- Author
- Byron Raal, CAS Founder-Editor About the author
- Checked against
- ISO 1217
- Date last checked
- 29 August 2026
Demand profiling means measuring how much compressed air your plant actually uses, minute by minute, across a normal production cycle, instead of guessing from the compressor’s nameplate. Data loggers on the system, capturing flow, pressure and compressor power for one to two weeks on a single-shift site and two to three weeks on multi-shift, produce a demand profile: the base load, the shift pattern, the peaks that cause pressure drops, and the off-hours flow, which on a site with no genuine 24/7 loads is dominated by leakage. The profile separates the two. It answers the questions that matter before any money is spent: is the compressor over- or under-sized, would a variable speed machine or a bigger receiver fix the pressure dips, what’s the leak load really costing, and where should the pressure set-point sit. It’s the first step of any honest energy audit, and it’s why audits done without measurement are quotes, not assessments.
What gets measured, and for how long
A demand profile runs on three logged signals, and all three matter. Drop any one and you’re back to guessing.
Demand profiling inputs
Three signals on one timeline, logged long enough
Drop any one of the three and you are back to guessing. Each tile gives the signal and the resolution it needs.
Signal 1
-
Not the nameplate number: the meter sees the whole package. Atlas Copco’s GA37 pairs a 37.3 kW motor with 43.3 kW package input, about 16 per cent higher. Log 15-minute intervals as the audit baseline.kW
Compressor electrical input
Signal 2
-
Sampled fast enough for the transient: for dip diagnosis that means sampling of about a second. A dip that trips a packaging line lasts seconds; a weekly gauge reading has next to no chance of catching it.bar
Header pressure
Signal 3
-
Flow at the compressor discharge is compressor supply. Main-header or downstream flow is what represents plant demand. Match the measurement point to the question you are asking.L/s
Flow, on a calibrated meter
One to two weeks for a single-shift site, two to three weeks for multi-shift operations. Cover at least one representative operating cycle plus a genuine low-load or non-production window: that floor is where leaks show.
Source: CAS demand profiling guide; GA37 figures from the Atlas Copco CAGI data sheet.
Compressor electrical input, in kW. Not the nameplate number. The kW in a compressor’s name is the nominal motor rating, and your meter sees the whole package: motor, cooling fan, drive and control losses together. Total package input runs higher, and by how much varies by model, so the CAGI performance data sheet is the reference: Atlas Copco’s GA37, for example, pairs a 37.3 kW nominal motor with a 43.3 kW package input at full load, about 16 per cent higher. Our energy audit guide works to 15-minute logged intervals on electrical input as the audit baseline.
Header pressure. Sampled fast enough for the transient you’re investigating, which for dip diagnosis means sampling of about a second. A pressure dip that trips a packaging line lasts seconds; a weekly spot reading with a gauge has next to no chance of catching it. One-minute stored records are only acceptable where the logger samples faster internally and preserves the minimum and maximum excursions within each interval.
Flow. Measure it with a calibrated flow meter, and be clear about what the meter is telling you. Flow at the compressor discharge is compressor supply, what the machines deliver, and on most sites it’s the practical place to put the meter. Plant demand is what your production equipment draws, and where storage sits between the compressors and the plant, main-header or downstream flow is the measurement that represents it. During a short burst the receiver gives up part of the air, so discharge flow on its own understates the instantaneous peak. Match the measurement point to the question. Receiver sizing needs plant-demand flow and the compressor’s contribution for the same event, because the receiver is charging and discharging between the two. CAGI-comparable specific power needs compressor or package kW paired with flow at compressor discharge. A downstream demand meter and compressor kW can’t simply be read together instant by instant across those charge and discharge transients. Where flow can’t be instrumented, it can be estimated from compressor loaded hours and nameplate free air delivery, but only on fixed-speed or load/unload machines running near their rated loaded output. It doesn’t hold for VSD or modulating compressors, whose delivered flow varies continuously while the machine runs. Even where it does hold, it gives you volume totals, not a time-resolved demand trace, and the profile should say so.
