A novel method for quality analysis of power flow samples

This article presents a series of quality analysis methods for power flow samples to support artificial intelligence learning algorithms. These mechanisms extract adequate physical quantities from a large amount of flow physical information, simplifying the complexity of sample quality analysis. The proposed method can effectively evaluate the redundancy of power flow samples and suggests a practical design for generating representative flow samples. The complete mechanism includes quality analysis, physical information selection, and power flow sample dataset re-integration, which helps data-dependent AI algorithms to be more convenient and efficient in the application to power systems.