Dr.Vakuum
全部应用案例

Energy: VSD, central versus point-of-use, and the cost of oversizing a vacuum pump

阅读 8 分钟

A 7.5 kW vacuum pump running 5,000 hours a year at 4.5 THB/kWh costs 170,000 THB a year in electricity — more than it cost to buy, every two years, for fifteen years. Across a plant with twenty pumps, vacuum is often 5–10 % of the electricity bill, and it is the part of the bill that nobody owns: the pumps were bought by the machine suppliers, sized for the worst case, and they run whether or not the process needs them. This note is about where that energy goes and the three decisions that cut it: the right size, variable speed, and the right architecture.

1. Where the energy goes

Every vacuum pump converts shaft power into gas compression and heat. How much of each depends on the technology and on how far from its design point it is running:

Technology Typical specific power at design point Power when the process needs no vacuum (idling at ultimate or closed inlet) Notes
Oil-sealed rotary vane (V, DV) 0.05–0.08 kW per m³/h 60–80 % of full load Power nearly constant across the pressure range; oil shear is a fixed loss
Dry rotary vane (DRV, DRK) 0.05–0.07 kW per m³/h 50–70 % Similar to oil-sealed; vane friction
Dry claw (CL) 0.03–0.05 kW per m³/h 40–60 % Non-contact; low fixed losses; power peaks at mid-range
Dry screw (DSI) 0.04–0.06 kW per m³/h 50–70 % Power rises as inlet pressure rises (gas load); VSD-friendly
Liquid ring (LM, LS) 0.06–0.10 kW per m³/h + seal water pumping and cooling 70–90 % Water churning is a large fixed loss; power peaks around 200–400 mbar
Side channel blower (1SC, 2SC) 0.005–0.01 kW per m³/h at shallow vacuum 80–100 % — power rises toward closed inlet Efficient only in the upper half of the flow range
Roots blower (RB, RBV) Proportional to flow × differential Low at low differential Positive displacement; power tracks the pressure difference

Two lessons jump out. First, positive-displacement pumps draw most of their full-load power when they are doing nothing — a vane pump at its ultimate against a closed valve is 70 % of a vane pump at full load. Second, side channel blowers are the opposite: they draw more power as the inlet closes, and a blower against a blocked filter is at its worst.

Why does an oversized pump cost more than its extra kilowatts?

Because of how it is controlled. A pump sized for 150 % of the duty does not run at 67 % power; it runs at 100 % power 67 % of the time (start-stop on a vacuum switch, wearing contactors and motor insulation), or at 100 % power all the time against a throttle or a bleed valve that throws the surplus away as heat, or — worst — pulls the process deeper than it needs, which on a drying or packaging duty costs ultimate-pressure-related problems as well as power. The extra capacity also brings a larger motor with a poorer power factor at part load, a larger inlet filter, more oil. The honest sizing margin is 25 % for a known duty and 50 % for an estimated one, as set out in How to size a vacuum pump; beyond that, the margin should be a variable-speed drive, which delivers capacity when needed and is not paid for when it is not.

2. What a VSD saves, by pump type

A variable-speed drive slows the pump to match the process's demand, holding the vacuum level at the set point rather than cycling or throttling. The saving depends on how the pump's power varies with speed:

  • Dry screw and claw pumps: power is roughly proportional to speed at a given pressure, and the minimum speed is 30–40 % — a plant whose average demand is 50 % of peak saves 40–50 % of the energy. These are the pumps VSDs were made for, and the DSI Series is designed for speed control as standard.
  • Roots blowers: power is proportional to speed × differential; speed range 30–100 %; the saving on a varying-flow duty (conveying, aeration, central systems) is 30–60 %.
  • Oil-sealed vane pumps: speed range is limited (oil pumping and vane sealing need a minimum speed, typically 50–60 %), and the fixed oil losses do not fall with speed as fast; saving 15–30 %. Often the better answer is a smaller pump with a receiver.
  • Liquid ring pumps: speed range 50–100 %; the ring's water dynamics limit it; saving 20–35 %, plus the seal water pumping falls too.
  • Side channel blowers: VSD works well for flow control at constant differential (conveying), poorly for vacuum-level control (the blower's maximum differential falls with speed squared) — saving 20–40 % on the right duty, nothing on the wrong one.

The control: a pressure transmitter on the receiver or the header, a PID loop holding set point, a minimum-speed floor, and a stop-after- delay when the demand is zero for more than a few minutes (the "sleep" that saves the idling power that no VSD can).

