Compressed Air Leak Detection 2026 — Stopping Leaks That Keep Coming Back

Compressed air leak programs usually start life as an energy-saving activity. The plant is shut down over a weekend, someone walks the lines with an ultrasonic detector, tags go onto every fitting that hisses, and a report goes up counting how many leaks were repaired. Nothing about that is wrong. The problem shows up a year later, when the same measurement is repeated and the leak rate has drifted back to roughly where it started. That outcome is common in Japanese-owned plants in Thailand. This article reframes compressed air leaks as something they genuinely are — not a stock you clear once, but a flow that returns if nothing keeps watching — and sets out how to measure them, what order to fix things in, and how to make the improvement stick, using Thai electricity tariffs and Thai regulations as the working context.

Three numbers to confirm before starting any compressed air leak program

When a plant asks for help with air leaks, there are three numbers worth confirming before anyone talks about detectors. Starting a leak hunt without them means you cannot express the result in money, and the following year’s budget request goes nowhere.

The third one is the one that gets skipped. Discussion of compressed air leaks pulls naturally toward driving leakage to zero, but the same electricity saving is sometimes available first through a pressure setpoint change that costs almost nothing. Get the order wrong and the payback arithmetic on the whole program turns against you.

The rest of this article works outward from those three numbers, arranging measurement, repair and monitoring into a sequence that holds.

What a compressed air leak actually is — and where plants leak from

Compressed air looks like free air, but it is one of the most expensive utilities in any factory. Most of the energy put into drawing in atmosphere, compressing it, cooling it, drying it and piping it around the site is lost as heat, so the electricity cost carried by each cubic metre delivered at the point of use is not light compared with water or steam. A hole in that system draws electricity 24 hours a day while contributing nothing to production.

Leak locations feel plant-specific, but in practice they cluster hard. The usual suspects:

• Quick-connect fittings and couplers. Anywhere air tools or blow guns are connected and disconnected repeatedly, O-ring wear runs fast. The more changeovers a line has, the higher the rate.
• Drain traps and automatic drain valves. A valve that never quite closes makes noise continuously, so the shop floor stops hearing it as a fault and it never gets reported.
• Bends and pinch points in polymer hose. Hose routed along moving parts of a machine develops fatigue cracks that are too fine to see.
• Solenoid valves and cylinder rod seals. These hide inside the machine’s own operating noise and cannot be picked up by ear while the line is running.
• Pipe branches and taped joints. Plants that have been extended repeatedly carry a lot of branches, and variation in installation quality converts directly into leakage.
• Continuous air blowing. Strictly this is artificial demand rather than leakage, but if it does not stop during non-production hours, it costs exactly what a leak costs.

In Thailand there is a climate layer on top of this. Dry-season dust drives filter differential pressure up, which creates a standing incentive to raise the discharge pressure to compensate. Wet-season humidity increases condensate volume, which raises both the cycle count and the failure rate of drain traps. At sites where piping runs outdoors in direct sun, polymer components degrade faster than the same parts do on identical equipment in Japan. Inspection intervals imported unchanged from the parent company in Japan are sometimes simply too long for these conditions.

How to measure leak volume — the no-load test and flow meter monitoring

Turning a qualitative sense that “we probably leak a lot” into a number you can convert into money can be done two ways. Which one you pick determines what you are able to do afterwards.

The first is the no-load test, also called a load/unload test. Production stops completely, every isolation valve to air-consuming equipment is closed, one compressor is run alone, and the loaded (compressing) and unloaded time are timed. Leak rate comes out of a simple ratio.

Leak rate (percent) = total loaded time / (total loaded time + total unloaded time) x 100

Take at least 5 cycles and average them. One or two cycles get pulled around by receiver tank volume and setpoint variation. The advantage of this test is that it needs no instrumentation at all — a stopwatch and the run log are enough, and half a day of weekend overtime completes it.

The second is flow meter monitoring. A flow meter is installed permanently on the main header, and flow during non-production hours — weekends, night shifts, lunch breaks — is logged continuously as a baseline. Air flowing while nothing is being produced is, by definition, either leakage or artificial demand.

