Definition
Preventive maintenance inspects and replaces on a defined interval or condition threshold, before equipment fails.
Preventive maintenance inspects and replaces on a defined interval or condition threshold, before equipment fails.
Equipment-specific rules
Check condition and issues
Preventive replacement and repair
Refine intervals and checks
Preventive maintenance inspects and replaces on a defined interval or condition threshold, before equipment fails.
Maintenance strategies fall into three groups: reactive (fix after failure), preventive (intervene on a schedule) and predictive (intervene when condition data says so).
The hard part of preventive maintenance is setting the interval. Too frequent and you spend maintenance cost and downtime; too sparse and you miss failures. That is why real failure data such as MTBF is needed.
For it to take hold there must be a loop: work orders are raised and the results recorded. Checklists without recorded outcomes give you no basis for adjusting the interval.
Preventive works on time or usage. Predictive works on condition data such as vibration or temperature. Predictive is more efficient but needs sensors and accumulated data, so most plants move to it from preventive rather than starting there.
Terms that make more sense read together.
Mean Time Between Failures (MTBF) is the average running time between one failure and the ne…
MetricsMean Time To Repair (MTTR) is the average time from a failure until the machine is running a…
MetricsOverall Equipment Effectiveness (OEE) expresses how much of its planned time a machine actua…
What IYULAB provides in this area.
Questions that come up when evaluating Preventive Maintenance.
Start from the manufacturer's recommendation and adjust with your own failure history. Before that history exists, following the recommendation is the safe choice.
The inspection items may not match the actual failure modes. Comparing failure locations against checklist items usually exposes the mismatch.
You need a period of sensor data first. Identifying an anomaly requires knowing what normal looks like.
A look at your equipment and current workflow is usually enough to scope it.
Talk to an engineer