What Is Peak Shaving?

Peak shaving is the practice of discharging stored energy during a site's highest-demand periods so the peak the utility measures — and bills for — never reaches its natural height. The load still gets served; the battery, not the grid, serves the top slice of it.

That single sentence hides the part that decides whether a project pays: a peak-shaving battery has to be sized for two things at once, and most disappointing installations got one of them right and the other wrong.

What peak shaving actually does

A commercial site's load is not flat. A factory might sit at 120 kW most of the day and touch 200 kW for twenty minutes when a compressor and a production line start together. Peak shaving puts a battery in parallel with that load. An energy management system watches the incoming power in real time, and when the site's draw approaches a set threshold, the battery discharges the difference — covering the top 50 kW so the meter never records more than 150 kW.

Load curve showing a 200 kW site peak capped at a 150 kW threshold, with battery discharge of 50 kW filling the clipped slice

The battery recharges later, during off-peak hours when the site's load is low and, on a time-of-use tariff, when energy is cheapest. Nothing about the site's total consumption changes. What changes is the shape of the demand curve the utility sees.

This is also why peak shaving is sometimes called load shedding in older literature — though load shedding more properly means switching load off, while peak shaving serves the same load from a different source. The load never notices.

The bill line it targets

A commercial and industrial electricity bill has two parts that matter here, and they answer different questions. Energy billing asks how much did you use in total? Capacity and demand billing asks how hard did you pull at your worst moment?

That distinction is the whole game. Peak shaving does almost nothing for the first question and a great deal for the second.

How the second question is priced depends on where you are, and it is worth reading your own bill before modelling anything:

• In Great Britain, network costs reach business consumers through use-of-system charges — distribution (DUoS) and transmission (TNUoS) — rather than a single US-style monthly demand charge. Distribution includes a capacity or availability component billed per kVA of agreed supply capacity, plus unit rates that vary by time band, so both your agreed capacity and when you draw power matter. Your distributor's charging statement is the authoritative document; National Grid publishes its use-of-system charges and SSEN's explainer walks through the components. Transmission charging is set out by NESO's TNUoS pages.
• In Germany, grid fees include a capacity price alongside the energy price, and consumers whose draw pattern helps the network can apply for individually calculated grid fees under §19(2) of the StromNEV ordinance — TenneT documents the mechanism and its notification procedure. Reshaping a peak can therefore change not only what you pay but which tariff regime you qualify for.

The practical consequence: do not model peak shaving on a generic rate. Take the capacity charge and time-band rates off your own invoice. Our demand-charge calculator is built to take those numbers rather than assume them.

Peak shaving vs load shifting

These two get used interchangeably and they are not the same thing. They can run on identical hardware and still pull in opposite directions.

The tension is real. Peak shaving wants reserve held back for a peak that might come at any moment; load shifting wants that same energy spent during the expensive hours. A site with both a steep capacity charge and a wide tariff spread has to co-optimise, and that is a control problem, not a hardware problem.

Solar self-consumption is a third neighbour: it targets the difference between your export price and your import price, and it is driven by when the sun shines rather than by either your peak or the tariff clock.

Why you size for power and duration

This is where undersized systems go wrong. A peak-shaving battery has two independent specifications and they are not interchangeable.

• Power (kW) — how deep a peak the battery can clip. To shave 50 kW off a peak, the inverter must be able to deliver at least 50 kW. A battery with plenty of stored energy but an undersized inverter simply cannot cover the spike.
• Energy (kWh) — how long it can hold that clip. A 50 kW shave sustained for two hours needs roughly 100 kWh of usable energy — and usable is not the same as nameplate, because depth of discharge and round-trip losses both take a slice.

The ratio between the two is the C-rate: a 100 kWh battery discharging at 50 kW runs at 0.5C for two hours. Get the C-rate wrong and you have either an inverter that cannot reach the peak or cells being pushed harder than they are rated for. If you want to work the duration side quickly, the backup-duration calculator does the kW-to-hours arithmetic.

A worked example

Take that mid-sized factory: 120 kW typical daytime load, 200 kW peak when the compressor and production line coincide, peak lasting about two hours across the working day.

To clip 50 kW for two hours the system needs ≥50 kW of power and ~100 kWh of usable energy — call it a 120 kWh nameplate unit at 90% usable, with round-trip efficiency meaning you must store a little more than you intend to deliver.

The saving side is deliberately left as a variable, because writing a rate here would make this page wrong the moment tariffs change:

With 50 kW clipped, every unit of capacity charge on your bill is multiplied by 50. That is the whole sensitivity: the same hardware on a site with a steep capacity charge pays back in a few years and on a site with a shallow one may never pay back at all. The tariff, not the battery, decides whether this project is viable — which is why the first thing to do is put your own numbers into the demand-charge calculator, not to shop for kWh.

