Ask a plant manager what electricity costs and you will get an answer to the tenth of a cent. Ask what a thousand cubic feet of compressed air costs and the room usually goes quiet. Compressed air is one of the most expensive ways a plant delivers energy, it is often unmetered, and its biggest loss — leaks — is silent under production noise and runs through nights and weekends.
The fix is not a heroic annual survey. It is a measured loop: meter compressor power against demand, find leaks with ultrasound, tag and repair them, verify the repair, and stop pressure set-points from creeping. This is how we approach it.
Why air is expensive per unit of useful work
A compressor turns electricity into pressure, and most of that electricity leaves as heat — in the aftercooler, the oil cooler and the compressor room. The pressurized air that remains then loses more on its way to the point of use: pressure drop across dryers, filters, piping, hoses and regulators. At the end of the chain, air motors and blow-offs convert what is left into work inefficiently. Put a pneumatic grinder beside an electric one doing the same job and the difference in input energy is large.
That chain is why waste is so costly. A leak wastes not just the air but all the electricity spent compressing, drying and distributing it. The US Department of Energy’s Compressed Air Tip Sheet #3 describes leaks as “a significant source of wasted energy in a compressed air system, often wasting as much as 20%-30% of the compressor’s output,” and suggests 5%–10% of total system flow as a typical target for cost-effective leak reduction in industrial facilities.
A worked example: from leak to dollars
Leak flow depends mainly on hole size and pressure; the same tip sheet notes that leakage rates rise with supply pressure and are proportional to the square of the orifice diameter. Turning flow into money takes three assumptions. For this hypothetical plant:
- Specific power: 20 kW of compressor input per 100 cfm delivered. Check your own compressor’s data sheet; well-controlled systems do better, poorly controlled ones worse.
- Hours: 8,760 per year, because the system stays pressurized around the clock.
- Electricity: $0.12 per kWh, blended.
| Leaks found (hypothetical) | Flow each | Total flow | Power | Energy per year | Cost per year |
|---|---|---|---|---|---|
| 40 small: worn couplers, loose fittings | 1 cfm | 40 cfm | 8 kW | 70,080 kWh | $8,410 |
| 15 medium: cracked hoses, bleeding FRLs | 5 cfm | 75 cfm | 15 kW | 131,400 kWh | $15,768 |
| 3 large: open drains, failed fittings | 20 cfm | 60 cfm | 12 kW | 105,120 kWh | $12,614 |
| Total | 175 cfm | 35 kW | 306,600 kWh | $36,792 |
Two lessons. The three large leaks cost more than the 40 small ones combined, so fix in order of flow, not order of discovery. And the savings are only real if the compressors respond: a compressor that keeps running unloaded after the leaks are fixed still draws power. The tip sheet makes the same point — compressor run time has to come down to match the reduced demand.
Where leaks hide
Leaks concentrate where air systems are easy to install and easy to forget:
- Fittings and threaded joints, where sealant dries out and vibration loosens connections.
- FRLs (filter-regulator-lubricator units), at bowl seals, auto-drains, gauge ports and regulator diaphragms that bleed continuously.
- Quick-connects and couplers, with worn O-rings, cheap bodies, or tools left disconnected while the line stays pressurized.
- Hoses, cracked at the crimp, abraded against frames or flattened by forklifts.
- Condensate drains, especially timer drains set to open too long or stuck open, and manual drains cracked open “to keep the water out”. Zero-loss drains that open only when condensate is present remove this one entirely.
- Point-of-use devices, such as cylinders with worn seals and solenoid valves exhausting continuously.
- Dead legs, piping to machines that were removed years ago but never isolated.
Finding leaks with ultrasound
Most leaks are inaudible in a running plant. Air escaping through a small opening creates turbulence with strong ultrasonic content, and a handheld ultrasonic detector — typically listening in a band around 40 kHz — shifts that sound down into the audible range. Because ultrasound is directional and fades quickly with distance, a technician can pinpoint a leak among dozens of fittings with production running. Parabolic attachments reach overhead piping, and acoustic imaging cameras, which overlay sound sources on a video image, speed up surveys of crowded machine areas.
