- UK heat networks are built to CIBSE/ADE CP1 (Heat Networks: Code of Practice), which sets the design flow and return temperatures, differential-pressure limits and commissioning duties that dictate valve selection at every substation.
- Modern hot-water networks run flow temperatures of 90–120°C and returns as low as 40–50°C, so buried valves see thermal swings above 60°C every load cycle — steel bodies, welded ends and PN25 ratings are the norm, not PN16.
- Pre-insulated buried valves carry the same copper leak-detection wires (Nordic or impulse system) as the pipe, so the alarm loop runs unbroken through the valve — break that continuity and you blind the whole network to water ingress.
District heating valves live a harder life than most people credit, because they combine buried, unmaintainable installation with hot water, aggressive thermal cycling and a demand for decades of leak-free service. A heat network moves heat from a central energy centre — a CHP engine, a biomass boiler, a large heat pump or an energy-from-waste plant — to hundreds or thousands of buildings through a closed loop of flow and return pipes. Every isolation point, branch and substation on that loop is a valve, and getting those valves wrong is expensive in a way you cannot easily dig up and fix.
I have seen a network commissioned beautifully on paper crippled by a single wafer butterfly valve chosen for a chilled-water spec and dropped into 110°C flow. The EPDM seat took a set, the disc stuck, and the balancing was never achievable. The lesson is that district heating is its own discipline: the pressures look modest, but the temperature, the cycling and the buried, welded, walk-away nature of the installation change everything about how you specify.
This guide walks through the network from energy centre to dwelling: buried pre-insulated isolation valves, differential-pressure and flow control at substations, direct versus indirect connection, the material and elastomer choices that hot-water and chilled-glycol duties force, and the CP1 commissioning discipline that ties it all together.
District Heating Valves and the Network Architecture
A heat network is a tree. A primary distribution main leaves the energy centre, branches into secondary mains through district chambers, and finally reaches building substations or, in a communal scheme, individual heat interface units (HIUs). Each branch needs isolation so that a fault or a new connection can be handled without draining the whole network, and each substation needs control valves to draw exactly the heat the building demands and no more.
The primary side is the utility’s asset and is designed for reliability and low heat loss. Secondary sides — the building’s own heating and hot-water circuits — are hydraulically separated by a plate heat exchanger in an indirect scheme. The interface between the two is where most of the interesting valve engineering happens, because that is where control accuracy, differential pressure and return-temperature performance are won or lost.
Networks are graded low-, medium- or high-temperature. Traditional schemes ran 120°C flow at PN16 or PN25; the fourth-generation networks now being built target 70–90°C flow to suit heat pumps and cut heat loss, and fifth-generation ambient-loop schemes run near ground temperature with heat pumps in every building. Lower temperature is kinder to valves, but it demands better control, because a low delta-T at the substation wrecks network economics.
Buried Pre-Insulated Isolation Valves
Out in the ground, the workhorse is the fully welded, pre-insulated ball or butterfly valve buried directly in the trench alongside the pipe. There is no chamber, no gland to repack and no annual inspection — the valve is designed to be welded in, backfilled and forgotten for the 30-year design life of the network. That single fact drives everything about its construction.
Buried ball valves dominate the smaller sizes, typically up to DN200. They are full-bore or reduced-bore floating or trunnion designs with a welded steel body, a hard-chromed or nitrided ball, and reinforced PTFE or PTFE-carbon seats rated for the flow temperature. The stem extends up through a pre-insulated riser to a surface box, so the only maintenance point — the operating spindle — is accessible while the body stays buried and sealed. Above DN200, buried butterfly valves take over on cost and weight grounds, usually double-offset with a metal or high-temperature seat.
The critical detail is the transition between the valve body and the pre-insulated pipe. The factory applies the same polyurethane (PUR) foam insulation and high-density polyethylene (HDPE) casing over the valve as over the pipe, and the field joint at each end is made with a shrink sleeve or welded casing joint exactly as a pipe joint would be. A poorly made casing joint at a valve is a classic water-ingress point, and once water is in the foam the steel corrodes and the leak-detection alarm trips.