How long? Long enough to capture a representative cycle: one to two weeks for a single-shift site, two to three weeks for multi-shift operations. Cover at least one representative operating cycle plus a genuine low-load or non-production window, because that floor is where your leak load identifies itself. On a Monday to Friday plant the weekend is what provides that window, so include one. On a 24/7 plant, a seven-day roster, or a site that takes its downtime midweek, it’s whatever period production actually stops or drops away.
What does a good profile look like? Picture a week of flow data on one chart. Overnight, a flat low line that shouldn’t be there but is. At 6 am, a ramp as production starts. Through the day, a sawtooth of peaks riding on a steady base, with a dip at lunch. Friday night, everything shuts down and the trace falls, but not to zero. The gap between that weekend floor and zero is the number this whole page is about.
What the profile reveals
Base load versus peaks. The profile splits your demand into the steady load your production draws all shift and the short bursts stacked on top. That split is the sizing question in one picture: base load determines the machine that runs all day, and peak behaviour determines whether you need storage, a trim machine, or both.
The off-hours signature. Flow at 2 am on a Sunday, on a site with no night processes, isn’t production. It’s leakage, plus whatever genuine 24/7 loads you can name and net off. The profile makes that number visible.
Pressure-band behaviour. The pressure trace tells you when your dips happen and how deep they go. Dips that coincide with short demand bursts point at storage and control settings. A pressure line that sags all shift points at a capacity shortfall or a distribution restriction, and separating those two takes more than system totals: you need per-compressor state, or load/unload and power data, to tell whether the machines were already flat out. A single header-pressure point can flag abnormal pressure behaviour, but on its own it can’t separate a distribution restriction from inadequate supply or control, and it can’t locate a restriction either. That takes supply-state data alongside pressure logged at more than one point across the distribution system. Buying a bigger compressor to fix a storage problem is an expensive way to keep the wrong problem.
Sizing truth. With measured flow and measured kW on the same timeline you can compute specific power, the kW your system burns per litre a second of free air delivered. Our energy audit guide publishes the benchmark ranges by machine class, and cost per cfm converts them between kW per m³/min, kW per 100 cfm and kW per L/s; if your measured figure sits well above the achievable benchmark, the profile has just quantified your gap to best practice before anyone opens a catalogue. Specific power is a useful system-efficiency KPI, and you read it alongside demand, pressure, controls and leakage rather than on its own.
Independent guidance backs the measurement-first position. The Australian Government’s energy.gov.au compressed air guidance lists metering and monitoring first among its quick wins and states that metering data is what lets you develop demand profiles of the system; the same page carries the Australian Alliance for Energy Productivity’s (A2EP) compressed air systems guide as further reading.
The decisions a real profile unlocks
Each feature of the chart maps to a capital or operating decision, and without the chart you’re making that decision on a salesperson’s assurance.
VSD or fixed speed. A variable speed drive matches motor speed to demand, so it earns its premium on sites where demand genuinely swings, and buys little on a site running flat out at steady load. That’s why a blanket “VSD saves 35 per cent” claim deserves a raised eyebrow: demand variability is the major screening input, and that’s exactly what the logger measures before anyone quotes you. What you actually save also depends on the full-load and part-load efficiency curves of the machines you’re comparing, your pressure settings and control band, and how your existing compressors are sequenced. The profile tells you whether a VSD compressor is your answer or an expensive irrelevance.
Receiver sizing. The profile gives you two of the inputs straight off the trace: the excess burst flow above what the compressor supplies, and how long the burst lasts. Correct receiver sizing combines those with the usable pressure band between your normal operating pressure and the minimum your processes accept, and with how your compressor and its controls respond during the event. Sized from measured inputs, receivers come out right. Sized from folklore, they come out wrong by up to a factor of ten, and our energy audit guide shows the arithmetic. The air receiver guide covers selection once the profile has given you real numbers to size from.
Set-point reduction. The pressure trace shows the pressure minima that actually occurred while you were logging, and how the system behaves as pressure falls. It doesn’t tell you what your processes could tolerate. That allowable minimum has to come from your process and equipment requirements, then be verified by a controlled trial at the lower set-point. Reducing the set-point by 1 bar typically saves 6 to 7 per cent of compressor energy, per our energy audit guide, and lower pressure shrinks every leak on site at the same time. The profile turns “can we lower the pressure?” from an argument into something you can test.
