Research On The Method Of Squeeze Casting Process Data Extraction From Periodical Literature

Squeeze casting technology can obtain high-performance and high-precision complex metal parts,which has a good development prospect.In order to adapt to the trend of big data,intelligent manufacturing and digital design of materials,it is urgent to build the squeeze casting process database to support the sharing and utilization of existing squeeze casting process data and knowledge.Periodical literature is the main source of process data.In the face of the challenge of the variety of periodical literature and the huge amount of data,this paper studies the key technology of squeeze casting process data extraction from periodical literature to establish the method and technology of squeeze casting process data extraction from periodical literature.It lays a foundation for the construction of squeeze casting process database system and the development of data-driven design method of squeeze casting process parameters.It is of great value to the popularization of squeeze casting technology,the application and sharing of squeeze casting technology data,and the effective reuse of periodical literature data.(1)By analyzing the research and production characteristics of squeeze casting process data,the main categories and attributes of the data are determined.According to the layout structure of periodical literature,combined with statistical methods,this paper analyzes the position and presentation form characteristics of squeeze casting process data in periodical literature,and the correlation characteristics between different categories of data.To provide guidance for the optimal design of data extraction algorithm.(2)Based on the results of(1),the ontology model and semantic scenario model of squeeze casting process knowledge are constructed,and a method of squeeze casting process data extraction from periodical literature is designed by using the above two models.The feasibility and effectiveness of the proposed method are verified by examples.(3)To meet the demand of tabular process data extraction from periodical literature,firstly,the visual model of table in periodical is abstracted,and a method of tabular data extraction on demand from PDF periodical literature based on the text state characteristic parameters is designed,which can filter the irrelevant tabular data.And a new performance evaluation index is proposed.The effectiveness of the proposed method is verified by examples,and the performance comparison with the existing table extraction methods and tools.(4)The structural storage structure of squeeze casting process database system is designed for the purpose of constructing squeeze casting process database system.The storage requirements of extracted data are analyzed and expanded,the storage structure and storage method are established,and the function and hierarchical structure of squeeze casting process database are designed.The Squeeze Casting Process Data Extraction System(SCPDES)based on B/S architecture is developed by Java + mySQL + JavaScript.The system integration test is completed,which shows the data in the periodical literature can be automatically extracted and stored into the database.