Data-Driven Minute-Ahead Forecast of PV Generation with Adjacent PV Sector Information

This paper proposes and validates a data-driven minute-ahead forecast model for photovoltaic (PV) generation, which is essential real-time micro-grid scheduling. Unlike day-ahead PV forecasts that heavily rely on weather information, our proposed does not require such data as it operates in an ultra-short-term time domain. Instead, the leverages generation of target sector its adjacent sectors to capture short-term factors affect electricity movement clouds. The employs long memory (LSTM) network process data. By conducting experiments with real site data, we demonstrate information from improves accuracy by 3.66% mean squared error index 1.19% absolute compared without information.