Self-Powered or Time-Based? On Forty Solar Homes, One Battery Setting Saved Nearly Four Times the Other

Forty simulated California solar homes, hour by hour: with a low export credit, self-powered reached 97% of the best possible saving and a peak-only rule 25%.

Most home batteries offer two settings. Self-powered stores solar the house isn't using and spends it whenever the house would draw from the grid. Time-based holds the charge for the expensive evening hours. Under a time-of-use rate, time-based sounds like the obvious choice. We built a home battery planner that replays a battery through a year of a household's own smart-meter data, then ran forty homes through it. Under a low export credit, time-based as we modelled it left three quarters of the money on the table.

What was run

Forty single-family homes with 5, 7 or 9 kW of rooftop solar, picked by a fixed rule from NREL's ResStock model of California housing. ResStock simulates each building on real 2018 weather; these are simulations, not meters. Each home's hourly use minus its solar was written out as a Green Button CSV and loaded through the tool's own reader. The battery was 13.5 kWh usable, 5 kW, with a 90% round trip and 20% held back for outages. The two rate plans have the same invented 4-to-9 pm prices ($0.33 to $0.55 a kWh) and differ only in what an exported kWh earns:

  • Plan A: $0.04 most hours, more on summer evenings. This is roughly how California's net billing tariff works.
  • Plan B: the hour's import price minus $0.03, the way older net metering worked.

Each home was billed four ways: with no battery, with each setting, and at the perfect-foresight ceiling. The ceiling is the cheapest schedule possible if the whole year were known in advance.

Low export credit

Plan A, median homesaving a yearshare of the ceiling
Self-powered$69197%
Time-based (discharge only in the day's top-price hours)$19225%
Perfect-foresight ceiling$732100%

Self-powered saved more in all forty homes. The reason is simple. When exports earn $0.04 and imports cost $0.33 or more, every kWh of solar kept in the house is worth about $0.30, whatever hour it's used. A battery that waits for 4 pm covers the evening but goes into the night partly full. The next day it has less room for surplus solar, and the overflow goes to the grid for almost nothing. A self-powered battery empties every night and fills every day.

Our time-based rule is deliberately strict: it only discharges in the day's top-price hours. Real controllers often forecast and fall somewhere between the two. The point stands, though. Under a low export credit, simple self-consumption already gets within a few percent of perfect.

Near-retail export credit

Plan B, median homesaving a year
Self-powered−$12
Time-based$39
Perfect-foresight ceiling$239

Here the order flips. When exports earn almost what imports cost, storing solar gains next to nothing, and the battery loses 10% of whatever passes through it. Self-powered costs the median home money, and the only savings left come from the time-of-use spread. On a plan like this, a battery is mostly backup power.

Size

Under Plan A, going from 13.5 to 20 kWh added a median $7 a year per extra kWh. From 20 to 27 kWh added $1.60. Once a battery holds the evening's use, more capacity buys backup, not savings. Where that point falls depends on how much the house uses after dark, and that's exactly what a household's own meter file shows.

How we know the ceiling is the ceiling

The ceiling comes from a dynamic programme over stored energy. Our first version kept energy on a fixed grid of levels. It lost 3 to 4% of the saving because a midday surplus of 0.37 kWh couldn't be stored exactly. The shipped version interpolates between levels and also tries the points where an hour's cost bends: stay put, store exactly the spare solar, cover exactly the house's draw. We then solved 32 full-year cases a second way, as one linear programme with the HiGHS solver. The programme came within 0.21% of that exact optimum in every case (worst: $1.51 a year), and usually within 0.01%.

Try it on your own year

The planner reads Green Button CSV or XML (solar exports included) and Smart Meter Texas files. You write your rate plan in a few lines, for example import jun-sep weekdays 16-21: 0.55, and the tool shows every setting and size, compares several plans, and gives a payback if you enter a cost. It all runs in the browser, and the file never leaves your computer. The full write-up has the tables, and the scripts that produced them are in the repository.

Not financial advice. The homes are simulations and the prices are invented. Your tariff, incentives and interconnection agreement decide what a real battery may do.

All writing · More in Tools and projects · All tools and research