When a facility's risk needs to be expressed as a number rather than a colour on a grid, a quantitative risk assessment is the tool that does it. QRA combines detailed consequence modelling with failure-frequency data to produce hard measures of risk, such as the chance of harm at a given location per year. This guide explains how a QRA is built, how it presents results as individual-risk contours and F-N curves, and how it differs from the qualitative methods used in most day-to-day process safety work.
Quantitative Risk Assessment (QRA) in one line: A quantitative risk assessment (QRA) is a method that expresses risk numerically by combining the estimated frequency of hazardous events with modelled estimates of their consequences. It models how releases disperse, ignite, and cause fires or explosions, multiplies those outcomes by how often each release is expected to occur, and sums the results into measures such as individual risk per year at a location and societal risk shown as an F-N curve. It is more detailed and resource-intensive than qualitative methods like a risk matrix or HAZOP.
A QRA is assembled from two halves that are eventually multiplied together: how often things go wrong, and how bad it is when they do. The frequency half begins with a set of representative release cases - a pipe rupture, a flange leak, a vessel failure - each assigned a frequency drawn from failure-rate data for that kind of equipment. The consequence half takes each release and models its physical evolution: how the material disperses into a cloud, whether and how it ignites, and the resulting thermal radiation from a fire, overpressure from an explosion, or toxic dose from a cloud. Consequence modelling is where much of the technical effort goes, because the outcome depends on release size, weather, terrain, ignition sources, and the vulnerability of people at each location.
For every combination of release case and outcome, the analysis estimates both how likely it is and the harm it would cause at points around the facility. Weather conditions and wind directions are weighted by how often they occur, ignition probability is applied, and the vulnerability of people to the resulting hazard is estimated. Summing across all the cases, weather states, and directions produces, for any given location, a total frequency of harm from all the modelled scenarios combined. This is the core calculation of a QRA: an aggregation of many individual event frequencies and their consequences into an overall risk picture.
Because a QRA rests on so many inputs, its credibility depends on the quality of the data and models behind it. Failure frequencies come from industry databases and, where available, plant-specific experience; consequence models are validated dispersion, fire, and explosion tools; and the population and ignition assumptions must reflect the real site. A QRA is only as good as these inputs, which is why it is reserved for situations where the effort is justified - major hazard installations, land-use planning near a facility, or scenarios where qualitative methods cannot resolve whether the risk is acceptable.
The most recognisable output of a QRA is the individual-risk contour map. Individual risk is the frequency, per year, at which a hypothetical person continuously present at a specific location would be harmed by the facility's hazards. Because that frequency varies with distance and direction from the hazards, it can be plotted as contour lines on a site plan - like elevation lines on a map - each joining points of equal risk. Reading these contours shows how risk falls off with distance from the plant and lets planners compare the risk at a location against a tolerability criterion, which is why individual-risk contours are central to land-use decisions around hazardous sites.
Individual risk, however, says nothing about how many people an event could affect at once, and for that a QRA produces a societal-risk measure, most commonly the F-N curve. An F-N curve plots the frequency (F) of events causing at least a given number of fatalities (N) against that number, on logarithmic scales. It answers a different question from the contour map: not how risky is a particular spot, but how the facility's events distribute across small-but-frequent and large-but-rare outcomes. A curve that climbs toward the high-N end signals the potential for a small number of very large incidents, which societal-risk criteria treat with particular concern because of the aversion to catastrophic multi-fatality events.
Together, the contour map and the F-N curve give complementary views. The contours protect the individual by showing where the risk to a person is too high, and the F-N curve protects against catastrophe by revealing the potential for large-scale harm. Both are compared against tolerability criteria - upper limits above which risk is intolerable and lower thresholds below which it is broadly acceptable - so that a QRA does not just describe risk but supports a defensible judgement about whether it is acceptable and, if not, where reduction is needed.
Most process safety analysis is qualitative or semi-quantitative: a HAZOP identifies hazards, a risk matrix ranks them by judgement, and a LOPA uses order-of-magnitude frequencies to check specific scenarios. QRA sits at the more rigorous end of that spectrum. Where a risk matrix compresses risk into a coloured cell, a QRA produces an actual number with units. This precision is powerful but costly - a full QRA takes specialist tools, extensive data, and considerable time - so it is not a replacement for everyday qualitative work but an escalation reserved for the situations that need it, such as major hazard sites, novel designs, or disputes that qualitative methods cannot settle.
The two approaches are complementary rather than competing. Qualitative methods are the efficient front line, catching most hazards and prioritising them quickly, and they flag the minority of scenarios where the consequences are severe enough or the acceptability marginal enough that a quantitative answer is warranted. A QRA then puts firm numbers on those scenarios. Understanding QRA as the high-resolution instrument you reach for selectively, rather than the default, is the key to placing it correctly among the other tools.
A QRA leans heavily on release and inventory information, and this is one place operating data contributes. The size and duration of a modelled release depend on how much material is in a vessel or line and at what pressure and temperature, which is exactly the sort of inventory and process condition data a control system records. Better knowledge of real operating pressures, temperatures, and inventories tightens the consequence estimates, and plant-specific equipment history can refine the failure frequencies away from generic database values. Merobix, as cloud SCADA for oil and gas, keeps process conditions and equipment status from many remote sites visible in one browser and retains that history, giving a QRA team factual inputs on inventories and operating states to ground the modelling rather than relying solely on nameplate assumptions.
Individual risk is the yearly frequency at which a person at a specific location would be harmed, and it is shown as contour lines on a site map. Societal risk describes the potential for an event to harm many people at once and is usually shown as an F-N curve of the frequency of events causing at least a given number of fatalities. Individual risk protects the person at a location, while societal risk guards against catastrophic multi-fatality events.
A HAZOP identifies hazards and a risk matrix ranks them by qualitative judgement, giving a fast but approximate picture. A QRA instead produces numerical risk by modelling consequences and multiplying by event frequencies, yielding measures like individual risk per year. It is far more detailed and resource-intensive, so it is reserved for major hazard sites or scenarios that qualitative methods cannot resolve, rather than used as the everyday default.
An F-N curve is a plot of the frequency (F) of events causing at least a given number of fatalities (N) against that number, drawn on logarithmic scales. It expresses societal risk by showing how a facility's potential incidents spread across frequent-but-small and rare-but-large outcomes. Curves that extend toward high fatality numbers are treated with particular concern because of the aversion to catastrophic events, and they are compared against societal-risk tolerability criteria.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.