FMEA or Failure Mode and Effects Analysis

Sign up for our weekly newsletter for updates, articles and free giveaways of case studies, templates and training materials every issue! Simply enter your e-mail on the right

Design FMEA

1 - PURPOSE

This document defines how to apply the method of FMEA (Failure Mode and Effects Analysis) to a product.

The purpose of the Design FMEA is to analyze the potential modes of failure of a product, caused by its design and definition.

This technique of preventative analysis makes it possible to do the following:

ƒ{ Look for product malfunctions caused by its design and definition,

ƒ{ Evaluate their potential effects on the customer,

ƒ{ Identify the possible causes connected with the design and definition of the product,

ƒ{ Look for appropriate preventative actions.

2 - SCOPE

This method of analysis applies to the design and development of any new product during the quality planning phase.

3 - DEFINITIONS

DATE OF ORIGINAL OF THE STUDY:

This is the date the study specifications were written.

REQUESTER:

Person who, as part of his or her functions, has the power to request that an FMEA be carried out, in agreement with the "decider".

DECIDER:

Person who, as part of his or her functions, has the power to employ the human and financial means necessary to carry out the FMEA and to implement the plan of action resulting from it. The requester and the decider can be the same person.

FUNCTION:

Service to be provided by each element of the product. Example : To ensure air-tightness.

FAILURE MODE:

Malfunction of an element of the product leading to a partial or total failure to satisfy its function. It is judged in relation to each function. There are one or more potential failure modes for each function, e.g., cracked (split), non-airtight, oxidized.

EFFECTS:

These are the possible consequences of the failure mode on the product¡¦s performance during use, as perceived by the customer, e.g., slow decrease in pressure, irregular wear of the cover.

CAUSES:

These are the causes, relating to the choice and definition of a product element, which can provoke a failure mode. These causes can involve:

ƒ{ The material (nature and quantity),

ƒ{ Shape,

ƒ{ Position,

ƒ{ Links with adjacent elements,

ƒ{ Identification.

There can be several possible causes for a failure mode. Likewise, one cause can generate several failure modes. These causes can be:

ƒ{ Independent: A cause generates a failure mode for itself. Example: Poorly defined shape.

ƒ{ Combined: Several causes must arise together for a failure mode to be generated. Example: Poorly defined shape and position.

DETECTIONS:

These are all the different means (human and/or material) which exist during the design period (before production of the product) which could prevent a non-conform product from reaching the customer, e.g., verification by calculating in terms of finished elements, trials, tests.

CLASSIFICATION:

This is left to the initiative of the product lines and serves to identify failure modes, which could affect a special or main characteristic of the finished product. This identification can be in the form of an asterisk.

SEVERITY (S):

This is the severity of the consequences of the potential effects as felt by the customer. Severity is scored from 1 to 10 using a pre-determined scale.

A failure mode can have consequences of varying degrees of severity. The score it is given corresponds to the most serious consequences, which could reasonably be observed.

.

OCCURRENCE (O):

This is the probability that the potential cause of failure will occur and that it will bring about the failure mode being considered. This is irrespective of the severity of the consequences. Occurrence is scored from 1 to 10 using a pre-determined scale.

DETECTION (D):

This is the probability that a non-conform product will reach the customer, assuming that the cause and failure mode have arisen. Only the means of detection, which existed at the design and development stage, should be taken into account. Detection is scored from 1 to 10 using a pre-determined scale.

.

RISK PRIORITY NUMBER (RPN):

Hierarchical organization index resulting from the product of the different scores for severity, occurrence and detection allocated to each cause plus failure mode.

RPN = S x O x D. RPN can be between 1 and 1000.

RECOMMENDED ACTIONS:

Nature of the action(s) which will allow the probability of the cause plus failure mode appearing to be reduced, as a priority. Then those which will allow the probability of detection to be adapted to the minimum necessary. They must ensure that the system is made permanent.

4 - APPLICATION OF THE TOOL

The application of the Design FMEA must be led by a validated FMEA leader and requires methodical preparation. The purpose of the study must be clearly defined.

4.1 - PREPARATION OF THE FMEA IN TERMS OF QUALITY PLANNING

Before the FMEA is carried out, the leader must:

ƒ{ Acquire knowledge of the product and how it operates.

ƒ{ Design FMEA

ƒ{ Draw up the specifications for the study with the help of the requester, as the study is prepared. The standard printout in appendix 1 can be used.

ƒ{ Define the product study limits (entire finished product or limited to certain sub-assemblies or components).

ƒ{ Define the area where the effects will be felt.

ƒ{ Break down the product into sub-assemblies or components, depending on the level of detail required for the purpose of the study. Make a list of the different sub-assemblies or components.

ƒ{ Estimate the number of meetings needed to carry out the study. Set out the frequency (min. 1, max. 5 per week) of the meetings (recommended duration = 3 hours), taking into account the deadline for the end of the study and the constraints placed on those taking part. The FMEA must be carried out in a maximum of 6 months, otherwise break it down into sub-studies each lasting 6 months max. Obtain the agreement of the requester and decider as regards the workload to be undertaken.

ƒ{ Create the work group, gathering together all knowledge of the product (research and development).

