SHARE Knowledge Hub

Quantitative Risk Assessment (QRA)

Quantitative risk assessment (QRA) uses numerical scenario, frequency and consequence estimates to inform major-hazard choices. A useful QRA states its scope, data limits and uncertainty, compares real controls, and is updated when conditions change. It supports proportionate risk management; a calculated number never replaces a working barrier or permits unsafe work.

Updated 2026-10-19Great Britain and Northern IrelandVersion 3.3 · review candidate

Understanding Quantitative Risk Assessment (QRA)

Quantitative risk assessment (QRA) uses numerical scenario, frequency and consequence estimates to inform major-hazard choices. A useful QRA states its scope, data limits and uncertainty, compares real controls, and is updated when conditions change. It supports proportionate risk management; a calculated number never replaces a working barrier or permits unsafe work.

Why this matters at work

A concentration contour, exposure dose, thermal flux, overpressure, injury probability and expected fatality are different quantities. State units, duration, receptor position, indoor/outdoor assumptions and vulnerability. A labelled fatality radius without its dose-response and exposure assumptions obscures the decision. Quantify environmental harm through the relevant pathways and receptors rather than treating a human fatality contour as an ecological assessment.

Match the physical model, material properties, release conditions, geometry, weather and exposure assumptions to the actual scenario. Record exclusions and applicability limits, including escalation. Seek competent verification rather than trusting software output because it is detailed. A visually impressive contour cannot establish that the assumed release, occupancy or protection is correct.

How organisations manage this in practice

A credited barrier needs a defined function, effective response time, capacity and availability under the scenario conditions. The historical HSE control-systems technical page distinguishes performance from readiness and relates both to design, testing, maintenance, survivability and independence. Use its method context with its age visible; its old numerical examples or standards references are not current universal criteria. Current COMAH predictive assessment guidance separately examines justified failure data and effective measures.

When deciding between two isolation arrangements, quantify only after identifying how each changes release duration, failure paths, exposed people and uncertainty. A number adds value if it resolves a real option or reveals the mechanism requiring control. When accepted good practice already identifies a necessary measure, implement that measure and establish whether further analysis changes the remaining decision.

Start with the decision and the required level of evidence, using a proportionate method. Quantification is useful when it clarifies scenario contributions, options or uncertainty; it should not postpone obvious good-practice controls. Specify plant states, receptors and harm endpoints so the resulting number has a clear meaning and can be challenged.

Keep numerical comparison, applicable good practice and consideration of further effective risk reduction as separate parts of the case. Use the correct sector and national decision context. A risk criterion is not universal design permission, and cost-benefit work cannot compensate for an invalid scenario set, poor data or an unavailable credited safeguard.

Build the scenario before multiplying probabilities

A quantitative result depends on the defined scenario, evidence and assumptions. An apparently precise number cannot repair missing coverage or dependence between barriers.

Process operators monitoring equipment from a UK industrial control room. Photorealistic illustration.
Illustrative UK workplace scene. Read the practical checks below for responsibilities, evidence and limits. Download image · Download smaller image
  1. Define the frequency basis

    Identify the event, unit, period, operating state and source of the initiating-event estimate.

  2. Trace the branches

    Consider protective-function success or failure, ignition where relevant and consequences with appropriate occupancy and exposure assumptions.

  3. Challenge dependencies

    Identify shared supplies, instruments, maintenance or other common causes before claiming independent risk reduction.

  4. Record uncertainty and authority

    Retain sources, limits, review date and competent interpretation. This symbolic teaching tree supplies no risk-acceptance criterion or automatic operating permission.

Read the complete task-specific control examples

ERICPD is a planning cross-map: Discipline verifies and sustains every control. ESEAP compares effectiveness. A label does not authorise the task; retain the owner, evidence and stop or change decision.

