Research On Transformer Winding Deformation Monitoring Based On Parameter Identification

The safe and stable operation of power transformer is of vital significance for the sustainable and reliable power supply.But transformers are often under the impact of short-circuit current caused by external faults,which leads to the permanent winding deformation and ultimately reduces the anti-short circuit ability and causes insulation damage.Therefore,the on-line monitoring of transformer windings has important practical significance.This article has done several works in online deformation monitoring,deformation type distinguishing and deformation degree diagnosing.Based on the theoretical analysis of winding inductance and capacitance parameters calculation method,a two-dimensional finite element model is established by ANSYS software.Verification shows the finite element model and the calculation process are reasonable and accurate and can be used as the basis for subsequent analysis.A global sensitivity analysis method of winding electrical parameters is proposed.Response surface method and Sobol 'analysis were used to quantitatively analyze the sensitivity of each electrical parameter to the deformation of different types of windings.The sensitivity of different electrical parameters to the same deformation was further analyzed by the variation coefficient of each parameter.The analysis results provide a reference for the establishment of identification model.A new winding parameter identification model and online identification method are proposed.Based on the results of sensitivity analysis,an identification model with higher sensitivity is established,which is more abundant than the traditional model and is more conducive to the judgment of winding status.Integral model recursive least square method is used to identify parameters online.The calculation speed is fast,the precision is high,and the stability is strong.Based on the parameters value of on-line identification,a classification method of winding deformation and a diagnosis method of winding deformation degree are proposed.The nonlinear support vector machine is used to classify the deformation types.A support vector regression model is established to diagnosis the winding deformation degrees.The verification shows that the method has high accuracy in the classification and accurate to the degree of deformation.