ƒ{ In agreement with the requester and the decider, determine who shall be responsible for the FMEA dossier.

ƒ{ Determine the study characteristics in order to give all the necessary details on the context of the study:

ƒ{ the finished product taken as a reference in the study;

ƒ{ the control conditions and tests;

ƒ{ other particularities not covered by the above.

4.2 - CARRYING OUT THE FMEA

The FMEA is carried out by the work group led by the FMEA leader, and takes place during work meetings (number and duration set out in the study specifications). It follows the sequence below:

4.2.1 - Kick-Off of the Work Team

The participants are informed of the study to be carried out and how it will be done, and are trained in FMEA during the first meeting of the work group.

4.2.2 - Description of the Functions, Failure Modes, Cause and Effects, Detection and Classification

Use a table for each element of the product (sub-assembly or component) and hold a brainstorming session within the work group to list:

ƒ{ The FUNCTIONS of the sub-assembly or component being analyzed.

ƒ{ For each function, the possible FAILURE MODES and potential EFFECTS.

ƒ{ The CLASSIFICATION where relevant.

ƒ{ The possible CAUSES which can generate the failure mode being considered.

These will be more or less detailed depending on the degree of detail required.

ƒ{ For each cause, the DETECTIONS which already existed at the design stage, from the moment the cause appears to the moment the product leaves for use in current production. If no means of detection exists, write ¡§None.¡¨

4.2.3 - Evaluation of Risk Priority Number Once the failure mode and effects analysis of all the sub-assemblies or components have been carried out, evaluate the risk priority number of each cause plus failure mode, using the scale given in the study specifications:

SEVERITY:

This is different for each failure mode, whatever the causes are. It is estimated by the product technical function. In the impossibility to evaluate the severity, the maximum score (10) is given. Specify this reason in the comments column in the table (to remove any doubt later on). If there is any doubt as to which of two scores to give, give the higher one.

OCCURRENCE:

This is different for each cause plus failure mode. First estimate the frequency with which the cause occurs, then the probability that the cause will generate the failure mode. The occurrence score is obtained by the calculation: frequency of occurrence of the cause multiplied by the probability that the cause will generate the failure mode.

Take into account the preventative controls which stop the cause from appearing or which stop a cause, which has already appeared, from generating the failure mode. Do not take into account controls, which intervene after the failure mode has appeared.

If there is any doubt as to which of two occurrence scores to give, give the higher one. If the occurrence is not known at all, the maximum score (10) is given. Specify this reason in the comments column in the table (to remove any doubt later on).

DETECTION:

This is different for each cause plus failure mode. To estimate the risk of a non-conform product reaching the customer:

ƒ{ Assume that a failure mode has appeared.

ƒ{ Take into account the manner in which the malfunction of the product becomes visible.

ƒ{ Take into account the performance of all the existing means of detection (described in the "detection" column in the table) after the failure mode has appeared.

ƒ{ If there is any doubt as to which of two detection scores to give, give the higher one. If the detection is not known at all, the maximum score (10) is given.

Specify this reason in the comments column in the table (to remove any doubt later on).

RISK PRIORITY NUMBER:

For each cause plus failure mode. Multiply the three scores: S, O and D.

4.2.4 - Hierarchical Classification and Description of Corrective Action

Once the risk priority number of each cause plus failure mode has been evaluated, classify them in descending order of risk priority number.

In subsequent meetings, the work group:

ƒ{ Verifies that the order of problems is logical (possible deviations between the different risk priority number evaluation meetings) and corrects those risk priority number values, which are recognized as being incorrect.

ƒ{ Recommends preventive actions (recommended actions) to reduce any risk priority number values which are felt to be too high.

ƒ{ Prepares the conclusion (the main outstanding facts arising from the FMEA) which will appear in the study report.

4.2.5 - Study Report

Makes a report on the FMEA once the study is complete, to include:

ƒ{ A reminder of the causes and objectives of the study.

ƒ{ A summary of the study specifications.

ƒ{ A summary of the way the study was carried out (composition of the work group, list of the sub-assemblies or components analyzed, the cost and duration of the study).

ƒ{ The results of the study:

ƒ{ Histogram of causes plus failure modes for each class of risk priority number (attach as an appendix):

ƒ{ Summary of the causes plus failure modes with the actions recommended (attach as an appendix):

ƒ{ The conclusion summarizing the main facts brought to light by the analysis.

ƒ{ The scale used to evaluate the risk priority number is attached to the report.

After agreement with the study¡¦s requester, the report is distributed to the members of the work group, their managers, the requester and the decider. It can also be commented on in a final summary meeting held within a few days of its distribution. This meeting marks the end of the original study. The list of recommended actions will be used when compiling operational plans, depending on the objectives set out elsewhere.

4.2.6 - FMEA Dossier

The FMEA dossier is created by the group leader. It includes:

ƒ{ A page on the administration of updates,

ƒ{ The study specifications,

ƒ{ The study report,

ƒ{ The reports on revisions,

ƒ{ The FMEA tables of all the sub-assemblies or components analyzed.

It is passed to the person delegated as being "responsible for the dossier" in the study specifications.

