Dissertation > Excellent graduate degree dissertation topics show

Design and Simulation of Decoupling Control on Multivariable Coupling System

Author: LongZuoZuo
Tutor: JiangPinQun
School: Guangxi Normal University
Course: Circuits and Systems
Keywords: Multivariable coupling system Class feedforward decoupling Neural Network Decoupling PID control
CLC: TP273
Type: Master's thesis
Year: 2010
Downloads: 530
Quote: 6
Read: Download Dissertation

Abstract

In recent years, with the rapid development of modern industry, the production process more complex, their control systems often have multiple input and output, the variable coupling between the control system has become a widespread phenomenon. Coupled there not only to make the system very difficult to control, but also greatly reduce the system control quality, serious and even cause the system can not run. Decoupling has become coupled variables increase the level of automation of the production process to meet the growing requirements of an important means of control. Therefore, multivariable decoupling control of coupled system has great practical significance and application prospects. In addition, as China's rapid economic development, the film more widely used, their quality requirements are also increasing. Film thickness control system is the extremely important film production line component of its quality control determines whether the quality of the film. As the film thickness system is a typical multi-variable coupling system decoupling control more practical significance and application value. This paper introduces the topic of sources of research background and significance, outlining the research status decoupling control, variable degree of coupling analysis method, three kinds associated decoupling method, PID controller and a neural network infrastructure; then in the full investigation based on the design and simulation of the film thickness of the class system feedforward decoupling PID control, as well as a two-variable coupling system RBF neural network decoupling PID control; Finally, this paper summarizes the work and outlook. Completed the following two main tasks: (1) the class system film thickness feedforward decoupling PID control design and simulation using the relative gain matrix of the film thickness of the coupling system was analyzed using multivariable class feedforward decoupling and PID control a combination of methods to achieve a fully dynamic decoupling control system. And inverse Nyquist array decoupling PID control compared to class feedforward decoupling PID control has not changed the main control channel characteristics, completely decoupled control and good effect, with great application value. (2) one pair of variable coupling system for RBF Neural Network Decoupling PID control design and simulation of complex algebraic algorithms nonlinear optimization problem into a set of linear algebraic equations to solve, is a new neural network learning algorithm, using algebraic algorithm training RBF neural network to achieve the precise mapping of the sample network, and to ensure faster convergence. This paper uses a dynamic RBF neural network a two-variable coupling system for online identification, and will receive information on the sensitivity of self-tuning PID controller parameters to achieve a decoupling control system. Simulation results show that the decoupling control with high precision, real-time, robustness and adaptability, etc., with great application value. In summary, this thesis class feedforward decoupling PID control and RBF neural network decoupling PID control can effectively achieve multi-variable coupling decoupling control system, greatly improve the control quality, with great practical significance and application value.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Automation technology and equipment > Automation systems > Automatic control,automatic control system
© 2012 www.DissertationTopic.Net Mobile