Short-term cluster wind power prediction method based on VMD-EMD-WT signal decomposition and SDAE deep learning model

Abstract

The invention provides a short-term cluster wind power prediction optimization method based on VMD-EMD-WT signal decomposition and SDAE deep learning, and the method comprises the following steps: carrying out the preprocessing of multi-dimensional NWP data and wind power plant historical power data in an original feature database, and dividing the data into a training set and a test set, using VMD, EMD and WT to decompose the wind speed and wind direction characteristic quantity of the training set to serve as a new training set, inputting the wind speed and wind direction characteristic quantity of the new training set and the test set into SDAE for deep learning, and establishing VMD-SDAE, EMD-SDAE and WT-SDAE prediction sub-models; randomly dividing output results of the three prediction sub-models into several sets, and integrating each set by using an SVM algorithm to generate a single integration result; and randomly dividing all single integration results into several sets, integrating each set by using an SVM algorithm, establishing a multi-integration learning model, and outputting a prediction result. The method has higher accuracy and better robustness, and the short-term wind power prediction precision is effectively improved.

Description

technical field [0001] The invention relates to a short-term cluster wind power prediction method based on Variational Mode Decomposition-Empirical ModeDecomposition-Wavelet Transform (VMD-EMD-WT) signal decomposition and a Stacked Denoising AutoEncoder (SDAE) deep learning model, belonging to the field of cluster wind power short-term prediction. Background technique [0002] The intermittence, randomness and volatility of wind power pose great challenges to the safe and stable operation of the power grid. Wind Power Prediction (WPP) is one of the effective ways to solve this problem. Wind power forecasting is to predict the future output of wind power generation through parameters such as weather forecast data and wind farm operating status data, so as to improve the predictability of wind power generation. [0003] At present, most wind power forecasts around the world focus on the power forecast of a single wind farm, but the power forecast of a single wind farm cannot m...

AI Technical Summary

Problems solved by technology

[0003] At present, most wind power forecasts around the world focus on the power forecast of a single wind farm, but the power forecast of a single wind farm cannot meet the needs of power grid dispatching. On the one hand, the power system as a whole, its uncertainty The total amount of power is more concerned by dispatchers, and the output change of a single wind farm does not play a prominent role in dispatching; Off-grid events caused by power increase require effective prediction of cluster wind power power

Images