4.2.7 ¡V Monitoring and revision

In case of a non-conformity being detected or following modification of the model, the FMEA is revised. Revision is carried out in work meetings by a re-convened work group, gathering together the same expertise as for the original study and led by the FMEA leader.

The summary of the causes plus failure modes from the last update will be used as support material. For each cause plus failure mode, list the actions taken the person responsible for each of them, the time frame and their date of application. Evaluate the new risk priority number, taking into account the actions taken (generally O and/or D change, unless the product has been redesigned; S does not change), using the same scale as for the original study.

Describe new recommended actions if those taken do not reduce the risk priority number sufficiently. Use the old recommended actions when no action has been taken. List the increases in risk priority number in relation to the last update by carrying out a sub-assembly-by-sub-assembly or component-by-component examination in co-operation with the members of the work group.

Write a revision report (using the same basic outline as for the original study). The summary of causes plus failure modes with recommended actions (attached as an appendix) is replaced by a summary of causes plus failure modes with the actions taken and the new recommended actions . The histogram of causes plus failure modes for each class of risk priority number is updated and attached as an appendix to the revision report.

Distribute this revision report and comment on it in a summary meeting (process identical to the original study).

Update the FMEA dossier by completing: ƒ{ The page on the administration of updates.

ƒ{ The FMEA tables for the study (action taken, person responsible, time taken, dates of application of actions taken, new risk priority number, new recommended actions if appropriate).

Note: The FMEA studies are updated periodically and systematically, and the dates of update are given as described above including when there has been no change.

Process FMEA

1 - PURPOSE

This document sets out how to apply the method for a Failure Mode and Effects Analysis to a process.

The purpose of the Process FMEA is to analyze the potential modes of failure of a product, caused by a process (production, conditioning, transport, warehousing, control, distribution etc. . . .).

This technique of preventative analysis makes it possible to do the following:

ƒ{ Look for product non-conformities caused by a process. ƒ{ Evaluate their potential effects on the customer (from the following operation or from a downstream or final factory). ƒ{ Identify the possible causes. ƒ{ Look for appropriate preventative action and pertinent improvement.

2 - SCOPE

This method of analysis applies to: ƒ{ All the transformation or assembly processes of a product (semi-finished or finished), or equipment in the Group (consent can be given to our suppliers for it to be used) in the quality planning phase, to contribute to the creation of a quality plan:

ƒ{ For new processes once they have reached a sufficient level of definition in the project stage. ƒ{ For existing processes retained for the production of new products in the project stage. ƒ{ For existing processes used to produce products in current production, in order to allow them to evolve continuously to keep up with customer requirements.

ƒ{ All basic operations making up the process: ƒ{ All transformation operations which add value to the product. ƒ{ Any other operations which do not add value to the product:

ƒ{ Reception of material.

ƒ{ Identification.

ƒ{ Handling.

ƒ{ Storage.

ƒ{ Control,

ƒ{ Transport,

ƒ{ Conditioning,

ƒ{ Re-standardization,

ƒ{ Dispatch.

3 - DEFINITIONS

OPERATION:

Breakdown of the process into elements which do or do not transform the product. Examples of operations: ƒ{ Installation,

ƒ{ Unrolling,

ƒ{ Cutting,

ƒ{ Evacuation.

PHASE OF PRODUCTION:

This is a noteworthy part of the process. It contributes to the acquisition of the product¡¦s characteristics. A phase of production is made up of one or more operations. Examples of phases of production: ƒ{ Changing a tire,

ƒ{ Laying a road surface,

ƒ{ Recording a record.

DATE OF ORIGINAL STUDY:

This is the date the study specifications were written.

REQUESTER:

Person who, as part of his or her functions, has the power to request that an FMEA be carried out, in agreement with the "decider".

DECIDER:

Person who, as part of his or her functions, has the power to employ the human and financial means necessary to carry out the FMEA and to implement the plan of action resulting from it. The requester and the decider can be the same person.

PURPOSE:

Added value to be obtained by the product or process during the operation or phase of production being considered (it is the service to be provided by the process during use).

FAILURE MODE:

Non conform state of the product as it leaves the phase or operation being considered. It is judged in relation to each purpose. There are one or more potential failure modes for each purpose.

EFFECTS:

These are the possible consequences of the failure mode, as suffered by the customer. Depending on the final objective of the study, the consequences relate to: ƒ{ The user of the finished product. Example: abnormal wear of the tire. ƒ{ The user of the product in the following operation or in a factory later in production. The effects are expressed as: rejection, needing repair, disruption of flow and risk for operator safety.

CAUSES:

These are the circumstances in which one or more elements of the process can provoke a failure mode. There can be several possible causes for a failure mode. Likewise, one cause can generate several failure modes. These causes can be: ƒ{ Independent: A cause generates a failure mode for itself. Example: failure of guide system. ƒ{ Combined: Several causes must arise together for a failure mode to be generated. Example: failure of product cooling system and ambient temperature in the workshop > 30¢XC.

DETECTION:

These are all the different means (human and/or material): ƒ{ Prevention which can help to avoid cause(s) occurring, or to prevent the cause(s) which have occurred from generating the failure mode:

ƒ{ Prevention interlock,

ƒ{ Preventative maintenance,

ƒ{ Verification of the incoming product before its used.