Staging and sequencing. On multi-compressor sites, sequencing problems look like machines fighting each other: two compressors trading load where one would do, or a large machine idling to cover a demand a small one should carry. Seeing that takes per-compressor logging, load state and power on each machine, because one system total can’t separate them, so specify it up front. Sequencing fixes cost little, and per-machine data is the evidence they’re needed.
System design. If the profile points at a distribution restriction, or shows demand that has outgrown the installed system, that’s a system design conversation, and the profile is the first document a good designer will ask for.
The off-hours leak signature, in dollars
Here’s the worked example, using the constants from our 2026 leak-cost report.
Say your profile shows weekday peak demand of 50 L/s of free air, and the overnight and weekend trace sits flat at 10 L/s, which is 20 per cent of peak. You walk the site list and net off everything downstream of your meter that uses air without being a leak. No air-operated process runs overnight, but there’s one small instrument load you can name, and if your meter sits at compressor discharge it also sees the air treatment. A heatless desiccant dryer diverts a meaningful share of rated flow as purge air, enough on a system this size to account for much of a 10 L/s residual on its own, and timed condensate drains and continuous instrument air add to it. Quantify those, take them out, and call what’s left, roughly 10 L/s here, the leakage, now visible on a chart instead of hiding inside the power bill. Corroborate it with an isolation test or an ultrasonic survey before you spend money on the figure, because a residual that still contains purge air is a repair case built on air you were always going to use.
Costing it uses the leak-cost report’s model: flow, times hours, times specific power of about 0.11 kWh per cubic metre of free air, times tariff. At the report’s indicative rate of $0.30 per kWh (an illustrative 2025-26 figure, not a national benchmark; your contract rate might be $0.22 or $0.38, so use your actual number):
- 10 L/s is 36 cubic metres of free air an hour, which takes about 4 kWh to compress, which is about $1.20 an hour.
- Across the report’s 4,000 to 8,000 operating hours a year, that’s roughly $4,800 to $9,500 a year of electricity spent compressing air that does nothing.
Cross-check it against the report’s per-leak table: a single 3 mm leak at 7 bar gauge discharges about 7.3 L/s and costs roughly $3,500 to $6,900 a year on the same basis. That flow is a choked-orifice calculation in the form given by ISO 6358, using a discharge coefficient of 0.65 and referred to ISO 1217 Annex C reference conditions of 20 degrees Celsius, 100 kPa absolute and 0 per cent relative humidity. That is the same basis our leak cost calculator discloses; ISO 6358 supplies the flow form, not the coefficient. Our 10 L/s off-hours load is about 1.4 times that single leak, and 1.4 times the cost range lands on the same answer. Same model, two routes, one number. That’s what a reproducible figure looks like, and it’s the standard you should hold any auditor’s numbers to.
Two honesty notes before that number goes in a business case. First, netting off genuine 24/7 loads isn’t optional; a profile that books legitimate instrument air as leakage is padding the case. Second, leakage is a flow figure, and it only becomes a bill saving when your control system converts the reduced demand into reduced power by unloading, slowing or switching off compressors after the repairs. Fix the leaks and the bill falls only as far as your compressor and its controls cut input power at the lower demand. Well-set start/stop, load/unload and VSD systems do that automatically; poorly sequenced or modulating systems may need control changes to capture the full saving.
Put your own leak sizes and tariff into the leak cost calculator and get the per-leak dollar figure for your repair register. And if the profile is telling you your machine runs long hours at heavy load, the running costs guide shows what those hours cost by compressor size.
Profiling versus the walk-through audit
Plenty of things get sold as compressed air audits. The distinction that matters is simple: was anything logged?
A walk-through audit is a person, a clipboard and an opinion. It can spot open blowing, hear the worst leaks and read your set-point off the gauge, and that has some value. What it cannot do is tell you your base load, your peak behaviour, your off-hours flow or your specific power, because none of those exist without a logger. When a walk-through ends in a proposal to buy equipment, you’re reading a quote dressed as an assessment. A vendor’s free audit isn’t free if it steers the answer before any data is collected; an independent audit optimises for your lowest-cost outcome regardless of brand, which is the whole case for measurement-first auditing.
