Photovoltaic controller
A service that stops PV farms at negative energy prices
| Years | 2022–now |
|---|---|
| Client | own project, developed further at Leysa |
| Category | Energy |
| Technology | Python (FastAPI, pandas, plotly, cfgrib, scikit-learn, skforecast), Nginx, JavaScript |
A web service that collects and analyses energy market data, weather forecasts (including GRIB numerical models) and solar radiation. Every 15 minutes it checks prices and the state of the farms; at negative prices it stops production automatically and resumes it when it pays again.
The models predict energy price spikes and weather-driven fluctuations with 50–90% accuracy. The system runs in production for several PV farms and is developed further at Leysa.