적응 알고리즘과 내부 모델 제어를 이용한 주기적 외란 억제 기술에 관한 연구

Abstract

Periodic disturbance involves various engineering applications in input parts of control systems. Adaptive disturbance cancellation is a generic approach to reject such disturbances by using control engineering techniques. In recent years, complicated dynamic systems have been constructed in industrial fields and correspondingly advanced disturbance cancellation strategies are presented for enhancing control performances. This dissertation presents a novel adaptive internal model control method for effective depression of disturbance due to dynamic vibration against mechanical systems. The proposed control system is composed of a generic IMC configuration and adaptive feed-forward control mechanism. The former includes a dynamic system model for mathematical representation of a controlled plant and a forward control law configured with its inverse model and control parameter. And an AFC system is constructed with sine and cosine functions and its parameter vector is optimally determined from an adaptive estimation approach. This control mechanism is effective to mitigate vibrated disturbance excitation with unknown frequency and amplitudes particularly. We accomplish computer simulation study for demonstrating reliability of the proposed control approach with numerical examples. The proposed control algorithm is coded by a well-known Matlab software and various simulation scenarios are adopted to test its robustness under changeable system environments in terms of external vibrated disturbances with different frequencies and amplitudes. Finally, we carried out a real-time experiment to prove practicability of the proposed control system by utilizing a test-bed system. This test-bed includes a mechanical vibrator to generate a physical disturbance excitation to a controlled plant, and a magneto-rheological (MR) damper utilized as an mechanical actuator to reduce vibration disturbance.