- Author
- Byron Raal, CAS Founder-Editor About the author
- Checked against
- ISO 8573-1
- Date last checked
- 21 August 2026
Looking for a compressed air provider? Tell us what you need. There is no cost to enquire.
Any proposed introduction is discussed with you first. We do not sell equipment.
Your compressor gets specified for the few weeks you are crushing and you pay to run it for fifty-two. On the illustrative 30 kW winery costed below, about as much of the leak bill burns outside vintage as during it: roughly $2,690 of wasted electricity in the months with no fruit in the building against roughly $2,565 across the crush, both figures conditional on your compressor’s controls actually converting avoided air into avoided power. That is why the assessment belongs in October to December, not in the week your presses are already running double shifts.
Air does real work in a winery, and one thing it must never do is protect wine from oxygen. That belongs to nitrogen, carbon dioxide and occasionally argon, and anyone who blurs the two hasn’t smelled an oxidised Semillon.
Work the energy in this order: leaks, then set point, then receiver volume and control. Alongside it, map every point where air reaches wine or a product-contact surface, and don’t park that behind the energy work.
Where compressed air works, and where gas takes over
Air, typically. Press membrane inflation at crush. Actuation of butterfly valves, sorting-table gates and destemmer controls. Air-operated diaphragm pumps on must, lees and juice. Packaging motion: filler, capper or corker, capsule spinner, labeller, reject arms. Air knives and bottle rinsers. Clearing residual water and cleaning solution from hoses, barrels and filter housings.
Nitrogen, carbon dioxide or argon, typically. Headspace blanketing on tanks and barrels. Purging empty tanks, hoses and bottles before fill. Sparging to strip dissolved oxygen or trim dissolved carbon dioxide before bottling. Moving finished wine anywhere oxygen pickup would cost you the batch. The Australian Wine Research Institute is explicit that the main sparging gases are nitrogen and carbon dioxide or combinations, that argon is occasionally used for headspace blanketing, and that all should be food grade at 99.5 per cent purity. Your compressor makes atmospheric air at roughly 21 per cent oxygen, and no filter train changes that. A membrane or PSA nitrogen generator does change it, by separating nitrogen out of a compressed air stream, but that is a separate machine carrying its own purity specification, not your shop air doing double duty.
The boundary is thinner than it looks in two places. A press membrane is one elastomeric barrier between shop air and fruit; a diaphragm pump is one diaphragm between shop air and must. Both are built to hold, neither is a purity system, and either one fails into product.
Pigging is the other, though here the risk lives in the configuration, not the method. Sanitary pigging systems routinely run on compressed air with the pig itself as the physical barrier between propellant and product, and double-pig arrangements are built for exactly that separation. What matters in oxygen-sensitive service is validation: a worn seal or an open bypass puts shop air against wine, so check the pig seals and bypass valving, or specify a double-pig or inert-gas propellant arrangement.
The seasonal problem: sized for the weeks you crush, paid for all fifty-two
Australian vintage runs from roughly January to April and moves with the region and the season. AWRI’s own compressed-vintage timeline puts white harvest across January and February and reds from March, then shows the compressed case dragging reds forward into February and closing out by April. Your intake window is much shorter than that national envelope, which is the whole problem. Through those weeks presses cycle continuously, must pumps run, valves fire every few minutes and plenty of wineries work around the clock. Then it stops. From May the load is barrel work, filtration, a bottling run or two, and a few actuators holding position.
The peak is sharpening, not flattening. Wine Australia’s vintage advancement and compression project describes the same squeeze in its own words: varieties that used to ripen over four to six weeks in the 1990s now mature over a much shorter time period. More air in fewer weeks buys a bigger compressor only if those peak demands genuinely land at the same time; some wineries sequence their presses and use receiver storage to shave the peak instead. Measure or log a vintage week before you size anything. Where the coincident peak is real, the bigger machine buys a deeper trough for the other eight or nine months.
For scale: Wine Australia puts refrigeration at as much as 50 to 70 per cent of winery electricity consumption. Compressed air is not the largest line on your bill and this page won’t pretend otherwise. It’s the line nobody meters, nobody profiles and nobody switches off, which is why it leaks for eight months unnoticed.
A fixed-speed machine chosen to hold your press peak then runs the rest of the year well below that peak, and what the quiet months cost depends on its control mode. On load/unload control without stop or timer logic, unloading isn’t switching off: the motor keeps turning and keeps billing you while nothing useful leaves the package. A machine with working stop or timer logic shuts down instead, so check how yours is set before you cost the year. The reliability argument is worth more again. A failure in September is a maintenance job. In March it stops the press, and fruit doesn’t wait.
