1. What problem does balancing solve?
Hydronic balancing aims to deliver each circuit's required design flow. Correct total pump flow does not prove correct distribution: short, low-resistance circuits may take too much while remote circuits receive too little. A larger pump can worsen that imbalance.
Calculate flow from heat demand, fluid properties and temperature difference. For 20 kW and 20 K, water flow is approximately 0.861 m³/h. Match this to actual terminal performance. Equal flow is not correct balance where terminal loads differ.
Check part-load behaviour as control valves close, not only one measurement with every valve open.
2. Static and dynamic methods
Manual balancing adds adjustable resistance and may provide measurement ports. A fixed setting does not mean fixed flow when differential pressure changes.
Automatic flow limiters cap flow within an approved pressure range. Differential-pressure control valves manage pressure between defined points. Pressure-independent control valves can combine flow presetting with modulation under their specified conditions. These functions are not interchangeable.
| Device | Selection question |
|---|---|
| Manual balancing valve | Are measurement range and design loss suitable? |
| Automatic flow limiter | Is required flow maintained within the pressure range? |
| DPCV | Are the controlled section and pressure setting correct? |
| PICV | Are preset flow, minimum differential and actuator suitable? |
Reference: https://www.caleffi.com/en-int/products/balancing-and-control-devices
3. Critical circuit and pump selection
Calculate each circuit's flow and loss. Common pipes carry combined connected demand; a terminal branch carries its own flow. Select pump flow from simultaneous total demand and head from the critical path.
Do not add parallel losses: branches needing 20/30/40 kPa do not require 90 kPa pump head. Add common-route losses and required control-device differentials to the relevant path, then balance lower-resistance routes.
Check pump curves, minimum flow and variable-speed strategy. Consuming unnecessary pump pressure in balancing valves wastes energy; link design requirements and operational settings in the same records.
4. Control-valve authority
A simplified authority definition is:
Define the elements included in Δprest. Mixing pump, common-pipe and terminal boundaries produces different results. Design losses and available differential with a closed valve are not the same condition.
Low authority can impair the relationship between valve motion and flow. Choosing excessive Kv from connection size reduces open-valve loss but does not automatically improve control. Select a target for the system, product and control objective. Background: https://www.caleffi.com/en-us/blog/2-balancing-valve-fundamentals
5. Step-by-step Kv example
For water under the manufacturer's applicable conditions, using m³/h and bar:
For a 20 kW terminal at 20 K, flow is 0.861 m³/h. Assume other terminal losses of 15 kPa and an illustrative authority target 0.40. Required valve loss is 0.40/(1−0.40)×15=10 kPa=0.10 bar, giving Kv≈0.861/√0.10=2.72.
| Candidate Kv | Valve loss at 0.861 m³/h | Authority within this boundary |
|---|---|---|
| 2.72 | 10.0 kPa | 0.40 |
| 4.00 | 4.63 kPa | 0.236 |
| 6.30 | 1.87 kPa | 0.111 |
Larger Kv lowers loss and authority in this example. The 0.40 target is not universal: verify characteristic, pressure arrangement and control objectives. Check the actual product setting and actuator separately.
6. Pressure changes at part load
At fixed Kv and fluid properties, local flow varies approximately with √Δp. Raising valve differential from 10 to 40 kPa can double its flow at the same opening. This describes a local valve, not a prediction that the entire system doubles its flow.
Two-way valve closure changes branch pressures and pump control. Pressure regulation or suitable PICVs may be needed. Sensor location at the pump or a remote circuit changes control behaviour; document what the setpoint represents.
For low-load noise, measure valve differential and pump settings before blaming pipe diameter. Further throttling a static valve can impair remote flow at full load.
7. PICV minimum differential
Pressure independence does not apply at zero differential or unlimited pressure. Verify minimum operating differential, maximum limit and flow-setting range. Include the required minimum in the critical-route calculation.
If other terminal losses are 18 kPa and an assumed PICV needs 25 kPa, the terminal boundary needs at least 43 kPa. The 25 kPa is educational; obtain the actual value from the selected product. Add common losses separately.
Reference: https://www.caleffi.com/en-us/blog/5-pressure-independent-control-valves-picv
8. Site balancing sequence
Clean and purge the system, check filters and flow directions, and verify actual actuator travel before balancing. Adjusting around trapped air or dirty filters produces settings that change after maintenance.
Label design and measurement points. Use the exact product's Kv/preset data, not another model's scale. Define balancing method and pump conditions first. Reassess other branches after adjusting one.
Record design/measured flow, preset, differential and pump mode. Acceptance tolerances come from the specification, not a universal percentage in this article.
9. Interpret water temperature difference correctly
Supply/return difference reflects flow and heat transfer at that moment. Small ΔT without design load does not automatically prove excess flow; large ΔT can indicate low flow, high demand or different equipment conditions.
Do not balance by making every terminal's ΔT equal. Review loads, types and conditions, combining flow/pressure with temperatures and capacity. Reference: https://www.caleffi.com/en-us/blog/9-hydronic-distribution-systems
10. Checklist and common mistakes
Check load-based design flow, correct product measurement data, critical/common boundaries, explicit authority target, PICV/DPCV limits, full/part-load pump control, filters/air/actuators/sensors and documented final settings.
Avoid treating large Kv as better control, wasting pump head in valves, drawing firm conclusions from unloaded ΔT and omitting dynamic-device minimum pressure.
11. Limits of flow and temperature measurements
Equal return temperatures do not prove correct terminal flow: loads, openings and inlet temperatures can differ. Compare calculated design flow with suitable valve measurements or meters; use temperatures to interpret transported heat.
An illustrative 4 m³/h at 10 K transports about 46.5 kW. If current load is 20 kW, the same flow produces a smaller difference without necessarily indicating bad balance. Separate full- and part-load acceptance conditions. Glycol changes density, heat capacity and viscosity, affecting both thermal and hydraulic calculations.
Flow inferred from valve differential requires Kv at the actual preset. Check hose connections, zero, air and clean ports. Uncertainty grows at low differentials; stay within the instrument and product measurement ranges.
12. Balancing order and part-load verification
After cleaning, purging and filter/valve checks, measure at defined pump/control conditions. Record adjustment order and repeat checks because branches interact. Continually increasing pump pressure can worsen overflow nearby; aim for the lowest suitable pressure meeting requirements.
Safely close selected controls during part-load tests and observe open-terminal flow, noise and stability. A PICV below its working range is not pressure independent. Check upper limits and actuators too. Incorrect DPCV sensing connections can invalidate control.
Handover schedules should contain terminal code, design/measured flow, valve model, preset, pressure, pump mode and date. Use specification tolerances. Reassess controls and critical circuits after pump replacement or zone additions.
13. Sources and calculation limits
https://www.caleffi.com/en-us/blog/2-balancing-valve-fundamentals ; https://www.caleffi.com/en-us/blog/5-pressure-independent-control-valves-picv ; https://www.caleffi.com/en-int/products/balancing-and-control-devices
Examples use approximate water properties and stated boundaries. Verify final selections for actual fluid, temperature, valve characteristic, actuator and control strategy.