A Neural Network Short-term Wind Speed ​​Prediction Method Based on Improved Difference Algorithm

Abstract

The invention discloses a neural network short-term wind speed prediction method based on an improved difference algorithm. The neural network short-term wind speed prediction method is used to solve technical problems of a prior art such as low accuracy and poor reliability of wind power plant short-term wind speed prediction, and large error fluctuation, easy falling into local minimum, long training time, low verification precision, and poor generalization capability caused by a single model or a prediction model combined with a heuristic formula algorithm. The neural network short-term wind speed prediction method comprises steps that wavelet packet decomposition is used to decompose an original wind speed signal into subsequences having different frequencies; a parent population X is generated by calculating the subsequences, and a progeny population S is generated according to the parent population X, and the progeny population S is used to update the parent population X to acquire the group of weights and thresholds of the parent population; neural network prediction models corresponding to the subsequences are established according to the group of weights and thresholds, and the neural network prediction models are used for the wind speed prediction to acquire the prediction results of the subsequences.

Description

technical field [0001] The invention relates to the field of wind speed prediction, in particular to a neural network short-term wind speed prediction method based on an improved difference algorithm. Background technique [0002] Wind energy, as a clean and renewable energy, has received extensive attention from all over the world in recent years. Vigorously developing wind power generation is the need of my country's energy construction to implement a sustainable development strategy, and it is of great significance to speed up the development of the national economy, promote the adjustment of the power industry, reduce environmental pollution, and promote scientific and technological progress. my country's wind energy reserves are large, widely distributed, and have great potential. Therefore, under the environment of national policy support and tight energy supply, the development prospects of China's wind speed power generation industry are very broad, and it is expect...

AI Technical Summary

Problems solved by technology

[0006] The embodiment of the present invention provides a neural network short-term wind speed prediction method based on the improved difference algorithm, which solves the defects of low accuracy and poor reliability in the short-term wind speed prediction of wind farms in the prior art. A single model or a prediction model combined with a heuristic formula algorithm has large error fluctuations, and is prone to technical problems such as local minimum, long training time, low verification accuracy, and insufficient generalization ability

Images