Research On Key Technology Of Mechanical Product Configuration Design In Big Data Environment

With the extensive application and deep integration of information technology such as the Internet and manufacturing enterprise information systems,manufacturing data has beenaccumulatedwith mass,high-speed,diversity and value.The effective analysis and study of the big data of manufacturing industry is of great significance to promote intelligent manufacturing.In mechanical industries,information in enterprise information system such as CAD/ERP/PLM/MES of product configuration design is important part of Big Data.According to the current research on the key technology of large data under the environment of mechanical product configuration design information modeling,processing and evaluation problems,mainly reflected in:(1)how to mechanical product configuration design information in the massive unstructured data model building and data expression;(2)the traditional design knowledge reuse method cannot meet the design staff in machinery the conceptual design of products and parts of the detailed design stage of configuration design information retrieval requirements;(3)the characteristics of product configuration design cannot translate into data mining;(4)how to modular mechanical product configuration design is divided into modules analysis and objective evaluation.According to the characteristics of configuration design information in Big Data environment of mechanical products,research on product configuration design information model,redundant data processing,feature retrieval,knowledge mining method and modular product configuration design and evaluation method,the main research contents are as follows:(1)Based on the analysis of mechanical product configuration design information of data sources and characteristics,put forward the galaxy model for mechanical product configuration design information,structured data and information in product configuration design of non-structured data and semi-structured data into a unified expression of JSON(JavaScript Object Notation)with the keyword data metadata;the source of sensitivity and semantic similarity,puts forward the calculation method of JSON data file text similarity,realizes processing redundant data mechanical product configuration information.(2)The mechanical product configuration design features are divided into structural features and associated features.Based on product function ontology,ontology similarity calculation method is studied.Component model retrieval method is proposed to meet the retrieval requirements of product conceptual design.Based on the structural features and correlation characteristics of components,the weighted similarity evaluation algorithm and component matching algorithm are studied.Component model retrieval method is proposed to meet the needs of detailed design retrieval of product parts.(3)Taking the passenger car products as an example,the standard parameters of the key parts and components configuration design were converted.The customer's most concerned functional requirements are extracted from different customer orders and configured as key components.Based on the data of the historical product configuration design of the passenger car,a configurable product feature template is established.The decision tree algorithm based on MapReduce parallel computing is studied,which is used to mine product configuration design knowledge.The feasibility of the algorithm is verified by example data,and the product configuration constraint rules are extracted.(4)The modular non-uniform granular clustering method is studied.Fuzzy clustering of product components is carried out.The hierarchical structure of product is established.Different modular product configuration design schemes are obtained by non-uniform assembly of modules with different particle sizes.Taking passenger car products as an example,the characteristics of product materials,functions,structures and processes are analyzed.The entropy weighted AHP(Analytic Hierarchy Process)method is used to evaluate the modular heterogeneous granular clustering scheme.Finally,the optimal scheme of modular design for passenger car products is obtained.