模糊控制技术在过程控制中的应用研究

【作者】 朱良红；

【导师】 王永初；

【作者基本信息】 华侨大学 ， 测试计量技术及仪器， 2004， 硕士

【摘要】 自1965年Zadeh教授创建模糊集理论和1974年Mamdani成功地将模糊控制应用于锅炉、蒸汽机的控制以来，模糊控制技术得以迅速的发展并在工程中得以广泛地应用。模糊控制技术之所以具有如此强大的生命力，是因为它具有其他控制方法不可替代的优点：其一是不需要被控对象的精确数学模型；其二是控制速度快、鲁棒性好，模糊控制的上升特性比其它控制方法的好，这一点和人的反应是相近的，虽然说人对细微的变化不是很敏感，但是对于大的变化却能够快速，准确的判断和加以控制；其三是模糊控制本身具有预测功能，这一功能可以从模糊控制规则中体现出来。模糊控制规则来源于专家的经验知识，任何一个专家在制定某个计划和控制策略的时候，都是在对未来有充分的估计和预测的基础上制定出来的，因而可以说模糊控制本身就具备预测功能，这点相比于其它控制方法是最为难能可贵的。 但是模糊控制也有其自身的缺陷：其一，数学机理还不是很清楚，总体来说模糊控制的理论研究是落后其应用的。其二，模糊控制的核心部分——控制规则，过分依赖于专家的经验知识，如果说没有专家的经验，或者是不全的经验，那么该模糊控制器的控制规则是有漏洞的，就有可能出现意想不到的结果。其三，模糊控制的控制精度不高，主要原因是模糊控制的稳态误差和零点极限环振荡问题，这一缺陷直接制约了其在高精度控制领域的应用。本文正是出于对这些问题的考虑，做了一些试探性工作，具体如下： 1．自调整模糊控制器与规则自适应模糊控制器设计，在模糊控制系统中，模糊控制的性能在很大程度上取决于模糊控制规则的确定是否合理以及模糊控制器的有关参数大小选择是否合适。对于一般的控制系统，采用相同的模糊控制规则以及一组固定不变的参数，往往控制性能不够理想。本论文分别探讨了比例参数K_u的自调整模糊控制和规则自适应模糊控制两种方法，实验表明，控制效果良好。 2．除稳态误差，提高控制精度，我们在探索新的控制方法的同时，也不能忽视了传统控制方法的优点。把模糊控制和其他的控制方法相结合，相辅相成，是现代控制智能控制研究的一种方法。本论文探讨了把模糊控制和PID相结合和在模糊控制中引入智能积分器两种方法，在实验中得到很好的控制效果。 3．模糊控制在MIMO系统中的应用，当前大多关于模糊控制的文献都集中在SISO系统中的应用，其实模糊控制的更大的优势是在MIMO系统中的应用。本论文探讨了模糊控制器本身的解耦特性，并且举出一个实例说明模糊控制在MIMO系统中的优越性。更多还原

【Abstract】 Since Prof. L.A.Zadeh of California University established the "Fuzzy Set" theory in 1965 and E.H.Mamdani firstly successfully applied the fuzzy control technology to the boiler and the steam engine control systems in 1974, the fuzzy control technology has been developed quickly and applied widely in control project. The fuzzy control technology therefore has the so formidable vitality, is because it has the merit which other control methods cannot be substituted: First is it needs not the precise mathematical model of the controlled object; Second is its fast controlling speed, good robustness, fuzzy control’s rising character is better than other control methods, this point is close to person’s response, although the person is not very sensitive to the slight change, they can fast and accurately judge and control to the big change; Third is fuzzy control technology itself has the forecast function, this function may manifest from the fuzzy control rule. The fuzzy control rule originates from expert’s experience knowledge. Any expert has full estimate and forecast to the future when they formulates some plan and control strategy, thus we can say fuzzy control itself has the forecast function, this spot will compare to other control methods is most commendable.But fuzzy control also has its own flaw: First, mathematics mechanism is not very clear, generally speaking the fuzzy control fundamental research is falls behind its application. Second, the fuzzy control’s core partial - controls rule, excessively relies on expert’s experience knowledge, if be short of expert’s experience, or has not the entire experience, then this fuzzy controller’s control rule has the loophole and maybe appear the unexpected result. Third, the fuzzy control’s precision is not high. The main reason is the fuzzy control static error and the zero vibration, which has directly restricted it in the high accuracy control domain application. This article has done some exploratory work for these questions, specifically as follows:1.The design of auto-adjusted fuzzy controller and the rule auto-adapted fuzzy controller. In the fuzzy control system, the performance is decided greatly by the fuzzy control rule and related parameter. Regarding the general control system, the performance is not ideal with the same fuzzy control rule as well as group of fixed invariable parameters, The paper has separately discussed two design methods of auto-adjusted proportion parameter Ku fuzzy controller and the rule auto-adapted fuzzy controller, the experiment indicated that, the control effect is good.2.Eliminating static error, increasing the control precision. While we explore the new control method, we shouldn’t neglect the traditional control method’s merit. Unifying the fuzzy control and other control methods is good research way in field ofmodern intelligence control. The paper has discussed two methods of unifying the fuzzy control with PID and the intelligent integrator. We obtain the very good control effect in the experiment.3.The application of fuzzy control in the MIMO system. Current literatures about the application of fuzzy control mostly all concentrate in the SISO system. Actually the fuzzy control’s bigger superiority is applied in the MIMO system. The paper has discussed fuzzy controller’s characteristic of eliminating coupling, and pointed out its merit by an example.更多还原