Maximum demand and EV charging: what a demand cap is worth

A demand cap is a controller setting that keeps a site's total power below a chosen ceiling by slowing EV charging when the building is busy. Its value comes from network demand charges, which can bill one 15- or 30-minute peak for a whole year, and from deferring a supply upgrade.

This guide is for commercial, industrial, fleet and building-level supplies on business network tariffs. To test your own site, use the EV charging capacity calculator.

What does "maximum demand" mean?#

"Maximum demand" means two different things in Australian electrical work. Design maximum demand is the peak load an electrical designer determines under AS/NZS 3000 to size the supply, mains and switchboards. Network maximum demand is the highest interval reading on a business electricity bill, which demand and capacity charges price. EV charging affects both.

Design maximum demand decides whether the supply is big enough. AS/NZS 3000:2018 clause 2.2.2 lets the designer determine it by calculation, assessment, measurement or limitation [1]. The Electric Vehicle Council (EVC) notes that a load-management system can ensure the "definite duty cycle" the assessment method relies on, with limitation devices as backup [1]. The designer chooses the method and carries the responsibility.

Network maximum demand is what the bill prices: the highest power your site draws in a metering interval, not the energy you use. A well-set cap serves both, keeping EV load inside the design limit and inside the billed peak you choose.

What is the difference between kW and kVA demand?#

kW measures real power, the part that does work. kVA measures apparent power, which also counts the reactive power that motors, transformers and some electronics draw without doing useful work. Power factor is kW divided by kVA. Many Australian business network tariffs bill demand in kVA, so a poor power factor raises the bill.

Powercor's pricing document spells out the arithmetic for a 15-minute interval: kVA = √(kW² + kVAr²), where kW is the kWh used in the interval × 4 and kVAr is the kVArh × 4 [3]. Two consequences follow:

• At a power factor of 0.9, 180 kVA of billed demand is only 162 kW of real power. A controller that measures kW and assumes unity power factor underestimates the billed kVA by 10%.
• Improving power factor lowers kVA demand. Powercor will consider resetting a site's rolling 12-month demand after power factor correction is installed and shows an observable improvement [3].

How do Australian networks measure demand?#

Australian networks measure demand as the highest average power in a fixed interval, usually 15 or 30 minutes, inside a tariff window, and many charges look back over a rolling 12 months. The rules differ by network and tariff and are approved by the Australian Energy Regulator each pricing year, so always check the current document.

Two patterns matter for EV charging:

• Windows overlap arrivals. Ausgrid's 3–9 pm window covers the hours when fleet vehicles return and residents get home; Powercor's 7 am–7 pm window covers a working day.
• The look-back makes one peak expensive. On rolling 12-month charges, one bad evening sets the charge for a year. The billed demand stays until a higher peak replaces it or 12 months pass.

Other networks differ, and definitions change each pricing year. The AER publishes every network's annual pricing proposal and its decision [5].

Why does EV charging create demand peaks?#

EV charging creates demand peaks because charger power adds up and arrivals coincide. Ten 22 kW chargers that start when a fleet returns add 220 kW in one step. If that lands inside the network's window on top of the building's own peak, billed demand jumps, even though the cars may need only modest energy.

The arithmetic:

• Ten 22 kW three-phase chargers (32 A per phase at 400 V) = 10 × 22 kW = 220 kW, if every car can accept 22 kW.
• Twelve 7.4 kW single-phase chargers on vans returning at 5:30 pm = 12 × 7.4 kW = 88.8 kW, inside Ausgrid's 3–9 pm window.
• Interval demand is energy ÷ interval length. Ten 22 kW chargers running for a full 15-minute interval use 55 kWh, and 55 kWh × 4 = 220 kW. A 10-minute burst inside a 30-minute interval uses 36.7 kWh, adding 73.3 kW of half-hour demand.

