感应电机无速度传感器直接转矩控制的智能优化设计

摘要

在一些特定作业环境中,需要对电机转速进行精确调控.为了改善直接转矩控制系统低速时常规PI调节器带来的转速脉动,设计了基于模糊神经网络的新型PI调节器,改进后的转速调节器具有快速自整定、微超调的特性.针对MARS转速观测器低速时的不稳定性,设计了基于人工神经网络的模型参考自适应速度观测器,改进后的转速观测器具有良好的动、静态特性.仿真结果表明:智能优化设计使得直接转矩控制系统的控制效果明显提高,且有更好的鲁棒性和自适应性.%In some special working environment, the speed of motor should be controlled accurately. In order to reducing rotor speed ripple caused by general regulator of direct torque control system in low speed, a new PI regulator based on fuzzy neural network was designed,and the modified speed regulator has fast self-regulating, micro over modulation characteristic. Aiming at improving the stability of MARS speed observer in low speed, a model reference speed-adaptive observer based on artificial neural network was designed, and the modified observer has good tendency , static characteristic. Simulation results showed that intelligent optimization design makes direct torque control system a obviously improved performance, better robustness and self-adaptive.