Search

A method based on extreme learning machine (ELM) is suggested to locate faults in multi-terminal high-voltage direct current systems. S-transform and wavelet transform are used for extraction of the features used for the learning. The accuracy of the technique for various types of input signals and different lengths of the analyzed window is investigated. Two different approaches are considered for employing the ELM in this application. In the first approach, an ELM is used for total length of the line. In the second one, a multi-ELM technique is applied to different sections of the transmission line. In this approach, one ELM is considered for each of the divided sections. It is proved that the performance of the method is improved by the multi-ELM approach in comparison with the single ELM one. The performance of the ELM approach is compared with the artificial neural network and support vector regression techniques. [ABSTRACT FROM AUTHOR]

Multi-layer perceptron neural networks (MLP-NN) are widely utilized in forecasting applications. However, optimal training of these networks is still a challenge. A comprehensive assessment of MLP training approaches comprising of three stages is performed in the present paper. First, the prediction performance is evaluated using twelve training algorithms. Next, optimization algorithms are utilized to enhance the best obtained network parameters obtained from the first step and the performance of eight optimization algorithms is evaluated. Finally, a novel modification is used to improve the performance of the optimization algorithms. The proposed methodologies are applied to two case-studies and statistical metrics are employed for their efficiency evaluation. Wavelet transformation is used to extract the features which will be fed to the MLP-NN as input data. [ABSTRACT FROM AUTHOR]

• Two improved types of accurate models based on the Schavemaker model are developed. • Unlike the original Schavemaker model which uses invariant parameters, the time-varying nature of the arc's parameters is considered in the proposed models. • The first improved model has time-varying coefficients. • In the second improved model, a new term has been added to the schavemaker model that includes time-varying coefficients. • A large number of the actual data are used. Electric arc furnace (EAF) is more and more applied in large industrial systems and represent one of the major sources of perturbations for the power system because it is highly non-linear and time-varying. New models are suggested in this research based on instantaneous EAF voltage and current measurements taken from EAFs at the Mobarakeh steel mill in Isfahan, Iran. This paper aims to develop two improved types of accurate models based on the Schavemaker model. The original Schavemaker model has three constant parameters. Unlike the original Schavemaker model which uses invariant parameters, the time-varying nature of the arc's parameters is considered in the proposed models. The first improved model has time-varying parameters. Additionally, to further enhance the model's accuracy, in the second improved model, a term has been added to the schavemaker model that includes time-varying parameters. Compared to the original Schavemaker model, all of the developed Schavemaker models are more precise. Finally, auto-regressive moving average (ARMA) models are used to attain the time variation characteristic of the parameters. [ABSTRACT FROM AUTHOR]

Load flow analysis is an essential part of system operation and planning. Up to now, load flow can be categorized as deterministic and probabilistic. In the deterministic load flow, the generations and loads are fixed. However, the state variables of power systems have uncertain nature due to the random load variations and stochastic distributed generations. In this regard, the probabilistic load flow (PLF) is implemented. The inputs of PLF are probability density functions (PDFs) of buses powers and the outputs are PDFs of system states. Therefore, the relation between the system variables over the time will be lost. To overcome this issue, analytic time series load flow (analytic TLF) is introduced in this work. TLF considers the effect of time synchronization and correlation between different loads in the network. Auto regressive moving average (ARMA) models are used to model the time varying loads and generations. The inputs in analytic TLF are ARMA model parameters of active and reactive powers of load buses and the active power of generation buses. The outputs are ARMA parameters of the voltage magnitude and phase of load buses as well as voltage phase of generation buses. The performance of the proposed analytic TLF is evaluated using several examples. [ABSTRACT FROM AUTHOR]

