Fault Diagnosis Of Partial Discharge In Power Transformer Based On Wavelet Analysis

Power transformer is to ensure the normal operation of the safe operation of the power system prerequisite. Transformers accident will lead to local or even a large area power outage, it would seriously affect the normal production and life. Therefore, it is related to the operating state of electricity generation. This paper makes judgment on the current situation of power transformer. It can improve the reliability and safety of the transformer, and reduce the blind overhaul, and the traditional method in diagnosis of precision is not enough in time, bring greater security risks.Firstly this paper introduces the domestic transformer fault diagnosis method of partial discharge current development. And point out their shortcomings. The method is rapid and accurate diagnosis of partial discharge in transformer fault. Application of wavelet analysis and analytic signal transformer partial discharge signals are processed. In this paper, Haar wavelet basis functions, construct a frequency-domain model compactly supported wavelets to deal with frequency of occurrence of each scale decomposition domain overlap. Extract the signal in each scale amplitude envelope, phase, and instantaneous frequency, and this characteristic parameter is used to determine the initial troubleshooting.Then, according to the concept of transformer partial discharge faults with uncertainty information entropy. Combined with the theory of entropy in information theory, established based on the singular spectrum entropy time domain, frequency domain power spectrum entropy, time-frequency domain wavelet energy spectrum entropy and wavelet spectral entropy spatial characteristics of the new fault diagnosis method and as a comprehensive evaluation of quantitative characteristics of the transformer discharge failure indicator. The information entropy to be normalized and wavelet analysis will be added to the information entropy in time-frequency analysis that wavelet energy spectrum entropy, it is relative to the power spectrum entropy reflects the differences in the distribution of the partial feature of the signal and change. Reuse entropy quantitative diagnostic methods from more accurate diagnosis of partial discharge fault. Using the signal processing toolbox of MATLAB to realize simulation of the diagnostic process, proved that this fault diagnosis method of transformer partial discharge is effective.