Abstract:Geometric deep learning has gained tremendous attention in both academia and industry due to its inherent capability of representing arbitrary structures. Due to exponential increase in interest towards renewable sources of energy, especially wind energy, accurate wind speed forecasting has become very important. . In this paper, we propose a novel deep learning architecture, Multi Scale Graph Wavenet for wind speed forecasting. It is based on a graph convolutional neural network and captures both spatial and temporal relationships in multivariate time series weather data for wind speed forecasting. We especially took inspiration from dilated convolutions, skip connections and the inception network to capture temporal relationships and graph convolutional networks for capturing spatial relationships in the data. We conducted experiments on real wind speed data measured at different cities in Denmark and compared our results with the state-of-the-art baseline models. Our novel architecture outperformed the state-of-the-art methods on wind speed forecasting for multiple forecast horizons by 4-5%.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is **accurate wind speed prediction**, especially short - term wind speed prediction in multiple time ranges (such as 6 hours, 12 hours, 18 hours and 24 hours). With the increasing global attention to renewable energy (especially wind energy), accurate wind speed prediction is of great significance for improving the efficiency of wind energy utilization, optimizing the management of wind farms and formulating sustainable energy policies. ### Specific problem background: 1. **Limitations of traditional methods**: - Traditional numerical weather prediction (NWP) relies on complex physical models, which consume a large amount of computing resources and are time - consuming, and it is difficult to meet the rapid prediction needs in a short time. - Data - driven methods (such as deep learning) can use historical data for fast and accurate prediction, especially in short - term prediction tasks (such as within a few hours). 2. **Deficiencies of existing methods**: - Most of the existing wind speed prediction methods fail to fully capture the spatio - temporal relationship, that is, the spatial correlation between different locations and the temporal correlation in the time series. - Many existing models perform poorly in handling multivariate time - series data and cannot effectively integrate data from multiple sensors or stations. ### Solutions proposed in the paper: To solve the above problems, the author proposes a new deep - learning architecture - **Multi - Scale Graph Wavenet**. This model combines the graph convolutional neural network (GCN) and the Wavenet architecture, aiming to capture the spatio - temporal relationship in multivariate time - series data, so as to achieve more accurate wind speed prediction. ### Main contributions: 1. **Propose a new multi - scale spatio - temporal graph convolutional network**, which is based on Graph Wavenet and Inception network and can capture spatio - temporal relationships simultaneously. 2. **Conduct experiments on the Danish wind speed prediction benchmark data set**, and the results show that the prediction performance of this model in multiple time ranges is better than that of the existing state - of - the - art methods, with an average improvement of 4 - 5% in prediction accuracy. 3. **Introduce a learnable adjacency matrix** to explain the relative importance between different nodes, making the model more interpretable. Through these improvements, Multi - Scale Graph Wavenet not only improves the accuracy of wind speed prediction, but also provides valuable insights for understanding and analyzing the meteorological impacts between different cities.

Multi Scale Graph Wavenet for Wind Speed Forecasting

Deep Graph Convolutional Networks for Wind Speed Prediction

Wind Speed Multi-Step Forecasting Using High-Resolution and Multi-Objective Ensemble Model with Error Correction

Multi-Wind Turbine Wind Speed Prediction Based on Weighted Diffusion Graph Convolution and Gated Attention Network

Short-Term Marine Wind Speed Forecasting Based on Dynamic Graph Embedding and Spatiotemporal Information

Short-Term Forecasting Based on Graph Convolution Networks and Multiresolution Convolution Neural Networks for Wind Power

Promoting wind energy for sustainable development by precise wind speed prediction based on graph neural networks

WindTrans: Transformer-Based Wind Speed Forecasting Method for High-Speed Railway

Graph WaveNet for Deep Spatial-Temporal Graph Modeling

Interpretable multi-graph convolution network integrating spatio-temporal attention and dynamic combination for wind power forecasting

Wavelet decomposition and neural networks: a potent combination for short term wind speed and power forecasting

A Spatiotemporal Directed Graph Convolution Network for Ultra-Short-Term Wind Power Prediction

Spatiotemporal analysis of bidimensional wind speed forecasting: Development and thorough assessment of LSTM and ensemble graph neural networks on the Dutch database

Deep learning approach for wind speed forecasts at turbine locations in a wind farm

Multi-step short-term wind power prediction based on spatio-temporal graph convolutional networks

Spatial correlation learning based on graph neural network for medium-term wind power forecasting

Deep Spatio-Temporal Wind Power Forecasting

Deep Spatial-Temporal 2-D CNN-BLSTM Model for Ultrashort-Term LiDAR-Assisted Wind Turbine's Power and Fatigue Load Forecasting

A spatio-temporal forecasting model using optimally weighted graph convolutional network and gated recurrent unit for wind speed of different sites distributed in an offshore wind farm

A novel temporal–spatial graph neural network for wind power forecasting considering blockage effects

Wind Power Forecasting Based on a Spatial–Temporal Graph Convolution Network With Limited Engineering Knowledge