Title :
A Finite-Element-Based Domain Decomposition Method for Efficient Simulation of Nonlinear Electromechanical Problems
Author :
Wang Yao ; Jian-Ming Jin ; Krein, Philip T. ; Magill, Matthew P.
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Illinois at Urbana-Champaign, Urbana, IL, USA
Abstract :
The dual-primal finite-element tearing and interconnecting (FETI-DP) method is combined with the Newton-Raphson method to expand the capability and improve the efficiency of 3-D finite-element analysis (FEA) of nonlinear electromechanical problems. Despite its modeling capability and high degree of accuracy, FEA has high computational complexity, especially for nonlinear analysis. The FETI-DP method is a robust domain decomposition method, which has been enhanced and applied to solve electromechanical problems involving linear materials. In this paper, the FETI-DP method is extended with the Newton-Raphson method to address problems involving nonlinearity and saturation. Using parallel computing techniques, the total computation time is reduced significantly. Linear and nonlinear regions are separated using the FETI-DP method. This further improves simulation efficiency and flexibility. Cubic splines and relaxation techniques are adopted to ensure stable and fast convergence of the Newton-Raphson method. The performance of the proposed method is compared with infolytica´s MagNet, a commercial 3-D FEA solver.
Keywords :
Newton-Raphson method; computational complexity; convergence of numerical methods; electric machines; finite element analysis; splines (mathematics); 3D finite-element analysis; FETI-DP method; Newton-Raphson method; commercial 3D FEA solver; computational complexity; convergence; cubic splines; dual-primal finite-element tearing and interconnecting method; infolytica MagNet; linear materials; linear regions; nonlinear electromechanical problems; nonlinear regions; parallel computing techniques; relaxation techniques; robust domain decomposition method; rotating electric machines; Computational modeling; Convergence; Finite element analysis; Magnetic domains; Magnetic separation; Splines (mathematics); Vectors; Domain decomposition; electric machines; finite-element analysis (FEA); nonlinear magnetic; parallel algorithms;
Journal_Title :
Energy Conversion, IEEE Transactions on
DOI :
10.1109/TEC.2014.2303987