【作者】 陆秀银；

【导师】 刘星桥；

【作者基本信息】 江苏大学 ， 电力电子与电力传动， 2006， 硕士

【摘要】 本文针对交流感应电机同步控制系统多变量解耦控制问题，应用神经网络控制方法对由两台交流电机和变频器组成的两电机同步系统的速度和张力解耦控制进行研究。首先，在对两电机同步系统模型分析的基础上，依据同步系统的结构特点和控制要求，结合人工神经网络的非线性映射、自适应、自学习等能力，提出了一种新的基于神经网络的两电机同步系统控制方案，其中神经网络控制器由基于RBF网络整定的自适应PID控制器和神经元解耦补偿器两部分组成。两个自适应PID控制器分别对速度控制回路和张力控制回路进行自适应控制，使系统具有更强的适应能力、更好的实时性和鲁棒性；神经元解耦补偿器综合两控制回路的耦合作用，通过训练网络权值，补偿各回路之间的耦合影响，实现速度和张力的解耦。整个神经网络控制算法通过西门子可编程控制器S7-300实现，STEP7软件采用结构化编程方法，可以高效灵活地编写神经网络控制程序。另外还在系统内部建立了PLC与变频器之间的Profibus-DP现场通讯，上位机监控软件WinCC与PLC之间MPI数据通讯，完成了整个两电机同步系统的远程化智能控制。最后，在实验室自行设计的多电机同步系统实验平台上进行了实际的控制实验。大量的实验结果表明：采用神经网络控制方法实现了两电机同步系统中速度和张力的解耦控制，系统具有良好的动静态性能。论文提出的控制方法具有很强的适用性，可应用于许多实际的工业控制场合。更多还原

【Abstract】 The paper focuses on the multi-variable decoupling control of synchronization system of the AC induction motors. With the application of neural network control method, we make researches on the decoupling control between the velocity and tension of the two-motor synchronization system composed of two AC induction motors and transducers.On the basis of model analysis of the two-motor synchronization system, according to the structural character and controlling requests of the synchronization system, we put forward a new control strategy of the two-motor synchronization system based on artificial neural networks combined with its nonlinear mapping capability, adaptive and learning capability. The neural network controller is composed of adaptive PID controller using the RBF networks to modulate and neuron decoupling compensator. The two adaptive PID controllers are used in the velocity control loop and tension control loop respectively, that make the system possessing stronger adaptive capability and other better performances. The neuron decoupling compensator integrates the effects of the two loops each other and realizes the decoupling control between velocity and tension by training the weights of networks to compensate the coupling relation.The whole arithmetic of neural network control is carried out by the S7-300 PLC of Siemens.The software of STEP7 may compile neural network control program high effectively with its structural compiling method. In addition, we still set up the PROFIBUS-DP communication between PLC and transducers, MPI communication between WINCC and PLC. So we realize the long-distance intellectualized control of the whole two-motor synchronization system.Finally, we perform experiments on the platform of multi-motor synchronization system. Lots of experimental results indicate that the neural network control scheme has realized the decoupling control between the velocity and tension of the two-motor synchronization system with better dynamic and static characteristics. The control method proposed in the paper is promising and can be applied in many industrial control environments.更多还原