A Practical Algorithm for Fault Classification in Distribution Grids

The ongoing transition from conventional to smart grids has created the necessity and, at the same time, the opportunity for the automation of the fault isolation process. Fault isolation includes the classification of the type of fault, followed by the location of the fault. Many methods have been developed over the years on the topic of fault isolation, however their practical implementation is still limited, and considered a challenge for many Distribution System Operators (DSOs). Although fault classification is not always necessary, it is essential for many fault location methods. Therefore, the present research introduces a practical method that can accurately and quickly identify the type of fault occurring in a distribution grid. Several algorithms consisting of complex current and voltage criteria were developed and compared. The algorithms were tested in Matlab, on the IEEE 13 node test feeder, simulated in Simulink, regarding the effect that parameters like the fault's location, the noise in the measurements and the fault resistance have on their accuracy. The final method showed satisfactory results and the ability to accurately identify all kinds of faults in various cases.