Fault Discrimination Research Articles

Impulse fault diagnosis in power transformers using self-organising map and learning vector quantisation

An artificial intelligence approach is proposed to an impulse fault diagnosis problem in oil-filled power transformers. The experiment focuses on the distinction between the effects caused by faults of a different nature and the different physical location of occurrences in a transformer winding. The proposed method involves an artificial neural network-based pattern recognition technique, to recognise the frequency responses of the winding admittance of a typical high-voltage transformer under healthy and different faulty conditions of winding insulation. It attempts to establish a correlation between the nature and site of the internal insulation fault and its associated frequency response. A self-organising neural network model has been employed as the basic pattern recogniser, to discover the significant patterns and to extract the hidden information from a set of frequency response patterns obtained from an EMTP model of the transformer with artificially simulated faults. A learning vector quantisation-based classification technique has been applied to efficiently classify visually indistinguishable response patterns. The method applied to a winding model of a high-voltage transformer, with tap changer winding, exhibited high diagnostic accuracy by successful detection and discrimination of faults of a different nature and site of occurrence.

Threshold-based mechanisms to discriminate transient from intermittent faults

This paper presents a class of count-and-threshold mechanisms, collectively named /spl alpha/-count, which are able to discriminate between transient faults and intermittent faults in computing systems. For many years, commercial systems have been using transient fault discrimination via threshold-based techniques. We aim to contribute to the utility of count-and-threshold schemes, by exploring their effects on the system. We adopt a mathematically defined structure, which is simple enough to analyze by standard tools. /spl alpha/-count is equipped with internal parameters that can be tuned to suit environmental variables (such as transient fault rate, intermittent fault occurrence patterns). We carried out an extensive behavior analysis for two versions of the count-and-threshold scheme, assuming, first, exponentially distributed fault occurrencies and, then, more realistic fault patterns.

Ground Fault Discrimination based on Wavelet Transform using Artificial Neural Networks

Ground Fault Discrimination based on Wavelet Transform using Artificial Neural Networks

Feature extraction methods for neural network-based transmission line fault discrimination

The suitability of conventional distance relays to operate correctly under variations in such factors as source impedance, prefault load and fault resistance is still a problem. This paper describes an alternative approach to nonunit protection of transmission lines using artificial neural networks (ANNs). Particular emphasis is placed on describing a methodology whereby the extraction of the input features (from the measured voltage and current signals) to the ANNs is near optimal; with this approach, the results presented clearly demonstrate that the protection technique gives satisfactory performance under a wide variation in practically encountered system operating and fault conditions.

Derivation of signatures for faults in gas turbine compressor blading

A procedure for the derivation of signatures for faults in the blades of a gas turbine is presented here. A variety of blade faults, corresponding to changes in the angle or spacing of one or more blades, are examined. A fluid dynamic simulation model is used to derive the unsteady pressure signals sensed by a stationary transducer for the cases of faulty blades. The blade fault signatures are derived by processing these signals using Fourier techniques. The features of blade fault signatures are studied. Finally, the possibilities they offer for the discrimination and identification of different possible blade faults are also examined.

Comparison of the Performances of Parametric Estimation, Time-Frequency, Wavelet and Segmentation with Detection of Abrupt Changes in Non-Destructive Evaluation

Laboratory studies of absorption frequency behavior in rocks often use spectral ratios of digitally recorded ultrasonic signals which have been transmitted through a rock sample and a reference sample of a very low absorption, respectively. It is proposed to treat the digitally recorded signals as autoregressive AR process which, using recursive adaptive filter concepts. The purpose of this paper is to show the advantages of the parametric methods over Fourier analysis method for characterization of materials, and by use of time frequency, time-scale, and abrupt changes detection (Linear Predictive Error algorithm (LPE)) the choice of the signal length in spectral ratio method is solved by segmentation. Also, as a tool of analysis time-frequency and time-scale are very adapted for faults discrimination.

ウェーブレット変換による配電線地絡故障の原因判別

ウェーブレット変換による配電線地絡故障の原因判別

Derivation of Signatures for Faults in Gas Turbine Compressor Blading

A procedure for derivation of signatures for faults in blades of a gas turbine is presented. A variety of blade faults, corresponding to changes in angle or spacing of one or more blades are examined. A fluid dynamic simulation model is used to derive the unsteady pressure signals sensed by a stationary transducer for the cases of faulty blades. The blade fault signatures are derived by processing these signals using Fourier techniques. The features of blade fault signatures are studied. Finally, the possibilities they offer for discrimination and identification of different possible blade faults is also examined.

