Smart Grid Fault Detection Using Locally Optimum Unknown or Estimated Direction Hypothesis Test

Abstract The security of smart grid systems is under threat as a consequence of sophisticated intrusion and imperceptible faults. To make the smart grid systems more secure, the development of an efficient fault detection approach is significantly important. Considering that the time evolution of a smart grid system can be modeled by a discrete-time linear state space model, we focus on faults that can be described by changes in system matrices of the state space model. Hypothesis testing-based approaches are employed to detect the faults and a few new locally optimum testing procedures are introduced. We present numerical results that verify the superiority of the new approaches in detecting the change of the system and show that this advantage is particularly obvious when the change is small.