Classification of power quality disturbances based on random matrix transform and sparse representation

A novel ensemble neural network structure is presented for automatic classification of power quality disturbances. Power quality (PQ) disturbances analysis is the focus of power quality control. The characteristics of PQ disturbances include short duration, variety of types and so on. Power quality disturbances classification is the foundation of power quality control automation. Different types of Neural network, such as BP neural networks, RBF neural networks and probabilistic neural network etc, is already applied in the area of PQ disturbances classification and recognition. The researches about the neural network for PQ disturbances recognition are mainly focused on the optimizing for the signal type of neural network. But the accuracy rate of the classification is still needed to be improved. Ensemble and hybrid algorithms research is currently flourishing in pattern classification machine learning and decision sciences. Compare to traditional NN, the ensemble and hybrid NN classifier achieves higher classification rate. In this paper, a novel PQ classification system using S-transform and ensemble and hybrid NN is designed. There are 2 stages in the novel system. Firstly, the PQ disturbances signals are transformed by S-transform and the subset of features extracted from the result of S-transform is used as the input vector of the ensemble and hybrid NN. Secondly in the pattern classification process, BP network and RBF neural network are utilized as two classification agents. Through choosing different parameters and different samples, every agent includes a group of neural networks. The classification results, generated by different agents, are fuzzified into fuzzy numbers. The centroid of all fuzzy numbers is compared with the threshold. Finally we obtain the classification results. In the simulation, 6 types of disturbances signals which are simulated by Matlab 7.0 use for test the new classification system. Simulation result shows that when the new system has better classification rate especially at the high noising environment.

Power quality disturbances (PQDs) can significantly affect the reliability of electrical power systems, leading to potential equipment damage and operational inefficiencies. Accurate classification of these disturbances is essential for ensuring continuous and reliable service. The study proposes a deep transfer learning (TL) approach for PQD classification. In the proposed work, various single and multiple PQD signals pertaining to 15 different classes have been generated using mathematical models of PQDs adhering to the guidelines of the IEEE 1159 and IEC 61000-4-30 standards. Time-domain PQD signals are first converted into 2D color images using continuous wavelet transform (CWT). These images are then used to re-train modified pre-trained models such as GoogleNet, SqueezeNet, ResNet-18, and ShuffleNet on synthetic PQD data. Various single and combined PQDs are then classified using trained models. Moreover, the performance of the trained models is evaluated with the PQD signals containing noise of various signal-to-noise ratios (SNR), as well as PQD signals collected from the experimental setup in the laboratory. The results signify that the GoogleNet model exhibits consistent performance for classifying PQDs under various conditions, achieving classification accuracy of 99.8% for synthetic noiseless signals, 98.87% for signals with 20 dB SNR, and 98.89% for signals acquired through experimental setup. Furthermore, the trained models were tested using real-world PQD signals, comprising 26 sag and 42 impulse signals. The GoogleNet model achieved the highest classification accuracy, correctly identifying 23 sag and 34 impulse events, thereby demonstrating real-world applicability and robustness of the proposed approach.

ABSTRACTThis paper presents a newly developed compressive sensing (CS), maximum likelihood (ML), and rule-based method for classification of power quality disturbances (PQDs). The first PQD signals are sampled using compressive sampling method and reduced PQD signal dimensionally, and then the data are directly used as features for the classification. PQD signals are divided into two groups: at first, impulsive, oscillatory, harmonics notch, flicker and DC offset, and the second, sag, swell, and interruption. Compressive sensing maximum likelihood (CSML) classifier is used for the classification of the first group and the rule-based method is used for the classification of the second group. CSML classifier and the rule-based method are tested utilizing 1000 PQD events under noiseless and noisy environment. About 100% correct classification rate is achieved without noise and 99.8% correct classification result is obtained with noise. In both methods, noise was considered from 20 to 40 dB. Results show that the proposed CS and ML algorithms and rule-based method can be efficiently used for PQD classification.

This paper deals with a modified technique for the recognition of single stage and multiple power quality (PQ) disturbances. An algorithm based on Stockwell's transform and artificial neural network-based classifier and a rule-based decision tree is proposed in this paper. The analysis and classification of single stage PQ disturbances consisting of both events and variations such as sag, swell, interruption, harmonics, transients, notch, spike, and flicker are presented. Moreover, the proposed algorithm is also applied on multiple PQ disturbances such as harmonics with sag, swell, flicker, and interruption. A database of these PQ disturbances based on IEEE-1159 standard is generated in MATLAB for simulation studies. The proposed algorithm extracts significant features of various PQ disturbances using S-transform, which are used as input to this hybrid classifier for the classification of PQ disturbances. Satisfactory results of effective recognition and classification of PQ disturbances are obtained with the proposed algorithm. Finally, the proposed method is also implemented on real-time PQ events acquired in a laboratory to confirm the validity of this algorithm in practical conditions.

