Discrimination of power transformer inrush and internal fault currents using signal processing and soft computing techniques

Abstract

Power transformers are important equipment in power systems and their protection schemes are of vital significance to provide continuous power supply. This thesis presents several signal processing and soft computing methods for the discrimination of the inrush current from internal fault currents. The proposed protection scheme has been analyzed for varying switching angles, saturation points, different winding configurations, and faults between the winding terminals and the energization of the transformer. These sampled differential currents collected at the relaying point are used for the discrimination of the inrush and internal fault currents. In the first approach the criterion function is defined, based on the slope patterns and the amplitude of the wavelet coefficients from detail 5, using Daubechies (db 9) mother wavelet in a specific frequency band. In another approach, different operating conditions are analyzed, using the wavelet transform using the moving window concept. Then, features like the energy and Standard Deviation (STD), are computed from the wavelet coefficients using Parseval s theorem in both SVM and PNN, for fault classification. The proposed scheme studied for popular mother wavelet like Daubechies, Symmetry, Coiflets Biorthogonal and Haar. The classification accuracy for the combination wavelet and SVM gives better results. The decomposition signals are analyzed only by approximation coefficients. The data of the different operating conditions are processed to the TTtransform; then the features are extracted and used in the PNN and SVM for distinguishing the inrush and fault currents. The TT-transform is a superior way of distinguishing between the inrush and fault currents. The classification results of the TT-transform and PNN gives excellent results. The TTtransform and HS-transform in combination with the PNN and SVM, is therefore, recommended for the protection of large transformers.