Signal parameter estimation and classification using mixed supervised and unsupervised machine learning approaches

Abstract

The increasing use of modern power electronics raises the issue of harmonics in power systems which ultimately deteriorate its optimal performance in terms of: increased power loss, breaker failure and mal-operation of equipment. It has been found that the most severe harmonics in the system are odd ones due to their unsymmetrical nature. This work presents the new framework for estimation and classification of harmonics using machine learning approaches. Initially, a shallow neural network and fuzzy logic systems are used to estimate the harmonics contents in the voltage and currents signals. Based on the sequence components and IHD level of source signals, the estimation of harmonic content is achieved. The obtained results are compared with the analytically computed data for validating the performance of designed networks. The results from neural and fuzzy systems are then used to train the explainable convolutional neural network (xCNN) for harmonics classification. The xCNN consists of pertained ALEXNET network which trains the standard binary support vector machine (SVM) for classification of harmonics. The dictionary-based approach is used to add the explanations to the SVM classifier output as a prototype. The performance of proposed framework is measured in-terms of accuracy and loss function and evaluated on the basis of its scalability and computability. The proposed approach is called a Human with Machine-In-Loop (HMIL). INDEX TERMS Power quality (PQ), individual harmonics distortion (IHD), neural network (NN), fuzzy inference system (FIS), explainable convolution neural network (xCNN)

A Comprehensive Review of Harmonic Issues and Estimation Techniques in Power System Networks Based on Traditional and Artificial Intelligence/Machine Learning

“…According to the signal processing of power quality in the harmonic domain, the origin of harmonics is due to deviation in waveshape from its fundamental frequency component. A novel framework for harmonic estimation and classification utilizing ML techniques is presented in [82]. The harmonic contents of the voltage and current signals are initially estimated using a shallow neural network and fuzzy logic systems.…”

Wavelet-Based Harmonic Magnitude Measurement in the Presence of Interharmonics

“…Methods based on cosine and mimic filters [12], [13] have demonstrated better performance than the DFT, especially for estimating the fundamental component in signals with decaying direct current (DC) offset [14]. Methods based on artificial intelligence [15]- [17], recursive last squares or least mean-square (LMS) [18], [19], Kalman filter [20]- [22], and other techniques [2], [23] have been used for harmonic estimation. However, none is affirmed to be effective in signals with interharmonics.…”