Leak-detection continuity through the valve
Pre-insulated systems carry surveillance wires in the foam — typically two copper conductors running the length of every pipe (the Nordic/BRANDES or the impulse/Nordic resistance systems are the two common families). A monitoring unit measures resistance and capacitance along the loop; moisture in the foam changes the reading and pinpoints the fault to within a metre or two. The valve is part of that loop, so its pre-insulated risers and body section carry the wires straight through, joined to the pipe wires at each field joint.
Get the wiring wrong at a valve — cross the conductors, leave a joint open, or nick a wire during backfill — and you either blind a whole leg of the network or generate permanent false alarms. On commissioning I always insist on a full loop-resistance test after the valve field joints are made and before backfill, because tracing a broken surveillance wire under a road afterwards is a genuinely miserable job.
Thermal Expansion and Cyclic Loading
Buried heat-network pipe is not free to move, so thermal expansion becomes a structural design problem that reaches right into the valves. A steel main heated from 10°C ambient to 120°C flow wants to grow roughly 1.3 mm per metre; over a long straight run that is tens of millimetres of movement, restrained by soil friction and managed with expansion loops, cushions and, on older systems, expansion bellows.
A valve welded into that main sees the axial thrust and bending that the expanding pipe delivers. This is why buried network valves are full-penetration butt-welded into the line rather than flanged: a welded joint carries the load and cannot loosen or leak under cyclic stress the way a bolted flange can. Body designs are proven for the combined pressure, thermal and bending loads, and the valve should never be used as an anchor point unless the designer has explicitly accounted for it.
The cycling is what quietly kills components. A network does not sit at 120°C; it modulates with heat demand, so seats, seals and stems see thousands of thermal cycles over their life. Seat materials must tolerate the swing without cold-flowing or taking a compression set, stem seals must stay tight through the expansion and contraction of the stem itself, and any elastomer has to survive both the peak temperature and the fatigue of the cycle. Rating a valve for the peak static temperature is not enough; it must be rated for the cyclic duty.
Differential-Pressure and Flow Control at Substations
At the building substation the game changes from isolation to control. The network delivers a differential pressure across the connection that varies with the building’s position on the network and with overall demand, and the substation has to draw its design flow regardless of that varying differential. This is the core hydraulic challenge of any heat network, and it is solved with differential-pressure control valves (DPCVs) and pressure-independent control valves (PICVs).
A DPCV holds a constant differential across the substation’s control circuit, so the two-port control valve behaves predictably no matter what the network pressure does upstream. It is a self-acting valve, driven by an impulse line sensing the pressure it controls, and it removes the interaction between neighbouring substations that otherwise makes a network impossible to balance. Without differential-pressure control, connecting a new building can starve or over-pressure the ones already on the leg.
The PICV combines this into one device: a control valve, a differential-pressure regulator and a flow limiter in a single body. It holds the design maximum flow independent of network pressure fluctuations, so the modulating actuator on top only has to position for load, not fight pressure swings. For heat networks the PICV is now close to standard at the substation because it guarantees the maximum flow will never be exceeded — which is the single most effective way to protect network delta-T and return temperature.
Why return temperature rules everything
CP1 hammers on low return temperature for a reason: the colder the water going back to the energy centre, the more heat each litre has delivered, the smaller the pipes can be, and the better a CHP or heat pump runs. A control valve that lets flow run on when the building does not need it — because it is oversized, hunting, or not pressure-independent — pushes hot water straight back to the return and destroys the delta-T. Specifying accurate, correctly sized PICVs and commissioning them properly is worth more to a network’s economics than almost any other single decision.
Direct Versus Indirect Connection
How the building couples to the network decides the valve duty on the secondary side. In a direct connection the network water flows through the building’s own emitters, with only pressure-reduction and mixing valves between. It is cheap and thermally efficient but exposes the building circuit to full network pressure and temperature, and any leak in the building loses network water and treatment chemistry. Direct connection demands robust pressure control and is increasingly reserved for lower-temperature networks.
The indirect connection puts a plate heat exchanger between network and building, hydraulically separating the two. The building circuit runs its own pressure, its own water treatment and its own fill, and the network sees only the exchanger. This is the norm for most UK schemes because it contains leaks, simplifies billing at the interface, and lets the building operate at a pressure and temperature suited to its emitters. The penalty is the few degrees of approach temperature the exchanger costs, which nibbles at return temperature.
