A maintenance plan sets out, for each asset, which maintenance tasks are needed, at what interval (time, usage or condition), which trade carries them out, with which parts and tools, and under which safety measures. To create a maintenance plan, work in a fixed order: map your assets, set priorities, collect tasks for each asset from the manual, legislation, failure history and experience, choose an approach and interval for every task, and group the tasks into rounds. The result makes preventive work predictable and helps reduce the risk of unplanned downtime.
This article covers machines and installations in industry, energy, waste-to-energy and logistics. A long-term maintenance and replacement budget for a building is a different tool; the difference is explained in the questions at the end.
What a maintenance plan is, and what it is not
The European maintenance terminology standard, EN 13306, describes a maintenance plan as a structured, documented set of tasks covering the activities, procedures, resources and timescale needed to carry out maintenance. In practice it is a list of recurring tasks per asset: what is done, how often, how long it takes, which trade does it and what has to be ready beforehand.
Three terms tend to get mixed up. It pays to keep them apart, because each one answers a different question:
| Term | Answers | Example |
|---|---|---|
| Maintenance strategy | Why, and which approach? | Cooling water pumps get vibration monitoring |
| Maintenance plan | What, and how often? | Measure vibration on pump P-201 every month |
| Maintenance schedule | When, and by whom? | Tuesday of week 14, by the measurement crew |
You choose the strategy per asset or per failure mode, the plan turns that choice into concrete tasks, and the schedule puts those tasks in the calendar within the capacity of your crew. Getting from a pile of tasks to a realistic weekly schedule is covered in maintenance planning and scheduling. What that looks like on a planning board is shown under maintenance planning software.
What each task in a maintenance plan should contain
"Check pump" is not a maintenance task; it is an intention. A usable task is written so that two technicians carry it out the same way and report the same things afterwards. Record at least the following for every task. The example is regreasing the bearings of a cooling water pump.
| Element | What you record | Example |
|---|---|---|
| Task and work instruction | What happens, step by step | Clean the nipples, apply the specified amount of grease, refit the caps |
| Interval | Time, usage or condition, with a second interval if needed | 2,000 running hours or 6 months, whichever comes first |
| Duration and trade | Estimated time and discipline | 20 minutes, mechanical technician |
| Materials and tools | What must be ready | Bearing grease as per lubrication schedule, labelled grease gun |
| Safety | Isolation, permits, protective equipment | Grease through extended lines outside the guard |
| Downtime | Does the installation have to stop? | No |
| Acceptance criterion | When the task has been done properly | No lasting rise in bearing temperature |
| Source | Why the task exists | Pump manual, maintenance chapter |
Two of these fields are often skipped. The acceptance criterion turns "check" into a measurement with a limit: which value is acceptable, and what do you do when it is exceeded? The source tells you later why a task exists. If you want to decide in three years' time whether a task can go, you need to see whether it came from the manual, a legal requirement or a single failure long ago.
How to create a maintenance plan in seven steps
The sequence below works for a single machine and for a whole plant. On a large site, you run through steps 3 to 6 for one group of assets at a time, starting with the most critical.
- Map your assets. A plan is attached to an asset or a position (functional location); without a register you do not know where a task belongs or where its history ends up. Tasks that belong to the function at a fixed location go on the position. Tasks that belong to the physical unit, such as an overhaul after a set number of running hours, go on the asset itself. Setting up that structure is covered in building an asset register; an example is shown under asset management software.
- Decide where to start. An asset criticality analysis ranks assets by the risk of failure: the consequences for safety, the environment and production, and the likelihood. Start with the critical ones. For cheap assets whose failure has little impact, repairing after failure can be a defensible choice, as long as you record that you made it.
- Collect the sources for each asset. These are the manufacturer's manual, legislation, the failure history and the experience of technicians and operators, supplemented by analyses such as an FMECA. Write down every candidate task together with its source. Pruning comes later. Manuals and drawings are easier to find when they are linked to the asset, as in technical document management.
- Choose an approach for each task. Replacing or overhauling at a fixed interval in time or usage only works for wear that increases with age or use. If a failure mode gives a measurable warning, measuring is often better than replacing. And if failure has few consequences, repairing afterwards is sometimes the cheapest option.
