Design And Study Of Transformer On-line Monitoring System Based On Vibration Signal Analysis

Power transformer is the most important equipment in the power system,which directly affects the safe operation of the entire power grid.So it is important to ensure the transformer’s safe operation.The vibration signal analysis is a kind of method which can diagnose the transformer status by measuring and analyzing the vibration signal of transformer.It is a powerful supplement to the traditional electric parameter monitoring method.The vibration monitoring method is used to obtain the vibration signal of the transformer through the vibration acceleration sensor which is stuck on the transformer body,and then the vibration signal is analyzed to monitor the operation of the transformer.During the whole monitoring process,there is no electrical connection with the power system,which greatly reduces the danger.In this paper,the transformer vibration is analyzed from transformer core and windings.It focus on the study of the force analysis for windingand core,andthe influence of winding deformation and core compression condition on vibration.By means of theoretical analysisand experimental verification,the vibration characteristics of the winding and the core is researched.In this paper,a new method of joint time-frequency analysis is proposed to deal with the particularity of non-stationary random signal,with the basic idea of designing the joint function of time and frequency.The advantages and disadvantages of each method are analyzed and compared by simulation signals.During the analysis of transformer vibration signal,the traditional Fourier analysis method and wavelet packet analysis method,Hilbert-Huang transform are used to do the analysis and contrast,which is validated for the analysis of transformer vibration data.Finally,according to the effect of endpoint effect on the accuracy of empirical mode decomposition method,the BP neural network is constructed for the transformer vibration signal,and this network will be extended at both ends of the signal,and thendecompose the new signal by EMD.The results prove the effectiveness of the method.A set of transformer vibration online monitoring system is developed,and the transformer vibration monitoring system hardware is selected.LabVIEW language is used to compile the program of transformer on-line monitoring,which can realize the real-time acquisition of transformer vibration signal.The relevant data are analyzed to determine the operation state of the transformer and alarm when the transformer fails.The main functions of the system include:Real time data acquisition,spectrum analysis of vibration data,time-frequency analysis of vibration data,Data storage and playback,the working state of the transformer,and alarming when it is beyond the threshold.