A plant maintenance program is built in an order, and plants that skip a step find it waiting for them later. The register first, because the programme is against assets. The strategy per class second, because the tasks depend on it. The schedule third, levelled to the hours. The work orders fourth, closed with findings. The reports fifth, read monthly. The review sixth, adjusting intervals from the history. This page walks the order, says what each step needs, and gives the load a small plant can carry from this site's worked example.
Steps one and two: register and strategy
Walk the plant and list every asset worth a work order, in a tree by line. Rank the classes by the consequence of failure. For each class from the top, decide: interval tasks, inspections, condition checks, run to failure with a spare, or redesign. Write the reason down. A plant that starts with software before this step spends a year configuring screens for assets it has not listed.
Steps three and four: schedule and work orders
Write the tasks per class with triggers and checklists, apply them to the assets, and add up the hours. Level the load to the hours the team has after the reactive work; this site's worked example carries 60 preventive jobs at 66 hours a month across 180 assets. Raise the work orders from the schedule, close them at the machine with findings and times, and raise corrective work from what inspections find. The maintenance log worksheet on this site holds steps one to four for a small plant.
Steps five and six: reports and review
Monthly: the reactive share, schedule compliance, the overdue list, the repeaters and the cost split. Quarterly: shorten the intervals on assets with failures between tasks, lengthen or drop the tasks that find nothing, add spares where downtime is waiting, and revisit the strategy on any asset whose consequence has changed. A programme without the review is a schedule that drifts; with it, the example's 84 breakdowns a month have somewhere to go.
The load a small plant can carry
The programme has to fit the team. If the strategy produces more preventive hours than exist, cut from the bottom of the ranking, not from the top: lengthen intervals on the low-consequence classes and keep the critical ones whole. A programme that fits and runs beats one that is complete and deferred, and the reactive share over the next two quarters is the test of whether the fit was right.
Questions people ask about plant maintenance program
How long does it take to build a plant maintenance program?
Weeks for the register and the strategy, a quarter to see the schedule run, a year before the history is deep enough to adjust intervals with confidence.
Should the programme be built in the CMMS from the start?
The register and tasks can be, if the software is chosen. The thinking is done on paper first, and the worksheets on this site are that paper.
Who owns the programme?
The maintenance manager, with the supervisor running the week and the technicians closing the work orders. A programme with no owner drifts within a quarter.