Weird growatt min schedule for "selling to grid"

[johanzander]

Version
8.4.2

Problem:
dashboard shows schedule and energy flow as expected:

but battery schedule shows weird fluctuations between load support / export during period 72-95 that I cannot explain.

Battery Schedule:
╔════╦═══════════╦══════╦═══════╦╦═════╦══════╦══════╦═════╦═══════╦═══════════════╦═══════╦══════╦══════╗
║ Hr ║ Buy/Sell ║Cons. ║ Cost ║║Sol. ║Sol→B ║Gr→B ║ SoE ║Action ║ Intent ║ Grid ║ Batt ║ Save ║
║ ║ (SEK) ║(kWh) ║ (SEK) ║║(kWh)║(kWh) ║(kWh) ║(kWh)║(kWh) ║ ║ (SEK) ║(SEK) ║(SEK) ║
╠════╬═══════════╬══════╬═══════╬╬═════╬══════╬══════╬═════╬═══════╬═══════════════╬═══════╬══════╬══════╣
║ 47 ║ 1.21/0.49 ║ 0.5 ║ 0.61 ║║ 1.0 ║ 0.5 ║ 0.1 ║ 19 ║ 0.6 ║ SOLAR_STORAGE ║ 0.10 ║ 0.29 ║ 0.22 ║
║ 48 ║ 1.20/0.48 ║ 0.5 ║ 0.61 ║║ 1.2 ║ 0.7 ║ 0.0 ║ 20 ║ -0.0 ║ SOLAR_STORAGE ║ -0.00 ║ 0.33 ║ 0.28 ║
║ 49 ║ 1.24/0.51 ║ 0.5 ║ 0.63 ║║ 1.2 ║ 0.7 ║ 0.0 ║ 21 ║ -0.0 ║ SOLAR_STORAGE ║ -0.00 ║ 0.33 ║ 0.30 ║
║ 50 ║ 1.27/0.54 ║ 0.5 ║ 0.65 ║║ 1.2 ║ 0.7 ║ 0.0 ║ 22 ║ -0.0 ║ SOLAR_STORAGE ║ -0.00 ║ 0.33 ║ 0.32 ║
║ 51 ║ 1.32/0.58 ║ 0.5 ║ 0.67 ║║ 1.2 ║ 0.7 ║ 0.0 ║ 22 ║ -0.0 ║ SOLAR_STORAGE ║ -0.00 ║ 0.33 ║ 0.34 ║
║ 52 ║ 1.24/0.52 ║ 0.5 ║ 0.63 ║║ 1.2 ║ 0.7 ║ 0.0 ║ 23 ║ 0.7 ║ SOLAR_STORAGE ║ 0.05 ║ 0.34 ║ 0.24 ║
║ 53 ║ 1.27/0.54 ║ 0.5 ║ 0.64 ║║ 1.2 ║ 0.7 ║ 0.0 ║ 24 ║ 0.7 ║ SOLAR_STORAGE ║ 0.05 ║ 0.34 ║ 0.25 ║
║ 54 ║ 1.29/0.56 ║ 0.5 ║ 0.66 ║║ 1.2 ║ 0.7 ║ 0.0 ║ 24 ║ 0.7 ║ SOLAR_STORAGE ║ 0.06 ║ 0.34 ║ 0.26 ║
║ 55 ║ 1.27/0.54 ║ 0.5 ║ 0.64 ║║ 1.2 ║ 0.7 ║ 0.0 ║ 25 ║ 0.7 ║ SOLAR_STORAGE ║ 0.05 ║ 0.34 ║ 0.25 ║
║ 56 ║ 1.25/0.52 ║ 0.5 ║ 0.64 ║║ 1.0 ║ 0.5 ║ 0.0 ║ 26 ║ 0.5 ║ SOLAR_STORAGE ║ -0.00 ║ 0.24 ║ 0.40 ║
║ 57 ║ 1.29/0.56 ║ 0.5 ║ 0.66 ║║ 1.0 ║ 0.5 ║ 0.0 ║ 26 ║ 0.5 ║ SOLAR_STORAGE ║ -0.00 ║ 0.24 ║ 0.42 ║
║ 58 ║ 1.30/0.56 ║ 0.5 ║ 0.66 ║║ 1.0 ║ 0.5 ║ 0.0 ║ 27 ║ 0.5 ║ SOLAR_STORAGE ║ -0.00 ║ 0.24 ║ 0.42 ║
║ 59 ║ 1.34/0.59 ║ 0.5 ║ 0.68 ║║ 1.0 ║ 0.5 ║ 0.0 ║ 27 ║ 0.5 ║ SOLAR_STORAGE ║ -0.00 ║ 0.24 ║ 0.44 ║
║ 60 ║ 1.21/0.49 ║ 0.5 ║ 0.61 ║║ 0.9 ║ 0.4 ║ 0.0 ║ 27 ║ 0.4 ║ SOLAR_STORAGE ║ 0.06 ║ 0.19 ║ 0.37 ║
║ 61 ║ 1.27/0.54 ║ 0.5 ║ 0.65 ║║ 0.9 ║ 0.4 ║ 0.0 ║ 28 ║ 0.4 ║ SOLAR_STORAGE ║ 0.06 ║ 0.19 ║ 0.39 ║
