Vibration Monitoring for Motors and Pumps

Motor vibration monitoring: a practical guide for plant engineers

Vibration is the most direct mechanical signal a rotating machine produces, and motor vibration monitoring turns that signal into actionable maintenance information. Every motor and pump converts electrical energy into rotation, and any defect in balance, alignment, bearings or hydraulic condition leaves a fingerprint in the vibration signature. For electrical engineers and plant buyers specifying condition monitoring systems, the engineering task is not simply to buy sensors, but to define what to measure, where to mount transducers, how to set alarm thresholds, and which acquisition architecture fits the criticality of the asset. This guide covers the fundamentals of motor vibration monitoring and pump vibration monitoring without product-specific claims, so it can be used as a specification aid during system design and vendor evaluation.

What motor vibration monitoring reveals about machine condition

Vibration analysis is diagnostic because different failure modes excite different frequencies. A well-designed monitoring system separates these components so that maintenance teams can act before functional failure. The main failure modes are well understood in rotating machinery practice.

Imbalance

Mass imbalance produces a strong component at the rotational frequency (1×). It is common after fan or impeller fouling, coupling changes or rotor modification. In a spectrum, a dominant 1× peak with low harmonic content typically points to imbalance rather than misalignment. Balance quality affects bearing life directly, because the centrifugal force grows with the square of speed.

Misalignment

Shaft misalignment between motor and driven pump or fan usually appears as elevated 2× rotational frequency, often with 1× and higher harmonics, and may show axial vibration of significant amplitude. It is one of the most common causes of premature bearing and coupling wear, and it is correctable by alignment procedures rather than parts replacement.

Bearing defects

Rolling element bearing faults generate characteristic frequencies — ball pass frequency outer race (BPFO), ball pass frequency inner race (BPFI), ball spin frequency (BSF) and fundamental train frequency (FTF). These are calculated from bearing geometry and shaft speed, so the monitoring system needs the bearing designation and running speed to interpret the spectrum. Early-stage defects appear as low-amplitude, high-frequency impact energy, which is why envelope or high-frequency techniques are used alongside overall level measurement.

Looseness

Mechanical looseness — soft foot, loose foundation bolts, worn bearing fits — typically produces a series of harmonics of running speed (1×, 2×, 3× …) and sometimes sub-harmonics. It is often found during installation checks and should be corrected before it accelerates bearing damage.

Cavitation and hydraulic faults in pumps

In pump vibration monitoring, cavitation appears as broadband, high-frequency random energy, often described as noise rather than discrete peaks. It indicates insufficient net positive suction head (NPSH), restricted suction, or operation far from the best efficiency point. Flow turbulence, recirculation and vane pass frequency effects also show in the pump spectrum. Because cavitation damages impellers quickly, early detection has direct maintenance value.

Sensor placement on bearing housings

Transducer location determines whether the data is useful. General practice for motor vibration monitoring is to mount accelerometers on the bearing housings, as close as possible to the bearing load zone, in the radial direction (horizontal and vertical) and, where axial thrust is a concern, in the axial direction. For a typical motor-pump set, this means drive-end and non-drive-end motor bearings, plus pump inboard and outboard bearings, with at least one axial point per machine train.

• Mount on a rigid, machined surface; avoid thin covers, fan shrouds and flexible panels.
• Use stud mounting where permanent measurements are required; adhesive mounting is acceptable for route-based surveys but reduces high-frequency response.
• Keep cable runs protected from heat, oil and mechanical abrasion, and strain-relieve connectors.
• Record the mounting orientation and location in the asset register so spectra remain comparable over time.
• For wireless sensors, verify that the selected mounting method and location give repeatable readings before committing to a full deployment.

Overall level versus spectrum: choosing the right measurement

Overall vibration level (often expressed as RMS velocity in mm/s over a defined frequency band, for example 10–1000 Hz) is a simple scalar that is easy to trend and to alarm on. It is well suited to general surveillance of many machines. The spectrum, by contrast, resolves amplitude against frequency and is what allows fault identification: 1×, 2×, bearing frequencies and broadband noise are separated and can be compared with ISO 10816/20816 evaluation zones for machines of similar type and power.

