AirSelect3D Blog
AHU Airflow Sizing: From Heat Load (kW) to Supply-Return ΔT
A load calculation hands you a number in kW. An AHU schedule needs a number in m³/h. The step between them is one equation and one design decision — the sensible heat equation, and the supply-return ΔT you choose to solve it with — and getting that decision wrong is what makes a "correctly sized" coil deliver a draughty, oversized, or high-SFP unit.
The equation that connects load, airflow and ΔT
For dry air at standard density, the sensible heat balance across the room is:
Q (kW) = 0.336 × V (m³/h) × ΔT (K) / 1000
Rearranged for the number every AHU schedule actually needs:
V (m³/h) ≈ 3000 × Q (kW) / ΔT (K)
This is the "rule of 3000" HVAC engineers use for a sanity check before the software runs the full psychrometrics: airflow is inversely proportional to the ΔT you're willing to accept between supply air and room air. Double the ΔT, halve the airflow — and halve the casing face area, the fan duty, and (all else equal) the SFP that goes on the data sheet.
Why ΔT isn't a free variable
If airflow scales inversely with ΔT, the obvious move is to push ΔT as high as possible and shrink everything downstream. Three constraints stop that:
- Draught risk. EN 16798-1 ties supply-air temperature difference to the room's ventilation effectiveness and occupied-zone velocity. Mixing ventilation typically tolerates a higher ΔT (commonly up to 8–10 K) than displacement or low-velocity diffusion strategies (often capped nearer 4–6 K), because a cold, low-momentum supply jet falls into the occupied zone before it mixes.
- Coil feasibility. A larger ΔT on the cooling side means a colder supply temperature, which pushes the coil closer to its apparatus dew point (ADP) — the same ADP/bypass-factor ceiling covered in our cooling coil selection guide. A ΔT the coil's row count and fin pitch can't actually reach is a paper number, not a spec.
- Latent load coverage. The sensible-only equation above says nothing about moisture. A zone with a high latent fraction (dense occupancy, process moisture) needs enough airflow to also carry the latent load, which can push the required V above what the sensible-only ΔT calculation implies — the two have to be checked together, not just the sensible one.
Worked example
Office zone, 45 kW sensible cooling load, three ΔT choices:
| ΔT (K) | Airflow V (m³/h) | Effect |
|---|---|---|
| 6 | ≈ 22,500 | Safer for displacement/low-velocity diffusion; larger casing, higher fan duty |
| 8 | ≈ 16,875 | Common mixing-ventilation compromise |
| 10 | ≈ 13,500 | Smallest casing and fan duty; supply temperature drops further toward the coil's ADP — verify it's reachable |
The airflow swing between the 6 K and 10 K columns is 67% — the single biggest lever in the whole sizing exercise, decided before a single coil row or fan curve is picked.
From airflow to the rest of the schedule
Once V is fixed, it drives every downstream selection: casing width and height come from dividing V by the target face velocity, fan duty comes from V plus the system's total pressure drop, and the coil's actual leaving-air state — the number that has to match the ΔT you assumed — comes from the psychrometric chain described in our five state points guide. If any of those come back inconsistent with the ΔT you started from, the airflow has to be re-run, not patched.
Where the tooling matters
AirSelect3D takes the load and ΔT as inputs and carries the resulting airflow through casing, coil and fan selection in one pass — face velocity, ADP feasibility and SFP are computed from the same V, not re-derived by hand at each stage. When a consultant asks why the schedule shows a particular airflow for a particular load, the ΔT decision behind it is already documented in the chain.
Design your next AHU in 3D — in five minutes.
AirSelect3D runs certified manufacturer engines (Camfil, Ziehl-Abegg, eBM Papst, Friterm, Hoval) and ships an ErP-compliant Eurovent dossier with every selection.
Launch the 3D Designer →