Research On Component Analysis And Configuration Technologies Of Mechanical Product

In this dissertation, applying complex network theory and its algorithms to mass customization of complicated mechanical product. Component analysis method and configuration technology of complicated mechanical product are studied comprehensively and deeply. Component analysis method and configuration module selection and modification method based on complex network theory are proposed. As an example, the methods are applied to industrial steam turbine and other products, the methods are verified.In chapter 1, the research background of the dissertation, the current situation and existing problems of the research are expatiated. The main content and framework of the dissertation are given.In chapter 2, complex network theory is discussed briefly, theoretical and methodological framework of mechanical product network is proposed. The detecting algorithm for simple path of complex network is studied and a universal algorithm for detecting all simple paths between any two nodes in a weighted directed acylic network is presented.In chapter 3, the constructing method of product structure network and computing method of statistical parameters are proposed. The statistical parameters are calculated. Bi-logarithmic coordinate distribution curves of statistical parameter are obtained. Correlation between the parameter and parameter's affecting degree to universality of component are analyzed. A universal analysis method of component is proposed. The characteristics of the network are analyzed. Combining the degree distribution rule, the evolving rule and dynamic equation of component relation network are proposed. The in-degree evolving rule is analyzed. Analytic expression of in-degree probability density of nodes is derived. Relation between the component increment and product increment is analyzed. Based on the in-degree evolving rule, a forecast method of increment of component number is presented.In chapter 4, the constructing method of extended product main structure network is proposed. An algorithm of correlative coefficient, between any two nodes of filiation or fraternal relation of extended product main structure, is presented. An algorithm of transferring means of correlative coefficient is put forwarded. A selecting method of configuration module based on correlative coefficient is proposed. An accessorial decision-making system of how to select configuration module is designed and realized.In chapter 5, method of how to extract dimension relation from CAD model is studied and proposed. The constructing method of dimension relation network is proposed. A dimension parameter transferring method whether within a single component or between two different components is proposed. A modifying method of dimension parameters based on dimension parameter transferring method is proposed.In chapter 6, relation between node's increasing speed and slope of degree distribution curve is studied. A compare method of modularization of product family based on slope of degree distribution curve of component relation network is presented. A module disassemble method of product is proposed. A quantificational evaluating method of modularization of product based on dimension relation network is proposed.Chapter 7 reviews the achievements and innovations of this dissertation, and gives a prospect for future work.