Sizing a Grid Battery: Power, Energy, Duration and the Augmentation Problem

A battery plant is described by two quantities that people routinely conflate. Power, in megawatts, is how fast it can charge or discharge. Energy, in megawatt-hours, is how much it holds. A 100 MW / 200 MWh plant and a 50 MW / 200 MWh plant contain the same amount of electricity and behave completely differently.

Divide energy by power and you get duration — how long the plant can hold its rated output before it is empty. Two hours in the first case, four in the second. Duration is the single most useful shorthand for what a plant is for.

• •Under an hour. Frequency response and fast reserve. Lots of power, very little energy, cycled shallowly and often.
• •Two hours. The workhorse of early Indian tenders. Enough to cover a sharp evening peak, and the cheapest way to claim a megawatt of capacity.
• •Four hours. Increasingly the default for standalone storage procurement, because a two-hour plant stops being useful once several of them are chasing the same narrow peak.
• •Six hours and beyond. Shifting midday solar into the evening in bulk, which is where solar-plus-storage procurement has been heading.

The trend is unmistakably towards longer duration. That matters commercially, because the cost of adding hours is close to linear in cells while the cost of adding power is concentrated in the inverters and the connection. A longer plant is not proportionally more expensive — it is mostly more cells behind the same electrical interface.

Several deductions sit between the number on the purchase order and the energy a plant can actually deliver into the grid, and every one of them has to be in the model.

• •Depth of discharge. Cells are not run from full to empty. The usable window is narrower than the nameplate, by design, to protect life.
• •Round-trip efficiency. What goes in does not all come out. Losses appear in the cells, the inverters and the transformer, and they are charged to you as energy you bought and cannot sell.
• •Auxiliary load. HVAC, controls, fire systems and lighting run whether or not the plant is cycling. In a hot climate the cooling load is not a rounding error.
• •Availability. Time spent unavailable for maintenance or faults is time the plant cannot earn.

Cells lose capacity. They lose it through cycling, and they lose it merely by existing at temperature. A plant that delivers its contracted energy on day one will not deliver it in year ten unless somebody has planned for the gap.

There are two honest ways to handle this, and the choice has to be made before the plant is designed rather than discovered in year six.

• •Oversize on day one. Install more energy than the contract requires, so that the plant is still compliant at end of life. Simple, and you pay for all of it up front at today’s cell prices.
• •Augment later. Install close to requirement and add cells periodically. Cheaper at the start, and it bets on cell prices continuing to fall — but it needs physical space, spare electrical capacity and a control system that can handle mixed-age cells, all reserved from the beginning.

The ratio of power to energy determines how hard each cell is worked. A two-hour plant discharges at roughly 0.5C; a one-hour plant at 1C. Higher C-rates mean more heat, more stress and faster ageing, which feeds straight back into the degradation assumption and the warranty.

This is why a supplier’s warranty is written against a specific operating profile — so many cycles a year, within a temperature band, at a stated C-rate. Operate outside it and the guarantee that underpinned your financial model no longer applies. Our explainer on reading a battery spec sheet covers the same numbers at pack scale.

Establish what the plant is being paid to do. Convert that into a required duration and a required power. Add the deductions above to get from useful energy back up to installed energy. Decide oversize-or-augment and reserve the space accordingly. Only then ask a supplier for a number.

Sized the other way round — starting from a container count someone quoted — the plant usually turns out to be the wrong shape for the revenue it was meant to earn. For the smaller behind-the-meter case, see our note on sizing a C&I system, where the same distinction decides the answer.