Heat network leak detection: the meter you already have

TL;DR: A heat network loses water before it loses anything else, and it replaces that water automatically. The pressurisation unit tops the system up on its own, so a buried leak can run for months while the building stays warm and nobody raises a ticket. The draft national technical standard for heat networks now requires a water meter on that top-up connection and requires it to feed early warning of leaks. Most UK schemes already have the pipe. What they do not have is the reading. That is what heat network leak detection comes down to.

Last updated: 4 August 2026

Key takeaways

• Heat network leak detection starts at the top-up line, not at the buried main.
• Heat networks stopped being an unregulated utility on 27 January 2026, the date Ofgem’s rules went live in Great Britain.
• Around 3% of UK heat comes from heat networks today. Government expects that to reach almost 20% by 2050.
• The draft Heat Network Technical Standard requires a water meter on the top-up connection, and requires that meter to be wired to the remote monitoring system “for early warning indication of leaks”.
• In operation it goes further: top-up water shall be monitored to detect leaks, with daily recording on larger systems.
• Wire-in-insulation surveillance is specified for buried steel pipe only. It says nothing about plastic distribution, plant rooms or risers.
• A leak is self-reinforcing. Every litre of fresh top-up water carries dissolved oxygen into a system designed to have none, and water quality drives leaks and component failure.

What is heat network leak detection?

Heat network leak detection is the practice of spotting water escaping from a closed heating or cooling circuit before it shows up as a failure. On a district or communal scheme that circuit is sealed and pressurised, so the useful evidence is not a puddle. It is the volume of fresh water the system has to swallow to hold its pressure.

That distinction matters because a heat network is not a mains water system. Nobody rings to say the pressure has dropped. The pressurisation unit does its job silently, and the first visible symptom of a slow leak is often a corrosion failure somewhere else entirely, months later.

The commercial context changed this year. The Ofgem rules for heat networks went live on 27 January 2026, and the explanatory memorandum to the regulations is blunt about the starting point: heat networks were “currently largely unregulated, despite being an essential utility”. Operators now sit inside a framework with service quality obligations and a special administration regime behind them. Unexplained water loss is no longer a private engineering matter.

The direction of travel is not marginal either. Government zoning material puts heat networks at around 3% of UK heat in 2026, rising to almost 20% by 2050, unlocking more than £80 billion of investment and around 35,000 jobs. Every scheme in that expansion is another buried asset nobody will dig up to inspect.

Why make-up water is the first thing a leak touches

Because the circuit is closed, its water is a fixed volume. Nothing should leave it. So the quantity of make-up water entering through the pressurisation unit is a direct mass balance on the whole network, and it is the only one you get without instrumenting every branch.

This is the part operators underrate. A leak on a heat main does not announce itself by turning the heating off. It announces itself by making the top-up valve open more often. If nobody counts how often, the leak has no voice at all.

There is a second reason to care about volume rather than heat. Heat losses across the distribution pipework are already the single largest inefficiency on most schemes. The government’s own optimisation guidance for the sector states that “typically, the greatest source of inefficiency in a heat network are heat losses within the heat distribution network and energy centre”, and it benchmarks that against CIBSE CP1, which sets 100 W per dwelling as the maximum acceptable communal distribution heat loss, with high performing networks under 70 W per dwelling. A leak pours hot treated water into the ground and never appears as a bill anyone reads.

The loop that makes a small leak grow

Here is the mechanism that turns a nuisance into a capital problem. A closed system is deliberately deoxygenated and chemically conditioned. Fresh make-up water is not. Every litre that replaces a leaked litre brings dissolved oxygen and fresh minerals into a circuit engineered to have neither.

The draft national standard is explicit that this is a failure pathway, not a theoretical concern: “the water quality of a Heat Network can be a significant contributor to outcomes such as leaks and component failure”. It requires top-up water to meet defined water quality limits, and allows conditioning equipment such as demineralisation to bring it into range.

So the sequence runs: water leaks out, oxygenated water comes in, corrosion accelerates, pinholes appear, more water leaks out. A network that tops up more than it should is not only wasting water and heat. It is shortening the life of its own pipework, and the meter that would have caught it costs a rounding error against a replaced section of main.

What the draft technical standard actually requires

The Heat Network Technical Assurance Scheme is the sector’s coming technical rulebook, and its first standard was published in draft in 2025 so the industry could see it early. It settles the argument about whether make-up water monitoring is best practice or a requirement. It is written as a requirement, three times over.

At design stage it calls for a water meter on the top-up equipment, positioned “upstream of the quick-fill connection” and downstream of any conditioning kit, so the meter sees everything that enters the system, including manual filling. It then requires that meter to be connected to the automatic and remote monitoring system “for early warning indication of leaks”.

At operation and maintenance stage it is stronger still. Top-up water “shall be monitored to detect leaks by recording of make-up water or daily recording for larger systems”, and where significant unexpected quantities appear, an investigation into where the leak is occurring “shall be undertaken with an appropriate level of urgency”, with remedial actions logged.

The implication is uncomfortable. A monthly clipboard reading technically satisfies “recording of make-up water” on a small scheme, but it cannot separate a burst that happened last Tuesday from a drip that has run since spring. Daily data can. Hourly data also separates a leak from a legitimate refill after maintenance, which is the question that actually wastes engineers’ time.

What buried pipe surveillance does not cover

Ask about heat network leak detection in the UK and most answers describe wire-based surveillance: sensing wires embedded in the insulation of pre-insulated pipe, registering moisture ingress and reporting a fault distance. It is good technology, and the draft standard calls for it, referencing BS EN 14419:2019 for surveillance systems on buried district heating pipes.

