Extended Parameter "Net calculation" not flowing through to Daily TOU / Tariff billing (Peak import still counted despite battery covering load)

[kemyers95963]

System: Tesla PW3 18.86kW (site 111702)

Setup:

• Battery (Tesla Powerwall 3, Solar Export mode / self-consumption only, no battery export to grid) discharge power is posted as Extended Parameter v10 (Battery Power, W), signed convention: positive = discharging, negative = charging.
• On the Extended Data page for v10, I set Net calculation: Power (W), which per the on-page help text implements Net = Consumption - Generation - v10.
• On Edit Tariffs, both Export Model and Import Model are set to Net.
• Change was saved today (8/11) at approximately 9:27am Pacific.

Expected behavior:
From 9:27am onward, any interval where the battery is discharging to cover household load (i.e., v10 positive, grid import = 0) should be netted out of the Import/Peak calculation, so the Daily TOU chart and Credit/Debit/Balance should reflect near-zero import during those intervals.

Observed behavior:
Today’s Peak window (4-9pm) fell entirely after the 9:27am settings change. My battery was discharging to cover the home’s evening load for the entire Peak window (confirmed via the Tesla app’s own Powerwall/Grid monitoring, which is independent of PVOutput). Specifically:

• Tesla app’s Grid “Importing” graph shows essentially flat-zero import from mid-afternoon through the evening.
• Tesla app reports a $28.89 Net Credit for the day.
• Despite this, PVOutput’s Daily TOU chart for the same day (08/11/26) shows 13.261 kWh of Peak import and a $14.15 debit.

This suggests the v10 “Net calculation” setting on the Extended Data page may be applied to the Net Production/Live chart display, but is not being incorporated into the Daily TOU / Tariff billing aggregation, even though the Import/Export Model is set to Net.

Questions:

1. Is the Extended Parameter “Net calculation” setting expected to feed into the Daily TOU/Tariff billing totals (Credit/Debit/Balance/Peak-Off Peak breakdown), or only into the Net Production chart’s own display math?
2. If it should feed into TOU billing, is there a delay/batch process for the Daily TOU aggregate to recalculate after a Net calculation setting change, or should it apply immediately per-interval as data is ingested?
3. Is there a known interaction issue between per-interval Net calculation (via an Extended Parameter) and the Tariff engine’s own “Net” Import/Export Model that I might be missing?

Happy to provide screenshots (Extended Data settings, Edit Tariffs settings, Daily TOU chart, and the Tesla app’s Grid import graph for the same day/window) if useful for troubleshooting. Thanks in advance for any insight.

[kemyers95963]

Follow-up with a clean reproduction case — Net calculation (v10) confirmed not applying despite correct extended data

System: Tesla PW3 18.86kW (site 111702)

Picking this back up with harder evidence than my original post. I’ve now confirmed every step of the pipeline up to PVOutput’s Net chart, and the discrepancy is isolated entirely to how the Net calculation is applied.

Setup (unchanged):

• v10 (Battery Power, W) has “Net calculation” set to Power (W), which per the on-page help text should compute Net = Consumption − Generation − v10.
• Export Model and Import Model on Edit Tariffs are both set to Net.

Event: 08/15/26, 7:00–8:00 PM, a deliberate ~1 hour battery discharge (arbitrage), sustained around 26-27kW.

Verified data at every stage:

1. Home Assistant’s own history for the source sensor shows a clean, sustained discharge from ~7:03 PM to ~8:03 PM, peaking around 27kW.
2. PVOutput’s own Extended Data graph for v10 confirms this arrived correctly — hovering the graph at 19:20 shows Battery Power: 27,732W, matching HA almost exactly. So the correct value was uploaded and is stored correctly.
3. Despite that, PVOutput’s Net Energy/Net Power chart for the same evening shows Net Power of only 505W at 19:00 — nowhere near reflecting a 27kW discharge being netted against consumption/generation. If the Net calculation were applying, Net Power should have swung sharply negative (net export) during this window.

Conclusion: This rules out any upload, sensor, or automation issue on my end — the correct v10 value is confirmed present in PVOutput’s own stored data. The Net calculation setting is either not being read by the Net Power/Net Energy chart, or is applying incorrectly. This looks like a bug isolated to how the Extended Data “Net calculation” setting feeds the Net Power/Net Energy chart specifically (separate from whatever’s happening with Daily TOU/Tariff billing, which I raised in my earlier post and which is a possibly related issue).

Happy to provide raw exported data or additional timestamps if that helps anyone dig into this.

[bankstownbloke]

@kemyers95963

Is the Extended Parameter “Net calculation” setting expected to feed into the Daily TOU/Tariff billing totals (Credit/Debit/Balance/Peak-Off Peak breakdown), or only into the Net Production chart’s own display math?

The former, daily debit/credits should match the 5-minute calculation on system live page.

If it should feed into TOU billing, is there a delay/batch process for the Daily TOU aggregate to recalculate after a Net calculation setting change, or should it apply immediately per-interval as data is ingested?

When an update is received on either consumption/generation or v10 then the net value should be re-calculated.

Is there a known interaction issue between per-interval Net calculation (via an Extended Parameter) and the Tariff engine’s own “Net” Import/Export Model that I might be missing?

Possibly, will have to revisit the code to ensure that the extended data power value is being correctly added to the net data.

Note that net data is not stored and the daily tariff credit/debit calculation occurs every time an update is received when the net data is re-calculated.

[kemyers95963]

Thanks for the detailed reply, this is really helpful — good to have confirmation on the intended behavior (real-time recalculation on every update, no batching, and Daily TOU/tariff figures expected to match the live 5-minute page).

If it’s useful for tracking down the code issue, I have a clean, concrete test case from 08/15/26 that isolates the discrepancy:

• At 19:20 that evening, the v10 Extended Parameter (Battery Power) is confirmed correctly stored at 27,732W (I hovered PVOutput’s own Extended Data graph directly to confirm this — it matches my source data almost exactly).
• At 19:00 the same evening, the site’s Net Power on the live/Net Energy chart shows only 505W — nowhere near reflecting a ~27kW discharge being netted against consumption/generation. If the Net calculation setting (Power (W), “net = consumption - generation - v10”) were applying correctly, Net Power should have swung sharply negative (net export) during that window instead.

So the correct v10 value is confirmed present and stored in PVOutput’s own data at that timestamp, but doesn’t appear to be reflected in the Net Power figure for the same moment. Happy to pull additional timestamps, raw exports, or anything else that would help you verify a fix once you’ve had a chance to look at the code.

Whenever you get a chance to dig into it — no rush — would you mind following up on this thread with whatever you find? Even just a note on what the failure turns out to be (and whether it looks fixable, or if there’s a workaround in the meantime) would be really valuable. Appreciate you taking a look at this.