联系我们

论文查询结果	返回搜索

论文编号：	16068
作者编号：	2120243873
上传时间：	2026/6/4 22:43:33
中文题目：	面向智慧开发利用的三维模型档案元数据设计研究
英文题目：	Research on the Design of 3D Model Archive Metadata for Intelligent Development and Utilization
指导老师：	王芳
中文关键字：	三维模型档案；档案元数据；机械工程；MBD；智慧开发利用
英文关键字：	3D model archives; archival metadata; mechanical engineering; Model-Based Definition (MBD); intelligent development and utilization
中文摘要：	随着制造业数字化转型的不断推进，基于模型定义技术的三维模型正逐渐成为机械工程领域的重要记录对象。与传统文本档案相比，三维模型档案在信息构成上具有更明显的复合性，其几何信息、装配关系、产品制造信息（PMI）以及管理属性往往分散在模型本体、关联图纸等多个载体中，现有档案著录实践对这类对象的系统描述仍不足。 本文以机械工程数字三维模型档案元数据设计为研究对象，围绕如何构建一套兼顾管理、保存与后续利用需求的元数据方案这一问题，综合运用访谈法、文献研究法和案例分析法，遵循“概念界定—要素识别—方案构建—案例检验”的研究路径，尝试提出一套面向智慧开发利用的三维模型档案元数据方案。 在要素识别阶段，本文结合SolidWorks零件模型、装配体及其关联零件样本，对三维模型档案中可直接获得的信息进行分析，初步提炼出涵盖基本标识、几何物理、材料信息、建模工艺、格式保存、装配关系和图纸关联等维度的元数据要素，并发现STEP AP203前端属性面板存在大量字段未录入的情况，STEP文件中Author/Organization字段呈UNSPECIFIED状态，反映出前端属性规范化不足造成信息缺失。在需求修订阶段，本文对档案管理人员和三维设计工程师共7人开展半结构化访谈，归纳出检索粒度不足、版本管理混乱、装配关系缺乏表达等方面的需求，并对初始要素进行补充调整。 在方案构建阶段，本文提出“三层架构，八维要素”的元数据体系：以通用基础层对接通用元数据标准，以三维扩展层回应几何、材料、装配等三维对象特征，以语义智慧层为后续知识组织和智慧利用预留描述接口，最终形成119项元数据字段，并按必填、推荐、可选三级进行约束划分。 在案例检验阶段，本文选取“自动打铆钉系统机”三维装配体档案开展著录实践。结果表明，该方案能够较完整地支持装配体层级关系的结构化描述，格式转换信息丢失标注字段和PMI完整性状态字段能够在案例著录中得到体现，并可为凭证查证、设计复用、关联发现等场景提供初步的描述基础。同时，著录结果也表明，前端属性规范化程度会直接影响字段填写情况，部分必填字段的完整著录率仍然偏低。 本文的主要贡献在于：针对机械工程三维模型档案的信息结构特点，提出了分层化、对象化的元数据设计思路；补充设计了格式迁移风险记录、PMI完整性状态标注和装配关系分层表达等更贴近三维对象特征的关键字段；形成了样本观察—访谈修订—案例著录相结合的元数据方案设计路径。本文属于探索性设计研究，所提方案已通过真实案例开展初步著录检验，后续仍需在样本扩展、系统集成和智慧利用效果验证等方面进一步深化。
英文摘要：	As the digital transformation of manufacturing continues to advance, three-dimensional (3D) models developed through Model-Based Definition (MBD) technology are increasingly becoming significant archival objects in the field of mechanical engineering. Compared with traditional text file, 3D model archives exhibit a more pronounced composite character in their information structure: geometric data, assembly relationships, Product Manufacturing Information (PMI), and administrative attributes are often distributed across multiple carriers, including the model entity itself and its associated drawings. Existing archival description practices remain insufficient for the systematic documentation of such objects. This study takes mechanical engineering digital 3D model archives as its research subject and addresses the question of how to construct a metadata scheme that comprehensively serves the requirements of management, preservation, and subsequent use. Employing a combination of semi-structured interviews, literature review, and case study methods, and following a research trajectory of "conceptual definition–element identification–scheme construction–case validation," the study proposes a metadata scheme for 3D model archives oriented toward intelligent development and utilization. During the element identification phase, the study analyzed information directly extractable from SolidWorks part models, assembly models, and their associated components. This analysis yielded a preliminary set of metadata elements spanning dimensions including basic identification, geometric and physical properties, material information, modeling processes, format and preservation, assembly relationships, and drawing associations. The analysis also revealed that custom attributes are largely unfilled at the front-end design stage; Author/Organization fields in STEP AP203 files consequently appear as UNSPECIFIED, reflecting information gaps caused by insufficient standardization of front-end properties. During the requirements revision phase, seven stakeholders, comprising archival managers and 3D design engineers, were consulted through semi-structured interviews. The interviews identified deficiencies in retrieval granularity, inconsistencies in version management, and an inadequate representation of assembly relationships, on the basis of which the initial set of metadata elements was supplemented and refined. During the scheme construction phase, the study proposes a metadata framework characterized by a "three-tier architecture with eight dimensions." The foundational general layer aligns with universal metadata standards; the 3D extension layer addresses the distinctive characteristics of 3D objects, including geometry, materials, and assemblies; and the semantic intelligence layer provides descriptive interfaces reserved for future knowledge organization and intelligent utilization. The resulting scheme comprises 119 metadata fields, which are classified into three levels of obligation: mandatory, recommended, and optional. During the case validation phase, the scheme was applied to the archival description of a 3D assembly model of an automatic rivet-driving machine. The results demonstrate that the scheme is capable of supporting a reasonably complete structured description of hierarchical assembly relationships. Fields designated for recording format-conversion information loss and PMI integrity status were successfully instantiated in the case description, providing a preliminary descriptive foundation for scenarios including evidentiary verification, design reuse, and associative discovery. At the same time, the case study reveals that the degree of upstream attribute standardization directly affects field completeness, and that the completion rate of certain mandatory fields remains relatively low. The principal contributions of this study are as follows: a layered, object-oriented metadata design approach tailored to the information structure of mechanical engineering 3D model archives is proposed; key fields more closely aligned with the characteristics of 3D objects—including format migration risk records, PMI integrity status annotations, and layered expressions of assembly relationships — are designed and incorporated; and a methodological pathway is established that integrates sample observation, interview-based revision, and case-based archival description. This study is exploratory in nature. Although the proposed scheme has undergone preliminary validation through real-world case description, further work is required in areas including sample expansion, system integration, and empirical verification of intelligent utilization outcomes.
查看全文：	预览  下载(下载需要进行登录)