The Psychology of Quality and More
Failure Mode and Effects Analysis (FMEA)
What's it for?
Failure Mode Effects Analysis, or FMEA as it is commonly called, is a simple method for finding out the real cost of potential failures in any product or system.
FMEA can be used during design or later analysis of a product or process to help identify potentially significant failure risks. For example, an engine casing may be found to be at risk of cracking under harsh vibration or an order entry system may lose customer details if the wrong computer key is pressed.
It is a scalable tool that can be used to examine failures in complete systems, subsystems or on individual components. The level and depth of analysis should depend on what is being examined and on the importance of finding all key risks.
How does it work?
Failures in products or processes can often occur in unexpected and unpleasant ways, such as when an electrical failure results in your car catching fire and burning to the ground. Such potentially catastrophic problems may be identified before they occur by using a simple series of questions.
First of all, the item that can fail is selected. Then the ways in which it can fail (the failure modes) are identified simply by asking, 'How can this fail?'. Finally, the real cost of potential failure is identified by asking 'What could happen as a result of this failure?. This is the failure effect which results from the failure mode.
Failure Mode and Effects
Deciding on which failure effect is the most important is not always easy, so 'criticality' measurement is often used. This simply includes a consideration the probability of the effect occurring, often in combination with other measures of importance, such as a customer-allocated severity figure. FMECA (where the C stands for 'criticality') is the proper name, but it is still often referred to with the generic name of FMEA.
How to do it
1. Select the item to be analysed. This can be as large as a car or as small as a door catch. Of course, you cannot analyse the car down to the level of detail that you can achieve when examining the catch, so it is sensible to focus your attention on a specific area of interest.
2. Identify the scope of the examination. For example, you may be looking at what may happen in a crash or fire, or just in normal wear and tear.
3. Ask, 'What may fail?'. This can be components, sub-assemblies, coatings, etc.
4. Find out how likely each item from step 3 is to fail. You can estimate, but this may well lead to faulty conclusions. If you want a definitive answer, this may well required extensive experimentation (or field failure data).
5. In the experiments of step 4, identify the different ways each item may fail (ie. the failure modes).For example, a wire can split, melt, crack, etc.
6. Find the probability of each failure mode occurring, given that the item from step 4 has failed.
7. For each failure mode in step 5, find the different actual results that the failure mode can cause. For example, a molten wire can cause fire or power loss.
8. Find the probability of each failure effect occurring, given the failure mode from step 7 has occurred.
9. Calculate the criticality of each failure effect by multiplying the probability figures from steps 4, 6 and 8.
10. Optionally, modify the criticality figure by multiplying by additional weighting factors, such as a customer-allocated severity figure. This will allow better comparison of failure effects and better decision-making in step 12.
11. Optionally, add the criticality for each failure effect across modes or parts. For example, 'power loss' may be caused by melting wire and by failed components.
12. Examine the final figures and identify specific areas for action.
Next time: The Tree Diagram
This article first appeared in Quality World, the journal of the Institute for Quality Assurance
And the big