DISTRIBUTION OF VERY FAST FRONTED SURGES IN TRANSFORMER WINDINGS

Imad  R. SUKAR

The increasing usage of Gas Insulated Substations (GIS) has led to anomalous failures of GIS components and connected plant. Such failures are thought to be the results of very fast fronted surges generated by switching operations performed within GIS. Transformers are especially considered at risk as they are increasingly used at higher voltages.

The problem of very fast fronted surge distribution interacts. Thus, adequate representation of all components is of utmost importance in order to investigate  the characteristics (in particular, wavefront nature) of the surges impinging on transformer windings and the way they distribute themselves inside such windings. In this thesis, the latter aspect of the problem is addressed through the development of a computer model of the section in the line-end region of the winding.

The model is based on multiconductor transmission line theory together with modal analysis (including two‑port network representation) and considers the individual turn as the basic element. The electrical parameters of the coil under study are obtained in a rather simple way. The capacitance’s matrices are calculated by approximating the walls of the different turns to parallel plate capacitors and the inductance matrices are calculated from the relation between the inverse capacitance matrices in air, velocity of light and loss factor. The series losses are due to skin effect.

The model accurately predicted the surge distribution in the top four sections of the HV winding, where insulation failure is most likely to occur, in the first few hundred nanoseconds after surge arrival. The validity of the model was demonstrated through comparison with experimental results obtained from measurements on the real HV winding. Such accuracy also justified the derivation of the electrical parameters.

A sensitivity analysis of various parameters illustrated the ability of the developed model to accurately predict insulation status of different transformers under different operating conditions.

The computer model developed in this thesis could be used to adequately represent transformer windings in future studies allocated to the investigation of the‑ wavefront characteristics of the surges created by switching operations in GIS. As a result, the exact nature of the surges reaching the transformer will be obtained together with the way they distribute themselves inside the windings leading to an accurate assessment of the transformer insulation status.