DC Bias Identification And Evaluation Of Transformer Based On Vibration Signal

Due to the influence of HVDC power transmission,urban rail transit stray current,geomagnetic storm and so on,DC bias of transformer is more and more common,which has more and more serious influence on transformer itself and power system,so it is urgent to carry out relevant research on it.Therefore,it is great significance to identify and evaluate the DC bias of transformer to ensure the safe and reliable operation of power system.There is no electrical connection between vibration monitoring and transformer,so it is suitable for online identification and evaluation of transformer DC bias.This thesis focuses on the analysis of vibration characteristics,extracts vibration characteristic values under DC bias.In order to realize the goal of DC bias states identification and evaluation of transformers,the following research works are carried out:Firstly,the vibration characteristics of transformer under normal state and DC bias are studied respectively.Based on the field measured transformer vibration signals,the characteristic values in time domain and frequency domain of transformer vibration signals under normal state and DC bias state are analyzed.Secondly,based on the analysis and research of transformer vibration characteristics,the characteristics of transformer vibration signals in frequency domain under DC bias,short circuit fault and power network harmonic,are further compared and analyzed respectively.Then a DC bias identification method of transformer based on frequency domain analysis is proposed.The energy of the 50 Hz frequency doubling component of the vibration signal except 100 Hz M is extracted to identify the abnormal state of the transformer,and the duration of the sum of energy is used to eliminate the interference of short circuit fault.At the same time,the ratio of the energy of 50 Hz odd frequency doubling component to M is extracted to eliminate the harmonic interference of the power grid.The validity and feasibility of the method are verified by the measured signals of 500 k V and 220 k V main transformers.Then,the wavelet singular entropy of transformer vibration signals is extracted by integrating continuous wavelet transform,singular value decomposition and information entropy theory,and the wavelet singular entropy of vibration signals with and without DC bias is compared and analyzed.Based on wavelet singular entropy of vibration signal,a DC bias identification method is proposed.Then the accuracy and feasibility of the identification method are verified by the measured vibration signals.Finally,in order to further evaluate the severity of DC bias,the characteristics of vibration signals in time domain,frequency domain and time-frequency domain,under different DC bias severity are compared and analyzed.Based on RBF artificial neural network and improved analytic hierarchy process,an evaluation method for DC bias of transformer is proposed.The characteristics of vibration signals in time domain,frequency domain and time frequency domain were input to the RBF neural network respectively,and then the information fusion of the three features was carried out to realize the evaluation of the severity level of DC bias of transformer.The feasibility of the proposed method is verified by the measured vibration signals of 500 k V transformer.Based on the analysis of transformer vibration properties and the study of vibration signal characteristics,transformer DC bias identification and evaluation methods based on vibration signal are proposed in this thesis.The identification and evaluation methods of DC bias proposed in this thesis are applied to the transformer in actual operation,and the results show that the method is feasible and effective.It provides theoretical reference and technical guidance for the treatment of DC bias,and is of great significance to ensure the safe and reliable operation of power system,and has certain engineering application value.