Intelligent Situation Sensing and Multi-dimensional Presentation for Power Emergency Communications

Abstract

Power emergency communications are very important to smart grids when some natural disasters happen. In this case, situation sensing can provide key scenario information to decision makers and disaster rescuers. Furthermore, the situation information should be presented from different dimensions. This paper studies intelligent situation sensing issues and provides key solutions from theoretical and practice aspects. The tensor theory is used to model multidimensional situation information, and a new definition of situation sensing is provided for power emergency communications. Some artificial intelligent (AI) schemes such as multi-head attention mechanism and federated learning (FL) are used to construct situation sensing schemes. Simulation results indicate that the proposed situation sensing schemes can achieve better sensing performance.