Compressor controls, sequencing and multi-compressor optimisation

Why several compressors on one header still waste energy

Most plants grow one compressor at a time. A new machine is added when production expands, the old one is kept “for standby”, and both are left on their factory pressure settings. The result is a compressor house where each machine reacts to its own discharge pressure sensor, nobody has decided which machine should lead, and two or three units spend long periods loaded at part capacity or running unloaded.

Typical symptoms are easy to spot once you look for them:

• Two fixed-speed machines both cycling between load and unload at the same time.
• A variable-speed drive (VSD) machine sitting at full speed while a fixed-speed machine runs unloaded beside it.
• Header pressure set high “to be safe” because pressure dips when a second machine is slow to load.
• Running hours that pile up on one machine while another is hardly used, then both need overhaul together.
• A standby compressor that fails to start on the day it is needed because it has not run or been tested for months.

An unloaded oil-injected screw compressor still draws a substantial share of its full-load power while delivering no air, and every extra bar of generated pressure costs about 7% more power. Sequencing attacks both losses at once: it removes unnecessary unloaded running and it lets the plant run at a lower, steadier pressure.

How compressor control modes work

Before machines can be sequenced, each one’s individual control behaviour must be understood. The common modes are summarised below.

A VSD machine is efficient only within its speed range. Below its minimum speed it begins to load and unload like a fixed-speed machine, and near maximum speed its specific power may be no better than a fixed-speed unit. Those limits are product-specific and should be read from the manufacturer’s data for the actual model, not assumed.

Base, trim and standby: assigning the roles

Every well-run multi-compressor plant gives each machine a clear job.

• Base load: these machines run fully loaded for long periods and carry the sustained part of demand. They should be the machines with the best full-load specific power, measured in kW per m³/min of free air delivered.
• Trim: the trim machine follows the variation in demand. A VSD screw is the usual choice, sized so that the normal swing in demand sits inside its efficient speed range.
• Standby: these machines cover a breakdown, maintenance or an unusual peak. They must be ready, not merely present.

The largest compressor is not automatically the base machine. If demand at night falls below the output of the largest unit, that unit will spend the night loading and unloading. The right choice comes from the demand profile, the corrected capacity of each machine, its specific power, its condition and its maintenance schedule.

A common mistake is sizing the trim machine too small or too large. If the gap between base machine steps is larger than the trim machine’s useful range, there will always be a band of demand where the plant hunts, starting and stopping machines. The rule of thumb used by most system designers is that the trim machine’s controllable range should be at least as large as the largest step change produced by bringing a base machine on or off.

Cascade settings versus a central sequencer

There are two ways to make several compressors share demand.

Cascade (staggered pressure bands)

Each compressor is given its own load and unload pressure, staggered so the lead machine loads first and the others follow as pressure falls. It needs no extra hardware, but every extra machine in the cascade pushes the top of the band higher. With three or four machines the plant can easily run 1 bar or more above what is actually needed, which is roughly 7% extra power for that bar. Cascades also drift as pressure switches and transducers age.

Central sequencer (master controller)

One controller reads a single pressure signal, usually on the dry-air header, and decides which machines load, unload, start, stop or change speed. All machines work within one narrow band. A good sequencer also rotates duty on hours, respects minimum run times and starts per hour, holds standby machines ready, and logs what happened. Most major manufacturers offer their own system controllers, and vendor-neutral controllers exist for mixed-brand fleets; compatibility with older machines depends on what each machine’s controller can accept, from a simple remote load/unload contact up to a full digital interface.

Where to measure the controlling pressure

The pressure that matters is the pressure your process receives, not the pressure at each compressor outlet. Discharge pressure at a compressor includes the drop across its own aftercooler, separator, the dryer and filters. If each machine controls on its own discharge sensor, a dirty filter on one line will make that machine behave differently from the rest.

For a central sequencer, the usual practice is to measure on the common header downstream of the dryers and filters. In critical plants, a second validating transducer is worthwhile, because a drifting or failed sensor can make the controller hold the wrong pressure for weeks without anyone noticing. Where a remote process area is the true limiting user, some plants add a remote sensing point, but this needs a defined fallback if the signal is lost.

