Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).
Session Chair: Zhuoxiang Ren Session Chair: David Lowther
9:30am - 9:50am
Start Value Estimation Using Gaussian Process Regression for Transient Nonlinear Electro-Quasistatic Field Simulations
Dudu Zhang, Christian Richter, Fotios Kasolis, Markus Clemens
University of Wuppertal, Germany
For high-dimensional transient electro-quasistatic (EQS) field simulations, large sparse nonlinear systems of equations need to be solved iteratively at each time step using a solver, such as a preconditioned conjugate gradient method (PCG). If an improved start value is available, the solver can converge faster and thus, the simulation time can be reduced. In this paper, Gaussian process regression (GPR) is used in order to predict start values for next time steps using data that is collected from previous time steps. Numerical results show that GPR can efficiently predict solutions with good accuracy.
9:50am - 10:10am
Adaptive Stopping Criterion of PGD for Edge Elements based on Equilibrated Error Estimates in Magnetostatic Problems
Shuai Yan1,2, Zuqi Tang3, Thomas Henneron3, Zhuoxiang Ren1,4
1Institute of Electrical Engineering, Chinese Academy of Science, 100190, Beijing, China; 2Institute of Microelectronics, Chinese Academy of Science, 100029, Beijing, China; 3Univ. Lille, Arts et Metiers Paris Tech, Centrale Lille, HEI, EA 2697 - L2EP - Laboratoire d'Electrotechnique et d'Electronique de Puissance, F-59000 Lille, France; 4Sorbonne University, UR2, L2E, F-75005 Paris, France
The proper generalized decomposition (PGD) is an a priori model order reduction (MOR) method based on a variable-separated expression of the problem. Two iterative loops are needed in the PGD algorithm, namely the outer loop for enriching the reduction basis progressively, and the inner loop for solving each basis by fixed point iterations. Setting the stopping criterion of these two loops blindly can cause either the inaccuracy of the PGD or a waste of iterations. In this work, we implement a spacial variable-separated PGD with edge elements on a hexahedron mesh. The calculation of PGD modes are carried out simultaneously with a pair dual formulations on magnetostatics, so that an adaptive strategy is incorporated to provide appropriate tolerances for stopping the outer and inner loops of PGD.
10:10am - 10:30am
Time-Domain Analysis of Magnetically Shielded Wire Coils Using Homogenized Finite Element Method
Shogo Fujita1,2, Hajime Igarashi1
1Hokkaido University, Japan; 2Research Fellow of Japan Society for the Promotion of Science
A magnetically shielded wire (MSW) coil is effective for reduction of the eddy current loss due to proximity effect. When using the finite element method (FEM), it is still challenging to analyze electric apparatus including MSWs because its radius is much smaller than the overall machine size. In order to circumvent this problem, the MSW coil is modeled as a uniform material with a macroscopic complex permeability defined in frequency domain. In this work, the macroscopic permeability is represented as the Cauer-equivalent circuit for time-domain analysis. The power evaluated by the proposed approach with course FE discretization using recursively convolution scheme is shown to agree well with that obtained by conventional FEM with fine FE discretization.