Designing for maintainability: what a maintenance regime needs from the design team

Maintenance is usually discussed as an operations problem: a schedule to keep, a contract to award, a budget line to defend. But by the time a building is handed over, most of what determines whether its systems can be maintained at all has already been decided — in the layouts, the clearances, the valve schedule and the data that came with the model.

A plant room that cannot be worked in does not get worked in. That is the practical link between design and operating cost, and it is worth being deliberate about.

Access is a design decision, not a site problem

Every serviceable component has a movement envelope that is larger than the component itself: a filter needs withdrawal space, a coil needs a pull length, a tube bundle needs room to come out, an actuator needs a hand on it. These envelopes rarely appear on a layout unless someone models them, and they are the first thing to be lost when a ceiling void or a riser gets tightened.

• Filter withdrawal and coil pull space modelled as reserved volumes, not assumed
• Access panels and ceiling hatches coordinated against the actual valve and damper positions
• Isolation valves, strainers and test points reachable without a ladder over a stairwell
• Safe standing and lifting positions for anything heavier than one person can hold
• Replacement routes out of the plant room — the item has to leave the building one day

A federated model is the cheapest place to find these conflicts. Resolving a clearance clash in coordination costs a modeller's afternoon; discovering it once the equipment is set costs a variation and, often, a permanent compromise the operator lives with for twenty years.

The model is the asset register

Planned maintenance depends on knowing what you own, where it is and what it needs. That information already exists at design stage — it is just often trapped in a format nobody downstream can use. Carrying consistent tagging and structured asset data through the model into handover turns the BIM deliverable into something the facilities team can actually operate from, rather than a set of pretty geometry that stops being true the first time a component is swapped.

The test is simple: can the operator search for an asset, find its location, its service requirement and its replacement part without opening a drawing? If not, the maintenance regime will be rebuilt from scratch by whoever inherits it.

Design out the failure modes you can

Some maintenance burden is unavoidable, and some is self-inflicted. Systems that cannot be isolated without shutting down a floor get maintained late, or not at all. Drainage that relies on a fall nobody protected blocks. Condensate trays without access do not get cleaned. Filters that are awkward to change are changed less often than the schedule says, whatever the schedule says.

For critical facilities the same logic drives redundancy. Concurrent maintainability — the ability to take equipment out of service without losing the load — is a design property, not an operational one. It has to be decided when the topology is set.

Commissioning sets the baseline that makes drift visible

Performance degrades quietly. Fans do not announce that they are moving less air than they were; chillers do not report that their approach temperature has widened. Detecting that drift requires knowing what the system did when it was correct — recorded airflows, pressures, temperatures and absorbed power at handover.

Without that baseline, every subsequent reading is uninterpretable. With it, a maintenance regime can be evidence-led rather than calendar-led, and the case for intervention can be argued with numbers.

Documentation carries the warranty

Manufacturers commonly make warranty cover conditional on documented servicing at defined intervals by competent people. That paperwork obligation lands on the operator, but it is set up — or undermined — at handover, by whether the O&M information identifies what needs servicing, how often, and to what standard. An asset nobody knew existed does not get its documented service.

What this means for owners

If maintainability is treated as a design requirement — reviewed, coordinated and verified like any other — the resulting regime is cheaper to run and far more likely to be followed. If it is left to be sorted out later, the building will still be maintained, but on the terms the design imposed rather than the ones the owner wanted.