Pump Vane-Pass Vibration: Resonance, Not Cavitation

The narrow 255 °C vibration peak is a temperature-tuned resonance amplifying vane-pass excitation, not a broadband rotor fault or a simple loss of suction margin. At 2975 rpm with six impeller vanes, the forcing frequency is 297.5 Hz. The diagnostic task is to determine whether the resonant path is acoustic in the liquid-filled piping or structural in the pump, baseplate, supports, and connected pipework.

Symptom Signature

The Goulds 3620 between-bearing pump circulates Syltherm 800 in a closed loop at 2250 m³/h. It heats from ambient to 300 °C while manually set valves hold the flow stable. Ordinary thermal change raises vibration from about 2 mm/s to 3.5 mm/s, but a separate response appears near 250 °C, exceeds 10 mm/s at 255 °C, and falls to 3–4 mm/s by 265 °C.

The term here means a resonance crossing: a natural acoustic or structural frequency moves through a fixed excitation frequency as temperature changes. The narrow temperature interval matters more than the absolute temperature.

Observation	Diagnostic meaning	Next measurement
Vibration rises gradually from 2 to 3.5 mm/s	Normal thermal changes may alter alignment, stiffness, viscosity, and hydraulic loading	Trend overall and frequency-filtered vibration during the complete heat-up
Amplitude rises sharply near 250 °C and peaks above 10 mm/s at 255 °C	A narrow amplification band points to resonance rather than a proportional process effect	Record amplitude and phase at vane-pass frequency in small temperature increments through the band
Vibration returns to 3–4 mm/s by 265 °C	The forcing remains, but the system moves away from the resonant condition	Compare spectra, phase, and pressure pulsation below, at, and above 255 °C
The dominant component is vane-pass frequency	The impeller-volute interaction supplies the excitation	Confirm that the peak tracks six times shaft speed
Operation improves as temperature continues toward 300 °C	A simple temperature-driven cavitation explanation does not match the observed trend	Retain suction-pressure checks, but prioritize acoustic and hot structural tests

Vane-Pass Excitation Mechanism

Shaft rotational frequency is:

2975 rpm / 60 = 49.58 Hz

For a six-vane impeller, vane-pass frequency is:

6 × 49.58 Hz = 297.5 Hz

Each vane produces a periodic pressure disturbance as it passes a volute tongue or splitter. A double-volute casing with the splitter and cutwater 180 degrees apart changes the spatial distribution of that force, but it does not by itself explain why the amplitude exists only in a narrow temperature band. It supplies a persistent 297.5 Hz input that a temperature-sensitive system can amplify.

Vane count and rotational speed set the frequency. Machining the impeller outside diameter at an angle cannot change the mathematical vane-pass frequency while speed and vane count remain fixed. Such machining may change the magnitude and phase distribution of the hydraulic forcing, which can reduce its coupling into a mode, but describing it as a frequency change is incorrect.

Temperature-Sensitive Resonance Paths

An acoustic mode depends on the liquid’s speed of sound, density, piping geometry, boundary conditions, and the effective compliance of vessels, valves, branches, and pipe walls. Syltherm 800 properties change during heat-up, so an acoustic natural frequency can pass through 297.5 Hz near 255 °C. Pressure waves reflect at changes in area, branches, valves, equipment connections, and other impedance discontinuities. A pressure measurement at only one location can miss the mode if that point lies near a pressure node.

A structural mode can move for a different reason. Thermal growth changes support contact, hold-down behavior, pipe loads, casing restraint, baseplate stiffness, and alignment. A cold impact test characterizes the cold, stationary structure under its test boundary conditions. It does not rule out a mode created or shifted by hot pipe loading, sliding supports that bind, clearances that close, or contact conditions that change during expansion.

Hydraulic excitation can also change with temperature because density and viscosity change even when volumetric flow remains stable. That may change forcing amplitude, but a rapid rise followed by a rapid fall over roughly 15 °C still calls for a resonance search. Separate excitation from amplification by measuring dynamic pressure, casing motion, support motion, and phase at the same time.

Interpretation of Completed Modifications

Change or check	What it addresses	What remains open
NPSH available calculation	Checks calculated suction margin	Confirm actual suction pressure and temperature during the event; the improvement above 255 °C weighs against suction-margin loss as the primary cause
Cutwater and splitter profile modification	Changes impeller-volute interaction and forcing amplitude	A resonant piping or structural path can still amplify the remaining vane-pass force
Impeller reduced from 416 mm to 410 mm; maximum diameter 432 mm	Changes developed head, operating point, clearances, and excitation strength	At unchanged speed and six vanes, vane-pass frequency remains 297.5 Hz
Inlet straight run increased from two to five pipe diameters	Reduces inlet distortion caused by nearby fittings	The outlet straight run remains 1.5 diameters, and acoustic reflections elsewhere in the closed loop remain possible
Piping-flow CFD	Examines mean and potentially unsteady flow behavior within the modeled domain	A steady-flow model does not establish the acoustic natural frequencies of the complete liquid-filled loop
Early bump test	Searches for accessible structural modes in the tested condition	The hot running structure requires operating-shape, phase, or hot modal measurements
Support assessment	Identifies weak or thermally changing restraint	Adding stiffness without locating the active mode can move the resonance into another operating condition

