Safety-critical barriers, instrumentation and assurance

Free online training · Approximate learning time: 90–150 minutes · Self-paced

Full public teaching and readers are free without an account. A free learner account is required for saved LMS progress and recorded assessment. Completion and any course-completion certificate depend on the configured LMS conditions. Full course outline and access routes.

Lesson 1

Start with the demand and the required safe outcome

Describe a barrier duty before inspecting its label or test record.

Start with the demand and the required safe outcome

A protective device only makes sense in relation to an event it must prevent or limit. In the fictional receipt case, ask what hazardous transfer condition creates a demand, how much inventory remains connected and what safe outcome is required. “There is an emergency-stop button” describes equipment. It does not establish which valves close, who stops the upstream supply or whether the response is timely enough for the assessed event.

Use a plain-language function statement: when this condition occurs, this system must achieve this action within the specified opportunity and under the relevant circumstances. The approved performance requirement may also define availability, capacity, environmental resistance and evidence. Learners do not calculate a safety integrity level from the model. They learn to find the competent design basis and challenge an assurance claim that lacks one.

Imagine two case records. Record A shows a red button and a recent inspection sticker. Record B connects a credible event to an initiating signal, logic, communications, final isolation and a test result, with exceptions identified. Record B allows a meaningful question about protection; neither record alone authorises a live test. Tests require their own controlled plan so proving protection does not create an unsafe process condition.

During practice, separate prevention, detection and mitigation. A bund may limit liquid spread after release; it does not stop an overfill. An alarm may require a timely human response; acknowledging it is not the same as taking that response. State the claimed barrier in functional terms and ask who owns it. This establishes what subsequent architecture, inspection and testing evidence must actually demonstrate.

How does FARSI help define barrier performance?

FARSI turns a barrier name into measurable performance criteria.

Learners often memorise barrier names but miss performance. A fire pump, gas detector, trip, alarm or bund is only useful if it performs the required function at the time and under the conditions of the scenario. HSE offshore KP3 asset integrity report.

FARSI also helps challenge overclaiming. A human response to an alarm may be useful, but it should not be credited casually unless response time, workload, alarm clarity, training, staffing and recovery controls are credible.

FARSI elementMeaningExample test question
FunctionalityWhat the element must do.What exactly must the gas detector detect, at what location and alarm level?
AvailabilityWhen the element must be ready.Can the deluge system be unavailable during hot work or plant start-up?
ReliabilityHow likely it is to work on demand.Does proof testing support the claimed risk reduction?
SurvivabilityWhether it survives the event conditions.Will the cable, detector or valve survive fire, blast, flood or impact long enough?
Interaction/independenceWhether interfaces or common causes can defeat it.Does it share power, software, instrument air or human response with another credited barrier?

Applied case

The receipt team offers a photograph of a red ESD station as evidence that the incoming pipeline can be isolated. No linked function or upstream response is described.

Decision and explanation

The credible demand, required safe outcome and complete function that achieves it.

Begin with the required safety function and design basis. A visible device cannot show that the complete response achieves the intended safe outcome.

Look closer

Show how the barriers prevent major harm
Explore the numbered equipment and control points. Existing SHAREhub teaching schematic; illustrative and not to scale.
Prevent the release

Find the route check and independent trip on the left. Ask which initiating event each barrier interrupts, who owns it and what current evidence supports it.

Limit the consequences

Follow the right side from loss of containment to fire, pollution and people exposed. Containment and isolation have specific jobs; their presence on a diagram does not prove availability.

Challenge shared failure

The dashed connection asks whether protection shares power, sensors or human dependencies. Several symbols can still depend on one vulnerable part of the system.

01 · Read the situation

The receipt team offers a photograph of a red ESD station as evidence that the incoming pipeline can be isolated. No linked function or upstream response is described.

02 · Explain the principle

A protective device only makes sense in relation to an event it must prevent or limit. In the fictional receipt case, ask what hazardous transfer condition creates a demand, how much inventory remains connected and what safe outcome is required. “There is an emergency-stop button” describes equipment. It does not establish which valves close, who stops the upstream supply or whether the response is timely enough for the assessed event.

03 · Carry it into work

Begin with the required safety function and design basis. A visible device cannot show that the complete response achieves the intended safe outcome.

Lesson 2

Trace the complete path, including the final element

Distinguish component checks from evidence that the whole protection works.

Trace the complete path, including the final element

A complete protective path can include a sensor or push button, signal wiring, logic, communications, a final valve or other action, and supporting services. The fictional boundary ESD check proved that a local lamp changed state. It did not show the intended final isolation or upstream signal. A partial check may be useful maintenance evidence, but its scope must be stated honestly when deciding what protection is available.

Ask what a failure could hide between the tested point and the required outcome. A blocked connection can prevent a sensor seeing the condition; incorrect configuration can send the wrong command; an actuator can move without the required valve response. The course does not teach learners to manipulate these components. It teaches them to request the approved test method and understand which parts of the claim its results support.

For a gas detector, calibration addresses instrument response under specified conditions. Coverage asks whether the actual substance and release can reach a suitable detector in time. The linked action asks what happens next and whether people or equipment respond as required. A calibrated device can still leave a scenario undetected. Siting, environment, maintenance state, failure signals and the consequence of a lost detector all belong in the assurance discussion.

Use an evidence table in the exercise: claimed step, evidence seen, limitation, responsible person and operating implication. Record a gap as a gap rather than assuming an untested section passed. Competent personnel determine appropriate restrictions and verification. When the missing evidence concerns an essential function, the safe response is the approved impairment arrangement, not an improvised live demonstration during transfer.

What would a full safety critical performance standard look like?

A full example standard makes the barrier expectation testable for operations, maintenance, engineers and managers.

For learners, the important shift is from naming a barrier to defining the standard it must meet. A high-high level trip is not simply present or absent. It has a purpose, set point, response time, final element, owner, test method, proof interval, impairment rule, bypass authority and evidence trail. HSE L65 offshore fire, explosion and emergency response guidance.

The exact figures must come from the site design basis and competent engineering judgement. The example below shows the structure, not a universal set point.