ƒ{ Detection used after the failure mode has appeared, which could prevent a non-conform product from reaching the customer:

ƒ{ Verification of the product / process,

ƒ{ Controlling the product / process / procedure,

ƒ{ Operator in the following operation,

ƒ{ Difficulties or impossibility of continuing the production process,

ƒ{ Detection interlock.

CLASSIFICATION:

This is left to the initiative of the product lines and serves to identify failure modes, which could affect a special or main characteristic of the finished product. This identification can be in the form of an asterisk.

SEVERITY (S):

This is the severity of the consequences of the potential effects as felt by the immediate and/or final customer (depending on the final objective of the study). Severity is scored from 1 to 10 using a pre-determined scale.

A failure mode can have consequences of varying degrees of severity. The score it is given corresponds to the most serious consequences, which could reasonably be observed.

Scale of scores depending on the final objective of the study: ƒ{ Technical performance of the finished product during use: see Appendix 8. ƒ{ Technical performance of the finished product during use and/or a later operation or in a customer factory: see Appendix 8. ƒ{ Other final objectives: scales to be created depending on the purpose of the study.

OCCURRENCE (O):

This is the probability that the potential cause of failure will occur and that it will bring about the failure mode being considered. This is irrespective of the severity of the consequences. To calculate this, take into account the means of detection said to be for ¡§Prevention.¡¨ Occurrence is scored from 1 to 10 using a pre-determined scale.

DETECTION (D):

This is the probability that a non-conform product will reach the customer (immediate or final, depending on the final objective of the study), assuming that the cause and failure mode have arisen. Only the means of detection, which intervene after the cause and failure mode has appeared, should be taken into account. Detection is scored from 1 to 10 using a pre-determined scale. Scale of scores applicable to any production:

RISK PRIORITY NUMBER (RPN):

Hierarchical organization index resulting from the product of the different scores for severity, occurrence and detection allocated to each cause plus failure mode.

RPN = S x O x D. RPN can be between 1 and 1000.

SERIES EFFECT (SE):

Number of finished products likely to be affected by a failure (taking into account existing controls) from the moment it occurs to the removal of the cause, which generated it. SE must be limited to the cases where a cause plus failure mode could affect a large number of finished products.

RECOMMENDED ACTIONS:

Nature of that action(s) which will allow the probability of the cause plus failure mode appearing to be reduced, is the priority. Then, those actions, which will allow the probability of detection to be increased to the minimum necessary, are the next priority. They must ensure that the system is made permanent.

4 - APPLICATION OF THE TOOL

The application of the Process FMEA must be led by a validated FMEA leader and requires methodical preparation. The purpose of the study must be clearly defined.

4.1 - PREPARATION OF THE FMEA IN TERMS OF QUALITY PLANNING

Before the FMEA is carried out, the leader must: ƒ{ Acquire knowledge of the process to be studied and of the product produced. ƒ{ Draw up the specifications for the study with the help of the requester, as the study is prepared. The standard printout in Appendix 7 can be used. ƒ{ Define the process study limits (incoming and outgoing state of the product). ƒ{ Define the area where the effects will be felt (e.g.: technical performance of the product during use, flow of products etc¡K). ƒ{ Break down the process into phases of production or production operations, depending on the level of detail required for the purpose of the study. Use an existing material circuit or, if necessary, draw a diagram of the process:

ƒ{ Estimate the number of meetings needed to carry out the study. Set out the frequency (min. 1, max. 5 per week) of the meetings (recommended duration = 3 hours), taking into account the deadline for the end of the study and the constraints placed on those taking part. The FMEA must be carried out in a maximum of 6 months, otherwise break it down into sub-studies each lasting 6 months maximum. Obtain the agreement of the requester and decider as regards the workload to be undertaken. ƒ{ Choose the risk priority number evaluation scale, depending on the choice of area of effects. ƒ{ Create the work group, gathering together all knowledge of the process, product during stages of manufacture and finished product. For an existing process, for example, it will be made up of 3 to 7 permanent members:

ƒ{ Operator and/or supervisor,

ƒ{ Process technician,

ƒ{ Workshop quality technician,

ƒ{ Product technician,

ƒ{ Maintenance technician,

ƒ{ Control technician,

ƒ{ Other specialists who may be consulted on request.

ƒ{ For example, for a process in the project stage, it can be made of: ƒ{ The process¡¦s designer

ƒ{ The central process technician for the operational unit, ƒ{ The quality technician for the operational unit, ƒ{ The product technician for the operational unit, ƒ{ The maintenance technician for the operational unit, ƒ{ The control technician for the operational unit.

ƒ{ In agreement with the requester and the decider, determine who shall be responsible for the FMEA dossier. ƒ{ Determine the study characteristics in order to give all the necessary details on the context of the study:

ƒ{ Dimension of the finished product taken as a reference in the study, ƒ{ Machines or equipment taken into account,

ƒ{ Particularities of the workforce,

ƒ{ Particularities of the method,

ƒ{ Particularities of the environment,

ƒ{ Conditions of control,

ƒ{ Other particularities not covered by the above.

