Random Forest Based Fault Classification Technique for Active Power System Networks

In recent times, integration of distributed energy resources with conventional power networks has been increased rapidly and with that several interlinking converters and power electronic devices are there. This increases the complexity of the system. In this paper, a fault classification technique based on random forest classifier has been proposed. As the random forest tree is an artificial intelligence tool, therefore, it is guaranteed the results obtained are of high accuracy value. The high accuracy in fault detection and classification is highly needed for a power system network to eradicate the fault from the system. This method has been tested over both transmission and distribution networks to show the efficacy of this proposed method, where the distribution network is a modified practical Indian distribution grid. Also a comparative study of this method with existing classification techniques like SVM, KNN and others has been done.