Key takeaways
Walk most plants and you can hear the leaks, the hiss at a fitting, the whistle at a quick-disconnect, the sigh from a worn cylinder seal. Everyone hears them. Almost nobody fixes them. The reason is an attribution gap.
The compressor runs harder to maintain system pressure against the leaks. That extra running shows up as electricity cost on the compressor's meter. But nobody connects the electricity cost to the specific leaks causing it.
The maintenance team that could fix the leaks does not pay the electricity bill; the energy team that pays the bill does not fix leaks. The leak sits in the gap, audible and unaddressed, for years.
The math is stark. A single small leak, the kind you can hear but barely feel, can cost a meaningful amount of electricity per year, running 24/7. A plant with dozens of these is losing real money continuously.
Yet because no single leak is dramatic and the cost is never attributed, they never make the priority list. For choosing which maintenance metrics to track, see our guide to maintenance KPIs.
Compressed air is expensive because the generation is inefficient. A large fraction of the electrical energy that goes into a compressor is lost as heat; only a portion ends up as usable compressed air. So every unit of compressed air leaked represents several units of electrical energy wasted upstream. This is why a leak that seems small in air terms is large in money terms.
Industry studies consistently put compressed-air leak losses at 20-30% of total compressed-air output in plants without an active leak program. For a plant running multiple compressors continuously, that 20-30% is a substantial annual electricity cost, usually the single largest avoidable energy waste in the plant, larger than lighting, larger than most process inefficiencies. The article on production loss analysis covers how this connects to the broader loss picture.
Compressed-air leaks emit ultrasonic sound, frequencies above human hearing, even when too small to hear directly. An ultrasonic leak detector turns these into an audible signal and pinpoints the leak location. A technician with a detector can survey a plant's compressed-air system in a day or two, finding leaks that are inaudible to the naked ear.
The survey produces a list: every leak, its location, an estimated severity. The estimate matters because it lets the program prioritise the biggest leaks first.
Each leak found gets a physical tag at the location and a work order in the CMMS. Tagging is what prevents the leak from being forgotten between the survey and the repair. The work order carries the location, the severity estimate, and the estimated annual cost, so the repair has a visible value attached. The piece on work order management systems covers how these survey-generated tickets get prioritised.
Safety first: isolate the section and vent it to zero pressure before any repair, never tighten or unscrew a fitting that is still under pressure, and block any machine motion that a loss of air could cause.
Most compressed-air leak repairs are trivial, tighten a fitting, replace a quick-disconnect, swap a worn seal, replace a length of degraded hose. They do not need a specialist or a long window; they slot into the opportunistic maintenance tier. The barrier was never the repair difficulty; it was the leak never making it onto a list.
After the repair round, the compressor's energy consumption is measured against the pre-repair baseline. The reduction is the program's value, attributed back to the maintenance team that did the work.
This attribution is what keeps the program funded, without it, the energy saving is invisible and the program competes for maintenance time against more visible work. The piece on the preventive maintenance schedule covers how the quarterly survey becomes a standing program rather than a one-off.
Three reasons compressed-air leak repair outperforms most maintenance investments:
The combination, continuous saving, trivial repair cost, cheap detection, makes leak repair pay back faster than almost any other maintenance activity. The only reason it is not universal is the attribution gap that keeps the leaks off the priority list. The article on root cause analysis covers how recurring leaks at the same fittings point to upstream causes worth fixing permanently.
Leaks recur. New ones appear as fittings vibrate loose, seals age, hoses degrade. A one-time survey produces a one-time saving that erodes over the following year as new leaks develop. The program has to be standing:
Plants that run the quarterly cycle keep their leak loss in the single digits (DOE puts a well-run program at under 10% of output); plants that do a one-time survey and stop drift back toward the 20-30% baseline over the following year or two. The recurring discipline is what makes the saving durable.
The leak program works in any plant with an ultrasonic detector and a CMMS.
Where a unified MES and OEE platform with CMMS built in helps is in keeping the evidence together: the leak work orders and the compressor readings sit in one place, so the before and after picture is visible rather than a claim the maintenance team has to argue for.
Fabrico keeps every leak work order in its CMMS and can show compressor readings from PLC or IoT sensor connections on a real-time dashboard. To see what a leak-program tracking view looks like, book a demo .
What one leak costs. The US Department of Energy (DOE) leak table gives about 25 cfm through a 1/8 inch (3.2 mm) hole at 100 psig (about 7 bar). For a sharp-edged hole DOE applies a factor of 0.61, which gives about 15 cfm. At DOE's rule of thumb of about 18 kW of compressor power per 100 cfm, that is 2.8 kW, or roughly 24,300 kWh over a year of round-the-clock running (8,760 hours). At an electricity price of 0.15 per kWh, in dollars or euros, that single 1/8 inch leak costs about 3,600 a year. On the same basis a 1/16 inch leak costs about 900 a year and a 1/4 inch leak about 14,600.
Make the compressors notice. Once leaks are repaired, have the compressor controls re-evaluated so that run time actually falls to match the lower demand. DOE's guidance is that, to capture the energy saving, compressor run time must be reduced to match the reduced demand.
Put a number on the setpoint. In a system around 100 psig, every 2 psi increase in discharge pressure raises compressor energy use by about 1 percent at full output, and by about 1.6 to 2 percent in total once the extra air lost through leaks and other unregulated uses is counted. The same arithmetic works in reverse, which is why lowering the setpoint and fixing leaks belong in the same program.
A basic industrial ultrasonic detector is an inexpensive tool relative to the first survey's savings. It pays for itself in the first repair round on most plants. There is no ongoing cost beyond the technician time to run the survey.
Audible inspection finds the big leaks but misses the many small ones that are inaudible over plant noise. The small leaks collectively are often a larger loss than the few big ones. Ultrasonic detection finds both; the naked ear finds only the obvious.
Lowering system pressure reduces leak loss (leaks flow less at lower pressure) and is worth doing where the process tolerates it. But it treats the symptom, not the cause. The leaks are still there, still wasting energy; lowering pressure just wastes slightly less. Repair the leaks and optimise the setpoint, they are complementary.
Leak severity (estimated from the ultrasonic signal and the orifice size) maps to an airflow loss, which maps to a compressor energy cost. The estimate is approximate but good enough to prioritise. The exact number matters less than ranking the leaks by severity.
Running a one-time survey, celebrating the saving, and not making it a standing quarterly program. The leaks come back. The one-time saving erodes within two years. The recurring survey is what makes the program a durable energy reduction rather than a one-quarter win.
Programați o întâlnire individuală cu experții noștri sau înscrieți-vă direct în planul nostru gratuit.
Nu este nevoie de card de credit!
Login
Blog