Measured baselines also underpin some project-based incentive pathways, while other scheme activities are calculated on deemed, benchmark or prescribed methods instead. Some Australian schemes reward measured, verified savings, on a per-state, per-scheme basis. In Victoria, the Victorian Energy Upgrades program runs a project-based activities pathway in which savings are measured before and after the project under the Essential Services Commission’s published measurement and verification method, a public methodology you can read. A demand profile can contribute useful baseline evidence to a VEU project like that, but the formal baseline and the certificate evidence must meet the Commission’s applicable measurement and verification requirements. In New South Wales, the Energy Savings Scheme is a continuous certificate scheme administered by IPART, with new scheme rules that commenced on 1 July 2026, and energy.gov.au’s business equipment upgrades listing names compressed air systems among the upgrades available under it. Eligibility, rules and incentive values are scheme-specific and change; our grants and incentives map tracks each scheme’s current status with the date we last verified it. The practical point stands on its own: a logged demand profile is evidence you can take to a scheme, a lender or your own finance team. A walk-through opinion is not.
Frequently asked questions
What is compressed air demand profiling?
Demand profiling is logging a compressed air system’s flow, pressure and compressor power over a representative production period, typically one to two weeks on a single-shift site and two to three weeks on multi-shift, to measure how much air the plant actually uses and when. The result is a time-series profile showing base load, peak demand, pressure behaviour and off-hours consumption, which is the evidence base for sizing, control and leak-repair decisions.
How long does a demand profile need to run?
Long enough to capture a full representative cycle of your operation: one to two weeks for a single-shift site and two to three weeks for multi-shift operations. The logging period must also cover a genuine low-load or non-production window, because the off-hours flow floor is the clearest view of the leak load. On a Monday to Friday plant that means including a full weekend; on a 24/7 plant or a seven-day roster it means capturing whatever period production actually stops or drops away.
What instruments does demand profiling use?
Three logged measurements: compressor electrical input in kW from a power logger, header pressure from a pressure transducer logged at short intervals, and system flow from a calibrated flow meter, sited at the compressor discharge for compressor supply or in the main header where you need plant demand. Ask any provider what instrumentation they’ll use and for its calibration certificate. If the answer is a clipboard, it isn’t a demand profile.
Can off-hours air consumption tell me my leak load?
Largely, yes. Air flowing at 2 am on a site with no overnight production is leakage plus any genuine 24/7 loads such as instrument air, and those known loads can be identified and netted off. The remaining off-hours flow is a measured estimate of the leak load. Converting that flow to dollars uses compression energy of about 0.11 kWh per cubic metre of free air multiplied by your electricity tariff.
Do I need a demand profile before buying a VSD compressor?
It’s the honest way to decide. A variable speed drive saves energy by matching motor speed to a varying demand, so how much your demand actually varies is the major screening input. A logged profile shows the swing directly. What you actually save also depends on the full-load and part-load efficiency curves of the machines you’re comparing, your pressure settings and control band, and how your existing compressors are sequenced. On a site with steady, near-full load a VSD delivers little, and on a site with big swings it can close most of the part-load gap. Buy the data before the drive.
Is a demand profile the same as an energy audit?
No, it’s the foundation of one. A full compressed air energy audit adds a leak survey, an end-use assessment, specific power benchmarking and a costed opportunity register on top of the logged profile. The profile is the measurement layer that makes the rest of the audit evidence instead of opinion.
Get a measured demand profile, not a guess
A useful demand profile depends on measurements that represent your operation. Tell us your compressor setup, shift pattern, location and what you need to establish. CAS does not sell compressors or perform the logging. 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.
Tell us your installed compressor kW, your shift pattern and your postcode. We can review your requirements and discuss a provider if a suitable one is available. We ask for written permission before an introduction.
Form not loading? Email us directly at byron@compressedairsolutions.com.au with your enquiry.