Abstract: To solve the low accuracy of simulation results due to the single model form of the CNC machine tool feed system built by traditional modeling methods and the lack of online real-time interaction with the physical space,an adaptive updating method of Digital Twin (DT) model for feed system based on hybrid drive of mechanism model and data-driven model was proposed.With the characteristics of multi-domain coupling of the feed system,the mechanical and electrical integration DT mechanism model was established by employing the physical modeling tool Simulink/Simscape.At the same time,based on the Long Short-Term Memory (LSTM) neural network,the DT data drive model was obtained by using a large amount of historical data training,and then the DT model of the feed system was obtained by combing the two models.On this basis,the recursive least squares algorithm with forgetting factor was used to update the model and improve the adaptability of the model.The effectiveness of the proposed method was verified by experiments on the virtual and real synchronous motion of the biaxial feed system.

Key words: digital twin, feed system, adaptive update, visual simulation, virtual reality interaction

摘要： 为了解决传统建模方法搭建的数控机床进给系统模型形式单一且缺乏与物理空间的在线实时交互,导致仿真结果精度低的问题,提出一种基于机理模型和数据驱动模型混合驱动的进给系统数字孪生模型自适应更新方法。该方法针对进给系统多领域耦合的特点,利用物理建模工具Simulink／Simscape建立了机电一体化数字孪生机理模型,同时基于长短记忆神经网络建立了数字孪生数据驱动模型,二者混合构成进给系统数字孪生模型。在此基础上,采用带有遗忘因子的递推最小二乘算法更新模型,提高了模型的适应性。通过双轴进给系统的虚实同步运动实验验证了所提方法的有效性。

关键词: 数字孪生, 进给系统, 自适应性更新, 可视化仿真, 虚实交互