基于改进蚁群优化神经网络反推控制的IM鲁棒控制

Advanced Ant Colony Optimized Neural Network Based Backstepping Robust Control for Induction Motors

针对六相铜转子感应电机(SpCRIM)在不确定扰动条件下的鲁棒控制,提出一种基于改进蚁群优化(AACO)递归罗曼诺夫斯基多项式神经网络(RRoPNN)的反推控制策略。首先,基于反推控制理论设计了SpCRIM的控制律,并提出了一种改进的具有自适应律的RRoPNN,以实现对反推控制律中的总不确定度进行估计;然后设计了相应的误差估计律对网络观测误差进行补偿,同时实现在线参数调节;为防止早熟并加快所提RRoPNN的收敛速度,提出了AACO算法对RRoPNN连接权值学习率进行调整;通过Lyapunov稳定性理论证明了所提控制方法的鲁棒性;最后,对所提出控制器的位置跟踪性能进行了实验验证,并与经典PI控制器及基于开关函数的反推控制器进行了对比。结果表明,所提控制方法具有更为良好的位置跟踪精度和鲁棒性。

Aiming at robust control of six-phase copper rotor induction motors(SpCRIM)under uncertain disturbances,a new backstepping control strategy based on advanced ant colony optimized(AACO)recursive Romanovski polynomial neural network(RRoPNN)was proposed.Based on the theory of backstepping control theory,the control law for SpCRIM was firstly designed,and an improved RRoPNN with adaptive law was proposed to estimate the lump uncertainty in the backstepping control law.The error estimation law was then designed to compensate the network observation error and to realize on-line parameter adjustment.In order to prevent precocity and accelerate the convergence rate of the proposed RRoPNN,an AACO algorithm was proposed to adjust the learning rate of RRoPNN connection weights.The robustness of the proposed control method was proved based on Lyapunov stability theory.Finally,the position tracking performance of the proposed controller was verified by experiments and compared with the classical PI controller and the switch function based backstepping controller.The results show that the proposed control method has better position tracking accuracy and robustness.