FMEA in maintenance: How to predict failures before they become a costly problem

FMEA, or Failure Mode and Effects Analysis, is a systematic method for identifying possible failure modes, their effects and causes, and for prioritising corrective and preventive actions. In manufacturing and maintenance, FMEA helps answer the question: what can fail, why, what will the effect be, and what needs to be done before the problem occurs in reality.

In practice, FMEA is not just a quality table. It is a proactive risk management tool that combines engineering knowledge, historical data, operator experience, maintenance information and an assessment of failure consequences. A well-executed FMEA supports decisions on condition monitoring, inspections, upgrades, preventive actions and maintenance priorities.

What Is FMEA?

FMEA is a risk analysis method that identifies potential failure modes, that is, the ways in which a product, process, system or component can stop performing its function.

The term failure mode analysis refers to one of the key elements of FMEA: recognising the ways in which a function can be lost. FMEA itself goes further, because it also analyses effects, causes, current controls, the level of risk and the actions required.

FMEA organises three basic questions:

1. What can fail? — the failure mode.
2. What happens if it fails? — the failure effect.
3. Why can it fail? — the failure cause.

The team then assesses the risk, usually on the basis of three factors:

• severity — how serious the effect of the failure will be;
• occurrence — how likely or how frequent the cause or failure mode is;
• detection — how likely it is that current controls will catch the problem before the effect occurs.

Traditionally, these three ratings are combined into the RPN, or Risk Priority Number:

RPN = severity × occurrence × detection

In the newer AIAG VDA FMEA approach from 2019, Action Priority is increasingly used as well, expressed as high, medium or low on the basis of the combination of severity, occurrence and detection.

The most important principle: FMEA is a preventive method. Its purpose is not to describe a failure after the fact, but to point out potential problems before they lead to a production stoppage, a customer complaint, a safety risk or costly repairs.

Where does FMEA come from and where is it used?

FMEA originates from military procedures developed in the United States in the 1940s. The method was later adopted by NASA, the automotive industry, aviation, manufacturing, energy, medicine, regulated sectors and reliability engineering.

A typical timeline of FMEA’s development looks as follows:

Period	Event	Significance
1949	MIL-P-1629	The beginning of a formal approach to failure mode, effects and criticality analysis
1960s	NASA	Use of FMEA in space programmes and high-reliability systems
1970s	Ford and the automotive industry	Extension of FMEA to product design and manufacturing processes
1993	AIAG FMEA Handbook	Standardisation of FMEA in the automotive industry
2008	SAE J1739	A standard approach to FMEA in the automotive context
2009	IEC 60812	International procedure for failure mode and effects analysis
2019	AIAG VDA FMEA Handbook	The seven-step FMEA process and the Action Priority method
Today	Many industries	Manufacturing, aviation, automotive, medical devices, energy, process industries and maintenance

In manufacturing, FMEA is used for product design, process planning, reducing quality risk, improving asset reliability, choosing a maintenance strategy and preparing preventive actions.

Why use FMEA in manufacturing and maintenance?

FMEA is used in order to recognise potential failures early, assess their effects and set priorities for preventive action.

In a manufacturing context, the method helps reduce:

• unplanned downtime;
• repeat failures;
• quality problems;
• process errors;
• safety risks;
• losses caused by non-conformities;
• excessive repair costs;
• ineffective maintenance activities;
• secondary component damage;
• misplaced priorities for technical work.

FMEA is particularly useful when a team has to decide which risks matter most. Not every potential failure requires the same response. A minor cosmetic defect should be treated differently from a failure that stops the line, and differently again from a problem that could affect people’s safety or regulatory compliance.

In maintenance, FMEA helps determine which failure modes should be monitored, which can be reduced through preventive maintenance, which require condition monitoring, and which should be eliminated by changing the design, the procedure or the operating conditions.

