Summary
In short
- The AIAG and VDA harmonised FMEA Handbook, first edition 2019, replaced the Risk Priority Number with Action Priority and introduced a seven-step method.
- RPN multiplies ordinal rating scales, which is mathematically unsound. Identical RPNs can describe situations with entirely different severity, and the number conceals which.
- Action Priority makes severity dominant, so a high-severity failure mode cannot fall below a threshold because occurrence is low and detection is good.
- AP gives an absolute priority for each item rather than a rank order, so a list where everything scores similarly still tells you what to act on.
- The threshold habit is the specific failure: acting on everything above an RPN of 100 systematically excludes low-frequency catastrophic failures, which is the category you least want excluded.
- The handbook also added FMEA-MSR, covering monitoring and system response to maintain a safe or compliant state during operation.
What it is
What it is
What is an equipment FMEA?
A structured analysis of how equipment can fail, what the effects of each failure would be, what causes it, what controls prevent or detect it, and what action is warranted. It produces a prioritised set of actions rather than a description of the equipment.
What is Action Priority?
The prioritisation method introduced in the AIAG and VDA FMEA Handbook published in June 2019, replacing the Risk Priority Number. Severity, occurrence and detection ratings are looked up in fixed tables producing a priority of High, Medium or Low, rather than being multiplied together.
When to use it
When to use it, and when not to
This analyses failure modes for equipment. The maintenance plan it informs sits separately.
Use it for
- Analysing failure modes for new, critical or problematic equipment
- Designing or revising a preventive maintenance regime from the failure modes identified
- Following repeated failures where the existing regime is not preventing them
- Introducing new equipment, before the maintenance strategy is set
- Assessing the consequences of failure for equipment whose criticality is unclear
Not for
- The preventive maintenance plan, which the FMEA informs
- Root cause analysis, which examines a failure that has occurred rather than modes that could
- The asset register and criticality assessment, which this may inform
- Process FMEA and design FMEA, which address the process and the product rather than the equipment
- Reliability-centred maintenance analysis, which uses FMEA output within a wider decision framework
Standards
What it is built against
FMEA methodology comes from standards and industry handbooks rather than from regulation.
| Clause | Requirement | Where it lands |
|---|---|---|
| AIAG-VDA FMEA Handbook 2019 | Seven-step method with Action Priority replacing the Risk Priority Number | Failure modes |
| AIAG-VDA step 5 | Risk analysis assigning severity, occurrence and detection and determining Action Priority | Result |
| AIAG-VDA step 6 | Optimisation: actions to reduce risk, with responsibility, target dates and effectiveness confirmation | Result |
| IEC 60812 | Failure modes and effects analysis procedure, the international standard for the technique | Header |
| SAE J1739 | Design, process and machinery FMEA guidance | Failure modes |
| IATF 16949 cl.6.1.2.1 | Risk analysis including lessons learned from recalls, field returns and nonconformities | Header |
| ISO 55001 cl.6.2.2 | Planning to achieve asset management objectives, informed by failure analysis | Result |
| SAE JA1011 | Reliability-centred maintenance criteria, within which FMEA output is used | Result |
What it does not cover
- The preventive maintenance plan, which this analysis informs.
- Root cause analysis, which examines failures that occurred rather than modes that could.
- Asset criticality assessment, which this may support.
- Process and design FMEA, addressing the process and the product.
- Reliability-centred maintenance decision logic, which consumes FMEA output.
Filling it in
Filling it in well
Analyse the failure mode rather than the component, use Action Priority, and require actions with owners.
Not the pump fails, but the seal leaks, the bearing seizes, the impeller erodes, the coupling fractures. Each has a different cause, a different effect, a different detection method and a different appropriate action. A failure mode written at component level produces an analysis that cannot generate a specific task.
The effect is what happens as a consequence: production stops, product is contaminated, someone is exposed. The cause is what produces the failure mode: contamination, misalignment, fatigue, incorrect assembly. Severity attaches to the effect and occurrence to the cause, and conflating them makes both ratings unstable between teams.
Look up severity, occurrence and detection to produce High, Medium or Low. This makes severity dominant and gives an absolute priority for each item rather than a rank against the rest of the list. Where an organisation retains RPN for tracking risk reduction over time, that is reasonable, and it should not be what decides what to fix.
Step six of the method is optimisation: actions taken, responsibility, target completion and confirmation that the action was effective, with the ratings reassessed afterwards. An FMEA producing a rated list and no actions has documented risk, which is the outcome most FMEAs actually reach.
