机械对称性知识挖掘方法及其在机械结构旋转对称性中的应用

发布时间：2018-06-28 03:27

本文选题：机械旋转对称性 + 概念体系　；参考：《浙江大学》2012年硕士论文

【摘要】：对称性广泛存在于机械领域中,镜射对称性、旋转对称性和平移对称性是三大基本对称性。本文聚焦旋转对称性,搜集了大量设计实例,在实例标准化分析的基础上应用知识挖掘方法研究了机械旋转对称性在机械中的存在方式和作用机理,主要研究内容及成果体现在以下四个方面： (1)建立了机械结构旋转对称性概念体系。借鉴课题组机械对称和机械对称破缺的研究成果,在实例分析的基础上提出了机械结构旋转对称性的分类方法,建立了包含基本旋转对称、旋转复合对称、旋转多对称和旋转对称破缺等类型的机械结构旋转对称性的概念体系。 (2)提出了改进的Eclat关联规则挖掘算法。为了克服课题组前期提出的比较模式映射关联规则挖掘算法(CPM算法)事务空间伸缩性不佳的缺点,提出了改进的Eclat关联规则挖掘算法——Eclat_opt算法。与其它Eclat改进算法和CPM算法的对比实验表明,Eclat_opt算法适用的数据集类型范围较广、运行效率较高,在事务空间和支持度空间上的伸缩性较好,对于当前及以后特征参数可能发生变化的机械对称性实例数据库都具有很强的适用性。 (3)机械旋转对称性知识挖掘与应用。从知识挖掘的一般步骤出发给出了机械对称性知识挖掘中的数据准备、数据挖掘以及结果解释和评估的流程,采用Eclat_opt算法挖掘了机械旋转对称性与设计需求、设计原理和设计准则之间的关联知识,重点分析了其中支持度和置信度较高的关联知识,归纳了若干旋转对称性在机械设计中的效用的应用知识并给出一个设计实例。 (4)开发了计算机辅助机械对称性知识挖掘与应用的软件平台。在上述理论和课题组前期开发的机械对称实例知识挖掘(KDDMSI)原型系统的基础上继续开发了机械对称性知识挖掘与应用系统(MSKMAP),改进了原系统的实例管理模块和知识挖掘模块,增强了实例分析、录入和导出的灵活性,以及知识挖掘的多样性和用户友好性。
[Abstract]:Symmetry widely exists in the field of machinery, mirror symmetry, rotation symmetry and translation symmetry are three basic symmetries. In this paper, we focus on rotational symmetry, collect a large number of design examples, and apply knowledge mining method to study the existence and mechanism of mechanical rotational symmetry in machinery on the basis of case standardization analysis. The main research contents and achievements are as follows: (1) the concept system of rotating symmetry of mechanical structure is established. Based on the research results of mechanical symmetry and mechanical symmetry breaking, the classification method of rotating symmetry of mechanical structure is put forward on the basis of example analysis, and the basic rotational symmetry and rotating compound symmetry are established. The concept system of rotational symmetry of mechanical structures such as rotation-polysymmetry and rotation-symmetric breaking is presented. (2) an improved algorithm for mining Eclat association rules is proposed. In order to overcome the disadvantage of poor scalability in transaction space of the association rule mining algorithm of comparative pattern mapping (CPM), an improved Eclat association rule mining algorithm named EclatStat opt algorithm is proposed. Compared with other improved Eclat algorithm and CPM algorithm, the experimental results show that the Eclatopt algorithm is suitable for a wide range of data sets, high efficiency, and good scalability in transaction space and support space. It is very applicable to the case database of mechanical symmetry, which may change the characteristic parameters. (3) knowledge mining and application of mechanical rotational symmetry. Starting from the general steps of knowledge mining, the data preparation, data mining, result interpretation and evaluation flow of mechanical symmetry knowledge mining are given, and the mechanical rotation symmetry and design requirements are mined by using Eclatopt algorithm. The relevance knowledge between the design principle and the design criterion is analyzed, and the correlation knowledge with high degree of support and confidence is analyzed. In this paper, some applied knowledge of the utility of rotational symmetry in mechanical design is summarized and a design example is given. (4) A software platform for computer-aided knowledge mining and application of mechanical symmetry is developed. On the basis of the above theory and the prototype system of mechanical symmetry case knowledge mining (KDDMSI) developed by the research group, a mechanical symmetry knowledge mining and application system (MSKMAP) is developed, which improves the case management module and knowledge mining module of the original system. The flexibility of case analysis, input and export, diversity and user friendliness of knowledge mining are enhanced.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH12;TP311.13

【参考文献】