Research Of Fault Diagnosis System Of Power Transformer Based On Wavelet Transform

As the most important equipment of power system, there is a close relationship between the Power system's operation and the transformer's operating status. Because of the long time operation, the transformer is unavoidable with fault. In order to find the hidden danger in time, avoid sudden fault and improve the power quality, the research of on-line monitoring system and new diagnosis method is very important.An online monitoring system of power transformer has been developed by author. The system's overall design and the achievement of each function module are introduced in this paper. The system gives a synthesis diagnosis for the transformer by analyzing several kinds of signal in real time such as current, voltage, pressure of cool fans, vibration signal delivered through the Ethernet. There are several advanced technologies applied in this system, such as Ethernet communication, virtual instrument, database and digital signal processing. The Ethernet communication makes the system has an advantage of real time ability and high reliability; the application software based on the virtual instrument has the feature of friendly interface and high-efficiency; the signal processing based on the wavelet analysis method can find the feature of the signal effectively and give a diagnosis for the operating transformer.The method based on the vibration signal of the transformer is a new method studying in civil and abroad in recently years. The author hopes to find an effective method to diagnose the status of the core and winding of the transformer by analyzing the vibration signal. The key technology is to find the feature of the vibration signal and gives a criterion for condition diagnosis.Taking advantage of the powerful function for signal processing of the wavelet analysis method, the vibration signal is analyzed both in time and frequency domain. At last, a diagnosis method based on frequency band-energy is proposed in this paper. The analysis result indicates that using wavelet method, the feature of the signal can be found effectively and the diagnosis of mechanical status of the core and winding based on the frequency band-energy method is applicable.