End User
EMOT
Description
This service delivers high-resolution forecasts of solar PV generation tailored to the specific configurations of energy communities, including individual rooftops, shared PV systems, and EV-coupled infrastructure. Using machine learning models trained on historical generation data, weather forecasts, and system metadata (e.g., tilt, azimuth, shading), it provides short- and medium-term predictions. The forecasts enable intelligent scheduling of battery charging, EV loads, and peer-to-peer energy trading. Designed for real-time operations and long-term planning, it supports enhanced self-consumption, grid-aware dispatch, and community-level flexibility management.
Core Capabilities
Predictive & Prescriptive Analytics
Business Need
Energy communities rely on variable solar PV generation as a primary energy source, but its intermittency challenges stability, especially when integrated with EV charging and battery storage. Accurate local PV forecasting is essential for optimizing energy flows, reducing curtailment, and coordinating flexible assets like EVs. This service addresses the need for granular, site-specific PV forecasts to support demand-response, storage scheduling, and market participation, improving overall energy resilience and autonomy at the community level.
Key Performance Indicators
RMSE and normalized MAE for 1-hour, 6-hour, and day-ahead horizons
Forecast bias
Latency and uptime of forecast service
Data Provided
Historical PV output data (internal, community-owned systems)
Weather forecasts (Open Meteo, ECMWF, MeteoLux – open/public)
PV system metadata (tilt, azimuth, capacity – internal)
TEF
TEF EV