AXILINEMultistage Pump Rotor Balancing: What On-Site Work Can and Can't Do
A sectional pump has racked up running hours, and vibration has risen without a single repair. A feedwater pump shakes harder after an overhaul than it did before it was opened up. A multistage rotor is the most awkward object for on-site balancing among all pumps: a long shaft, several impellers, imbalance distributed across the stages, while the available correction planes are bunched at the ends. We arrive with a two-channel instrument, measure, and honestly split cases into "we balance on site" and "this needs a workshop." Either way, you get numbers and an assessment.
Symptoms: what a multistage pump complains about
On a single-stage pump, imbalance more often comes from outside: buildup, impeller erosion. On a multistage pump it accumulates internally and shows up slowly. The rotor is long, the impellers are hidden inside the casing, and the first things to report a problem aren't the impellers themselves but what surrounds them: the wear rings, the thrust bearing, the balancing device.
- Vibration rises gradually with running hours, over months, with no repairs or triggering events. The classic picture of wear-ring wear — the narrow clearances between the impeller and the casing: the clearances open up, the rotor loses hydrostatic support (the water in the clearances stops centering the shaft), and the same amount of unbalance produces an ever-higher level.
- After an overhaul the pump vibrates more than it did before. A common sign of reassembly: the impellers seated on the shaft with different eccentricities, and the distribution of imbalance along the length has changed.
- Axial vibration and thrust-bearing temperature rise together with the radial vibration. That's a reason to check the balance drum or balancing disk and the wear of the seals: the axial force has drifted from its design value.
- On a variable-speed drive there's a speed band where the level jumps sharply and the phase drifts. A sign of a nearby critical speed, and it changes the whole approach to the work.
- The bearing at the free end vibrates noticeably differently from the bearing at the coupling end. Sharp asymmetry points to a couple-unbalance component from impellers spaced along the shaft.
- Subsynchronous components below the rotation frequency have appeared in the spectrum. That's how a loss of rotor stability shows up in worn seals and at the balance drum.
- The pump runs for long stretches below the minimum allowable flow. Recirculation shakes the machine on its own and accelerates wear on the first stages.
No single item here is a verdict of imbalance on its own. But each one means it's time to measure, not to start turning bolts at random.
Sources: ISO 13373-3:2015 · ISO 281:2007
Design: the long rotor, and where imbalance accumulates in it
A multistage rotor is a shaft with impellers, spacer sleeves, and a balance drum fitted onto it in sequence, all clamped together by nuts. Each fit is made with a small clearance or interference and has its own eccentricity — a small offset of the part's center from the axis of rotation — within tolerance. Individually, these microns are harmless. Stacked along the rotor's length at random angles, they produce a distributed imbalance that can't be reduced to a single point.
The second feature: the rotor isn't supported by just the two bearings. The stage wear rings and the balance-drum clearance act as additional hydrostatic supports and center the shaft in the water. As long as the clearances are at their design values, the system is rigid. As they wear, this support disappears, the first critical speed drifts downward, and the machine vibrates harder with the very same masses.
Sectional (ring-section) pumps
Ring-shaped casing sections are clamped together by tie bolts, one impeller per section. Used in water supply, irrigation, mine dewatering. Weak points: the stack-up of fit eccentricities along the length, and rapid wear-ring wear on water carrying abrasives.
Volute-casing multistage pumps
Impellers sit in a single cast or forged casing, often arranged back-to-back to cancel out axial thrust. This layout unloads the thrust bearing, but the couple unbalance from impellers spaced along the shaft doesn't go anywhere.
Feedwater pumps
High speed, hot water, a balancing disk. The rotor almost always operates near or above the first critical speed, meaning it behaves as flexible. Add thermal distortion on top of that: shaft alignment here is checked with a correction for the hot running state.
Borehole pumps
A long column, dozens of stages, a shaft running in the pumped water, rubber-metal bearings. There's nothing to balance on site: there's no access to the rotor, and a measurement at the wellhead mostly shows the column and the motor.
Rigid or flexible: the key question before any balancing
For a short rotor, two-plane balancing is enough at any speed. A long multistage rotor starts to bow as it approaches the first critical speed, and its unbalance stops being describable by two numbers. Weights fitted for the operating speed can drive up vibration at other speeds, and vice versa. We covered the theory in the article on critical speed and flexible rotors; here are just the practical markers.
