Study On Iterative Learning Control Of Ring Permanent Magnet Torque Motor

Taking the project supported by Science and Technology Council of Liaoning Province and Education Council of Liaoning Province as background, this thesis researches on the Ring Permanent Magnet Torque Motor servo system used in NC Rotary Table. For the characteristic of the ring permanent magnet torque motor used in the direct drive NC rotary table servo system and the requirements of stable low-speed and high precision work at low-speed for the NC rotary table, it is very difficult to obtain the requirements using traditional control scheme, therefore in the thesis the proposed scheme used in the the Ring Permanent Magnet Torque motor servo system is to combine iterative learning control, fuzzy control and sliding model control to improve the servo performance of system at low speeds.Firstly, the Ring Permanent Magnet Torque Motor mathematical model is established, and then the torque ripple of the Ring Permanent Magnet Torque Motor is analyzed.Secondly, briefly describes the principle of iterative learning control (ILC). In view of the periodic nature of torque ripples, design an ILC controller to minimize torque ripples, which records the torque error and the compensating reference current of last cycle to update the compensating reference current of this cycle. Considering the learning gain is the key factor which impacts the speed of the ILC controller, a fuzzy controller is designed to tune the learning gain. Combine it with the ILC controller to form the fuzzy iterative learning controller (F-ILC). In contrast with the traditional ILC controller, the convergence rate of F-ILC controller is faster and steady-state torque error is slighter.Finally, because the sliding mode control is insensitive to disturbance and inertia variations, adopt it to design the position controller for the system. However, the robustness and invariance properties can not be guaranteed as the continuous time case. Select the appropriate reaching law is conductive to reduce chattering and improve the system response, so a discrete-time iterative learning sliding mode controller is designed. It tunes the parameters of the reaching law of the sliding mode to improve the performance of discrete sliding mode control system. With the proposed algorithm, the resulting system servo performance can be improved and the chatting effect can be suppressed.