HTS coil design using artificial neural network and fuzzy interference system

Abstract

The design of High Temperature Superconducting (HTS) RF coil highly relies on the computer simulation because HTS materials are very expensive and the coil fabrication requires high accuracy. Normally, the simulation for the HTS coil design is time consuming and not straightforward. In this paper, two novel approaches for HTS coil design, the electromagnetically (EM) trained artificial neural networks (EM-ANN) and the electromagnetically trained fuzzy inference systems (EM-FIS) are presented. These two models can simplify the normal simulation and speed up by millions of times. Therefore, the difficult tuning procedure of HTS coil can be easily simulated before its fabrication.