How to Size C&I Battery Storage for Demand

Start with the Right Data

Demand charge management fails when the input data is wrong. The first step is to pull at least twelve months of interval metering records from your utility provider or internal submetering system. The standard interval is five minutes, not hourly. Hourly data averages out short spikes and masks the true peak behavior of the site. You need the raw data to see exactly when and how long the load surges.

Extract the monthly maximum demand value from these records. This is the figure the utility bills against. Do not rely on the annual average. Look at the worst month, usually driven by extreme heat or cold, or a specific production run. Note the time of day and day of the week for these peaks. A facility that peaks at 2 p.m. on weekdays has a very different profile than one that peaks at 9 a.m. on Saturdays. This baseline establishes the actual load curve you are trying to flatten. Without this data, any battery sizing calculation is a guess.

Identify the Demand Target

The demand target is the maximum power level you want the battery to keep the facility under. Most commercial and industrial buyers start by targeting a reduction of 20 to 40 percent off the historical peak. A higher reduction means a larger battery, higher capital cost, and more frequent cycling. The target must be realistic and grounded in the utility rate structure.

If the building has a large HVAC compressor that kicks on at 2 p.m. daily, the battery must be ready to cover that exact spike. If the utility rate has a high tier after 500 kW, and the historical peak hits 650 kW, a target of 500 kW might save money. A target of 300 kW might not, because the cost of the extra battery exceeds the savings. A 100 percent reduction is almost never economically viable for a base load. Set the target based on the utility rate structure and the specific peaks you want to eliminate.

Calculate the Energy Required

Demand is measured in kilowatts, but batteries are sized in kilowatt-hours. To find the energy needed, multiply the demand reduction by the duration of the peak. A 50 kW reduction for 30 minutes equals 25 kilowatt-hours of energy. This is the raw energy requirement.

You must then apply efficiency losses. The inverter, the battery management system, and the power electronics all consume some power during the discharge cycle. A typical round-trip efficiency for modern lithium-ion systems is in the low 80 percent range. This means you need more stored energy than the raw calculation suggests. If you need 25 kWh of raw energy, you must store roughly 31 kWh to account for losses. Ignoring this step leads to an undersized system that will fail to hold the demand target when the load spikes.

Account for Thermal and Cycle Constraints

Battery chemistry and operating temperature affect performance significantly. Lithium-ion packs have lower self-discharge than lead-acid, but they are sensitive to high heat. In a hot warehouse or an unconditioned electrical room, the inverter may derate its output to protect the pack. Check the manufacturer’s temperature curves. If the average ambient temperature exceeds 35 degrees Celsius, you may need to size the battery larger to compensate for derating.

Also consider the number of cycles per year. A facility that peaks daily will cycle the battery 250 to 300 times per year. This is well within standard warranties, but it is a factor in long-term cost. Deep cycling reduces life, while shallow cycling is gentler on the pack. The dispatch strategy you choose will determine the depth of discharge. A shallow cycle might be 20 percent, while a deep cycle could be 80 percent. You must select a strategy that balances energy cost against battery longevity.

Model the Dispatch Strategy

A battery does not just discharge a fixed amount. It must be programmed to respond to the load. The most common strategy is a set-point dispatch. The battery charges when the load is low and discharges when the load approaches the demand target. Model this using a simple spreadsheet or a simulation tool. Plot the load curve and the battery output over a representative day.

Ensure the battery can hit the target before the meter records the peak. The utility often measures demand on a 15-minute window, so the battery must be active for the full duration, not just at the exact peak instant. If the battery only kicks in for two minutes, the 15-minute average will still show the high load. The dispatch logic must anticipate the peak. It should start discharging slightly before the load hits the target, allowing the inverter to ramp up to full power. This requires a clear communication path between the battery management system and the inverter.

Size the Battery Capacity

Now combine the numbers. Take the total energy required from the previous step. Divide by the round-trip efficiency. That is your required stored energy in kilowatt-hours. Convert that to capacity. For lithium-ion, this is often close to the rated kilowatt-hour value. For lead-acid, the capacity is often measured in amp-hours at a specific rate.

You must match the rate. A battery rated for a 2-hour discharge is not the same size as one rated for a 15-minute discharge. Use the manufacturer’s capacity tables. If you are using a 15-minute peak, a 2-hour rated battery will show a much lower usable capacity. This is a common source of undersizing. The Peukert effect means that high discharge rates reduce the effective capacity of the battery. Always verify the capacity at the specific discharge rate your application requires.

Verify with a Load Test

Before finalizing the purchase, run a simulation with the selected battery. Feed the 12 months of historical data into the model. Check the results against the demand target. If the battery misses the target on even a few days, increase the capacity or adjust the dispatch set-point. Do not assume the model is perfect. Field conditions vary. A new HVAC system, a production line change, or a utility rate adjustment can shift the load profile. A conservative margin of 10 to 15 percent is standard practice. This margin protects against data gaps and future changes.

Common Mistakes in Sizing

• Using hourly data. Hourly averages hide the true peak. You need interval data.
• Ignoring efficiency. Calculating raw energy and buying that exact capacity will result in an undersized system.
• Assuming constant load. A facility with a single large motor is different from a building with many small loads. The shape of the curve matters.
• Forgetting thermal derating. A battery in a hot room will not deliver its rated power.
• Not checking the utility measurement window. If the utility measures on a 15-minute average, the battery must sustain the output for 15 minutes, not just a few seconds.

Final Verification and Commissioning

After installation, do not just turn it on and walk away. Run a commissioning test. Set the battery to discharge for a known period and measure the actual energy delivered. Compare it to the nameplate. Check the inverter logs for any derating events. Verify that the dispatch logic triggers correctly when the load approaches the target. If the system is connected to a building management system, confirm that the signals are clean and reliable. A misconfigured set-point can cause the battery to discharge too early or not at all, defeating the purpose of the investment.

Table of Typical Sizing Parameters

Parameter	Typical Range	Notes
Demand Reduction	20-40%	Higher reduction requires larger capacity
Round-Trip Efficiency	75-90%	Varies by technology and temperature
Cycle Life	2000-10000	Depends on chemistry and depth of discharge
Daily Cycles	1-2	Most C&I facilities cycle once per day
Margin	10-15%	Recommended buffer for real-world variability

Conclusion

Sizing a battery for demand is a mathematical exercise based on real data. It is not a one-size-fits-all decision. Each facility has a unique load profile, a specific utility rate structure, and a particular operating environment. By following the steps above, you can move from guesswork to a defensible engineering calculation. The goal is to reduce the demand charge without over-investing in hardware that will sit idle. Start with the data, set a realistic target, and verify the result with a model. That is how you get the right system for the job.