Fazel Mohammadi - Department of Electrical and Computer Engineering, University of Windsor, Windsor, ON N۹B ۱K۳, Canada
Chuyi Zheng - Department of Civil and Environmental Engineering, University of Windsor, ON N۹B ۱K۳, Canada
Rumei Su - College of Hydraulic and Environmental Engineering, Changchun Institute of Technology, Changchun ۱۳۰۰۰۰, China

In this paper, a Wavelet Transform (WT) based on data analysis is proposed to extract the features from real-time active power and RMS (Root Mean Square) voltage of the power grid and use a hybrid classification technique based on Particle Swarm Optimization (PSO) and Support Vector Machine (SVM) to classify the features and diagnose different types of faults in smart grid system. Different PSO-SVM models have been used for training to detect the fault according to P-t characteristics and determine the type of fault based on V-t characteristics of the power grid. Simulations were carried out on the IEEE 9-bus test system considering the temporary and permanent open-circuit faults on the power system. The simulation results show the accuracy, effectiveness, and robustness of the proposed method.

Fault Detection, Fault Identification, Particle Swarm Optimization (PSO), Smart Grid (SG), Support Vector Machine (SVM), Wavelet Transform (WT)