Fire · emergency power · fixed CO₂
A fuel hose that should not have been there, and three safety systems that did not answer
A departure fire from a hose fitted in an unapproved modification, an emergency generator breaker that faulted, and a CO₂ system that did not fully operate.
Source of record
UK Marine Accident Investigation Branch (MAIB) — Report 13/2025 — Finnmaster
Occurrence 19 September 2021 · report published 18 September 2025
Read the official reportWhat the investigator found
At 2013 on 19 September 2021 a fire broke out in the auxiliary engine room of the Finnish registered ro-ro cargo ship Finnmaster during departure from Hull, England. The vessel lost power. The fire was later extinguished and the vessel safely re-berthed with tug assistance. The auxiliary engine room was significantly damaged; there were no injuries. Twelve recommendations were made.
- The fire started after mechanical failures on one of the auxiliary engines allowed fuel to leak from a flexible hose onto a hot surface, where it ignited.
- The flexible hose was installed in the fuel system during an unapproved modification and did not meet the required standard.
- A fault in the emergency generator circuit breaker prevented it from supplying power to the emergency systems on board.
- The fixed carbon dioxide fire-extinguishing system failed to fully operate due to defects in the system.
- The crew's response, affected by the loss of critical safety systems, was ineffective.
The teardown
Everything below this line is our reading of the published record — a Chief Engineer's interpretation, not a finding of the investigating authority.
Read the four issues as one sentence
An unapproved part started it, a faulty breaker removed the power to fight it, a defective fixed system removed the means to smother it, and the crew — who did nothing wrong — were left responding to a fire with the tools written into the plan already gone. The report is careful to call the response ineffective and then to explain why. That order matters, and it is usually reversed in the retelling.
An unapproved modification is not a bad part, it is an invisible one
The hose was in the fuel system and did not meet the required standard. Nothing on board would have shown that, because the modification never entered the approval trail — and a part that is not in the trail is not in anybody's inspection scope either. This is the most common shape of engine-room risk: not a defect somebody ignored, but a change nobody recorded, which therefore never became inspectable.
The tested-but-not-proven systems
Emergency generators are tested routinely. Fixed fire systems are surveyed. Both were, and both failed at the moment of use. The gap is between testing that a system starts and proving it will carry the load, or discharge, under the conditions of an actual casualty. That gap is exactly why the investigator went to the IMO on emergency power testing and to class on fixed-system service suppliers. When the recommendations climb that high, the finding is that the routine itself was insufficient — not that the ship skipped it.
The human factor sits at the end, not the beginning
The crew fought a fire without emergency power and without a fixed system that worked. Any competent team is ineffective in that position. Treat 'crew response' as a root cause here and you learn nothing; treat it as the last domino and you learn where the real failures were — in change control and in system proving, weeks and years earlier.
Which record already held the warning
A modification with no approval record on a fuel or hydraulic system
Lives in
Change/modification control records against the class-approved system drawings
Would have shown
Non-standard components in a fuel line, absent from the drawing set — the single cheapest thing to audit and the most commonly skipped.
Emergency generator proven to start, never proven to carry the emergency load
Lives in
The emergency source test record and its stated test method
Would have shown
A repeated test that exercises starting only, with the breaker's ability to actually supply the board untested.
Fixed fire-extinguishing system defects carried across service visits
Lives in
Service supplier reports and the outstanding defect list
Would have shown
Items noted and rolled forward rather than closed, which is what 'failed to fully operate due to defects' looks like beforehand.
Hot surfaces and fuel lines in the same space without a verified screening record
Lives in
Engine-room inspection rounds
Would have shown
Insulation and shielding gaps logged as observations that never became work — the standard precondition for a fuel-on-hot-surface fire.
What a maintenance system should be doing
- Give modifications their own record with an approval status, and surface unapproved changes on critical systems as open compliance items, not as engine-room folklore.
- Separate 'started' from 'carried the load' in emergency source tests, and treat a test that cannot demonstrate the second as incomplete.
- Carry service-supplier defects forward as visible open items on the equipment, so a rolled-forward defect keeps costing attention until it is closed.
- Put fire-risk rounds in the same evidence chain as maintenance, so an observation about lagging becomes a job with a due date.
Method
- — Only published official investigations. Report number and link on every case.
- — The investigator's findings are quoted as theirs; our reading is labelled as ours.
- — No vessel the author has attended, ever — no exceptions, no de-identified exceptions.
- — Every case ends the same way: which record already held the signal, and what would have shown.