【作者】 曹刚；

【导师】 褚健； 苏宏业；

【作者基本信息】 浙江大学 ， 控制理论与控制工程， 2004， 硕士

【摘要】 随着工业生产的发展，于20世纪30年代，美国开始使用PID调节器，它比直接作用式调节器具有更好的控制效果，因而很快得到了工业界的认可。至今，在所有生产过程控制中，大部分的回路仍采用结构简单、鲁棒性强的PID控制或改进型PID控制策略。PID控制作为一种经典的控制方法，几乎遍及了整个工业自动化领域，是实际工业生产过程正常运行的基本保证；控制器的性能直接关系到生产过程的平稳高效运行以及产品的最终质量，因此控制系统的设计主要体现在控制器参数的整定上。随着计算机技术的飞跃发展和人工智能技术渗透到自动控制领域，近年来出现了各种实用的PID控制器参数整定方法。 PID控制算法作为最通用的控制方法，对它的参数整定有许多方法；对于不同的控制要求、不同的系统先验知识，考虑用不同的方法；这些算法既要考虑到收敛性、直观、简单易用，还要综合负载干扰、过程变化的影响，并能根据尽可能少的信息和计算量，给出较好的结果。 论文在较为全面地对PID控制器参数自整定方法的现状分析研究的基础上，针对基于继电器反馈和最小二乘的自整定方法以及其应用的可行性进行了相关的研究，主要的工作和结果概括如下： 为克服一自由度PID控制器无法兼顾目标跟踪和外扰抑制的缺点，结合二自由度控制器的结构和基于幅值最优化的控制器参数整定方法，并通过分析得到控制器参数求解公式，实现了二自由度PID控制器参数整定和二自由度PI控制器参数整定。与常规控制方法相比，该方法得到的控制器具有更好的闭环响应性能，并且由于二自由度系数的半固定性，在整定PID控制器参数之前就可以确定，因此，对控制器参数的求解难度无影响。 针对一类一阶大时滞不稳定特殊对象，普通的PID控制器很难满足要求，甚至不能实现系统的稳定。基于首先引入内环状态反馈，以改善对象动态特性的思想，采用双环控制结构，先将对象状态反馈镇定，然后按照内模控制原理设计外环的控制器。只要选择适当的可调参数λ的值，通过该方法得到的PID控制器对不稳定对象具有较好控制效果及鲁棒性。 通过仿真比较研究，对于连续对象，综合得到几种较好的基于继电器反馈的控制器参数整定方法，对离散采样数据采用基于最小二乘模型辨识的参数整定方 摘要法，提出并设计基于Matlab/simulink仿真工具的PID控制器参数整定仿真应用软件。介绍了PID整定控制器的应用框架、辅助设计与仿真软件的功能、特点，并给出了仿真实例。 最后是对论文的综述和展望。更多还原

【Abstract】 With the development of industrial technology, PID tuner, which appeared in 1930s, was accepted for its better control performance than direct-action tuner. Because of simple structure and strong robustness, PID controllers and transformative PID controllers have been widely used in industrial control systems. The capability of controller directly influences the qualities of producing process and products. Therefore, parameter tuning of controllers is the most important step during system design. With the development of computer technology and artificial intelligence in automatic control field, auto-tuning PID controllers began to be used in recent years.Parameters of PID controllers can be obtained with various algorithms. Excellent control algorithms are astringent, intuitionistic and simple. Load disturbance and set-point variety also need to be considered.This thesis studies various methods of parameter auto-tuning of PID controllers. Furthermore, it focuses on the feasibility of auto-tuning methods based on relay-feedback and the least squares. The main results obtained in this thesis are as follows:To avoid the shortcoming of traditional PID controllers that they are unable to give attention to reference response and disturbance response at one time, a novel auto-tuning algorithm, which combines two-degree-of-freedom structure with parameter tuning method based on magnitude optimum criterion, is proposed in this thesis. Compared with traditional methods of controller design, it has better closed-loop response performance. Furthermore, because the coefficients of two-degree are semi-fixation, they are irrelevant to the difficulty of the controller parameters’ tuning.Normal PID controllers can hardly satisfy the design requirements of unstable first-order systems with large dead time, and they cannot even ensure the stability of systems. Considering that inner-loop state feedback can improve systems’ dynamic characters, this thesis adopts the double-loop control structure. Firstly, the inner loop is introduced to get better systems’ dynamic performance; secondly, according to internal model control principle, the outer loop is designed to cater for systems’ requirements.By setting the parameter " A." properly, this double-loop PID controller will obtain better control performance and robustness.According to the simulation results, several relay feedback control methods are adopted in the continue-time control systems; for the discrete sample data, controller parameter is obtained by identify the parameters of process model using the least squares. This thesis also presents an auto-tuning software based on the platform of Matlab/Simulink. The functions and characters of this software are mentioned, and simulation examples axe also given.The conclusion and perspective are given in the end of the thesis.更多还原