Solar Tracker
EcoFlow Stream Ultra + 3 panels (1,390 Wp: 2x465 W added 17 Apr, 1x460 W added 30 Jun) + 1.92 kWh battery. AC output capped at 1,200 W. Capex £1,000. Import rate 20.51p/kWh, standing 49.49p/day (Fuse, fixed to ~Aug 2027).
Updated 2026-11-05 · source: HA statistics, read-only
Headline
| Metric | Value |
|---|---|
| Online since | 17 Apr 2026 (203 days online · 191 with data) |
| Generated | 503 kWh |
| Battery discharged | 241 kWh |
| Import avoided | 498 kWh = £102.09 |
| Solar share of household use | 40% |
| Payback progress | 10.2% of £1000 |
| Projected annual saving | ~£144/yr (seasonally adjusted) |
| Payback period | ~6.9 years on £1000 capex (range 6.0-7.7 yr) |
Monthly
| Month | Days w/ data | Generated kWh | Battery out kWh | Grid in kWh | Avoided £ | Fill days | Curtailed kWh (upper bd) |
|---|---|---|---|---|---|---|---|
| 2026-04 | 11 | 22.9 | 16.7 | 60.8 | 4.56 | 1 | 1.6 |
| 2026-05 | 22 | 69.5 | 43.8 | 66.2 | 13.98 | 15 | 8.2 |
| 2026-06 | 30 | 93.8 | 43.0 | 57.6 | 19.03 | 17 | 10.7 |
| 2026-07 | 31 | 106.4 | 50.1 | 94.9 | 21.63 | 25 | 16.0 |
| 2026-08 | 31 | 93.3 | 42.8 | 97.3 | 19.09 | 20 | 15.5 |
| 2026-09 | 30 | 76.3 | 31.0 | 156.0 | 15.59 | 5 | 1.4 |
| 2026-10 | 31 | 37.5 | 12.7 | 189.4 | 7.64 | 0 | 0.0 |
| 2026-11 | 5 | 2.8 | 0.7 | 39.0 | 0.56 | 0 | 0.0 |
How to get more out of it
- The battery hits a ~95% charge ceiling on roughly a third of summer days; after that the panels throttle back to cover house load and the rest is lost. Running flexible loads (washer, dishwasher, dryer, immersion) during those ceiling afternoons uses energy that is otherwise thrown away. Prize is ~£11 this season and the fix is free.
- The Stream’s AC output is capped at 1,200 W, so a 1.8-2.0 kW washer heating element always pulls the remainder from the grid. A 30C/eco wash cycle is the single biggest lever on a laundry load. The EcoFlow Smart Plug lets the Stream see the appliance load and ramp up to that cap.
- Array growth: worth ~£14-19/yr per extra panel, all of it in the shoulders and winter (Nov-Mar + Oct), where the array is the constraint. Payback 5-8 yr and no replacement cost, so a panel is the best-value part of the kit. But it buys ~NOTHING in Jun-Aug: clear-day harvest is ~3.5 kWh whether 2 panels or 3, because the battery, not the array, is the summer ceiling. It also buys peak (needed to approach the 1,200 W AC cap on a heavy appliance). More battery helps summer but costs far more than the curtailed energy is worth. No winter data exists yet (online 17 Apr) - winter figures are modelled, not measured.
- Feed-in is off; a SEG tariff would earn only a few pounds a year on the curtailed volume, so it is not the lever.
Method notes
- PV is integrated from hourly statistics; the sensor’s daily sum is unreliable (froze at ~502 W on 27-29 Apr 2026, reporting 7.5 kWh/day from a 502 W peak). Auto-rejected flat days.
- Annual projection calibrates the array against the typical-yield curve over all usable history, not a trailing 30-day window: a November window samples the darkest weeks and then applies that ratio to June and July, which understated the year by ~25% (£106 vs £144).
- Import avoided = (PV + battery out - battery in) x 20.51p. No derived house-load figure involved, so it cannot double-count.
- Curtailment is modelled (per-month 90th-percentile envelope x morning clearness), so treat it as +/- 30%.
- Flagged and excluded as sensor faults: 2026-04-27, 2026-04-28, 2026-04-29