Pump Condition Monitoring for Critical Industrial Pumps
Pump condition monitoring is most useful when a change in vibration, temperature or process behaviour leads to a defined maintenance decision. addanode scopes instrumentation, operating context and addaNet reporting around the pumps your site depends on. Start with a critical pump register, relevant failure history and a named maintenance owner, then verify the measurements and response workflow on representative operating conditions before extending the installation.
Choose measurement capability around the pump question
Pump condition monitoring should connect an operating question to the measurements that can answer it. A trend summary may support a change review, while a specialist may need waveform or frequency information, speed and hydraulic context for a particular diagnosis.
Ask a supplier to state what the selected device actually records and exports: scalar summaries, waveform length, frequency information and timing are different capabilities. Confirm mounting, access and the range of operating conditions represented in the first data collection.
Compare periodic inspection, reuse of existing process measurements and additional instrumentation for the target pump. Price measurement installation and verification separately from specialist interpretation and site work. Bring the pump/drive details, duty changes, existing pressures/flow and maintenance history; these make the trial more useful than an unqualified alarm demonstration.
| Required decision | Capability to verify | Scope implication |
|---|---|---|
| Has behaviour changed? | Repeatable measurement under comparable duty | Baseline and operating-context capture |
| What might explain the change? | Evidence needed by the responsible mechanical reviewer | Qualified acquisition and interpretation |
| What should maintenance do? | Work ownership and inspection route | A response process alongside the data service |
Select pumps by the consequences of losing the service
A duty pump with a proven standby arrangement presents a different maintenance problem from the only transfer pump feeding production. Record the service each pump provides, the available standby capacity, access constraints and the time needed to restore operation. Include the consequences for the process, maintenance workload and availability of parts. This makes the first installation a decision about operational priorities rather than simply instrumenting the easiest machines to reach.
Review the failure history with the maintenance team. Repeated seal work, bearing replacement, suspected suction problems or recurring trips point to different information needs. Retain the observed symptom, work performed and inspection findings where available. A previous work-order description is a useful starting point, but the project should distinguish an observed failure from an unconfirmed explanation recorded during an urgent repair.
Build a register of the complete pump train
Record pump type, manufacturer, duty, motor, coupling or drive arrangement, bearing locations and the relevant operating-speed range. Identify variable-speed equipment and how speed will be obtained. Include the fluid, usual temperature range, suction source, discharge arrangement and the process conditions that change during a shift. The condition record needs this context to make comparisons meaningful.
Assign identifiers that match the site's maintenance system and drawings. A pump and its motor may need separate measurement points while remaining part of one train. Mark sensor locations with photographs and an orientation reference. A technician returning after an overhaul should be able to find the same point and understand whether a changed mounting arrangement affects comparison with the earlier baseline.
| Asset information | Why it matters to the monitoring scope | Evidence to bring |
|---|---|---|
| Pump and drive configuration | Establishes the components and accessible measurement points | Equipment list, drawings and nameplate details |
| Speed and operating duty | Supports comparisons between relevant running conditions | Drive records, schedules or observed speed range |
| Fluid and suction arrangement | Helps interpret hydraulic changes alongside mechanical observations | Process diagram and available pressure or level records |
| Failure and repair history | Identifies the decisions the team needs to improve | Work orders and inspection findings |
| Standby and access conditions | Sets the practical response and installation priorities | Maintenance procedure and permitted work windows |
Choose each signal for a question the team can answer
Vibration can reveal a change in mechanical behaviour, provided the sensor, mounting and acquisition settings suit the task. Temperature can show a developing trend at the measured location. Motor current and operating state provide load and running context. Suction and discharge pressures, flow and source level can help explain whether a change coincides with a hydraulic operating condition.
Choose the required combination with the reliability and process owners. An overall vibration trend may support a screening decision; a more detailed investigation may require suitable waveforms, frequency information, speed context or specialist measurements. The wireless vibration sampling specification guide gives procurement questions for checking exactly what a proposed sensor measures, retains and makes available for review.
Assess mounting and measurement access before ordering
The installation survey should identify a suitable measurement surface, the expected orientation and the means of fixing the sensor. Keep the sensor within its specified environmental and operating conditions. Access to the bearing area, guarding, washdown, corrosion exposure and the installation work window can influence the practical choice. The site engineering owner should approve the physical arrangement and any work around the pump train.
Record the mounting method with the baseline. A change from one surface or fixing method to another may alter the measured response even when the machine has not changed. After sensor replacement or overhaul, review that configuration before interpreting a step in the trend. The handover should include enough information for maintenance staff to preserve the intended measurement arrangement during routine work.
Build baselines around actual running conditions
Compare readings taken under sufficiently similar conditions. A pump starting, running steadily, changing speed or operating at a different flow can produce different observations. Record the state and relevant process context alongside the condition measurement. Where several operating modes are normal, retain examples of each and agree how the report separates them.
The initial baseline also needs a maintenance judgement about the asset's condition. A stable trend from a machine already known to have a problem is a record of that condition, not proof of a healthy reference. Agree what inspection or historical evidence supports the baseline. Following repair, retain the before-and-after records with the work order so a new reference can be introduced with an explanation.
Investigate hydraulic symptoms with process evidence
For a centrifugal pump, a change in noise or vibration may coincide with a change in suction conditions, fluid temperature or operating point. KSB's pump lexicon describes vapour cavitation in terms of local pressure falling to the fluid's vapour pressure, with bubble collapse as pressure subsequently rises. That mechanism explains why fluid and pressure context belong alongside mechanical observations when the team investigates suspected cavitation.
