Search

A convolution neural network (CNN) based deep learning method has been proposed for automatic classification and localization of nonlinear loads present in an interconnected power system. The identification of nonlinear loads has been previously dealt with the use of Nonlinear Auto Regression neural network with eXogenous inputs (NARX), Backpropagation Neural Network (BPNN), Probabilistic Neural Network (PNN), Artificial Neural Networks (ANN) and Fuzzy Logic (FL). However, these techniques had not explored the area of classification of industrial and domestic nonlinear loads in an interconnected power system. Also, a Deep learning-based solution for identification of the type of nonlinear load has not been reported in the literature to date. Hence, to address these shortcomings, an IEEE-9 Bus system with industrial nonlinear loads has been used to obtain various current waveforms with distortions. The recorded current waveforms are transformed into a time-frequency (TF) domain plane, and the obtained images are then fed to the deep learning algorithm. The colored images of the TF plots of each type of nonlinear load in Red-Green-Blue (RGB) index provide the best visual features for extraction. The TF domain signatures of individual events are scaled to a standard size before feeding to the algorithm. Through these TF signatures, unique features were extracted with the deep learning algorithm, and then passed on to different stages of convolution and max-pooling with fully connected layers. The softmax classifier at the end classifies the input data into the type of nonlinear present in the power system. The algorithm, when run at different buses, also identifies the location of the nonlinear load. The proposed methodology avoids the usage of any additional fusion layer for obtaining unique features, reduces the training time and maintains the highest accuracy of 100%. [ABSTRACT FROM AUTHOR]

This paper proposed a hybrid intelligent approach based on empirical mode decomposition (EMD), artificial neural network (ANN) and J48 algorithm of machine learning for real-time harmonics analysis of digital substation's equipment based on IEC-61850 using explanatory input variables based on laboratory proto-type real-time recorded database. In the proposed hybrid model, these variables are first extracted then diagnostic of power transformer harmonics of digital substation is evaluated/analyzed to perform the long term as well as the short term goal and planning in the electrical power network. In this paper, firstly, experimental analysis is performed to validate the laboratory prototype setup using FFT (fast Fourier transform), STFT (short-time Fourier transform) and CWT (continuous wavelet transform). Then, features are extracted from experimental dataset using EMD (empirical mode decomposition) method. The IMFs (intrinsic mode functions) have generated from EMD, which are used as an input variable to the two different diagnostic models, i.e., ANN and J48 algorithm. In order to validate the performance and accuracy of the proposed hybrid model, a comparative analysis is performed by using ANN and J48 method (with and without EMD method) and the results are compared. Obtained results shows that the proposed hybrid diagnostics approach for harmonics analysis has outperformance characteristics. [ABSTRACT FROM AUTHOR]

In this paper, two accurate hybrid islanding detection schemes are proposed based on Wavelet Transform and Stockwell transform (S-transform). The proposed methods use the potential of sequence voltage (negative) retrieved at the target Distributed Generation (DG) location of the distribution network under study. In one of the schemes, Discrete Wavelet transform (DWT) is applied to process the negative sequence voltage signal and for its decomposition, which is further used to extract six statistical features like energy, entropy, mean, kurtosis, standard deviation, and skewness from the reconstructed DWT coefficients. Test and train data sets are generated with the wide variation of loading conditions, and optimal features are chosen from the full feature set by forward feature selection method (FFS) during the training process by an artificial neural network (ANN). After that, the trained system is tested to get the detection result. Another scheme presented in this paper for islanding detection is based on S-transform, which is used to decompose the negative sequence voltage signal. Amplitude, frequency, and phase are the three coefficients acquired from the pre-processing of the raw signal by S-transform. Then the cumulative sums of the energy content of the S-transform coefficients are determined and are compared with a threshold value to get the detection result. The proposed schemes are tested in a distribution network consisting of two 9 MW wind farm driven by six 1.5 MW wind turbine connected to 120 kV main grid through a 25 kV, 30 km feeder. Several cases have been investigated like normal condition, islanding, DG line trip, disconnection of point of common coupling, and sudden change in load to test the performance of the proposed schemes. It can be observed from the results that both the approaches gave high accuracy in the detection of islanding conditions and demarcates properly from the non-islanding state. However, results show that the S-transform based approach provides a better resolution and quick detection of islanding than the wavelet transform approach. [ABSTRACT FROM AUTHOR]

