Abstract
Solar power is one of the most attractive green energy sources and plays a vital role in daily electricity supply. Since the amount of available solar power is uncontrollable, it is essential to forecasting its availability so that power plants can arrange power supply in advance. Global horizontal irradiance (GHI) is the key indicator of available solar power, highly accurate forecasts for which are required to successfully integrate solar energy into the power grid. In this letter, we propose probabilistic solar irradiance transformer (ProSIT); a novel deep learning model for multihorizon probabilistic GHI forecasting based on both historical GHI and natural irradiance impact factors. ProSIT is able to capture the correlations of the heterogeneous impact factors of GHI as well as the long- and short-term temporal dependencies across the historical sequence. ProSIT also features residual connections and gating mechanisms to suppress superfluous components <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ad hoc</i> . Rather than simply providing single forecasting values as in existing methods, ProSIT employs a probabilistic output module to afford auxiliary forecasting confidence and bounds, which are particularly desirable in practical applications. We conduct experiments on two real-world datasets and demonstrate that ProSIT can achieve considerably better performance than state of the art.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have