The condition-monitoring screen alternates between GOOD, FAIR, and CRITICAL while recorded speed remains 1482 rpm. That variation is the first diagnostic clue: record the operating point and suction conditions for each reading before replacing the overhung pump with a between-bearings design.
What is the condition screen actually telling you?
The recorded condition frequency is 24.7 Hz, which corresponds to the running speed because 1482 rpm / 60 = 24.7 Hz. A running-speed component can result from rotor imbalance, misalignment, excessive hydraulic radial load, looseness, or structural response. Frequency alone does not distinguish among them.
| Date | Displayed condition | Frequency | Speed |
|---|---|---|---|
| 27 Mar 2013 | GOOD | 24.7 Hz | 1482 rpm |
| 20 Dec 2013 | CRITICAL | 24.7 Hz | 1482 rpm |
| 27 Dec 2013 | CRITICAL | 24.7 Hz | 1482 rpm |
| 14 Jan 2014 | FAIR | 24.7 Hz | 1482 rpm |
| 3 Nov 2014 | GOOD | 24.7 Hz | 1482 rpm |
| 27 Nov 2014 | FAIR | 24.7 Hz | 1482 rpm |
Trace each screen indication back through the displayed condition tag, its historical record, the condition-monitoring channel, and the physical measurement point. Confirm that the tag represents the same bearing location, direction, vibration quantity, and alarm basis on every date. Then retrieve the waveform and spectrum rather than diagnosing from GOOD, FAIR, or CRITICAL labels alone.
Because the classification changed while the recorded speed did not, correlate each record with flow, suction pressure, discharge pressure, level, valve position, noise, and parallel-pump status. A permanently weak rotor arrangement would not by itself explain large condition changes at the same speed; changing hydraulic conditions can.
Is the pump operating at its selected duty point?
The stated duty is Q = 1350 m3/h at an observed head of 70 m, while the pump was reportedly selected for 73 m. A three-metre difference cannot be classified without the certified pump curve, measurement accuracy, liquid density, and actual flow. Locate the measured point on the installed impeller curve.
| Reading | Location | What the result means |
|---|---|---|
| Actual flow | Verified flow instrument or independent test | Shows whether the pump is operating near its rated point, toward runout, or at low flow. |
| Suction and discharge pressure | Corrected to a common datum at the pump | Provides developed differential head after velocity and elevation terms are handled. |
| Input power | Pump driver | Supports the flow-and-head result and exposes overload or an implausible instrument reading. |
| Liquid density and temperature | Operating process | Required to convert pressure to liquid head and calculate hydraulic power. |
| Valve positions and system lineup | Suction and discharge system | Identifies the configuration associated with each vibration state. |
Calculate pump head from verified suction and discharge measurements, not from a single gauge difference when pipe elevations or velocities differ. For the stated flow, 1350 m3/h = 0.375 m3/s. Hydraulic power is P_h = rho × g × Q × H; use the measured amine-solution density and measured head before comparing power with the curve.
If the measured point lies to the high-flow side of the intended operating region, increased internal velocity, higher NPSH demand, and off-design radial loading become primary suspects. If flow is near the rated point, continue to the suction-system and NPSH checks. If flow is low, investigate recirculation, throttling, obstruction, and operation below the pump's minimum continuous stable flow.
Does added discharge resistance change the vibration?
A controlled discharge-throttling test can determine whether excessive flow contributes to the noise and vibration. It changes system resistance and moves the operating point left on the pump curve without disturbing the suction piping.
- Record baseline flow, developed head, suction pressure, discharge pressure, driver power, vibration spectrum, bearing condition, and noise.
- Increase discharge resistance gradually under the site's operating procedure while remaining within the allowable operating region.
- Stop at the selected head of 73 m, or earlier if any operating limit is approached.
- Repeat the vibration and noise measurements at stabilized conditions.
If vibration and noise fall as head rises and flow falls, the branch points toward high-flow operation, inadequate system resistance, runout, reduced NPSH margin, or excessive off-design hydraulic load. If they do not change, inspect suction geometry, mechanical condition, foundation response, alignment, and the vortex plate. Throttling is a diagnostic test unless the pump curve, power, thermal behavior, and operating limits support it as a permanent control method.
Is the suction connection disturbing the inlet flow?
The suction pipe appears larger than the pump inlet, but the actual pipe and nozzle diameters and transition geometry must be measured. A poor transition, nearby elbow, asymmetric approach flow, entrained gas, or a damaged internal plate can deliver swirl and uneven velocity to the impeller eye. The resulting pressure field can create noise, cavitation-like symptoms, fluctuating radial force, and cyclic loading of internal components.
| Observation | Likely branch | Next check |
|---|---|---|
| Noise and vibration change with flow | Hydraulic operating point or suction margin | Plot flow, head, NPSHA, and vibration together. |
| Noise persists near the intended duty | Inlet distortion, gas, mechanical defect, or resonance | Inspect piping geometry and compare spectra by direction and location. |
| Vortex plate cracks from one edge or weld | Local stress concentration or cyclic plate motion | Perform fracture and attachment inspection. |
| Plate shows broad deformation or repeated fatigue | Unsteady hydraulic loading or structural response | Measure plate condition, flow pattern, and pressure pulsation. |
| Bearing distress follows high-vibration periods | Vibration may be driving bearing damage | Correlate bearing records with operating conditions and spectrum changes. |
The described “voltrx plate” is a vertical plate at the pump suction, apparently intended to interrupt or condition incoming flow. Its formal function and design basis are not identified. Confirm whether it is a vortex breaker, flow straightener, or another inlet component from the drawing. Treat repeated breakage as a failure requiring fracture analysis, not as proof that the plate caused the original vibration. Determine whether the fracture initiated at a weld, attachment, sharp corner, corrosion site, or impact mark.