Smart city has come a long way since the development of emerging technology like Information and communications technology (ICT), Internet of Things (IoT), Machine Learning (ML), Block chain and Artificial Intelligence. The Intelligent Transportation System (ITS) is an important application in a rapidly growing smart city. Prediction of the automotive accident severity plays a very crucial role in the smart transportation system. The main motive behind this research is to determine the specific features which could affect vehicle accident severity. In this paper, some of the classification models, specifically Logistic Regression, Artificial Neural network, Decision Tree, K-Nearest Neighbors, and Random Forest have been implemented for predicting the accident severity. All the models have been verified, and the experimental results prove that these classification models have attained considerable accuracy. The paper also explained a secure communication architecture model for secure information exchange among all the components associated with the ITS. Finally paper implemented web base Message alert system which will be used for alert the users through smart IoT devices. [ABSTRACT FROM AUTHOR]

In this paper, nature inspired search algorithms, namely particle swarm optimization (PSO) and genetic algorithm (GA) are used to design fractional order proportional and integral (FOPI) and artificial neural network (ANN) controller based distribution static compensator (DSTATCOM) and electronic load controller (ELC) for power quality improvement. Improvement in power quality is achieved using DSTATCOM and an ELC. DSTATCOM is designed using FOPI and ANN based controllers, as opposed to conventional PI controllers which are comparatively less efficient. PSO and GA techniques are employed to determine the optimal parameters for the controllers. The improvement in the performance of the ANN and FOPI as compared to PI controller for the DSTATCOM and ELC is validated using MATLAB based modeling and simulations. Linear consumer loads were used to perform a comparative study in terms of maximum percentage error. Further, we analyzed the system for a nonlinear load and demonstrated decrease in the harmonic distortion. [ABSTRACT FROM AUTHOR]

Quality management is for providing better resources to the customers. It is applicable for hospitals for giving best services. QM has an initiative i.e. FMEA, which is applicable in the purchasing process of the hospital (PPH) for calculating the RPN, which determines the risks linked with the problems occurring in purchasing the particular equipment for the hospital. RPN is conquered from past experience and engineering decisions which leads to errors and discrepancies. In this study, Neuro-Fuzzy approach based technique is applied to improve the purchasing process in Indian private hospitals. Neuro-Fuzzy approach eliminates insufficiencies in the assessment of the RPN, resulting in saving of time. PPH in India has never been improved before by applying neuro-fuzzy scheme based FMEA technique. Analyzed results show that the applied neuro-fuzzy method is able to solve the problems which arise from conventional FMEA approach and will effectively find out RPN. Proposed method provides quality assurance in the process. [ABSTRACT FROM AUTHOR]

Rolling bearing is an important mechanical element therefore its condition monitoring is necessary to ensure the steadiness of industrial machineries. In this paper, the open source vibration data have been processed using advanced signal processing techniques such as EMD method to extract more symmetric waves (IMFs) out of non-linear and non-stationary vibration signals. In addition to this, statistical time-domain and frequency-domain features are calculated and then J48 Decision Tree Algorithm is used for feature selection. The processed input signals have been used for comparative study of five different types of Artificial Neural Network (ANN) classifiers. The performance characteristics of MLP, PNN, GRNN, RBF and LVQ are shown in results and discussion section. [ABSTRACT FROM AUTHOR]

Weight initialization is the most important component which affects the performance of artificial neural network during training the network using Back-propagation algorithm. The initial starting weights have significant effect on the training. If the weights are too large then the sigmoid will saturate, that makes learning slow. If weights are too small then gradients are also too small. In this paper a new weight initialization method has been proposed. The results for the proposed weight initialization technique are compared against the random weight initialization method. In this paper the proposed weight initialization method is statistically analyzed. Ten different data sets out of which five sets of data are taken from UCI machine learning repository and five sets of data are generated using function approximation problems that are used. Resilient Back Propagation training algorithm is used for training the feed forward artificial neural network. The proposed weight initialization method gives better results when compared with random weight initialization technique. [ABSTRACT FROM AUTHOR]

This paper deals with MATLAB/SIMULINK simulation and analysis of a position sensor-less field oriented control of permanent magnet synchronous motor. Adaptive position estimators are required as the parameters of the machines like rotor resistance, inductance changes sometimes. Adaptive position and speed estimators viz. SMO, MRAS are much discussed in literature but the artificial neural network, adaptive neuro-fuzzy inference based estimators are least discussed. In this paper a MATLAB study of MRAS, ANN and ANFIS based position estimator in a Field oriented control of a permanent magnet synchronous motor drive is being done. MRAS, ANN, ANFIS estimators adaptive in nature so these estimators can adapt if there is any parameters change online. The performances of these three drives are analyzed, and results are compared. It is seen that ANFIS based system performance is better even when the parameters of the machines vary with time. This work is limited to analysis and simulation only and could be extended to a practical realization in future work. [ABSTRACT FROM AUTHOR]

