Detection and classification of power quality disturbances in wind-grid integrated system using fast time-time transform and small residual-extreme learning machine

Abstract

The proposed work accomplishes detection and classification of simple and complex power quality disturbances occurring in the realm of wind-grid integrated system using fast time-time transform and small residual-extreme learning machine. Taking the advantage of time-time transform, this work has further tailored time-time transform by accommodating the dyadic scaling to make it a computationally less complex and faster technique. Here, threephase power quality signals are first segmented into single phases to be analyzed with fast time-time transform for getting the event information in timetime domain. The event information is further passed through two-stage feature extraction process to bring out twenty two characteristic curves; each of them is further used to compute five statistical features to produce 107 features after ignoring three redundant features. Thus, a high-dimensional feature set of size 321 is obtained corresponding to three phases of voltage signal. Because, high-dimensional feature set demands a robust classifier that can accurately classify the disturbances even in the presence of noise, small residual-extreme learning machine has been trained to classify the power quality signal into 12 classes (symmetrical sag, asymmetrical sag, swell, unbalance, harmonics, notch, momentary interruption, normal voltage signal, sag with harmonics, swell with harmonics, sag and notch with harmonics, and swell and notch with harmonics). This simulation-based study has been proved effective in dealing with noisy power quality signals also with reduced computational complexity and higher classification rate. KEYWORDS fast time-time transform, power quality, small residual-extreme learning machine classifier, wind-grid integrated system List of symbols and abbreviations: h(t), a time signal; w(t), Gaussian window; σ, standard deviation; f, frequency; β, output weight of SR-ELM network; G, output of hidden layer of SR-ELM network; X, input to SR-ELM; L, number of neurons in hidden layer of SR-ELM; M, number of neurons in output layer of SR-ELM; e, error; C. B.

A new method with minimum number of monitoring points for flicker source tracing by wavelet transform

“…However, several unanticipated irregularities happen in the power supply due to different reasons, which in turn can cause many problems for consumers like malfunction and failure of their equipment. Consequently, the power quality has been turned into a growing concern in power grids . Among power quality problems, voltage fluctuations (flickers) have attracted more attention because of their numerous and frequent happenings .…”

“…Consequently, the power quality has been turned into a growing concern in power grids. [1][2][3][4] Among power quality problems, voltage fluctuations (flickers) have attracted more attention because of their numerous and frequent happenings. [5][6][7][8][9] In general, voltage flicker is a periodical or fortuitous deviations of the voltage waveform.…”

Power Quality Disturbance Monitoring and Classification Based on Improved PCA and Convolution Neural Network for Wind-Grid Distribution Systems

“…WT and least squares SVM [64] considered only four types of three-phase PQ disturbances with simulated dataset having classification accuracy of 99.71%, but in the proposed method 12 types of three-phase PQ disturbance are considered with a significantly higher classification rate. Fast time-time transform (FTT) and small residual extreme learning machine (SR-ELM) were studied [4]. The classification rate for 12 types of three-phase simulated PQ disturbances was achieved as 99.59, and 107 features were selected, but no real data was evaluated in this paper.…”

“…Some of the principal manufacturers of PQ analyzers are Fluke (Everett, WA, USA) Yokogawa (Tokyo, Japan), and FLIR (Wilsonville, OR, USA). PQ analyzer solutions with basic functionality are expensive yet they cannot analyze the complex and extensive data [4]. Non-stationary PQ disturbances occur due to fluctuations and loads, which change the capability of the signals.…”

Classification of multiple power quality events via compressed deep learning

“…Generally, in literature after extraction of PQD signal, the mean root‐mean‐square (rms) values of current and voltage, the coefficient of skewness, standard deviation, entropy, energy, etc, are extracted, and used as a feature vector and input to a classifier. Nevertheless, these techniques are not good enough for multiple disturbances …”