[Total: 0    Average: 0/5]

The Failure Mode and Effect and Criticality Analysis (FMECA) makes it possible to ensure the operational safety of the processes and the products by the control of the failures.

Introduction

According to the Afnor X60-510 standard, FMECA is an inductive system analysis method used for systematic investigation of causes and control of the effects of failures likely to affect the components of this system. The FMECA is performed for each operating phase of the system and assigns a criticality note to the identified effects.

Historical

It was the American army that developed the FMECA in the years 1940. The military reference to the procedure was MIL-P-1629, entitled ” Procedures for analyzing failure Modes, their effects and criticity” dated November 9, 19491. This method was used as a failure assessment technique to determine the reliability of an equipment and system. The failures were then classified according to their impact on the success of the missions and on the safety of the personnel and equipment. The wide spread of the tool will only be a few decades later.

In 1988, the International Organization for Standardization issued ISO 9000 series standards. This new standard pushed companies to develop the formalisation of quality system management. QS 9000, the equivalent of ISO 9000 for the automotive sector, was developed by a working group represented by Chrysler, Ford and General Motors. In an issue of standardising the quality systems of suppliers, they developed the Advanced Product Quality Planning (APQP), including the FMECA tool and developing control plans.

The automotive industry Action Group (AIAG) and the American Society for Quality Control (ASQC) emit the FMECA standards in February 19932. The standards are presented in a FMECA manual approved and supported by three car manufacturers. This manual provides the general principles for preparing a FMECA. These pages are intended to understand the use of FMECA in the American automotive industry. A FMECA is defined as “a systematic process to identify potential modes and treat failures before they occur, with the intention of eliminating them or minimizing associated risks.

FMECA, FMEA…

The FMECA tool has several possible denominations3 . The FMEA is a version of the FMECA where the criticality of each failure mode is not measured. It translates into FMEA, Fealure Modes, effects and analysis.

The FMEA does not allow a prioritization of the different modes of failure. It just allows you to do a preliminary work on the analysis and is used primarily when you want to take only one of the 3 measurement indicators.

The principle: The analysis of “failures”

By default, we mean that a system:

  • does not work
  • Does not work as expected
  • Works unexpectedly
  • Works but at an insufficient performance level

Of these failures, we identify the modes of failure, which are the different ways in which a failure occurs. We can find the leaks, the stops…

The FMECA evaluates the criticality of the different modes of failure via 3 criteria:

  • Occurrence: the probability of occurrence of the events.
  • Detectability: our ability to detect the problem when it appears.
  • Severity: the impact of the problem on quality, performance…

Step 1: Prepare the project

  1. Create the team: the FMECA requires teamwork to have complete information and a global view of the elements judged. To do this, the team must be composed of a pilot in charge of guiding the group and ensuring the proper follow-up of the methodology, and of people with a mastery of the product or process studied.
  2. Define the perimeter: a FMECA will be complex to manage when there is too much information. It is necessary to limit the study to a well defined framework from the outset of a project.

Step 2: Develop the rating grid

The first step is to develop your own rating grid. In fact, depending on your subject and its complexity, the rating grid is not the same. We assume that:

  • The lowest note indicates either no severity or little likelihood of occurrence or is readily detectable.
  • The highest score indicates either a significant severity, a high probability of occurrence or an almost impossible detectability.

Some ways of thinking:

  • If variables are all easily measurable variables, the grid can be accurate, with a scale of 1 to 10 increment 1.
  • If the variables are more difficult to measure, it is advisable to take a simplified scale of type 1, 5, 10 or 1, 3, 6, 9.
  • To make the differences significant, it is necessary to set up grids with non-regular increments (type 1, 3.9).
  • It should never be a scale starting from 0, in which case the criticality will be null which will result in not treating this failure.

Level

Severity

Occurrence

Detectability

1

No discernible effect

The failure is eliminated via preventive control

The cause of the failure cannot appear because eliminated by preventive design solutions

2

Minor disturbance (noise…), less than 25% of customers note

No failures observed with a design of this type or during simulations and tests.