Due to increase in penetration of Distributed Generations (DGs) in power systems, fault current level is being increased, which results in some problems in the systems. Fault Current Limiters (FCLs) are attractive devices to tackle these problems for transmission and distribution systems. The utilized FCLs may have considerable impact on the signals used for differential protection of power transformers, which leads to mal-operation of these protections. It seems a comprehensive analysis is necessary for performance evaluation of differential protection algorithms in presence of FCLs. This paper deals with investigation of FCLs impact on power transformers’ differential protection. The performance of some well-known differential protection algorithms for discrimination between internal fault current and magnetizing inrush current with and without presence of FCL are evaluated. [ABSTRACT FROM AUTHOR]

Stand-alone micro-grids (MGs) and grid-connected MGs with high penetration level of Distributed Generation (DG) are growing at a fast rate. In these grids, the power quality disturbances such as harmonics, inter-harmonics and deviation from the fundamental frequency are widespread. The influence of these disturbances on two categories of digital metering algorithms is evaluated in this paper. The first category is related to the measurement of the fundamental frequency component of current and voltage signals, which are used as inputs of the digital protection algorithms. Most of the relays only use fundamental frequency component of their input voltage and current signals to fulfill the desired protective functions. The accuracy of the protective relays highly depends on the performance of the applied algorithms for extracting the fundamental components. The second evaluated category is related to the reactive power calculation algorithms, which are used in the static VAr compensator (SVC) control system. Utilizing SVC is growing due to the DGs such as wind turbines magnify voltage fluctuations called flicker. Since the SVC compensates only the reactive power related to the fundamental frequency component, the reactive power signal should not be sensitive to the harmonics. In addition, the reactive power calculation should be fast enough that SVC can follow the abrupt changes to mitigate the flicker, efficiently. In this paper, frequency deviation, harmonic, and inter-harmonic phenomena are considered as the most causes for power quality problems. Their influence in digital metering algorithms is evaluated. It is confirmed that defining new indices is mandatory to fully reflect algorithms׳ accuracy and efficiency. Furthermore, the field measurements of instantaneous voltage and current of a wind farm and electric arc furnaces are used for the evaluation. [ABSTRACT FROM AUTHOR]

In this study main reactive power calculation methods are evaluated and compared. Furthermore, a novel simple method that can be used for calculation of reactive power in electric arc furnace (EAF) plants is proposed. The focus here is on their application in the control system of static VAr compensators (SVCs). SVC can compensate only the fundamental harmonic component. So the reactive power signal used in the control system should be sensitive only to the main harmonic component. In addition, the reactive power calculation should be fast enough that SVC can follow the quick changes of the load to mitigate the flicker efficiently. To compare various reactive power calculation methods and to evaluate the proposed method, the simulated known loads are used beside huge amount of practical data recorded in Mobarakeh Steel Company in Isfahan/Iran. Results show the proposed method has the best performance for SVC control system. [ABSTRACT FROM AUTHOR]

Islanding is one of the most important concerns of the grid connected distributed resources due to personnel and equipment safety. Many approaches have been proposed for islanding detection, which can be categorized into passive and active schemes. The main concern of the passive schemes is related to their large Non Detection Zone (NDZ), while the main problem of the active methods is related to their negative impact on power quality. This paper propose an efficient and intelligent islanding detection algorithm using combination of an optimal Artificial Neural Network (ANN) based on Particle Swarm Optimization (PSO) with a simple active method. The intelligent islanding detection method based on ANN, may have mal-detection in the case of change in the power network structure. In the proposed scheme, ANN is adapted with change in power network structure to reduce NDZ. Optimal parameters of the ANN such as weight coefficients and biases are derived using the PSO in order to minimize the technique NDZ. Also the performance of the various structures of ANN such as Multilayer Perceptron (MLP), Radial Basis Function (RBF) and Probabilistic Neural Network (PNN) in combination with PSO is compared for islanding detection purpose. The proposed method is simulated and tested in various operation conditions such as islanding conditions, motor starting, capacitor bank switching and nonlinear load switching. The test results showed that it correctly detects the islanding operation and does not mal-operate in the other situations and has a small NDZ. [ABSTRACT FROM AUTHOR]

