How to Size a Battery for Solar Plus Storage

What do kW and kWh mean for a battery?

A battery has two ratings. Power, in kW, is the maximum rate it can charge or discharge. Energy, in kWh, is how much it can store. Energy divided by power is duration in hours. The NREL Annual Technology Baseline describes commercial batteries by an energy to power ratio from 1 to 8 hours.

The two ratings carry different costs. The Annual Technology Baseline splits system cost into a battery pack cost in dollars per kWh and a balance of system cost in dollars per kW, so adding hours is cheaper per kWh than adding kW.

What inputs does battery sizing need?

Most sizing errors trace back to a missing or approximated input.

What is a step by step method for sizing?

The method below is the logic an optimization model runs.

1. Align a year of interval load data with a solar profile for the same hours; net load, load minus solar, is what the battery sees.
2. Build the baseline bill from the tariff on net load, month by month, including any ratchet.
3. Set the objective, lowest bill, self consumption, resilience, or a mix, and rank the value streams.
4. Find the monthly peak reduction target; that kW sets the minimum power rating.
5. Find the midday solar excess; its largest rate is a second power candidate and its daily energy an energy candidate.
6. Multiply the evening gap duration by the kW covered; this is delivered energy.
7. Divide delivered energy by round trip efficiency and usable depth of discharge, then add degradation margin.
8. Simulate a full year for two or three candidate sizes; the bill difference from baseline is the annual value.
9. Compare candidates on net present value and check interconnection, space, and incentive rules.
10. Pick the smallest system whose value holds up when load or tariff changes.

How many hours of duration do you need?

Duration follows from the job. Removing a demand peak that lasts one to two hours needs a short duration battery. Shifting midday solar into an evening time of use window needs a matching duration. A resilience goal sets a floor from the critical load in kW and the outage hours you plan for.

The laboratories treat duration as a design variable: the Annual Technology Baseline models 1 to 8 hours with 4 hours as a reference. The right answer is the duration at which the next hour of storage stops paying for itself, which only a simulation over the actual load and tariff can show.

What does a worked example look like?

Consider a facility with a 400 kW solar array and a 500 kW peak from 5 p.m. to 8 p.m., after solar output has declined. At midday the array exceeds load by 150 kW for about four hours, or 600 kWh. The goal is to hold the evening peak to 400 kW using that excess.

Power: the evening gap needs 100 kW, but absorbing the midday excess without curtailment needs 150 kW, so the power rating is 150 kW. Energy: the evening job is 100 kW for three hours, or 300 kWh delivered. At 85 percent round trip efficiency that requires about 353 kWh stored; at 90 percent usable depth of discharge the nameplate becomes about 392 kWh; and if the design must still work after 20 percent capacity fade, it rises to about 490 kWh. A 150 kW, 500 kWh battery, roughly 3.3 hours, fits the job.

How does behind the meter sizing differ from front of the meter sizing?

Behind the meter, the battery serves one customer and its value comes from the tariff: demand charges, time of use spreads, self consumption, and incentives with performance requirements. NREL research on this problem minimizes the lifecycle cost of electricity to the site and finds that valuing resilience during outages leads to larger systems. Lawrence Berkeley National Laboratory adds that demand charge savings from PV with storage are almost always greater than the sum of the savings from either alone.

Front of the meter, the battery sells into a market and its value comes from arbitrage, capacity, and ancillary services under ISO rules. Duration is set by what the market pays for, and power is often set by the interconnection capacity rather than by load. Coupling matters at this scale: DOE reports that colocated DC coupled solar plus storage costs 8 percent less than siting the two separately and AC coupled costs 7 percent less.

What are the common sizing mistakes, and when does a model beat a rule of thumb?

The recurring mistakes are simple to state. Sizing to the solar nameplate instead of to net load. Treating nameplate kWh as usable kWh. Ignoring degradation, so a system that works in year one misses the peak in year eight. Ignoring round trip efficiency. And sizing from a single peak day rather than a full year.

A rule of thumb is adequate when load is flat, the tariff is simple, and there is one value stream. A model is needed when demand charges, time of use windows, solar variability, degradation, and incentive rules interact across 35,040 intervals in a year. Intellect PLAN runs that simulation with the same optimization engine used to dispatch batteries in the field and reports the kW, kWh, and financial return for each candidate size.