VESSEL MACHINERY HEALTH GUIDE

Vessel Machinery Condition Monitoring: Baselines, Trends and Alarm Decisions

A useful machinery-health programme does more than collect readings. It links each asset and probable failure mode to a repeatable measurement, comparable operating condition, defensible alert and named maintenance decision.

Technical guidance12 min readPublished 2026-08-16Vessel Doctor Technical Team
Quick answer

Trend only comparable measurements: the same asset, point, direction, sensor method and documented speed/load state. Use absolute criteria, change from a healthy baseline and rate of change as separate evidence. Validate an alert before diagnosing a fault, then state urgency, confidence, action, verification method and next monitoring interval.

01

Start with the maintenance decision and asset criticality

Condition monitoring is valuable when a change can trigger a defined decision: continue operation, shorten the interval, inspect, lubricate, align, repair at the next opportunity or stop for an authorised technical review. Build the asset register around safety, redundancy, failure consequence, repair lead time and detectability—not around which instruments happen to be available.

Programme rule

For every monitored asset, write the failure mode, detectable indicator, measurement interval, alert owner and permitted response. A data route without an owner or decision is only data collection.

02

Match the probable failure mode to the sensing method

Failure-mode-to-evidence planning
Condition questionUseful evidenceLimitation to control
Rotor, coupling or structural responseOverall vibration, spectrum, waveform, phase, speed and load.One high value cannot identify imbalance, misalignment or resonance by itself.
Bearing or lubrication deteriorationHigh-frequency/envelope data, lubricant evidence, temperature and inspection.Mounting, speed and lubrication changes can imitate a developing fault.
Electrical or thermal problemTemperature pattern, current/electrical evidence, load and ambient condition.A hot surface is not automatically an internal defect.
Loss of hydraulic or process performancePressure, flow, temperature, valve position, power and vibration together.Process changes must be separated from mechanical deterioration.

ISO 17359:2018 gives general procedures for setting up a machine condition-monitoring programme. ISO 13373-1 addresses vibration measurement and data collection, including parameters, transducers, locations, attachment and operating conditions. The applicable project standard must still be selected for the actual machine and decision.

03

Design a repeatable onboard measurement route

  • Give every machine and measurement point a stable identifier.
  • Record bearing position, horizontal/vertical/axial direction and sensor attachment.
  • Use the same measurement quantity, bandwidth, sampling setup and instrument chain.
  • Capture RPM, load, running auxiliaries, valve state and vessel operating context.
  • Separate fixed online monitoring from periodic handheld routes and one-time diagnostics.
  • Record access constraints, changed foundations, overhaul work and sensor relocation.

A shipboard route may cover engines, generators, pumps, fans, compressors, gearboxes and auxiliaries at berth, anchorage, harbour trial, sea trial or routine operation. Measurements from those states should not be merged as if they were identical.

04

Build a healthy baseline before judging the trend

A baseline is a documented reference condition, not merely the first value collected. Confirm the machine was believed healthy, the operating state was representative and the signal was technically valid. Use multiple comparable readings when possible so normal variability is visible.

Comparable trend

  • Same point, direction and attachment.
  • Same quantity and frequency range.
  • Comparable speed, load and process state.
  • Known maintenance and operating history.

Non-comparable change

  • Different point, mounting or sensor.
  • Different RPM/load or valve position.
  • Temporary nearby work or vessel event.
  • Data-quality, clipping or unit error.
05

Use three alarm views instead of one generic limit

  1. Absolute level: compare the defined quantity with the machine-specific manufacturer, project or applicable standard basis.
  2. Change from baseline: identify a material departure from that machine’s known healthy condition.
  3. Rate of change: distinguish a stable elevated condition from accelerating deterioration.

ISO 20816-1 provides general conditions for measurement and evaluation using vibration magnitude and change, but machinery-specific parts, manufacturer criteria and the actual measurement location can matter. Do not apply a single universal alarm to every onboard machine.

06

Validate an alert before declaring deterioration

Repeat the reading, check units and sensor contact, confirm speed/load, compare related points and examine the raw spectrum or waveform. Review overhaul, lubrication, process and environmental changes. A valid condition change should survive these checks.

Keep measured fact separate from interpretation: “velocity increased from the documented baseline at the drive-end horizontal point under comparable load” is evidence; “bearing failure” is a hypothesis requiring corroboration.

07

Escalate from monitoring to focused diagnostics

When a trend needs a different scope
Observed changeFocused next stepDecision sought
Running-speed or harmonic growthMarine machinery vibration analysis with spectrum, waveform and phase.Test imbalance, alignment, looseness, eccentricity or response hypotheses.
Narrow RPM-sensitive increaseRun-up/coast-down and source-path-response comparison.Determine whether resonance or transmitted forcing is credible.
High-frequency bearing indicationEnvelope/high-frequency measurement plus lubricant and temperature evidence.Confirm severity and maintenance timing without relying on one indicator.
Cabin or deck responseStructural/hull or habitability vibration survey.Separate machine condition from transmission and receiver response.
08

What an action-ready condition report should contain

  1. Asset register, criticality and monitored failure modes.
  2. Point/direction map, instrument chain, settings and traceability.
  3. RPM, load, process and vessel condition for every data set.
  4. Baseline and trend plots with non-comparable readings identified.
  5. Alert basis and the exact evidence that triggered review.
  6. Measured fact, diagnostic interpretation, confidence and limitations.
  7. Urgency, responsible action, verification measurement and next interval.
09

Class, flag and supplier-qualification context

Ordinary maintenance diagnostics, a machinery-condition class notation, formal acceptance testing and an approved service-supplier activity are not automatically the same scope. Requirements may depend on the vessel, flag administration, classification society, notation, contract, manufacturer and attending surveyor. Confirm the route before mobilisation for an ABS-, BV-, CCS-, CRS-, DNV-, IRS-, KR-, LR-, ClassNK-, PRS-, RINA- or Türk Loydu-classed vessel.

This guide does not state that Vessel Doctor is approved or recognised by a classification society or flag administration and does not guarantee report acceptance.

10

Five items to send for a fast programme review

  1. Vessel, machinery register and critical equipment.
  2. Available history, alarms, failures and recent overhaul work.
  3. Normal RPM, load, duty cycle and operating states.
  4. Existing points, sensors, instruments, routes and limits.
  5. Required maintenance decision, report route and deadline.

Attendance can be assessed for Singapore, Malaysia, India, Sri Lanka, Indonesia, Thailand, the Philippines, the Middle East, Asia-Pacific and worldwide, subject to access and availability.

NEXT STEP

Turn the machinery question into a controlled onboard scope.

Send the vessel, machine, operating condition, symptom, available history and required decision. The scope will define the measurement route, analysis evidence and deliverables before mobilisation.

Review the canonical serviceRequest a technical scope