Quantitative Risk Assessment (QRA) · ERICPD

Eliminate
At Crownway, remove an unnecessary flammable-liquid transfer or choose a process that avoids the delivery entirely where the business and technical need allow. This removes the modelled overfill/release pathway; existing inventory hazards still need assessment.
Reduce
Where a compatible, less hazardous material or lower inventory is technically possible, assess its new fire, health and environmental hazards. Otherwise reduce batch quantity or transfer rate within engineered design limits; less material can limit consequence but does not eliminate ignition.
Isolate
Use suitable bund/containment and separation of ignition sources and occupied areas, based on the actual layout. Keep drains and emergency routes within the design decision. These physical barriers interrupt spill spread or exposure, not the initiating release frequency.
Control
Specify independent overfill prevention, level indication and emergency isolation appropriate to the design; test their functional response before use and maintain them. Transfer limits, permit, driver–operator communication and emergency plan are administrative residual controls, not equivalent to an interlock.
PPE
Select chemical-resistant gloves, eye/face protection and protective clothing from the actual liquid safety data and task assessment for connection/disconnection splashes. Check compatibility, size, condition, cleaning and replacement. PPE cannot protect neighbours from a major fire or compensate for lost containment.
Discipline
Train and observe the actual transfer, supervise contractors, exercise emergency response, investigate near misses and verify corrective actions. Keep the model and barriers updated after inventory or layout change. This sustains engineered controls rather than replacing them.

Quantitative Risk Assessment (QRA) · ESEAP

Elimination
Remove an unnecessary flammable-liquid transfer or adopt a process avoiding the delivery where technically and operationally feasible. Existing inventory hazards remain within the assessment.
Substitution
Consider a compatible less hazardous liquid or a process that replaces the hazard, assessing new fire, health and environmental risks. Reducing the quantity or transfer rate alone is source reduction, not substitution; it still needs the engineered design envelope.
Engineering controls
Use suitable bunds, containment and physical separation from ignition sources and occupied areas. Provide independent overfill protection, level indication and emergency isolation for the actual design; these physical protection systems remain engineering controls when their testing is managed administratively.
Administrative controls
Apply approved inventory and transfer limits, permits, driver–operator communication, supervision, emergency plans and change control. Manage proof-test schedules, bypasses, maintenance evidence, competence and near-miss learning so the engineering barriers remain available.
PPE
Choose chemical-resistant gloves, eye/face protection and protective clothing using actual liquid safety data and task assessment. Check compatibility, size, condition, cleaning and replacement; PPE cannot compensate for lost containment or protect neighbours from a major fire.

What is quantitative risk assessment?

Quantitative risk assessment (QRA) models defined harmful scenarios with numerical estimates of frequency and consequence. It helps compare options and expose the assumptions behind a major-hazard decision. It does not itself make a plant safe, replace basic risk assessment or grant permission to operate.

Start with a question the decision-maker can act on: which loss-of-containment events matter, who can be harmed, which barriers prevent or limit them, and what changes if one barrier fails? A QRA combines credible scenarios, event frequencies, consequence modelling and exposure or vulnerability assumptions. The output may be an event frequency, individual risk estimate, group-risk presentation or a comparison of alternatives. Those measures answer different questions; never use one as if it were another. See HSE: highly flammable liquids safety-report guide.

A qualitative assessment describes likelihood and severity in words; a semi-quantitative method orders them with bands or scores; QRA estimates frequencies and consequences numerically with stated units and assumptions. A risk-matrix score is not a calibrated annual probability. The appropriate depth follows the hazard, decision and evidence available. A robust qualitative assessment can be preferable to a precise-looking QRA built from unsupported data.

How does QRA differ from a 5 × 5 assessment?

SHARE’s 5 × 5 multiplies ordinal likelihood and severity ratings to prioritise action. A quantitative risk assessment (QRA) estimates quantities such as event frequency or consequences from data and models. Use the 5 × 5 guide for everyday scored assessments and the examples to compare the scored and qualitative methods.

Use the SHARE 5 × 5 guide · Download the worked examples

HSE guidance on qualitative, semi-quantitative and quantitative assessment

How is a credible QRA assembled?

Define the decision and system boundary first, then trace hazard to event to harm, quantify only the branches that the evidence supports, and test how results change with assumptions.

For a fuel-storage installation, a loss of containment may be caused by overfill, corrosion, incorrect transfer or an impact. Ignition may lead to a pool fire or another event depending on the material, geometry and conditions. People may be exposed by location, occupancy and escape time; environmental effects require a separate appropriate assessment. Analysts should record excluded scenarios, common causes, maintenance states and the probability of each branch conditioned on its preceding events, and dependencies such as shared power or a common maintenance error. Multiplying consistently conditioned branch probabilities does not require mutual independence. Independence is needed before substituting unconditioned marginal probabilities for those conditional values. See HSE: highly flammable liquids safety-report guide.

