1. Why heating pipes cannot be sized from a table alone
A heating pipe carries energy from the heat source to radiators, fan coils, heating coils or underfloor manifolds at the required flow. An undersized pipe increases velocity and friction at the same flow. The pump may need more differential pressure, remote circuits may receive insufficient water, and valves may become noisy. Oversizing increases pipe, fitting, insulation and installation costs. The objective is an economical diameter that carries design flow within accepted losses.
This guide addresses a water-based, pumped, closed heating circuit. Domestic hot water, open systems and steam need different assessments. The examples teach a method; actual pipe dimensions, temperature, equipment losses and project criteria remain final design inputs.
Follow this sequence: heat load → design temperature difference → flow → candidate bore → velocity and friction per metre → index circuit → pump and control verification. Skipping steps can produce a satisfactory-looking layout that performs poorly.
2. Assemble the design inputs
Determine each space's design heat loss and terminal demand. Boiler nameplate capacity is not the load carried by every section. In a building with 100 kW total load, a branch serving a 20 kW floor is sized for that branch.
Select supply and return temperatures with emitter performance. A 70/50 °C regime has a 20 K difference; 45/35 °C has 10 K and needs greater flow at the same load. These are examples, not compulsory differences for radiators, fan coils, underfloor heating or heat pumps.
Record the following for every section:
- Served heat load and circuit reference.
- Supply/return temperatures and mean fluid temperature.
- Density, specific heat and viscosity of water or glycol.
- Manufacturer, material, pipe series, outside diameter, wall thickness and actual bore.
- Supply/return lengths and types/counts of bends, tees, reducers and valves.
- Equipment pressure drops at design flow.
- Project velocity, friction, noise and control criteria.
Preliminary sizing can precede final routing. Recalculate with installation lengths; pump pressure based on an approximate distance is not a final result.
3. Calculate water flow from heat load
The energy balance relates heat output Φ, mass flow ṁ, specific heat cp and temperature difference ΔT. Separate symbols avoid confusing thermal power with volume flow V̇.
With Φ in W, density ρ in kg/m³, cp in J/(kg·K) and ΔT in K, flow is m³/s. Multiply by 1,000 for L/s or 3,600 for m³/h. A temperature difference has the same numerical value in K and °C; this is not an absolute-temperature conversion.
Assuming approximately ρ = 1,000 kg/m³ and cp = 4,180 J/(kg·K):
At 100 kW and 20 K, flow is approximately 4.31 m³/h or 1.196 L/s. A 20 kW branch needs 0.861 m³/h. These coefficients use approximate water properties; hot-water density differs. For glycol, use the manufacturer's temperature-dependent mixture properties.
4. What happens when the temperature difference decreases?
Halving ΔT approximately doubles flow and velocity at the same heat output. In turbulent flow, assuming little change in friction factor, loss can rise approximately fourfold because it depends on velocity squared. An existing diameter is therefore not automatically suitable for a new regime.
| Heat load | Design temperature difference | Approximate flow |
|---|---|---|
| 100 kW | 20 K | 4.31 m³/h |
| 100 kW | 10 K | 8.61 m³/h |
| 100 kW | 5 K | 17.22 m³/h |
A heat-pump retrofit must check emitter output at the new temperatures, equipment flow requirements, pipe losses and pump duty together. The 5 K value is a calculation scenario, not a universal heat-pump requirement.
Increasing ΔT on paper to disguise insufficient flow is not a solution. It changes mean emitter temperature and potentially output, and must remain compatible with equipment limits.
5. Nominal size, outside diameter and actual bore
Hydraulics use the actual internal diameter. DN is a nominal designation and need not equal measured bore. Different wall thicknesses at the same nominal size give different areas. Plastic and multilayer pipes are commonly identified by outside diameter.
A pipe with 32 mm outside diameter and 3 mm walls has an approximate 26 mm bore. Calculating a 32 mm bore overestimates area and understates velocity and friction. Confirm dimensions in manufacturer tables.
Fittings can have smaller passages despite matching outside dimensions. Use pressure-loss data for inserts and restrictions. Temperature/pressure ratings, oxygen permeability and equipment compatibility matter alongside hydraulics.
6. Check velocity in a candidate diameter
Use v in m/s, V̇ in m³/s and Di in metres. Enter a 41 mm bore as 0.041 m. Substituting m³/h and millimetres without conversion creates a unit error.
The 1.196 L/s example gives approximately 0.91 m/s in a 41 mm bore, but 1.94 m/s in a 28 mm bore. This increases friction too.