This paper presents simultaneous allocation of distributed generations (DGs) and fast directional protection switches (FDPSs) to decrease energy losses and enhance reliability of the network. The main contribution of the paper is restricting the allowable size of DGs to ensure the correct operation of upstream installed FDPS in all conditions. The paper introduces a method based on genetic algorithm to solve the optimization problem. Finally, to confirm the effectiveness of the proposed method some simulations considering a 33 bus test network are performed. The optimization problem with and without applying protection constraint is solved. Customer’s load is modeled by a three level yearly load. Time value of money and load growth rate are also considered. To assess the importance of the proposed protection constraint, fault studies after adding DGs and FDPSs to a 33 bus test network are performed. Results confirm the importance of the proposed protection constraint. [ABSTRACT FROM AUTHOR]

Linear models are capable of capturing the Linear Deterministic (LD) component of the time series. In order to benefit from both Nonlinear Deterministic (ND) and LD components during the prediction procedure, it is necessary to employ nonlinear models. The complexity of the prediction algorithm increases when nonlinear models are utilized. Hence, before applying nonlinear models the presence of nonlinear component should be confirmed. Although surrogate data technique uses various tests to indicate the nonlinearity, in many cases its test results are different and in conflict with each other. The reason is time series include LD and ND components together and giving a strict answer about nonlinearity cannot be applicable. Here instead of such a strict answer, by quantizing the ND component, a new index (a number between 0 and 1) is proposed (the closer to 1 the more ND components). In this method first we use ARMA models. The residual series is used to calculate the proposed index which it contains all components of the original series except LD. The proposed procedure is applied to three different case studies. Furthermore, the performance of some nonlinear prediction methods (Markov, Grey, Grey–Markov, EMD–Grey, NARnet and ARMAX) is compared with the proposed index. [ABSTRACT FROM AUTHOR]

Highlights: [•] The use of Modified Firefly Algorithm to estimate the SVR parameters. [•] Introduction of a new modification method for the firefly algorithm. [•] Comparative forecasting with ARMA, ANN, SVR-GA, SVR-HBMO, SVR-PSO and SVR-FA. [ABSTRACT FROM AUTHOR]

• Proposing discrimination index based on fundamental phase angle of the current signal. • The proposed index's threshold does not require transformer parameters. • Classifying internal/external faults and inrush currents in sub-cycle considering CT saturation. • Enhancing reliability and response time in the various fault/inrush conditions. Power transformer differential protection may confront mal-functioning in authentic discrimination between inrush and internal faults. To tackle the latter mal-functioning, a new two-stages algorithm based on phase content of the current signal of the current transformers (CTs) is put forward. The proposed algorithm is designed based on the fact that the fundamental phase angle of a fault signal ideally remains constant during the fault. However, during inrush cases, the phase angle varies. Also, during the internal fault, the phase angles of the current signals of CTs are in phase while during the external fault, the phase angles of the current signals of CTs are 180° out of phase. In the first stage, the proposed algorithm calculates the fundamental phase angles of the current signals of the CTs using sub-cycle modified recursive least squares (MRLS). Afterward, normalized mean residue (NMR) is employed to measure distance between the estimated phase angles of the CT's currents. MRLS and NMR algorithms require limited samples (i.e. 10 and 5 samples respectively) for executing their calculations. Performance evaluation with simulated and experimental recorded current signals shows the ability of the proposed method in discrimination of the internal faults from inrush and external fault currents. [ABSTRACT FROM AUTHOR]

In order to increase production of an EAF (electric arc furnace) by reduction of melting time, one can increase transferred power to the EAF. In other words a certain value of energy can be transferred to the EAF in less time. The transferred power to the EAF reduces when series reactors are utilized in order to have stable arc with desired characteristics. To compensate the reduced transferred power, the secondary voltage of the EAF transformer should be increased by tap changing of the transformer. On the other hand, after any tap changing of the EAF transformer, improved arc stability is degraded. Therefore, the series reactor and EAF transformer tap changing should be simultaneously determined to achieve arc with desired characteristics. In this research, three approaches are proposed to calculate the EAF system parameters, by which the optimal set-points of the different series reactor and EAF transformer taps are determined. The electric characteristics relevant to the EAF for the all transformer and series reactor taps with and without SVC (static VAr compensator) are plotted and based on these graphs the optimal set-points are tabulated. Finally, an economic evaluation is also presented for the methods. [ABSTRACT FROM AUTHOR]

