Search

In order to reduce the position overshoot of the permanent magnet synchronous linear motor (PMSLM) vector control system and improve the system response, on the basis of given mathematical model and mechanical motion equations of PMSLM in the d-q rotating coordinate system, controlled auto regressive integrated moving average (CARIMA) model have been introduced. A new generalized predictive controller of PMSLM system have been designed based on this model and Sun Jun’s proposed generalized predictive control algorithm that takes full advantage of predictive information. Simulation results show that this controller has the advantages of small overshoot and excellent tracking performance.

Permanent magnet synchronous linear motor (PMLSM) has excellent performance and wide application, but its special structure makes it sensitive to external disturbance. This paper aims at the disturbance problem of X-Y motion platform driven by PMLSM under actual operating conditions. In the cross-coupling control strategy, the input command is compensated by the coupling control quantity between axes, and the linear active disturbance rejection controller (LADRC) is applied to the position loop in the three-loop control strategy to suppress the external disturbance and reduce the influences caused when the coupling control quantity is compensated to the uniaxial servo system. Finally, it is verified by Simulink. The simulation results show that the LADRC-based disturbance compensation method has better dynamic performance than the traditional cross-coupling control, and the contour control accuracy is higher.

There are some problems accompanied by detecting rotor parameters of the traditional method on the permanent magnet linear synchronous motor (PMLSM), such as inaccurate detection and poor anti-interference ability. Based on this, this paper proposes a speed closed-loop control strategy based on PID neural network (PIDNN) without speed sensor. The algorithm obtains the error between real current and the given current of the q-axis by calculating the observation angle and the actual angle. The algorithm obtains the error between the given current and the feedback value of the q-axis, and uses the gradient descent algorithm to update the parameters and to obtain the fitted electrical angle. When the current error converges to zero, the output is the velocity of the mover. The simulation and experimental results show that the algorithm can track the speed very well, which reduces the cost and improves the reliability of the system.