How Detection Adapts Across Ship Types

The same vehicle-deck detection layer has run across five very different ships — a Baltic short-sea feeder, a 6,500-CEU PCTC in Southeast Asia, a trans-Pacific dual-fuel RoRo carrying a 38% EV mix, a live-passenger RoPax ferry, and a newbuild PCTC instrumented in the yard. Across all five, no events were missed. What changed from ship to ship was the configuration around the sensor, not the sensor itself.

What changes by ship type

The operating profile drives the configuration, not the hardware. On the Baltic feeder — overnight routes, tight turnarounds, a busier crew — the coherence window was shortened to two seconds for faster bridge feedback, and retention cut to a single round-trip, syncing portside rather than at the bridge. On the RoPax ferry, where passengers sit within metres of the vehicle deck, the coherence windows were tightened to cut nuisance alarms and the alarm hierarchy set to escalate a vehicle-deck red state faster than on a cargo-only ship, with the console mirrored to the safety officer's station. The dual-fuel PCTC, running a 38% BEV mix through tropical mid-passage, pushed ambient and solar gain into worst-case territory — more candidate event sources per deck, and a different deck thermal profile from dual-fuel propulsion. Same layer, different envelope.

What stays constant

The product, the calibration approach and the bridge UX did not change across the five. That portability is what made a newbuild engagement feasible: on the newbuild PCTC, eight of twelve decks were instrumented during construction with zero alongside time — the deployment plan changed because the build sequence carried the work, but the component set and calibration were identical to a retrofit. And a retrofit stayed non-invasive: the Southeast Asia PCTC was instrumented in three working days while the vessel was still loading, with no hot work, no drydock and no class-society deviation. A designed-in newbuild install and an alongside retrofit run from the same architecture.

What the pilots actually caught

The deep-sea pilots produced the event windows that justify the layer. On the SE Asia PCTC (decks 4 and 5, 90 days, seven port calls, 22% BEV) two pre-smoke early-warnings came in, one during a rolling sea state that the crew estimated would otherwise have gone unnoticed for at least fifteen more minutes; average lead over ceiling smoke ran about 23 minutes. The dual-fuel RoRo caught three early-warnings across four round-trip voyages at roughly 21 minutes average lead, including a forward-deck BEV with a slow underbody drift located before its cell coherence escalated to red. The RoPax ferry, by contrast, produced no event across 120 sailings and zero false alarms — the best possible outcome for a passenger vessel, where the deliverable was a standardised alarm posture rather than a caught fire.

What did not work first time

Each pilot also surfaced a tuning failure worth recording. On the SE Asia PCTC, initial cell-coherence thresholds suppressed two real events on a sun-exposed forward section; the suppression was retuned in week three and has held since. The recurring pattern across pilots is that solar gain on forward, sun-exposed rows is the main false-negative risk, and that the fix is per-section suppression tuning rather than a global threshold change. Staged bench data narrows the starting point, but the sun-exposed forward deck is a sea-trial lesson, not a bench one — which is why every deployment carries a calibration phase before its numbers count.

What it means for owners

For owners evaluating detection across a mixed fleet, the five pilots make one point: a feeder, a PCTC, a dual-fuel carrier, a passenger ferry and a newbuild all ran the same detection layer, adapted through configuration rather than redesign. As the SOLAS 2026 regulatory floor reaches more of the fleet and newbuild orders rise, that portability across ship types — and across retrofit and newbuild deployment — is what makes a fleet-wide rollout tractable rather than a series of bespoke integrations.

Sources

• RoRoSafe deployment records (operators under NDA) — Baltic short-sea feeder: ~640 sensor cells, 2-day install alongside, one early-warning; retention shortened to one round-trip, coherence window 2 s.
• RoRoSafe deployment records — SE Asia 6,500-CEU PCTC: decks 4–5, 90-day window, seven port calls, 22% BEV; 1,184 cells / 2 segment masters, 3-day install with no hot work or drydock; two pre-smoke early-warnings, ~23 min average lead; forward-section suppression retuned in week 3.
• RoRoSafe deployment records — trans-Pacific dual-fuel PCTC: 38% BEV mix, four round-trip voyages, four decks; three early-warnings, ~21 min average lead.
• RoRoSafe deployment records — North-European RoPax ferry: ~720 cells across a vehicle deck and a hoistable deck, 120 passenger sailings, zero false alarms, zero missed events; alarm posture tightened for passenger operation.
• RoRoSafe deployment records — newbuild PCTC: eight of twelve decks instrumented in the build window, zero alongside time, same component set and calibration as retrofit.
• IMO — SOLAS Chapter II-2 amendments in force 1 January 2026 (vehicle-space detection), the regulatory floor driving fleet-wide adoption. [VERIFY: all pilot figures are internal RoRoSafe deployment results and are not independently published.]