The majority of battery storage projects are doomed from the start. An individual glances at a utility bill, selects a nice, round, doable number, then reaches out to a vendor with an estimate to design a system around. This strategy will lead to a battery that is too small to achieve saving goals or too large for repayment within a sensible window.

Start with real load data, not a monthly average

A monthly bill provides information about the total amount of energy used. It gives you little to no information about when the energy was used, which is what matters for determining the size of the battery. To determine the appropriate battery size, you require a 15-minute or hourly load profile. This profile should cover at least one year. The best source of this data is interval metering data or a data logging device that has been connected to the equipment at the site.

Averages conceal the peaks that drive up costs. While a site may average 200 kW over a month, it could peak at 450 kW over a 20-minute period as three compressors engage on a sweltering day. If you base your sizing on the average, you are designing a battery system that will never come close to that peak. The load profile is used to design everything else and is the easiest step to skip.

Pick the use case before you pick a size

Sizing a battery for behind-the-meter demand charge reduction means ignoring the facility's consumed kilowatt-hours and focusing solely on the half-dozen or so 15-minute windows each month when demand was highest. You're looking to shave those spikes, so you'll want enough battery capacity to cover the kilowatt spike during the 15-minute interval of the month you're aiming for - plus a conservative margin of error to ensure you never hit your true peak and get saddled with a high demand charge again.

Sizing a battery for self-consumption adds a second dimension. Enough energy storage capacity has to be added to store the amount of kilowatt-hours generated that you can't use locally. In the US, that typically means measuring how much of your monthly peak solar generation is happening when you're off-peak, tracking how many of those behind-the-meter solar electrons you tend to export outside the property when the grid could use them most, and converting it to a comparable kilowatt-hour tally.

The overlap is going to be more conceptual than practical right now. For example, solar-plus-storage applications still only form a small fraction of the total behind-the-meter market and typically need you to vastly overprovision the battery because the battery is there for lots of jobs the PV needs to shave the kilowatt peak on an indirect basis. Most batteries built to this day are utility-scale systems and are of interest mainly to power engineers. These utility-scale systems find it pretty easy to size for one application alone.

Separate kW and kWh before you touch a spec sheet

This is where most sizing errors happen. Power (kW) and energy (kWh) are different numbers, and a battery has to satisfy both independently before you reconcile them.

Your kW rating comes from the peak demand reduction target: how much power do you need to shave off the highest spikes? Your kWh rating comes from duration: how long does that discharge need to run, and how many times a day? A battery rated for 200 kW but only 200 kWh will run out in an hour. A battery rated for 800 kWh but only 100 kW output won't touch a sharp demand spike no matter how much energy is sitting in it. The C-rate is what links the two - a 1C battery can discharge its full kWh capacity in an hour, while a 0.5C battery needs double the energy capacity to hit the same power target.

Derate everything before you finalize a number

The capacity described in a specification sheet isn't the number you want to aim for when sizing a battery system to shave your peaks. There are three main things that clip into the amount of energy you can practically pull in and out of a battery, and you need to account for all of them up front.

That spec sheet cost/kWh? Don't forget about round-trip efficiency, which mops up somewhere between 5% to 15% of the energy you put in that you can't use. Then there's depth of discharge - if you're not abusing your batteries, you're probably not cycling them all the way from 100% to 0% every charge, which knocks another 20% off right there. Finally, you've got to account for the fact that your batteries are going to lose capacity faster and faster as they age.

Plainly put, if you don't want to undersize your project and push your batteries to 80% degradation in the first five years, the battery capacity you're actually going to be able to use is going to be about 60% of what's on the label. This is also where it's worth comparing actual equipment specs side by side rather than relying on marketing capacity numbers. Sites like bessbase.com let you cross-reference BESS product data and supplier specs so you can see what a system actually delivers after derating, not just what's printed on the label.

If there's solar involved, model the overlap, not just the array size

One common shortcut here is to simply pick a battery size based on simple rules like 2 hours of storage per kW of solar. While you might end up in the right neighborhood with rules of thumb like this, you can't know unless you do the careful profile analysis. Optimize solar battery storage size because every dollar into storage is a dollar not going into generation, and get the most generation you can for your money before worrying about storage.

Build in room for growth, and check the physical constraints

Facilities are dynamic. New equipment is installed, floor space is reassigned, and adding staff is taken into account. A battery that is exactly the size of today's load profile with no margin will be undersized the following years. Therefore, most engineers add 10-20% headroom both on the kW and the kWh numbers.

The best specification is not always the one with the biggest number. In fact, anyone who adds headroom and then says "as long as it fits in the room" has missed the point. Making the battery too big wastes CapEx, forces inefficient partial charge/discharge cycling on a daily basis, and excessive cycling reduces the lifetime warranty, plus creates recycling issues with unused materials and undersized thermal management systems.

As you finalize the performance spec, it's important to walk the kW and kWh requirements against their physical constraints: available floor space, weight loading, ventilation and thermal requirements, and how the interconnection limits your utility or site service allows affect the use of interconnection cables and thermal paths. It's time to start counting batteries, DC connections, HVAC modules, etc.