Worked example: a 30 kW winery compressor, vintage and off-season
Assumptions. One rotary screw compressor, 30 kW nominal motor rating, an illustrative proxy for electrical input; true package input runs higher because it carries cooling fan, drive and control losses, so pull the CAGI data sheet for your model. Vintage 10 weeks at about 100 hours a week, so 1,000 hours; off-season 42 weeks at about 25 hours, so 1,050 hours; total 2,050 hours. Electricity $0.30 per kWh, our illustrative rate rather than a national benchmark; if your bill reads 22 or 38 cents, use that. Specific power 6.5 kW per cubic metre per minute of free air. Grid factor 0.62 kg CO2-e per kWh, the national location-based Scope 2 figure from NGA Factors 2025, with state factors varying materially.
Step 1. Baseline and capacity. 30 kW x 2,050 hours x $0.30 = $18,450 a year, split $9,000 across vintage and $9,450 outside it. Capacity is 30 divided by 6.5 = 4.62 cubic metres per minute, or 76.9 litres per second of free air.
Step 2. The leak register. An illustrative result from a site not surveyed since last vintage: three 3 mm equivalent leaks at 7 bar gauge, across press couplings, a valve manifold and the bottling line. Our leak model, a choked-orifice calculation at 7 bar gauge with a discharge coefficient of 0.65, puts a single 3 mm leak at 7.31 L/s and 2.85 kW, so three give 21.93 L/s and 8.55 kW, or 28.5 per cent of capacity. That sits at the top of the band the US Department of Energy describes when it says leaks can sometimes waste 20 to 30 per cent of a compressor’s output, so treat it as a bad, long-unaudited site, not a forecast for yours.
Step 3. Turning flow into dollars, conditionally. Those 8.55 kW are the value of the air at the assumed full-load specific power, not 8.55 kW guaranteed off your meter. What you avoid turns on how the machine answers a drop in demand, whether load/unload, start/stop, modulation or variable speed; on load/unload an unloaded compressor still draws power. For illustration only, assuming controls convert avoided flow proportionally into reduced input power: 8.55 kW x 2,050 hours = 17,528 kWh, at $0.30 that is $5,258 a year and about 10.9 tonnes of CO2-e. By season, roughly $2,565 across vintage and roughly $2,690 outside it. About half of that buys air that never went near fruit.
Step 4. What repair gets back, and what it costs. Do not model to zero; leaks return, and a few sit on machines you cannot isolate while fruit is coming in. Well-maintained systems with regular surveys hold leak rates below 5 to 10 per cent. Take 10 per cent of 76.9 L/s as the residual: 7.69 L/s, or 3.0 kW, 6,150 kWh and $1,845 a year, so $5,258 less $1,845 leaves about $3,410 a year recovered. Our published band for a professional ultrasonic survey is $1,500 to $8,000 depending on plant size and scope, with smaller single-compressor or single-shift sites at $1,500 to $5,000. For this worked example we model a single-compressor winery toward the lower band and, with fittings and labour, call the program $2,000 to $6,000. That is a modelling assumption, not a quote; get a site quotation and put it in place of that band before the payback below carries any weight. Against $3,410 a year that is roughly 7 to 21 months, and real payback has to be built from your measured or control-specific kW reduction.
Two cheap levers sit beside it. Every unnecessary bar of header pressure costs compressor energy and makes every leak flow harder, and that one is a commissioning visit. And a compressor turns nearly all the electricity you feed it into heat. The US Department of Energy Sourcebook puts the recoverable share at up to 90 per cent of the heat of compression, which is a different basis from input power, so don’t read it as 90 per cent of your bill coming back. Plenty of factories have nowhere to put the heat anyway. A winery has barrel washing and CIP.
Air quality where air reaches wine or a product-contact surface
The standard won’t pick a class for you. Your process risk does. ISO 8573-1 is a shared vocabulary for specifying and auditing three contaminants, particles, water and oil, each carrying its own numbered class. It is not a regulation, it names no class for food or beverage contact, and holding a class number is not proof of food-safety compliance. Viable microbiological contamination falls outside those classes and is measured under ISO 8573-7 in colony-forming units per cubic metre. In Australia the legal baseline is the Food Standards Code as applied and enforced in your state or territory, which makes this a per-state, per-scheme question rather than one national answer. A HACCP-based food safety program may be demanded by your jurisdiction, a customer, or a certification scheme such as SQF or ISO 22000; it is not a universal national legal requirement for every winery.