4.2 - CARRYING OUT THE FMEA

The FMEA is carried out by the work group led by the FMEA leader, and takes place during work meetings (number and duration set out in the study specifications). It follows the pattern below:

4.2.1 - Kick-Off of the Work Team

The participants are informed of the study to be carried out and how it will be done, and are trained in FMEA during the first meeting of the work group.

4.2.2 - Description of the Functions,

Failure Modes, Cause and Effects, Detection and Classification

Use a table for each element of the process (phase or operation) and hold a brainstorming session within the work group to list:

ƒ{ The PURPOSES of the phase or operation analyzed, ƒ{ The possible FAILURE MODES for each purpose, ƒ{ For each failure mode: ƒ{ The POTENTIAL EFFECTS (depending on the area of effect defined in the study specifications). ƒ{ The CLASSIFICATION where relevant.

ƒ{ The POSSIBLE CAUSES from the beginning to the end of the phase or operation being analyzed. These will be more or less detailed depending on the degree of detail required. It is recommended to remain at the sub-assembly level (workforce, equipment, method, environment) if the break-down of the process is into production phases. Examples of sub-assembly causes: ƒ{ method of work and/or consignes not respected, ƒ{ failure of rolling system, ƒ{ packaging equipment in poor condition. Note: With the exception of product reception phases or operations and or the initial phase or operation of the process being analyzed, the incoming product must be considered to be conforming.

ƒ{ For each cause, list the CONTROLS which already exist for the product and/or the process and/or the procedures, from the moment the cause occurred in the phase or operation being considered to the moment the product is delivered to the customer (immediate or final, depending on the choice set out in the study specifications). If no means of detection exists, write ¡§None.¡¨

4.2.3 - Evaluation of Risk Priority Number

Once the failure mode and effects analysis of all the phases or operations has been carried out, evaluate the risk priority number of each cause plus failure mode, using the scale given in the study specifications:

SEVERITY:

This is different for each failure mode, whatever the causes are. It is estimated by the product technical function. In the impossibility to evaluate the severity, the maximum score (10) is given. Specify this reason in the comments column in the table (to remove any doubt later on). If there is any doubt as to which of two scores to give, give the higher one.

OCCURRENCE:

This is different for each cause plus failure mode. First estimate the frequency with which the cause occurs, then the probability that the cause will generate the failure mode. The occurrence score is obtained by the calculation.: frequency of occurrence of the cause multiplied by the probability that the cause will generate the failure mode. To do this use the practical experience of the operators and relevant technicians and cross-check this information against the available quality data. Take into account the preventative controls which stop the cause from appearing or which stop a cause, which has already appeared from generating the failure mode. Do not take into account controls, which intervene after the failure mode has appeared. Take into account any events, which may occur during the year (seasonal or climatic phenomena, changes of dimension, start-up at the beginning of the week, seasonal employment etc. . . .).

If there is any doubt as to which of two occurrence scores to give, give the higher one. If the occurrence is not known at all, the maximum score (10) is given. Specify this reason in the comments column in the table (to remove any doubt later on).

Note: If there is a risk of a series of non-conform products (whether the occurrence score is high or low), mention this in the comments column of the table and re-iterate it in the study¡¦s conclusion.

DETECTION:

This is different for each cause plus failure mode. To estimate the risk of a non-conform product reaching the customer: ƒ{ Assume that a failure mode has appeared (non-conform product(s) have been produced). ƒ{ Take into account the manner in which the non-conformity of the product becomes visible (non-conform series of random or intermittent appearance, very visible or not very visible). ƒ{ Take into account the performance of all the existing means of detection (described in the ¡§Detection¡¨ column in the table) after the failure mode has appeared. ƒ{ Take into account the rules for treating the non-conform product and the application of these rules, as well as the reliability of the non-conform circuits. Comment: Just because a non-conform product has been detected does not mean it will not be sent to the customer. ƒ{ Use the practical experience of the operators and relevant technicians and cross-check this information against the available quality data (indication of leak rates). ƒ{ If there is any doubt as to which of two detection scores to give, give the higher one. If the detection is not known at all, the maximum score (10) is given. Specify this reason in the comments column in the table (to remove any doubt later on).

RISK PRIORITY NUMBER:

For each cause plus failure mode, multiply the three scores -- S, O and D.

SERIES EFFECT (SE):

Fill in if appropriate.

4.2.4 - Hierarchical Classification and Description of Corrective Action

Once the Risk Priority Number of each cause plus failure mode has been evaluated, classify them in descending order of Risk Priority Number. In subsequent meetings, the work group: ƒ{ Verifies that the order of problems is logical (possible deviations between the different risk priority number evaluation meetings) and corrects those risk priority number values, which are recognized as being incorrect.

Recommends preventive actions (recommended actions) to reduce any risk priority number values and/or risks of a series of non-conform product which are felt to be too high.

ƒ{ Prepares the conclusion (the main outstanding facts arising from the FMEA) which will appear in the study report.

