Compressed air is the fourth utility in most manufacturing facilities — after electricity, gas, and water — yet it is the one where energy loss is least visible. A 6-bar distribution network running through a 200-person discrete manufacturing plant will lose, on average, somewhere between 20 and 30 percent of its total compressed volume through leaks before the air reaches a single tool or actuator. At typical industrial electricity rates in Sweden and Germany, that translates to €15,000–€60,000 per year disappearing into fittings, worn couplings, and corroded valve seats. The plant keeps buying electricity to run compressors harder to maintain system pressure. Nobody notices, because the machines still work.
This article is about why that pattern persists, what the physics of leak losses actually look like in load-curve data, and what continuous pressure monitoring changes about the economics of finding and fixing those leaks.
The Physics: Why Small Leaks Add Up Faster Than You Expect
A compressed air leak's flow rate increases roughly with the square root of the pressure differential across the orifice. At 6 bar gauge pressure and a standard 6 mm hole diameter — the kind you get from a cracked push-in fitting on a pneumatic cylinder — you are looking at approximately 8–10 m³/hour of free air loss. Run that 24 hours per day and you need roughly 1.5–2 kW of continuous compressor input to compensate. Across an entire distribution network with 15–20 such fittings in various states of wear, the compressor load impact becomes substantial.
The harder problem is detection. The human ear can detect a loud 6 mm leak near a quiet zone, but most plant floors are not quiet zones. A 1 mm leak at 8 bar generates a high-frequency ultrasonic signature between 38–42 kHz — above human hearing range — and requires an ultrasonic leak detector to find. Manual ultrasonic surveys are typically done once every 12–18 months during planned maintenance shutdowns. Between surveys, leaks develop and compound. Fittings that were tight in January are weeping by June.
What Load-Curve Data Actually Shows
The diagnostic signature of an increasing leak burden is readable in compressor load curves, if you sample at sufficient resolution. At 1-second data capture, the pattern looks like this: baseline compressor load gradually rises over weeks as network pressure settles lower; the compressor cycles more frequently between cut-in and cut-out pressures (a shorter duty cycle at the same average load); and in extreme cases, the compressor transitions from intermittent operation to near-continuous operation to maintain set-point pressure.
Consider a plausible scenario: a precision machining facility in central Sweden — call it Arktis Precision Components — operating two 55 kW rotary-screw compressors in a lead-lag configuration. At commissioning in 2021, the lead compressor ran at roughly 62% average load. By late 2023, that same compressor was running at 81% average load during production hours, while system pressure had drifted from a set-point of 7.2 bar down to 6.6 bar. No alarms had fired. No work orders had been raised. The operators had simply noted that the second compressor sometimes kicked in during peak production periods, which they attributed to higher production volumes.
Load-curve analysis of that 24-month drift tells a different story. The gradual load creep — 19 percentage points over roughly 28 months — combined with the pressure set-point drift, is the textbook signature of cumulative leak growth. An ultrasonic survey in early 2024 identified 23 individual leak points, with the six largest accounting for approximately 14,200 kWh/month in wasted compression work. That is roughly €1,600/month at Swedish industrial electricity rates, for leaks that had been growing undetected for over two years.
The Detection Gap: Why Annual Surveys Are Not Enough
We are not saying that ultrasonic leak surveys are ineffective. They are the correct tool for physical location and quantification of individual leak points. But a survey is a snapshot. The typical maintenance cycle for a mid-size discrete manufacturer means leak surveys happen once per year, often tied to a summer or Christmas shutdown. Between surveys, leaks develop on newly installed components, fittings work loose under thermal cycling, and polymer push-in connectors harden and crack. By the time the next survey occurs, the network has accumulated another year of load growth.
Continuous pressure monitoring at the compressor discharge, header, and key distribution branch points creates a different capability: not the ability to find individual leak locations, but the ability to detect the rate of leak development in near real-time. A pressure decay test run automatically during a scheduled night-shift idle period — when all tools and actuators are known to be closed — gives a direct measurement of leak flow rate across the entire network. If that measurement trends upward week over week, the maintenance team knows a survey is warranted before annual schedule, not after.
