基于神经网络的感应电动机命令滤波反步控制
本文选题:感应电动机 切入点:命令滤波技术 出处:《青岛大学》2017年硕士论文 论文类型:学位论文
【摘要】:感应电动机以其简单的结构,低廉的制造成本,运行时较低的损耗和较高的稳定性等优势使其逐渐在工业应用中扮演了重要角色。然而,由于感应电机驱动系统为高阶、强耦合、参数时变的非线性系统,使其在运行过程中容易受负载扰动和参数时变的影响而降低运行效率。所以,研究新颖有效的控制方法来提升感应电机驱动系统的动态和静态特性,是一个具有重要理论意义与实际生产价值的研究方向。针对经典的感应电机控制策略中存在的一些不足,本文结合命令滤波自适应反步原理研究了感应电机的神经网络速度调节和位置跟踪控制方法。运用神经网络逼近感应电机系统中的非线性项,通过引入命令滤波技术和自适应反步原理,实现了对感应电机驱动系统的有效控制。论文的主要研究成果如下:1.研究了具有严格反馈结构的非线性系统模型的命令滤波自适应反步控制策略。利用神经网络系统逼近系统中的非线性项,通过引入命令滤波器有效地避免了“计算爆炸”问题,根据反步原理构造RBF网络自适应控制器,并使用李雅普诺夫方法给出了系统的稳定性证明。2.研究了基于命令滤波神经网络自适应反步技术的感应电动机速度调节和位置跟踪控制问题。利用神经网络系统逼近感应电动机系统中的非线性项,通过引入命令滤波技术来避免对虚拟控制信号的反复求导问题,采用反步法构造整个系统模型的真实控制器,根据李雅普诺夫原理对其稳定性进行分析。本文构造出的真实控制律只有一个自适应参数,因此更容易应用于工程实际,并且能够克服参数不确定及负载扰动的影响,模型中的每个状态信号都是有界的。最后在MATLAB仿真环境下,对所提策略的有效性给出了直观证明。3.研究了基于滤波误差补偿机制的命令滤波自适应神经网络速度调节控制器。利用RBF网络逼近感应电机驱动模型中的非线性项,命令滤波器被用来解决由于对虚拟控制律反复求取导数而大大增加计算负担的不足,通过引入误差补偿信号,对命令滤波器的滤波误差进行有效补偿,从而减少了滤波器误差对控制系统产生的影响,同时运用反步原理构造整个系统的真实控制器。所构造的控制器可以使感应电机的转子速度快速准确地对期望信号进行有效跟踪,且所有状态变量均有界。从仿真的结果可以看出,在参数未知和负载扰动的影响下,所提出的控制方法仍能使系统获得很好的鲁棒性和跟踪效果。
[Abstract]:Induction motors have gradually played an important role in industrial applications because of their simple structure, low manufacturing cost, low loss while running and high stability. However, the induction motor drive system is of high order. The nonlinear system with strong coupling and time-varying parameters makes it easy to be affected by load disturbance and parameter time-varying during operation. To study new and effective control methods to improve the dynamic and static characteristics of induction motor drive system is a research direction with important theoretical significance and practical production value. In this paper, the neural network speed regulation and position tracking control method of induction motor are studied based on the principle of command filter adaptive backstepping. The nonlinear term in induction motor system is approximated by neural network. By introducing command filtering technique and adaptive backstepping principle, The main research results of this paper are as follows: 1. The command filtering adaptive backstepping control strategy of nonlinear system model with strict feedback structure is studied. The system approximates the nonlinear term in the system, By introducing a command filter, the problem of "computational explosion" is effectively avoided, and the adaptive controller of RBF network is constructed according to the principle of backstepping. The stability proof of the system is given by using Lyapunov method. Secondly, the speed regulation and position tracking control of induction motor based on command filtering neural network adaptive backstepping technique are studied. Approaching the nonlinear term in Induction Motor system, The command filtering technique is introduced to avoid the repeated derivation of the virtual control signal, and the real controller of the whole system model is constructed by using the backstepping method. The stability of the real control law is analyzed according to Lyapunov principle. The real control law constructed in this paper has only one adaptive parameter, so it is easier to be applied to engineering practice and can overcome the influence of parameter uncertainty and load disturbance. Each state signal in the model is bounded. Finally, in the MATLAB simulation environment, The effectiveness of the proposed strategy is proved intuitively. 3. The command filter adaptive neural network speed regulator based on filter error compensation mechanism is studied. The nonlinear term in the driving model of induction motor is approximated by RBF neural network. The command filter is used to solve the problem that the derivative of the virtual control law is repeatedly obtained and the computational burden is greatly increased. By introducing the error compensation signal, the filter error of the command filter can be compensated effectively. Thus reducing the influence of filter error on the control system, At the same time, the real controller of the whole system is constructed by using the principle of backstepping. The controller can make the rotor speed of induction motor track the desired signal efficiently and accurately. And all state variables are bounded. The simulation results show that the proposed control method can still achieve good robustness and tracking performance under the influence of unknown parameters and load disturbance.
【学位授予单位】:青岛大学
【学位级别】:硕士
【学位授予年份】:2017
【分类号】:TP273;TM346
【参考文献】
相关期刊论文 前10条
1 于金鹏;于海生;林崇;;考虑铁损的异步电动机模糊自适应命令滤波反步控制[J];控制与决策;2016年12期
2 王晓玲;马玉梅;于金鹏;宋修彬;刘丽超;;异步电动机的命令滤波自适应速度调节控制[J];青岛大学学报(工程技术版);2016年01期
3 史发涛;马玉梅;于金鹏;李琪炜;赵全文;;基于动态面技术的异步电动机模糊自适应速度控制[J];青岛大学学报(工程技术版);2015年03期
4 赵帅;于海生;;基于自适应模糊控制的异步电动机DTC方法[J];青岛大学学报(工程技术版);2015年01期
5 李琪炜;于金鹏;史发涛;曲子扬;刘加朋;李伟;;基于Micro830的异步电动机控制系统设计[J];青岛大学学报(工程技术版);2014年03期
6 张立嘉;于海生;;双馈感应电动机的状态误差端口哈密顿建模和控制[J];青岛大学学报(工程技术版);2012年04期
7 刘星平;黄峰;;基于神经网络的感应电动机直接转矩控制研究[J];电力电子技术;2012年08期
8 刘进;于海生;魏晓晨;于珊珊;于金鹏;;异步电动机的端口受控哈密顿控制与L2增益扰动抑制[J];电机与控制应用;2010年10期
9 苗敬利;李华德;胡广大;赵仁涛;;感应电机效率优化的自适应反步控制研究[J];电机与控制学报;2009年05期
10 樊扬;瞿文龙;陆海峰;程小猛;张星;伍理勋;蒋时军;;一种基于磁链误差的异步电动机定子电阻校正方法[J];电工技术学报;2009年01期
,本文编号:1622734
本文链接:https://www.wllwen.com/kejilunwen/dianlidianqilunwen/1622734.html