On the indirect primary side you need a modulating control valve (usually a PICV) to draw heat through the exchanger, tight-shutoff isolation each side for maintenance, and often a strainer to protect the exchanger and the valve trim. The control valve here must give tight shutoff at low load to protect return temperature; a leaking two-port valve bleeds hot water through the exchanger continuously and shows up as a stubbornly warm return that no amount of balancing will cure.
Materials, Ratings and Elastomers for Hot and Chilled Duties
Body materials follow the temperature and the buried duty. Buried network valves are carbon steel — typically forged or a cast grade equivalent to A216 WCB — with welded ends and a corrosion-protected exterior under the insulation. Bronze and gunmetal appear on smaller building-side fittings, and stainless (304/316) is used where water chemistry or a chilled/condensing duty demands it. The pressure rating is usually PN16 or PN25; older high-temperature schemes and tall buildings with high static heads push to PN25 as standard.
Elastomer choice is where heating people and cooling people part company, and mixing them up is the classic failure. For hot water up to around 120°C, EPDM is the default seat and O-ring material — it handles hot water and steam well but is destroyed by mineral oils. Where temperatures or glycol concentrations climb, FKM or PTFE-based seats take over. The one rule you never break: do not put hot-water EPDM into an oil-contaminated system, and never let a cooling contractor’s oil-resistant NBR seat end up in a hot flow line, because NBR hardens and fails quickly above about 100°C.
Chilled and ambient-loop networks change the emphasis. The temperatures are gentle, but the fluid is usually a water-glycol mix for frost protection, and glycol is more viscous, mildly attacks some elastomers, and corrodes certain alloys if the inhibitor package is depleted. Chilled valves must also handle condensation and the risk of external corrosion and, on ambient loops, the same buried pre-insulated construction as heating but insulated to stop condensation and gain rather than loss. EPDM remains a safe seat choice for glycol; verify inhibitor compatibility with the valve internals.
Valve Types by Position on the Network
Each part of the network calls for a different valve. Buried mains isolation is ball (to DN200) or double-offset butterfly (above), welded and pre-insulated. District chambers and branch tees use the same, with actuation where remote operation is wanted. Substation isolation is usually flanged or threaded ball or butterfly valves in stainless or bronze, sized for tight shutoff.
Control is PICV or DPCV plus two-port modulating valves on the primary; the secondary building side adds its own balancing valves, three-port mixing or diverting valves on radiator and underfloor circuits, and thermostatic control at the emitters. Safety and ancillary duties bring in pressure-relief valves sized for the closed-loop expansion, automatic air vents at high points, drain and fill valves at low points, and dirt separators or strainers protecting exchangers and control trim. None of these is exotic on its own; the skill is matching each to the network’s temperature, pressure and cycling regime.
Commissioning, Balancing and CP1 Compliance
A heat network is only as good as its commissioning, and CP1 makes this explicit with objectives for design temperatures, differential pressure and, above all, low return temperature. The commissioning engineer’s job is to prove that every substation draws its design flow at the available differential and returns water as cold as the design promised, across the range of demand the network will actually see.
Balancing a pressure-independent network is different from balancing a traditional variable-volume system. Where PICVs are used, each is set to its design maximum flow and the differential-pressure control does the rest, which removes most of the iterative flushing-and-measuring of a classic proportional balance. The commissioning effort shifts to verifying flow settings, confirming the DPCVs hold their setpoint, proving control-valve authority, and hunting down any valve leaking hot water through to the return.
Two practical points close this out. First, flush and clean the network thoroughly before the control valves and exchangers see full flow — construction debris destroys PICV cartridges and PHE channels, and a strainer differential is worth monitoring for the first weeks. Second, record every valve’s set flow, the leak-detection loop readings and the isolation-valve positions in the network model at handover, because a heat network is a 30-year asset and the person who has to extend it in 2040 will need to know exactly what you buried and where.
District heating rewards engineers who respect the difference between a modest pressure and a punishing duty. The pressures rarely exceed PN25, but the combination of high temperature, relentless cycling, buried construction and a demand for decades of leak-free, low-return-temperature performance makes valve selection unforgiving. Choose welded pre-insulated bodies for the buried mains, protect the leak-detection loop through every valve, put pressure-independent control at every substation, and commission to CP1 as if the network’s whole business case depends on it — because it does.