【摘要】 游动芯头拉拔是制备高质量GH4169高温合金毛细管的关键工艺。本文以外径D 6.0 mm×壁厚t 0.2 mm的GH4169毛细管为研究对象，通过建立宏观有限元模型，研究了模具锥角、道次延伸系数、定径段长度、道次减径量和道次减壁量之比、摩擦系数及管坯尺寸因子等工艺参数对拉拔力与残余应力的影响规律，并采用人工神经网络与遗传算法相结合的方法，对多道次拉拔工艺参数进行优化设计。研究结果表明：模具半锥角为12°、道次延伸系数为1.26～1.29、芯头定径段长度为3.5 mm、模具定径段长度为1 mm、减径/减壁比为35时，拉拔力最小且残余应力分布均匀；通过神经网络模型预测拉拔力误差低于0.1%；结合遗传算法，确定了将初始管坯拉拔至目标尺寸（D 4.0 mm×t 0.08 mm）所需的最少拉拔道次为6个道次，并获得了各道次的工艺参数。更多还原

【Abstract】 Floating plug drawing is a critical process for manufacturing high-quality GH4169 superalloy capillary tubes. This study investigates GH4169 capillary tubes with an initial specification of 6.0 mm in outer diameter and 0.2 mm in wall thickness. A macroscopic finite element model was developed to systematically analyze the effects of key process parameters including die cone angle, pass extension coefficient, sizing section length, pass diameterreduction/wall-reduction ratio, friction coefficient, and tube blank size factor on drawing force and residual stress. Subsequently, an integrated approach combining an artificial neural network(ANN) and a genetic algorithm(GA) was adopted to optimize the multi-pass drawing process parameters. The results demonstrate that the drawing force is minimized and the residual stress is uniformly distributed when the die semi-cone angle is 12°, the pass extension coefficient ranges from 1.26 to 1.29, the plug sizing section length is 3.5 mm, the die sizing section length is 1 mm, and the diameter-reduction/wall-reduction ratio is 35. The ANN model exhibits high prediction accuracy for drawing force, with a relative error of less than 0.1%. Combining with the genetic algorithm, the minimum number of drawing passes required to reduce the initial tube blank to the target dimensions(D 4.0 mm × t 0.08 mm) was determined to be six, and the process parameters for each pass were obtained.更多还原