SDWPF: A Dataset for Spatial Dynamic Wind Power Forecasting over a Large Turbine Array

Area of Science:

• Renewable Energy Systems
• Data Science
• Grid Integration

Background:

• Wind power is a clean, renewable energy source facing grid integration challenges due to its inherent variability.
• Accurate Wind Power Forecasting (WPF) is essential for managing grid stability and maximizing renewable energy utilization.
• Existing WPF datasets are limited in scope, lacking detailed spatial and dynamic contextual information for individual turbines.

Purpose of the Study:

• To introduce the Spatial Dynamic Wind Power Forecasting (SDWPF) dataset, a comprehensive resource for advancing WPF research.
• To provide detailed spatial distribution and dynamic contextual factors for each wind turbine, addressing limitations of prior datasets.
• To facilitate improved predictive analysis and grid integration strategies for wind energy.

Main Methods:

• Development of the SDWPF dataset, incorporating power generation, wind speed, turbine spatial distribution, and dynamic contextual factors.
• Inclusion of turbine-specific weather information and internal operational status within the dataset.
• Leveraging the SDWPF dataset to host the ACM KDD Cup 2022 data mining competition.

Main Results:

• The SDWPF dataset enriches WPF research with granular, multi-faceted data.
• The ACM KDD Cup 2022, utilizing SDWPF, attracted over 2400 global teams, indicating significant interest and potential for novel forecasting solutions.
• The dataset's comprehensive nature supports more accurate and robust wind power prediction models.

Conclusions:

• The SDWPF dataset represents a significant advancement in resources available for wind power forecasting research.
• The successful ACM KDD Cup 2022 demonstrates the dataset's value and potential to drive innovation in data mining and renewable energy.
• Enhanced WPF capabilities through datasets like SDWPF are critical for the reliable integration of wind energy into power grids.