Exogenous Parameters in Solar Forecasting

摘要: The ability to predict solar radiation reliably is crucial in optimizing energy integration, ensuring grid stability and regulating markets. One way improve accuracy forecasting with time series modeling use exogenous variables (e.g. temperature, humidity, pressure, wind speed, direction) addition measurements. Evidence from existing studies indicates that the extent which such can largely dependent on type of algorithm used. Our results indicate scope prediction target (lag duration, number steps ahead) also plays an important role determining results. More specifically, accurate pairing algorithms help achieve improvements longer lags at diverse horizons. These argue favor a multi-modeling approach where specific configurations are determined dynamically for each choice input.