Day-Ahead PV Power Forecasting Based on MSTL-TFT

Energy demand is increasing dramatically as global urbanization progresses.Solar energy is a clean energy source with low production and maintenance costs.Accurately predicted PV generation is of great importance for grid integration.Recent day-ahead PV forecasting studies mainly include generation data decomposition, additional meteorological and equipment features, improvement and integration of ANN-based models.We proposed a MSTL-TFT method for day-ahead PV forecasting. The results are better than any of the other studies we have surveyed on day-ahead DKASC PV forecasting.