QRA steps and the decision each supports
StepEvidence and methodCritical challenge
1. FrameQuestion, asset boundary, life-cycle state, people and neighbouring receptors.Does the boundary include deliveries, maintenance, simultaneous work and foreseeable change?
2. Identify scenariosHazard study, incident learning and credible release or failure paths.What was screened out, by whom and why?
3. Estimate frequencyRelevant operational data, equipment failure data and justified event/fault trees.Is the data comparable and current; are shared causes or human actions double counted?
4. Model consequencesMaterial properties, release size, weather, ignition, dispersion, fire or other effects as relevant.Do assumptions cover credible conditions and physical barriers?
5. Estimate exposure and riskOccupancy, vulnerability, emergency response and suitable risk measure.Is this event frequency, individual risk or population risk, and in what units?
6. Test and decideSensitivity, uncertainty, good-practice comparison and alternative controls.Would the decision change under plausible high/low assumptions?

Write the sequence explicitly: release frequency × P(ignition given release) × P(isolation failure given release and ignition). Shared power may make isolation failure more likely on an ignition branch; use a defensible conditional value or model the common cause. Do not silently reuse an unconditional failure figure. NIST probability tutorial, sections 3.1–3.2 explains the general multiplication rule and the independent-event case.

What harm chain should a tank-farm QRA make visible?

A loss of containment is only the initiating event. The assessment must follow credible pathways through failed or successful barriers to people and the environment, while the operating team acts on early warning signs before a modelled outcome occurs.

In an illustrative flammable-liquid transfer, overfill can release fuel. Vapour or a spill may find an ignition source, with burns, smoke inhalation, blast or secondary injury depending on the scenario. Drainage and bund performance affect off-site or environmental consequences. Emergency isolation, detection, separation and response may change the outcome, but only if they are available during the actual task. A changed level alarm, bypassed interlock, unexplained inventory difference, leaking coupling, impaired bund or loss of communications calls for a hold and competent assessment, not a new number on a spreadsheet. See HSE: COMAH Regulations 2015.

Consider vulnerable people, contractors, control-room occupancy, neighbours and emergency responders where relevant. A QRA cannot make an unmodelled person disappear: occupancy assumptions and emergency plans must match actual arrangements.

When is numerical modelling useful, and when should work pause?

Use QRA when a major-hazard choice depends on scenario frequency, consequences or the comparison of alternative safeguards. Use a proportionate qualitative method for simpler work when it can identify and control significant hazards. Pause any job whose required physical barrier or agreed operating assumption has failed.

Ask first whether two credible options would lead to different decisions if their relative frequency or off-site effect were known. If the answer is yes, define the decision, scenarios and data before commissioning a model. If the only task is to control a familiar maintenance hazard, an observable task assessment may be more direct. Where data are sparse, a bounded range and explicit assumptions may be more useful than a single annual number. The choice of method should be recorded and revisited if new evidence appears.

Assessment depth and operating permission are separate decisions. A low estimated frequency is conditional on the barriers represented in the model. If an overfill alarm, isolation path or containment route is impaired now, the current task has crossed a field hold point regardless of the long-run estimate.

Choosing an assessment depth
SituationApproachDecision boundary
Routine task with known hazards and established controlsSite-specific qualitative assessment, competent review and field checks.Do not order QRA merely to generate a score.
Major-accident installation with several credible escalation pathsStructured hazard study; quantitative work where it helps distinguish options or support a safety report.Set model scope and data needs before purchasing software.
Uncertain rare event with weak dataScenario analysis, ranges and sensitivity; a qualitative explanation may be more honest.Do not present a single point estimate as certainty.
Failed critical control during live transferStop/hold, make safe, isolate and escalate under the site plan.Do not wait for a recalculated annual risk number.

How can the hierarchy of controls turn quantitative risk estimates into stronger process controls?

A useful QRA compares prevention and mitigation options against the physical harm path, then checks that selected barriers work in service. Training and personal protective equipment cannot substitute for a feasible design or engineering control.

