Research On Cleaning Methods Of Turning Machining Data And Design And Implementation Of System

Machine tool is the main supporter of tool,and tool is an indispensable cutting tool for machine tool.The data quality of cutting parameters not only affects the accurate design of processing scheme and the correct selection of processing parameters,but also affects the normal operation of cutting and the processing quality of products.Therefore,the introduction of data cleaning and database technology in turning can not only provides accurate cutting parameters for workers,but also realizes the sharing of knowledge and experience among workers.More importantly,it can improve turning efficiency and product quality.In view of the shortcomings of traditional processing parameter database system and the necessity of data cleaning,combined with the requirement of database development and data retrieval of turning process parameters,a data cleaning and retrieval system of turning process parameters is designed by using My SQL database,data cleaning and Java application development technology etc.and has realized the functions of cleaning automation of error,similar duplicate and vacancy data before multi-condition data retrieval,and it provides reference for the design and optimization of data retrieval system.The main research contents and conclusions of this paper are as follows:(1)The basic theory of data cleaning.The related theories of data cleaning for turning parameters are analyzed and summarized,and a new definition of data quality is given based on data quality dimension and user satisfaction.(2)Improvement of similar or duplicate data cleaning method.The traditional algorithm is improved by combining ranking keyword formatting,multiple keyword sorting,dynamic scaling window and multiple execution of SNM algorithm,through the analysis of the reasons why the recall and precision of traditional SNM algorithm are relatively low in the process of similar duplicate data cleaning,and the comparative experiment of similar duplicate data cleaning with traditional algorithm is carried out in MATLAB.The experimental results show that with the increase of the detected records,the recall,precision and F value are significantly reduced,but compared with the traditional algorithm,the recall,precision and F value of the improved algorithm are both improved,they are maintained above 86.3%,and the running time is also reduced.(3)Improvement of vacancy data cleaning method.Aiming at the dimension limitation in calculating sample similarity by Euclidean distance and the deficiency of filling missing field values only depending on the average value or maximum value of missing value attribute types,The traditional KNN algorithm is improved by using the distance calculation method of average value of Mahalanobis distance and grey relational degree,and the filling method of attribute weights,and the cleaning experiment of vacancy data with the traditional algorithm is carried out in MATLAB.The experimental results show that with the increase of data missing rate,the filling accuracy decreases;with the increase of K value,the filling accuracy first increases and then decreases.When k=11,the highest filling accuracy is 0.208,and the filling accuracy of the improved algorithm is higher.(4)Design and implementation of data retrieval and data cleaning system for turning parameters.The database of processing parameters is established by My SQL based on the research of similar duplication and vacancy data cleaning,and combined with the needs of user data retrieval.the design of turning parameters retrieval system with data cleaning function has been completed by using Java language,data cleaning technology and Java application development technology,and the functions of data filtering and purification,multi-condition data retrieval and data verification are realized.