基于知识进化与实例推理的机械产品三维建模技术

发布时间：2018-02-01 16:45

本文关键词： 遗传算法实例推理智能设计空间分解三维参数化建模　出处：《沈阳理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：伴随着机械装备制造业的快速发展零件的不断复杂化、高精度化对机床行业也提出了高效率、高精度的要求,机床要想有所改变就要从设计入手,机床的设计是机床发展的基础,所以智能化设计机床是一个重要的研究方向。在数控车床的基础上研究数控车床的智能设计系统(ICAD),本文的开始构建了智能设计系统的总体框架,依次划分为多级实例的空间描述、基于遗传算法的数控车床方案设计、基于实例推理的数控车床详细方案设计和三维参数化建模这四部分,要对关键部分的技术进行分析研究,然后将各模块的功能实现。多级实例的空间描述是系统的基础部分。本文首先对数控机床做了空间分解,将复杂的实例分解成由各个功能模块组成,这使得在设计中能较为准确的确定每个模块所能实现的功能。在分解时要分析相关的关联度并将数控车床按功能分解与按结构分解,它分解的好坏直接影响了遗传算法与实例推理系统的性能。遗传算法与实例推理系统是本文的核心。在遗传算法系统部分要对设计要求转换、分析各项指标建立相应的数学模型作为适应度函数,然后进行选择、交叉、变异优选初步设计方案。在得到基本方案之后使用实例推理得到能满足方案要求的近一步详细方案设计,分析影响结构部分的要求参数确定各个设计参数部分,使用层次分析法与专家法相结合确定特征权重,利于最近邻法计算相似度,在检索出相似的实例之后使用遗传算法修改,最终形成满足用户要求的方案。在本文里对UG进行二次开发将产品的轮廓特征表达出来,使用表达式驱动图形模板进行三维参数化建模,图形的尺寸特征与表达式之间要相互依赖,能在实例推理结束后将实例推理得到的尺寸数据传递到表达式中,使表达式能够驱动参数生成三维模型。本文在多层空间实例的分解基础之上运用遗传算法与实例推理相结合实现数控车床的创新型方案设计然后进行三维模型,这样就可以模拟设计者的思维模式去设计避免过多的重复劳动、提高了设计效率。
[Abstract]:With the rapid development of machinery and equipment manufacturing parts are becoming more and more complicated, and the high accuracy of the machine tool industry has also put forward the requirements of high efficiency and high precision. If the machine tool wants to change, it must start with the design. The design of machine tools is the basis of the development of machine tools, so intelligent design machine tools is an important research direction. At the beginning of this paper, the general framework of intelligent design system is constructed, which is divided into space description of multi-level examples and numerical control lathe scheme design based on genetic algorithm. The detailed scheme design and 3D parameterized modeling of NC lathe based on case-based reasoning (CBR) should be analyzed and studied in this paper. Then, the function of each module is realized. The spatial description of multi-level examples is the basic part of the system. Firstly, the spatial decomposition of NC machine tools is done, and the complex examples are decomposed into each functional module. This makes it possible to determine the function of each module accurately in the design. When decomposing, we should analyze the correlation degree and decompose the NC lathe by function and structure. Its decomposition directly affects the performance of genetic algorithm and case-based reasoning system. Genetic algorithm and case-based reasoning system are the core of this paper. In the part of genetic algorithm system, the design requirements should be changed. The corresponding mathematical model is established as fitness function by analyzing each index, and then it is selected and crossed. After obtaining the basic scheme, we can get the further detailed scheme design that can meet the requirements of the scheme by case-based reasoning, and analyze the required parameters that affect the structural part to determine the design parameters. AHP and expert method are used to determine the feature weight, which is conducive to the nearest neighbor method to calculate the similarity. After retrieving similar examples, genetic algorithm is used to modify them. Finally form a solution to meet the requirements of users. In this paper, the second development of UG to express the contour features of the product, using expression driven graphics template for three-dimensional parametric modeling. The dimension feature of the graph depends on the expression and can transfer the dimension data of the case reasoning to the expression after the case reasoning is finished. The expression can drive the parameters to generate 3D model. In this paper, based on the decomposition of multi-layer spatial cases, genetic algorithm and case-based reasoning are used to realize the innovative scheme design of NC lathe and then carry out 3D model. In this way, the designer's thinking mode can be simulated to avoid too much duplication of work, and the design efficiency can be improved.
【学位授予单位】：沈阳理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TG502.1

【参考文献】