Quick answer: 82% of preventive maintenance programs fail to reduce failures because they're time-based instead of condition-based. The Nowlan-Heap reliability research showed that only 11% of equipment shows age-related wear; the other 89% fails on random patterns that a calendar can't predict. The fix is condition-based PM driven by OEE signals or sensor data.
Related deep-dives: improve PM compliance · PM optimization software · eliminate PM overruns · closing the OEE-CMMS loop.
If you ask a Maintenance Manager why a machine broke, they often say:
"It was old. We should have serviced it sooner."
This assumes a direct link between Age and Failure.
It assumes that machines are like car tires—they wear out smoothly over time.
If this were true, Preventive Maintenance (PM)—servicing machines on a calendar—would prevent 100% of breakdowns.
But it doesn't. Machines still break the day after a service.
Why?
Because the "Age Theory" is wrong.
According to the foundational studies of Reliability Centered Maintenance (Nowlan & Heap), only 18% of assets fail due to age.
The other 82% fail randomly.
This is the 82% Rule. If you are building your maintenance strategy entirely around Calendar PMs, you are using the wrong tool for 82% of your problems.
Here is why "More Maintenance" isn't the answer, and how to fix the random failures.
1. The 6 Patterns of Failure
Engineering studies classify failure into 6 patterns. Only one of them looks like "Wearing Out."
-
The Bathtub (4%): High failure at start, low middle, high at end.
-
Wear Out (2%): Constant until a sudden increase at the end.
-
Fatigue (5%): Slowly increasing failure probability.
The vast majority (Patterns D, E, F) show Random probability.
-
What this means: A motor is just as likely to fail on Day 100 as on Day 1,000.
-
The Cause: Random failures are caused by Events, not Time.
2. Why Calendar PMs Fail the 82%
If a machine fails randomly (e.g., due to a voltage spike), changing the oil every month does nothing to prevent it.
In fact, Calendar PMs can hurt you.
The "Intrusion" Risk:
Every time a technician opens a machine to do a PM, there is a risk they will:
-
Strip a bolt.
-
Introduce dirt.
-
Bump a sensor.
This creates "Infant Mortality" in an old machine. You took a healthy asset, opened it up to "Maintain" it, and accidentally introduced a defect.
For random-failure assets, "Hands-Off" is often the best policy—until the data says otherwise.
3. The Strategy Shift: From "When" to "If"
For the 18% of assets that wear out (Tires, Belts, Brake Pads), keep using Calendar/Usage PMs.
For the 82% of assets that fail randomly (Electronics, Hydraulics, Pneumatics), you need Condition Monitoring.
You stop asking: "When is it due?"
You start asking: "Is it healthy right now?"
The Digital Approach:
-
Monitor the Variables: Random failures leave clues. Heat. Noise. Vibration.
-
Connect the Data: Use Fabrico to read the PLC tags (Amps/Temp).
-
The Trigger: Instead of a "Monthly Service," set an alert: "If Temp > 60°C, Inspect."
This allows you to catch the Random Event (the voltage spike or the jam) the moment it happens, rather than waiting for next month's schedule.
4. How Software Manages the Mix
You cannot manage a complex factory with a simple calendar. You need a Hybrid System.
How Fabrico handles the 82% Rule:
-
For Wear Parts (The 18%): We use Cycle Counts. "Conveyor Belt has run 10,000 hours. Replace."
-
For Random Parts (The 82%): We use Trend Analysis. "Motor current is trending up. Something changed. Investigate."
This moves you from "Blind Maintenance" (guessing) to "Evidence-Based Maintenance."
Conclusion: Stop Disturbing Healthy Machines
The goal of maintenance is reliability, not activity.
If you are over-maintaining healthy machines "just in case," you are wasting money and introducing risk.
Respect the 82% Rule. Listen to the machine, don't just look at the calendar.
Switch to condition-based.
[Request a Demo] and let Fabrico help you move from Calendar to Condition.
Zaloguj
Blog
