How to Size Commercial Battery Storage: A Worked Example
Author
Hisham Abdalla
Date Published

Disclaimer: Research and analysis by the engineering team. Sources referenced below.
Commercial battery storage sizing comes down to five moves. Name the job, take the shortest-interval load data you can get, find the power and energy the job needs at the meter, convert that to a nameplate kWh and a converter kW that still deliver after losses, ageing and heat, then check the C-rate. The last move is the one buyers skip, and in the example below a 0.5C product lifts the battery from 325 kWh to 500 kWh.
Our guide to commercial battery storage covers what a BESS is, the four jobs it does and the safety standards around it. This article is the arithmetic.
Which Job Sets the Size?
Each job has one constraint that binds. Peak shaving depends on the shape of the worst peaks, in kilowatts and in minutes. Backup depends on the critical load multiplied by the hours you promise. Load shifting, which moves energy from a cheap tariff period to an expensive one, depends on the price spread and how long it lasts.
Solar shifting is capped by the smaller of two numbers: the surplus the array cannot use at midday, and the load after sunset. A campus with its own array should not size the battery on the array's kWp. Suppose the surplus on a poor-sun day is 180 kWh and the evening load is 260 kWh. The battery is worth 180 kWh, and another 80 kWh of capacity is a box with nothing to fill it.
Diesel displacement is set by the generator. The battery carries the site while the set is off, and the set refills it while running near its best loading. Our guide to hybrid solar systems covers how the three sources share a load.
The rest of this article works through peak shaving, because it has the most arithmetic and the most ways to go wrong.
Why Interval Data Beats the Monthly Bill
A monthly bill gives one maximum demand figure and a total of kWh. It cannot say whether the peak lasted fifteen minutes or two hours, and that duration is the energy you are about to buy.
Demand charges are usually set by the highest average over a single interval. A 2015 study of 98 commercial facilities by the US National Renewable Energy Laboratory, now the National Laboratory of the Rockies, Deployment of Behind-The-Meter Energy Storage for Demand Charge Reduction, assumed charges based on peak demand over 15 minutes. So ask the utility, or the meter, for 15-minute data covering twelve months, and size on the worst day. A typical weekday hides the one hot week when chillers and production overlap.
Even the worst day will not repeat exactly. The same study found that a battery should not be expected to cut the metered peak by its full rated power: for a four-hour system it suggests about 85 percent is a reasonable expectation, and the figure falls for shorter systems. Build a margin in, and say so on the drawing.
If the site has its own array, load the PV output into the same 15-minute record and look at net demand at the meter. Solar trims the sunny-day peaks and does nothing on the overcast day with the highest load, which may be the day that sets the demand charge. The study found that small, short-duration batteries were the most cost-effective with or without solar, so an array leaves the method alone and changes the data you feed in.
How Much Power and Energy Does the Peak Need?
Suppose a packhouse has a 15-minute meter record, and its worst day in twelve months shows seven consecutive intervals at 880, 940, 990, 970, 920, 860 and 830 kW, with every other interval that day under 700 kW. The bill's maximum demand is 990 kW. Management wants a cap of 800 kW.
Every interval above the cap is an excess: 80, 140, 190, 170, 120, 60 and 30 kW. The largest excess, 190 kW, is the power the battery must deliver. The energy is the sum of the excesses multiplied by 0.25 hours, because each interval is a quarter of an hour. The sum is 790 kW, so the energy is 790 × 0.25 = 197.5 kWh.
The cap is a choice, and the energy price of that choice is not linear. Run the same sum on the same day and a 900 kW cap, a 90 kW cut, needs 55 kWh. An 850 kW cap, cutting 140 kW, needs 115 kWh. At 800 kW the answer is the 197.5 kWh above for a 190 kW cut. A 750 kW cap takes 240 kW off and needs 285 kWh, and a 700 kW cap takes 290 kW off and needs 372.5 kWh.
A little over double the power, 90 kW to 190 kW, takes the energy from 55 kWh to 197.5 kWh, about 3.6 times as much. The study found that small, short-duration batteries paid back fastest in its simulations, for this reason. The last kilowatts of a peak are the most expensive to shave, so set those five answers beside the tariff before you settle on a round number.

What each demand cap costs in battery energy on the same worst day of a hypothetical packhouse. The last kilowatts of a peak are the most expensive to shave. Source: MIMAH engineering practice, worked from hypothetical 15 minute data.
From Usable kWh to Nameplate kWh
The packhouse needs 197.5 kWh and 190 kW at the meter. Nobody sells that. The battery has to be bigger, and four things make it so.
Margin. Add 10 percent to both numbers for the peak you have not seen. That is a judgement and a low one: the study's 85 percent would point to about 18 percent on power (1 ÷ 0.85), so take yours from the spread in your own data. Energy becomes 197.5 × 1.10 = 217.25 kWh and power 190 × 1.10 = 209 kW.
Round-trip efficiency. The 2022 cost and performance assessment from Pacific Northwest National Laboratory (PNNL) uses 86 percent AC to AC for lithium iron phosphate, measured at the inverter, and about 83 percent with a transformer in the path. It works at 1 MW and above, so treat 86 percent as a starting point for a smaller cabinet. If the losses split evenly between charging and discharging, the discharge leg is the square root of 0.86, which is 0.9274. The cells must supply 217.25 ÷ 0.9274 = 234.3 kWh.
Depth of discharge. This is the share of the rated energy you may use in a cycle. Suppose the warranty allows 90 percent. Then 234.3 ÷ 0.90 = 260.3 kWh of nameplate.
End-of-life capacity. Suppose the warranty guarantees 80 percent of nameplate energy at the end of its term. Then 260.3 ÷ 0.80 = 325.4 kWh. The same PNNL report defines end of life for lithium-ion at 60 percent, so read your own warranty for what it promises.
Said the other way round, 0.90 × 0.80 = 72 percent of the nameplate is usable at the end of the term. A 325.4 kWh battery then gives 234.3 kWh to the converter and 217.25 kWh at the meter.
You can instead start smaller and add modules as capacity fades. That defers cost and leaves a second purchase, with a supplier who may have stopped making your module.

