The accuracy of short‐term PV power prediction is a vital factor to improve PV accommodation capacity. To address the low forecast precision of typical PV power models, this essay suggests a short‐term PV power prediction methodology according to quadratic mode decomposition (QMD) and hybrid bidirectional gated recurrent unit (HBiGRU). Firstly, to cope with PV power’s unpredictability, the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), sample entropy (SE), and variational mode decomposition (VMD) are employed to break down the PV power information, and a run of smooth intrinsic mode function (IMF) components is obtained; secondly, the HBiGRU model is established to express the connection between every IMF component and the impact factors of PV power, and the forecast results of all component are then obtained; finally, the short‐term photovoltaic power forecast outcomes are acquired by superimposing the forecast outcomes of every component. Testing is conducted using data from a photovoltaic power station in Australia, and the simulation results indicate that the proposed integrated forecasting model can effectively improve the accuracy of short‐term photovoltaic power forecasting. Compared to other forecasting models, its normalized mean absolute error and root mean square error are reduced by 3.21% and 5.04%, respectively, while the coefficient of determination increases by 22.7%.
Ye et al. (Thu,) studied this question.