Acceptable velocity depends on circuit, diameter, noise sensitivity and controls. Caleffi discusses 2–4 ft/s, approximately 0.61–1.22 m/s, as guidance for certain applications. It is not a universal regulatory limit and should not automatically govern small emitter or underfloor circuits. Source: https://www.caleffi.com/en-us/blog/5-physical-operating-conditions-hydronic-systems
Low velocity alone does not prove incorrect sizing: air management, maintenance and cost matter. High velocity is not acceptable merely because the pipe withstands it; consider valve noise, fitting loss and pump energy.
7. Calculate straight-pipe friction
Darcy–Weisbach can be used:
Δp is in Pa, L in m and friction rate R in Pa/m. The Darcy factor λ depends on Reynolds number and relative roughness. Do not substitute a Fanning factor unchanged; the factors differ.
Re = ρ × v × Di / μ, where μ is dynamic viscosity. For laminar flow, the Darcy factor is 64/Re. Use an appropriate turbulent correlation or validated table for turbulent flow; one constant factor is unreliable in the transition region.
A project friction target makes preliminary selection systematic. The example assumes 100–300 Pa/m, not a mandatory range. Long routes may warrant lower losses for energy efficiency; underfloor circuits need their own assessment.
Check chart temperature, glycol concentration, pipe series and units. Pa/m and kPa/m differ by 1,000. Establish whether loss covers straight pipe only or also fittings.
8. Compare candidate bores for 100 kW
Flow is 4.306 m³/h, density 1,000 kg/m³ and λ is fixed at 0.025 for teaching. In practice, recalculate λ for each Reynolds number and roughness. The bores are example geometries, not a manufacturer's DN schedule.
| Actual bore | Approximate velocity | Straight-pipe loss | Preliminary assessment |
|---|---|---|---|
| 28 mm | 1.94 m/s | 1,685 Pa/m | Well above the assumed target |
| 35 mm | 1.24 m/s | 552 Pa/m | Above the assumed target |
| 41 mm | 0.91 m/s | 250 Pa/m | Meets the assumed target |
| 52 mm | 0.56 m/s | 76 Pa/m | Lower loss; compare cost and space |
A 41 mm bore is a candidate for further checking, not proof that DN40 is always required. Verify actual bore, pressure/temperature class and connection losses. Lower friction does not automatically make the largest pipe most economical.
The table concerns the 100 kW main. A 40 kW branch has different flow. Assign load, flow and diameter to every section rather than calculating all at 4.306 m³/h.
9. Include fittings, valves and equipment
Straight-pipe friction is insufficient for pump sizing. Bends, tees, reducers, check valves, strainers, control valves and terminals add losses.
Check which cross-section defines the coefficient's reference velocity. Straight-through and branch tee flow differ. Counting a fitting by both ζ and equivalent length counts it twice.
For water under stated manufacturer conditions, V̇ = Kv × √Δp applies with flow in m³/h and loss in bar. At 0.861 m³/h and assumed Kv = 2.72, loss is approximately 0.10 bar or 10 kPa. Distinguish fully open Kv from the effective value at the control/preset position.
Obtain radiator, fan-coil, coil, exchanger and heat-source drops from manufacturer curves at design flow. With hydraulic separation or multiple pumps, identify each pump's circuit equipment; select it for that boundary.
10. Index circuit and pump duty: a complete example
The index circuit is the path from discharge, through a terminal and back to suction requiring the highest differential pressure at design flow. It need not be the longest: a high-resistance terminal can make another path critical.
Here, the common main serves 100 kW, branch 40 kW and terminal connection 20 kW. All use 20 K. Each length includes supply plus return already. The calculation retains λ = 0.025 and approximate water properties.
| Section | Flow | Bore | Supply + return length | Friction rate | Straight-pipe loss |
|---|---|---|---|---|---|
| Common main | 4.306 m³/h | 41 mm | 30 m | 250 Pa/m | 7.50 kPa |
| Floor branch | 1.722 m³/h | 28 mm | 20 m | 270 Pa/m | 5.40 kPa |
| Terminal connection | 0.861 m³/h | 22 mm | 30 m | 225 Pa/m | 6.75 kPa |
Straight pipe totals 19.65 kPa. Assume fittings 4 kPa, control valve 10 kPa, terminal 12 kPa and strainer 3 kPa: total approximately 48.65 kPa. These are example inputs; actual components require calculated or catalogue values.
