In almost every plant, a handful of assets cause most of the trouble. In the example area of this series, the ten worst assets caused 71 percent of 212 hours of unplanned downtime in three months. The bottleneck machining center alone caused 17 percent.
Improvement starts with knowing these assets by name.
The bad actor list
Sort the unplanned downtime per asset over the last three months from high to low and calculate the cumulative share:
| Rank | Asset | Downtime (h) | Share | Cumulative |
|---|---|---|---|---|
| 1 | Machining center 7 | 36 | 17 % | 17 % |
| 2 | Washing machine 3 | 28 | 13 % | 30 % |
| 3 | Machining center 4 | 22 | 10 % | 40 % |
| 4 | Press 1 | 16 | 8 % | 48 % |
| 5 | Test bench 2 | 12 | 6 % | 54 % |
Update the list every month. Each of the top three gets an improvement action with an owner. A list that does not change for months means the actions are not working.
When to dig for the root cause
Start a root cause analysis when:
- the same failure mode hits the same asset three times in six months,
- a single event causes more than eight hours of downtime or touches safety,
- an asset stays in the top three of the bad actor list.
Use a defined method: describe the failure with facts, find the cause with 5 Why or a cause and effect diagram, confirm it at the asset, fix it, and change the strategy or the design. The Problem Solving module of the Production System describes the method in depth.
In the example, the encoder cable of the bottleneck machine broke four times in a year. The root cause was a bend radius below the specification of the cable. A new routing took 18 hours and removed the failure mode.
Lubrication first
Many bearing failures start with lubrication: the wrong lubricant, the wrong quantity, contamination. Before investing in sensors, check the basics: a list of lubrication points with lubricant, quantity and interval, color codes at points and containers, clean storage and filtered filling. The example plant had 1,240 lubrication points in 14 routes.
Condition monitoring with limits
| Method | Detects | Typical use |
|---|---|---|
| Vibration | Bearing wear, imbalance, misalignment | Spindles, pumps, fans, gearboxes |
| Thermography | Loose connections, overload, friction | Control cabinets, motors |
| Oil analysis | Wear particles, contamination | Hydraulics, gearboxes |
| Ultrasound | Leaks, early bearing damage | Compressed air, slow bearings |
Every measuring point needs a warning limit, an alarm limit and the job that follows. A measurement without a limit and an action is a hobby. Predictive models come later, on the few assets where continuous data and expensive failures justify them.
The whole system
Criticality decides where effort goes. Strategy decides what to do per failure mode. The work order path, planning and the weekly schedule turn that into prepared jobs. Spare parts and skills make the jobs possible. Key figures and the bad actor list show what works and where to dig next. Together they turn a maintenance team from firefighting into a team that decides which failures it accepts.