Photovoltaic power generation power prediction method based on Copula function

The invention discloses a photovoltaic power generation power prediction method based on a Copula function. The used algorithm cannot comprehensively measure the nonlinearity and trend correlation between photovoltaic power generation and meteorological factors. The invention provides a CNN-Transform photovoltaic power generation power prediction method based on a Copula function, and the method is used for measuring the nonlinear relation and trend correlation between meteorological variables and photovoltaic power generation, so that meteorological factors with high correlation are selected, and the training amount of a model is reduced. Meanwhile, the data passes through a convolutional neural network and a self-attention module to respectively capture short-term and long-term dependency relationships. The CNN can effectively reduce network training parameters, shorten model training time and identify the internal relation between important features and generated power, the self-attention module directly p