Dr.Vakuum
All applications

Sizing a central vacuum system

7 min read

The strongest argument for a central vacuum system is not that it is tidier. It is that a central plant is the only architecture that can exploit diversity — and diversity is usually a factor of three to six.

1. Diversity is the number that decides the plant size

Drops on a factory header are not all open at once. The naive sum of every connected drop overstates real peak demand badly.

The cleanest published treatment does not use a flat percentage. It uses a marginal form: the first drop counts at full flow, each subsequent one at a fraction 1/k:

Q = Q₁ + (n − 1) · Q₁ / k

with k chosen by how the drops are actually used. That is better engineering than a guessed coefficient, because the resulting diversity factor emerges from n rather than being asserted.

For a 40 L/min drop:

Drops k = 2 k = 4 k = 8
5 120 L/min 80 60
10 220 130 85
20 420 230 135
Naive sum at n = 20 800 800 800

Twenty intermittent drops at k = 8 size to 135 L/min, not 800 — a diversity factor of 0.17 and a 5.9× oversize avoided.

The critical caveat: diversity is a behavioural assumption, not a property of the pipe. Where every station genuinely can be in use at once — continuous-duty process equipment rather than intermittent handling — the correct k is 1 and no diversity applies. Getting this wrong in the optimistic direction is the one error that leaves a plant permanently short.

For handling and cleaning drops, the practical trade method is to ask how many simultaneous operators there really are, then size on that. Interviewing production about shift patterns beats any published coefficient.

2. Pressure drop hurts far more in vacuum than in compressed air

This is the part most often underestimated, and it has three separate causes.

There is no headroom. A compressed-air system at 7 bar feeding 6 bar tools can absorb a 0.3 bar line loss — about 4% of absolute pressure — and you can simply raise the compressor setpoint. In vacuum, the entire available differential is bounded by atmospheric pressure, and you cannot set the source below zero. If the process needs 100 mbar absolute at the drop, a 30 mbar line loss is 30% of your working differential, not 4%.

The gas is rarefied. In ΔP = f·(L/D)·(ρV²/2), vacuum gas at 100 mbar has a tenth of atmospheric density, so the same mass flow needs roughly ten times the volumetric flow and ten times the velocity in the same pipe. Loss scales roughly inversely with absolute pressure: halve the operating pressure and the same duty roughly doubles the line loss.

It compounds. Higher loss means the pump sees a lower inlet pressure than the process does, so the gas at the pump is less dense still, and the pump's volumetric rating buys less mass flow.

The practical corollary: oversizing vacuum pipe is cheap insurance and undersizing is nearly unfixable. You cannot buy your way out later with a bigger pump, because the loss grows with the flow the bigger pump draws.

Useful figures for sizing:

  • Keep the pump inlet size for the first 100 ft, then step up one increment per additional 100 ft. On the exhaust, step up every 30 ft — exhaust restriction directly raises compression work.
  • Convert every fitting to equivalent length. A short-radius 4″ elbow is worth 100 inches of straight pipe and a long-radius one 50 — so a dozen sloppy elbows can silently double an apparent pipe run. Branch tees cost about 3× a through tee, and every drop on a header is a branch tee.
  • Budget 5% total loss as a design target.

Capacity falls hard with distance. Published tables for a fixed allowable loss show the same 28 mm line carrying 350 L/min at 8 m but only 42 L/min at 305 m — an 8.3× collapse. Capacity scales as √(Δp) and as 1/√(L).

3. Sizing the receiver — and why vacuum is not compressed air

A receiver stores far less usable energy under vacuum than the same vessel under pressure. A 1 m³ air receiver at 7 bar holds about 8 m³ of atmospheric-equivalent air. A 1 m³ vacuum receiver at 100 mbar holds a deficit of only 0.9 m³ — and it shrinks as you go deeper.

You cannot buffer your way out of an undersized vacuum system. Size the receiver to prevent short-cycling, not to supply demand. A defensible rule from the medical-gas standards is a volume equal to one minute of design flow.

With a variable-speed lead pump the receiver requirement drops sharply, because the drive absorbs demand swings by modulating instead of cycling. Receiver volume and VFD control are substitutes, not complements.

