Nerual Observer를 이용한 PMSM의 정밀 속도 제어

This paper presents neural observer that used to deadbeat load torque observer Most practical systems are nonlinear, and It is general practice to use linear models to simplify their analysis and design However, the locally linearized model is invalid for a large signal change The neural observer is suggested to increase the performance of the load torque observer and main controller. The output error and estimeted state is trianed by neural network of neural observer. As a result, the state estimation error is minimised and deadbeat load torque observer make use of corrected esimation state. To reduce of the noise effect of deadbeat load torque observer, the post-filter which is implemented by MA process, is adopted As a result, the proposed control system becomes a robust and precise system against the load torque. A stability and usefulness, through the verified computer simulation, are shown in this paper