FMEA without the forty-page spreadsheet

An FMEA asks three questions about a piece of equipment: how can it fail, what happens when it does, and how likely are you to catch it first. Failure modes and effects analysis has a reputation for producing enormous spreadsheets nobody reads. Done narrowly it takes a morning and changes the maintenance schedule the same week.

The five columns you actually need

• •Function: what this equipment is for, in one line
• •Failure mode: how it stops doing that, specifically. Not "pump fails" but "seal leaks", "bearing seizes", "impeller wears"
• •Effect: what happens to the building or the line when it does
• •Cause: why that failure mode occurs
• •Current control: what you do today that would catch it, if anything

How the risk number works

Each failure mode gets three scores from 1 to 10: severity of the effect, how often it occurs, and how likely you are to detect it before it bites. Multiply the three and you get a risk priority number between 1 and 1,000.

A seal leak scored severity 5, occurrence 6, detection 7 gives 210. A catastrophic but well-monitored failure scored 9, 2, 2 gives 36. The second is more dramatic and the first is where your attention belongs.

The number is the least useful part

Do not rank a hundred rows by RPN and work down the list. The scores are judgements and a difference between 210 and 190 means nothing. What the exercise is really for is the conversation, and two things always come out of it.

• •Failure modes with high severity and no detection at all, which is where a new check belongs
• •Existing checks that address a failure mode nobody has ever seen, which is where an hour comes back

Detection is the score people get wrong

A high detection score means you probably will not catch it in time. That feels backwards and it trips up almost every first-time team. Say it out loud each time: 10 means invisible until it fails, 1 means obvious well beforehand.

Do it narrow, and do it with the right people

One critical asset, ninety minutes, with a technician who works on it, an operator who uses it, and somebody who takes notes. Whole-plant FMEA programmes are how this technique got its reputation. One machine at a time is how it actually pays.

Turn it into schedule changes the same week

The output is not the spreadsheet. It is a short list of checks to add, checks to remove, and spares to hold. If nothing in the maintenance schedule changed within a week, the session was a meeting.