节点文献
摩阻材料智能设计系统的开发
Development Intellectual Design System for Tribology Materia
【作者】 柳丽娜;
【导师】 高诚辉;
【作者基本信息】 福州大学 , 机械设计及理论, 2004, 硕士
【摘要】 摩阻材料的设计实际上是一个摩擦学系统的设计。鉴于中小型企业的设计能力和水平,难以胜任传统的设计方法,开发摩阻材料的智能设计系统既可以节约大量的人力和物力又可以提高企业的开发水平,并且对于改变传统的设计方法,提高摩阻材料设计领域的智能水平也有积极的意义。 本文根据专家系统,人工神经网络以及遗传算法的基本理论,将人工神经网络的知识获取和表现方法引入专家系统,并且结合遗传算法的优点,在此基础上建立具有智能的摩阻材料设计系统。该系统能够从已有的数据中获取知识,从而达到知识自学习和扩展的目的。它具有自组织,自学习,模糊性和容错性等特点,能够弥补单纯的专家系统的不足。 本文详细讨论了人工神经网络的BP算法以及将其应用于专家系统知识的获取,提出了相似度的概念并且完成了部分组分相似度的定义。根据系统元素相似,结构相似,性能也相似的原理,结合遗传算法,设计完成了可以根据已有的配方优化设计符合用户要求的新配方的智能设计系统。分析了该摩阻材料智能设计系统的数据特点,建立了适合该系统的数据库。
【Abstract】 The design of tribology material in fact is the design of tribology system. As far as the developing new material ability of small and moderate factories is concerned, it is difficult for them to design a new formula by traditional method, and the development of intellectual design system for tribology material can not only save a lot of manpower and material resources but also improve the small and moderate factories’ developing ability. In addition, it is significative to change the tradional design method and improve the intellectual design level in the domain of tribology material design.Basing on the fundamental theory of expert system (ES) and artificial neutral network (ANN), taking advantage of ANN’s good ability of knowledge acquisition and expression and the merits of genetic algorithm (GA) as an optimize, a mingle intellectual design system for tribology material is presents by integrating ES, GA and ANN. it can acquire knowledge from the existing dada and can get knowledge itself. So it can learn itself, organize itself, and tolerate mistakes to overcome the shortcomings resulting from using ES or neutral network independently.In this paper, the author discussed the BP learning algorithm of ANN, apply it to knowledge acquisition of ES, put forward the conception of comparability and define the comparability of some components. Based on the theory that if the component and the structure is similar of two system, the two system should have the similar function, combining GA, develop the intellectual design system which can make good use of the existed formula in the system to design the new formula met with the user’s requests. In addition, the author analyzed the data characters of the intellectual design system and built the rudiment of suitable the data-base for the system.
【Key words】 tribology material; intellectual design system; artificial neutral network(ANN); genetic algorithm;
- 【网络出版投稿人】 福州大学 【网络出版年期】2004年 03期
- 【分类号】TB39
- 【被引频次】1
- 【下载频次】92