Wind power interval prediction method, system and storage medium

Abstract

The embodiment of the invention provides a wind power interval prediction method, a system and a storage medium. The method comprises the steps of comprehensively utilizing the one-step prediction interval structure characteristic of the LUBE boundary estimation theory and the time sequence modeling capability of a GRU neural network, establishing an interval prediction model based on the neural network based on the LUBE boundary estimation theory, and enabling the output layer of the interval prediction model to comprise two neurons; directly generating a prediction interval according to theinput tensor, The neural network comprises a GRU recurrent neural network and a full-connection neural network. The GRU performs time-dependent modeling on the multivariable time sequence, and the full-connection neural network is used for obtaining the upper and lower bounds of a prediction interval. Secondly, aiming at the defect that the evaluation function of the prediction interval cannot bedifferentiated, the evaluation function of the prediction interval is improved, the difficulty that a traditional LUBE cannot adopt a gradient descent method is overcome, a GRU deep learning algorithmis effectively integrated into an LUBE model system, and the prediction precision is improved under the condition that calculation steps are reduced.

Description

technical field [0001] The embodiments of the present invention relate to the technical field of wind power generation, and in particular to a method, system and storage medium for wind power interval prediction. Background technique [0002] In recent years, in response to climate change and the depletion of traditional energy sources, renewable energy power generation has developed rapidly. Among them, wind power, as a clean and pollution-free form of energy, is increasingly integrated into the power system. However, the uncertainty and instability of wind speed bring certain difficulties to the safe and reliable operation of wind power system, which limits the rapid popularization and application of wind power generation. Therefore, high-quality wind power forecasting is of great significance for formulating optimal power system planning and rationally arranging energy storage systems. [0003] Wind power forecasting methods are mainly divided into two categories: deter...

AI Technical Summary

Problems solved by technology

[0005] The embodiment of the present invention provides a wind power interval prediction method, system and storage medium, which overcomes the difficulty that the traditional LUBE cannot adopt the gradient descent method, effectively integrates the GRU deep learning algorithm into the LUBE model system, and improves the Prediction accuracy

Images