- Set the intervals. The starting point is usually the manufacturer's advice, adjusted to your usage and operating conditions. How to do that is explained further down.
- Group tasks into rounds. Tasks with the same interval, the same trade and the same conditions (with or without a shutdown) are combined into one round. Rounds in the same area become a route: the weekly check of ten pumps in one building turns into one route instead of ten separate work orders.
- Review and adjust. Have findings and readings recorded every time a round is carried out, and give each asset an owner who keeps the plan up to date. If a task never finds anything after several rounds, the interval can be extended. If it keeps finding damage, or the asset fails between rounds, shorten the interval or change the approach.
The diagram below shows steps 3 to 6 for a single asset: four sources produce six tasks for cooling water pump P-201, and those tasks end up in four rounds.
Where the tasks come from: four sources
The manufacturer's manual
For machinery, the manual is the first source, and what it has to contain is laid down in law. The EU Machinery Directive (2006/42/EC) requires the instructions to describe the adjustment and maintenance operations the user should carry out and the preventive maintenance measures to observe, how to carry them out safely and, where safety is affected, the specifications of the spare parts to be used (Annex I, section 1.7.4.2). The Machinery Regulation (EU) 2023/1230, which applies from 20 January 2027, carries these requirements over. A manual is written for an average situation, though: the manufacturer does not know your operating hours, dust, moisture or loads. Treat it as a starting point, not the final word.
Legislation
The EU directive on the use of work equipment, Directive 2009/104/EC, requires employers to keep work equipment at the required safety level throughout its working life by means of adequate maintenance, and to keep any maintenance log up to date. Equipment exposed to conditions that cause dangerous deterioration must be inspected periodically by competent persons, with the results recorded. In the Netherlands these duties sit in articles 7.4a and 7.5 of the Working Conditions Decree (Arbobesluit), and pressure equipment and cranes have their own inspection rules. Add these inspections to the plan as tasks, with the legal basis as their source; statutory inspections explains which ones apply to your installations.
Failure history
Notifications and work orders with a recorded cause show which failure modes actually occur on your installation, and how often. A bearing that keeps failing because of misalignment calls for a task that is not in the manual. The history also shows which tasks never find anything.
Experience and FMECA
Technicians and operators know which points are hard to reach, where dirt builds up and which noises matter. Write that knowledge down so that it does not leave with one person. For critical assets, an FMECA is the systematic version: for each failure mode you assess the effect and criticality, and that tells you which task can detect the failure mode in time or deal with it.
Setting intervals: calendar, usage or condition
You can express an interval in three ways. In EN 13306 terms, the first two count as predetermined maintenance: the interval is fixed in advance, without first measuring the condition.
- Calendar. Suits deterioration that continues while the machine is idle, as with rubber, grease and corrosion, and inspections with a fixed term. It is also the simplest choice when usage is constant.
- Usage. Wear on bearings, seals and filters mainly follows running hours, starts, cycles or tonnes processed. When operating time varies, usage tracks wear better than the calendar; running-hours monitoring with IoT sensors explains how to record running hours reliably.
- Condition. Here the interval is the measurement frequency. You derive it from the P-F interval, the time between the first measurable sign of trouble and the actual failure; see condition-based maintenance.
Calendar and usage are often combined: "2,000 running hours or 6 months, whichever comes first". A pump running around the clock reaches 2,000 hours after about twelve weeks. A standby pump that runs 500 hours a year has only reached 250 hours after six months, so it is due on the calendar first. That calendar limit is needed because grease also ages while the pump stands still.
Start from the manufacturer's advice unless you have a good reason to deviate, and record that reason. Check whether following the schedule is a condition of the warranty. From then on, your findings set the interval.
If an asset fails shortly after maintenance, look at how the work was done before you look at the interval. Opening equipment up and putting it back together carries risks of its own, such as a wrongly fitted seal or dirt in a bearing housing.
