MODELING AND ANALYSIS OF POWER TRANSFORMER USING FINITE ELEMENT METHOD

Abstract

The problem of evaluating the various performance characteristics of power transformer such as temperature distribution, flux distribution, and losses, is an age long issue in electrical engineering, and an attempt to manually or analytically evaluate them is very difficult and subject to errors. Hence the development and application of Finite Element Method to complex engineering analysis of this nature. This research presented enhanced finite element model application to Power Transformer. The method applied was modeling and simulation. The Finite Element Analysis of a 30/40MVA Power Transformer Model was created using ANSYS MAXWELL Analysis Software. The input, output voltage, and frequency of operation of the power transformer was defined and inputted into the design parameter section from where the phase current was computed. Other parameters of the Power Transformer were selected as appropriate in the design stage before the mesh generation was carried out by process of adaptive discretization. The Solution of the Model, covered Finite Element Force Calculation, Thermal Analysis, Short Circuit Analysis, and Insulation Analysis. In summary, the adaptive discretization result for winding loss was 98.4kW, and that of core was 19.62kW, produced a maximum flux density of 1.84T. For thermal model computation, the result showed 94.30 was the hot spot temperature on the winding while 75.84 on the core. The electrodynamics force distribution along the windings during the short circuit condition, translating to 1.25kN for low voltage radial force and 652.69kN high voltage axial force. The electrostatic field distribution between the high voltage and low voltage was 10.48kV/mm, while that exiting between core and low voltage was 3.74kV/mm. The performances of the designed models were compared with the performance of the transformer, which was determined analytically, and the performance of the transformer was also determined experimentally and the results were confirmed. When the results were compared, it was seen that the designed models gave more optimum results.