摘  要：针对传统感应电动机伺服驱动系统的位置与速度外环PI控制的结构复杂、双闭环耦合及对参数等不确定性扰动鲁棒性差的问题，在直接转矩控制理论将感应电动机的转矩与磁链解耦的基础上提出了基于动态神经网络的自适应控制方案，简化了控制系统结构，它可随着伺服驱动系统的运行工况而改变控制系统的结构参数，大大提高了伺服驱动系统对参数变化的鲁棒性，同时,也较好地改善了伺服驱动控制系统的动态及稳态性能。最后通过实验验证了该控制系统的有效性和可行性。
关键词：伺服驱动系统；感应电动机；自适应控制；动态神经网络；直接转矩控制;鲁棒控制

Abstract: For the problems of structural complexity,double close loop coupling,and sensitivity to parameter change about the classic outer-loop PI control of position and rotor speed for servo system of induction motor,on the basis of decoupling between torque and flux by DTC theory,this dissertation presented an adaptive control-scheme based on dynamic neural network,which simplified control system structure,it could change structure parameter of the control system along with function condition of the system,greatly enhanced robustness to parameter-changes,at one time,and improved preferably dynamic and stable-state performance of the control system.Finally,the experiment results show its effectiveness and feasibility.
Key Words: Servo drive system;Induction motor；Adaptive control；Dynanic neural network；Direct torque control(DTC);Robust control