Finite element analysis of 3D electric machine problems

Yao, Wang

Permalink

https://hdl.handle.net/2142/16475 Copy

Description

Title

Finite element analysis of 3D electric machine problems

Author(s)

Yao, Wang

Issue Date

2010-06-22T19:37:16Z

Director of Research (if dissertation) or Advisor (if thesis)

Jin, Jianming

Department of Study

Electrical & Computer Eng

Discipline

Electrical & Computer Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Date of Ingest

2010-06-22T19:37:16Z

Keyword(s)

• Finite element
• Electric Machine
• Simulation

Abstract

The objective of this research is to develop an efficient numerical simulation tool to analyze electric machine problems. As one of the largest fossil energy consumers, electric machines have to be designed properly with robust simulation tools that are capable of handling the inherent complexity, such as multiple materials, complicated geometry, eddy currents, and saturation. In this thesis, we adopt the finite element method, one of the most powerful numerical schemes, to analyze electric machine problems. Starting from Maxwell’s equations, we derive the governing equations using both magnetic vector and electric scalar potentials. We then discuss the finite element implementation, which includes the use of higher-order elements and isoparametric elements. Since electric machines usually operate in a low frequency band, it is necessary to handle the low-frequency breakdown problem properly. In this thesis, we propose two possible solutions. One is to solve the singular system directly by using an iterative solver. Another is to regularize the singular matrix with tree-cotree splitting. An algorithm of finding the minimum spanning tree is given. The two solutions have different effects on the convergence of iterative solvers, which is important to the efficiency of the simulation tool. Besides, the convergence of iterative solvers can be affected by other factors, such as different formulations and different preconditioners. In order to further improve the efficiency of the algorithm, we conduct a detailed convergence discussion. Finally, we model the nonlinear problem using the Newton-Raphson method. We utilize cubic splines and relaxation factors to improve the convergence of the Newton-Raphson iteration.

Graduation Semester

2010-5

Permalink

http://hdl.handle.net/2142/16475

Copyright and License Information

Copyright 2010 Wang Yao

Owning Collections