Novel Multi-Time Scale Deep Learning Algorithm for Solar Irradiance Forecasting

Solar irradiance forecasting is an inevitable and most significant process in grid-connected photovoltaic systems. power highly non-linear, thus to manage the grid operation efficiently, with for various timescales, such as hour ahead, a day week strategies are developed analysed this article. However, single time scale model can perform better that specific but cannot be employed other forecasting. Moreover, data consideration limited. In work, multi-time solar proposed based on multi-task learning algorithm. An effective resource sharing scheme between each task presented. The algorithm implemented long short-term memory (LSTM) neural network performance investigated hyperparameter estimation of LSTM made by hybrid chicken swarm optimizer combining best features both optimization (CSO) grey wolf (GWO) validated, comparing existing methodologies timescale forecasting, strategy demonstrated consistent all improved metric results.