A chiller plant can have a good chiller and still run badly because the pump is oversized, the tower is selected on nominal tonnage, or the condenser water flow is set by a rule of thumb. The chiller then runs at a low evaporating temperature or a high condensing temperature, and the extra energy is charged to the compressor rather than to the pump or tower.
This article sets out the three calculations that decide the water side: flow from the temperature difference, head from the circuit, and tower approach from the wet-bulb. It uses SI units throughout and states the assumptions, so the numbers can be checked against a real project.
1. Flow comes from heat and ΔT, not from tonnage
The basic relation is V = 0.86 × Q ÷ ΔT, where V is flow in m³/h, Q is heat in kW and ΔT is the water temperature difference in K. The constant 0.86 comes from 3,600 ÷ (4.187 × 1,000).
For a 350 kW cooling load at a 5 K chilled-water difference: 0.86 × 350 ÷ 5 = 60.2 m³/h. At 7 K, the same load needs 43.0 m³/h. A 7 K design reduces pump and pipe size, but it also raises the evaporating temperature and can improve chiller COP; the optimum is not always 5 K.
Pipe size follows velocity, not a fixed rule. For 60 m³/h at 1.5–2.5 m/s, the internal diameter is 92–119 mm. A DN125 pipe gives about 1.4 m/s; DN100 gives about 2.1 m/s. Above 2.5 m/s, erosion and noise rise quickly; below 1.0 m/s, air separation and fouling become problems.
On the condenser side, heat rejection is roughly 1.2 times the cooling load for a standard vapor-compression cycle. For 350 kW, that is 420 kW. At a 5 K condenser-water rise, the flow is 0.86 × 420 ÷ 5 = 72.2 m³/h. If the tower is selected at a 4 K range, the flow rises to 90.3 m³/h — and so does the pump, the pipe and the tower distribution.
| Load (kW) | Chilled water at 5 K | Chilled water at 7 K | Condenser water at 5 K |
|---|---|---|---|
| 100 | 17.2 m³/h | 12.3 m³/h | 20.6 m³/h |
| 200 | 34.4 m³/h | 24.6 m³/h | 41.3 m³/h |
| 350 | 60.2 m³/h | 43.0 m³/h | 72.2 m³/h |
| 500 | 86.0 m³/h | 61.4 m³/h | 103.2 m³/h |
| 700 | 120.4 m³/h | 86.0 m³/h | 144.5 m³/h |
| 1,000 | 172.0 m³/h | 122.9 m³/h | 206.4 m³/h |
2. Build the head budget, then add a margin
Pump head is the sum of the pressure losses in the circuit, converted to meters of water: 1 kPa ≈ 0.102 m. For the 350 kW plant, a realistic budget is:
| Item | Pressure loss | Equivalent head |
|---|---|---|
| Evaporator | 60 kPa | 6.1 m |
| Terminal units and control valves | 40 kPa | 4.1 m |
| Straight pipe, 200 m at 150 Pa/m | 30 kPa | 3.1 m |
| Fittings, 40% of straight pipe | 12 kPa | 1.2 m |
| mechanical room valves, strainer and header | 20 kPa | 2.0 m |
| Subtotal | 162 kPa | 16.5 m |
| 12% margin | 181 kPa | 18.5 m |
Select a pump with 20 m head at 60 m³/h, not 30 m. The margin should cover calculation uncertainty and filter loading, not compensate for a guessed pipe length. A 20% margin on top of a 40% overestimate is the most common cause of valve throttling and noise.
Shaft power is P = ρgQH ÷ (3,600 × 1,000 × η). At 60 m³/h, 20 m and 65% efficiency: 1,000 × 9.81 × 20 × (60 ÷ 3,600) ÷ 0.65 ÷ 1,000 = 5.0 kW. A 5.5 kW motor is adequate; a 7.5 kW motor is not needed unless the pump curve has a steep rise or the duty point is uncertain.
With a variable-speed pump, the head falls with the square of speed and the power with the cube. At 80% speed, head is 64% and power 51%. But the pump must still meet the minimum flow through the evaporator — usually 60–70% of design — and the chiller’s minimum flow switch must not trip. A bypass valve controlled by flow, not by differential pressure alone, is the safer arrangement.
3. The tower is selected on wet-bulb and approach, not nominal tonnage
A cooling tower’s “nominal tonnage” is a catalog value at a defined condition, usually 35 °C entering water, 29.4 °C leaving water, 25.5 °C wet-bulb and 3 K approach. If the project’s wet-bulb is 28 °C and the approach is 4 K, the tower must be larger than the nominal figure.
Approach is the difference between leaving water and wet-bulb. A 4 K approach at a 28 °C wet-bulb gives 32 °C leaving water. With a 5 K range, the return is 37 °C. If the wet-bulb falls to 26 °C, the same tower can leave water at 30 °C; the condenser temperature falls by 2 K and the chiller COP rises by roughly 5%. That is why a tower selected for the design wet-bulb performs better for most of the year.
The condenser-water flow must match the tower’s distribution. A tower selected for 72 m³/h at a 5 K range but supplied at 90 m³/h will overflow the distribution troughs and splash out. A tower selected for 90 m³/h but supplied at 72 m³/h will have poor distribution and scale in dry areas. Select the tower on flow and range together, then check the spray pressure and the basin volume.
Water balance is often neglected. For 72 m³/h and a 5 K range, evaporation is about 72 × 5 × 0.0017 = 0.61 m³/h, or 0.85% of flow per 5 K. With blowdown at 3–5 cycles of concentration, total makeup is about 0.8–1.0 m³/h. A DN25 makeup line is adequate; a DN15 line will not keep up during a hot afternoon. The basin should hold 1.5–2 minutes of flow, and the makeup valve should be a mechanical float or a level probe with a slow-close solenoid, not a manual valve.
