P. Purkait

Time and frequency domain analyses based expert system for impulse fault diagnosis in transformers

The presence of insulation failure in the transformer winding is detected using the voltage and current oscillograms recorded during the impulse test. Fault diagnosis in transformers has several parameters such as the severity of fault, the kind of fault and the location of the fault. Detection of major faults involving a large section of the coils have never been a big issue and several visual and computational methods have already been proposed by several researchers. The present paper describes an expert system based on re-confirmative method for the diagnosis of minor insulation failures involving small number of turns in transformers during impulse tests. The proposed expert system imitates the performance of an experienced testing personnel. To identify and locate a fault, an inference engine is developed to perform deductive reasoning based on the rules in the knowledge base and different statistical techniques. The expert system includes both the time-domain and frequency-domain analyses for fault diagnosis. The basic aim of the expert system is to provide a non-expert with the necessary information and interaction in order to make fault diagnosis in a friendly windowed environment. The rules for fault diagnosis have been so designed that these are valid for the range of power transformers used in practice up to a voltage level of 33 kV. The fault diagnosis algorithm has been tested using experimental results obtained for a 3 MVA transformer and simulation results obtained for 5 and 7 MVA transformers.

Impulse fault classification in transformers by fractal analysis

Transformers are usually subjected to lightning impulse tests after assembly for assessment of their insulation strength. In the case of a fault the resulting winding current gets changed to a certain extent. The pattern of the fault currents depends on the type of fault and its location along the length of the winding. This paper describes the application of the concept of fractal geometry to analyze the properties of fault currents. Fractal features such as fractal dimension, lacunarity used for image surface recognition and the sliding window algorithm used for fractal analysis of waveform have been employed for classification of transformer impulse faults. Experimental results obtained for a 3 MVA transformer and simulation results obtained for 3 MVA, 5 MVA and 7 MVA transformers are presented to illustrate the ability of this approach to classify insulation failures. The results indicate that this new approach possesses reasonable abilities for waveform pattern discrimination.

Monitoring of inter-turn insulation failure in induction motor using advanced signal and data processing tools

Detection of stator winding inter-turn insulation failure at early stages is crucial for promoting safe and economical use of induction motors in industrial applications. Whereas major insulation failures involving larger percentages of winding are easily discernible from magnitude of supply current, minor inter-turn insulation failures involving less than 5% of turns often go undetected. The present contribution reports experimental results of minor faults due to inter-turn insulation failures in stator windings of induction motor under different loading conditions being analyzed using data and signal processing tools combining Parks Transform and Cross Wavelet Transform. Rough Set Theory (RST) based classifier has been used for fault severity monitoring.

Performance of a load-immune classifier for robust identification of minor faults in induction motor stator winding

Reliable detection of induction motor stator winding insulation failure at its early stages is a challenging issue in modern industry. Insulation failure between small number of turns, involving less than 5% turns of phase winding are often indiscernible and detection becomes even more complicated when motor operates at varying load levels. In line-fed motors, supply voltage unbalance is another inadvertent issue which may tend to exhibit current signature similar to stator winding inter-turn insulation failure case. The proposed work presents a robust system, to identify severity of stator winding insulation faults when an induction motor with random wound stator winding works under such operating conditions. In the present work, various features obtained from time, frequency, timefrequency, and non-linear analysis of stator currents at various stator winding short circuit faults and supply voltage unbalance conditions for different load levels have been studied. A Support Vector Machine based Recursive Feature Elimination (SVM-RFE) algorithm is used to identify the features which can provide discrimination information related to severity of fault level, independent of supply voltage unbalance and immune to load level variations. Among the extracted features, features obtained through Detrended Fluctuation Analysis (DFA) are found to be most robust for this purpose. Finally a Support Vector Machine in Regression mode (SVR) has been formed to identify winding failures employing the optimum number of features selected through SVM-RFE technique.

Separating induction Motor Current Signature for stator winding faults from that due to supply voltage unbalances

Statistical spreads of the surveys suggest that stator winding faults are one of the most prevailing faults in induction motor. Most of the methods for stator winding inter-turn fault diagnosis are based on Motor Current Signature Analysis (MCSA) combined with signal-and-data processing tools. Fault diagnosis using MCSA becomes more challenging when stator current signatures due to winding short circuit fault and supply voltage unbalance appear identical. The present paper proposes a method through analysis of Parks Vector Modulus (PVM) to discriminate stator winding inter-turn fault cases, from supply voltage unbalance conditions where both cases exhibit apparently similar kind of current signatures. A series of experiments have been performed on a motor with different degrees of stator winding inter-turn faults. The same motor under healthy condition was also tested while operating under unbalanced supply voltages that caused similar current unbalances as in the case of inter-turn short circuit faults. This work aims at identification of the motor voltage unbalance conditions separately from inter-turn fault cases through detection of high frequency signals present in different PVM profiles. Signal processing tools such as Fast Fourier Transform (FFT), Discrete Wavelet Transform (DWT) and Power Spectral Density (PSD) calculation have been employed to discriminate inter-turn short circuit faults from supply voltage unbalance conditions of the motor at different load levels. Entire analysis presented in this paper is based on experimentally obtained motor current signatures.

