Energy Monitor

Custom Home Assistant integration for forecasting home battery state of charge until the next solar takeover window and recommending when to run flexible loads.

This first version is recommendation-only. It does not directly control EV chargers, spa heating, or appliances.

Installation

HACS

Requires Home Assistant 2025.3 or newer for the programmable load UI.

1. In HACS, open the menu and choose Custom repositories.
2. Add https://github.com/TheBarnHome/Energy-Monitor as repository type Integration.
3. Install Energy Monitor from HACS.
4. Restart Home Assistant.
5. Go to Settings > Devices & services > Add integration and search for Energy Monitor.

Manual

Copy custom_components/energy_monitor into your Home Assistant custom_components directory, restart Home Assistant, then add the integration from Settings > Devices & services.

Features

• Battery SoC forecast at 30 minute resolution.
• forecast time-series attribute suitable for ApexCharts-style dashboards.
• Baseload profile learned from historical home consumption.
• Optional grid export support. If export is unavailable or disabled, surplus solar that cannot be stored is treated as clipped.
• Configurable flexible loads via integration options.
• Flexible loads can be added as Home Assistant config subentries from the UI.
• Recommendation calendar for viable load windows.
• energy_monitor.recalculate service to force a refresh.

Home Assistant entities

• sensor.energy_monitor_battery_forecast
• sensor.energy_monitor_min_soc_before_solar
• sensor.energy_monitor_solar_takeover_time
• sensor.energy_monitor_recommendation_summary
• calendar.energy_monitor_recommendations

The battery forecast sensor exposes:

{
  "forecast": [
    {
      "datetime": "2026-05-28T22:30:00+02:00",
      "soc_percent": 62.1,
      "battery_energy_kwh": 6.21,
      "solar_forecast_kwh": 0.0,
      "baseload_kwh": 0.24
    }
  ],
  "actual_entity": "sensor.battery_soc",
  "generated_at": "2026-05-28T22:29:10+02:00",
  "resolution_minutes": 30,
  "horizon_end": "2026-05-29T18:00:00+02:00"
}

Solar forecast input

The configured solar forecast entity should expose a list attribute named one of forecast, forecasts, or data. Each item should include a datetime-like key (datetime, start, or period_start) and an energy-like value (energy_kwh, solar_forecast_kwh, pv_estimate, estimate, or value).

Flexible loads

Preferred setup: open the Energy Monitor integration in Settings > Devices & services, use the entry menu, and choose the subentry action to add a programmable load. Each load gets its own recommendation sensor.

Configure:

• Name and optional stable ID.
• Priority, where 1 is evaluated first.
• Duration in minutes.
• Either power in kW or total energy in kWh.
• Optional earliest start and latest end in HH:MM format.

The legacy JSON options format is still supported as a fallback:

[
  {
    "id": "ev",
    "name": "EV",
    "priority": 1,
    "duration_minutes": 180,
    "energy_kwh": 7.2,
    "earliest_start": "22:00",
    "latest_end": "08:00"
  },
  {
    "id": "spa",
    "name": "Spa",
    "priority": 2,
    "duration_minutes": 90,
    "power_kw": 2.0
  }
]

For EVs, calculate energy_kwh from the car battery capacity and the difference between current SoC and target SoC. A future version can read those entities directly.

Development

Run the pure Python forecast tests:

python3 -m unittest discover -s tests