The principle
Instead of changing parts at set intervals, or waiting for them to break, a parameter that reflects the health of the equipment is measured, and action is taken when that measurement departs from normal. The intervention is scheduled according to the real condition and not an average assumption.
What is monitored
| Variable | What it can reveal |
|---|---|
| Vibration | Imbalance, misalignment, bearing wear. |
| Temperature | Friction, high-resistance connections, overload. |
| Insulation resistance | Moisture, contamination or aging of the insulation. |
| Oil analysis | Internal wear and lubricant degradation. |
| Current | Load changes or defects in a motor. |
How a program works
- Choose the equipment where monitoring pays off: equipment whose failure is costly or dangerous and that gives prior warning signs.
- Establish a baseline while the equipment is healthy.
- Measure at intervals consistent with how quickly the variable degrades.
- Define alert and action levels, and what to do at each one.
- Analyze the trend, not just the isolated value.
What an engineer sees in practice
The trend is what is valuable. A value still within limits but growing steadily is a more useful alert than a stable high value. That is why it matters to always measure under the same conditions and with the same method.
When it does not work well
If the failure occurs suddenly, with no measurable prior degradation, monitoring the condition gives no timely warning. Nor is it worthwhile on cheap, easily replaced equipment, where the cost of monitoring exceeds that of the breakdown.
Nor does it work if the measurements are not repeatable: changing the instrument, the measurement point or the load condition between readings introduces variations that are mistaken for real deterioration and generate false alarms or, worse, false reassurance.