The last row of that table is the doorway into this article’s main argument. A no-load test tells you today’s leak rate. It cannot tell you that leakage started rising last week. As long as management runs on a single annual data point, up to a year can pass between leaks growing and anyone noticing. The electricity for that year is paid quietly, by everyone, without anyone deciding to.

Why repaired compressed air leaks always come back

The classic failure mode in a compressed air leak program is executing “find and fix” as a one-off event. The first pass always delivers, because fittings neglected for years turn up in bulk. The trouble starts afterwards.

Leakage is a flow, not a stock. Repair reduces the stock, but as long as the sources keep running, new leaks keep being supplied. The sources look like this:

• Air tool couplers connected and disconnected at every changeover
• Branches and temporary piping added during equipment upgrades and layout changes
• Age-driven degradation of polymer hose and O-rings
• Wear proportional to the cycle count of drain traps
• Provisional pipework left in place after a new line is commissioned

In other words, the rate at which leaks appear is a function of how volatile that plant’s production is and how old its equipment is on average. Until you change that function, repair only outruns the inflow temporarily, and the level drifts back.

There is a second reason, and it is organisational. When a compressed air leak program is set up as “an activity owned by the energy team”, nobody owns the repair. The people who actually swap fittings are the maintenance department, and what they need is a work order raised, spare parts in stock and hours allocated. If the energy engineer tags a leak but the tag never becomes a maintenance work instruction, that tag is still there six months later.

At sites in Thailand this gap opens wider by default. The person driving energy savings is often a Japanese expatriate, while the people holding the tools are Thai maintenance staff. If work procedures and leak judgement criteria are not documented in Thai, the activity ends when the expatriate’s posting ends. That is what is meant by the rule that an air leak program has to be designed as maintenance work, not as an energy initiative. Concretely, three things have to be true:

• Discovering a leak has to result in a work order raised in the CMMS or on the daily inspection sheet
• Fittings, O-rings and hose have to be stocked as standard consumables, so a tagged leak is not left for weeks waiting on a purchase order
• Leak rate has to be reported monthly as a maintenance KPI, not as a slide in the energy committee pack

Buy a detector into a plant where those three are absent and the first inventory sweep will go well, but by year two the instrument is asleep in a cupboard.

Four compressed air leak detection methods and where each one fits

Once you know how much you are leaking, the next question is where. There are broadly 4 methods, and each covers different ground. You do not need all of them; the right mix depends on plant size and how bad the leak rate is.

The principle for combining them is simple. A flow meter answers how much you are leaking; ultrasonic answers where it is leaking from. You need both, and neither closes the loop alone. Flow meters alone will not tell you what to repair; ultrasonic alone leaves no numerical record that the repair worked.

Whether to buy an ultrasonic detector or outsource the survey comes down to frequency. If all you do is one annual sweep, outsourcing is cheaper and the surveyor will be more skilled than your own staff. If layout changes are frequent and you want a check every month, owning the instrument and training maintenance staff works better. A reasonable dividing line is whether you will survey more than 4 times a year. Owning it, though, comes bundled with obligations: standardising the measurement procedure and keeping a record format that survives staff transfers.

There is also an international standard for assessing the energy performance of compressed air systems, ISO 11011, published in 2013. It provides a framework covering the supply side, the distribution system and the demand side as one assessment, and it works well as the basis of a scope of work when commissioning an external audit. If you have not written down how far the audit should reach, the common outcome is a report covering only the supply side, with artificial demand on the shop floor left completely untouched.

Three measures that often pay off before fixing any leak

Leak repair is patient work that consumes maintenance hours before it shows results. Before starting it, check whether savings are still available from setpoint changes and operating rules. There are 3 worth checking.

Lowering discharge pressure is the most reliable of the three, and it reduces leak volume as a side effect. The mass flow escaping through an orifice depends on upstream pressure, so a lower setpoint means less air lost through the same hole. Over-reducing it, however, causes equipment to misbehave. The correct procedure is to log the inlet pressure at the most distant and most pressure-hungry machine while it is running, then add back the margin it needs — typically the variation in pipe friction loss and filter differential pressure — and set the compressor from there. At plants that have been raising the setpoint for years instead of cleaning fouled filters, filter and drain maintenance has to come first.

