Study on Application of Fisher Information for Power System Fault Detection

The ability to accurately detect power system faults is of vital importance for the purpose of isolating malfunctioning equipment and resuming normal operation as soon as possible after a fault occurs. People have used a variety of electric parameters as metrics to identify faults for a long time. The method proposed by this paper departs from the traditional approach by introducing Fisher information (FI) as a measure of the stability of electric signals and as a criterion for making fault decisions. In this way, a non-dimensional positive parameter is used as a single criterion to deliver fault detection for power distribution networks. Firstly, we simplified the formula of FI and then adopted a practical method for calculating values of FI. We demonstrated the application of FI to measure the stability of electric signals. Finally, we combined FI with wavelet analysis to propose a novel technique for phase selection of a power distribution network with a grounding short-circuit fault, namely the wavelet-based Fisher information (WFI). Simulation studies were then carried out to show the feasibility of the proposed method.