4. Control

Lead/lag with the variable-speed machine as lead is the standard arrangement: the lead modulates, lag units stage in as it reaches capacity, and runtime alternates to balance wear.

Anti-short-cycling settings that matter:

  • Hysteresis of about 10% between stage-on and stage-off
  • Minimum run and minimum off times of 60–300 s
  • Restart delay of 30–300 s

One trap: the VFD's turndown range must overlap the capacity step of one fixed-speed machine. If it does not, the system hunts between stages. Aim for overlap of at least 5% of a machine's flow so the operating envelope has no gaps — a plant with a coverage gap hunts, cycles and blows off, and none of that appears on a datasheet.

5. Why central beats point-of-use

Point-of-use pumps run at full power whenever they run, regardless of demand. A central plant with variable-speed control generates only what is needed at that moment — this is diversity monetised, and it is the only architecture that can capture it.

A documented consolidation replaced 30 machine-mounted pumps with 2 central units and cut energy by 75%. Note the 30 → 2 ratio is itself a diversity factor made visible.

The other gains are real but secondary:

  • Heat leaves the production hall instead of loading the air conditioning. Recovery of 50–70% of motor energy is achievable.
  • Noise moves to a plant room.
  • Maintenance happens in one place, and can be done while the system runs.
  • Redundancy gets cheaper. With 30 point-of-use pumps, full cover needs 30 spares. With a central bank sized on diversified demand, n+1 covers the whole factory — often four or five machines where thirty stood.

6. Injection moulding: do not put every duty on one header

Three quite different duties exist in a moulding plant:

  • Mould venting — the deepest, around 100 mbar absolute, extracting cavity air before and during injection. Published work reports active vacuum venting delivering 9–17% shorter cycle times and 8–12% lower energy, which is the strongest economic case on this page.
  • Material conveying — high flow, shallow vacuum. The level only has to sustain conveying velocity, and that velocity is a floor set by saltation, not something diversity can reduce.
  • Handling and clamping — shallow, flow-light once sealed, but needing high flow to achieve seal.

These should generally not share a header. A deep-vacuum venting duty and a high-flow shallow conveying duty have incompatible optimum operating points, and combining them forces the whole system to the deepest level at the highest flow — the worst of both.

7. Common mistakes

  1. Summing all drops with no diversity. The most expensive error, and usually a 3–6× oversize.
  2. Applying diversity where it does not belong — continuous-duty equipment is k = 1.
  3. Undersized pipe and ignored fittings, which cannot be corrected later by a larger pump.
  4. Confusing actual and standard flow. Pumps are rated at inlet conditions; a machine "rated 500 m³/h" delivers a fraction of that at the working setpoint.
  5. Mixing incompatible duties on one header.
  6. Sizing the receiver as if it were an air receiver.
  7. No growth allowance. Because loss scales inversely with pressure, adding drops to a header sized exactly for today degrades the whole system, not just the new branch.

Recommended Dr.Vakuum machines

Dry claw pumps (CL Series) suit central systems in the 150–400 mbar band — handling, clamping and conveying — running oil-free and continuously, with roughly 30% better efficiency than comparable rotary vane machines at these levels.

Oil-sealed rotary vane (V and DV Series) remain the more efficient choice at deeper levels, including mould venting near 100 mbar absolute.

Dry screw pumps (DSI Series) where oil-free operation is required across the whole range:

Model Flow (50/60 Hz)
DSI300 300 / 360 m³/h
DSI630 630 / 756 m³/h
DSI1200 1,200 / 1,440 m³/h

Specify n+1 modular rather than one large machine: it captures diversity, gives redundancy at the cost of one small module, and keeps every unit near its efficient operating point.

Sizing help. Send us the number and type of drops, how many genuinely run at once, the duty cycle of each, and the header geometry, and our engineers will work the diversified demand and the line losses through with you. Talk to an engineer.

See also: the lifecycle-cost case for central versus point-of-use, with VSD savings by pump type, in Energy: VSD, central versus point-of-use; and the harder vacuum duty on the compounding and recycling extruders in Extruder vent degassing and PET recycling.