The classification of power quality (PQ) disturbances to improve the PQ is an important issue in utilities and industrial factories. In this paper, an approach to classify PQ disturbances is presented. First, a signal containing one of the PQ disturbances, like sag, swell, flicker, interruption, transient, or harmonics, is evaluated using the proposed approach. Afterwards, S-transform and TT-transform are applied to the signal and an artificial neural network is used to recognize the disturbance using S-transform and TT-transform data, like the variance and mean values of S-transform and TT-transform matrices. The main features of the proposed approach are the real-time and very fast recognition of the PQ disturbances. Finally, the proposed method's results are compared with the support vector machine and k-nearest neighbor classification methods to verify the results. The results show the effectiveness of this state-of-the-art approach.

The widespread integration of renewable energy sources (RESs) into power systems using power electronics-based interface devices has led to a substantial rise in power quality (PQ) issues. There is an immediate requirement for effective monitoring, detection, and classification of power quality disturbances (PQDs) that is needed to take remedial measures and design planning of the system architecture. This study presents a hybrid approach with an objective for the feature extraction and classification of PQDs. The proposed hybrid approach is comprised of an optimized tunable-Q wavelet transform (OTQWT) for the feature extraction and incremental support vector machine (ISVM). A four-stage approach is suggested for the PQ detection and classification in this study. In the first stage, the various data are retrieved both in the form of synthetic data by mathematical formulations and real-time data with prototype design setup. In the second stage, regardless of the specified wavelet function, the PQD signals are decomposed into low-pass and high-pass sub-bands using the tunable-Q wavelet transform (TQWT). However, the utilization of default decomposition parameters to address nonstationary PQ signals may lead to information loss and reduced performance of the system. To avoid this limitation, an OTQWT as an enhanced technique to TQWT based on an Adaptive Particle Swarm Optimization (APSO) is suggested. A modified objective function based on the mean square error (MSE) is used to improve the decomposition process. In the third stage, an efficient classifier is suggested based on the ISVM. Lastly, to test and evaluate the performance of the proposed approach, twelve types of PQDs including noise and multiple occurrences are considered. The comparative analysis with other popular methods reflects the better performance of the proposed approach and justifies its use for PQ detection and classification purposes in real-time​ conditions.

This paper develops a knowledge-based neural network (KBNN) for the classification of power quality (PQ) disturbances. Initially, the tunable-q wavelet transform (TQWT) is employed for the extraction of the 50 Hz component from a voltage signal with any sort of disturbance. This is achieved by varying the quality factor of wavelet according to the signal information. The KBNN is a combined model of neural network and rule-based approach. This paper explores the potential of the KBNN for classification of the most common power quality disturbances. The efficacy of the KBNN approach is evaluated on a wide range of time-varying signals with noise, fundamental frequency deviation, and variation in signal parameters. The performance analysis elucidates the efficiency and robustness of the proposed approach using KBNN classifier for classification of the normal and eight PQ disturbances.

Accurate classification of power quality disturbance is the premise and basis for improving and governing power quality. A method for power quality disturbance classification based on time-frequency domain multi-feature and decision tree is presented. Wavelet transform and S-transform are used to extract the feature quantity of each power quality disturbance signal, and a decision tree with classification rules is then constructed for classification and recognition based on the extracted feature quantity. The classification rules and decision tree classifier are established by combining the energy spectrum feature quantity extracted by wavelet transform and other seven time-frequency domain feature quantities extracted by S-transform. Simulation results show that the proposed method can effectively identify six types of common single disturbance signals and two mixed disturbance signals, with fast classification speed and adequate noise resistance. Its classification accuracy is also higher than those of support vector machine (SVM) and k-nearest neighbor (KNN) algorithms. Compared with the method that only uses S-transform, the proposed feature extraction method has more abundant features and higher classification accuracy for power quality disturbance.

Sparse representation in compressed sensing is a hot topic in signal processing and artificial intelligence due to its success in various applications. A general classification algorithm based on sparse representation theory named Sparse Representation Classification (SRC) was successfully applied in face recognition. In this paper, we research the issue of facial expression recognition (FER) via SRC. Extensive experiments of FER via SRC algorithm are carried out to study the performance of sparse representation theory for FER. The comparison of SRC algorithm with various traditional algorithms such as two-dimensional PCA as well as curvelet transform for FER is also given. Support vector machine is used for expression classification. The paper also studies the robustness of SRC algorithm for FER to noises. Satisfactory results are obtained for FER via sparse representation. The experiment results show the effectiveness of SRC algorithm on FER.