Neural Network SVS Gain Scheduling for Improving Power Systems Stability

Abstract Static Var Sources (SVS) have been widely used in electrical power systems to keep terminal voltages within bounds. This paper proposes a new approach to design such a controller, which is realised by using a gain scheduled PID controller. A radial basis function neural network (RBFNN) is utilised as the gain controller. With the help of a fault discriminator which has been trained to reflect the different character of power system oscillations, the gain schedule is generated automatically through a learning process. A three-machine test system, that becomes longitudinal after fault clearance, assesses the performance of the proposed controller. Simulation results have shown considerable damping improvements from the proposed scheme after training of the RBFNN during the post-fault period.

A hybrid physical model—neural network approach to machine diagnostics

Neural network based diagnostic algorithms have been an effective approach to machine diagnostics. A feature extraction routine calculates parameters indicating the relative changes in signal phase along different paths to multiple accelerometers as might be due to localized damage along one path. These parameters were calculated for accelerometer data from the Westland seeded fault helicopter gearbox experiment. Perfect fault discrimination was achieved using a conventional feed forward neural network classifier on only a small training set, with no special treatment for the confounding influence of varying torque levels. A hybrid neural network was designed, which implemented a simple springs and masses mechanical model of a more complex structure, where sensor locations served as the mass points. This approach is able to localize structural damage using only a small training set on data generated by a finite elements model, and exhibits excellent interpolation and extrapolation. Extending the hybrid neural network approach, a method is proposed for solving for the external forces applied to surface-mounted sensors, from vibrating machine components. The derived forces will then serve as inputs to the signal phase change feature extraction routine. It is hoped this method will train well on small data sets and reduce ambiguities due to signal multipath effects. [See NOISE-CON Proceedings for full paper.]

Practical high-impedance fault detection on distribution feeders

Distribution protection systems must balance dependability with security considerations to be practical. This is quite difficult for high-impedance faults. Only highly sensitive algorithms can achieve absolute dependability in detecting very low current faults. This high sensitivity results in a propensity for false tripping, creating a less secure, system and resulting in the potential for decreased service continuity and lower reliability. Researchers at Texas A&M University have balanced fault detection with fault discrimination, resulting in a practical combination of detection algorithms in a commercially viable system. This device has many "intelligent" features, including the ability to analyze and correlate numerous fault characteristics in real time, so that a correct determination of the status of the feeder can be made with a high probability of accuracy. This paper describes the use of multiple algorithms to detect various types of faults and the use of an expert decision maker to decipher incoming data, to determine the status and health of a distribution feeder. Requirements for a practical, secure high-impedance fault relay are also discussed. Finally, Texas A&M has licensed this technology to a commercial partner, which manufactures a device that detects high-impedance faults, in addition to performing numerous other monitoring and protection functions.

A new method for discrimination of ground faults causes in distribution lines

AbstractSince most problems in distribution lines are ground faults, the technology to distinguish the causes in the distribution lines will be very useful in preventing faults beforehand or, at least finding any problems at an early stage.The waveforms of ground faults which were collected from data concerning problems in the distribution lines and ground fault experiment are analyzed. The technology which could predict classified groups drawn from the scale of I0 (the waveform of ground fault current) and V0 (the waveform of ground fault voltage), and both of them in the ground faults is studied.The results demonstrated that I0 could be classified into three groups, that is, leaking ground faults are sine curve shaped, gap discharge ground faults are triangular, and intermittent discharge ground faults are needle shaped.Also, simulated equivalent impedance of the causes of ground faults by EMTP is calculated. Furthermore, the influence caused by the condition of the distribution lines is investigated. With these results, a technique is proposed which could predict the cause of problems in ground faults.

A digital technique for stator winding protection of synchronous generators

This paper describes a digital technique for detecting internal faults in stator windings of synchronous generators. The technique uses positive- and negative-sequence models of the synchronous generator, and voltages and currents measured at the generator terminals. It does not need information concerning the machine parameters and is applicable to all types of generators, irrespective of their size. Fault discrimination characteristics and a digital algorithm based on the proposed technique are described. The performance of the proposed technique was evaluated using fault data generated by simulations using an electromagnetic transient program, EMTDC. Fault data were also acquired by subjecting a laboratory generator to internal and external faults. Results of the studies using simulated and experimental data are included in the paper. The results indicate that the proposed technique can detect faults including open circuits and faults that short-circuit windings of one phase.

Analysis of distribution disturbances and arcing faults using the crest factor

Transient analysis of distribution disturbances for fault discrimination and event identification can be an important tool for the secure protection of power systems. A simple parameter which quantifies wave distortion has been used to analyze many different behaviors. Theoretical perspectives of the transients were studied and their impacts on changes in the crest factor were thoroughly examined. After testing the scheme with real transient data, discrimination and identification guidelines based on the crest factor were developed. The identification method may not be complete, but it successfully discriminates arcing faults from other disturbances, providing security enhancement for distribution protection.

