Chilled Water Flow Calculator for Data Centers | AKCP
Chilled water flow depends on the heat carried by the loop and the temperature rise between supply and return. For plain water, a useful metric estimate is flow in L/s = cooling duty in kW ÷ (4.18 × ΔT in °C). A 100 kW load at 5°C ΔT needs approximately 4.8 L/s, or 76 US GPM.
That calculation is a useful starting point when checking a CRAH coil, reviewing a chiller circuit or planning a liquid cooling connection. It also explains why a plant can become short of available flow even when the installed cooling capacity appears sufficient: the operating temperature rise may be smaller than the design assumed.
The AKCP Chilled Water Flow Calculator lets you work in metric or imperial units, calculate cooling duty from measured flow, and explore the effect of different temperature rises. Optional checks cover pipe velocity, pump electricity and measured performance.
Use the Chilled Water Flow Calculator
Select the quantity you want to calculate, then enter your loop conditions. The tool supports kW, refrigeration tons and BTU/h, with flow in L/s, L/min, m³/h or US GPM. Changing units converts the existing values.
Chilled Water Flow Calculator
Calculate water flow, cooling duty or temperature rise for a chiller, CRAH or liquid cooling loop. Check pipe velocity, pump electricity and measured performance.
1. Choose your calculation
Use the heat actually carried by this loop. Thermal kW is heat removal, not chiller electrical input. Evaluate facility and CDU secondary loops separately.
2. Coolant and distribution
3. Optional operating checks
Pipe velocity and minimum inside diameter
Pump electrical demand and annual cost
Uses total modeled flow and the entered pressure differential at that operating point. Include the complete active circuit's losses. This does not calculate pressure drop or select a pump.
Compare measured flow and ΔT
Enter simultaneous readings from the same loop and location. This check uses the same coolant properties. Compare measurements taken at similar load and operating conditions.
How the chilled-water calculator works
The calculator uses a steady-state energy balance for a single-phase liquid in a full pipe. Enter the heat carried by the selected loop, rather than all facility power or cooling electrical demand. For a CRAH coil this may include sensible and latent heat transferred to the water. CDU facility and secondary loops can have different temperatures, fluids and flow rates.
Flow (L/s) = 1,000 × Q ÷ (density × specific heat × ΔT)
ΔT (K) = 1,000 × Q ÷ (density × specific heat × flow)
Planned duty = base duty × (1 + flow planning allowance / 100), in Required flow mode only
Pipe velocity (m/s) = flow (m³/s) ÷ [π × inside diameter (m)² / 4]
Minimum inside diameter (m) = √[4 × flow (m³/s) ÷ (π × selected velocity ceiling)]
Hydraulic power (kW) = flow (m³/s) × differential pressure (kPa)
Electrical input (kW) = hydraulic power ÷ overall efficiency
Annual electricity (kWh) = electrical input × annual hours; cost = kWh × entered tariff
Plain water uses fixed approximate properties: density 998 kg/m³ and specific heat 4.18 kJ/(kg·K). The familiar Q = 500 × US GPM × ΔT°F formula uses rounded water properties; small differences are expected. Custom coolant properties are held constant at the user-selected operating condition. The equivalent imperial heat-transfer factor is reported for checking against supplier data.
1 US gallon = 3.785411784 L; 1 kW ≈ 3,412.142 BTU/h; 1 refrigeration ton = 12,000 BTU/h. ΔT°F = 1.8 × ΔT°C; absolute °F = 1.8 × °C + 32. Values convert before calculation and round only for display. US GPM is used throughout, not UK gallons.
Temperature pair inputs are limited to 0–80°C for this planning tool. The ΔT-only and inverse modes do not establish an absolute supply temperature or check freeze/boil limits. A positive heat load with zero flow cannot yield a finite ΔT. Pipe sizing requires friction, material and equipment checks beyond velocity. Pump cost assumes constant flow, pressure and efficiency for the stated hours, excluding idle or spare pumps.
Using measurements to investigate low ΔT
A smaller water temperature rise requires more flow to carry the same heat. For example, reducing ΔT from 5°C to 3°C increases theoretical flow by 67% at constant duty and coolant properties. This does not establish the actual pump electricity increase because circuit pressure and pump efficiency also change.