Isolating idle circuits is particularly effective in Thailand. Where the plant runs two shifts and some lines are down overnight and at weekends, the leakage during those hours can be eliminated by isolation without repairing a single fitting. Making it stick as routine practice means labelling valve positions and operating steps in Thai and adding one line to the end-of-shift checklist. It has to be embedded in the daily work procedure, not posted on an energy campaign noticeboard.

Reducing artificial demand comes last because it is the hardest to see. Continuous air blowing is air “in use”, so no detector reacts to it and the shop floor does not perceive it as a fault. The only way to find it is per-circuit flow metering compared against a non-production-hour baseline. Once again, monitoring has to come first.

Putting a Thai baht figure on the loss at 3.95 baht per kWh

Now the money. What follows models a Japanese-owned metalworking plant in Chonburi province. The numbers move site by site, so replace them with your own measured values.

Assumptions

• Compressors rated 75 kW, two units, one running continuously and one on standby. The calculation counts only the continuously running unit and excludes the second unit’s contribution at peak
• Average load factor 70 percent, giving average power draw of 75 x 0.7 = 52.5 kW
• Running hours of two shifts, 16 hours per day x 25 days per month x 12 months = 4,800 hours per year
• Annual energy consumption of 52.5 x 4,800 = 252,000 kWh per year
• Electricity price of 3.95 baht per kWh, the average tariff for the May to August 2026 billing period, with an Ft of 16.23 satang per kWh

The calculation

Annual electricity cost attributable to compressed air is 252,000 x 3.95 = 995,400 baht per year. Applying the general leak rate of 20 percent, the amount walking out through leaks is 995,400 x 20 percent = 199,080 baht per year.

Here it matters not to overstate the result. Suppose the program halves the leak rate from 20 percent to 10 percent. The energy saved is 252,000 x 10 percent = 25,200 kWh per year, worth 25,200 x 3.95 = 99,540 baht per year. Against compressor electricity that is a 10 percent reduction, but if the compressors account for 25 percent of the plant’s electricity, total plant consumption is 252,000 / 0.25 = 1,008,000 kWh per year, and the reduction is 25,200 / 1,008,000 = about 2.5 percent. If senior management has been told that an air leak program will make a large dent in the electricity bill, the results report is going to disappoint. Saying up front that this is a 2 to 3 percent story at plant level is what keeps the program alive.

Unit price varies with the tariff class at each site. Working rates for factories in Thailand sit broadly in the 4.10 to 5.50 baht per kWh band, higher than the average because of demand charges and power factor adjustment, and large TOU contracts are structured around roughly 3.80 baht per kWh off-peak and 5.27 baht per kWh on-peak. The same 25,200 kWh saving therefore moves like this.

What this table shows is that the investment case for a compressed air leak program turns not only on how much you leak but on when you leak. A plant that produces only during on-peak hours and a plant running mainly night shifts will differ by more than 30 percent in loss value at the same leak rate. At sites on a TOU contract, time-of-day flow data raises the precision of the investment case directly.

For comparison, run the pressure reduction through the same assumptions. If dropping 0.1 MPa cuts power consumption by 4 to 5 percent, that is 252,000 x 4 to 5 percent = 10,080 to 12,600 kWh per year, worth 39,816 to 49,770 baht per year at 3.95 baht per kWh. That is roughly half the benefit of halving the leak rate, available from a setpoint change and some measurement. The argument for looking at pressure first is right there in the numbers.

Cost and payback across three layers — audit, repair, permanent monitoring

Compressed air leak spending falls into 3 layers that behave quite differently. Collapsing them into a single figure when comparing quotations leads to bad decisions, because the cost structure and the payback logic differ layer by layer.

Calculate payback separately for each layer. Layers 1 and 2 can be judged on single-year cost effectiveness. Using the model plant above with an annual saving of 99,540 baht, the payback period in months is straightforward.

