Akademik Veri Yönetim Sistemi
14th International Conference on Electrical and Electronics Engineering, ELECO 2023, Virtual, Bursa, Türkiye, 30 Kasım - 02 Aralık 2023
A lightweight forecasting method was developed for day-ahead (12 to 72-hours) and intra-day forecasting of photovoltaic (PV) array power output for use with energy management of agricultural microgrids. The proposed method allows the PV output to be forecast without requiring extensive computing resources or high-bandwidth communication channel to the PV site. The proposed day-ahead forecasting method combines historical PV output data with historical weather data to infer the relationship between weather and PV output by fitting a simple regression model. Based on the assumption that updated weather forecasts are in general not obtained throughout the day, the current-day PV output forecast was updated using time-series techniques. The proposed forecasting methods were applied to a dataset of rooftop PV output and compared to reference persistence forecasts. The method was demonstrated by using the forecasts in the optimized operation of an agricultural microgrid and simulating with actual PV output.