Performance standard fieldExample contentEvidence expected
PurposePrevent tank overfill during receipt or transfer by detecting high-high level and stopping inflow before loss of containment.Scenario register, HAZOP/bowtie link and design basis.
Performance requirementSensor, logic solver and final valve act together to close inflow within the defined response time at the agreed high-high set point.Cause and effect chart, set point register and loop test procedure.
OwnerAsset owner owns risk; technical authority owns engineering standard; operations owns daily availability; maintenance owns proof testing and defects.RACI, asset register and maintenance plan.
Testing methodProof test checks sensor, alarm, logic, final element, valve closure and operator indication, not only the instrument reading.Approved proof-test procedure and completed test sheet.
Proof intervalInterval is justified by risk assessment, reliability data, manufacturer guidance, failure history and the credited risk reduction.Proof-test schedule and interval basis.
Impairment ruleTank filling is restricted or stopped if the trip is failed, overdue, bypassed or cannot be proven before receipt.Impairment assessment and operating restriction.
Bypass authorisationTime-limited bypass requires operations manager, technical authority and process safety review with compensating controls.Bypass log, authorisation and handover note.
EscalationFailed proof test, repeated nuisance trip, open bypass, changed set point or changed operating envelope triggers MOC or senior review.MOC record, defect escalation and restart approval.
Evidence packDesign basis, performance standard, C&E, set point record, proof-test evidence, defect close-out, bypass history and MOC trail.Retrievable barrier file and dashboard entry.

A device exercise is not automatically a proof test of the credited complete function. Where demand-condition testing is unsafe or incomplete, the approved method must state coverage, untested dangerous failures, alternative evidence and residual uncertainty. Retain as-found/as-left results and restoration checks; a partial-stroke exercise alone does not demonstrate the full safety function.HSE COMAH control systems

Applied case

A detector certificate is current, but no record explains coverage of a low-level petrol release or tests the alarm and final action. A boundary ESD record stops at the panel lamp.

Decision and explanation

The evidence covers only part of each claim; seek coverage and complete-function verification and apply the approved restrictions.

Each check has a defined scope. Coverage, linked response and final action need suitable evidence; untested parts cannot be assumed effective.

Look closer

Process operators monitoring equipment from a UK industrial control room. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Tank high-high trip bypassed during receipt: The level rises past the safe limit and liquid spills into a flammable-vapour area; ignition can injure workers and the public. Fire-water pump or utility unavailable: A fire grows while mitigation lacks pressure or supply; people cannot rely on the planned response. Common-cause failure of detection and shutdown: A shared power or instrument-air failure defeats more than one credited layer when demand occurs.
01 · Read the situation

A detector certificate is current, but no record explains coverage of a low-level petrol release or tests the alarm and final action. A boundary ESD record stops at the panel lamp.

02 · Explain the principle

A complete protective path can include a sensor or push button, signal wiring, logic, communications, a final valve or other action, and supporting services. The fictional boundary ESD check proved that a local lamp changed state. It did not show the intended final isolation or upstream signal. A partial check may be useful maintenance evidence, but its scope must be stated honestly when deciding what protection is available.

03 · Carry it into work

Each check has a defined scope. Coverage, linked response and final action need suitable evidence; untested parts cannot be assumed effective.

Lesson 3

Two labels can hide one common failure

Recognise shared dependencies without inventing independent protection.

Two labels can hide one common failure

The tank drawing in this course gives one level sensor three labels: display, high alarm and independent trip. The labels imply three opportunities to prevent overfill, yet one sensor fault may defeat all three. Distinct names or screen symbols do not establish independence. The relevant question is whether the assumed separate protection remains effective when the initiating failure or a common cause occurs.

Draw the actual dependencies before counting barriers. Include sensing connections, logic, software, power, communications, instrument air, final elements and the physical route exposed to fire or flooding. A change that moves separate cables into one vulnerable route may alter an important assumption even if neither circuit’s label changes. Competent engineering assessment determines the required separation and integrity; this lesson does not assign a design or authorise a bypass.

The vapour-recovery carryover case makes the same point through a process connection. A recovery system designed for vapour can face abnormal liquid inflow from connected operations. Ask where that liquid originates, what stops it and whether the proposed protection shares the failed control. “The recovery unit is running” does not answer those questions. Trace a credible abnormal path, not only the normal flow shown by arrows.

In the utility case, a service being switched on says little about quality, reserve or behaviour on failure. An instrument-air fault, loss of backup power or unsafe automatic recovery can affect multiple protections together. Record dependent users and their required safe response. Water used for welfare also requires its own appropriate hygiene controls; being supplied from the same utility area does not make its health risks identical to process protection.

What makes a protection layer independent and credible?

A protection layer is credible when it is effective for the scenario, independent from the cause and other layers, auditable and maintained.

A safeguard should not be credited simply because it exists. It must act in time, withstand the event, be available when demanded and have enough reliability for the risk reduction claimed. It should also be testable, owned and understood by the people who depend on it.

Human response can be valuable, but it is often over-credited. The team should ask whether the alarm is clear, the operator is present, the response time is sufficient, the diagnosis is realistic, the procedure is usable and the action has been practised.

Protection layer credibility
Criterion Meaning Example challenge
CriterionEffectiveness MeaningThe layer can prevent or mitigate the scenario. Example challengeWould it actually stop overfill or only warn that it is happening?
CriterionIndependence MeaningIt is not defeated by the same failure as other layers. Example challengeDoes the alarm use the same transmitter as the trip?
CriterionSpecificity MeaningIt is designed for this hazard and consequence path. Example challengeIs the layer relevant to the scenario being analysed?
CriterionAuditability MeaningIts condition can be checked and evidenced. Example challengeWhere is the test, inspection or drill evidence?
CriterionAvailability MeaningIt is ready when demanded. Example challengeAre bypasses, overrides or failed proof tests controlled?
CriterionSurvivability MeaningIt can function during the event conditions. Example challengeWill fire, power loss, flood or vibration defeat it?

Stop or restrict activity that depends on an unavailable credited barrier. A competent assessment must establish and verify any suitable time-limited alternative before reliance; permit, senior approval or manager acceptance alone does not establish safety.

Specific official context: COMAH assessment criteria; functional safety guidance; control-systems guidance. Method tables, scenarios and invented numbers are teaching applications, not approved designs. Competent scenario data, layer independence/performance, full-function lifecycle and licensed standards approval remain required.

Applied case

TK-102 uses a common sensor and logic for display, alarm and the credited independent trip. The drawing also omits loss of instrument air to the final valve.

Decision and explanation

As unproven until competent assessment addresses architecture, shared dependencies and the specified function.

Independence is a property of the actual architecture and failure behaviour. A separate label or display cannot remove a shared failure.