Types of FMEA: DFMEA, PFMEA, SFMEA and FMECA

The most commonly used types of FMEA relate to the design, the process and the system. FMECA is a further variant. They differ in the scope of the analysis, the point at which they are used and the type of input data required.

Type	Full name	Scope	When to use	Key input	Example
DFMEA	Design FMEA	Risks associated with the product design	Product design and development phase	Specifications, drawings, bill of materials (BOM)	A bearing seal may admit moisture, leading to corrosion and premature failure
PFMEA	Process FMEA	Risks associated with the manufacturing process	Process planning, launch, improvement	Process flow diagram, work instructions, operation parameters	A tightening step may deliver insufficient torque, leading to a loose joint and vibration
SFMEA	System FMEA	Interactions at system level	Concept, architecture or integration phase	System architecture, interfaces, dependencies	Loss of power in subsystem A causes loss of signal in subsystem B and a system stoppage
FMECA	Failure Mode, Effects and Criticality Analysis	FMEA extended with criticality analysis	Safety-critical systems, aviation, defence, nuclear power	FMEA results, failure rate data, criticality matrix	Quantitative assessment of the criticality of failure modes based on probability and effects

DFMEA — design FMEA

DFMEA, or Design FMEA, focuses on how a product can fail as a result of design decisions.

The analysis covers product functions, requirements, components, materials, interfaces, loads and operating conditions. The goal is to detect risks before the design is approved or production is launched.

Example: a design team analyses a bearing seal. The potential failure mode is moisture ingress. The failure effect is corrosion, increased running resistance and reduced service life. The cause may lie in an unsuitable seal material or inadequate protection in a given environment.

PFMEA — process FMEA

PFMEA, or Process FMEA, concerns the manufacturing process. It analyses how a specific operation, workstation, parameter, tool, control or instruction can lead to a defect or a failure.

PFMEA is particularly important in manufacturing because it links quality risk to the way work is actually carried out on the line.

Example: in an assembly process, a bolt should be tightened to 45 Nm ± 5%. The potential failure mode is insufficient tightening torque. The failure effect is a loose joint, increased vibration and the risk of component damage. The cause may be an out-of-calibration tool, a missing control or an unclear work instruction.

PFMEA can draw on DFMEA results, because the product design often indicates which functions and components are particularly sensitive to process errors.

SFMEA — system FMEA

SFMEA, or System FMEA, analyses risks at system level and in the interactions between system elements.

It is used where the failure of one module can affect other subsystems, safety functions, power supply, control, communication or the operation of the entire installation.

Example: loss of power in one subsystem causes loss of the sensor signal in another. The effect may be an incorrect control decision or a shutdown of the whole system.

FMECA — FMEA with criticality analysis

FMECA, or Failure Mode, Effects and Criticality Analysis, extends FMEA with a quantitative or semi-quantitative criticality analysis.

While FMEA identifies failure modes, effects, causes and actions, FMECA adds a criticality assessment, often using failure rate data, probability and a criticality matrix.

FMECA is used above all in environments where the consequences of failure are particularly significant: aviation, defence, energy, critical infrastructure, process industries, nuclear power and safety-critical systems.

The FMEA process step by step

The modern FMEA process can be described as seven steps: planning and preparation, structure analysis, function analysis, failure analysis, risk analysis, optimisation and results documentation.

Step	Name	Main activities	Result
1	Planning and Preparation	Defining the scope, the team, the type of FMEA, the schedule and the documentation	FMEA plan, team composition, scope boundaries
2	Structure Analysis	Defining the structure of the system, product or process	Block diagram, process flow diagram, functional structure
3	Function Analysis	Defining the functions and requirements for each element or step	List of functions linked to the structure
4	Failure Analysis	Identifying failure modes, effects and causes	Chain: cause → failure mode → effect
5	Risk Analysis	Assessing severity, occurrence and detection; determining RPN or Action Priority	S/O/D ratings, RPN or Action Priority
6	Optimisation	Defining actions for the highest-priority risks	Action plan, owners, deadlines
7	Results Documentation	Updating the worksheet, reassessing risk after actions	Updated FMEA, new RPN or Action Priority, decision history

Step 1. Planning and preparation

At the outset, you need to define what is being analysed and why.