Setting the threshold and the peak period

Two questions come up constantly once a system is specified, and they are settings questions rather than hardware ones.

What the peak shaving setting is

The threshold (sometimes labelled the peak shaving setpoint or limit) is the site draw, in kW, above which the battery starts discharging. Set it just below the peak you want to avoid recording — but not so low that the battery is discharging all day.

Sizing the threshold sensibly means looking at a real load profile, not a nameplate: take the highest demand intervals over a representative period, decide which of them you intend to cover, and leave headroom for the load growing. A threshold set aggressively low turns every ordinary afternoon into a discharge event and leaves nothing in reserve for the peak that actually costs you money.

What the peak period is

The peak period is the window in which peaks are worth shaving — determined by two things that do not always agree: the time bands your tariff prices highest, and the hours your own load actually spikes. Where they overlap is where a battery earns most. Where your spike falls outside the expensive band, shaving it still protects your agreed capacity, but the payback is thinner.

Good control blends live measurement with forecasting — learning the site's daily and weekly rhythm so the battery is charged before the window it needs to cover. That forecasting-and-dispatch job is what an energy management system exists to do; without it, peak shaving degrades into a battery that reacts too late to matter.

Where peak shaving pays best

Peak shaving pays where the load is spiky — a high, brief peak sitting well above the average draw. The bigger the gap between peak and average, the more there is to clip and the less energy it takes to clip it.

• Factories with large motors, presses, arc furnaces or compressors that start intermittently and set the monthly peak in minutes
• EV charging depots, where a few simultaneous fast-charging sessions can double site demand
• Cold storage and food processing, where defrost cycles and pull-down loads stack up predictably

A site with a flat, steady load — running near its peak all day — has almost nothing to shave, and a battery there is solving the wrong problem. If your load profile is the spiky kind, the commercial and industrial storage solutions page covers how these systems get deployed.

The failure mode: undersized duration

The failure that ruins peak-shaving projects is not an undersized inverter — that shows up immediately during commissioning. It is undersized duration, and it hides.

Picture the two-hour example with only one hour of usable energy behind it. The battery clips the peak beautifully for an hour, empties, and the site's demand jumps back to 200 kW for the remaining hour. The utility bills on the highest measured demand, so a peak survived for one hour out of two produces the same bill as no battery at all. The system works, the meter reads exactly as before, and the fault looks like a control problem when it is an energy-capacity problem.

This is why duration deserves the same scrutiny as power, and why the daily-cycling assumption matters: peak shaving typically means one cycle every working day, so cycle life at the depth of discharge you actually plan to use — not at a headline figure — is the number to check.

How Hua Power builds for peak shaving

Peak shaving is a sizing problem before it is a hardware problem, so our range is built to be matched to a load profile rather than sold in fixed steps.

Sizing power and duration independently. 17 standardised C&I SKUs from 64 kWh / 30 kW up to 1.2 MWh / 500 kW — enough granularity to match the department-level peak you are actually clipping instead of rounding up to the next big unit. Cabinet formats are covered on the energy storage cabinet page and larger deployments on the containerised systems page.

The control layer. Our in-house Visual EMS handles peak forecasting, threshold setting and real-time dispatch, and co-optimises peak shaving against tariff arbitrage where a site's tariff makes both worth chasing.

Cells and cooling for daily cycling. LFP cells rated for ≥6,000 cycles to 80% state of health, with a 90% depth-of-discharge rating, because peak shaving usually means one cycle every working day for the life of the system rather than the occasional deep discharge of a backup application.

Peak shaving is the clearest business case in stationary storage — but only if the battery holds through the whole peak, and only if the tariff underneath it is worth shaving. Both are checkable before you buy.

FAQ

What is meant by peak shaving?

Discharging stored energy during a site's highest-demand periods so the peak the utility measures stays below its natural level. The load is fully served; the battery covers the top slice.

What is the peak shaving process?

An energy management system meters the site's incoming power continuously. When draw approaches a preset threshold, the battery discharges the difference for as long as the peak lasts, then recharges during off-peak hours.

What is a peak shaving setting?

The threshold, in kW, above which the battery begins discharging. It is set just below the demand level you want to avoid recording, with headroom for load growth — low enough to catch real peaks, high enough not to drain the battery on ordinary days.

What is a peak shaving period?

The window in which peaks are worth shaving: where your tariff's most expensive time bands overlap with the hours your own load actually spikes. Outside that overlap, shaving still protects agreed capacity but returns less.