The New Intelligent Controller Applied Research In The The Bdcm Ac Servo System

~I~盩h~ Abstract rn industrial control process, conventional control theory (including classical control theory and modem control theory) has played a vital part for its full-blown theory and accuracy control. The primary characteristic of conventional control theory is that control is based on accuracy mathematical model and primarily adapts to linear system. But it is incapacity to strong coupling, time-varying, nonlinear system, and so on whose mathematical model is uncertainty. In fact, the controlled plant and process become more and more complexity so that it is very difficult to control such system. The localization of conventional control stands out increasingly. In recent years, with the advancement of science technology and the development of production technology, adaptive control techniques developed at very fast speed, and intelligent techniques have also been applied to every aspect of automation fields. The combination of both of them can satisfy the requirement of complexity, uncertainty and high precision of system very well. Since intelligent control does not need to acquaint with precise model of system and is simple, it catches more and more attention in home and abroad and becomes research hotspot in control fields. In this paper, a special control methou is adopted that is the combination of neural network adaptive control and PID output feedback control, and DBP is adopted in the study of brushless direct current motor (BDCM) alternating current (AC) servo system. Higher performance AC servo system is applied in a lot of fields. The main purpose of this paper is how to use neural network in BDCM, how to decrease torque ripple in wide ranges of speed and how to improve the characteristics of current modulation so as to realize higher performance of speed control and more accuracy of position control. In this paper, the combination of recurrent neural network controller and PID controller is introduced. From the viewpoint of the practical application, neural network adaptive control adapted to motors in real time is studied. And adaptive control of dynamic recurrent neural network with an identification network is applied to control a number of nonlinear plants to evaluate the effectiveness of RNC. The principle of AC servo system and mathematical model of BDCM are analyzed. The reason of toque ripple, modeling and simulation of BDCM are also discussed. In terms of hardware, the 8098 single chip processor is taken as the core, and AC servo system controlled by computers is made up of the BDCM and the IGBT-PWM inverter. Considering the control algorithm of software, the programs of 8098 single-chip processor are adopted to achieve speed and position closed loop control. The theory analysis, simulation and running on Site show that the BDCM has wonderful performances: quick response, almost non-overshoot, wider speed modulation range, higher locating accuracy, smaller torque ripple, good anti-disturbance, and stronger robustness, satisfying static and dynamic characteristics.