The scope should be specific: a product, component, line, process, operation, system or asset group. The team should include people who know the design, the process, quality, maintenance, automation, production and the historical data.

Important inputs include:

• technical drawings;
• design documentation;
• the bill of materials (BOM);
• process flow diagrams;
• work instructions;
• piping and instrumentation diagrams (P&ID) in process industries;
• failure data;
• customer complaints;
• quality reports;
• maintenance work order history;
• the experience of operators and technicians.

Step 2. Structure analysis

Structure analysis organises the object under review.

For a Design FMEA this may be the product structure and its components. For a Process FMEA it will be the process flow, operations, workstations and steps. For a System FMEA it is the system architecture and its interfaces.

Without this step, FMEA easily turns into a loose list of problems. Structure helps ensure that the team is analysing the right level of detail and is not mixing up components, functions, failures and actions.

Step 3. Function analysis

Function analysis defines what function a given element, process or system is supposed to perform.

Examples of functions:

• a pump is to maintain the required flow rate;
• a bearing is to allow the shaft to rotate under a defined load;
• a sensor is to deliver a correct temperature signal;
• an assembly operation is to deliver a tightening torque of 45 Nm ± 5%;
• a filter is to remove contaminants from the medium.

Only once the function has been defined can the failure mode — the way in which the function may not be fulfilled — be identified reliably.

Step 4. Failure analysis

Failure analysis identifies potential failure modes, effects and causes.

In this step the team describes:

• the failure mode — how the function can fail;
• the failure effect — what the consequence of the failure will be;
• the failure cause — why the failure may occur.

Example:

Cause: lubricant degradation or contamination.
Failure mode: bearing seizure.
Failure effect: pump shutdown and line downtime.

It is worth describing the effect at several levels: local, system and end. Locally, a component may be damaged. At system level, the machine stops. At the end level, production is interrupted, a delivery is delayed, a quality non-conformity arises or a safety risk emerges.

Step 5. Risk analysis

Risk analysis assigns priority to individual risks.

In the traditional approach, severity, occurrence and detection are rated on a scale of 1 to 10 and the RPN is then calculated:

RPN = severity × occurrence × detection

In the AIAG VDA approach, the S, O and D ratings are used to determine the Action Priority: high, medium or low.

Rating should follow a scale agreed before the FMEA session. The team should not change its interpretation of the values during the analysis, as this leads to inconsistent priorities.

Step 6. Optimisation

Optimisation means selecting actions for the most important failure modes.

Actions may involve:

• a design change;
• a process change;
• an additional control;
• improved work instructions;
• tool calibration;
• a material change;
• condition monitoring;
• preventive maintenance;
• automated inspection;
• training;
• a change to the workstation layout;
• component redesign;
• additional protection.

An action should have an owner, a deadline and a criterion for effectiveness. Simply writing “monitor” or “train” is too general if it is not clear who is to do it, when, how and to what standard.

Step 7. Results documentation

FMEA is a living document. It should be updated after actions are implemented, after a process change, after a failure, after a change of supplier, after a complaint, after a change in operating parameters or after equipment modernisation.

Once actions have been implemented, the team should reassess severity, occurrence and detection, or the Action Priority. This keeps FMEA from becoming an archived workshop record and turns it into a risk management tool across the life cycle of the product, process or asset.

How do you calculate RPN in FMEA?

RPN, the Risk Priority Number, is the traditional risk priority indicator calculated as the product of severity, occurrence and detection.

RPN = S × O × D

Where:

• S — severity, the seriousness of the failure effect;
• O — occurrence, the probability or frequency with which the cause or failure mode arises;
• D — detection, the ability of current controls to detect the cause or failure mode.

