Key takeaways
Short answer: Preventive maintenance schedules service by the calendar or a usage counter — every 500 hours, every quarter — whether or not the asset needs it. Condition-based maintenance watches the asset's real health through sensors or inspections and only acts when the data says intervention is due. PM trades some wasted effort for simplicity; CBM trades simplicity for efficiency and earlier warning. Most plants run both, matched to each asset's cost and failure pattern. For the broader picture, see OEE for manufacturing.
Preventive maintenance is the workhorse of most maintenance programmes. You define a fixed interval — by time (every 90 days) or by usage (every 10,000 cycles) — and the work order fires automatically, regardless of the asset's condition that day. Its strengths are real: it is simple to plan, easy to resource, and it dramatically reduces the random failures that come from running to breakdown. Its weakness is just as real: a fixed interval is a guess. Set it too long and you still get failures; set it too short and you spend labour and parts servicing assets that were perfectly healthy — and every intervention is itself a chance to introduce a fault.
Condition-based maintenance replaces the guess with measurement. Instead of service every quarter, the rule becomes service when vibration exceeds this level — or temperature, oil particle count, motor current, or ultrasonic signature. The work order fires on evidence of degradation, not on the calendar. Done well, CBM both cuts unnecessary service and catches developing faults earlier than any fixed schedule could. The cost is infrastructure and discipline: you need the right sensors or inspection routes, a baseline of what normal looks like, and someone — or something — watching the trend and acting before the threshold becomes a breakdown.
The choice is not old versus modern. It is a trade between wasted effort and required infrastructure. PM wastes some service on healthy assets but needs almost no technology. CBM eliminates most of that waste and warns you earlier, but only pays off where the asset is critical enough, expensive enough to over-service, or dangerous enough to fail that the sensors and analysis earn their keep. Spending on continuous vibration monitoring for a cheap, redundant pump is as wasteful as running a critical, single-point-of-failure compressor to breakdown. Match the strategy to the asset, not to fashion.
Take a critical pump with a fixed quarterly PM. Over a year that is four interventions — say roughly 16 labour-hours and four parts kits — whether or not the pump needed them, and it still suffered two unplanned failures from bearing wear that struck between services. Move it to condition-based monitoring: vibration trending flags the bearing degradation about ten days out, you schedule one planned repair into a gap in the run, and you skip two of the four blind services. Fewer interventions, no surprise stops, and the failure is caught while it is still a planned job rather than a 2 a.m. emergency. That swing from unplanned to planned is where the money is.
The most common error is treating the interval as fixed truth and never revisiting it against actual failure data — intervals should tighten or relax as evidence accumulates. The second is buying condition-monitoring sensors but never defining the thresholds or the response, so alerts pile up unactioned and the programme quietly dies. The third is applying one strategy plant-wide; a uniform policy always over-serves some assets and under-serves others. A CMMS that records both the schedule and the failure history is what lets you tune the mix instead of guessing.
Both strategies attack the same enemy: unplanned downtime, the largest availability loss in OEE for most plants. Preventive maintenance converts some random failures into planned work; condition-based maintenance converts even more, and earlier, with less wasted labour. The metric to watch is the ratio of planned to unplanned work: as it climbs, availability climbs with it and your big losses shrink. Maintenance strategy is not a cost centre disconnected from production — it is one of the most direct levers on OEE you have.
Fabrico links maintenance to the losses it is meant to prevent. It tracks PM schedules and completion, records failure history so intervals can be tuned with evidence, and connects that maintenance data to live OEE — so you can see whether a recurring downtime loss is a maintenance problem and whether your strategy is actually moving availability. That closed loop between condition, action, and result is what separates a maintenance calendar from a maintenance strategy. Book a demo to see it on your assets.
Preventive maintenance is scheduled by time or usage regardless of condition; condition-based maintenance is triggered by measured asset health indicators such as vibration or temperature. PM acts on a calendar, CBM acts on evidence.
No. CBM is more efficient on critical, instrumented, high-cost assets, but it requires sensors, baselines, and analysis. For cheap, predictable wear items a simple preventive schedule is often the better-value choice.
They are closely related. CBM acts when a current indicator crosses a threshold; predictive maintenance goes further, using models to forecast when failure will occur. Predictive maintenance is essentially CBM plus prediction.
Yes, and most plants should. Apply preventive schedules to low-cost predictable items, condition-based monitoring to critical assets, and deliberate run-to-failure to cheap redundant ones. A long-interval PM backstop on CBM assets is good practice.
Both strategies reduce unplanned downtime, the biggest availability loss in OEE for most plants. Shifting work from unplanned to planned raises availability directly and shrinks the six big losses.
Zaloguj
Blog
