Date of Award

4-2019

Degree Name

Doctor of Philosophy

Department

Electrical and Computer Engineering

First Advisor

Dr. Pablo Gomez

Second Advisor

Dr. Damon A. Miller

Third Advisor

Dr. Azim Houshyar

Abstract

Power transformers are static devices capable of transferring energy from one electrical circuit to another by means of electromagnetic induction at the same frequency. The most common application of transformers in an electric power system is to step up or down the voltage level of an electric grid. This is essential for the efficient transmission and distribution of electricity from generating plants to different types of consumers. Electrical energy used in residential, commercial and industrial settings is delivered to consumers by means of transformers, which normally operate 24 hours a day, 365 days a year. Therefore, design improvements that result in an increase of efficiency and reliability of these devices will be reflected in an enhanced overall performance of electric power systems.

Modern transformers are subjected not only to fast front excitations due to lightning and switching conditions, but also to fast and repetitive pulses related to the widespread connection of power electronic components. These phenomena can produce large transient overvoltages and dielectric stresses that can damage the transformer or significantly reduce its life expectancy. The definition of models able to accurately and effectively predict these phenomena is essential for the insulation design of transformers.

In this dissertation, a winding model to predict the transient voltage distribution in a disk-type transformer is described. The presented model is a function of frequency, with electrical parameters computed using the Finite Element Method and analytical formulas. In order to increase the accuracy of the presented model, frequency-dependent core and proximity losses are calculated and validated using the Finite Element Method. In addition, this dissertation introduces a simple and accurate formula for the computation of frequency-dependent proximity impedance. Furthermore, a complete model of a prototype disk-type transformer with 2000 turns is implemented and verified experimentally.

Access Setting

Dissertation-Campus Only

Restricted to Campus until

4-2021

Share

COinS