Reliability centered maintenance in practice: how to choose the right maintenance strategy for failures, risks, and costs

RCM, or reliability centered maintenance, is a structured method for determining which maintenance activities an asset genuinely needs in its specific operating context. It does not start with the question “what should we service, and how often?” It starts with a different one: what function is this asset supposed to deliver, and what happens if it fails to deliver it?

In practice, RCM is not a single maintenance strategy. It is a decision framework that helps you select the right approach for each failure mode: condition monitoring, preventive maintenance, predictive maintenance, a failure-finding test, run-to-failure, or a design change.

What is RCM?

RCM (reliability centered maintenance) is a systematic engineering method used to determine the maintenance requirements of physical assets in their current operating context.

Its most important principle is this: the goal of maintenance is not to preserve the equipment, but to preserve the function the equipment is meant to deliver. A pump is not serviced because it is a pump. It is serviced because it has to deliver a defined flow rate, a defined pressure, process availability, or a safety function.

RCM analyses:

• the functions of the asset;
• the required performance standards;
• functional failures;
• failure modes;
• failure effects;
• failure consequences;
• the technical feasibility of each task;
• the cost-effectiveness of each task;
• the appropriate type of maintenance task.

The output of RCM is a documented maintenance program: specific tasks, intervals, ownership, and a rationale explaining why a given strategy was assigned to a given failure mode.

The core principle of RCM: function matters more than equipment

RCM is function-oriented rather than equipment-oriented. The analysis begins with what the user expects the asset to do in a particular working environment.

The same type of equipment may require entirely different strategies in different plants. An auxiliary pump in a non-critical system can reasonably be run reactively. An identical pump in a critical cooling loop may require condition monitoring, periodic testing, redundancy, or a design change.

RCM therefore takes the operating context into account – the real conditions under which the asset works:

• its function within the process;
• load;
• working environment;
• the consequences of a shutdown;
• the availability of standby capacity;
• safety requirements;
• impact on production;
• environmental impact;
• repair cost;
• availability of spare parts and people;
• historical failure data.

Without the operating context, an RCM analysis loses its meaning. The equipment model alone is never enough to define a maintenance strategy. What decides is the function, the consequence of failure, and the real conditions of operation.

A short history of RCM

RCM originated in commercial aviation, where growing aircraft complexity and rising maintenance costs forced the industry to abandon the assumption that more inspections always mean higher reliability.

In the 1960s, the reliability programmes run by United Airlines showed that many failures are not the result of simple age-related wear. In 1978, Stanley Nowlan and Howard Heap published the report Reliability-Centered Maintenance for the US Department of Defense. That document became one of the foundations of the method.

A key finding of their analysis was that only a small share of failure patterns is genuinely wear-related. According to the aviation data they cite, only around 11% of the cases analysed showed a clear relationship with component age, while roughly 89% were independent of any simple time interval.

This called into question the reflex of “replace every X hours”. RCM developed as a method that matches the task to the specific failure mode and its consequences, instead of applying one schedule to every component.

Key milestones in the development of RCM:

Period	Milestone	Significance
1960s	United Airlines reliability programmes	The first move away from a purely time-based inspection model
1978	Nowlan and Heap report, “Reliability-Centered Maintenance”	The foundation of modern RCM
1968–1980	MSG-1, MSG-2 and MSG-3 documents (Maintenance Steering Group)	Development of maintenance principles in aviation
1991	John Moubray publishes “RCM II”	Adaptation of RCM to general industry and popularisation of the 7 questions
1999	SAE JA1011	Evaluation criteria for RCM processes
2002	SAE JA1012	A guide to applying the JA1011 standard
Today	Manufacturing, energy, infrastructure, process industries	RCM as the method for selecting maintenance strategy

The 7 RCM questions

The seven RCM questions form the backbone of the analysis: from the function of the asset, through failures and their consequences, to the selection of the right maintenance task.

