This article is for plant managers, production managers, maintenance managers and continuous improvement or energy leads who pay for compressed air without knowing how much of it actually goes into making parts. You probably know what your compressors consume in a month. The question here is narrower: how much air do your lines use while they are stopped, in micro-stops or in changeovers, and how can you measure it without launching a heavy project?
- A stopped line often keeps consuming compressed air.
- The site meter shows nights and weekends, not daytime stops.
- Matching machine states with air flow reveals air used without producing.
- Act on the most expensive state first, then measure again.
Compressed air does not stop when the line stops
Maxime Laffers, an industrial energy efficiency expert at OXA Groupe, a French firm focused on energy performance for mid-sized and small manufacturers, published Air comprimé industrie : 7 leviers pour -20 % d’énergie (in French). He reviews the levers on the supply and network side: hunting leaks, lowering the pressure setpoint, variable-speed drives, compressor sequencing, shutting compressors down outside production, matching air treatment to real needs and recovering heat. He also presents the compressed air energy performance indicator, expressed in kWh per Nm³, and the French energy savings certificates (CEE) that fund these actions. On leaks, which compressors absorb without triggering any alarm, he sums it up: “Pendant ce temps, les kilowattheures s’accumulent.” (in the meantime, the kilowatt-hours keep piling up).
That line applies to machines too. The OXA Groupe article deals with producing and distributing air. This one looks at the other end of the network: machine demand, depending on machine state. A site meter tells you how much air the plant uses. It does not tell you whether that air made parts, or was blown while a line waited for material.
What a line consumes when it is not producing
When a machine stops, not all of its air consumers stop with it. Depending on the equipment, you will often find:
- Blow-off guns and nozzles left open or running continuously for cleaning, drying or ejection.
- Machine-level leaks: worn fittings, hoses, valves and cylinders that leak whether the machine runs or not.
- Venturi vacuum generators, which use air as long as they are supplied.
- Drains, air cooling and air curtains, often left on by default.
- Repressurizing a branch at every restart, when the isolation valve was closed.
At night and at weekends, this consumption shows up on the site meter because nothing else is running. During the day, it disappears into production consumption. Yet that is where unplanned stops, waiting time, micro-stops and changeovers hide.
Machine state and air consumption: a reading grid
Each machine state leaves a different signature on air consumption. Use this table as a reading grid when you put both data series side by side.
| Machine state | What still consumes air | What the data shows | Typical lever |
|---|---|---|---|
| Long stop (night, weekend, end of a run) | Network and machine leaks, blowing left on | A steady background flow while no machine is running | Isolate the branch with an interlocked valve, shut compressors down outside production |
| Unplanned daytime stop (breakdown, waiting for material or an operator) | Blow-off, venturis, cooling still on | Branch flow barely different from running flow | Automatically cut unneeded consumers after a time delay |
| Micro-stop | Everything that stays supplied between two cycles | Consumption spread over many short events, invisible in hourly averages | Reduce micro-stop frequency, tie blowing to the machine cycle |
| Changeover | Cleaning blow-off, drains, setup trials | A flow peak during the changeover, with a duration that varies by shift | Standardize cleaning, shorten the changeover (SMED) |
| Normal running | Air used by the process | A reference flow for each rate | Baseline for comparing the other states |
A step-by-step method to find air consumed without producing
Test the following method on one line or cell before rolling it out across the shop floor.
- Pick a measurable scope. Ideally a network branch fitted with a flow meter that feeds an identified line or cell. Otherwise, start from your existing compressor metering and a line that weighs heavily in demand.
- Timestamp machine states automatically. Running, stopped, micro-stop, changeover, with start, end and reason. Manual entries at the end of the shift miss exactly the short stops.
- Align the two data series. Same clock, same time step. Without this alignment, you cannot link a micro-stop to a change in flow.
- Split air consumption by state. For each period, assign the volume consumed to the machine state at that moment. You get a volume while running, while stopped, during micro-stops and during changeovers.
- Calculate two simple indicators. The share of air consumed without producing (non-running volume divided by total volume), and the average flow while stopped compared with the average flow while running. For example, a line whose flow while stopped stays close to its running flow points to consumers that do not follow the machine state. This is an example of how to read the data: every line has its own baseline.
- Act on the most expensive state, then measure again. Isolation valve interlocked with the stop signal, timed cut-off of blow-off, repair of leaks found on the machine, fewer micro-stops. Measure again over a comparable period and product mix to judge the effect on a like-for-like basis.
Micro-stops and changeovers: the hardest part to see
Long stops are easy to spot on a night-time flow curve. Micro-stops leave almost no trace in an hourly average. One by one, they look harmless. Added up over a shift, they can amount to a significant time during which the machine stays supplied without producing. For example, a line that stops dozens of times per shift on short jams keeps its blow-off and venturis running every time. Treat this as a mechanism to check with your own measurements, not as an average.
Changeovers raise a different question. Their duration often varies from one shift to another, and cleaning air follows that duration. Shortening changeovers with SMED reduces both the lost time and the air blown during that time.
This work complements the indicator approach described in our article Energy performance indicator in manufacturing: measuring kWh per part and energy used during downtime. kWh per part tells you whether energy per good part is drifting. Splitting air by state tells you during which state the air is lost. Both should be read alongside OEE: if OEE drops while the share of air consumed during stops rises, the issue is first and foremost a downtime issue.
Where TEEPTRAK fits: machine states, captured automatically
TEEPTRAK measures OEE, downtime and micro-stops in real time. A module wired into the electrical cabinet and a sensor at the edge of the line take about one hour to install per machine, with no connection to the PLC, and OEE is available within 48 hours. Every state (running, stopped, micro-stop, changeover) is timestamped automatically, with its reason. You can then match this state data against your compressed air flow readings or energy metering to split consumption by state. TEEPTRAK is not an energy metering tool: it supplies the production side of the equation. It is used in more than 700 plants in over 30 countries, with an average gain of 10 to 15 OEE points.
Further reading
- Air comprimé industrie : 7 leviers pour -20 % d’énergie (Maxime Laffers, OXA Groupe, in French)
- Energy performance indicator in manufacturing: measuring kWh per part and energy used during downtime
- The 6 big losses: what your manual log doesn’t see
FAQ
Why measure compressed air consumption during downtime rather than total consumption? Because total consumption mixes useful air and wasted air. Split by machine state, it shows how much air goes out during stops, micro-stops and changeovers, and therefore which lever to tackle first.
Do I need a flow meter on every machine? No. Most plants start with a network branch or cell that is already metered, matched against the state of the machines it feeds. Finer metering is then justified where the share of air consumed without producing is highest.
How does this relate to OEE? OEE measures lost time; splitting air by state measures the energy lost during that time. Reducing stops and micro-stops improves both, provided you measure them on the same scope.
0 Comments