So rank your air points by consequence. A press membrane, a diaphragm pump on juice, a bottle rinser, and air clearing a filter housing or a cleaned transfer line are where a purity failure reaches wine or a product-contact surface. An air knife drying sealed bottles before capsuling blows across the outside of closed glass, not exposed wine, so it sits below the direct product-contact tier unless your line runs it against open product. Actuator air on a valve that never vents into product is a different risk class. Specify to the point, not to the header.
Book it in October, not in March
Between October and December the press is idle, the cellar is quiet, and a fitter can isolate a line without anyone shouting. By late January that window has closed. By March nobody is letting a contractor near the press.
Five things worth doing first. None of them cost you anything but a bit of time, and they change what you get quoted.
- Note the compressor nameplate: kW, model, year and control type. Control type decides whether any of the savings above are real for you.
- Log genuine running hours across one vintage week and one off-season week. Estimated hours produce estimated business cases.
- Identify your header set point, then the highest pressure any single machine truly requires. A wide gap between those two is money.
- Establish whether your press carries its own onboard compressor. One current manufacturer specification lists a vane compressor for the high-pressure stage plus a separate fan for low-pressure membrane inflation, quoted at up to 0.2 bar. If the press is self-contained, sizing the house machine around it is a mistake.
- Write down every point where air can reach wine, juice, must or a product-contact surface. That list, not a catalogue, sets your air-quality spec.
Frequently asked questions
When should a winery book a compressed air assessment?
October to December. You want leak repairs, the set point and any reliability work closed out before the first fruit arrives, because from February nobody will release the press for a fitter and a breakdown becomes a fruit-quality problem. Booking pre-vintage also buys a clean off-season demand profile, which is what decides whether new plant is worth the capital.
Can compressed air replace nitrogen or CO2 in the cellar?
No. Air is roughly 21 per cent oxygen, and blanketing, purging and moving finished wine all exist to keep oxygen away from wine, while sparging removes dissolved oxygen from it or adjusts its dissolved carbon dioxide. AWRI names nitrogen and carbon dioxide as the main sparging gases, with argon occasionally used for headspace blanketing, food grade at 99.5 per cent purity. Air handles the mechanical side: presses, actuators, pumps and packaging motion. Compressed air can feed an on-site nitrogen generator, which separates nitrogen out of the air stream, but that is a separate machine with its own purity specification rather than shop air doing the job.
What ISO 8573-1 class does winery air need?
None is prescribed, and a supplier who names a class before asking what your air touches is guessing. ISO 8573-1 classifies particles, water and oil, and names no class for food or beverage production; microbiological contamination is handled separately under ISO 8573-7. In Australia the legal baseline is the Food Standards Code as applied and enforced in your state or territory, and a HACCP-based program may be required by a jurisdiction, a customer or a certification scheme without being a universal national requirement.
What do leaks actually cost a winery?
On the illustrative case above, about $5,258 of electricity a year and roughly 10.9 tonnes of CO2-e, with about $2,690 of it burning outside vintage. Pulling the leak load to 10 per cent returns about $3,410 a year against a program of roughly $2,000 to $6,000, so payback lands around 7 to 21 months. All of that assumes your controls convert avoided flow proportionally into reduced power.
Is a variable speed compressor worth it for a winery?
Often, because the seasonal turndown is severe, but not before you have logged data. The screening input is how far and how often demand swings, and a winery swings hard between a February press cycle and an August Tuesday. The answer also turns on the efficiency curves of the machines you are comparing, your set point and control band, and how any second compressor is sequenced. A brochure promising a fixed percentage before it has seen your data is a sales document.
Does my pneumatic press run off the house compressor?
Check, because plenty do not. Some presses are self-contained: one current specification, the VP Standard membrane press listed by Vitikit, carries an onboard compressor for the high-pressure stage plus a separate fan for low-pressure inflation, quoted at up to 0.2 bar. Others run from house air, so check the manual for your exact press before you count press duty in or out. If your press is self-contained then your house machine is not sized by the press, and your vintage peak sits elsewhere: valve actuation, diaphragm pumps and blow-down. That’s a common and expensive sizing error in seasonal plants.
Get an independent pre-vintage assessment
CAS does not sell, install or service compressors. Describe your winery and what you need from a leak survey, energy audit or system assessment. We can review your brief and discuss a suitable Australian provider if one is available. We ask for your written permission before an introduction and explain any referral payment arrangement. There is no cost to enquire.
Send us your annual crush in tonnes, your compressor kW, your vintage and off-season hours and your postcode. Book it before vintage.
Form not loading? Email us directly at byron@compressedairsolutions.com.au with your enquiry.