【作者】 张金超；

【导师】 任彦硕；

【作者基本信息】 东北大学 ， 控制理论与控制工程， 2008， 硕士

【摘要】 矢量控制技术属于交流调速领域的高性能变频调速技术,目前,在该领域得到了深入的研究,并以其优良的性能得到了广泛应用。但是转子磁链计算精度受易于变化的转子电阻的影响,转子磁链的角度精度影响定向的准确性。针对矢量控制技术的不足,有很多研究者从多方面进行了改进。以人工神经网络控制为代表的智能控制方法的出现为提高交流调速系统性能提供了有效的控制方法。把智能控制技术和矢量控制方法结合起来,成为国内外电气传动工作者研究的热点,但都没有到大规模实用的程度,将神经网络控制方法和传统控制方法结合起来应用到交流调速系统当前正处于研究阶段。因此,在这些方面的研究是很有意义的。PID神经网络是一种动态神经网络,具有结构简单和系统稳定的特点,克服了传统神经网络的稳定性差和设计复杂等缺点。本论文将PID神经网络控制方法引入交流传动系统,并与常规控制方法相结合,力图以新的控制方式提高系统的动态性能和鲁棒性。论文在深入分析PID神经网络原理的基础上,研究将PID神经网络控制算法应用于感应电机调速控制,并和矢量控制相结合以设计性能更优良的调速系统。针对目前神经网络应用于自动控制系统的实用性方法,设计研究了两种神经网络控制策略：其一是直接运用SPIDNN作为速度控制器,组建一个自适应速度控制系统；其二是用MPIDNN神经网络来构建交流感应电机的动态逆系统,实现三相感应电机调速系统的解耦和线性化。在详细阐述控制原理和算法的基础上,利用具有交互功能的MATLAB/Simulink仿真工具平台,对其进行了仿真研究,仿真结果表明加入了PID神经网络的应用使控制系统具有更好的控制效果。更多还原

【Abstract】 Vector control technology is belang to high-performance frequency conversion speed control technology in the field of AC speed control, nowadays, it get embedded study, and broad application based on its high-quality. However, the easy-changed rotor resistance has an influence on the account accuracy of rotor magnetic linkage, the angle accuracy of rotor magnetic linkage affect the orient accuracy.The emergence of intellect control method based on artificial neural network provides AC speed control system performance with effective control mode. To combine intellect control technology with vector control mode has become the study hotspot of operators on electric drive in both domestic and other counties, but none of them have reached the cosmically application, the study of neural network controller combine conventional controller is still going on. Therefore, researches on these aspects are significant. PID neural network is a kind of dynamic network, unlike conventional neural network, it’s sample in stuctrue and stable. This thesis bring the PID neural network control mode into AC drive system, and combined with conventionality control mode, Trying to improve system’s dynamic-state behavior and robustness on new control mode.This thesis, based on PID neural network theory, trying to use PID neural network control arithmetic into induction machine speed control, and combined with vector control to design more excellent speed control system. According to the application of neural network in auto control system at the present time, two neural network speed controllers were designed in this thesis, one is to make the SPIDNN as speed controller, and bring forward SPIDNN self-adaptive controller, the other use the MPIDNN neural network to construct dynamic-inverse system of AC induction machine, realize the nonlinear system’s decoupling and linearization. Basing on particular expatiate on control theory and algorithm, this paper utilize MATLAB/Simulink emulated tool platform with interactive function and make a emulation study of it, the result indicate that the application of PID neural network makes the control system more effective.更多还原