Solax-EMHASS

Automated energy management for a Qcells/SolaX X3 Hybrid inverter using EMHASS (Energy Management for Home Assistant) and Home Assistant automations.

This repository contains Home Assistant configurations, sensors, MPC (Model Predictive Control) payload templates, and battery management automations that optimize solar usage with price forecasts via EMHASS.

---

🧠 Overview

This project integrates:

• EMHASS for energy optimization with a 15-minute MPC horizon
• Solax Power Hybrid Inverter (X3-Hybrid) controlled via the Home Assistant solax integration
• Home Assistant Automations to execute remote control modes based on MPC forecasts
• Dynamic scheduling of battery charge/discharge based on price, PV forecast, and load

The goal is to maximize self-consumption, minimize grid cost, and keep the battery within safe state-of-charge (SOC) bounds.

---

⚙️ Requirements

This setup assumes:

• Home Assistant (Container or OS)
• EMHASS installed and reachable via REST API
• Solax Power integration for the inverter & battery (Modbus or built-in API)
• Qcells/SolaX inverter with firmware that supports Mode 8 remote control
• Forecast sources such as Solcast and/or price feeds

---

• sensors.yaml: Template sensors for forecasts and MPC inputs
• rest_commands.yml: REST definitions for EMHASS API calls
• *.yml automations: MPC control & battery automations

---

🧩 Key Features

🔄 MPC Integration

• 15-minute MPC horizon using EMHASS
• Cost-optimized scheduling based on:
  • PV forecast
  • Load forecast
  • Day-ahead electricity prices
  • Battery SOC constraints
• Automatic REST-based triggering from Home Assistant

The MPC controller uses:

• PV forecast (e.g., Solcast)
• Load forecast (ML-predict sensor)
• Price forecast (EPEX/Nord Pool)
• Battery SOC (Solax Modbus sensor)
• Cost minimization objective (costfun = cost)

The payload sensor builds the EMHASS API call and sends it every 15 minutes.

🧠 Dictionary-Based Sensor Design

All forecast and MPC output sensors are stored in dictionary (dict) format instead of flat time-series sensors.

This design choice significantly simplifies:

• Start-time alignment
• Time-index calculations
• MPC horizon handling
• Synchronization between PV, load, grid and battery forecasts

Each dict entry contains explicit timestamps, allowing robust time arithmetic without relying on implicit sensor update intervals.

🔋 Battery Closed-Loop Control

• Minute-based closed-loop execution
• Follows MPC scheduled battery power (p_batt_forecast)
• Supports:
  • Grid charging
  • PV charging
  • No-discharge periods
  • Normal operation fallback
• Additional SOC safety stop outside EMHASS

This automation reads:

• sensor.p_batt_forecast
• sensor.soc_batt_forecast from dayahead
• PV & load sensors

and adjusts the inverter remote control accordingly every minute.

📌 Battery Safety We enforce SOC bounds both in EMHASS config and within automations. For example, a safety stop is triggered when SOC < 22%. In Version v1.1.0 the min SOC is set to 20%, in addition a recovery loop is implemented to recharge battery back to 25% to keep the loop alive.

📊 Dashboard Example

For visualization, an ApexCharts card shows:

• Price forecast
• PV forecast
• Load forecast
• Grid & battery power forecasts

Below is an example configuration snippet:

📈 ApexCharts – MPC Forecast Visualization

type: custom:apexcharts-card
experimental:
  color_threshold: true
graph_span: 6h
span:
  start: minute
now:
  show: true
  label: now
  color: red
yaxis:
  - id: first
    decimals: 2
    apex_config:
      forceNiceScale: true
  - id: second
    opposite: true
    min: -4000
    max: 8000
    decimals: 0
    apex_config:
      forceNiceScale: true
header:
  show: true
  title: Energy flow
  show_states: true
  colorize_states: true
series:
  - entity: sensor.epex_spot_data_total_price_3
    yaxis_id: first
    curve: stepline
    float_precision: 2
    show:
      in_header: after_now
    stroke_width: 1
    name: price kWh
    data_generator: |
      return entity.attributes.data.map((entry) => {
        return [new Date(entry.start_time), entry.price_per_kwh];
      });
  - entity: sensor.p_pv_forecast
    yaxis_id: second
    curve: stepline
    show:
      in_header: before_now
    stroke_width: 1
    name: PV Forecast
    color: rgb(255,215,0)
    data_generator: |
      return entity.attributes.forecasts.map((entry) => {
        return [new Date(entry.date), entry.p_pv_forecast];
      });
  - entity: sensor.p_load_forecast
    yaxis_id: second
    curve: stepline
    show:
      in_header: before_now
    stroke_width: 1
    name: Load Forecast
    color: rgb(255,0,0)
    data_generator: |
      return entity.attributes.forecasts.map((entry) => {
        return [new Date(entry.date), entry.p_load_forecast];
      });
  - entity: sensor.p_grid_forecast
    yaxis_id: second
    curve: stepline
    opacity: 0.1
    color: rgb(0,255,255)
    type: area
    show:
      in_header: before_now
    stroke_width: 1
    name: Grid Forecast
    data_generator: |
      return entity.attributes.forecasts.map((entry) => {
        return [new Date(entry.date), entry.p_grid_forecast];
      });
  - entity: sensor.p_batt_forecast
    yaxis_id: second
    curve: stepline
    opacity: 0.1
    color: green
    type: area
    show:
      in_header: before_now
    stroke_width: 1
    name: Batt Forecast
    data_generator: |
      return entity.attributes.battery_scheduled_power.map((entry) => {
        return [new Date(entry.date), entry.p_batt_forecast];
      });

...

You can add additional panels for day-ahead MPC outputs and SOC forecast.

🐞 Known Issues & Notes ⚠️ Solax Mode 8 There is a discrepancy in how “Enabled No Discharge” behaves in Mode 8 when PV is present: According to documentation, Mode 8 submode should prevent discharge, supply house load from PV, and store surplus to battery. In practice, the inverter may still draw from the battery or export to grid unexpectedly.

🛠 EMHASS Add-on Compatibility Notes

The Home Assistant EMHASS Add-on installation is not fully compatible with the file path structure used in the EMHASS standalone setup.

Several paths (e.g. for forecast CSV files) differ from the upstream documentation and cannot be used reliably inside the add-on environment.

Workaround Used in This Project

To avoid filesystem path issues, this project:

1. Executes EMHASS predict via REST
2. Publishes the load forecast as a Home Assistant sensor
3. Converts the forecast into dictionary format
4. Feeds the resulting dict sensor back into subsequent MPC and battery control logic

This approach:

• Avoids direct file access inside the EMHASS container
• Keeps all data exchange within Home Assistant
• Improves robustness and portability across installations

While this deviates from the standalone EMHASS workflow, it has proven to be reliable in long-term operation with the Home Assistant add-on.

🛠 Usage Enable EMHASS server and configure Home Assistant to publish data Schedule dayahead optimization at ~05:30 Run regular MPC optimizations every 15 minutes Battery automation runs every minute to follow the MPC plan

📬 Contributions If you have improvements, examples, or fixes, feel free to submit a Pull Request.

🔌 SolaX Modbus Integration

This project relies on the excellent Home Assistant SolaX Modbus integration:

👉 https://github.com/wills106/homeassistant-solax-modbus

It provides detailed inverter and battery metrics and allows reliable remote control of operating modes required for MPC-based optimization.

📜 License

This project is released under the MIT License.