Skip to content

Latest commit

 

History

History
63 lines (57 loc) · 3.78 KB

File metadata and controls

63 lines (57 loc) · 3.78 KB

Platform / provider maturity

Tracks which inverter platforms and price providers are real-world validated versus experimental. The experimental marker (README/USER_GUIDE tag + this file) is the stability flag for this codebase — see feature-lifecycle skill.

Experimental / not yet real-world validated

  • Growatt VPP control mode (inverter.control_mode="vpp", on top of the solax_modbus_growatt_min (GEN4) and solax_modbus_growatt_sph (GEN3) platforms) — shipped per issue #118. Confirmed writing real VPP commands to hardware by multiple testers (nholmgaard, ridax67, jdungen), and several real bugs found in the field have been fixed: LOAD_SUPPORT forcing a fixed rate instead of releasing control (#413), a spurious power=0% write on every schedule reload (#421/#423), IDLE periods draining the battery overnight (#466), and VPP Remote Control continuing to override the inverter after switching away from VPP mode (#479). A real-world regression scenario built from ridax67's live config is locked in as ci-wizard-growatt-vpp-ridax-118 (backend discovery test + Playwright wizard E2E). Still marked experimental: the most recent fix (#479) has only shipped in the beta channel (v10.1.0b4) so far, not yet in a stable/prod release, and no tester has given a clean "this now works end-to-end" confirmation since it landed. Move to the validated list below once both happen (feature-lifecycle Stage 5/6).
  • solax_modbus_growatt_sph (GEN3) — monitoring-only, schedule control not implemented.
  • solax_modbus_native (SolaX VPP).

Real-world validated

(Populate as platforms/providers graduate through feature-lifecycle Stage 6. Candidates not yet formally tracked here: Growatt cloud MIN/SPH, GEN4 Growatt TOU via solax_modbus — all in production use prior to this file's creation.)

  • solis_modbus (Solis hybrid via Pho3niX90/solis_modbus, added for issue #130) — confirmed working against real Solis installations by two beta testers (tatusbar on an S6-EH3P10K-NV-YD-L, andys1802 on an S6-EH3P15K): entities auto-detected, schedule control running successfully. A debug log from the field surfaced one real bug (grid export power never auto-configured, #475), fixed and locked into the ci-wizard-solis regression scenario (backend discovery test + Playwright wizard E2E).
  • ENTSO-e / Belpex price provider (entsoe_source.py, added for issue #126) — confirmed working against a real Belgian Belpex/Luminus Dynamic contract by Frank-Leysen over an extended live-test period spanning many beta builds. Several real bugs surfaced and were fixed along the way: wizard defaulting to SEK instead of EUR, ems_discharging_rate writes overriding Load First mode, and the terminal-value arbitrage cap using an unrelated slot's price as its ceiling (#422/#425). Frank's real (anonymized) config is locked into the regression suite as ci-wizard-entsoe-frank-126 (backend discovery test + Playwright wizard E2E). Unrelated live-test findings from the same thread were split out into their own issues (#337, #269, #352, #330, #345, #362, #363, #428, #429) rather than tracked here.