基于知识的机械产品快速设计方法研究

发布时间：2018-09-13 16:51

【摘要】：随着全球制造业的飞速发展,机械产品更新换代的速度也随之加快,企业只有不断提升对市场需求的快速反应能力,缩短产品设计周期,降低设计与制造成本,才能在激烈的市场竞争中谋得一席之地。本研究在国家科技支撑计划(2012BAF12B08)和国家科技型中小企业技术创新基金(10C26212110396)支持下,针对机械产品快速设计过程中存在的问题,在分析国内外关于知识管理和快速设计理论、方法和应用现状的基础上,以多主轴数控车床为研究对象,综合运用系统、集成的思想,对快速设计的理论知识体系和实现方法进行了深入的研究。针对现有产品设计理论体系对客户需求的动态性和创造性设计规律进行充分表达方面尚显不足的问题,采用基元模型和基元蕴含树模型对客户需求进行描述和分解,获得共性需求和个性需求；利用变精度粗糙集进行设计规则的抽取和知识约简,提取出共性设计规则；以客户需求为输入,以设计规则为标准,以产品最终模型为输出,构建了基于客户需求的机械产品快速设计框架。针对机械产品知识结构复杂,产品异构信息具有多维度、多层次性等特点,研究了快速设计过程中的知识获取与表示方法。利用公理化设计理论对产品的知识模块进行划分；采用自适应遗传算法提取出知识模块的分类特征,运用本体语言将产品需求、功能特性、设计参数等产品特征信息建立在统一的知识表示模型中。通过实例对比表明,所提出的知识获取与表示方法能够有效解决快速设计中的瓶颈问题,实现新产品的快速响应设计和创新设计,为产品设计知识重用奠定了基础。针对传统设计过程是从设计空间到性能空间的正向映射过程,致使已有的知识难以在新产品设计中得到重用的问题,构建了基于多级实例分解的产品知识重用模型。该模型以从性能空间到设计空间的逆映射方法为导向,按照设计分解的粒度和顺序,通过多级实例分解技术实现功能域和结构域间的相互映射,使没有耦合或冗余的实例在全新的设计方案中得以重用,方便了各部门设计人员之间的知识共享。该方法有助于设计者快速得到与设计需求相匹配的知识,有效进行设计知识重用,从而降低设计成本,提高知识重用效率。针对产品快速设计各阶段产生的设计成果缺陷、设计反复迭代的问题,对设计过程中的知识库进化方法进行了研究。利用朴素贝叶斯分类器建立机器学习模型进行监督学习,可将非结构化数据处理为结构化数据；采用领域本体技术进行信息抽取、运用相似度算法进行本体演进、将产品结构进行聚类分析并用于知识挖掘：从而实现产品知识库和设计规则库的持续优化和升级。该知识库进化方法可以减少设计过程中的重复性工作,提高快速设计质量。最后,将研究成果与企业工程实际相结合,进行了多主轴数控车床快速原型系统的开发,研究了系统的总体框架和功能结构以及产品进行快速设计的具体过程。该系统可以逐步实现用户对产品快速设计的全面管理,满足了企业对快速设计系统设计自动化、智能化的需求。通过系统在企业中的应用,验证了论文研究的快速设计方法的可行性与有效性。
[Abstract]:With the rapid development of global manufacturing industry, the speed of mechanical product renewal is also accelerating. Only by continuously improving the rapid response ability to market demand, shortening the product design cycle, reducing the design and manufacturing costs, can enterprises get a place in the fierce market competition. 08) Supported by the National Science and Technology Innovation Fund for Small and Medium-sized Enterprises (10C26212110396), aiming at the problems existing in the process of rapid design of mechanical products, based on the analysis of the knowledge management and rapid design theory, methods and application status at home and abroad, the multi-spindle CNC lathe is taken as the research object, and the idea of integrated system and integration is applied comprehensively. The theoretical knowledge system and implementation method of rapid design are studied in depth.
In view of the insufficiency of the existing product design theory system in fully expressing the dynamic and creative design rules of customer needs, the elementary model and the elementary implication tree model are used to describe and decompose the customer requirements to obtain the common and individual requirements; the variable precision rough set is used to extract the design rules. The common design rules are extracted by knowledge reduction, and the rapid design framework of mechanical products based on customer requirements is constructed with customer requirements as input, design rules as standard and final product model as output.
In view of the complex knowledge structure of mechanical products and the multi-dimensional and multi-level characteristics of heterogeneous information of products, the knowledge acquisition and representation methods in the process of rapid design are studied. The product feature information, such as product requirements, functional characteristics, design parameters and so on, is established in a unified knowledge representation model. The comparison of examples shows that the proposed knowledge acquisition and representation method can effectively solve the bottleneck problem in rapid design, realize rapid response design and innovative design of new products, and reuse product design knowledge. Laid the foundation.
Aiming at the problem that traditional design process is a positive mapping process from design space to performance space, which makes it difficult for existing knowledge to be reused in new product design, a product knowledge reuse model based on multi-level case decomposition is constructed. The granularity and order of the model are mapped between the functional domain and the structural domain by the multi-level instance decomposition technique, so that the uncoupled or redundant instances can be reused in the new design scheme, and the knowledge sharing among designers in different departments is facilitated. The reuse of design knowledge can reduce the design cost and improve the efficiency of knowledge reuse.
Aiming at the shortcomings of design results produced in various stages of product rapid design and the problem of iterative design, the knowledge base evolution method in the design process is studied. Line information extraction, ontology evolution using similarity algorithm, clustering analysis of product structure and knowledge mining are used to achieve continuous optimization and upgrade of product knowledge base and design rule base.
Finally, combining the research results with the engineering practice of enterprises, the rapid prototyping system of multi-spindle NC lathe is developed, and the overall framework, functional structure and the specific process of rapid product design are studied. The system can gradually realize the overall management of rapid product design by users and meet the needs of enterprises. The design automation and intellectualization of the design system are required. The feasibility and validity of the rapid design method are verified by the application of the system in the enterprise.
【学位授予单位】：大连理工大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TH122;F273

【参考文献】