Compare matched supply/return sensors and a flow reading from the same circuit and time. A low ΔT may reflect low load, bypass mixing, control behavior, distribution problems or measurement error. The comparison identifies a difference from your model; it does not diagnose the cause.
Small ΔT makes heat calculations sensitive to temperature sensor uncertainty. Trend flow, differential pressure and supply/return temperatures alongside cooling duty. Review minimum flow and controls before changing pump speeds. Check insulation and local dew point for condensation at cold pipes, valves and CDU connections.
Calculation references
- Caleffi — Physical operating conditions in hydronic systems: liquid heat transfer, fluid properties, velocity and head.
- Caleffi — On-site measurements of circuit performance: measured flow and temperature difference.
- Grundfos — What is a pump curve?: flow, head, hydraulic power and efficiency.
References explain the underlying methods; they do not endorse this tool or establish project-specific limits.
Validate cooling performance with AKCP
Trend temperature, humidity and leak detection alongside supported third-party chiller, pump and CDU data in your AKCP monitoring system. Quicklime can bring compatible SNMP and Modbus readings into the operating picture.
How to Calculate Chilled Water Flow
The underlying energy balance is:
Cooling duty = mass flow × specific heat × temperature rise
For volumetric flow in metric units:
Flow (L/s) = 1,000 × cooling duty (kW) ÷ [density (kg/m³) × specific heat (kJ/kg·K) × ΔT (°C)]
Here, ΔT is return temperature minus supply temperature. The calculator uses approximate plain-water properties of 998 kg/m³ and 4.18 kJ/(kg·K), giving:
Water flow (L/s) = cooling duty (kW) ÷ (4.17164 × ΔT°C)
The familiar imperial estimate is:
Water flow (US GPM) ≈ cooling duty (BTU/h) ÷ (500 × ΔT°F)
For refrigeration tons, this becomes US GPM ≈ tons × 24 ÷ ΔT°F. The 500 factor uses rounded water properties, so it can differ slightly from the calculator. For other coolants, use density and specific heat at the operating temperature and concentration. Caleffi’s hydronic engineering reference explains this relationship.
Worked Example for a 100 kW Cooling Load
Consider a water loop carrying 100 kW of heat, supplied at 7°C and returning at 12°C. These are illustrative conditions, not recommended setpoints.
The water temperature rise is 12 − 7 = 5°C. Using the calculator’s default water properties:
Flow = 100 ÷ (4.17164 × 5) = 4.794 L/s
| Flow unit | Calculated requirement |
|---|---|
| L/s | 4.79 |
| L/min | 287.66 |
| m³/h | 17.26 |
| US GPM | 75.99 |
This is the flow needed to carry the stated heat at that ΔT. Equipment minimum flow, available pump pressure and distribution requirements still need checking.
Why Low ΔT Increases Flow Demand
At a constant heat load, required flow increases as the water temperature rise falls. The same 100 kW duty produces these results:
| Water temperature rise | Required flow in L/s | Required flow in US GPM |
|---|---|---|
| 3°C / 5.4°F | 7.99 | 126.65 |
| 5°C / 9°F | 4.79 | 75.99 |
| 7°C / 12.6°F | 3.42 | 54.28 |
| 10°C / 18°F | 2.40 | 38.00 |
Reducing ΔT from 5°C to 3°C increases theoretical flow by about 67%. A system constrained by flow may therefore struggle to deliver its expected cooling duty.
Persistent low ΔT at comparable operating conditions warrants investigation. Possible contributors include bypass mixing, valve behavior, coil performance and controls. Trane’s chilled-water design guidance discusses these issues.
A low reading alone does not identify a fault. Check the current load and the temperature measurements before changing operation. A larger ΔT is also subject to the coil, chiller or CDU performance limits; entering a higher value in the calculator does not make that operating condition achievable.
Calculate the Heat Your Loop Is Actually Removing
If you have a flow reading and matched supply and return temperatures, select Cooling duty from flow + ΔT. This reverses the calculation and estimates the heat carried by the water.
For example, plain water flowing at 6 L/s with a 4°C rise carries approximately:
Cooling duty = 4.17164 × 6 × 4 = 100.12 kW
Use readings from the same circuit and time, preferably during stable operation. Flow from one branch and temperatures from a mixed header may describe different heat balances. Small temperature differences make the result particularly sensitive to sensor error. Caleffi’s guidance on circuit measurements covers the measurement approach.