We check this right on site. With a variable-speed drive we record the 1x amplitude and phase across the whole speed range; at fixed speed we record the coast-down. An hour of work that spares you either a pointless balancing job or an unnecessary teardown.
- The operating speed is roughly 70% of the first critical speed or more. For feedwater pumps this is almost always the case.
- On run-up and coast-down, the 1x amplitude passes through a peak, and the phase swings through nearly 180 degrees.
- The level and phase at the operating point change noticeably with a small change in speed.
- As the wear rings wear down, the critical speed drops, and a machine that ran quietly for years enters an amplification zone without a single change in mass.
The formal distinction between rigid and flexible rotors and the methods for balancing them are described in ISO 21940, with a caveat about the applicable part and edition. We state in the report which marker we used to place your rotor in one category or the other.
Sources: ISO 21940-12:2016
What we check before any talk of weights
We place accelerometers on both of the pump's bearing housings, radially, magnet-mounted on a cleaned spot, plus an axial reading at the thrust bearing. We stick the reflective marker on the coupling half, which the laser phase sensor reads. After that come the checks specific to a multistage machine.
- Axial rotor position and axial float. A worn balancing disk or drum lets the rotor drift along the axis, and the thrust bearing runs overloaded.
- Thrust-bearing temperature and spectral signs: with elevated axial force, it's the first candidate for failure.
- Operating mode: flow not below the minimum allowable, with the valve position and speed fixed. Under recirculation, trial runs aren't comparable to each other.
- Shaft misalignment and a soft foot — poor contact of one of the feet against the frame. On hot pumps we assess shaft alignment accounting for thermal growth: a cold shaft alignment set to zero is a mistake here.
- The share of the 1x running-speed component in the overall level, and the repeatability of its phase from run to run.
- The full spectrum: stage vane-passing frequencies, subsynchronous components, and the high-frequency bearing region.
- The machine's history: what was changed at the last repair, whether new wear rings were fitted, and whether a factory rotor-balancing report exists.
Sources: ISO 13373-3:2015
How a visit to a multistage pump proceeds
- 01
Data before the visit
You send us the pump type, number of stages, speed and its control range, repair history, and photos of the coupling assembly and both bearing housings. From this we assess in advance the chances of balancing on site. What to prepare on site is described in our article on preparing for on-site balancing.
- 02
Baseline measurement
Two accelerometers on the bearing housings, an axial reading at the thrust bearing, and the phase sensor on the marker. We record the overall level, 1x with phase, speed, spectrum, and time waveform at a stable operating point, with flow and pressure logged.
- 03
Checking for flexibility
1x amplitude and phase across the speed range on a variable-speed drive, or a coast-down recording. We look for the critical speed and check how close it sits to the operating speed.
- 04
The fork in the road
Rotor rigid, 1x dominant, phase stable: we balance in two planes. Rotor flexible, or imbalance distributed across the stages: we stop at diagnostics and prepare an assessment for the workshop. We make the decision by the numbers and show them to you.
- 05
Balancing, if it's possible
A trial weight in each plane in turn, calculation by the influence-coefficient method — from the measured relationship of "weight added, vibration changed like this" — fitting the masses on the coupling-half bolts and the free end of the shaft, and a check run. Usually four runs with stops of 15-30 minutes each.
- 06
Trim and coefficients
We save the machine's influence coefficients. The next correction — say, after a season on abrasive water or after a workshop repair — goes ahead in a single run, with no trial weights.
- 07
Report
Before-and-after numbers, spectra, run-up curves, weight masses and positions. Or an assessment with a work order for the workshop, if fitting weights on site would be pointless.
Sources: Balanset-1A operation manual
Correction planes: what's accessible on site, and why it may not be enough
Single-plane correction on a multistage rotor is pointless, and that's not caution speaking, it's geometry. The impellers are spaced along the shaft's length, and their eccentricities add up into static and couple components. A single mass only removes the static part. The couple pair remains, keeps rocking the rotor about its center of mass, and the vibration simply redistributes between the bearings.
The problem with on-site work lies elsewhere: both accessible planes sit at the ends of the rotor, while the imbalance sits in the middle, in the stages.
- Coupling-half bolts. The working plane on the drive side: the mass is built up from washers under the standard bolts, the instrument calculates fixed positions, and nothing is welded or drilled.