Look closer

Workers reviewing a permit beside safely isolated industrial equipment. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Elimination: Remove an unnecessary transfer or hazardous stock through a validated supply or process redesign. Substitution: Evaluate a less hazardous stored material or suitable alternative supply arrangement through change review. Engineering controls: Separate delivery operations from people and incompatible plant; control the credible spill and escalation area. Administrative controls: Confirm tank/product identity, available ullage, transfer limits and the receiving operator’s communication and stop authority. PPE: Select splash and other protection for the actual product; PPE is not protection against a major tank release.
01 · Read the situation

TK-102 uses a common sensor and logic for display, alarm and the credited independent trip. The drawing also omits loss of instrument air to the final valve.

02 · Explain the principle

The tank drawing in this course gives one level sensor three labels: display, high alarm and independent trip. The labels imply three opportunities to prevent overfill, yet one sensor fault may defeat all three. Distinct names or screen symbols do not establish independence. The relevant question is whether the assumed separate protection remains effective when the initiating failure or a common cause occurs.

03 · Carry it into work

Independence is a property of the actual architecture and failure behaviour. A separate label or display cannot remove a shared failure.

Lesson 4

Turn impairment information into an operating decision

Identify the owner, restrictions and restoration evidence for a degraded barrier.

Turn impairment information into an operating decision

A protection impairment needs an explicit operating decision. In the fictional terminal, “engineer informed” is an action but leaves dispatch unsure which transfers can continue. State the affected function, credible exposure, current plant condition and responsible authority. The competent decision must establish whether the operation can proceed under defined restrictions or must remain unavailable. Learners do not choose compensating measures by intuition.

A useful impairment record includes its start, affected users, authorised controls, communication, duration limits, review point and restoration conditions. Cross-check it against permits, planned transfers and emergency arrangements. A temporary bypass that remains invisible to the next shift can turn a known defect into a hidden vulnerability. Only authorised people may introduce, extend or remove such an arrangement under the actual site system.

Testing should preserve both the condition found and the condition left, with exceptions and repairs traceable. A repaired component may need relevant repeat verification; a passed component check may still leave another part of the function unresolved. These records help competent specialists compare observed failures with the assumptions supporting the protection claim. They also reveal whether repeated “temporary” fixes conceal a wider design, maintenance or operating problem.

Close the exercise by writing a release statement that someone else can audit: what was restored, the evidence and limits, who accepted it and when affected teams were informed. Do not close an impairment merely because a deadline arrived or a control was reset. A clear unresolved status can be the correct outcome, provided the plant remains in its authorised safe condition and ownership of the next action is explicit.

How should bypass and impairment be controlled?

A safety critical barrier that is bypassed, failed or degraded should trigger formal risk review and clear authority.

Barrier impairment is a live risk condition. The plant may look normal, but the protection layer has changed. Controls should define what can continue, what must stop, what temporary controls are acceptable and who can authorise exceptions. HSE HSG254 process safety indicators.

The hardest failures are often tolerated impairments: alarms inhibited for nuisance, trips bypassed for production, fire systems isolated during maintenance or proof tests deferred because shutdown is inconvenient.

Impairment conditionDecision ruleEscalation point
Barrier failed proof testKeep the affected operation stopped until restored, or until competent assessment demonstrates an acceptable restricted state with effective compensating protection, defined limits and authorisation.Barrier is credited for a major accident scenario.
Safety trip bypassedRequire time-limited authorisation, compensating measures and operator briefing.Bypass extends beyond agreed time or becomes routine.
Fire/gas detection unavailableReview hot work, occupancy, ignition sources and emergency response.Detector protects occupied or high-hazard area.
Relief or vent system defectCheck safe operating limits and loading conditions immediately.Overpressure scenario can occur before repair.
Procedure unavailable or unvalidatedDo not rely on informal memory for safety critical tasks.Start-up, line breaking, isolation or abnormal operation is affected.
Common utility failureReview all barriers dependent on the same power, air, network or cooling source.Multiple credited barriers share the same vulnerability.

Applied case

A liquid-carryover trip is under investigation. The defect log says “attention requested”, while the loading plan assumes normal availability and the next shift has not been briefed.

Decision and explanation

Explicit affected operations, authorised restrictions, ownership, communication and restoration evidence.

A defect notice must connect to a current operating decision. Restrictions, responsibilities and evidence-based restoration prevent a known impairment being silently treated as normal protection.

Look closer

Process operators monitoring equipment from a UK industrial control room. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Function: High-high trip detects level at defined set point and closes import valve within required time. Independence: Trip independence from normal level control is defined and verified. Testing: Proof test covers sensor, logic, final element, alarm and valve response.
01 · Read the situation

A liquid-carryover trip is under investigation. The defect log says “attention requested”, while the loading plan assumes normal availability and the next shift has not been briefed.

02 · Explain the principle

A protection impairment needs an explicit operating decision. In the fictional terminal, “engineer informed” is an action but leaves dispatch unsure which transfers can continue. State the affected function, credible exposure, current plant condition and responsible authority. The competent decision must establish whether the operation can proceed under defined restrictions or must remain unavailable. Learners do not choose compensating measures by intuition.

03 · Carry it into work

A defect notice must connect to a current operating decision. Restrictions, responsibilities and evidence-based restoration prevent a known impairment being silently treated as normal protection.

Lesson 5

Verify the emergency shutdown function from demand to safe state

An emergency-stop device is the start of a safety function; verify the final process outcome and the dependencies in between.

Understand this fuel-industry task

The red ESD station is easy to recognise, but its purpose is not merely to light a lamp or change a display. In this receipt scenario, the required safety action includes the intended isolation and the relevant upstream interface. A healthy input switch can coexist with a failed actuator, a bypassed output or a communications fault. Testing the initiating button alone therefore leaves important parts of the function unproven. The success criterion must describe what the process actually does following the demand.

Define that function from the accident scenario: the condition that initiates it, the required safe state, appropriate response performance and the dependencies needed to achieve and maintain it. Functional safety involves specification, design, validation, operation, testing, maintenance and modification across the lifecycle. An end-to-end proof test needs a competent planned method that avoids introducing an unsafe plant state. Test coverage, exclusions, failed first demands and restoration of bypasses all matter; repeated operation until a valve finally moves should not erase evidence that it initially failed.

Trace the initiating event and loss of control

StepCase pathway
initiatingEventAn emergency demand occurs after incomplete testing has overlooked a failed final element or supplier interface.
lossOfControlThe command is acknowledged but the required isolation or coordinated cessation of inflow is not achieved.
consequencesFuel continues feeding the developing incident, increasing release duration, escalation and the difficulty of safe evacuation or response.

Choose effective safeguards

Control priorityPractical application
EliminateRemove unnecessary hazardous material, energy and transfer operations.
SubstituteChoose a less hazardous product or simpler process where practicable.
EngineerPrefer effective engineered preventive and mitigative barriers.
ManageUse suitable hazard studies, verify assumptions and act on weak-signal indicators.
ProtectKeep PPE as an additional exposure control within its actual capabilities.