No.	RCM question	What it establishes
1	What are the functions of the asset and the required performance standards in its current operating context?	Asset functions and measurable performance standards
2	In what ways can the asset fail to fulfil its functions?	Functional failures – total or partial loss of function
3	What causes each functional failure?	Failure modes – the specific damage mechanisms
4	What happens when each failure occurs?	Failure effects visible to the process, the user or the system
5	In what way does each failure matter?	Failure consequences: hidden, safety, environmental, operational or economic
6	What can be done to predict or prevent each failure?	Proactive tasks: condition monitoring, restoration, discard, failure-finding tests
7	What should be done if no suitable proactive task can be found?	Default actions: run-to-failure, redesign, or a one-off change

The first question anchors the analysis in function and context. The second and third move to functional failures and failure modes. The fourth and fifth establish effects and consequences. The sixth and seventh lead to the decision: what is worth doing, and what is not.

Questions 1–2: functions and functional failures

A function describes what the asset is supposed to do. Functional failures describe the ways in which the asset can stop delivering that function.

Example:

Element of the analysis	Example
Asset	Process pump
Function	Deliver a flow of 120 m³/h at a defined pressure
Functional failure	No flow
Partial functional failure	Flow below the required level
Context	The pump operates in a system without full standby capacity

In RCM, a function should be stated in measurable terms wherever possible. “The pump must pump” is far too vague. A better statement defines flow, pressure, operating conditions and the required performance standard.

Questions 3–4: failure modes and failure effects

A failure mode is the specific cause or mechanism that leads to a functional failure. A failure effect describes what happens once that failure mode occurs.

Example for a pump:

Failures and effects	Example
Functional failure	The pump does not deliver the required flow
Failure mode	Bearing seizure
Technical cause	Lubricant degradation, misalignment, contamination
Local effect	Rising vibration and temperature
System effect	Pump shutdown
End effect	Reduced output or a line stoppage

This part of the analysis closely resembles Failure Mode and Effects Analysis (FMEA). RCM uses the same logic of identifying functions, failures, modes and effects – but it goes further, adding consequence assessment and the selection of a maintenance task.

Question 5: failure consequences

Failure consequences determine how seriously a given failure mode should be treated and what type of action makes sense.

RCM normally distinguishes four categories of consequence:

Consequence category	Meaning	Example
Hidden failure consequences	The failure is not evident during normal operation	Failure of a standby system, a safety sensor or a protective device
Safety and environmental consequences	The failure may harm people, the environment or regulatory compliance	Loss of a protective function, a leak, an emission, injury risk
Operational consequences	The failure affects output, performance, quality or operating cost	Line stoppage, reduced throughput, quality deterioration
Non-operational (economic) consequences	The failure mainly generates repair cost, with no material impact on production	Damage to an auxiliary component with no process impact

Assessing consequences matters more than counting failures. A rare failure with a safety consequence can outrank a frequent failure that is cheap and easy to fix.

Questions 6–7: proactive tasks and default actions

The sixth question concerns actions that can predict or prevent a failure. The seventh deals with the situation in which no proactive action is technically feasible or economically justified.

In RCM, a proactive task must satisfy two conditions:

• it must be technically feasible – it has to genuinely detect or mitigate the failure mode in question;
• it must be worth doing – the benefit has to justify the cost and the burden placed on the organisation.

If no proactive task meets both conditions, RCM may point to run-to-failure where consequences are low, or to redesign where the consequences are unacceptable.

The RCM process step by step

The RCM process runs from the selection of a system through to a documented maintenance program and its subsequent revision in the light of operating experience.

Step	Activity	Output
1	Select the system or asset to analyse	A prioritised list of systems based on criticality, cost and safety impact
2	Define the system boundaries	A boundary document: what is in scope and what is not
3	Identify functions	A list of primary and secondary functions with measurable standards
4	Identify functional failures	A list of the ways in which function can be lost, fully or partially
5	Identify failure modes	A list of the mechanisms that cause each functional failure
6	Analyse failure effects	A description of what happens when each failure mode occurs
7	Assess consequences	A consequence category assigned to every failure mode
8	Select maintenance tasks	A maintenance task assigned to each failure mode
9	Implement the program	Task schedules, ownership, procedures and records in the maintenance system
10	Keep the program alive	Periodic review and revision based on operating data

RCM should not be a one-off workshop that ends with a spreadsheet. The analysis needs to return to the team after failures, process changes, upgrades, new data, load changes, or any change in the consequences of a shutdown.