ERICPD and HSE’s five-level hierarchy are related planning lenses, not identical statutory ladders: ERICPD reduction includes substitution, isolation maps to physical engineering where it separates people, and its discipline step is administrative assurance. The table tests every level against the actual task, including why a higher-order choice may not be feasible. HSE explains the five-level order and PPE as last resort

Crownway is an illustrative fuel-transfer installation. Confirm the actual substance, COMAH tier and GB or NI instrument before applying a control or modelling claim; the two COMAH regimes are separate. Northern Ireland source

The six-prompt ERICPD sequence is used here as a SHARE planning mnemonic: eliminate, reduce, isolate, control, PPE and sustain the controls through supervision and learning. It is not a statutory QRA algorithm. HSE’s public five-level hierarchy provides the control order. For this tank farm, a competent engineer must select the actual design against the substance, equipment, site and applicable duties.

Classification check: containment, physical separation, overfill protection, level indication and emergency isolation are engineering measures. Proof-test management, permits, communication, supervision and change control are administrative. Lower inventory reduces the source; only a suitable replacement of the hazardous material or process establishes substitution.

How should a tank-farm decision work before, during and after transfer?

At Crownway, the QRA supports a choice among safeguards, but the transfer is released only when the real barriers and operating conditions match the approved method. A failed level alarm or containment barrier stops the task even if the last model output was low.

Crownway tank-farm transfer

Task, material and setting: Crownway Fuels operates an illustrative tank farm receiving flammable liquid by road tanker. The teaching case assumes a transfer line, tank-level controls, bund and people nearby. It is not a design basis, permit or approved site method.

People and prerequisites: Operators, tanker driver, maintenance workers, emergency responders and neighbours may be affected. The dutyholder identifies the actual liquid, inventory, layout, occupied areas, weather and emergency interfaces before any assessment is applied.

Before: The team maps overfill, hose failure, valve error and ignition scenarios; checks incident history and comparable data; verifies overfill and isolation barriers; and records transfer states, dependencies and exposed people. A competent reviewer challenges screening and estimates.

During: The operator verifies level indication, connection, communications and transfer limits. Supervisor checks critical barriers and any simultaneous maintenance or delivery interface. Change in material, occupancy, equipment or method triggers reassessment rather than blind reuse of a model.

Hold or stop: Stop the transfer under the site plan for a leak, unexplained level, failed alarm or isolation, unsafe ignition control, impaired containment or mismatched permit. Restrict exposure, use trained emergency arrangements and obtain competent authority to restart.

Completion and hand-back: A competent owner confirms defect correction, physical tests, revised assessment/permit and briefing. The receiving and sending parties agree the hand-back and any residual restrictions; emergency readiness is restored.

Evidence: Retain scope, hazard-study record, model version, input provenance, units and uncertainty, barrier test results, work authorisation, change record, hold/restart decisions and decision-maker sign-off.

If the installation is subject to the Control of Major Accident Hazards (COMAH) regime, the dutyholder must address applicable major-accident prevention and, for the upper tier, safety-report obligations. That does not mean every transfer needs a standalone full QRA. The evidence must be sufficient for the actual hazards, tier and regulatory route.

What does a small QRA calculation actually show?

In a deliberately hypothetical branch, 0.001 initiating events per year × 0.10 probability of ignition given the event × 0.20 probability of isolation failure given that branch gives 0.00002 outcomes per year. This is an outcome frequency, not a person’s annual fatality risk or an acceptance threshold.

The arithmetic is 1 × 10−3/year × 0.1 × 0.2 = 2 × 10−5/year as an annualised model estimate. It does not predict an actual event date. The two conditional factors may be multiplied only if the event-tree branches are defined consistently; shared causes, plant states and dependencies must be checked. None of these invented values is a recommendation for a real installation. See HSE: highly flammable liquids safety-report guide.

A person’s risk would require the event’s physical effects, that person’s location and time present, protective factors and vulnerability. A societal-risk presentation also needs a distribution of possible numbers harmed. If isolation failure is 0.40 rather than 0.20, this branch becomes 4 × 10−5/year; if the initiating frequency is also ten times higher, it becomes 4 × 10−4/year. The spread shows why input provenance and sensitivity matter more than decimal places. It still cannot justify operating with a known failed critical barrier.