• Recursive and straightforward formulation based on KF is presented. • An index based on the standard deviation of the residual signal is introduced. • Also an auxiliary index based on the voltage signal is utilized. • The performance of the proposed indices is investigated with experimental data. Today, due to the widespread utilization of Induction Motors (IMs) in different industries, their condition monitoring is of great importance. Concentrating on the failures of IMs, it has been acknowledged that the stator inter-turn faults (SITFs) are the most frequent electrical failures. This paper puts forward an algorithm for SITF detection founded on Kalman Filter (KF). More specifically, the proposed algorithm employs KF to extract motor current signatures (MCS) and motor voltage signatures (MVS). Afterward, a statistical SITF index is used, based on the standard deviation of the extracted signatures. The proposed SITF index is technically robust against non-fault conditions including voltage quality problems and heavy load changes as well as has a significant performance in the presence of high measured noise-impregnated signals due to utilization of KF algorithm-based. Moreover, during source harmonic pollutions, the proposed algorithm has a very robust performance. Also, due to straightforward and low-computational mathematical basis, the proposed method is computationally efficient. The performance of the proposed method is validated with numerous simulation and experimental scenarios. The results indicate the proposed SITF index has robust performance, promising accuracy and good speed of convergence. [ABSTRACT FROM AUTHOR]

• Proposing discrimination index based on the rate of phase angle change of current signal. • The proposed discrimination index does not depend on transformer parameters. • Classifying fault and inrush currents in sub-cycle considering CT saturation. • Enhancing reliability and response time in the various fault/inrush conditions. Magnetizing inrush currents are known as the most critical potential threat in maloperation of the transformer differential protection. Therefore, there is a need to detect the inrush current from the internal fault current to avoid the differential protection maloperation. Due to the possibility of a current transformer (CT) saturation at a high level of inrush and fault current, the discrimination algorithm should function appropriately in CT saturation conditions. Here, a method is developed based on the rate of phase angle change (RoCoPA) to perform the discrimination process. The proposed discrimination index employs the RoCoPA of the three-phase differential currents and calculates the standard deviations of the RoCoPA of the three-phase differential currents. The discrimination procedure is designed based on the fact that during fault scenarios, the signal remains almost sinusoidal, and as a result, the RoCoPA has the minimum variations. On the contrary, owing to the non-sinusoidal wave shape of the inrush current, the RoCoPA has notable variations. Based on the various simulated and practical scenarios, the proposed index's performance is evaluated considering challenging scenarios. The results reveal that the proposed discrimination index has promising accuracy with average response delay with about half a cycle. [ABSTRACT FROM AUTHOR]

• The proposed method utilizes the direct features of the electric arc. • Our method detects incipient fault considering the minority and momentary features. • Fault data are obtained from both simulation and experimental recorded data. • Our method guarantees simplicity in implementation, speed, and accuracy. • It distinguishes transients such as inductive load change and capacitor switching. • It discriminates the incipient fault from the harmonic load. Incipient faults are known as the faults mainly occurring on the junctions of underground power cables. Clearing quickly, these faults are difficult to detect by conventional protection equipment, leading to insulation degradation and permanent fault in the power cable over time. In this paper, employing similarity functions, a new method is put forward to distinguish the existing fault from other similar events with both high speed and accuracy. The proposed algorithm consists of a two-stages algorithm including change detection, disturbance identification, and fault detection. The proposed method performs the calculation on the voltage signal that is recorded at the sending end of the line. The cross-correlation function is employed for identifying change detection and determining disturbance time. Afterward, using sinusoidal curve fitting, the algorithm can distinguish the incipient fault from sudden load change, capacitor switching, and harmonic load. The accuracy of the proposed approach is verified by some experimental data and 6272 simulation data obtained from four types of well-known arc models, capacitor switching, and also load variations. [ABSTRACT FROM AUTHOR]

