Optimal design of a transformer and the impact of the accuracy of the adopted modeling

• Grant number: 14/26845-0
• Support type: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): June 01, 2015
• Effective date (End): May 31, 2016
• Field of knowledge: Engineering - Electrical Engineering - Electrical, Magnetic and Electronic Circuits
• Principal Investigator: Luiz Lebensztajn
• Grantee: Caoe Ferreira Magalhães
• Home Institution: Escola Politécnica (EP). Universidade de São Paulo (USP). São Paulo , SP, Brazil

Abstract

The design of a transformer is an activity that includes a number of decision-making to the engineer. When seeking a good design of a transformer, there is essentially two distinct problems: I) what is the model that should be adopted to describe the physical problem? and II) what is the optimization method which appears more appropriate to address the problem? From the modeling point of view, a transformer can be described by at least one of three ways: I) by a simple equation for draft purposes only; II) by an equation of electrical circuits, which has a more refined level of detail and III) by the method of finite element elements. An effective way to understand the project, mainly when optimize it, is through models of evolution, i.e., adopting a simple set of equations to draft purposes, then a refinement, from electrical circuit equation and finally by using the finite element method for a detailed design. The advantage: it is possible to delawith many design variables in the simple model and few variables in the more sophisticated model, which makes it easier to optimization task, since the finite element models have higher computational cost. In this project we intend to adopt genetic algorithms as optimization method without any loss of generality. (AU)