CAJ | 학술논문

当永磁同步电机速度控制系统面对变负载和大范围调速时，PID控制器参数需要频繁调整且速度跟踪不理想。通过结合滑模理论和神经网络方法，本文提出一种变负载自适应速度控制策略。控制器可以分为PI控制和神经网络控制两部分。其中，PI控制器用于保证系统前期收敛和稳定性，其输出信号作为误差反馈信号输入到神经网络，并基于线性滑模理论设计权值参数的调节律。当系统负载发生变化时，神经网络控制器的输出将取代PI控制器，逐渐成为系统的主要速度控制信号。仿真结果显示，本文提出的速度控制方法能够提高系统的速度跟踪性能和抗干扰能力。
당영자동보전궤속도공제계통면대변부재화대범위조속시，PID공제기삼수수요빈번조정차속도근종불이상。통과결합활모이론화신경망락방법，본문제출일충변부재자괄응속도공제책략。공제기가이분위PI공제화신경망락공제량부분。기중，PI공제기용우보증계통전기수렴화은정성，기수출신호작위오차반궤신호수입도신경망락，병기우선성활모이론설계권치삼수적조절률。당계통부재발생변화시，신경망락공제기적수출장취대PI공제기，축점성위계통적주요속도공제신호。방진결과현시，본문제출적속도공제방법능구제고계통적속도근종성능화항간우능력。
When wide range of speed and variable load are faced in permanent magnet synchronous motor(PMSM)speed control system,frequently tuning of PID controller parameters is needed and the tracking performance is unsatisfactory. In this paper,an adaptive speed control scheme is proposed by combining the sliding mode control(SMC)principle and a neural network(NN). The controller can be devided into two parts:a PI controller and a NN controller. The former is provided to guarantee system convergence and stability in early stage,and its output is the input of the NN as a feedback error signal. The update law of NN weight is designed based on SMC principle,and the output of the NN controller gradually replaces that of the PI controller when the system load varies. Simulation results show that the proposed speed control method can improve the system speed tracking performance and anti-disturbance ability.