RCM decision logic: how a maintenance task is selected

The RCM decision logic runs from the failure mode, through its consequences, to the selection of the right task: condition monitoring, restoration, discard, a failure-finding test, run-to-failure, or redesign.

The team first establishes whether the failure is hidden or evident. It then assesses whether the failure carries safety, environmental, operational or purely economic consequences. Only then is the task type selected.

The logic in short:

failure mode → failure effect → consequence category → technical feasibility of the task → cost-effectiveness of the task → selected action

If a failure has hidden consequences, a failure-finding test is often the appropriate task. If degradation is measurable, an on-condition task may be the answer. If a component shows a clear wear-out pattern, scheduled restoration or discard becomes viable. If consequences are low, run-to-failure can be accepted. And if consequences are unacceptable while no effective proactive task exists, redesign may be the only remaining option.

The task categories RCM can select

RCM does not impose a single strategy. For each failure mode it selects the task that can actually be justified.

Task type	Description	When RCM selects it	Example
On-condition task	Monitor a condition parameter and act once degradation is detected	When degradation can be detected before functional failure and the P-F interval (from the first detectable symptom of a potential failure to functional failure) is long enough	Vibration monitoring on a pump bearing
Scheduled restoration	Restore the asset to its initial capability at a fixed interval	When an age-related wear pattern exists and restoration is feasible	Overhaul of a compressor valve assembly every 8,000 hours
Scheduled discard	Replace the component at a fixed interval regardless of its condition	When the item wears out with age and restoration is not appropriate	Seal replacement every 12 months
Failure-finding task	A periodic test confirming that a hidden function still works	When the failure is not evident during normal operation	Quarterly test of the fire protection system
Run-to-failure	No scheduled maintenance; repair or replace after failure	When consequences are low and no proactive task can be justified	Replacing a light bulb once it burns out
Redesign	A change to the design, the process or the configuration	When consequences are unacceptable and no effective proactive task exists	Adding a standby pump in a critical cooling loop

On-condition tasks are closely related to condition-based maintenance (CBM) and condition monitoring. RCM may also point towards predictive maintenance (PdM) where the available data and analytics make it possible to forecast failure risk.

The six failure patterns: why a schedule alone does not always improve reliability

One of the foundations of RCM is the observation that many failures have no simple relationship with component age. This is why fixed replacement “every X hours” does not automatically improve reliability.

Nowlan and Heap described six failure patterns:

Pattern	Description	Share of components in Nowlan and Heap’s aviation data	Age-related?
A – bathtub	High initial failure rate, a period of constant failure rate, then wear-out	4%	Yes (in the wear-out phase)
B – wear-out	Constant or slowly rising failure rate, followed by rapid wear-out	2%	Yes
C – gradual increase	Failure rate rises gradually with age	5%	Yes
D – initial break-in	Low failure rate when new, rising to a constant level	7%	No
E – random	A constant failure rate independent of age	14%	No
F – infant mortality	High initial failure rate, then a constant or slowly rising level	68%	No

The key conclusion for RCM: patterns D, E and F – roughly 89% of the cases in the aviation data – are not classically age-related. For those failures, rigid time-based restoration or discard may simply fail to deliver the expected effect.

This does not mean preventive maintenance is wrong. It means preventive maintenance belongs where it fits the actual degradation pattern and the consequences of failure. For many failure modes, condition monitoring, a hidden-function test, run-to-failure or a design change will be the better answer.

RCM and FMEA – similarities and differences

RCM contains elements that closely resemble FMEA, but it does not stop at identifying and scoring risk. RCM adds consequence assessment and a decision logic for selecting the maintenance task.

