期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2017
卷号：95
期号：22
页码：6215
出版社：Journal of Theoretical and Applied
摘要：In this paper, proposed an intelligent technique for diagnosing the internal faults conditions in power transformers. The proposed intelligent technique is the composites of wavelet transform and RNN based FFA optimization technique. Initially, the normal signals are analyzed at the particular time instant. After that, investigate any faults occurred or not in the power transformer with the help of proposed technique. With the utilization of the proposed technique, the current signals of the power transformer is monitored and detected. Initially, the MWT is utilized to extract the features of the signal. In wavelet transform, the feature approximation of the signal is depends on the decompose levels of high and low frequency components. The extracted features are applied to the input of the FFA. The FFA is selected the optimized training dataset for training the RNN. After that the RNN testing process is evaluated the signal and classified the fault signal type. The effectiveness of the proposed technique is evaluated based on the statistical measures like accuracy, sensitivity and specificity qualities. The proposed method is implemented in MATLAB/Simulink platform and compared with the existing techniques.
关键词：Power Transformer; Fault Detection and Classification; Multi-Wavelet Transforms (MWT); Recurrent Neural Network (RNN) and Firefly Algorithm (FFA)
