Self-consumption, export and whether a battery earns its keep

A solar system generates most in the middle of a summer day. A house uses most in the morning and the evening, and in winter. The mismatch between those two shapes decides what a solar system is worth, and it is a bigger factor than the generation figure everybody concentrates on.

Self-consumption is the proportion of what a system generates that is used in the house instead of exported to the grid. It matters because the two are worth quite different amounts. A unit used at home saves whatever you would have paid to import it. A unit exported earns whatever your export tariff pays, which is generally a good deal less. The size of that gap is what determines whether a battery pays.

Self-consumption is not independence

Two figures get confused, and they answer different questions.

Self-consumption asks: of the electricity the panels made, how much stayed in the house? It is measured as a share of generation.

Grid independence, or self-sufficiency, asks: of the electricity the house used, how much came from the panels? It is measured as a share of consumption.

They move in opposite directions as a system gets bigger. A tiny array in a busy house has self-consumption near 100 per cent, because every unit it makes is instantly swallowed, but it supplies almost none of the house's needs. A very large array in the same house exports most of what it makes, so self-consumption is low, but it supplies a great deal more of the house.

Quoting one when the reader assumes the other is a common way of making a system sound better than it is.

How MCS estimates it

MGD 003 does not model anything. It publishes lookup tables built from half-hourly monitoring of real households by Loughborough University, with battery behaviour modelled on top and validated by the BRE National Solar Centre. You find your row and read the answer.

Three things index the tables.

Occupancy

Not how many people live in the house, but when they are in it on weekdays. MGD 003 uses three patterns:

• Home all day — at least one person in the house between 9am and 5pm.
• In half the day — the house is empty for a morning or an afternoon.
• Out all day — the house is empty on weekdays.

Where the pattern is not known, the standard says use the middle one. The number of occupants is not asked for separately, because annual consumption already stands in for it.

Annual consumption

Taken from a bill based on actual meter readings, in bands of 500 kWh from 1,500 to 5,999. Where no figure is available, MGD 003 falls back on 3,500 kWh, which is the figure the guidance uses; it is not a current average, and Ofgem's typical household is nearer 2,700. If the house already has solar, the metered figure is net of it and the solar contribution has to be added back in, for which the guidance has its own table.

Annual generation

The figure from the MIS 3002 estimate, in bands of 300 kWh.

What the tables say

A representative case: a household consuming 3,500 kWh a year with a system generating 3,600 kWh, which is roughly a 4 kWp array in the Midlands.

Three things stand out.

Without a battery, most of the output leaves the house. Even in the household that is home all day, seven tenths of what the panels make is exported; in the household that is out all day it is more than eight tenths. The old Feed-in Tariff assumption that half of it stayed home was generous, and the difference between 30 per cent and 50 per cent is a large part of why some early systems disappointed.

A battery changes the picture completely, roughly doubling self-consumption in every case. This is the strongest argument for storage, and the tables support it plainly.

Occupancy matters much less once a battery is fitted. Without one, the spread across the three patterns is 16 to 30 per cent, nearly a factor of two. With 10 kWh of usable storage, all three sit between 70 and 73. The battery does for the empty house what being at home does for the occupied one, which is why the household that is out all day gains most from it: 54 points against 43.

Where it stops paying

The second half of the table is the important half. In the empty house, going from no battery to 5 kWh takes self-consumption from 16 to 51 per cent, a gain of 35 points. Going from 5 to 10 kWh adds 19 more, and going from 10 to 15 adds only two.

MGD 003 stops tabulating at 15.1 kWh of usable capacity, and by then the figures have essentially stopped moving. The reason is simple: there is only so much surplus in a single day for a battery to catch. A 4 kWp array on a good June day might export 15 kWh; on a January day it might export nothing at all, because nothing is left over after the house has taken what it needs. A battery larger than the daily surplus spends most of the year part empty.

The practical implication is that battery sizing is bounded by the array and the household, not by the household's appetite. Doubling a battery does not double its benefit, and past a certain point adds almost nothing.

Usable capacity, not the number on the box

The tables are indexed on usable capacity, which is smaller than nominal capacity. A battery advertised as 10 kWh might have 9.5 usable, or 8, depending on the depth of discharge the manufacturer allows in order to protect cell life. Using the headline figure puts you a band or two too high.

Round-trip efficiency is a separate deduction: a few per cent are lost putting energy in and taking it out again, so a battery returns rather less than it stored.

Doing the arithmetic

Take the household above, out all day, 3,600 kWh generated, import at 27p and export at 15p.

Without a battery: 16 per cent self-consumed is 576 kWh saved at 27p, which is £156. The remaining 3,024 kWh exported at 15p earns £454. Total £610.

With 5 kWh usable: 51 per cent self-consumed is 1,836 kWh at 27p, which is £496. Exported 1,764 kWh at 15p earns £265. Total £761.

The battery is worth about £150 a year at these prices. Against an installed cost of several thousand pounds, that is a long payback, quite possibly longer than the warranty on the cells.

The arithmetic turns on the gap between the two prices. At 27p import and 15p export the gap is 12p, and a battery is buying you 12p per unit it shifts. On a tariff with a very cheap overnight rate and a high day rate, or with an export rate near zero, the gap widens and the case improves sharply. It is worth doing the sum with your own two prices instead of accepting a general answer, because the general answer changes with the tariff.

Reasons that are not the arithmetic

Batteries are bought for things the tables do not measure, and it is more honest to say so than to bend the payback until it works.

Backup during a power cut, which requires a battery specified for it and wired for it, since most grid-tied systems shut down in an outage by design.

Arbitrage on a time-of-use tariff: charging from the grid overnight at a cheap rate and running the house on it by day. That can be worth more than the solar shifting is, and it works whether the sun shines or not.

And the wish to use one's own electricity instead of selling it cheaply and buying it back dear, which is a perfectly ordinary thing to want.