【摘要】 针对主轴系统固定结合部传统弹簧阻尼单元法等效建模精度较低的问题，采用虚拟材料层法建立主轴系统固定结合部的孪生物理模型，对模型进行有限元分析计算，以虚拟材料层相关参数(弹性模量、密度、泊松比)作为输入，计算机床主轴的固有频率，并以此数据样本训练深度神经网络，建立主轴系统的孪生有限元模型。采用粒子群优化算法，以虚拟材料层相关参数为优化变量，以孪生有限元模型计算的理论固有频率与对应试验值的相对误差最小为目标函数，优化确定虚拟材料层的相关参数。以VMC850E型立式加工中心主轴的刀柄-夹头-刀具系统两个固定结合部为实例进行了建模、试验、参数识别等分析，分析结果表明，该方法是可行的、有效的，建模精度达到了1%以内。更多还原

【Abstract】 Here, aiming at lower accuracy of traditional spring damping element method for equivalent modeling of fixed joint of a spindle system, the virtual material layer method was used to establish the twin physical model of a fixed joint of a spindle system. The model was calculated and analyzed with finite element method. Relevant parameters of virtual material layer including elastic modulus, density and Poisson’s ratio were taken as inputs, natural frequencies of machine spindle were calculated, and a deep neural network was trained with the calculated data sample to establish the twin finite element model of the spindle system. The particle swarm optimization algorithm was used to optimizeand and determine relevant parameters of virtual material layer by taking relevant parameters of virtual material layer as variables to be optimized, and taking the minimum relative error between theoretical natural frequencies calculated using the twin finite element model and the corresponding test values as the objective function.Two fixed joints of tool handle-collet-tool system of spindle of VMC850E vertical machining center were taken as actual examples to perform modeling, tests and parameter identification, the analysis results showed that the proposed method is feasible and effective; the modeling error is less than 1%.更多还原