║ 62 ║ 1.39/0.63 ║ 0.5 ║ 0.70 ║║ 0.9 ║ 0.4 ║ 0.0 ║ 28 ║ 0.4 ║ SOLAR_STORAGE ║ 0.06 ║ 0.19 ║ 0.45 ║
║ 63 ║ 1.52/0.74 ║ 0.5 ║ 0.77 ║║ 0.9 ║ 0.4 ║ 0.0 ║ 29 ║ 0.4 ║ SOLAR_STORAGE ║ 0.07 ║ 0.19 ║ 0.51 ║
║ 64 ║ 1.20/0.48 ║ 0.5 ║ 0.61 ║║ 0.7 ║ 0.2 ║ 0.1 ║ 29 ║ 0.3 ║ SOLAR_STORAGE ║ 0.10 ║ 0.15 ║ 0.36 ║
║ 65 ║ 1.35/0.60 ║ 0.5 ║ 0.68 ║║ 0.7 ║ 0.2 ║ 0.0 ║ 29 ║ 0.2 ║ SOLAR_STORAGE ║ -0.01 ║ 0.10 ║ 0.59 ║
║ 66 ║ 1.64/0.83 ║ 0.5 ║ 0.83 ║║ 0.7 ║ 0.2 ║ 0.0 ║ 29 ║ 0.2 ║ SOLAR_STORAGE ║ -0.01 ║ 0.10 ║ 0.75 ║
║ 67 ║ 1.82/0.98 ║ 0.5 ║ 0.92 ║║ 0.7 ║ 0.2 ║ 0.0 ║ 30 ║ 0.2 ║ SOLAR_STORAGE ║ -0.01 ║ 0.10 ║ 0.84 ║
║ 68 ║ 1.26/0.53 ║ 0.5 ║ 0.64 ║║ 0.5 ║ 0.0 ║ 0.0 ║ 30 ║ -0.0 ║ IDLE ║ 0.00 ║ 0.00 ║ 0.64 ║
║ 69 ║ 1.77/0.94 ║ 0.5 ║ 0.90 ║║ 0.5 ║ 0.0 ║ 0.0 ║ 30 ║ -0.0 ║ IDLE ║ 0.00 ║ 0.00 ║ 0.90 ║
║ 70 ║ 2.08/1.19 ║ 0.5 ║ 1.05 ║║ 0.5 ║ 0.0 ║ 0.0 ║ 30 ║ -0.0 ║ IDLE ║ 0.00 ║ 0.00 ║ 1.05 ║
║ 71 ║ 2.24/1.31 ║ 0.5 ║ 1.13 ║║ 0.5 ║ 0.0 ║ 0.0 ║ 30 ║ -0.0 ║ IDLE ║ 0.00 ║ 0.00 ║ 1.13 ║
║ 72 ║ 2.21/1.29 ║ 0.5 ║ 1.12 ║║ 0.2 ║ 0.0 ║ 0.0 ║ 29 ║ -0.3 ║ LOAD_SUPPORT ║ -0.05 ║ 0.00 ║ 1.17 ║
║ 73 ║ 2.42/1.46 ║ 0.5 ║ 1.23 ║║ 0.2 ║ 0.0 ║ 0.0 ║ 29 ║ -0.3 ║ LOAD_SUPPORT ║ -0.06 ║ 0.00 ║ 1.29 ║
║ 74 ║ 2.62/1.62 ║ 0.5 ║ 1.33 ║║ 0.2 ║ 0.0 ║ 0.0 ║ 29 ║ -0.3 ║ LOAD_SUPPORT ║ -0.07 ║ 0.00 ║ 1.39 ║
║ 75 ║ 2.77/1.73 ║ 0.5 ║ 1.40 ║║ 0.2 ║ 0.0 ║ 0.0 ║ 28 ║ -0.9 ║ EXPORT_ARBITRAG ║ -1.11 ║ 0.00 ║ 2.52 ║
║ 76 ║ 2.44/1.47 ║ 0.5 ║ 1.24 ║║ 0.1 ║ 0.0 ║ 0.0 ║ 27 ║ -0.5 ║ LOAD_SUPPORT ║ -0.08 ║ 0.00 ║ 1.31 ║
║ 77 ║ 2.52/1.53 ║ 0.5 ║ 1.28 ║║ 0.1 ║ 0.0 ║ 0.0 ║ 27 ║ -0.5 ║ LOAD_SUPPORT ║ -0.08 ║ 0.00 ║ 1.36 ║
║ 78 ║ 2.66/1.65 ║ 0.5 ║ 1.35 ║║ 0.1 ║ 0.0 ║ 0.0 ║ 26 ║ -0.7 ║ EXPORT_ARBITRAG ║ -0.42 ║ 0.00 ║ 1.77 ║
║ 79 ║ 2.95/1.88 ║ 0.5 ║ 1.50 ║║ 0.1 ║ 0.0 ║ 0.0 ║ 22 ║ -3.8 ║ EXPORT_ARBITRAG ║ -6.22 ║ 0.00 ║ 7.71 ║
║ 80 ║ 2.68/1.67 ║ 0.5 ║ 1.36 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 21 ║ -0.9 ║ EXPORT_ARBITRAG ║ -0.67 ║ 0.00 ║ 2.02 ║
║ 81 ║ 2.80/1.76 ║ 0.5 ║ 1.42 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 17 ║ -3.8 ║ EXPORT_ARBITRAG ║ -5.72 ║ 0.00 ║ 7.14 ║
║ 82 ║ 2.73/1.71 ║ 0.5 ║ 1.39 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 16 ║ -0.9 ║ EXPORT_ARBITRAG ║ -0.68 ║ 0.00 ║ 2.07 ║