Treat unsupported claims of availability as an impairment and apply the approved assessment and operating restriction rather than assuming the visible ESD station is effective.

Specify the complete function, including initiating devices, logic, communications, final elements, utilities, safe state and response requirements derived from the actual receipt scenario.

Plan competent tests with suitable coverage and safe controls, addressing relevant failure modes and dependencies without creating uncontrolled isolation or surge conditions.

Manage defects, bypasses and restoration through controlled authority; retain first-test results and confirm the function is returned to its intended operating configuration after testing.

Verify effectiveness before release

Transfer the decision to the workplace

Apply end-to-end reasoning to high-level trips, firewater starts, gas-detection shutdowns and loading permissives, rather than equating a healthy sensor with successful protection.

What would a full safety critical performance standard look like?

A full example standard makes the barrier expectation testable for operations, maintenance, engineers and managers.

For learners, the important shift is from naming a barrier to defining the standard it must meet. A high-high level trip is not simply present or absent. It has a purpose, set point, response time, final element, owner, test method, proof interval, impairment rule, bypass authority and evidence trail. HSE L65 offshore fire, explosion and emergency response guidance.

The exact figures must come from the site design basis and competent engineering judgement. The example below shows the structure, not a universal set point.

Performance standard fieldExample contentEvidence expected
PurposePrevent tank overfill during receipt or transfer by detecting high-high level and stopping inflow before loss of containment.Scenario register, HAZOP/bowtie link and design basis.
Performance requirementSensor, logic solver and final valve act together to close inflow within the defined response time at the agreed high-high set point.Cause and effect chart, set point register and loop test procedure.
OwnerAsset owner owns risk; technical authority owns engineering standard; operations owns daily availability; maintenance owns proof testing and defects.RACI, asset register and maintenance plan.
Testing methodProof test checks sensor, alarm, logic, final element, valve closure and operator indication, not only the instrument reading.Approved proof-test procedure and completed test sheet.
Proof intervalInterval is justified by risk assessment, reliability data, manufacturer guidance, failure history and the credited risk reduction.Proof-test schedule and interval basis.
Impairment ruleTank filling is restricted or stopped if the trip is failed, overdue, bypassed or cannot be proven before receipt.Impairment assessment and operating restriction.
Bypass authorisationTime-limited bypass requires operations manager, technical authority and process safety review with compensating controls.Bypass log, authorisation and handover note.
EscalationFailed proof test, repeated nuisance trip, open bypass, changed set point or changed operating envelope triggers MOC or senior review.MOC record, defect escalation and restart approval.
Evidence packDesign basis, performance standard, C&E, set point record, proof-test evidence, defect close-out, bypass history and MOC trail.Retrievable barrier file and dashboard entry.

A device exercise is not automatically a proof test of the credited complete function. Where demand-condition testing is unsafe or incomplete, the approved method must state coverage, untested dangerous failures, alternative evidence and residual uncertainty. Retain as-found/as-left results and restoration checks; a partial-stroke exercise alone does not demonstrate the full safety function.HSE COMAH control systems

Applied case

The maintenance record says 'ESD tested: lamp changed to red'. The tank operator cannot find a record of final valve action or the upstream stop signal.

Decision and explanation

The complete function remains unproven; apply the approved impairment process and arrange suitable verification of the missing stages.

The record supports only the observed part of the test. It does not establish final isolation or the required supplier response. Availability must be supported by evidence covering the claimed function, with safe operating decisions made while that evidence is incomplete. An improvised live test can create other hazards and may still fail to establish test coverage, acceptance criteria or safe restoration.

Look closer

Process operators monitoring equipment from a UK industrial control room. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Tank high-high trip bypassed during receipt: The level rises past the safe limit and liquid spills into a flammable-vapour area; ignition can injure workers and the public. Fire-water pump or utility unavailable: A fire grows while mitigation lacks pressure or supply; people cannot rely on the planned response. Common-cause failure of detection and shutdown: A shared power or instrument-air failure defeats more than one credited layer when demand occurs.
01 · What to notice

The red boundary ESD station is visible. The case record shows a push-button check but no evidence that the intended final isolation and upstream signal operated.

02 · Trace the harm

A lamp or command acknowledgement can work while a final element or communications interface fails, allowing continued inflow during an emergency.

03 · Test the safeguard

Review a suitable witnessed test from initiating device through logic, communications and final elements, including loss of power or signal where relevant.

Lesson 6

Check independence across the entire overfill function

Count effective independent functions only when their performance and dependencies support that claim; do not count screen objects as protection layers.

Understand this fuel-industry task

The exercise drawing for TK-102 gives a level display, a high alarm and a trip three different names, but all depend on the same sensor and common logic. A dangerous shared fault can therefore make the display look reasonable while also preventing the alarm and trip from detecting the rising level. Counting those labels as three independent layers creates false confidence. Redundancy and independence are different questions: additional devices help only when their common dependencies and failure mechanisms are understood for the scenario being assessed.

Follow the complete safety function through sensing, logic, communications, utilities and final action. Assess shared installation, environment, maintenance and bypass arrangements as well as electronics. An alarm requiring human response also depends on usable information, competence, workload and sufficient time for the action. The performance standard should describe the required outcome and how it is maintained and tested. FARSI is a teaching prompt for functionality, availability, reliability, survivability and interaction or independence; it is not a statutory checklist or a method that assigns a SIL to equipment.

Trace the initiating event and loss of control

StepCase pathway
initiatingEventA shared sensor or common-system fault occurs in an architecture incorrectly credited as several independent overfill barriers.
lossOfControlThe same failure defeats level control, warning and protective action, leaving less risk reduction than the assessment assumes.
consequencesReceipt can continue unnoticed towards overfill, followed by fuel and vapour release with fire, explosion and off-site consequences.

Choose effective safeguards

Control priorityPractical application
EliminateAvoid unnecessary filling and minimise hazardous inventory.
SubstituteUse a simpler receipt arrangement with less dependence on human intervention.
EngineerProvide an independently assessed overfill function and effective final stopping element.
ManageVerify capacity, maintain proof-test evidence and control impairments and bypasses.
ProtectKeep people away from a potential release; PPE will not protect against a major vapour-cloud event.

Do not credit unsupported independence; refer the architecture to competent assessment and apply the approved impairment or operating restrictions for the actual protection status.

Define the specific safety function and justified integrity need before selecting architecture, considering the accident scenario, demand and complete final process action.

