Research On The High-Precision Control Of Permanent Magnet Linear Servo System

In modern machining industry, laser-cutting, high-speed milling and scanning machines require fast and accurate linear motions. But traditional drives need to use rotational motors and lead screw or toothed belts to obtain linear motions, which leads to some mechanical problems such as backlash, large frictional, inertial loads and structural flexibilities. With the development of linear motors, recently, they are becoming increasingly popular in such applications. Due to the direct-drive technology, linear motor exhibits the property of high accuracy, high speed and excellent dynamic responsivity, which just satisfies the requirement of high speed machining (HSM). In order to explore the linear motor applications in machining tools, here, we designed a linear servo system via selecting a permanent magnet synchronous linear motor (PMSLM) and a linear encoder. Our aim is to develop some algorithms to realize high accuracy and high performance control for this linear servo system.First, we applied field-oriented principle and the strategy of i_d = 0 to the state space model of PMSLM in d-q coordinates, and obtained a decoupling model and an approximate linearization model in frequency domain of the motor and driver system. Based on these models, we designed a robust PID controller of speed-loop by means of H_∞closed-loop gain shaping techniques. Moreover, a speed feed-forward block could be obtained easily from the frequency domain model of PMSLM and its driver. The robust PID controller exhibits good robustness against external disturbances and internal parameter perturbation by simulation results. When the feed-forward block is added, the responsive speed and control accuracy of the system will be improved more. Moreover, a pseudo derivative feed-forward (PDFF) method is analyzed in this thesis. As a type of PI control with feed-forward, PDFF can be designed independently from the plant model which is very suitable for engineering applications. Due to the pseudo derivative item of speed feedback, the dynamic performance and noise control of PDFF is improved greatly.In order to realize the high accurate position control, firstly, we made a speed plan through which a speed curve can be obtained from a displacement order. According to speed loop control, we design a P/PI position and speed control system with speed and acceleration double feed-forward path. The feed-forward paths provide the main control signals, speed order signal to the input terminal of speed loop and acceleration signal to the input terminal of current loop, and the main path controls the error of position and speed. Simulation results show that the double closed-loop system has strong robustness and excellent control performance with the dynamic error less than 10μm and static error near to zero. There are five major control parameters in this system, here, these parameters are optimized with genetic algorithm and the satisfied results are achieved. Furthermore, the method of PDFF is investigated, but its accuracy is less than that of above method.This thesis has completed the main design work of the linear servo system from the selections and installations of equipments to the design of control algorithms and the optimization of their parameters. After these work, we will continue to apply these algorithm to the experiment system and realize the design target of system.