User activity classification is one of the most popular research topic in the domain of health care and social care, since this automated technology can provide monitoring and understanding of activities of patients. Smartphone inbuilt sensors based User Activity Classifier (UAC) recognizes user activities using features extracted from sensors like accelerometer and gyroscope in build in smartphones. In this research paper, we are proposing a new user activity classifier system using Layer Recurrent Neural Network (LRNN) which is Artificial Neural Network (ANN). We utilize synthesized data, containing features of user activity classification system, extracted from the raw data recorded in smartphones. With these derived features, we train and test Layer Recurrent Neural Network classifier for user activity classifier. In order to evaluate this system, we have compared the performance of this Layer Recurrent Neural Network based user activity classifier against the convention Multilayer Perceptron (MLP) and Naive Bayes based user activity classifier. Test results show that the proposed Layer Recurrent Neural Network -based user activity classifier is able to recognize user activities reliably and outperforms the Multilayer Perceptron based user activity classifier. We have achieved the classification accuracy of 98.56% for the activities. The results are much more accurate than Multilayer Perceptron based classifier and Naive Bayes classifier. [ABSTRACT FROM AUTHOR]

Development of power through wind with the enhancement of renewable energy resources, frolics/romps a principal role in a developing country like India due to its censorious locations. Wind speed prediction in long term scenario has become a key research area in distinct applications (i.e., management of energy, optimal designing of wind farm, restructuring of electricity marketing, load-shedding and load forecasting). However, forecasting of accurate wind speed data for installation of wind turbine is very difficult due to its deterministic and probabilistic characteristics. The presented technique in this study may bridge the research gap related with the long term wind speed forecasting as resolve the previously indicated problems. Thence, two basically distinct techniques, k-nearest neighbors (kNN) algorithm and artificial neural network (ANN), have been implemented to forecasting of monthly wind speed of Indian cities. The uniqueness of the presented paper is to predict the wind speed in common form of incoming month by implementing the kNN algorithm. A dataset of current wind speed recorded specimen from 168 cities of India is utilized to train and test the proposed approach. Obtained results through the proposed approach have been validated by using ANN technique, which shows very small MSE. [ABSTRACT FROM AUTHOR]

Radial basis function network (RBFN) is used in this paper for predefined trajectory control of both one-link and two-link robotic manipulators. The updating equations for the RBFN parameters were derived using the gradient descent principle. The other advantage of using this principle is that it shows the clustering effect in distributing the radial centres. To increase the complexity, the dynamics of robotic manipulator is assumed to be unknown, and hence, simultaneous control and identification steps were performed using the RBFNs. The performance of the RBFN is compared with the multilayer feed-forward neural network (MLFFNN) in terms of mean square error, tolerance to disturbance and parameter variations in the system. The efficacy of RBFN as a controller and identification tool is verified by performing the simulation study, and the results obtained reveal the superior performance of RBFN over MLFFNN in both identification and control aspects for one-link and two-link robotic manipulators. [ABSTRACT FROM AUTHOR]

This paper performs the comparative study of two feed-forward neural networks: radial basis function network (RBFN), multilayer feed-forward neural network (MLFFNN) and a recurrent neural network: nonlinear auto-regressive with exogenous inputs (NARX) neural network for their ability to provide an adaptive control of nonlinear systems. Dynamic back-propagation algorithm is used to derive parameter update equations. To ensure stability and faster convergence, an adaptive learning rate is developed in the sense of discrete Lyapunov stability method. Both parameter variation and disturbance signal cases are considered for checking and comparing the robustness of controller. Three simulation examples are considered for carrying out this study. The results so obtained reveal that RBFN-based controller is performing better than that of NARX- and MLFFNN-based controllers. [ABSTRACT FROM AUTHOR]

This paper proposes a diagonal recurrent neural network (DRNN) based identification model for approximating the unknown dynamics of the nonlinear plants. The proposed model offers deeper memory and a simpler structure. Thereafter, we have developed a dynamic back-propagation learning algorithm for tuning the parameters of DRNN. Further, to guarantee the faster convergence and stability of the overall system, dynamic (adaptive) learning rates are developed in the sense of Lyapunov stability method. The proposed scheme is also compared with multi-layer feed forward neural network (MLFFNN) and radial basis function network (RBFN) based identification models. Numerical experiments reveal that DRNN has performed much better in approximating the dynamics of the plant and have also shown more robustness toward system uncertainties. [ABSTRACT FROM AUTHOR]

