Definition
Predictive Maintenance is a maintenance strategy in which the timing of a maintenance task is predicted from the actual condition of a machine. Sensors continuously record operating data such as vibration, temperature, pressure or power consumption. Algorithms detect deviations from the normal state and estimate when a component is likely to fail – its remaining useful life (RUL).
The core principle: maintenance happens when it is needed. Not on a calendar, and not after the damage is done.
How it differs from other maintenance strategies
| Strategy | What triggers maintenance | Data basis |
|---|---|---|
| Reactive maintenance | The failure itself | none |
| Preventive maintenance | A fixed schedule or operating-hours counter | experience, manufacturer specifications |
| Condition-based maintenance | A measured value exceeds a threshold | real-time values |
| Predictive maintenance | A forecast that a threshold will be exceeded | real-time values plus historical data plus a model |
The distinction between condition-based and predictive maintenance is often blurred, but it matters:
condition monitoring tells you that something is currently abnormal.
Predictive maintenance tells you when that will turn into a failure.
Typical data sources
- Vibration
- Temperature
- Power consumption
- Pressure and flow
- Control and axis data
Open standards such as OPC UA and MQTT have become established for transmitting this data.
Related terms
Condition Monitoring · OEE · Industrial Internet of Things (IIoT) · Edge Computing · Retrofit
In practice: How predictive maintenance is actually implemented on existing CNC and cutting machines – retrofitModernization and upgrading of existing machines with new technology for digitalization without complete replacement. options, costs and limitations – is covered in our practical guide: Predictive Maintenance for CNC and legacy machines