Technical decision guide · Updated 2026-08-22
Torsional, Machinery and Structural Vibration: Understanding the Difference
A shaft twisting cyclically, a bearing housing moving, a deck resonating and a cabin rattling are different technical problems even when a crew member describes all of them as vibration. The fastest route to a defensible answer is to define the decision, reproduce the operating condition and measure the correct response with a common speed and time reference.
1. Choose the measurement from the engineering question
| Observed need | Primary evidence | Strongest canonical route | What it does not prove |
|---|---|---|---|
| Validate a propulsion torsional calculation, investigate a damper or coupling, or verify a restricted-speed concern. | Angular displacement/velocity, dynamic torque or strain at a defined shaft-train location, synchronised with RPM, load and order. | Marine torsional vibration measurement and analysis | Casing acceleration alone does not establish shaft alternating stress. |
| Assess an engine, generator, pump, fan, gearbox or bearing for abnormal vibration. | Repeatable casing/bearing measurements, waveform, spectrum, phase or order evidence under controlled load. | Marine machinery vibration analysis | A high overall value does not identify the defect without frequency and operating context. |
| Map a deck, bulkhead, foundation, superstructure or hull response. | Multi-point structural response, direction, phase/order reference and operating-deflection or modal evidence where justified. | Structural and hull vibration survey | Structural response does not automatically identify the forcing source. |
| The vessel is noisy, shaking, rattling or failing repeatedly and the mechanism is uncertain. | Synchronous acoustic, vibration and operating evidence using source-path-receiver tests. | Vessel noise and vibration troubleshooting | No ethical investigation can guarantee a root cause before the symptom is reproduced. |
2. Torsional vibration: validate the rotating train, not the casing
Torsional vibration is cyclic variation of angular motion superimposed on steady shaft rotation. The decision basis normally starts with the approved or project-controlled torsional-vibration calculation, engine and shaftline data, coupling/damper properties, applicable class rules and ISO 20283-4:2012 with its applicable amendment. ISO lists the standard as current after confirmation in 2023.
Trace calculation to measurement
- Identify the exact shaft, crankshaft, coupling or train location represented by the calculation.
- Confirm whether the comparison quantity is angle, angular velocity, torque, strain or derived stress; these are not interchangeable without a valid model.
- Record engine firing arrangement, gear ratio, inertias, shaft/coupling stiffness, propeller or driven load, damper condition and relevant retrofit changes.
- Use a reliable tachometer or angular reference so orders and critical speed bands are traceable.
- Agree steady holds, controlled run-up/run-down sweeps and the relevant propeller/load curve before the trial.
Changing a barred-speed range, damper setting or acceptance conclusion is a design and review decision. A measurement report supplies evidence; it does not independently authorise the change.
3. Turn a complaint into a controlled diagnostic plan
Before selecting sensors, write the symptom as a testable statement: where it occurs, what it sounds or feels like, which RPM/load/configuration triggers it, when it started, what was repaired or changed, and whether the response disappears when one operating variable changes.
| Symptom pattern | Controlled comparison | Useful evidence |
|---|---|---|
| Cabin or deck shakes only in a narrow RPM band. | Stepped speed and safe run-up/run-down with common tachometer reference. | Orders, spectra, phase/response map and adjacent-space comparison to test resonance and transmission hypotheses. |
| Noise appears when one generator, pump, fan or compressor runs. | One-at-a-time machine/load comparison while other conditions remain stable. | Source casing vibration, foundation/structure response, airborne sound and transfer-path correlation. |
| Rattle is local but the measured vibration source may be remote. | Local restraint or temporary isolation test where safe, plus synchronous reference channels. | Receiver response before/after the controlled change; avoid treating the rattle as the original source. |
| Problem follows propeller load, heading, draught or sea condition. | Repeat matched propulsion states and document water/weather/loading changes. | Shaft speed/order, hull response, acoustic data and exact navigation/operating log. |
| Earlier reinforcement, balancing or mount replacement did not solve the complaint. | Reproduce the original condition and compare pre/post evidence on the same basis. | Confirmed change, unchanged response and remaining hypotheses ranked by evidence. |
4. Separate source, path and receiver
A defensible investigation does not jump from a loud or high-vibration location directly to a repair. It tests three layers:
- Source: propulsion orders, engine firing, unbalance, misalignment, gear mesh, pumps, fans, exhaust pulsation, propeller pressure pulses, cavitation, flow or intermittent impacts.
