Search results

The purpose of this study is to demonstrate the novel approach in treating multiply connected problems in magnetostatic.

The new double layer approach (DLA) to be proposed is based on the use of the exciting double layer on the cut-surface. Applying Ampere’s circuital law to the circuital path along a toroidal core of M–C model, this paper derives unified exciting potential (UEP) from the common exciting potential. The UEP is applicable to the simply or M–C analysis. To check the effectiveness of the UEP, this paper analyze typical M–C problems and compares the results with those of other benchmark problems and also those obtained by surface charge method (SCM). Because the SCM encounters a cancellation error, this paper overcomes this problem by using the concept of direct boundary element method (BEM).

Using the improved DLA, this paper analyzed a typical multiply connected model and compared the results with those of the SCM, which has been improved to overcome cancellation errors. This paper has confirmed that the results obtained by the improved DLA are the same as those obtained by the improved SCM and Steklov–Poincaré operator formulation, tested at the well-known benchmark problems given in Andjelic et al. (2010). From these results, this paper concluded that the Improved DLA works well and that the improved SCM becomes available for analyzing both the simply and multiply connected problems.

Expanding a concept of the exciting double layer on the cut-surface, this paper improve the DLA to analyze the M–C problems. Applying Ampere’s circuital law to the full circuital path along the toroidal core of M–C problem, this paper derive UEP from the original exciting potential to get the governing BIE. The BIE is applicable to either simply or multiply connected analysis.

The purpose of this paper is to supply a numerical analysis tool to solve eddy currents induced in nonlinear materials such as steel by boundary element method (BEM), and then apply it to design and analysis of power devices.

Utilizing integral formulas derived on the basis of rapid attenuation of the electromagnetic fields, the paper formulates eddy currents in steel. In the formulation, nonlinear terms are regarded as virtual sources, which are improved iteratively with the electromagnetic fields on the surface. The periodic electromagnetic fields are expanded in Fourier series and each harmonic is analyzed by BEM. The surface and internal electromagnetic fields are obtained numerically one after the other until convergence by the Newton‐Raphson method.

It is confirmed that this approach gives accurate solutions with meshes much larger than the skin depth and therefore is adequate to apply to a large‐scale application.

The eddy current is formulated by utilizing the impedance boundary condition in order to meet a large‐scale application, and so solutions near the edge are poor. In the case of better solutions being required, some modifications are necessary.

To lessen computer memory consumption, the parallel component of the currents to the steel surface is analyzed as a 2D problem and the normal component is obtained from the parallel component. One 2D equation for one analyzing region is discretized by dividing the region into layers adaptively and then solved. Next, another is solved sequentially. This method gives a compatible numerical analysis tool with finite element method.

The purpose of this paper is to solve generic magnetostatic problems by BEM, by studying how to use a boundary integral equation (BIE) with the double layer charge as unknown derived from the scalar potential.

Since the double layer charge produces only the potential gap without disturbing the normal magnetic flux density, the field is accurately formulated even by one BIE with one unknown. Once the double layer charge is determined, Biot‐Savart's law gives easily the magnetic flux density.

The BIE using double layer charge is capable of treating robustly geometrical singularities at edges and corners. It is also capable of solving the problems with extremely high magnetic permeability.

The proposed BIE contains only the double layer charge while the conventional equations derived from the scalar potential contain the single and double layer charges as unknowns. In the multiply connected problems, the excitation potential in the material is derived from the magnetomotive force to represent the circulating fields due to multiply connected exciting currents.