Pressure band, storage and response time

A narrow pressure band reduces average pressure but gives the system less time to react. Storage buys that time. When demand rises suddenly, the receiver and pipework supply the shortfall while the next machine starts and loads. The required volume depends on the size of the step, the acceptable pressure fall and how quickly the machine can respond; the usable stored air is proportional to receiver volume multiplied by the allowed pressure drop. Our receivers and storage page explains the calculation.

Practical points that a sequencing design must handle:

• Staging delay. If the controller waits too long before calling another machine, pressure dips; too short and it starts machines it does not need.
• Simultaneous loading. Two machines loading at once can overshoot and then both unload, producing a saw-tooth pattern.
• Start limits. Motors have a maximum number of starts per hour; the controller must respect them.
• Unload-to-stop timers. A machine left running unloaded “just in case” is a direct energy loss; a machine stopped too soon may restart minutes later.
• Dryer purge and drains. Heatless desiccant purge and timer drains are real demand and must be included in the load the fleet sees.

Setting up sequencing: a practical method

1. List every machine. Record make, model, technology, rated and corrected free air delivery, pressure range, control type, unloaded power if known, controller type and condition.
2. Measure the demand. Log header pressure, compressor power and states for at least a full production week, including a weekend or holiday. See our demand assessment and load profiling page.
3. Find the minimum required pressure. Identify the user that genuinely needs the highest pressure and the losses between it and the header.
4. Remove waste first. Fix the major leaks and stop artificial demand; sequencing an oversized load only optimises waste.
5. Assign roles. Choose base, trim and standby machines from the profile and the specific power of each unit.
6. Choose the control method. Decide between re-tuned cascade bands and a central sequencer, considering the brands and controller interfaces available.
7. Commission and test. Test demand steps, a machine trip, sensor loss and controller loss, and confirm the fallback is safe.
8. Verify. Repeat the logging under similar production conditions and compare kWh per m³ of air, unloaded hours and pressure stability.

Worked example: two machines, one smart change

Consider a plant with two identical 22 kW fixed-speed screw compressors and a typical day-shift demand that one machine could almost carry. Both are set with overlapping bands, so both load and unload through the day and each runs unloaded for long periods. Header pressure swings between 6.5 and 7.5 bar (about 94–109 psig).

A sequencing review would make three changes: one machine becomes the lead and is set to carry the base load, the other is set to start only when pressure falls below a defined level and to stop after a short unloaded run, and the whole band is narrowed and lowered after confirming the highest real user needs only 6 bar (87 psig) at the point of use. The benefit comes from fewer unloaded hours and lower average pressure, which can be measured directly from logged kWh before and after. We do not predict a percentage in advance; we measure it.

If demand varies widely, the next step is often to replace or supplement one fixed-speed unit with a VSD trim machine sized to the swing, which is covered in our base, trim and VSD optimisation page.

Mixed fleets, failures and standby readiness

Real compressor houses often contain several brands, a piston unit or two, and machines of very different ages. Each brings its own limits:

• Older machines may accept only a remote load/unload or start/stop signal, so the sequencer cannot read their state digitally.
• Piston compressors deliver capacity in steps and suit a base role only where those steps match demand.
• Two VSD machines trimming together can fight each other unless the controller gives one clear priority.
• Oil-free and oil-injected machines feeding the same header must be checked against the air-quality requirement before being treated as interchangeable standby.

Plan for failure as carefully as for normal running. Decide what happens if the master controller fails, if communication to a machine is lost, or if the header sensor fails. The safe default is usually for each compressor to revert to its own local pressure control at a band slightly below the sequencer band, so the plant keeps air. Independent high-pressure protection, safety valves and emergency stops must never depend on the sequencer.

Standby must be tested. “N+1” means the plant can meet its peak demand with the largest machine out of service, and that only holds if the standby machine starts, loads and holds pressure when asked. Rotate it into service on a schedule and record the result.