配電線地絡故障時の波形と原因判別法

配電線地絡故障時の波形と原因判別法

光CTを用いた送電線故障様相判別に関する基礎検討

光CTを用いた送電線故障様相判別に関する基礎検討

Multiple filter methods for detection of hazardous states in an industrial plant

AbstractModel‐based detection systems are presented and classified according to their complexity. The main emphasis is put on multiple Kalman filters. A two‐stage approach including Bayes formula and inhomogeneous Markovian processes is suggested to achieve a fault discrimination. Methods are presented to use the information of nondisjunctive models in this context. To show the proper working of the detection algorithms simulation studies and on‐line laboratory experiments with runaway reactions are considered first. A report on an implementation at a large‐scale industrial process unit concludes the discussion.

SOLUTIONS INCORPORATING MAGNETIC HYSTERESIS TO THE SUBHARMONIC PROBLEM IN POWER SYSTEMS

ABSTRACT One possibility of taking care of the subharmonics in high speed protection is to filter the fundamental frequency components from the fault transients and use them for fault detection and discrimination. This is correct only if the phases and amplitudes of fundamental components during and post-subharmonic periods remain the same in transient waveforms. An attempt is therefore made to establish this aspect by a detailed study of a typical electrical network presenting subharmonics. The Dual Input Describing Function is adopted to show the subharmonic boundaries and the influence of system parameters on subharmonic oscillations. Magnetic hysteresis is incorporated in the analysis. A fifth order nonlinearity and one-third subharmonic are assumed. The theory could be easily extended to any order of nonlinearity and to any subharmonic frequency and also to a combination of subharmonics. The analysis gives a set of nonlinear algebraic equations. These equations are solved by Newton's technique using ...

Application of protective devices to radial overhead line networks

The paper reviews the need for overhead line switching equipment and its use in developing countries for the control of long radial feeders into remote districts. The requirements for pole-mounted equipment and relaying characteristics are discussed showing the need for sensitive earth fault discrimination, particularly to detect broken conductors in regions of high earth resistance. Designs of pole-mounted, autoreclosers are discussed. Sectionalising equipment is also reviewed in the context of fault management on spur lines.

Metal-clad switchgear, automatic protection, and remote control, with particular reference to developments during the last seven years

This paper outlines a picture of the progress during the last seven years in the field it deals with, and also sets out the problems of to-day and the lines of possible future practice. The period under review has seen a change of national policy in Great Britain, which practically puts an end to isolated plants for public service, and enhances the scope for economy by the running of the largest plants with interconnections between them and all other plants and loads. New conditions therefore arise, demanding a reconsideration of the problems dealt with by the author in their earlier form in 1925. The construction of the various types of metal-clad switch-gear is diagrammatically represented by illustrations brought upto date and concluding with a forecast of the full application of the principle of complete enclosure and immersion of all conductors and insulators for all voltages as the means of safeguarding life and property, and securing continuity of supply. Certain basic constructional features are dealt with more fully. These are, in general, typical of changes, some brought about by research and practical experience, and others evolved in the quest by designers for something ?different? in a period of intensified competition. An attempt is made to classify the prevalent forms of arcing contacts and busbar enclosures, and to state the effects of such details, particularly of the latter, on the general lay-out of the whole switching equipment. The use of metal-clad switchgear has of itself effected such economy in buildings that outdoor switchgear has not been widely used in Great Britain; but reasons are given in favour of outdoor construction. A review is given of the present position of rating, performance, and selection of circuit breakers, in order to lead up to a better understanding of the interpretation of these terms, and in view of the need for an international agreement on standard conditions for assessing ratings. Results of tests in a British 1500000-kVA testing station are recorded as examples. The need for economy is accepted as a fundamental object for the designer, but it should be effected by the reduction of components rather than by lowering the quality of details. Some ways and means of effecting economy along these lines are therefore proposed. The automatic isolation of faults has a particular bearing on the protection of property and supply; in the latter its performance in procuring stability is considered to be of paramount importance, and in the former its speed of operation is a criterion of equal merit. In the light of these principles the known protective systems are reviewed, and some new developments are mentioned. This study is in the main a history of the quest of the ideal, which, for the sake of economy alone, excludes the pilot cables of the earlier systems, and thus an opening occurs for the use of telephone pairs. Some original research is recorded for the avoidance of interference with the other uses of such telephone conductors arising from accidents on main lines. Further, the resistance of heavy arcs has been experimentally ascertained, with a view to making clear the necessity of using reactance instead of impedance as the true means of measuring distance for the accurate loaction and discrimination of faults in the distance types of proctection. The evolution of automatic and remote-control systems is traced from the earlier power-station control-room to the present-day design of regional control from a central room, and the forecast for the future favours the increasing application of mechanical principles, which in the first place tends to avoid errors, and in the second place takes over the responsibility of action in an emergency.