Each factor is usually rated on a scale of 1 to 10. The theoretical range of RPN is therefore 1 to 1,000. The higher the RPN, the higher the priority for analysis and action.

Severity, occurrence and detection scales

The table below shows an indicative S/O/D scale in a manufacturing context. Each organisation should adapt the definitions to its own process, risk profile, quality requirements and regulatory obligations.

Rating	Severity	Occurrence	Detection
10	Hazard without warning; safety or regulatory risk	Very high: ≥1 in 2	No known detection method
9	Hazard with warning	Very high: 1 in 3	Controls are unlikely to detect the problem
8	Very high; equipment inoperable, major loss	High: repeat failures, 1 in 8	Low chance of detection
7	High; reduced capability, significant impact	High: 1 in 20	Very low chance of detection
6	Moderate; operation in a degraded state	Moderately high: 1 in 80	Low; controls may detect the problem
5	Low; limited performance	Moderate: 1 in 400	Moderate chance of detection
4	Very low; minor defect	Moderately low: 1 in 2,000	Moderately high chance of detection
3	Minor; some operators will notice the problem	Low: 1 in 15,000	High chance of detection
2	Very minor; noticeable only to experienced users	Very low: 1 in 150,000	Very high chance of detection
1	No noticeable effect	Remote: ≤1 in 1,500,000	Detection almost certain

Scales should be defined before the FMEA workshop and applied consistently. Otherwise two teams may assign different ratings to the same risk and the results cease to be comparable.

An example RPN calculation

Suppose we are analysing a pump bearing assembly in a critical section of a production line.

Item	Value
Element	Pump bearing assembly
Function	Maintaining stable operation of the shaft and pump
Failure mode	Bearing seizure
Failure effect	Pump shutdown and line downtime
Failure cause	Lubricant degradation or contamination
Current control	Periodic inspection and vibration monitoring
Severity	7
Occurrence	4
Detection	5

Calculation:

RPN = 7 × 4 × 5 = 140

In many organisations, a result above a defined threshold — 100 or 200, for example — triggers a mandatory action. Such a threshold is not universal, however. It should follow from the organisation’s risk policy, asset criticality, customer requirements and quality rules.

An example action:

Implementing automatic lubrication, reviewing oil quality, refining the vibration monitoring frequency and updating the inspection instruction.

Once the action has been implemented, the team reassesses occurrence and detection. Severity usually stays the same, because the effect of the failure can still be serious, but probability and detectability may improve.

The limitations of RPN and the action priority method

RPN is simple and useful, but it can lead to misleading comparisons because it treats severity, occurrence and detection as equivalent numerical factors.

The problem is that different combinations of S, O and D can produce the same RPN even though they represent entirely different risks.

Example:

S = 10, O = 1, D = 1 → RPN = 10
S = 1, O = 2, D = 5 → RPN = 10

Both cases give the same numerical result, but the first involves a potentially extremely serious effect. In practice, a high-severity risk should not be ignored simply because occurrence and detection ratings are low.

A second limitation follows from the nature of the scales. Ratings of 1 to 10 are ordinal scales, not precise physical values. An RPN of 200 does not automatically mean that the risk is twice as large as an RPN of 100.

This is why the AIAG VDA FMEA approach from 2019 uses Action Priority, abbreviated to AP.

What is action priority?

Action Priority is a method of prioritising actions in FMEA that uses the combination of severity, occurrence and detection without reducing them to a single simple product.

The AP result is expressed as:

• high — action required or strongly expected;
• medium — action recommended and requiring a team decision;
• low — action optional or to be documented.

Action Priority protects high-severity risks better, because it does not allow them to be “lost” inside a low RPN. Many organisations still use RPN, particularly outside the automotive sector, but it is worth understanding the limitations of the method and stating clearly whether priorities in the organisation come from RPN, AP, expert judgement or a combination of these approaches.

The FMEA worksheet: What does the working document look like?