Payback in months = investment / (annual saving / 12)

If layers 1 and 2 together come in at 100,000 baht, payback is 100,000 / (99,540 / 12) = about 12.1 months, so a little over a year. The trap in that calculation is that it assumes the saving persists into the following year. As covered above, without monitoring the leak rate returns. If year two drifts back to the original level, the real payback period more than doubles.

Layer 3, permanent monitoring, is the investment that prevents that return. Its payback therefore has to be built not on a single year’s saving but on how many years the saving can be held. The return on layers 1 and 2 alone looks attractive, but it is worth understanding that this is the first-year picture only.

The figures above are illustrative. Actual audit fees, parts costs and instrumentation costs in Thailand vary widely with specification and site conditions, so obtain quotations against your own configuration.

Watching compressed air leaks continuously — connecting to utility monitoring

To avoid ending on “visibility is important”, decide what gets measured, at what granularity, by whom, and when. The minimum viable setup for continuous leak monitoring looks like this.

What to measure, and at what granularity

• Compressed air flow on the main header, at 1-minute values. This is the primary leak indicator
• Compressor room electrical power, at 1-minute values. Cross-referenced against flow, this yields specific energy consumption per cubic metre of air, in kWh per cubic metre
• Pressure at 2 or 3 representative points at the far end, at 1-minute values. A growing pressure drop signals either restriction in the piping or rising demand
• Discharge pressure setpoint, as a change history, so you can see whether someone has quietly raised it

The alerting rule

Log the minimum flow during non-production hours — Sunday and overnight idle periods — as a monthly baseline. In practice the rule should be expressed as a rate of change rather than an absolute value, for example alerting when the figure rises 15 percent against the previous month’s baseline. Absolute targets lose meaning as seasons and product mix shift.

Who looks, and when

• Daily, the maintenance lead checks the previous night’s minimum flow. One dashboard screen, one minute
• Weekly, the maintenance meeting reviews where alerts fired and how many repair tags remain open
• Monthly, specific energy consumption in kWh per cubic metre and the leak rate trend are reported as a maintenance department KPI

That monthly maintenance KPI is the load-bearing part. Reporting it only in the energy committee pack will never raise the priority of the repairs.

Compressed air flow monitoring pays back better when it is designed as one part of a utility monitoring platform covering electricity, water and gas, rather than deployed on its own. The measurement points differ, but the logging platform, dashboards and alert workflow can all be shared. If your site has already started monitoring electricity, adding a flow meter to the same platform is the shortest route. The way to structure power measurement from the incoming supply down to individual machines is covered in our article on factory power monitoring systems.

The other practical question is how to get signals out of the compressors you already own. Even on older machines with no communication capability, running status can be captured with external CTs and a pressure transmitter. Retrofit methods that do not require opening the control panel are described in our article on IoT retrofit for legacy equipment. And if drain trap cycle counts or compressor vibration are going to be brought into scope as condition monitoring, aligning the design with the approach in predictive maintenance systems avoids rework when the systems are integrated later.

Thai regulation — the 150 percent deduction for energy-efficient machinery and designated factory duties

Two Thai frameworks are worth knowing about when planning an energy investment.

The first is the tax incentive under Royal Decree No. 805 (B.E. 2569). High-efficiency machinery acquired and put into use between 3 March 2026 and 31 December 2028 qualifies for a deduction of 150 percent of the investment amount. Eligibility is tied to the DEDE and EGAT five-star label scheme, so the qualifying equipment determination follows that labelling system. One caveat: it cannot be combined with businesses already receiving corporate income tax exemption under BOI or EEC promotion, because during an exemption period the deduction itself carries little benefit.

In relation to compressed air leaks specifically, leak repair is not an acquisition of machinery and therefore falls outside the scheme. Replacing a compressor with a high-efficiency inverter-driven unit, on the other hand, can qualify if the labelling requirement is met. Because eligibility involves interpreting the qualifying equipment list, confirmation with a tax adviser and with DEDE is essential. The description here is general information drawn from published commentary and is not a basis for deciding an individual case.

