【作者】 金培；

【导师】 刘振娟；

【作者基本信息】 北京化工大学 ， 控制理论与控制工程， 2008， 硕士

【摘要】 PID控制器的特点是结构简单,鲁棒性强,规则容易理解等,而神经网络以其很强的适应于复杂环境和多目标控制要求的自学习能力,并能以任意精度逼近任意非线性连续函数的特性引起控制界的广泛关注。对于强非线性系统,利用神经网络的优点,设计PID控制器的在线调整控制系统,改善系统性能,无论在理论还是实践上都将具有重要意义。本文在吸取传统的经典控制理论强大的分析能力基础上,结合神经网络控制的特点,将神经网络控制与传统PID控制相结合,明确了神经网络对于解决传统过程控制问题的重要地位。设计了一种基于神经网络瞬时线性化的在线自校正PID控制器,对PID控制与神经网络相结合的几种神经网络PID算法进行了仿真分析与研究。对于pH中和反应控制系统,将神经网络用于其PID参数的在线整定,给出了设计神经网络在线自校正控制器的一般算法,并讨论了在MCGS中用VB语言实现该控制过程的技术方法。论文的工作充分表明基于神经网络的智能PID控制器具有良好的控制效果,有着广阔的发展前景。更多还原

【Abstract】 The PID controller is simple in structure, strong in robustness, and can be understood easily. Nevertheless, neural networks have great capability in solving complex mathematical problems since they have been proven to approximate any continuous function as accurately as possible. Hence, it has received considerable attention in the field of process control. Taking the advantage of neural network and applying it to controller design which PID parameters are adjusted on-line for highly nonlinear system, which can improve controller performance. So the research of PID control based on neural network is significant in either theory or application.Based on the analyzing capability of the classical control theory and the character of the neural network control, this paper combines the traditional PID control and the neural network control, this paper explicates the significant position for traditional process control. Subsequently, this paper designs an on-line self-tuning PID controller based on the instantaneous linearization neural network, discusses and simulates two kinds of PID algorithm based on neural network. What’s more, an on-line updated PID algorithm is proposed, and one simulation example of nonlinear pH neutralization system is discussed to demonstrate the applicability of the proposed algorithm. In the end, we discuss the technology method by VB realization based on MCGS.As is shown in the paper, the intelligent PID controller based on neural network has better control performance and the expansive foreground.更多还原