• Overall level: low data volume, simple alarms, good for trending and prioritisation.
• Spectrum: needed for diagnosis, bearing condition assessment and root-cause analysis.
• Envelope / high-frequency acceleration: sensitive to early rolling element defects, useful when bearing damage is suspected before it appears in velocity RMS.
• Time waveform: useful for impacting, gear meshing and transient events.

In practice, a two-tier approach works well: overall level for continuous surveillance and spectrum collection at intervals or on alarm for diagnosis. The measurement band and units should be fixed in the specification so that historical data remains comparable.

Alarm setting philosophy

Alarm thresholds should be based on a baseline of the machine in good condition, not on a single generic number. A practical philosophy uses several layers:

• Baseline: record overall level and a reference spectrum after commissioning or overhaul, at defined load and speed conditions.
• Alert level: a change of typically 2–4 dB above baseline (or a defined percentage increase), prompting inspection and a repeat measurement.
• Alarm level: a larger change, or an absolute value approaching the ISO 10816/20816 zone boundary for the machine class, prompting planned intervention.
• Rate-of-change alarms: catch fast-developing faults such as lubrication loss or looseness, which may cross thresholds between scheduled intervals.

Alarms should account for normal operating variability — load changes, speed changes, ambient conditions and transient starts — so that false triggers do not erode confidence in the system. Where speed varies, speed-normalised or order-based analysis is preferable to fixed frequency bands.

Wired online, wireless, or route-based: trade-offs

There is no single correct architecture; the choice depends on criticality, accessibility and the maintenance organisation.

• Wired online systems: continuous data, high bandwidth for spectra and waveforms, best for critical machines where downtime cost is high. Installation cost and cable routing are the main constraints.
• Wireless sensors: lower installation effort, suitable for assets that are difficult to cable or where continuous high-frequency data is not required. Battery life, sampling interval and data volume must be planned.
• Route-based (portable) collection: lowest capital cost, flexible, and well suited to large populations of less critical machines. It depends on a disciplined collection schedule and consistent measurement points.

A common pragmatic strategy combines route-based collection for general machinery with online monitoring on the most critical motor-pump sets, and wireless where cabling is impractical.

What a first deployment on a critical motor/pump set looks like

A first deployment should be scoped as an engineering project, not a purchase. A workable sequence is:

• Define the asset boundary: motor, coupling, pump, driven equipment, and the criticality ranking that justifies monitoring.
• Collect machine data: rated power, speed, bearing designations, coupling type, pump impeller data, and operating envelope.
• Select measurement points and mounting methods, and document them on a machine drawing.
• Record a baseline at representative load and speed, including overall level and spectra.
• Configure alert and alarm thresholds from the baseline and the applicable ISO zone.
• Define the response workflow: who reviews alarms, what inspection is triggered, and how corrective work is recorded.
• Review data after the first weeks of operation and adjust thresholds if normal variability causes nuisance alarms.

Comparison and checklist

Where to go next

For teams specifying hardware and systems, the manufacturer's condition monitoring pages provide the product-level detail that complements this guide: see vibration monitoring for sensor and system options, and motor equipment monitoring for motor-focused monitoring configurations. LOGGON TECH supplies industrial metering and monitoring equipment from Shenzhen, China; technical enquiries should reference the machine data listed in the deployment checklist above.

FAQ

How often should vibration data be collected?

It depends on criticality and fault development rate. Route-based surveys are often monthly or quarterly for general machinery; critical machines are monitored continuously or at short intervals. The interval should be short enough that a developing fault is detected before it reaches the alarm level.

Can one sensor cover both motor and pump?

Generally no. Motor and pump bearings are separate measurement locations with different fault signatures. A machine train normally requires points on each bearing housing, plus axial measurement where thrust is a concern.

What units should be used for overall vibration level?

Velocity RMS in mm/s over a defined band (commonly 10–1000 Hz) is widely used for machine condition evaluation and is the basis of ISO 10816/20816 zones. Acceleration is used for high-frequency and bearing assessment, and displacement for low-speed machines.