A novel ARX-based multi-scale spatio-temporal solar power forecast model

In this paper an autoregressive with exogenous input (ARX) based spatio-temporal solar forecast model is proposed. Compared with conventional temporal-only autoregressive models, the proposed formulation takes into account both spatial and temporal solar power correlations. The overall root mean squared error (RMSE) percentage can be reduced by 12.72% using spatio-temporal solar forecast compared with persistence forecast model for hour ahead intervals.