FMEA, or Failure Mode and Effects Analysis, is a systematic analysis of how an item or process may fail to perform its function, which effects that failure produces, what may cause it, and which controls already exist. In industrial maintenance, the method turns operational, maintenance, and engineering knowledge into actions addressing specific failure modes. Scoring supports the decision, but it does not replace safety criteria, legal requirements, asset criticality, or documented engineering judgment.
What FMEA is and when to use it
FMEA is used before failure or when reviewing an existing strategy. It organizes functions, functional failures, failure modes, effects, causes, and controls to identify prevention and detection gaps. Root cause analysis starts from an event that has occurred and seeks to explain why it happened. The two methods can complement each other, but they answer different questions.
IEC 60812 presents FMEA as a method applicable to hardware, software, processes, human actions, and interfaces. When the analysis includes a formal criticality assessment, it may be treated as FMECA. The name used, the structure of the scales, and the prioritization criteria must follow the procedure defined by the organization or the technical reference it has selected.
Start with the function and operating context
A component list is not enough. Before discussing ratings, the team must record what the asset is expected to do, under which conditions, and to what performance standard. A pump may need to deliver a specified flow and pressure, but the effect of losing that function changes with the fluid, available redundancy, process stage, and ability to isolate the equipment.
This context prevents two common errors: copying an FMEA from similar equipment as if the risk were identical, and treating every physical defect as a failure mode without relating it to the function. The scope also needs a clear boundary, the versions of the documents used, and the people accountable for the analysis.
Severity, occurrence, and detection depend on the selected scales
Many procedures use three factors known as severity, occurrence, and detection. Scales from 1 to 10 are common, but they are not universal. The organization must define in writing what each level means, which evidence is required, and how the scale is to be used. A rating cannot be safely transferred between plants or methods that use different descriptions.
- Severity: represents the consequence of the effect for people, the environment, operations, quality, property, or compliance, according to the categories defined by the organization.
- Occurrence: represents the likelihood or frequency of the failure mode or its cause within the selected horizon. History, comparable populations, and changes in operating regime must be considered.
- Detection: represents the ability of current controls to identify the cause or failure mode in time to prevent or reduce the effect. Having a sensor or inspection does not ensure strong detection if coverage, frequency, threshold, and response are inadequate.
The team should record the rationale for every evaluation. A rating without evidence becomes an opinion that is difficult to review and prevents the next update from showing what actually changed.
RPN is a decision aid, not an automatic work queue
When the selected method calculates a Risk Priority Number, or RPN, the value is commonly the product of severity, occurrence, and detection. That formula only makes sense with the scales and rules defined by the organization. The result can order part of the discussion, but it has important limitations.
- Different combinations can produce the same RPN.
- A failure mode with severe consequences can have a lower RPN because occurrence is low or detection is strong.
- Values from analyses built with different scales are not directly comparable.
- Reducing a rating in the worksheet does not reduce risk if the corresponding control has not been implemented and verified.
A serious safety, environmental, or compliance risk should not wait for its place in the ranking. The organization needs separate triggers for these cases, even when the RPN is not among the highest. Legal requirements and risk acceptance criteria take precedence over the multiplication.
RPN and Action Priority are not the same
The AIAG-VDA automotive handbook uses Action Priority, or AP, to guide the need for action based on the combination of severity, occurrence, and detection. AP is not the product of the three ratings and cannot be converted directly into RPN. It classifies combinations according to the logic and tables of the selected method.
This does not make AP a universal requirement for every maintenance FMEA. An organization may use RPN, AP, a criticality matrix, severity criteria, or a documented combination, as long as it states which method governs the analysis. This summary explains the conceptual difference without reproducing tables or requirements from the AIAG-VDA handbook. Organizations adopting that reference need the licensed edition and the applicable internal procedure.
How to apply FMEA step by step
- Define the objective, scope, and boundary. Record the system under analysis, operating context, input documents, and exclusions.
- Build a team that knows the asset. Operations, maintenance, engineering, safety, and process specialists participate according to the possible effects.
- Describe functions and performance standards. State what the asset must deliver before listing how it can fail.
- Identify functional failures and failure modes. Separate the loss of function from the physical or logical way in which it occurs.
- Record effects, causes, and current controls. Distinguish controls that prevent from those that only detect.
- Apply the selected scales and criteria. Justify each evaluation with history, inspection, requirements, or identified engineering knowledge.
- Prioritize with the complete rule set. Address safety and compliance triggers first, then use RPN, AP, or another defined criterion to organize the remaining actions.