The calculator’s measured comparison shows the difference from your base modeled duty. Use it to check assumptions and identify conditions worth investigating.
Using Glycol or a CDU Secondary Coolant
A water-based glycol mixture has different fluid properties from plain water. Select Custom coolant and enter supplier density and specific heat for the actual concentration and mean operating temperature.
The tool calculates the thermal flow requirement from those values. It does not apply a viscosity correction or calculate the resulting pressure drop.
For a cooling distribution unit, calculate the facility loop and secondary loop separately. A heat exchanger transfers heat between the two, but their temperature differences and fluid properties can produce different flow requirements. Include relevant equipment losses when defining each loop’s duty.
Check Pipe Velocity and Pump Electricity
The pipe check uses the actual inside diameter and modeled flow to calculate mean velocity. Nominal pipe size alone is insufficient because the bore varies with pipe material and wall thickness. You can check a main carrying the total flow or one branch in an assumed equal split.
Enter your project’s velocity ceiling to calculate a theoretical minimum bore. The example ceiling of 2 m/s is editable. Final selection also needs pressure-drop, noise, material and equipment checks.
The pump estimate uses:
Electrical input (kW) = flow (m³/s) × differential pressure (kPa) ÷ overall efficiency
For the 100 kW example, 4.794 L/s at 150 kPa and 65% overall efficiency gives approximately 1.11 kW of pump electrical input. At a constant operating point for 8,760 hours, that is about 9,692 kWh per year, costing 1,163 currency units at a tariff of 0.12 per kWh.
Use the pressure differential and combined pump, motor and drive efficiency at the modeled operating point. Actual annual consumption changes with load, pressure control and efficiency. Grundfos explains why pump efficiency depends on the duty point.
Monitor Cooling Performance with AKCP
A useful operating record combines supply and return temperatures, flow, differential pressure, cooling load and equipment status. Review those trends alongside rack inlet temperatures and IT power to understand whether cooling is reaching the equipment.
Quicklime DCIM can bring AKCP sensor data together with compatible third-party equipment readings through protocols including SNMP and Modbus. Available chiller, pump and CDU points depend on the equipment interface and configuration.
Use AKCP sensorProbe+ monitoring for environmental visibility and suitable water leak detection around water-bearing equipment. Check coolant compatibility when selecting leak detection for glycol or other fluids.
The calculator is a manual browser tool. Copy operating readings into it to check flow and heat removal, then save the summary or CSV for comparison. It does not connect directly to Quicklime or your sensors.
Contact AKCP to discuss monitoring for your chilled-water plant, CRAH units and liquid cooling infrastructure.
Frequently Asked Questions (FAQ)
How much chilled water flow is required per kW?
For plain water, flow is approximately 1 ÷ (4.18 × ΔT°C) L/s per kW. At 5°C ΔT, that is about 0.048 L/s, or 2.88 L/min per kW. The required flow changes with temperature rise and fluid properties.
How many GPM are required per ton of cooling?
Using the rounded water formula, US GPM per refrigeration ton = 24 ÷ ΔT°F. At 10°F ΔT, the requirement is approximately 2.4 GPM per ton; at 12°F ΔT, it is 2 GPM per ton.
What is a good chilled water ΔT for a data center?
Use the temperature rise specified for your equipment and system design. The correct value depends on the cooling duty, supply temperature, heat-exchanger performance and flow limits. The calculator’s default 5°C is an editable example.
Can I use IT electrical power as the cooling load?
IT electrical power is a useful starting point for heat generation. Enter the portion of heat carried by the selected loop, plus relevant additional loads. In a facility using both air and liquid cooling, the full IT load may not belong to either individual circuit.
Does the calculator select a pump or pipe size?
It estimates thermal flow, pipe velocity and a theoretical bore at your selected ceiling. It also estimates pump electricity from entered pressure and efficiency. Pump and pipe selection require a hydraulic design and equipment performance data.
Should I add spare chiller capacity to the heat load?
Use the duty that the operating system must carry. Spare capacity does not create an additional simultaneous cooling load. Evaluate equipment failure scenarios and minimum-flow requirements separately.