In this paper adaptive control of nonlinear dynamical systems using diagonal recurrent neural network (DRNN) is proposed. The structure of DRNN is a modification of fully connected recurrent neural network (FCRNN). Presence of self-recurrent neurons in the hidden layer of DRNN gives it an ability to capture the dynamic behaviour of the nonlinear plant under consideration (to be controlled). To ensure stability, update rules are developed using lyapunov stability criterion. These rules are then used for adjusting the various parameters of DRNN. The responses of plants obtained with DRNN are compared with those obtained when multi-layer feed forward neural network (MLFFNN) is used as a controller. Also, in example 4, FCRNN is also investigated and compared with DRNN and MLFFNN. Robustness of the proposed control scheme is also tested against parameter variations and disturbance signals. Four simulation examples including one-link robotic manipulator and inverted pendulum are considered on which the proposed controller is applied. The results so obtained show the superiority of DRNN over MLFFNN as a controller. [ABSTRACT FROM AUTHOR]

This research proposes a pattern/shape-similarity-based clustering approach for time series prediction. This article uses single hidden Markov model (HMM) for clustering and combines it with soft computing techniques (fuzzy inference system/artificial neural network) for the prediction of time series. Instead of using distance function as an index of similarity, here shape/pattern of the sequence is used as the similarity index for clustering, which overcomes few of the shortcomings associated with distance-based clustering approaches. Underlying hidden properties of time series are captured with the help of HMM. The prediction method used here exploits the pattern identification prowess of the HMM for cluster selection and the generalization and nonlinear modeling capabilities of soft computing methods to predict the output of the system. To see the validity of the proposed method in the real-life scenario, it is tested on four different time series. The first is a benchmark Mackey-Glass time series, which is tested for delay parameters τ = 17 and τ = 30. The remaining time series are monthly sunspot data time series, Laser data time series and the last is Lorenz attractor time series. Simulation results show that the proposed method provide a better prediction performance in comparison with the existing methods. [ABSTRACT FROM AUTHOR]

In this paper, a hybrid of Neural Network (NN) and Fast Traversal Filter (FTF) based controller in each area is used to determine the optimal parameters of Load Frequency Control (LFC) of a realistic two area power system. The two area power system is modeled considering the various non -linearities like governor dead band, generation rate constraint (GRC) and boiler dynamics. Input to the controller i.e. the error signal is divided into two parts- linear and non- linear. The linear part of the input signal is minimized by the FTF algorithm, whereas the non- linear part is minimized by the NN algorithm. The output of the controller is the sum of the outputs of NN and FTF networks. The proposed hybrid controller requires less number of samples for training of weights, thus making the system fast. This is highly desirable in power quality problems. The various components of power system are reduced to transfer functions and the system performance is analyzed for 1% step load perturbation in area1 with different controllersproportional and integral (PI), neural network (NN) and NN+FTF based controllers. The simulations demonstrate the fast and smooth performance of the power system with the proposed controller. Simulated results evince the superiority of the proposed hybrid controller. [ABSTRACT FROM AUTHOR]

Abstract: Design and learning of networks best suited for a particular application is a never-ending process. But this process is restricted due to problems like stability, plasticity, computation and memory consumption. In this paper, we try to overcome these problems by proposing two interval networks (INs), based on a simple feed-forward neural network (NN) and Choquet integral (CI). They have simple structures that reduce the problems of computation and memory consumption. The use of Lyapunov stability (LS) in combination with fuzzy difference (FD) based learning algorithm evolve the converging and diverging process which in turn assures the stability. FD gives a range of variation of parameters having the lower and the upper bounds within which the system is stable thus defining the plasticity. Effectiveness and applicability of the NN and CI based network models are investigated on several benchmark problems dealing with both identification and control. [ABSTRACT FROM AUTHOR]

Abstract: By utilizing some of the important properties of wavelets like denoising, compression, multiresolution along with the concepts of fuzzy logic and neural network, new two fuzzy wavelet neural networks (FWNNs) are proposed for approximating any arbitrary non-linear function, hence identifying a non-linear system. The output of discrete wavelet transform (DWT) block, which receives the given inputs, is fuzzified in the proposed two methods: one using compression property and other using multiresolution property. We present a new type of fuzzy neuron model, each non-linear synapse of which is characterized by a set of fuzzy implication rules with singleton weights in their consequents. It is shown that noise and disturbance in the reference signal are reduced with wavelets and also the variation of somatic gain, the parameter that controls the slope of the activation function in the neural network, leads to more accurate output. Identification results are found to be accurate and speed of their convergence is fast. Next, we simulate a control system for maintaining the output at a desired level by using the identified models. Self-learning FNN controller has been designed in this simulation. Simulation results show that the controller is adaptive and robust. [Copyright &y& Elsevier]