Assess common causes across sensing, logic, power, environment, communications, maintenance and final elements; specify the independence or other measures justified by the risk method.

Maintain lifecycle evidence through suitable design verification, validation, proof testing, controlled bypass and change management, including failures and test-coverage limitations.

Verify effectiveness before release

Transfer the decision to the workplace

Use the same dependency map for duty and standby pumps, fire detectors sharing power, redundant communications and multiple checks based on one incorrect information source.

What makes a protection layer independent and credible?

A protection layer is credible when it is effective for the scenario, independent from the cause and other layers, auditable and maintained.

A safeguard should not be credited simply because it exists. It must act in time, withstand the event, be available when demanded and have enough reliability for the risk reduction claimed. It should also be testable, owned and understood by the people who depend on it.

Human response can be valuable, but it is often over-credited. The team should ask whether the alarm is clear, the operator is present, the response time is sufficient, the diagnosis is realistic, the procedure is usable and the action has been practised.

Protection layer credibility
Criterion Meaning Example challenge
CriterionEffectiveness MeaningThe layer can prevent or mitigate the scenario. Example challengeWould it actually stop overfill or only warn that it is happening?
CriterionIndependence MeaningIt is not defeated by the same failure as other layers. Example challengeDoes the alarm use the same transmitter as the trip?
CriterionSpecificity MeaningIt is designed for this hazard and consequence path. Example challengeIs the layer relevant to the scenario being analysed?
CriterionAuditability MeaningIts condition can be checked and evidenced. Example challengeWhere is the test, inspection or drill evidence?
CriterionAvailability MeaningIt is ready when demanded. Example challengeAre bypasses, overrides or failed proof tests controlled?
CriterionSurvivability MeaningIt can function during the event conditions. Example challengeWill fire, power loss, flood or vibration defeat it?

Stop or restrict activity that depends on an unavailable credited barrier. A competent assessment must establish and verify any suitable time-limited alternative before reliance; permit, senior approval or manager acceptance alone does not establish safety.

Specific official context: COMAH assessment criteria; functional safety guidance; control-systems guidance. Method tables, scenarios and invented numbers are teaching applications, not approved designs. Competent scenario data, layer independence/performance, full-function lifecycle and licensed standards approval remain required.

Applied case

A presentation claims three independent barriers: displayed level, high alarm and trip. The diagram shows one sensor feeding shared logic for all three.

Decision and explanation

Challenge the independence claim and obtain a competent whole-function and common-cause assessment before crediting protection.

Names do not prevent the shared sensor or common logic failure from defeating all outputs together. Duplicating the display does not remove its dependence on the same failed measurement or shared system. Protection claims need scenario-specific architecture and lifecycle evidence. Current operation must reflect the protection actually available.

Look closer

Workers reviewing a permit beside safely isolated industrial equipment. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Elimination: Remove an unnecessary transfer or hazardous stock through a validated supply or process redesign. Substitution: Evaluate a less hazardous stored material or suitable alternative supply arrangement through change review. Engineering controls: Separate delivery operations from people and incompatible plant; control the credible spill and escalation area. Administrative controls: Confirm tank/product identity, available ullage, transfer limits and the receiving operator’s communication and stop authority. PPE: Select splash and other protection for the actual product; PPE is not protection against a major tank release.
01 · What to notice

Tank TK-102’s exercise drawing uses one sensor and common logic for the displayed level, high alarm and claimed independent trip.

02 · Trace the harm

A shared dangerous failure can defeat several apparent safeguards together; counting labels as independent layers overstates risk reduction.

03 · Test the safeguard

Review the approved architecture and full functional test evidence against the scenario performance standard; ensure shared dependencies and common causes are explicitly addressed.

Lesson 7

VRU overfill: a vapour system can receive liquid

Count demonstrated functions with justified independence, not the number of alarms or shutdown labels on one controller.

Understand this fuel-industry task

This plant receives displaced hydrocarbon vapour from loading operations, but its name does not rule out abnormal liquid entry. Depending on the recovery design, credible routes can include connected tanker transfers and the unit’s own absorbent supply and return arrangements. A liquid source under head pressure can continue feeding an unwanted route even if a particular local pump has stopped. The hazard review must trace the complete connected system rather than assuming the normal process stream is the only possible one.

HSE’s 2023 VRU safety notice describes an incident where the basic process control system and overfill prevention shared a failure dependency. This scene asks the learner to find that reasoning trap: two displayed functions may fail together if they share the initiating fault, sensors, logic, utilities or final elements. Identify the required protective action and justify independence and reliability through competent assessment. Review overfill, over-temperature and emergency shutdown functions as applicable to the actual unit. Neither this model nor a device certificate can assign a safety-integrity level or prove the complete installed protection.

Trace the initiating event and loss of control

StepCase pathway
initiatingEventAn abnormal liquid path feeds the recovery unit during a control or valve fault.
lossOfControlThe same failure disables the routine control and credited protection, leaving liquid inflow without an effective independent stopping action.
consequencesRecovery equipment can overfill and release fuel through an unintended outlet, creating fire, explosion and environmental consequences away from the original loading point.

Choose effective safeguards

Control priorityPractical application
EliminatePrevent releases and remove unnecessary flammable material and ignition sources.
SubstituteAssess a less hazardous substance or transfer arrangement.
EngineerProvide suitable containment, recovery, ventilation and assessed static/overfill protection.
ManageMaintain hazardous-area records, control equipment and work permits, and investigate faults.
ProtectUse clothing and equipment selected for the assessed electrostatic and exposure risks.

Identify all credible liquid sources, driving heads and routes through the tanker-to-recovery interface and recovery plant’s own circuits.

Establish the required protection through competent scenario and dependency assessment, with independent measures where needed rather than counting shared functions twice.

Physically verify installed sensing, logic and final protective elements against current drawings and cause-and-effect, with suitable maintenance and proof testing.

Control changes and impairments across connected loading, tank and VRU operations; retain the required restrictions until restoration is demonstrated and authorised.

Verify effectiveness before release

Transfer the decision to the workplace

Carry this dependency test to tank overfill protection and electric/diesel firewater backup: diversity in names or equipment does not eliminate a shared failure path.

What evidence should prove a safety critical element is healthy?

Evidence should show design basis, performance criteria, maintenance, proof testing, impairment control and management review.

A performance standard is only as good as the evidence behind it. A certificate or inspection record may show something was checked, but it does not always prove the right thing was checked against the right scenario. HSE offshore KP3 asset integrity report.

For major hazards, evidence should be traceable from the scenario through the credited barrier and into the live maintenance, inspection, operations and assurance systems.

