What Is Predictive Maintenance?

Predictive maintenance is a maintenance strategy that uses condition data - sensor readings, operating hours, vibration, temperature, oil analysis - to forecast when an asset is likely to fail, so it can be serviced just before it breaks. Instead of fixing equipment on a fixed calendar or waiting for a breakdown, the asset’s own behaviour decides when attention is due. It sits at the data-heavy end of a spectrum that runs from reactive (fix it when it breaks) through preventive (fix it on a schedule) to predictive (fix it when the evidence says failure is near), and its success is usually measured with reliability metrics like MTBF.

How predictive maintenance works

Every predictive programme has three ingredients:

• Condition data - something measurable that degrades before failure: vibration on a bearing, temperature on a motor, pressure in a hydraulic circuit, contamination in oil, or simply accumulated operating hours.
• A trigger - a rule that turns readings into action. This can be as simple as “vibration above this threshold” or as elaborate as a model trained on past failures.
• A response - the trigger raises a work order so a technician intervenes while the fault is still small, cheap, and schedulable.

The whole value is in the timing. A bearing replaced a week before seizure is a planned one-hour job; the same bearing replaced after seizure can take the shaft, the housing, and a day of downtime with it.

Predictive vs preventive maintenance

Preventive maintenance services equipment at fixed intervals regardless of condition. It is simple and predictable, but it over-services machines that were fine and still misses the failures that arrive between visits. Predictive maintenance services on evidence instead, which means fewer unnecessary jobs and fewer surprises - at the price of monitoring equipment, data plumbing, and enough failure history to calibrate the triggers.

The two are not rivals in practice. Most maintenance programmes run preventive servicing as the baseline and add predictive monitoring only on the assets where unplanned downtime hurts most.

Examples of predictive maintenance

• A vibration sensor on a forklift drive motor flags bearing wear weeks before it seizes, so the repair happens during a quiet shift instead of mid-load.
• Periodic hydraulic oil sampling on excavators catches metal particles and water contamination before they destroy a pump.
• A compressor’s hours meter and discharge temperature are reviewed together; a temperature creep at constant load triggers an inspection rather than waiting for the scheduled service.
• A delivery van’s battery voltage trend predicts a no-start morning before it happens.

What predictive maintenance means for smaller teams

Full predictive programmes are sensor projects, and for most small and mid-size fleets the economics only work on one or two critical machines. The realistic first rung on the ladder is history: meter or estimate operating hours, log every fault as a service ticket against the specific asset, and record what the fix was. After a year of honest logging, the patterns that sensors would have found often surface anyway - the machine whose MTTR keeps growing, the component that fails every 400 hours, the unit that eats batteries. A system like AMPthilly keeps that raw material automatically, because every reported issue and repair stays on the asset’s record permanently. That history will not predict a bearing failure to the day, but it tells you exactly where a sensor would pay for itself.

Related terms

• MTBF (Mean Time Between Failures) - the reliability metric predictive maintenance aims to raise
• MTTR (Mean Time to Repair) - how long repairs take once a failure occurs
• Work Order - the job record a predictive trigger should create
• Service Ticket - the logged fault reports that build failure history
• Warranty Tracking - knowing which predicted repairs the supplier should pay for