Illustrative event-tree calculation and its limits
InputIllustrative valueMeaning or caveat
Initiating loss of containment0.001/yearInvented frequency; use comparable, justified plant data for real work.
Ignition given release0.10Conditional probability; weather, source and material affect it.
Isolation failure given ignition branch0.20Conditional branch; common causes and sequence need checking.
Modelled branch frequency0.00002/yearOutcome frequency only; no receptor or fatality calculation.

An illustrative conditional event tree

Invented teaching values, per year: loss of containment (0.001/year) branches first on ignition and then on isolation. Each probability below is conditioned on the events named before it.

  • No ignition given release: 0.90 → 0.0009/year. This branch is not a claim of no harm; unignited release and environmental consequences are outside this illustration.
  • Ignition given release: 0.10 → 0.0001/year.
    • Isolation succeeds given release and ignition: 0.80 → 0.00008/year. Other consequences and residual effects are not modelled here.
    • Isolation fails given release and ignition: 0.20 → 0.00002/year for the illustrated release–ignition–isolation-failure branch.

The success/failure alternatives at each illustrated split sum to one. Other initiating scenarios and consequence paths are excluded. This original SHARE text event tree illustrates conditioning; it supplies no real plant data or acceptance threshold.

Separate receptor step

Branch frequency → physical effects at the receptor → the person’s location and time present → protection and vulnerability → defined individual outcome. These extra relationships are needed before estimating individual fatality risk. Event frequency alone is insufficient; do not insert an arbitrary fatality factor.

How the existing sensitivity results compare

Base inputs
0.001/year × 0.10 × 0.20 = 0.00002/year (2 × 10−5/year).
Isolation failure doubled
0.001/year × 0.10 × 0.40 = 0.00004/year (4 × 10−5/year).
Initiating frequency also ten times higher
0.010/year × 0.10 × 0.40 = 0.0004/year (4 × 10−4/year).

The changed assumption is named in every case. All three are scenario frequencies, not individual fatality risks or permission to run a failed barrier. The same values and limits remain explained in the prose and input table above.

Why can a numerical result mislead?

A model can be internally correct yet decisionally wrong when scenarios are missing, data are imported from unlike plant, dependencies are ignored or people are placed where the model assumes they are not.

Document epistemic uncertainty (what is not well known) separately from variation in weather, occupancy or operation. Explore plausible ranges and ask whether control ranking changes. Treat a very low value with high uncertainty as a prompt to inspect the scenario and barrier, not as a reason to dismiss it. Record whose data, what edition, what site modifications and what time period underlie the estimate.

Do not transplant HSE land-use planning contours or regulator risk tolerability numbers into a company operating limit. HSE’s Reducing Risks, Protecting People (R2P2) framework describes the regulator’s own decision approach; it is not an instruction to dutyholders to accept an exposure. First meet applicable law and relevant good practice, then justify any residual choice in its actual context.

What common QRA failures should a reviewer challenge?

Challenge the model at its weakest boundary: omitted scenario, optimistic barrier, outdated occupancy, hidden common cause or a risk metric used for the wrong decision.

A reviewer should trace one important result backward to the initiating event, data source, conditional branches, effect model and person exposed. If the branch assumes that the alarm and isolation valve are independent but both lose power together, the calculated benefit can be exaggerated. If driver occupancy is averaged across a year, a loading-bay risk at the moment of transfer can be understated. These are model-boundary errors that a polished chart cannot reveal.

Challenge also what the chosen metric leaves out. An individual fatality estimate may say little about non-fatal injury, environmental damage or response difficulty. The decision record should state which outcomes were assessed elsewhere, who owns them and what evidence would trigger a revision.

QRA failure modes and corrective questions
FailureWhy it mattersAsk or do
Generic release frequenciesPlant condition and maintenance may differ.Show applicability, age and sensitivity of data.
Independent barriers assumed without proofShared power, maintenance or alarm logic can defeat several layers.Map common causes and test dependencies.
Average occupancy hides peaksDrivers, contractors or neighbours may be present at high-risk times.Model credible patterns and transient tasks.
Only a fatality metricInjury, environmental harm or emergency complexity can vanish.State outcome scope and assess other significant harm separately.
Report never revisitedChanged inventory, layout or operating method invalidates assumptions.Set change triggers and a model owner.

