Why Your FMEA Is Sitting in a Binder (And What It Should Be Doing Instead)

There’s usually a three‑ring binder somewhere on site with “FMEA” (Failure Mode and Effects Analysis) on the spine. It was built during a shutdown or a quality push, printed, filed, and likely forgotten while the day-to-day work on site kept moving.

Needless to say, an FMEA only matters if it changes decisions on the ground. And the ground gets expensive fast. In ABB’s “Value of Reliability” survey of 3,215 maintenance decision‑makers, over two‑thirds of industrial businesses said they experience unplanned outages at least once a month, with a typical hourly cost close to $125,000. Yet 21% still rely on run‑to‑fail maintenance.

In this article, we’ll delve into the importance of FMEA, and how you SHOULD be using it on a daily basis, including:

• What FMEA Does
• The Purpose of a FMEA
• When a FMEA doesn’t get used
• FAQ

What FMEA Does

Failure Mode and Effects Analysis (FMEA) is a step‑by‑step way to identify failure modes, understand their effects, and prioritize what to prevent first. “Failure mode” is plain language: how a thing fails – belt mistracking, plugged chute, cracked weld, wrong lockout point, missed lubrication. Most teams rank each item with a risk priority number (RPN), often Severity × Occurrence × Detection.

Two distinct branches of FMEA matter in industrial operations:

• DFMEA looks at design decisions: a new crusher, a retrofit, a guard redesign, a dust collector upgrade.
• PFMEA looks at how you run and maintain: start‑up/shutdown, PM routines, inspection points, handovers, and material flow.

These are formalized heavily in sectors like automotive (AIAG & VDA publish a DFMEA/PFMEA handbook), but the thinking is universal and applies to any type of industrial setting.

The Purpose of a FMEA

A living FMEA is not a compliance tick-box exercise. It’s a prioritized list of “how this plant breaks,” connected to controls your crew can execute. Keep it simple. Start with one constraint asset (primary crusher line, batch plant, your worst conveyor) and do this:

1. List failure modes in shop terms. “Tail pulley bearing cooks.” “Belt slices on skirt steel.” “Hydraulic hose rubs through.” “Screen blinds in wet feed.”
2. Make Scoring Detection honest. If you only catch the problem when you smell burning rubber, Detection is poor. If a five‑minute walkaround, a temperature‑gun route, vibration check, or oil sample catches it early, score it better.
3. Use RPN to rank work, then force an action. An RPN number without a control is just math. Convert your top items into something real, such as an added inspection point, a changed PM interval, a stocked seal kit, a training step. Whatever closes the gap between the score and the fix.
4. Don’t let RPN hide severity. High-severity safety and compliance items deserve action even if occurrence feels 'rare.' A low probability doesn't make the consequence acceptable.
5. Close the loop after breakdowns. If the same failure mode hits twice, update occurrence and update the control plan. If you solved it, capture what worked so the fix survives the next crew and the next shift.
6. Make it accessible where the work happens. If your FMEA, PFMEA, or DFMEA lives in a spreadsheet on someone’s desktop, it won’t survive shift change. The highest-risk failure modes should show up inside daily inspections, PM tasks, and deviation reports – on the tablet in the loader cab or the screen in the shop. When crews can see the top risks, log photos, and update actions in the same place they document the work, and the analysis with resolutions stays alive.

When a FMEA Doesn’t Get Used

Most site FMEAs die because they’re disconnected from work. They’re written in “meeting room language,” they don’t tie to pre‑shift checks and PMs, and nobody owns updates after the next changeout or workaround.

And that’s where most FMEAs fall apart.

If high-risk failure modes aren’t built into daily inspections, they get forgotten. If repeat breakdowns don’t automatically trigger a review of Occurrence scoring, the RPN never changes. If deviations live in one system and PMs in another, nobody sees the pattern forming.

A living FMEA needs to sit where the work actually happens – in the same flow as pre-shift checks, workplace exams, PM tasks, and incident reporting. When a tech logs a failed bearing with a photo, that data should feed the risk picture. When a supervisor closes a deviation, it should affect how Detection is scored next time.

That’s when FMEA stops being a workshop exercise.