Abstract: Based on the requirements about size, insulation structure, constant pitch and constant tension wrapping of CRAFT toroidal field (TF) coil, the D-shaped helix equation of single turn conductor was deduced.Aiming at the uniform motion of the wrapping machine along the conductor, based on the traditional three-phase PMSM vector control, an adaptive neuro-fuzzy inference controller (ANFIC) was designed to optimize the motion control of the wrapping machine. The simulation results showed that the three-phase PMSM by ANFIC had better speed and torque following ability and the three-phase current fluctuations were small, which provided the reference for the debugging and optimization of the CRAFT TF coil turn insulation wrapping system.

Key words: CRAFT, TF coil, Wrapping machine, Adaptive neuro-fuzzy inference

摘要： 根据 CRAFT TF 线圈的尺寸、绝缘结构、恒节距和恒张力包绕的要求，建立了单匝导体 D 形螺旋线方程。针对包绕机沿导体运动的匀速控制，在传统三相永磁同步电机(PMSM)矢量控制的基础上，设计了自适应神经模糊推理控制器(ANFIC)，并对包绕机的运动控制进行了优化仿真。结果表明，使用 ANFIC 的三相 PMSM 具有更好的转速和转矩的跟随性，同时三相电流波动较小，为 CRAFT TF 线圈匝间绝缘包绕系统的调试与优化提供参考。

关键词: CRAFT, TF 线圈, 包绕机, 自适应神经模糊推理