- Define actions addressing causes and controls. Every action needs an owner, due date, completion criterion, and effectiveness check.
- Reassess and version the analysis. Update controls, rationale, and residual risk after implementation without erasing the decision history.
Worked example: centrifugal process pump
The example below uses a hypothetical internal scale from 1 to 10 and calculates RPN only to demonstrate the mechanics. The values are not benchmarks and should not be copied into a real operation.
Mode 1: bearing degradation associated with inadequate lubrication
- Affected function: maintain rotation and transmit power with stability.
- Effect: increased vibration, loss of availability, and possible secondary shaft damage.
- Current control: a periodic route without a consolidated condition trend.
- Illustrative evaluation: severity 7, occurrence 5, detection 6, RPN 210.
- Candidate action: review lubrication practice and specification, confirm the dominant cause, and define monitoring capable of detecting degradation within the available response window.
Mode 2: seal wear with loss of containment
- Affected function: contain the fluid during operation.
- Effect: leakage, unavailability, and consequences that vary with product properties and the affected area.
- Current control: visual inspection and leakage monitoring.
- Evaluation: must use process data. If the consequence for people or the environment is severe, the severity trigger requires treatment even when occurrence and detection keep the RPN below another item.
- Candidate action: verify the dominant mechanism, operating condition, alignment, seal plan, and containment barriers before selecting the task.
Mode 3: impeller erosion caused by cavitation
- Affected function: deliver the required flow and pressure.
- Effect: performance loss, vibration, and progressive damage.
- Current control: periodic inspection and operating readings without a validated alarm for the condition.
- Illustrative evaluation: severity 6, occurrence 4, detection 7, RPN 168.
- Candidate action: review suction conditions, operating range, instrumentation, and response criteria.
The example shows why the team should not simply execute the first item in the RPN column. The bearing may lead the operational ranking, while a severe loss of containment is addressed first under a safety criterion. The final decision combines consequence, evidence, controls, obligations, and the ability to reduce risk.
From analysis to the maintenance plan
An FMEA action does not automatically become a time-based preventive task. Depending on the failure mode and consequence, the response may be a design change, an operating change, a condition-based task, a functional test, an inspection, contingency stock, or a documented run-to-failure decision.
After the engineering decision, the preventive maintenance plan template in Portuguese can help structure tasks that are genuinely applicable. In an SAP environment, strategy and frequency remain in the maintenance plan. PM Run supports scheduling and execution of the resulting work orders without replacing the risk analysis that defined the work.
FMEA within reliability engineering
FMEA becomes more valuable when it is connected with the asset criticality matrix, failure history, causal analysis, inspections, and reliability engineering. The analysis should be reviewed when the function, process, material, operating regime, control, or knowledge of the failure changes.
For deeper discussions with maintenance planning and reliability professionals, the PM Run Advanced Study Group, GEA, holds invitation-only sessions focused on customers and partners. Customers and partners can request an invitation from their account executive. Companies that are not yet customers can contact the commercial team to learn about the initiative.
Technical references
- IEC 60812:2018, Failure modes and effects analysis, the official page with the standard's public scope and summary.
- NASA GSFC-HDBK-8004, a public guide to FMEA, FMECA, and risk assessment as a living document.
- AIAG-VDA FMEA Handbook, the official catalog for the automotive reference. The complete tables and rules require the licensed publication.
Frequently asked questions
What does FMEA mean?
FMEA means Failure Mode and Effects Analysis. The method identifies how an item may fail, its effects, its causes, and the associated controls.
Do FMEA ratings always range from 1 to 10?
No. Scales from 1 to 10 are common, but the range, descriptions, and prioritization method depend on the selected procedure or reference. The criteria must be defined, applied consistently, and supported by a recorded rationale.
How is RPN calculated?
When the selected method uses RPN, it commonly multiplies severity, occurrence, and detection. The result should not be interpreted in isolation or compared across analyses that use different scales.
What is the difference between RPN and Action Priority?
RPN is a numeric product. Action Priority is a classification obtained by combining the evaluations according to the tables in the AIAG-VDA method. There is no direct conversion between the two, and AP only governs the analysis when that reference has been adopted.
Can a safety risk with a low RPN wait?
It should not wait merely because the numeric product is low. Severe safety, environmental, or compliance consequences must follow the triggers and acceptance criteria defined by the organization and applicable law.
Who should participate in an FMEA?
The team needs operational, maintenance, and engineering knowledge. Safety, process, quality, and other specialists participate according to the scope and possible effects.
