How do you facilitate a maintenance FMEA workshop?
Begin with a bounded asset and its required functions, agree on the functional-failure definitions, identify technically credible failure modes, document effects and consequences, assess existing prevention and detection controls, and assign evidence-backed reliability actions. A useful workshop leaves a controlled register with owners, task justifications, uncertainty and review triggers, not just severity × occurrence × detection scores.
Select a system boundary and bring the records that can resolve disagreements.
Fictional example: a packaging conveyor CV-21 must move specified products at a defined rate while satisfying engineered guarding and safety functions. The equipment boundary includes motor, VFD, gearbox, coupling, bearings, belt, sensors and relevant controls. Exact capacity, acceptance values and protective requirements must be obtained from site-approved design information.
Invite operators, mechanical and electrical technicians, a planner, controls or engineering personnel and someone accountable for safety or quality where those functions are involved. Bring CMMS failures, alarm history, equipment drawings, OEM instructions, existing PMs, lubrication records, spare lead times and defect photographs.
Write at the level where effects and controls can differ.
A mode is not the same as a symptom or the failed function. 'No belt motion' is a functional failure; 'gearbox output bearing seizes from contamination' is a candidate mode; 'high motor current' may be a symptom. Too generic a row produces an unhelpful universal inspection.
Use the following illustrative rows. They are hypotheses for a fictional plant, not validated design findings or specified protective measures.
| Function & mode | Effects / consequence | Existing-control question | Candidate action |
|---|---|---|---|
| Convey product; gearbox bearing degrades due to contamination | Heat, vibration and eventual loss of belt motion; production interruption | Can contamination be excluded, detected and corrected in useful time? | Inspect failure evidence, validate seal suitability and lubrication/condition strategy. |
| Maintain belt tracking; tracking wanders from worn idler or misalignment | Spillage, edge damage and unscheduled stop; quality/availability loss | Are acceptance limits and detection methods adequate? | Engineer approved tracking standard, job plan and defect-trigger thresholds. |
| Detect guarded access; interlock fails dangerously | Potential failure of protective function; safety consequence | Is the protective system validated and tested under governing requirements? | Qualified functional-safety/engineering review; do not improvise test or bypass. |
| Sense product presence; photoeye fouls | False stop or missed count; throughput/quality consequence | Is contamination evident and cleaning effective? | Improve approved cleaning/inspection criteria or modify shielding where supported. |
Treat consequence and technical feasibility separately from scoring.
The guarding mode requires applicable safety standards and site engineering authority; a low observed event count or composite RPN does not make it acceptable to defer. A nuisance photoeye fouling mode might be managed through cleaning, shielding or controlled corrective maintenance depending on operational consequences.
Before proposing vibration monitoring for a gearbox bearing, establish detectable degradation behavior, route capability and a work-response path. Before choosing calendar replacement, establish a relevant age relationship or governing requirement. Not every row needs another PM.
A maintenance FMEA can identify a required engineering decision without itself authorizing safety changes, intervals, test procedures or protection-system modifications.
Convert the selected action into a traceable maintenance implementation package.
Each accepted action should include an asset tag, failure-mode ID, function, operating context, effects, consequence, supporting evidence, control effectiveness, proposed task or change, approval authority, owner, due date, linked work order, and a review trigger.
For example, a confirmed lubrication ingress mechanism may call for a seal redesign, contamination-control procedure and appropriate condition checks. Any inspection needs an actionable finding definition and an escalation to planned corrective work, not a checkbox labeled 'inspect gearbox'.
Return operating experience to the FMEA.
Compare events per operating hour, confirmed repeat mode, severity, finding-to-corrective conversion, and failed-control observations. Require meaningful exposure before declaring an action successful. Check for changed materials, new products or new duty cycles that invalidate original assumptions.
Where frequency data are sparse, preserve unknown occurrence and the engineering rationale. The next significant failure or approved change should trigger revision to the specific rows concerned, not an indiscriminate annual rewrite.
What should the facilitator deliver?
A version-controlled failure-mode register with defensible consequence statements, evidence confidence, existing controls, owned proposed actions, approval path and CMMS implementation links. A completed workshop without task execution and follow-up is incomplete.
- Separate function, functional failure, failure mode and symptom.
- Use site-specific evidence and safety authority.
- Select controls appropriate to each mode, not one PM per row.
- Assign implementation and effectiveness owners.
Technical references
- IEC 60812:2018, FMEA and FMECA — Failure modes, effects and criticality analysis methods.
- SAE JA1011_202411: Evaluation Criteria for RCM — Evaluation criteria for genuine RCM processes.
Make the workshop result useful in the CMMS.
Translate each accepted failure-mode decision into a controlled, executable improvement with verified follow-up.
RCM learning path See facility-specific XRVE services