What question did this study set out to answer?

The aim is to develop a more accurate wind power forecasting model that addresses limitations of existing methods.

February 8, 2026Wind EnergyOpen Access

TSG‐Net: A Multiscale Decomposition and Spatio‐Temporal Graph Neural Network Framework for High‐Precision Wind Power Forecasting

View Full Paper

Ask AI

Bookmark

Authors

This framework demonstrates improved wind power forecasting accuracy in dynamic conditions, suggesting enhanced energy management.

The aim is to develop a more accurate wind power forecasting model that addresses limitations of existing methods.
Introduced TSG-Net framework integrating multiscale decomposition and graph neural networks.
Employed MDLinear for adaptive temporal feature extraction through learnable weights.
Constructed dynamic adjacency matrices for improved wake effect modeling.
Implemented adaptive attention mechanisms to adjust weight allocation based on scenarios.
Achieved MAE of 36.53 kW and RMSE of 45.17 kW.
Demonstrated improvements of 8.3% in MAE and 5.7% in RMSE compared to the best baseline.
Showed enhanced performance during wind direction changes and extreme weather events.

YuChen Zhang (2026) studied this question.

View Full Paper

Ask AI

Bookmark