A useful maintenance FMEA connects failure modes to real consequences and real maintenance decisions.

Start with required function, not a generic list of component failures.

The same component can require different maintenance in different service. A pump protecting a critical process constraint and an identical spare pump in noncritical service do not necessarily carry the same consequence or strategy.

Define what the asset or system must do, under what operating context, and what loss of function matters. This keeps the analysis tied to the reason the equipment exists instead of producing a catalog of theoretical defects.

Write failure modes at a level that can drive a decision.

Too broad and the analysis cannot select a useful control. Too narrow and the team creates hundreds of rows with no decision value. The right level is specific enough to connect the mechanism to evidence and action.

For a rolling-element bearing, bearing failure is usually too broad. Contamination-related damage, lubrication degradation, overload, misalignment, looseness, installation damage, or fatigue may require different prevention, detection, or redesign decisions.

Separate what happens locally from what it means to the operation.

Failure effect describes what occurs when the mode happens: leakage, heat, vibration, loss of flow, trip, degraded accuracy, loss of containment, or another observable outcome. Consequence describes why the failure matters to safety, environment, quality, throughput, mission, cost, or another business objective.

This distinction supports asset criticality and prevents high-frequency nuisance failures from automatically outranking rare but unacceptable consequences.

Existing controls must be technically capable of controlling the failure mode.

A PM task is not automatically a control merely because it exists. Ask whether the task can prevent, detect, or reveal the failure at a useful point, whether the interval or trigger has a defensible basis, and whether a finding reliably becomes corrective action.

Where maintenance cannot effectively manage the failure, the next decision may be redesign, operating control, protective function, spare strategy, or accepted run-to-failure under appropriate authority.

Use scoring to support judgment, not to manufacture certainty.

Severity, occurrence, detection, or risk-priority scoring can help structure discussion, but the ordinal numbers are not physics. Keep the underlying consequence, evidence, assumptions, and data confidence visible.

For maintenance use, observed work history, operating exposure, criticality, failure evidence, and control effectiveness are often more actionable than a single composite number.

Technical references

• SAE JA1011_202411 — Current SAE evaluation criteria for Reliability-Centered Maintenance processes.
• NASA Reliability-Centered Maintenance Guide — Connects FMEA, predictive inspection, maintainability, reliability analysis, and maintenance strategy.

Related XRVE decisions