Power Quality Disturbance Recognition Based on Wavelet Transform and Convolutional Neural Network

2021 ◽  
Author(s):  
Wenhui Hong ◽  
Ziwen Liu ◽  
Xuyan Wu
Keyword(s):  
Neural Network ◽  
Wavelet Transform ◽  
Convolutional Neural Network ◽  
Power Quality ◽  
Power Quality Disturbance

scholarly journals Classifying Power Quality Disturbances Based on Phase Space Reconstruction and a Convolutional Neural Network

2019 ◽  
Vol 9 (18) ◽  
pp. 3681 ◽  
Author(s):  
Kewei Cai ◽  
Taoping Hu ◽  
Wenping Cao ◽  
Guofeng Li
Keyword(s):  
Neural Network ◽  
Phase Space ◽  
Convolutional Neural Network ◽  
Power Quality ◽  
Hybrid Approach ◽  
Phase Space Reconstruction ◽  
Feature Maps ◽  
Image File ◽  
Mode Decomposition ◽  
Power Quality Disturbance

This paper presents a hybrid approach combining phase space reconstruction (PSR) with a convolutional neural network (CNN) for power quality disturbance (PQD) classification. Firstly, a PSR technique is developed to transform a 1D voltage disturbance signal into a 2D image file. Then, a CNN model is developed for the image classification. The feature maps are extracted automatically from the image file and different patterns are derived from variables in CNN. A set of synthetic signals, as well as operational measurements, are used to validate the proposed method. Moreover, the test results are also compared with existing methods, including empirical mode decomposition (EMD) with balanced neural tree (BNT), S-transform (ST) with neural network (NN) and decision tree (DT), hybrid ST with DT, adaptive linear neuron (ADALINE) with feedforward neural network (FFNN), and variational mode decomposition (VMD) with deep stochastic configuration network (DSCN). Based on deep learning algorithms, the proposed method is capable of providing more accurate results without any human intervention for PQDs. It also enables the planning of PQ remedy actions.

scholarly journals Application of Extension Neural Network with Discrete Wavelet Transform and Parseval’s Theorem for Power Quality Analysis

2019 ◽  
Vol 9 (11) ◽  
pp. 2228 ◽  
Author(s):  
Shiue-Der Lu ◽  
Hong-Wei Sian ◽  
Meng-Hui Wang ◽  
Rui-Min Liao
Keyword(s):  
Neural Network ◽  
Wavelet Transform ◽  
Discrete Wavelet Transform ◽  
Power Quality ◽  
Discrete Wavelet ◽  
Correlation Degree ◽  
Power Quality Disturbance ◽  
Propagation Network ◽  
Parseval’S Theorem

The development of renewable energy and the increase of intermittent fluctuating loads have affected the power quality of power systems, and in the long run, damage the power equipment. In order to effectively analyze the quality of power signals, this paper proposes a method of signal feature capture and fault identification, as based on the extension neural network (ENN) algorithm combined with discrete wavelet transform (DWT) and Parseval’s theorem. First, the original power quality disturbance (PQD) transient signal was subjected to DWT, and its spectrum energy was calculated for each order of wavelet coefficients through Parseval’s theorem, in order to effectively intercept the eigenvalues of the original signal. Based on the features, the extension neural algorithm was used to establish a matter-element model of power quality disturbance identification. In addition, the correlation degree between the identification data and disturbance types was calculated to accurately identify the types of power failure. To verify the accuracy of the proposed method, five common power quality disturbances were analyzed, including voltage sag, voltage swell, power interruption, voltage flicker, and power harmonics. The results were then compared with those obtained from the back-propagation network (BPN), probabilistic neural network (PNN), extension method and a learning vector quantization network (LVQ). The results showed that the proposed method has shorter computation time (0.06 s), as well as higher identification accuracy at 99.62%, which is higher than the accuracy rates of the other four types.