What evidence should accompany a QRA-based decision?

A decision file should allow another competent person to reproduce the question, assumptions, control choice and reasons for accepting uncertainty.

  • A dated scope and hazard-study record with scenarios included and excluded, plus the screening rationale.
  • Input register with data source, site applicability, units, range and version; model file and change history.
  • Consequence and exposure assumptions, including people off site where relevant, weather and emergency response.
  • Sensitivity results and explicit limits on what the model can conclude.
  • Good-practice and legal comparison, options considered, chosen safeguards and decision-owner record.
  • Field evidence that critical barriers are installed, tested, maintained and understood.

A polished report without a tested isolation system is not a safe decision. HSE process-safety indicator guidance can help design checks that warn of barrier degradation before an accident.

How should QRA assumptions be maintained after the study?

Assign an owner to each critical assumption and barrier, check leading indicators, and review the model after changes, incidents or evidence that the controls behave differently from prediction.

Use change control for inventory, operating pressure, layout, people, contractors, protective systems and emergency arrangements. Verify the physical barrier before treating it as a probability reduction. Trend overdue tests, bypasses and degraded alarms; investigate near misses and unplanned releases. Re-run only the relevant portion of the model when the change affects its boundary, and record the decision to retain or revise the full assessment.

A QRA often supports a safety report or investment decision, but daily control relies on operating limits, inspections, competence, maintenance and a working stop process. Supervisors should have a simple list of assumptions that can be recognised in the field.

Scheduled safety-report review is a separate boundary: where QRA supports an upper-tier GB COMAH safety report, HSE states the report must be reviewed after five years even without change, and earlier when relevant changes or new information affect it. Notify the Competent Authority of revisions and of a five-year review that produces none. For NI, regulation 10 of the separate COMAH (NI) 2015 instrument specifies safety-report review, including the five-year review and earlier triggers; confirm the currently applicable instrument and any amendments for the site.

The safety-report owner schedules the statutory review and the model owner checks affected QRA inputs, scenarios, barriers and assumptions against it; they record the decision and reconcile the resulting report/model updates. This is a safety-report duty, not a universal five-year expiry for every QRA. Non-COMAH assessments still need review when their own circumstances or requirements demand it. NI regulation 10, original official instrument.

How do standards relate to QRA?

HSE’s public COMAH, risk-assessment and major-hazard guides support the scenario, barrier and review decisions taught here. A private risk-management or technique standard is not used as a reader-facing authority for QRA method or acceptance.

For a live fuel-transfer decision, use the GB COMAH route or the NI COMAH route, and test the actual barriers and assumptions. HSE’s management-system guidance explains why a voluntary certificate cannot establish legal compliance or validate a model. Edition, clause and technique claims from private standards remain held pending controlled-text rights and competent review.

What competence does QRA work require?

Match competence to the task: operators must recognise and act on barrier failure, supervisors must manage change and release, and analysts must defend the scenario, data and model. A generic course certificate alone does not establish any of these abilities.

The dutyholder sets the competence standard and appoints a person capable of assessing each role. Record the observed task or challenged work product, assessor, date, deficiencies and authorised scope. Reassess when equipment, material, limits, model method or emergency arrangements change, after a relevant incident or when observed performance shows a gap. A course may prepare a person for the check; it is not the check itself.

Competence and verification
RoleCompetence to demonstrateWho assesses and recordRecheck trigger
OperatorKnow transfer limits, alarm response, isolation and emergency action.Competent supervisor observes a drill and failed-barrier response; record result and authorised task.Changed alarm/transfer method, failed drill or incident.
SupervisorUnderstand assumptions in the approved method, interface and change triggers.Process-safety/operations lead challenges stop/restart and hand-back; retain signed observation.Changed operating boundary or deficient field decision.
QRA analystUse hazard methods, event/fault trees, consequence and exposure models, units and uncertainty.Independent competent peer reviews a model case, sources and limitations; retain review record.New model method/data, major site change or material model error.
Decision ownerCompare good practice, options, uncertainty and actual control performance.Governance lead or competent adviser challenges signed rationale and funded action.New evidence, changed risk profile or overdue critical action.

How does the lesson transfer beyond a tank farm?

