What size battery for your solar array

The common instinct is to size a battery against the solar array: bigger array, bigger battery. That is the wrong reference point, and it leads to buying capacity that never gets cycled.

A battery’s job is to cover demand when the panels are not producing. So the figure to size against is how much electricity you use between sunset and sunrise, and you can measure that this week.

Measure your overnight use

If you have a smart meter, this is a measurement rather than a guess.

1. Get your half hourly data from your supplier’s app or data portal.
2. Pick a fortnight that represents your normal life, avoiding holidays.
3. For each day, total the consumption from roughly 4pm to 8am.
4. Look at the spread across the fortnight, not just the average.

That daily total is the most a battery could usefully discharge on those days. Buying much beyond the upper end of your range means paying for capacity that will rarely be touched.

If you have no half hourly data, an approximation: take your annual consumption, divide by 365 for a daily average, then estimate the share falling outside daylight hours from your own routine. A household using 3,600 kilowatt hours a year averages about 10 a day; if roughly 60% of its life happens outside generating hours, its evening and overnight demand is around 6 kilowatt hours. Cruder than measuring, and still better than sizing off the array.

Then check what there is to store

The battery also cannot store more than the array produces in surplus.

Take the same illustrative household with a 4kW array generating about 3,400 kilowatt hours a year on our model’s UK average. In high summer a clear day can put well over 15 kilowatt hours on the roof against perhaps 4 of daytime consumption, so surplus is abundant. In midwinter the whole day may generate 2 or 3 kilowatt hours and the house uses it as it arrives, leaving almost nothing to store. In those months the battery’s usefulness shifts to charging from a cheap overnight window, such as the 8p off peak rate on Intelligent Octopus Go, rather than from the roof, which is a return in its own right.

So there are two ceilings, and the useful size sits under both:

• Demand ceiling: your overnight consumption, about 6 kilowatt hours in the example.
• Supply ceiling: your typical daily surplus generation, which varies enormously by season and only exceeds the demand ceiling in the brighter half of the year.

For that household, a unit with 5 to 6 kilowatt hours usable is the honest size. The 12 kilowatt hour unit the array-based rule of thumb suggests would spend most of the year half empty.

Why oversizing costs more than undersizing

An undersized battery still cycles fully every day. It captures less than it could, but every kilowatt hour of capacity you bought is working.

An oversized battery has capacity that sits idle. You paid for it, it degrades with age whether or not you use it, and it earns nothing. The economics of storage are driven by cycles: a cycled kilowatt hour is worth roughly 10 to 15 pence on current rates, the gap between the capped 26.11p import rate and what you gave up to charge, after round trip losses, as worked through in the battery cost guide. An uncycled kilowatt hour is worth nothing at all. In the example above, the spare 6 kilowatt hours of the oversized unit would need to cycle to justify their share of the price, and for perhaps eight months of the year they have nothing to cycle with.

This is the opposite of the sizing logic for panels, where a large fixed cost means the marginal panel is cheap. For batteries, the marginal kilowatt hour is cheaper per unit but only earns anything if it gets used.

Power rating matters as well as capacity

Capacity is how much it holds. Power rating is how fast it can deliver.

A battery with plenty of stored energy but a modest power rating cannot cover a large simultaneous load: an oven, a kettle and an electric shower running together draw several kilowatts at once, and whatever the battery cannot supply is imported at the full rate even though the battery is charged. If your evenings involve big simultaneous loads, check the continuous power rating on the datasheet rather than only the capacity, and ask the installer what happens above it. The answer, import at your normal rate, is not a failure, but it does trim the savings the capacity figure implied.

Round trip efficiency is a quiet deduction

Not everything you put in comes back out. The loss is modest but real, around the low teens in percentage terms for typical home systems, and it applies every cycle.

It matters most for arbitrage, where you are earning the spread between two rates. Losses eat directly into that margin, so a wide spread tolerates them comfortably and a narrow one may not. It is also why our model applies a round trip deduction explicitly instead of letting stored units pretend to be free.

A worked approach

Rather than a number that will not fit your house:

1. Overnight consumption, from your own half hourly data.
2. Typical daily summer surplus, from your generation estimate and your daytime consumption.
3. Take the smaller, and look for a unit whose usable capacity is close to it.
4. Check the power rating against your largest simultaneous evening load.
5. Compare candidates on price per usable kilowatt hour, which is the only fair basis for comparing storage quotes.

Then model it

Sizing is a starting point, not a payback figure. Put the usable capacity, your consumption pattern and your import and export rates into the ROI calculator, which applies degradation and a mid-life replacement over the projection, and you will see whether the size you landed on actually earns its price. If two candidate sizes come out close, the smaller one is usually the safer bet: it will cycle harder, and cycles are what you are paying for.