Abstract:SUMMARY
In this paper, a new method is introduced for identification of transformer winding fault through transfer function analysis. For this analysis, vector fitting and probabilistic neural network are used. The results of transfer functions estimation through vector fitting are employed for training of neural network, and consequently, probabilistic neural network is used for classification of faults. The required data for fault type identification are obtained by measurements on two groups of transformers… Show more
“…The first category includes methods according to which faults' classification is solely based on the rate of variations in numerical indices (statistical and mathematical indices) in specific frequency ranges [9]- [18]. The second category includes methods that use intelligent classifiers to distinguish faults [19]- [25]. In these methods, the necessary features of frequency response (mainly the statistical and numerical indices) are extracted, and these features are used for training and testing classifiers.…”
Section: Introductionmentioning
confidence: 99%
“…In [19], [20], based on the TF estimation with the help of vector fitting, the necessary features of the measured TFs for four faults AD, RD, DSV, and SC were extracted and these faults are classified by using PNN [19] and SVM [20]. A distinction has been made between electrical and mechanical faults (AD and RD only) and the inrush current through ANN and the DT in [21].…”
With the expansion of the use of frequency response analysis (FRA) as a reliable tool for fault detection in transformers, more capabilities of this method are discovered every day. So that today the number of transformer faults that can be identified by FRA method has also increased. One of the most critical steps in fault detection with FRA is to distinguish faults and classify them in different classes. In this paper, well-known intelligent classifiers (probabilistic neural network, decision tree, support vector machine, and k-nearest neighbors) are used to classify transformer faults. For this purpose, the necessary measurements are performed on the model transformers under the healthy condition and under different fault conditions (axial displacement, radial deformation, disc space variation, short-circuits, and core deformation). Then, by dividing the frequency ranges of the measured transfer functions of the transformer, a new feature based on numerical and statistical indices for training and validation of classifiers is proposed. After completing the training process, the performance of the classifiers is evaluated and compared by applying the data obtained from real transformers. INDEX TERMS Transformer, fault type detection, frequency response analysis (FRA), intelligent classifiers, measurement, numerical indices.
“…The first category includes methods according to which faults' classification is solely based on the rate of variations in numerical indices (statistical and mathematical indices) in specific frequency ranges [9]- [18]. The second category includes methods that use intelligent classifiers to distinguish faults [19]- [25]. In these methods, the necessary features of frequency response (mainly the statistical and numerical indices) are extracted, and these features are used for training and testing classifiers.…”
Section: Introductionmentioning
confidence: 99%
“…In [19], [20], based on the TF estimation with the help of vector fitting, the necessary features of the measured TFs for four faults AD, RD, DSV, and SC were extracted and these faults are classified by using PNN [19] and SVM [20]. A distinction has been made between electrical and mechanical faults (AD and RD only) and the inrush current through ANN and the DT in [21].…”
With the expansion of the use of frequency response analysis (FRA) as a reliable tool for fault detection in transformers, more capabilities of this method are discovered every day. So that today the number of transformer faults that can be identified by FRA method has also increased. One of the most critical steps in fault detection with FRA is to distinguish faults and classify them in different classes. In this paper, well-known intelligent classifiers (probabilistic neural network, decision tree, support vector machine, and k-nearest neighbors) are used to classify transformer faults. For this purpose, the necessary measurements are performed on the model transformers under the healthy condition and under different fault conditions (axial displacement, radial deformation, disc space variation, short-circuits, and core deformation). Then, by dividing the frequency ranges of the measured transfer functions of the transformer, a new feature based on numerical and statistical indices for training and validation of classifiers is proposed. After completing the training process, the performance of the classifiers is evaluated and compared by applying the data obtained from real transformers. INDEX TERMS Transformer, fault type detection, frequency response analysis (FRA), intelligent classifiers, measurement, numerical indices.
“…One approach to interpreting FRA data has been the application of Black Box modelling [17][18][19]. This approach is based on the development of mathematical models which accurately represent the frequency response between various terminal combinations.…”
Section: Introductionmentioning
confidence: 99%
“…Any change in the FRA will result in model parameter changes which can then be interpreted for indications of a structural change within the transformer. Researchers such as Bigdeli [18, 19] have used neural networks to classify the fault relative to the change in parameters. Such an approach is dependent upon training data for different transformer topologies and their corresponding fault conditions.…”
Frequency response analysis (FRA) is a tool that can be used to detect changes within the structural geometry of a transformer. One application of FRA is to detect winding displacement after the transformer experiences an overcurrent event. This paper proposes a methodology which will facilitate both locating and quantifying winding deformation. The procedure is based on the fitting of a Gray Box transformer model to FRA measurements which were recorded both before and after the fault. Subtle variation in key parameters of the model can then be used to quantify winding deformation severity. The research will demonstrate the applicability of this approach to a wide range of transformer applications by using FRA from a modified 1.3 MVA distribution transformer, and assessing the change in key parameters relative to the emulated buckling introduced to the windings. The authors propose that this approach is ideally suited for use within an automated diagnostic system which will locate, diagnose and quantify winding deformation within power transformers. High Voltage
“…At high current values, these induced currents cause overheating of the metal structure, which can degrade the properties of transformer oil, damage the painting of the tank, and damage the insulation of nearby cables . All these harms can be a cause of major fault in power transformers .…”
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.