摘要
Abstract
An AC permanent-magnet synchronous linear motor(PMLSM)offers high response speed and control accuracy and is widely used in high-precision manufacturing applications.However,due to its unique physical structure,PMLSM is inevitably affected by thrust fluctuations,including detent and friction forces during movement,which deteriorate the motor's control performance.To address thrust fluctuation problems in high-precision PMLSM,a composite compensation strategy is designed that combines disturbance model feedforward(DMF)and a linear extended state observer(LESO).
Firstly,based on the physical structure and magnetic field distribution of the PMLSM,the force of the motor mover during motion,along with its position and frequency characteristics of thrust fluctuation,are analyzed.It is found that a high-frequency component of thrust fluctuation occurs at high speed and high dynamic response,leading to an increase in servo error.Therefore,a thrust-fluctuation composite compensation strategy is proposed.Secondly,a neural network is used to model the detent force in the position domain,a piecewise linear friction model is used to model the friction force in the velocity domain,and the residual disturbance after modeling is further estimated and compensated by LESO.In addition,to enhance the tracking performance of the instructions,a two-degree-of-freedom control structure combining differential feedforward and PI control is adopted.Considering the real-time requirement of the servo system,the computational amount of the proposed thrust fluctuation composite compensation strategy is also analyzed.
Finally,the effectiveness of the proposed strategy is evaluated through position and velocity-servo experiments on a PMLSM prototype platform.Trapezoidal-wave-position servo experiments verify the proposed strategy for thrust fluctuation suppression.Additionally,experiments conducted at different speeds confirm that the proposed strategy suppresses thrust fluctuations across the full frequency range.Furthermore,sinusoidal position experiments show that the proposed control strategy provides adequate thrust fluctuation compensation in both steady-state and dynamic conditions,improving the servo performance of the PMLSM system and meeting the real-time computing requirement.The suppression rate of thrust fluctuations reaches 87.39%,with a steady-state position control accuracy of 2.72 μm and a speed control accuracy of 4.46 mm/min at 4 m/min.In addition,the strategy is robust to the mass change in the motor mover.关键词
永磁同步直线电机/推力波动/神经网络建模/扩张状态观测器Key words
Permanent magnet linear synchronous motor/thrust fluctuation/neural network modeling/extended state observer分类
信息技术与安全科学