One of the prominent problems in wind farms is voltage flicker emission. To prevent flicker emission or mitigate the impact as best as possible, a static VAr compensator (SVC) is a great candidate both economically and technically. However, SVCs cannot completely compensate the fast-changing reactive power due to delays caused by the reactive power calculation unit and the triggering fire angle of the SVC. This paper proposes a predictive control system for SVCs, by merging an additional predictive control block into the conventional control system. It is constructed based on deep neural networks, namely adaptive one-dimensional convolutional neural network (1D-CNN). The training process is conducted based on the adaptive learning weights process to enhance the prediction accuracy and training computational complexity of the 1D-CNN. Numerical results on the actual dataset in a wind farm in Manjil, Iran, have verified the forecasting accuracy and flicker mitigation of the proposed controller. [ABSTRACT FROM AUTHOR]

Abstract: Prediction of electric arc furnace (EAF) reactive power is an appropriate solution to compensate for static VAr compensator delay and improve its performance in flicker reduction. A linear autoregressive moving average (ARMA) can only pull out the linear deterministic (LD) component of EAF reactive power time series. For the prediction to be made through both nonlinear deterministic (ND) and LD components, employing nonlinear models is necessary. However, before developing the nonlinear models for prediction, the necessity of the employing them should be verified by investigating the significance of the ND components in the process. This paper presents a novel approach for wide analysis of nonlinear behavior of EAFs reactive power time series related to eight ac EAFs installed in Mobarakeh steel industry, Isfahan, Iran to answer the question about the importance of their ND components. In the approach, a suitable linear auto regressive moving average (ARMA) model with order (4,4) is used for the time series to extract the residual time series. Then, a number of well established nonlinear analysis techniques such as time delay reconstruction, surrogate data, delay vector variance and recurrence plot methods are applied to the original and residual time series. To describe the nonlinear characteristics of the time series, some new indices are defined. They quantify the significance of the ND component in compare with LD component. [Copyright &y& Elsevier]

Abstract: The time varying nature of electric arc furnace (EAF) gives rise to voltage fluctuations which produce the effect known as flicker. The ability of static VAr compensator (SVC) in flicker reduction is limited by delays in thyristor ignition. To improve SVC performance in flicker compensation, EAF reactive power can be predicted for a half-cycle ahead, by using appropriate autoregressive moving average (ARMA) models. This paper uses huge field data collected from ac arc furnaces, and demonstrates that the EAF reactive power models coefficients are different from one data record to another and do not follow any specific law. Therefore, it is necessary to update the model coefficients for prediction purposes. For this purpose, two major adaptation algorithms, the least mean square (LMS) and recursive least square (RLS) are used to determine online the prediction relationship coefficients. By applying the methods to the data records and using some indices such as newly defined indices based on concepts of flicker frequencies and power spectral density, the transient and steady state performances of the methods are studied in EAF reactive power prediction. A simulation example on the application of the predictive models in a SVC control system is presented. [Copyright &y& Elsevier]

• A hybrid time-frequency model for EAF. • A four-winding model for the EAF's transformer. • Modeling of the tap changer mechanism. • An efficient differential protection algorithm. This paper deals with the modeling of EAF feeding system taking into account the tap transition for assessment of a suggested scheme of differential protection. There are several challenges in the implementation of the differential protection in the EAF transformer which reveal the importance of the system modeling. Large current in the secondary side of the transformer, similarity of normal and fault conditions, and frequent tap variation in the transformer during the melting procedure are the most serious concerns. Considering the mechanism of the tap changers, the new tap setting is readjusted in the algorithm to prevent mal-operation of the protection. These challenges can be rectified by suitable modeling of EAF feeding system and an efficient differential protection scheme. To this end, the EAF as an electrical load, the EAFs' transformer, and the mechanism of the tap changing are properly modeled in this work. The proposed model of the EAF is based on real recorded data from EAFs used in a steel company, to fully simulate the nonlinear and time-variant characteristics of this electrical load. The proposed models of the EAFs' transformer and tap changer enable us to simulate different tap variations and internal faults in the system. The performance of the suggested differential protection algorithm is evaluated using plenty of simulations in different conditions of the system. [ABSTRACT FROM AUTHOR]

