Application of ANN to numerical field calculation methods for an accurate simulationas well as field optimization of electrode systems

摘要

The paper introduces a new method foraccurate field calculation as well as electrode fieldoptimization. The method is based on combining theArtificial Neural Network (ANN) with the numericalfield method under use. Data of well-chosen orgenerated equipotentials and their resulting electricfield intensities are used to train ANN. The trainedANN is then used to determine the electric field forany predetermined electrode configuration in thespecified region with acceptable accuracy. Moreoverthe trained ANN can be used for field optimization byevaluating the desired, but originally unknown,equipotential surfaces with constant electric field inthat region.