A research team has recently published a study in the Journal of Remote Sensing (February 2024) that utilizes data augmentation and the LightGBM machine learning model to estimate diffuse and direct solar radiation. Unlike traditional methods that rely on sparsely and unevenly distributed ground-based observations, this innovative approach leverages data from over 2,453 weather stations across China. By using this extensive dataset, the researchers were able to overcome the limitations of sparse and irregularly collected ground-based data.
The key aspect of this study is the novel use of machine learning algorithms, trained on augmented datasets, to
Additionally, the researchers were able to create a new satellite-based dataset as a result of this study that outperforms existing datasets in terms of accuracy. This dataset provides a detailed spatial distribution of solar radiation components, which is crucial for further advancements in solar energy research and deployment. It offers valuable insights that can lead to more efficient and optimized solar energy production.
The newly developed satellite-based dataset boasts superior precision compared to previous datasets and offers a comprehensive spatial analysis of solar radiation components. This advancement is crucial for the solar energy industry as it enables strategic site selection and system optimization, particularly in areas with high solar energy potential.