At ρ = 1,000 kg/m³ and g = 9.81 m/s², this is 4.96 m head. The common pump's preliminary duty is 4.31 m³/h at 4.96 m. Pump flow sums simultaneous circuits; pressure is the index requirement, not summed parallel losses.
Verify duty, efficiency, motor power, speed control and operating limits on the manufacturer curve. Resolve missing data instead of stacking arbitrary safety factors. Document any allowance.
11. Should building height be added to pump head?
In a completely water-filled closed loop, ascending and descending water columns balance statically. Adding building height directly to friction to select the circulator is incorrect. Reference: https://handbook.ashrae.org/Handbooks/S20/IP/s20_ch13/s20_ch13_ip.aspx
Height still matters for fill pressure, expansion-vessel precharge, positive pressure at the highest point, relief settings and equipment ratings. Absolute suction pressure and fluid temperature affect pump operation.
Open-tank transfer or different free surfaces involve static head differently. Do not generalise the closed-loop result to all water systems.
12. Balancing and control-valve selection
Correct diameters do not ensure distribution. Low-resistance circuits may over-flow while others receive too little. Balancing and controls distribute design flow. Reverse return helps but does not automatically equalise different terminals and valves.
Valve authority describes pressure distribution within the controlled hydraulic section, whose boundaries must be explicit. Conventionally, a = Δpv / (Δpv + Δprest), using design valve loss and other relevant circuit losses. Low authority can distort travel-to-flow response; a large Kv chosen only from connection size can worsen this.
Verify minimum differential pressure and flow-setting range for PICVs. They are not zero-loss devices. Avoid needless pump pressure while providing design flow. Reference: https://www.caleffi.com/en-us/blog/5-pressure-independent-control-valves-picv
13. Part load, glycol and existing systems
Closing two-way valves reduces flow; fixed-speed pump differential can rise. Investigate control mode and pressure setting before blaming diameter for noise or instability. Variable-speed settings must respect equipment minimum flow.
Glycol changes density, specific heat and viscosity, so flow and losses differ from water, particularly at low temperature. Recalculate for concentration and temperature rather than adding an arbitrary pump margin to water calculations.
Confirm existing pipes/fittings match drawings. Closed valves, blocked strainers, air and pump settings also cause underheating. Compare design and measurements before enlarging pipework.
14. Installation and commissioning checklist
Clean and vent the system, establish valve positions, then measure flow and differential pressure. Record open circuits and pump mode. Temperature difference with little load does not prove design heat output.
- Are heat load and ΔT recorded for every section?
- Are flows stated clearly in m³/h or L/s?
- Is manufacturer bore used rather than DN?
- Are velocity and friction based on consistent fluid properties?
- Does the index path include supply, return and components?
- Are parallel losses kept separate?
- Are balancing/control operating ranges verified?
- Are pressure/temperature ratings, insulation and access checked?
- Are pump duty and part-load settings recorded?
- Are measured flows, strainer condition and presets reported?
15. Common errors and symptoms
Remote radiators remain cold: Check flow, valves, air and strainers, then the index calculation. Excess nearby flow may indicate balancing; a larger pump is not the first remedy.
Noisy valves: Investigate differential pressure, selection and velocity. Valve noise does not prove the whole pipeline is undersized.
Insufficient flow after a heat-pump retrofit: Compare new flow and ΔT with the previous regime. Recalculate equipment connections, exchangers and strainers as well as mains.
Continuously high pump speed: Check settings, sensor position, index resistance and controls. Measure before assigning a single cause.
Frequent errors include confusing outside diameter with bore, using m³/h as m³/s, omitting return or equipment losses, and adding parallel pressure drops. Traceable calculations make errors easier to identify.
16. References and limits
These are independent teaching calculations. Approximate properties and fixed friction are explicit simplifications. Example bores, velocities and targets are neither a product selection table nor mandatory limits.
- ASHRAE Handbook, Hydronic Heating and Cooling. https://handbook.ashrae.org/Handbooks/S20/IP/s20_ch13/s20_ch13_ip.aspx
- Caleffi, Physical Operating Conditions in Hydronic Systems. https://www.caleffi.com/en-us/blog/5-physical-operating-conditions-hydronic-systems
- Caleffi, Pressure Independent Control Valves. https://www.caleffi.com/en-us/blog/5-pressure-independent-control-valves-picv
Final design uses adopted standards, local requirements and manufacturer data. Record loads, temperatures, fluid properties, bores, losses and selected duty together.