What we check on a sequencing review

• Nameplate and controller data of each compressor and dryer.
• Current pressure settings, timers and control modes on every machine.
• Location and condition of pressure sensors and gauges.
• Logged header pressure, power and machine states over a representative period.
• Pressure drop across dryers, filters and main headers.
• Receiver volume, location and drain condition.
• Leakage indicators such as non-production demand.
• Running hours, service history and condition of each machine.
• Fallback behaviour on sensor, communication or controller failure.

How Reddy Compressors helps

Reddy Compressors has worked with industrial compressed air since 2002 from our office at Tarnaka, Secunderabad. We are an independent multi-brand specialist, so we look at a mixed fleet of ELGi, Atlas Copco, Ingersoll Rand, KAESER, Chicago Pneumatic, Sullair, Doosan, Kirloskar and other machines without favouring a brand. We are not an authorised dealer of any of them. We have tie-ups with all major and reputed brands.

For sequencing and controls we can review settings, log pressure and machine states, re-tune cascade bands, advise on a central controller suited to the machines you have, and help commission and test it. Where the right answer is a different machine, we supply new and used compressors, including VSD units for trim duty. Our team of more than 20 service engineers and technicians covers the whole of Telangana and Andhra Pradesh, and equipment, rentals and spares are supplied across India. If a machine must come out of service while controls are changed, a rental compressor from our own fleet can cover the gap.

Related work is handled through our air audit and measurement services, AMC and rotary screw service.

What to send us

• Make, model and nameplate photo of every compressor and dryer.
• Current load/unload or setpoint pressures on each machine.
• The minimum pressure your most demanding process actually needs.
• A simple sketch of the compressor room, receivers and main header.
• Shift pattern and any known peaks, such as start-up or batch processes.
• Any logged data or electricity readings you already have.

Frequently asked questions

What is compressor sequencing?

Compressor sequencing is the control of several compressors from one pressure signal so that each has a defined role. Base machines run fully loaded, a trim machine follows changes in demand, and standby machines start only when needed. A sequencer or carefully set cascade bands decides which machines load, unload, start or stop, keeping header pressure stable while reducing unloaded running and wear.

Should the VSD compressor be base load or trim?

In most plants the VSD compressor should be the trim machine, because its efficiency advantage comes from following varying demand within its speed range. Running a VSD at full speed as base load wastes that advantage, and its full-load specific power may be no better than a fixed-speed machine. The exception is a plant whose entire demand fits within one VSD’s efficient range.

Do I need a master controller for two compressors?

Not always. For two machines, correctly staggered pressure bands with a clear lead machine, sensible unload-to-stop timers and regular changeover can work well. A master controller becomes more valuable with three or more machines, mixed sizes, a VSD trim unit, or where you need automatic duty rotation, standby testing and recorded data on machine states and pressure.

Can compressors of different brands be sequenced together?

Yes, in most cases. Vendor-neutral controllers can manage mixed brands, and many manufacturer controllers can also accept other machines through simple load/unload or start/stop signals. The limit is what each compressor’s own controller can accept. Older machines may need a small interface or relay arrangement, and their status may not be visible digitally.

Why does my plant pressure drop when the second compressor starts?

Usually because the second machine is called too late or takes time to start and load, and there is not enough storage to cover the gap. Other causes include a controlling sensor in the wrong place, high pressure drop across dryers or filters, or a cascade band set with too little margin. Logging pressure and machine states shows which cause applies.

How much energy can sequencing save?

It depends entirely on how the plant runs today. The savings come from fewer unloaded hours, fewer machines running at part load and a lower average pressure, where each bar removed reduces compressor power by about 7%. The honest approach is to log power and pressure before the change, make the change, then log again under similar production and compare kWh per m³ of air.

Related pages

• Base, trim and VSD energy optimisation
• Reducing unload and idle losses
• Compressed-air receivers and storage
• Pressure management at the point of use
• Demand assessment and load profiling
• Power and flow logging