A Python-based simulator for optimizing solar PV systems with battery storage. The simulator allows testing different battery management strategies to maximize cost savings and efficiency by analyzing historical generation and consumption data.
Important
Some metrics are yet to be validated, this is more of a proof of concept than anything else. Code is messy and there are no tests, so use at your own risk. 💃😄
The project aims to optimize a PV installation with battery by exploring different charging strategies:
- Self-consume strategy: Battery charges from solar and discharges when load exceeds production
- Force charge at night strategy: Battery charges from solar and discharges when load exceeds production, but it also forces the battery to charge at night to take advantage of lower prices
- Other strategies are possible like predictive charging, weather-aware charging, etc.
You can also explore what would happen if you increased the number of batteries, in my case, adding another battery increases the maximum power of the battery system from 2kW to 5kW. Or what happens when export limit is changed...
Inputs of the simulation are solar power and house load, outputs are battery SoC (kWh) and import/export grid (kW). Simulation granularity is 1 minute, and the shown scenario is force charging the battery at night.
Estimated metrics are almost spot on when compared with Home Assistant's energy dashboard. Some discrepancies might arise because we are not simulating battery behavior in detail, but rather a simplified model. End to end efficiency of solar panels, battery, and inverter isn't modeled, but rather a simplified model.
| Simulation | Ground Truth |
|---|---|
 |
 |
git clone https://github.com/leandroalbero/powerflow-simulator.git
cd powerflow-simulator
docker compose up --buildOpen http://localhost:8000 — the image includes one month of sample data (January 2024, 1-minute resolution).
To use your own full dataset, uncomment the volume mount in docker-compose.yml:
volumes:
- ./data:/app/data# Create virtual environment
python -m venv .venv
source .venv/bin/activate
# Install dependencies
pip install -r requirements.txt
# Install frontend dependencies
cd web/frontend && npm install && cd ../..
# Start backend + frontend dev servers
make web-devBackend runs on http://localhost:8000, frontend dev server on http://localhost:5173.
Input CSVs need a last_changed datetime index and a state column (values in Watts):
"last_changed","state"
"2024-01-01T00:00:00.000+01:00","1424.67"The data/ folder contains real data — high-res (1-minute) and low-res (1-hour) versions. The data/sample/ folder has a January 2024 subset used by the Docker image.
tariff = PowerTariff(
import_rate_schedule={(0, 8): 0.085, (8, 10): 0.134, ...},
export_rate_schedule={(0, 24): 0.08}
)
battery = Battery(capacity=5, max_charge_rate=2.0, max_discharge_rate=2.0)
sim = EnergySimulator(battery, load, grid, tariff, solar)
for timestamp in load_data.index:
sim.step(timestamp, prev_timestamp, strategy='self_consume')
results = sim.get_metrics()- Implement additional battery management strategies:
- Time-of-use optimization
- Weather-aware charging
- Peak demand reduction
- Create Home Assistant integration for strategy execution
- Implement predictive modeling capabilities
- Add support for varying solar capacity
- Better UX/UI for data visualization
- Better modeling of entities in the code, code cleanup/refactor is needed
Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.