Pembangunan Satu Kaedah Cekap dalam Pengesanan dan Pengkelasan Gangguan Kualiti Kuasa

2012 ◽  
Author(s):  
Ramizi Mohamed ◽  
Azah Mohamed ◽  
Aini Hussain
Keyword(s):  
Neural Network ◽  
Wavelet Transform ◽  
Discrete Wavelet Transform ◽  
Power Quality ◽  
Predictive Coding ◽  
Discrete Wavelet ◽  
Linear Predictive Coding ◽  
Power Quality Disturbance ◽  
Voltage Swell ◽  
Power Quality Disturbances

Pengesanan dan pengkelasan data gangguan kualiti kuasa secara automatik telah menjadi penting terutamanya untuk menangani masalah gangguan pangkalan data yang besar. Kertas kerja ini membentangkan satu kaedah cekap dalam pengesanan dan pengkelasan gangguan kualiti kuasa. Kaedah yang dicadangkan untuk mengesan gangguan adalah berdasarkan penjelmaan anak gelombang diskrit dan pengekodan ramalan lelurus manakala kaedah yang telah dibangunkan untuk mengkelaskan gangguan adalah berdasarkan rangkaian neural tiruan (RNT). Sebelum pelaksaan RNT, isyarat gangguan dikesan terlebih dahulu untuk mendapatkan pekali anak gelombang kuasa dua dan pekali pengekodan ramalan lelurus. Pekali ini mewakili sifat bagi berbagai jenis gangguan dan digunakan sebagai data masukan kepada RNT yang telah dibina. Oleh itu, anak gelombang dan pengekodan ramalan lelurus digunakan sebagai prapemprosesan isyarat gangguan yang kemudiannya disambungkan kepada RNT. Dalam pelaksanan RNT, model rangkaian neural lapisan berbilang dengan algoritma perambatan ke belakang telah dipertimbangkan. Reka bentuk RNT yang telah dibangunkan adalah berbentuk hierarki dan modular supaya RNT yang berasingan dikhaskan untuk mengkelas berbagai jenis gangguan dan juga gangguan dengan kadar persampelan yang berbeza. Keputusan yang diperolehi menunjukkan bahawa kaedah anak gelombang dan pengekodan ramalan lelurus adalah sangat berkesan untuk mengesan gangguan kualiti kuasa dan kaedah RNT pula dapat mengkelaskan dengan jitu gangguan kualiti kuasa seperti lendut voltan, ampul voltan, fana dan takukan. Kata kunci: Kualiti kuasa; anak gelombang; pengekodan ramalan lelurus; rangkaian neural Automated power quality disturbance detection and classification is preferred so as to enable faster and more efficient analysis of a disturbance large database. This paper presents an efficient method to detect and classify some power quality disturbances. The proposed method for detecting the disturbances is based on discrete wavelet transform and linear predictive coding whereas the method for classifying the disturbances is based on artificial neural network (ANN). Prior to the ANN implementation, the disturbance signals are first detected by the discrete wavelet transform and the linear predictive coding techniques to obtain the squared wavelet transform coefficients and the linear predictive coding coefficients. These features represent the various disturbances and serve as inputs to the developed ANNs. Therefore, wavelets and linear predictive coding are employed as a preprocessing stage and is connected to the ANN. In the ANN implementation, the multilayer perceptron neural network model and the backpropagation algorithm are considered. The design of the developed ANNs are hierarchical as well as modular in nature so that separate ANNs are dedicated to classify the various types of disturbances and to handle the disturbances with different sampling rates. The results obtained show that the wavelets and the linear predictive coding methods are effective in detecting power quality disturbances and the ANNs can accurately classify the disturbances such as voltage sag, voltage swell, transients and notching. Key words: Power quality; wavelets; linear predictive coding; neural networks

A Practical Approach to Detection and Analysis of Time-Varying Waveform

2011 ◽  
Vol 48-49 ◽  
pp. 1-4 ◽  
Author(s):  
Jing Bo Liu ◽  
Gui Fen Zhao
Keyword(s):  
Neural Network ◽  
Wavelet Transform ◽  
Power Quality ◽  
Scaling Properties ◽  
Fundamental Component ◽  
The Neural Network ◽  
Power Quality Disturbance ◽  
Novel Method ◽  
Transform Coefficients ◽  
Disturbance Pattern