The FMEA worksheet is the working table that documents the structure of the analysis: the element, the function, the failure mode, the effect, the cause, the risk ratings, the actions, the responsibility and the outcome after implementation.

Column	Description	Example
Element / process step	Component, asset or process step	Pump bearing assembly / tightening step
Function / requirement	Required function or parameter	Maintaining shaft alignment / 45 Nm ± 5%
Potential failure mode	The way in which the function can fail	Bearing seizure / insufficient torque
Potential failure effect	Local, system or end effect	Pump and line shutdown / loose joint
Severity	Rating of the effect on a scale of 1 to 10	8 / 6
Potential failure cause	Cause or failure mechanism	Lubricant degradation / out-of-calibration tool
Occurrence	Rating of frequency on a scale of 1 to 10	3 / 4
Current detection controls	Existing ways of detecting the problem	Vibration monitoring / torque audit
Detection	Rating of detectability on a scale of 1 to 10	4 / 5
RPN or AP	Risk priority result	RPN = 96 / AP = medium
Recommended action	Action to reduce the risk	Automatic lubrication / daily tool check
Responsibility	Action owner	Reliability engineer / process engineer
Deadline	Date or period for completion	2026-Q3
Action taken	Information on the change implemented	Lubrication system implemented
Reassessed S/O/D/RPN	Rating after the action	S = 8, O = 1, D = 2, RPN = 16

A good FMEA worksheet should be specific. Entries that are too general, such as “machine failure”, “operator error” or “quality control”, make risk assessment harder and lead to actions that cannot be implemented effectively.

An FMEA example in manufacturing

The example below shows a simplified Process FMEA for a bolt-tightening operation in an assembly process.

Field	Example entry
Process step	Tightening the bolt that fixes the component
Function / requirement	Apply a torque of 45 Nm ± 5%
Potential failure mode	Insufficient tightening torque
Potential failure effect	Loose joint, increased vibration, component damage
Severity	6
Potential failure cause	Out-of-calibration tool, missing control, unclear instruction
Occurrence	4
Current controls	Periodic torque audit, weekly tool check
Detection	5
RPN	6 × 4 × 5 = 120
Recommended action	Daily tool calibration check and a digital work instruction
Responsibility	Process engineer
Deadline	End of quarter
Reassessment	S = 6, O = 2, D = 3, RPN = 36

Interpretation: the initial RPN is 120, so the team concludes that action is needed. Once a daily check has been introduced and the work instruction clarified, both the probability of occurrence and the risk of non-detection fall. Severity remains at the same level, because the effect of a potential failure can still be significant.

FMEA, FMECA and FTA — what is the difference?

FMEA, FMECA and FTA are all risk and reliability analysis methods, but they differ in the direction of the analysis, the degree to which they are quantitative, and their starting point.

FTA, or Fault Tree Analysis, starts from an undesirable event and breaks it down into possible causes and combinations of causes. FMEA starts from a component, a function or a process step and examines how they can fail.

Dimension	FMEA	FMECA	FTA
Full name	Failure Mode and Effects Analysis	Failure Mode, Effects and Criticality Analysis	Fault Tree Analysis
Direction of analysis	From element or process step to system	From element or process step to system, with criticality assessment	From undesirable event to causes
Starting point	Component, function or process step	Component, function or process step	Undesirable event, for example a system failure
Risk assessment	RPN or AP, qualitative or semi-quantitative	Criticality number or matrix	Probability of the top event
Result	List of failure modes and actions	FMEA plus criticality ranking	Fault tree with causal logic
Typical use	Design, process, maintenance, quality	Safety-critical systems	Cause investigation, system reliability
Standards	SAE J1739, IEC 60812, AIAG VDA	MIL-STD-1629A	IEC 61025

FMEA and FTA are complementary. FMEA starts from the elements and asks how they can fail. FTA starts from an undesirable event and asks which combinations of causes can lead to it.

