Research On Key Technologies Of Data Mining In Manufacturing Execution System

It is the development of Manufacturing informatization and intelligent and the increasing use of the Manufacturing Execution System that leads the explosive growth of production datas.it is difficult to solve the problem of dealing with large amounts of data rapidly with the way of purely manual analysis.So,there is an urgent need for new technologies to solve the problem.It is undeniable that the mass production datas stored in the MES requires to be processed intelligently and quickly by data mining techniques,which is an effective means of improving the value of production datas and MES intelligence.In this paper,the real-time data processing and its application of the key techniques of data mining in MES are studied.Firstly,It proposes a bottom-up data-push mode to realize the real-time performance on data processing and display of MES system based on B/S structure with the Oracle DCN Mechanism and WebSocket in HTML5,which can be used in processing of data mining.The K-means algorithm and BP neural network algorithm in data mining are studied,and,genetic algorithm is used to optimize the number of hidden layer,the weight and threshold in BP network.Then the MapReduce technology is used to carry on parallel design and implementation to them.A data mining paltform based on Hadoop distributed framework,playing the role of a data processing platform for cloud manufacturing system,is built.Finally,Taking the process of quality prediction in actual machining as an example,the application of data mining technology in quality prediction is introduced in detail.This paper combines the data real-time push technology and algorithm parallelization technology in data mining.The MES system has accumulated a lot of valuable historical data and quality forecasting model integrates real-time data push technology and algorithm parallelization technology in data mining.With the purpose of ensuring the product quality and reducing the missed inspection,the quality forecasting module in MES system,combined with the history production datas that are used to train the module,and real-time parameters of the current production,is designed.To a certain extent,it improves the MES system intelligence.In this paper,J2 EE technology is used to develop a manufacturing execution system with quality prediction module,which makes the relevant techniques of data mining to be practical applications in the MES system.It is a solution to improve the intelligence of the MES system.