面向多域复杂机电产品系统设计的功能表示与功能分解方法

发布时间：2018-05-02 06:31

本文选题：多域复杂机电产品 + 系统设计　；参考：《浙江大学》2012年硕士论文

【摘要】：多域复杂机电产品由机械、电子、控制等多个领域的众多构件通过能量、物质、信息的交互而形成。随着复杂性的不断增加,如何实现多域复杂机电产品的自动设计是十分具有挑战性。由于用户给出的总是总体性的功能需求,因此,将粗粒度的产品总功能不断分解成细粒度的子功能、直到子功能足够小能够直接由相应构件实现、并由此得到产品的功能结构模型的功能分解(Functional decomposition)过程,将是多域复杂机电产品系统设计过程中的第一步,十分关键。目前虽然已有一些这方面的研究,但是依然没有有效、实用的支持系统设计的方法与工具,甚至对最基本的功能的理解和定义都仍有歧义。本文从功能知识的表示方法出发,为产品系统设计的自动化功能分解提供了一套较完整的解决思路。为了支持自动化的功能分解,首先需要功能知识的表示方法。功能知识的表示包括对功能的表示和对工作原理的表示。功能总体上被表示为输入流和输出流关系,其作用是对输入流施加一些影响和效果,然后产生想要的输出流,在这个过程中流的类型或者属性产生改变,这些影响和效果被称之为功能效应。工作原理包括物理效应和几何材料特性,它表述了功能是按照什么原理进行工作的,其核心是物理效应。使用什么样的工作原理很大程度上决定了功能的创新程度。基于以上的功能知识表示,三种功能分解方法被提出,分别称为效应分解、原理分解和反向搜索分解,分别适用于不同的功能分解情况。效应分解把拥有多个功能效应的功能分解成多个独立的子功能；原理分解对于仅有一个功能效应的子功能,寻找能够实现其功能效应的工作原理,依照不同的工作原理进行分解；对于无法使用效应分解和原理分解的情况,使用反向搜索分解从功能知识库中搜索满足待分解功能的输入流和输出流关系的子功能序列。通过这样的功能分解方式,大量的原理解会产生,设计人员可以从中找到具有创新原理的设计方案。在上述工作的基础上,本文基于SysML开发了功能建模和分解的工具,并以汽车发动机为例,展示了产品功能建模的过程。
[Abstract]:Multi-domain complex electromechanical products are formed by the interaction of energy, substance and information in many fields, such as machinery, electronics, control and so on. With the increasing complexity, how to realize the automatic design of multi-domain complex electromechanical products is very challenging. Since the user always gives the overall functional requirements, the coarse-grained total function of the product is continuously decomposed into fine-grained sub-functions until the sub-function is small enough to be directly implemented by the corresponding component. The functional decomposing process of the functional structure model of the product will be the first step in the design of the multi-domain complex electromechanical product system, which is very important. Although there has been some research on this aspect, there are still no effective and practical methods and tools to support the design of the system, and even the understanding and definition of the most basic functions are still ambiguous. Based on the representation of functional knowledge, this paper provides a complete solution for the automatic functional decomposition of product system design. In order to support automatic functional decomposition, it is necessary to express functional knowledge. The representation of functional knowledge includes the representation of function and the representation of working principle. The function is generally represented as the relationship between the input stream and the output stream, which exerts some influence and effect on the input stream, and then produces the desired output stream, during which the type or attribute of the stream changes. These effects and effects are called functional effects. The working principle includes the physical effect and the geometrical material characteristic, which describes the principle according to which the function works, the core of which is the physical effect. The degree of functional innovation is largely determined by how well you work. Based on the above functional knowledge representation, three functional decomposition methods are proposed, which are called effect decomposition, principle decomposition and reverse search decomposition, respectively, which are suitable for different functional decomposition cases. Effect decomposition decomposes a function with multiple functional effects into multiple independent sub-functions; principle decomposition for a sub-function with only one functional effect, looking for a working principle that can achieve its functional effect, Decomposing according to different working principles; for the cases where effect decomposition and principle decomposition can not be used, the reverse search decomposition is used to search the sub-function sequence satisfying the relationship between input stream and output stream from the functional knowledge base. Through such a functional decomposition, a large number of original understanding will be generated, designers can find innovative design principles. Based on the above work, a function modeling and decomposing tool based on SysML is developed in this paper. Taking the automobile engine as an example, the process of product function modeling is demonstrated.
【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】：TH122

【参考文献】