From energy at the meter to nameplate kWh, with a 10 percent margin, round-trip efficiency, depth of discharge and end-of-life capacity. Source: MIMAH engineering practice; 86 percent AC to AC efficiency from PNNL-33283 (2022).
What Does C-Rate Do to the Size?
C-rate is converter power divided by nameplate energy. A 1C system empties in one hour, a 0.5C system in two hours and a 0.25C system in four. A 500 kWh battery at 0.25C delivers 125 kW. The same battery at 0.5C delivers 250 kW.
The converter comes first. It must deliver 209 kW at the hottest hour, and a converter that derates in heat has to be bigger than that. Suppose a 10 percent derate: 209 ÷ 0.90 = 232.2 kW, so the next standard size is 250 kW. The heat section below explains why that figure is an assumption. The energy need is 325.4 kWh, so the duty works out at 250 ÷ 325.4 = 0.77C. Suppliers sell fixed ratios, and the ratio changes the answer. A 2026 data sheet from FIMER lists one 108 kW converter paired with a 215 kWh pack rated at 0.5C and with a 107.5 kWh pack rated at 1C. Applied to 250 kW:
0.25C. 250 ÷ 0.25 = 1,000 kWh of nameplate. This is a four-hour product, and it suits solar shifting better than a short peak.
0.5C. 250 ÷ 0.5 = 500 kWh.
1C. 250 ÷ 1 = 250 kWh, which is less than the 325.4 kWh the energy needs. Energy rules, so the battery is 325.4 kWh at 0.77C.
At 0.5C the power requirement, not the energy, sets the size. The 500 kWh battery carries more energy than the peak needs. At the end of its term it delivers 500 × 0.90 × 0.80 = 360 kWh to the converter, or 360 × 0.9274 = 333.9 kWh at the meter, against the 217.25 kWh the peak needs. The spare 116.6 kWh is where a backup reserve can sit, and a 40 kW critical load would run about 2.9 hours on it.
Which is cheaper, 1C at 325 kWh or 0.5C at 500 kWh, depends on the supplier's price per kW and per kWh. Ask for both prices, and for the maximum continuous C-rate the warranty allows. The National Laboratory of the Rockies' battery lifespan page lists charge and discharge rates among the things that drive battery ageing, so a product pushed to its C-rate limit every day may not last as long as one with headroom.

How the C-rate sets the minimum battery size for a 250 kW converter when the energy need is 325.4 kWh. Source: MIMAH engineering practice; the 0.5C and 1C pairings follow the FIMER PVX-215 and PVX-107 data sheet.
How Does Heat Change the Answer?
Heat enters in three places.
Power comes first, because converters and cells derate when hot. The FIMER data sheet rates its cabinet for ambient temperatures up to 55 °C and says it derates above 50 °C. It does not give the curve, so ask for one. The 10 percent used above is an assumption, and the real curve replaces it.
Efficiency moves too. PNNL notes that round-trip efficiency depends on charge and discharge power, ambient temperature and state of charge range. A figure measured in a cool test hall is optimistic for a roof in July, and the cooling load is part of the loss. The FIMER sheet lists only a maximum efficiency, with no word on whether that is round trip, at what power or at what temperature. Ask for efficiency at the meter, at your design temperature, with auxiliaries included.
Then there is life. Temperature is the first factor the lifespan page names. A hot container that cycles daily will wear faster than the warranty's test conditions assume, and the warranty may exclude it, so read the temperature clause before the capacity figure.
What to Check Before You Order
Recharge. Delivering 217.25 kWh at the meter takes 217.25 ÷ 0.86 = 252.6 kWh from the grid, about 2.5 hours at 100 kW. Charging counts toward the demand cap as well, so site load plus charging power has to stay under 800 kW or the battery creates a new peak.
Warranty throughput. One cycle a day for a year is 365 cycles. Check the warranty's energy throughput and its years against that. Our guide to solar PV warranties covers how to read those terms on the rest of the equipment.
Competing jobs. If the same battery also provides backup, the reserve state of charge has to be protected from the peak-shaving logic. The 116.6 kWh of spare energy above only exists if the controller keeps it.
Campuses with generators. A site with critical IT load adds reliability rules that the battery must not disturb. Our article on solar power for data centres covers that for one kind of campus.
What to Do Next
Get twelve months of 15-minute data before you ask anyone for a quotation. Work out the energy at several caps on the worst day and decide the cap with the tariff in front of you. Then send suppliers the power, the energy, the margin and the site temperature, and ask for kW and kWh priced separately, round-trip efficiency at the meter, end-of-term capacity, the derating curve and the C-rate limit. Compare 0.5C and 1C options on that basis, and keep the 0.25C product for solar shifting.
For the arithmetic of the solar side, see our guide to sizing a solar system for a business.
MIMAH's renewable energy team designs off-grid and hybrid systems with battery storage, sized from measured load and interval data. If you have a quotation and want it checked against your own peaks, get in touch.