This paper presents a novel hybrid approach combining Gramian Angular Summation Field (GASF) method with a convolutional neural network (CNN) to classify power quality disturbances. Firstly, a 1-D Power quality disturbance signal is transformed into a 2-D image file using GASF. Subsequently, CNN is implemented for features extraction and image classification. In this work, the synthetic power quality (PQ) disturbances are considered including nine single disturbances and five mixed disturbances. Further, to capture multi-scale aspects of power quality disturbances problem and reduce overfitting, a unit is designed using 2-D convolutional, pooling, and batch-normalization layers. The classification study is further supported by experimental signals obtained on a prototype setup of PV system. The obtained results demonstrate the efficiency and reliability of the proposed method. The proposed method is compared with the other advanced CNNs and other conventional methods to illustrate its effectiveness.

The power quality (PQ) disturbance signals have the characteristics of short duration and strong randomness, and often form complex disturbances, which make the disturbance signals difficult to detect and identify. In this paper, EMD algorithm is introduced to decompose the PQ disturbance signals and calculate the intrinsic mode function (IMF) of the disturbance signals. Then, Hibert transform are performed for each IMF to obtain the characteristic information of the disturbance signal. EMD transform is used to detect the type, duration, frequency and amplitude of PQ disturbances. To verify the effectiveness of the algorithm, several kinds of PQ disturbance signals are simulated with transient harmonic, voltage interruption, voltage drop and voltage surge and complex disturbances. Experimental results show that the algorithm can accurately detect power quality interferences. This paper provides a new method for the detection of PQ disturbances and a new idea for the power management.

Nowadays, due to the increase in the demand for electrical energy and the development of technology, the electrical devices have a more complex structure. This situation has increased the importance of concept of the power quality in the electrical power system. This paper presents a deep learning-based system to recognize the power quality disturbances (PQDs) in the solar photovoltaic (SPV) plant integrated with power system networks. The PQDs are analyzed using continuous wavelet transform (CWT) and image files are obtained from scalograms of CWT. Then, these image files are used to recognize PQDs with the help of a hybrid deep learning approach based on convolutional neural network (CNN), neighbor component analysis (NCA), and support vector machine (SVM). In this hybrid deep learning approach, the image files are given as input to AlexNet and GoogLeNet. The NCA is applied to the features obtained from the last dropout layer of each architecture. The distinctive features obtained from the NCA process are classified using the SVM algorithm. In order to evaluate the proposed approach, PQD data are obtained from a modified IEEE 13-bus test system including the SPV system. Several analyses and comparisons are carried out to verify the success of the proposed approach. It has been found that the proposed hybrid deep learning approach has the ability to accurately recognize the PQDs even if the SPV plant integrated power system has a negative effect on power quality.

This paper presents a S-transform effective feature extraction method for power quality (PQ) disturbance classification problem. Firstly, S-transform technique is used to extract the significant features of distorted signal. Then, an optimum combination of the most useful features is identified for increasing the accuracy of classification. Features extracted by using the S-transform are applied as input to support vector machines (SVM) for automatic classification of the power quality (PQ) disturbances that solves a relatively complex problem. Seven single disturbances and six complex disturbances selected as reference are considered for the classification. Sensitivity of proposed classifier under different noise conditions is investigated. The analysis and results show that the classifier can effectively classify different PQ disturbances.

Detection and classification of multiple power quality disturbances in Microgrid network using probabilistic based intelligent classifier

Power Quality (PQ) monitoring in a systematic and automated way is the important issue to prevent detrimental effects on power system. The development of new methods for the automatic classification of PQ disturbances is at present a major concern. This paper presents a novel approach of automatic detection and classification of single and hybrid PQ disturbances using Discrete Wavelet Transform (DWT) and Modular Probabilistic Neural Network (MPNN). The automatic classification of the PQ disturbances consists of three stages i) data generation, ii) feature extraction and iii) disturbance classification. The data is generated by synthetic models of single and hybrid PQ disturbance signals based on IEEE 1159 standard. DWT with multiresolution analysis was applied for feature extraction from the PQ waveforms. The entropy and energy features extracted from the detail and approximation coefficients were applied as the training and testing data to MPNN in order to accomplish the automatic classification process. The effectiveness of the proposed algorithm has been validated by using a typical real-time underground distribution network in Malaysia which was simulated in PSCAD/EMTDC power system software to generate PQ disturbances. The simulation results show that the classifier has an excellent performance in terms of accuracy and reliability even in the case of PQ signals under noisy condition.