How do you select maintenance tasks during an RCM analysis?
For each credible failure mode, identify its operating consequence and determine whether a proactive task is technically applicable and worthwhile under the governing RCM decision process. Candidates include on-condition monitoring, scheduled restoration or replacement, and failure-finding for hidden functions. Where no effective task exists, an authorized default action may involve redesign, operating controls or corrective maintenance, subject to safety, environmental and other mandatory requirements.
Lock the function, failure mode and consequence before discussing frequency.
Define the required function, failure mode, local and system effects, how failure becomes evident, safety or environmental exposure, quality and production consequences, and governing technical authority. A recurring malfunction and a hidden safety-system failure are not interchangeable categories.
Use SAE JA1011 evaluation logic to test whether the analysis is truly function and consequence driven. An existing calendar PM is a proposed control to evaluate, not automatic evidence that it is correct.
Each policy requires its own evidence of technical applicability.
Choose at the failure-mode level. The same pump or conveyor may have bearings monitored by condition, a hidden standby-start function subjected to authorized failure-finding, and a noncritical indicator repaired when it fails.
| Failure behavior | Candidate policy | Technical applicability test | Implementation evidence |
|---|---|---|---|
| Progressive detectable deterioration | On-condition task / condition monitoring | Reliable precursor exists; detection-and-response time beats functional failure | Method, acceptance trigger, route capability, linked corrective work |
| Age- or cycle-related mode | Scheduled restoration or replacement | Defensible age relationship or controlling requirement; intervention reduces consequence | Engineering/OEM basis, age exposure, work standard, review trigger |
| Hidden protective or standby function | Failure-finding test | Task can reveal hidden failure without unacceptable test risk | Authorized test procedure, coverage, demand basis, approval, defect response |
| High consequence with no effective task | Redesign or other controlled action | Proactive maintenance cannot meet required protection | Qualified engineering/safety decision, change control and validation |
| Evident, tolerable consequence | Deliberate corrective maintenance | Downtime, restoration, spares and risk are accepted by authority | Documented acceptance, contingency plan and review condition |
One functional system, three distinct decisions.
Fictional cooling skid: a running pump bearing has detectable progressive wear, the standby pump's automatic start function can fail silently, and a local noncritical running lamp can burn out while redundant verified indication remains available.
The bearing could be a condition-monitoring candidate only if a validated signal and response window exist. The standby auto-start function is a hidden-function issue, requiring appropriate authorized failure-finding analysis and test governance. The lamp might be managed by corrective maintenance if engineering and operations accept its actual consequence. No single PM frequency addresses these three cases.
A critical safety or environmental function can require redesign or mandatory controls even when no economically attractive proactive task is available. A spreadsheet cannot waive mandatory protective requirements.
Write an executable task with observation and action criteria.
Every selected task needs a precise purpose and a finding standard: what to inspect or measure; instrument and competence; approved safe access and isolation; permitted trigger or acceptance range from engineering or OEM; frequency justification; evidence to capture; and the corrective work created when abnormal.
Bad PM: 'Check bearing condition monthly.' Better structure: 'At the engineering-approved interval, collect the specified bearing condition data using the approved procedure, compare with approved limits, log the measurement and initiate corrective work on an actionable alarm.' Only plant authority can insert actual technical values.
Protective-function failure-finding needs special governance because tests can themselves create risk or disturb protection. Do not infer proof-test intervals, bypass permissions or acceptance values from an educational article.
Optimize labor only after technical effectiveness and consequence screening.
Estimate inspection labor, false alarms, missed-event burden, required spares, work access, corrective lead time and expected consequence. Compare strategies on a compatible basis, accounting for uncertainty and operating exposure.
PMs that repeatedly find nothing can still be necessary where a hidden consequence is high; conversely a frequently executed task may be ineffective if it never detects the mode it is claimed to control. Use evidence, not compliance alone, to justify deletion or redesign.
Carry the RCM rationale into CMMS and review.
Preserve function, mode, consequence, task type, reason for technical applicability, evidence confidence, governing authority, asset/PM IDs, owner, acceptance criteria, feedback requirement and material-change trigger.
After implementation, compare actual failures and findings with the original assumptions. If a mode recurs, distinguish a missing failure mode, weak detection, wrong task, insufficient response capacity, execution deficiency or changed context before changing its interval.
What is a legitimate RCM task-selection decision?
A failure-mode-specific policy that is technically applicable, consequence-appropriate, authorized, executable in the real facility and connected to feedback. Where these criteria cannot be met, record the missing information and the responsible technical decision rather than inventing a PM.
- Consequence and governing requirements come before task economics.
- Test detection capability and response before choosing CBM.
- Use failure finding only with approved safety and engineering procedures.
- Document accepted corrective maintenance rather than silently doing nothing.
Technical references
- SAE JA1011_202411: Evaluation Criteria for RCM — Evaluation criteria for genuine RCM processes.
- NASA Reliability-Centered Maintenance Guide — Failure-based maintenance strategies, inspection principles and program review.
Make every scheduled task defend its purpose.
Connect functions and failure modes to precise, governed work packages and observable effectiveness.
RCM learning path See facility-specific XRVE services