Global Wavelet Spectrum based fault detection for Self-Excited Induction Generator

This work proposes a fault detection procedure for Self-Excited Induction Generator (SEIG) in power generation standalone applications. The fault detection methodology proposed is based on digital signal processing techniques applied to the available SEIG electrical signals. In order to get the enough data to perform the analysis, the dynamic model of SEIG was simulated in Matlab/Simulink ® platform. The simulated models describe the machine behavior under healthy and several faulty conditions. The main characteristics of the signals were obtained by means of the Continuous Wavelet Transform and the Global Wavelet Spectrum applied on the stator voltages of the generator. A particular pattern, for each operating condition, was observed and quantified. Consequently, the fault detection technique is characterized by a low computational cost for practical implementation.