Diagnostic Procedure

1. Establish a repeatable thermal run. Hold speed, valve positions, and flow at the existing operating condition. Record temperature at the pump and relevant loop locations rather than assigning one bulk temperature to the whole system.
2. Acquire synchronous vibration data. Measure orthogonal radial directions and axial vibration at both bearing locations. Add casing, baseplate, support, and nearby pipe measurements. Retain spectra, waveform, tachometer reference, amplitude, and phase.
3. Resolve the resonance band. Collect closely spaced measurements below 250 °C, through the 255 °C peak, and above 265 °C. Avoid relying on widely spaced route readings that can skip the narrow maximum.
4. Measure dynamic pressure. Use temperature- and service-rated pressure instrumentation at multiple loop positions. Compare pressure amplitude and phase at 297.5 Hz with casing vibration. Multiple locations distinguish a genuine low-amplitude response from a transducer placed near an acoustic node.
5. Map operating deflection. Compare the motion and phase of the pump, bearing housings, baseplate, supports, and connected piping at the critical temperature. A structural mode produces a repeatable motion pattern with relatively stationary points and high-motion regions.
6. Track thermal restraint. Measure thermal movement, support gaps, sliding points, hold-down conditions, and nozzle displacement from cold through 300 °C. Correlate each contact change with the vane-pass amplitude and phase.
7. Classify the dominant path. Strong coherent pressure pulsation throughout the loop points toward an acoustic mode. Large relative structural motion with weak dynamic-pressure response points toward a structural mode. Both can coexist, so retain both data sets.
8. Model the confirmed mechanism. For an acoustic result, use a pulsation model containing actual pipe lengths, diameters, branches, valves, equipment boundaries, and temperature-dependent fluid properties. For a structural result, model the measured hot support and pipe-load conditions rather than only the cold geometry.

Corrective Actions by Confirmed Cause

For an acoustic resonance, shift the acoustic natural frequency or reduce reflection and excitation at 297.5 Hz. Candidate changes include relocating a pipe-area transition, altering a branch or effective acoustic length, changing a reflective boundary, or installing an engineered pulsation-control device. A pulsation damper is not the first selection for 300 °C hot-oil service; its thermal rating, materials, pressure boundary, maintenance requirements, and effect on the closed loop require a dedicated design review.

For a structural resonance, change the active load path identified by the operating-deflection data. Correct a binding sliding support, restore the intended thermal expansion path, reduce nozzle loading, or modify stiffness at the measured high-motion region. Adding material to a convenient support without modal data is wrong practice because it may leave the active mode unchanged or relocate it.

If reducing excitation is necessary, consult the pump manufacturer about an alternate impeller or casing modification. Five- or seven-vane designs would move vane-pass frequency at the same speed, while a split or staggered arrangement could change force cancellation if the actual double-suction impeller construction permits it. Inspect the impeller before specifying stagger because the presence and relationship of vane rows must be established. A proposed volute-lip change, including a 168-degree relationship, requires manufacturer hydraulic and mechanical review before machining.

Verification Checks

1. Check 1: Frequency identity. Expect the dominant component to remain at six times shaft frequency: approximately 297.5 Hz at 2975 rpm. A peak that does not track speed at that order requires a different diagnosis.
2. Check 2: Thermal repeatability. Before modification, expect the amplitude rise to begin near 250 °C, peak near 255 °C, and subside by about 265 °C when operating conditions match.
3. Check 3: Acoustic correlation. For an acoustic mode, expect coherent dynamic-pressure activity at 297.5 Hz and a repeatable pressure-phase pattern across multiple measurement locations.
4. Check 4: Structural correlation. For a structural mode, expect a repeatable hot operating-deflection shape and a phase transition or strong relative motion near the resonant temperature.
5. Check 5: Post-change result. Repeat the complete ambient-to-300 °C ramp at the same speed, flow, and valve settings. Expect the narrow peak above 10 mm/s to disappear or move outside the permitted operating range, with vibration remaining within the applicable API 610 and site acceptance criteria.

Recurring Diagnostic Pitfalls

Do not treat a cold bump test as proof that no hot structural resonance exists. Do not use stable volumetric flow as proof that hydraulic forcing is constant, because fluid properties and pressure-wave behavior change with temperature. Do not declare cavitation solely from high vibration when the measured response improves as the liquid becomes hotter.

Do not infer the acoustic field from one pressure location. Do not interpret impeller diameter trimming as a change in vane-pass frequency when speed and vane count are unchanged. Finally, do not modify supports, pipework, volute geometry, or impeller vane count before recording a repeatable baseline; each change must target either the measured excitation or the measured resonant path.

FAQ