节点文献
基于工艺知识网格的可重构CAPP系统关键技术研究
Research of Key Technologies in Reconfigurable CAPP System Based on Process Knowledge Grid
【作者】 张格伟；
【导师】 廖文和；
【作者基本信息】 南京航空航天大学 ， 航空宇航制造工程， 2008， 博士
【摘要】 计算机辅助工艺设计(CAPP)系统可以提升制造企业的工艺设计和管理能力。在制造模式的快速化和网络化趋势下,CAPP系统存在着快速应变的需求,但由于自身的复杂性和个性化,CAPP系统的可重构一直是该领域研究的难点。本文针对国家863项目等项目的科研目标,结合企业工艺设计过程的网络化和工艺知识快速重用需求,研究支持灵活应变和快速重构的CAPP系统以及工艺知识管理体系的相关理论、方法和关键技术。本文主要研究内容和成果如下:1)在总结现有CAPP系统的研究现状、存在问题和发展趋势的基础上,提出分别从工艺知识管理理论和可重构软件技术实现的两个方面进行研究,实现CAPP快速重构的方法。2)研究了工艺知识的来源、层次、分类和特点,提出了一种动态工艺知识网格模型(APKGA),该模型可实现工艺知识的统一描述,在工艺知识的动态使用过程中获取活跃知识和分布规律,实现工艺知识从低级形态向高级形态的进化。基于XML定义了模型中知识对象的形式化方法,给出了常用工艺知识集的形式化描述实例。3)为了在工艺知识网格中动态获取活跃知识,借鉴了人脑的记忆和遗忘的规律,建立了基于持续激励和长效学习机制的复合记忆-遗忘模型(CMFM),该模型可以实现知识网格中的工艺知识单元在持续激励学习环境下的生命周期演化。在模型求解的基础上,研究了模型的参数估值和数学特性、适用于计算机模拟的等效和简化算法,以及工艺知识的获取、更新和遗忘的处理技术。结合新旧知识的对比实例,验证了模型的有效性。4)在工艺知识单元演化的基础上,研究了知识网格中复杂工艺知识的进化和知识提取,该流程可分为知识元分解、知识整理、知识聚类和知识聚合等过程。通过知识视图和知识篇章,可以实现知识网格中工艺知识的动态聚合和超文本方式浏览,实现工艺知识从无序到有序的自组织进化过程。给出了工艺知识提取和知识篇章的实例。5)与网络化技术相结合,研究并开发了CAPP.Net系统,该系统以面向服务方式封装CAPP的功能单元,通过基于动态服务列表的应用系统软装配技术,实现CAPP系统的快速重构。结合实例研究了系统实现中的服务分解与软装配、工艺路线设计、资源管理、工艺文档电子签名等关键技术。该系统已在多家企业应用,效益显著。本文的工作为CAPP中工艺知识管理与重用,网络化、可重构的快速系统定制等的理论研究和系统开发提供了方法、模型和基础框架。
【Abstract】 Computer Aided Process Planning (CAPP) system will promote the capability of process design and management of manufacture enterprises. The trends of the rapid production and globalization in manufacture require on-demand CAPP systems. However, reuse-orient CAPP is one of the difficulties in this research field because of the complexity and individuation. Supported by National 863 Projects, and integrated with the requirement of distributed process designing and rapid reuse of process knowledge, relevant theories, key technologies and applications of reconfigurable CAPP system and dynamic process knowledge management architecture are studied to support the rapid reconstruction of CAPP applications. The main research contents and achievements are as follows:1. The theory on the Process Knowledge Management and the reconfigurable CAPP application structure are proposed based on the analysis on the current situation of the research, existing problems and trends of CAPP systems.2. Based on the analysis of the origination, hierarchy, classification and characters of the process knowledge, the Active Process Knowledge Grid Architecture (APKGA) is proposed. Through APKGA, process knowledge can be defined in uniform descriptions. The model can produce hot knowledge lists and distribution rules, which will be a basis in simulating knowledge evolution. The formalization standards based on XML of the APKGA objects are defined, and the samples of common process knowledge assemblies are provided.3. The disciplines of the memory and the forgetting of human brains are utilized for the process knowledge acquisition. The Compound Memory-Forgetting Model (CMFM) based on the mechanism of iterative excitations and long term learning is proposed. The mathematical solution, parameter estimation and features of the model are studied, as well as the equivalent and simplify for computer simulation, and the acquiring, updating and forgetting procedure in simulation are provided. On this basis, the evolvement of a single process knowledge unit is inspected and verified with test cases.4. Combined with the features of APKGA, complex process knowledge evolution procedures, including process knowledge decomposition, process knowledge self-organization, clustering and aggregation, are defined and studied. Cases of process knowledge extraction and process document are provided.5. Combined with service-oriented technologies, the CAPP.Net system is proposed and developed. With encapsulations and hierarchical platform structures, common CAPP functions are decomposed into service units and reassembled together through dynamic service list. Key technologies, including process generating, resource management, and embedding digital signatures, are studied. The CAPP.Net system has been applied in manufacture enterprises.The research on process knowledge reuse and reconfigurable network based CAPP systems provided theories, methods and models for the implementation of on-demand process knowledge management and rapid reorganization of CAPP applied systems.