and flow-conditioning arrangement based on the pump inlet requirements.
A suction stabilizer with automatic degassing controls can address recurring gas breakout where piping changes cannot prevent accumulation. Vent discharge must go to a suitable closed flash or recovery system. A flash accumulator is another engineered option when the released gas volume exceeds the casing vent’s intended duty.
Discharge-to-suction recirculation may increase inlet flow and move gas through low-velocity regions, but it also adds heat and can raise the suction temperature. Evaluate both pressure gain and thermal effect before adopting it. Likewise, reducing suction line diameter can increase gas-carrying velocity but also increases friction loss; select it only from a complete hydraulic calculation showing a net improvement in pressure margin.
Removing the suction strainer is a diagnostic option only when process cleanliness, downstream equipment protection, and site authorization permit it. A safer first step is differential-pressure measurement across the installed strainer during both normal and failing operation.
Mechanical checks and flushing recovery
The first impeller is keyed to the shaft, so a slipping collet is not the applicable failure mode. The key, keyway, shaft fit, retaining hardware, and impeller position can still be inspected during a planned outage if hydraulic tests do not explain the abrupt head loss. Confirm that coupling condition, rotation, speed, and axial position remain stable.
Hours of flushing can restore suction by cooling the casing, condensing vapor, displacing accumulated gas, changing fluid properties, or re-establishing a liquid-filled flow path. Turn that recovery into a controlled diagnostic: measure flush temperature and pressure, casing temperature, vented phase, flushing duration, and the pressure developed at restart. If a brief closed-system gas removal produces the same recovery as prolonged flushing, the corrective action should target gas accumulation rather than treating flushing as the permanent operating method.
Inspect the impeller inlet and casing during the next outage. Pitting concentrated near low-pressure inlet surfaces supports repeated vapor collapse. An absence of visible wear does not exclude intermittent cavitation, especially when the event duration is short.
Verification and escalation criteria
Verify the correction with a trend that includes suction absolute pressure, temperature, flow, discharge pressure, motor load, and vibration. The corrected system should cross the former event duration and operating point without a progressive suction-pressure decline, sudden head loss, gas accumulation, or vibration increase. Compare repeated runs because crude composition and vapor pressure can change even when indicated temperature remains similar.
Stop a test if vibration, temperature, leakage, motor load, or pressure crosses the site’s operating limit. Repeated loss of prime, inability to vent through a closed system, or unstable operation near the calculated vapor margin requires an engineering review before another restart.
Frequently asked questions
Why does the crude-oil pump cavitate even below 165 °C?
165 °C is the maximum design temperature, not a cavitation threshold. Cavitation depends on local absolute inlet pressure relative to vapor pressure, which was reported as 3.28–3.84 kg/cm²a.
Why does throttling the discharge reduce cavitation noise?
Lower flow usually reduces suction-system losses and the pump’s required suction head. If noise, vibration, and head instability fall together during controlled throttling, insufficient cavitation margin is the leading cause.
Why does flushing for hours restore pump suction?
Flushing can cool the casing, condense vapor, displace gas, and refill an interrupted liquid path. Record temperature, vented phase, pressure, and recovery timing to identify which effect restores operation.
Why can a clean suction strainer still cause loss of flow?
Visual cleanliness does not measure operating pressure loss. Trend differential pressure across the strainer because trapped gas, partial fouling, or local flashing can reduce suction pressure at the pump.
When should repeated pump cavitation be escalated?
Stop when operation cannot remain inside site vibration, temperature, leakage, motor-load, or pressure limits, or when the pump repeatedly loses prime. Escalate through the pump manufacturer’s official support channel with the pump curve, casing and impeller configuration, as-built suction sketch, fluid-property data, and synchronized operating trends. Request a hydraulic and mechanical review before further restarts if closed venting and controlled flow reduction do not identify the mechanism.