Fault Detection Using Artificial Neural Network and Wavelet Transforms for Power Transformer

Abstract: The requirement for a stable supply of electrical energy to meet the needs of the modern world has expanded dramatically, necessitating near-faultless power system functioning. The main goal is to reduce the frequency and duration of undesired power transformer outages by imposing a high point demand that includes criteria for dependability (no false tripping) and operating speed (quick fault detection and clearance time). For many years, the second harmonic restraint concept has been widely applied in industrial applications. It employs the discrete Fourier transform (DFT) and frequently confronts issues like lengthy restrain times and the inability to distinguish internal defects from magnetizing inrush circumstances. As a result, artificial neural networks (ANNs), a strong tool for artificial intelligence (AI) that can imitate and automate information, have been suggested for defect identification and tracking in normal and inrush conditions. For the investigation of power transformer transient conditions under diverse settings, the wavelet transform (WT) is utilized, which has the capacity to extract information from transient signals in both the time and frequency domains at the same time. In the MATLAB/SIMULINK environment, all of the above-mentioned conditions of a power transformer to be investigated in a power system are modelled.