3. Central or point-of-use

A plant with many vacuum consumers can give each its own pump (the machine builder's default) or run a central system — a few large pumps in a plant room, a receiver, a header, a valve at each machine. The energy case for central rests on diversity: the consumers are not all at peak together, so the central capacity is 50–70 % of the sum of the individual pumps, and it runs on a VSD at the actual demand instead of twenty pumps idling at 70 % power. The cases against it are real and must be checked:

Favours central Favours point-of-use
Many similar consumers at one vacuum level (moulding hall, packaging hall, print hall, labs) Few consumers, or very different vacuum levels (a 5 mbar packer and a 300 mbar blower duty do not share a header)
Steady or diverse demand — the VSD has something to follow Short sharp cycles that a header cannot deliver without a huge receiver (chamber packers, die casting) — though a receiver per machine on a central pump often solves it
Dust, heat and noise better in a plant room Processes that contaminate the gas differently (one machine's solvent in everyone's pump)
Spares, maintenance and monitoring in one place A single failure must not stop the whole plant (so central = duplex, always)
Pipe runs short enough that pressure drop < 30–50 mbar Long runs at deep vacuum — conductance kills a 5 mbar header

The architecture that usually wins in a mixed plant: a central system for the dominant shallow-to-medium duty (moulding, handling, packaging-line conveyors, cleaning) and point-of-use pumps for the deep, dirty or cycle-critical machines.

4. A worked hall at Thai tariffs

An injection-moulding hall with 16 machines, each fitted by the machine builder with a 100 m³/h oil-sealed vane pump (2.2–3 kW) for mould venting and hopper loading, running 6,000 h/year. Measured average demand on the header after a trial: 40 % of the sum of nominal capacities, peaks at 70 %.

Option Installed Energy per year Cost at 4.5 THB/kWh Notes
A. As built: 16 × vane, start/stop 16 × 2.6 kW = 42 kW ~42 kW × 0.75 × 6,000 = 189 MWh 850,000 THB Pumps idle at 70 % power; 16 oil changes
B. Central, 2 × fixed-speed vane (1 duty, 1 standby) 2 × V630 (15 kW) 15 kW × 0.85 × 6,000 = 77 MWh 345,000 THB Oil mist out of the hall; standby included
C. Central, 2 × dry screw on VSD 2 × DSI630 (22 kW) 22 kW × 0.42 × 6,000 = 55 MWh 250,000 THB Oil-free, follows demand; the standby is on VSD too
D. Central, MB-boosted vane for the deep venting + 1SC blowers for loaders mixed ~60 MWh 270,000 THB Splits the duties by depth; more pipework

Option C saves 600,000 THB a year against the as-built plant, removes sixteen oil-sealed pumps from the hall, and pays back its installed cost in about three years. The receiver and header cost less than the first year's saving. That is the typical shape of the result; the central vacuum for an injection moulding facility note works a similar plant in detail.

5. Ten cheap kilowatts

Before any capital project, the things an electrician and a fitter can do in a week:

  1. Stop idling pumps. A timer or a "no demand" stop on every pump that runs through breaks and shift changes.
  2. Raise the set point to what the process needs. A packer set at 2 mbar that seals fine at 5 saves the slowest, most expensive decade.
  3. Open the pipe. A hose one size up is a free 20–30 % on any short-cycle duty — see How to size a vacuum pump.
  4. Change the inlet filter on differential pressure, not calendar; a loaded filter is a permanent throttle.
  5. Fix the leaks. A leak is a gas load the pump compresses for nothing; on a central system, a pressure-rise test on the header with the machines isolated, quarterly.
  6. Cool the seal water on liquid ring pumps — every 5 °C is 8–10 % of capacity, which is a smaller pump or a slower one.
  7. Match blowers to the flat part of the curve; a side channel blower at 80 % of its maximum differential is at its least efficient.
  8. Separate deep and shallow duties onto their own pumps; the deep pump is not pulling the shallow duty to a depth it does not need.
  9. Receivers on cyclic duties so the pump runs steadily instead of in bursts.
  10. Measure. A current clamp and a vacuum gauge on each pump for a week tells you more than any catalogue. We lend them.

Recommended Dr.Vakuum machines for energy

Dry screw (DSI Series) — central systems on VSD; the widest speed range and the best part-load behaviour: DSI300, DSI630, DSI1200.

Dry claw (CL Series) — non-contact, low fixed losses, for medium vacuum central systems: CL150, CL300.

Roots blowers (RB / RBV Series) — VSD-controlled conveying, aeration and shallow central systems: RBV100, RB100.

Combined vacuum-pressure dry vane (DRK Series) — one motor for two duties on presses and converting: DRK140.

Energy survey. Tell us how many pumps you have and their hours, or let us put meters on them for a week, and we will return a measured demand profile and the central/VSD option with its payback at your tariff. Talk to an engineer.