The pressure loss of a heating network is worked out section by section: linear loss in Pa/m multiplied by the length, plus the fitting losses of elbows, tees, valves and heat exchangers, all set against the head available at the circulator on the least favoured circuit. In domestic design, 10 to 20 mmWC/m is the usual target, roughly 100 to 200 Pa/m, which comes down to holding water velocity between 0.5 and 1.0 m/s in distribution. Since 1 August 2015, Regulation (EC) No 641/2009, as amended by Regulation (EU) No 622/2012, has required an energy efficiency index of EEI ≤ 0.23 for glandless circulators from 1 to 2,500 W: the oversizing margin that once absorbed a poor calculation is gone.
Understanding pressure loss in heating hydraulics
What pressure loss covers: friction, fittings, singularities
Pressure loss is the drop in pressure produced by water circulating, not to be confused with the static pressure read on the plant room gauge, which depends only on water column height. It comes from friction inside the pipe, a function of roughness, diameter and velocity, and from the fittings. A valve, a filter, a heat exchanger or a fouled circulator each add their own resistance, just like elbows, tees and reducers.
Linear and fitting losses: two calculations, one total
Linear losses depend on length, diameter and velocity. Fitting losses attach to each hard point, with a coefficient per fitting or, in a site approach, an equivalent length of straight pipe. In renovation work, it is the accumulation of connections along a constrained route that breaks the calculation, rarely one isolated section.
Why pressure loss unbalances the network
Water takes the least resistant path. When the diameter is too tight, velocity rises, pressure loss grows with its square, and the short branches capture the flow at the expense of the emitters at the end of the network. Sizing, balancing and circulator settings are handled together, never one without the others.
Calculating pressure loss on site
Flow rate first: from output to flow per section
The calculation starts with flow, not with diameter. Q (m³/h) = P (kW) / (1.163 × ΔT). A section feeding 12 kW at ΔT 15 K carries 0.69 m³/h; the same section at ΔT 7 K carries 1.47. That is why moving to low temperature, at equal output, can make a diameter inadequate that was perfectly comfortable at 80/60.
Flow and transmissible output by diameter
At a fixed velocity, every diameter has a ceiling flow rate, and therefore a ceiling output that depends on the design ΔT. The table below is calculated at 0.8 m/s, the top of the distribution band, with Q = area × velocity and P = Q × 1.163 × ΔT.
| Pipe (internal Ø) | Flow at 0.8 m/s | Output at ΔT 10 K | Output at ΔT 20 K |
|---|---|---|---|
| Copper 14 × 1 (12 mm) | 0.33 m³/h | 3.8 kW | 7.6 kW |
| Copper 16 × 1 (14 mm) | 0.44 m³/h | 5.2 kW | 10.3 kW |
| Copper 18 × 1 (16 mm) | 0.58 m³/h | 6.7 kW | 13.5 kW |
| Copper 22 × 1 (20 mm) | 0.90 m³/h | 10.5 kW | 21.0 kW |
| Copper 28 × 1 (26 mm) | 1.53 m³/h | 17.8 kW | 35.6 kW |
| Steel DN 32 (35.9 mm) | 2.92 m³/h | 33.9 kW | 67.8 kW |
It reads both ways: it gives the minimum diameter of a section from its output, and the maximum output an existing section can still carry before velocity turns noisy.
The fittings allowance and the ΔP reading that confirms it
Without manufacturer K coefficients, count the fittings as equivalent lengths of straight pipe, then add them to the linear figure. These values are site orders of magnitude, to be confirmed by a ΔP reading across the components.
| Fitting | Equivalent length of straight pipe |
|---|---|
| Standard elbow | 0.5 to 1 m |
| Tee on a straight run | 1 to 2 m |
| Open shut-off valve | 0.5 to 2 m |
| Heavily branched network, overall allowance | + 10 to 30% of the linear loss |
Sizing diameters and limiting pressure loss without extra cost
Choose the diameter on velocity, not on habit
A diameter that is too small is paid for in water velocity, and therefore in noise, erosion and pumping power. A diameter that is too large is paid for in copper, in installation time and in pointless inertia. The right trade-off is read off the table above, section by section, starting with the most heavily loaded.
Cutting lengths and elbows: the route is a free lever
Every metre and every elbow adds a resistance that nothing later compensates for. A direct route lowers pressure loss without changing a single component. Move the manifolds closer to the zones they serve, favour straight runs, and replace a string of small elbows with a simpler path wherever the building allows it.
Choose fittings on Kv, not on diameter
A valve is chosen on its Kv at nominal flow, not on the diameter of the pipe around it. An undersized filter becomes the hard point of the network as soon as it starts to load. The check valve, often left out of the calculation, weighs as much as a tee. The goal is low ΔP at nominal flow, to keep some control margin in hand.
Adjusting the hydraulics once the network is installed
Field readings: ΔP, flow, temperatures, valve positions
On an existing network, measure ΔP at the key points, circulator, filter, heat exchanger, then the actual flow rate and the flow and return temperatures by zone. These readings say whether the pump is working within its curve and whether the balancing holds. Note the positions of the balancing valves and lockshields as you go: it is the only way to step back.
Spotting the hard points: filter, sludge, valve, heat exchanger
An abnormal ΔP across a single component signals a local restriction. Look for the clogged filter, sludge at the foot of a riser, the partly closed valve, the scaled heat exchanger, or a section whose route was lengthened during an earlier job. A return temperature that is abnormally high points to a shortage of useful flow far more often than to a generator fault.
Circulator: read the curve before touching the speed
The operating point sits at the intersection of the circulator curve and the network curve. A pump that is too strong creates noise, wear and needless auxiliary consumption, without correcting the imbalance. Set proportional pressure on a variable-flow network, and read the detail of the choice in the article on variable-speed circulators.
Practical cases: pressure loss by network type
Radiator networks: single-pipe, two-pipe, risers
On single-pipe systems, pressure loss climbs as soon as the tappings are throttled, and the last radiator on the loop pays for the others. On two-pipe systems, distribution is set on the lockshields, with velocities held and a variable-speed circulator. On risers, track down the pointless loops inherited from successive reworks and stabilise differential pressure at the foot of the riser.
Underfloor heating: loop length and admissible ΔP at the manifold
On underfloor heating, everything hinges on the manifold. Keep loop lengths close to each other, without exceeding the order of 80 to 120 m depending on pipe diameter, and stay within a per-loop ΔP band that leaves margin at the circulator. The layout detail is covered in the article on sizing a low-temperature underfloor heating system.
Retrofit with a heat pump: guaranteeing the minimum flow
With a heat pump, excessive pressure loss translates directly into insufficient flow, a ΔT outside its band and fault shutdowns. The manufacturer's minimum flow is a design constraint, not a recommendation. When the existing network cannot guarantee it, decouple with a low-loss header or secure the flow with an adjusted bypass, and record the choice in the heat loss report built from the survey readings.
