What is predictive maintenance?
Predictive maintenance schedules work based on the actual condition of a machine, not on a fixed calendar. How it relates to reactive and preventive maintenance, what you need to get started and which machines it pays off for.
updateLast updated: 16 September 2026
Predictive maintenance is maintenance based on the measured condition of a machine. Sensors continuously track vibration and temperature; those measurements reveal when wear or damage is developing. You intervene at the moment it is needed, not earlier and not later. That way you do not replace parts that are still fine, and you are not brought to a standstill by a failure you could have seen coming.
PdM is the common abbreviation for predictive maintenance. In Dutch, voorspellend onderhoud and predictief onderhoud are used interchangeably for the same thing. PdM stands in contrast to PM, preventive maintenance, where the maintenance moment is fixed in advance.
The four types of maintenance
Maintenance strategies are usually divided into four forms. They do not exclude one another: in most plants they exist side by side, depending on how critical a machine is.
| Type of maintenance | When to act | Based on | Cost of unplanned downtime | Insight into wear |
|---|---|---|---|---|
| Corrective (reactive) | After the failure | Nothing; you wait and see | High | None |
| Preventive | At a fixed interval | Running hours or calendar | Medium | Limited |
| Condition-based | On a deviating measurement | The current measurement | Medium | Snapshot |
| Predictive | Before the failure develops | The trend in the measurement data | Low | Continuous |
Classifications differ by field; some sources count condition-based and predictive as a single form, others add prescriptive maintenance as a fifth.
The difference from preventive maintenance
Preventive maintenance is the most widely used strategy and works well, right up to the moment the interval no longer matches reality.
A fixed interval replaces parts that are still perfectly good. That costs material, labour and production time you need not have spent.
A machine loaded more heavily than assumed fails before the planned moment. The interval then offers false certainty.
With predictive maintenance the measurement data sets the moment. The interval does not disappear, but is driven by what actually happens.
What do you need to get started?
Predictive maintenance calls for a measurement chain: from the machine to a specialist who reads the data. Four parts, and you do not have to build them all yourself.
Sensors
Wireless or wired vibration sensors on the measurement points of your machine. Wireless saves cabling work in existing installations.
View the sensorsData collector
Collects the measurements from the sensors and forwards them. This is the link between the shop floor and the platform.
View the data collectorsPlatform
The measurement data arrives in SuperCare, where the trend per measurement point is kept and deviations raise a signal.
Specialist
Someone who reads the data and turns it into advice. Without that step, measuring produces charts but not a decision.
View the serviceMaintenance based on data
Software makes the difference between isolated measurements and a usable picture, but it does not take over the assessment.
Trend over snapshot
A single measurement says little. Only the trend over weeks shows whether a value is stable or rising, and how fast.
A signal on deviation
The platform responds when a measurement point moves outside its normal pattern, so nobody has to watch charts continuously.
Assessed by a person
A deviation is not yet a diagnosis. Our specialists look on remotely and determine what the measurement means for your machine.
Maintenance software such as a CMMS plans and records the work; a condition monitoring platform measures the state of the machine. They complement each other and do not replace one another.
When does it pay off?
Predictive maintenance pays for itself on machines whose failure affects your process. Not every machine needs to be monitored.
Does the process come to a halt if this machine fails? Then a measurement point almost always outweighs the downtime it prevents.
Machines at height, in a pit or in a zone you cannot simply enter are particularly suited to continuous remote monitoring.
Pumps, motors, blowers, compressors and gearboxes reveal their condition through vibration. That is where the measurement says the most.
How we carry out and assess those measurements is explained on the page about condition monitoring.