Also called DFMEA, PFMEA, AMDE, risk priority number
What is failure mode and effects analysis?
Working through every way a design or process could fail, before it does, and fixing the ones that matter most.
What else it is called
- Other names
- DFMEA, PFMEA, AMDE, risk priority number
- In French
- analyse des modes de défaillance et de leurs effets
Related entries
Understanding failure mode and effects analysis
A team lists the ways each step could fail, what the consequence would be, and what would cause it. Each one gets three scores: severity, how bad the effect is; occurrence, how likely the cause is; and detection, how likely you are to catch it before it escapes. The three are multiplied into a priority number used to decide what to work on. Newer versions replace that single number with an action priority table, because multiplying three ordinal scores produces some odd rankings.
The value is almost entirely in who is in the room. A maintenance technician, a setter and a quality inspector together will produce a list of realistic failures in an afternoon. The same document produced by one engineer to satisfy a customer requirement will be thorough, plausible and of no use to anybody.
One specific habit worth taking from it: detection scores should be honest. "The operator would notice" is usually optimistic, and rating it as reliable detection is how a known failure mode ends up at a customer.
In practice
Severity cannot be designed down by better inspection. If the consequence of a failure is a safety recall, the only acceptable answer is making the failure impossible, not catching it more often.
Frequently asked
What is the difference between a design and a process version?
The design version asks how the product itself could fail in the customer’s hands. The process version asks how your manufacturing could produce a part that does not match the design. Automotive customers usually want both, and they are written by different people.
What replaced the risk priority number?
The harmonized automotive standard moved to action priority, a table that looks at severity, occurrence and detection together rather than multiplying them. It avoids the problem of a high-severity failure getting a low score because it is rare.