{"id":4897,"date":"2026-10-02T05:58:04","date_gmt":"2026-10-02T05:58:04","guid":{"rendered":"https:\/\/sayiair.com\/?p=4897"},"modified":"2026-09-23T06:19:27","modified_gmt":"2026-09-23T06:19:27","slug":"compressed-air-leak-cost-audit","status":"publish","type":"post","link":"https:\/\/sayiair.com\/it\/compressed-air-leak-cost-audit\/","title":{"rendered":"What a Compressed Air Leak Really Costs, and How to Run Your Own Leak Audit"},"content":{"rendered":"

Nobody budgets for leaks. That’s the whole problem.<\/p>\n\n\n\n

Energy efficiency gets a capital line item. New compressors get a business case with a payback calculation. Leaks just sit there, hissing at frequencies too high to hear, quietly consuming somewhere between a fifth and third of everything the compressor produces. It doesn’t appear on any invoice, so it never appears on any meeting agenda either.<\/p>\n\n\n\n

Then someone does the arithmetic for the first time. A 3 mm hole, small enough to cover with a fingertip, can cost more per year than a genuine maintenance budget. This article is that arithmetic, plus how to find the leaks yourself with equipment you already own.<\/p>\n\n\n\n

\"Industrial<\/figure>\n\n\n\n

The arithmetic almost nobody does<\/h2>\n\n\n\n

Here’s the calculation, and it’s simpler than people expect.<\/p>\n\n\n\n

Step one, how much air a hole loses.<\/strong> At around 7 bar, choked flow through a round hole works out roughly like this:<\/p>\n\n\n\n

Hole diameter<\/th>Approximate leak rate (free air)<\/th><\/tr><\/thead>
1 mm<\/td>~1.3 L\/s (0.08 m\u00b3\/min)<\/td><\/tr>
2 mm<\/td>~5 L\/s (0.3 m\u00b3\/min)<\/td><\/tr>
3 mm<\/td>~12 L\/s (0.7 m\u00b3\/min)<\/td><\/tr>
5 mm<\/td>~33 L\/s (2.0 m\u00b3\/min)<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Note the shape of that progression. It goes with area, so doubling the hole diameter quadruples the loss. A hole that’s slightly bigger costs a lot more than “slightly”.<\/p>\n\n\n\n

Step two, turn air into kilowatts.<\/strong> Making compressed air at 7 bar takes roughly 6 to 7.5 kW of electrical input for every m\u00b3\/min delivered, depending on machine efficiency and how well it’s maintained. Use 7 kW per m\u00b3\/min for a rough estimate. Older or poorly maintained packages sit higher.<\/p>\n\n\n\n

Step three: multiply by hours and tariff.<\/strong><\/p>\n\n\n\n

\n

Annual cost \u2248 leak rate (m\u00b3\/min) \u00d7 specific power (kW per m\u00b3\/min) \u00d7 running hours \u00d7 electricity price<\/p>\n<\/blockquote>\n\n\n\n

Run the 3 mm hole through it. Leak rate 0.7 m\u00b3\/min, specific power 7 kW, a normal two-shift plant at 6,000 hours a year, electricity at $0.12 per kWh:<\/p>\n\n\n\n

0.7 \u00d7 7 \u00d7 6,000 \u00d7 0.12 = $3,528 a year.<\/strong> A two-shift plant, nothing exotic. And that’s for one hole you can hide under your thumb.<\/p>\n\n\n\n

Scale that to reality. Utility-sponsored compressed air programmes in Europe and North America have repeatedly measured that plants with no leak management lose 20\u201330% of total compressed air production<\/strong> to leaks. Take a plant running a 90 kW compressor for 6,000 hours: that’s 540,000 kWh of input energy a year, of which the leak share at 20% is around 108,000 kWh. At $0.12\/kWh, that’s roughly $13,000 a year<\/strong> in air nobody used. Every year, until someone walks the line with a piece of soap solution.<\/p>\n\n\n\n

We’ve watched smaller versions of this play out dozens of times. The reaction is always the same little pause while it sinks in.<\/p>\n\n\n\n

Why leaks multiply rather than stay put<\/h2>\n\n\n\n

Two mechanisms are worth knowing, because they explain why “we fixed leaks last year” isn’t a defence.<\/p>\n\n\n\n

Leaks grow.<\/strong> A pinhole in a threaded joint erodes. Vibration walks fittings loose. Moisture in the line corrodes everything downstream of a dryer that isn’t quite doing its job. An unmanaged leak that starts at 1 mm quite cheerfully becomes 3 mm over a couple of years.<\/p>\n\n\n\n

Higher pressure pushes more through.<\/strong> Leak flow rises with system pressure, roughly linearly and sometimes worse. Which means the standard fix for weak tools, nudging up the setpoint, increases leakage as well as energy cost. If you solve tool performance with pressure instead of by fixing the restriction, you’ve built a machine that manufactures leaks.<\/p>\n\n\n\n

There’s a third, softer factor. Once a line has audible leaks, people stop hearing them. New staff don’t even register them as abnormal. This is normal human behaviour, not negligence \u2014 and it’s why the fix has to be a scheduled programme rather than a one-off heroic weekend.<\/p>\n\n\n\n