Dimension	RCM	FMEA
Full name	Reliability Centered Maintenance	Failure Mode and Effects Analysis
Primary goal	Select the right maintenance strategy for each failure mode	Identify and prioritise failure modes by risk
Scope	Functions → functional failures → failure modes → effects → consequences → maintenance tasks	Failure modes → effects → causes → risk assessment
Output	A maintenance program: tasks, intervals, ownership	A list of risks and recommended actions
Consequence assessment	Yes, split into hidden, safety, environmental, operational and economic	Usually via severity scoring, without RCM’s full consequence logic
Maintenance task selection	Yes, through a decision logic	Not as a core element of the method
Standards	SAE JA1011, SAE JA1012	IEC 60812, SAE J1739
Relationship	RCM uses an FMEA-like analysis, then goes further	FMEA can feed inputs into RCM

Put simply: FMEA helps you understand what can fail and how much risk it carries. RCM helps you decide which maintenance task makes sense for that failure mode and its consequences.

RCM and TPM – how do they differ?

RCM and TPM share a goal: better reliability and better equipment performance. Their nature, however, is different. RCM is an analytical method; TPM is an organisational approach.

Total Productive Maintenance (TPM) engages the entire organisation in improving availability, quality, safety and the stability of machine operation. It covers autonomous maintenance, work standards, skills development, loss reduction and operator involvement.

Dimension	RCM	TPM
Nature	An engineering and decision-making method	An organisational philosophy and working system
Core question	Which maintenance task fits this failure mode?	How do we engage the organisation in improving equipment effectiveness?
Level of analysis	Asset, function, failure mode, consequence	Line, plant, teams, working culture
Output	A program of maintenance tasks	An organisational system: standards, roles, operator engagement
Relationship	Defines what to do	Helps define who performs part of the work, and how

RCM can identify tasks that operators will carry out as part of autonomous maintenance. TPM, in turn, supports the implementation of those tasks through work standards, training and day-to-day operational discipline.

RCM standards: SAE JA1011 and SAE JA1012

The two most important standards associated with RCM are SAE JA1011 and SAE JA1012. The first defines the criteria a process must meet before it can legitimately be called RCM. The second helps you apply those criteria in practice.

Standard	Scope
SAE JA1011	Evaluation Criteria for Reliability-Centered Maintenance Processes
SAE JA1012	A Guide to the Reliability-Centered Maintenance Standard

SAE JA1011 sets the minimum requirements for an RCM process. The process must answer the seven questions, take the operating context into account, define functions together with their performance standards, analyse functional failures, failure modes and effects, and apply a decision logic for selecting tasks.

SAE JA1012 explains how to apply the JA1011 criteria. It can be treated as an implementation guide: JA1011 defines what an RCM process must satisfy, while JA1012 helps you understand how to meet those requirements.

Data, analytics and RCM

RCM can be run as a workshop exercise, but its quality rises sharply when the team works from reliable data on failures, work orders, costs, availability, downtime and asset condition.

Typical data that supports an RCM analysis:

• failure history;
• work orders;
• repair times;
• mean time between failures (MTBF);
• mean time to repair (MTTR);
• failure cause codes;
• condition monitoring data;
• alarms;
• spare part costs;
• labour costs;
• production data;
• the process impact of downtime;
• asset criticality.

An industrial data platform can integrate information from CMMS, SCADA, production systems, sensors, quality and energy to support the assessment of failure modes and their consequences. Industrial analytics can help identify recurring failures, MTTR and MTBF trends, costs, and the impact of failures on availability.

RCM also supports operational decisions: whether to keep a preventive maintenance schedule, move to condition monitoring, accept run-to-failure, change the design, or simply tighten a working procedure. In that sense, RCM is a practical decision framework for maintenance.

The impact of RCM on availability, OEE and the maintenance program

RCM improves asset availability by aligning maintenance tasks more closely with functions, failure modes and consequences.

Availability is one of the three components of OEE, alongside performance and quality, but RCM does not replace OEE analysis. RCM acts earlier: it helps select the maintenance tasks that reduce unacceptable failures, improve reliability and cut the share of unplanned downtime.

The output of RCM should be a maintenance program that documents:

• the assets covered by the analysis;
• functions and performance standards;
• functional failures;
• failure modes and effects;
• failure consequences;
• the task types selected;
• intervals;
• task owners;
• required competencies;
• required spare parts;
• work instructions;
• the rules for keeping the program up to date.

