The Transformer Winding Deformation Detection System Based On Vibration Analysis

The normal work of the power transformer is the important guarantee for the safe operation of the power system.It is significant to detect the transformer winding deformation for the reliability of the operation,because the winding is relatively prone to fault in transformer.At present,the common detection methods applied with transformer off line cannot immediately detect the state of winding.Therefore,a detection method is proposed based on ensemble empirical mode decomposition(EEMD),principal component analysis(PCA)and K-Nearest Neighbor(KNN)classification and identification of transformer winding deformation detection.The detection system of transformer winding deformation based on vibration analysis is designed in this thesis.The intuitively classified and identified defection of the transformer winding is implemented.First of all,the basic principle of the source and the propagation of transformer vibration is introduced in this article.The core vibration is mainly derived from the magnetostriction of the silicon steel sheet by analyzing the vibration mechanism of transformer core.With the model of transformer winding vibration based,the basic characteristics of the transformer winding vibration signal are studied,which lays the foundation for transformer winding vibration signal acquisition and detection.Secondly,the transformer winding vibration signal acquisition system is designed in the thesis.The suitable appropriate sensor,data acquisition board and data acquisition system are selected,based on the characteristic of the vibration signal,reducing the surrounding interference to ensure the reliability of vibration signals.The optimum position of the sensor is determined by the experimental study of the different placement of the senor,to improve the repeatability of the data acquisition.Comparing the vibration signal of transformer in different time and conditions,the characteristics of the vibration signal in the working state of transformer winding are achieved,which provides the basis for the transformer winding deformation detection software system.Finally,a software system of transformer winding detection is designed.The characteristic vectors extracted from transformer vibration signals by ensemble empirical mode decomposition are classified based on classified by principal component analysis and K-nearest neighbor method and identified the states of transformers intuitively.A transformer winding detection software system is developed using MATLAB GUI in this thesis.The system can analysis the transformer vibration signal in time-domain and frequency-domain and realize the classification and identification of the states of the transformer winding intuitively.