双容水箱系统的辨识与滑模变结构控制

发布时间：2018-05-28 19:44

本文选题：多容水箱系统 + 滑模变结构控制　；参考：《西安理工大学》2017年硕士论文

【摘要】：在现代化工业生产中,过程自动控制起着非常重要的作用,液位控制技术广泛应用于过程工业。在实际生产中,液位控制的性能与整个装备的生产安全、效益和成本直接相关,为确保过程控制的安全性和平稳性,开发先进的液位控制技术和策略具有重要意义。多容水箱系统是工业生产中许多复杂控制系统的原型,是过程控制中的典型对象,也是控制算法验证的平台。多容水箱系统是典型的非线性系统,本文根据机理建模推导了系统的数学模型,通过系统辨识方法得到系统线性化模型参数以及非线性辨识法得到非线性模型参数,为后续控制器的设计奠定了基础。滑模变结构控制(SMC)作为一种处理具有不确定性系统的有效方法,具有响应快、对系统参数变化以及外部扰动不敏感、控制器结构简单等优点,在液位控制系统中得到了广泛的应用。然而,滑模变结构控制存在抖振的缺点,为了降低抖振改善系统性能,引入了“准滑动模态”的概念,即基于边界层的滑模变结构控制。设计了单容水箱系统和双容水箱系统一般滑模变结构控制器和边界层的滑模变结构控制器,并利用Lyapunov函数证明了所设计控制器的稳定性,仿真和实验结果验证了该方法的有效性。边界层的滑模变结构控制可以有效减弱抖振,但这是以增大跟踪误差为代价的,通过在滑模面中引入积分作用可以提高控制精度,但同时也导致瞬态响应超调大,时间长的缺点。采用具有“小误差放大,大误差饱和”功能的光滑非线性函数设计了一种非线性积分滑模变结构控制器,控制器的设计考虑了输入受限的影响,可以保证系统的输入限制在一定范围内,利用Lyapunov函数证明了所设计控制器的稳定性。该控制器比具有积分作用的常规SMC改进了瞬态性能,并且还能够保持传统SMC的期望属性,稳定调节,有效减弱抖振,非线性积分SMC相比条件积分SMC,调节时间更短,稳态误差更小,具有更好的性能。考虑输入受限情况时系统的动态性能没有降低。最后,提出了一种基于NN-SANARX模型的辨识与控制方法,这种方法的优点是无需对象的数学模型,只需通过实际对象的输入输出数据,利用神经网络模型辨识对象模型,进而利用动态输出反馈线性化方法设计基于神经网络辨识模型的控制器,实验结果表明了该方法的有效性。这种方法对于未知模型的对象是一种有效控制方法。
[Abstract]:Automatic process control plays an important role in modern industrial production, and liquid level control technology is widely used in process industry. In actual production, the performance of liquid level control is directly related to the production safety, benefit and cost of the whole equipment. In order to ensure the safety and stability of process control, it is of great significance to develop advanced level control technology and strategy. Multi-tank system is a prototype of many complex control systems in industrial production. It is a typical object in process control and a platform for verification of control algorithms. The multi-tank system is a typical nonlinear system. In this paper, the mathematical model of the system is derived according to the mechanism model, and the parameters of the system linearization model and the nonlinear identification method are obtained by the system identification method. It lays a foundation for the design of subsequent controller. Sliding mode variable structure control (SMC), as an effective method for dealing with uncertain systems, has the advantages of fast response, insensitivity to system parameters and external disturbances, simple controller structure, and so on. It is widely used in liquid level control system. However, sliding mode variable structure control has the disadvantage of buffeting. In order to reduce chattering and improve system performance, the concept of "quasi-sliding mode" is introduced, that is, sliding mode variable structure control based on boundary layer. The general sliding mode variable structure controller and the sliding mode variable structure controller of the boundary layer are designed for the single water tank system and the double water tank system. The stability of the designed controller is proved by using the Lyapunov function. Simulation and experimental results show the effectiveness of the proposed method. The sliding mode variable structure control in the boundary layer can reduce the chattering effectively, but this is at the cost of increasing the tracking error. The control accuracy can be improved by introducing the integral action in the sliding mode surface, but the transient response is overshoot at the same time. The drawback of a long time. A nonlinear integral sliding mode variable structure controller is designed by using a smooth nonlinear function with the function of "small error amplification and large error saturation". The design of the controller takes into account the effect of input constraints. The input of the system is limited to a certain range, and the stability of the designed controller is proved by using the Lyapunov function. The controller improves the transient performance compared with the conventional SMC with integral effect, and can maintain the expected properties of traditional SMC, stabilize the adjustment, effectively reduce the buffeting. The nonlinear integral SMC has shorter adjusting time than the conditional integral SMC. The steady-state error is smaller and has better performance. The dynamic performance of the system does not decrease when the input constraints are considered. Finally, a method of identification and control based on NN-SANARX model is proposed. The advantage of this method is that it does not need the mathematical model of the object, but only by the input and output data of the actual object, and the neural network model is used to identify the object model. Then the dynamic output feedback linearization method is used to design the controller based on neural network identification model. The experimental results show the effectiveness of the method. This method is an effective control method for the object of unknown model.
【学位授予单位】：西安理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP273

【参考文献】