Digital work instructions can help standardise the tasks that come out of RCM, but execution documentation is only one part of implementation. What matters most is that every task is justified by the function of the asset and the consequence of its failure.

Implementing RCM in a plant

It is worth starting RCM with critical assets rather than the whole machine park. A full analysis of every item can become prohibitively time-consuming without a clear set of priorities.

A practical implementation sequence:

1. Set the business and technical objectives of the programme.
2. Select systems or assets based on criticality.
3. Define the boundaries of the analysis.
4. Assemble a cross-functional team.
5. Prepare data on failures, work orders, costs and downtime.
6. Describe functions and performance standards.
7. Identify functional failures.
8. Identify failure modes and effects.
9. Assess consequences.
10. Apply the RCM decision logic.
11. Define tasks, intervals and responsibilities.
12. Load the tasks into the maintenance system.
13. Train the users.
14. Track results through KPIs.
15. Update the program as new data and new failures come in.

An RCM team should combine maintenance, production, automation, process engineering, safety, quality and data competencies. Technical knowledge alone is not enough if the analysis lacks information about the real consequences of failure for the process.

The tooling and vendor landscape in 2026

RCM can be supported by training providers, CMMS and EAM systems, data platforms, analytics, and specialist reliability engineering tools. The table below is deliberately neutral and is not a ranking.

Vendor	Scope	Typical context
MaintainX	Mobile CMMS supporting schedules, tasks and compliance tracking	Executing the maintenance programs that come out of an analysis
Reliability Academy	RCM training and certification programmes based on SAE standards	Building reliability competencies within teams
ReliablePlant / Noria	Industry articles, methodology examples and practitioner material	Practical knowledge for maintenance and reliability teams
Smart RDM	Industrial data and operational analytics platform	Integrating failure, work order, production, condition monitoring and KPI data

Smart RDM is not an RCM system in itself. It can, however, support both the analysis and the implementation of RCM by integrating data from maintenance, production, quality, energy and condition monitoring, and by providing dashboards that track the results of maintenance decisions.

The most common mistakes in RCM

The most frequent mistake is reducing RCM to a longer list of preventive tasks. RCM is not about adding more inspections; it is about matching the right task to the function, the failure mode and the consequence.

Typical pitfalls:

• analysing equipment without defining its function;
• ignoring the operating context;
• failure modes described too generically;
• confusing a functional failure with a failure mode;
• skipping consequence assessment;
• defaulting to preventive maintenance for every problem;
• overlooking hidden failures;
• working without historical data;
• excluding production and operators from the analysis;
• leaving tasks without an owner;
• never loading the results into the maintenance system;
• failing to update the program after failures;
• treating RCM as a one-off project.

A good RCM analysis should end in a decision: which task we perform, why that task, at what interval, who owns it, and which data will tell us whether it works.

FAQ

What is reliability centered maintenance?

Reliability centered maintenance, or RCM, is a method for determining the maintenance requirements of physical assets on the basis of their functions, operating context, failure modes and failure consequences.

RCM helps you select the right strategy for each failure mode: condition monitoring, preventive maintenance, predictive maintenance, run-to-failure or redesign.

What does the abbreviation RCM stand for?

RCM stands for reliability centered maintenance.

In practice, the abbreviation denotes an analytical method focused on preserving the function of an asset rather than the asset itself as a piece of hardware.

What does RCM focus on?

RCM focuses on the function of an asset in its current operating context.

The central question is: what is the asset supposed to do, to what performance standard, in what ways can it lose that function, and what will the consequences of failure be? Only then is a maintenance task selected.

What is an RCM analysis?

An RCM analysis is the process of answering seven questions about functions, functional failures, failure modes, effects, consequences, and proactive or default actions.

The result is a maintenance program with specific tasks, intervals and responsibilities.

What are the 7 RCM questions?

