Spatial Distance-based Uniformity Evaluation Method for Power Flow Datasets

Research on the application of artificial intelligence (AI) methods in power grid analysis has been widely conducted. The flow dataset required for training AI models is not uniformly distributed, and to improve quality dataset, related research generally applied tuning that tend have a more uniform sample distribution, resulting improved performance models. However, lack evaluate uniformity distribution poses an obstacle judgment validity adjustment impact distributed datasets This paper proposes method by using calculation distances high-dimensional feature space plotting minimum distance statistics as way datasets. It was also tested 36-node evaluated separately two levels, which can clearly represent space.