Núbia Silva Dantas Brito

Fault detection and classification in transmission lines based on wavelet transform and ANN

This paper proposes a novel method for transmission-line fault detection and classification using oscillographic data. The fault detection and its clearing time are determined based on a set of rules obtained from the current waveform analysis in time and wavelet domains. The method is able to single out faults from other power-quality disturbances, such as voltage sags and oscillatory transients, which are common in power systems operation. An artificial neural network classifies the fault from the voltage and current waveforms pattern recognition in the time domain. The method has been used for fault detection and classification from real oscillographic data of a Brazilian utility company with excellent results

Real-Time Detection of Transients Induced by High-Impedance Faults Based on the Boundary Wavelet Transform

The development of modern protection functions is a challenge in the emerging environment of smart grids because the current protection system technology still has several limitations, such as the reliable high-impedance fault (HIF) detection in multigrounded distribution networks, which poses a danger to the public when the protection system fails. This paper presents the wavelet coefficient energy with border distortions of a one-cycle sliding window designed for the real-time detection of transients induced by HIFs. By using the border distortions, the proposed wavelet-based methodology presents a reliable detection of transients generated by HIFs with no time delay and energy peaks scarcely affected by the choice of the mother wavelet. The signatures of different HIFs are presented in both time and wavelet domains. The performance of the proposed wavelet-based method was assessed with compact and long mother wavelets by using data from staged HIFs on an actual energized power system, taking into account different fault surfaces, as well as simulated HIFs. The proposed method presented a more reliable and accurate performance than other evaluated wavelet-based algorithms.

Real-time classification of transmission line faults based on Maximal Overlap Discrete Wavelet Transform

The analysis of fault-induced transients in three-phase overhead transmission lines can provide extensive information about the fault type, detection, location, direction, and sustained time in satisfactory agreement with real application in protective relays. In this paper, the wavelet coefficient energies of the Maximal Overlap Discrete Wavelet Transform regarding the fault-induced transients were used for real-time fault classification in transmission lines. The performance of the real-time fault classification was evaluated by means of the Real Time Digital Simulator (RTDS) and good results were obtained. The performance of the method was also off-line evaluated with actual oscillographic records of a Brazilian Power System Utility and good results were also obtained.

Comparison between backpropagation and RPROP algorithms applied to fault classification in transmission lines

The computed results from implemented artificial intelligence algorithms, used to identify and classify faults in transmission lines, are discussed in this paper. The proposed methodology uses sampled data of voltage and current waveforms obtained from analog channels of digital fault recorders (DFRs) installed in the field to monitor transmission lines. The performances of resilient propagation (RPROP) and backpropagation algorithms, implemented in batch mode, are addressed for single, double and three-phase fault types.

Haar Wavelet-Based Method for Fast Fault Classification in Transmission Lines

The application of Haar wavelet transform for fault classification in transmission lines is proposed. The faulted phases identification is carried out by the analysis of the energy of the detail coefficients of the phase currents. In addition, by the analysis of the smooth coefficients of the neutral current, the ground faults can be distinguished from phase to phase ones. The algorithm was evaluated with EMTP-simulated faults. The obtained results reveal that the proposed method can accurately identify transmission line faults from a wide variety of power system operating conditions, within the first half-cycle after the beginning of the fault

High Impedance Fault Identification on Distribution Networks

This paper presents a transient-based algorithm for high-impedance fault identification on distribution networks. It uses the discrete wavelet transform to monitor high- and low-frequency voltage components at several points of the power system, being able to indicate the most likely area within which the disturbance has occurred, without requiring data synchronization nor the knowledge of feeder or load parameters. The proposed algorithm is evaluated through electromagnetic transients program simulations of high-impedance faults in a 13.8 kV system modeled from actual Brazilian distribution grid data. Solid faults, capacitor bank switching, and feeder energization are also simulated, considering the system with and without distributed generation. Obtained results show that the algorithm significantly reduces the search field of the high-impedance fault, reliably distinguishing it from other disturbances.

Sampling rate of digital fault recorders influence on fault diagnosis

A case study of fault classification in transmission lines using artificial neural networks (ANN) is presented. The database is built from current and voltage waveform samples obtained from fault simulations with the ATP. Utility companies usually have digital fault recorders with different sampling rates, so it is important to evaluate how good the classifier is when the sampling rate changes, this is the main purpose of the paper. A routine to reduce the sampling rate with no loss of accuracy in classifying faults was implemented.

A Method for Fault Classification in Transmission Lines Based on ANN and Wavelet Coefficients Energy

This paper proposes a novel method for transmission lines fault classification using oscillographic data. The scheme is based on the analysis of the current wavelet coefficients energy using an artificial neural network. In order to validate the proposed approach, both simulated and actual oscillographic data were used.

Detecção e classificação de faltas a partir da análise de registros oscilográficos via redes neurais artificiais e transformada wavelet

This paper presents a method for fault detection and classification in transmission lines, based on analysis of oscillographic data using artificial neural networks and wavelet transform. The fault detection and its clearing time are determined based on a set of heuristic rules obtained from the current waveform analysis in time and wavelet domains. The method is able to single out faults from other power quality disturbances such as voltage sags and oscillatory transients, which are common in power systems operation. An ANN classifies the fault by the voltage and current waveforms pattern recognition in time domain. The method was used for fault detection and classification from both simulated and real oscillographic data of Chesf, a Brazilian utility company, with excellent results.

An IEC 61850 network simulator

The IEC 61850 standard inserted new paradigms on communication between power system protection equipments. In this new environment the involved engineers need to learn new concepts. One way to help engineers in learning these concepts is using a simulator. In this context, this paper presents the main characteristics and functionalities of an IEC 61850 network simulator which is being developed by Federal University of Campina Grande (UFCG) and Hydro Electric Company of San Francisco (Chesf), an utility company of Brazil.