Alarm handling at sea – the human factor matters
An ongoing project on alarm management, conducted under the ShippingLab platform, investigates how watchkeeping engineering personnel respond to a range of abnormal situations – typically first signalled by alarms and indicators – both in simulators and on board some of the world’s largest vessels.
By applying analytical methods from the nuclear industry, the project’s preliminary results reveal a critical insight: the human factor plays a much greater role in successful alarm response than current maritime alarm system designs and classification rules typically account for.
As Asger C. Schliemann-Haug from Lloyd’s Register puts it:
“Traditionally, we have placed a great deal of faith in the value of alarms and indicators to save the day. Our data show that this faith is not unwarranted – that people are remarkably adept at solving problems they’ve never encountered before or for which they were never trained to handle. Nevertheless, we find that this only holds when the time it takes for humans to determine what to do – and to act upon it – is adequate.”
One of the interesting findings is that available time – the window in which operators can interpret and respond – appears to dominate the likelihood of success. This factor overshadows more traditional performance-shaping elements. Since available time is inherently a design feature, it should be considered upfront to ensure that seafarers are dealt a good hand of cards for the moments where ‘ship happens’.
In short: if you want to claim safety by design, the time it takes for humans to figure out what’s wrong – and respond – should be factored in.
Preliminary findings show:
- On large passenger ships, the noise-to-signal ratio of alarm systems is very high. Field observations show that only 4–10% of alarms trigger any crew action – often due to chattering, consequential, or false alarms, which are left unaddressed or stem from poor system commissioning and process tuning. This ‘noise’ delays response to genuine alarms, as operators are forced to seek confirmatory information before acting. Tackling alarm fatigue is the first step toward enabling prompt and accurate responses.
- Operators often take longer to diagnose issues than the system design anticipates – especially when a disturbance or trip initially throws them off balance.
- Compared to long-established models from the nuclear sector, maritime operators fall somewhere between nominal and pessimistic expectations of human error probability relative to available diagnosis time. This shows that, while there is room for improvement, we need not take an overly negative view.
The project underlines that alarm response is not merely an operational issue – it is fundamentally a matter of design.
In the nuclear sector, each alarm is backed by clear, field-validated response procedures. These include tested margins for human reaction time and recovery factors that account for slips and omissions. The project suggests that the maritime sector could benefit greatly from adopting similar principles.
Findings from this and related initiatives provide valuable input for both the practical design of maritime alarm systems and the future development of international safety regulations.
The project is conducted on the ShippingLab platform and supported by the Danish Maritime Fund. You can read more about the project here.
Project partners:
DFDS, Lloyd’s Register, Aarhus School of Marine and Technical Engineering, Fredericia School of Marine and Technical Engineering, Blue Stamp, and the IT University of Copenhagen.