The second is the designated factory obligation under Thailand’s Energy Conservation Promotion Act, commonly called the ENCON Act. Designated factories are required to appoint energy managers — 1 person where contracted demand or installed capacity is below 3 MW, and 2 where it exceeds 3 MW. At sites that fall under the designation, energy management reporting is already a legal duty, so compressed air flow data is not new overhead but evidence that supports reporting you are obliged to produce anyway.

Taken together, these two shift how a compressed air leak program should be positioned in Thailand. On its own it is a 2 to 3 percent plant-level measure. Designed as part of the measurement platform that satisfies designated factory reporting, the monitoring investment no longer has to be justified on compressed air economics alone, and the hours your energy manager spends assembling the annual report can come out of the same platform.

Implementation sequence — a 90-day roadmap

Here is the whole argument rearranged into execution order. The goal for the first 90 days is to reach a state where you can measure, and to have the skeleton of the mechanism that prevents regression. Do not try to finish the repairs in 90 days. Repair is ongoing work, not a project.

Two things are deliberately left out of this roadmap. The first is completing every repair. Leak locations can number in the dozens or the hundreds, and clearing them all within 90 days would stop maintenance doing its normal job. Rank them by leak volume multiplied by accessibility and work through them as routine monthly business.

The second is asking for an investment decision on day one. The flow meter appears on day 61 because the preceding 45 days of measured pressure-reduction results give you the material to justify spending on monitoring. A proposal that opens with “please buy us a measurement system” does not get approved at most plants. Deliver a result first with the numbers already to hand, then argue that monitoring is what keeps that result — that sequence is far more realistic.

Beyond day 90, operation runs on the three tiers described above: one minute daily, a weekly review and a monthly KPI. Once that is turning, leak rate stops being a number that shocks everyone once a year and becomes an indicator that ticked up slightly last week. That is what a compressed air leak program that does not regress looks like.

Frequently asked questions

Where should we start with a compressed air leak program?

Not by ordering a detector. Start by measuring the discharge pressure setpoint against the pressure actually required at the far end. If the gap is 0.1 MPa or more, a setpoint change alone may release 4 to 5 percent of power consumption, and the labour involved is orders of magnitude smaller than leak repair. In parallel, run one no-load test over a weekend to establish the current leak rate. Both cost almost nothing and both become the evidence base for later investment decisions.

How much does an air leak survey cost?

The three layers behave differently: the audit in layer 1 is charged per visit when outsourced, repair in layer 2 is parts plus maintenance hours, and permanent monitoring in layer 3 is upfront investment plus running cost. Going rates vary so much with site size, pipe length and the number of measurement points that this article does not quote figures. When you request a quotation, use the ISO 11011 framework to write the scope first, stating how far the audit reaches across supply, distribution and demand. Order without defining scope and you will get a supply-side report while demand-side issues such as continuous air blowing remain untouched.

Should we buy an ultrasonic detector or outsource the survey?

Survey frequency decides it. For one or two inventory sweeps a year, outsourcing is cheaper and the surveyor will be more practised. For a plant with frequent layout changes and equipment upgrades that wants a check 4 or more times a year, ownership works. If you do buy, plan the purchase together with standardising the measurement procedure, training the staff and setting up a record format. Without those, the instrument goes to sleep in a cupboard the moment the trained person transfers. Note that ultrasonic detectors pick up roughly the 38 to 42 kHz band, above the range of human hearing, which is precisely why they can survey a plant that is still running.

How much will the electricity bill fall from an air leak program alone?

Plan on 2 to 3 percent of total plant electricity. Compressors account for 20 to 30 percent of plant power, leakage takes 20 percent of that, and even halving the leak rate lands at 2 to 3 percent plant-wide. In money, the model plant above — 75 kW class, 252,000 kWh per year, 3.95 baht per kWh — comes to roughly 99,540 baht per year. Whether that reads as small or as a reliable annual return is a management judgement, but if the expectation is not set in advance, the program stalls at the results report.

Can we manage with no-load tests alone and skip the flow meter?

