What is the difference between FMEA and FMECA?
Failure Modes and Effects Analysis (FMEA) identifies potential failure modes and describes their effects. Failure Modes, Effects and Criticality Analysis (FMECA) builds on that analysis by explicitly evaluating failure-mode criticality for prioritization, at minimum considering consequence severity and often failure occurrence or other context. A numeric Risk Priority Number is one possible local method, not a universal requirement or a substitute for safety or regulatory obligations.
An FMEA explains failure behavior. An FMECA helps prioritize what to address.
Start with functions, functional failures, credible failure modes, local and system effects, detection, operating context, and current controls. Criticality analysis then evaluates how important each failure mode is to a defined consequence, using severity, likelihood or observed frequency where credible, and other explicitly defined factors.
IEC 60812:2018 describes both FMEA and the FMECA variant. Neither method grants authority to change safety functions, inspection intervals, or operating limits without the responsible engineering and regulatory review.
Create a repeatable criticality method before anyone begins scoring.
- Define one analysis boundary, function, duty, operating context, and exposure unit.
- Record failure-mode specificity and source: actual history, OEM analysis, expert judgment, or unknown.
- Define severity criteria by consequence category, including people, environment, regulatory, quality, production, and repair.
- Define occurrence basis separately: events per operating hour or cycle if reliable, or a labeled qualitative estimate where data is sparse.
- Document preventability, detectability, existing controls, uncertainty, and uncertainty owner.
- Use consequence-first screening for intolerable modes; prioritize remaining modes using the documented ranking method.
- Assign an action, policy owner, approval path, and review trigger, then confirm the change has been implemented.
Worked example: a conveyor line has three very different failure modes.
The following is a fictional screening example, not an engineered risk assessment. The frequency labels are deliberately qualitative because no validated exposure data is provided.
Do not compare frequency labels across assets with different operating exposures unless the denominator and boundaries match. A high-frequency nuisance event and a very-low-frequency safety event do not become equivalent because of a single arithmetic score.
| Failure mode | Consequence | Frequency evidence | Action priority |
|---|---|---|---|
| Guard interlock fails to detect an open guard | Potential safety function loss; governing safety controls apply | Not established | Immediate qualified safety/engineering assessment; do not use an RPN to downgrade. |
| Drive bearing degradation causes conveyor trip | Production downtime and possible equipment damage | Several relevant CMMS events, coding quality to verify | Evaluate load/alignment/lubrication failure modes and technically applicable monitoring. |
| Indicator lamp fails | Reduced local visibility; consequence depends on function | Sparse data | Determine whether an alternate indication and approved operating controls make corrective-on-failure acceptable. |
Severity, occurrence, and detection rankings are not physical probabilities.
An RPN calculated as severity × occurrence × detection can be a useful internal discussion aid if its scales are controlled, but multiplication of ordinal rankings does not yield an actual expected loss or universally comparable risk. Different triplets can have the same product despite radically different consequences.
Preserve each dimension separately and use explicit consequence gates before optimization. Unknown occurrence is not zero occurrence; a missing failure code is a data-quality problem, not evidence of a reliable asset.
Every significant FMECA row should produce a governed decision.
The useful deliverable is a failure-mode register linking function, effect, consequence, source confidence, existing controls, selected policy, action owner, evidence, and review date. Tasks might involve condition monitoring, failure finding, spares, engineering modification, operating limits, work instructions, or accepted run-to-failure where permitted.
Update the analysis when new work orders, condition findings, a process change, an OEM notice, or a confirmed RCA invalidates the assumption. Keep the previous assessment traceable instead of silently overwriting it.
What makes the FMECA actionable?
A defensible severity definition, transparent occurrence evidence, consequence gates that cannot be averaged away, and specific actions owned by maintenance, operations, engineering, or safety.
- Use FMEA to identify and describe credible failure modes.
- Use criticality analysis to prioritize with consequence explicit.
- Do not interpret a composite ordinal score as probability or expected dollars.
- Convert decisions into controlled tasks or engineering actions.
Technical references
- IEC 60812:2018, FMEA and FMECA — Defines systematic FMEA and criticality-oriented FMECA practices.
- SAE JA1011_202411, Evaluation Criteria for RCM — Current RCM process-evaluation standard, revised November 2024.
Make the top-priority modes actionable.
Preserve consequence, evidence strength and the decision owner so criticality changes actual maintenance strategy instead of becoming a frozen spreadsheet.
RCM learning path See facility-specific XRVE services