Planning a greenfield compressor station: the decisions that are cheap now and expensive later
A plant move or a new build is the only moment when the compressed air system is designed rather than inherited. Everything that will be awkward for the next fifteen years — the room, the pipe routes, the metering points, the pressure bands — is decided in a few weeks of design review, before anyone has measured the load.
The most expensive mistakes in a station are not equipment mistakes. They are sequencing mistakes: decisions taken in the wrong order, which later force the plant to pay for a workaround.
Decide by pressure band, not by machine count
The first split in a greenfield station is not how many compressors, but how many pressure levels the plant actually needs.
A representative case: a new metal-finishing plant needed 8 bar for general actuation and 5.5 bar for a single workshop whose process could run at a lower pressure. The station was designed with three 8 bar two-stage variable-speed units — two duty, one standby — and a separate low-pressure two-stage variable-speed unit dedicated to that workshop.
Two things make that arrangement worth the extra machine:
- The low-pressure workshop gets air at the pressure it uses. Serving a 5.5 bar process from an 8 bar header means every cubic metre delivered to it carries energy that gets thrown away at a regulator.
- The main station keeps a clean load shape. One outlier department no longer distorts the sequencing of the base machines.
The mistake to avoid is the opposite: building a single pressure level because it is simpler to pipe, then throttling everywhere downstream.
Size the standby explicitly, and say what it is for
“Two duty, one standby” is a statement about reliability, and it should be written down as one. Standby for a machine failure is a different requirement from standby for a demand peak, and the two produce different machine sizes. On a greenfield site, the peak is usually uncertain because the process is not yet running — which argues for keeping the standby unit at the same frame size as the duty units and covering peak by rotation, rather than buying a large unit for a peak that may never arrive.
Meter each department from day one
Every department in that station got its own flow meter. That single decision changes the operation of the plant more than the compressor model does:
- Consumption becomes attributable, so a leak or a misuse in one shop shows up as a number instead of an argument.
- The station can report specific consumption (kWh per m³) per department, which is the only way to track a saving after a retrofit.
- Capacity planning stops being a guess when the plant expands.
Retrofitting meters later means shutting down and re-piping — one of the reasons metering is so often skipped.
Automate the sequence, and keep the reports
The station runs unattended: machines start and stop on demand, the controller maintains pressure, and it produces shift and daily reports visible on a PC or phone, with alarms pushed out rather than displayed in the room.
The value is not the dashboard. It is that an unattended station with logging cannot quietly drift. A pressure set point that has been nudged up by 0.3 bar to stop a nuisance alarm, or a machine that has been left in fixed-speed mode, becomes visible in the specific-consumption trend within days.
Plan heat recovery at civil-works stage
Recovery was included in this station to supply domestic hot water for the plant. On a greenfield build it costs almost nothing to design in — a heat exchanger, a buffer tank and a pipe route. Retrofitting it later means breaking into a finished plant room and finding a place for a tank that was never planned for.
The same logic applies to the room itself: intake and discharge routes, ventilation sized for the actual heat rejection, and space for one more machine than the first phase needs.
The order that works
- Map demand by pressure band and by department, including the process that is not yet running.
- Size the station for that map, with standby stated as a requirement.
- Fix the room, the pipe routes and the metering points before equipment is ordered.
- Select machines for the duty that remains, favouring efficient part-load behaviour where the load will swing.
- Design recovery and future capacity into the building, not into a later project.
Skipping to step 4 is normal practice, and it is why so many three-year-old stations are already being retrofitted.
—
Industry practice note. Facts as reported by industry press; the project described is not a SIND project.
Planning a compressed-air upgrade? Talk to SIND’s engineers.
Facts as reported by industry press; the project described is not a SIND project.
Planning a compressed air project?
Talk to an engineerPublished November 16, 2026.