The seven RCM questions are:

1. What are the functions of the asset and the required performance standards?
2. In what ways can the asset fail to fulfil its functions?
3. What causes each functional failure?
4. What happens when each failure occurs?
5. In what way does each failure matter?
6. What can be done to predict or prevent each failure?
7. What should be done if no suitable proactive task can be found?

These questions lead from the function of the asset to the selection of the right maintenance task.

What are the steps of the RCM process?

The RCM process covers system selection, boundary definition, function identification, identification of functional failures, identification of failure modes, effects analysis, consequence assessment, task selection, program implementation and periodic revision.

In practice, what matters most is the right scope, a capable team, concrete failure modes, honest consequence assessment, and getting the results into day-to-day maintenance.

How does RCM differ from FMEA?

FMEA identifies failure modes, their effects, their causes and the associated level of risk. RCM uses a similar analysis but adds consequence assessment and a logic for selecting a specific maintenance task.

FMEA mainly answers the question “what can fail, and how risky is it?” RCM also answers “and what should maintenance do about it?”

How does RCM differ from TPM?

RCM is an analytical method for selecting maintenance tasks on the basis of functions, failure modes and consequences.

TPM, or Total Productive Maintenance, is an organisational approach that engages operators, maintenance and the wider organisation in improving equipment effectiveness. RCM helps define what to do; TPM helps organise who performs part of that work, and how.

Which standards apply to RCM?

The key standards are SAE JA1011 and SAE JA1012.

SAE JA1011 defines the criteria a process must meet to be called RCM. SAE JA1012 is a guide explaining how to apply those criteria in practice.

Which tasks can RCM select?

RCM can select an on-condition task, scheduled restoration, scheduled discard, a failure-finding task, run-to-failure or redesign.

The choice depends on the failure mode, the detectability of degradation, the consequences of failure, technical feasibility and cost-effectiveness.

What are hidden failures in RCM?

Hidden failures are failures that operators cannot see during normal operation.

They typically affect standby systems, protective devices, safety sensors and emergency functions. In RCM, such failures often call for failure-finding tests, because without a test the organisation may never learn that a protective function has been lost.

What are failure consequences in RCM?

Failure consequences define why a given failure matters.

RCM divides consequences into hidden, safety and environmental, operational, and non-operational (economic). That classification determines whether a proactive task, a test, run-to-failure or redesign is required.

What are the six failure patterns?

The six failure patterns described by Nowlan and Heap are: the bathtub curve, wear-out, gradual increase in failure rate, initial break-in, a random failure rate, and infant mortality.

The key RCM conclusion is that many failures are not driven by component age. This is why fixed replacement intervals do not always improve reliability.

Does RCM replace preventive maintenance?

No. RCM does not replace preventive maintenance; it helps you decide when preventive maintenance is the right answer.

RCM can select scheduled restoration or scheduled discard, but it may equally select condition monitoring, predictive maintenance, a hidden-function test, run-to-failure or redesign.

Does every plant need RCM?

RCM makes most sense where assets are critical, failures carry material consequences, and the organisation wants to match its maintenance strategy to risk in a deliberate way.

A full RCM analysis is not always worth running on every item. It is usually better to start with critical systems, bottlenecks, assets with a high cost of downtime, or equipment that affects safety.

How do you measure the results of RCM?

The results of RCM can be measured through availability, MTBF, MTTR, the share of planned work, the number of repeat failures, maintenance cost, task completion, and the impact of failures on production.

What matters is comparing the situation before and after implementation, and revising the analysis regularly against real operating data.

Sources

1. SAE JA1011 – Evaluation Criteria for Reliability-Centered Maintenance Processes.
2. SAE JA1012 – A Guide to the Reliability-Centered Maintenance Standard.
3. Nowlan, F. S. & Heap, H. F. (1978), Reliability-Centered Maintenance, US Department of Defense.
4. Moubray, J. (1997), RCM II: Reliability-Centred Maintenance, 2nd ed.
5. ATA Maintenance Steering Group documents MSG-1, MSG-2 and MSG-3.
6. Industry literature on RCM, FMEA, TPM and maintenance strategy.
7. Smart RDM internal material on industrial data, analytics, asset monitoring and maintenance KPIs.