Failure Mode and Effects Analysis (FMEA) Resource Center

10 Steps to Conduct a PFMEA

Step 1 | Review the process

Use a process flowchart to identify each process component.

Review the process components and the intended function or functions of those components.

  • Use of a detailed flowchart of the process or a traveler (or router) is a good starting point for reviewing the process.

There are several reasons for reviewing the process

  • First, the review helps assure that all team members are familiar with the process. This is especially important if you have team members who do not work on the process on a daily basis.
  • The second reason for reviewing the process is to identify each of the main components of the process and determine the function or functions of each of those components.
  • Finally, this review step will help assure that you are studying all components of the process with the PFMEA.

Using the process flowchart, label each component with a sequential reference number.

  • These reference numbers will be used throughout the FMEA process.
  • The marked-up flowchart will give you a powerful visual to refer to throughout the PFMEA.
  • With the process flowchart in hand, the PFMEA team members should familiarize themselves with the process by physically walking through the process. This is the time to assure everyone on the team understands the basic process flow and the workings of the process components.

For each component, list its intended function or functions.

  • The function of the component is the value-adding role that component performs or provides.
  • Many components have more than one function.

Step 2 | Brainstorm potential failure modes

Review existing documentation and data for clues.

Consider the potential failure modes for each component and its corresponding function.

  • A potential failure mode represents any manner in which the component or process step could fail to perform its intended function or functions.

Using the list of components and related functions generated in Step 1, as a team, brainstorm the potential failure modes for each function.

  • Don’t take shortcuts here; this is the time to be thorough.

Prepare for the brainstorming session.

  • Before you begin the brainstorming session, review documentation for clues about potential failure modes.

Step 3 | List potential effects of failure

There may be more than one for each failure.

Determine the effects associated with each failure mode.

  • The effect is related directly to the ability of that specific component to perform its intended function.
  • An effect is the impact a failure could make if it occurred.
  • Some failures will have an effect on the customers and others on the environment, the facility, and even the process itself.

As with failure modes, use descriptive and detailed terms to define effects.

  • The effect should be stated in terms meaningful to product or system performance.
  • If the effects are defined in general terms, it will be difficult to identify (and reduce) true potential risks.

Step 4 | Assign Severity rankings

Severity rankings are based on the severity of the consequences of failure.

Assign a severity ranking to each effect that has been identified.

  • The severity ranking is an estimate of how serious an effect would be should it occur.
  • To determine the severity, consider the impact the effect would have on the customer, on downstream operations, or on the employees operating the process.

The severity ranking is based on a relative scale ranging from 1 to 10.

  • A “10” means the effect has a dangerously high severity leading to a hazard without warning.
  • Conversely, a severity ranking of “1” means the severity is extremely low.

Step 5 | Assign Occurrence rankings

Occurrence rankings are based on how frequently the cause of the failure is likely to occur.

Next, consider the potential cause or failure mechanism for each failure mode; then assign an occurrence ranking to each of those causes or failure mechanisms.

We need to know the potential cause to determine the occurrence ranking because, just like the severity ranking is driven by the effect, the occurrence ranking is a function of the cause. The occurrence ranking is based on the likelihood, or frequency, that the cause (or mechanism of failure) will occur.

If we know the cause, we can better identify how frequently a specific mode of failure will occur. How do you find the root cause?

  • There are many problem-finding and problem-solving methodologies.
  • One of the easiest to use is the 5-Whys technique.
  • Once the cause is known, capture data on the frequency of causes. Sources of data may be scrap and rework reports, customer complaints, and equipment maintenance records.

The occurrence ranking scale, like the severity ranking, is on a relative scale from 1 to 10.

  • An occurrence ranking of “10” means the failure mode occurrence is very high, and happens all of the time. Conversely, a “1” means the probability of occurrence is remote.
  • See FMEA Checklists and Forms for an example PFMEA Occurrence Ranking Scale.

Your organization may need an occurrence ranking scale customized for a low-volume, complex assembly process or a mixture of high-volume, simple processes and low-volume, complex processes.

  • Consider customized occurrence ranking scales based on time-based, event-based, or piece-based frequencies.
  • See FMEA Checklists and Forms for examples of Custom PFMEA Ranking Scales. (Examples of custom scales for severity, occurrence, and detection rankings are included in this Resource Center.)

Step 6 | Assign Detection rankings

Based on the chances the failure will be detected prior to the customer finding it.

To assign detection rankings, identify the process or product related controls in place for each failure mode and then assign a detection ranking to each control. Detection rankings evaluate the current process controls in place.