Identification of flicker sources is necessary to find who is responsible for the measured flicker and improve power quality. This paper puts forward a new method for identifying flicker sources with minimum measurement units. Contrary to the previous works where flicker sources are considered a single-frequency signal, the autoregressive moving average (ARMA) is used to model active and reactive power variations. First, the envelope of the network voltage at the considered busbars is derived by the Hilbert transform. Then, appropriate flicker indices are extracted from the power spectral density (PSD) of the voltage envelope. A novel two-level structure of a set of ANNs is proposed, which needs a low number of voltage measurement units to locate the flicker sources. Using the captured data from different simulations of various scenarios, the Artificial Neural Networks (ANNs) are trained to categorize flicker sources. [ABSTRACT FROM AUTHOR]

Environmental concerns besides fuel costs are the predominant reasons for unprecedented escalating integration of wind turbine on power systems. Operation and planning of power systems are affected by this type of energy due to the intermittent nature of wind speed inputs with high uncertainty in the optimization output variables. Consequently, in order to model this high inherent uncertainty, a PRPO (probabilistic reactive power optimization) framework should be devised. Although MC (Monte-Carlo) techniques can solve the PRPO with high precision, PEMs (point estimate methods) can preserve the accuracy to attain reasonable results when diminishing the computational effort. Also, this paper introduces a methodology for optimally dispatching the reactive power in the transmission system, while minimizing the active power losses. The optimization problem is formulated as a LFP (linear fuzzy programing). The core of the problem lay on generation of 2m + 1 point estimates for solving PRPO, where n is the number of input stochastic variables. The proposed methodology is investigated using the IEEE-14 bus test system equipped with HVDC (high voltage direct current), UPFC (unified power flow controller) and DFIG (doubly fed induction generator) devices. The accuracy of the method is demonstrated in the case study. [ABSTRACT FROM AUTHOR]

Highlights: [•] Introduction of a critical issue for travelling wave based assessments in large distribution systems. [•] Application of Core Vector Regression (CVR) for the first time in fault location. [•] Application of Empirical Mode Decomposition (EMD) to extract the fault feature based on a peculiar method called PCF. [•] Comparison between the proposed CVR based and ANN based methods. [ABSTRACT FROM AUTHOR]

• We propose twofold discrimination indices based on kurtosis of the current signal. • The proposed discrimination indices do not require transformer parameters. • It allows sub-cycle classification of fault and inrush current under CT saturation. • It enhances reliability and response time in various fault/inrush conditions. Differential protection is a widely-used approach for power transformer protection, the performance of which may be, at times, challenged due to the inrush current and current transformer (CT) saturation. This paper presents an enhanced statistical algorithm for discrimination of fault and inrush current signals under CT saturation scenarios. This paper proposes twofold discrimination indices centered on the concept of fourth-order statistic moment, the so-called kurtosis on the current signal. Mathematically proven, the algorithm performance only depends on the current signal phase angle and it does not require transformer parameters. The mathematical closed form formulation in this paper classifies fault condition and inrush currents in sub-cycle timeframes even in the CT saturation conditions. The performance of the proposed algorithm is verified under various fault and inrush current scenarios with and without CT saturation. The numerical results demonstrate that the proposed algorithm can successfully classify the disturbances and detects saturations quickly with promising accuracy. Index Terms: Inrush current; internal fault; current transformer saturation; differential protection [ABSTRACT FROM AUTHOR]