Research On Iterative Learning Control Of Permanent Magnet Linear Servo System

This thesis is supported by National Natural Science Foundation of China(51075281), Doctoral Program of Higher Education of Special Fund (20112102110002)and Liaoning Province University Innovation Team (LT2010081). This thesis consideredthe design of a cascaded open-closed loop iterative learning controller (ILC) for arepetitive position servo system of permanent magnet linear synchronous motor(PMLSM) to meet control requirement of high precision linear servo system. The ILCwas presented for optimizing the position input signal to compensate for cogging effect,end effect, friction effect, thrust ripple etc, which are periodic or repetitive disturbances,and to reduce tracking error caused by servo lag without modification to the existingtracking controller.First of all, by referencing a large number of paper, this thesis analyzed linear servosystem and the developing trend of control strategy, some kinds of disturbances whicheffect dynamic performances of linear servo system were introduced. The importance oflinear servo system in modern society and the significance of ILC to PMLSM wereexplained.Secondly, the effects of the optimal input signal with open-loop ILC andclosed-loop ILC were researched. Not taking advantage of the present cycle output error,the iterative process only referencing to previous cycle motion information and theability of anti-jamming was poor and the iteration speed was slow. ILC of closed-loopstructure took advantage of the current moving information but abandoned the previousmoving information. In order to achieve high accuracy requirements, it required largefeedback gain. Control strategies of two structures both had some drawbacks.Finally, in order to improve the ILC position tracking performance and systemrobustness, position feedback controller was introduced to optimize the input signal onthe basis of the open-loop ILC. That scheme was the open-closed loop ILC that wasproposed in this thesis. The introduced feedback controller increased system stability andanti-interference ability, and accelerated the learning convergence speed. The system motion information of the previous and the current cycle were both utilized at the sametime. And the performance of the proposed ILC is superior to open-loop ILC andclosed-loop ILC. Simulation results verify that open-loop, closed-loop and open-closedloop ILC schemes can improve the accuracy of position tracking PMLSM servo systemeffectively and the open-closed loop scheme has better robustness and iteration speed.