Wind power prediction method based on deep learning and refined weather forecast

By combining deep learning and refined meteorological forecasting technology, using multiple models for wind power power feature extraction and meteorological downscale processing, the problem of insufficient precision and real-time performance of the stroke wind power is solved in the existing technology, and more efficient and accurate wind power prediction is achieved.

CN119723109APending Publication Date: 2025-03-28HUANENG CLEAN ENERGY RES INST
0 Cites 1 Cited by

Patent Information

Application Number
CN202411843774.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-14
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

The prior art has high accuracy and real-time requirements in wind power prediction, especially in the case of short-term prediction and data scarcity, making it difficult to build an accurate and stable model.

Method used

Using deep learning and refined meteorological forecasting methods, wind power power feature extraction, meteorological downscale processing and feature fusion are carried out by collecting wind farm operation data, NWP data and ERA5 data, using AlexNet, improved ResNet and spateGAN models, wind power power feature extraction, meteorological downscale processing and feature fusion, and finally output the wind power power prediction results.

Benefits of technology

It improves the accuracy and stability of wind power prediction, and can provide more accurate prediction results when data is scarce or volatile.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention belongs to the technical field of electric power prediction, and particularly relates to a wind power prediction method based on deep learning and refined weather forecast. The method specifically comprises the following steps: collecting operation data, NWP data and ERA5 of a wind power plant; according to the operation data of the wind power plant, obtaining a thermodynamic image, and based on an image processing technology, eliminating abnormal values in the wind power data to obtain historical wind power data; processing the historical wind power data to obtain wind power characteristics; inputting the NWP data into the trained weather downscaling model to obtain a refined weather forecast result; based on an improved ResNet model, constructing a refined meteorological feature extraction model; according to the refined weather forecast result, using the improved ResNet model to extract refined weather features; carrying out feature fusion on the wind power features and the refined meteorological features; and finally outputting a wind power prediction result after feature transformation. And the prediction of the wind power is realized.
Need to check novelty before this filing date? Find Prior Art