Fuzzy logic based impulse test analysis

Virtual Instrument for Lightning Impulse Tests

Abstract: We propose an objective formulation of the impulse analysis problem from a signal analysis viewpoint. The winding response is quintessentially that of a deterministic network to a finite energy signal, with breakdown and partial discharge being inherently nonlinear events. A significant improvement to the acquisition of waveforms is demonstrated by a virtual instrument approach. It retains the advantages of the time- and frequency-domain methods. The drawbacks of the transfer function method are highlighted and a new piecewise linear approach is proposed for analysis. Experiments on a discrete lumped parameter model of the winding are used to validate the PXI based instrument.

Wavelet power ratio signature spectrum analysis for prediction of winding insulation defects in transformer and shunt reactor

Abstract: The reliable analysis for identification of winding insulation defects in transformer and shunt reactor are necessary for power industries during lightning impulse test. In this paper, wavelet power ratio signature spectrum analysis is proposed to identify the short duration of partial discharge (PD) pulses of 170 ns and 30 ns within windings due to impulse voltage excitation. It provides a solution to the problems associated with several random features in the measured winding responses at neutral terminal of the winding due to PD pulse through visual assessment of mutual strength between normal winding response and one with PD response using appropriate mathematical formulation in wavelet transform. In this context, the normalized response due to reduced impulse voltage excitation is formed to be a basis function to identify the PD pulse. If the surge impedance characteristic of the winding is changed due to PD pulse at full wave impulse voltage excitation, then formulated wavelet power ratio signature spectrum analysis will extract the evidence of potentially interesting features through assessment between basis function response and one with PD response. To prove feasibility of the proposed analysis, experimental analysis is performed on a layer winding.

Nonlinear Interpretation Technique for Lightning Impulse Test

Abstract: The nonlinear interpretation technique is proposed to detect the fault in transformer windings during the impulse test. To prove feasibility of the proposed technique, 90-MVA 220/110-kV, 250-MVA 500/275/33-kV transformers, and a 22-kV interleaved winding are used.

Impulse Fault Detection and Classification in Power Transformers with Wavelet and Fuzzy Based Technique

Abstract: Impulse testing of transformers after assembly is a routine procedure carried out for the assessment of their winding insulation. During impulse test insulation failure may result in two classes of winding faults in a transformer namely series faults and shunt faults. Series faults are due to the short between turns in the section and the shunt faults are due to the short between turns in the section and the ground. This paper aims at deriving a technique for the detection and classification of impulse faults in a transformer winding using wavelet transform and a fuzzy Inference system. A specially designed 6.6 kV model layer winding is considered for the study. The entire winding comprising ten sections are divided into three regions namely sections near line end, sections near the neutral end and the sections in the middle of the winding. The neutral currents are recorded with series faults and shunt faults introduced in the sections belonging to the three regions. Continuous wavelet transform is applied on these neutral current records to extract the discriminating features. The features extracted from the wavelet transformed signal are the second most predominant frequency, the time range at which it occurs and the corresponding wavelet coefficient. A fuzzy Inference system is designed and implemented using Matlab software with these three features extracted from the wavelet transformed signal as inputs and generates an output that classifies the fault and no fault conditions. It is observed that the results are satisfactory.

Lecture Notes in Computer Science: S Transform for the Analysis of Impulse Faults in Transformer

Abstract: The tremendous and rapid growth of digital signal processing has motivated researchers to apply improved signal analysis techniques for the fault diagnosis of transformers. This paper demonstrates the application of Stockwell transform for the detection and analysis of impulse faults in transformers. Stockwell transform (S transform) is a time–frequency transformation method of signal analysis that conveys information directly in terms of time and frequency, and hence, interpretation of result becomes easier. Further, it produces the progressive resolution of the wavelet transform (WT) and maintains a direct link to the Fourier transform. The proposed method is validated through simulation of faults in a lumped parameter model of a layer winding transformer. The results are encouraging and pave way to develop an automated impulse fault classification system.

Fuzzy systems as universal approximators

Abstract: An additive fuzzy system can uniformly approximate any real continuous function on a compact domain to any degree of accuracy. An additive fuzzy system approximates the function by covering its graph with fuzzy patches in the input-output state space and averaging patches that overlap. The fuzzy system computes a conditional expectation E|Y|X| if we view the fuzzy sets as random sets. Each fuzzy rule defines a fuzzy patch and connects commonsense knowledge with state-space geometry. Neural or statistical clustering systems can approximate the unknown fuzzy patches from training data. These adaptive fuzzy systems approximate a function at two levels. At the local level the neural system approximates and tunes the fuzzy rules. At the global level the rules or patches approximate the function. >

Impulse testing of power transformers using the transfer function method

Abstract: The transfer function of a transformer winding is deconvoluted in the frequency domain from the digitally recorded neutral current and high voltage applied during impulse tests. The integrity of the winding insulation is determined by comparing the transfer function obtained at full and reduced test voltage. Differences between the transfer function plots reveal local breakdowns in the winding that can be dissociated from partial discharges. Thus the method permits unambiguous acceptance or rejection if the transformer and, since the transfer function is theoretically immune to changes in the applied impulse, also allows evaluation of the chopped-impulse test. Some 100 windings of large HV power transformers have been tested using the transfer function method, which on several occasions has revealed transformer faults as well a test setup problems that would have been missed or misinterpreted by conventional techniques. >

Fusion of soft computing and hard computing in industrial applications: an overview

Abstract: Soft computing (SC) is an emerging collection of methodologies which aims to exploit tolerance for imprecision, uncertainty, and partial truth to achieve robustness, tractability, and low total cost. It differs from conventional hard computing (HC) in the sense that, unlike hard computing, it is strongly based on intuition or subjectivity. Therefore, soft computing provides an attractive opportunity to represent the ambiguity in human thinking with real life uncertainty. Fuzzy logic (FL), neural networks (NN), and genetic algorithms (GA) are the core methodologies of soft computing. However, FL, NN, and GA should not be viewed as competing with each other, but synergistic and complementary instead. Considering the number of available journal and conference papers on various combinations of these three methods, it is easy to conclude that the fusion of individual soft computing methodologies has already been advantageous in numerous applications. On the other hand, hard computing solutions are usually more straightforward to analyze; their behavior and stability are more predictable; and, the computational burden of algorithms is typically either low or moderate. These characteristics. are particularly important in real-time applications. Thus, it is natural to see SC and HC as potentially complementary methodologies. Novel combinations of different methods are needed when developing high-performance, cost-effective, and safe products for the demanding global market. We present an overview of applications in which the fusion of soft computing and hard computing has provided innovative solutions for challenging real-world problems. A carefully selected list of references is considered with evaluative discussions and conclusions.

Identification and location of breakdown in windings

Abstract: It is possible to estimate the location of the fault based on a study of the resonant frequencies of the winding current. It is shown that the nearest neighbourhood rule based on a clustering approach provides a good framework for fault location. Additional signal acquisition in the form of the tank current can considerably simplify the identification of a fault to ground. Experimental investigations are performed in order to identify and locate breakdown in windings.

Programmable continuous time filters for impulse testing of transformers

Abstract: An objective formulation of the impulse testing procedure is proposed based on models for winding response as well as fault models. Experimental validation of the fault models for breakdown and partial discharge (PD) is provided. A new adaptive filtering approach is proposed for analysis during impulse tests. The procedure is validated with experimental investigation on a 10 section lumped parameter network.