CASIA OpenIR  > 毕业生  > 博士学位论文
基于迭代学习控制的精密给定系统关键问题研究
其他题名Research on Solid Accuracy Batching System Based on Iterative Learning Control
龚利
学位类型工学博士
导师王云宽
2007-06-11
学位授予单位中国科学院研究生院
学位授予地点中国科学院自动化研究所
学位专业控制理论与控制工程
关键词精密给定 学习控制 固体粉末染料 非线性系统 开放式系统 Accurate Batching Learning Control Solid Dyestuff Nonlinear System Distribute System
摘要染整配色是纺织印染行业的一道关键工序,但对于我国众多印染厂家来说目前仍然采用效率低可靠性也不高的人工配料,采用计算机技术的高精度配料设备已经成为该行业的发展趋势。本文以研制纺织印染行业精密配料设备为工程背景,将迭代学习控制理论和算法应用于开放式染整配色生产线硬件平台上,针对染整配色过程中的技术难点:多达上百种不同物理特性的固体粉末染料进行高精度称量,对染整配色设备的若干关键技术进行了研究。 本文首先对目前国内外自动固体粉末称重设备的特点、应用和发展趋势作了阐述,然后详细介绍了迭代学习控制理论的研究内容、现状、发展和应用等问题。并对本文的研究背景和主要内容作了介绍。 其次 对于构成染整配料设备最大非线性环节的螺旋输送机的加料段的固体输送机理进行了分析。并从力矩平衡方面提出对Darnell-Mol固体输送理论进行修正。此外,针对固体非塞流理论计算公式繁多,工程使用困难的缺点,建立了适合工程使用的较精确、方便和直观的五层模型法。 第三 针对一类非线性系统,把闭环PID控制加入高阶学习律,提出一种高阶开、闭环迭代学习控制。然后对于具有初始状态偏差的系统,研究了具有初态学习的高阶开闭环迭代学习律。两种情况都进行了推导迭代学习收敛条件,最后给出算法的仿真结果。 第四 针对以往迭代学习参数选取依靠试凑的情况,给出了一种PID型迭代学习律的参数优化设计方法。然后讨论一种利用系统输入输出数据构造“数据模型”来决定控制初始输入的方法。 第五 根据印染行业的特点和工艺要求,设计了基于总线结构的开放式的自动染整配色生产线控制系统,分别从硬件和软件的角度阐述了控制系统的开放性。然后以此设备为平台,设计了一套基于迭代自学习的控制方法,并在实际生产现场检验了算法的有效性。 最后,对取得的研究成果进行了总结,并展望了需要进一步研究的工作。
其他摘要Dyestuff accurate batching is a key process in the textile Industry. In our country, it mainly depends on human operation, which has low efficiency and reliability. So adopting the dyestuff-batching device controlled by computer is the tendency of textile Industry. The main result of this dissertation is to introduce some theory and method based on iterative learning control into the batching system. Aiming at the technical difficulty that is accurate weighing hundreds kinds of different physical character dyestuff, some key technique issue are researched and applied into the dyestuff accurate batching system. Firstly, the development and advantage of automatic dyestuff batching system are introduced. The research development and main research directions of iterative learning control are reviewed. In addition, the background and structure of this dissertation are also introduced. Secondly, the mechanism of solid transfer by the screw conveyor is analysed, which is the most nonlinear factor of the batching device. Then the famous solid convey theory presentation by Darnel and Mol from the aspect of torque balance is revised. Furthermore, this dissertation present the five-layers theory of solid conveying which has more accuracy, convenience and directness, compared with those theories which are based on too many formulas to be applied into the engineering. Thirdly, high order open and closed loop iterative learning control to solve a kind of nonlinear system is put forwardd. Then an initial state learning scheme is proposed, together with the high order open and closed loop iterative learning control. As far as the latter updating law owing to the tracking error bound is related to the initialization error bound. We separately present the convergence conditions and prove that this arithmetic is convergent under these conditions. The effectiveness of the proposed iterative learning control scheme is illustrated by a simulation. Fourthly, aiming to the case that determining the learning parameter just by trial in the past, we extend this parameter optimal design method. Then, we discuss a method which determines the initial input control by constructing so-called digital model. This metho uses previous input and output data. Fifthly, according to the textile Industry’s characters and craftwork, an automatic dyestuff batching product line control system is developed, which is based on a general field bus. The system generality is described separately from two aspects of hardware and software. Then a kind of iterative learning control algorithm is presented and its effect is testified in the factory based on the platform. Finally, the obtained results are summarized and future work is addressed.
馆藏号XWLW1111
其他标识符200318014603005
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/6015
专题毕业生_博士学位论文
推荐引用方式
GB/T 7714
龚利. 基于迭代学习控制的精密给定系统关键问题研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2007.
条目包含的文件
条目无相关文件。
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[龚利]的文章
百度学术
百度学术中相似的文章
[龚利]的文章
必应学术
必应学术中相似的文章
[龚利]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。