| Transformer is one of the most important equipments in electric power system. One of the most hackneyed faults on transformer is winding deformation. Frequency response analysis (FRA) is an effective method for diagnosing winding deformation, which are difficult to detect by conventional measurements.FRA test is a non-destructive method based on the fact that every transformer winding has a unique frequency characteristic that is sensitive to changes in the electrical parameters of the winding. If winding is mechanically damaged, the physical position, sections or shape of the winding will change, which in turn alters the electric parameters of the winding and changes its frequency response characteristic.In this thesis, a simulation-based method is adopted to produce FRA results and winding deformation is presented by changing electric parameters of the model. In this model, changes of parameters R, Rs and Rg will cause changes of the frequency response characteristics with no resonant frequency shift. The simulation, in which the values of inductance L , serial capacitance Cs and ground capacitance Cg are also changed to simulate the real winding deformation, it shows that the influence on magnitude of amplitude frequency response increases with the frequency increases; whereas the influence on resonant frequency is quite different: the inductance L has more significant effects in lower frequency band, the serial capacitance Cs has more significant effects in higher frequency band and the ground capacitance Cg has more significant impacts on lower and higher frequency bands than middle frequency band. Simplified simulation model can be used in different frequency bands. The HV bushing model is established and the simulation results show that the HV bushing has a significant impact on the FRA results. Meanwhile, a distribution transformer is used to study the effect of winding deformation. It is found that the repeatability and consistency of three phases are excellent and the effect of the test leads length on the FRA results is significant. |
