“The crew will respond to the alarm” can sound reassuring in a risk-assessment workshop. On board, that sentence leaves most of the question open. Which alarm, reaching whom, in what condition? What action does the claim depend on, and how long will the developing hazard allow?

Human response can be a legitimate part of a safety barrier. Crediting it means giving a reasoned account of how a specific action interrupts a specific accident sequence, under the conditions in which it would actually be needed. A procedure, a training record and a successful drill each contribute to that account. None of them completes it.

Define the response before assessing it

A safety function needs a success criterion before it can be assessed. “Take appropriate action” has none. A usable description names the initiating condition, the person who responds, the equipment to be operated and the end state that counts as success: a valve closed and confirmed shut, a transfer stopped, a space cleared.

The human contribution then divides into five stages, each able to fail on its own — detect, diagnose, decide, act and confirm.

Detection fails when an alarm is one of many, or annunciates in a space nobody occupies at that hour. Diagnosis fails when the indication is ambiguous or the supporting information is on a screen in another compartment. Decision fails when the correct action conflicts with another instruction, or when authority to take it is unclear. Action fails when a correct command reaches equipment that cannot respond. Confirmation fails most quietly: acknowledging an alarm establishes neither isolation nor control of the hazard, and an arrangement that treats acknowledgement as closure is recording the wrong event.

Crediting a response without identifying which of these stages it depends on leaves the claim untested. Human-factors guidance from other safety-critical industries is useful here — the UK Health and Safety Executive's material on human factors in risk assessment and on alarm management treats the operator and the supporting equipment as parts of one safety-related system. It is analytical guidance rather than a maritime requirement, but the discipline transfers.

Work backwards from the point of no return

The available time is a property of the scenario, not of the crew. It is established by working backwards: find the latest moment at which the specified action still prevents or limits the specified consequence, then subtract everything that happens before the operator can act at all.

That subtraction is where the interval is usually lost. The release has to develop to a concentration the detector registers. The detector has a response time and the logic a configured delay. The indication has to reach a position a person occupies. Only then does the human sequence begin. Starting the clock when someone notices the alarm omits all of it, and produces an interval longer than the hazard allows.

The required time is the mirror image: recognition, diagnosis, communication, movement, donning protective equipment, operating the control, and the equipment's own travel and closure time. Where verification or a recovery action forms part of the claimed function, that time belongs in the budget as well.

The comparison that matters is between the required time under credible demanding conditions and the available time in the scenario that drives the risk. A stopwatch result from a planned drill on a clear afternoon establishes little about the same task at night, in weather, with one watchkeeper and an alarm heard rarely.

Diagram of the time available for a human response compared with the time the developing hazard allowsFigure 1. The human response chain and its time budget. The timeline is conceptual and not to scale.Source: ICORA MED

The responder may be inside the hazard

Onshore process safety often separates the person from the process: the operator is in a control room and the hazard develops in a plant some distance away. At sea that separation is frequently absent. The space containing the release, the route to the isolating valve and the position of the person credited with closing it can lie inside the same affected area.

That changes the assessment in two ways. The time available to the responder may be shorter than the time available to the ship, because exposure limits are reached before the consequence to the vessel is. For a toxic fuel the margin can be very short. And the response may compete directly with the protection of the same person: a barrier requiring someone to move toward a developing release, to a control operable only locally, depends on a decision taken against the responder's own safety.

Where the assessment finds that pattern, the useful conclusion is rarely more training. It is a second means of achieving the same function from a protected position — remote isolation, shutdown initiated from a location the release does not reach, or an arrangement that removes the need for the local action.

Test the response in its operating context

Consider a methanol bunkering assessment that credits a watchkeeper with initiating emergency shutdown after a leak alarm. The claim is reasonable on its face. Establishing it means showing that the alarm reaches the position the watchkeeper occupies during transfer; that the shutdown control is usable from a location the release does not affect; that transfer stops within the time the scenario allows, including pump run-down and the contents of the line beyond the valve; and that competing duties at that moment do not prevent the action. It also means defining how successful shutdown is confirmed, and what follows if confirmation does not arrive. The IMO interim guidelines for ships using methyl/ethyl alcohol as fuel address bunkering arrangements and safety actions, alongside the equivalent guidelines for other fuels.

If the arrangement instead requires a crew member to reach a local valve through the area affected by the release, access and exposure can invalidate the assumption before competence becomes relevant. Training does not make an inaccessible valve accessible.

Walk-throughs, simulator scenarios and exercises surface what a desk assessment does not: competing demands, communication that takes longer than assumed, controls that are obvious on a drawing and obscure in the space. Use realistic manning, communications and protective equipment where that can be done safely, and examine degraded conditions rather than rehearsing the expected sequence. Record what was tested and what remained simulated. A successful exercise supports a judgement about feasibility; it does not, by itself, establish a probability of success.

Match the credit to the evidence

Naming a human response as a safeguard in a qualitative study is a different act from assigning it a numerical risk-reduction factor. The second calls for a human-reliability assessment suited to the task, documented assumptions, data relevant to the activity and an explicit treatment of uncertainty. IMO's guidelines for formal safety assessment carry an appendix on human reliability analysis for that purpose, and ask for the reliability, uncertainty and validity of data to be reported alongside the result. No general human-error probability attaches to a trained seafarer or a written procedure.

Independence deserves the same scrutiny. Where an alarm response and an automatic shutdown share a sensor, a power supply or a final valve, their contributions cannot be combined as separate layers. Two actions by the same person may share one missed cue. Dependencies are assessed against the initiating event and the other credited controls, including common causes able to disable several at once.

In practice a human response is creditable when five things hold together: the action is defined; the time budget closes under demanding rather than favourable conditions; the responder can reach and operate the equipment without entering the hazard; the outcome can be confirmed; and the conditions that make those true are carried into the safety management system and revisited when manning, equipment, procedures or operating experience change. The ISM Code provides the framework for that last step, and the IGF Code requires the risk assessment that identifies the claim in the first place.

Where one of the five does not hold, the honest course is to credit less — and to record which element is missing, so that the gap is visible to the next person who reads the assessment.

Sources and further reading

  1. International Maritime Organization. Revised Guidelines for Formal Safety Assessment (FSA) for Use in the IMO Rule-Making Process, MSC-MEPC.2/Circ.12/Rev.2. Includes an appendix on human reliability analysis, and the requirement to report the reliability, uncertainty and validity of the data used.
  2. International Maritime Organization. International Safety Management (ISM) Code. Assessment of identified risks and establishment of appropriate safeguards within the shipboard safety management system.
  3. International Maritime Organization. International Code of Safety for Ships Using Gases or Other Low-flashpoint Fuels (IGF Code), and the interim guidelines for individual fuels. Risk assessment, safety actions and bunkering arrangements for low-flashpoint fuel installations.
  4. Health and Safety Executive. Human factors in risk assessment. Task context, involvement of the people who do the work, and verification that a control functions as assumed. Cross-industry guidance, not a maritime requirement.
  5. Health and Safety Executive. Alarm management. Relevant alarms, defined responses and sufficient time to complete them. Cross-industry guidance, not a maritime requirement.

This article is provided for professional information only. It does not constitute project-specific technical, engineering or regulatory advice.