Time-Frequency and Artificial Neural Network Applications and Analysis for Electrical System Power Quality Disturbances in MATLAB
Aslam P. Memon1, M. Aslam Uqaili2, Zubair A. Memon3, Naresh K. Tanwani4
1Aslam P. Memon, PhD Scholar, Department of Electrical Engineering, Mehran University of Engineering, & Technology, Jamshoro, Sindh Pakistan.
2M. Aslam Uqaili, Meritorious Professor, Department of Electrical Engineering, Mehran University of Engineering, & Technology, Jamshoro, Sindh Pakistan.
3Zubair A. Memon, Meritorious Professor, Department of Electrical Engineering, Mehran University of Engineering, & Technology, Jamshoro, Sindh Pakistan.
4Naresh K. Tanwani, PhD Scholar, Department of Electrical Engineering, Mehran University of Engineering, & Technology, Jamshoro, Sindh Pakistan.
Manuscript received on 11 January 2014 | Revised Manuscript received on 20 January 2014 | Manuscript Published on 30 January 2014 | PP: 91-98 | Volume-3 Issue-8, January 2014 | Retrieval Number: H1454013814/14©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: In recent years due to increasing utilization of nonlinear loads and power electronic equipment, the issue of EPQD (Electrical power quality disturbances) has become the most important apprehension for suppliers and the users of electric power. It is imperative to detect the sources and causes of electrical power quality disturbances in order to improve EPQ problems. Traditional signal processing techniques permit mapping signals from time to frequency domains by decomposing the signals into several frequency components. Due to this transformation time information is lost. EPQ disturbances vary in the wide range of time and frequency, which means these traditional techniques are not suitable for EPQ problems. This problem can be solved with the application of WT (Wavelet transform) and feedforward neural networks as classifier. Statistical features extraction data is obtained using DWT (discrete wavelet transformation) and MRDA (multiresolution decomposition analysis) utilizing MATLAB/Simulink and Wavelet toolbox. This minimum feature vector data is used for training FFNN as input. Proposed FFNN classifier reduces training. The results obtained show the promising applicability and suitability of WT analysis with neural network for improved and an efficient methodology for automatic diagnosis of EPQ problems.
Keywords: Detection And Classification, Discrete Wavelet Transform, Electrical Power Quality Disturbances, Feedforward Neural Network, Wavelet Transform.
Scope of the Article: Adaptive Networking Applications