Where leaks actually hide<\/h2>\n\n\n\n

Roughly in order of how often we find them:<\/p>\n\n\n\n

    \n
  1. Quick-connect couplings<\/strong>, the single most common offender by a wide margin. Especially the cheap ones, especially where hose is dragged across the floor.<\/li>\n\n\n\n
  2. Threaded joints and unions<\/strong>, where vibration meets poor assembly and age.<\/li>\n\n\n\n
  3. Condensate drains.<\/strong> Float-type traps that stick open, manual drains left cracked, electronic drains that died months ago. A drain stuck open can be the biggest single leak in the plant.<\/li>\n\n\n\n
  4. Regulators and filter bowls<\/strong>, where seals harden, O-rings flatten and bowls crack.<\/li>\n\n\n\n
  5. Flexible hose itself<\/strong>, pinholed by abrasion, especially where it rubs a machine frame.<\/li>\n\n\n\n
  6. Flange gaskets and filter housing seals.<\/strong><\/li>\n\n\n\n
  7. Solenoid valve bodies and actuator diaphragms<\/strong> on automated equipment.<\/li>\n\n\n\n
  8. The compressor room itself<\/strong>, with easing valves, seepage at safety valves and loose piping.<\/li>\n\n\n\n
  9. Instruments and gauges<\/strong>, small individually and ridiculous in aggregate.<\/li>\n<\/ol>\n\n\n\n

    One practical note from experience: the biggest leaks are often not<\/em> where people look first. The really valuable finds are usually a stuck-open drain or a cracked manifold inside a machine cabinet, both of which the ear misses entirely.<\/p>\n\n\n\n

    Running an audit with nothing but the compressor and a clock<\/h2>\n\n\n\n

    You can get a solid number for total leakage without buying anything. This is the shutdown decay test, and it takes about half an hour.<\/p>\n\n\n\n

    Prepare.<\/strong> Pick a time when the plant genuinely isn’t using air, or when you can arrange for everything to be switched off. Note that this includes equipment that uses air intermittently: packaging machines with pneumatics, purge valves, anything on a timer.<\/p>\n\n\n\n

    Record.<\/strong> With the system at normal working pressure and all the air-using equipment off, close the valve isolating the distribution from the compressor and receiver. Start timing.<\/p>\n\n\n\n

    Measure.<\/strong> Record how long it takes for the system pressure to fall to a chosen lower reading, say from 7.0 bar down to 6.0 bar. Repeat the test once or twice; if the times vary wildly, something was still drawing air and the run doesn’t count.<\/p>\n\n\n\n

    Convert.<\/strong> As an approximation, the percentage leakage is:<\/p>\n\n\n\n

    \n

    % leakage \u2248 (pressure drop in bar \u00f7 elapsed minutes) \u00d7 (system volume in m\u00b3 \u00f7 absolute start pressure in bar) \u00d7 100<\/p>\n<\/blockquote>\n\n\n\n

    The one input people get wrong is system volume<\/strong>. It’s not just the receiver \u2014 it’s the receiver plus every metre of pipe<\/em>, plus the volume inside dryers, filters and aftercoolers. If you don’t know it, work it out from pipe dimensions, or estimate conservatively and treat the result as an upper bound on leakage. Any plant using tank-mounted packages is already sitting on a known figure; a unit like the 7.5 HP piston compressor with an 80-gallon tank<\/a> contributes exactly that tank volume.<\/p>\n\n\n\n

    \"Technician
    System volume = receiver plus pipework. Packages like this 80-gallon tank-mounted unit<\/a> make half of that number easy to find.<\/figcaption><\/figure>\n\n\n\n

    Worked example: system volume (piping plus receiver) 3 m\u00b3, drop from 7 bar(g) to 6 bar(g) \u2014 that’s 1 bar \u2014 taking 4 minutes, starting from about 8 bar absolute:<\/p>\n\n\n\n

    (1 \u00f7 4) \u00d7 (3 \u00f7 8) \u00d7 100 \u2248 9.4% leakage<\/strong><\/p>\n\n\n\n

    Do that after hours and again on the same system three months later and you have a real trend line, which is worth far more than a one-off number. Anything above about 10% deserves attention; plants that genuinely manage their air typically sit well under 5%.<\/p>\n\n\n\n

    If you want more precision, or the plant can’t be stopped, an ultrasonic leak detector changes everything \u2014 they find leaks 8\u201310 metres away in a noisy plant, including the small ones that never show in a decay test. Most distributors rent them. Honestly, if you have more than about 100 kW of installed compressor capacity, buying one is a hard argument to lose.<\/p>\n\n\n\n

    Turning findings into repairs<\/h2>\n\n\n\n

    Finding them is the fun part. Getting them fixed is where programmes die.<\/p>\n\n\n\n

    Tag everything.<\/strong> Physical tags with a number and location, photographed into a log. It sounds like over-administration until you try to hand twenty findings to a maintenance team and watch fifteen evaporate.<\/p>\n\n\n\n

    Rank by size, but allow for difficulty.<\/strong> A 2 mm leak at a reachable coupling gets fixed first. A huge leak on a production machine that can’t stop until the next shutdown gets scheduled \u2014 it doesn’t help to put jobs at the top of a list that can’t physically be done this month.<\/p>\n\n\n\n

    Track repair impact.<\/strong> Redo the decay test after a repair round. Seeing the number fall is what gets budget renewed next year. It’s also the only way to know the work was actually done properly.<\/p>\n\n\n\n

    Set a recurring slot.<\/strong> Quarterly or twice-yearly is normal. Some plants put leak checks on the same PM route as filter changes, which is efficient. Same walk, different listening.<\/p>\n\n\n\n

    Buy better next time.<\/strong> Every failed coupling replaced with the identical cheap part is a scheduled future failure. High-flow, high-spec couplings genuinely pay for themselves in air alone, before you count the performance improvement at the tool.<\/p>\n\n\n\n

    The mistakes that put the leaks back<\/h2>\n\n\n\n

    A short list, all of these seen recently:<\/p>\n\n\n\n