Ship Propulsion Vibration Troubleshooting Test Matrix

Separate the measurement quantity before selecting the instrument.

A complaint such as “the shaft vibrates,” “the vessel shakes at one RPM,” or “the noise started after propulsion work” is a starting observation, not a measurement definition. Torsional vibration is angular oscillation about a shaft axis. It may be expressed as angular displacement or velocity, dynamic torque, strain-derived stress or another quantity established by the approved method. Machinery vibration normally describes motion measured on casings, bearings or foundations. Structural vibration describes response of decks, hull, superstructure, panels or supports. Acoustic work addresses sound pressure or intensity and its airborne or structure-borne path.

Freeze the question, calculation and controlling document.

Before attendance, identify the decision owner, shaftline arrangement, engine or motor, gearbox, couplings, propeller or driven equipment, operating range, known restricted ranges, recent modifications and required witness or report. Where a torsional calculation exists, record its revision, assumed inertias and stiffnesses, damping basis, excitation orders, predicted resonances, stress or torque outputs, allowable or agreed limits and the exact quantities and locations available for field comparison. A calculation-to-measurement comparison is only meaningful when the definitions, reference locations, units, amplitude conventions and operating states are compatible.

Do not invent a universal pass criterion. The controlling basis may come from an approved calculation, maker instruction, class rule, flag requirement, contract, test memorandum or competent diagnostic plan. IACS itself does not approve suppliers, and a survey report should not imply class, flag or customer acceptance before the responsible reviewer confirms it.

Convert observations into testable alternatives.

Record the exact receiver, time, RPM or load band, direction, sound character, weather or sea condition, propulsion lineup, machinery state, onset history and any maintenance or damage event. Then list plausible alternatives without promoting one to a conclusion. Possible paths may include torsional resonance, shaft alignment or coupling behaviour, engine firing or motor excitation, propeller-order forcing, gear-mesh response, bearing or casing motion, foundation or structural resonance, looseness, piping response, cavitation, ventilation noise or a combination.

The purpose of the register is not to make the list long. It prevents confirmation bias and shows why each channel or operating run exists. Remove a hypothesis only when the available evidence justifies doing so, and keep untested alternatives visible in the report.

Make every run reproducible and safe.

Prepare the run sequence with the vessel operator and responsible technical parties. Capture engine, motor and shaft RPM; propeller pitch; gearbox state; power, torque or load where available; vessel speed; draught and trim; displacement or loading condition; generators, thrusters and auxiliaries; heading, sea state and water depth; and the stable record duration. Record whether the run is a steady hold, controlled step, run-up, run-down, barred-speed passage, berth test or sea-trial condition.

Keep the measured and derived quantities traceable.

For each torsional channel, record the transducer principle, shaft or component location, reference geometry, installation and pickup details, calibration or verification evidence, measured quantity, units, amplitude convention, sensitivity, range, sampling rate, bandwidth, filtering, integration or differentiation, pulse-per-revolution reference and any processing used to obtain torque or stress. If strain is converted to torque or stress, retain the material, geometry, bridge, gauge-factor and transfer assumptions used by the responsible analyst.

Measurement at one shaft location does not automatically represent response or stress everywhere in the shaftline. Compare the measured location with the calculation model and state the transfer or inference used. Do not use casing acceleration as a substitute for a required torsional quantity.

Synchronise channels before comparing cause and response.

A common clock and reliable rotational reference allow the analyst to relate shaft twist, casing motion, structural response, sound and operating logs. Record the time source, clock offset checks, tachometer or encoder location, pulses per revolution, missing-pulse handling, speed resolution, order-tracking method and any resampling. Note whether engine and shaft speed differ through gearing or slip and whether separate references are needed.

An order line indicates a relationship to rotational speed; it is not a source diagnosis by itself. Use phase, location, transfer paths, controlled changes, calculations, inspection and repeatability to increase or reduce confidence.

Map the wider vessel response without confusing it with shaft twist.