║ 83 ║ 2.70/1.68 ║ 0.5 ║ 1.37 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 16 ║ -0.9 ║ EXPORT_ARBITRAG ║ -0.67 ║ 0.00 ║ 2.04 ║
║ 84 ║ 2.87/1.82 ║ 0.5 ║ 1.45 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 12 ║ -3.8 ║ EXPORT_ARBITRAG ║ -5.89 ║ 0.00 ║ 7.35 ║
║ 85 ║ 2.71/1.69 ║ 0.5 ║ 1.38 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 11 ║ -0.9 ║ EXPORT_ARBITRAG ║ -0.66 ║ 0.00 ║ 2.04 ║
║ 86 ║ 2.67/1.66 ║ 0.5 ║ 1.35 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 10 ║ -0.9 ║ EXPORT_ARBITRAG ║ -0.65 ║ 0.00 ║ 2.00 ║
║ 87 ║ 2.60/1.60 ║ 0.5 ║ 1.32 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 9 ║ -0.6 ║ LOAD_SUPPORT ║ -0.15 ║ 0.00 ║ 1.47 ║
║ 88 ║ 2.77/1.74 ║ 0.5 ║ 1.40 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 7 ║ -1.9 ║ EXPORT_ARBITRAG ║ -2.33 ║ 0.00 ║ 3.74 ║
║ 89 ║ 2.71/1.69 ║ 0.5 ║ 1.38 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 6 ║ -0.9 ║ EXPORT_ARBITRAG ║ -0.66 ║ 0.00 ║ 2.04 ║
║ 90 ║ 2.55/1.56 ║ 0.5 ║ 1.29 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 6 ║ -0.5 ║ LOAD_SUPPORT ║ 0.02 ║ 0.00 ║ 1.27 ║
║ 91 ║ 2.46/1.49 ║ 0.5 ║ 1.25 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 5 ║ -0.5 ║ LOAD_SUPPORT ║ 0.02 ║ 0.00 ║ 1.23 ║
║ 92 ║ 2.68/1.67 ║ 0.5 ║ 1.36 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 5 ║ -0.9 ║ EXPORT_ARBITRAG ║ -0.66 ║ 0.00 ║ 2.02 ║
║ 93 ║ 2.50/1.52 ║ 0.5 ║ 1.27 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 4 ║ -0.5 ║ LOAD_SUPPORT ║ 0.02 ║ 0.00 ║ 1.25 ║
║ 94 ║ 2.43/1.47 ║ 0.5 ║ 1.23 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 4 ║ -0.5 ║ LOAD_SUPPORT ║ 0.02 ║ 0.00 ║ 1.21 ║
║ 95 ║ 2.39/1.43 ║ 0.5 ║ 1.21 ║║ 0.0 ║ 0.0 ║ 0.0 ║ 3 ║ -0.5 ║ LOAD_SUPPORT ║ 0.02 ║ 0.00 ║ 1.19 ║
╠════╬═══════════╬══════╬═══════╬╬═════╬══════╬══════╬═════╬═══════╬═══════════════╬═══════╬══════╬══════╣
║Tot ║ ║ 24.8 ║ 49.73 ║║24.1 ║ 10.1 ║ 0.5 ║ ║C:10.6 ║ ║-26.12 ║ 5.14 ║70.71 ║
║ ║ ║ ║ ║║ ║ ║ ║ ║D:26.0 ║ ║ ║ ║ ║
╚════╩═══════════╩══════╩═══════╩╩═════╩══════╩══════╩═════╩═══════╩═══════════════╩═══════╩══════╩══════╝
Summary:
Grid-only cost: 49.73 SEK
Optimized cost: -20.98 SEK
Total savings: 70.71 SEK
Savings percentage: 142.2 %

