Wavelet spectrum energy feature extraction based fault detection scheme for synchronous generators

This paper presents a wavelet transform based fault detection scheme for synchronous generators of power system equipment. The proposed method analyzes characterization of faults using a multi resolution analysis and defines a novel feature extraction, which is called wavelet spectrum energy. The multi-resolution signal analysis based on wavelet transform is utilized to decompose a given signal into approximate and detail signals of original signal. The detail signal coefficients are utilized for calculating wavelet spectrum energy. The fault detection technique utilizes the wavelet spectrum energy as feature extraction to extract information of fault signals for transient analysis. The simulation results show accurate discrimination of faults and also in characterization of internal and external faults.