You can, but you will be slow to notice leakage growing. With an annual no-load test, a leak that appears just after a test goes untreated for up to a year. The no-load test is a point measurement, so you also cannot tell whether a repair worked until the next test. If budget rules out a flow meter for now, the workable compromise is raising the test frequency to quarterly. Each test requires a full shutdown, though, so the coordination cost against the production plan accumulates. Compare the annual cost of that downtime coordination against the flow meter investment and most sites find the permanent instrument wins.

Can Thai energy-efficiency tax incentives be used for a compressed air leak program?

Leak repair itself is not an acquisition of machinery, so it is outside the scheme. The 150 percent deduction under Royal Decree No. 805 (B.E. 2569) applies to high-efficiency machinery acquired and put into use between 3 March 2026 and 31 December 2028, with the DEDE and EGAT five-star label as a requirement. Replacing a compressor with a high-efficiency unit can qualify, but the scheme cannot be combined with businesses receiving BOI or EEC corporate income tax exemption. Because eligibility involves interpretation of the qualifying equipment list, always confirm with a tax adviser and with DEDE.

Summary

The points worth holding onto from any compressed air leak program:

• Leakage is a flow, not a stock. A one-off find-and-fix returns to the original leak rate within one to two years, because production volatility and equipment ageing keep supplying new leaks
• Confirm three numbers first: the share of plant electricity taken by the compressors, typically 20 to 30 percent; compressed air flow during non-production hours; and the gap between discharge pressure and the pressure the far end actually needs
• Before repairing anything, check lowering the discharge pressure, where 0.1 MPa yields 4 to 5 percent of power consumption, plus isolating idle circuits at the valve and cutting artificial demand. Almost no labour, and reliable results
• Measurement comes in two forms — the no-load test as a point and flow meter monitoring as a line — and you need both. The test gives today’s leak rate; the meter tells you when it started climbing
• Do not overstate the benefit. Halving a 20 percent leak rate is about 2 to 3 percent of total plant electricity, roughly 99,540 baht a year in the model plant
• To make it stick, design it as maintenance work rather than an energy initiative. Without a raised work order, stocked spares and a maintenance KPI, the tags simply stay where they were stuck
• In Thailand, factory electricity sits in the 4.10 to 5.50 baht per kWh band depending on the contract, and under TOU pricing time-of-day data drives the accuracy of the investment case. Build the Royal Decree No. 805 deduction of 150 percent and the designated factory energy manager duty into the design as well

What decides whether a compressed air leak program succeeds is not detector performance or audit precision. It is whose job it is to keep measuring. Settle that, and surveying and repair become ordinary annual business.

TOMAS TECH builds monitoring platforms for compressed air, electricity and production data at Japanese-owned plants in Thailand. Capturing signals from compressors already installed, designing flow meter installations, and designing the dashboard the maintenance team actually looks at each day — we work through all of it against the conditions at your site. If you are unsure how to start measuring your own leak rate, or which layer to invest in first, get in touch through our contact page. We will start from your current equipment configuration and electricity contract and help you build the sequence.

Reference information

• ERC electricity tariff, May to August 2026, Nation Thailand — https://www.nationthailand.com/business/economy/40064550
• Thailand 150 percent deduction for energy-efficient machinery, Bizwings — https://www.bizwings.co/post/thailand-introduces-tax-incentives-for-solar-rooftop-installation-and-energy-efficient-machinery
• ISO 11011, energy efficiency assessment of compressed air systems — https://www.iso.org/standard/46580.html
• Keyence, energy-saving examples by equipment type covering fluid and production equipment — https://www.keyence.co.jp/ss/products/process/energy-saving/facility-examples/fluid-production.jsp
• IHI Rotating Machinery Engineering, reducing compressed air leakage — https://www.ihi.co.jp/compressor/solution/air_leak.html
• J-Net21, energy saving in compressed air systems — https://j-net21.smrj.go.jp/development/energyeff/Q1166.html
• Factory electricity cost in Thailand, tariff types 3 to 5, CapSolar — https://capsolar.co.th/en/knowledge/factory-electricity-cost-thailand
• ECCJ, energy efficiency policy in Thailand — https://www.asiaeec-col.eccj.or.jp/archive/eec-policy-country/thailand/