4.2.5 - Study Report

Makes a report on the FMEA once the study is complete, to include: ƒ{ A reminder of the causes and objectives of the study. ƒ{ A summary of the study specifications. ƒ{ A summary of the way the study was carried out (composition of the work group, list of the phases or operations analyzed, the cost and duration of the study). ƒ{ The results of the study: ƒ{ Histogram of causes plus failure modes for each class of risk priority number (attach as an appendix): ƒ{ Summary of the causes plus failure modes with the actions recommended (attach as an appendix):

ƒ{ The conclusion summarizing the main facts brought to light by the analysis. ƒ{ The scale used to evaluate the risk priority number is attached to the report.

After agreement with the study¡¦s requester, the report is distributed to the members of the work group, their managers, the requester and the decider. It can also be commented on in a final summary meeting held within a few days of its distribution. This meeting marks the end of the original study. The list of recommended actions will be used when compiling operational plans, depending on the purposes set out elsewhere.

4.2.6 - FMEA Dossier

The FMEA dossier includes:

ƒ{ A page on the administration of updates, ƒ{ The study specifications, ƒ{ The study report, ƒ{ The reports on revisions, ƒ{ The FMEA tables of all the phases or operations analyzed. It is passed to the person delegated as being ¡§responsible for the dossier¡¨ in the study specifications.

4.2.7 - Monitoring and Revision

In case of a non-conformity being detected or following modification of the model, the FMEA is revised. Revision is carried out in work meetings by a re-convened work group, gathering together the same expertise as for the original study and led by the FMEA leader.

The summary of the causes plus failure modes from the last update will be used as support material. For each cause plus failure mode, list the actions taken, the person responsible for each of them, the time frame and their date of application. Evaluate the new risk priority number, taking into account the actions taken (generally O and/or D change, unless the product has been redesigned; S does not change), using the same scale as for the original study.

Describe new recommended actions if those taken do not reduce the risk priority number sufficiently. Use the old recommended actions when no action has been taken. List the increases in risk priority number in relation to the last update by carrying out an operation-by-operation examination in co-operation with the members of the work group.

Write a revision report (using the same basic outline as for the original study). The summary of causes plus failure modes with recommended actions (attached as an appendix) is replaced by a summary of causes plus failure modes with the actions taken and the new recommended actions .

The histogram of causes plus failure modes for each class of risk priority number is updated and attached as an appendix to the revision report.

Distribute this revision report and comment on it in a summary meeting (process identical to the original study).

Update the FMEA dossier by completing:

ƒ{ The page on the administration of updates. ƒ{ The FMEA tables for the study (action taken, person responsible, time taken, dates of application of actions taken, new risk priority number, new recommended actions if appropriate).

Note: The FMEA studies are updated periodically and systematically, and the dates of update are given as described above including when there has been no change.

Machine FMEA

1 - PURPOSE

This document sets out how to apply the method for a Failure Mode and Effects Analysis to a machine, machine assembly or part of a machine.

The purpose of the Machine FMEA is to analyze the potential modes of failure of a machine, caused by its design and definition.

This technique of preventative analysis makes it possible to do the following: ƒ{ Look for machine malfunctions caused by its design and definition. ƒ{ Evaluate their potential effects on the machine¡¦s user. ƒ{ Identify the possible causes connected with the design and definition of the machine. ƒ{ Look for pertinent preventative actions.

2 - SCOPE

This method of analysis applies to any machine, part of a machine or machine assembly, both existing and in the project stage.

3 - DEFINITIONS

DATE OF ORIGINAL STUDY:

This is the date the study specifications were written.

REQUESTER:

Person who, as part of his or her functions, has the power to request that an FMEA be carried out, in agreement with the "decider".

DECIDER:

Person who, as part of his or her functions, has the power to employ the human and financial means necessary to carry out the FMEA and to implement the plan of action resulting from it. The requester and the decider can be the same person.

FUNCTION:

Service to be provided by each element of the machine

Example: Load covers onto the hooks of the chain.

FAILURE MODE:

Malfunction of each element of the machine, leading to a partial or total failure to satisfy its function. It is judged in relation to each function. There are one or more potential failure modes for each function. Example: Machine not loading.

EFFECTS:

These are the possible consequences of the failure mode on the sub-system and/or the system in use, as perceived by the machine¡¦s user. Example: Machine downtime, non- conforming products are produced, and/or the safety of employees is at risk.

CAUSES:

These are the circumstances, which provoke, or can provoke, the failure mode being considered. Their description varies depending on the purpose of the study. The following two cases are given as an example: ƒ{ 1st Case: The purpose of the study is to make the design and definition of the machine more reliable: these are the causes relating to the choice and definition of each element of the machine and of its contacts with the others. They concern the material (nature and quantity), form, position and identification of the machine which can provoke a failure mode. ƒ{ 2nd Case: The purpose of the study is to improve the availability of the machine: these are the causes related to the failure of elements of the machine.

There can be several possible causes for a failure mode. Likewise, one cause can generate several failure modes. These causes can be: ƒ{ Independent: A cause generates a failure mode for itself. Example: Poorly defined shape. ƒ{ Combined: Several causes must arise together for a failure mode to be generated. Example: Poorly defined shape and position.

CONTROLS:

These are planned actions and/or the existence of initial signs and/or warnings, which could prevent a failure from reaching the user of the machine.

