【作者】 胡春；

【导师】 李平；

【作者基本信息】 浙江大学 ， 控制科学与工程， 2003， 博士

【摘要】 在我国向“世界工厂”迈进的过程中，世界先进的制造技术和管理经验使工业企业生产水平有了很大的提高，企业间的协同和竞争也提升到新的高度，信息系统在工业企业中起着越来越重要的作用。经过几十年的发展应用，设备和流程的自动控制系统日益成熟，为工业企业的信息化发展奠定了良好的基础。作为企业信息化的主要内容之一，企业资源计划ERP系统的开发和实施有力推动了企业管理和业务的信息化和自动化。随着在企业中应用的不断深入，ERP系统在面对企业新的需求时遇到一些由于自身限制而无法解决的问题。 本文针对ERP系统在企业实施和应用中，出现的ERP系统所需信息的集成、工厂层管理支持和企业诊断支持等问题，就制造执行系统和智能诊断系统进行了相关研究。主要研究工作如下： (1)综述MES的产生、发展和研究概况，并分析了MESA功能模型对于多种不同工业生产企业内部集成或企业间集成的不合理性。 (2)从连续工业和离散工业生产特点的对比出发，分析了适应多种工业生产MES功能模型的必要性，并提出以计划执行、物料、设备、人员、文档、产品等生产活动管理主线作为MES功能模型各模块的划分依据。 (3)建立了由计划执行与工厂管理、数据采集与处理两大部分组成，包含生产调度和资源配置、执行和生产过程管理、设备管理、物料管理、产品管理、劳动力管理、技术和资料管理、数据采集和获取、数据综合、模拟分析等10个模块的MES功能模型。 (4)根据工厂内部各生产活动主线间的关系以及响应速度要求的不同，建立了MES递阶结构信息流模型，并提出MES数据的组织和集成架构。 (5)根据工业企业数据进行分析和诊断是企业信息系统的一个薄弱环节，管理者不得不在大量的企业数据中去发现存在的问题，而且管理者的这些知识和技能得不到有效的管理，容易遗失。因此我们根据企业管理的实际需求，提出与企业信息系统相集成的智能企业诊断系统，建立了基于数据仓库的智能企业诊断系统模型，并提出以模糊Petri网实现企业诊断的知识表示和模糊推理。浙江大学博士学位论文企业信息系统相集成的智能企业诊断系统，建立了基于数据仓库的智能企业诊断系统模型，并提出以模糊Petri网实现企业诊断的知识表示和模糊推理。 (6)分析了已有模糊Petri网模型在建模模糊产生式系统和模糊推理方面存在的问题。扩展了模糊有色Petri网系统和模糊Petri网系统，并分别提出了两者的推理建模算法，以解决己有模糊Petri网模型存在的问题，使模糊Petri网能够完全依靠自身机理实现复杂实际问题的正确推理。 (7)基于扩展模糊Petri网的建模算法，对机械工程工业企业中既包含定性关系又包含定量关系的业务和财务问题，进行企业诊断建模和推理的应用仿真。诊断结果表明诊断模型能够有效地发现企业中潜藏的问题，并对问题的产生原因作出正确的推理。应用仿真所取得较好的效果，为下一步在企业中的应用和实施奠定良好的基础。 下一阶段，我们将在某大型造纸企业和某大型医药化工企业，根据本阶段的研究结果分别展开MES和智能企业诊断系统的具体应用研究。制造执行系统建模研究为包含多种工业生产过程的企业内部的集成以及虚拟企业的实现奠定基础，通过企业信息系统的集成最终实现企恤集成:智能企业诊断系统的研究将弥补企业信息系统在企业诊断方面的不足，使企业数据转化为管理者所需的企业状况的信息，为企业的高速健康发展提供有力的支持。更多还原

【Abstract】 As China is becoming a ’global factory’, advanced manufacturing technologies and management experiences in the world promote greatly production level of our industrial enterprises. And coordination and competition among enterprises are boosted at the same time. So information systems play more and more important roles in industrial enterprises. Automatic control systems of machines and processes are well developed after development and application of several decade years. It establishes a well basis for developing informatization of industrial enterprises. As a major part of informatization of industrial enterprises, development and implementation of ERP systems impel informatization and automation of management and business of enterprises. As applied deeply in enterprises, ERP systems confront with issues that systems can’t meet with new requires of enterprise by themselves.In this article, relational researches on manufacturing execution system (MES) and intelligent enterprise diagnosis system are carried out to solve issues of integration of information required by ERP systems, supporting for management on factory floor, and supporting for enterprise diagnosis. These issues have occurred during implementation and application of ERP systems in enterprises. The principle contents and results of our researches are outlined as follows:(1) Reviewing the creation, development and present situations of researches of MES, and analyzing unreasonableness of MESA functional model for integration of or between enterprises, which have different types of industrial production.(2) Analyzing the necessity of modeling MES functional model for multi-industrial production based on comparison of characteristics between continuous and discrete industrial production. And proposing to divide functional modules of MES model according to major threads of management of production activities, such as plan execution, material, equipment, people, document, and product.(3) Establishing MES functional model comprised of two parts: plan execution and factory management, data collection and processing. The model includes 10 modules, such as scheduling and resource allocation, execution and process management, equipment management, material management, product management, labor management, technology and document management, datacollection and acquisition, data integration, simulation and analysis.(4) Establishing a hierarchical model of information flow of MES according to relationship between major threads of production activities and difference of responsibility, and proposing a model of data organization and integration of MES.(5) Analyzing and diagnosing based on data of industrial enterprises are weaknesses of enterprise information systems. Managers and manipulators have to find out problems from a lot of data in enterprises by themselves. Furthermore, their knowledge and skills about trouble-shooting can’t be managed effectively and lost easily. So we propose intelligent enterprise diagnosis system integrated with enterprise information systems to meet with actual requirements of enterprise management. A model of intelligent enterprise diagnosis system based data warehouse is established, and fuzzy Petri net is used to realize knowledge representation and fuzzy reasoning of enterprise diagnosis.(6) Analyzing issues of modeling fuzzy rule-based systems and fuzzy reasoning based existing fuzzy Petri net models. The definitions of fuzzy colored Petri net and fuzzy Petri net are modified. And two modeling algorithms based them are proposed to solve issues of existing fuzzy Petri net models. Therefore, fuzzy Petri net models can reason correctly complex problems according to mechanism in themselves.(7) Based on the modeling algorithm of modified fuzzy Petri net, business and financial fields of mechanical industrial engineering enterprises, which combine quantitative and qualitative relation更多还原