UCY offers AI capabilities as software-as-a-service (SaaS) to PV plant owners/operators, including Cyprus’s Distribution and Transmission System Operators, with primary focus on c-Si systems. Capabilities include AI-powered generation forecasting (day/hour-ahead), fault diagnostic algorithms (statistical/ML-based), and digital twin models. Services are gradually expanding to perovskite technologies — forecasting has already been demonstrated using 1 year of field data, though fault diagnostics remain more challenging depending on data quality.
Virtual Acces
Main purpose: PV generation forecasting, performance prediction, digital twin models, and some fault diagnosis/degradation analytics
- Achieves <5% error for day- and hour-ahead forecasts using ML + statistical post-processing; validated globally across varied weather/PV configurations
- Models rely on a single measured input (historical onsite power), combined with forecasted irradiance and temperature
- Numerical weather predictions feed the trained model; day-ahead forecasts corrected via statistical post-processing, hour-ahead via sliding-window training with error vector correction
- Deployed by EAC and TSOC to manage PV integration in Cyprus’s grid, avoiding high-voltage and reverse power flow issues