Integrating Services in Modules

Start with what actually goes into a module at the factory, because it is more than most people expect. At minimum, a module carries its first-fix services: the pipework, wiring, ductwork and containment (tray, trunking, conduit) that normally gets installed before walls are lined. Clipped to the frame, routed through the floor and ceiling cassettes, dropped to where outlets will sit. In a well-developed volumetric system the module goes much further — second-fix too: the accessories, the sockets and switches, the light fittings, the radiators or fan-coil units, the sanitaryware, a complete plumbed-and-wired kitchen or bathroom. A bathroom 'pod' is the purest example: a whole wet room delivered with tiling, fittings, pipes and drainage done and water-tested, needing only to be dropped in and connected.

The reason to push the fit-out this far is the same reason prefab exists at all. Bench work in a dry, lit, controlled space, done repeatedly with the right tools at the right height, is faster, safer and more consistent than the same work done overhead on a congested site in the weather. A soldered or pressed joint made at a bench and then pressure-tested in the factory is a joint you trust; a cable run proved with an insulation-resistance test before the module ships is a circuit that will not surprise you. The defect is caught at the bench, where fixing it is cheap, rather than behind a finished wall on site, where it is expensive and disruptive.

But how far to go is a real design and strategy decision, not a default. The more services you finish off-site, the more quality and speed you capture — and the more you must decide early, the heavier and more fragile the module becomes in transit, and the more you depend on the factory getting it exactly right. Some systems deliberately leave certain services as site work precisely because they cross module boundaries or connect to the building's mains, which we come to below. The honest framing is the one this whole course uses: factory-installed services are a *potential* gain you earn by designing for them, not an automatic one. The binding design of the services themselves — loads, sizing, sizing of risers, compliance — always belongs to a qualified building services engineer; your job is to design the module so that engineer's work can be built in cleanly and repeatedly.

Services in any building move in two directions, and in a modular building this becomes a hard spatial discipline rather than something the site works out. Horizontal distribution runs services across a floor — along corridors, through ceiling voids, within a dedicated service wall — to reach each room. Vertical distribution carries them up and down the building through risers: the shafts that take water, drainage, electrical rising mains, data and ventilation between floors. In a site-built job these zones are generous and adjustable; in a module they must be designed, dimensioned and coordinated up front, because the module is built to fixed internal dimensions and cannot be opened up later.

The elegant move in volumetric design is to give each module a clear, repeated service zone — most often a ceiling void or a service wall thick enough to carry the horizontal first-fix — and to place the riser where it will line up vertically as modules stack. When every module in a stack is built to the same layout, the risers align automatically: module upon module, the shaft is continuous, and the only work on site is joining one module's riser to the next at the joint. This is one of the quiet rewards of repetition — design the service zone once, get it right, and it repeats perfectly up the building.

The corollary is that distribution and architecture are now the same conversation. Where the riser sits determines which rooms are 'wet' and drives the plan; the depth of the ceiling service zone sets the floor-to-floor height and therefore the module height on the truck; the route of the horizontal run fixes where you can and cannot have a structural connection. A common pattern concentrates risers and wet services into a small number of heavily-serviced module types (the bathroom pod, the plant or riser module, the kitchen) and keeps other modules comparatively simple, so the complex, services-dense work is standardised and repeated. Getting this zoning right early is the difference between a system that builds smoothly and one that fights itself — and it is exactly the kind of coordination that must happen at concept and scheme, with the services engineer in the room, not left to resolve on site.

Here is the problem that is unique to modular MEP and that catches people out: a building's services are continuous, but a modular building is made of discrete boxes. Every pipe, duct and cable that needs to run further than one module must cross the joint into the next module — and that crossing is a designed connection, made and tested on site, not a seamless run. A module is manufactured with its services left off at the boundary: short lengths of pipe, lengths of cable, duct ends that stop just inside the module face, ready to be joined to the matching tails in the adjacent or stacked module once the boxes are in position.

The design questions follow directly. Where, physically, is the connection made? It needs an accessible connection zone — a reachable void, a ceiling space, a removable panel — because a joint you cannot reach is a joint you cannot make, test or maintain. How is it joined? Mechanical push-fit or pressed couplings, proprietary plug-and-play connectors, bolted duct flanges, or traditional site-made joints, depending on the system and the service. How is it proved? Each connection has to be tested after it is made — pressure-tested for water and drainage, continuity- and insulation-tested for electrical, leak- and flow-checked for ducts — because the factory's bench test only proved the services *within* the module, not the joints *between* them.

The number and position of these crossings is something the designer controls, and fewer is better. A services strategy that keeps long horizontal runs within a single long module, or that stacks wet modules so drainage drops straight down one aligned riser, minimises the crossings and therefore the site connections, the testing and the risk. A strategy that scatters services across many modules multiplies the joints. This is DfMA applied to MEP: design out the connections you do not need, make the ones you do need accessible and testable, and standardise the connection detail so the same quick, proven joint is repeated. The connection design itself — the actual couplings, the fire-stopping where services pass through compartment lines, the acoustic detailing of penetrations — is binding engineering that belongs to the services engineer and the manufacturer's tested system; your role is to give those connections room, access and as few of them as the design allows.

Everything above adds up to a single, unavoidable bargain, and naming it honestly is the point of this lesson. To get factory-installed, tested services — the quality, the speed, the parallel working while the foundations go in — you must commit the services design far earlier than a site-built project demands. On a conventional job the services engineer can keep developing routes, resizing ducts and moving outlets while the frame rises; there is slack, and site trades absorb late change. In a module, the moment the line starts building, the services are being physically installed. A late change to a duct size, a riser position or an outlet is not a revised drawing — it is a built module that now has to be opened up, reworked or scrapped. The factory cannot improvise.

So the off-site decision pulls the whole services design forward in time. Loads have to be estimated and frozen, the distribution strategy (where the risers sit, how deep the service zones are, which modules are heavily serviced) has to be fixed, the connection details and the interface with the building's mains have to be agreed, and the information has to be coordinated and released to the factory in a disciplined sequence. This is demanding, and it is where modular projects that treat MEP as an afterthought come unstuck — the design was not ready when the factory needed it, and the programme advantage evaporates in rework and change.

The flip side, and the reason good teams accept the bargain willingly, is that the early commitment is itself a discipline that improves the job. Freezing the services design early forces the coordination to happen early, when it is cheap, in the model rather than in the building. It rewards the repetition that prefab is built on: decide the serviced-module design once, prove it, and repeat it with confidence. And it produces a building whose services were installed and tested in controlled conditions, which is a better building. The craft for the designer is to see the services not as a layer added after the architecture but as something designed into the module from the first sketch, with the services engineer engaged from the start — and to defer, always, the binding sizing, compliance and performance of those services to that engineer and to the manufacturer's proven system.