Search results

The purpose of this paper is to present a numerical approach for the computation of 3D magnetic fields in rotating electrical machines. The technique is suitable for the computation of flux densities and forces in the end windings of large synchronous turbo generators (TG).

The magnetostatic FEM model of the generator end windings is carried out for different displacements of the rotor axis to the stator magnetomotive force (MMF) axis. The method is based on a parallel integral formulation allowing to substantially reduce the computational effort.

The computational model requires only the discretization of magnetic materials and conductors and is fast enough for carrying out 3D analyses on a time scale fast enough for the needs of the designer. As far as the present application is concerned, the analysis of a synchronous generator in the class of 300‐400 MVA has shown that the most stressed elements of the armature conductors are those closer to the stator ends. The study demonstrates that the maximum stress component on the end windings is axial and is achieved when the MMF is aligned to the direct axis.

The present approach combining an efficient integral formulation, the sparsification of the relevant matrices and the parallel implementation of the related algorithms gives rise to an original computational tool that allows a more accurate description of the machine in comparison to other numerical simulations that can be found in the literature.

The purpose of this paper is to present the use of the compensation theorem (CT), well known in the analysis of linear electric networks, to compute sensitivity of the performance functions used in the robust design or tolerance analysis of electromagnetic devices.

The CT is first illustrated in the case of a simple field analysis problem. Then, using numerical simulations, the effectiveness of compensation approach for assessing impact of the small modification of material properties is shown. The numerical simulations are performed with a finite elements code based on an integral formulation.

The complexity of additional computations to assess the effect of small variations involved in sensitivity analysis can be reduced.

The method can be applied only to linear systems; in addition, although compensation applies to any variations, the reduction of computational complexity is achieved only for small variations, giving localized effects.

The method proposed in the paper can speed up the computations of sensitivity arrays in the robust design and tolerance analysis of electromagnetic device, when numerical methods are applied.

The use of CT in field computations is not new, but its adoption in the sensitivity computation is new to the best knowledge of authors.

The paper aims to illustrate a numerical technique to calculate fields and inductances of rotating electrical machines.

The technique is based on an integral formulation of the nonlinear magnetostatic model in terms of the unknown magnetization. The solution is obtained by means of a Picard-Banach iteration whose convergence can be theoretically proved.

The proposed method has been used to build a model of a large turbine generator. In particular, the influence of end effects on flux linkages has been computed. It has been demonstrated that the 2D solution underestimates the flux linkages as well as the no load voltage of 2 per cent, while the leakage fluxes are computed by the 2D solution with errors as high as 20 per cent.

The method is advantageous in comparison to standard methods.

The paper aims to apply an innovative inversion method to the problem of imaging (location, direction and size) of concrete rebars by means of eddy current measurements.

An accurate numerical model of the probe‐rebar interaction, including eddy currents and skin effect, is considered. The inverse problem is approached with a very efficient inversion procedure previously introduced in a different context.

A critical analysis of the issues to be considered for the quantitative imaging of rebars is given, and the possibility of relevant simplifications in the numerical model outlined, allowing the development of an accurate and computationally efficient method.

The proposed formulation is applied for the first time to the problem of rebars imaging. Experimental tests have been carried out to validate the numerical model and its underlying hypothesis.