  • A control can relate to the failure mode itself, the cause (or mechanism) of failure, or the effects of a failure mode.
  • To make evaluating controls even more complex, controls can either prevent a failure mode or cause from occurring or detect a failure mode, cause of failure, or effect of failure after it has occurred.
  • Note that prevention controls cannot relate to an effect. If failures are prevented, an effect (of failure) cannot exist!

The Detection ranking scale, like the Severity and Occurrence scales, is on a relative scale from 1 to 10.

  • A Detection ranking of “1” means the chance of detecting a failure is certain.
  • Conversely, a “10” means there is absolute certainty of non-detection. This basically means that there are no controls in place to prevent or detect.
  • See FMEA Checklists and Forms for an example PFMEA Detection Ranking Scale.

Taking a lead from AIAG, consider three different forms of Custom Detection Ranking options. Custom examples for Mistake-Proofing, Gauging, and Manual Inspection controls can be helpful to PFMEA teams.

  • See FMEA Checklists and Forms for examples of Custom PFMEA Ranking Scales. (Examples of custom scales for severity, occurrence, and detection rankings are included in this Resource Center.)

Step 7 | Calculate the RPN

RPN = Severity X Occurrence X Detection.

  • The RPN is the Risk Priority Number. The RPN gives us a relative risk ranking. The higher the RPN, the higher the potential risk.
  • The RPN is calculated by multiplying the three rankings together. Multiply the Severity ranking times the Occurrence ranking times the Detection ranking. Calculate the RPN for each failure mode and effect.
  • Editorial Note: The current FMEA Manual from AIAG suggests only calculating the RPN for the highest effect ranking for each failure mode. We do not agree with this suggestion; we believe that if this suggestion is followed, it will be too easy to miss the need for further improvement on a specific failure mode.

Since each of the three relative ranking scales ranges from 1 to 10, the RPN will always be between 1 and 1000. The higher the RPN, the higher the relative risk. The RPN gives us an excellent tool to prioritize focused improvement efforts.

Step 8 | Develop the action plan

Define who will do what by when.

Taking action means reducing the RPN.

  • The RPN can be reduced by lowering any of the three rankings (severity, occurrence, or detection) individually or in combination with one another.
  • A reduction in the Severity ranking for a PFMEA is often the most difficult. It usually requires a physical modification to the process equipment or layout.
  • Reduction in the Occurrence ranking is accomplished by removing or controlling the potential causes.
  • Mistake-proofing tools are often used to reduce the frequency of occurrence.

A reduction in the Detection ranking can be accomplished by improving the process controls in place.

  • Adding process fail-safe shut-downs, alarm signals (sensors or SPC), and validation practices including work instructions, set-up procedures, calibration programs, and preventative maintenance are all detection ranking improvement approaches.

What is considered an acceptable RPN? The answer to that question depends on the organization.

  • For example, an organization may decide any RPN above a maximum target of 200 presents an unacceptable risk and must be reduced. If so, then an action plan identifying who will do what by when is needed.

There are many tools to aid the PFMEA team in reducing the relative risk of failure modes requiring action.

Among the most powerful tools are Mistake-Proofing, Statistical Process Control, and Design of Experiments.

Mistake-Proofing (Poka Yoke)

  • Techniques that can make it impossible for a mistake to occur, reducing the Occurrence ranking to 1.
  • Especially important when the Severity ranking is 10.

Statistical Process Control (SPC)

  • A statistical tool that helps define the output of a process to determine the capability of the process against the specification and then to maintain control of the process in the future.

Design of Experiments (DOE)

  • A family of powerful statistical improvement techniques that can identify the most critical variables in a process and the optimal settings for these variables.

Step 9 | Take action

Implement the improvements identified by your PFMEA team.

  • The Action Plan outlines what steps are needed to implement the solution, who will do them, and when they will be completed.
  • A simple solution will only need a Simple Action Plan while a complex solution needs more thorough planning and documentation.
  • Most Action Plans identified during a PFMEA will be of the simple “who, what, & when” category. Responsibilities and target completion dates for specific actions to be taken are identified.
  • Sometimes, the Action Plans can trigger a fairly large-scale project. If that happens, conventional project management tools such as PERT Charts and Gantt Charts will be needed to keep the Action Plan on track.
  • Most Action Plans identified during a PFMEA will be of the simple “who, what, & when” category. Responsibilities and target completion dates for specific actions to be taken are identified.

Step 10 | Calculate the resulting RPN

Re-evaluate each of the potential failures once improvements have been made and determine the impact of the improvements.

  • This step in a PFMEA confirms the action plan had the desired results by calculating the resulting RPN.
  • To recalculate the RPN, reassess the severity, occurrence, and detection rankings for the failure modes after the action plan has been completed.