But read the requirement precisely. It applies to steel systems, on buried district distribution pipework. That leaves three gaps that matter on real estates.

• Plastic distribution. Many smaller and retrofitted schemes are not pre-insulated steel with a sensing pair running through them.
• Everything inside the building. Risers, plant rooms, substations and heat interface unit connections sit outside the buried-main scope entirely.
• Legacy assets. A design-stage standard does not retrofit itself into a scheme commissioned fifteen years ago, which is precisely the population most likely to be leaking.

The make-up water signal has none of those blind spots. It does not care what the pipe is made of, where it runs, or who installed it. What it cannot do is tell you where the leak is, which is exactly why complete heat network leak detection uses both: volume tells you that you have a problem and roughly how big, surveillance and survey work tell you where to dig.

What normal top-up looks like

A healthy closed system should need almost nothing. Real top-up demand comes from venting after works, small losses at seals and glands, and sampling. It should be occasional, explainable and traceable to something in the maintenance log.

Practical heat network leak detection is therefore not about an absolute number, because a two-block communal scheme and a city district network are not comparable. You are looking for shape. Three patterns are worth naming.

• A step change. Top-up volume moves to a new baseline and stays there. That is a leak that opened on a date you can now identify.
• A continuous trickle. Small volume, every day, including days when nobody was on site. Closed systems have no legitimate reason to drink overnight.
• A ramp. Slowly rising demand over weeks. Usually corrosion doing its work, and the cheapest of the three to fix.

None of those are visible in a monthly total. All three are obvious in a daily series within a fortnight. It is the same discipline estates teams already apply to night flow analysis on cold water, pointed at the one pipe on a heat network that should never flow at all.

Where to put the meter, and why clamp-on

One position does most of the work: the cold water feed into the pressurisation unit or fill set, upstream of the quick-fill connection so that manual top-ups are captured too. That single point is the whole-network mass balance the standard is asking for.

The obstacle is rarely the argument. It is the shutdown. Cutting into a live fill line means isolating, draining, fitting and recommissioning, and on an occupied residential scheme in heating season that is a job nobody wants to programme. So it stays on the list.

A clamp-on ultrasonic flow meter removes that objection. The transducers strap to the outside of the pipe, so there is no cutting, no drain-down and no shutdown, and the AQUAIOT unit covers DN8 to DN100, which comfortably spans the fill lines on communal and district schemes. It reads flow and volume in real time and integrates with telemetry, so the reading becomes a series rather than a number someone writes down.

On a multi-building scheme one more position earns its place: a meter on each substation fill point, so a rising total can be attributed without a site walk. Estates already metering cold water run both on one dashboard, which is why this belongs inside existing water balance monitoring rather than as a new project.

Five steps to start heat network leak detection

1. Find the fill point. Locate the pressurisation unit and the quick-fill connection, and confirm whether anything is currently measuring what passes through them.
2. Meter it without a shutdown. Fit a clamp-on ultrasonic meter upstream of the quick-fill connection and downstream of any conditioning equipment, in line with the draft standard’s positioning.
3. Log daily, review weekly. The standard asks for recording, and daily recording on larger systems. Aim for hourly data and a weekly glance at the shape.
4. Write down what normal is. Agree a baseline and an alarm threshold after four weeks of clean data, and record planned refills so maintenance never looks like a leak.
5. Join it to water quality. If top-up demand rises, sample. Rising make-up water and drifting water quality are the same story told twice.

Frequently asked questions

How does heat network leak detection using make-up water work?

A closed heating circuit holds a fixed volume of water. If water escapes, the pressurisation unit automatically replaces it from the mains. Metering that replacement flow gives you a whole-network mass balance, so an unexplained rise in make-up water is direct evidence of a leak somewhere in the system.

Is make-up water monitoring a legal requirement for heat networks?

Ofgem’s rules, live since 27 January 2026, cover consumer protection, billing, service quality and transparency rather than pipe-level detail. The technical requirement for heat network leak detection sits in the Heat Network Technical Assurance Scheme, whose first standard is published in draft and requires both a meter on the top-up line and monitoring of top-up water to detect leaks.

How much make-up water is too much?

There is no universal figure, because it scales with system volume. The workable test is behavioural: any top-up you cannot trace to a recorded maintenance activity is worth investigating, and any consumption continuing overnight on a closed system is worth investigating urgently.

Can a clamp-on meter be fitted without draining the system?

Yes. Clamp-on ultrasonic transducers mount on the outside of the pipe and measure through the wall, so there is no cutting, no drain-down and no interruption to heat supply. That is why it suits occupied schemes where a planned shutdown is hard to arrange.

Does this replace buried pipe leak detection?

No. Wire-based surveillance in pre-insulated steel pipe locates a fault along the route. Make-up water monitoring tells you whether a fault exists anywhere at all, including in plastic pipe, risers and plant rooms that surveillance does not cover. Most schemes need the second before they can justify the first.

The leak is already telling you. Something has to be listening

Heat networks are moving from an unregulated corner of the energy market to a regulated utility supplying a fifth of the country’s heat. The assets going in now will be in the ground for decades, and the standard being written around them treats heat network leak detection as a metering duty, not a plumbing detail.

The uncomfortable part is that most schemes already have every component of that alarm except the instrument. The pressurisation unit is finding the leak for you, hour after hour, and then quietly fixing it so you never hear about it. For related reading see our non-invasive flow monitoring guide, the water sub-metering guide for how the same data structure works across an estate, and our water leak detection service.

Running a communal or district heating scheme and unsure what your top-up demand is doing? Talk to AQUAIOT about clamping a meter to the fill line and watching it for a month before committing to anything else.

By GP, Digital and IoT lead at AQUAIOT.