Wind power interval prediction method based on bidirectional time convolution and gating circulation unit

Abstract

The invention provides a wind power interval prediction method based on bidirectional time convolution and a gating cycle unit, and belongs to the field of artificial intelligence application of new energy power prediction.The method comprises the steps that a historical data set containing wind speed, temperature, season and wind power is obtained; on the basis of a network structure of a bidirectional time convolution combined gating cycle unit network, constructing a wind turbine generation power processing network and training the wind turbine generation power processing network; performing clustering segmentation on the historical data set, taking the wind speed data set, the temperature data set and the season data set as input quantities, and inputting the input quantities into the trained fan power generation power processing network to obtain a point prediction result of wind power; calculating errors between the point prediction results of the wind power and the actual wind power to obtain error values, and sorting the point prediction results by adopting a self-service method to obtain wind power interval prediction results under different confidence intervals; the wind power prediction method solves the problem that the difference between the wind power prediction value and the actual value of an existing wind power prediction method is large.

Images