Evidence itemWhat it provesWeak signal
Major accident scenario referenceThe barrier is linked to a credible hazard.Barrier exists with no scenario owner.
Performance standardFunction, criteria and verification route are defined.Standard uses vague words such as adequate or available.
Design basisCapacity, set point, rating or response time is justified.Design basis cannot be found or has changed.
Proof-test procedureThe test checks the actual safety function.Test only checks part of the loop.
Inspection and maintenance recordTasks are completed and defects managed.Overdue tasks or repeat failures occur.
Bypass and impairment logUnavailable barriers are authorised and risk assessed.Bypasses are open-ended or poorly justified.
MOC recordChanges to plant, software, alarms, set points or procedures are reviewed.Change occurs without reassessing the performance standard.

Applied case

A review finds that the VRU’s process controller also provides its overfill shutdown using the same fault-sensitive components. Routine operation and recent calibration are satisfactory.

Decision and explanation

Apply the approved restriction and review the complete abnormal-liquid route, dependencies and required independent protection.

Different labels do not establish independence. The failure that creates the demand may also defeat the claimed protection. Abnormal liquid routes must be assessed. Successful vapour service does not show that liquid carryover or absorbent inflow is safely controlled. This addresses the failure mechanism identified in HSE’s safety notice and requires physical and functional evidence before relying on the protection.

Look closer

Process operators monitoring equipment from a UK industrial control room. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Function: High-high trip detects level at defined set point and closes import valve within required time. Independence: Trip independence from normal level control is defined and verified. Testing: Proof test covers sensor, logic, final element, alarm and valve response.
01 · What to notice

The vapour-recovery exercise has not examined abnormal liquid inflow from connected transfer or absorbent circuits, shared control failures or effective independent protection.

02 · Trace the harm

Liquid can overfill or damage recovery equipment and cause a release at another location; successful routine vapour service does not prove abnormal-duty protection.

03 · Test the safeguard

Trace the credible source through the actual plant and verify the specified detection and final protective action with suitable test and design evidence.

Lesson 8

Detector assurance: calibration, coverage and final action are different

Ask three questions separately: can it sense the event, will the signal cause the right response, and does the final action work?

Understand this fuel-industry task

Calibration checks or adjusts a detector’s response against a known input under a specified method. It is important, but cannot establish that a credible release will reach the sensor, that the technology suits every relevant substance or that the linked shutdown will succeed. A functioning sensor can sit outside the likely plume, be screened by new equipment or respond too late for the protective purpose. A zero indication must therefore be understood within the detector’s duty and limitations.

This VRU case provides a calibration record but no coverage or cause-and-effect justification. Review the credible sources, weather or ventilation, obstruction, target substances, response time and sensor failure behaviour. Then follow the signal through the operator alarm or automatic logic to the final equipment that achieves the defined safe state. Functional and proof testing need a controlled method, success criteria, management of bypasses and verified reinstatement. Some tests necessarily cover only parts of the chain; the assurance plan must identify those limits and the other evidence needed, rather than calling every calibration a complete proof test.

Trace the initiating event and loss of control

StepCase pathway
initiatingEventA release occurs where the detector cannot recognise it in time, or a downstream protective element fails.
lossOfControlThe installation is credited as an effective safeguard solely because the sensor has a current calibration record.
consequencesPeople remain exposed or fuel continues flowing, allowing a release, fire or explosion to escalate despite a healthy-looking display.

Choose effective safeguards

Control priorityPractical application
EliminateRemove unnecessary hazardous material, energy and transfer operations.
SubstituteChoose a less hazardous product or simpler process where practicable.
EngineerPrefer effective engineered preventive and mitigative barriers.
ManageUse suitable hazard studies, verify assumptions and act on weak-signal indicators.
ProtectKeep PPE as an additional exposure control within its actual capabilities.

Design detection from credible scenarios and target substances, with justified siting, response performance, environmental suitability and fault indications.

Define the complete required action, including operator response where credited, logic, final elements, utilities and communication dependencies.

Use suitable inspection, calibration, functional and proof-test arrangements that reveal relevant dangerous failures and explicitly state partial-test limitations.

Control detector outages, inhibitions and restoration through the impairment process and update the design review after layout, product or ventilation changes.

Verify effectiveness before release

Transfer the decision to the workplace

Use this chain at tank level protection, ESD stations and fire-pump initiation. A test certificate should identify exactly which part of the safety function it demonstrates.

What can make a gas or flame detector miss the event?

A detector only works for the substance, release path and environmental conditions it was designed and maintained to sense.

Gas or vapour can disperse away from the sensor, stratify, be diluted or be blocked by equipment. Flame detection can depend on fuel and line of sight; smoke and heat devices respond differently in ventilation, dust, steam or high ceilings. A failed-zero reading, contaminated sensing head, lost power or unavailable communication path can look reassuring unless faults are made visible.

Use the actual scenario, fluid properties, ventilation, obstructions, weather, occupancy and response time in competent design and coverage assessment. Test that the device, logic, annunciation and final response work as intended; the page gives no generic placement, threshold, voting or proof-test interval.

Official context: HSE leak and gas detection; HSE functional-safety guidance. Check the stated jurisdiction and premises; practical examples require a competent site assessment.

Applied case

A sensor passes calibration. A new enclosure now stands between the likely release and sensor, and no record shows the linked isolation valve has been tested.

Decision and explanation

Escalate the coverage and final-action gaps, apply the approved restrictions and verify the complete credited function.

Calibration addresses only a limited part of the function. The changed release path and unverified final action remain significant gaps. The review separates sensor health from coverage and effective response. Each needs appropriate technical and test evidence. Relocation changes the detection design. Competent assessment must consider the source, atmosphere, response and all connected functions before approving a new position.

Look closer

Process operators monitoring equipment from a UK industrial control room. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Initiating scenario: A release can create fire, toxic exposure or escalation, depending on the actual substance and process. Protective function fails: A detector, logic, power supply or final element may not deliver its required action. Possible harm: Delayed detection or isolation can leave people exposed and allow escalation; determine consequences through the actual scenario assessment.
01 · What to notice

The VRU has a gas detector. The case provides a calibration record but no review of siting, target substances, response, failure signals or linked actions.

02 · Trace the harm

A healthy sensor can miss the credible release if poorly located or selected; a detected condition can still escalate if the required action fails.

03 · Test the safeguard

Check coverage justification and suitable functional tests through alarm and final action, including maintenance, impairment and restoration evidence.

