关键词: 数字孪生模型, 惯量摄动, 状态观测器, 参数同步方法, 复合控制

Abstract: With the advantages of compact structure, small loss of momentum and no gap transmission, direct drive components are widely used in high-end CNC machine tools, aerospace precision servo mechanisms and other major equipment. The dynamic response characteristics and robust performance of direct drive components are important indexes of the servo system. However, the perturbation of load inertia and changes in assembly characteristics will cause the system's servo performance to decrease or even lose stability. In order to improve the system's servo performance and motion safety, A composite control method based on digital twin model is proposed, which uses physical entity data to drive virtual model parameter synchronization, and uses virtual model to optimize physical entity servo parameters. Firstly, the key characteristic parameters of the direct drive system are extracted, the linear link is described by the lumped parameter model, the nonlinear feature is described by the distributed parameter model. Then, the virtual physical model of the direct drive component is established. The parameter updating method of the digital twin model based on the recursive augmented least squares algorithm is proposed to realize the adaptive synchronization of the model features. The control parameters of the system are optimized by using synchronous parameters. The design method of the speed controller under index constraints and the extended state Kalman filter is proposed to form the compound control strategy. Experiments show that the method proposed has better resistance to inertia perturbation, and the virtual model parameters can identify the inertia change of the direct drive system accurately and quickly; the root mean square of the velocity residual following the 20°/s &1Hz sinusoidal reference signal is reduced from 5.88°/s to 1.76°/s, which is optimized by 70%; the adjustment time of the step response process always meets the design index, the Pearson correlation coefficient of the actual response curve and the theoretical response curve has been increased from 0.957 to 0.993. The research result is to further refine the digital twin model and promote the application of digital twin model in motion control.

Key words: digital twin, inertia perturbation, state observer, parameter synchronization method, compound control