Ultra-short-term prediction of solar irradiance with multiple exogenous variables by fusion of ground-based sky images

Developing and using solar energy has become an important strategic decision for sustainable development in many countries. Short-term changes in solar irradiance can affect the safety and stability of photovoltaic and solar thermal power plants, so the accuracy of solar irradiance prediction has attracted significant attention. This paper proposes a short-term irradiance prediction method based on an improved complete ensemble empirical mode decomposition with adaptive noise and the partial differential equation model. Image feature information is obtained from ground-based sky images, and two ordinary differential equation (ODE) networks are used to process historical irradiance information and exogenous variables, including historical meteorological and sky images information. Using the ODE solver, the temporal pattern of the target sequence and the serial correlation between the exogenous variables are obtained, and an irradiance prediction model based on multivariate time series is established. The proposed method is evaluated using a public dataset from California, USA, and locally collected datasets. The experimental results show that the proposed method has high prediction accuracy and significantly improves the estimation of solar irradiance.