多变量系统模糊神经网络解耦的精编.doc

华北电力大学硕士论文摘要 摘 要 复杂的工业生产过程中，被控对象动态特性往往表现出非线性、时变性、大迟延、和多变量强耦合等特点，这使得常规的单回路及传统的多变量解耦设计难以达到预期的控制效果。目前这一主要的研究课题受到了广泛的关注，一些学者将模糊逻辑系统、神经网络、模糊神经网络等智能方法应用到多变量过程控制中。 本文在这些研究成果的基础上，提出了一种多变量系统模糊神经网络解耦的控制方法。该方法结合了基于模糊自适应整定PID控制算法和改进型模糊神经网络解耦算法，用于球磨煤机制粉系统和循环流化床锅炉床温-主汽压力系统中，对解耦设计方案进行了仿真研究，结果表明，该方案比传统的PID解耦控制方案的控制效果明显要好。 关键词：多变量系统，模糊自适应，模糊神经网络，解耦控制，床温-主汽压力系统，球磨煤机制粉系统 ABSTRACT It is difficult to achieves anticipated control effect with traditional Single-loop or muti-variable decoupling design methods. At present, many scholars have fans on this topic, many intelligent control methods such as fuzzy control, neural network, fuzzy neural network have been used in multivariable system. In the foundation of these results, a kind of multivariable fuzzy neural network decoupling control method is put forward, fuzzy auto-adapted PID control and advanced fuzzy network decoupling algorithm is used. The method has been used in ball mill system and CFB control. It is shown by simulation results that the control effect with this method is obviously better than traditional PID control method. Ping Yuhuan (School of control science engineering) Directed by prof. Yu Xining KEY WORDS: Multivariable system, fuzzy auto-adapted, FNN, decoupling control, bed temperature-main steam pressure control system; ball mill system 华北电力大学硕士论文目录 目 录 中文摘要 英文摘要  TOC \o 1-3 \h \z \u  HYPERLINK \l _Toc186526820 第一章 绪论  PAGEREF _Toc186526820 \h 1  HYPERLINK \l _Toc186526821 1.1课题的研究背景及意义  PAGEREF _Toc186526821 \h 1  HYPERLINK \l _Toc186526822 1.2多变量系统解耦理论的发展与现状  PAGEREF _Toc186526822 \h 1  HYPERLINK \l _Toc186526823 1.2.1传统解耦  PAGEREF _Toc186526823 \h 1  HYPERLINK \l _Toc186526824 1.2.2智能解耦方法  PAGEREF _Toc186526824 \h 2  HYPERLINK \l _Toc186526825 1.2.3模糊神经网络解耦  PAGEREF _Toc186526825 \h 3  HYPERLINK \l _Toc186526826 1.3论文的主要研究内容  PAGEREF _Toc186526826 \h 4  HYPERLINK \l _Toc186526827 第二章 多变量系统解耦的基本理论  PAGEREF _Toc186526827 \h 5  HYPERLINK \l _Toc1865