Start with the question, not the battery

Sizing a battery for a charging station is not a matter of picking a round number of kilowatt-hours. It starts with four questions: How many charging ports, at what power, will run simultaneously? How much continuous power can the grid supply? What is the busiest realistic day? And what tariff structure applies? Get these four answers and the battery size almost follows.

The DOE First Hour rule

The US Department of Energy's battery-buffered fast-charging guide offers a simple starting rule: usable battery kilowatt-hours should be at least 150 kWh times the number of ports, minus the grid kilowatts available times one hour. The logic is that the battery must be able to cover one hour of worst-case, all-ports-at-full-power demand that the grid cannot supply. A four-port 150 kW site on a 200 kW grid therefore needs at least 150*4 - 200*1 = 400 kWh usable.

The NREL buffer recommendation

National Renewable Energy Laboratory research suggests that grid-constrained sites should plan for a two-to-three-hour buffer, raising the figure for the same four-port, 150 kW example to 800-1,200 kWh. For sites where the grid is adequate and the battery only needs to shave demand charges, a one-hour buffer is usually sufficient.

Worked examples

  • Highway corridor: 4 x 350 kW (1,400 kW peak) on a 400 kW grid - battery 800-1,200 kWh for a 0.6-0.9 hour buffer, demand-charge cut 65-80%.
  • Fleet depot: 8 x 50 kW (400 kW peak) on a 300 kW grid - battery 300-500 kWh, buffer 0.75-1.25 hours, demand-charge cut 50-70%.
  • Urban plaza: 6 x 180 kW (1,080 kW peak) on a 350 kW grid - battery 700-1,000 kWh, driven mainly by transformer deferral.
  • Retail DCFC: 4 x 100 kW (400 kW peak) on a 150-200 kW grid - battery 400-600 kWh, demand-charge cut 55-70%.

Chemistry, power and efficiency

LFP chemistry dominates charging-station batteries because of its cycle life (often 6,000+ cycles), thermal stability and lower cost, even though it is bulkier than NMC. The battery's power-to-energy ratio (its C-rate) matters as much as its capacity: a battery sized only for energy may not discharge fast enough to help a 400 kW session. Most suppliers quote a system round-trip efficiency of about 90% and a depth of discharge of 90%, so usable energy is roughly 80% of nameplate.

Data you need before you buy

Collect at least a year of session data if the site exists, or a demand model if it is new. You need the distribution of simultaneous sessions, average energy per session, dwell times, and the 15-minute peak profile. Add the local tariff: energy price by time-of-use, demand charge per kW, and any capacity or connection fee. Model the battery against these and iterate until payback meets your hurdle rate.

Mistakes to avoid

The most common errors are sizing for nameplate power rather than usable energy, ignoring round-trip losses, assuming 100% simultaneous use when real sites peak at around 65%, and forgetting that a battery adds cooling, fire suppression and space requirements. Oversizing by 20% for growth is prudent; oversizing by 100% usually kills the business case.

Bottom line

Size from the grid gap and the tariff, not from a catalogue. Use the DOE rule as a sanity check, NREL's buffer guidance for constrained sites, and your own session data for the final number. The right battery is the smallest one that covers your worst realistic hour and still pays back.