Two cautions. Estimating flow from loudness is approximate — distance, angle and hole shape all change the reading — so use flow estimates to rank leaks, not to invoice them. And survey at normal operating pressure: a Sunday survey with the header turned down finds fewer, quieter leaks than exist on Tuesday.
Tag, repair, verify: the loop that pays
A survey that ends in a spreadsheet saves nothing. The value comes from closing the loop:
- Tag each leak at the location with an ID, the estimated flow and the date.
- Log it in the CMMS with a photo and a priority based on estimated annual cost.
- Fix the easy ones on the spot. Many leaks need a wrench, not a work order.
- Plan the rest. Leaks that need depressurizing go into the next planned shutdown, with parts staged in advance.
- Verify with the detector after repair, remove the tag and close the work order.
- Re-baseline compressor demand once a batch of repairs is complete, and adjust compressor controls to match.
Track open tags as a number. A wall of faded tags six months after the survey tells you the loop is broken, usually at step 4.
Measure compressor kW against demand
Without measurement, savings are a claim. Three signals cover most plants: electrical power on each compressor, delivered flow downstream of the dryers and receivers, and header pressure. Together they give specific power — kW per 100 cfm — and a demand profile you can read like a production schedule.
The most useful view is the non-production baseline. Measure flow or compressor power during hours when nothing should be using air: Sunday night, a holiday, a planned stop. Whatever the compressors deliver then is mostly leaks plus a few legitimate continuous users. In a hypothetical plant that draws 70 kW on a weekday and 22 kW on a Sunday night, almost a third of the weekday load runs whether or not anything is being made. Track that number monthly; when it creeps up, leaks are coming back.
A clamp-on electrical node such as Pulsenode C on each compressor, plus a flow meter on the main header, is enough to start. Keep the metering permanent. A baseline is cheap to collect as you go and expensive to reconstruct after the fact.
Pressure set-point creep
A machine at the far end of the plant starves for air during a busy shift. Someone raises the compressor set-point by 5 psi, the complaint stops, and the set-point never comes back down. Two years and three complaints later, the header runs 15 psi above what most equipment needs.
Higher pressure costs twice. The compressor works harder for every cubic foot it delivers, and every leak and unregulated use passes more air. The DOE tip sheet’s table shows a 1/16-inch orifice passing 5.72 cfm at 90 psig and 6.31 cfm at 100 psig — about 10% more air through the same hole for 10 psi.
The real fix is usually in distribution: a clogged filter, an undersized drop, a long hose or a restrictive quick-connect at the machine that starves. Measure pressure at the point of use, fix the restriction, then bring the set-point down in small steps while watching the far end. Log set-point changes and alert on them, so the next increase is a decision rather than a reflex.
Permanent monitoring or periodic surveys?
| Periodic ultrasound survey | Permanent monitoring | |
|---|---|---|
| What it finds | Every accessible leak on the route, on the day | New leaks in monitored areas as they appear; the system-wide leakage trend |
| Cost profile | Labor for every survey | Sensors and installation, then low running cost |
| Blind spots | Everything that starts the day after | Areas without sensors; pinpointing still needs a handheld |
| Best for | Whole-plant sweeps and pinpointing | Recurring leak zones, proving repairs hold, baseline tracking |
Most sites need both. Meter the compressor room permanently, survey the whole plant on a route, and add permanent ultrasonic sensors such as Pulsenode U where leaks recur — machine banks with dozens of quick-connects, or aisles where forklifts meet hoses. Our distribution center example engagement combines a survey route, permanent leak monitors and compressor metering; in that modeled scenario, compressed-air energy falls by about 25%.
Compressed air will never be cheap. It can be measured, and once it is measured it stops being the utility nobody watches. Our facility and energy monitoring work puts air alongside electricity, HVAC and refrigeration in one view, and the same ultrasound sensing doubles as an early check on bearing lubrication in condition monitoring.