522f47e821
Fix G provoquait un yoyo EDF de 2-4 min (coupure a SOC 21 en bypass min_on contre rallumage a SOC 20 en bypass cooldown) : son garde `soc > gridSocOn` ne protegeait que l'instant de l'engagement sur SOC bas, pas la phase de remontee 20 -> 30. jsCode restaure byte-identique a la version pre-Fix-G. Le fixture de test portait les fallbacks du code (22/35) au lieu des valeurs de production lues via MCP (20/30), et le scenario de non-regression testait SOC 18 %, le seul point que Fix G epargnait. Suite reecrite avec les garde-fous manquants h-recovery-band-no-cut-21/-25. Mutation : Fix G reapplique -> 4/7 echecs, retire -> 7/7. Documentation du diagnostic des deux oscillateurs (yoyo EDF et oscillation de la charge EVCS en plein soleil) dans MANUEL_COMPLET 11.8 et HANDOFF 25. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KKzkkVTG6KXnGXFFC7X9XJ
33 lines
695 B
JSON
33 lines
695 B
JSON
{
|
|
"soc_percent": "50",
|
|
"grid_state": "off",
|
|
"pv_power_w": "2000",
|
|
|
|
"house_ac_power_w": "700",
|
|
"grid_power_w": "0",
|
|
|
|
"evcs_ac_power_w": "0",
|
|
"evcs_charging": "off",
|
|
"evcs_status_raw": "6",
|
|
|
|
"hysteresis_enabled": "on",
|
|
"notifications_enabled": "on",
|
|
"grid_latch": "off",
|
|
|
|
"grid_soc_on": "20",
|
|
"grid_soc_off": "30",
|
|
"grid_min_on_minutes": "25",
|
|
"grid_cooldown_minutes": "15",
|
|
"ac_loads_max_safe_w": "7800",
|
|
|
|
"last_on_at": "",
|
|
"last_off_at": "",
|
|
"last_signature": "",
|
|
|
|
"evcs_priority_active": "off",
|
|
"night_grid_takeover_needed": "off",
|
|
"evcs_source_mode": "Bloqué",
|
|
"user_strategy": "auto_solar_plus",
|
|
"evcs_surplus_state": "idle_wait_conditions"
|
|
}
|