A Distributed Real-~Time Short-~Term Solar Irradiation Forecasting Network for Photovoltaic Systems

Michael Adelbert Gacusan


Solar irradiation forecasting is essential for PV connected electrical grids to maintain reliability, stability, and effective matching of real-time demand to power distribution. This research paper develops and evaluates proposed forecasting methods using wireless sensor networks. Each node of the network is capable of monitoring illuminance data and communicate it through RF and/or WiFi. The nodes are calibrated with respect to irradiance data from an industry-standard pyranometer. Power consumption of each node type is also collected at different operating states. The proposed sensor network can estimate a cloud motion vector or a cloud shadow’s speed and direction from the data collected. By processing the collected data further, a forecasted solar irradiance ramp-down time-of-arrival is possible. The results are evaluated for both artificial and on-site cloud shadows.