- Path: shaftline, mounts, foundations, piping, ducts, hull structure, bulkheads, resilient connections and airborne openings.
- Receiver: bearing housing, deck, cabin panel, accommodation position, bridge, control room or sensitive equipment.
DNV notes that ship noise and vibration can arise from machinery and propellers as well as cranes, rudders, ventilation and flow-induced phenomena. That breadth is why multidisciplinary, synchronous evidence is more useful than an isolated overall reading.
5. Build a synchronised measurement chain
- List every sensor, channel, serial number, calibration status, range, orientation and mounting method.
- Synchronise tachometer/order, torsional, casing, structural and acoustic channels when cross-domain causation is being tested.
- Preserve time histories and sufficient spectral/order resolution; overall values alone can hide narrow resonances, modulation and transients.
- Record RPM, torque/load, pitch, generator combination, thruster/HVAC/cargo equipment state, draught, trim, speed, heading, sea state, water depth and manoeuvres as relevant.
- Document clipping, sensor movement, telemetry loss, background interference, unavailable operating states and every deviation from the agreed plan.
6. Match the operating context to the problem
| Context | Best use | Limitation to record |
|---|---|---|
| Berth or anchorage | Auxiliary machinery, HVAC, local rattles and controlled equipment combinations. | Cannot reproduce propulsion loading required for many shaftline or sea-trial questions. |
| Harbour trial/commissioning | System checks, generator combinations and defined quay-side operating tests. | Mooring, water depth and restricted operating range may change the response. |
| Sea trial or pre-delivery | Propulsion load curve, speed sweeps, acceptance points and synchronised vessel-wide evidence. | Weather, draught, heading and manoeuvres must be controlled or disclosed. |
| In-service investigation | Reproduce the actual complaint under normal cargo, hotel-load or operational configuration. | Access and safe isolation options may be limited. |
| Dry dock or repair period | Inspection, sensor preparation and before/after corrective-work planning. | Dynamic symptoms may not be reproducible without the operational load. |
7. Produce evidence that supports the next decision
A useful report should contain the vessel and system identity, purpose, drawings and measurement locations; instrument/calibration chain; controlled operating matrix; calculation or comparison basis; time, spectrum and order results; source-path-receiver findings; confirmed observations versus hypotheses; uncertainty and limitations; corrective options ranked by evidence; and exact retest conditions.
For torsional work, show how the measured location, quantity and operating point correspond to the calculation or review basis. For troubleshooting, state the confidence of each conclusion and which additional test would resolve any remaining ambiguity.
8. Class, flag and acceptance boundaries
ABS, BV, CCS, CRS, DNV, IRS, KR, LR, ClassNK, PRS, RINA and Türk Loydu may apply society-specific calculation, measurement, witnessing or supplier requirements. Flag administrations including Singapore, Panama, Liberia, Marshall Islands, Bahamas, Malta, Cyprus, Hong Kong and Isle of Man may have project-dependent review interests. Confirm the controlling class rule, flag requirement, contract, approved calculation, maker limit and attending surveyor expectation before testing.
Project-specific report acceptance is not the same as blanket supplier approval. IACS does not approve service companies, and Vessel Doctor does not claim class or flag approval, recognition, appointment or guaranteed acceptance on this page.
Authoritative references
- ISO 20283-4:2012 — measurement and evaluation of vibration of ship propulsion machinery.
- ISO 20283-4:2012/Amd 1:2014 — published amendment to Part 4.
- ISO 20283-2:2008 — structural-vibration measurement, diagnostic evaluation and reporting; it explicitly directs torsional shaft/crankshaft matters toward class rules and ISO 20283-4.
- DNV noise and vibration overview — examples of multiple marine sources and multidisciplinary troubleshooting context.
9. What to send before attendance
Send the vessel/IMO number, vessel type, class and flag where relevant, propulsion and machinery arrangement, engine/shaft RPM and load range, coupling/damper/gear/propeller data, approved torsional calculation if applicable, symptom video or log, affected locations, previous spectra and reports, recent damage/repair/retrofit, operating state that reproduces the problem, port or shipyard, trial window, witness requirement and reporting 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.