LOG-015 ·
Eighteen Months of Solar: The Honest Numbers
Production, payback, and the day the grid went down and my panels shut off anyway. Energy independence, measured without the brochure.
- Words
- 636
- Est. read
- 2.8 min
- Confidence
- 0.88
- Topics
- energy, solar, measurement
My rooftop array was commissioned on 2024-12-19: 8.2 kW of panels, and as of this spring, a 13.5 kWh battery. I have 18 months of five-minute-interval data from a revenue-grade meter, which is 157,000 data points, which is enough to write the entry the solar brochures do not. The brochures are not lying, exactly. They are rounding in one consistent direction.
Production versus the promise
The installer's model projected 10,400 kWh/year. Measured production, first 12 full months: 9,620 kWh, a 7.5% shortfall. Sources, quantified from the data: a shading estimate that was optimistic about my neighbor's oak (4.1%), panel soiling between rains (measured by the production jump after cleaning: 2.2%), and inverter clipping on the brightest days (1.2%). None of this is scandalous. All of it compounds, and none of it was in the brochure.
Household consumption over the same 12 months: 11,300 kWh. So the array produced 85% of what we consumed, annually. That number is technically true and practically misleading, which brings us to the important section.
Self-sufficiency is not a yearly number
Annual netting hides the mismatch that actually defines energy independence: the sun makes power at noon and my household wants it at 19:00. The honest metric is self-consumption: what fraction of my usage was served directly by my own system at the moment of use.
| Period | Self-sufficiency (panels only) | With battery |
|---|---|---|
| June (best month) | 61% | 89% |
| December (worst month) | 22% | 31% |
| 12-month weighted | 38% | 64% |
Panels alone made me 38% independent, not 85%. The 13.5 kWh battery, added this spring, lifts the running figure to 64%, projected from three months of data plus last year's seasonal shape (so treat it as a forecast, not a measurement, until December proves it). Winter remains stubbornly grid-shaped. Anyone quoting you a single solar independence number without a month-by-month table is quoting the marketing department.
The day that taught the architecture lesson
2025-02-11, a storm took the grid down for 6.5 hours, and my panels, in full sun, produced nothing. This astonishes every homeowner the first time and it is by design: grid-tied inverters must shut down during outages to avoid electrocuting line workers. Reasonable rule, but the result deserves stating plainly: a solar array without islanding capability is not backup power. It is a grid accessory. Decentralized generation on a dependency it cannot escape.
The battery retrofit included an islanding gateway. Tested monthly since (the reader of this log knows I test my failovers): cutover in 1.2 seconds, essential loads carried for 11 to 19 hours depending on weather, indefinitely if the sun cooperates, since the panels recharge the battery once islanded. The February lesson, priced: the gateway was $1,700 of the battery project, and it is the line item that converts the whole system from bill reduction to actual resilience.
The money, without rounding
Total system cost after incentives: $19,400 (panels $12,100, battery and gateway $7,300). Value captured in 18 months from bill savings and export credits: $2,610. Straight-line payback: 11.1 years for the full system, though the panels alone pencil at 8.3 years and the battery honestly does not pencil at current rates. I bought the battery as resilience infrastructure, not as an investment, and the ledger records it under the same category as the 5G failover router: $19.20/month depreciated, purchased to make a failure mode boring.
Verdict
Energy independence at my latitude, measured rather than felt: 64%, seasonal, hard-won, and worth it. The remaining 36% is winter, and the options (more panels into a shaded roof, a generator, or acceptance) all have prices I have now computed. Acceptance is currently winning.
I put 0.88 on the claim that a household copying this configuration in a comparable climate lands within 10 points of my self-sufficiency numbers. The variance is mostly your oak tree. Measure your oak tree before you sign anything.