Peak Shaving EV Charging: How Demand Charges Change ROI

Demand charges can decide whether a charging site is profitable. Peak shaving changes the economics by controlling when power is pulled from the grid.

Many EV charging business cases are built on energy price and utilization. That is only half of the electricity bill. A commercial site may also pay a demand charge based on its highest power interval during the billing period. A single fifteen-minute spike can set that charge for an entire month, even if the site runs at low power for the rest of the time.

Peak shaving means managing the site so its maximum grid import stays below an agreed or economically sensible level. It can be done with charger control alone, with battery storage, or with a combination of both. The right answer depends on the tariff, the load profile, and how much service capacity the site already has.

What a Demand Charge Actually Measures

A demand charge is a price for power capacity rather than energy. The utility records the site load in short intervals, commonly fifteen or thirty minutes, and bills the highest average interval during the month. A charger that briefly draws 400 kW can produce a larger demand charge than a site that consumes more total energy at a steadier 150 kW.

The exact rule varies by utility and tariff. Some sites have a ratchet clause that carries a high seasonal peak into later months. Others measure demand only during defined hours. Some include a separate charge for reactive power or for exceeding a contracted maximum. The site owner should obtain the actual tariff sheet and recent bills before sizing any solution.

This is why a charging site can look profitable on an energy-only model and fail once demand charges are included. The risk is greatest at sites with low to moderate utilization, because the same peak can occur on days when only one or two vehicles arrive together.

Reading the Site Load Profile

Interval meter data is the foundation of a peak-shaving project. A month of fifteen-minute readings is enough to see the shape, but a full year is better when the site has seasonal demand. The analysis should identify the current peak, how often the site approaches it, and how much of the peak comes from chargers rather than from the building.

The next step is to forecast charger demand. A single DC fast charging station can add 80 kW to 240 kW depending on the model and the vehicles connected. If several units are expected to run at once, the combined load may exceed the existing service even if average daily energy is modest. The EV charging station ROI guide covers the revenue side; peak shaving is about protecting the cost side of that model.

The load profile also shows whether peaks are predictable. Fleet depots often have a repeatable morning or evening peak. Public hubs may have peaks driven by traffic, weather, and holidays. Predictable peaks are easier to manage with scheduling. Unpredictable peaks are better handled with real-time control or storage.

Peak Shaving Without a Battery

A site can often reduce demand by limiting total charger power. A controller measures the service import and tells the chargers to reduce output when the site approaches its target. OCPP charging profiles allow the reduction to happen without ending sessions or requiring a driver to restart.

The tradeoff is service quality. If every charger is slowed at the same time, drivers stay longer. A priority policy can avoid that problem by giving more power to vehicles that are nearly complete or to fleets with a fixed departure time. A dynamic power sharing strategy is usually a prerequisite for serious peak management.

• Scheduling moves flexible charging into off-peak hours without reducing peak power
• Power sharing caps total site demand while keeping every charger available at a reduced rate
• Load interlocks prevent charging from starting when the building load is already near the limit
• Storage absorbs the peak while chargers continue at full power

Sizing Storage for a Demand Charge Target

Storage sizing starts with a target. If the service limit is 500 kW and the forecast peak is 650 kW, the battery must cover at least 150 kW while the peak lasts. If the peak lasts thirty minutes, the battery must also provide 75 kWh of usable energy, plus losses and a margin for degradation.

That calculation is only the beginning. The battery must recharge before the next peak, and it may need to cover several peaks in one day. The charging strategy should use periods when the site has spare capacity or when energy is cheaper. A battery that recharges while the building is at its own peak can create a new demand problem.

The inverter rating and battery capacity are separate decisions. A power-heavy system can shave a short spike with a relatively small battery. A longer peak needs more energy capacity and may require a larger enclosure or multiple cabinets. The 261kWh energy storage cabinet is one building block, but the number of cabinets follows from the load analysis.

Control Logic and Measurement Points

The controller must measure power at the same point the utility measures it. If the meter is on the utility side of a transformer and the site controller measures only the low-voltage bus, the two readings may not match because of transformer losses or other loads. The design should identify the billing meter, the protection point, and the control point.

Control should be fast enough to keep a fifteen-minute average below the target, but not so aggressive that charger output oscillates. A deadband prevents constant adjustment. A local controller should retain authority when the cloud connection drops, and the battery dispatch should have a safe fallback when state of charge is low.

The site should also record data for verification. Without interval data before and after installation, the owner cannot prove that demand fell or calculate the actual payback. The monitoring system should log charger power, battery power, grid import, state of charge, and any limit events.

How to Model Payback

A peak-shaving model should separate savings by source. Demand savings come from reducing the billed peak. Energy savings come from shifting consumption to lower-price periods. Service-upgrade avoidance is a capital saving rather than an annual benefit. Resilience has value only if the site actually needs backup operation.

The cost side includes the battery, inverter, controls, installation, protection, fire safety, permitting, and maintenance. Battery degradation reduces usable capacity each year, so the model should not assume the original peak reduction forever. A tollgate or ratchet clause may also reduce savings if another month already set a high peak.

The strongest projects are usually those with a high demand charge, a persistent and measurable peak, and a site that cannot easily reduce power without hurting service. The weakest are sites with a low demand charge or a peak that occurs only once or twice a year, where a modest scheduling policy may capture most of the benefit at a fraction of the cost.