1 default for 1 000 000

The design or control allows a simple detection

3

Minor disturbance, less than 50% of clients note

Only isolated failures with a design of this type

1 Default for 100 000

Reliability tests with degradation measures were done before the design freeze

4

Significant disturbance, more than 75% of customers note

Isolated failures with a design close to it

1 default for 10 000

Destructive tests were done before the design freeze

5

Performance on secondary functions is reduced (comfort of the car…)

Occasional failures with identical design

1 Default for 2 000

Pass/pass tests were done before the design freeze

6

Loss of secondary functions

Frequent failures with identical design

1 Default for 500

Reliability tests with degradation measurements were made after the design freeze but before the production launch

7

Problems with primary functions are present

The new design is uncertain, the production conditions need to be changed

1 Default for 100

Destructive tests were done after the design freeze but before the production launch

8

Primary functions are inoperable but do not affect security

The new design causes many faults, the production conditions must be changed

1 Default for 50

No pass/pass tests were done after the design freeze but before the production launch

9

Potential effects exist regarding safety or compliance with the regulations

Failures are inevitable with this new design, the production conditions must be changed

1 Default for 20

The analysis of the design and controls have a low capability.

10

Important effects exist regarding the safety or compliance with the regulations

We have no history on this new technology/design

1 Default for 10

There are no controls or they cannot detect faults

Source: QS 9000

Another factor

In certain sectors of activity where the concept of risks is particularly important (nuclear, pharmaceutical…), another criterion is sometimes used: knowledge.

At the most we have knowledge of the risk in question, at the most it can be noted correctly. The idea is to highlight the fact that to control the risk is therefore know to evaluate it correctly, we need to have a good knowledge of it.

At the end of the FMECA, we calculate the GAP of knowledge that allows to prioritize the actions. This Gap is calculated via: Severity * knowledge. At the most this score is high, the more we have to work on it to be sure to evaluate it correctly.

Step 3: Identify failure modes

For each of the failures, one or more failure modes are identified. In other words, we identify each of the possible variations that the failure may have. For example, for the filter lubricant function, there may be 2 modes of failure: improper filtration or clogging of the filter.

Experience shows that the risk list is “infinite“. There is no clear rule on the subject, apart from stopping by common sense and when consensus is achieved.

Step 4: Identify the effects of different modes of variation and their severity

For each of the failure modes, one or more effects are identified on the end-customer. At this point, we must ignore whether this can happen or not. The challenge is to make an exhaustive list of all the consequences of a process failure.

Then, for each of them, we identify the severity level of the failure mode. This can go from no severity, to a significant severity that can cause injuries see fatal accidents.

This notation will be done in 2 times:

  1. A first pass using the evaluation grid
  2. Compare the results to see if the hierarchy matches what you imagine.

 

Taking our example.

Purpose

Failure Mode

Effect

Gravity

Filter the Lubricant

Clogging

Machine shutdown

3

Bad filtering Pump Wear 2

Step 5: Identify potential causes and their likelihood of appearances

For each of the failure modes, we identify the different root causes possible. It is strongly recommended not to exhaustively enumerate possible causes, but only to list the clearly identified root causes. In which case, the FMECA will not be readable.

The probability of occurrence of the cause is estimated in terms of the rating grid. As before, this quotation will be a 2-time:

  1. A first pass using the evaluation grid
  2. Compare the results to see if the hierarchy matches what you imagine.

 

Taking our example.

Purpose

Failure Mode

Effect

Gravity

Causes

Probability

Filter the Lubricant

Clogging

Machine shutdown

3

Presence of impurity during filling

1
Bad filtering Pump Wear 2 Deterioration Strainer 1

Step 6: List the means of control and the level of detectability

For each of the cases, the means of control identified to remove the cause of the variation (Poka Yoké…) or detect it (Andon…).

As before, we will evaluate the level of detectability in 2 strokes:

  1. A first pass using the evaluation grid
  2. Compare the results to see if the hierarchy matches what you imagine.

Taking our example.

Purpose

Failure Mode

Effect

Gravity

Causes

Probability

Means of control

Detectability

Filter the Lubricant

Clogging

Machine shutdown

3

Presence of impurity during filling

1 Grid on filler cap 3
Bad filtering Pump Wear 2 Deterioration Strainer 1 Changing the strainer every quarter 3

Step 7: Calculating criticality

Finally, the last phase is to calculate the level of criticality. This level is calculated by multiplying the 3 criteria: C = severity * probability of occurrence * detectability.

Step 8: Prioritization

Following step 6, a table with a criticality level is obtained for each of the possible causes of failures.