- The free end of the shaft: the end face, a nut, or a standard removable element, if the design gives access to it. The radius is small, so the masses end up sizeable.
- The balance drum and the impeller shrouds. This rotor's proper correction planes, the ones a workshop balancing machine works on. They're only exposed when the casing is disassembled, and are inaccessible on site.
- Welding on the shaft and any weights inside the flow path are ruled out: the pressure and flow would tear off an applied weight, and welding would damage the shaft and the seals.
The method's honest limit: as long as the rotor is rigid, correction using the two end planes reduces vibration at the bearings, and that's often enough. But it doesn't remove the bending moment inside the rotor. On a flexible rotor, end weights can improve behavior at one speed and worsen it at another, so we don't propose them there.
Sources: ISO 21940-11:2016 · ISO 21940-12:2016
When on-site work won't work, and what we do instead
On multistage pumps we decline on-site balancing more often than on any other equipment, and we consider that decline an honest part of the job. Here are the typical cases.
- The rotor is flexible: the operating speed is at or above the first critical speed. Two end planes can't reproduce a correction distributed across the stages.
- Imbalance is distributed along the length after wear or reassembly: end correction shuffles vibration between the bearings instead of reducing it.
- Wear-ring wear has eaten away the rotor's stiffness. Restoring the clearances comes first, then a conversation about balancing: weights don't bring back hydrostatic support.
- Axial problems: a worn balancing disk, an overloaded or damaged thrust bearing. That's a repair, not a balancing job.
- A borehole pump in the column: there's no access to the rotor at all, and balancing the motor at the wellhead only fixes the motor.
- The operating point won't hold steady: flow below the minimum, cyclic operation. Trial runs can't be compared with each other.
Instead of useless weights, you get work that moves things forward. A full vibration diagnosis: how much of the level comes from 1x, what the seals contribute, and where the critical speed sits. An assessment for the repair: which clearances and runouts to check, what to replace. A work order for the workshop: balancing the assembled rotor on a balancing machine, with individual balancing of the impellers during assembly, a final correction on the assembly, and a tolerance to balance quality grade G — the balance quality grade from ISO 21940. After reinstallation we come back, take a check measurement at the operating point, and confirm the result with a report. Any small residual unbalance left after the workshop we take up with trim balancing on site: that's exactly the legitimate territory for on-site work.
Sources: ISO 21940-11:2016 · ISO 20816-1:2016
What you get, and how to book
The result of a visit is numbers, not impressions, and they're just as valuable whether balancing was carried out or the conclusion was "this needs a workshop."
- Overall vibration and 1x with phase at both bearings and in the axial direction, before and after, in mm/s RMS (root-mean-square value).
- A run-up or coast-down characteristic marking the critical speed, if it was recorded.
- Annotated spectra: running-speed, vane-passing, and subsynchronous components, plus the high-frequency region.
- Masses, radii, and positions of the fitted weights, and saved influence coefficients for future trim corrections.
- A condition assessment by zone, stating the applicable part and edition of ISO 20816, and, if on-site balancing is declined, an assessment with a work order for the workshop.
- Pump type and number of stages, speed, power, and whether the drive is variable-speed.
- Running hours since the last overhaul, and what was changed in it: impellers, wear rings, drum, bearings.
- How the vibration changed: gradually with running hours, in a sudden jump, or after a repair.
- Photos of the coupling assembly, both bearing housings, and the free end of the shaft.
- Whether the speed can be varied or a coast-down recorded, the allowable number of runs, and the duration of stops.
- The site address and a convenient time window.
Packages start from 550 EUR per unit (diagnostics 300 EUR + balancing from 250 EUR), with a minimum invoice per visit of 500 EUR: the travel, sensor setup, and diagnostics take time regardless of the outcome. A diagnostic visit ending in an assessment is priced by the same rules. The calculator on our website gives an exact estimate. We design and manufacture the Balanset instruments and do the on-site balancing with them ourselves. We're based in Vila Nova de Gaia near Porto and work throughout Portugal.
Sources: Balanset-1A operation manual · Balanset-1A manufacturer specification
Frequently asked questions
Can a feedwater pump be balanced on site?
More often no than yes. Feedwater pump rotors run near or above the first critical speed and behave as flexible, and end-plane correction on a rotor like that can improve behavior at one speed while worsening it at another. We record the amplitude-phase characteristic across the speed range, show where the critical speed sits, and decide by the numbers: balance, repair, or send the rotor to the workshop for a balancing machine.
