基于神经网络广义逆的PMSM模型参考自适应解耦控制器  

Model Reference Adaptive Decoupling Controller Based on Neural Network Generalized Inverse for PMSM

作　　者：蒋彦[1] 刘国海[1] 董蓓蓓[1] 赵文祥[1] 

机构地区：[1]江苏大学电气信息工程学院,江苏镇江212013

出　　处：《信息与控制》2013年第6期657-663,共7页Information and Control

基　　金：国家自然科学基金资助项目(51077066;61273154;51277194);江苏省自然科学基金资助项目(BK2012711);高校博士点基金资助项目(20123227110012);江苏省高校优势学科建设工程项目

摘　　要：设计了一种应用于永磁同步电机(PMSM)的自适应解耦控制器,实现了调速系统的线性化及解耦.该智能控制器采用模型参考自适应解耦神经网络广义逆控制器(MRAD-NNGI),由模型参考自适应控制器构成其反馈环,2个前置控制器构成其前馈环.MRAD-NNGI控制器集合了自适应理论高精度跟踪与人工神经网络解耦的优点.前置控制器可以有效地提高系统的抗干扰性能.仿真与实验结果表明,即使在PMSM负载扰动以及系统参数变化的情况下,MRAD-NNGI控制器依然能表现出很好的稳定性与精确的响应性.An adaptive decoupling controller for permanent-magnet synchronous motor （PMSM） is proposed to linearize and decouple the speed regulation system. The proposed intelligent controller adopts a model reference adaptive decoupling- neural network generalized inverse （MRAD-NNGI） controller, in which the feedback loop is constructed by model reference adaptive （MRA） controller and the feed-forward loop is constructed by two front controllers. The MRAD-NNGI controller combines the merits of the high-precision tracking ability of self-adaptive theory and the decoupling ability of artificial neural networks. The feed-forward controller is employed to effectively improve the anti-interference performance. The simulation and experimental results show that the proposed MRAD-NNGI controller takes on good stability and precise response regardless of the load disturbances and the system parameters variations.

关 键 词：永磁同步电机 神经网络广义逆 模型参考自适应 解耦 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]