To prioritize actions, it should be noted that a criticality level of 250 is not considered to be different from 260. Thus, from a scoring scale of 1 to 10 per criterion, the modes of failure are prioritized according to the table below:

Priority

Criticality

A

More than 200

B

100 at 199

C

26 to 99

D

1 to 25

In the second classification, the modes of failure in each group can be prioritized according to the level of each criterion and in the following order:

  1. Deal with the most important levels of severity
  2. Deal with the most important probability levels
  3. Deal with the most important detectability levels

Step 9: The Action plan

Depending on the situation, variables… the types of actions are different. Here are some tips:

  • First try to lower the level of occurrence and detectability by treating the cause. In fact, it is only in rare cases that you will be able to lower the severity level. This is called a preventative plan.
  • If the severity level can be reduced, it should be the first action to reduce the consequences. This is called a contingency plan.
  • If you do not know how to adjust the different variables to control the causes, you need to set up a plan of experiments.
  • If the measurement system does not seem to be reliable, you must do an MSA.
  • If you are not on the level of correlation between the cause identified and the failure mode, an analysis must be done.

Step 10: Criticality re-evaluation

As a result of the actions, the results are verified by measuring the criticality index of the failure modes. This new level is called ” residual risk “. It must match your expectations and reach a level of criticality below the threshold you have set yourself.

The FMECA sheet

The FMECA sheet is the support of all the previous steps. This sheet contains all the elements used to assess the level of criticality and to control corrective actions.

FMECA and business lines

HACCP (Hazard analysis Critical Control Point) for agri-food: A method derived from FMECA, it is used to prevent, eliminate or reduce to an acceptable level any biological, chemical or physical hazards.

FMECA ATEX (explosive atmospheres) for areas with explosive atmosphere: Since the introduction of the new ATEX directive, manufacturers of machinery used in explosive atmospheres must carry out an ATEX FMECA , which will allow To identify the risks of warm-up or sparks, irrespective of their origin.

AEEL (Analysis of errors and their effects on the software) for computing: derived from the FMECA, it is a method to prevent software failures as soon as they are conceived. It allows to refine the work of the teams in charge of the validation tests.

HAZOP (HAZard and OPerability Study): Method created in the years 70 by the chemical industries, it focuses on the operational risks associated with industrial installations, in particular thermo-hydraulic type systems. This method is based on an analysis of malfunctions based on a set of guide words (so-called “functional”) that apply to the physical parameters of a fluid.

The different types of FMECA

The principle of FMECA can be applied wherever one wishes to study the modes of failure (recruitment process, operation of Software…). However, 2 FMECA are most commonly used.

FMECA Process

Called PFMEA (Process failures modes and effects analysis), allows to determine the possible failure modes of the means of production that will affect:

  • The productivity of the line
  • The quality of the products
  • Production times
  • Production costs

 

The PFMEA will be used to help define the technical solutions for a new way. For each of these solutions or potential solutions, we will be able to identify the different modes of failure. The tool will:

  • Either to focus on a certain number of failures and thus to conceive at best the means.
  • Or to choose between several possible solutions that has the least mode of failure.
  • Build the monitoring plan and quality controls as well as preventative maintenance ranges.

FMECA Product

Called DFMEA (Design failures Modes and effects analysis), can be used to verify the viability of a product developed in relation to customer or application requirements. For each of the functions of the product, one can identify:

  1. What are the modes of failure on the fact that the functions of the product are not at the level expected by the customer.
  2. If the solutions selected do not present specific failures and therefore responds well to the customer need.

Limitations of the FMECA

  • It does not allow to have a cross-vision of possible failures and their consequences: Two failures occur at the same time on two subsystems, what is the consequence on the whole system? In this case further studies are necessary.
  • FMECA is a ” heavy ” tool when applied to complex systems. The scope of the study must be precisely determined at the outset to avoid being subjected to a volume of information too important to be manageable.
  • The quality of a FMECA is linked to the completeness of the identified failure modes. This is strongly dependent on the experience of the authors of the study.
  • In addition, the FMECA tool must not become an end in itself. The actions recommended must be implemented and a monitoring of their effectiveness must be ensured.

Source

1-F. A. Meyer (2014)-Apply the TOC Lean Six Sigma in the services

2-N. Dufour, G. Teneau (2013)-risk management, a Border object

3-D. DD (2010)-Running Safety Course

V. Ozouf (1992)-Design and produce “safe operation”

M. Ramamurthy (2000)- FMEA -Medium

I. Poullain, F. Lespy (2002)-risk and quality management

Share This