The controls to be taken into account vary depending on the purpose of the study. Using the same examples for the description of causes, the controls would be:

ƒ{ 1st case: All the control measures, which exist during the machine, design stage (before it is produced). Example: Verification by calculation using finite elements, trials, tests.

ƒ{ 2nd case: All the means existing when the machine is used. Example: Preventative inspections, warnings when initial signs appear (noise, smoke etc.), call via beeper.

SEVERITY(S):

This is the severity of the consequences of the potential effects as felt by the user of the machine. Severity is scored from 1 to 10 using a pre-determined scale, which depends on the purpose of the study. This scale must be suitable for the subject being dealt with. It must be able to cover both the least serious consequences (score = 1) and the most serious (score = 10).

A failure mode can have consequences of varying degrees of severity. The score it is given corresponds to the most serious consequences, which could reasonably be observed.

OCCURRENCE (O):

This is the probability that the potential cause of failure will occur and that it will bring about the failure mode being considered. This is irrespective of the severity of the consequences. Occurrence is scored from 1 to 10 using a pre-determined scale, which depends on the purpose of the study. This scale must be suitable for the subject being handled. It must be able to cover both the least frequent consequences (score = 1) and the most frequent (score = 10).

DETECTION (D):

This is the probability that the failure will reach the user of the machine. Detection is scored from 1 to 10 using a pre-determined scale, which depends on the purpose of the study. This scale must be suitable for the subject being handled.

RISK PRIORITY NUMBER (RPN):

Hierarchical classification index resulting from the product of the different scores for severity, occurrence and detection allocated to each cause plus failure mode. RPN = S x O x D. RPN can be between 1 and 1000.

EMERGENCY ACTIONS:

These are the actions and/or plans of action which, once a machine failure has appeared, make it possible to reduce the severity of the consequences of this failure for the user of the machine.

RECOMMENDED ACTIONS:

Nature of the action(s), which will allow the probability of the cause plus failure mode appearing to be reduced, is the priority. Then, those actions, which will allow the probability of detection to be increased to the minimum necessary, are the next priority. They must ensure that the system is made permanent.

4 - APPLICATION OF THE TOOL

The application of the Machine FMEA must be led by a validated FMEA leader and requires methodical preparation. The purpose of the study must be clearly defined.

4.1 - PREPARATION OF THE FMEA

Before the FMEA is carried out, the leader must:

ƒ{ Acquire knowledge of the machine and its operation. ƒ{ Draw up the specifications for the study with the help of the requester, as the study is prepared. The standard printout in Appendix 14 can be used. ƒ{ Define the study limits (all machines, machine or part of a machine). ƒ{ Define the area of effect (determine in relation to what the severity will be expressed). ƒ{ Break down the system to be analyzed into ¡§equipment¡¨ or ¡§operational¡¨ sub-assemblies or into components, depending on the level of detail required for the purpose of the study. Draw up a list of sub-assemblies or components. ƒ{ Estimate the number of meetings needed to carry out the study. Set out the frequency (min. 1, max. 5 per week) of the meetings (recommended duration = 3 hours), taking into account the deadline for the end of the study and the constraints placed on those taking part. The FMEA must be carried out in a maximum of 6 months, otherwise break it down into sub-studies each lasting 6 months maximum. Obtain the agreement of the requester and decider as regards the workload to be undertaken. ƒ{ Create the work group, gathering together all knowledge of the system being analyzed. ƒ{ In agreement with the requester and the decider, determine who shall be responsible for the FMEA dossier. ƒ{ Determine the study characteristics in order to give all the necessary details on the context of the study: ƒ{ The different equipment making up the system being studied, ƒ{ The automatic failure detection systems, ƒ{ Existing preventative actions, ƒ{ Other particularities not covered by the above.

4.2 - CARRYING OUT THE FMEA

The FMEA is carried out by the work group led by the FMEA leader, and takes place during work meetings (number and duration set out in the study specifications). It follows the pattern below:

4.2.1 - Kick-Off of the Work Team

The participants are informed of the study to be carried out and how it will be done, and are trained in FMEA during the first meeting of the work group.

4.2.2 - Description of the Functions, Failure Modes, Cause and Effects, Detection and Classification

Use a table for each element of the system analyzed (sub-assembly or component) and hold a brainstorming session within the work group to list:

ƒ{ The FUNCTIONS of the sub-assembly or component being analyzed. ƒ{ For each function, the possible FAILURE MODES and potential EFFECTS. ƒ{ The possible CAUSES which can generate the failure mode being considered. These will be more or less detailed depending on the degree of detail required. ƒ{ For each cause: ƒ{ The existing DETECTIONS. If no means of detection exists, write ¡§None.¡¨ ƒ{ The existing EMERGENCY ACTIONS.

4.2.3 - Evaluation of Risk Priority Number

Once the failure mode and effects analysis of all the sub-assemblies or components has been carried out, evaluate the Risk Priority Number of each cause plus failure mode, using the scale given in the study specifications:

SEVERITY:

This is different for each failure mode, whatever its causes. In the impossibility to evaluate the severity, the maximum score (10) is given. Specify this reason in the comments column in the table (to remove any doubt later on). If there is any doubt as to which of two scores to give, give the higher one.