The economic argument for this is straightforward. A growing network leak burden that adds 5% to monthly compressor energy cost over 4 months represents roughly €2,000–€6,000 in additional electricity cost before the annual survey catches it. The cost of deploying a pressure sensor on each compressor and three or four distribution header points — the hardware investment is typically under €800 — and running continuous monitoring pays for itself within the first avoided month of undetected leakage growth.
Pressure Decay Testing: The Diagnostic You Are Already Running (Inefficiently)
Most plants with pneumatic systems have experienced a version of pressure decay testing, usually informally: someone shuts off the compressor at the end of a shift and comes back in the morning to see whether the system held pressure. If the pressure gauge reads significantly lower than the previous evening, there is a leak somewhere. This is the right intuition executed poorly.
A structured pressure decay test defines: (a) the starting pressure and ambient temperature, (b) the duration of the test window, (c) all open tool and actuator connections that need to be confirmed closed before the test, and (d) an acceptable pressure decay threshold in bar per hour. From those four parameters, you can calculate the approximate total leak flow in standard m³/hour, and from that, the electrical cost of running the compressor to compensate.
Automated overnight decay tests run by a monitoring platform provide something a manual test cannot: trend data over time. A single decay test tells you the current leak burden. Monthly decay tests over 12 months tell you whether the leak burden is growing, stable, or improving after a repair campaign. That trend is what drives maintenance prioritization. A facility that shows a 10% month-on-month growth in decay rate should be scheduling an ultrasonic survey. A facility that shows a stable decay rate after a repair campaign has evidence that the repairs held.
Where the Savings Actually Come From
The financial benefit of leak reduction is not simply the electricity saved by the compressor running less. There are three distinct cost buckets:
- Direct electricity reduction: Each 1 kW reduction in average compressor load saves approximately 8,760 kWh/year. At €0.11–0.16/kWh for Swedish industrial tariffs, that is roughly €960–€1,400 per kW per year.
- Demand charge reduction: In EU industrial tariff structures, including the Swedish effekttariff and the German Leistungspreis (capacity charge), the highest 15-minute demand interval in each billing period often determines a significant portion of the monthly bill. Compressors that cycle aggressively to compensate for leak losses can create demand peaks that would not otherwise occur. Reducing leak burden flattens those peaks.
- Compressor wear and maintenance cost: A compressor running at 85% average load versus 65% average load accumulates service hours faster, reaches oil change and valve inspection intervals sooner, and runs hotter — reducing bearing life. The maintenance cost difference between a compressor running in its designed efficiency band versus near its upper load limit can add 20–35% to annual maintenance spend over a five-year period.
A Note on Detection Methods and Their Limitations
Continuous pressure monitoring is a trend-detection tool, not a location tool. It tells you that the leak burden has increased; it does not tell you where the new leak is. This distinction matters for maintenance workflow. The correct operating model is: continuous monitoring triggers a survey when the decay trend crosses a threshold; the survey team uses ultrasonic equipment to locate and tag individual leak points; repair work orders are raised with the specific leak list and the estimated energy recovery per repair, so prioritization is cost-justified rather than ad hoc.
What continuous monitoring replaces is the expensive false negative: the scenario where a plant runs an annual survey, finds and repairs 15 leaks, and then the network slowly redevelops 12 new leaks over the next 11 months before the next survey. With a pressure decay trend running in the background, that redevelopment is visible within two to four weeks of onset, not 11 months later.
Compressed air systems are forgiving enough that most plants can run with elevated leak rates for years before a crisis occurs — the machines keep working, the compressors keep running. That tolerance is exactly what makes this category of energy waste so persistent. The cost is real, it is large, and it is almost entirely avoidable with metering that any plant engineer can deploy without external consultants or a major capital project.
