VMD-Transformer-Based Wind Speed Forecasting for Spot Market Trading

In the power spot market, renewable energy power generation usually show very big fluctuation, so accurate wind speed forecast becomes more and more important. To solve this problem, this study propose a mixed forecasting framework which combine Variational Mode Decomposition (VMD) with a Transformer-based predict network: first, using VMD to decompose the complex wind speed series into many components of different time-scale, so that can dig out more richer and more structural dynamic feature; then, these refine signal will be put into Transformer model, who is good at catching long-short term time dependency, in order to exactly extract the micro hidden pattern in wind speed fluctuation. The experiment results show that the VMD-Transformer framework has quite strong generalization ability and also high prediction accuracy, so overall it works pretty nicely. The proposed method provides a reliable forecasting tool for renewable energy operation. It also gives effective decision support for power dispatching and participation in the electricity spot market.