Transfer Pump Sizing (Ground → Roof)

Unlike a domestic supply pump, where the flow is a peak demand you have to estimate, a transfer pump's flow is a straightforward division. You know the volume to move — typically the working capacity of the roof tank between its fill and draw-down levels — and you choose how long to take over it. The flow is one divided by the other, converted into whatever units the pump curve uses.

The transfer time is the real design lever. Choose it too short and the flow rockets, demanding a large pump and a fat discharge pipe to keep the velocity sane. Choose it too long and the pump may not refill the roof tank fast enough during a busy period, letting the level sag. A good time is one that comfortably outpaces the worst draw-down the roof tank sees, with some margin, but no more — there is no prize for filling the tank in ten minutes if it drains over four hours.

Because the flow is fixed by this choice, it cascades into everything else. The friction losses depend on it, the power depends on it, and the discharge pipe size depends on it. That is why it pays to settle the transfer time deliberately rather than reaching for a round number, and to sanity-check the resulting flow against the roof tank's real usage pattern before sizing the rest.

The head for a transfer pump is usually dominated by the static lift — the vertical distance from the ground tank up to the roof tank, which on a tall building is most of the work. To that you add the friction losses along the riser and through its valves and fittings, and, if the pump discharges into anything pressurised, a head term for that pressure. When the destination is an open roof tank vented to atmosphere, that last term is simply zero, which is the common case.

With flow and head in hand, the power is the familiar chain: hydraulic power ρgQH, then divided by the pump and motor efficiencies to reach the electrical input. For a transfer duty this tends to land at a modest motor, because while the lift can be large, the flow is usually moderate. The number you get is what sizes the starter and the cable, and what you round up to a standard motor frame.

The discharge pipe comes almost for free. Having fixed the flow, you choose a velocity — around 2 m/s is a sensible target for a pumped water line, fast enough to keep the pipe small but slow enough to limit friction and water hammer — and continuity gives the minimum bore directly. Round it up to a standard size and you have a first-pass discharge diameter to carry into the detailed loss calculation. It is an estimate, but a useful one to have alongside the pump duty.

Sizing a transfer pump for flow and head tells you the duty point and the motor, but it says nothing about whether the pump can actually draw water cleanly from the ground tank. That is the job of an NPSH check, and it is deliberately separate. The suction side has its own physics: if the available net positive suction head falls below what the pump requires, the water flashes to vapour at the impeller and the pump cavitates, losing performance and eroding itself.

For many transfer arrangements the suction is benign — the ground tank sits below the pump with a short, generous suction line, giving a flooded suction and plenty of NPSH. But the moment the pump is mounted above the tank, or the suction run is long, restrictive, or carries warm water, the margin can vanish. Those are exactly the situations where a pump sized perfectly for flow and head still fails in service, because nobody checked the suction.

The practical rule is simple: size the pump here for duty, then run an NPSH calculation whenever the suction is anything other than short and flooded. Feed the suction-side friction loss from a head loss calculation into the NPSH check, compare the available figure against the pump's required NPSH from its curve, and keep a margin. Do both, and the transfer pump that looks right on the duty point will also behave on the suction — which is where many otherwise-sound selections come undone.