Example: FMEA may identify pump bearing damage as a critical failure mode. FTA can then analyse which combinations of causes lead to a shutdown of the entire pumping system.

FMEA in the context of RCM and maintenance strategy

In maintenance, FMEA answers the question “what can fail and why”, while RCM helps decide “what should be done about it”.

RCM, or Reliability-Centred Maintenance, uses asset functions, functional failures, failure modes, failure effects and failure consequences to select a maintenance strategy. FMEA supplies part of that knowledge: it identifies failure modes, effects and causes.

On the basis of FMEA, a team can decide whether the right approach for a given asset is:

• run to failure;
• preventive maintenance;
• condition-based maintenance;
• predictive maintenance;
• condition monitoring;
• redesign;
• an additional control;
• a procedural change;
• an operator task within autonomous maintenance.

FMEA can also indicate which failure modes are worth monitoring. If the analysis shows that a critical bearing failure mode is preceded by rising vibration, then vibration monitoring is a logical part of the strategy. If the problem concerns lubricant degradation, oil analysis may be the more appropriate data source.

Input data for FMEA

The more an FMEA rests on real data and experience rather than the team’s general intuition, the better it is.

Typical data sources include:

Data source	What it contributes to FMEA
CMMS (Computerised Maintenance Management System)	Failure history, failure codes, work orders, repair times
MTTR and MTBF data	Information on repair times and failure frequency
Industrial data platform	Aggregation of data from many machines, lines and systems
Condition monitoring	Signals on technical condition and early symptoms of degradation
Operators and technicians	Practical knowledge, recurring problems, process workarounds
Engineering documentation	Drawings, P&IDs, specifications, materials, design requirements
Quality data	Scrap, complaints, deviations, inspection results
Previous FMEA analyses	Knowledge from similar products, processes or assets
Supplier data	Information on component service life and known failure modes

MTTR, or Mean Time To Repair, helps assess how long it takes to restore operability after a failure. MTBF, or Mean Time Between Failures, helps assess how often failures occur during operation.

This is where industrial data plays a decisive role. If failure history is inconsistent, cause codes are ambiguous, and production and maintenance data are not linked to assets, FMEA will be more of an expert workshop than an evidence-based analysis.

Smart RDM can act as the data and context layer: connecting information from production, maintenance, quality and monitoring systems in order to provide a more reliable input for assessing occurrence, detection and action priorities.

FMEA, data, OEE and operational decisions

FMEA supports maintenance strategy decisions because it shows which failure modes have the greatest impact on production, safety, quality and cost.

FMEA results can influence:

• the choice of assets for condition monitoring;
• the scope of preventive maintenance;
• predictive maintenance priorities;
• RCM decisions;
• actions to improve MTBF;
• the reduction of repeat failures;
• improved availability as a component of OEE;
• updates to procedures and work instructions;
• decisions on redesign or modernisation;
• process optimisation activities.

OEE, or Overall Equipment Effectiveness, measures the availability, performance and quality of equipment operation. In the FMEA context, availability matters most, because critical failure modes can directly drive unplanned downtime and lost production time.

FMEA should not, however, be treated as a substitute for these areas. FMEA identifies risks and priorities; dedicated methods and systems then carry out the specific work: monitoring, planning, analytics, reporting, workflow and the maintenance of standards.

Standards related to FMEA

FMEA is applied across many industries, which is why it appears in several standards and industry handbooks.

Standard or document	Significance
SAE J1739	The FMEA standard used in the automotive industry
IEC 60812	International procedure for failure mode and effects analysis
AIAG VDA FMEA Handbook	The joint AIAG and VDA handbook, the seven-step process and Action Priority
MIL-STD-1629A	The classic approach to FMECA in defence applications
AS9100	Aerospace context and quality systems in the aviation industry
IATF 16949	Quality context in the automotive industry
ISO 14971	Risk management for medical devices

Standards help bring order to terminology, analysis steps, documentation structure and the approach to risk. Each organisation should nonetheless adapt the scales, thresholds and ways of working to its own process, industry, customer requirements and asset criticality.