The power quality has attracted considerable attention from both utilities and users due to the development of sensitive electronic equipment in power system network. A novel method for detecting and localizing power quality disturbance based on wavelet neural network is presented. The wavelet transform is established on decomposition of a signal according to time-scale using basis functions with adaptable scaling properties known as multiresolution analysis. The wavelet transform expands a signal not in terms of a trigonometric polynomial but by wavelets, generated using transition and dilation of a fixed wavelet function. The feature obtained from wavelet transform coefficients are inputted into the neural network for power quality disturbance pattern recognition. The improved network algorithm is utilized to complete the network structure initialization. The fundamental component of the signal is estimated to extract the mixed information using wavelet network, and then the disturbance is acquired by subtracting the fundamental component. The simulation results show that the proposed method is more sensitive to signal singularity, and achieves better performance over traditional method.

scholarly journals Automatic Detection of Power Quality Disturbance Using Convolutional Neural Network Structure with Gated Recurrent Unit

2021 ◽  
Vol 2021 ◽  
pp. 1-11
Author(s):  
Enes Yiğit ◽  
Umut Özkaya ◽  
Şaban Öztürk ◽  
Dilbag Singh ◽  
Hassène Gritli
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Power Quality ◽  
State Of The Art ◽  
Current State ◽  
Frequency Domains ◽  
Power Quality Disturbance ◽  
Neural Network Structure ◽  
Short Time ◽  
Gated Recurrent Unit

Power quality disturbance (PQD) is essential for devices consuming electricity and meeting today’s energy trends. This study contains an effective artificial intelligence (AI) framework for analyzing single or composite defects in power quality. A convolutional neural network (CNN) architecture, which has an output powered by a gated recurrent unit (GRU), is designed for this purpose. The proposed framework first obtains a matrix using a short-time Fourier transform (STFT) of PQD signals. This matrix contains the representation of the signal in the time and frequency domains, suitable for CNN input. Features are automatically extracted from these matrices using the proposed CNN architecture without preprocessing. These features are classified using the GRU. The performance of the proposed framework is tested using a dataset containing a total of seven single and composite defects. The amount of noise in these examples varies between 20 and 50 dB. The performance of the proposed method is higher than current state-of-the-art methods. The proposed method obtained 98.44% ACC, 98.45% SEN, 99.74% SPE, 98.45% PRE, 98.45% F1-score, 98.19% MCC, and 93.64% kappa metric. A novel power quality disturbance (PQD) system has been proposed, and its application has been represented in our study. The proposed system could be used in the industry and factory.

scholarly journals Automatic ECG Classification Using Continuous Wavelet Transform and Convolutional Neural Network

Entropy ◽  
2021 ◽  
Vol 23 (1) ◽  
pp. 119
Author(s):  
Tao Wang ◽  
Changhua Lu ◽  
Yining Sun ◽  
Mei Yang ◽  
Chun Liu ◽  
...  
Keyword(s):  
Neural Network ◽  
Wavelet Transform ◽  
Convolutional Neural Network ◽  
Continuous Wavelet Transform ◽  
Continuous Wavelet ◽  
Time Frequency ◽  
Ecg Signals ◽  
Rr Interval ◽  
Frequency Components ◽  
Fully Connected

Early detection of arrhythmia and effective treatment can prevent deaths caused by cardiovascular disease (CVD). In clinical practice, the diagnosis is made by checking the electrocardiogram (ECG) beat-by-beat, but this is usually time-consuming and laborious. In the paper, we propose an automatic ECG classification method based on Continuous Wavelet Transform (CWT) and Convolutional Neural Network (CNN). CWT is used to decompose ECG signals to obtain different time-frequency components, and CNN is used to extract features from the 2D-scalogram composed of the above time-frequency components. Considering the surrounding R peak interval (also called RR interval) is also useful for the diagnosis of arrhythmia, four RR interval features are extracted and combined with the CNN features to input into a fully connected layer for ECG classification. By testing in the MIT-BIH arrhythmia database, our method achieves an overall performance of 70.75%, 67.47%, 68.76%, and 98.74% for positive predictive value, sensitivity, F1-score, and accuracy, respectively. Compared with existing methods, the overall F1-score of our method is increased by 4.75~16.85%. Because our method is simple and highly accurate, it can potentially be used as a clinical auxiliary diagnostic tool.

Sign in / Sign up

Export Citation Format

Share Document