【作者】 王军；

【导师】 孙惠学；

【作者基本信息】 燕山大学 ， 机械电子工程， 2010， 博士

【摘要】 CAPP不仅提高企业工艺设计的自动化水平,同时也是实现CIMS、FMS和制造业信息化的关键技术。由于工艺设计仍依赖于大量的经验知识和实验数据,难以建立工艺设计过程的数学模型,用计算机处理工艺问题还有很多困难,使得CAPP技术远远不能满足企业的需求,并且已经成为制约CAD和CAM集成、以及实现CIMS的瓶颈问题。另一方面,制造业信息化的发展也对CAPP技术提出了更高更新的要求,CAPP的智能化、集成化已经成为主要发展方向。CAPP将产品的设计要求与制造环境提供的可用制造能力相匹配,将产品设计信息转化为制造工艺信息,其关键问题是对大量的设计信息和制造信息的处理问题,本文研究了智能集成CAD/CAPP的一些亟待解决的问题,包括智能集成CAD/CAPP系统体系结构、CAPP系统与CAD系统集成、人工智能在CAPP中的应用、智能工艺设计中知识的获取和利用等问题。首先,体系结构研究是开发和研究CAD/CAPP系统的基础。本文将CAD和CAPP系统作为一个整体来规划CAD/CAPP的体系结构,满足CAPP多任务、多层次、多知识源的功能特点,研究了系统的各组成部分及相互关系,通过基于统一的产品模型与制造资源和工艺数据库实现CAD/CAPP集成,采用人工智能的多知识库协同工作方式和黑板推理模式实现CAPP的智能推理。通过对零件设计过程和制造过程的分析,在特征建模思想的基础上,提出了以特征的加工刀具为主要属性的过程特征的描述,给出了过程特征的定义、分类、和属性描述方法,并建立了零件特征信息的EER模型和数据结构,这个模型不仅为CAPP提供完整的零件信息描述,也建立了设计信息与制造信息之间的联系,为构建统一的零件信息与制造资源模型来实现CAD/CAPP集成提供技术支持。为构建统一的零件信息模型与制造资源数据库,基于过程特征的概念,采用计算智能的神经网络实现了特征与刀具技术参数的多对多的智能映射,不仅解决刀具选择的问题,还基于智能映射关系建立了零件信息模型与制造资源和工艺的关联模型,并设计了程序,实现了多对多映射、开发了数据库原型系统,验证了模型的合理性,为智能集成的CAD/CAPP系统提供了集成的数据环境。最后,还研究了采用计算智能中的神经网络技术进行工艺知识表达和利用的方法,将不能用显式表达的知识用神经网络的权值来表示,文中以切削力修正系数知识的建模为例,研究了神经网络的构建和训练问题,并选取了大量样本,编制了程序进行验证,为工艺知识中大量存在的同类问题提供了知识建模的可行方法,由权值构成数据库(ANN DB)和规则库、实例库一起作为知识库支持智能工艺决策,文中还研究了基于过程特征演变的工艺推理方法,为实现智能集成CAD/CAPP系统奠定基础。更多还原

【Abstract】 CAPP can automate the process planning and also CAPP is a key technology for CIMS, FMS and informationization of manufacturing industry. Because process is planned based on a lot of experience knowledge and the mathematical model is hard to build, there are many problems remains unsolved, CAPP can not meet the need of plants and it has become the bottle neck for CAD CAM integration and CIMS. On the other hand, with the development of manufacturing informationization, intelligent and integrated CAPP has become a developing trend.CAPP matches the design specification with the available manufacturing ability, transfer the design information into manufacturing process, the key problem is the process of design and manufacturing information. This thesis aims at solving some urgent problems, such as the intelligent and integrated CAD/CAPP structure, CAPP and CAD integration, the combination of artificial intelligence and CAPP, knowledge acquisition and utilization.First, the CAD/CAPP structure is designed by considering CAD and the CAPP as a whole, the components and its relations of the structure is presented, also the integration of CAD and CAPP is realized by integrating product model with database, and the multiple knowledge bases collaborating working mode and inference based on blackboard is implemented to meet the requirement of CAPP with multitask, multilevel and multi-knowledge function.CAD and CAPP are integrated by integrating product model into database. Based on the feature modeling and the analysis of design and manufacturing process, a new concept process feature is presented, and the definition, classification and attribute representation are given. With this concept, the EER model and data structure is proposed for part modeling, this model can be used as a tool for connecting part information with manufacturing information to build a unified database to support the integration of CAD/CAPP.Based on this concept, the mapping between features and cutters can be made with the computational intelligence of AI so that we can construct the connection among features, cutters and manufacturing resources, the matching is not only for the cutting tool selection, but also for the construction of unified database. A prototype is developed to validate the model and for supporting the realization of CAD/CAPP structure proposed above.At last, the knowledge acquisition and utilization with ANN technology is presented for modeling such knowledge as with complex data structure which is hard to construct a unified knowledge model. The ANN DB can be built with some ANN knowledge model to support the intelligent inference of CAPP together with other knowledge representation. In addition to this, this thesis also introduces the feature evolution approach to facilitate the inference ability of CAPP.更多还原