高速磁悬浮片梭驱动装置参数优化设计

Parametric Optimization Design for High-speed Maglev Projectile Propulsion Device

中文关键词: 夹持装置  电磁线圈  深搜迭代  Maxwell 仿真  BP 神经网络

英文关键词: Clamping mechanism  Electromagnetic coil  Depth-first search iteration  Maxwell simulation  BP neural network

中文摘要:

      聚焦于高速磁悬浮片梭引纬驱动装置及电磁驱动线圈设计相关问题展开探讨。首先介绍了新型磁悬浮片梭结构，与传统扭轴式驱动相比，磁悬浮片梭的电磁驱动能实现入纬速度为 40 m/s，提高引纬率。基于磁悬浮片梭的引纬过程，设计夹持机构、电磁线圈等装置的协同控制方法。同时，基于深搜迭代原理计算电磁线圈驱动片梭时的变化电磁力、加速度、速度，获取片梭引纬位移，实现高速非接触式近零损耗驱动。建立 BP 神经网络设计驱动线圈匝数、内径、外径、线径并应用序列二次规划 (SQP) 算法进行参数优化。结果表明：当电磁线圈匝数为 1500、内径为 28 mm、外径为 70 mm、线径为 1 mm 时，基于神经网络构建的电磁驱动装置能准确预测电磁力，电磁驱动效率最高，能实现 40.99 m/s 的片梭引纬。

英文摘要:

      This paper focuses on the design challenges of high-speed maglev projectile weft insertion clamping mechanisms and electromagnetic drive coils. A novel magnetically levitated projectile structure is introduced. Compared to conventional torsion-bar drives, its electromagnetic propulsion achieves a weft insertion speed of 40 m/s, significantly enhancing the insertion rate. Subsequently, a systematic control methodology is presented to harmonize the clamping mechanism and electromagnetic coils during projectile weft insertion. A depth-first search iterative algorithm is proposed to enable real-time calculation and monitoring of electromagnetic force, acceleration, velocity, and displacement during coil excitation, thereby facilitating dynamic position parameter acquisition for the projectile. Furthermore, a backpropagation neural network was employed to design coil parameters (number of turns N, inner diameter D₂, outer diameter D₁, wire diameter d), with subsequent optimization via the Sequential Quadratic Programming (SQP) algorithm. Optimal performance was achieved at N=1500, D₂=28 mm, D₁=70 mm, and d=1 mm. Under this configuration, the neural network-based electromagnetic drive model accurately predicted magnetic forces while maximizing drive efficiency, enabling projectile acceleration to 40.99 m/s during weft insertion.