bess-debug-2026-05-03-115301.md

PS. also BESS version missing in debug export file.

[bess-manager-claude-bot]

User reports fluctuating battery actions (load support / export arbitrage alternating) during hours 72-95 in the schedule, despite stable prices in that period. Debug report and detailed schedule table attached. Reply @claude-bot analyze to start deep code-level analysis.

[johanzander]

@claude-bot analyze

[bess-manager-claude-bot]

Root cause

The alternating LOAD_SUPPORT / EXPORT_ARBITRAGE pattern in hours 72–95 is not a bug — it is the correct, intentional output of the DP optimizer making fine-grained per-quarter-period decisions. Two mechanisms drive what appears to the user as "weird flickering":

1. Intent classification is based solely on discharge volume vs. home consumption. Any quarter-period where battery discharge exceeds 0.6 kWh gets EXPORT_ARBITRAGE; ≤ 0.6 kWh gets LOAD_SUPPORT. With consumption ≈ 0.5 kWh/period, the threshold is simply: discharge − 0.5 > 0.1.

2. The DP optimizer assigns different discharge amounts to different quarters based on sell price. Higher-priced quarters get larger discharges (up to the 3.75 kWh physical max); lower-priced quarters get only enough discharge to cover home load. The interleaving of these two tiers creates the visible alternation.

The -3.8 kWh "dump" periods (hrs 79, 81, 84) are the optimizer hitting the maximum discharge limit: 15.0 kW × 0.25 h = 3.75 kWh, displayed as -3.8 by the :.1f format string.

Evidence

• core/bess/decision_intelligence.py:421-424 — intent classification:

  if power < -0.1:  # Discharging
      if energy_data.battery_to_grid > 0.1:
          return "EXPORT_ARBITRAGE"
      return "LOAD_SUPPORT"

• core/bess/models.py:116 — battery_to_grid computation:

  battery_to_home = min(self.battery_discharged, remaining_consumption)
  # ...
  battery_to_grid = self.battery_discharged - battery_to_home

  With solar=0 and consumption=0.5 kWh, this simplifies to: if discharge > 0.6 → EXPORT_ARBITRAGE, else LOAD_SUPPORT. Every single hour in the table confirms this exactly (hr 87 at -0.6 kWh stays LOAD_SUPPORT because is false).

• core/bess/min_schedule.py:139-145 — hardware mode mapping:

  INTENT_TO_MODE = {
      "LOAD_SUPPORT": "load_first",   # inverter default — no export
      "EXPORT_ARBITRAGE": "grid_first",  # enables battery→grid export
  }

  The mode difference is real and necessary: during LOAD_SUPPORT periods, discharge ≤ consumption so no grid export is possible; during EXPORT_ARBITRAGE periods, the inverter must be in grid_first mode to allow the planned export.

• core/bess/settings.py:45 — max power:

  BATTERY_MAX_CHARGE_DISCHARGE_POWER_KW = 15.0

  15.0 kW × 0.25 h = 3.75 kWh → displayed as -3.8 kWh (rounding artifact of format).

• core/bess/min_schedule.py:170 — segment limit is enforced:

  self.max_intervals = 9  # Growatt supports up to 9 TOU intervals

  Only grid_first (EXPORT_ARBITRAGE) periods require explicit TOU segments; load_first is the inverter default. The 72–95 range produces 4 distinct grid_first TOU segments, well within the 9-segment hardware limit.

Proposed fix

There is no code defect to fix. The schedule is mathematically correct and the hardware modes are assigned appropriately. The "weird" visual appearance arises because the display shows per-quarter-period intents, making the optimizer's fine-grained sell-price decisions look like rapid flickering.

If the goal is a less-confusing display, the schedule table could aggregate consecutive same-intent blocks and show a single row for each run rather than individual quarter-periods. That would be a UI improvement, not a correctness fix.

Risks / open questions

• No behavioral or economic risk identified — the optimizer is correctly maximizing export revenue at peak-price quarters while conserving charge for higher-priced periods later.
• One minor cosmetic issue: the EXPORT_ARBITRAG truncation (15-char column width) in the schedule display cuts off the last character; this is unrelated to the reported problem.
• The "PS" note in the issue (BESS version missing from debug export) is a separate, minor issue.

---

Reply @claude-bot fix to implement and open a draft PR.