【作者】 李琳；

【导师】 江志斌；

【作者基本信息】 上海交通大学 ， 工业工程， 2007， 博士

【摘要】 随着科技的进步和经济全球化的深入,当今制造业所处的环境发生了巨大的变化。各制造企业不仅需要向用户提供高质量、高性能的产品,而且要能及时对动态多变的市场环境做出准确的响应,为用户提供多品种小批量生产直至大规模定制。以实现柔性和对多变市场需求作出及时响应为目标的面向客户的敏捷制造日益成为21世纪企业的主要模式。在敏捷制造的研究中,如何使企业生产系统的资源结构(资源组织方式)和控制结构(资源调度方式)快速、高效和经济地适应市场需求是敏捷思想得以实现的核心部分。对于我国大部分制造企业而言,尽管已在硬件方面营建了实施敏捷制造的条件,但还是普遍缺乏有效和适合的资源组织、调度以及控制等软性技术的支持,因此,从敏捷制造的角度研究企业车间级生产的资源结构和控制结构及其复杂的优化问题,将对我国制造企业进一步发展敏捷制造、提高企业竞争力起到有力的指导和推进作用。本文通过对生产系统资源结构和控制结构研究现状的比较,分析和归纳了它们的发展趋势,并预见了结合众多新型资源结构理论而提出的虚拟生产系统(VPSs)具有潜在的研究和应用价值。在分析了VPSs研究现状后,以提高控制效率、动态调度的实时能力、稳定性以及降低变化成本等为目标,提出了在虚拟空间上构造一种具有可重构性、鲁棒性、自适应性、经济性,适合于单件小批量生产环境的VPSs监控调度集成系统。首先,针对现有基于Petri网方法的VPSs模型在控制效率上的不足,本文构筑了兼顾控制和调度的基于自治与协调的VPSs监控调度系统,并在UPPAAL软件平台上加以实现。该系统主要包括:采用自动机/形式语言方法并结合面向对象的技术以及层级的方法建立的VPSs可重构DEDS模型;针对VPSs逻辑(虚拟)和物理的双重特点设计的基于自治与协调的混合式监控框架;以及进一步应用自动机反馈监控原理及网络技术具体构建的基于自治与协调的VPSs两层级混合式监控结构。将赋时自动机(网络)理论与所设计VPSs闭环控制系统相结合,最终在UPPAAL上搭建了基于自治与协调的VPSs监控调度系统,并进行了无死锁性分析以及时间最优调度的计算和仿真。其次,针对现有基于整体重调度方法的VPSs动态调度在实时能力、稳定性和经济性上的不足,本文采用局部重调度方法,在专家系统软件平台CLIPS上开发了VPSs自适应动态调度专家系统。它主要包括:结合VPSs自身的特点,构建基于事件驱动的VPSs自适应动态调度基本框架;进一步系统地制定VPSs自适应动态调度的机理所包含的各项内容;并提出改进的启发式受影响操作重调度算法,作为VPSs自适应动态调度的算法;最终在CLIPS上构建VPSs自适应动态调度专家系统。最后,建立了由这两个系统构成的VPSs监控调度集成系统,并以典型而复杂的FMS为例,在正常和异常情况下分别进行了实时监控调度和动态调整调度的实验,从而验证了本文所提出的方法和所建立的系统是可行和有效的。更多还原

【Abstract】 With the development of technology and economic globalization, today’s manufacturing environment has changed quite a lot. Besides providing high-quality and high-performance products, manufacturing enterprises should be able to react quickly to the dynamic and changing market environment in a cost-effective manner and satisfy customers’demands with diverse variety and small batch size production or even mass customization. Customer-oriented Agile Manufacturing has been the main pattern for the 21-century enterprises, which aims at realizing flexibility and response to the changing market requirements as quick as possible. In the researches on agile manufacturing, how to make the resources structure (organization pattern of resources) and the control structure (scheduling method of resources) of enterprise’s production system adapt to the market requirements quickly, high-efficiently and economically is the core to realize agile ideas. To Chinese manufacturing enterprises, most of them have had the hardware conditions to implement agile manufacturing. However, they still lack effective and proper solftware technologies to support resources organization, scheduling, control and etc. Therefore, to research resources structure and control structure in the level of enterprises’plants and their complex optimization problems from agile manufacturing viewpoint will be a powerful guaidunce and propulsion for Chinese manufacturing enterprises to further develop agile manufacturing and increase therir competitive abilities.On the basis of comparisons of the existing researches, the development trends of resources structure and control structure of production systems are analyzed and summarized. Furthermore, Virtul Production Systems (VPSs), which is proposed through combing the theories of new resources structures, are anticipated to have the potential research and application values. In order to increase control efficiency, real-time ability and stability of dynmic scheduling, and reduce changeover cost, on the basis of analysis on the existing researches about VPSs, an integrated system for supervisory control and scheduling of VPSs is proposed to establish on virtual space. It is charazterized by reconfiguration, robustness, self-adaption and cost-effectivity, and suitable for the environment of diverse variety and small batch size production.First, in order to improve the control efficiency of VPSs modeled by Petri Nets, a supervisory control and scheduling system based on autonomy and coordination is built for both controlling and scheduling VPSs and realized on the software platform of UPPAAL. The system mainly consists of the following parts: the reconfiguable DEDS model of VPSs which is built by using Automata and Formal Language method in combination with object-oriented technology and hierarchical ideas; the hybrid supervisory control framework which is then designed based on autonomy and coordination according to the logical (virtual) and physical characteristics of VPSs; and furthermore, the two-layer hybrid supervisory control structure which is established for VPSs based on autonomy and coordination by applying the feed-back supervisory control theory of automata and the network technology. On UPPAAL, the supervisory control and scheduling system based on autonomy and coordination is eventually implemented for VPSs by jointly using the timed automata (networks) theory and the designed close-loop control system of VPSs. In addition, deadlock-free analysis, and simulation of schedule optimization are made.Second, in order to overcome the limitation of the whole rescheduling in real-time ability, stablility and cost-effectivity of VPSs’dynamic scheduling, a self-adaptive dynamic scheduling expert system is proposed for VPSs by using local rescheduling method and developed on the software platform of CLIPS, an expert system platform. It is mainly composed of following parts. The basic framework of self-adaptive dynamic scheduling based on event-driven is built for VPSs according to its own characteristics. Furthermore, the contents of self-adaptive dynamic scheduling mechanism are systemically specified for VPSs. The modified heuristic rescheduling algorithm for affected operations is developed to act as the self-adaptive dynamic scheduling algorithm of VPSs. The self-adaptive dynamic scheduling expert system of VPSs is eventually built on CLIPS.Finally, an integrated system for supervisory control and scheduling of VPSs is established, which consists of the above two systems. A case study is made in a typical and complex FMS. In the case study, real-time supervisory control in normal situations and dynamic adjustment to schedule in unexpected situations are respectively made. The results show that the proposed approaches and developed system are feasibile and efficient.更多还原