The vendor landscape in 2026

The market for FMEA tools spans educational organisations, PLM solutions, specialist reliability engineering software, requirements management systems and data platforms that supply the input for the analysis.

PLM, or Product Lifecycle Management, supports FMEA work in connection with the product design, the bill of materials, requirements and engineering changes.

Vendor	Area	Typical context
ASQ	Quality knowledge, training and materials	Education, standards, preparing quality and reliability teams
PTC Windchill	PLM with an FMEA module	FMEA linked to the BOM, product design and change management
Relyence	Reliability engineering software	FMEA, FMECA, FTA, RBD and reliability analyses
APIS IQ-RM	Specialist FMEA software	Automotive industry, AIAG VDA, VDA, structured FMEA facilitation
Jama Software	Requirements management and risk analysis	Linking FMEA to requirements and traceability
Siemens Teamcenter	PLM and quality management	Enterprise PLM, quality, FMEA and product lifecycle management
Smart RDM	Industrial data and analytics layer	Failure, production, maintenance and condition monitoring data as input for FMEA

The table is not a ranking. Not every tool plays the same role. Some systems support the running of the FMEA itself, some manage requirements and the product life cycle, and some — such as a data platform — supply the empirical data needed for a sound risk assessment.

The most common mistakes in FMEA

The most common mistake is treating FMEA as a document for the auditor rather than as a living risk management tool.

Typical problems:

• a scope of analysis that is too general;
• the absence of the right people on the team;
• confusing the failure mode, the failure effect and the failure cause;
• S/O/D ratings assigned without shared scales;
• applying RPN thresholds mechanically without considering severity;
• no action for high-severity risks;
• entries such as “operator error” with no root cause analysis;
• no action owner;
• no reassessment after an action is implemented;
• failure to update the FMEA after a process change or a failure;
• copying an old FMEA without verifying the data;
• not taking CMMS, quality and production data into account.

A good FMEA should be specific, evidence-based and connected to a real decision-making process. The worksheet on its own does not reduce risk. Risk is reduced only by well-chosen actions that are actually carried out.

How do you start implementing FMEA?

The safest place to start is a process, asset or product with a clear impact on safety, quality, availability or production cost.

A practical sequence:

1. Choose the scope of the analysis.
2. Decide the type of FMEA: Design FMEA, Process FMEA, System FMEA or FMECA.
3. Assemble a cross-functional team.
4. Prepare the technical documentation and historical data.
5. Define the functions and requirements.
6. Identify failure modes, effects and causes.
7. Agree the severity, occurrence and detection scales.
8. Assess the risk using RPN or Action Priority.
9. Define actions for the most important risks.
10. Assign owners and deadlines.
11. Implement the actions.
12. Update the FMEA and reassess the risk.

The most important question at the outset is this: which decision is the FMEA meant to support? If the answer is unclear, the analysis quickly becomes a formal exercise. If the decision is concrete — selecting assets for monitoring, changing a process or reducing repeat failures, for example — FMEA can become a practical maintenance and quality tool.

FAQ

What is FMEA?

FMEA, or Failure Mode and Effects Analysis, is a method for analysing potential failure modes, their effects, their causes and the priority of actions.

Its purpose is to prevent problems before they occur. In manufacturing, FMEA helps identify which failures, process defects or design errors may have the greatest impact on quality, safety, availability and cost.

What does the abbreviation FMEA stand for?

FMEA stands for Failure Mode and Effects Analysis.

A failure mode is the way in which a function can fail. A failure effect is the consequence of that event. The analysis is the systematic assessment of these risks and the selection of actions to limit their impact.

What does FMEA mean in manufacturing?

In manufacturing, FMEA most often means Process FMEA, or PFMEA.

PFMEA analyses risks in the manufacturing process: operations, parameters, tools, controls, work instructions, operator activities and potential execution errors. It helps reduce scrap, complaints, downtime and quality problems.

