【作者】 罗超；

【导师】 左铁镛； 彭颖红；

【作者基本信息】 中南大学 ， 材料加工工程， 2004， 博士

【摘要】 挤压铝合金型材以其重量轻、强度高、加工性能好、耐腐蚀等良好的特性,在航空航天、铁道车辆、汽车和民用装饰材料中得到了非常广泛的应用,而挤压成形是铝型材最主要的生产方式。目前,我国的大多数铝型材生产厂家在设计制造中仍然采用传统的方法,工艺分析及模具设计主要依靠经验,需要反复调试,生产效率低,成本高,周期长,而且产品质量难以保证。短周期、高质量、低成本是面向21世纪全球化制造业市场竞争的三大要素,而围绕这三大竞争要素的产品创新能力的形成则是应对制造业市场竞争的关键所在。产品创新能力的形成依赖于有效的知识供应和高新技术的驱动,而其基础在于新的理论体系和相应关键技术的研究和开发。 本文结合铝型材挤压特点,将基于知识的工程(Knowledge-based Engineering, KBE)方法引入型材挤压模具设计,在基于知识工程的铝型材挤压模具智能设计系统框架内,对型材挤压模具智能化设计方法进行了研究,在分析弹塑性有限元法和有限体积一步模拟法基础上,提出了薄壁型材挤压过程的有限体积分步模拟技术,并对数值模拟与人工智能技术相结合的工作带优化方法进行了的研究。 本文分析了铝型材挤压产业的发展状况及面临的问题、型材挤压成形研究现状;针对知识经济条件下的先进制造技术与产品创新设计理论体系,研究了智能技术的发展状况,介绍了KBE技术的产生、发展和应用情况;提出了将KBE技术引入型材挤压模具设计。 对KBE技术进行了较深入的研究,给出了KBE的定义和内涵,从几何适用性、知识表现形式、知识管理方法、自学习能力等方面分析了KBE技术与传统专家系统的区别;研究了知识表示、知识推理、知识获取和繁衍等KBE关键技术。 提出了型材产品特征模型及挤压模具设计要素特征模型的构造方法;采用面向对象的知识语言AEKL实现了知识框架的描述,对设计知识集成建模技术及知识推理的实现技术进行了研究,并通过AEPCRL知识语言实现了型材挤压事例检索的方法。 在分析有限元法和有限体积法模拟薄壁铝型材挤压成形不足之更多还原

【Abstract】 Aluminum profile parts, with the advantages of lightweight, high strength, good workability and anticorrosion, are widely used in aircraft, automotive and decoration. Extrusion is the dominating manufacturing method of aluminum profiles. At present, most domestic factories that manufacture aluminum profiles still adopt traditional method in the design and manufacturing. The process and mould design mainly depend on experience, and the fabrication of mould needs trial-and-error tests, which results in low efficiency, high cost and long developing period. Moreover, the product quality cannot be guaranteed. Short developing period, high quality and low cost are the three key elements for the competition in the global manufacturing market in 21st century. The ability of product innovation is the most important factor for coping with the competition. The innovation ability is resulted from effective knowledge supply and the driving of high-tech, which are based on the research and development of new theoretical system and corresponding key technology.In present study, Knowledge-based Engineering is introduced into die design of aluminum profile extrusion. In the framework of knowledge-based intelligent design system of aluminum profile extrusion, the intelligent design method is studied. The finite element method and finite volume method are compared, and a FVM multi-stage simulation method is put forward. A die bearing optimization method based on numerical simulation and artificial intelligence is also proposed.The production status of aluminum profile extrusion is analyzed and the state-of-art of corresponding research is summarized. In the framework of advanced manufacture technology and theory of product design innovation, the evolution of intelligence technology is studied. The establishment, evolution and application of KBE are introduced. Then a viewpoint is pointed out that introduction of KBE into die design of aluminum extrusion is an effective way to improve the designintelligence and innovation.A deep research on KBE technology is carried out. The definition and connotation of KBE are given. The difference between KBE and expert system is analyzed from the aspects of geometrical applicability, knowledge-based modeling, knowledge management and self-studying. Several key technologies of KBE, knowledge expression, knowledge reasoning, knowledge acquisition and multiplying, are studied.The methods of constructing feature models for aluminum profile and the key elements of die design are developed. Knowledge framework is constructed with an object-oriented knowledge language AEKL, which is used to study integrated knowledge modeling and knowledge reasoning technology. The case retrieving method of aluminum profile extrusion is developed with knowledge language AEPCRL.Based on the shortcoming analysis of finite element method (FEM) and finite volume method (FVM) in simulating thin-walled aluminum extrusion, a finite volume multi-stage simulation technology is proposed to simulate extrusion process of thin-walled aluminum part. With the FVM multi-stage simulation method, mesh rezoning in FEM simulation is avoided. Moreover, the problem of computer resource shortage in FVM one-step simulation can be solved efficiently. Finer Euler mesh can be obtained through reducing meshing scope in single step. Therefore the simulation problem of thin-walled aluminum profile extrusion is solved successfully.An optimization model for die bearing design of aluminum extrusion is presented, which integrates ameliorated BP neural network, numerical simulation and genetic algorithm. The area of extrusion section is divided into several elements and the bearing values of them are given as the input parameters of network training specimen by using the orthogonal method. The target value of the model is mean-squared error of velocity after forming. Finite volume method is used in the numerical simulation to get the target value of specimen and the general optimization solution is attained through genetic algorithm. Theoptimization of process parameters can not only be realized with this method, but also the time of numerical simulation can be reduced greatly through prediction of artificial neural network.The architecture of KBE system for aluminum extrusion die design is studied. Then an intelligent design KBE system for aluminum profile extrusion system KBAES is established, taking AutoCAD as developing platform and Autodesk ObjectARX API as developing tool. The functions of the system are demonstrated with an example. The design results of die dimension, die profile layout, die bearing, die strength, extrusion press and extrusion barrel are obtained. Then the design scheme is verified with numerical simulation. So, the effectiveness of the system is verified.更多还原