Research And Application Of Quality Data Analysis Method In Customized Production Of General Parts

With the development of the Internet,informatization has penetrated into all walks of life.The general-purpose parts manufacturing enterprise is no exception,and its production mode is also undergoing tremendous changes.The main performance is the development of customized production to the network.The production line of the enterprise has shifted from large-scale mass production to small-volume customized production.In this transformation process,a large amount of product quality related data is generated.Traditional quality data analysis methods have been difficult to unearth the full value of data.Therefore,this paper applies data mining technology to the quality-related data of general-purpose parts manufacturing enterprises.Through association rules mining,locate the products and processes that are most prone to quality problems in the production process of the enterprise,and then provide data support for process improvement in the production department and assist enterprise managers to formulate corporate strategies.In addition,the paper predicts the number of inputs for a production task of the enterprise,and reduces the cost of production caused by excess or insufficient production.This paper mainly does the following research work.Through the investigation of the informationization status of the general-purpose parts manufacturing enterprises,the data related to the existing production quality of the enterprises are collected and collected.Because the manufacturing process of the universal parts has its own characteristics,the data form of the relevant quality data also has its uniqueness.The paper firstly analyzes the characteristics and organization forms of the quality data,and then mines the data in two aspects.First of all,the use of association rules mining technology to locate the frequent occurrence of scrapped products in the previous production and the specific reasons for scrapping.The association rule algorithm uses the FP-growth algorithm,which uses an extended prefix tree to store the compressed data set in main memory.The algorithm has excellent performance,generates rules quickly,and can be applied to distributed environments.Secondly,predicting the number of scraps of a production task of the enterprise,the paper tried a variety of algorithms,the effect is not ideal,and due to the characteristicsof the manufacturing process of the common parts,the usual prediction algorithm can not be directly applied,so this paper proposes a regression-based A comprehensive prediction algorithm based on sex-clustering solves this problem.The algorithm only needs the product coding and the number of materials to be fed,so it can predict the number of scraps of the production task through historical data,and combine the production line experience to improve the accuracy of the prediction.Experiments show that the prediction effect of the algorithm is in line with expectations.Finally,this paper designs a quality data analysis software system for enterprises,and completes the mining of association rules for quality data and the prediction of scrap counts.The system adopts B/S architecture as a whole,and association rules mining is implemented by Spark.The comprehensive prediction algorithm based on regression-association-cluster is implemented by sklearn,and each module interactively completes all functions of the system.