The four building blocks
The planned maintenance system is the single most consequential dataset on a ship. It defines what gets maintained, when, by whom, and with what evidence — and it is the artefact a class surveyor, a PSC officer and a vetting inspector will all ask to see.
Done well, it is a working tool. Done badly, it is a compliance theatre generator: thousands of jobs auto-closing on schedule, none of them describing what actually happened in the engine room.
1. The equipment hierarchy
Everything hangs off this. A workable structure is four levels:
Vessel → System → Equipment → Component e.g. MV Example → Propulsion → Main Engine → Cylinder Unit No. 3
Rules that save years of pain:
- Code once, centrally, and propagate to sister vessels. Never let each ship invent its own.
- Every component that can be maintained, or that can consume a spare, needs an identity.
- The same hierarchy feeds procurement and inventory. Two catalogues means wrong parts.
- Retain retired equipment with a status flag; do not delete. History is the asset.
2. Job library
| Attribute | Why it matters |
|---|---|
| Job description and procedure | The actual work; should reference maker's manual section |
| Interval basis | Calendar, running hours, condition, or event |
| Criticality | Drives compliance measurement and class scrutiny |
| Required spares | Links to inventory; drives min/max |
| Required competency | Which rank may sign off |
| Evidence required | Readings, photographs, measurements |
| Class relevance | Whether the job supports a survey item |
Typical libraries run 2,000–6,000 jobs per vessel. That number is less important than the criticality split: if more than about 15% of jobs are flagged critical, the flag has stopped meaning anything.
3. Running hours and condition inputs
Calendar-based intervals are simple and often wrong. A generator running 6,000 hours a year and one running 1,200 should not share an overhaul interval. Running hours must come from the machine, be entered on a disciplined cycle, and be visible ashore.
4. Evidence
The difference between a maintained ship and a documented ship is evidence: readings recorded, clearances measured, photographs attached. Evidence capture is also what makes a PMS defensible when a claim arrives.
Class-approved maintenance schemes
Where a vessel operates under a class-approved planned maintenance scheme for machinery, the PMS moves from being an internal tool to being part of the survey regime. That brings obligations:
| Obligation | Practical effect |
|---|---|
| Approved job scope and intervals | Changes need class agreement |
| Chief engineer competence requirements | Named, qualified sign-off |
| Annual audit by class | Records must be complete and retrievable |
| Condition monitoring where credited | Data must be genuine and trended |
| Deviation reporting | Overdue jobs must be declared, not hidden |
The reward is deferred opening-up surveys and reduced downtime. The price is that the records must be true. A class-approved scheme with fabricated completions is a far worse position than no scheme at all.
Where PMS data goes wrong
Migration. Moving 15 years of history and an inaccurate equipment register into a new system is the largest single risk. Budget a physical equipment verification on board — someone with a torch and a clipboard walking the machinery spaces — before migration, not after.
No evidence capture. If the system accepts "completed" with no reading, no measurement and no photograph, it will receive exactly that.
Coding drift. Sister vessels diverge within eighteen months unless master data is governed centrally.
Crew overload. A PMS that takes an engineer forty minutes of typing per job will be completed in bulk on a Sunday. Design for mobile capture at the machine.
No shore review. Overdue and deferred jobs need weekly review by a named superintendent, with deferrals justified in writing.
The overdue-job discipline
| Status | Required action |
|---|---|
| Approaching due (30 days) | Spares confirmed available |
| Due | Scheduled into work plan |
| Overdue, non-critical | Recorded, reason logged |
| Overdue, critical | Escalated to superintendent within 7 days |
| Overdue, critical, class-relevant | Reported to class; deviation recorded |
Fire safety accounted for 16.8% of all Paris MoU deficiencies in 2025 and structure/machinery/electrical for a further 11.6%. A large share of those findings trace back to routine jobs that were carried, and carried, and carried.
Migration checklist
- Physically verify the equipment register on board before extraction.
- Freeze the source system; extract, do not re-key.
- Map criticality deliberately — do not inherit an inflated legacy flag set.
- Migrate history, not just open jobs. Trend data is the value.
- Reconcile spares to the new equipment codes before go-live.
- Run parallel for one full month on one vessel.
- Obtain class agreement before changing any approved scheme intervals.
deficiency shares from the Paris MoU Annual Report 2025. The root-cause chart is an indicative practitioner view for illustration, not a formal survey. Reviewed by the Zeaclub Editorial Team, 24 August 2026.
Frequently asked questions
How many jobs should a PMS contain?
Between roughly 2,000 and 6,000 per vessel depending on segment and automation. Quality of criticality classification matters more than count.
Can maintenance intervals be changed?
Yes, on justification — running hours, condition data, maker guidance. If the vessel is on a class-approved scheme, class must agree first.
Who signs off a PMS job?
A crew member holding the required competency for that job, recorded by name. Blanket sign-off by a single officer is a common audit finding.
Is condition-based maintenance replacing PMS?
It is supplementing it. Condition monitoring changes when jobs are done; the PMS remains the record of what exists, what is required, and what was performed.