Wind speed interval prediction method based on probabilistic long and short term memory model

Abstract

The invention provides a wind speed interval prediction method based on a probabilistic long short-term memory model. The method comprises the following steps: firstly, constructing an LSTM time sequence prediction model and a Gaussian process regression probabilistic prediction model; then, according to the LSTM time sequence prediction model and a Gaussian process regression probabilistic prediction model, constructing a probabilistic LSTM time sequence prediction model; training to obtain probabilistic LSTM time sequence prediction model parameters through the extracted prediction factors and prediction objects, and obtaining an optimal probabilistic LSTM time sequence prediction model; and for any group of wind speed forecast factors, extracting a mean value and a variance of the forecast wind speed, and constructing normal distribution of the wind speed according to the mean value and the variance so as to predict and obtain a high-precision wind speed result. The method has the beneficial effects that a wind speed interval prediction result can be quickly provided, and the prediction result has a high-precision fitting characteristic of deep learning.

Description

technical field [0001] The invention relates to the technical field of wind speed prediction, in particular to a wind speed interval prediction method based on a probabilistic long-term and short-term memory model. Background technique [0002] Large-scale wind energy development has become one of the important strategies to solve energy and environmental problems. Accurate forecasting of wind energy can provide a reasonable reference for power generation planning. Data mining models have been widely used in wind speed forecasting, including multiple linear regression, time series analysis, fuzzy clustering, and artificial neural networks. In recent years, inspired by the rapid development and successful application of deep learning in the fields of human perception, image classification, and environment simulation, researchers have begun to introduce various deep network models into wind speed prediction. Long Short-Term Memory (LSTM) is an improved recurrent neural networ...

AI Technical Summary

Problems solved by technology

It can be seen that the existing technology cannot completely solve the problem of wind speed interval prediction based on the long short-term memory model

Images