Research On Knowledge Discovery Method Of Aircraft Engine Collaborative Development Based On Big Data

Aircraft engines are the crown of the manufacturing industry.Among the national major special projects launched during the 13 th Five-Year Plan period,aircraft engines rank first.In the past eight years,new models related to aircraft engines have been launched one after another,and the historical models have been continuously optimized iteratively.A large amount of research and development data and related process data have been accumulated in the research and development process.However,since the application of data mining technology in the field of aircraft engine has just begun,the three major aircraft engine giants in the world are still at the stage of exploration and experiment,and the research and development process technology is highly confidential,and there is no way to learn and reference.Therefore,how to improve the efficiency of engine collaborative development through data processing,analysis and mining is a very concerned issue in the industry.At present,the cooperative development mode is adopted for aircraft engines.Under this mode,a large number of 2D or 3D CAD models,technical reports or change orders and other documents are generated for various types of products.How these massive structured or unstructured data can be precipitated into effective knowledge and transferred to upstream and downstream links such as design and manufacturing,and can be reused and continuously improved in different models,so as to accelerate the development progress and improve the development quality;Based on the massive,multi-source and heterogeneous data accumulated in the process of aircraft engine collaborative development,this dissertation carries out research and Exploration on relevant knowledge discovery methods,realize the discovery and precipitation of relevant knowledge,and provide theoretical and technical support for the independent research and development,inheritance and optimization of aircraft engine related products.The main research contents are as follows:(1)Establish a big data modeling method for knowledge discovery.Firstly,the knowledge discovery of aircraft engine collaborative development is defined in detail and expressed formally;The hierarchical framework of knowledge discovery in aircraft engine collaborative development is established;Based on the connotation and characteristics of aircraft engine collaborative development,in view of the large amount of industrial multi-source and heterogeneous data generated and accumulated in the process of collaborative development of new models and historical models of engines,the data galaxy model and vector space data model are proposed.Aiming at the massive unstructured data in the product knowledge data of aircraft engine collaborative development,this dissertation analyzes the multi-source heterogeneous data model and solves the knowledge discovery of aircraft engine collaborative development.At the same time,the dissertation completes the establishment of an example with JSON,carries out the word dimension reduction of vector space,and carries out the redundancy processing of collaborative development scenarios.(2)The similarity calculation algorithm of feature data and the weighted evaluation algorithm of similarity are proposed.The collaborative development of aircraft engine Based on big data carries out data mining and knowledge discovery of feature knowledge,and establishes a feature extraction algorithm based on functional ontology in combination with the difference of feature data;Combined with the structural features and association features of components / parts,data mining is carried out through weighted evaluation and other algorithms,so as to meet the knowledge discovery in collaborative development.According to the requirements of product configuration design on component data in aircraft engine collaborative development,the data mining method based on product requirements,functions and configuration is improved to better meet the data mining requirements of knowledge discovery in collaborative development.In the conceptual design stage,the standardized description is carried out based on the product function;In the product design stage,the knowledge expression is based on the structural characteristics of components.According to the demand configuration and engineering configuration,establish the construction process of the target information table.A dependency analysis model is designed for requirement configuration;The association rule mining method is established for engineering configuration.Finally,the feasibility of this method is proved by an example of commercial aircraft engine.(3)To explore the knowledge discovery and prediction method of aircraft engine collaborative development based on big data.Therefore,a fuzzy clustering data analysis method based on triangular modules is proposed,and a rough set dependency model is established based on the hierarchy of aircraft engine configuration data.According to the configuration examples and module types,the configuration process of aircraft engine configuration performance prediction is established,and the initial information table is established.A prediction model based on BP neural network is proposed to predict the basic performance parameters of the new derived products in the collaborative development based on attribute reduction,so as to realize the knowledge discovery and prediction in the collaborative development of aircraft engine.The dissertation also makes an attempt and test on the optimization and confirmation of the configuration scheme of an aircraft engine.