一种基于神经网络逆的感应电动机解耦线性化改进策略  

An Improving Method to Decoupling and Linearization of Induction Motor Based on Neural Network Inverse

作　　者：王新[1] 戴先中[1] 

机构地区：[1]东南大学自动化学院,南京210096

出　　处：《电工技术学报》2008年第4期32-38,共7页Transactions of China Electrotechnical Society

摘　　要：由于感应电动机运行过程中的参数变化,磁场定向控制和解析逆控制所实现的解耦线性化遭到破坏。为此,基于输出为转子磁链幅值和转速的电流控制型感应电动机模型,本文提出了一种神经网络逆解耦线性化方法,理论分析表明,此方法可以实现感应电动机系统的自适应解耦线性化,弱化转子磁链与转速之间的耦合,从而简化外环控制器的设计,进一步提高整个系统控制性能。最后,对采用所提解耦线性化方法的整个感应电动机控制系统进行仿真研究,仿真结果对比表明该解耦线性化方法是有效的。Due to parameters’ variation during operation of induction motor, the decoupling and linearization implemented by field oriented control （FOC） and analytical inverse control（ANIC） is destroyed. For that, based on current-fed induction motor model whose outputs are rotor flux magnitude and speed, a neural network inverse decoupling and linearization method is proposed, theoretically analysis shows that the method can implement the adaptive decoupling and linearization of induction motor, and make the couple between the rotor flux and speed weaker and the design of outer loop controller easier, so the whole system control performance is further improved. At last, taking induction motor system which adapted proposed method as object, the study of simulation is done. The comparison simulation results show that the proposed decoupling and linearization method is valid.

关 键 词：转子磁链 神经网络逆系统 自适应解耦线性化 感应电动机 鲁棒性 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]