Should Detection Fire the Drencher?

This pilot asked the question that sits one step past detection: if the system can tell you which vehicle is developing, should it be allowed to open the drencher on that zone? The trial connected detection evidence to the release decision on two enclosed vehicle decks of an in-service PCTC — and deliberately stopped short of letting it act. The recommendation went to a human, every time. [VERIFY: this pilot is a representative composite pending a confirmed operator dataset — founder to verify or replace figures before publish.]

What the trial actually tested

Not the suppression engineering, and not the detection hardware — the decision between them. The vessel already had zoned drencher coverage and a per-vehicle detection layer from an earlier retrofit, so neither was the variable. What had never been tested was the handover: whether detection evidence, presented as a zone-level release recommendation, actually helped the officer of the watch make the call faster and with more confidence than the fitted panel alone. The trial ran over a working liner season, logging every alarm, every recommendation, and every human decision on a common time base. The mechanics of how a recommendation is derived are outside the scope of this note; what was measured was whether the crew's decision improved.

• In-service PCTC on a North Europe / Mediterranean rotation, two enclosed vehicle decks instrumented.
• Existing zoned drencher coverage and a previously installed per-vehicle detection layer — neither modified for the trial.
• Output was a zone-level release recommendation surfaced to the bridge, never an actuation.
• Class was engaged from the outset, on the basis that no automatic release path was being sought.
• Every alarm, recommendation and human decision logged on a common time base for review.

Why it stopped at advisory

Because releasing a drencher zone is not a reversible act, and the evidence bar for committing it is higher than the bar for investigating. A system good enough to say 'look here now' is not automatically good enough to say 'flood this deck now'. Those are different claims with different costs attached, and conflating them is the easy mistake in this area. There is also a governance answer: any automatic release path would need class and flag approval on its own merits, and the trial did not seek that — it set out to measure whether the human decision got better, not to remove the human. On a crewed ship the officer of the watch remains the cheapest and most flexible safeguard available, and the honest finding of the pilot is that improving that person's information was where the value sat.

The cost of being wrong, in both directions

A drencher decision is expensive whichever way it fails, which is precisely why the evidence has to be good. Release too early or on a false signal and you are applying water across a loaded deck at the FSS Code rate — soaking cargo that is someone's insured consignment, adding weight high in the ship, and introducing free surface. That last consequence is not theoretical: the Höegh Osaka casualty turned on how badly a vessel's real weight distribution can diverge from the assumed one, and a flooded deck moves the same variables in the same unhelpful direction. Release too late and you are past the point where a fixed system extinguishes anything — the Auto Banner burned 67 hours alongside a pier with hundreds of firefighters available, because once a dense vehicle stow is alight, suppression becomes containment. The decision window between those two failures is narrow, and narrowing it further is an argument for better evidence, not for taking the human out.

What operators, class and underwriters should take from it

• Separate the two claims. 'Investigate this vehicle' and 'flood this zone' need different confidence levels, and a detection system should be assessed against each separately rather than as one capability.
• Treat advisory operation as a required phase, not a limitation. A season of recommendations that a crew agreed or disagreed with is the evidence base any future automatic path would have to be argued from.
• Engage class early and say plainly what you are not asking for. The approval conversation is materially simpler when no automatic actuation is in scope.
• For underwriters: a logged record of recommendation, decision and outcome is claims-grade evidence about how a casualty was handled — the same evidentiary point that matters after any car-deck fire.
• The reduced-crew case changes this calculus, and not yet. Under the IMO MASS Code the trigger decision starts to become a sensor decision by necessity; on a crewed vessel today it does not have to be, and there is no reason to spend that safeguard early.

Sources

• RoRoSAFE detection-to-release pilot record (operator under NDA) — in-service PCTC, North Europe / Mediterranean rotation, two enclosed vehicle decks with existing zoned drencher coverage and a previously installed per-vehicle detection layer; advisory-only zone-level release recommendations with human authorisation retained throughout. [VERIFY: figures and operating description are a representative composite pending a confirmed operator dataset — founder to verify or replace before publish. No automatic actuation was trialled and none should be implied.]
• IMO — FSS Code Chapter 7: drencher system discharge rate for vehicle carrier decks. [VERIFY: the 5.0 L/min/m² figure is carried from RoRoSAFE's 'PCTC Drencher Zoning and Fixed-First'; confirm against the FSS Code text before publish, and note the applicable rate varies by space category.]
• IUMI — position and best-practice material advocating a 'Fixed First' firefighting approach for vehicle carriers, and the limits of manual intervention on vehicle decks — iumi.com.
• IMO — MASS Code, adopted at MSC 111 and in effect from 1 July 2026 (non-mandatory), covering manning and remote operations alongside fire protection, detection and extinction — the framework under which the release decision becomes a sensor decision on reduced-crew ships — imo.org.
• RoRoSAFE case studies — 'Höegh Osaka: The Numbers That Weren't' (how far actual weight distribution can diverge from assumed, the risk a flooded deck compounds) and 'Auto Banner: 67 Hours Alongside a Pier' (what a fixed system faces once a dense stow is alight).
• Companion RoRoSAFE analysis — 'PCTC Drencher Zoning and Fixed-First' (the suppression engineering this pilot sits on top of), 'IUMI 2025: The Fixed-First Approach' (the doctrine), and 'Can an Autonomous Ship Fight a Deck Fire?' (why reduced crewing eventually forces this question).