A potable expansion vessel is a steel bottle bolted to a tee off the cold or hot supply, holding a litre or several litres of water against a rubber diaphragm. Water enters and leaves through one connection. That single fact makes it one of the most reliable dead-volumes in the building — warm, still, and almost never inspected.
It is also one of the easiest things for a risk assessment to walk past. The vessel is at high level, painted the same colour as the pipework, and the survey sheet has no obvious box for it. So it gets logged, if at all, as “pressure vessel” and never as what it is: a branch of stored water with effectively no turnover.
Here are the mistakes that let expansion vessels quietly fail a tidy water system, and what to do about each.
Mistake: treating the vessel as plumbing, not as stored water
What it looks like: the assessor records the calorifier, the cold water storage tank and the sentinel outlets, but the expansion vessel on the hot circuit is absent from the asset register and the schematic.
Why it happens: a vessel has no tap, no float and no obvious “water in it” until you think about how it works. The connection is single-ended, so water only exchanges when pressure cycles. Between draw-offs, it sits.
The fix: put every potable expansion vessel on the asset register as a discrete item, with its location, connection type and the circuit it serves. BS 8580-1 frames a Legionella risk assessment as covering the components of the water system, not just the headline plant [1]. If it holds water and connects to outlets people drink from or wash in, it is in scope. Treat it as any other length of low-turnover pipework — that is, a dead leg waiting to be found.
Mistake: assuming the diaphragm keeps the water clean
What it looks like: “It’s a sealed bladder, the water’s separated from the air, so it’s fine.” The vessel gets a tick.
Why it happens: people picture the diaphragm as a barrier that protects the water. It does separate water from the pre-charge gas. It does nothing to keep the water moving or cool.
The fix: recognise that the wetted side of the diaphragm is the problem, not the gas side. Water pools against a large area of flexible EPDM or butyl rubber — a surface that biofilm colonises readily and that you cannot brush, drain fully or see. Legionella does best in still water in the tepid band, roughly 20-45°C in the guidance usually cited, and a vessel on a hot-water circuit frequently sits inside that window as it loses heat to the plant room [2]. The diaphragm is not a safeguard. It is extra surface area in a warm, stagnant pocket.
What nobody tells you about the geometry
The detail that gets lost is that a standard expansion vessel is the worst possible shape for water hygiene, by design.
It is a single-port vessel. Flow does not pass through it; it pushes in and the same water pulls back out of the same hole. So the water nearest the diaphragm barely participates in the daily exchange — a cul-de-sac inside a cul-de-sac. When the system is quiet overnight or over a weekend, nothing moves.
That geometry has two consequences people rarely connect. First, you cannot meaningfully flush a closed diaphragm vessel from an outlet — running a nearby tap exchanges the pipe, not the vessel contents. Second, a thermal disinfection carried past the tee may not penetrate the dead volume behind the diaphragm, so the one place you most want to reach is the place least likely to hold temperature. The vessel can read as “treated” while its core never got there.
Mistake: leaving the vessel with no isolation or drain
What it looks like: the vessel is teed straight into the main with no isolating valve and no drain point, so there is no way to take it out of service, recharge it, or empty it for inspection without shutting down the circuit.
Why it happens: it was installed for pressure control by a heating engineer, not for water hygiene by anyone thinking about Legionella, and the cheapest install is a bare tee.
The fix: fit a serviceable arrangement — an isolation valve and a drain on the vessel branch — so it can be isolated, drained and recharged as a planned task. This is the same gap that turns up repeatedly in survey findings, where unserviceable fittings are flagged but never made good; it sits alongside the other recurring issues found in Legionella risk assessments. Without isolation you have no practical way to manage the vessel at all.
Mistake: ignoring the pre-charge until the diaphragm has failed
What it looks like: nobody checks the gas pre-charge for years. Eventually the diaphragm perishes or waterlogs, the vessel fills entirely with water, and turnover drops to near zero while pressure problems start showing up elsewhere.
Why it happens: pre-charge is a heating-side maintenance item that falls between the water-hygiene contractor and the plumber, so neither owns it.
The fix: put a periodic pre-charge check and diaphragm-condition assessment into the planned maintenance schedule, with the result recorded. A waterlogged vessel is not just a pressure fault — it is a litres-scale slug of static water on your potable system. The pragmatic call is to log the check against the asset, with date and reading, so a failing vessel is caught as a trend, not a surprise.
Mistake: cleaning around the problem instead of removing it
What it looks like: the vessel gets added to a flushing sheet and someone runs the nearest tap weekly, ticking “flushed” without exchanging a drop of its contents.
Why it happens: a flushing line is easy to add and looks like control. Reaching the dead volume is hard.
The fix, and the single highest-impact correction here: where a potable expansion vessel is needed, specify a flow-through (or through-flow) type. These are built with an inlet and an outlet so that normal draw-off passes water through the vessel body rather than letting it stand behind a closed diaphragm. In my view, designing the dead-volume out with a flow-through vessel beats every downstream control you could layer on a closed one. Where an existing vessel cannot be replaced immediately, the interim measures are isolation, scheduled drain-down and recharge, and bringing it inside the temperature regime; the same principles apply to any low-turnover hot water storage component.
A note on judgement
None of this replaces a competent, site-specific assessment. Whether a particular vessel is a real risk depends on its volume relative to the system, where it sits, the temperatures it runs at, and how the building is used — figures and frequencies belong to your own risk assessment and the cited guidance. Use the points above to make sure expansion vessels are seen and judged, not to pre-judge them.
The concrete next step today: walk the plant rooms with the schematic and physically tag every expansion and pressure vessel against the asset register. If any are missing — and on most sites at least one is — that gap, and the planned checks that follow, are exactly what a digital logbook keeps visible instead of letting it slide off a paper survey sheet.
FAQ
Do small domestic-style expansion vessels on unvented cylinders count?
Yes. The same single-port, low-turnover geometry applies whether the vessel is a few litres on a domestic unvented cylinder or a larger unit on a commercial circuit. Smaller volume reduces the slug of static water but does not change the principle, and these are the ones most often left off the register entirely.
Can I just disinfect the expansion vessel periodically instead of replacing it?
You can include it in a disinfection, but treat the result with caution. Holding conditions in the dead volume behind a closed diaphragm is difficult, so a planned isolate-drain-recharge cycle, or replacement with a flow-through type, is more dependable than a disinfection that may not penetrate [3].
Is a flow-through vessel always the right answer?
Not automatically — it depends on system design, pressures and space. It is the strongest option for water hygiene because it removes the dead volume, but the choice still sits with the designer and the risk assessment. The point is that the hygiene implications should be on the table when the vessel is specified, not discovered years later.
Sources
[1] BSI, “BS 8580-1:2019 - Risk assessments for Legionella control. Code of practice”. https://knowledge.bsigroup.com/products/water-quality-risk-assessments-for-legionella-control-code-of-practice-1 [2] HSE, “Legionnaires’ disease: Technical guidance (HSG274)”. https://www.hse.gov.uk/pubns/books/hsg274.htm [3] HSE, “Hot and cold water systems”. https://www.hse.gov.uk/legionnaires/hot-and-cold.htm