When the complaint extends beyond the shaftline, record reference channels at candidate sources, along suspected transmission paths and at affected receivers. Candidate source points may include engine or motor casings, gearbox bearings, thrust bearing, shaft bearing supports, pumps, fans and hydraulic equipment. Path points may include foundations, chocks, deck panels, pipe supports, resilient mounts, bulkheads and structural junctions. Receiver evidence may include cabin floors, control positions, panels, fittings or sound-pressure positions.

Use consistent identifiers on drawings, photographs, channel files, plots and tables. A large receiver response can result from strong excitation, efficient transmission, a local resonance or several mechanisms together. Preserve that boundary until the evidence separates them.

Change one authorised condition at a time where practicable.

A controlled change can test a mechanism more efficiently than adding unrelated measurement points. Examples may include an authorised machinery combination, load step, pitch or RPM hold, ventilation state, support condition, removable panel state or temporary diagnostic treatment approved by the responsible parties. Record the baseline, exact change, safety approval, expected response, channels held constant, result and restoration. Do not adjust protection settings, restraints, alignment, balancing weights, supports or operating restrictions solely for diagnosis without proper authority and risk control.

Qualify weak runs before they reach the conclusion.

Check stable speed and load, required propulsion configuration, adequate duration, shaft pickup quality, sensor security, calibration or field checks, clock alignment, tachometer dropout, telemetry quality, clipping, overload, electromagnetic interference, unexpected transients, sea or depth changes and unauthorised operating changes. Record unavailable or incomplete conditions rather than silently replacing them with a convenient run.

Uncertainty includes more than instrument calibration. Location transfer, model assumptions, mounting, derived quantities, speed variation, operating-state mismatch and incomplete access can dominate the confidence of a comparison. State the material contributors and their effect on the decision.

Separate measured facts from interpretation.

“A peak occurred at order X” is a measured or processed observation when the method is traceable. “The coupling caused the complaint” is an interpretation that needs location, phase, transfer, controlled-change, calculation, inspection or other corroboration. State alternative mechanisms and explain the confidence assigned. Do not convert a diagnostic comparison into a formal pass or class-acceptance statement.

Prepare publishable first-party proof without inventing a result.

Before corrective work, freeze the shaft location, transducer setup, reference channels, clock and RPM method, run sequence, propulsion configuration, vessel loading state and outputs that must be repeated. After the action, list every difference and explain whether the records remain comparable. A percentage change needs the same quantity, amplitude convention, processing and sufficiently matched condition.

Make the result reviewable by the actual decision owner.

• The question, vessel configuration, calculation or governing basis, recipients and exclusions match the agreed scope.
• Every result has a traceable quantity, shaft or point location, instrument, setup, units, amplitude convention, run and operating-state record.
• Torsional quantities and derived stress or torque retain their assumptions, processing and reference geometry.
• Every plot and table identifies engine, shaft, propeller or gear order correctly and uses a documented RPM reference.
• Torsional, casing, structural and acoustic results remain distinct while their time relationship is preserved.
• Invalid runs, unavailable states, uncertainty, conflicting evidence and untested alternatives remain visible.
• Measured facts, evaluation, diagnosis, recommendation and acceptance decision are clearly separated.
• A matched retest repeats the material channels and conditions or explicitly qualifies the differences.

The report may support an owner, yard, designer, maker, manager, class or flag review. Those parties control their own review and acceptance. Supplier attendance, a calibrated instrument or a standards citation does not establish approval by itself.

Confirm applicability for the actual shaftline and decision.

• ISO 20283-4:2012 addresses measurement and evaluation of vibration of ship propulsion machinery, and ISO lists it as published and confirmed in 2023.
• ISO 20283-4:2012/Amd 1:2014 is the published amendment recorded by ISO.
• ISO 13373-1:2002 gives general procedures for machinery vibration measurement and data collection.
• ISO 20816-1:2016 provides general machinery-vibration guidance but explicitly excludes torsional vibration from its scope.
• ISO 3046-5:2001 addresses torsional vibrations in reciprocating internal-combustion-engine driven sets.