Application of wavelet transform to discriminate induction motor stator winding short circuit faults from incipient insulation failures

Stator winding insulation faults in induction motor can be classified in two categories namely, direct inter-turn short circuit faults and incipient insulation failures. Both these two types of faults, when involving less number of turns, may remain undetected by normal protection schemes since such minor faults do not hamper normal operation of the motor. However, if these faults are not caught early they can lead to major failures in stator winding. The fault detection problem becomes more complicated when direct inter-turn short circuit faults and incipient insulation failures exhibit similar fault current magnitudes. The present contribution reports experimental results on an induction motor where both these two types of faults have been emulated. Parks Transformation has been used to extract AC components of the Parks Vector Modulus (PVM) of three phase line currents under different operating conditions of the motor. Continuous Wavelet Transform (CWT) has been used to extract several features from the non-stationary AC components of PVM. These features have been used to discriminate stator winding inter-turn faults from equivalent incipient insulation failures.

Relating stator current Concordia patterns to induction motor operational abnormalities

With industrial growth it has become necessary to monitor the condition of three phase squirrel cage induction motor, as it is the most sensitive part in the production line of industry. Special attention has been given to non-invasive methods which are capable of detecting fault by acquiring major data without disassembling the machine. The present paper discusses a few basic yet important aspects of condition monitoring of the induction motor based on investigation of Concordia patterns obtained through the Parks transformation of stator currents. The current Concordia pattern is greatly influenced by the internal condition of the motor as well as the external supply condition. The prime focus of the diagnostic technique is on identifying the change in shape of current Concordia patterns and shifting of its major axis orientation corresponding to stator inter-turn fault conditions with varying severity and varying power quality problems of supply system, like voltage sag and swell. This proposed method can be efficiently employed to discriminate and assess the operating condition of induction motor due to internal stator winding fault and power quality problems of the external supply system.

Wavelet and SFAM based classification of induction motor stator winding short circuit faults and incipient insulation failures

Inter-turn fault in stator windings is one of the prime reasons behind a significant percentage of induction motor failures. Inter-turn short circuit fault may develop either due to complete failure of insulation between turns causing direct interturn short circuit fault or due to partial degradation of insulation between turns resulting in incipient insulation fault. Common protection schemes accompanied with induction motors in industrial applications normally fail to detect these faults especially when a minor number of turns are involved in such faults. Detection of these faults at their early stage can substantially reduce the possibilities of serious damage to the motor and consequently financial losses, environmental damage and probable personnel injury etc. Detection of such faults becomes a tricky task when these two different types of faults exhibit identical unbalance in motor supply currents. In the present study, two different series of experiments were carried out on an induction motor which was subjected to operate under direct short circuit fault and also under equivalent incipient insulation fault in stator winding. Using Extended Parks Vector Approach, Parks Vector Modulus (PVM) have been estimated from captured three phase current signals under different operating conditions of the motor. Several fault features were extracted from AC components present in PVMs by applying Continuous Wavelet Transform (CWT). Then, the fault features were fed to a Simplified Fuzzy Art-Map (SFAM) based classifier which has been found to perform accurately in identifying separately the direct inter-turn short circuit fault levels and equivalent incipient insulation failure conditions.

Anomalies in harmonic distortion and Concordia pattern analyses in induction motors due to capacitor bank malfunctions

It is a common practice in industries these days to incorporate properly sized capacitor banks across induction motors for several reasons including power factor correction, reducing distortions, increasing capacity, etc. Total harmonic distortion (THD) and power factor (PF) are used in such cases to quantify the improvements obtained through connection of the external capacitor banks. On the other hand, one of the methods for assessing the motor internal condition is by the use of Concordia pattern analysis. Though adequate precautionary measures are adopted, the capacitor banks may sometimes malfunction, leading to damage of one or more of the individual capacitors in the whole bank. Such a minor fault in the capacitor bank is often not apparently discernible. This may however, give rise to substantial degradation of power factor correction performance and may also damage the supply profile. The case is more severe with the fact that the Concordia pattern gets distorted due to such external capacitor faults, and can give anomalous results about motor internal fault analyses. The aim of this paper is to present laboratory and field test results to have a close look at the anomalies in harmonic distortion and Concordia pattern analyses in induction motors due to capacitor bank malfunctions.

A combined image processing and Nearest Neighbor Algorithm tool for classification of incipient faults in induction motor drives

Incipient faults in Inverter of Induction motor drive are detected and classified.Concordia pattern is obtained from 3-phase output current of faulty inverter.Concordia pattern is converted to binary image.Shape context based Nearest Neighbor algorithm is used to classify the Concordia image. Switching Devices such as IGBT used in Pulse Width Modulation (PWM) Inverter feeding an induction motor often suffer from different types of incipient faults like improper contact points, poor connections and problematic solder joints. These are due to ageing or prolonged operation in unfriendly environments. These faults need to be detected at their initial stages to prevent subsequent spreading of faults. In the present work, different variations of the above mentioned faulty cases in a PWM-Inverter have been studied by recording three phase inverter output current profiles and converting them to Concordia patterns. It has been observed that the Concordia patterns are quite different in shapes for different types of faults. A suitable image based shape descriptor has been applied to extract relevant information from these Concordia patterns. Finally, Nearest Neighbor Algorithm is employed on this information to identify the nature and location of faults. Performance of the algorithm is found to be quite satisfactory when its results are compared with two more related algorithms. Display Omitted