In the heart of China, researchers at Northeast Electric Power University are tackling a critical issue in the wind energy sector: the prevalence of abnormal data in wind power measurements. This problem, often overlooked, significantly hampers the accuracy of wind power predictions and the overall efficiency of wind farms. Led by Mao Yang, a key figure in the university’s Key Laboratory of Modern Power System Simulation and Control & Renewable Energy Technology, a groundbreaking study has been published that promises to revolutionize how we handle and interpret wind power data.
Wind farms generate vast amounts of data, but not all of it is reliable. Abnormal data points can skew analyses and reduce the predictive accuracy of wind power systems. “The high availability of accurate wind power data is crucial for understanding wind power patterns and improving prediction models,” Yang explains. “However, the presence of abnormal data points can seriously affect these analyses, leading to inefficiencies and increased costs.”
The research team identified two main types of abnormal data: outlier abnormal points and accumulation abnormal points. To address these, they developed a sophisticated method that combines the Cluster-Based Local Outlier Factor (CLOF) algorithm with conditional constraints based on physical background knowledge. This approach allows for the precise identification of abnormal data points, ensuring that only high-quality data is used in analyses.
But identifying the problem is only half the battle. The team also developed a method to reconstruct the abnormal data segments. They divided the data into fluctuation and stationary segments based on wind speed characteristics. For fluctuation segments, they employed a Bidirectional Gate Recurrent Unit (BiGRU) model, which uses wind speed as input to reconstruct the data. For stationary segments, they utilized a bi-directional weighted random forest model. This dual approach ensures that both types of data segments are accurately reconstructed, leading to a more reliable dataset.
The implications of this research are vast. Accurate wind power prediction is essential for grid stability and the efficient integration of renewable energy sources. By improving the quality of wind power data, this method can enhance the reliability of wind power predictions, reduce operational costs, and increase the overall efficiency of wind farms. “This method not only identifies abnormal data but also reconstructs it to a high standard, significantly improving the accuracy of wind power predictions,” Yang states.
The study, published in the Chinese Society for Electrical Engineering Journal of Power and Energy Systems, translates to the English title of ‘Journal of Power and Energy Systems,’ marks a significant step forward in the field of wind energy. As the world continues to shift towards renewable energy sources, the ability to accurately predict and manage wind power will become increasingly important. This research provides a robust framework for addressing one of the key challenges in this area, paving the way for more efficient and reliable wind power systems.
The energy sector is on the cusp of a transformation, and this research from Northeast Electric Power University is at the forefront of that change. As wind power continues to grow in importance, the ability to accurately predict and manage this resource will be crucial. This study offers a promising solution, one that could shape the future of wind energy and contribute to a more sustainable and efficient energy landscape.