Audit findings
Common audit findings
FMEA findings concentrate on method and on whether anything followed.
| Finding | Clause | What fixes it |
|---|---|---|
| RPN threshold used to decide what to act on. | AIAG-VDA 2019 | Use Action Priority; thresholds systematically exclude high-severity low-frequency modes. |
| Failure modes written at component rather than mode level. | IEC 60812 | Specify how it fails; a component-level entry cannot generate a task. |
| Effect and cause conflated, making ratings unstable. | AIAG-VDA step 4 | Severity attaches to the effect, occurrence to the cause. |
| Analysis produces ratings with no actions. | AIAG-VDA step 6 | Optimisation with owners, dates and effectiveness confirmation is part of the method. |
| Ratings not reassessed after actions were completed. | AIAG-VDA step 6 | Reassessment is what shows the action worked. |
| FMEA not revisited after failures the analysis did not anticipate. | IATF 16949 cl.6.1.2.1 | An unanticipated failure is direct evidence the analysis was incomplete. |
| Detection rated on whether a failure would eventually be noticed. | AIAG-VDA step 5 | Detection concerns controls detecting the cause or mode before the effect occurs. |
| Hidden failures in protective systems not analysed. | SAE JA1011 | Standby and protective functions fail unnoticed and need failure-finding tasks. |
| FMEA performed by engineering alone without operators or maintainers. | IEC 60812 | The people who see the failures hold the occurrence data. |
| Document produced for audit and never used to set the maintenance regime. | ISO 55001 cl.6.2.2 | The output should be the maintenance plan; otherwise it is a paperwork exercise. |
Worked case
Case in point: the mode that scored sixty
A site ran FMEAs on its critical equipment and acted on every failure mode with an RPN above 100. The approach was consistent, documented and defensible, and it generated a steady programme of maintenance improvements.
One entry described the failure of a rotating shaft guard retention. Severity was rated 10, because failure exposed a rotating shaft at working height. Occurrence was rated 2, because the fasteners were substantial and had never failed. Detection was rated 3, because a missing guard would be visible on inspection. The RPN was 60, and it fell below the threshold every year.
Under Action Priority, a severity of 10 returns a high or medium priority regardless of the other two ratings, because the method treats severity as dominant rather than as one of three equal factors.
Definitions
Definitions and key terms
- Failure mode
- The specific manner in which an item fails, distinct from the component that fails.
- Effect
- The consequence of the failure mode, to which severity is assigned.
- Cause
- What produces the failure mode, to which occurrence is assigned.
- Action Priority
- The AIAG-VDA prioritisation returning High, Medium or Low from lookup tables, replacing the Risk Priority Number.
- Risk Priority Number
- Severity times occurrence times detection, criticised for multiplying ordinal scales and superseded in 2019.
- Detection
- The rating for controls detecting the cause or failure mode before the effect is realised, not whether failure is eventually noticed.
- FMEA-MSR
- Monitoring and System Response, added in the 2019 handbook to address maintaining a safe or compliant state during operation.
- Optimisation
- Step six of the method: actions with owners and target dates, followed by reassessment of the ratings.
FAQ
Frequently asked questions
What replaced the Risk Priority Number?+
Action Priority, introduced in the AIAG and VDA FMEA Handbook published in June 2019. Severity, occurrence and detection are looked up in fixed tables producing High, Medium or Low rather than being multiplied. Some organisations retain RPN alongside it for tracking risk reduction over time, which is reasonable, but it is no longer the method for deciding what to act on.
What was wrong with RPN?+
It multiplies ordinal rating scales, which is mathematically unsound. An RPN of 200 arising from severity 10, occurrence 5 and detection 4 describes a very different situation from an RPN of 200 arising from severity 5, occurrence 8 and detection 5, and the number gives no indication which. The rating scales are ranks, not quantities, and multiplication treats them as quantities.
Why does a threshold approach fail?+
Because it lets high-severity modes fall below the line. A failure mode rated severity 10, occurrence 2, detection 3 produces an RPN of 60 and describes something rare, hard to detect and catastrophic. Under a threshold of 100 it is never actioned. Action Priority makes severity dominant precisely so that this cannot happen.
Who should be in the FMEA team?+
People who know how the equipment actually fails, which means operators and maintainers alongside engineers. Occurrence ratings depend on failure history that lives with the people who repair the machine, and detection ratings depend on what is actually noticed on shift rather than what a control is designed to catch.
What should the FMEA produce?+
Actions with owners, target dates and effectiveness confirmation, and a maintenance regime derived from the failure modes identified. An FMEA that produces a rated table and nothing else has documented risk without reducing it, and that is the outcome most reach because step six is treated as optional.
The agents
What the agents do with it
The analysis prioritises failure modes. What fails is an RPN threshold that filters out the severe ones and a table nobody acted on.
Holds failure modes against assets, applies Action Priority rather than an RPN threshold, and derives maintenance tasks from the modes identified.
Compares actual failures against the modes anticipated, so an unanticipated failure raises the analysis for revision.
Ensures high-severity modes with safety consequences carry priority regardless of occurrence and detection ratings.
Connects equipment failure modes to product quality and food safety effects, where severity ratings are frequently understated.
This template lives in KnowMaintain — asset maintenance. Work orders, planned maintenance, calibration, reliability and shutdowns.
Meet KnowMaintain→Sources
Sources
- AIAG and VDA FMEA Handbook, first edition, June 2019
- IEC 60812, failure modes and effects analysis procedure
- SAE J1739, potential failure mode and effects analysis
- SAE JA1011, evaluation criteria for reliability-centred maintenance processes
- IATF 16949:2016 clause 6.1.2.1, risk analysis