Neural mesh iteration for the FE computation of charged-particle trajectories in vacuum electron devices

Author

Coco, Salvatore

Author_Institution

Dipt. Elettrico, Elettronico e Sistemistico, Catania Univ., Italy

Volume

38

Issue

2

fYear

2002

fDate

3/1/2002 12:00:00 AM

Firstpage

1193

Lastpage

1196

Abstract

In this paper, a neural mesh-adaptation iterative procedure is presented for the finite-element solution of coupled steady-state electromechanical problems describing the motion of charged particles inside vacuum electron devices. In the procedure at each step a neural algorithm performs a mesh refinement in which the mesh "density" is related to the current estimate of the three-dimensional (3-D) spatial charge density distribution. Convergence to an accurate solution is obtained with a reduced computational effort in comparison with conventional fixed mesh iterative techniques. An example of application to the 3-D calculation of electron trajectories inside the collector of a travelling wave tube is shown in order to illustrate the advantages of the iterative scheme

Keywords

convergence of numerical methods; electrodynamics; finite element analysis; iterative methods; mesh generation; neural nets; space charge; travelling wave tubes; vacuum tubes; charged particle trajectory; convergence; coupled steady-state electromechanical problem; finite element method; iterative technique; mesh adaptation; neural network algorithm; three-dimensional spatial charge density distribution; travelling wave tube; vacuum electron device; Electrodes; Electron beams; Electron devices; Equations; Finite element methods; Iron; Iterative algorithms; Space charge; Steady-state; Voltage;

fLanguage

English

Journal_Title

Magnetics, IEEE Transactions on

Publisher

ieee

ISSN

0018-9464

Type

jour

DOI

10.1109/20.996305

Filename

996305