Fault Diagnosis in Smart Distribution System Using Smart Sensors and Entropy

Power quality and reliability of electric power supply have become indispensable in today’s digital world. Advanced measurement, sensing and communications are available in smart grid environment. Using advanced infrastructure, tools and techniques, modern researchers are trying to device sophisticated methods for fault diagnosis. This paper proposes a technique using nonconventional symlet mother wavelet function to carry out fault diagnosis process. The technique is discussed for extracting entropy of fault transient signal and is used for pattern recognition. The algorithm is presented which is developed using MATLAB software for fault identification, classification and location tasks. The method is implemented on a 9-bus system model, and the results are discussed. The results show effectiveness of symlet nonconventional wavelet function for feature extraction task. The performance indicates the applicability of the method to fault identification, classification and location tasks. The result proves superiority of the method over other methods of feature selection. The method is useful for real-time monitoring and automation purpose of power system if developed further.