Design And Implementation Of Intelligent Manufacturing Big Data Platform Based On Apriori Improved Algorithm

The manufacturing industry at home and abroad is currently facing new technological reforms,and traditional production methods need to make new changes,not only to upgrade the production methods,but also to dig deep into the core assets of the industry,that is,production information.At present,the domestic manufacturing industry has basically established primary smart factories and introduced intelligent production lines and production tools.However,there are two problems in enterprise smart factories.One is that the information of the smart production line is not interactive,and the other is the abnormal problem of the smart station of the smart production line.Slow to solve and difficult to predict.Therefore,it is urgent to solve the problem of information island of intelligent production line and the problem of abnormal analysis and prediction.This paper takes the intelligent manufacturing S enterprise as the research object,facing the intelligent production line and intelligent station,firstly studies the key technologies and core algorithms of the big data platform;secondly,designs the general architecture of the big data platform and proposes the use of K-means clustering algorithm The algorithm model was constructed with the improved Apriori association rule algorithm,and the big data platform for intelligent manufacturing based on Hadoop was realized.The main research contents are as follows:(1)Carry out demand analysis for the intelligent manufacturing big data platform based on the current situation of the smart factory and intelligent production line of S enterprise,and determine the main five functional modules of the big data platform: data collection,data storage,data algorithm,data analysis and data visualization At the same time,the functional requirements and non-functional requirements of the big data platform are studied.According to the demand analysis,the overall architecture of the intelligent manufacturing big data platform is designed,and the platform is divided into four layers: the production line layer,the acquisition storage layer,the computing engine layer and the application layer.(2)Detailed design of platform function modules and optimization of Apriori classic algorithm.According to the data characteristics,the data acquisition module divides the data into cold and hot data for collection design.The data algorithm module contains two algorithm models,and the model design includes the K-means clustering algorithm model and the improved Apriori association rule algorithm model.The data storage module uses the combination of HBase and My SQL to design the database according to the characteristics of the data object.The data visualization module is designed based on the Python Flask framework.In this paper,the Apriori algorithm is improved,mainly from the two aspects of target data and storage method.Firstly,the target database is optimized,secondly,a 0-1matrix is constructed to store database data and filter non-empty data,then the matrix is calculated by column-wise subregional matrix to calculate frequent itemsets,and finally the subregional matrices are fused to obtain the maximum frequent itemsets.(3)Big data platform implementation and algorithm model experimental analysis.Firstly,building a cluster of intelligent manufacturing big data platform includes platform environment building and Hadoop cluster building,and realizes the business requirements of five functional modules.Through the experimental analysis of the K-means algorithm model,the optimal silhouette coefficient and K value are obtained,and the main factors affecting the downforce of the robot are excavated.Experiments were carried out on the improved Apriori algorithm,and it was found that the efficiency of Apriori after the improvement was increased by 27.8%.