Study On Vibration Characteristics Monitoring And Abnormal State Detection For Power Transformer

The normal operation of power transformer, which is one of the most important equipment of electric power system, is directly related to the security and stability of the entire electric power system. In the long running of transformer, the core and winding will be abnormal avoidably, which will cause power transformer failure. In order to detect hidden faults of power transformer and improve the operational reliability of power transformer, conducting research on on-line monitoring of transformer vibration signals is significant.Firstly, this thesis designs the system of on-line monitoring vibration of the power transformer and develops system platform of online vibration monitoring by using Lab VIEW software. The platform is able to continuously monitor the vibration signals of twelve positions in the transformer surface, as well as observe the frequencies, spectrums, trend charts and other relevant informations. In addition, the platform, with historical data analysis module, achieves the all-round monitoring to vibration signals of power transformer.Moreover, the vibration signals at different locations of transformer are analyzed based on frequency spectrums. The study found the vibration signals at the same measuring point of transformers with the same type have similar characteristic frequencies, but different amplitudes, so do the vibration signals at the same measuring point of three phases of one transformer. Besides, the vibration signals at upper location of one phase and lower location of the same phase of one transformer have similar characteristic frequencies and amplitudes, so do the vibration signals at the measuring point of one phase and the vertical measuring point of the same phase of one transformer. Therefore, under the the above conclusions, we can reduce the measuring point in the case of less precision, save costs and greatly reduce the workload of the data analysis.The vibration signals of transformer are analyzed based on wavelet packet transform. By using three levels wavelet packet transform, the original signal is decomposed into eight signals of different bands. Next step, we conduct a vector whose elements are energy of the eight signals. By comparing this vector and the standard vector of vibrations database, we can diagnose the status of core and winding of transformer. And then, we analyze a group of vibration signals through this method and confirm the validity of the method.