Evaluate the severity without taking emergency actions into account, and then repeat the process taking existing emergency actions into account. Enter the scores in the appropriate columns.

OCCURRENCE:

This is different for each cause plus failure mode. First estimate the frequency with which the cause occurs, then the probability that the cause will generate the failure mode. The occurrence score is obtained by the calculation: frequency of occurrence of the cause multiplied by the probability that the cause will generate the failure mode.

Take into account the preventative controls, which stop the cause from appearing.

If there is any doubt as to which of two occurrence scores to give, give the higher one. If the occurrence is not known at all, the maximum score (10) is given. Specify this reason in the comments column in the table (to remove any doubt later on).

Enter the score in the appropriate columns, without emergency actions and with emergency actions.

DETECTION:

This is different for each cause plus failure mode. Estimate the risk of the failure reaching the user of the machine, taking into account initial signs and existing warnings.

If there is any doubt as to which of two detection scores to give, give the higher one. If the detection is not known at all, the maximum score (10) is given. Specify this reason in the comments column in the table (to remove any doubt later on).

Enter the score in the appropriate columns, without emergency actions and with emergency actions.

RISK PRIORITY NUMBER:

For each cause plus failure mode. Multiply the three scores S, O and D and enter the result in each of the appropriate columns, without emergency actions and with emergency actions.

4.2.4 - Hierarchical Classification and Description of Corrective Action

Once the risk priority number of each cause plus failure mode with emergency actions has been evaluated, the leader classifies them in descending order of risk priority number.

In subsequent meetings, the work group, together with the Project Leader: ƒ{ Verifies that the order of problems is logical (possible deviations between the different risk priority number evaluation meetings) and corrects those risk priority number values, which are recognized as being incorrect. ƒ{ Determines the risk priority number thresholds beyond which preventative actions will be recommended. These are set based on the S, O, D to be taken into account to achieve the purpose of the study. ƒ{ Makes a precise description of the nature of the actions recommended in the summary for each cause plus failure mode for which the risk priority number is higher than the thresholds determined previously . The choice of the nature of the recommended actions is made based on first S, then O, then D. ƒ{ Prepares the conclusion (the main outstanding facts arising from the FMEA) which will appear in the study report.

4.2.5 - Study Report

Makes a report on the FMEA once the study is complete, to include: ƒ{ A summary of the causes and purposes of the study, ƒ{ A summary of the study specifications, ƒ{ A summary of the way the study was carried out (composition of the work group, list of the sub-assemblies or components analyzed, the cost and duration of the study), ƒ{ The results of the study: ƒ{ Histogram of causes plus failure modes for each class of risk priority number (attach as an appendix): ƒ{ Summary of the causes plus failure modes with the actions recommended (attach as an appendix): ƒ{ The conclusion summarizing the main facts brought to light by the analysis. ƒ{ The scale used to evaluate the risk priority number is attached to the report.

After agreement with the study¡¦s requester, the report is distributed to the members of the work group, their managers, the requester and the decider. It can also be commented on in a final summary meeting held within a few days of its distribution. This meeting marks the end of the original study.

The list of recommended actions will be used when compiling operational plans, depending on the objectives set out elsewhere.

4.2.6 - FMEA Dossier

The FMEA dossier includes: ƒ{ A page on the administration of updates, ƒ{ The study specifications, ƒ{ The study report, ƒ{ The reports on revisions, ƒ{ The FMEA tables of all the sub-assemblies or components analyzed.

It is passed to the person delegated as being ¡§responsible for the dossier¡¨ in the study specifications.

4.2.7 - Monitoring and Revision

In case of a non-conformity being detected or following modification of the model, the FMEA is revised.

Revision is carried out during work meetings by a work group gathering together the same expertise as for the original study and led by the FMEA leader, with the participation of the Project Leader.

The summary of the causes plus failure modes from the last update will be used as support material. For each cause plus failure mode, list the actions taken, the person responsible for each of them, the time frame and their date of application. Evaluate the new risk priority number, taking into account the actions taken, using the same scale as for the original study. Describe new recommended actions if those taken are not sufficient to achieve the risk priority number objectives. Use the old recommended actions when no measure has been taken. Where risk priority number thresholds fixed at the onset do not allow the purpose of the study to be achieved, or if additional progress has become necessary, set new thresholds and describe the resultant new recommended actions. List the increases in risk priority number in relation to the last update by carrying out a sub-assembly-by-sub-assembly or component-by-component examination in co-operation with the members of the work group.

Write a revision report (using the same basic outline as for the original study). The summary of causes plus failure modes with recommended actions (attached as an appendix) is replaced by a summary of causes plus failure modes with the actions taken and the new recommended actions . The histogram of causes plus failure modes for each class of risk priority number is updated and attached as an appendix to the revision report. Distribute this revision report and comment on it in a summary meeting (process identical to the original study).

Update the FMEA dossier by filling in: ƒ{ The page on the administration of updates, ƒ{ The FMEA tables for the study (measure taken, person responsible, time taken, dates of implementation of actions taken, new risk priority number, new recommended actions if appropriate).

Note: The FMEA studies are updated periodically and systematically, and the dates of update are given as described above including when there has been no change.