【摘要】 基于考虑工件变形的薄壁零件铣削加工变形模型，分析加工变形影响参数；依据正交切削试验加工数据，采用遗传算法改进的BP神经网络建立基于主轴切削转矩相对变化值的薄壁零件加工变形预测模型；通过加工过程数据与加工变形量的实时映射，基于镜像补偿原理重构零件三维模型，实现薄壁零件加工变形补偿。对预测及补偿后的三维零件进行对比实验，结果表明：在零件刚度最低处，相对于未补偿的涡旋盘零件，补偿后的零件平均变形量减小42.9%,最大变形量减小36.2%,且加工变形量均匀度较高。更多还原

【Abstract】 Based on the milling deformation model of thin-walled parts considering workpiece deformation, this paper analyzes the influencing parameters of machining deformation, adopts genetic algorithm-improved BP neural network to esablish a deformation prediction model for thin-walled parts machining based on the relative variation value of spindle cutting torque, according to orthogonal cutting experimental machining data, and by real-time mapping of machining process data and machining deformation, compensates for thin-walled parts machining deformation and reconstructing 3D model of parts based on the principle of mirror compensation. Through comparative experiments on the predicted and compensated 3D parts, the results showed that at the lowest stiffness of the parts, compared to the uncompensated vortex plate parts, the average deformation of the compensated parts decreased by 42.9%, the maximum deformation decreased by 36.2%, and the uniformity of the machining deformation was relatively high.更多还原