Conference Agenda

A shape sensitivity analysis is presented for the optimal design of a time-harmonic electromagneto-quasistatic (EMQS) system. The design variable is the interface boundary, and its equation of motion is described using the concept of material derivatives. The shape sensitivity formula is derived from a continuous functional, which represents an augmented objective function that characterizes the EMQS system and its performance through state variables, rather than from discretized differential-algebraic equations. The derived formula is obtained analytically as a surface integral, composed of state and adjoint variables, along with design velocity on the boundary. The deformation of the design variable is computed by combining the design velocity with the level set method. Two numerical examples are presented to demonstrate the feasibility of the proposed sensitivity analysis.

This paper proposes a fast equivalent-circuit-based analysis method for PCB-based magnetic components with high-aspect-ratio coils. Circuit parameters are derived using quasi-static and homogenized FE analysis, significantly reducing computation time compared to full-wave FE analysis. The method effectively models dielectric losses by incorporating parasitic resistance into the equivalent circuit, achieving high accuracy in frequency-dependent resistance simulations.

We present a method for modeling arbitrarily shaped anisotropic magnetoelectric objects immersed in a homogeneous isotropic medium and exposed to an arbitrary electric field. The method requires the discretization of only boundary layers and solves the problem directly, without transforming it into an isotropic one. We investigate anisotropic magnetoelectric materials of the Tellegen type, characterized by nine parameters for each of the permittivity, permeability, and coupling matrices. Results are compared against an analytical solution for the case of a magnetoelectric anisotropic sphere placed in air and exposed to a uniform electric field. We achieve a mean relative error within the sphere below 1.4% and a normalized root mean square error outside the sphere below 0.1%. With a slight modification, the method can be applied to magnetoelectric materials exposed to a magnetic or combined electric and magnetic fields.

In this paper, we investigate the parallel Block Jacobi ICCG (BJ-ICCG) solver. It is well-known that the BJ-ICCG method suffers from the convergence degradation as the number of threads increases. We analyze this phenomenon by computing the eigenvalue distribution of the preconditioned coefficient matrix. Because the calculation of the eigenvalues requires significant computational effort, it has been rarely performed for three-dimensional eddy current analyses with large problem sizes. In our analysis, we observed that many eigenvalues are shifted in an unfavorable direction, resulting in the decline in convergence as the number of

threads increases.

This paper proposes a derivation of a physically interpretable RC Cauer Ladder Network (CLN) for heat transfer analysis. Originally,CLN method was formulated for eddy current problems, represented as a voltage-driven RL-CLN. Although a generalized CLN formulation for parabolic partial differential equations was later established, it was also based on RL-CLN and lacked a clear physical interpretation. This paper revisits the CLN formulation for thermal circuits, deriving a current-driven RC-CLN representation that offers a more explicit physical meaning than a simple transfer function based CLN. Additionally, we demonstrate that an analytical expression for the RC-CLN can be derived for simple two-dimensional thermal problems, and verify its accuracy by reproducing results obtained through FEM analysis. By utilizing this thermal RC-CLN, we present its effectiveness in achieving efficient computation through reduced calculation time, particularly in cases with complex pulsed input heat waveforms.

This study investigates model order reduction of a cage induction motor (IM) based on the multiport Cauer ladder network (CLN) method considering nonlinear magnetic properties. First, the permeability distributions in the stator and rotor are obtained using nonlinear time-harmonic eddy-current analysis as a function of the primary current and the slip. Then, the circuit parameters of the parametrized multiport Cauer circuits for the stator and rotor are derived under the conditions of the calculated permeability distributions as the nonlinear model order reduction of the cage IM. The effectiveness of the proposed method is validated by comparing the numerical results obtained by the CLN method and a finite-element method.

This study proposes a novel AC copper loss modeling for a FEM-parametrized plant model of a permanent magnet synchronous motor (PMSM). In the proposed method, an equivalent circuit for the PMSM is combined with Cauer circuit derived from the Cauer ladder network (CLN) method to take account of frequency-dependent AC resistance due to armature windings composed of parallel round wires or flat wires. The effectiveness of the proposed plant model is validated though the loss analysis of PMSMs with full-pitch and short-pitch armature windings composed of parallel round wires.

This paper presents a methodology based on Field Reconstruction Method (FRM) and Active Harmonic Elimination (AHE) techniques enabling development of an advanced low-speed surface-mounted Permanent Magnet Synchronous Motor (PMSM) drive scheme. The FRM enforces periodicity between rotor and stator fields, significantly enhancing the precision of Finite Element Analysis (FEA) by accurately representing space harmonics and changes in local saturation effects due to loading variations. This way harmonic periodicity is ensured while torque ripple and motor losses are minimized, resulting in optimization of the AHE strategy and significant enhancement of the computational efficiency. Experimental validation on a prototype system has confirmed the proposed method effectiveness and demonstrated its superior characteristics in terms of motor performance and harmonic mitigation. This methodology is particularly beneficial for electric vehicle applications, where high efficiency and dynamic control are essential under various loading conditions.

The digest proposes a gradient-based topology optimization method to design chip inductors. The density method is used to ensure that the material distribution is continuous and differentiable, and the adjoint variable method is employed to calculate efficiently the

gradient of the objective function, combined with the finite element method. It has been verified that this method can efficiently design inductors that meet the target inductance value while minimizing Joule losses.

The evaluation of eddy current losses in windings is a significant task for designing efficient electric machines . Previous research has highlighted challenges with iterative methods for solving linear systems in eddy current problems with current and voltage excitations , primarily due to poor convergence characteristics. This study investigated a design of preconditioners to improve the convergence of iterative methods for this problem.

This paper proposes a method for time-domain analysis using the Darwin model. In this approach, the second time derivative term is removed using the redundant variables to apply the standard backward Euler method. We analyze a parallel plate capacitor model and compare the results with the frequency-domain analysis, thereby demonstrating the validity of the proposed method.

The liquid discharge was examined through the finite element method with a lossy dielectric material immersed in insulating oil considering both the dielectric permittivity and conductivity. Until now, the liquid discharge analysis with the lossy dielectric material has rarely been reported with the 3-carrier system consisting of positive ion, negative ion, and electron. To verify our numerical setup, one carrier analysis model, first, was constructed to test the space and surface charges with the lossy dielectric barrier. The charge propagation was validated by comparing it with theoretical results. Based on this model, we extended it to 3-carrier model for analyzing the effects of lossy dielectric material in a streamer channel. This study provides insights into streamer propagation characteristics and charge accumulation behavior as a function of dielectric permittivity and conductivity, offering a foundation for further analysis of discharge phenomena in insulating oil.