What a QRA is used for
A QRA estimates the risk of defined scenarios numerically. It can combine the frequency of a loss of containment, the probability of different outcomes and their consequences for people or facilities. It supports complex decisions, such as comparing design alternatives or studying the location of facilities containing hazardous substances. It does not mean assigning arbitrary numbers to a qualitative matrix.
Related concept: Risk assessment.
Input data
The study defines the facilities, activities and population considered. It requires information about substances, quantities, process conditions, equipment and protective measures. Depending on the scenario, weather, building occupancy and ignition-source locations may also be relevant.
Frequencies may come from databases or failure models. The study must justify whether those references actually represent the equipment, maintenance and operating conditions being examined. Using a published figure without considering its context creates an appearance of precision.
From scenarios to results
Initiating events and possible developments are defined, including detection, isolation, ignition and dispersion. Fault trees help examine causes; event trees describe outcomes. Consequence models estimate physical effects, always within their range of validity.
Results may be expressed as scenario frequencies, contributions to risk or spatial representations. Each indicator answers a different question. An estimated effect distance must not be confused with the annual probability of experiencing that effect.
Uncertainty and review
The study needs to show the assumptions that drive the result and explore how the conclusion changes when uncertain inputs vary. If the result depends on a barrier always working, that assumption deserves particular scrutiny. Independence between failures cannot be assumed where they share power, maintenance or environmental conditions.
An independent review may identify omitted scenarios, double-counted barriers and unit errors. Model traceability matters as much as the final chart.
Hypothetical example
Two locations for a storage tank are compared. One shortens the pipework but places the installation closer to an occupied building. The study examines leaks, cloud development and people’s exposure for both options. Alongside central results, it presents sensitivity analyses for occupancy and leak frequency. The decision combines that evidence with safe design and applicable requirements.
What it cannot demonstrate
A QRA does not prove that an accident is impossible or remove preventive obligations. Acceptance criteria depend on the regulatory framework and the purpose of the study. They must be explained before the numbers are interpreted. For straightforward tasks, a less complex technique may provide a clearer and more proportionate preventive decision.
