Application of statistical interpretation technique for frequency response analysis and detection of axial displacement in transformer winding

Transformers are one of the most important and expensive equipment in high voltage electrical system. Any fault in power transformer may reduce power system efficiency and any sever fault may led to complete failure of electric network. Thus regular maintenance of power transformer is of utter importance. Frequency response analysis(FRA) is an emerging non destructive condition monitoring and fault diagnostic tool. It is an offline condition monitoring technique used to check the mechanical and geometrical integrity of transformer winding, core and clamping structures. FRA is accurate and highly sensitive and it is conducted by measuring the transfer function of the transformer winding over a wide range of frequency. The graphical nature of FRA requires experienced expert for interpretation and analysis of the result. In this paper a statistical evaluation of FRA result has be proposed which can be easily interpreted by untrained personnel. For this purpose different statistical indicators has been used to determine fault condition and its severity. The experimental studies were conducted on a 3 MVA, 33/11 kV, three phase power transformer.