Example: a maintenance plan for a cooling water pump
Take cooling water pump P-201 in a power plant: a horizontal centrifugal pump with an electric motor and a flexible coupling, running continuously. The four sources produce six tasks.
| Task | Interval | Source |
|---|---|---|
| Check for leaks, noise and bearing temperature | Weekly | Experience & FMECA |
| Measure vibration on both bearing housings | Monthly | Experience & FMECA |
| Regrease the bearings | 2,000 running hours or 6 months | Manufacturer's manual |
| Check pump and motor alignment | Yearly, during shutdown | Failure history |
| Inspect the coupling element | Yearly, during shutdown | Manufacturer's manual |
| Check the coupling guard | Yearly, during shutdown | Legislation |
The six tasks fall into four rounds. The weekly operator round takes a few minutes per pump and brings leaks and hot bearings to light early. The monthly vibration measurement comes from the FMECA: bearing damage shows up in vibration, often well before the pump fails. Regreasing follows the manual on running hours, with the calendar as a safety net. The three yearly tasks all need a shutdown and isolation, so they are bundled into a single round during the planned outage.
Two of the tasks come from sources other than the manual and the FMECA. The alignment check comes from the failure history: two bearing failures were traced back to misalignment after earlier work. The guard check comes from the rules on work equipment; in the Netherlands, article 7.7 of the Arbobesluit requires moving parts that pose a danger to be guarded, so the guard is part of what you keep in a safe condition. Each round becomes a work order when it falls due. You can see what that looks like in practice under maintenance and preventive rounds.
Common mistakes when creating a maintenance plan
- Copying the manual one-to-one. You end up with tasks that do not fit your usage, and you miss the failure modes the manufacturer never saw.
- Too many tasks. Every task costs capacity, and every time you open something up there is a chance of an assembly error. A task should detect or deal with a failure mode, or meet a requirement. If nobody can say which, drop it.
- Tasks without a criterion. "Visual check" without saying what to look for and where the limit lies produces ticks, not information.
- No owner. Without someone who reads the findings and updates the plan, it starts going out of date on day one.
- Rounds without materials. A round that falls due while the filter or seal is not in stock gets pushed back or half done. Link the materials to the round and make sure they are in stock before the round falls due.
- Lubrication outside the plan. Lubrication is the task that comes round most often. If it lives on a separate list, you cannot see what actually happens; how to create a lubrication schedule explains how to organise it properly.
Frequently asked questions
Is a maintenance plan a legal requirement?
Not as a document with a prescribed format, but the duties behind it are. EU rules on work equipment, which in the Netherlands sit in articles 7.4a and 7.5 of the Arbobesluit, require adequate maintenance, an up-to-date maintenance log where the equipment has one, and inspections where wear could create a hazard. Evidence of those inspections must be available at the workplace and shown to the inspectorate on request. A maintenance plan with recorded execution is the practical way to organise and demonstrate all that. Pressure equipment and cranes come with additional inspection rules.
Can you create and manage a maintenance plan in Excel?
Yes, for a limited number of assets Excel or another spreadsheet works: one row per task with the asset, the interval, and when it was last done and is next due. It becomes hard to manage as the plan grows. Nobody gets an automatic reminder, running hours have to be updated by hand, and whatever was found during each round becomes detached from the plan. At that point, a maintenance system that turns rounds into work orders and keeps the history per asset is the logical next step.
Who should write the maintenance plan?
Usually a maintenance planner, maintenance engineer or reliability engineer, together with the technicians who do the work and the operators who see the installation every day. Technicians know which tasks make sense and how long they take; operators are the first to notice when something changes. Give every asset one owner who reads the findings, makes the changes and updates the plan whenever a machine is added or modified.
How often should a maintenance plan be reviewed?
As a rule of thumb, review the plan for each asset at least once a year, and also whenever something prompts it: a failure the plan should have flagged, a change to the installation or the way it is used, new instructions from the manufacturer or a new legal requirement. In the annual review, focus on tasks that have not found anything for a long time and on failures the plan did not see coming.
What is the difference between a maintenance plan and a long-term building maintenance plan?
A long-term building maintenance plan (in the Netherlands an MJOP) schedules and budgets major work on a building over a number of years, such as replacing the roof, the facade or the lifts. A maintenance plan for machines and installations describes recurring tasks per asset, with intervals ranging from days to years. The two can exist side by side: replacing an entire installation belongs in a multi-year budget, while its periodic tasks belong in the maintenance plan.