QRA adds most value where major-hazard scenarios and choices require numerical comparison; the same discipline of defining hazards, controls and evidence applies more widely without forcing a full model onto every task.

In chemical manufacture, a proposed vessel or pipe change may alter the set of credible releases and which shutdown option is effective. Offshore, a change in occupied areas or escape arrangements may change the exposure and response assumptions in a safety-case analysis. A warehouse or logistics site first needs to establish whether dangerous-substance inventory and location bring it into a major-hazard regime; an ordinary warehouse task still needs a practical assessment even when no QRA is justified.

The transfer lesson is the decision chain, not a shared numerical threshold. Define the local hazard, identify people and barriers, choose proportionate analysis, then verify the selected control in the real operation. The applicable GB or NI law and any sector-specific regime must be checked for that installation.

Sector transfer without false equivalence
SettingPotential QRA questionImmediate practical control
Chemical manufacture: process operators and neighboursA vessel or pipe modification can change credible fuel or toxic releases; fire, inhalation or off-site injury depends on the actual substance and location.Retain engineered containment and independent isolation; test scenarios before approving the modification. Stop on an unassessed release path. Process-safety lead checks containment and isolation at modification release. Source: HSE: COMAH Regulations 2015 / HSENI: Control of Major Accident Hazards
GB offshore installation: crew and emergency respondersChanged occupancy or escape arrangements can expose more people during a fire or gas-release scenario.Verify detection, isolation, escape and emergency response against the safety case; hold a change that invalidates those assumptions. Offshore safety-case owner checks detection, escape and changed occupancy. Source: HSE: assessment principles for offshore safety cases
Dangerous-substance storage: staff, drivers and neighboursMore stock or a changed layout can increase a major-accident fire, toxic or environmental consequence, depending on the substance.Check tier, segregation, containment and emergency access; stop an unapproved inventory or route change. Storage manager checks stock, separation and emergency access. Source: HSE: COMAH Regulations 2015 / HSENI: Control of Major Accident Hazards
Routine valve inspection: maintenance staffA live or inadequately isolated pipe can release fuel, causing splash injury or ignition; this task needs direct control even if no new full QRA is justified.Isolate and verify the line under a site-specific task assessment; hold for uncertain contents, failed isolation or changed work. Maintenance supervisor checks isolation and the task assessment before work. Source: HSE: steps needed to manage risk / HSENI: risk assessment

What do QRA terms mean?

Name the measure and its units before drawing a conclusion.

QRA acronyms and terms
TermMeaningUse or caution
QRAQuantitative risk assessmentNumerical analysis of defined harmful scenarios, with documented assumptions.
Initiating eventEvent that starts a scenarioFrequency is not automatically a harm or fatality frequency.
Event treeBranching sequence of conditional outcomesBranch probabilities need consistent definitions and dependence checks.
Fault treeLogic model of combinations that can cause a top eventUseful for failure analysis; validate its inputs.
Individual riskEstimated risk to a defined individual/location under stated exposureRequires consequence, occupancy and vulnerability assumptions.
Societal riskDistribution of events affecting groups of peopleDifferent from individual risk and from a single event frequency.
F–N curveCumulative annual frequency F of events with N or more people suffering the specified outcome, usually fatalities in major-hazard practiceInterpret with its outcome definition and population assumptions.
ALARPAs low as reasonably practicableContext-specific legal/risk principle; a modelled number alone does not demonstrate it.
COMAHControl of Major Accident HazardsGB and NI have separate 2015 instruments.
MAPPMajor-accident prevention policyA COMAH management requirement, not a QRA output.
HSE / HSENIHealth and Safety Executive / Health and Safety Executive for Northern IrelandUse the regulator for the site jurisdiction.
ERICPDEliminate, reduce, isolate, control, PPE, disciplineTeaching mnemonic for considering controls; not a legal calculation method.
PPEPersonal protective equipmentResidual protection after stronger feasible controls.
GB / NIGreat Britain / Northern IrelandUse the correct legal instrument for the site location.
ISOInternational Organization for StandardizationHSE describes ISO 45001 as a voluntary management standard; no private clause is relied on here.
R2P2Reducing Risks, Protecting PeopleHSE regulator risk framework, not a dutyholder operating permission.