Progressively Detect Faults in a Large Power System: A Visual Analytics Approach

With rapid expanding in power system equipment, enormous attention is now paid on developing methods to detect and reason faults in large power system data set. Existing techniques mainly focus on data analyzing methods, while little work studies visual analytics approaches. In this paper, we propose a progressively visual analytics procedure (i.e., overview, detecting candidate faulted buses, locating buses and check reasons) to detect fault origins and check possible reasons. In particular, an efficient clustering algorithm associated with a novel visualization for time-varying multivariate as well as multiple and coordinated views are developed to locate faulted buses easily and quickly. Case studies have indicated that our approach is much beneficial to real-life applications.