Flow alone is meaningless: how to actually specify a fan duty

A buyer who says “I need a 50,000 CMH fan” has not specified a fan — only half of one. A centrifugal fan delivers a given flow only against a given resistance: the operating point is the intersection of the fan curve and your system’s resistance curve. State one without the other and you have described a wish, not a duty. The first question our estimating desk asks any flow-only enquiry is always the same — at what static pressure? — followed immediately by at what temperature and altitude?, because both change the air itself.

The eight parameters that define a duty

A complete duty is fixed by eight physical parameters. Miss one and your vendor either assumes it in their own favour or emails you a week of clarifications.

Those eight fix the physics. A quotation-ready datasheet then adds the commercial layer — margins, installation arrangement, drive (direct / belt / VFD), materials, accessories, test standard and noise limit — but the fan cannot be selected until the eight above are on the page.

The actual-vs-standard flow trap

Catalogue and test performance is quoted at standard air (1.2 kg/m³, 20 °C, sea level). Process engineers, however, routinely state the mass-equivalent standard flow — Nm³/h or SCMH. The fan doesn’t move mass; it moves the actual volume at operating conditions. Confuse the two and you undersize by the temperature ratio:

• Enquiry: “10,000 Nm³/h of flue gas at 200 °C.”
• Actual volume = 10,000 × (273 + 200) / 273 = 17,326 ACMH.
• Quote a fan for 10,000 CMH and you have undersized it by ~42%. This is exactly how the notorious “up to 2×” sizing error happens.

So declare the basis on every line of the datasheet. It is the single most common way a duty goes wrong before a wheel is ever chosen.

Why a hot-gas fan is sized for the cold start

Density corrects pressure and power in lock-step: volume flow stays put (a fan is a volume machine), but static pressure and shaft power both scale with density — P₂ = P₁ × (ρ₂/ρ₁). Take a system needing 450 mmWC of air-like gas at 200 °C. Density there is 1.2 × 293/473 = 0.743 kg/m³, so the fan curve at standard density must show 450 × (1.2/0.743) = 727 mmWC to deliver 450 mmWC hot.

Now start that fan cold. At 20 °C the gas is denser, and the same wheel absorbs ~1.61× the hot-duty shaft power. Size the motor for the hot duty alone and it trips — or cooks — on the first cold start. This is the trap on almost every induced-draft (ID) fan on a kiln / furnace exhaust: rated hot, started cold. The fix is one of two: size the motor for the cold-start power, or specify a damper-closed starting procedure. Forced-draft (FD) fans handling near-ambient air escape it; hot ID fans never do. We size this in explicitly — see hot-start furnace-fan design.

Say which pressure you mean

Two vendors quote “100 mmWG”; one means fan static pressure, the other fan total (static + velocity). Those are two different fans, not two prices for the same one. Industrial practice rates fans in static pressure — but a number only means something when it says which. State the required static pressure at the fan outlet, and say it is static. If your consultant’s schedule gives total pressure, pass it on as total pressure — don’t relabel it. The full breakdown is in static, velocity & total pressure.

One more thing the datasheet rarely mentions: system effect. An elbow or a drop-box hard against the fan inlet robs pressure the fan never sees — a badly-fed drop-box inlet can cost up to 45% of flow. State the duct layout near the fan honestly so the allowance goes in before the frame is chosen, not after commissioning.

The enquiry checklist — print this

• Flow — value, unit, and actual or standard basis?
• Static pressure — at that flow, in mmWC, at operating conditions?
• Temperature — operating, maximum, cold-start?
• Altitude / site location?
• Gas — composition, density, humid or corrosive?
• Dust — loading (g/Nm³), type, abrasive?
• Duct layout near the fan — straight run or elbow / drop-box?
• Drive, electrical supply, starting method, VFD?
• Materials, accessories & options — dampers, IGV, silencer, bellows, instrumentation?
• Test standard, certification, noise limit?

Every duty we receive is sanity-checked against our proven served range — up to about 2,25,000 CMH, 2,000 mmWC, 425 HP and 500 °C — and then put on the actual fan curve with our selection software. Note those are independent corners of the envelope, not a combinable box: maximum flow pairs only with the pressure that fits the power cap, and the highest static isn’t available at the highest temperature. Every quoted duty gets a fresh selection, no exceptions. And on test wording: we perform to the IS 4894 / ISO 5801 / AMCA 210 method, and the test record names the setup used — no fan, from anyone, carries an AMCA product stamp, so treat any AMCA product-certification claim on a rival quote as a red flag, not a credential.

Send us the eight parameters — in any format — and you get a committed selection back: family, size, duty point on the curve, absorbed power and the wheel rationale in writing. For the classic ways a good spec still goes wrong at site, read seven mistakes buying an industrial fan.

Jitamitra Electro Engineering · Fan-engineering notes, written for the engineer.

Why is the motor bigger than my calculation?

You worked out the absorbed power at the duty point, added a margin, and got a smaller motor than the one we quoted. The calculation is not wrong — it is answering only half the question.

A fan impeller is a high-inertia load. A large wheel stores a great deal of rotational energy, and the motor has to put that energy in during start-up while the fan is already taking air. Two things follow:

• Starting time is set by inertia, not by duty. It depends on the GD² (WK²) of the whole rotating assembly, the motor's accelerating torque, and the counter-torque the fan itself develops as it speeds up — which rises with the square of speed.
• The motor must survive that time. Every motor has a permissible starting time; run beyond it and the rotor overheats. That figure comes from your motor maker, and it is what turns a torque calculation into a motor size.

So the honest answer to "can we use one frame smaller" is: sometimes yes, and it is an engineering question rather than a commercial one. It can be settled by reducing wheel inertia — a lighter impeller of the same duty — and proving the starting time on the rebuilt wheel. That is real work with a real test behind it, and we would rather do it than argue about a nameplate.

The starting method matters here too: star-delta gives roughly one third of the direct-on-line accelerating torque, so the same wheel takes substantially longer to reach speed on star-delta than on DOL. Tell us the starting method with the duty — it can change the motor.