Lesson 9

Utility assurance: quantity, quality, duration and safe recovery

For every utility ask: what must arrive, where, in what condition, for how long—and what happens when it fails or returns?

Understand this fuel-industry task

The utility indicators show that services are switched on, but safety-critical users need more than an on/off state. Instrument air may have to reach the right pressure and quality at a valve under demand, while reserve must last long enough for the intended safe action. Backup electrical supplies likewise need the required capacity, duration and changeover performance. Partial loss or contamination can be as important as total failure, and multiple safeguards can depend on the same apparently minor service.

Trace source, treatment, storage, distribution and the actual user, then define the safe response before reserve expires. A valve’s failure position and ability to complete its duty must come from the approved design, not an assumption that all valves close safely when air is lost. Returning air or power is a separate recovery task; it must not restart hazardous plant around open work or unresolved faults.

Welfare and process-water systems add another kind of duty. Water hygiene needs a competent risk-based programme for relevant organisms and exposure routes, with suitable system control, monitoring and corrective action. Where Legionella presents a risk, inhalation of contaminated aerosols is a relevant exposure route; potable-water quality and cross-connection control need their own appropriate checks. This is distinct from instrument-air functional safety, but both illustrate why a healthy supply light cannot establish the quality and performance that each user needs.

Trace the initiating event and loss of control

StepCase pathway
initiatingEventUtility supply partly fails, becomes contaminated or exhausts reserve while dependent functions are needed.
lossOfControlValves, controls or backup systems cannot complete their required duty, or unsafe water conditions remain unrecognised.
consequencesA process can lose containment or protection; unsafe restoration can energise work, and relevant water exposure can cause illness.

Choose effective safeguards

Control priorityPractical application
EliminateRemove unnecessary stored energy and hazardous service dependencies.
SubstituteUse a simpler or lower-hazard service where suitable.
EngineerProvide assessed safe states, separation and reliable backup for critical functions.
ManageMap dependencies, inspect pressure equipment and test loss and restoration of supply.
ProtectUse task-specific protection for electrical, pressure, heat and noise exposure.

Map each safety-critical source-to-user dependency and define required quantity, quality, duration, safe failure action and recovery authority.

Provide justified monitoring, reserve, backup and protection against shared failures, with alarms and procedures that allow safe action before reserve is exhausted.

Manage relevant water-hygiene risks through competent assessment, control, monitoring and prompt corrective action rather than a generic green supply indication.

Control maintenance and restoration across dependent users, verifying the cause is corrected and preventing automatic unsafe restart when a service returns.

Verify effectiveness before release

Transfer the decision to the workplace

Use this question set for firewater, tank-level protection power and customer-site pumping. A utility’s importance is determined by the functions and people that depend on it.

Which services need the highest assurance?

Classify by credible harm and the safety function served, not by the utility name alone.

A compressor serving office tools is not equivalent to instrument air needed for a protective valve. A fire-water supply is only a credited barrier if it can deliver its required duty when demanded. Engineering sets site-specific capacity, autonomy, quality and test criteria; this page does not invent pressures, durations or test intervals.

Service and contextPotential critical failureAssurance question
Power to control and detectionLoss of alarms, trips or visibility.Does backup transfer and carry all credited loads? Source: HSE: reliability of utilities, HSE: control systems technical measures
Instrument air to safety actuatorsValve cannot reach required safe position.Is actual fail action tested at the equipment? Source: HSE: reliability of utilities, HSE: control systems technical measures
Cooling for heat-generating processTemperature rise or loss of containment.What is time to limit and tested recovery? Source: HSE: reliability of utilities
Nitrogen for inertingAir ingress or oxygen-deficient release.What confirms composition, isolation and safe access? Source: HSE: inerting
Fire waterMitigation unavailable during incident.Are flow, pressure and route available under demand? Source: HSE: reliability of utilities

Applied case

Instrument air has returned after a supply problem. The panel is green, but air quality, dependent valve status and maintenance handback have not been checked.

Decision and explanation

Use the approved recovery process to verify quality, dependent functions, work clearance and authorised safe restart.

A restored service may carry contamination or leave equipment in an unsafe configuration. Readiness includes users, work status and verified protection. Fail actions and reserve depend on the actual design and demand. A common assumed valve response cannot replace user-specific verification. This establishes what the service delivers and what the users will do, rather than equating a green source indication with safe plant readiness.

Look closer

Air receiver, relief fittings and protected compressed-gas storage. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Remove exposure: Stop or redesign the affected operation while required protection is unavailable. Verify the actual operating boundary. Reduce demand: Consider a safer process or lower hazard with competent design input; document the limitations. Separate and engineer: Design suitable separation and isolation; examine shared power, headers and controls before claiming independence. Verify the function: Test the complete specified function and actual final-element state through the approved method. Remaining exposure: Record restrictions and suitable residual protection; PPE cannot restore a failed supply or trip.
01 · What to notice

The utility exercise links instrument air, backup power and welfare water to users, but checks only whether each service is currently switched on.

02 · Trace the harm

Partial loss, contamination or reserve depletion can disable dependent valves and controls; poorly managed water systems can also expose people to harmful organisms.

03 · Test the safeguard

Check actual dependent users and relevant quality, demand and failure tests, plus water-hygiene controls; verify recovery does not automatically restart unsafe plant.

Lesson 10

Can receipt resume with the evidence provided?

Make a supported availability decision across linked protection claims.

Bring the evidence together

Investigate five fictional protection claims, from a boundary emergency stop to tank-level trips, vapour recovery and utility dependencies. Practise asking whether a barrier can meet its intended duty, rather than whether an indicator is green. This is supervisory awareness; engineering design and integrity allocation remain competent specialist work.

Consider the whole sequence of interfaces and degraded conditions. A local repair or completed permit cannot by itself show that every required barrier is available. Separate immediate protection, the competent technical resolution, the operating decision and independent verification.

Use the worked cases and the linked primary guidance to identify the missing information. State who owns the decision and what would keep the operation restricted. The fictional scene illustrates choices; it cannot establish actual process limits, emergency capability or permission to operate.

What would a full safety critical performance standard look like?

A full example standard makes the barrier expectation testable for operations, maintenance, engineers and managers.

For learners, the important shift is from naming a barrier to defining the standard it must meet. A high-high level trip is not simply present or absent. It has a purpose, set point, response time, final element, owner, test method, proof interval, impairment rule, bypass authority and evidence trail. HSE L65 offshore fire, explosion and emergency response guidance.

The exact figures must come from the site design basis and competent engineering judgement. The example below shows the structure, not a universal set point.

