HOME BATTERY PLANNER · GUIDE · THE TOOL

Self-powered or time-based: which home battery setting saves more?

Most home batteries offer two settings. Self-powered stores solar the house is not using and spends it whenever the house would otherwise draw from the grid. Time-based saves the stored energy for the expensive hours. Which one saves more depends almost entirely on one line of your tariff: what an exported kWh earns.

Measured on building simulations, not meters, with invented example prices. Not financial advice. Your own data and your own tariff are the real answer: run them in the tool.

What was run

Forty single-family homes with 5, 7 or 9 kW of rooftop solar, drawn by a fixed rule from NREL's ResStock model of California's housing stock (AMY2018 release 2, real 2018 weather, CC BY 4.0). Each home's hourly use minus its solar was written as a Green Button CSV and read by the tool's own reader. The battery: 13.5 kWh usable, 5 kW, 90% round trip, 20% kept back for outages, no charging from the grid. Two rate plans with the same invented 4 to 9 pm time-of-use import prices ($0.33 to $0.55 a kWh) and different export credits:

Each home was billed with no battery, with each setting, and at the perfect-foresight ceiling: the cheapest schedule possible if the whole year were known in advance, found by a dynamic programme that an exact linear-programme solver (HiGHS) confirms to within 0.21% on every test case.

Low export credit: self-powered wins, by a lot

Plan A, per home25th pctmedian75th pct
Saving, self-powered$562$691$878
Saving, time-based (discharge only in the day's top-price hours)$100$192$326
Saving, perfect-foresight ceiling$592$732$900
Self-powered as a share of the ceiling94%97%98%
Time-based as a share of the ceiling13%25%59%

When an exported kWh earns $0.04 and an imported one costs $0.33 or more, every kWh of solar kept in the house is worth about $0.30, whatever the hour it is used. A battery that holds its charge for 4 pm covers the evening, but stays partly full overnight and has no room for the next day's surplus, which is exported for almost nothing. Self-powered empties every night and refills every day. It saved more than the peak-only rule in all forty homes.

Near-retail export credit: the order flips

Plan B, per home25th pctmedian75th pct
Saving, self-powered−$50−$12$18
Saving, time-based$22$39$78
Saving, perfect-foresight ceiling$142$239$303

When exports earn nearly what imports cost, storing solar earns almost nothing and the battery loses 10% of everything it stores, so self-powered costs the median home $12 a year. The only money left is the time-of-use gap, which the peak-only rule captures part of. Under this kind of plan a battery is mostly an outage backup.

Size: past the evening's use, extra kWh earn little

Plan A, usable kWh / kWceiling, median $/yrself-powered, median $/yrextra $/yr per extra kWh
5 / 3.842641885.1
10 / 566964548.5
13.5 / 573269120.8
20 / 7.67737167.0
27 / 107807241.6

A second battery is often quoted for backup; on savings alone it earns a few dollars a year per extra kWh once the first covers the evening. Where the curve bends depends on how much the house uses after dark, which is exactly what your own meter data shows.

Not decided here

Real controllers forecast, they do not know the year; real time-based modes are often smarter than the strict peak-only rule modelled here, so expect them between its figure and the ceiling. No battery wear, outages, demand charges, incentives, netting rules beyond hour-by-hour, or tariff changes are modelled. The prices are invented; the homes are simulations. The full tables, and the scripts that make them, are in the source repository.

Run your own year in the tool: your meter file and your rate plan, every mode, every size.