Study On New Diagnosis Method Of Transformer Winding Deformation

Transformer is one of the core equipment of power system,the operation conditions of which directly affect the safe and stable operation of power system.Winding deformation is one of transformer faults of the main types,which threatens the normal operation of power system.In order to diagnose transformer winding deformation accurately and reduce transformer faults,the research status and some existing problems of the detection methods of transformer winding deformation are summarized and analyzed.On this basis,the diagnosis method of transformer winding deformation is studied from different angles,which uses the electrical information measured in the process of transformer operation and maintenance.The research content includes the following several aspects.A detection method of transformer winding deformation based on impedance-based frequency response analysis is found.In actual detection,the diagnostic accuracy of frequency response analysis is limited because of the noise interference and the influence of wire length,etc.Impedance frequency characteristics of transformer winding in low and intermediate frequency are further studied.Based on the wiring of frequency response analysis,relevant data of voltage and current is measured and processed to make impedance frequency curves.Transformer winding deformation is diagnosed through the comparison among the difference of response curves.The diagnosis results can be validated with the results of frequency response analysis,which can reduce the misjudgment of frequency response analysis caused by environment interference.The simulation results show that the proposed method can identify different types of transformer winding deformation effectively.An on-line fault location method of transformer winding deformation based on complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)and variable predictive model based class discriminate(VPMCD)is proposed.The steep rising edge of transient overvoltage contains a large number of high frequency components.The energy loss of traveling wave in winding deformation position is different from normal position when the traveling wave hit the transformer.The difference will reflect on certain frequency components of traveling wave.Transient overvoltage signal is measured in the end of the winding.Intrinsic mode function is obtained through CEEMDAN of the signal.Its correlation coefficients are calculated as fault feature.Variable predictive model is obtained to identify the position of deformation with the train of VPMCD.The simulation results show that the proposed method can identify the position of transformer winding deformation reliably,which has high practical value.An on-line detection method of transformer winding deformation based on variational mode decomposition(VMD)and probability density estimation is proposed.Short-circuit reactance is an important criterion of transformer winding deformation.But it presents certain randomness caused by noise interference at on-site measurement,influencing judgment of winding state.At first,VMD is used to de-noise electrical signal and extract fundamental mode component.Then short-circuit reactance is calculated on-line with fundamental mode component.Finally,with samples of short-circuit reactance calculated in each testing period,probability density function of normal distribution is obtained with parameter estimation.Mean of normal distribution is estimated to calculate deviation factor of short-circuit reactance,reflecting current state of windings.Results of simulation show that the proposed method can get estimate value of short-circuit reactance steadily,eliminating influence of noise and measurement error of equipment and detecting winding deformation reliably.