Long intervals dilute short bursts, but charging sessions are rarely short: a car taking 30 kWh at 22 kW charges for about 82 minutes. The energy is not the problem; the coincidence is. If ten cars each need 30 kWh, that is 300 kWh, and spread evenly over a 10-hour night it averages 30 kW.

What does a demand cap do?#

A demand cap keeps a site's total power below a chosen ceiling by measuring the incoming supply and reducing EV charging current whenever the building's own load rises. The building's loads are never touched; the chargers get whatever headroom is left. There are two kinds: an instantaneous cap and a rolling-average cap over a window.

Instantaneous cap#

An instantaneous cap holds site power below the ceiling at every control step. It is simple and conservative: a brief spike from a lift or compressor immediately squeezes the chargers, even though a few seconds barely move a 15- or 30-minute average.

Rolling-average cap#

A rolling-average cap holds the average over a sliding window, such as 15 or 30 minutes, below the ceiling. Chargers can use dips in building load, and brief spikes are absorbed. In principle, if every sliding window of the network's interval length stays under the cap, each fixed network interval does too. In practice the controller's samples, meter and clock differ from the network meter's, so keep a margin.

Either way, a cap governs only EV charging. If the building alone exceeds the cap, chargers fall to their minimum or pause, and the peak happens anyway.

How do you set a kW cap when the network bills kVA?#

To set a kW demand cap for a kVA-billed tariff, convert the kVA target to kW using the worst power factor the site sees in the tariff window, then subtract a margin for measurement error and control response. The formula is: kW cap = kVA target × worst power factor × (1 − margin).

Hypothetical example, not a design or a quote. A depot on a kVA capacity tariff has a building peak of 180 kVA in the tariff window, at a power factor of 0.9 (162 kW). It plans ten 22 kW chargers. The building load is at power factor 0.9 and the EV load is near unity (1.0).

1. Unmanaged worst case (every car drawing 22 kW). Building: 180 kVA at 0.9 is 162 kW and √(180² − 162²) = 78.46 kVAr. EVs: 220 kW, 0 kVAr. Site: 382 kW and 78.46 kVAr, so √(382² + 78.46²) = √152,080 ≈ 390 kVA
2. Chosen target: 230 kVA (the 180 kVA building peak plus 50 kVA for EVs)
3. kW cap = 230 kVA × 0.9 × (1 − 0.05) = 196.65 kW, so set 196 kW
4. Check, conservatively at 0.9 for the whole site: 196 kW ÷ 0.9 = 217.8 kVA, under the 230 kVA target (with EVs near unity it is √(196² + 78.46²) ≈ 211 kVA)
5. EV allowance at the building's peak = 196 kW − 162 kW = 34 kW

Site demand at the building's peakIllustrative

Building loadEV chargingAvoided by the capkVA target

Then check energy, not just power. If the building drops to 100 kW overnight, the same cap leaves 96 kW for EVs, or 864 kWh over 9 hours.

Where does the power factor come from? The network's meter records kVArh to compute kVA, as Powercor's formula shows [3], so ask your retailer or metering provider for interval kW and kVAr. A power analyser at the incomer gives the same picture in real time.

What is a demand cap worth?#

A demand cap is worth the demand charge it avoids plus the supply upgrade it defers. The first part is arithmetic from your own tariff: demand rate × kVA avoided × 12 months. The second depends on your site and network and is often larger. We publish no "typical" rate: none holds across networks.

Annual demand-charge value = demand rate ($/kVA/month) × kVA avoided × 12. For tariffs quoted in cents per kVA per day, convert first: $/kVA/month = c/kVA/day × 365 ÷ 12 ÷ 100.

In the hypothetical depot, the cap avoids 390 − 230 = 160 kVA. At a hypothetical $10.00 per kVA per month (about 32.9 c/kVA/day), that is 160 × $10.00 × 12 = $19,200 a year. Take your real rate from your network tariff or bill, not from this example.