The first review should align the event times with available suction level, pressure, discharge behaviour, valve state and speed records. The responsible pump or process specialist then evaluates the evidence against the equipment and operating arrangement. A vibration alert becomes more useful when the reviewer can see what the pump was doing and what changed around it, rather than receiving a single severity number without context.
Preserve the distinction between screening and diagnosis
Use report labels that describe the evidence collected: changed vibration, elevated measured temperature, abnormal operating pattern or missing data. The maintenance workflow can then identify what needs checking and who will perform it. A detailed fault assessment may require additional measurements or inspection, particularly when several possible causes could produce similar symptoms.
Keep the review conclusion linked to its supporting evidence. Record whether an issue was confirmed through inspection, resolved through an operating change or left under observation. If the initial hypothesis changes, retain that history. This gives the team a growing record of which measurements support useful decisions on its own machines and where the monitoring or diagnostic process needs improvement.
Design collection and communications separately
Specify the local acquisition rate, measurement duration, calculated features and reporting interval for the selected devices. A sensor may collect a short waveform and transmit a summary later; another arrangement may make more detailed records available through an approved interface. The reporting design should preserve the time and operating conditions of the underlying measurement.
LoRaWAN and 4G can carry monitoring information within the agreed architecture. The scope identifies the gateway or edge arrangement, retained records and what happens during an interruption. Missing data should appear as a data-quality condition that prompts investigation. Define how the team checks whether a quiet trend means stable operation, a stopped pump or a source that has stopped reporting.
Turn an alert into an accountable maintenance action
Agree alert rules and review priorities with the asset owner, using the available baseline and equipment context. Name the first recipient, the expected review route and the escalation contact. The notification should provide the asset identifier, measured change, event time and links to the relevant history. The person receiving it needs enough context to decide the next observation or work request.
The work record should retain acknowledgement, inspection or additional measurement, the decision made and the eventual outcome. Link it to a work-order reference where the site uses one. After repair or an operating adjustment, inspect the subsequent trend under comparable conditions. This closes the operational loop and creates evidence for revising the alert rule or baseline when the asset owner considers a change justified.
Commission the measurements and the user's workflow
| Commissioning exercise | Evidence retained | Acceptance decision |
|---|---|---|
| Verify asset and point mapping | Equipment identifier, location photograph and orientation | Each trend belongs to the correct component and point |
| Compare representative readings | Suitable reference method and operating condition | Measurements meet the agreed comparison criteria |
| Observe multiple running states | Speed, duty and process context for each interval | Normal modes remain distinguishable in reporting |
| Exercise an approved test alert | Trigger, notification, acknowledgement and review record | The named team can follow the response procedure |
| Test interruption and recovery | Missing-data indication and retained or recovered records | The actual configured recovery behaviour is understood |
| Complete handover | Configuration, open defects and support responsibilities | The customer can operate the accepted scope |
Choose reference methods and numerical tolerances before testing. An authorised simulated alert can verify routing without forcing the pump into an abnormal condition. Separate that workflow test from the evidence needed to assess real machine behaviour. Keep untested conditions visible and assign a follow-up action. Acceptance should demonstrate both the collection path and the process for acting on the information it produces.
Define the quotation around the whole delivery
Initial costs depend on the number of pump trains and measurement points, sensor type, mounting work, access requirements, available process signals and the proposed communications arrangement. Configuration includes asset mapping, operating-state context, reports, alert routing and commissioning. Any connection to an existing control or maintenance system needs an agreed interface, data owner and mapping scope.
Ongoing responsibilities include battery or power maintenance, instrument checks, replacement configuration, platform and connectivity charges, specialist review and site support. Identify which activities the customer's team performs and which are included in the service agreement. Expansion can reuse the reporting approach while still reviewing different pump types, speeds, fluids and access conditions. Compare proposals using these defined responsibilities as well as the equipment list.
Prepare the first pump group for review
Send the critical pump list, known recurring problems, relevant work orders and the available process records. Include photographs of potential measurement points and identify the maintenance owner who can witness operation. These inputs help define a first stage whose baseline, acceptance tests and response procedure can be agreed before installation.
The parent asset condition-monitoring service places pump trains within a wider maintenance programme. The manufacturing overview connects the requirement with production and shared utilities. For common project-scoping mistakes and the importance of an actionable response, see the vibration-monitoring project guide.
Frequently asked questions
Which pumps should be included first?
Prioritise the pumps whose loss of service matters most and for which the team can define a useful maintenance response. Provide duty, standby arrangements, repair history and access information to select representative assets.
How is variable-speed operation handled?
Record speed or an appropriate operating-state reference alongside the measurement. Compare relevant conditions and select the acquisition and analysis approach around the actual speed range and failure questions.
Can pressure and level data be included?
Yes. Available approved instrument or system interfaces can supply process context. The scope identifies measurement locations, units, timestamps and ownership so those records can be compared with the condition event.
Who decides what maintenance action to take?
The site's named maintenance and engineering owners review the evidence through the agreed workflow. The installation provides measurements, history and alert routing, with specialist assessment and work-order responsibilities defined in the project scope.
Technical reference
KSB: cavitation in centrifugal pumps provides the hydraulic mechanism referenced above. Reviewed 5 November 2026. Measurement selection, baseline development and the acceptance matrix are engineering planning methods to be agreed for the actual pump train and operating conditions.
Define a first stage your team can verify.
Tell us the site, equipment, existing records and the decision you need to improve. We will use those inputs to discuss the scope and acceptance evidence.