← The Signal Report Work with me

Report 257 · Energy Storage

How much battery a 3-day outage takes

"How many batteries do I need?" has no single answer, and the main variable is not the battery. Berkeley Lab and NREL modeled about 10,000 homes in ten U.S. metro areas riding out a three-day outage on solar plus storage. Keeping the fridge, lights and water going is cheap. Keeping the heating and cooling on is what decides the size.

The study is Gorman, Barbose, Miller, Carvallo, Baik and White, Solar+Storage for Household Back-up Power: Implications of building efficiency, load flexibility, and electrification for backup during long-duration power interruptions, a 61-slide Lawrence Berkeley National Laboratory report from November 2023, done with the National Renewable Energy Laboratory and funded by the DOE Solar Energy Technologies Office. It is three years old, but it answers the question installers get asked every week better than any sales sheet I have seen.

What they modeled

For each of ten locations (Boston, Washington D.C., Denver, Dallas/Fort Worth, Duluth, Los Angeles, Memphis, Phoenix, Seattle and Tampa), the team built about 1,000 statistically representative single-family homes in NREL's ResStock platform, with 15-minute load data broken out by appliance. They then asked a simple question for every home: what is the smallest battery that keeps a chosen set of loads running through the outage?

The base case is worth knowing, because every number below depends on it:

  • A three-day outage starting at midnight on each home's 90th-percentile net-load day, the 36th hardest day of the year. The authors call it "challenging but not extreme."
  • Solar sized to cover 100% of the home's annual use, capped at 70% of roof area.
  • The battery starts the outage full, with 85% round-trip efficiency. Results are in usable kilowatt-hours.
  • It is a purely technical analysis: no prices, and no limits on space.

The essentials are cheap

The study defines three backup levels. "Limited critical load" means fridge and freezers, nighttime lighting, well pump, water heating, cooking and 70 watts of plug load. For that, the required battery was "quite small (<15 kWh) across all locations." That is one typical home battery, in every climate studied.

Heating and cooling are the whole story

Add heating and cooling to the backup list and the answer spreads out fast. For today's housing stock, the median required battery ran from roughly 10 kWh in Los Angeles to 90 kWh in Dallas/Fort Worth and Phoenix. In half the locations the median was 40 to 90 kWh. Those are medians, so plenty of homes in each city need more.

The worst case is electric-resistance heat (baseboards, electric furnaces, the heat strips in some systems). For those homes, backing up heat over three days would take "extremely large batteries in most locations," which the authors call "not a practical solution." Replacing resistance heat with a heat pump cut the median in many locations from 100 to 250 kWh down to 30 to 50 kWh.

Whole-home backup, meaning everything including pool pumps, hot tubs and the dryer, needed about 30 kWh more on average than critical loads with heating and cooling.

A rule of thumb you can use

The most practical finding is on slide 22. Across every home and location, the relationship between battery size and energy need was close to a straight line: "roughly 1.1 kWh increase in battery size for each kWh increase in net load." Net load here means what the backed-up loads draw over the outage minus what the solar makes during it. The authors suggest it "can also serve as a useful heuristic for sizing battery systems."

So the way to size your own system is not to start from a product. Start from your own data: what your house actually uses on a bad weather day with the loads you would keep on, minus what your panels make on that kind of day, times three days, plus about 10%. Utility interval data or a whole-home energy monitor will give you the first number. The study's homes are a representative sample, not large estates, so a big house with a pool will sit above these medians. The rule still applies.

What shrinks the battery, and what grows it

Thermostat setbacks. Raising the cooling set point 5 °F and lowering the heating set point 6 °F during the outage cut the median battery by 14 to 25 kWh in the five hot or electric-heat locations (D.C., Memphis, DFW, Tampa, Phoenix). In Phoenix, setbacks plus insulation and air sealing took the median from 90 to 50 kWh.

Heat pumps cut both ways. In hot climates they replace inefficient air conditioning and shrink the battery: in Phoenix, from 50 to 20 kWh. In cold climates, replacing a gas or oil furnace with a heat pump puts the whole winter heating load on the battery. That added 10 to 30 kWh to the median in Denver, Boston and Seattle and about 50 kWh in Duluth. Boston went from 10 to 40 kWh. On the coldest days (99th percentile) the median Boston home needed roughly 60 to 200 kWh, depending on how the heat pump was sized and whether its backup heat was electric. The authors' own fix: keep the existing fossil heating for occasional use during outages.

Duration. Battery size scales "more-or-less linearly" with outage length, because the solar usually cannot refill the battery each day. Each added day took 1 to 31 kWh more in the baseline stock. Electrifying heat in Denver and Boston shortened how long a 30 kWh system could carry a home from 7+ days to about 2.

What the study does not cover

Three limits matter for a real purchase. First, the model works in 15-minute steps and does not consider start-up current, so it says nothing about whether a battery's inverter can start your air conditioner. That is a separate question, covered in Report 120. Second, it assumes a full battery at the start, which is realistic for a forecast storm and not for a surprise outage. Third, most results use typical-year weather; when the team reran Phoenix with actual 2011 to 2021 weather, required batteries got larger. On power, the news is good: most homes needed a battery-hours ratio over 4, so typical 2-hour home batteries would rarely run short of kilowatts before kilowatt-hours.

And sizing is not the same as capacity on day one. Usable kilowatt-hours shrink as the battery ages (Report 205), so a system sized exactly to the line today falls short in a few years.

A note on my angle: I help design the AI battery-cycling systems for a veteran-owned (HUBZone) energy-storage integrator, run by fellow veterans. I do not own that company and earn nothing from this link. Full policy here. Nothing in this report involves them, and the study takes no position on any product.

The bottom line

For essentials, one home battery covers three days almost anywhere. With heating and cooling, the honest answer is "measure your load first," and the study gives you the multiplier: about 1.1 kWh of battery per kWh of net load. The cheapest kilowatt-hours are the ones you never need, which is why thermostat setbacks and insulation did more in hot climates than an extra battery would.

Sources

  1. W. Gorman, G. L. Barbose, C. Miller, J. P. Carvallo, S. Baik and P. White, Solar+Storage for Household Back-up Power: Implications of building efficiency, load flexibility, and electrification for backup during long-duration power interruptions, Lawrence Berkeley National Laboratory with NREL, November 2023 (report PDF). (Primary source, all 61 slides read; the accompanying building-level data file was not used. Source for the ten locations, ~1,000 ResStock homes each and 15-minute data; the base-case assumptions; the three backup load definitions; "<15 kWh"; the 10 to 90 kWh medians; the electric-resistance results; the ~30 kWh whole-home increment; the 1.1 kWh heuristic; the set-point, envelope, heat pump, duration and 99th-percentile results; the start-up current and 15-minute limits; the AMY weather check; and the kWh/kW ratio. All quotes verbatim.)
  2. Lawrence Berkeley National Laboratory, Energy Markets & Planning, "Study shows how required storage sizing changes as homes become more efficient, flexible, and electrified," 28 November 2023. (Read. Source for the "40-90 kWh in half of the locations" summary, the DFW and Phoenix figures, and the 6% to 90% and ~60% share of homes a 30 kWh system could carry.)
  3. Prior reporting in this publication: Report 120, Report 205. (Context only.)
Onur Oncer
Onur Oncer

U.S. Army combat veteran (Counter-IED / Electronic Warfare), peer-reviewed researcher in microwave spectroscopy, and founder & CEO of Shroombiosis. Consults on laboratory operations, AI, and supplement formulation.

← All reports