The factory balanced each impeller individually, but the pump shakes after assembly. Why?
Balancing each part individually doesn't cancel out fit eccentricities. Each impeller seats on the shaft with its own offset within tolerance, the angles are random, and a distributed imbalance builds up along the rotor's length that wasn't there on the balancing arbor. That's why a sound procedure calls for a final balance of the assembled rotor, plus alignment marks on the parts, so the next teardown doesn't undo the result.
Vibration rose over months with no repairs at all. Is that imbalance?
On a multistage pump this points first to suspected wear-ring wear. Their clearances act as hydrostatic supports, centering the rotor and adding stiffness. As they wear, the support weakens, the critical speed drops, and the same old residual unbalance produces an ever-higher level. Weights here only treat the symptom. We measure, assess the seals' contribution, and say plainly when it's a ring replacement that's needed, not balancing.
Only the coupling half is accessible. Will a weight there help?
Only partly, and only on a rigid rotor. One plane removes the static component of the imbalance, while the couple pair from impellers spaced along the shaft remains and redistributes vibration between the bearings. The minimum for a multistage rotor is two planes: the coupling half and the free end of the shaft. If the second end isn't accessible, we honestly set expectations before the visit even happens.
A borehole pump is vibrating. What can you do?
The pump rotor itself can't be balanced on site: it's down in the borehole, inside the column, with no access. At the wellhead we can measure the vibration, separate the motor's contribution from the column's, and check the motor itself, including balancing its rotor if needed. If the source is in the pump end, there's only one solution: pull it, inspect the stages and bearings, and balance the rotor on a machine.
What's the point of your visit if on-site balancing isn't possible?
You get a measurement and a diagnosis instead of guesswork: the share of 1x, the condition of the seals and the thrust bearing, where the critical speed sits, and a zone assessment per ISO 20816 with a note on the applicable part. Plus an assessment with a work order for the workshop: what to check during teardown and what grade G to specify when balancing the assembled rotor on a machine. After the repair we take an acceptance measurement, and we take up any small residual unbalance with trim balancing on site using the saved coefficients.
Related content
On-site pump balancing: impellers, rotors, and shafts of pump units
Yes, we balance pump rotors on site, in their own bearing supports, with no dismantling and no sending the impeller to a machine shop. Three conditions apply. First: most of the vibration has to come from the 1x running-speed component — vibration at the rotor's rotational frequency, the main sign of imbalance. Second: at least one correction plane has to be accessible, meaning a spot where a weight can be fitted — usually the coupling bolts or the impeller nut. Third: the pump has to be able to hold a stable mode with a constant flow rate. If the level is being driven by cavitation, operation away from the best-efficiency point, shaft misalignment, or worn wear rings, we'll show that with a measurement and say plainly: weights won't help.
On-Site Pump Impeller Balancing: Closed, Semi-Open and Open Impellers
Yes, we balance pump impellers on site, in their own bearing supports, without removing the impeller from the shaft. Three conditions apply: the running-speed component accounts for most of the vibration level, and its phase repeats from start to start; the impeller has been cleaned of deposits; and there is at least one correction plane — a place where mass can be added or removed — that is actually reachable. If erosion has eaten the blades down to uneven thickness, the disc has cracked, or the clearance on a semi-open impeller has gone beyond the manufacturer's tolerance, balancing will only postpone replacement. In that case we say so directly, before the visit.
Hydraulic Causes of Pump Vibration: Cavitation, Vane-Pass Frequency, and Operating Point
Hydraulics produces vibration that doesn't live at running speed, and weights have no effect on it. Cavitation produces a "gravel" crackle and a broadband rise across roughly 1 to 10 kHz, recirculation at low flow produces unstable subsynchronous components (vibration at frequencies below running speed), and vanes passing the diffuser or volute tongue produce a peak at the vane-pass frequency f_v = z · n / 60, where z is the number of vanes and n is the speed in rpm. Telling all this apart from imbalance is straightforward: hydraulics responds to valve position, flow rate, and suction head, while imbalance responds to nothing except speed. First restore the operating point and the NPSH margin (the cavitation pressure margin at the suction), then re-measure, and only then decide whether balancing is needed.
Describe your equipment and the problem
We'll answer your questions, clarify the details, and let you know what's needed for an estimate and a visit.