Hazards, exposure, risk and consequence analysis

Connect source, exposure and consequence

A hazard is a potential source of harm. Exposure describes how a person, asset or environment can encounter it. A scenario connects conditions and events to possible consequences. For an engineering system, include ordinary use, cleaning, maintenance, foreseeable changes and other affected people, rather than considering only the intended operator during normal production.

Risk combines likelihood and consequences under a stated scenario and time basis. A probability per mission differs from a frequency per year. An injury category, lost service duration and an environmental release are different consequences. Retain that meaning whenever the result is summarized; a bare number called “risk” cannot identify whose outcome it represents.

Keep important consequences visible

Expected loss multiplies each numerical consequence by its probability and adds the results. It is useful when the quantity and decision purpose are appropriate. The figure shows why the same expectation need not mean equivalent outcomes: a rare prolonged outage differs from a more frequent short one. An average does not describe the tail of the distribution.

Qualitative rankings can help organize discussion when data are weak, but category numbers are not automatically ratio measurements. A severity label 4 is not necessarily twice as much harm as label 2. Multiplying ordinal category labels cannot create a calibrated probability, and equal products do not establish equally acceptable scenarios.

Ask what would change the assessment

Separate uncertainty about event likelihood from uncertainty about consequences and the completeness of the scenario set. A precise-looking input may come from few events or from a different population. Unknown dependencies, exposure patterns and future changes can dominate the result. Sensitivity analysis asks how conclusions change when defensible input ranges change.

Record the existing controls, evidence about their effectiveness and the conditions in which they act. Then compare ways to remove the source, change exposure or limit consequences. Risk evaluation also involves affected people, responsibilities and applicable requirements. An expected-loss calculation alone cannot authorize exposing people to a severe hazard.

Compare two supplied service-loss distributions

Consider one hypothetical mission. Scenario A has a 20% chance of a 10 h outage and otherwise no loss. Scenario B has a 2% chance of a 100 h outage and otherwise no loss. Expected outage duration is 0.20×10=2 h for A and 0.02×100=2 h for B.

If a service question asks about outages longer than 24 h, their answers differ: probability zero in A and 0.02 in B. At a 5 h threshold they instead give 0.20 and 0.02. Neither distribution is universally preferable without the service consequences, uncertainty and decision criteria. These invented outage probabilities are not probabilities of injury.

Try the reasoning

1. Can a hazard exist when immediate exposure is absent?

   Show the answer and reasoning

   Yes. A source of potential harm can remain present, while separation or other conditions prevent the current exposure path.

2. Why do the two 2 h expectations not establish equal risk for every purpose?

   Show the answer and reasoning

   The durations and probabilities differ. Consequences such as missing a critical service window depend on that distribution, not just its mean.

3. Can a qualitative risk score be read as a measured annual probability?

   Show the answer and reasoning

   No. Category rankings need their own definitions and evidence; their numerical labels do not supply a calibrated time-based probability.

What to carry forward

State the scenario, affected party, time basis and consequence distribution before interpreting any risk summary.