4. Four mistakes that show up as chiller faults
Oversized head. A pump selected at 30 m for a 16.5 m circuit runs at 55–65% valve opening, with the valve throttling 10–14 m. The energy is lost at the valve, and the pump runs at a false duty point. Measure the actual differential pressure before accepting the selection.
Flow margin on top of flow margin. Adding 10% to the load and 10% to the flow gives 121% of design flow, a 46% rise in pump power, and a 20% rise in pipe velocity. The chiller may then trip on high condenser pressure in summer because the tower is undersized relative to the actual flow.
Tower sized on chiller tonnage. A 350 kW chiller does not need a “350-ton” tower. It needs a tower that rejects 420 kW at the project wet-bulb and approach. In a hot, humid climate, the correct tower may be two sizes larger than the nominal label.
Multiple cells without water balance. Two towers in parallel need a common basin or a balancing pipe sized for the full flow. Without it, one cell runs dry while the other overflows; the dry cell scales and loses performance, and the wet cell carries the load.
5. Commissioning checks
Measure chilled-water flow at the evaporator and at the farthest terminal; the difference should be within 10%. Measure the pump differential pressure and compare it with the calculated head; a 20% difference usually means a closed valve, a blocked strainer or a wrong impeller trim.
On the tower, measure entering and leaving water temperature, wet-bulb and spray pressure. Calculate the approach; if it is more than 1 K above the catalog value, check distribution, fill fouling and air flow. Measure makeup and blowdown; a rising basin level with the makeup closed means the fill is plugged or the blowdown valve is passing.
Record the condenser-water flow and the chiller’s condensing pressure at full load. A high condensing pressure at design flow points to a tower problem; a high condensing pressure at low flow points to a pump or valve problem.
Key criteria for the water side
The flow and head tables size the pipes; the table below is the acceptance envelope that keeps the plant stable after the contractor leaves. It applies to a conventional chilled-water plant with a 5 K design ΔT and a cooling tower selected on the design wet-bulb.
| Criterion | Recommended value | Effect of deviation | Check method |
|---|---|---|---|
| Supply-to-return ΔT | 5 K | Below 4 K, flow is too high and pump energy rises | Supply and return thermometers |
| Pump flow margin | 10% | Too much margin shrinks ΔT and can make the chiller cycle | Compare with design flow |
| Pump head margin | 10–15% | Too much margin throttles valves and wastes energy | Head budget and valve position |
| Cooling-tower approach | 3–5 K | A higher approach raises condensing temperature and compressor power | Tower curve at design wet-bulb |
| Evaporator minimum flow | 60–70% of rated | Lower flow risks poor heat transfer and freezing | VFD lower limit and flow switch |
These are not catalog values. The 5 K ΔT is a design choice; if the terminals are selected at 7 K, the flow is lower and the pump can be smaller, but the coil must be selected for that ΔT. The 10% flow margin is for control stability, not for future load.
Check the tower approach at the design wet-bulb, not on a mild day. A 3–5 K approach is normal for a well-selected tower; 1 K above the catalog may be distribution, fill fouling or air flow. If the condensing pressure is high at design flow, the tower is the first suspect; if it is high only at low flow, the pump or a bypass is the first suspect.
Record the criteria in the commissioning sheet and trend them for the first month. Longcom’s water chiller page gives the pump, tower and control options behind these criteria.
Common questions
Should chilled water be designed at 5 K or 7 K?
7 K reduces flow and pump energy, and raises the evaporating temperature, which improves chiller COP. It also requires larger terminals and may not suit a low-humidity room. Run both options through the annual energy calculation; do not choose by habit.
How much pump head margin is right?
10–15% on a well-calculated circuit. If the pipe length is unknown, measure it or use a conservative length, but do not add 30–40% and then throttle.
Can a variable-speed pump save energy on a chiller?
Yes, but only if the terminals have variable flow and the chiller allows a minimum evaporator flow. Keep a bypass and a flow switch; otherwise the chiller will trip at low load.
Why does the tower approach worsen over time?
Fill fouling, distribution nozzle blockage, fan belt slip and air recirculation. A 1 K rise in approach raises the condensing temperature by about 1 K and the compressor power by 2–3%.
How much makeup water does a tower need?
Evaporation is about 0.5% of flow per 5 K, plus blowdown and drift. For 72 m³/h, plan 0.8–1.0 m³/h and a DN25 line. Check the actual blowdown rate; too much wastes water, too little scales the fill.
Can two chillers share one tower?
Yes, with a common basin or a balancing pipe, and with the tower flow matched to the number of running chillers. A fixed-flow tower on a variable-flow plant will either overflow or starve the fill.
How do I calculate pump head?
Add the chiller, tower, valves, filter, pipe and static head at design flow. Use the critical branch, not the pump header. Add a margin for fouling and balance. A head budget without the control valve is incomplete.
What tower approach should I assume?
Use the manufacturer’s approach at the design wet-bulb, typically 3–5 °C. A tighter approach needs a larger tower. Check the annual energy, not just the capital cost.
Related reading
Chiller cooling-load calculation · Cooling-tower selection for chillers · Chilled-water piping and insulation · Water-chiller commissioning
Longcom Environmental Protection Equipment Co., Ltd. can check the flow, head and tower approach for a chiller plant, and review the pump and tower selection against the actual wet-bulb. For project data, call 400-878-0280 or 13002001280.
Related reading
Cooling tower selection · Chiller water treatment · Chiller plant sensors