Performance standard fieldExample contentEvidence expected
PurposePrevent tank overfill during receipt or transfer by detecting high-high level and stopping inflow before loss of containment.Scenario register, HAZOP/bowtie link and design basis.
Performance requirementSensor, logic solver and final valve act together to close inflow within the defined response time at the agreed high-high set point.Cause and effect chart, set point register and loop test procedure.
OwnerAsset owner owns risk; technical authority owns engineering standard; operations owns daily availability; maintenance owns proof testing and defects.RACI, asset register and maintenance plan.
Testing methodProof test checks sensor, alarm, logic, final element, valve closure and operator indication, not only the instrument reading.Approved proof-test procedure and completed test sheet.
Proof intervalInterval is justified by risk assessment, reliability data, manufacturer guidance, failure history and the credited risk reduction.Proof-test schedule and interval basis.
Impairment ruleTank filling is restricted or stopped if the trip is failed, overdue, bypassed or cannot be proven before receipt.Impairment assessment and operating restriction.
Bypass authorisationTime-limited bypass requires operations manager, technical authority and process safety review with compensating controls.Bypass log, authorisation and handover note.
EscalationFailed proof test, repeated nuisance trip, open bypass, changed set point or changed operating envelope triggers MOC or senior review.MOC record, defect escalation and restart approval.
Evidence packDesign basis, performance standard, C&E, set point record, proof-test evidence, defect close-out, bypass history and MOC trail.Retrievable barrier file and dashboard entry.

A device exercise is not automatically a proof test of the credited complete function. Where demand-condition testing is unsafe or incomplete, the approved method must state coverage, untested dangerous failures, alternative evidence and residual uncertainty. Retain as-found/as-left results and restoration checks; a partial-stroke exercise alone does not demonstrate the full safety function.HSE COMAH control systems

Applied case

The receipt push button passes a lamp check; TK-102 still credits common sensing as independent protection; the VRU detector has calibration but no coverage assessment; instrument-air reserve is unknown. A supervisor asks whether the next receipt may start. Produce a concise claim/evidence/gap/restriction table and identify the competent owners. Do not count equipment labels as multiple barriers or create a temporary bypass yourself.

Decision and explanation

Do not credit the unproved functions; apply the authorised safe-state and impairment decision until competent evidence supports the required protection.

The appropriate limits depend on the actual risk and authority. Smaller inventory or closer observation cannot automatically replace unproved safety functions.

Look closer

Process operators monitoring equipment from a UK industrial control room. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Function: High-high trip detects level at defined set point and closes import valve within required time. Independence: Trip independence from normal level control is defined and verified. Testing: Proof test covers sensor, logic, final element, alarm and valve response.
01 · Read the situation

The receipt push button passes a lamp check; TK-102 still credits common sensing as independent protection; the VRU detector has calibration but no coverage assessment; instrument-air reserve is unknown. A supervisor asks whether the next receipt may start. Produce a concise claim/evidence/gap/restriction table and identify the competent owners. Do not count equipment labels as multiple barriers or create a temporary bypass yourself.

02 · Explain the principle

Investigate five fictional protection claims, from a boundary emergency stop to tank-level trips, vapour recovery and utility dependencies. Practise asking whether a barrier can meet its intended duty, rather than whether an indicator is green. This is supervisory awareness; engineering design and integrity allocation remain competent specialist work.

03 · Carry it into work

The appropriate limits depend on the actual risk and authority. Smaller inventory or closer observation cannot automatically replace unproved safety functions.

Lesson 11

Turn the learning into a verified workplace action

Choose a bounded action, owner, task-specific evidence and escalation trigger within your role.

Make the course useful at work

Awareness and supervisory assurance of terminal protective systems; no functional-safety design, SIL assignment or bypass authorisation.

Choose one relevant task from this course. Describe the material, equipment, people and interfaces, then explain the credible harm pathway before choosing the control. Follow the actual site process for obtaining missing technical evidence and deciding whether work may proceed.

Use the linked resource with a specific question. A briefing should test understanding; a checklist should identify missing facts; an audit should compare claims with observations. Keep this exercise fictional and separate from live operating records.

Record the planned action, responsible role, verification evidence and a trigger for escalation or reassessment. Completion of this course is a learning record, not an operating instruction, engineering design, specialist licence or authority to enter, isolate or fight a fire.

What evidence should prove a safety critical element is healthy?

Evidence should show design basis, performance criteria, maintenance, proof testing, impairment control and management review.

A performance standard is only as good as the evidence behind it. A certificate or inspection record may show something was checked, but it does not always prove the right thing was checked against the right scenario. HSE offshore KP3 asset integrity report.

For major hazards, evidence should be traceable from the scenario through the credited barrier and into the live maintenance, inspection, operations and assurance systems.

Evidence itemWhat it provesWeak signal
Major accident scenario referenceThe barrier is linked to a credible hazard.Barrier exists with no scenario owner.
Performance standardFunction, criteria and verification route are defined.Standard uses vague words such as adequate or available.
Design basisCapacity, set point, rating or response time is justified.Design basis cannot be found or has changed.
Proof-test procedureThe test checks the actual safety function.Test only checks part of the loop.
Inspection and maintenance recordTasks are completed and defects managed.Overdue tasks or repeat failures occur.
Bypass and impairment logUnavailable barriers are authorised and risk assessed.Bypasses are open-ended or poorly justified.
MOC recordChanges to plant, software, alarms, set points or procedures are reviewed.Change occurs without reassessing the performance standard.

Applied case

Your handover action says improve fuel safety but names no operation, control owner, evidence or stopping condition.

Decision and explanation

Name the actual task, competent control owner, evidence required and the hold/escalation trigger.

An effective plan connects a real task and control with a responsible role, objective evidence and a clear decision boundary. A course badge or signature does not establish the plant condition.

Look closer

Process operators monitoring equipment from a UK industrial control room. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Tank high-high trip bypassed during receipt: The level rises past the safe limit and liquid spills into a flammable-vapour area; ignition can injure workers and the public. Fire-water pump or utility unavailable: A fire grows while mitigation lacks pressure or supply; people cannot rely on the planned response. Common-cause failure of detection and shutdown: A shared power or instrument-air failure defeats more than one credited layer when demand occurs.
01 · Read the situation

Your handover action says improve fuel safety but names no operation, control owner, evidence or stopping condition.

02 · Explain the principle

Awareness and supervisory assurance of terminal protective systems; no functional-safety design, SIL assignment or bypass authorisation.

03 · Carry it into work

An effective plan connects a real task and control with a responsible role, objective evidence and a clear decision boundary. A course badge or signature does not establish the plant condition.