DWT with Enhanced ANN Technique for Detecting and Classifying the Faults of Synchronous Generator

Abstract

In the paper, a hybrid technique based fault detector algorithm for synchronous generator is proposed. The hybrid technique is the combination of Artificial Neural Network (ANN) and Gravitational Search Algorithm (GSA). The GSA is used to train the ANN and improve the performance of ANN. Initially, the synchronous generator is analyzed in the normal condition. After that, the fault is created in the synchronous generator and the system behaviors are monitored and signals are measured which can be seen as distorted waveforms. These distorted waveforms are composed of different frequency components and which are needed to be represented in timefrequency domain for fault analysis. For this representation of signal, discrete wavelet transform (DWT) is presented. It extracts the features and forms the datasets which are forwarded to ANN classifier for classifying the type of fault occurred in the stator winding of the synchronous generator. In order to evaluate the effectiveness of the proposed method, the internal faults are analyzed. The proposed technique is implemented in MATLAB/simulink platform and this is validated using statistical measures such as accuracy, sensitivity & specificity. The proposed method is compared with the existing techniques DWT-ANN with GA and DWT-ANN methods.

Optimizing Torque Ripples In Switched Reluctance Motor Via ANN

“…[24] In conclusion, even though SRMs have many benefits, such as fault tolerance and a straightforward design, resolving torque ripple is still a major obstacle to their wider use. [17] A thorough understanding of the many contributing elements, such as dc-link voltage, load current, phase voltage, inductance, rotor position, and speed, is necessary for the design of an ideal output power management system for SRGs. [25] II PROPOSED SCHEME: The fast speed and lower cost of the three-phase, six-four-pole SR motor are its benefits.…”