The look-back cuts both ways: a cap that holds avoids 12 months of charges, and one unmanaged evening costs 12 months.

The larger prize is often the supply. If a cap keeps design maximum demand within the existing connection, the site can defer or avoid new consumer mains, a larger main switchboard or network augmentation. Those costs depend on the site and network, so get a quote rather than trusting a rule of thumb. The cap has a cost too: when the building is busy, EVs charge slower, so size it so vehicles still get the energy they need.

What does a demand-management system need?#

A demand-management system for EV charging needs a meter at the incoming supply that reports voltage, current and power per phase, a control loop much faster than the network's demand interval, and documented fail-safe behaviour for when the meter, controller or a charger misbehaves. Chargers must accept remote power limits, ideally through an open protocol.

• A meter at the incomer. It should see the whole site, as close to the network's point of supply as practical, per phase. Without voltage, a controller cannot turn measured current into kW. A meter that also reports power factor helps you set and check a kVA-based cap.
• A fast loop. Seconds, not minutes, so a rising building load is absorbed well inside a 15-minute interval.
• Chargers that take limits. The OCPP Smart Charging profile lets one controller manage mixed brands; see hardware-agnostic EV load management. For a building management system, see integrating EV charging with your BMS.
• Fail-safe behaviour. When the meter goes silent, the controller stops or a charger ignores its limit, there should be a safe default current, an alarm and a record of what happened.
• Protection underneath. Breakers and the design maximum demand remain the hard limits. The cap sits on top.

Where Nohmad Edge CSMS fits#

Nohmad Edge CSMS is an on-site EV charging management system that can enforce a demand cap on a site's incoming supply. It runs an OCPP 1.6J charge-point management system and a load-management engine on one industrial PC at the site, measures the building's real load, and gives chargers only the headroom under the cap.

• Two demand caps on the incomer. An instantaneous kW cap and a rolling-average kW cap with a window set in minutes. Caps apply to the incoming supply only.
• A meter reporting voltage at the incomer. For example an Algodue UPM209 (built-in preset over Modbus-TCP), or an Algodue UPM309 or any other Modbus-TCP meter through a configurable register map. Meters are polled every second by default.
• kW, not kVA. Caps work in kW, assuming a power factor of about 1, so set the cap with margin below your kVA target using the method above. Edge CSMS does not align caps to the network's fixed 15- or 30-minute intervals or produce demand-charge reports.
• Measured, cascading, per phase. Under the cap, each board's limit minus the measured non-EV load is shared per phase where meters and charger phase mappings exist (otherwise on the worst phase), re-allocated every 10 seconds through standard OCPP 1.6J Smart Charging profiles.
• Fails safe, on record. If the incomer meter goes stale (60 s by default), active chargers are set to their configured safe current and a critical alarm is raised, with alarm webhooks to your service desk. Every allocation is logged and can be replayed exactly. Load management sits on top of your protection, never instead of it.
• Runs on site. No internet connection is needed, and data stays on the box.

Edge CSMS is available for early sites by enquiry. See commercial and industrial sites, fleet depots and workshops, apartments and strata or the Edge CSMS overview.

Sources#

These public documents support the facts above; all were checked on 23 October 2026. Network tariff definitions change every pricing year, so confirm the current version for your network and tariff, and take rates from your own bill or tariff schedule rather than from any example here.

1. Electric Vehicle Council, Determining maximum demand for an apartment complex with EV chargers, March 2024 (quotes AS/NZS 3000:2018 clause 2.2.2).
2. Ausgrid, 2026–27 Pricing Proposal, Statement of compliance, 31 March 2026, published by the AER.
3. Powercor, 2026/27 Pricing, 7 May 2026, published by the AER (section 2.5, kVA demand).
4. SA Power Networks, Annual Pricing Proposal Overview 2026–27, April 2026, published by the AER.
5. Australian Energy Regulator, Pricing proposals.