Does the pump have adequate NPSH margin?
Noisy operation combined with marginal suction performance makes NPSH the next decision point. Calculate NPSHA at the pump suction for the worst credible liquid level, pressure, temperature, line loss, flow, and dissolved-gas condition. Compare it with the manufacturer's NPSHR curve at the actual flow, not only at rated flow.
The cited Shell requirement specifies the following checks:
| Service condition | Required comparison | Applicable range |
|---|---|---|
| General service | NPSHA − NPSHR ≥ 1 m |
Minimum continuous stable flow through rated capacity |
| Suction pressure below atmospheric | NPSHA − NPSHR ≥ 2 m |
Apply where suction pressure is sub-atmospheric |
| Above rated capacity | NPSHR ≤ NPSHA |
Rated capacity through 125% of BEP
|
| Liquid containing dissolved gases |
NPSHA = 1.5 × NPSHR, with at least 5 m difference |
Operating range restricted to 70%–110% of BEP flow
|
The same requirement states that NPSHR is based on water without correction for other liquids. For sulfinol and carbonate service, it limits maximum impeller tip speed to 45 m/s and calls for evidence of successful similar applications plus principal approval. Apply those service-specific provisions only when the pumped service matches them.
If the calculated margin fails the applicable criterion, correct the suction condition or operating point before blaming the bearing arrangement. If the arithmetic passes but noise remains, check for gas release, local inlet losses, swirl, a blocked passage, and transient suction-pressure depression. A steady pressure reading can miss pulsation and short-duration loss of margin.
Is an overhung pump inherently too small for 1350 m3/h?
There is no universal maximum capacity for all overhung pumps. The acceptable envelope depends on the selected hydraulic design, speed, impeller diameter, shaft stiffness, bearing arrangement, nozzle loads, rotor dynamics, NPSH characteristics, liquid properties, and the manufacturer's published operating limits.
| Configuration decision | Engineering effect | Limitation |
|---|---|---|
| Retain the overhung pump | Can satisfy the duty when its hydraulic and mechanical margins are demonstrated. | Requires acceptable shaft deflection, bearing loads, vibration, NPSH margin, and operation on the approved curve. |
| Select a between-bearings pump | Supports the rotor between bearings and may improve shaft and rotor-dynamic behavior for a large hydraulic element. | Does not correct inadequate NPSHA, distorted inlet flow, gas entrainment, or an incorrect system operating point. |
Both configurations can work when correctly selected and applied. At 1350 m3/h and approximately 70–73 m, the overhung selection is large enough to justify a formal hydraulic, mechanical, and rotor-dynamic review, but flow alone is not a rejection criterion. Request the certified curve, allowable operating region, NPSHR curve, shaft-deflection calculation, bearing-load calculation, and vibration acceptance data for the supplied pump. Compare a between-bearings alternative on the same duty and suction basis.
How should the resolving branch be executed and verified?
- Validate the displayed condition tag against the correct physical channel, measurement point, direction, vibration quantity, spectrum, and waveform.
- Measure actual flow, suction pressure, discharge pressure, liquid density, temperature, driver power, valve positions, noise, and vibration at one stabilized condition.
- Calculate developed head and plot the operating point on the certified pump curve. Resolve the difference between the measured 70 m and selected 73 m.
- Repeat the data set during a controlled discharge-throttling test. Use the response to separate a high-flow hydraulic problem from a condition that remains at reduced flow.
- Measure the suction-pipe and pump-inlet diameters. Inspect the transition, upstream fittings, liquid level, gas entry paths, internal obstructions, and the exact function and mounting of the vertical plate.
- Perform fracture analysis on the broken plate and inspect bearings, shaft, coupling alignment, hold-downs, base, and foundation. Compare axial, horizontal, and vertical spectra at the pump and driver.
- Calculate NPSHA for every tested state and compare it with NPSHR using the applicable margin and operating-range requirements.
- Correct the branch identified by the measurements: restore the intended operating point, improve the suction condition, eliminate inlet distortion or gas entry, repair the plate design or attachment, or correct the mechanical defect. Reassess pump configuration only after these checks.
- Run the corrected installation at the required duty and at the approved operating-range boundaries. Record flow, head, suction pressure, power, noise, bearing condition, and vibration using the same channels and settings as the baseline.
Frequently Asked Questions
How do I know whether 1350 m3/h is too much for an overhung pump?
Do not apply a universal flow cutoff. Plot 1350 m3/h on the supplied curve and verify the allowable operating region, NPSH margin, shaft deflection, bearing loads, rotor dynamics, power, and vibration limits.
How do I check whether the pump is running beyond its intended flow?
Measure flow and developed head at the same stabilized condition, then compare the point with the certified curve. A controlled move from 70 m toward the selected 73 m can show whether lower flow reduces noise and vibration.
How do I diagnose vibration at 24.7 Hz?
At 1482 rpm, 24.7 Hz is running speed. Compare amplitude and phase by direction and location, then correlate the component with flow, alignment, hydraulic load, looseness, and structural response.
How do I determine whether the broken vortex plate is a cause or a consequence?
Confirm the component's intended function from the drawing, inspect the fracture origin and attachment, and compare its failure timing with vibration and operating data. Repeated fatigue can result from inlet pulsation or structural motion even when the original fault lies elsewhere.
How do I verify that the pump problem is resolved?
Repeat the baseline test at the required duty and approved range limits using the same vibration channels and settings. Accept the correction only when flow, head, NPSH margin, power, noise, bearing condition, and vibration all remain within their documented limits.