Power Quality Event Classification Under Noisy Conditions Using EMD-Based De-Noising Techniques

The paper deals with the application of two empirical mode decomposition (EMD)-based de-noising techniques in power quality assessment. Distinct threshold parameters at a distinct level of decomposition are employed for noise exclusion. Each of the thresholded intrinsic mode functions (IMFs) are then combined together to yield a de-noised version of the signal. For enhanced performance, various noisy versions from the original signal are created and then de-noised using soft thresholds. The resultant de-noised signals are then averaged to obtain the de-noised version of the signal. In other procedures based on EMD, the signal is chopped into smaller portions each having equal length. Each signal portion is then separately subjected to EMD and then de-noised and later combined to obtain noise-free signal. These de-noised signals are then subjected to Hilbert transform for feature extraction. Fuzzy Product Aggregation Reasoning Rule-based intelligent classifier is used here for classification purpose. A comparative study is also made between S-transform-based and wavelet-transform-based de-noising techniques paper. Real-time implementation of the algorithm on a noisy harmonic signal is also given in this paper.