신경회로망을 이용한 교류서보전동기 제어에 관한 연구

Abstract

In this paper, neural networks is used to control AC servo motor and is finding widespread applications in different areas of science and engineering problems. AC servo systems are competed with DC servo systems for motion control. Because of their favorable electrical, mechanical properties, good dynamics and a high efficiency, AC servo motor has been covered most servo-applications in robotics, machine tools and positioning devices. Permanent magnet synchronous motor is receiving attention for servo drive applications in recent years because of its high torque to inertia ratio, superior power density and high efficiency. Recent developments of high energy materials such as neodymium have focussed attention on wide variety of speed, position control, applications. In order to reduce with mechanical parameter variation, and to obtain a good dynamic performance at the viewpoint wide speed range, an intelligent controller is employed. The neural networks is adopted widely, however the convergence speed is limited by its constant learning rate. The neural networks based self tuner of coefficient parameter consisting position value. In this paper, neural networks is proposed to AC Servo Motor. It adopted back-propagation algorithm and executed in a DSP. For the implementation of high performance speed control system. It is used a 32-bit DSP, TMS320C31, which is capable of the high speed and floating point calculation. The design technic of neural networks proposed learning algorithm was presented, and the designed controller was applied to the accurate position control of AC servo system. The proposed controller for good tracking command values for motor driving is verified through the experimental results.