【作者】 伍铁斌；

【导师】 刘祖润； 王俊年；

【作者基本信息】 湖南科技大学 ， 控制理论与控制工程， 2007， 硕士

【摘要】 PID控制器是最早发展起来的控制策略之一,因为结构简单,容易实现,并且具有较强的鲁棒性,因而被广泛应用于各种工业过程控制中。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,各种先进PID控制器参数整定方法层出不穷,给PID控制器参数整定的研究带来了无限活力和契机。自整定技术的发展一方面减轻了控制工程师现场调试的工作量,节省了大量的时间,另一方面也使整定的结果更加理想,并使一些复杂但是更加精细的设计方法得以应用于实际工业过程。然而很多先进的PID参数整定方法并没有像预期的那样产生完美的控制效果。所以研究PID控制器参数自整定具有很高的学术和工程应用价值,本文就对这个问题进行一些较为深入的研究。论文第一章讨论了PID参数整定的意义和发展状况,并介绍了数字PID的几种改进型。第二章详细介绍了遗传算法,利用遗传算法具有不用求导数,不必对问题局部线性化,对初始模型要求较低和鲁棒性强等优点,进行PID参数整定,仿真表明自适应遗传算法具有较强的寻优能力。第三章介绍了DNA遗传算法的基本原理,利用DNA遗传算法十分灵活,DNA染色体长度的可变性,使插入和删除碱基序列的操作更易实现等特点,应用于PID参数整定,仿真表明,采用DNA遗传算法在进化代数相同时能找到比常规遗传算法更优的控制参数,该算法对PID控制参数寻优是实用的和有效的,具有很好的应用前景。第四章第二节在传统的混沌算法的基础上,引入微粒群算法的寻优思想,形成了一种混沌微粒群算法,并应用在PID控制器的参数优化上,仿真证明了该算法能有效地实现PID参数最优整定,寻优速度快,容易实现,为解决PID控制器参数全局最优设计提供了一种新的有效方法。第四章第三节将遗传算法和混沌优化方法智能集成,利用混沌序列的“遍历性、随机性、规律性”的特点生成初始种群,在遗传操作中加入混沌细搜索,大大提高了局部搜索能力,能有效防止遗传算法陷入局部最优和发生早熟现象,仿真表明,混沌遗传算法优化结果相当理想,效果令人满意,优于常规的遗传算法。第五章是对论文的综述和个人的一点展望。更多还原

【Abstract】 PID is one of the earliest control measures, it is used widely in kinds of industry circumstance for its simple structure, easy implementation and strong robustness. 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, all kinds of parameters tuning methods of PID controller have emerged in endlessly, which bring much energy for the study of PID controller. The development of auto-tune release control engineers from field configuration, and save a large amount of time. On the other hand, it makes tuning result more reliable, and some refined but complex methods can be used in practical industrial process control. But many advanced tuning methods behave not so perfect as to be expected as to be expected. So there are academic value and engineering application value which study PID parameter auto-tuning. The paper go deep into studying the question.In the chapter one, We discusses the meanings of PID parameter auto-tuning methods and researching achievements on this subject. In the chapter two, genetic algorithm was described in detail. GA has such good qualities as no differential coefficient requiring, no local linearization, low request for initial model, robust and so on. It is applied in PID para--meter auto-tuning, simulation results show that the adaptive genetic algorithm has perfect optimization effect. In the chapter three, the basic principles of DNA-GA is presented. DNA-GA is very agile, the length of the DNA chromosome is veried which makes insert and delete DNA sequence easy to realize. It is applied in PID parameter auto- tuning, simulation results show that DNA-GA can search more excellent parameters than GA in the same evolution and that the algorithm for optimizing parameter is applied and effective, and is much better than that of common Genetic Algorithm, and has good perfectible application future. In section two of the chapter four, a new CPSO algorithm is formed by combining the traditional chaos algorithm and Particle Swarm Optimization, and used in PID controller to optimize parameters. Simulation results show that the algorithm is efficient to realize self-tuning global optimal parameters of PID controller, which have the advantages of stability and small overshoot, and it is easy to realize, highly effective and speediness. The algorithm supports a effective method for searching for global optimal PID parameters. In section three of the chapter four, we integrate GA and chaos optimizing, by the use of the chaos serial’s property of "ergodicity, randomicity, regularity", original population is generated; adding chaos operator to simple genetic algorithm greatly improves the local search ability, which avoids local optimization and premature convergence in effect. The results of the examples demonstrate that the chaos genetic algorithm has ideal and satisfied optimization result and much better than that of common Genetic Algorithm. The summary of the paper and personal perspective are given in the chapter five.更多还原