What are the 5 steps of FMEA?

The simplified five-step FMEA process covers:

1. Defining the scope and the functions.
2. Identifying failure modes.
3. Determining effects and causes.
4. Assessing risk through severity, occurrence and detection.
5. Selecting actions and updating the assessment after implementation.

The AIAG VDA approach uses a more detailed seven-step model.

What are the 7 steps of FMEA?

The seven steps of FMEA are:

1. Planning and preparation.
2. Structure analysis.
3. Function analysis.
4. Failure analysis.
5. Risk analysis.
6. Optimisation.
7. Results documentation.

This model takes the analysis from scope and structure, through functions and failures, to risk assessment, actions and the documentation of results.

How do you calculate RPN in FMEA?

RPN is calculated using the formula:

RPN = severity × occurrence × detection

Each factor is usually rated on a scale of 1 to 10. Example: if severity = 7, occurrence = 4 and detection = 5, then RPN = 7 × 4 × 5 = 140.

What is a good RPN?

There is no single universal threshold for a good or bad RPN.

Some organisations trigger action at an RPN above 100 or 200, but the threshold should follow from the risk policy, the industry, safety considerations, customer requirements and process criticality. It is also worth paying attention to high severity regardless of the RPN itself.

What is Action Priority in FMEA?

Action Priority is the prioritisation method used in AIAG VDA FMEA.

Instead of a simple RPN threshold, it uses the combination of severity, occurrence and detection to set the priority as high, medium or low. The method reduces the risk of misreading RPN, particularly for high-severity failure modes.

What is the difference between FMEA and FMECA?

FMECA is FMEA extended with criticality analysis.

FMEA identifies failure modes, effects, causes and actions. FMECA adds a quantitative or semi-quantitative criticality assessment, often using failure rate data and a criticality matrix. It is used above all in safety-critical systems.

What is the difference between FMEA and FTA?

FMEA is an element-to-system analysis: it starts from a component, a function or a process step and examines how a failure can arise and what its effects will be.

FTA, or Fault Tree Analysis, is an event-to-cause analysis: it starts from an undesirable event, such as a system shutdown, and looks for the combinations of causes that can lead to it. The two methods can complement each other.

What is the difference between FMEA and FEA?

FMEA and FEA are entirely different methods.

FMEA, or Failure Mode and Effects Analysis, is used to analyse failure risk and preventive actions. FEA, or Finite Element Analysis, is an engineering simulation method used for, among other things, stress, deformation, temperature and structural strength analysis.

Can you explain FMEA with an example?

Example: in an analysis of a pump bearing, the failure mode is bearing seizure, the failure effect is a pump and line shutdown, and the failure cause is lubricant degradation.

The team rates severity = 7, occurrence = 4 and detection = 5. The RPN is 140. The action might be automatic lubrication, oil analysis, vibration monitoring or a change in inspection frequency. Once the action has been implemented, the FMEA should be updated.

What data is needed for FMEA?

FMEA requires the functions of the element under analysis, the requirements, technical documentation, failure history, data from the maintenance management system, customer complaints, quality results, MTTR and MTBF data, input from operators and engineering knowledge.

The better the input data, the more realistic the assessment of occurrence and detection.

Is FMEA a one-off document?

No. FMEA should be a living document.

It should be updated after a change to the design, the process, the supplier, the material, the operating parameters or the maintenance strategy, and after a serious failure, a complaint or the implementation of corrective actions.

Sources

1. SAE J1739 — the FMEA standard for automotive applications.
2. IEC 60812 — the international procedure for failure mode and effects analysis.
3. AIAG VDA FMEA Handbook — the seven-step approach and Action Priority.
4. MIL-STD-1629A — the classic approach to FMECA.
5. ASQ — educational materials on FMEA.
6. Smart RDM internal materials on industrial data, maintenance and failure analysis.