Research On Multi-version Data Management For Intelligent Manufacturing

With the development of Internet technology and the upgrading of industrial technology,the industrial manufacturing technology in the advanced manufacturing countries represented by the Industry 4.0 in Germany and the Industrial Internet in American has developed rapidly.As one of the major industrial powers,China has formulated the "Made n China 2025" plan to reform China’s manufacturing industry.In the process of this transformation,the manufacturing data is derived,which plays an important role in realizing production management and analyzing the design of production schemes.In the manufacturing industry,from the aerospace field,automotive processing and manufacturing,to the production of daily use parts,it is inseparable from the machine tools.Therefore,the effective management of manufacturing tool data in machine tool processing is particularly important for the development of intelligent manufacturing.Existing tool data management schemes have the following limitations:1)Lack of a data model that accurately expresses the meaning and semantic relationship of tool data;2)Lack of storage solutions that support the efficient storage and retrieval of multi-version tool data;3)Inefficient specific queries for multi-version tool data.In order to manage manufacturing tool data more efficiently,this paper studies the above limitations and provides solutions.Firstly,we propose a tool data modeling method based on the object deputy model.Given the defects of the existing tool data management model and combined the features of tool data with the requirements of application scenarios,the characteristics of classes and objects are used to accurately express the features of tool data and the semantic relationship between them.Unlike the traditional relational model,objects are connected by two-way pointers in the object deputy model.An object can have one or more deputy objects that inherit the properties and methods of the source object,and the deputy object can have its own properties and methods.This model can make the characteristics and relationships of data with flexible expression and strong extensibility.Meanwhile,it makes feature-based data retrieval and association query simple and efficient,avoiding a large number of complex join operations in the relational data model.Secondly,for multi-version tool data management problems,we designed a multi-version data storage scheme based on the combination of full and incremental storage.The main idea of this strategy is to use the hybrid storage of full and incremental storage of version data,taking into account the impact of four factors(depth of version,number of children of the version,number of version differences and frequency of version access)on storage space and version retrieval.Selecting a version except for the root one for full storage minimizes the cost of recovery for different versions within a certain amount of information redundancy.Then,the indexing method is proposed to solve the problem of low efficiency in the demand for querying the maximum version difference of multi-version tool data.The maximum version difference query is to find the version with the largest difference between the child version and the parent version in the version evolution history.Its index records the difference value of the parent-child version and the corresponding version pair,sorting it as an index keyword,which effectively reduces the retrieval traversal time,thus improving query efficiency.Finally,the above storage strategies and index optimization methods are implemented in the object proxy database TOTEM,and the experiments are tested on these programs.The experimental results show the efficiency and functional accuracy of the tool data model based on the object deputy model.Meanwhile,it also proves that the storage strategy based on the combination of full storage and incremental storage can improve the storage space utilization and query efficiency of multi-version data.The MDI method can effectively improve the query efficiency of the maximum version difference of multi-version data.